Trade & Industry Monitor Vol 38 2007.pdf - tips
Trade & Industry Monitor Vol 38 2007.pdf - tips
Trade & Industry Monitor Vol 38 2007.pdf - tips
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
TIPSTRADE&INDUSTRY MONITORMONITORvol <strong>38</strong> 2007
TIPSTRADE & INDUSTRYMONITOR2007volume <strong>38</strong>
TIPS EXECUTIVE DIRECTOR: STEPHEN HANIVALTRADE & INDUSTRY MONITOR EDITOR: LUCILLE GAVERAWhile copyright in the journal as a whole is vested in TIPS, copyright in the articles rests with the individual authors and contributors.Opinions expressed are the responsibility of the authors and not of TIPS.ISSN: 1010-4<strong>38</strong>XCover photographs: Greatstock/Corbis & X Collection
CONTENTSTRADE LIBERALISATIONIntegrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process .............................1Helena McLeod, Department for International Development (DFID) South AfricaThe Impact of the Asian Drivers on Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>......................................................35Raphael Kaplinsky, Universities of Sussex and Brighton and Mike Morris, Universities of KwaZulu-Natal and Cape TownDEVELOPMENT AND POVERTY REDUCTIONCan Small Farmers Survive, Prosper, or be the Key Channel to Cut Mass Poverty?...........................................................55Michael Lipton, Poverty Research Unit, University of Sussex, UK<strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation ......69Eric Schaling, South African Reserve Bank Chair, University of Pretoria and Center for Economic Research, Tilburg UniversityINDUSTRIAL POLICYThe Industrial Challenge Facing Africa in the Global Trading System...............................................................................95United Nations Industrial Development Organisation (UNIDO)SOUTH AFRICA’S GROWTH TRAJECTORYAn Assessment of the Accelerated and Shared Growth Initiative ..................................................................................121Rob Davies, University of Zimbabwe and Dirk Ernst van Seventer, TIPSSECTOR STRATEGIESThe South African Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports? .................................131Anthony Black, University of Cape Town and Sipho Bhanisi, Department of Agriculture (Western Cape)Cassava <strong>Trade</strong> Information Brief .................................................................................................................................153Alison Goldstuck, <strong>Trade</strong> and Industrial Policy Strategies (TIPS)TRADE AT A GLANCEMmatlou Kalaba, TIPS
ANNUAL FORUM 2007The Sustainability of South Africa’s Current Growth TrajectoryWednesday 3 - Friday 5 October 2007Zebra Lodge, PretoriaThe TIPS Forum, now in its 11th year, is an annual event bringing together members of the research, academic and policycommunities to discuss economic policy issues of South African and regional interest. The 2007 event promises to continuedeveloping these links and, indeed, to further strengthen the flow of relevant research into policy-making.South Africa has experienced the longest continuous period of economic growth in its recorded history. Moreover, inthe last two years, economic growth accelerated to rates of just over 5% per annum. However, this much improvedgrowth performance has coincided with the appearance of an increasing number of constraints, in particular supply-sideconstraints. These include the supply of:Appropriate workforce skills;Intermediate manufactured products, such as cement and steel;Basic infrastructure, such as electricity, water and industrial land; andLogistics infrastructure, such as roads and ports.The effects of these constraints are beginning to be seen in a number of areas, for example, the widening currentaccount deficit, high credit extension and environmental damage due to the rapid growth in industrial and residentialdevelopments.Forum 2007’s focus is therefore on the sustainability of South Africa’s current growth trajectory .Some of the pressing research questions that Forum 2007 would like to interrogate are:To what extent do South Africa’s data reflect growth in the ‘real’ sector?What have been the key contributors to South Africa’s recent improved growth performance?What effect has government investment had on growth performance?To what extent, or through which mechanisms, has the exchange rate influenced growth?What are the underlying limitations to South Africa’s growth potential?Are there measures which policy-makers could consider to alleviate supply-side constraints?What is the role of trade policy in a supply-side constrained growth phase?What is the appropriate fiscal and monetary policy stance during periods of high growth?Are there likely to be environmental externalities stemming from the current growth path?Is the current growth path likely to have significant income distribution, or more generally poverty, effects?Is the current growth pattern pro-poor?Closing date for final papers:Friday 14 September 2007Please submit abstracts, enquiries about papers and general enquiries about the Forum to:amanda@<strong>tips</strong>.org.zaFor further information, visit our websites:www.<strong>tips</strong>.org.zawww.commark.org
tradeliberalisation1INTEGRATING AFRICATHROUGH A SACU AND COMESA FTA:SPEEDING UP THE REGIONALINTEGRATION PROCESS<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>BY HELENA MCLEODDFID SA REGIONAL INTEGRATION ADVISOR 1 21. Africa’s vision for an integrated continentSOUTHERN AFRICA AND AFRICA have committed themselves to economic integrationas part of a growth and prosperity strategy for the Continent. The Southern AfricaDevelopment Community (SADC) and the Common Market of East and Southern Africa(COMESA) members have committed to implementing customs unions, while theSouthern African Customs Union (SACU) is already a customs union. SACU, SADC and COMESA alloverlap in membership and mandate. The Africa Union’s vision is for an integrated Africa. Immediatedesired outcomes of the New Partnership for Africa’s Development (Nepad) include that “regionalintegration is further accelerated and higher levels of sustainable economic growth in Africa is1I would particularly like to thank Owen Willcox (previously from TIPS) for peer input and data provision. The methodologicalframework has been taken from a series of papers by Owen Willcox and TIPS research fellow Dirk Van Seventer. Iwould also like to thank Themba Munalula from COMESA for data provision.2The views expressed in this piece do not reflect those of DFID. The author writes in her personal and not official capacity.This paper looks at the complicated overlappingtrade arrangements in Southern Africa andexplores how to simplify these arrangements,unlock scarce capacity and speed up regionalintegration in Africa. The paper responds toquestions including whether regional integrationis beneficial for Africa, whether overlappingmembership of Regional Economic Communities(RECs) are hampering or helping integration andwhat the practical reality of a customs unionin SADC is. In looking for the quickest way tointegrate markets in Africa, would a unionbetween the largest REC in Africa (COMESA) andthe oldest customs union in the world (SACU) bea viable option? In the context of South Africa’sexternal trade policy and commitment to Africathe paper also explores economic and politicalreasons why such a union might be worthpursuing.
<strong>Trade</strong> Liberalisation2<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>achieved” 3 . Yet what currently prevails is a complex web of overlappingmembership of RECs, the so-called building blocks of Nepad.Other regions of the world clearly believe economic advancement will beachieved through market integration. The EU is considerably advanced interms of the integration of goods markets, labour markets and monetaryintegration. Whilst nation states in Europe retain a high level of nationalautonomy, significant powers have been divested to the European Parliament.<strong>Trade</strong> policy amongst other areas is a European Commission lead.And the European Court of Justice is the highest legal entity in the EU.The North American Free <strong>Trade</strong> Agreement (NAFTA) has combined theeconomic strengths of the US, Canada and Mexico into a free trade areawith the removal of substantial tariff restrictions and ongoing removalof non-tariff barriers. And in Asia the ASEAN group have agreed that by2015 they will have formed a trading bloc with China, India and Japan.This mammoth economy will comprise 25% of the world gross domesticproduct (GDP) and on the current growth paths of China will be set toovertake the GDP of both the EU and the US. By 2025 there are likelyto be three economic power blocks in the world – Asia, the Americasand Europe. At the current pace of economic development and regionalintegration, Africa will be left a distant fourth.1.1. What are the costs and benefits ofregional integration?The first question to ask is why would we want to integrate Africa’s marketsanyway? The answer lies in orthodox trade theory which encouragescountries to reduce their trade barriers so they can specialise and exportgoods that use the factor they are endowed with most abundantly,whether that be land, labour or capital. Production is more efficient andthe freed-up resources can be used to import other goods more cheaplythan before trade liberalisation. Longer term benefits such as lower pricescan also be realised through increased competition, technology transferand lower production costs as scale economies are pursued. Regional integrationhas these economic benefits and also non-economic benefits,including locking in policy reform, increasing the bargaining power ofsmall countries and improved peace and stability.Whether these benefits are attained also relates to the level of integration.In economic theory, the higher the level of integration the greater thegains will be. A free trade area (FTA) where countries reduce their tariffbarriers but maintain their own trade policy is the lowest level of integra-3Nepad official website, front page, 11 April 2006.tion. The SADC <strong>Trade</strong> Protocol, when it is fully implemented, will form anFTA. Within a customs union (CU), the next level of integration, countriesmust agree to a common external tariff. To be an improvement on an FTA,CUs should also allow the free movement of goods within the CU. Somegroups such as the East Africa Community (EAC) and SACU have commonexternal tariffs but do not allow the full free movement of goods, and solimit the benefits of a CU.If countries continue to integrate their markets by agreeing a commonmarket where factors of production, including workers, can move freelyamongst the members, then a fairly deep level of economic union hasbeen attained. The EU is an example of a Common Market and is also apolitical union, ceding some of its members’ national powers to the Europeanlevel. Some members of the EU have also agreed a common currency,the Euro. Although both SADC and COMESA members have agreedin principle to aim for economic and monetary union, given the currentpace of integration this seems a long way off and a very political journey.Regional integration is not a substitute for multilateral or unilateral tradeliberalisation. A successful completion to the current Doha <strong>Trade</strong> Round inthe World <strong>Trade</strong> Organisation (WTO) negotiations will bring the greatestbenefits to most developing and African countries, as it would includeliberalisation of important agricultural and industrial markets.<strong>Trade</strong> diversion – where members of a free trade bloc switch trading froma low-cost producer outside the bloc to a higher cost producer inside thebloc – is a potential negative consequence of regional integration. Therecan be an overall loss in welfare in the importing country as the savingin price does not compensate the loss in tariff. Lowering external tariffsmultilaterally or unilaterally can help to avoid this problem. Although, asmost favoured nation (MFN) tariffs get closer to zero, the benefits of anFTA fall as compliance costs come into play, such a rules of origin.Complementary policies are also necessary to ensure all members benefitfrom regional integration. Such policies include improving the domesticenabling environment, diversifying the revenue base to make up for losttariff revenue, considering the free movement of labour, reducing transportcosts and adopting a revenue-sharing formula that helps to redistributethe benefits of being in a customs union evenly.1.2. Africa’s economic marginalisationAfrica’s share in international trade has fallen from 6% in 1990 to 2%in 2002. Sub-Saharan Africa (SSA) is currently the poorest region in theworld and off track to meet any of the Millennium Development Goals.
Economies in Africa are small and population density is low. Even SouthAfrica, which comprises 40% of the GDP of SSA, is relatively small in Europeanterms. According to the International Monetary Fund (IMF), SouthAfrica is the 28th-largest economy in the world in terms of GDP (2005),and between Denmark and Greece in size. Transport costs in SouthernAfrica are on average 73% higher than in the US or Europe. And 30% ofthe population of Africa is living in landlocked countries, compared to theworld average of 1%. Africa hardly trades with itself – 5% compared to46% in NAFTA, 55% in East Asia and 62% in the EU15.One reason for the lack of intra-African trade is the internal barriersto trade, both tariff and non-tariff. If Africa wants to integrate its smallmarkets successfully, removing barriers to its internal trade and reducingtransport costs are essential. That is why Nepad’s vision prioritises regionalintegration. The irony is that whilst Africa is most in need of economicintegration due to the small size of its markets, the number of countriesthat need to participate in negotiations is high, making negotiations slowand laborious.1.3. Overlapping membership of regionaleconomic groupingsIf regional integration and moreover an integrated Africa is what we wantto see, what is the fastest way we can achieve this?SADC and COMESA members have agreed to become customs unionswithin the next six years – COMESA by 2008 and SADC by 2010. Customsunions have a common external tariff and therefore for legal reasons twocustoms unions cannot overlap unless they have the same common externaltariff. A full customs union (with free movement of goods) and a freetrade area cannot overlap unless all members of the customs union aremembers of the free trade area as well.Currently, South Africa, Botswana, Lesotho, Namibia and Swaziland aremembers of SACU. SADC membership includes these countries as well asMozambique, Tanzania, Angola, Madagascar, Mauritius, Zimbabwe, Zam-bia, the Democratic Republic of Congo (DRC) and Malawi. Membership ofCOMESA includes eight countries that are members of SADC – Malawi,Zambia, Zimbabwe, Swaziland, Angola 4 , the DRC, Mauritius and Madagascar.In addition, Kenya, Uganda, Ethiopia, Djibouti, Eritrea, Sudan,Egypt, Libya, Seychelles, Comoros, Burundi and Rwanda are also members.If this was not complicated enough, Uganda, Tanzania and Kenyahave formed a customs union called the East Africa Community, but whilstUganda and Kenya are members of COMESA, Tanzania is a member ofSADC. The membership of RECs like SACU, SADC and COMESA in Southernand Eastern Africa overlap so extensively that these overlaps havecome to be known as a ‘spaghetti bowl’, as Figure 1 illustrates.4Angola is in the process of withdrawing from COMESA.3<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>FIGURE 1: OVERLAPPING GROUPINGS IN SOUTHERN AFRICA, 2006COMESACBIDjiboutiEritreaSudanBurundiRwandaComorosSeychellesEgyptEthiopiaLibyaAngolaMalawi Congo DRZambiaZimbabweMauritiusMadagascarIOCMozambiqueSADCCBI:COMESA:EAC:IOC:SADC:SACU:Cross-Border InitiativeCommon Market for East and Southern AfricaEast African CommunityIndian Ocean CommissionSouthern African Development CommunitySouthern African Customs UnionKenyaUgandaTanzaniaEACSACUSACUSwazilandBotswanaLesothoNamibiaSouth AfricaSource: Author’s own compilationIntegrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
(2006) find that giving South Africa duty-free access above their currentTDCA preferences into the EU is unlikely to lead to a supply responsegreat enough to threaten either EU producers or other SADC country exports,except in a handful of products. They find that the inclusion of SouthAfrica in the EPA is essential to strengthen regional integration as the EPApromises to do. Breaking up the SACU would be contrary to this.The SADC Strategic EPA position identifies the ideal SADC EPA arrangementas two-fold. First, South Africa would be included and the SACUcountries would negotiate an enhanced TDCA. Secondly, Mozambique,Angola and Tanzania would sign the SADC EPA based on contractuallybinding duty-free and quota-free access, as they already receive on agifted basis by the EU. This implies that there can be no SADC commonexternal tariff unless the EPAs are re-negotiated at a later date to agree acommon external tariff which is identical for Mozambique, Angola, Tanzaniaand the SACU countries.Another concern voiced by some is that the EPAs have the potential tofragment SADC. This is because five of the countries belonging to bothSADC and COMESA – Malawi, Zambia, Madagascar, Mauritius and Zimbabwe– are negotiating with the COMESA grouping. This is the choiceof the countries. Inevitably if they are members of two RECs they wouldhave to choose to sign an EPA with either SADC or COMESA, unless SADCand COMESA negotiate an EPA together, which is currently not the case.Although it is possible that at the last minute they could choose to sign anEPA with the SADC grouping, it seems more likely they shall remain withthe COMESA grouping. COMESA is currently progressing fairly quickly towardsa customs union, with a common tariff nomenclature having beensigned. If dual members sign an EPA with COMESA, it is likely that theywill adopt the COMESA customs union, and legally they will have to leaveany SADC customs union arrangement. However, these countries couldchoose to remain with the SADC <strong>Trade</strong> Protocol and free trade arrangementif internal barriers were maintained within the COMESA customsunion (losing many of the benefits of a full customs union) or if all membersof the customs union became members of the SADC <strong>Trade</strong> Protocol.1.5. So what are Southern Africa’s options?Generally, four options for regional integration in Southern Africa arediscussed:1. All SADC members negotiate a new common external tariff. Allexisting bilateral arrangements that members have, includingSACU, are renegotiated and aligned to the new SADC commonexternal tariff. Countries must choose between SADC or COME-SA customs union membership.2. All SADC members sign up to the current SACU common externaltariff so the SADC customs union is essentially an expandedSACU. Countries must choose between SADC or COMESA customsunion membership.3. SADC and COMESA become a single customs union and havea single external tariff. Members do not have to choose as allgroups have integrated.4. SACU expands to include some but not all SADC members. Membersof SADC and COMESA choose between either the SACU orCOMESA customs union.One major obstacle exists in terms of Option 1 and 2 – SACU, ledby South Africa, has already negotiated a number of bilateral tradeagreements, including the TDCA with the EU and with the EuropeanFree <strong>Trade</strong> Association (EFTA), is about to sign with Mercosur and isexploring agreements with China and India. Why is this a problem?SACU has already negotiated a common external tariff and tariffphase-down arrangement with these groups of countries. Either allother SADC countries have to sign up to these trading arrangementsalready negotiated by SACU or SACU will have to re-negotiate all itsexisting trade agreements.It is unlikely (although not impossible) that either SACU will wish to renegotiatethese hard-won arrangements or that non-SACU countries willbe content to sign up to a customs union whose negotiations they havenot been party to. Whether these options can be realised depends on howfar the SACU trade arrangements would need to change to accommodatenew members, and whether joining a ‘SACU+’ arrangement brings otherbenefits, such as revenue sharing. Before SADC can start negotiating acustoms union, members will have to decide whether the SACU commonexternal tariff will be adopted or SACU will have to decide to give up itscommon external tariff and start negotiating a SADC common externaltariff from square one. It is therefore extremely unlikely if not impossiblethat SADC will implement a customs union by 2010.Option 3 is perhaps the ideal, but it would face similar problems tooptions 1 and 2 in terms of negotiating a new common external tariff,as both SACU and COMESA have common external tariffs whichwould require integrating.A commonly suggested trading arrangement is option 4. SACU alreadyincludes five countries – South Africa, Botswana, Lesotho, Swazilandand Namibia. Mozambique has been rumoured to be investigatingthe possibility of joining. This would only leave Tanzania that is notalready a member of COMESA. Given the fact that Tanzania is already5<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> Liberalisation6<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>a member of the East Africa Community (EAC), a customs union withUganda and Kenya who are members of COMESA but not of SADC, itwould make sense for Tanzania to join COMESA. There would then betwo trading blocs – SACU and COMESA. SADC’s trade capacity couldbe merged with that of SACU and COMESA and would continue tolead on other areas, such as energy, peace and security, and water.Alternatively, the agendas of SACU, SADC and COMESA could beintegrated with much closer working together and based aroundtwo common external tariffs – that of the current SACU and that ofCOMESA.This perhaps begs the question of why all the fuss on agreeing customsunions. This paper asserts that the real benefits of a customs union comewhen internal trade barriers, including rules of origin, are fully removed.This would require a revenue-sharing mechanism to be put in place, whichis sensitive given the dependency of countries on trade taxes. It would alsorequire divesting revenue collection to the customs union members whoform the border with the rest of the world. Although internal barriers canbe reduced in a free trade area they cannot be fully removed, as withouta common external tariff, goods would flow in through the border of themember with the lowest external tariff and then flow to all other membersin the group. Given that a customs union with free movement of goods isthe ideal and SACU and COMESA are the closest to being fully functioningCUs, this paper explores the possibility of a free trade agreement betweenthem, with the aim of increasing the pace of integration in Africa.2 Current and potential tradebetween COMESA andSACUThis section provides a detailed analysis of current and potential trade betweenSACU and COMESA, using South Africa data as a proxy for SACU.The analysis is useful to inform policy-makers and trade negotiators whatthe gains and risks would be to negotiating a COMESA SACU FTA.SACU’s trade is dominated by South Africa’s trade by a ratio of about10:1. Therefore, in focusing on South Africa’s trade we are capturing mostof SACU’s trade. Where relevant we have also compared COMESA SACUtrade with China and India SACU trade, two countries where SACU is currentlynegotiating preferential trade arrangements. Two data sources havebeen used, one presented in Rand and the other in US dollars. We havekept the currency denomination of the original data used; at the time ofwriting (April 2006), US$1 equalled R6.South Africa’s GDP, population and trade dwarf that of its SACU neighbours.Its GDP in 2002 was US$182.28-billion compared to the next highest,Botswana, at US$7.02bn (see table 1).The COMESA members are also economically heterogeneous, althoughless so than SACU. South Africa dominates not just SACU but SADC as awhole. South Africa is both the largest economy in Africa and constitutesabout 60% of the total GDP of the SADC region.2.1. Aggregate bilateral tradeTable 2 shows that South Africa’s imports from COMESA have increasedby 11.55% over the past 10 years (1995-2005), using a semiLog regressionto estimate the underlying trend. During 2002-2005, imports fromCOMESA doubled to R7.445-million. This seems to have been driven byEgypt, Angola and Zimbabwe and commodity price increases. The 1995-2005 trend growth rate of 11.55% is less than the growth rate of SouthAfrica’s imports from the rest of the world, which have increased by13.<strong>38</strong>%. This accounts for the marginal 1.61% drop in COMESA’s shareof South Africa’s imports. South Africa’s exports to COMESA tell a similarstory. Although exports have grown over the same period by 9.98%, theyhave grown fractionally faster to the rest of the world (by 12.57%), leadingto a 2.30% drop in COMESA’s share of South Africa’s exports.In comparison, South Africa’s trade with the two largest Asian economies,China and India, shows a significantly faster increase in both cases. Importsfrom China grew on average 32.4% from 1993-2003 and from Indiaat 18.5% from 1995-2004. Exports to China also showed a high growthrate, increasing by 29.7% from 1993-2003 and to India by 19.4% from1995-2004 6 . What is interesting is the actual level of trade between SouthAfrica and these different trade partners. Whilst South Africa’s importsfrom China were quadruple those from COMESA in 2003, at R16,582mcompared to R4,354m, South Africa imports from COMESA in 2004, atR6,601m, were significantly greater than imports from India at R4,547m.In terms of South African exports, COMESA was a more important partnerby far than either China or India. In 2003, exports to COMESA wereR22,545m, to China R6,570m and to India R3,662m. So even thoughtrade with China in particular but also with India is growing at a very highrate, the COMESA countries will continue to be the most significant exportmarket for South African goods. Reduced trade barriers between SouthAfrica and COMESA could lead to even higher levels of trade.6Owen Willcox and Dirk van Seventer (TIPS), October 2004 (China) and March 2005(India).
TABLE 1: BASIC SACU INDICATORSPopulation Gross domestic product Per capita gross national income (GNI)2002 (million) 2002 (US$-billion) 2002 (US$)Botswana 1.71 7.02 3,010Lesotho 1.78 1.15 470Namibia 1.99 4.37 1,960Swaziland 1.09 1.69 1,240South Africa 45.35 182.28 2,6007Table 3 shows that total trade between South Africa and COMESA hasgrown by 10.33% over the past 10 years, although this has not keptpace with growth in South Africa’s trade with the world, resulting in a2.34% fall overall. What is interesting to note is that South Africa’s tradebalance (that is, the amount it exports minus the number of imports) isnegative with the world (it has a trade deficit). However, it was positivewith COMESA (it had a trade surplus) at R18,391m in 2005. This comparesto a trade deficit with China of -R10,012m in 2003 and with Indiaof -R886m in 2004.Table 4 shows South African imports from COMESA per COMESA membercountry. A handful of countries are significant in trade terms. Imports in2005 were greatest from Zimbabwe at R3,131.5m, followed by Angolaat R1,891m, Zambia at R1,303m and Malawi at R455m. All of thesecountries also belong to SADC. The surge in imports from Angola from2003-2004, resulting in a 37.9% growth rate over the period 1995-2005,is due to the oil price increase and not related to the SADC <strong>Trade</strong> Protocol.The significant increase in imports from Zimbabwe from 2002 is causedby a R1bn increase in imports of nickel ore.<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>2.2. <strong>Trade</strong> with COMESA disaggregatedby countryBefore we go any further, it is imperative to analyse both which countriesin COMESA are most important in terms of trade with South Africa andthe impact of the SADC <strong>Trade</strong> Protocol and overlapping membership ofSACU, SADC and COMESA on the results of this paper. As noted earlier, ofthe 20 countries that are members of COMESA, eight are also membersof SADC.This complicates our calculations and analysis because an FTA has alreadybeen agreed between the SADC member countries through the SADC<strong>Trade</strong> Protocol. However, the 2005 mid-term review of this Protocol revealedthat most of the agreed tariff reductions have not been implemented.This is largely due to back-loading by non-SACU SADC countries,where they have postponed the time when they have to reduce their tariffsas agreed in the FTA. This means that the impacts of the SADC FTAare yet to be felt.In terms of our analysis of potential trade from a SACU COMESA FTA, thismeans our results will still be relevant. However, what we must bear inmind is that without a SACU COMESA FTA, if SADC members fulfil theircommitments to implement the SADC FTA, a significant percentage ofCOMESA member country trade (those countries that belong to SADC)would be liberalised anyway.In terms of non-SADC COMESA countries (marked with an asterisk, ‘*‘),Kenya and Egypt are by far the most important, although with importsat R204m and R176m, respectively, their importance to South Africa areseemingly far weaker. If we turn our attention to the bottom half of the tablewe see that non-SADC COMESA trade as a proportion of all COMESAtrade has fallen from 22% in 1995 to 6% in 2005. In terms of non-SADCCOMESA imports as a proportion of South Africa’s world imports, it isinsignificant,falling from 0.5% to 0.1% over that period. All-COMESA importsare similarly insignificant, falling over the period to just 2% of SouthAfrica’s total imports.To some extent what we are seeing here is a competitive advantagegained by countries geographically close to South Africa. However, the figurescould also disguise trade diversion, where imports are diverted awayfrom the most efficient producer towards less efficient producers withinthe trading bloc – demonstrated to occur in FTAs and customs unionsthat maintain high trade barriers against non-member countries. BothSACU and COMESA would probably fall into this category. For instance,the majority of SACU’s trade is with the EU where a far-reaching FTA exists.Tariffs between SACU and the EU have been reduced but maintainedwith other countries not party to an FTA. The question is: if South Africais serious about supporting Africa as a continent, should developing freetrade with Africa not be a priority – and perhaps a greater priority thandisadvantaging African countries in favour of those in Latin America, Asia,Europe or the US where FTAs have been signed or are being explored.Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> Liberalisation8<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>TABLE 2: SOUTH AFRICA’S AGGREGATE TRADE WITH COMESA AND THE WORLD, 1995-2005 (IN R-MILLION)Growth1995-2005 *(%)Year 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005SA imports from COMESA 2,203 2,818 2,823 2,296 2,363 2,188 2,693 3,970 4,354 6,601 7,445 11.55% growth per year 27.9 0.2 -18.7 2.9 -7.4 23.1 47.4 9.7 51.6 12.8SA imports from world 101,054 116,903 129,834 143,976 147,<strong>38</strong>3 188,064 215,441 274,458 258,431 306,368 350,661 13.<strong>38</strong>% growth per year 15.7 11.1 10.9 2.4 27.6 14.6 27.4 -5.8 18.5 14.5COMESA share of SA imports 2.2 2.4 2.2 1.6 1.6 1.2 1.3 1.4 1.7 2.2 2.1 -1.61SA exports to COMESA 9,090 12,483 14,566 14,252 13,793 17,030 19,839 25,769 22,545 22,651 25,836 9.98% growth per year 37.3 16.7 -2.2 -3.2 23.5 16.5 29.9 -12.5 0.5 14.1SA exports to world 100,447 114,133 137,339 142,740 161,508 208,285 215,248 277,993 255,560 291,129 320,005 12.57% growth per year 13.6 20.3 3.9 13.1 29.0 3.3 29.2 -8.1 13.9 9.9COMESA share of SA exports (%) 9.0 10.9 10.6 10.0 8.5 8.2 9.2 9.3 8.8 7.8 8.1 -2.30Source: TIPS database and own calculationsNote: * Throoughout the analysis, growth is estimated by a semiLog regression which attempts to trace the underlying growth curve despite yearr-by-year fluctuations.TABLE 3: TOTAL TRADE BETWEEN SOUTH AFRICA AND COMESA (IN R-MILLION)Growth1995-2005(%)Year 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005SA trade balance with COMESA (X-M) 6,887 9,665 11,743 11,956 11,430 14,842 17,146 21,799 18,191 16,050 18,391SA trade balance with world (X – M) -607 -2,770 7,505 -1,236 14,126 20,220 -193 3,535 -2,871 -15,239 -30,656SA total trade with COMESA (X and M) 11,293 15,301 17,390 16,548 16,157 19,218 22,532 29,740 26,899 29,252 33,280 10.33Annual growth (%) 35.5 13.7 -4.8 -2.4 18.9 17.2 32.0 -9.6 8.7 13.8SA total trade with world (X and M) 201,501 231,036 267,173 286,715 308,891 396,349 430,689 552,451 513,991 597,497 670,666 12.985.6 6.6 6.5 5.8 5.2 4.8 5.2 5.4 5.2 4.9 5.0 -2.34COMESA’s share of SA total trade (% overallX and M)Source: TIPS database and own calculations
TABLE 4: SOUTH AFRICA IMPORTS FROM COMESA (IN R-MILLION)Growth1995-2005(%)Year 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005World 101,054 116,903 129,834 143,976 147,<strong>38</strong>3 188,064 215,441 274,458 258,431 306,368 350,661 13.4Angola 3.6 261.7 210.0 9.2 196.8 67.9 12.5 128.5 28.8 1,684.3 1,891.1 37.9Burundi* 0.0 0.2 2.2 0.7 5.4 0.4 0.7 1.5 4.3 0.5 3.1 36.9DRC 365.5 480.2 453.1 25.0 17.9 9.5 20.7 17.9 28.6 44.3 26.8 -24.8Djibouti* 0.0 0.1 0.0 0.2 0.1 0.0 0.0 0.4 0.6 0.3 1.0 0.0Egypt* 369.0 228.1 189.5 298.8 82.3 52.2 72.8 109.9 358.0 108.6 175.8 -6.1Eritrea* 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.5 0.2 0.8 0.5 0.0Ethiopia* 2.7 1.3 4.5 4.1 3.8 3.3 3.1 10.7 11.7 23.4 9.3 22.7Kenya* 112.7 126.7 89.6 62.5 <strong>38</strong>.5 44.1 89.6 110.1 106.6 329.2 203.5 8.8Comoros* 0.3 0.4 0.4 0.1 0.2 0.1 0.2 1.4 0.3 0.4 1.8 13.4Libyan Arab Jamahiriya* 0.0 0.0 0.0 0.1 16.8 0.1 0.0 0.0 0.1 0.5 0.1 0.0Madagascar* 10.0 6.3 17.7 39.6 16.1 21.2 17.4 35.3 14.9 10.5 11.2 1.8Mauritius 37.9 17.3 25.6 29.4 52.1 46.5 153.3 93.0 124.7 103.8 167.8 23.0Malawi 206.7 317.2 400.0 459.4 467.2 285.8 328.2 483.6 <strong>38</strong>1.9 435.6 455.8 4.5Rwanda* 1.2 0.3 2.4 2.2 0.1 26.1 75.9 4.2 1.5 0.9 6.4 18.6Seychelles* 2.6 4.7 6.3 8.7 9.4 21.7 33.9 11.9 26.0 27.7 17.9 23.2Sudan* 1.0 4.0 2.3 1.1 1.7 2.0 1.4 3.5 2.6 5.6 0.6 1.5Swaziland 2.9 1.6 0.3 0.7 0.6 2.1 0.5 1.1 0.3 0.0 0.3 -21.5Uganda* 3.4 1.9 3.2 10.8 21.6 4.7 18.1 20.7 36.1 36.9 36.4 33.9Zambia 95.4 179.0 181.3 230.5 221.8 301.7 421.2 776.6 571.3 992.3 1,303.6 27.2Zimbabwe 988.0 1187.1 1,235.1 1,113.4 1,210.9 1,298.3 1,443.6 2,159.5 2,655.8 2,795.7 3,131.5 12.5Total SA imports from non-SADC COMESA 492.8825 367.564 300.3<strong>38</strong> <strong>38</strong>9.2077 179.9096 154.652 295.6934 274.832 547.7514 534.842 456.511 2.5Total SA imports from all COMESA 2,202.858 2,817.96 2,823.42 2,296.394 2,363.256 2,187.6 2,693.214 3,970.33 4,354.204 6,601.44 7,444.62 11.522.37468 13.0436 10.6374 16.94864 7.612786 7.06947 10.9792 6.92214 12.57983 8.10189 6.13209 -8.1Proportion of SA imports from non-SADCCOMESA (%)2.179886 2.41051 2.17464 1.594987 1.603484 1.16322 1.250092 1.44661 1.684863 2.15474 2.12303 -1.6Imports from COMESA as % of imports fromworld0.487742 0.31442 0.23132 0.270329 0.12207 0.08223 0.13725 0.10014 0.211953 0.17457 0.13019 -9.6Imports from non-SADC COMESA as % ofimports from worldKey: ‘*‘ = non-SADC COMESA countriesSource: TIPS and COMESA database and own calculations9<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> LiberalisationIn terms of South Africa’s exports to COMESA countries in 2005 (see Table5), Zimbabwe is the largest market at R7,165m, followed by Zambia atR5,054m and Angola at R2,978m. Kenya, a non-SADC COMESA country,is next at R2,854m.During this period, South Africa liberalised its tariff structure somewhat,including towards certain COMESA countries which participate in theSADC <strong>Trade</strong> Protocol. This may explain some of the shifts and the movementtowards the comparative advantage of the two groups.10<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The proportion of South Africa exports to non-SADC COMESA as a proportionof total South Africa exports to COMESA has grown over the period1995-2005 from 14.2% to 18.3%. As a proportion of South Africa’sworld exports, non-COMESA exports form a minor 1.5% and that of allCOMESA, 8.1%. Once again, whilst the proximity of a country to SouthAfrica heavily influences the direction of South Africa’s exports and thesmall size of African markets in general limits demand, there is the possibilitythat trade diversion is taking place.2.3. <strong>Trade</strong> with COMESA according tobroad classificationsTo understand what drives trade between South Africa and COMESA it isimportant to look at a more disaggregated level at what types of goodsare being traded. This may reveal natural comparative advantage of SouthAfrica and COMESA countries and also disclose any trends in changesof goods traded. Figure 3 looks at changes in the composition of SouthAfrica’s trade with COMESA between 1997 and 2004. The top two piecharts show that in 1997, basic processed goods were the largest sectorof imports from COMESA at 42% of total imports, closely followed bymineral products including oil comprising 34% of the total. Both agricultureand advanced manufactured goods formed smaller shares – 8% and16% respectively. By 2004, imports have become dominated by mineralsand oil at 47% of the total, followed by basic processed goods at 37%.The share of agriculture in imports has fallen to 8%, equalling advancedmanufactured imports, the share of which remains unchanged.In terms of exports to COMESA the shares have remained fairly similarfrom 1997-2004, with basic goods dominating exports, increasing from48% in 1997 to 54% in 2004. Advanced manufacturing is the secondlargestexport sector, although the share has fallen slightly from 32% in1997 to 26% in 2004.Finally, agriculture and mining make up smaller shares, both totalling 10%in 1997 and then 7% for agriculture and 13% for mining in 2004. Theresults are not surprising. South Africa’s economy has a more developedmanufacturing base than the rest of Africa, accounting for the exports ofadvanced manufactures. Due to the size of its market it has probably beenbetter able to take advantage of economies of scale, which could explainthe net export of basic processed goods.Appendix 1 illustrates the full table of trade between South Africa andCOMESA at the 23-chapter level 7 . We have captured the information inTables 6 and 7 by grouping chapters by the size of their share in all importsor exports and by their growth rate. This allows us to see chapterswith particularly promising trade potential. Chapters are defined as highgrowth if their growth rate is higher than 20% per annum. Low growthchapters grow by less than 10% per annum, with medium growth in between.A chapter is considered high share if it forms more than 10% oftotal imports or exports, low share if below 5% and medium share if it isbetween 5% and 10% of imports or exports.Table 6 reflects the slow growth of South Africa’s exports to COMESA.Most chapters fall into the low growth category, with most also low share.Only vehicle parts, optical equipment and base metals exhibit growthabove 10% per annum.Table 7 is a more upbeat story, with a number of chapters exhibiting highgrowth rates of COMESA exports to South Africa. High share and highgrowth chapters are base metals and prepared foodstuffs. Textiles exports,where we might expect COMESA countries to have a comparative advantagecompared with South Africa, exhibit a low growth rate even thoughthey represent a high share.2.4 Intra-industry trade (IIT)Conventional trade theory arising from the Heckscher-Ohlin Model assumesconstant returns to scale where the cost of production remains thesame at any level of industry size or output level. Reducing trade barrierswill lead to countries taking advantage of their comparative advantage, inother words specialising in particular industries that use the factors theyhave available in large quantities, such as land or labour.However, where branding is important and the cost of production falls asthe industry size or level of production increases (if scale economies exist),intra-industry predominates. European trade is characterised by this typeof trade where a country may import and export fizzy drinks but it is thebrand name, such as Coke or Pepsi, that influences the demand.7 Chapter 23 is the level of product groupings we use to measure the trade. Globally, chapterheadings are more usually measured as 21 different chapters.
TABLE 5: SOUTH AFRICA’S EXPORTS TO COMESA I(IN R-MILLION)Growth1995-2005(%)1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005World 100,447 114,133 137,339 142,740 161,508 208,285 215,248 277,993 255,560 291,129 320,005 12.6Angola 391.7 663.4 876.9 1,079.6 1,223.5 1,289.7 2,472.1 3,189.6 3,130.5 2,932.1 2,978.6 23.0Burundi 13.7 10.1 39.2 23.0 19.0 34.0 20.5 43.2 29.8 28.0 34.4 8.7DRC 632.8 946.6 906.3 1,045.7 774.9 857.6 864.4 1,424.5 1,230.2 1,329.1 1,747.5 7.6Djibouti 2.8 1.4 2.4 2.1 18.0 22.6 44.6 43.7 26.9 37.7 35.5 44.0Egypt 89.6 185.4 145.0 132.8 139.0 119.6 259.7 183.3 247.0 159.2 243.6 6.8Eritrea 0.0 0.0 0.0 1.4 1.0 0.5 11.5 14.0 21.8 15.2 24.5 0.0Ethiopia 14.1 62.5 56.4 66.8 71.8 68.6 92.1 130.8 113.7 230.5 125.3 19.8Kenya 835.4 906.5 1,584.3 1,252.7 1,180.0 1,429.0 1,700.4 2,192.4 2,129.9 2,855.7 2,853.7 12.7Comoros 73.6 83.9 110.5 93.8 75.5 59.8 76.1 132.7 122.5 75.1 47.0 -1.5Libyan Arab Jamahiriya 1.2 0.8 18.9 14.7 14.3 48.3 31.1 69.3 40.9 44.6 <strong>38</strong>.9 43.4Madagascar 161.4 214.6 302.7 254.5 245.5 282.0 410.7 419.1 722.2 557.4 523.4 13.4Mauritius 655.0 927.1 1,181.1 1,026.9 1,185.0 1,931.6 2,005.2 2,590.7 2,006.1 1,713.3 2,076.0 11.7Malawi 627.3 1,034.0 1,127.1 1,213.2 1,342.9 1,555.6 1,679.0 2,187.0 1,610.7 1,481.4 1,555.0 8.0Rwanda 15.0 12.3 36.6 84.8 45.9 35.0 60.9 73.0 83.4 67.5 59.0 15.8Seychelles 120.7 162.8 188.9 179.0 198.2 208.4 214.0 349.1 290.9 218.6 327.5 8.4Sudan 18.2 78.2 23.3 60.8 29.2 47.8 124.1 294.9 275.2 283.9 426.0 34.9Swaziland 0.1 0.9 0.0 0.0 0.0 0.9 0.0 0.0 0.0 0.0 0.0 0.0Uganda 103.7 109.0 182.0 302.7 2<strong>38</strong>.6 327.4 444.4 528.2 450.0 497.4 520.2 18.4Zambia 1,235.1 1,757.2 2,208.9 2,126.3 2,218.3 4,168.1 4,362.9 5,002.5 3,799.8 4,274.7 5,054.4 14.2Zimbabwe 4,098.4 5,326.0 5,575.7 5,291.3 4,772.9 4,543.7 4,965.6 6,901.5 6,213.5 5,849.6 7,165.3 3.8Total SA exports to non-SADC COMESA 1,288.0 1,612.9 2,<strong>38</strong>7.5 2,214.4 2,030.5 2,400.9 3,079.4 4,054.5 3,832.2 4,513.4 4,735.6 13.2Total SA exports to all COMESA 9,089.9 12,482.6 14,566.2 14,252.0 13,793.5 17,030.0 19,839.2 25,769.4 22,545.0 22,651.1 25,835.7 10.014.2 12.9 16.4 15.5 14.7 14.1 15.5 15.7 17.0 19.9 18.3 3.0Proportion of SA exports to non-SADCCOMESA (%)9.0 10.9 10.6 10.0 8.5 8.2 9.2 9.3 8.8 7.8 8.1 -2.3Exports to COMESA as % of exports toworld1.3 1.4 1.7 1.6 1.3 1.2 1.4 1.5 1.5 1.6 1.5 0.6Exports to non-SADC COMESA as % ofexports to worldSource : TIPS database and own calculations11<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> LiberalisationFIGURE 3: CHANGES IN COMPOSITION OF SOUTH AFRICA AND COMESA’S TRADE, 1997-200412<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: TIPS and COMESA database and own calculationsThe Grubel-Lloyd Index (GLI) estimates the level of intra-industry tradebetween two groups of countries. It is simply total trade less net trade,divided by total trade. We identified the 100 most traded HS6 commoditygroups and then ranked them by the GLI. Only 54 commodity groups in the100 largest groups by total trade show evidence of positive IIT; these arecaptured in Table 8. We then calculated the overall level of intra-industrytrade by the weighted GLI. IIT is low especially when compared to SouthAfrican trade with Europe where the weighted GLI is 0.13. The weightedGLI for South Africa and the world is 0.18. However, IIT with India andChina is even lower, where the weighted GLI is 0.01 for both countries.The low IIT between South Africa and COMESA suggests that a reductionin trade barriers will lead to structural adjustment, and whilst it mightbe less severe than adjustment if barriers were removed with India andChina, it could still be significant for both South Africa and COMESA.determine whether there is a bias, positive or negative, towards importingfrom this source. The import intensity index for South African imports fromCOMESA is the proportion of COMESA imports in total South African imports,divided by the total exports from COMESA, divided by world exportsexcluding South Africa. It is written:EQUATION 1: (INTENSITY OF IMPORTS)M ij= South Africa’s imports from COMESAM i= Total imports of South AfricaX w= Total world exports (trade)X i, X j= Total South African exports and total exports of COMESA,respectivelyThe index of intensity of COMESA’s import trade with South Africa is defined as:2.5. <strong>Trade</strong> intensities between South Africaand COMESAThe previous sections looked at the trade flows between the two groupswithout placing it in a wider context. By examining the proportion ofSouth African trade with COMESA and comparing it to total trade we canEQUATION 2: (INTENSITY OF EXPORTS)X ij= South Africa’s exports to COMESAX w= Total world imports (trade)M j= Total COMESA imports
TABLE 6: GROWTH SHARE MATRIX FOR SOUTH AFRICA’S EXPORTS TO COMESALow share Medium share High shareHigh growth Vehicle parts (23)Medium growth Optical equipment (18) Base metals (15)Low growth Live animals (1)Animal or vegetable fats (3)Leather products (8)Wood products (9)Wood pulp and paper (10)Textile and textile articles (11)Foot wear (12)Stone and glass (13)Precious metals (14)Arms (19)Misc manufacturers (20)Art and antiques (21)Unclassified goods (22)Source: TIPS and COMESA database and own calculationsTABLE 7: GROWTH SHARE MATRIX FOR SOUTH AFRICA’S IMPORTS FROM COMESAVegetable products (2)Prepared foodstuffs (4)Plastics (7)Vehicles (17)Mineral products (5)Chemicals (6)Machinery (16)13<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>High growth Machinery (16)Chemicals (6)Live animals (1)Leather products (8)Optical equipment (18)Medium growth Wood pulp and paper (10)Foot wear (12)Vehicle parts (23)Low share Medium share High sharePrepared foodstuffs (4)Base metals (15)Low growth Animal or vegetable fats (3)Plastics (7)Wood products (9)Stone and glass (13)Precious metals (14)Vehicles (17)Arms (19)Misch manufactures (20)Art and antiques (21)Unclassified goods (22)Vegetable products (2) Mineral products (5)Textile and textile articles (11)Source: TIPS and COMESA database and own calculationsIf the value of the index is greater than 1, this indicates South Africanconsumers are biased towards COMESA imports or COMESA exportersare biased towards exporting to South Africa. If the value of the index isequal to one, trade is not geographically biased. If the value is less thanone, South African consumers prefer non-COMESA goods or COMESA exportersprefer exporting elsewhere.Figure 4 reveals that for both South Africa and COMESA, there is a biastowards trading with each other. However, whilst the import intensity ofCOMESA exports to South Africa is approximately 2, the export intensity isover 30. COMESA countries therefore have a strong preference for SouthAfrican imports, far more than South African consumers do for COMESAimports. This may conceal a number of issues, including trade diversion.2.6. Tariff barriersTariffs can reduce the amount of trade that takes place, and if tariffs arevery high, stop trade taking place completely. Traditionally, tariffs havebeen used to protect domestic industries. Inadvertently this can lead touncompetitive industries and restricting downstream industries. SouthIntegrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> LiberalisationTABLE 8: THE GRUBEL LLOYD INDEX OF INTRA-INDUSTRY TRADE14<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>HS6 code anddescriptionSA exportsto COMESA (US$)SA imports fromCOMESA (US$)1 H854459: Electric conductors, 80-1,000 volts, no connectors 40,118,843 57,800,880 0.822 H170111: Raw sugar, cane 29,815,840 46,490,789 0.783 H520512: Cotton yarn >85% single uncombed 714-232 dtex,not ret. 23,770,622 41,274,378 0.734 H270400: Coke, semi-coke of coal, lignite, peat & retort carbon 17,542,528 62,804,679 0.445 H271290: Mineral waxes nes 37,586,157 8,432,892 0.376 H880230: Fixed wing aircraft, unladen weight 2,000-15,000 kg 33,345,000 190,090,594 0.307 H240120: Tobacco, unmanufactured, stemmed or stripped 23,201,520 135,059,516 0.298 H721049: Flat rolled i/nas, coated with zinc, width >600mm, nes 62,119,139 9,393,142 0.269 H440710: Lumber, coniferous (softwood) thickness < 6 mm 12,091,554 83,090,631 0.2510 H721420: Bar/rod, i/nas, indented or twisted, nes 45,183,859 5,558,622 0.2211 H870323: Automobiles, spark ignition engine of 1500-3000 cc 70,598,636 7,188,372 0.1812 H170199: Refined sugar, in solid form, nes, pure sucrose 288,270,562 29,292,789 0.1813 H240220: Cigarettes containing tobacco 75,078,169 7,055,791 0.1714 H090240: Tea, black (fermented or partly) in packages > 3 kg 4,321,679 102,476,594 0.0815 H847490: Parts for mineral sort, screen, mix, etc machines 13,986,122 5,454,062 0.0816 H210690: Food preparations nes 98,602,621 3,657,956 0.0717 H390120: Polyethylene - specific gravity >0.94 in primary forms 72,529,568 2,540,483 0.0718 H870333: Automobiles, diesel engine of >2500 cc 48,925,353 1,698,764 0.0719 H852812: Color television receive 45,763,355 1,492,575 0.0620 H401120: Pneumatic tyres new of rubber for buses or lorries 181,824,150 5,539,0<strong>38</strong> 0.0621 H721310: Hot rolled bar/rod grooved i/nas in irregular coils 105,774,901 2,862,704 0.0522 H340111: Soaps, for toilet use, solid 79,368,444 2,045,357 0.0523 H100590: Maize except seed corn 418,180,495 10,480,873 0.0524 H490199: Printed reading books, except dictionaries etc 54,126,429 1,231,835 0.0425 H300490: Medicaments nes, in dosage 155,099,354 3,458,655 0.0426 H841391: Parts of pumps for liquids 76,847,192 1,565,294 0.0427 H847330: Parts and accessories of data processing equipment nes 119,332,597 2,240,185 0.0428 H100190: Wheat except durum wheat, and meslin 65,375,623 992,275 0.0329 H732690: Articles of iron or steel, nes 57,899,759 863,140 0.0330 H852520: Transmit-receive apparatus for radio, TV, etc. 44,847,952 608,790 0.0331 H100510: Maize (corn) seed 161,964,2<strong>38</strong> 2,151,323 0.0332 H870190: Wheeled tractors nes 62,704,328 759,000 0.0233 H392690: Plastic articles nes 48,159,286 542,1<strong>38</strong> 0.0234 H330499: Beauty, makeup and suntan preparations nes 46,331,740 502,525 0.0235 H848180: Taps, cocks, valves and similar appliances, nes 54,525,431 535,943 0.0236 H<strong>38</strong>0810: Insecticides, packaged for retail sale 1<strong>38</strong>,422,679 1,3<strong>38</strong>,642 0.0237 H740811: Wire of refined copper > 6mm wide 1,608,782 192,894,413 0.02<strong>38</strong> H730890: Structures and parts of structures, iron or steel, nes 235,103,081 1,487,688 0.0139 H843149: Parts of cranes, work-trucks, shovels, constr machine 78,241,555 464,421 0.0140 H480100: Newsprint 87,553,255 467,080 0.0141 H841<strong>38</strong>1: Pumps nes 63,626,067 283,740 0.0142 H<strong>38</strong>0830: Herbicides, sprouting and growth regulators 13,622,7862 549,842 0.01GLI
43 H940600: Prefabricated buildings 62,194,367 232,841 0.0144 H330590: Hair preparations, nes 54,463,561 197,478 0.0145 H841370: Centrifugal pumps nes 71,224,839 244,064 0.0146 H630533: Sacks&bags,f/pckg polyet 50,856,476 166,812 0.0147 H360300: Safety or detonating fuses, detonators, igniters 80,635,285 260,246 0.0148 H390760: Polyethylene terephthalate, in primary forms 67,307,113 205,064 0.0149 H360200: Prepared explosives, except propellent powders 83,479,401 247,834 0.0150 H481910: Cartons, boxes & cases, of corrugated paper or board 75,852,637 220,479 0.0151 H270900: Petroleum oils, oils from bituminous minerals, crude 4,607,120 1,671,750,746 0.0152 H<strong>38</strong>2490: Chemical prep, allied in 70,582,370 191,069 0.0153 H870899: Motor vehicle parts nes 110,326,347 297,660 0.0154 H220210: Beverage waters, sweetened or flavoured 242,839,308 647,908 0.01Weighted average GLI for all HS 6 Groups 0.060331Source: TIPS and COMESA database and own calculations15<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Africa was forced into protection as a development strategy when sanctionswere imposed during the apartheid era, but it is becoming moreliberal.The level of protection amongst COMESA members has varied, rangingfrom some very liberal countries such as Uganda to others with muchmore protective tariff structures. A number of countries in COMESA operatein a free trade area. Recently, COMESA countries have agreed a commonexternal tariff structure, which is being implemented by members.COMESA will be a customs union by 2008. For most members this willmean a simplification and liberalisation of their tariffs.We have looked at the tariff structures of South Africa and COMESA atthe HS6 level. Table 9 shows the tariff structure of South Africa towardsCOMESA countries that are not in SADC. This is essentially the MFN statusgiven to other WTO members as the basic level of preference available.The tariff structure is complicated. We have split tariffs into the eightbands in Table 9, but there is actually no organised approach to tariffsin South African trade policy. The tariff structure is fairly liberal, with over50% of all tariffs lines between 0% and 0.9% and less than 30% oftariff lines greater than 15%. There are, however, a number of tariff peaksabove 30%. 67% of COMESA’s exports into South Africa enter at the0%-0.9% tariff level because they consist of commodities. This is representedvisually in Figure 5 and reveals a fairly liberal tariff structure.COMESA has agreed a common tariff nomenclature, although it is notyet implemented. The tariff structure has been organised into four bands:0% for raw materials, 0% for capital goods, 10%-15% for intermediategoods and 25%-40% for finished goods. On the implementation of thecustoms union, an external tariff will be agreed within these bands. InTable 10 we have taken two scenarios – a high tariff scenario wheremembers of COMESA choose a common external tariff at the maximumpoints in the bands of 0%, 15% and 40% and a low tariff scenario wheremembers agree a common external tariff at the minimum point of thebands of 0%, 10% and 25%.48% of South Africa’s exports to COMESA enter at the 0% tariff but32% enter at the highest tariff band of 25%-40% tariff. Figure 5 and 6,when compared, reveal that South African exports face higher tariffs intoCOMESA than COMESA’s exports into South Africa.2.7. Revealed comparative advantage andrevealed trade barriersSo far we have looked at which sectors have been growing and at thetariff structures facing trade between COMESA and South Africa. This sectionbegins to explore the idea of comparative advantage and suggestswhich commodity groups would be traded if no trade barriers existed betweenthe two. We apply the concept of revealed comparative advantage(RCA).Comparative advantage predicts that in a world without trade barriers, acountry will specialise in the production of goods that uses its abundant‘factor of production’, such as labour or land, because it can produce theIntegrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> Liberalisation16<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>good more efficiently as it uses its abundant factor. It will then exportthis good, using the surplus created to purchase imports of other goods.A comparative advantage is revealed in a particular commodity groupif its share in the country’s export basket is larger than the share of thecommodity’s world trade in total world trade; in other words, whetherthe commodity is more important to South Africa’s exports than to worldtrade:EQUATION 3:in which X ikis equal to exports of country i in product k. The results for this calculationare reported in Table 11. The value for the RCA index is the average of the RCA index forthe years 1999-2004. A value above 1 indicates comparative advantage, with a highervalue indicating a stronger advantage.Table 11 shows the commodity groups at the HS2 level for which SouthAfrican exports reveal a comparative advantage. They fall into the categoriesof precious stones and precious and base metals, fruit and vegetables.Tobacco and beverages also reveal a high comparative advantage.COMESA countries reveal a strong comparative advantage in primarycommodities such as coffee, tobacco and mineral oils (Table 12). Althoughsome commodity groups such as tobacco are repeated in both the SouthAfrica and COMESA RCA Index, which could suggest some intra-industrytrade, it is more likely to be showing value added taking place in SouthAfrica of COMESA raw imports, which are then being exported back outto COMESA in the more processed form.2.8 Revealed trade barriers (RTBs)RTBs start to unpick the overall barriers to trade, including non-tariff barrierssuch as transport costs, taste and technical barriers to trade. If theindex is below 1 we can conclude South Africa is exporting more to theworld than to COMESA and that trade barriers may exist. The same logicapplies when we consider COMESA’s exports to South Africa. Equation 4calculates the RTB index:EQUATION 4:jin which M ikis country j’s imports from country i of product k. The results of the computationsare shown in full in Appendix 2.The results are interesting. Over 1999-2004, COMESA exported tobacco,cotton, ores, copper, coffee, oil seeds and live plants into South Africa inmuch greater proportions than to the rest of the world. For these groups,the RTB index is greater than 13% and in the case of tobacco, cotton, oreand copper between 23% and 29%. Tariffs on these imports are relativelylow, between 0% and 8% except for tobacco where the unweighted averagetariff is high at 27.9%. For 60% of commodity groups the index isbelow 1, indicating some RTB exists.In terms of South African exports to COMESA, the highest RTB index scoreis 3.3. This indicates in general that South Africa’s exports to COMESAcompared to the rest of the world are more evenly balanced. However,fruits and nuts, live animals, base metals and explosives are amongst thegroups that are exported in greater proportions to COMESA than to therest of the world. In term of RTBs, 51% of commodity groups exhibit somefrom of trade barrier.2.9. Market access gainsThe last part of our trade analysis will assess the likely market accessgains (MAG) to both South Africa and COMESA if respective trade barrierswere reduced – essentially if they agreed an FTA. This is importantfor us to see whether an FTA between the two can be justified on solelyeconomic grounds.We assume that an FTA will result in reductions of all tariffs to 0 on bothsides. Although this is not wholly realistic it does allow us to frame potentialbenefits. The market access gains are calculated assuming that importsupply is perfectly elastic and therefore equal to infinity, and we assumethat import demand elasticity is 2 so that a reduction in a tariff by 1% willlead to an increase in import demand of 2%. This is of course a simplifyingassumption, as the demand and supply elasticity of goods varies; however,we have no data on individual elasticities that can be called upon.The equation for COMESA MAG from South African tariffs falling to zerobecomes:EQUATION 5where t iis the tariff for good i.Table 13 summarises the MAG gains for COMESA exports into South Africafor the products that gain most from the introduction of zero tariffs(see Appendix 3 for full tables). Over 80% of the gains come from thecommodity groups of live animals, meat, fish, dairy, trees, vegetables andcoffee. Live animals make up 35.6% of total MAG from an FTA. Total MAGfor COMESA to enter an FTA with South Africa are US$121.076m in 2004alone. This is equivalent to a present discounted value of US$818.356m,assuming an interest rate of 10% and no change to the current level of
FIGURE 4: TRADE INTENSITIES BETWEEN SOUTH AFRICA AND COMESA, 1999-200417<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: TIPS and COMESA database and own calculationsTABLE 9: SOUTH AFRICAN TARIFFS AND THE ASSOCIATED IMPORTS FROM COMESA AND THE WORLD#HS6 linesImports fromCOMESA (US$)Imports fromthe World (US$)% of all HS6lines% imports fromCOMESA% totalimports40%+ 233 199,974,021 3,881,098,658 4.46 3.05 1.2730-39% 154 182,518,656 4,159,276,077 2.95 2.78 1.3620-29% 602 248,978,827 54,970,055,765 11.52 3.79 17.9715-19% 421 374,637,883 6,981,120,368 8.06 5.71 2.2810-14% 441 865,818,923 21,404,594,010 8.44 13.19 7.005-9% 545 259,955,743 28,080,198,451 10.43 3.96 9.181-4% 65 21,095,725 10,452,826,140 1.24 0.32 3.420-0.9% 2,763 4,410,031,818 176,048,771,429 52.89 67.20 57.54Total 5,224 6,563,011,596 305,977,940,898 100 100 100Source: TIPS and COMESA database and own calculationsimports or tariff schedule. If the growth in imports continues, this figurein reality is much higher.The potential gains for South Africa are more complicated to assess. Firstwe must look at the range depending on whether the countries choosethe top or bottom of the tariff bands allowed under the COMESA commonexternal tariff. Table 14 illustrates the top 10 products that willgain from COMESA liberalising its tariffs to zero. Mineral fuels and ironand steel stand to gain the most from zero tariffs, at US$98.735m andUS$85.815m, respectively, if the CET was originally at the uppermost endof the band. Overall South Africa stands to gain between US$599.46mand US$874.98m in MAG from entering an FTA with COMESA. This translatesinto a gain of between US$4.052bn and US$5.914bn over the next10 years if we again discount future wealth by 10%. These gains aresignificant and should not be ignored.Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> LiberalisationFIGURE 5: SOUTH AFRICA’S TARIFFS AND THEIR ASSOCIATED IMPORTS FROM THE WORLD AND COMESA, 200418<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: TIPS and COMESA database and own calculationsTABLE 10: COMESA TARIFFS AND THE ASSOCIATED IMPORTS FROM SOUTH AFRICA AND THE WORLD (2004), US$HIGH LOW Lines Imports from SAImports fromWorld% of all HS6lines% imports fromSA% total imports40 25 1,551 901,909,545 4,760,280,068 30 32 3515 10 2,044 878,137,268 2,995,272,994 40 20 340 0 1,520 798,182,719 7,173,776,906 30 48 315,115 2,578,229,532 14,929,329,968 100 100 100Source: TIPS and COMESA database and own calculationsHowever, before we get too excited in the case of South Africa exportingto COMESA we need to take into account the fact that South Africaalready has an FTA with SADC and therefore some of the members ofCOMESA have already agreed to reduce tariffs to zero. Using COMESAdata which are somewhat different to the TIPS database but allow us todisaggregate by country, we break MAG down, excluding the COMESAmembers who are currently members of SADC. Table 15 illustrates that ifwe exclude the SADC COMESA members, the MAG are much lower – inthe range of US$47.768m and US$68.459m – and would lead to a gainof between US$200m to US$400m over the next 10 years, given a 10%discount rate.3. An FTA with COMESAand SACU – solving theproblem?The fastest path to integration is probably for SACU and COMESA to signan FTA and for SADC’s trade capacity to merge with SACU and COMESA,leaving SADC to lead in other areas such as water, energy and peace andsecurity. This is because it is highly unlikely that SADC with its currentmembership will be able to find agreement on a customs union, and thispaper asserts that a customs union with the free flow of goods is a bettermechanism for integrating markets than an FTA. Figure 7 indicates thecurrent members of SACU and COMESA with a star and reveals that anFTA between the two RECs would unite at least half of Africa, in land
mass and the number of countries if the customs unions became fullyoperational.Apart from Swaziland, COMESA and SACU have no overlapping members.An FTA between the two groups would be desirable for two mainreasons:1. Half of Africa would be united in an FTA.2. It would remove the problem of overlapping membership; jointmembership of SADC, COMESA and SACU is only undesirableon the trade side or where mandates are being duplicated. TheSADC <strong>Trade</strong> Protocol would inform and be replaced by the SACUCOMESA FTA, leaving SADC to take the lead on areas in which itcurrently has a comparative advantage.3. The AU REC survey found that overlapping membership is drivenby political and economic motives. Agreeing an FTA would accommodateall economic interests, as countries would have accessto both South Africa and COMESA markets.This of course assumes that all countries belonging to COMESA wouldjoin its customs union arrangement – currently a major assumption althoughthere is a growing number of countries committed to the COMESAcustoms union.4. Non-tariff barriers (NTBs)According to the study commissioned by the South African Departmentof Transport, known as Moving South Africa, a load travelling fromSouth Africa to a destination within SADC costs between 46% and 119%more than it does moving within the Republic. A recent World Bank studyfound that a shipment from Baltimore to Durban costs US$2,500, whilsta shipment from Baltimore to Maseru via Durban cost US$12,000. Thisis a problem of NTBs. A plethora of NTBs exist in the region, such asnon-acceptance of certificates of origin, temporary bans on products, visarequirements, non acceptance of standards, customs blockages and poorinfrastructure.As tariff rates fall through WTO negotiations or bilateral agreements,NTBs becomes increasingly important to facilitate trade. Mold (2005)looks at the prevalence and relevance of NTBs to African countries andfinds that they have frequently suffered consequences through lost exportmarkets of NTBs imposed by industrialised countries, and in particular dueto anti-dumping measures and standards compliance. Yet African countriesare also guilty of putting up administrative impediments to intraregionaltrade. An inventory of SADC countries’ NTBs was carried out byImani (2004). They found that NTBs in the SADC region were a serious im-19<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>FIGURE 6: COMESA’S TARIFFS AND THEIR ASSOCIATED IMPORTS FROM THE WORLD AND SOUTH AFRICA, 2004Source: TIPS and COMESA database and own calculationsIntegrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> LiberalisationTABLE 11: REVEALED COMPARATIVE ADVANTAGE OF SOUTH AFRICAN EXPORTS, 1995-2004 (US$’000)20<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>ProductProduct nameSA RCA index1999-2005SA exports to COMESA2004SA exports to World200471 Precious metals (natural/cultured pearls, prec stone) 11.7 7,597 158,849,62426 Ores, slag and ash. 9.5 404 59,779,6888 Edible fruit and nuts; peel of citr 8.2 18,860 46,954,10424 Tobacco and manufactured tobacco su 5.0 23,781 23,4<strong>38</strong>,66722 Beverages, spirits and vinegar. 4.4 16,629 53,267,83272 Iron and steel. 4.2 341,091 240,330,<strong>38</strong>451 Wool, fine/coarse animal hair, hors 3.3 2,899 12,133,37681 Other base metals; cermets; article 3.0 55,189 10,157,56517 Sugars and sugar confectionery. 2.9 53,812 18,6<strong>38</strong>,31625 Salt; sulphur; earth & ston; plaste 2.9 28,861 26,212,12220 Prep of vegetable, fruit, nuts or o 2.9 14,905 27,639,46336 Explosives; pyrotechnic prod; match 2.8 24,345 2,194,34697 Works of art, collectors’ pieces an 2.5 358 11,676,82314 Vegetable plaiting materials; veget 2.5 37 523,18147 Pulp of wood/of other fibrous cellu 2.3 4,641 27,447,60576 Aluminium and articles thereof. 2.2 31,909 87,639,91375 Nickel and articles thereof. 2.2 224 18,775,81028 Inorgn chem; compds of prec mtl, r 2.0 57,295 59,496,8923 Fish & crustacean, mollusc & other 1.7 5,748 55,107,04493 Arms and ammunition; parts and acc 1.6 492 4,525,90141 Raw hides and skins (other than fu 1.6 634 23,470,45631 Fertilisers. 1.4 69,698 25,122,178<strong>38</strong> Miscellaneous chemical products. 1.3 86,127 80,239,48644 Wood and articles of wood; wood ch 1.1 19,045 97,176,31729 Organic chemicals. 1.1 39,464 245,161,24974 Copper and articles thereof. 1.1 49,425 58,029,120Source: TIPS and COMESA database and own calculationspediment to trade. They found the biggest barriers faced by traders to beagricultural commodities. Commodities worst affected were sugar, maize,meat products, dairy, tea, timber and seasonable vegetables. The worstNTBs in the region are caused by customs administration bottlenecks.Whilst the relative cost to intra-regional trade caused by NTBs as opposedto tariffs is not quantified, NTBs are without doubt a major impediment.The moves now being taken by COMESA and SARS to implement onestopborder posts and by SADC to agree and implement an NTB RemovalAction Plan are a priority. As the level of tariffs fall within the region, tradewill increase, and investment in physical and human capacity to facilitatetrade will be required to ensure the economic benefits are reaped.5. ConclusionThis paper has included a detailed analysis of the current and potentialtrade between South Africa (proxying for SACU) and COMESA. The purposeof this exercise is to provide practical information on one option formore rapid integration in Africa.South Africa in particular has played a strong leadership role in the AU’sNepad, which commits to integration in Africa. If it is serious about thisend, it should consider the merits of an FTA with COMESA on the potentialrole it would play in integrating Africa, as it would contribute towardsintegrating half of Africa, and towards faster economic development.
TABLE 12: REVEALED COMPARATIVE ADVANTAGE OF COMESA EXPORTS, 1995-2004 (US$’000)ProductProduct nameCOMESA RCA index1999-2005COMESA exports to SA2004COMESA exports to World20049 Coffee, tea, mate and spices 14.5 23,519 1,314,53324 Tobacco and manufactured tobacco su 7.4 23,908 795,37627 Mineral fuels, oils & product of th 6.4 273,054 39,493,84717 Sugars and sugar confectionery. 5.9 14,614 758,9016 Live tree & other plant; bulb, root 5.6 1,140 481,71181 Other base metals; cermets; article 4.4 777 429,31813 Lac; gums, resins & other vegetable 4.0 228 99,24514 Vegetable plaiting materials; veget 3.7 182 10,24852 Cotton. 3.5 139,850 803,65716 Prep of meat, fish or crustaceans 2.7 4,249 <strong>38</strong>4,1687 Edible vegetables and certain roots 2.7 3,752 556,84225 Salt; sulphur; earth & ston; plaste 2.7 4,357 733,11757 Carpets and other textile floor co 2.4 1,337 170,06753 Other vegetable textile fibres; pap 2.4 780 49,63531 Fertilisers 2.2 86 325,85661 Art of apparel & clothing access, 2.2 10,750 1,682,62275 Nickel and articles thereof. 2.1 26,030 178,81041 Raw hides and skins (other than fu 1.8 3,895 286,95571 Natural/cultured pearls, prec stone 1.8 1,860 931,08512 Oil seed, oleagi fruits; miscell gr 1.8 10,941 369,6671 Live animals 1,7 163 41,<strong>38</strong>674 Copper and articles thereof. 1.6 61,851 585,9253 Fish & crustacean, mollusc & other 1.6 615 578,92362 Art of apparel & clothing access, n 1.6 14,592 1,331,9678 Edible fruit and nuts; peel of citr 1.2 2,378 <strong>38</strong>9,34621<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>63 Other made up textile articles; set 1.2 666 194,72446 Manufactures of straw, esparto/othe 1.1 104 11,11626 Ores, slag and ash 1.1 209,891 598,28721 Miscellaneous edible preparations. 1.1 660 175,767Source: TIPS and COMESA database and own calculationsTABLE 13: COMESA MARKET ACCESS GAINS INTO SA, US$’000 PER YEAR, 2004 IMPORTSProduct Product name MAG (US$’000)% of totalMAG52 Live animals 43,097 35.627 Meat 18,756 15.524 Fish 10,418 8.662 Dairy prods; birds eggs; honey; 7,777 6.474 Products of animal origin 6,021 5.061 Live trees, plants, cut flowers 5,810 4.844 Edible vegetables 3,865 3.217 Edible fruit & nuts 2,782 2.394 Coffee, tea, mate & spices 2,421 2.0Total MAG 121,076 100.0Source: TIPS and COMESA data base and own calculationsIntegrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> LiberalisationTABLE 14: SOUTH AFRICA MARKET ACCESS GAINS INTO COMESA, US$’000 PER YEAR, 2004 IMPORTS22<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>ProductProduct nameMAGlowMAGhighTABLE 15: SOUTH AFRICA MARKET ACCESS GAINS INTO ALL COMESA AND NON-SADC COMESA, US$%MAGlow%MAGhigh27 Mineral fuels, oils & product of th 67,341 98,735 11.2 11.372 Iron and steel 59,714 85,815 10.0 9.839 Plastics and articles thereof 45,241 65,727 7.5 7.573 Articles of iron or steel 28,163 41,081 4.7 4.787 Vehicles o/t railw/tramw roll-stock 27,957 42,729 4.7 4.948 Paper & paperboard; art of paper pu 27,633 40,063 4.6 4.685 Electrical mchy equip parts thereof 24,274 37,168 4.0 4.2<strong>38</strong> Miscellaneous chemical products. 17,667 25,506 2.9 2.940 Rubber and articles thereof. 15,983 23,476 2.7 2.715 Animal/veg fats & oils & their clea 15,180 22,046 2.5 2.5Total MAG 599,455 874,984 100.0 100.0Source: TIPS and COMESA database and own calculationsMAG - All COMESAlow CETMAG - All COMESAhigh CETMAG - non-SADCCOMESAlow CETMAG - non-SADCCOMESAhigh CETTotal US$ 520,425,140 744,456,170 47,768,175 68,458,476Source: TIPS and COMESA database and own calculationsFIGURE 7: AFRICA.Stars indicate countries belonging to SACU or COMESA– they span half of Africa, stars in the sea indicate theisland nations of Mauritius, Comores and Seychelles
At the moment, Southern Africa seems to have lost momentum in thetrade integration agenda. One reason may be the dominance of SouthAfrica and potential fear countries outside of SACU have in terms of movingtowards a customs union. But another clear reason is that SACU hasagreed a number of bilateral trade agreements that will inevitably makenegotiating a SADC customs union more difficult, unless the SACU commonexternal tariff is adopted or SACU accepts that its own commonexternal tariff can be renegotiated.COMESA is a significantly more important export market for South Africaexporters than either China or India. In 2003, exports to COMESA werethree times more than exports to China and six times more than exportsto India. Making trade work better in the region is imperative for growthin Africa. The low growth in trade between COMESA and SACU over thepast 10 years and compared to countries like China and India shouldalso be taken as a reason to speed up regional integration in Africa withwhatever complementary policies may be necessary.The suggestion of an FTA between SACU and COMESA will have implicationsfor the SADC <strong>Trade</strong> Protocol and for membership of SADC in termsof the trade agenda. Countries can still be members of an REC even if theyare not part of a customs union of that REC, and one possibility would bethat the SADC FTA becomes the SACU COMESA FTA.This paper should not be taken as a criticism of the SADC trade agendaso far; it has been a significant achievement to reach its current stage.However, implementation of the SADC <strong>Trade</strong> Protocol has been slow anda number of difficulties stand in the way of implementing a customs unionthat should not be ignored. In addition, SADC is leading in a numberof fundamental areas in regional integration, such as energy, water andfinance, and a division of labour between the three RECs of SACU, SADCand COMESA could have many benefits outside the narrow trade integrationagenda.BibliographyWillcox, O. & Van Seventer, D.E.N. (2004) <strong>Trade</strong> between South Africa andChina: current and future potential, TIPS, October.Willcox, O. & Van Seventer, D.E.N. (2005) <strong>Trade</strong> between South Africa andIndia: current patterns and potential, TIPS, March.Overlapping membership in COMESA, EAC SACU and SADC: trade policyoptions for the region and EPA negotiations.Mills, L. (2006) Overcoming the EPA-TDCA confusion in Southern Africa,DFID funded.No author cited (2006) The case for rationalisation. Experts’ meeting onthe rationalisation of the RECs. Ghana, Accra, 27-28 October, AU Commissionand Economic Commission for Africa.Imani Development Austral Pty Ltd (2005) Inventory of non-tariff barriersin the SADC region. <strong>Vol</strong>ume 1, synthesis report, February.Mold, A. (forthcoming) Non-tariff barriers – their prevalence and relevancefor African Countries. Economic, Commission for Africa. ATPC Workin Progress No.25.Olympio, J., Robinson, P., Cocks, M. & Mills, L. (2004) Study to assess thelikely impact on Southern Africa and EU producers of further liberalisingthe TDCA by granting SA duty-free access to the EU. DFID-funded.Imani Development. SADC non-tariff barrier road map. Pretoria.23<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>An FTA between SACU and COMESA should be considered by both sidesand by SADC who can play a strategic role in informing discussions. Itcould remove the problematic overlap of membership of REC <strong>Trade</strong> Protocolsand it will pave a clear and fast way towards a more integrated Africawhich is able to emerge as a fourth economic pillar against those of theAmericas, Europe and Asia.Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> Liberalisation24<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Appendix 1TABLE 16: TRADE BETWEEN SOUTH AFRICA AND COMESA AT THE 23 CHAPTER LEVELImports from COMESA Exports to COMESA Total <strong>Trade</strong>Ave share2000-2004(%)Ave share1997-2000(%)Ave annualgrowth2000-2004(%)Ave share2000-2004(%)Ave share1997-2000(%)Ave annualgrowth2000-2004(%)Ave share2000-2004(%)Ave share1997-2000(%)Ave annualgrowth2000-2004(%)C01: Live animals 32.07 1.27 0.71 -5.07 1.65 1.46 -8.50 1.58 1.33C02: Vegetables -40.84 9.53 6.67 8.08 6.28 6.29 6.18 6.70 6.29C03: Animal or vegetable fats 1.48 0.23 0.09 -0.93 1.03 0.66 -0.42 0.92 0.57C04: Prepared foodstuffs and tobacco 36.79 12.79 10.12 -5.31 7.84 8.83 -5.13 8.54 8.94C05: Mineral products -4.17 15.99 27.57 -1.50 10.45 11.53 22.50 11.32 14.66C06: Chemicals 130.63 2.52 1.81 3.40 15.03 14.24 3.33 13.29 12.22C07: Plastics -0.22 0.82 0.69 5.30 6.51 6.85 5.81 5.71 5.85C08: Leather 34.03 1.08 1.23 -12.23 0.09 0.09 -4.82 0.22 0.25C09: Wood Products -1.59 4.09 3.67 -0.26 0.62 0.66 5.79 1.10 1.11C10: Wood Pulp and Paper 11.75 0.90 0.89 1.54 4.13 4.43 1.56 3.69 3.85C11: Textiles 2.01 23.94 21.80 -9.59 2.71 2.73 3.08 5.70 5.64C12: Footwear 12.37 0.68 0.63 -2.99 0.30 0.26 -7.32 0.35 0.31C13: Stone and glass -15.94 1.31 1.46 0.88 1.28 1.16 -1.62 1.28 1.18C14: Precious metals -12.73 4.09 0.35 9.45 0.09 0.30 16.09 0.73 0.31C15: Base metals 54.06 11.18 11.69 18.04 11.91 13.73 20.00 11.74 13.31C16: Machinery 32.04 4.92 4.92 4.57 17.93 15.79 4.43 16.06 13.98C17: Vehicles 1.58 1.44 2.51 -7.39 8.75 8.13 -5.17 7.70 7.28C18: Scientific Equipment 25.96 0.60 0.43 11.02 1.35 1.34 10.78 1.24 1.19C19: Arms & ammunition 6.55 0.00 0.00 0.00 0.05 0.00 0.00 0.04 0.00C20: Miscellaneous manufactures 0.00 1.96 1.74 2.15 1.55 1.47 1.81 1.60 1.50C21: Art and antiques 0.65 0.12 0.08 -8.11 0.02 0.01 -26.45 0.03 0.02C22: Unclassified -36.26 0.55 0.93 -47.98 0.39 0.08 -18.35 0.41 0.21C23: Special class; vehicle parts 14.96 0.00 0.00 24.32 0.04 0.01 24.41 0.04 0.01Total 100.00 100.00 2.67 100.00 100.00 7.06 100.00 100.00Source: TIPS and COMESA database and own calculations
Appendix 2: Full tables of RTBs of SA’s exports into COMESA and COMESA’s exports into SATABLE 17: REVEALED TRADE BARRIERS OF SOUTH AFRICA’S EXPORTS TO COMESA, 1994-2005 (US$’000)Product Product name RTB indexSAs exports toCOMESASA’s exports to theWorldUnweighted averagetariff (%)8 Edible fruit and nuts 3.3 114,447.5 8,664,409 7.41 Live animals 3.2 75,590.71 139,922.9 0.036 Explosives 2.7 87,911.65 194,491.6 3.181 Base metals 2.5 59,691.19 684,090.4 1.820 Prep of vegetable, fruit, nuts 2.4 93,789.53 1,679,700 16.378 Lead and articles thereof 2.4 7,974.827 21,751.72 5.179 Zinc and articles thereof 2.3 51,644.49 146,716.6 0.017 Sugars and sugar confectionery 2.3 279,696.3 1,745,083 10.547 Wood pulp 2.1 45,954.46 2,093,810 0.022 Beverages, spirits and vinegar 2.1 149,240.5 3,092,507 17.072 Iron and steel 1.9 1,017,739 19,321,052 7.718 Cocoa and cocoa preparations 1.9 25,583.65 171,379.9 9.331 Fertilisers 1.9 425,800.7 797,036.9 0.086 Railw/tramw locom, rolling-stock & 1.8 34,911.32 225,308.6 9.848 Paper & paperboard; art of paper pu 1.7 543,312.2 2,603,679 16.728 Inorganic chemicals 1.7 262,829 3,303,822 0.9<strong>38</strong> Miscellaneous chemical products 1.7 423,802.6 1,877,091 2.916 Prep of meat, fish or crustaceans, 1.7 34,533.27 188,516.5 10.465 Headgear and parts thereof 1.6 7,090.089 28,212.9 25.176 Aluminium and articles thereof 1.6 141,559.3 5,919,476 6.739 Plastics and articles thereof 1.5 786,<strong>38</strong>0.7 2,324,732 8.233 Essential oils & resinoids; perf, 1.5 166,670 718,346.6 9.344 Wood and articles of wood; wood ch 1.4 110,762.4 3,232,883 10.732 Tanning/dyeing extract; tannins & 1.4 131,993.3 703,113.3 2.843 Furskins and artificial fur 1.4 142.393 9,535.11 9.421 Miscellaneous edible preparations. 1.4 120,562.8 373,104.6 12.473 Articles of iron or steel. 1.4 440,884.5 2,306,699 4.468 Art of stone, plaster, cement, asbe 1.4 42,148.6 360,542.6 25.034 Soap, organic surface-active agents 1.4 103,206.5 308,409.9 13.557 Carpets and other textile floor co 1.3 15,568.04 147,375.3 16.980 Tin and articles thereof 1.3 2,137.461 16,247.52 5.723 Residues & waste from the food indu 1.3 57,441.06 247,023.7 4.035 Albuminoidal subs; modified starche 1.3 26,698.64 101,547.8 2.819 Prep.of cereal, flour, starch/milk; 1.3 79,202.88 220,265.4 18.126 Ores, slag and ash 1.3 9,009.703 11,245,678 0.074 Copper and articles thereof 1.3 74,230.75 1,637,<strong>38</strong>1 3.27 Edible vegetables and certain roots 1.3 94,414.25 319,976.4 10.694 Furniture; bedding, mattress, matt 1.2 162,218.7 3,231,783 0.440 Rubber and articles thereof. 1.2 301,145.7 1,396,235 9.229 Organic chemicals 1.2 189,172.3 2,829,316 1.525<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> Liberalisation26<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>12 Oil seed, oleagi fruits; miscell gr 1.2 49,166.19 416,678.8 6.85 Products of animal origin, nes or 1.1 4,440.497 64,524.3 0.061 Art of apparel & clothing access, 1.1 50,903.27 1,011,818 11.759 Impregnated, coated, cover/laminate 1.1 34,071.23 102,272.4 26.782 Tool, implement, cutlery, spoon & f 1.0 67,934.97 400,753.2 0.013 Lac; gums, resins & other vegetable 1.0 6,742.248 26,166.86 5.684 Nuclear reactors, boilers, mchy & m 1.0 1,648,613 13,184,757 0.04 Dairy prod; birds’ eggs; natural ho 0.9 108,823 298,934.1 20.<strong>38</strong>7 Vehicles o/t railw/tramw roll-stock 0.9 1,249,615 15,159,873 12.114 Vegetable plaiting materials; veget 0.9 954.308 10,156.8 1.369 Ceramic products 0.9 77,499.05 235,396.6 17.549 Printed books, newspapers, pictures 0.9 210,342 475,518.5 0.083 Miscellaneous articles of base meta 0.9 63,880.57 232,543.2 0.02 Meat and edible meat offal 0.9 35,894.88 474,550 17.375 Nickel and articles thereof 0.9 603.215 873,281.9 13.827 Mineral fuels, oils & product of th 0.9 2,071,307 19,939,099 3.645 Cork and articles of cork. 0.9 516.051 10,535.82 0.064 Footwear, gaiters and the like; par 0.8 62,279.35 214,504.2 36.367 Prepr feathers & down; arti flower; 0.8 1,646.785 14,785.46 20.095 Toys, games & sports requisites; pa 0.8 20,948.44 108,721.8 0.011 Prod.mill.indust; malt; starches; 0.8 107,974.9 213,150.7 5.862 Art of apparel & clothing access, n 0.8 76,474.77 1,121,149 17.337 Photographic or cinematographic goo 0.8 27,023.47 83,336.19 5.470 Glass and glassware 0.8 63,609.44 476,085.4 5.685 Electrical mchy equip parts thereof 0.8 84,1007.9 4,666,057 0.025 Salt; sulphur; earth & ston; plaste 0.7 145,555.9 232,0052 1.090 Optical, photo, cine, meas, checkin 0.7 16,0012.1 881,797.6 0.492 Musical instruments; parts and acce 0.7 1,781.843 6,948.057 0.024 Tobacco and manufactured tobacco su 0.6 39,436.1 303,952.6 27.966 Umbrellas, walking-sticks, seat-sti 0.6 25,64.619 48,408.82 22.646 Manufactures of straw, esparto/othe 0.6 637.769 11,647.47 8.015 Animal/veg fats & oils & their clea 0.6 232,296.8 40,6851.4 7.410 Cereals 0.6 436,922.6 798,312.7 0.571 Natural/cultured pearls, prec stone 0.6 41,<strong>38</strong>7.39 57,792,184 8.656 Wadding, felt & nonwoven; yarns; tw 0.5 14,142.27 114,721.9 15.551 Wool, fine/coarse animal hair, hors 0.5 19,294.59 1,128,<strong>38</strong>2 4.330 Pharmaceutical products. 0.5 195,746.6 664,763.4 0.689 Ships, boats and floating structure 0.5 5,505.175 426,697.5 6.196 Miscellaneous manufactured articles 0.5 28,082.93 99,257.82 0.06 Live tree & other plant; bulb, root 0.5 4,366.878 309,982.6 8.39 Coffee, tea, matn and spices 0.5 28,652.75 35,8491.4 3.152 Cotton 0.5 141,274.8 517,699.2 0.088 Aircraft, spacecraft, and parts the 0.4 98,277.52 419,055.1 2.642 Articles of leather; saddlery/harne 0.4 9,798.<strong>38</strong>4 153,665 25.255 Man-made staple fibres 0.4 61,846.65 233,291 3.054 Man-made filaments 0.4 31,553.97 5,181,86.9 18.2
97 Works of art, collectors’ pieces an 0.4 941.009 102,661.6 11.863 Other made up textile articles; set 0.3 56,526.81 227,155.3 37.041 Raw hides and skins (other than fu 0.3 1,526.193 1,243,633 3.953 Other vegetable textile fibres; pap 0.3 1,437.534 20,951.25 8.13 Fish & crustacean, mollusc & other 0.3 29,518.73 2,763,161 2.391 Clocks and watches and parts thereo 0.3 5,055.1<strong>38</strong> 17,985.84 12.058 Special woven fab; tufted tex fab; 0.2 11,068.03 111,714.9 14.660 Knitted or crocheted fabrics 0.1 8,241.599 54,753.06 17.150 Silk 0.1 558.958 1,667.674 7.793 Arms and ammunition; parts and acc 0.1 2,432.145 242,993.9 2.2Source: TIPS and COMESA database and own calculationsTABLE 18: REVEALED TRADE BARRIERS FOR COMESA EXPORTS TO SOUTH AFRICA (1999-2004), US$’000Product Product name RTB indexCOMESA’s exports toSouth AfricaCOMESA’s exports tothe WorldUnweightedaveragetariff (%)24 Tobacco and manufactured tobacco su 29.7 201,261 5,631,553 27.952 Cotton. 27.8 425,6<strong>38</strong> 4,253,364 0.026 Ores, slag and ash 25.5 425,427 1,530,742 0.074 Copper and articles thereof 23.7 148,322 2,427,395 3.29 Coffee, tea, matn and spices 17.5 102,768 7,490,742 3.112 Oil seed, oleagi fruits; miscell gr 13.2 54,411 1,575,835 6.86 Live tree & other plant; bulb, root 10.3 5,503 2,119,035 8.379 Zinc and articles thereof 9.5 9,366 15,093 0.017 Sugars and sugar confectionery 8.0 35,864 3,522,583 10.562 Art of apparel & clothing access, n 7.8 98,211 6,526,971 17.361 Art of apparel & clothing access, 7.6 60,917 7,631,967 11.775 Nickel and articles thereof 6.3 72,903 791,124 13.844 Wood and articles of wood; wood ch 6.3 101,531 539,509 10.77 Edible vegetables and certain roots 4.7 16,091 2,521,153 10.663 Other made up textile articles; set 4.5 21,759 888,425 37.08 Edible fruit and nuts; peel of citr 4.2 11,059 1,671,664 7.425 Salt; sulphur; earth & ston; plaste 4.2 32,832 2,262,840 1.014 Vegetable plaiting materials; veget 3.8 1,057 67,067 1.368 Art of stone, plaster, cement, asbe 3.0 22,<strong>38</strong>0 174,471 25.089 Ships, boats and floating structure 2.9 9,769 95,580 6.153 Other vegetable textile fibres; pap 2.8 2,970 248,791 8.116 Prep of meat, fish or crustaceans, 2.8 9,868 1,728,429 10.480 Tin and articles thereof 2.8 5,469 7,797 5.794 Furniture; bedding, mattress, matt 2.7 47,556 580,006 0.423 Residues & waste from the food indu 2.6 33,188 255,850 4.01 Live animals 2.4 1,701 536,727 0.081 Other base metals; cermets; article 2.3 6,067 1,197,898 1.82 Meat and edible meat offal 2.2 16,702 297,<strong>38</strong>7 17.341 Raw hides and skins (other than fu 2.1 19,825 1,441,066 3.942 Articles of leather; saddlery/harne 2.0 12,363 72,081 25.227<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> Liberalisation28<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>72 Iron and steel 1.9 54,086 4,135,736 7.746 Manufactures of straw, esparto/othe 1.8 743 58,<strong>38</strong>6 8.086 Railw/tramw locom, rolling-stock & 1.7 3,006 9,473 9.897 Works of art, collectors’ pieces an 1.7 2,467 192,825 11.876 Aluminium and articles thereof 1.6 15,424 1,313,794 6.720 Prep of vegetable, fruit, nuts or o 1.6 4,705 792,949 16.319 Prep.of cereal, flour, starch/milk; 1.4 4,979 79,207 18.173 Articles of iron or steel 1.3 49,698 474,236 4.457 Carpets and other textile floor co 1.1 2,725 759,168 16.951 Wool, fine/coarse animal hair, hors 1.0 1,959 47,689 4.331 Fertilisers 0.9 8,052 1,501,126 0.070 Glass and glassware 0.8 9,245 276,482 5.627 Mineral fuels, oils & product of th 0.8 427,078 - 3.678 Lead and articles thereof 0.8 872 21,488 5.164 Footwear, gaiters and the like; par 0.8 16,590 137,464 36.356 Wadding, felt & nonwoven; yarns; tw 0.7 2,481 77,891 15.543 Furskins and artificial fur; manuf 0.7 11 2,842 9.434 Soap, organic surface-active agents 0.6 5,512 278,016 13.54 Dairy prod; birds’ eggs; natural ho 0.6 2,403 152,763 20.321 Miscellaneous edible preparations 0.6 3,914 792,884 12.43 Fish & crustacean, mollusc & other 0.5 1,981 2,955,055 2.349 Printed books, newspapers, pictures 0.4 6,235 182,604 0.048 Paper & paperboard; art of paper pu 0.4 16,401 464,401 16.782 Tool, implement, cutlery, spoon & f 0.4 6,232 95,500 0.060 Knitted or crocheted fabrics 0.4 2,308 50,420 17.110 Cereals 0.4 9,884 853,286 0.569 Ceramic products 0.4 7,414 351,487 17.513 Lac; gums, resins & other vegetable 0.3 584 396,897 5.688 Aircraft, spacecraft, and parts the 0.3 <strong>38</strong>,753 322,336 2.640 Rubber and articles thereof 0.3 12,359 234,615 9.211 Prod.mill.indust; malt; starches; 0.3 816 153,151 5.896 Miscellaneous manufactured articles 0.3 2,217 111,847 0.028 Inorgn chem; compds of prec mtl, r 0.3 14,344 981,048 0.936 Explosives; pyrotechnic prod; match 0.2 373 26,645 3.171 Natural/cultured pearls, prec stone 0.2 9,678 7,989,170 8.685 Electrical mchy equip parts thereof 0.2 60,121 1,048,200 0.084 Nuclear reactors, boilers, mchy & m 0.2 77,720 1,789,093 0.029 Organic chemicals 0.2 9,714 1,472,409 1.530 Pharmaceutical products 0.1 10,223 467,530 0.658 Special woven fab; tufted tex fab; 0.1 330 25,274 14.622 Beverages, spirits and vinegar. 0.1 1,817 142,810 17.055 Man-made staple fibres 0.1 1,585 170,548 3.047 Pulp of wood/of other fibrous cellu 0.1 4<strong>38</strong> 159,928 0.015 Animal/veg fats & oils & their clea 0.1 3,153 240,097 7.490 Optical, photo, cine, meas, checkin 0.1 13,012 482,908 0.491 Clocks and watches and parts thereo 0.1 476 161,221 12.0
33 Essential oils & resinoids; perf, 0.1 1,990 521,126 9.35 Products of animal origin, nes or 0.1 328 156,963 0.083 Miscellaneous articles of base meta 0.1 917 54,<strong>38</strong>5 0.039 Plastics and articles thereof 0.1 8,929 1,129,308 8.287 Vehicles o/t railw/tramw roll-stock 0.1 16,995 354,663 12.132 Tanning/dyeing extract; tannins & 0.1 2,374 117,752 2.850 Silk 0.1 <strong>38</strong> 1,436 7.7<strong>38</strong> Miscellaneous chemical products 0.1 4,163 204,719 2.945 Cork and articles of cork 0.1 139 910 0.065 Headgear and parts thereof 0.1 84 22,193 25.135 Albuminoidal subs; modified starche 0.1 468 22,134 2.818 Cocoa and cocoa preparations 0.1 191 99,448 9.354 Man-made filaments 0.0 811 157,737 18.292 Musical instruments; parts and acce 0.0 40 6,311 0.095 Toys, games & sports requisites; pa 0.0 581 120,144 0.059 Impregnated, coated, cover/laminate 0.0 134 51,197 26.793 Arms and ammunition; parts and acc 0.0 9 7,528 2.237 Photographic or cinematographic goo 0.0 44 13,377 5.467 Prepr feathers & down; arti flower; 0.0 2 3,528 20.066 Umbrellas, walking-sticks, seat-sti 0.0 1 2,823 22.6Source: TIPS and COMESA database and own calculations29<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Appendix 3TABLE 19: MARKET ACCESS GAINS FOR COMESA’S EXPORTS INTO SOUTH AFRICA (US$’000), 2004Product Product name MAG gains % of total MAG52 Live animals 43,097 35.627 Meat 18,756 15.524 Fish 10,418 8.662 Dairy prods; birds eggs; honey; 7,777 6.474 Products of animal origin 6,021 5.061 Live trees, plants, cut flowers 5,810 4.844 Edible vegetables 3,865 3.217 Edible fruit & nuts 2,782 2.394 Coffee, tea, mate & spices 2,421 2.085 Cereals 1,676 1.473 Milling products; malt; starch; inulin; wht gluten 1,468 1.29 Oil seeds etc 1,406 1.212 Gums & resins 1,394 1.242 Vegetable plaiting materials 1,160 1.072 Animal or vegetable fats 985 0.864 Edible preps of meat, fish 905 0.716 Sugars and sugar confectionary 800 0.784 Cocoa and cocoa preparations 742 0.67 Prep cereal, flour 717 0.640 Prep vegetables, fruit, nuts 700 0.6Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> Liberalisation30<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>87 Misc. edible preparations 579 0.557 Beverages & spirits 564 0.548 Food industry residues & waste; prep animal feed 533 0.439 Tobacco 487 0.460 Salt; sulphur; stone; & cement 456 0.419 Ores, slag and ash 440 0.420 Mineral fuels & oils 417 0.323 Inorganic chemicals <strong>38</strong>1 0.368 Organic chemicals 348 0.376 Pharmaceutical products 344 0.<strong>38</strong> Fertilizers 327 0.341 Tanning & dye extracts 293 0.270 Essential oils etc; perfumery, cosmetic etc preps 285 0.234 Soap and waxes 272 0.263 Albuminoidal subst; starch; glue; enzymes 267 0.24 Explosives 258 0.26 Photographic or cinematographic goods 175 0.169 Misc. chemical products 174 0.171 Plastics and articles thereof 157 0.121 Rubber and articles thereof 146 0.156 Raw hides, skins and leather 121 0.115 Leather art; saddlery 111 0.133 Wood and articles of wood 103 0.196 Basketware & wickerwork 99 0.125 Wood pulp etc 84 0.151 Paper & paperboard 58 0.0<strong>38</strong> Printed books, newspapers 57 0.055 Silk 53 0.022 Wool & animal hair 47 0.053 Cotton 46 0.054 Veg text fib 43 0.028 Manmade filaments 42 0.049 Manmade staple fibres 34 0.030 Wadding, felt etc; sp yarn; twine, ropes etc. 33 0.032 Carpets 33 0.029 Spec wov fabrics 32 0.089 Impregnated etc text fabrics 30 0.046 Knitted or crocheted fabrics 30 0.083 Apparel articles and accessories 29 0.03 Apparel articles and accessories 28 0.02 Textile articles NESOI 27 0.013 Footwear 24 0.082 Headgear and parts thereof 23 0.010 Umbrellas, walking-sticks, riding-crops 22 0.090 Prep feathers, down etc 20 0.058 Art of stone, plaster, cement, asbestos 12 0.0
18 Ceramic products 7 0.036 Glass and glassware 5 0.014 Precious metals 4 0.059 Iron and steel 4 0.095 Articles of iron or steel 4 0.035 Copper and articles thereof 3 0.086 Aluminium 2 0.011 Lead 1 0.065 Zinc 1 0.066 Tin 0 0.043 Tools & cutlery 0 0.01 Misc. articles of base metal 0 0.05 Machinery 0 0.026 Electric machinery 0 0.031 Railway or tramway stock 0 0.037 Vehicles 0 0.047 Aircraft & spacecraft 0 0.050 Ships, boats and floating structures 0 0.067 Medical and scientific equip. 0 0.078 Clocks and watches 0 0.079 Musical instruments 0 0.080 Arms and ammunition 0 0.088 Furniture & bedding 0 0.091 Toys, games & sport equipment 0 0.092 Misc. manufactured articles 0 0.093 Art & antiques 0 0.097 Nickel and articles thereof 0 0.075 Base metals 0 0.081 Furskins and artificial fur 0 0.045 Cork and articles of cork 0 0.0121.076 100.0Source: TIPS and COMESA database and own calculations31<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>TABLE 20: MARKET ACCESS GAINS FOR SOUTH AFRICA’S EXPORTS INTO COMESA (US$’000), 2004Product Product name MAG low MAG high % MAG low % MAG high27 Mineral fuels, oils & product of th 67,341 98,735 11.2 11.372 Iron and steel 59,714 85,815 10.0 9.839 Plastics and articles thereof 45,241 65,727 7.5 7.573 Articles of iron or steel 28,163 41,081 4.7 4.787 Vehicles o/t railw/tramw roll-stock 27,957 42,729 4.7 4.948 Paper & paperboard; art of paper pu 27,633 40,063 4.6 4.685 Electrical mchy equip parts thereof 24,274 37,168 4.0 4.2<strong>38</strong> Miscellaneous chemical products 17,667 25,506 2.9 2.940 Rubber and articles thereof 15,983 23,476 2.7 2.715 Animal/veg fats & oils & their clea 15,180 22,046 2.5 2.530 Pharmaceutical products 13,799 19,863 2.3 2.3Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> Liberalisation32<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>17 Sugars and sugar confectionery 13,009 18,862 2.2 2.281 Other base metals; cermets; article 12,746 18,926 2.1 2.231 Fertilisers 12,451 17,979 2.1 2.149 Printed books, newspapers, pictures 12,258 17,512 2.0 2.074 Copper and articles thereof 10,484 15,175 1.7 1.728 Inorgn chem; compds of prec mtl, r 10,417 14,947 1.7 1.736 Explosives; pyrotechnic prod; match 9,7<strong>38</strong> 13,911 1.6 1.610 Cereals 8,349 13,120 1.4 1.594 Furniture; bedding, mattress, matt 8,057 11,588 1.3 1.376 Aluminium and articles thereof 7,646 11,184 1.3 1.321 Miscellaneous edible preparations. 7,632 10,902 1.3 1.219 Prep.of cereal, flour, starch/milk; 7,319 10,476 1.2 1.229 Organic chemicals. 7,175 10,295 1.2 1.232 Tanning/dyeing extract; tannins & 6,893 9,997 1.1 1.133 Essential oils & resinoids; perf, 6,630 9,594 1.1 1.122 Beverages, spirits and vinegar. 6,209 8,909 1.0 1.024 Tobacco and manufactured tobacco su 6,204 9,171 1.0 1.020 Prep of vegetable, fruit, nuts or o 5,962 8,517 1.0 1.069 Ceramic products 5,812 8,304 1.0 0.934 Soap, organic surface-active agents 5,673 8,1<strong>38</strong> 0.9 0.982 Tool, implement, cutlery, spoon & f 5,105 7,328 0.9 0.852 Cotton. 4,786 6,892 0.8 0.883 Miscellaneous articles of base meta 4,7<strong>38</strong> 6,817 0.8 0.884 Nuclear reactors, boilers, mchy & m 4,040 6,443 0.7 0.770 Glass and glassware. 3,998 5,815 0.7 0.744 Wood and articles of wood; wood ch 3,965 5,814 0.7 0.790 Optical, photo, cine, meas, checkin 3,951 6,065 0.7 0.763 Other made up textile articles; set 3,922 5,620 0.7 0.611 Prod.mill.indust; malt; starches; 3,562 5,111 0.6 0.659 Impregnated, coated, cover/laminate 3,547 5,142 0.6 0.688 Aircraft, spacecraft, and parts the 3,427 5,156 0.6 0.662 Art of apparel & clothing access, n 3,278 4,686 0.5 0.568 Art of stone, plaster, cement, asbe 2,892 4,145 0.5 0.54 Dairy prod; birds’ eggs; natural ho 2,615 3,847 0.4 0.464 Footwear, gaiters and the like; par 2,358 3,<strong>38</strong>4 0.4 0.437 Photographic or cinematographic goo 2,256 3,222 0.4 0.479 Zinc and articles thereof 2,177 3,171 0.4 0.416 Prep of meat, fish or crustaceans, 1,767 2,528 0.3 0.355 Man-made staple fibres 1,756 2,530 0.3 0.371 Natural/cultured pearls, prec stone 1,595 2,347 0.3 0.396 Miscellaneous manufactured articles 1,586 2,289 0.3 0.39 Coffee, tea, matn and spices 1,585 2,272 0.3 0.361 Art of apparel & clothing access, 1,560 2,230 0.3 0.335 Albuminoidal subs; modified starche 1,526 2,214 0.3 0.318 Cocoa and cocoa preparations 1,399 2,0<strong>38</strong> 0.2 0.295 Toys, games & sports requisites; pa 1,312 1,889 0.2 0.2
56 Wadding, felt & nonwoven; yarns; tw 1,034 1,495 0.2 0.223 Residues & waste from the food indu 978 1,443 0.2 0.254 Man-made filaments 967 1,395 0.2 0.27 Edible vegetables and certain roots 928 1,372 0.2 0.225 Salt; sulphur; earth & ston; plaste 883 1,334 0.1 0.257 Carpets and other textile floor co 864 1,234 0.1 0.178 Lead and articles thereof. 721 1,050 0.1 0.147 Pulp of wood/of other fibrous cellu 688 994 0.1 0.142 Articles of leather; saddlery/harne 544 777 0.1 0.158 Special woven fab; tufted tex fab; 543 785 0.1 0.151 Wool, fine/coarse animal hair, hors 453 663 0.1 0.13 Fish & crustacean, mollusc & other 396 601 0.1 0.186 Railw/tramw locom, rolling-stock & 379 560 0.1 0.165 Headgear and parts thereof 313 450 0.1 0.191 Clocks and watches and parts thereo 247 359 0.0 0.093 Arms and ammunition; parts and acc 162 234 0.0 0.060 Knitted or crocheted fabrics 144 207 0.0 0.06 Live tree & other plant; bulb, root 117 177 0.0 0.012 Oil seed, oleagi fruits; miscell gr 98 147 0.0 0.066 Umbrellas, walking-sticks, seat-sti 98 141 0.0 0.080 Tin and articles thereof 91 133 0.0 0.092 Musical instruments; parts and acce 83 120 0.0 0.053 Other vegetable textile fibres; pap 77 113 0.0 0.041 Raw hides and skins (other than fu 74 108 0.0 0.02 Meat and edible meat offal 58 87 0.0 0.075 Nickel and articles thereof 54 79 0.0 0.089 Ships, boats and floating structure 51 78 0.0 0.097 Works of art, collectors’ pieces an 25 39 0.0 0.046 Manufactures of straw, esparto/othe 24 34 0.0 0.045 Cork and articles of cork 20 29 0.0 0.067 Prepr feathers & down; arti flower; 17 25 0.0 0.050 Silk 5 7 0.0 0.043 Furskins and artificial fur; manuf 1 2 0.0 0.01 Live animals 0 0 0.0 0.05 Products of animal origin, nes or 0 0 0.0 0.08 Edible fruit and nuts; peel of citr 0 0 0.0 0.013 Lac; gums, resins & other vegetable 0 0 0.0 0.014 Vegetable plaiting materials; veget 0 0 0.0 0.026 Ores, slag and ash 0 0 0.0 0.0599,455 874,984 100.0 100.0Source: TIPS and COMESA database and own calculations33<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Integrating Africa through a SACU and COMESA FTA – Speeding up the Regional Integration Process
<strong>Trade</strong> Liberalisation34<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>
tradeliberalisation35THE IMPACT OF THEASIAN DRIVERSON SUB-SAHARAN AFRICA’S WOODENFURNITURE INDUSTRY*BY RAPHAEL KAPLINSKYINSTITUTE OF DEVELOPMENT STUDIES AND CENTRE FOR RESEARCH ININNOVATION MANAGEMENT, UNIVERSITIES OF SUSSEX AND BRIGHTONAND MIKE MORRISSCHOOL OF DEVELOPMENT STUDIES, UNIVERSITY OF KWAZULU-NATALAND SCHOOL OF ECONOMICS, UNIVERSITY OF CAPE TOWN1. Introduction1.1 Why is furniture important?THE WOOD FURNITURE VALUE CHAIN, a large and rapidly globalising sector, has thepotential to play an important role in promoting growth and alleviating poverty insub-Saharan Africa (SSA). There is scope for labour-intensive production, but at thesame time, the ability to mechanise production and innovate new products lays outa path for upgrading and innovation over time. This not only supports growing producer incomes inthe furniture value chain, but also acts as an example to other sectors. It is also a resource-intensivesector, providing opportunities for many SSA countries where timber grows rapidly and cheaply,and helps to facilitate productive linkages between industry and agriculture and urban and ruralareas. Consequently, the furniture industry has been one of the primary stepping stones used byindustrialising economies in promoting growth and diversifying economic structures. It is potentiallya key industry in SSA’s future development.* This article is an excerpt from the report ‘Dangling by a thread: How sharp are the Chinese scissors?’, prepared for DFID <strong>Trade</strong>Division. The views and opinions expressed in the report are those of the authors and do not reflect any official DFID position.The wood furniture value chain, a large andrapidly globalising sector, has the potential toplay an important role in promoting growthand alleviating poverty in SSA. It is a a resourceintensivesector, providing opportunities forcountries where timber grows rapidly andcheaply, and helps to facilitate productivelinkages between industry and agriculture andurban and rural areas. However, the evidencepresented here on the SSA industry makes forblunt reading. Most of the continent’s furnitureindustry is inward oriented and involves theinformal sector using low-grade timber toproduce basic products for middle- and lowincomedomestic consumers. Moreover, Chinahas become a growing presence in globalfurniture markets, and SSA firms are unable tocompete effectively with it. As a consequence,SSA furniture exports have come to bethreatened by a combination of falling prices,reduced profitability and evaporating marketshares. Further, the liberalisation of imports intoSSA has meant growing domestic competitionfrom low-cost Asian-sourced furniture, whichalso threatens the survival of the formalfurniture-manufacturing sector in SSA.<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>
<strong>Trade</strong> LiberalisationFIGURE 1: WOOD FURNITURE VALUE SYSTEMInputsMachineryForestrySawmills36<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>DesignMachinery andother technologyDomesticwholesaleFurnituremanufacturersBuyersLogistics, qualityservicePaint,adhesivesForeignwholesaleDomesticretailForeignretailConsumersRecyclingWaste1.2 The wood furniture value chainFigure 1 maps the broad features of the wood furniture value chain. Rawmaterials such as seed inputs, chemicals, equipment and water feed intothe forestry sector. Wood flows to the sawmills, and sawn timber andintermediate wood products move to the furniture manufacturers who,in turn, obtain inputs from the machinery, textiles, plastics, adhesives andpaint industries.The furniture industry also draws on logistical, design and branding skillsfrom the service sector. Depending on which market is served, the furniturepasses through various intermediary buying stages until it reaches thefinal customer. (The buying function is represented by a dotted box in thisfigure to emphasise that several organisation types, including wholesalers,retailers and independent buyers, can manage this function). Finally,customers will either recycle or dispose of the furniture.1.3 Factors contributing to theglobalisation of the wood furniturevalue chainThe wood furniture value chain is an increasingly global industry. Seenfrom the buyer-end of this chain, four distinct globalising strategies can beobserved. The first of these are firms who have little or no involvement inproduction or in the organisation and co-ordination of global productionnetworks. They are satisfied to buy furniture either directly from producersin arm’s-length relationships or to work through specialised buying firms.They comprise independent furniture stores, many of which are small ormedium-sized and which serve local sub-national markets. The secondcategory are the much larger national or international firms which makeextensive use of marketing and brand names, and purchase directly fromsuppliers, often providing assistance with upgrading and the sourcing ofinputs. Examples of these transnational corporations (TNCs) are IKEA andB&Q. Third are manufacturing firms in the importing country who buy in
semi-finished components. This has become, for example, a major elementin the burgeoning furniture trade in rubber-wood products betweenThailand and Japan (Mitsuhashi, 2006) and is referred to as ‘productionsharing’. Finally, there are the manufacturing firms in the importing countrieswhich have established subsidiaries in low-income economies, suchas Steinhoff of Germany expanding production in Poland, Ukraine andSouth Africa.Underlying this fragmenting global division of labour in the furniture valuechain has been a series of changes in technology and organisation. Themajor components of these technological advances reflect the transfer ofpractices from other industries and include: 1 Computer-numerically-controlled (CNC) wood working machinery, which enhancesproductivity, reduces waste, improves time-to-market and facilitates modular productionof non-standardised items. Computer-aided design and manufacturing (CAD/CAM) allows designs to be fedfrom designers to manufacturing firms (anywhere in the world), with significantimprovements in quality and productivity. The introduction of flat-pack or RTA (ready-to-assemble) furniture led to an importantchange in furniture production methods. RTA-designed furniture, with standardshapes and sizes and high volume demand, led factories to take advantage ofdesign for manufacturing processes. It also dramatically cut the transport costs ofshipping bulky products. The development of flat-pack furniture was critically dependent on advances inmaterial technology, such as MDF (multi-density fibre) which, in addition to usingoff-cuts and ‘waste’, makes for the optimal use of forestry products. Flexible manufacturing systems (FMS) and cellular plant layout improve the efficientflow of furniture parts through the plant, enhancing flexibility and quality, and reducinginventories and costs. Made-to order and just-in-time distribution systems reduce inventory levels of rawmaterial inputs and finished-but-not-sold items.These technological and organisational innovations have allowed for thegrowing globalisation of this sector. It has enabled producers in highwageeconomies to reduce their costs significantly, such that the share ofwage in company sales in European manufacturing firms fell from around50% in the 1960s to 28% in the mid-1990s (European Commission,1997). But simultaneously, it has also provided a role for low-income andresource-intensive economies to become increasingly active participantsin the global chain.Another factor facilitating the globalisation of this chain has been thegrowth of concentrated buying power in final markets. This is part of amuch larger phenomenon, spanning many sectors of final consumption,including the clothing sector (Feenstra and Hamilton, 2005; Kaplinsky,2005). For example, in Germany, although there are more than 15,000furniture stores employing more than 110,000 people, the buying groups(Einkaufsverbände) and their affiliates control 60% of the market. Themajority of retailers and manufacturers are connected with these groups.Similar buyer concentration occur in many other countries, including theUK where retail multiples control 40% of the furniture market.1.4 Global trade in furniture intensifiesThese evolving corporate strategies, changes in technology and increasinglyconcentrated buying structures have led to the rapid globalisation ofthe wood furniture value chain. At the three-digit SITC 2 level, in 2002 thefurniture industry was the 16th largest out of 141 traded manufacturingproduct groups (SITC 5 to 8, excluding SITC 68), with a total value ofglobal trade of manufactures of US$66.1bn (www.unctad.org, last accessedon 20 June 2005). It was the largest traditional, low-technologytraded sector in 2002, exceeding the value of trade in the footwear industry(US$43.6bn) and the toys and sporting goods industry (US$49.8bn).The growth of world trade between 1993 and 2002 in the furniture sector(94%) exceeded that for all world trade of manufactures (75%), as well asthat of toys and sporting goods (52%) and footwear (26%).Although furniture is a resource- and labour-intensive product, it is strikingthat the major exporting countries are the industrially advanced economies.As can be seen from Table 1, of the 20 largest exporters, only five(China, Mexico, Indonesia, Malaysia and Thailand) are low-wage economiesand two (Poland and the Czech Republic) are transitional economies.However, given that emerging, transitional and developing countries tendto be small importers of furniture, their participation in the group of thelargest net exporters is much more significant – only five advanced industrialcountries (Italy, Canada, Denmark, Spain and Sweden) of the top 20gross exporting countries are also positive net exporters. Italy is by somedistance the largest gross and net exporter of furniture.Table 1 also shows the very rapid growth of furniture exports by lowincomeeconomies. Between 1993 and 2002, China’s exports increasedby 6.2 times, Mexico by a factor of 5.3, Poland by 5, Malaysia by 2.6 andIndonesia by 2.3. Although to some extent these very high growth ratesare a result of low export volumes in 1993, they highlight that these countrieshave become leading global exporters of furniture. It is also strikingthat, other than Canada, all of the major dynamic furniture exporters arelow-income or transitional economies.37<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>1Company interviews and industry reports. See Spalding, 2001 and European Commission,1997.2Standard International <strong>Trade</strong> ClassificationThe Impact of the Asian Drivers on Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>
<strong>Trade</strong> LiberalisationTABLE 1: VALUE OF GLOBAL FURNITURE TRADE (SITC 821): THE LEADING 20 EXPORTING COUNTRIES, 1993 AND 2002 (US$’000)Gross exports, 1993 Gross exports, 2002 % change Net exports, 2002<strong>38</strong><strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Italy 5,797,350 8,885,963 53 7,792,247China 1,080,566 6,696,259 520 6,377,748Germany 3,758,009 5,<strong>38</strong>4,098 43 -1,342,364Canada 1,672,760 4,795,399 187 1,694,281US 3,027,889 4,703,943 55 -18,889,234Mexico 649,941 3,409,<strong>38</strong>1 425 2,267,533Poland 580,978 2,893,334 398 2,452,535France 1,648,592 2,355,577 43 -1,706,937Denmark 1,589,971 2,057,728 29 1,323,356Bel.-Lux. 1,409,314 1,817,374 29 -361,084Indonesia 675,588 1,519,572 125 1,500,561Spain 548,801 1,502,148 174 313,670Malaysia 564,853 1,489,940 164 1,310,355UK 904,498 1,350,189 49 -3,322,<strong>38</strong>9Austria 650,646 1,278,178 96 -152,753Sweden 825,930 1,269,427 54 154,205China, (Taiwan) 1,833,173 1,198,911 -35 952,679China, (Hong Kong) 563,632 1,141,647 103 -159,480Czech Republic 191,733 1,102,751 475 686,169Thailand 593,849 914,334 54 840,290Other countries 5,447,541 10,240,568 88 -6,469,599World 34,041,110 66,086,250 94Source: Calculated from UNCTAD, 2005 (www.unctad.org, last accessed on 21 June 2005)FIGURE 2: SHARE OF GLOBAL FURNITURE IMPORTS (SITC 821), 1993-2002Source: Calculated from UNCTAD, 2005 (www.unctad.org, last accessed on 6 July 2005)
TABLE 2: LEADING 20 PRODUCERS OF WOOD-BASED PANELS BY VALUE AND MARKET SHARE, 1993 AND 2002Value (US$’000) Market share (%)1993 2002 % change 1993 2002 % changeCanada 909,457 2,404,937 164 7 14 7Germany 675,461 2,010,540 198 5 11 6Indonesia 4,311,515 1,818,219 -58 33 10 -22Malaysia 1,354,653 1,404,271 4 10 8 -2US 921,886 927,617 1 7 5 -2China 402,129 873,705 117 3 5 2Austria 370,789 801,994 116 3 5 2Bel.- Lux. 492,481 791,527 61 4 4 1France 434,985 680,916 57 3 4 1Finland 441,835 621,260 41 3 4 0Brazil 433,823 501,601 16 3 3 0Italy 308,599 450,126 46 2 3 0Poland 55,739 392,700 605 0 2 2Spain 193,574 371,465 92 1 2 1Russian Federation 180,270 335,352 86 1 2 1New Zealand 160,615 280,320 75 1 2 0Switzerland 163,750 210,522 29 1 1 0Ireland 57,823 165,759 187 0 1 0Chile 61,429 164,886 168 0 1 0Portugal 122,653 158,110 29 1 1 0Other countries 1,172,470 2,348,463 100 9 13 4World 13,225,936 17,714,290 34Source: Calculated from FAOSTAT data, 2005 (www.fao.org, last accessed on 21 June 2005)39<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Furniture imports have increased in most high-income countries, with thevalue of global trade almost doubling between 1993 and 2002 (see Table1). As we illustrate below, prices during this period fell sharply (Table 3and Figure 2). This pervasive fall in global prices of furniture has clearlybenefited consumers in trade-liberalising economies who have access tocheaper, better quality and more varied products. Consumers in the UShave particularly taken advantage of low-priced imports: the share of theUS in global imports rose from 21% to 33% between 1993 and 2002,whilst that of the EU fell from 46% to 37% in the same period (see Figure3). 3Intermediate products such as wood-based panels consisting of veneersheets, plywood, particle board, hardboard, medium-density fibreboard(MDF), compressed fibreboard and insulating board also experienced rapidgrowth during this period. Global exports of these wood-based panelswere US$17.7bn in 2002, an increase of 34% since 1993 (see Table 2).3EU imports include EU-extra and EU-intra trade.Although some low-income economies have seen export growth (notablyPoland and China), the stellar performers tended to be high-incomeeconomies (Canada, Germany and Austria). In some cases, low-incomeeconomies have seen a fall or stagnation in the value of export of theseintermediate products; for example, Indonesia and Malaysia saw theirmarket share decline by 22% and 2%, respectively, while Brazil’s shareremained unchanged during the 1993-2002 period.1.5 Price trends in the global woodfurniture value chainIn the measurement of unit price performance in the furniture industry, wedraw on the data for EU imports because the EU is the major destinationfor extra-continental furniture exports from SSA. Since the EU accountsfor more than one-third of global imports, it is a reasonable surrogate forglobal price performance and for assessing the future export prospects ofthe SSA furniture industry.The Impact of the Asian Drivers on Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>
<strong>Trade</strong> LiberalisationTABLE 3: AVERAGE UNIT PRICES (TWO-YEAR MOVING AVERAGES) AND THE NUMBER OF COUNTRIES HOLDING 1% OF MARKET SHARE OFWOOD FURNITURE IMPORTS TO THE EU, 1989 AND 200140<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Average unit price(US$1,000/metricton)Averageunit price (%change)Unit price standarddeviationTotal no. of countryexportersNo. of low-wagecountry exporters1989 2001 1989-01 1989 2001 1989 2001 1989 2001Kitchen furniture 3.63 2.51 -31 4.26 1.83 15 14 2 4Bedroom furniture 2.34 1.94 -17 2.36 1.74 18 25 6 11Upholstered seats with wooden frames 7.<strong>38</strong> 4.42 -40 4.03 3.16 19 26 6 12Seats with wooden frames 3.26 3.06 -6 2.77 4.44 24 31 10 18Wooden office desks 3.13 2.51 -20 4.23 2.16 19 19 5 6Wooden office furniture (=< 80 cm.) 4.41 2.68 -39 3.84 2.41 19 25 3 7Wooden office cupboards (>80 cm.) 4.09 3.09 -24 1.76 1.90 14 18 1 6Wooden office furniture(> 80 cm. ex. cupboards) 3.52 2.88 -18 2.48 2.50 17 20 2 4Wooden furniture (dining/ and living room) 3.26 2.07 -37 3.32 1.99 20 35 6 18Wooden furniture (shops) 5.31 4.73 -11 2.51 4.64 14 23 1 7Other wooden furniture 2.90 2.19 -25 2.47 2.44 23 31 8 16All wooden furniture (aggregate) 2.72 2.17 -36 28 48 11 28Source: Calculated from Eurostat COMEXT databaseOver the 1990s, there were a number of significant trends in the unitprices of furniture imported into the EU (see Figure 3 and Table 3) 4 :For the industry as a whole, there was a pervasive trend towards a decline in unitprices of 36% between 1989 and 2001.The growing globalisation of the furniture industry meant that there was a tendencyfor a ‘world price’ to emerge, that is, for a growing convergence in the price ofproducts originating from different types of economies. (In eight of the 11 productcategories, the standard deviation of prices – measured as average unit prices ofdifferent exporting economies – fell between 1989 and 2001,There were sub-sector variations: the largest fall in unit price was for upholsteredwooden seats, wooden office furniture (= < 80 cm. in height) and wooden diningand living room furniture (40%, 39% and 37%, respectively).There was a growing dispersion in global sourcing. The number of countries with atleast 1% market share of one of the 11 imported product markets increased from28 in 1989 to 48 in 2001.The total number of low-wage countries with at least 1% market share in one ofthe 11 imported product markets more than doubled during this period, from 11countries in 1989 to 28 countries in 2001.1.6 The demand for timber and thegrowth of environmental concernsBoth for reasons of maintaining biodiversity and because of concerns withcarbon depletion, the global wood furniture industry has increasingly had4Unit prices are calculated as a two-year moving average to even out possible currency fluctuations.This means that ‘1989’ is an average of 1988 and 1989 unit prices and ‘2001’ isan average of 2000 and 2001 unit prices.to deal with the problems of the regulation of logging. The primary concernsrelate to the depletion of hardwoods, and this has led to two majorresponses. The first has been the growth of certification schemes such asthe Forestry Stewardship Council (FSC), which regulates not only the needto replant cut timber, but also the procedures used in logging and therights of indigenous peoples. The second response to these environmentalconcerns has been the search for substitutes for hardwoods, such as rubberwood(predominantly from Thailand and Malaysia) (Mitsuhashi, 2006)and eucalyptus/saligna from Brazil and South Africa (Kaplinsky, Morrisand Readman, 2001).As a consequence of these developments, there have been significantchanges in the type of timber used in the global furniture industry 5 : Hardwood production is shifting from Asia to Latin America, with Asia’s share forecastedto fall from 60% to 10% in 2000. In terms of volume, the tropical timbers most in demand are lighter hardwoodssuch as lauan, meranti and seraya from Indonesia, Malaysia and the Philippines,and increasingly rubberwood and saligna. Over the last decade, rubber wood, a tropical timber, has gained strong markets asa material for furniture production and saligna is also gaining ground rapidly. There has been an Increasing use of logs of smaller dimensions (rubberwood, melina,acacia, saligna, teak and pine) from fast-growing plantations. As teak obtained from natural stands have become more scarce in Myanmar, thekey Asian producer country, more producers are turning to plantation-grown teak,if end-use specifications allow it.5International <strong>Trade</strong> Forum – Issue 2/2001.
FIGURE 3: SHARE OF GLOBAL FURNITURE IMPORTS (SITC 821), 1989-2001Average unit priceEU imports – seatsAverage unit priceEU imports – office furniture$1 000/metric ton (TYMA)Upholstered seats with wooden framesSeats with wooden frames$1 000/metric ton (TYMA)Wooden office desksWooden office furniture < 80cm heightWooden office cupboards > 80cm heightWooden office furniture > 80 cm height41<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Average unit priceEU imports – kitchen and bedroom furnitureAverage unit priceEU imports – dining, livingroom, shop furniture$1 000/metric ton (TYMA)$1 000/metric ton (TYMA)Kitchen furnitureBedroom furnitureWooden furniture for shopsWooden furniture for dinning and living roomsOther wooden furnitureSource: Calculated from UEurostat COMEXT data base New bio-composite boards – extracted from oil palm residues, coconut shells orflattened bamboo – are being developed to help overcome the raw material shortages.These developments illustrate an industry in the process of global reconfiguration– the rapid growth of exports of labour-intensive final products,often incorporating capital- and technology-intensive intermediatesimported from high-income economies. It is also an industry that is experiencingintense competition, with new entrants and escalating pricepressure, and subject to growing concerns and regulation on the environmentalimplications of logging and deforestation. It is further an industrywhere China is showing a growing presence, both with respect to finalproducts and in relation to intermediate products.2. The Chinese and SSAfurniture industriesData on the SSA furniture industry are very poor. The overwhelming bulkof the continent’s production occurs in informal sector workshops, usinga combination of locally-sourced timber, offcuts and timber suppliesfrom the formal sector, which are largely cut to meet the needs of theconstruction industry. These workshops target the domestic market formiddle- and especially lower-middle and low-income consumers. Designsare unsophisticated, but often incorporate ingenious use of local materials(such as tyre-offcuts for springs). With the exception of a very small craftsegment, this informal sector has no presence in export markets.The Impact of the Asian Drivers on Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>
<strong>Trade</strong> LiberalisationTABLE 4: TARIFFS ON FURNITURE IN MAJOR IMPORTING REGIONS, 2005HS Number US EU Japan Canada Australia9401 Logs 0 0% - 5.6% 0% - 3.8% 0% - 15.5% 0% - 10%9403 Timber 0 0% - 5.6% 0 0% - 9.5% 5%9404 Furniture 0% - 12.8% 3.7% 3.2% - 3.8% 8% - 15.5% 0% - 7.5%Source: World Tariff Online (www.tdctrade.com/main/industries/t2_2_15.htm – accessed on 9 January 2006)42TABLE 5: CHINESE FURNITURE INDUSTRY (US$BN)<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>1995 2000 2002Compoundannual growth rate1995–2000 (%)Compoundannual growth rate2000–2002Furniture production 6.78 12.74 19.95 13 25Furniture exports 1.10 3.65 5.30 27 21Furniture imports 0.08 0.10 0.10 5 0Apparent domestic consumption of furniture 5.76 9.19 14.75 10 27Source: Elaboration of CSIL data (Anon, 2001) and CNFA data (Anon, 2003), cited in Robb and Bin Xie (2003)The major SSA furniture exporters are South Africa, the Ivory Coast (beforethe recent political turmoil) and Ghana. However, SSA’s main role in theglobal wood furniture value chain (as opposed to the furniture sector) isas a provider of timber to other producers, often at the cost of its ownfurniture sector (see below). In the context of the global economy, SSA’sfurniture industry can therefore be characterised as an industry with potential,rather than one with a discrete presence.Although no data are available, SSA’s share of global furniture exports isminiscule. South Africa, overwhelmingly the largest exporter, had woodenfurniture exports of less than US$60m in 2004 (NPI, 2005), and evenat its peak, Ghana’s furniture exports were less than US$14m in 2001.This compares with global exports of US$66bn in 2002, So, even withthe most generous assumption that total exports are in the region ofUS$100m, SSA’s share of global furniture exports were somewhat lessthan 1.5% in 2002. Since then, furniture exports from the largest exporter(South Africa) have fallen, and global trade continues to grow, so that thecurrent share of SSA in global markets is probably around 1% – muchlower than its share of global clothing and textiles exports. This is because,unlike the clothing sector where tariffs into the US and the EU are high(between 16.6% and 32.2%), in the furniture sector tariffs are relativelylow (see Table 4). The real measure of this disparity between clothing andtextiles and furniture is not to be found in the comparison of nominalrates of protection in the US and Europe. This is because the furnitureindustry uses domestic inputs, and there is little difference between thederived nominal and ‘effective rates of subsidy’ provided to SSA producersby these tariff preferences. By contrast, the clothing and textiles sectorsmake extensive use of imported inputs, so that nominal rates of subsidyare much higher than effective rates of subsidy. In other words, in thefurniture sector, SSA competes ineffectively on a much less uneven playingfield, which is not adequately sloped in its favour. 6Compare this SSA continental performance with the dynamism of China’sfurniture industry. The Chinese furniture industry consists of around50,000 companies and five million employees (CNFA, 2003). Most ofthese companies are small- to medium-sized operations with annual salesless than US$36m (or CNY ¥300M). 90% of the companies are privatelyowned stock-holding companies, as well as various joint ventures (Xu,2004a) 6 . Taiwanese ‘triangular manufacturers’ have helped to revolutioniselocal furniture-making sectors with state-of-the-art manufacturing facilitiesand modern business concepts (Wall Street Journal, 2004; Lei andMcgowin, 2002). It is estimated that Taiwanese companies are contributing75% of the furniture export shipments from China (CNFA 2003).Between 1995 and 2002, Chinese furniture production grew in valuefrom US$6.8bn to US$20bn (Table 5), a compound annual growth rateof 17%. In the same period, exports grew from US$1.1bn to US$5.3bn,a compound annual growth rate of 27%. The export ratio increased from16% to 26.5%. Imports of furniture were negligible and showed littlegrowth. Double-digit annual growth is expected during this decade (Sunand Bean, 2001), driven both by export markets and increases in domesticfurniture consumption 7 .6 Xiaozhi Cao, Hansen, Meiqi Xu and Boming Xu (2004).7Robb and Bin Xie (2003).
To fuel this boom in production, Chinese forest products imports aregrowing quickly, rising from US$6.4bn in 1997 to US$11.2bn in 2002(Sun et al., 2004). As the world’s second-largest wood products importerbehind the US, China imported 106.7m 3 in round wood equivalent (RWE)volume of wood products in 2003, more than double the 40.2m 3 of 1997.Russia, Malaysia, Indonesia, New Zealand and Thailand are the top fivetimber suppliers to the Chinese market. Much of these timber imports runagainst international conservation agreements and are illegal, both in theexporting economies and in China (Watts, 2005). It is forecasted that by2010, domestic Chinese wood supplies will be 114m 3 , creating a shortfallof 99m 3 that needs to be filled by imports (Lague, 2003). The furnituremakingsector will account for at least 18% of the country’s total wood consumptionby 2010 (Xu, 2004; China National Furniture Association, 2004).3. A case study of the gardenfurniture industries in SouthAfrica and GhanaBearing in mind that this is a limited pilot project which identifies theemerging impact of China on SSA manufacturing exports, we focus oninterviews with key informants in a particular segment of the wood furnitureindustry – the garden furniture sector. We chose this sector since itrepresents the major wood furniture export from Ghana. Here we interviewedkey informants which included:Major global buyers of garden furniture;Producers of garden furniture in Ghana and South Africa;Producers of non-garden furniture in Ghana;Furniture retailers in Ghana; andGovernment officials and other key informants in both countries.Although a restrictive sample in terms of numbers (especially in SouthAfrica), these interviews identify important emerging trends.What do global buyers of garden furniture think about future trends inglobal sourcing?A group of 15 firms exhibiting garden furniture at GLEE 2005, the internationalgarden and leisure exhibition at Birmingham, were interviewedfrom 18-20 September 2005. They ranged in size from small family businessesto large-scale firms. All specialised in the import of garden furniture,predominantly for the UK market, but also for the continental Europeanmarket. Most of this furniture was made from hardwoods. Some of thefirms had product ranges covering wood and non-wood products, and alloffered accessories such as cushions and parasols as part of their service.The major source of furniture imports was Indonesia, followed by China.Indonesia benefits from indigenous sources of certified teak, and most ofthe importers had long-standing relationships with Indonesian suppliers,ranging from three to 15 years. Unlike China where they had no contactwith manufacturers themselves and purchased through third parties,there are generally direct links to Indonesian producers, who were visitedon a frequent basis. Indonesian firms posses both the required technologyand the skills. However, importers paid little attention to labour standardsand were much more focused on the use of certified timber.Notwithstanding Indonesia’s overall dominance as a supplier, China is arapidly growing source of supply, based in part on the use of hardwoodsprocured from West Africa and substitute hardwoods such as rubberwoodobtained from Malaysia. China’s particular strength lies in products combiningdifferent types of materials (for example, aluminium and wood)and cushions and parasols. Importers are concerned that some of China’sfurniture exports use uncertified illegal timber sourced from Myanmar, andthis is one factor holding back the rapid progress of China as a source ofsupply. One advantage which China clearly holds is its flexibility as a supplierand also its design capabilities.Whilst Indonesia and China dominate this sector, other suppliers are alsoimportant, notably Poland (a niche provider of hand-made oak swings),Sweden (low-cost softwood pine furniture), Thailand and Vietnam. Noneof these 15 importers either sourced from SSA or considered it likely thatthey would do so in the future, notwithstanding the fact that much of China’sgarden furniture was made from timber sourced from West Africa.In addition to these importers, we interviewed the largest importer ofGhanaian garden furniture into Europe, using iroko, a teak-like hardwood.It also sources hardwood furniture from China, Vietnam and Indonesia,softwood pine furniture from Poland and contemporary furniture (a hybridwood and metal product series) from Central and Eastern Europe. Productsare designed for the high-income niche market and are sold into boththe UK and continental EU markets.This importer has had a long-term relationship with Ghana’s major exporterof garden furniture, accounting at its peak for more than two-thirdsof the Ghanaian firm’s exports. In the past it held a share of equity in theGhanaian furniture manufacturing plant and a linked sawmill. Its link withthe Ghanaian exporter began in the 1980s when it acted as a buyer forUK retailers. In 1992, the company developed a product line of gardenfurniture under its own brand name. Crucially, it rather than the Ghanaianproducer was responsible for designing the products sold under thisbrand name.43<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The Impact of the Asian Drivers on Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>
<strong>Trade</strong> Liberalisation44<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Despite the co-evolution of the UK and Ghanaian firms and the link inequity, the UK-based importer gradually began to diversify its sourcingafter the late 1990s. By 2001, only 56% of its furniture was sourced fromGhana, and this dropped further to 35% in 2005. Its imports from Ghananow account for over 95% of Ghanaian wood furniture exports. It is expectedDelivery times work in Ghana’s favour, with an average shipping time of18 days (from the date of placing an order to reaching a UK port). Shipmentsfrom China and Vietnam take 28 days and 22 days from SouthAfrica. However, speed of delivery is much less important than on-timedelivery. Ghana has no advantage over China in this respect.that the absolute value of orders from Ghana will remain constant,but market growth for garden furniture will be met from its other primarysuppliers, China and Vietnam. It used to source from South Africa, butwith the strengthening of the Rand, South Africa has been dropped as asource of supply. The major reason for diverting imports from Ghana toChina and Vietnam is cost. The price of furniture from Asia is 24%-40%lower than similar products from Ghana. South African costs are 20%higher than those of Ghana (see Table 6).We asked this UK importer to rank the importance of national conditionfactors to its sourcing decision. We also asked the company to comparethe national conditions of Ghana, China, Vietnam and South Africa. Theresults are presented in Figure 4. China, and to a lesser extent Vietnam,import raw wood for production but access to material does not have anyinfluence on the sourcing decision of the buyers. Red tape is an importantissue but local governments do not hinder the shipping of goods. AlthoughTABLE 6: PRICE OF A SIMILAR PIECE OF GARDEN FURNITURE FROM SELECTED COUNTRIES, 2005Average price Price index (%)Ghana £50 100South Africa £60 120China £30 60Vietnam £<strong>38</strong> 76Source: Company interviewFIGURE 4: BUYERS’ VIEWS ON NATIONAL CONDITIONSImportance scale (1 = not important; 5 = critical)Performance scale (1 = poor; 5 = excellent)Source: Company interview
there is differential performance between these supplying countries (withChina and Vietnam generally being seen as more desirable locations), allof these four countries meet the firm’s minimum expectations.We also asked the UK buyer to rank the importance of physical infrastructureand human capital factors and to evaluate the performance of suppliers.Figure 5 shows that the buyer considers the labour capabilities ofChina in general to be superior to that of other suppliers. Its workforce isbetter educated than the Ghana workforce. Corporate social responsibilityis becoming an important issue for this buyer but the company was not ina position to comment on the performance of the suppliers.The buyer was also asked to consider operational factors (Figures 6 and7). Cost, quality and on-time delivery are critical-order winners. Productsfrom Ghana performed better than other suppliers in delivery and qualityfactors while Asian products performed better on price. Chinese firms appearto be more ambitious, however. The MD observed that:“On more than one occasion, our Chinese suppliers haveoffered an improvement on... our designs and did not chargeme for it. Other suppliers would probably not be able to makethis change. And if they could, they would send me a bill.”3.1 What is happening on the ground inGhana?Unfortunately there appears to be no structured analysis of the Ghanaianfurniture value chain in general, or the garden furniture and exportingsegments in particular. (The Furniture Manufacturing Association ofGhana has unsuccessfully tried to produce a census of the industry inrecent years). A single firm is responsible for most furniture exports, andthis comprises a mix of garden furniture to the EU (via its UK buyer – seeabove) and parquet flooring to Italy. It is estimated that informal-sectorproducers account for at least two-thirds of the local market (interviewswith key informants).A number of factors define and dominate the Ghanaian furniture industry.The first arises from the global demand for certified hardwoods. TheGhanaian government has taken extensive steps to ensure that all of itshardwood timber supplies are of certified origin. There is no clear estimate45<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>FIGURE 5: BUYERS’ VIEWS ON PHYSICAL INFRASTRUCTURE AND HUMAN CAPITALImportance scale (1 = not important; 5 = critical)Performance scale (1 = poor; 5 = excellent)Source: Company interviewThe Impact of the Asian Drivers on Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>
<strong>Trade</strong> LiberalisationFIGURE 6: EVALUATION OF OPERATIONAL FACTORS, COST, QUALITY AND DELIVERYImportance scale (1 = not important; 5 = critical)Performance scale (1 = poor; 5 = excellent)46<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: Company interviewFIGURE 7: EVALUATION OF OPERATIONAL FACTORS, TECHNOLOGY AND LABOURImportance scale (1 = not important; 5 = critical)Performance scale (1 = poor; 5 = excellent)Source: Company interview
of how successful this has been, and the extent of illegal logging. Second,complementing this policy on certification has been a ban on the exportof unsawn logs. To preserve the local industry, local saw mills are requiredto allocate at least 20% of their output to the local market. However, as inthe case of South Africa (Kaplinsky, Morris and Readman, 2001), most ofthis timber is for the building industry, and furniture manufacturers haveto struggle for supplies and, more importantly, are unable to get the woodcuts in sizes or qualities which are suitable for furniture manufacture, particularlyfor the quality of furniture required to compete in external markets.Furniture manufacturers in Ghana therefore often resort to the use ofillegal timber, roughly and inefficiently chain-sawn (with high wastage) inrural areas. None of this is helpful in achieving high-quality products.A third problem confronting the local industry is the absence of complementaryinputs, particularly aluminium and other metal fittings, butalso soft furnishings (although these latter products are often acquiredindependently by external buyers). This is a growing problem as gardenfurniture becomes more innovative in design and wood is only part of amore complex product. A further problem inhibiting competitive production,particularly in export markets, are the short time horizon and highcost of capital (more than 20% per annum), as well as the fluctuatingexchange rate which makes exports a hazardous business.Almost all of the manufacturers we interviewed saw China as a problem.The major exporter described conditions in its external markets as beingvery tough. In the late 1990s China first entered global markets with poorquality products and suffered from an adverse reputation. SubsequentlyChinese producers began to use better woods, including hardwoods importedfrom West Africa, and have now become a very substantial threat,leading in a drop in its unit export prices. But manufacturers targeting thelocal market also complained about Asian competition in general, andChinese competition in particular. Some reported attempts to raise thematter with government in 2001, but to no avail; government’s over-ridingconcern has been with certification and the ban on exports of rawlogs, and the furniture industry has had little lobbying power. In the face ofthese pressures, manufacturing firms reported either that they had givenup all hope of exporting, or that they saw the emerging Economic Communityof West African States (ECOWAS) market as their primary target.We are unable to judge the realism of this response of targeting regionalmarkets and none of our respondents seem to have considered this issuevery thoroughly. It may, therefore, be a matter of clutching at straws. Anumber of furniture firms have either reduced their labour forces or expectto do so in the near future. Ghana’s furniture industry is not in a robuststate of health.3.2 What is happening on the ground inSouth Africa?The South African wood furniture industry too is not in a robust state ofhealth. The recent NPI Customised Sector Programme (NPI, 2005) showsa ‘furniture industry’ dominated by the export of leather automobileseats. After achieving steady growth between 1963 and 2003 (73% inreal terms), real output began to decline after 2002. This reflected both afall in exports (wooden furniture exports declined from R892m in 2003 toR629m in 2004) and an increase in imports. The share of imports in domesticconsumption rose from 4% in 1995 to 21% in 2002. Employmentin the furniture industry remained fairly stable after reaching its peak in1989 (43,000), but with the sharp fall in exports after 2002, fell back to<strong>38</strong>,000 in 2004. Profitability is low and many firms are on the edge ofshedding labour. (All data in this paragraph were taken from NPI, 2005).The NPI study ascribed these systemic weaknesses to a combination ofpoor manufacturing processes (with high inventories and outdated equipment),poor design and R&D capabilities, high wages and the absenceof government support. These corroborate the conclusions of Kaplinsky,Morris and Readman (2001), who also point to weaknesses in the wholechain, particularly in the links between furniture manufacturers, timbersuppliers and the forestry sector. There are additional problems whicharise from high levels of ownership concentration, not just in the furnituresector, but throughout the chain (Kaplinsky and Manning, 1999).The NPI study concludes with a depressing message:The future of the furniture industry may not immediatelyhang in the balance, but trends suggest that the SA furniture[industry] is moving in the direction of ruin” (NPI, 2005:18).Our interviews with two garden furniture exporters are illustrative of thesegeneral trends, and also point to the specific impact of China and otherAsian competitors. The first manufacturer specialises in garden furniture,which accounts for 70% of total sales, selling both into the domesticmarket and into the continental EU market. But falling global prices anda rising domestic currency have made export sales very unattractive andexports as a share of sales have fallen from 40% of sales in 2000 to 10%in 2005. Sales fell in current Rand prices by one-third between 2003 and2004. The major external competitors are Asian-based, with China and Indonesiabeing particularly active in global markets. There are plans to takeover an ailing South African producer and to target future export growthby diversifying into indoor furniture and developing a brand name.47<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The Impact of the Asian Drivers on Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>
<strong>Trade</strong> Liberalisation48<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The second manufacturer uses saligna hardwood substitutes and producesboth garden and indoor furniture. Sales rose between 2005 and 2004(from $1.5m to $1.7m), but on the back of falling profitability. In 2003,the labour force was halved, and mechanisation and improved manufacturingprocesses have seen a sharp rise in productivity. Export prices havenot risen in relation to either costs or domestic prices, and the share ofexports has fallen from 50% in 2000 to 30% in 2005. Two factors arereported as undermining competitiveness in external markets – the appreciatingexchange rate and competition from Vietnam and China. As theoperations manager observed:“Down the road from us there is a large saligna forest andsawmill. This facility is FSC certified and we buy from them.Lately they have been selling timber to Vietnam. It’s expensiveto send timber by ship. The Vietnamese manufacturers useAustralian designs and turn out good quality and cheapproducts. Furniture made from South African trees, producedin Vietnam, can be bought in Cape Town and Durban. And thiswill be cheaper than anything we can make.”4. Where to for the SSAfurniture sector?The evidence presented above on the SSA industry makes for blunt reading.Most of the continent’s furniture industry is inward oriented and involvesthe informal sector using low-grade timber and timber offcuts (ofteninvolving wasteful chain-sawn timber) to produce basic products formiddle- and low-income domestic consumers. In some cases, particularlyin West Africa, this involves the illegal use of uncertified timber, leadingto a depletion of scarce global hardwood supplies and contributing toglobal warming.The increasingly scarce hardwood timber in which West Africa has a comparativeadvantage and the hardwood substitute (saligna) in which SouthAfrica has a comparative advantage are being exported on a growingscale to Asia in general, and China in particular. Partly on the back of timbersourced from other countries, China has become a growing presencein global furniture markets, and SSA firms are unable to compete effectivelywith it. It is not just China which threatens SSA’s furniture industryin the global market but other Asian economies as well, such as Indonesia,Vietnam and Thailand. And it is not just SSA which is threatened in thisway – Thai rubberwood furniture manufacturers (who pioneered the useof rubberwood for furniture) are being squeezed out of Japanese marketsby Chinese competitors using rubberwood imported from Thailand andMalaysia (Mitsuhashi, 2006).As a consequence of this Asia-based competition, furniture exports fromSSA have come to be threatened by a combination of falling prices, reducedprofitability and falling and evaporating market shares. To be blunt,SSA’s export-oriented furniture industry is at a crisis point, if that crisispoint has not already been reached. But it is not just the export marketwhich is threatened. The liberalisation of imports into SSA has meantgrowing domestic competition from low-cost Chinese- and Asian-sourcedfurniture. This, too, threatens the survival of the formal furniture-manufacturingsector throughout SSA.The threat to the furniture industry has not had an adverse effect on allof the wood furniture value chain, since many SSA economies have seenlargely stable and even growing exports of timber, logs and chips andpulp for the paper industry. For example, in 2004, South Africa’s exportsof all forestry-related products had grown to US$1.3bn, from US$794min 2000. Almost all of this was pulp and chips for the paper industry, inwhich it is a major player, and wood furniture exports were less thanUS$50m. Similarly, in 2004, Ghana exported US$105m of sawn timber,compared to US$77m in 2000, and less than US$9m of wood furniture. 8This pilot study of the furniture sector has been hampered by the absenceof systematic and detailed data on these sectors. Even the relatively sophisticatedSouth African data set is based on highly aggregated figures,and provides little insight into the underlying processes which determinesystemic inefficiency. We know that this inefficiency is not confined toindividual segments in the wood and furniture value chain, but also arisesfrom poor levels of co-ordination between segments in the chain as awhole (Kaplinsky, Morris and Readman, 2001). As a consequence, we arenot only unsure which policy levers to tap, but also what the impacts areof this systemic weakness on economic growth, growth multipliers, the agriculturalsector, employment and poverty and urban and rural livelihoods.5. PolicyDuring the late 1980s and early 1990s there was a lively debate in SouthAfrican industrial policy on the necessity of ‘beneficiating’ raw materials(Joffee et al., 1995). This mirrored the debate in the development litera-8With the exception of the estimate for Ghana’s wood furniture exports (derived from interviewswith key industry informants), all figures are drawn from the FAOSTAT website, lastaccessed 9 January 2006.
ture on the nature of dependency in the developing world (for example,Girvan, 1987). On the one hand, it was argued that a primary route forindustrialisation in low-income economies was to take advantage of resourceendowments and to become processors of these raw materials.Thus, for example, the existence of a natural advantage in forestry (tropicaltrees grow more rapidly and more dense in a shorter period of time thanthey do in temperate climates), meant that developing countries shoulddevelop a comparative advantage in industries which use timber, such asfurniture. The policy implication of this was that local raw material processorsshould get priviledged access to these inputs. The counter-argumentwas that the post-war industrialising economies in Asia (Japan and thenthe Asian Tigers) had performed well despite (and perhaps because of)having no access to these raw materials. They imported these from thelowest-cost producers and added value to them very efficiently. There wasno reason why local processors should be favoured, and they should beforced to source their inputs at the landed cost of imports. Market pricesshould prevail.Clearly, the systemic weaknesses of SSA furniture manufacturers meansthat they are unable to add value to their timber as efficiently as theirChinese and other Asian competitors. This leaves two sets of related policyissues to be confronted. The first is whether SSA’s furniture industry is insome sense disabled compared to its competitors by factors exogenous tothe enterprise. This may be due to poor infrastructure, high interest rates,fluctuating and uncompetitive exchange rates or system-wide underinvestmentin public goods such as basic education. These are policies whichare of general relevance, targeting market failures which affect all industryand are not to be met via sectorally targeted industrial development.The second set of policy issues relates to the specific needs of the furnituresector. This involves targeted measures to promote its efficiency, such asfurniture training programmes, the promotion of furniture design skills,firm-level furniture upgrading initiatives, subsidised access to finance andthe facilitation of value chain co-ordination and efficiency. For these policiesto be justified it would need to be shown that the furniture industryis either specifically disabled compared to other sectors, or that it holdsa place of strategic importance in overall industrial development. Thesestrategic issues were considered above.of subsidy were much higher than the nominal preference rates, since SSAclothing exporters can access duty-free imported inputs and obtain tariffprotection on the nominal cost, insurance and freight (cif) of importedproducts rather than on their value added. This arises from a special derogationto the African Growth and Opportunity Act (AGOA) provisions,which is not only specific to that sector but holds less potential in a woodfurniture industry using domestic inputs.ReferencesChina National Furniture Association (CNFA) (2003) China Furniture AssociationAnnual Report, 2003. CNFA, Beijing, China. 120 pp. (in Chinese)Quoted in Xiaozhi Cao X., Eric N. Hansen, Meiqi Xu, Boming Xu (2004)China’s Furniture <strong>Industry</strong> Today, USA: Forest Products Journal <strong>Vol</strong>. 54,No. 11.European-Commission (1997) Panorama of EU Industries-1997: Furniture.Brussels, European Commission.Girvan, N. (1987) “Transnational Corporations and Non-fuel PrimaryCommodities in Developing Countries”, World Development, <strong>Vol</strong>. 15, No.3, pp. 713-740.Joffe A., D. Kaplan, R. Kaplinsky and D. Lewis (1995) Improving ManufacturingPerformance: The Report of the ISP, Cape Town: University of CapeTown Press.Kaplinsky, R., M. Morris and J. Readman (2001) “The globalisation ofproduct markets and immiserisiing growth: Lessons from the South Africanfurniture industry”, mimeo, Brighton: Institute of Development Studies,and Durban: School of Development Studies.Kaplinsky R. and C. Manning (1999) “Concentration, competition policyand the role of small and medium sized enterprises in South Africa’s IndustrialDevelopment”, Journal of Development Studies, <strong>Vol</strong>. 35, No. 1,pp. 139-161.Lei, S. and M. Mcgowin. (2002) If we can’t beat them, let’s work together.US: Wood & Wood Products Oct.:63-67.49<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>There is one final set of policy conclusions which can be derived fromthe experience of SSA’s clothing and textiles sector – that the only factorsustaining the labour- and export-oriented clothing sector was the availabilityof trade preferences in external markets. These trade preferenceswere designed to provide significant support to SSA industry, with impliciteffective rates of subsidy of between 28% and 84%. These effective ratesMitsuhashi, K. (2006) “The Furniture Value Chain from Thailand to Japan:Upgrading and the Roles of Buyers”, D Phil Dissertation, Brighton: Universityof Sussex.NPI (2005) Customised Sector Programme: Furniture, Pretoria, NationalProductivity Institute.The Impact of the Asian Drivers on Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>
<strong>Trade</strong> Liberalisation50Robb D., Bin Xie (2003) A Survey Of Manufacturing Strategy And TechnologyIn The Chinese Furniture <strong>Industry</strong>, Great Britain: European ManagementJournal <strong>Vol</strong>. 21, No. 4, pp. 484–496Spalding, Josephine, (2001) ‘<strong>Industry</strong> <strong>Trade</strong> Summary, Furniture and MotorVehicle Seats’, USITC Publication 3<strong>38</strong>2, January 2001, Washington,D.C.: US International <strong>Trade</strong> CommissionWatts, J. (2005) “China consumes forests of smuggled timber”, The Guardian,April 22nd, http://www.guardian.co.uk/china/story/0,,1466021,00.html, accessed 9th January 2006.Xiaozhi Cao X., Eric N. Hansen, Meiqi Xu, Boming Xu (2004) China’s Furniture<strong>Industry</strong> Today, USA: Forest Products Journal <strong>Vol</strong>. 54, No. 11.<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>
TIPS and economic regulationTIPS has added economic regulation as one of its research areas, focusing mainly onutility regulation (energy, telecoms, gas and water).Why is economic regulation important?51Infrastructure industries such as energy, telecommunications,water and transport are essential elements of oureconomy. These industries, which are in various stagesof restructuring, need to function and develop in an efficientand effective manner. They are crucial for generating economicgrowth, alleviating poverty and increasing international competitiveness.All of these large utilities have some natural monopoly elements.A natural monopoly can be defined as the supply of aservice/good where economies of scale are such that costs ofsupply for the whole market are lower if there is a single supplier.This is usually the case where large sunk costs are involvedand where a duplication of facilities would be seen aseconomically inefficient.These natural monopoly concepts present a dilemma: on the onehand a natural monopoly is more efficient if a single firm suppliesan entire market; on the other, this monopoly holder willbe tempted to exploit its monopoly power to maximise his/herprofits in the absence of any competition.From the 19 th century, these industries were brought under directstate ownership and control in many countries. The reasonfor state ownership was mainly seen as the strategic nature ofthe industry and also the fact that they were seen as naturalmonopolies. However, it was realised that some of the activitiesundertaken by these utilities were not naturally monopolistic,which led to unbundling of the potentially competitive areasfrom the monopoly parts.Since the 1980s, more and more countries, including developingcountries, have been privatising their public utilities in the beliefthat this would typically lead to increased efficiency, investment,higher quality of supply for consumers and as an avenueto raise revenue for government. The success of privatisationwas premised on the existence of competition; where there wasno competition, a regulator was created to safeguard and promotethe competitive process.Economic regulation has therefore developed as a tool or watchdogto ensure fair treatment to customers (wich is supposed tocome automatically with competition) and to ensure that competitorshave fair access to network facilities that are controlledby incumbent service providers. It is therefore applied whenthere is a case of market failure and to mitigate distortionsassociated with monopoly prices in situations where there arebarriers to entry. Economic regulation should provide the rightinvestment incentives to market participants and should protectconsumers from monopoly abuse.Policy-makers have to design these regulatory frameworks, whoseoutcomes are aimed to simulate that of a competitive market inthese industries. Regulators have to implement these policies,which can sometimes be conflicting and contradictory, whichmeans a regulator has to execute a balancing act to ensurean efficient achievement of all objectives – reasonable tariffs,third-party access, universal service obligations, potential forcompetition and creating an environment to attract investment.Economic regulation therefore has a crucial impact on how governmentpolicy is implemented, how infrastructure industriesdevelop in terms of investment and how other social objectivesof government are met.Regulation is a complex task and involves particular skills andexperience. Regulatory failure happens where economic welfareis reduced rather than increased. It is therefore crucial for aregulator to have the right skills and experience so that economicwelfare can be increased. The requirement for regulatorycapacity (consisting of regulatory governance and the regulatorysubstance) cannot be overestimated and is limited in SouthAfrica, especially in view of new legislation and the creation ofindependent regulators.South Africa has, over the past decade, seen the developmentof different regulatory entities in these infrastructure industries.All of these regulators differ in mandate, structure, regulatoryambit and funding. Although reasons can exist for piecemealdevelopment of sector regulators, coherence in policies for theseindustries is usually preferable.Most regulatory institutions in Africa are no more than a fewyears old, with none older than 12 years. Eberhard (2005) statesthat the challenges in establishing new public institutions indifficult contexts have been underestimated, as it takes time tobuild and entrench governance, management and organisationalsystems, in addition to the imperative of building new professionalcapacity. In South Africa (Teljeur, Gillwald, Steyn & Storer,2003) there is a particular lack of economic regulatory capacity<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The Impact of the Asian Drivers Announcementson Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>
Announcements52<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>on many levels, but especially at government, regulator, utility There is a saying by an unknown source: “if you don’t knowand consumer/customer level. This is not a unique phenomenon where you’re going, any road will get you there”. These thoughtsand can be found in many developing and transition economies lead one to realise the importance of coherence and the argumentthat there should be a co-ordinated focus on the achieve-(Kessides, 2004). One of the reasons is that economic regulationis a relatively new area of expertise and the necessary skills ment of certain policy goals. A large body of literature exists onare scarce.the experience of the developing world in this regard that can beextremely valuable for the South African context.Campbell-White & Bhatia (1998), in the context of Africa, statesthat, ‘regulation is being examined as part of individual sectorinitiatives, but these efforts are unco-ordinated, and implementationis being left to follow privatisation instead of being putin place concurrently’.TIPS’ involvement and focus in economic regulationTIPS is primarily a trade and industrial policy research institute.However, developments in regulated industries are importantsince it impacts on a whole range of industries and has aneffect on the economy as a whole. Research is needed to establishthe nature and consequences of economic regulation atnational, local (provincial, municipal) and international levels.TIPS has organised economic regulatory programmes in the pastand will continue with these programmes with the aim of providingRegulation is seen as a complex balancing act betweenadvancing the interests of consumers, competitors andinvestors, while promoting a wider, ‘public interest’agenda. How has this been addressed successfully elsewhere,in particular focusing on the developing world?What is the process and content of restructuring, privatisationand regulatory reform in developing countries?What can be learnt from it?policy-relevant research to policy-makers and engagingon a wide range of issues. It will also focus on experiences Regulation and poverty alleviationin economic regulation in developing and developed countriesand derive lessons that can be valuable for the South Africancontext.Research issues that will be covered include: A big challenge for government is universal access to essentialinfrastructure. Initiatives in this regard is mostlydriven by government and is not really seen as part ofthe mandate of regulators. What is the role and limitationsof regulation in poverty alleviation and what is the Regulation, restructuring and privatisationnecessary economic and social policies that must accompany this? What has been the experience elsewhere inThe sequence of reforms are important. Countriesthis regard and what lessons can be learnt from it?have gone about reform in many different ways andtheir experiences can be valuable. It is important to Effects of restructuring and regulation on the developmenttake stock of why and how policies developed. Questionsof essential infrastructurein this regard include the role of policy transfer, The role of economic regulation in the development ofespecially looking at the applicability to the specificeconomic infrastructure, such as telecommunications,country context; how the issue of ‘policy transfer’ hasports, airports, postal services, water and energy suppliescannot be underestimated.been handled in other developing countries and whatcan we learn from it. Policy transfer from developed todeveloping countries has not always been successfulbecause of differences in, for example, culture, marketrealities and political values, which have not alwaysbeen taken into account. Has an environment been created that will promote investment?What has been achieved in other countries?For regulation to be efficient and effective there should,amongst others, be a willingness by government to establishthe regulatory rules and allow regulators to operatewithin the rules, and there should be an ‘arms length’ The form of regulation is important to determine regulatoryefficiency. Under what circumstances is it preferablethat government should be the provider of a publicrelationship between the regulator and stakeholders (includinggovernment).service rather than a privatised, regulated business? Do our regulatory bodies have credibility and legitimacyin the eyes of customers and potential investors? If not,
what is missing? What has been the experience elsewhere?What have been done wrong and what can bedone to create efficient and effective regulatory institutions? Investment in a regulated environment is subject to thethreat of hold-up, leading to under-investment. Is thisa real or perceived threat?Regulation and competition policy Privatisation and the promotion of competition areseen to be important for development of infrastructureindustries. Government is a shareholder in the majorityof these industries. How does this affect the possibilityof competition and how has this type of situationbeen addressed in other developed and developingcountries? Effective legislation is crucial to facilitate competition.What constitutes poor implementation of this and whatleads to poor/good implementation? What is the experienceelsewhere? What has been the goal with the introduction of competitionand what has been the experience since it wasimplemented? Weak regulation of competition is likelyto undermine the potential gains to be made from restructuringand privatisation. What is the link between economic regulation and competitionpolicy and how it is currently dealt with? Whatlessons can be learnt from experiences elsewhere inthis regard?latory expertise in developing countries is generally inshort supply. What are the considerations in determining the formof a regulator? Sometimes transitional regulatory systemsare needed because a country may be unable toimplement the independent regulator model because ofa lack of capacity, commitment or both. It is unrealisticto expect that the requisite capacity and commitmentwill appear overnight when little or no prior experienceexists with autonomous regulation and when there isnot enough trust in an entity that it will execute itsfunctions efficiently. The requisite skills and experience are currently lackingin South Africa. What can be done about this and whathave been done in other countries in the region andother developing countries that faced the same challenges?The nature and scale of producer and consumer representationand whether there is a lack of participation shouldbe examined. How are regulators ensuring that all marketparticipants are included in the regulatory processes? Transparencyand participation typically lead to credibility andlegitimacy of regulators.There should be a focus on country and regional case studiesof economic regulation and deregulation to foster a widerunderstanding of the particular nature of regulation andmarket liberalisation in the context of developing countries.This could include building a knowledge base on the natureof regulatory regimes, especially in developing countries.53<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Policy formulation, coherence and the relationship withregulation The institutional context is critical to the processes andoutcomes of any regulatory regime. This context shouldclearly set out the legal and regulatory framework andshould then trust the regulator to continue with implementingits mandate. Checks and balances are importantto ensure that a regulator is not a ‘loose canon. A lack of policy coherence can create uncertainty whichcan impact the effectiveness of a regulator. Research should focus on current legislation and internationallyrecognised regulatory principles, overlapof government departmental jurisdictions, applicationof legislation, the position of regulators vis-à-vis statemonopolies, typical regulatory structures and the consequencesof regulatory failure.Capacity-buildingIn addition, TIPS will focus on the development of a capacitybuildingprogramme with the aim of improving capacity in economicregulation across a spectrum of stakeholders. Training toimprove capacity is crucial, since a lack of capacity threatensregulatory independence and also the credibility of policies,which might impact negatively on potential investment and thelong-term development of the sectors involved.ConclusionIn order to develop this research programme, TIPS will needbuy-in from stakeholders and will focus on areas and researchgaps that will assist policy-makers.Regulation need people who are passionate about makinga difference. The implementation of regulation isa human and not simply a technical function, so thequality of regulators is important. Unfortunately, regu-The Impact of the Asian Drivers Announcementson Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>
Development & Poverty Reduction54<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>
development& povertyreduction55CAN SMALLFARMERS<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>SURVIVE, PROSPER OR BE THE KEYCHANNEL TO CUT MASS POVERTY? *BY MICHAEL LIPTONPOVERTY RESEARCH UNIT, UNIVERSITY OF SUSSEX, UK1. IntroductionTHIS PAPER REVIEWS THE ROLE of small farms in development and poverty reductionin countries (or large regions within a country) with persistent mass dollar poverty– where, say, one-third or more of the population are below the purchasing powerparity (PPP) dollar-a-day poverty line. The conventional wisdom, set out with evidencein the International Fund for Agricultural Development’s (IFAD’s) 2001 Rural Poverty Report,is that in such places development with substantial, sustainable mass poverty reduction is achievable,but initially – save in very rare cases – only via rapidly accelerated productivity and incomegrowth on small farms. There are three reasons why this might be disputed: (1) Non-farm growth might be seen as a more credible option. One might question whether – given past progress,environmental limits, world price trends, falling share of farm workforce and so forth – agricultural growth, onsmall farms or large, is feasible in remaining mass poverty areas, let alone promising against poverty. (2) Alternatively, one might accept that accelerated agricultural growth is both feasible and a precondition forinitial mass poverty reduction, but believe that processes of national poverty reduction are consistent with largescalefarming.The paper reviews the role of small farms indevelopment and poverty reduction in countries(or regions within a country) with persistentmass poverty. It discusses the argumentssupporting the importance of agriculturaldevelopment for poverty and argues that initialsuccess in reduction of mass poverty requiresprior agricultural developments. It explores inparticular the role of small-scale farming andpolicy requirements to ensure the competitiveadvantage of small farms.* This paper first appeared in eJade, the electronic Journal of Agricultural and Development Economics, <strong>Vol</strong>ume 3, No. 1,pp. 58-85, 2006 and is available online at www.fao.org.es/esa/eJADE.
Development & Poverty Reduction56<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>(3) One might claim that small farms in (some) developing countries (a) never weremore equitable and efficient than large farms as a source of initial mass povertyreduction and development; (b) whether or not they were, are not now, becauseworld-wide processes of farm change – commercialisation of increasing proportionsof input and output, institutional developments such as supermarkets, privatisationof key aspects of technical progress and of output and process grades andstandards – now indicate a large-farm focus.This paper concentrates on issue (3), but first briefly reviews the validrange of assumptions on issues (1) and (2) above, which are logically priorissues. Has ‘agricultural progress’ in developing countries still a big part toplay in reducing initial, persistent mass poverty? If so, do the processes bywhich agricultural progress can cut poverty mandate a small-farm focus,exclude it or neither? Only after agreeing working assumptions on (1) and(2) can we proceed to our core question (3): what do theory, history andrecent experience tell us about whether small farms (to be defined) can,will or should remain central to initial mass poverty reduction? 1© Matthew de Gale2. Is farming a route to cuttinginitial mass dollar poverty?Until fairly recently, most developing countries acted as if the role of agriculturein mass poverty reduction was quite small. Terms of trade wereheavily turned against farming through selective industrial protection,subsidised food imports, compulsory procurement of farm products andexchange-rate over valuation. Public expenditure per person on health, educationand physical infrastructure was far higher in rural than urban areas.Development was seen as requiring rapid industrialisation, even withoutand before significant agricultural or rural progress. Two things havechanged this perception. First, it became clear that price and expenditurepolicies to bias development towards industry-led growth delivered neithergrowth nor poverty reduction. Terms-of-trade and public-expenditurebias in favour of urban industry, while cutting farm growth and harmingthe rural poor, generated little industrial growth (at border prices) and verylittle extra employment per unit of output. Second, technical progress – irrigation,fertilisers and new crop varieties (hybrid maize in Latin Americafrom the mid-1950s and semi-dwarf wheat and rice in much of Asia fromthe mid-1960s) – showed that, under the right conditions, agriculturalgrowth could bring hundreds of millions of people rapidly and sustainably1This paper draws, with regard to issues (1) and (2), on M. Lipton, ‘The family farm in aglobalising world: the role of crop science in alleviating poverty’, International Food PolicyResearch Institute, Washington, D.C., 2005; and, with regard to issue (3) on R. Eastwood,A. Newell and M. Lipton, ‘Farm size’, in <strong>Vol</strong>. IV, Handbook of Agricultural Economics, ed.R. Evenson, P. Pingali and P. Schultz, Rotterdam: Elsevier, forthcoming. On all three issuesand related matters, see International Fund for Agricultural Development, Rural PovertyReport 2001, Oxford: Oxford University Press, 2001.out of dollar poverty. This worked despite artificial measures to damageboth price incentives and stability 2 for farmers in developing countries,both via developed-country farm subsidies and tariffs and developingcountrypolicies biased, sometimes heavily, against the rural sector.Governments, analysts and the public in some developing countries nowappear to accept, in principle, that mass poverty initially has to be, andcan be, reduced by growth in agriculture.Globally, Ravallion has shown that over 70% of the dollar-poor are ruraland, on best projections of migration and growth, that the proportion willbe over half until 2035. In the large majority of mass-poverty countries(and sub-regions):More than two-thirds of the workforce has its main single income source in agriculture.In rural areas, though recent research shows that the rural poor derive significant,growing parts of income from non-farm work, 3 agriculture is by far the main incomesource.The few available studies show that 8%-15% of workers classified as urbanhave agriculture as the main income source; the proportion of the urban poor ishigher.Farming is what the great mass of the poor do. That, however, does notimply that poverty reduction should take an agriculture-first path (to raisepoor people‘s income from what they mostly do), any more than it impliesan anti-agricultural path (to get the poor into activities where poverty isless). But it demands that we confront and specify the choices betweenthose two routes.The poor’s income depends mainly on labour. To bid up demand for labour(and hence employment and/or the wage rate) in capital-constrainedeconomies, capital should go where capital costs per extra workplace – inself-employment or hired work – are relatively low, as in most agriculture.Though the response of employment to output growth has been falling insome poor areas (within India and China), it remains substantially largerthan in other sectors. 42EU long used a ‘reference price’ system so EU farmers, when world prices fell, were compensatedby higher support per unit of output, and so did not cut output. Hence they fell further,destabilising returns for non-EU farmers. See U. Koester, Policy Options for the GrainEconomy of the European Community: Implications for Developing Countries.(ResearchReport No. 35), Washington, D.C. : International Food Policy Research Institute, 1982.3The same research, however, shows that rural non-farm growth and employment intensitydepend largely on expanded demand from prior, nearby agricultural growth. Further, whilemost poor farmers derive significant income shares from non-farming, it is often forgottenthat the converse also applies.4Recent time-series from India have created an illusion: that the elasticity of employmentto output in agriculture has become very low or negative. Labour-displacing innovation(mainly tractors and weedicides) indeed cut the employment intensity of farm output,during agricultural growth. However, the innovation produced almost no growth, butreplaced labour on given land. With similar spread of tractors and weedicides but sloweragricultural growth, employment losses would have been substantially more.
Except perhaps for owner-occupied housing, farmland is the dollar-poor’smain major asset type (non-labour source of productive income) aswell as the asset type of which they have the largest share. For reasonsof arithmetic as well as work experience, it is therefore credible that morepoverty reduction is likely to be achieved by raising returns to farmlandthan to other assets – provided ownership of land is not more biasedagainst the poor than of other assets. 5The dollar-poor typically spend over two-thirds of income on food andover half of income on staples. Large, though not well documented, proportionsof the dollar-poor depend for these mainly on own or nearbyfarms. As long as internal or international transport costs for food (especiallystaples) loom large relative to value, local farming therefore restrainsand stabilises the price of the poor’s main consumables – which loommuch smaller in the spending of the rich.In assessing the case for or against small farms, we should ask whetherthe above three links from agricultural growth to poverty reduction – vialabour demand, asset ownership and food and staples prices – operatebetter through small or large farms and (which is an overlapping issue,but not quite the same one) through more or less equal distribution offarmland. 6 These static links from agricultural development to mass povertyreduction are over and above the intersectoral connections spelledout in classic work by Mellor, Johnston, Kilby 7 and others.Theoretical links apart, recent empirical evidence that farm growth is morepro-poor than non-farm growth in mass-poverty heartlands is strong. Historically,too, agricultural acceleration preceded non-farm growth and accompaniedinitial mass poverty reduction in developing Europe, Americaand Japan in the 18 th and 19 th centuries. This has also been true of late20 th century success stories, except a few unusually well-managed mineraleconomies and entrepot city states. 8Theory, history and empirics cannot disprove that some part of Africaor ‘inner’ South or East Asia should attack mass poverty while bypassingagriculture (environmental constraints or industrial opportunities mightrequire that). But they are not a good trio of enemies to choose.Ironically, it is in the countries that have accepted the priority of agriculturefor initial mass poverty reduction, and where governments haveacted accordingly to achieve such reduction – China, India, much of therest of Asia and some of Latin America – that later mass poverty reductionhas come to depend increasingly on non-farm growth. Yet even inmost of these successful countries there are regions with still high dollarpoverty (i.e. low local income growth, insufficiently offset by high inwardremittances). These poor regions within successfully developing countries– like the unsuccessful developing and poverty-reducing countries withina developing and poverty-reducing world – usually show weaker pastagricultural performance and more present dependence on agriculture.Moreover, they feature faster population growth than surrounding regionsor countries, so the concentration of the poor in these less successfulareas is increasing. 9Unfortunately, while many poor countries have shifted their price policiesaway from agricultural and rural extraction, their public-expenditure policieshave, if anything, shifted in the reverse direction, partly under fiscalstress, partly because the internal balance of rural-urban power requirescompensation for the price-policy shifts. On most measures, for developingcountries overall, the substantial rural-urban inequalities (in mean realincome, poverty incidence and health and education) show no downtrendsince the 1970s. 10 This may be related to the fact that, while urban bias indeveloping countries, at least via farm price repression, has been forceddown by economic realities, developed-country policies have increasinglydiscriminated against developing-country farm production. 11 This isshown both by the explosion of farm support in Organisation for EconomicCo-operation and Development (OECD) countries and by the collapse57<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>5Latin America, where land Ginis are exceptionally high, does not clearly follow the patternof Asia and Africa that farm growth is more pro-poor than is other growth, a patternshown econometrically in international cross-sections and in regional cross-sections andtime series for China, India, Indonesia, etc. See review in R. Eastwood and M. Lipton,‘Pro-poor growth and pro-growth poverty reduction: meaning, evidence, policy implications’,Asian Development Review 18 (2000): 22-58.6Growth in poor countries may be slowed by very unequal income and (more clearly) assets(Eastwood and Lipton 2000).7Johnston, Bruce F., and John W. Mellor, ‘The Role of Agriculture in Economic Development’American Economic Review 51 (1961): 566-93; Johnston, B. F. and P. Kilby, Agricultureand Structural Transformation: Economic Strategies in Late-developing Countries,New York: Oxford University Press, 1975.8Some countries once seen as exceptions have turned out not to be. For example, post-1945Korea’s manufacturing growth was not prior, but based on agricultural acceleration in the1930s and 1940s.9T. Dyson, R. Cassen & L. Visaria, Twenty-first Century India, Oxford: Oxford UnversityPress, 2004.10R. Eastwood and M. Lipton, ‘Rural-urban dimensions of inequality change’, in Inequality,Growth and Poverty in an Era of Liberalization and Globalisation, ed. G. Cornia (UNU-WIDER: Oxford: Oxford University Press, 2004.11OECD farm subsidies rose from US$182bn in 1995 (40% of production) to US$248bnin 1999-2001 (de Moor, A.P.G. ‘Perverse Incentives – Subsidies and Sustainable Development:Key Issues and Reform Strategies’, Earth Council, 1996; R. Ricupero, Report ofaddress to UN (ECOSOC), 30/6/2003. Third World Network http://www.twnside.org.sg/title/twe309a.htm These subsidies are claimed to help smaller OECD farms (which arerelatively labour-intensive) to survive, so more people can stay in farmwork: the peasantoutcome. Yet OECD farm support has not overcome the tendency of farm size to grow inrich countries, and of farm employment to decline – indeed, it may have worsened prospectsfor the peasant outcome. Between 1986-90 and 1996-97, farm employment fell from7.1% of the workforce to 4.9% – numbers fell by 14% – in the EU15 despite huge farmsupport, but fell much more slowly in Australia and New Zealand, with much less supportper unit of output [ILOSTAT].Can Small Farmers Survive, Prosper or be the Key Channel to Cut Mass Poverty?
Development & Poverty Reduction58<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>of international aid to agriculture, both in absolute terms (a fall of almosttwo-thirds since the mid-1980s) and as a share of sectorally allocable aid.In 1990-2015, the period in which developed countries purportedly prioritisehalving world dollar poverty as the first Millennium Development Goal(MDG), they have so far slashed aid to agricultural development – andintensified farm support policies that discourage it by false incentives toglut world markets. These incentives misdirect not only developed-worldfarmers, but also scientists everywhere who are increasingly stimulated todistort farm science to focus on increasing subsidised, capital-intensivecompetition against farms in the developing world, which is then lecturedabout free markets.There are some signs of improvement in developed countries. Many committeesand ministries promise modest reversal of the collapse of aid todeveloping-country farming and modest reduction of trade discriminationagainst it (though we have heard such promises before). Whether thishappens or not, more and more analysts and politicians in developingcountries with mass poverty – say 30% or more of their people belowa dollar a day – accept that faster growth of agricultural income, outputand productivity are extremely helpful, and usually necessary, to start asustainable process of mass poverty reduction.3. Do small farms fit processesof initial mass povertyreduction?Suppose we agree with that. Suppose we also believe, for most countriesand regions, that substantial sustainable farm growth acceleration isfeasible. 12 Do the likely processes of achieving this indicate a focus onsmall-scale or large-scale farming? Further, do world-wide processes offarm change – commercialisation of increasing proportions of input andoutput, institutional developments such as supermarkets, privatisation ofkey aspects of technical progress and of output and process grades andstandards – increasingly indicate large-farm focus? I suggest:(2) However, credible processes of initial, rapid mass povertyreduction require rapid growth in demand for productive labour,and in its command over staples. Both requirements indicate asmall-farm orientation, especially in very poor or remote countriesor sub-regions.(3) Processes of farm commercialisation and globalisation makethe small-farm route more difficult if there are substantialintermediation failures between small farms and theemerging customers and institutions of globalising farm change.How might issue (3) be reconciled with (2) in the context of an agriculture-leddevelopment policy? One might identify (a) efficient, affordableshifts in incentives or institutions so as to favour employment-intensityin a large-farm growth path; 13 or (b) efficient, affordable support for, ormarket development of, successful intermediation between small andfamily farms and growing, modernising export or supermarket outlets; or(c) paths by which small farms can efficiently grow without such access.I hypothesise that most rural poor and the world’s ‘smaller half’ of farmsare in substantial food deficit and can expand a good deal while achievingstaples self-sufficiency. 14Much progress has been made on (a), but it requires a higher priority, almostno progress has been made on (b) and modest progress is beginningon (c), with spreading awareness that mass-poverty reducing agriculturalgrowth usually starts with staples for national, even local or domestic, use,not with export horticulture, and that, without decrying the medium-termneed for smallholders to switch towards marketed crops in the mediumterm, ‘staples security first’ may well be the right sequence for smallholdersin much of inner Asia and sub-Saharan Africa.(1) Neither poverty-reducing nor farm-globalising processeseither mandate or exclude small farms as the basis of agricultureled,rapidly poverty-reducing development.12For sub-Saharan Africa and for water-scarce regions there is a contrary view. The issues aretoo complex to discuss here; see International Fund for Agricultural Development, RuralPoverty Report 2001, loc.cit., ch. 4.13If advocates of such a path showed that it might, with stated policies, affordably createdemand for the growing and under-employed poor workforce, it would be easier to takelarge-farm recipes for initial mass poverty reduction seriously. (Much the same can be saidof proposed non-farm-led escape routes). Unemployment data in Africa are notoriouslyfraught, but (for example) International Labour Organisation (ILO) and other enquiriesconfirm genuine time-rates of adult unemployment over 30% in South Africa – and morein rural areas.14The common claim that poor deficit farmers who do not market have no cash to buy inputsand therefore cannot much raise output, is incorrect. By definition such farm householdsbuy staples. If they grow more, they can divert cash (from non-farm activity, hired work onothers’ farms, etc.) from staples purchases to staples-input purchases and will have cash inhand. There may, of course, be a first-year credit problem.
Why do credible processes of poverty reduction strongly indicate smallfarmgrowth? The three arguments for agriculture as the main, normallythe only, feasible source of processes of early, rapid mass early povertyreduction re-emerge as arguments that such processes should have asmall-farm focus. In that form, the ‘small-farm logic’ appears sound.The ‘stylised facts’ behind it are not very controversial. Yet evidence foror against them is surprisingly scarce. So it remains possible that, underconditions to be empirically established, large-farm agriculture might be amajor source of mass poverty reduction.Because the ‘small-farm logic’ is rather strong, I hypothesise that such circumstancesare rare and cover only small proportions of persons in areas ofmass initial dollar poverty; but the issue needs to be more fully explored.The demand-for-labour argumentThat the poor’s income depends largely on labour and that agriculture haslowest costs per workplace and highest average and marginal capital/labourratios in early development in poor regions, is a central argumentbehind the case for agricultural growth as the key to development thatslashes initial mass poverty by bidding up the returns to (employmentand/or wage-rate of) poor workers. This process applies more convincinglyif such development is through small farms.For example, in Pakistan in 1972, farms above 60.7 hectares engaged0.12 workers per hectare and farms of 20.2-60.7 ha engaged 0.22 workers/ha,whereas farms below 0.4 ha engaged 9.15 workers/ha, and farmsof 0.4-1.0 ha 3.32 workers/ha. 15 Data for Bangladesh, Thailand, Indonesiaand India were comparable. Simulations showed that egalitarian redistributionwould raise labour demand and use – by only 9% in Java, butby 19%-24% in Bangladesh, Pakistan, Thailand and the Outer Islands ofIndonesia. 16 Plausible partial land redistribution on Brazil’s estate sub-sectorwould raise person-year equivalents of labour use in agriculture from2.6-million to 3.0-million over the 1978 base case, and World Bank evidencefrom the 1970s showed ‘employment per hectare higher... in thosecountries that have... more equal distribution of land ownership’. 17Most of these data are rather out-of-date. More recent data are scantypartly because higher labour-intensity on small farms, other things roughly15Workers-per-hectare for other size-groups, : 1-2ha: 1.72; 2-3ha: 1.12; 3-5.1ha: 0.82;5.1-10.1ha: 0.52; and 10.1-20.2ha: 0.32 w/ha (A. Booth and R. Sundrum, LabourAbsorption in Agriculture, Oxford University Press, Delhi, 1984: 101).16Ibid.: 100-9, 279-80.17G. Kutcher and P. Scandizzo, The Agricultural Economy of Northeast Brazil, Johns Hopkins,Baltimore, 1981: 201; 37.equal, is not seriously contested, resting not only on widespread observation– some by sceptics who see small farms as sources of ‘self-explotation’– but also on basic transaction-cost theory. It is often objectedthat many small farms owe part of their higher labour/land ratios to theirbetter land or water resources, but this is itself in part a consequence ofearlier choices of higher ratios, used to improve or maintain land/waterresources (terracing, bunding, manuring, etc.), and often done off-seasonwhen family labour can be used to build up ‘labouresque’ capital morereadily on small farms than on large. 18The demand-for-labour argument for small-farm emphasis in poverty reductionshould not rest only, or in some regions mainly, on small farms’greater capacity to generate income-per hectare for self-employed familyfarmers. For instance, in India, hired labour is the main income sourceof more rural households, and many more poor rural households, thanis own-account farming. 19 But world-wide, most evidence is that smallerfarm size goes with higher demand for hired labour per hectare, despitea lower ratio of hired to family labour. Further, better growth options on(or land transfer to) smaller farms provide incentives for farm families towithdraw labour supply, from hiring-out, into the family farm. This raisesthe proportion of work in the hired-labour market available to the landlessand thus their employment and/or wage rate.The reliance of the poor on rising demand for labour for poverty reduction– and also for bargaining power and dignity – is at the core of thecase for agriculture as the leading source of initial mass poverty reduction.Setting agricultural growth mainly into a context of small, labourintensivefarms greatly strengthens that argument, and has been shownby abundant worldwide experience to be feasible. That does not imply,however, that large farms are not suitable for some purposes, even wherecapital/labour ratios are low, nor that large farms cannot be stimulatedto be more employment-intensive and pro-poor. However, opponents ofsmall-farm approaches have so far done little to show where or how thishas been or can be achieved.It is often mistakenly believed that slower population growth weakensthe employment-related case for small-farm-centred poverty reductionprocesses. On the contrary, workforce grows rapidly after child populationgrowth has slowed or stopped, creating a ‘demographic window’ to workone’s way out of poverty – but only if labour demand expands, affordably,at least as fast as workforce, as happened in East Asia in 1960-90 thanks18A.K. Sen, Choice of Techniques (3rd ed.). Cambridge University Press, 1968.19K. Sundaram and S. Tendulkar, The Working Poor in India: Employment-Policy Linkagesand Employment Policy Options, International Labour Office, Recovery and ReconstructionDept (Geneva), 2002: 43.59<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Can Small Farmers Survive, Prosper or be the Key Channel to Cut Mass Poverty?
Development & Poverty Reduction60<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>mainly to small-farm-based growth. The effect of new farm technologiesin reducing poverty incidence in 1965-2000 was much amplified by thefalling dependency ratio. Lower proportions of dependants were supportedby fast-rising numbers of workers, for many of whom the new farmtechnology provided rising employment income. The fertility reductionstriggering this process 20 came somewhat later to South Asia and Africa,but are now sharply cutting dependency ratios for poor countries in theseregions too. In 2000, there were 99 dependants for every 100 people ofprime working age in Ethiopia; the projection for 2030 is 72. For Nigeriathe dependency ratio falls from 99 to 67; for Bangladesh from 79 to 55;and for India from 71 to 58. 21 If physical conditions allow and policiesstimulate small-farm-led poverty reduction in Africa and ‘inner Asia’, demographicsmakes 2000-2030 an ideal time for it. If those chances arenot provided by crop science and appropriate policy, the extra workerswill face downward pressure on rural wage rates or employment, and theopportunity will be lost after 2030-50, 22 as rises in the proportion of over-60s put the dependency ratio into reverse (it starts to rise again).Three challenges to the above ‘employment intensity’ case for small-scalefarming should be noted. First, some regions with high year-round averageunemployment (and low wage rates) nevertheless face extreme seasonallabour peaks, when wage rates rise sharply and/or absolute labourscarcity (in some sense!) persists at low wage rates. While research, extensionand water policy should certainly respond to such conditions, they donot justify (say) shifting agriculture, or its growth, towards larger and lesslabour-intensive farms.Second, large rural regions are afflicted by HIV/Aids, reducing effectivelabour availability. However, it cannot ameliorate this temporary thoughtragic demographic situation by policies to reduce demand for and hencewage rates of labour, for example by shifting agricultural growth awayfrom small employment-intensive units.Third, artificial stimulation of farm output in OECD countries has inducedtechnical change, and basic science, favourable to farms of ‘OECD’ sizeand factor endowments, and thus to labour replacement in farms of allsizes, even in poor countries with plenty of labour per unit of capital andland. This, however, is anyway reflected in factor markets, and hence farmsizes, in developing countries. We shall see that these have neverthelessfeatured steady increases in the proportion of land in small farms. It is unappealingto suggest that developing countries should respond to risingworld capital/labour ratios – given that lower ones will anyway be chosenby small farms than by large ones – with policies that increase marketincentives to shift land towards large farms.The assets argumentIn China and Vietnam, and to a somewhat lesser extent in most of the restof Asia and of central and western Africa, a large majority of the rural poorhave, at least, usufruct rights to enough farmland to provide a significantproportion of household income.The absolute value of these rights has been reduced over time by decliningfarm size, but except in Africa this has been more than offset atnational level, and in most regions, by offsetting rises in productivity offarmland. 23 Even where severe land inequality prevails, as in Southern andparts of East Africa and most of Latin America, many or most of the ruralpoor derive significant parts of income from owned, communal or rentedfarmland.Therefore, in most areas where widespread poverty prevails, anti-povertypaths via enhancing the physical assets of the poor, as well as via employmentincome and food entitlements, direct policy towards increasingthe resources and technology base – and the land base – of small farmsrather than large ones. Policies to enhance the poor’s farm machinery orlarge stock run against the fact that in most countries the dollar-poor owna smaller proportion of such assets than of farmland, for good reasons (divisibility,collateral, risk). Small stock assets are often distributed at leastas equally as land, and therefore pro-poor policy here has some leverage,but the absolute value of the poor’s small stock is much less than that oftheir farmland. The poor do normally have some housing assets, but theseare seldom (perhaps too seldom) seen as a part of pro-poor productionpolicy.The staples argument20Strictly, the process starts with big infant mortality falls during 1945-60 as malaria iscontrolled and nutrition improved. This first raises dependency ratios, but as the ‘saved’infants age into the workforce, the process slows down and then reverses. Later, fertilitydecline strengthens the reversal (fall in dependency ratios).21These data allow for HIV/Aids. It hits mainly (i) persons aged 15-30 and (ii) infants, withoffsetting effects on dependency ratios.22The dates of the turning point vary by country, as did the earlier changes that set the processgoing.Food staples comprise a substantially higher proportion, by value, of outputgrowth on small farms than on large ones. So farm growth is likelierto restrain and stabilise the price of staples if it is concentrated on23Falling world staples prices since 1980 reduce the value of this offset, but were themselvespartly offset by falling domestic policy bias against staples prices. Remaining net price fallsare less harmful for food deficit farmers, including most small farms.
small farms. The dollar-poor spend a considerably higher average incomeshare (typically well over half) on food staples than the non-poor. The poortherefore acquire a higher proportion of consumer gains, from a givenabsolute value of extra farm output, if it is concentrated in smaller farms.Also, the poor have a higher marginal propensity to consume staples;therefore, as the poor’s money income rises, their gains are increased tothe extent that extra food production comes from small farms with higherstaples focus. Freeing of global food markets (and large public foodgrainreserves, for example in India) may moderate the impact of domestic staplesoutput growth – and hence of the small-farm share of such growth– on prices facing the poor, but only to a modest extent due to the highratio of staples transport costs to production costs.The facts behind this logic, while perhaps not very controversial, are hardto verify.(a) I can find no database, at FAOSTAT, the World Bank or elsewhere,collating estimates (let alone those from reliable nationwideand/or successive, household surveys) of average or marginalstaples/consumption (or even food/consumption or food/income)ratios by income group. 24(b) There is limited evidence that staples/production ratios rise asfarm size falls. It is familiar that labour-intensity of crop-mix rises,but that does not help much, because labour-intensity in staplesis intermediate between that in other main land uses – morethan in pasture/grazing, trees and most fruit crops; less than inbeverage crops, cotton, rubber, sugar and most vegetables. However,bitty data and ‘anecdotal wisdom’ that smaller or poorerfarmers are likelier to devote given land to staples is supportedboth by food security considerations (absolute aversion fromhunger risk increases with the level of that risk) and by the factthat, in raising staples output, surplus farmers increase, but netfood-buying farmers reduce, marketing transaction costs. 25(c) That the rural (urban) poor are likelier than the rural (urban) nonpoor– and rural and remote people (with higher poverty incidence)than urban people – to source staples from own productionor locally and to face high unit cost of staples transport anddistribution from central stores or ports seems self-evident, butthe evidence again is thin. It is important to pinpoint exceptionsto any general rule that the poor are so placed, and therefore less‘delinked’ than others, by easy access to local or global staples24Grouping by income-per-household is useless. Income per consumer unit or adult equivalentis preferable to income per person.25Price risk, too, rises with rising output of a main, locally consumed staple for surplus farmers(deterring its production) but falls for deficit farmers (encouraging them to grow more ofthe staple): C. Barrett, ‘On price risk and the inverse farm size-productivity relationship’, J.Development Economics, 51, 1996: 193-215.markets, from any effects of farm size on the price or reliability oftheir consumption due to local supply.Crude absorption:A related point is that countries with initial, largely rural, mass povertypresent a further crude demand-side case for concentrating agriculturalgrowth on small farms. Such growth evades the demand/marketing problemto some extent, because much of small farms’ extra output comprisesstaples consumed locally – often in the farm household, or by (extra) farmlabour. The marginal propensity to consume farm products is high forsmall farmers and farmworkers, but low for large farmers and owners ofsubstantial land or farm capital. Hence, for example, in India there is muchpotential for poverty-reducing growth through expanded, labour-intensivestaples production on small deficit farms in the poorer states, but little,if anything, is gained when well-off surplus farmers in the Punjab (aresubsidised to) grow extra wheat and rice capital intensively. Absent muchdirect demand created by the incomes corresponding to the extra output– well-off farmers and tractor owners do not use much of their extraincome to buy food – it is then sold to the Food Corporation of India forlong-term storage (and deterioration); India’s poor have little extra incomewith which to buy such grain and export markets are protected and/or, forlarge sellers, price-inelastic.Agricultural priority and consistency of reducing poverty in allmain poor groupsOne criterion for good pro-poor policy is that it should not set largegroups of the poor against one another by achieving poverty reductionat the cost of creating many losers who are ‘stuck’ in a particular group.The conditions for growing agricultural income, output and productivityto benefit all main poverty groups – those whose income derives mainlyfrom small-scale farming, from rural labour or non-farm activity, and fromurban work – are in fact rather tight. (We were fortunate that the greenrevolution largely met these conditions.) In particular agricultural growth– for example, via new science or applications of existing science – has towalk two tightropes, and this is much easier if the growth is concentratedon small farms.The price/total-productivity tightropeTo help poor food consumers and poor farmers as marketers respectively,farm growth must cut staples prices, but must raise total factor productiv-61<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Can Small Farmers Survive, Prosper or be the Key Channel to Cut Mass Poverty?
Development & Poverty Reduction62<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>ity on small farms faster. If farm growth is concentrated on small farms,much of it comprises extra staples grown by poor households with a highmarginal propensity to consume them. Prices are still pressed downwards(the marginal propensity is less than one), but by less than if farm growthis on big farms in staples surplus, where farmers’ extra self-consumptionout of income is small and anyway can form only a small part of extraoutput.The wage rate-labour/land-productivity tightropeIn early development out of mass poverty – in the increasingly universalcondition that unused farmland is absent or very costly to develop – forfarm growth to help poor farmers as employers and poor farm labourers 26respectively, it must raise output per labour-hour, but output per hectaremore. Such concentration on employment-intensive extra production – especiallyproduct-mixes but also product-specific technical choice – is muchlikelier on small farms, for familiar transactions cost reasons.The case for small farms self-destructs, but only with success. It is thenthe very arguments for agricultural growth as the main channel of initialmass poverty reduction that point to small, not large, farms as the maincarriers of such growth. Agriculture within mass-poverty economies– and for the same reasons, small farming within agriculture – is the mostfavourable sector for enhancing employment-intensity, (land) assets of thepoor, (staples) consumption of the poor, demand absorption of the extraoutput and consistency of progress among main poverty groups.And just as the case for agriculture as the source of poverty-reducinggrowth, if successful, self-destructs, so does the case for small farms. Incountries or regions that have succeeded in slashing mass dollar poverty,the above arguments for concentrating farm growth on small farms areweakened. On equity grounds, major initial success against mass dollarpoverty induces urbanisation; shifts in rural and urban economic structuretowards non-farm work and urban life; and demographic declinesin workforce growth, even for the remaining dollar-poor. All three trends,by cutting the dependency of labour on agriculture, shift the emphasis ofanti-poverty policy away elsewhere, and hence away from farm size. Onefficiency grounds, initial success brings rising capital/labour ratios, so thattransaction-cost reductions become more important for costs associatedwith capital acquisition and use, as opposed to those associated with laboursearch, screening and supervision (for costs readily reduced by largeunits rather than small ones. Both sorts of change shift the emphasis fromsmaller to larger scales of farming.26The off-farm poor also suffer (wage rates off-farm also fall) if this condition is not met.However, countries and regions where initial development has been attemptedwith land heavily concentrated into large farms – those with veryunequal farmland, as in much of Southern and some of Eastern Africa andmuch of Latin America – move from low-income to middle-income status,if at all, with severe, and still largely rural, unemployment and dollar povertylevels far ahead of what is normal for their levels of PPP-GDP per person.There has not been ‘success’, even in initial mass-poverty-reducinggrowth, and hence the case for small-farm-based development has notself-destructed. It remains to be seen whether this can be corrected withoutrenewed emphasis on small-farm development and asset acquisition.The countries, chiefly Asian, that have achieved initial success in sharplyreducing mass poverty through farm growth – while they have largelymaintained, or even accelerated, growth as it shifted to the non-farmsector – have been much less successful in maintaining the rate of povertyreduction. The responsiveness of poverty reduction to economicgrowth in China and India, for example, has been much less since the mid-1990s than was the case in 1970-85. Further, middle-income counties inSouthern Africa and Latin America, despite quite low shares of agriculturein workforce and GDP, have been rather unsuccessful in bringing downmass dollar poverty, which is far above levels predicted from global regressionsagainst PPP income-per-person. Even after agriculture’s share ofworkforce has fallen below 20%-25%, inequality of farmland remains astrong predictor of overall inequality and absolute poverty – and presumablya fortiori, of weak transmission of growth into poverty reduction. 27So processes of mass poverty reduction through agriculture favour smallfarms (as more labour-intensive, as providing asset income to the poor andas supplying locally available food staples), but do not mandate them.While initial success in economic growth and/or poverty reduction clearlyreduces the role of farms, small or large, in carrying success further, experiencein the successful developing countries suggests big knowledgegaps about how to continue, or revive, rapid poverty reduction that is notbased on small-farm growth. More important, initial success in reducingmass poverty has not been achieved in large parts of the developingworld – and the historical evidence is that it cannot be achieved withoutprior agricultural development, save in special cases such as city-statesor (unusually well handled) mineral discoveries. But can that agriculturaldevelopment be attained where it has not been attained so far? And canor should it be on small farms?27A. de Janvry and E. Sadoulet, ‘Growth, poverty and inequality in Latin America: a causalanalysis’, Review of Income and Wealth, 46 (2000): 267-87.
4. Can, should and will smallfarms survive, grow and cutpoverty?All the above arguments would not ensure small farms’ survival, let alonetheir growth or contribution to poverty reduction, if they were competitivelyinefficient. The first issue is whether small farms have survived– have shrunk or increased their share of farmland – in developing countries.We measure smallness by land area, rather than gross or net outputor labour input. The policy issue is what farm sizes to encourage (or stopdiscouraging) by public action, including price policy, incentive-compatibleland reform and/or tenancy policy.Note that large households tend to operate more land, and size inequalityis usually much less per person than per household. In a study in theIndian Punjab, the per-person farmland Gini was half the per-householdGini. 28 It also matters whether land/water quality is factored in. In India(but often not elsewhere; certainly not in Southern Africa), smaller farmstend to have better-quality land/water endowments, so the Gini is muchsmaller if farmland is quality-adjusted. 29 However, “discussions of empiricalfacts are driven by the available comparative data. In the FAO farmcensuses, land area of holdings is available for most countries, but noother potential measure of scale is widely available”. 30Thus holdings with low area are ‘small farms’, given our concerns and thedata limitations. In deciding whether small farms are competitive or efficient(not quite the same thing) in a country, the first task is to see whetherthe proportion of farmland, found in small farms, is rising (falling). If so,it suggests either, dynamically, that small farms are becoming relativelymore (less) efficient or competitive, or that farmers are steadily findingways to approach a static farm-size optimum by evading or avoiding lawsor incentives impeding a move to optimal farm size. Suggestion is notproof; perhaps small farms’ proportion of area fell because, for example,new laws, taxes or price manipulations, artificially favouring small farmsagainst large ones, were introduced or old ones enforced, even thoughnot incentive compatible. However, such enforcement is usually very difficultin the long term. It seems likely that in a country or sub-region withsubstantial or steady falls in farm size (rises in the proportion of landoperated in small farms), small farms were either statically more efficient(so land was being shifted into them by market forces such as sale andtenancy and/or by land reforms which ‘stuck’ ex post in the marketplace),or were dynamically becoming relatively more efficient.Alternative explanations, commonly advanced for falling mean farm size,are unconvincing. Rural population growth, with non-partible inheritance,increases the number of owned farmholdings. However, if larger farm sizeis efficient (or becoming more so), one would expect many new ownersinheriting increasingly subdivided and tiny holdings to sell, rent orotherwise transfer and amalgamate them, so that land was increasinglyoperated in large holdings.Technical progress in agriculture decreases the operated area from whicha given household can achieve a given total household income (or a givenproportion of it), but that does not explain why land owners and operatorsmight increasingly choose smallness, unless it is efficient, or becomingmore so. Of course land and other factor markets work imperfectly, andland laws constraining tenancy, sales or farm size may be (partly) enforced.However, it is implausible – outside a Stalinist system – that low orfalling levels of small-farm efficiency relative to large farms are consistentwith long-term high and rising proportions of farmland in smallholdings.Tables 1 and 2 show such high and rising proportions in a very largemajority of developing countries. Table 1, based on data so far releasedfrom the FAO’s 1990 and 2000 World Censuses of Agriculture, covers allcountries with such data, with over 100,000 ha of farmland, and withmore than one set of estimates of distribution of holdings and farmlandby farm size groups between 1985 and 2003. In this period, acceleratedglobalisation and commercialisation were allegedly shifting competitiveadvantage, even capacity to survive, from small to large farms. Yet in Table1 Part A – for only eight developing countries, but including some withextensive agricultural area (Ethiopia, India, Pakistan and Thailand) – sevencountries show a rise in the proportion of farmland in the lowest two sizecategories of operation. 31 The changes in the proportions of area (andholdings) in different farm-size groups in these eight developing countriessince 1985 are prima facie evidence against the view that globalisation,or anything else, made small farms less competitive or survivable. Notethat the limited data for (five) developed countries show, if anything, declinesin the proportion of farmland in the smaller size groups.63<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>28A. Julka and P. Sharma, ‘Measurement of land inequality in India: a revision of theLorenz-Gini ratio’, Indian J. Agricultural Economics, 44, 4, 1989.29S.Bhalla and P.Roy, ‘Mis-specification in farm productivity analysis: the role of land quality’,Oxford Economic Papers 40, 1988: 55-73.30Eastwood, Newell and Lipton (forthcoming), loc. cit.31The exception is Ethiopia, where the proportion of (private) land in holdings below 1 hafell from 1989-82 to 2001-02. Continuing redistribution of collective and State landsafter the 1989-92 Agricultural Census (in conjunction with population growth) raised privatefarm area 2.3-fold, but the number of holdings only 1.8-fold, so that many holdingsin the smallest size category could be enlarged. Even though in Ethiopia, the proportion ofland in holdings below 5ha fell between successive agricultural censuses, the falling proportionof land in holdings below 1 ha was more than offset by the rising proportion of land in1-5 ha holdings.Can Small Farmers Survive, Prosper or be the Key Channel to Cut Mass Poverty?
Development & Poverty ReductionTABLE 1: SMALL AND MEDIUM FARMS: AGRICULTURAL CENSUSES FROM 198564<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>(Countries with > 100,000 ha farmland & censuses in 1990 and 2000 rounds with % area and holdings by size groups: FAO website)Part A: Africa, Asia and Central and South AmericaCountry Year Holdings (m) Ha (m) < 1 Ha 1–2 Ha 2–5 Ha 5–10 Ha 10–20 Ha%hdgs %ha %hdgs %ha %hdgs %ha %hdgs %ha %hdgs %haColombia 88 1.45 36.03 14.1 0.3 21.5* 1.4* 13.0* 1.9* 6.0 4.4 12.6 7.001 2.02 50.71 18.1 0.4 23.0* 1.7* 11.7* 1.8* 14.4 4.0 11.1 6.2Egypt 90 2.91 3.30 60.6’ 18.5’ 29.3’ 30.4’ 6.8, 15.9, 2.1, 10.1, 0.9| 9.8|99/00 3.72 3.75 81.1’ 33.5’ 13.9’ 24.0’ 3.3, 13.2, 1.2, 9.9, 0.5| 8.8|Ethiopia 89/92 6.09 4.87 72.1 36.9 20.2 25.2 7.4 25.4 0.3 2.2 .. ..01/02 10.76 11.05 62.8 27.1 24.3 33.3 11.9 32.6 0.9 5.5 .. ..India 86 97.16 164.56 57.8 13.4 18.4 15.6 13.6” 22.3” 8.1” 28.6” .. ..91 106.64 165.51 59.4 15.0 18.8 17.4 16.8 30.9 4.4 19.3 .. ..95/96 115.58 163.36 61.6 17.2 18.7 18.8 14.8 31.5 3.7 17.7 1.0 9.2Nepal 92 2.74 2.60 69.8 30.5 19.4 27.6 9.4 28.0 1.2 8.1 .. ..02 3.34 2.65 74.7 <strong>38</strong>.9 17.6 29.8 6.9 24.0 0.6 5.3 .. ..Pakistan 90 5.07 19.15 27.0 3.7 20.4 7.6 37.5 27.6 12.3 21.6 4.7 1 5.800 6.62 20.41 36.1 5.8 21.5 9.7 28.1 27.9 8.8 19.1 3.9 1 6.3Panama 90 0.21 2.94 46.7 0.5 11.4 0.9 13.5 2.7 7.6 3.5 7.1 6.701 0.24 2.77 52.7 0.6 .. .. .. .. .. .. .. ..Thailand 88 4.88 17.46 14.4 2.5 12.3^ 4.2^ 59.7^ 54.6^ .. .. .. ..93 5.65 19.00 19.7 3.0 13.2< 4.6< 45.0< 36.1< 17.2< 32.0< .. ..Turkey 91 3.97 23.45 15.9 1.4 19.0 4.3 32.1 16.5 18.0 19.0 9.7 21.001 3.02 18.43 15.5 1.3 17.9 4.0 31.5 16.0 18.5 20.7 10.8 23.5Uruguay> 90 0.05 15.80 > > 8.1 -------------- 0.1 12.1 0.3 13.2 0.600 0.06 16.42 > > 10.9 -------------- 0.1 12.5 0.3 12.5 0.6Notes: Holdings without farmland (Colombia 1988; Egypt, Nepal 2002; Panama 2001), area in them (Egypt) and government holdings (Pakistan) omitted. ‘m’ = million.*: 1-3 ha and 3-5 ha, not 1-2 ha and 2-5 ha ’: < 0.8 ha and 0.8-2.1 ha ,: 2.1-4.2 ha and 4.2-8.4 ha |: 8.4-21 ha ”: 2-4 ha and 4-10 ha ^: 1-1.6 ha and 1.6-6.4 ha: excludes holdings below 1 haTable 2 allows a longer view, for more developing countries, over AgriculturalCensuses from the 1970, 1980 and 1990 rounds (in practice, from1969 to 1993). These data are not yet available for the 2000 round ofagricultural censuses, so they may exclude the impact on farm size post-1990, accelerated farm commercialisation and globalisation, but Table1 showed that this was at least consistent with falling farm size. Care isneeded in interpreting the data – wars and other sources of non-comparabilityabound – but the message from two columns of Table 2 is unmistakeable.‘Median size for number’ is the size of the median holding, withall farms ranked in order of area. If instead we rank all farmed hectares,starting with each of the hectares in the largest holding and ending withthose in the smallest, ‘median size for area’ is the size of the holding inwhich the ‘median hectare’ is located. The vast majority of successive AgriculturalCensuses in both Asia and Africa between 1960 and 1990 showsteep or very steep falls in both median farm size for number and medianfarm size for area. Thus a large and growing proportion of farmland isbeing cultivated in small holdings, and a large and growing proportion offarm operators is small. 32 Why is this happening?First, there is no production-related reason why it should not. There isneither theory nor a weight of evidence from developing countries thatsuggests any of the following: economies of scale (output rising morethan 1% when all producer inputs rise 1%), unit production costs fallingas farm scale increases, or falling unit production costs as farm size (area)rises. A few studies in developing countries find significantly non-constantreturns or costs, but in most such cases either returns increase (or costs32This is in sharp contrast to the trends in developed countries shown in Table 2. There thesize of farm containing the ‘median hectare‘ is not only larger than in developing Asia orAfrica, but (in contrast to these) is falling in most cases, often sharply.
TABLE 2: SIZE (HECTARES) OF (A) MEDIAN FARM AND (B) FARM WITH MEDIAN HECTARE: DEVELOPING COUNTRY TRENDS 1Country Date Med. farm Med. ha Country Date Med. farm Med. haAfrica(Dvpg. Asia)DRC 1970 1.2 1.8 Turkey 1980 3.6 13.01990 0.39 0.76 1991 3.0 13.0Ethiopia 1977 1.0 2.31989/92 0.54 1.3 S/Cent. AmericaLesotho 1970 1.5 2.6 Brazil 3 1970 0.4 5201990 1.1 2.4 1980 9.8 730Malawi 1969 1.2 2.1 1985 8.6 6701993 0.52 1.8 Panama 1971 3.6 861981 1.7 95Dvpg. Asia 1991 1.2 110India 1971 0.98 5.5 Paraguay 1981 8.2 ---1977 0.85 4.8 1991 6.9 ---1991 0.74 3.4 Peru 1972 1.8 ---Indonesia 1973 0.56 1.8 1994 2.5 ---1993 0.54 1.8Korea, Rep. 1970 0.71 1.21980 0.75 .811990 0.81 1.4Nepal 1972 --- 2.41982 0.49 2.81992 --- 1.6Pakistan 2 1980 2.9 7.81989 2.1 7.2Thailand 1978 2.7 5.81993 2.4 5.5Source: FAO agricultural censuses, rounds for the 1970s, 1980s and 1990s, at www.fao.org/es/ess/census/gini/table2.asp65<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Notes: Countries with less than 25,000 hectares of farmland are omitted. The column headed ‘Med. farm’ shows the ‘median size for number’, that is, hectare size of the median farm rankedby size. ‘Med. ha’ shows the ‘median size for area’, that is, hectare size of the farm containing the ‘median hectare’ of farmland, with hectares ranked in order of the size of the farm where theyare found.Notes (from original FAO source):1Includes holdings without land [usually zero or very few]. Includes only countries with data for the 1990 census round and for the 1980 and/or 1970 census rounds.2Data exclude 149 Government holdings with 103,035 ha for 1989 and 192 Government holdings with 49,995 ha for 1980.3Due to lack of data for 1986–95, data from the 1985 Agricultural Census are presenteddecrease) uniformly as farm size falls, or there are increasing returns tofarm size only up to a very small threshold level of holding size (say 0.5haor 1 ha) and mildly decreasing returns thereafter. Perhaps most important,a large balance of evidence, though not all, favours the existence of an ‘inverserelationship’ between farm size and yield-per-hectare in developingcountryagricultures. 33 Though this predicts, and in some sense ‘justifies’,the observed shift towards smaller farm size – especially where muchland is concentrated in large farms – it tells us nothing about efficiencyor economies/diseconomies of scale, any more than does the similarlylarge balance of evidence of a direct relationship between farm size andyield-per-labour-hour. For reasons discussed in the next paragraph, smallfarms are choosing to farm given areas with more labour and less capital33H. Binswanger, K. Deininger & G. Feder in Handbook of Development Economics, <strong>Vol</strong>.IIIB (ed. J. Behrman & T. N. Srinivasan), North Holland, 1996; IFAD (2001), ch. 5;A. Berry and W. Cline, Agrarian Structure and Productivity in Developing Countries,Baltimore: Johns Hopkins, 1979; Booth and Sundrum, op. cit., 1985; Eastwood, Liptonand Newell, forthcoming Lipton, World Development, 1993.Can Small Farmers Survive, Prosper or be the Key Channel to Cut Mass Poverty?
Development & Poverty Reduction66<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>than otherwise similar large farms. Given that both large and small farmsselect from the techniques and crop-mixes likely to be efficient in developingrural areas, this means higher average yields (and probably averageproductivity of capital) for smaller farms, and higher average labour productivityfor large ones.Second, however, there is a credible account of transaction costs that explainsthe evidence in the above tables, and predicts small farms in developingcountries (and large farms in developed countries) (see Eastwood etal.). Unit transaction costs associated with labour search, supervision andscreening normally decrease as farm size falls, because household/familymembers are a larger part of the workforce, because there are morehousehold/family members to supervise each unit of hired labour input, 34and because the farm operator has a smaller space over which to supervise.On the other hand, unit transaction costs associated with capitalacquisition and use decrease as farm size rises: a loan of US$1,000 costsless to negotiate than 10 loans of US$100, and a tractor is somewhatcheaper to manage and schedule on a large holding than on many smallones.Therefore, where incentives lead farmers to choose high labour/capitalratios – in developing areas with initial mass poverty and, increasingly,land and/or water constraints in developed countries – smallness hasnet advantages because it cuts transaction costs associated with labour.Where farmers choose low labour/capital ratios – in developed countrieswhere labour, especially rural labour, is more costly relative to capital, sothat cutting unit transaction costs associated with capital matters morethan cutting those associated with labour – largeness has net advantagesbecause it cuts transaction costs associated with capital. This is consistentwith the data in tables 1 and 2, and helps account for the tendencyof farm size to be higher in developed than in developing countries. Oneneeds to factor in the effects of (a) ‘colonial land grab’ in creating and perpetuatingabove-optimal farm size in Latin America, Southern (and partsof Eastern) Africa and plantation regions of the Caribbean and small partsof rural Asia, and of enforced State and collective farming in (more briefly)over-enlarging farms in communist agricultures; (b) land reform, partlyto correct this, reducing median farm size in big areas of Latin Americaand Asia (and in transitional economies) – an effect far larger than suggestedby popular accounts of the evasion and general unsuccess of suchreforms.34M. Taslim, 1989. ‘Supervision problems and the size-productivity relation in Bangladeshagriculture’. Oxford Bull. Economics and Statistics, 51.A further transaction-cost effect, favouring small farms in early developmentand large farms later, concerns the transaction cost of disposing ofoutput. There are economies of scale, off-farm, in processing and transportingcash crops. These need not be translated into much higher distributioncosts for small farms, but there are costs of transactions and/or intermediationif it is to be avoided. On the other hand, if self-consumptionis a large part of total farm output, small farms have several transactioncost and associated advantages over large ones. In farming, a householdwith a small farm escapes – as a large commercial farm operator doesnot – costs of purchasing much of its own farm needs, and also costs ofmarketing its product, in both cases especially if that product is a staple,where transport costs for marketed products are typically large relative tofarm value added. Large farms also incur price risk, as self-consumers donot, or to a much smaller extent.More subtly, as hypothesised by Srinivasan and later established empiricallyby Barrett, staples-deficit farms reduce price risk by expanding production(because expansion cuts the effect on them of consumer pricerises), while staples-surplus farms increase that risk (because expansionraises the impact on them of producer price falls). These effects rendersunit transactions costs likely to rise with farm size in rural areas whereself-consumption is a big part of output, and to fall with farm size wheremarketings are overwhelmingly predominant in output – just as differentialfactor-specific transaction costs render total unit transaction costslikely to rise with farm size in rural areas where conditions favour techniquesand crop-mixes with high labour/capital ratios, and to fall withfarm size where low ratios are favoured.But do liberalisation/globalisation processes undermine the pressuresto small farms from transaction-cost theory and poverty-reduction processes?The data of Tables 1 and 2 suggest, but do not prove, that thisundermining has, at least so far, not been predominant. However, it is notabsent. The new institutions of farm-product exchange – supermarkets,national and foreign; grades and standards, often private ones, affectingprocess (for example, pesticide application and child labour) as well asproduct; and large horticultural export buyers – tend to favour overview,bulking-up and processing at a fairly large scale.Such institutions have spread since the mid-1980s in the developing world,far faster than historical precedent from the developed world would haveled us to expect – first in Latin America and South East Asia, but laterin China, Southern Africa and recently some other parts of Africa and ofSouth Asia. Some of these places include areas of mass dollar poverty.Do these new institutions of globalisation seriously impede a small-farm
path to poverty-reducing agricultural development? Many papers identifypossible paths, and point to success stories, though stressing the difficulties.The great successes in (for example) micro-scale sweet limes andpomegranates, following the spread of micro-drip irrigation in WesternIndia, is only the most recent of many illustrations that, with appropriateinitial information and support, extremely small farmers can hook onto,and meet standards for complex and distant markets. The problems of intermediationbetween small farms and supermarkets, export horticulture,etc. are perhaps not so different from past, and often solved, problems ofintermediation between small growers of tea, sugar, rubber and cottonand large processors, marketers and final users. 35These intermediation problems do, however, threaten the advantage ofsmallness in agricultures with high labour/capital ratios. That advantagecan be outweighed by new requirements (for example, grades and standards)for selling to wider markets, in commercialised or globalised agriculture.Such requirements create, for the (wholesale/supermarket) buyer, anextra unit cost of ensuring, say, non-excessive pesticide application, or forabsence of child labour (for example, to meet supermarket standards).The extra cost to the buyer of achieving a given degree of confidencethat, say, pesticides use on 100,000 kg of a vegetable is not above acertain level is normally less if the purchase is from one big farm ratherthan 100, let alone 10,000, small ones. That excess may be more than thecost reduction from small-farm procurement due to the large farm’s extraunit cost of labour supervision in direct production processes (ploughing,weeding, water management, harvesting and so on). Higher standardsensuring transaction costs to a buyer can outweigh the lower unit costs ofthe small/family units in supervising direct production.There are two ways to tackle this in modern enlarging, liberalising andglobalising farm markets – and both call up analogies to earlier concernsthat small farmers would be unable to compete in crops such as tea, rubberand sugar, requiring standardisation of product quality and timing byfarmers for synchronised and rapid crop collection and processing by orfor wholesalers.35T. Reardon, J. Berdegue & J. Farrington, ‘Supermarkets and farming in Latin America:pointing directions for elsewhere?’, Perspectives, 81 , Dec. 2002; T. Reardon, T, J-M.Codron, L. Busch, J. Bingen & C. Harris, ‘Global Change in Agrifood Grades andStandards: Agribusiness Strategic Responses in Developing Countries’, International Foodand Agribusiness Management Review, 2(3), 2001; T. Reardon, S. Rozelle, P. Timmer& Honglin Wang, ‘Emergence of supermarkets with Chinese characteristics’, DevelopmentPolicy Review, 2004; T. Reardon, P. Timmer, C. Barrett & J. Berdegue, ‘The riseof supermarkets in Africa, Asia and Latin America’, American J. Agricultural Economics,85, 2003; Binswanger, Deininger and Feder 1996, loc. cit.; IFAD 2001, loc. cit., ch. 5;Creating New Markets for the Poor with Micro-irrigation Technologies in Maharashtra,India: Final report to USAID, International Development Enterprises (India), New Delhi,2004.IntermediationAs has happened with tea, cotton, rubber and sugar over huge areas,intermediaries can find it profitable to check that grades and standardsare attained by smallholders’ products (with otherwise lower direct unitcosts), and then to bulk up approved products for sale to the large buyer.InternalisationAs happened with micro-irrigated horticulture in Western India, smallholdersmay be able to acquire and internalise, on terms competitive withlarge farmers, the information, incentives and methods to supervise thenew requirements of buyers for larger markets, so that buyers can checkthem with a lighter (and cheaper) hand.With internalisation, new markets with their associated standards can inprinciple favour smallness: pesticide and size control could have lowerunit costs if done by motivated family farmers or closely supervised hiredworkers on small farms, rather than by employees and/or machinery orchemical tests applied by buyers or large farmers. With intermediation,new markets need not disfavour small farms; intermediation costs can beless than small-scale supervision savings.While there are many instances of successful internalisation and/or intermediation– allowing smallholders to remain competitive in wider marketsas supermarkets, grades and standards, and wholesaling to exporthorticulture spread – in East and South Asia (especially China and India),this appears to be much rarer in Latin America and sub-Saharan Africa,especially for fruit and vegetable markets (Reardon et al.). In some LatinAmerican cases such as Chile, this may be because development has alreadyraised capital/labour ratios enough to undermine the advantagesof small farms.This is unlikely to be the explanation in sub-Saharan African countries,which are mostly starved of capital and with abundant labour at verylow productivity or even unemployed. This may also be, or may become,relevant for smallholder involvement in poverty heartlands in South andInner Asia, if there is a real prospect to enter into wider markets for cashcrops.The death of small farms has often been predicted. The evidence fromthe only source of large-scale, comparative data – the FAO World Censusof Agriculture – is that, if anything, small farms are occupying increasing67<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Can Small Farmers Survive, Prosper or be the Key Channel to Cut Mass Poverty?
Development & Poverty Reductionshares of farmland in developing countries, even during the period of (andin the areas exposed to) intense liberalisation and globalisation. Yet someof the institutions that grow alongside liberalisation and globalisation,unless accompanied by internalisation or intermediation, do reduce andcan reverse the competitive advantages of small farms.This extends well beyond pure public goods. It includes (at least initially)support for the institutions to intermediate between smallholders and thelarger national and global economy, and to internalise with smallholdersthe information and training required. It also includes agricultural research,roads and irrigation in many cases.68<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The policy issue is what governments in developed anddeveloping countries can and should do about this. For developedcountries, accelerated reduction of farm tariff and subsidy andother support is the obvious answer. It is less obvious that thiswill not favour large (as against small) farmers in developingcountries. I know of no theory or evidence that explores this.However, theory generally concludes that liberalisation andglobalisation, by raising demand for products incorporating acountry’s relatively plentiful factor, shifts income to labour – andto small farmers with higher labour/capital ratios – in developingcountries. For this reason (and because of the effect on thosewith high ratios of food to total consumption, normally the poor)I am doubtful about the suggestion, on its surface attractive, thatdeveloping countries should protect their agricultures againsteven subsidised competition from the OECD. A better optionis to provide far more rural and agricultural goods that areundersupplied by the market.That seems a preferable option for developing-country governments facedwith OECD countries that preach (and impose) liberalisation and globalisation,but practise the opposite. In such a context, however, even thoughsmall farmers have shown they can survive and grow, such a second best,for the world’s poor, is very second indeed.We also may need to ask: what are (region-specific) conditions where,to cut mass poverty via small farms, ‘best strategy’ is extra productionfor (a) ‘subsistence’/extended subsistence, (b) local or nearby markets, (c)urban domestic markets, (d) exports; and extra production of (A) mainstaples, (B) fruit and vegetables, (C) beverage and fibre crops, (D) animalproducts?In practice, policy-makers (and the set of small farmers and many individualsmall-farm households) mix strategies, and answers to the ‘for’ and‘of’ questions interact, but – except that (a) induces (A) – most combinationsare feasible. The outcomes, and the correct emphases, depend partlyon evolving global and national market incentives and transport costs,but also on public policy, including both macro-policy and policies on irrigation,research, transport and regional allocation of infrastructure andcurrent public spending.
development& povertyreductionTRADE AND POVERTYIN SOUTH AFRICA:THE ROLES OF THE EXCHANGE RATE,TARIFFS AND IMPORT PARITY PRICINGFOR CPI INFLATION69<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>BY ERIC SCHALINGSOUTH AFRICAN RESERVE BANK CHAIR, UNIVERSITYOF PRETORIA And Centre for ECONOMIC RESEARCH,1. Introduction 1TILBURG UNIVERSITYIN THIS PAPER WE ANALYSE how poor households’ real incomes are affected by internationaltrade. Here we study how the relevant consumption price index is influenced bytrade liberalisation and/or the opposite, namely protectionist trade policies. 2One of the expected benefits of trade liberalisation is increased levels of competitionbetween import substitutes producing industries and foreign exporters as the tariffs on the latter’sproducts fall or are phased out altogether.To evaluate the scale of the expected price effects, it is first important to identify the goods consumedby the poor (or key inputs in the domestic production process) and the extent to which1The views are those of the author and not necessarily those of the South African Reserve Bank. A previous version of thispaper was written in the context of the <strong>Trade</strong> and Poverty project. The latter was commissioned by the dti and partly fundedby USAID, DFID and TIPS. The project was managed by Matthew Stern (Sega II, USAID and DNA) and LawrenceEdwards (UCT). The authors thank both of them for helpful comments.2The real consumption basket will be equal to C/p , where pccis the price index associated with the household’s consumptionbasket. Other parts of the project look at how the numerator of this ratio is affected by trade liberalisation, via, for example,effects on economic growth and employment.This paper’s main conclusion is that protectionist policies– the presence of tariffs on imported goods – are very badnews for the consumer in general, and the poor consumerin particular. First, a tariff directly increases the rand priceof imported goods, much like a weaker exchange rate– following a nominal depreciation – would do. We findthat the presence of tariffs on outputs or final goods, say,is not the major threat to an orderly price transmissionin general, or to poverty reduction in particular. Rather,the crux of international trade in South Africa is trade inintermediate inputs, such as chemicals and steel. Therefore,the potential adverse impact on poverty via higher randprices of imported intermediate inputs is much larger thanthat of more expensive foreign final goods (as the lattereffect constitutes only 10% or so of the consumer priceindex, or CPI). So, the first policy recommendation is thatmore gains are to be made for the domestic consumer bylowering tariffs on intermediate goods than on final goods.Further, we show that the presence of IPP increases theadverse price impact of tariffs – either on intermediate orfinal goods – (or of a weaker rand) on the CPI and thereforeon poverty. So the second policy recommendation is thattariffs on intermediate inputs need to be scrapped as soonas possible as they invite domestic producers to engage inIPP, which not only increases the adverse price effects oftariffs on the CPI (more than three times as large as in theabsence of IPP) but also of a nominal rand depreciation(almost three times as large as in the absence of IPP).
Development & Poverty ReductionTABLE 1: FINAL GOODS: THEORYDomestic producers engage in autonomouspricing (AP)Domestic producers engage in import paritypricing (IPP)Foreign producers engage in producer currency pricing (PCP) Section 2.1 Section 2.270<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>consumers (and producers) should benefit from changes in border prices.But it is equally important to assess whether changes in border prices arepassed through to the rest of the economy. Sector-specific analyses arenecessary to map the movement of a select number of goods through theproduction chain from foreign producer to domestic consumer (and producer).This will help to show how prices are transmitted in the economyand identify key constraints (lack of competition; tariffs; regulatory barriersetc.).So, in order to understand actual price levels (and their implied evolution,or the implied inflation rate), we need to understand (or have a basicpicture of), value chains, or input-output tables associated with the SouthAfrican economy. The way this will be operationalised is to draw up aconceptual framework on price transmission that will be illustrated withthe case studies of the project.More specific, building on a macroeconomic decomposition of the CPIinto imported final and intermediate goods – and on the relevant casestudies – we present stylised price transmission models for the clothingindustry, the motor trade, the small household appliances sector and thewheat-flour-bread value chain.The remainder of this study is organised as follows. It appears sensible– in line with the method of declining abstraction or Occam’s razor – tostart out with final goods. Therefore, in Section 2 we discuss some theoryrelevant for trade in final goods. Relevant factors here are:i) Foreign price setting (or choice of invoicing currency); Producer Currency Pricing(PCP) versus Pricing to Market (PTM) and tariffs;ii) Market structure; andiii) The relevant share in the domestic consumption basket.If we assume there is no trade in intermediate goods but only in finalgoods then international price transmission is solely via the direct channelof the CPI. Here we distinguish two important cases with respect to domesticprice setting of home-produced final goods: what we call ‘autonomouspricing’ (AP) and import parity pricing (IPP). Table 1 illustrates.With respect to intermediate inputs we are thinking of traded intermediateinputs such as oil, steel and plastics. The relevant issues parallel thoseabove, namely:i) Foreign price setting (or choice of invoicing currency); PPCP versus PTM withrespect to both intermediate inputs and final goods, tariffs;ii) Market structure; andiii) The relevant share in the domestic consumption basket (final goods), and therelevant input share in the production process of the relevant good (intermediategood)What we have now is that also home-produced goods will have an imported‘inflation’ component. If we assume there is no trade in final goodsbut only in intermediate goods then international price transmission issolely via the indirect channel of the PPI. However, in Section 3 we workwith an analytical framework that allows for trade in both intermediateand final goods. This allows for a richer and more realistic menu of pricetransmission. Again, we distinguish between two important cases withrespect to domestic price setting of home-produced intermediate inputs:what we call ‘autonomous pricing’ (AP) and import parity pricing (IPP).Table 2 illustrates.Section 4 deals with the price implications of the case studies on themotor trade, the clothing industry, the wheat-flour-bread value chain andsmall household appliances. Table 3 illustrates. Section 5 contains conclusionsand policy recommendations.2. Final goodsBefore we start with our analytical framework we would like to introducea visual tool that will help the reader navigate through – and keep trackof – the conceptual framework. Paragraphs and equations have beencolor-coded to keep track of transmission effects. First, the consumer priceindex or CPI is always indicated in light grey. Medium green indicates theproducer price index, or PPI. Next, imported final goods are indicated inlight green. Home-produced final and intermediate goods are dark grey,whilst imported intermediate inputs have an dark green label.
TABLE 2: INTERMEDIATE GOODS: THEORYDomestic producers engage in autonomouspricing (AP)Domestic producers engage in import paritypricing (IPP)Foreign producers engage in producer currency pricing (PCP) Section 3.1 Section 3.2TABLE 3: FINAL AND INTERMEDIATE GOODS: CASE STUDIESGood Case study Author SectionHousehold appliances Small household appliances Taz Chaponda and Matthew Stern 4.1Bread The wheat-flour-bread value chain Simon Hobson 4.2Clothing Clothing <strong>Industry</strong> Christi van der Westhuizen 4.3Cars Motor trade Nzeni and Frank Flatters 4.42.1 Foreign firms engage in PCP,domestic price exogenousIn this Section we assume that producers of foreign final goods set theirexport prices in foreign currency, say euro or dollar, so that currency fluctuationsdo affect the rand export price of those goods (and hence competitivenessof SA producers of final goods).An example of a sector of the SA economy that operates in this way isthe consumer appliance sector. In the case study of Chaponda and Stern(2006) [hereafter CS] for small household appliances (toasters, hotplatesand kettles) it is found that the recent strength of the rand has hurt local(SA) manufacturers of small appliances (Nu World and Amalgamated Appliancesor AMAP in Gauteng) as the cost of Chinese imports continuedto fall. Prior to the strengthening of the rand they found that the sharprise of prices in 2002 was almost entirely a currency story. This suggests analmost uniform exchange rate pass-through effect (see further below 3 ).More precisely, we assume that the rand price of the (imported) foreignfinal good is set according to:Here is the rand price of the imported final good, is the dollarprice of the imported final good,is the rand dollar exchange rate (definedas rand per dollar), and is the proportional tariff (as a perunage) 4 .Define , then we get . Taking natural logsof this expression we get the central expression we need for this Section,namely:(2.1) 5We assume that in setting foreign, say, Chinese producers take thehome price of the final good in domestic currency ( ), the nominal exchangerate and the tariff as given. 6 Here, the sequence of events in termsof price-setting is as follows. First, the foreign producer decides on histarget price in his own currency (dollars or euros), then he observes theprevailing level of the exchange rate and declares a rand export price toSARS. 7 Finally, SARS adds on a proportional tariff. Figure 1 illustrates.CS (2006, p. 10) call a tariff on outputs as distinguished from a tariffon inputs or intermediate goods, which we label (see Section 3).Note that the former can be quite substantial. Tariffs on vacuum cleanersand irons (as examples of foreign-produced final goods) have remained71<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>4All exogenous (unexplained) variables are underlined.5And we use that3More precisely, in line with equation (2.1), the rand price of imported final goods respondsin a one-to-one manner to currency movements.6The choice of China as the exporter is intentional; by 2004 China was the leading exportingcountry to South Africa in all but one of six products, microwave ovens [CS (2006, p.7)].7Alternatively what may happen is that the foreign producer (and possibly also the domesticconsumer) are unaware of the domestic tariffs. The domestic consumer orders the foreigngood, paying in foreign currency. When the foreign final good passes through South Africa’sborders the tariff is added on by SARS and paid for by the domestic consumer.<strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty Reduction72unchanged, at 20% since at least 1988, while those on cooking equipment,toasters and kettles have been reduced from 30% to 20%. This isa high tax to pay on the consumption of basic household appliances [CS(2006, p. 8)].It is clear from equation (2.1) that under PCP the pass-through from boththe nominal exchange rate as the tariff to the rand price of the importedfinal good is 100% (as ). 8Let us now look at another crucial equation, namely the consumer priceindex, or CPI (here denoted by ). Assuming only final goods are traded, 9the latter is simply the weighted average of the prices of home producedand imported final goods. Of course the weight of foreign prices in the CPIis equal to the share of imported final goods in the domestic consumptionbasket. In South Africa this share is about 10%; that is 10 % of the CPIis made up of the prices of imported final goods. Denoting the latter by, we then get:<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>(2.2) 10Note that we have underlined the home price level, or producer priceindex (PPI). That is we have:(2.2.a)The reason is that here we have no theory of what determines , we simplyassume it to be exogenous (in this paper all exogenous variables areunderlined). In Section 3 we will come up with a theory for the PPI. Therewe will state that the South African PPI is a weighted average of (markedup)unit labor costs (ULC) and import prices of intermediate goods suchas steel and oil.Note that we can write the CPI (2.2) in a slightly different way:(2.3)Here we have expressed the domestic consumer price index as theweighted sum of the price of the home produced good and the relativeprice between foreign and home products . The latter variableis also known as the terms of trade, or the (producer price-based) realexchange rate.Note that changes in the pattern of trade would be picked up by a changein . For example, CS (2006, p. 11) show that for the case of small householdappliances in the mid-1990s South Africa opened up to the worldeconomy, and imports, particularly from Asia, began to increase. Anotherexample would be the sharp increase in imports of clothing from Chinasince 2001 [Van der Westhuizen (2006, p. 18)]. In terms of equation (2.3)this can be modeled by increasing , the share of imported final consumptiongoods in the total consumption basket.8From the case study on small household appliances [Chaponda and Stern (2006)] it appearsthat PCP also applied to the exports side. They state that increases in iron and kettleexports appear to be driven by increased competitiveness due to a depreciating currency.9Later on in the analysis we will also consider what happens in case of imported intermediateinputs, such as steel and oil.10In terms of micro foundations, equation (2.2) assumes Cobb-Douglas preferences, so the(implicit) elasticity of substitution between home and foreign-produced final goods is unity(1)!
FIGURE 1: THE TIMING OF EVENTSStage 1 Stage 2 Stage 3decided observed and follows SARS slaps on ; realisesTABLE 4: FOREIGN PRODUCERS ENGAGE IN PCP, HOME PRICE EXOGENOUS; *1 0 1 2 1.1 0.9 1.191 0 0 1 1.1 -0.1 1.091 0.5 0 1.5 1.1 0.4 1.141.5 0 0 1.5 1.6 -0.1 1.591 0 0 1 1.6 -0.6 1.54* All numbers are computed assumingTable 4 illustrates how things might work under PCP for final goods. the imported final good, so that the South African consumer will startpaying more for the imported good (compared with row 2, increasesNote that the numbers for in column four are computed using equationto 1.5 from 1, a 50% increase). Again, of the 0.5 increase 10% works its(2.1) and the assumptions outlined in columns 1-3 (the numbers incolumn four are simply the sums of those in columns 1-3). Then, given theway to the CPI as the CPI increases by 0.1 * 0.5 = 0.05 (the CPI was 1.09and has as a result of the appreciation increased to 1.14). 13assumption on the home price , the terms of trade follows.Finally, using equation (2.3) – and working with an imported shareof final goods in the consumption basket of 10% – we can then computeRow 4 shows the combined effect of higher dollar prices for the importedgood and higher rand prices for the domestic good. Because of the latterthe CPI, .increases to 1.6 (compared to 1.1 in row 2). goes up from 1 (row2) to 1.5. As a consequence of this the CPI increases to 1.59 comparedNote that in row 1 we have a 1% tariff, and this is dropped by the authoritiesto 1.09 in row 2. The increase of 0.5 can then simply be decomposed asin row 2. 11 Therefore, by comparing the numbers for the CPI in 90% of the increase in (0.45) plus 10% of the increase in as arows 2 and 1, we see the immediate beneficial effect on lower tariffs for result of the higher dollar price (0.05).the (poor) consumer. The CPI drops from 1.19 to 1.09; a 10 percentagepoints fall. The lower CPI is entirely the result of lower import prices for Finally, row 5 shows that the (partial) effects of lower tariffs can be easilythe final good, . This price falls from 2 to 1, and 10% of this price eroded by higher domestic prices. For compared to row 1 the tariff hasdecrease (being the import share of final goods in the CPI) feeds through dropped from 1 to 0, and as a consequence has fallen from 2 to 1 (ato consumer prices. Put differently, the lower tariff induces a lower price ofthe imported good, which induces an appreciation of the terms of trade,which then lowers the CPI. 12100% reduction). But the domestic price has increased from 1.1 (row 1)to 1.6 (row 5). This latter effect dominates as the PPI price level is 90% ofthe CPI. As a consequence, compared with row 2 the CPI increases by 45basis ppoints, which is equal to the increase in the PPI (+ 50 basis points)In row 3 we analyse the effect of a nominal depreciation of the rand. Wesee that foreign producers factor this depreciation into the rand price ofminus 10% of the real appreciation (at 0.5). We are now ready to statethe conclusions from this Section.73<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>11South Africa implemented a wave of trade reforms in the mid-1990s that saw tariffsdecrease on most traded goods reduced sharply between 1994 and 2004. The average tarifffell from 22% to 8% over this period [Edwards (2005)]. Also, South Africa commencedthe liberalisation of the clothing sector in compliance with the WTO’s agreement on Textilesand Clothing (ATC). More specific, South Africa decided unilaterally to halve ad valoremtariffs over a shorter period than required – seven instead of 12 years – to a maximum of40% by 2004 [Van der Westhuizen (2006, p. 17)].12Formally we have.13Note that this example is consistent with the evidence reported by the CS (2006) case studyfor small household appliances. They find that the recent strength of the rand has promptedconsumers to respond positively to the reduced cost of imported appliances.<strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty Reduction74The price of the final good is ‘stacked’; that is the beneficial effect of lower tariffs(lower level of ) could be undermined by higher foreign currency (dollar, say)denominated dock prices ( ), and/or a weaker rand dollar exchange rate (higher). 14There is a role for trade policy in engineering a lower level of tariffs as this will resultin lower prices of final imported consumer goods, which in turn will result in a lowerlevel of the CPI. We know the effect of a lower tariff on will be one-for-one,so that a 1% lower tariff will result in the price of the final good also droppingby 1%. In turn the lower price of the imported final good will then lower the CPIis no reason for domestic producers to charge below the cost (to theconsumer) of imported appliances, on which duties are charged. The onlyoptions available to consumers are to pay the duty-inclusive price on localgoods or purchase imports and pay the duty anyway. Another example isthe clothing industry [see Van der Westhuizen (2006)]. If Truworths canimport clothing from China (presumably with zero or minimal tariffs), whywould it pay a KwaZulu-Natal Cut Make and Trim (CMT) more than whatthey can import those goods for (including the cost of tariff, insurance and<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>( ) by 1 x = 0.1%. 15<strong>Trade</strong> and competition policy will work in parallel – in the sense that tradepolicy will achieve some of the objectives of competition policy, namely a lowerlevel of . 16 The reason is that given the level of prices for home-produced finalgoods ( ), lower rand prices for imported final goods as a result of the lowertariff will decrease the competitiveness of domestic producers as the terms of tradewill deteriorate. This then may put pressure on domestic producers toalso lower their prices ( ) in order to regain some of the lost ground. 17Finally, the beneficial effects of lower tariffs on consumer prices can be easilyeroded if domestic prices (the PPI) go up. This is especially threatening as importedfinal goods make up only a small fraction of the CPI (10%), so assuming PCP thethe exchange rate) from China?In case of IPP the price of home-produced final goods is equal to therand price of the imported good including the relevant tariff according toequation (2.4) below.(2.4)Combining equation (2.4) with the definition of the CPI (2.3) we find thatthe latter is now given by equation (2.5):drop in tariffs would have to be nine times as large as the increase indomestic prices! 18(2.5)2.2 Domestic producers of final goodsengage in IPPIn this Section we assume that domestic producers of final goods engagein IPP; that is, the rand price of the locally produced final good ( ) isequal to the rand price of the imported final good including the relevanttariff ( ).This case is very important for South Africa. For instance CS (2006, p. 8)use the assumption of IPP – domestic manufacturers mark-up their pricesfor (equivalent) products sold in South Africa – to estimate the impactof protection (the cost to the South African consumer) on purchases ofdomestically produced products. They motivate IPP by stating that there14This follows from equation (2.1). Note that all mentioned variables have the same effect on, i.e.15This follows from taking equations (2.1) and (2.2) together.From the expression for the CPI we see that now – due to the presenceof IPP – we have equality between the CPI and the PPI (as the sharedrops out)! The reason is that the imported good is priced similarly to thehome-produced good, thus for the consumption price index it does notmatter whether a good is imported or locally produced.Table 5 illustrates how things might work if domestic producers of finalgoods engage in IPP.Note that in row 1 we have a 1% tariff, and this is dropped by the authoritiesin row 2. Therefore, by comparing the numbers for the CPI inrows 2 and 1, we see the immediate beneficial effect on lower tariffs forthe (poor) consumer. The CPI drops from 2 to 1 (a fall of 100 percentagepoints). Compare this with the case where home producers of final goodsdo not engage in IPP; then the fall in the CPI is a mere 10 percentagepoints.16Obviously, things may also go too far here. If competitiveness deteriorates too much, domesticfirms may be put out of business completely.17This follows from equation (2.3). Note that another strategy followed by domestic producersof small household appliances [see CS (2006, p. 4)] was to reduce the portfolio of productsmanufactured domestically to focus on those where they remain competitive, and to investin technological innovations to retain their competitiveness in these products. This may includeusing local design skills and marketing skills to develop products to be manufacturedin a foreign country (say, China) for the domestic market.18This follows from the definition of the CPI (2.2). Taking first differences and imposing azero change we getRearranging and using thedefinition of (equation (2.1)) we get .
TABLE 5: FOREIGN PRODUCERS ENGAGE IN PCP, DOMESTIC PRODUCERS ENGAGE IN IPP; *1 0 1 2 2 0 21 0 0 1 1 0 11 - 0.5 0 0.5 0.5 0 0.51.5 0 0 1.5 1.5 0 1.51 0 0 1 1 0 175* All numbers are computed assumingThus, the benefits of cutting tariffs for the poor consumer are much largerif domestic producers of final goods engage in IPP; here the benefits are10 times as large! 19Further note that the reason why they engage in IPP in the first place, isthe presence of tariffs which raises above (the price of the homeproducedfinal good in the absence of IPP). This can be formalised asfollows. It makes sense for domestic producers to set according tothe rational pricing rule:(2.6)where is the (exogenous) price of the locally produced final good notlinked to the price of the imported final good and is the operatorthat returns the largest value in a set of values. 20 So, the presence of thetariff raises above and it therefore becomes lucrative – accordingto rule (2.6) – to price according to IPP. 21In row 3 we analyse the effects of a nominal depreciation of the randunder IPP for final goods. Note that the CPI now increases from 1 to 1.5;a 50 basis points increase! This increase is much larger than under thecorresponding case of AP (see row 3 of Table 4) where the increase wasonly 5 basis points (from 1.09 to 1.14). Thus, the presence of IPP has the19Note that IPP with respect to final goods – where the share of home-produced final goods is90% of the CPI – is a bit of a theoretical curiosity. The reason is that in reality part of thePPI – in South Africa about 30% – consists of the prices of imported intermediate inputs.More specifically, in the conceptual framework employed in this study the only domesticprices that can in the end be set according to IPP are the prices of home-produced intermediategoods. The latter constitute 0.90 x 0.70 x 100 = 63% of the CPI. If those goodsare priced to parity with the prices of imported intermediate inputs, the effects of tariffson those inputs on the CPI would increase from 0.27 (in the absence of IPP) to 0.9. Thismeans that with IPP the adverse effect on the CPI is now more than three times as large (or0.9/0.27 = 3.33)!20Of course, if the relevant price is not charged by the producer but has to be paid to asupplier the formula becomes , where is theoperator that returns the smallest value in a set of values.21Note that the effects of a weaker rand (higher number for ) are completely analogous.effect of causing the inflationary effect of the nominal depreciation to be10 times as large as the one under PCP.We are now ready to state the conclusions from this Section. The price of the final good is ‘stacked’; that is the beneficial effect of lower tariffs(lower level of ) could be undermined by higher foreign currency (dollar, say)denominated dock prices ( ), and/or a weaker rand dollar exchange rate (higher). There is a role for trade policy in engineering a lower level of tariffs as this will resultin lower prices of final imported consumer goods, which in turn will result in a lowerlevel of the CPI. We know the effect of a lower tariff on will be one-for-one, sothat a 1% lower tariff will result in the price of the final good also dropping by 1%.In turn the lower price of the imported final good will then lower the CPI ( ) by1%. Note that the benefits of cutting tariffs for the poor consumer aremuch larger if domestic producers of final goods engage in IPP; here thebenefits are 10 times as large! <strong>Trade</strong> and competition policy will work in parallel – in the sense that tradepolicy will achieve some of the objectives of competition policy, namely a lowerlevel of . The reason is that given the level of prices for home produced finalgoods (), lower rand prices for imported finalgoods as a result of the lower tariff will decrease the competitiveness of domesticproducers as the terms of tradewill deteriorate. This then may putpressure on domestic producers to also lower their prices ( ) in order to regainsome of the lost ground. Finally, the inflationary effects of a nominal depreciation are much more pronounced(here 10 times as large) than under AP. The reason is that the increase inrand prices of final goods is immediately followed by a similar increase in the pricesof home-produced final goods.3. Intermediate goodsNow it is time to start thinking about what determines the domestic pricelevel or PPI ( ), as so far this has been assumed to be exogenous. Inparticular, in reality an economy not only imports final goods, but alsointermediate inputs such as steel and oil that show up as the importedcomponent of the producer price level (PPI or ), not directly in the CPI.<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong><strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty Reduction76<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The causality is then as follows: the imported intermediate input componentaffects the PPI, and the PPI affects the CPI via the domestic finalgoods component of the latter (where in order to produce final goodsdomestically one needs to import intermediate inputs). 22Let’s now make the above intuition a bit more explicit by introducingsome equations. We assume each variety of the final consumption goodin the home country (say, South Africa) is produced using domestic andforeign intermediate goods aggregates. For instance, the final goodis produced using the home and foreign intermediate good aggregates,respectively and , with the production function:(3.1)In turn linearising (3.1) we get:(3.4) 23The home-produced final good is associated with a producer price index(PPI) given by:(3.2)where , the relative price of foreign to domestic intermediate goodsmeasured in domestic currency is defined as, and.Taking natural logs of (3.2) we get:3.1 Foreign firms engage in PCP,domestic producers mark-updomestic unit labour costsWe assume that the price of the domestic intermediate input aggregateis determined as a fixed mark-up over unit labour costs (ULC). That is, weassume domestic producers engage in local currency pricing, or producercurrency pricing.More specific, the price of the domestic intermediate input aggregateobeys:where(3.5)is labour’s share in the aggregate value of output ofdomestically produced intermediate goods ( ), and isthe average productivity of labour.Taking logs of (3.5) we get:where .Combining (3.6) with (3.3) we get:(3.6)(3.7)(3.3)Equation (3.3) is the central expression in this Section. It is completelyanalogous to the definition of the CPI (2.3). So, is the share of importedintermediate inputs needed for home-produced final goods, andis the (remaining) share of local inputs used in production. A sensiblenumber for the South African economy for is 0.30. Note that this equationnow replaces equation (2.2.a) which we have used in Section 2 andwhich assumed the home PPI to be completely exogenous. 24Equation (3.7) states that the domestic price level, or the PPI, is a weightedaverage of (marked-up) ULC and import prices of intermediate inputs.Here the relevant weight is determined by the share of imported intermediateinputs in the production of the home produced final good, . Assaid before, a sensible number for this share in the South African contextis 0.30.So far we have not said anything about what determines import pricesof intermediate inputs. Following the analysis in Section 2, here we willwork with various assumptions, namely AP and PTM. In this Section weassume foreign producers of intermediate inputs engage in producer currencypricing.22Another example, based on the case study on the motor trade [see Nzeni and Flatters(2006)] would be imported vehicle (spare) parts such as clutches, brake pads, etc.23For an example of a similar set-up (involving labour and oil) see Agénor (2004, p. 197).24Note that as the share of imported intermediate inputs goes to zero, that is if ,then. Also, in terms of micro foundations equations (3.1)-(3.3) assumeCobb-Douglas preferences, so the (implicit) elasticity of substitution between the home andforeign intermediate goods aggregates is unity (1)!An example of a sector of the South African economy that seems to operatein this way is the wheat-flour-bread value chain [Hobson (2006)].
FIGURE 2: THE TIMING OF EVENTSStage 1 Stage 2 Stage 3decided observed and follows SARS slaps on ; realisesTABLE 6: FOREIGN PRODUCERS ENGAGE IN PCP, HOME PRICE MARKED-UP ULC;Hobson points out that consumers have benefited from the appreciatingrand; suggesting a substantial degree of exchange rate pass-through, andhence PCP.Then, in the case of a tariffsteel, 25 we get:on imported intermediate inputs, say(3.8)Note that equation (3.8) is completely analogous to equation (2.1) in thecontext of PCP by foreign producers of imported final goods.**1.1 0 1 2.1 1 1.1 1 1.4 -0.4 1.361.1 0 0 1.1 1 1.1 0 1.1 -0.1 1.091.1 0.5 0 1.6 1.5 1.1 0.5 1.25 0.25 1.2751.1 0 0 1.1 1.5 1.1 0 1.1 0.4 1.14* All numbers are computed assuming =0.30 and = 0.10.** We assume here that is exogenous, so terms of trade movements are not driven by rand prices of imported final goods. Inaddition, in this case we are not interested in changes in , or .Here, the sequence of events in terms of price-setting is as follows. First,the foreign producer decides on his target price in his own currency (dollarsor euros), then he observes the prevailing level of the exchange rateand declares a rand export price to SARS. Finally, SARS adds on a proportionaltariff. Figure 2 illustrates.It is clear from equation (3.8) that under PCP the pass-through from boththe nominal exchange rate as the tariff to the rand price of the importedintermediate good is 100% (as ).Note that we can write the PPI (3.3) in a slightly different way:77<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>It is important to analyse tariffs on inputs (as distinct from tariffs on outputsor final goods) as well, as tariffs on outputs ( ) do not show the fullimpact of trade protection on local products because they do not showhow tariffs on inputs ( ) affect the cost of intermediate inputs used inthe production process [CS (2006, p. 10)].We assume that in setting , say the ‘free on board’ (FOB) price, foreignproducers take the home price of the intermediate good in domesticcurrency , the nominal exchange rate and the tariff as given. 2625Another example of imported intermediate inputs that are (substitutes for home-producedintermediate inputs) subjected to tariffs are vehicle parts (disc-brake pads, bumpers, etc.).See Nzeni and Flatters (2006, pp. 9-10).26For the case of wheat the mechanism of the tariff did not work exactly like equation (3.8).Instead, according to the system of tariffs which was effectively implemented during February1998 and repealed in July 2006, the tariff was in the form of a variable import levythat worked as follows. A reference price, say, which was tied to the internationalimport price of wheat ( ), was set. The levy then only became effective when the worldprice fell below the reference price, i.e. when. This had the effect of settinga floor price for wheat in South Africa. In a similar fashion the import duty on flour wasset at 1.5 times the duty on wheat.(3.9)Here we have expressed the domestic producer price index as the sum ofthe price of the home produced intermediate good and the relative pricebetween foreign and home intermediate products. Thelatter variable is also known as the terms of trade of intermediate goods(compared to the terms of trade of final goods ).Table 6 illustrates how things might work under PCP for intermediateinputs. From the first two rows of Table 6 we can see the first interestingaspect of our analysis of trade and tariffs regarding intermediate goods:the effect on the CPI ( ) of a cut in tariffs is much larger if the tariffused to be on intermediate goods than if the tariff was on final goods. Bycomparing the first and second rows it can be seen that drops from1.36 to 1.09, a drop of 27 basis points. This effect is much larger than theequivalent reduction of the tariff on the final good (in both cases of courseassuming the foreign producer obeys PCP) at 10 basis points (as can be<strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty Reduction78<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>seen from inspecting the first and second rows of Table 4). 27 In the lattercase the size of the effect is simply the drop in the rand-denominatedimport price (at 1%) multiplied by the share of imported final goods in theSouth African consumer basket (at 10%).Here, the mechanism is a bit more complicated, but can be neatly split interms of the effect on the PPI and the terms of trade of the final good,in line with:(2.3)The first effect of the lower tariff is that the PPI ( ) drops by 0.3 x 1= 0.3% (or 30 basis points). The PPI drops by this amount as accordingto equation (3.9), the local component, , remains unchanged but theterms of trade of intermediate goods appreciates by 1%,30% ( ) of which passes through to the PPI. The second effectruns via the terms of trade of the final good . The rand priceof the imported final good does not change (as the tariff change on importedintermediate inputs has no effect on . But, as said above, thedrop in at 0.3% does imply a depreciation of the real exchange rate(terms of trade of final goods) of 0.30% (-0.40 minus -0.10). This realdepreciation then according to equation (2.3) pushes up the CPI by 0.10(being the number we use for , the share of imported final goods in theSouth African consumption basket) multiplied with 0.30 equals 0.03%.The total effect on the SA CPI is then the sum of these two effects, that is,-0.30 plus 0.03 equals -0.27%.pend on the calibrated shares and ) 29 suggest that those gains are almostthree times as large as an equivalent drop on tariffs on final goods.The price of the intermediate imported good is ‘stacked’; that is the beneficial effectof lower tariffs (lower level of ) could be undermined by higher foreign currency(dollar, say) denominated dock prices ( ), and/or a weaker rand dollar exchangerate (higher ).<strong>Trade</strong> and competition policy will work in parallel – in the sense that tradepolicy will achieve some of the objectives of competition policy, namely a lowerlevel of . The reason is that given the level of prices for home produced intermediategoods ( ), lower rand prices for imported intermediate goods as aresult of the lower tariff will decrease the competitiveness of domestic producerswith respect to supplying home-produced intermediate inputs as the relative pricebetween foreign and home intermediate productswill deteriorate.This may then put pressure on domestic producers to also lower their prices( ) in order to regain some of the lost ground. 30Finally, the beneficial effects of lower tariffs on consumer prices can be easilyeroded if domestic prices of intermediate inputs ( ) go up. This is especiallythreatening as imported intermediate goods still make up only the minority of thePPI (30%), so assuming PCP, the drop in tariffs would have to be more thantwo times as large as the increase in domestic prices! 313.2 Domestic producers of intermediategoods engage in IPPIn this Section we assume that domestic producers of intermediate goodsengage in IPP. That is, the rand price of the locally produced intermediategood ( ) is equal to the rand price of the imported intermediate goodincluding the relevant tariff ( ). That is:Equation (3.10) generalises the discussion above by deriving the generalexpression for the CPI by combining the equations for the PPI (3.9) andthe CPI (2.3):(3.10) 28Here the first term in curly brackets is the effect on the PPI (here -0.30),and the second term is the terms of trade effect (here + 0.03).Here are our conclusions: The first – and most important – conclusion that can be drawn from the case ofPCP for imported intermediate goods is that more gains are to be made forthe domestic consumer by lowering tariffs on intermediate goods thanon final goods. The numbers in the exercise conducted (which do solely de-27Chaponda and Stern (2006, p. 10) show that for the case of small household appliances(more specifically vacuum cleaners and irons) the fall in output tariffs ( ) has been accompaniedby an even faster fall in the tariffs on key inputs ( ).28This can be further expanded (by using the expression for again) as:(3.11)(3.12)This case is very important for South Africa. CS (2006) find that SouthAfrican producers of plastics, steel and aluminium follow the IPP model.In addition, Hobson (2006) finds that South African producers of wheatalso follow IPP calculations to determine prices. 3229The chosen numbers for and are closely aligned with the national (aggregate) accountsof the South African economy.30In the set up we have here for the determinants of , that is(3.6), this would most likely take the form of a lowermark-up over domestic unit labour costs.31This follows the definition of CPI (2.2). Taking first differences, imposinga zero change, settingand using (3.3) we getRearranging and using the definition of(equation (3.8)) we get .32In the latter case this actually benefits the South African consumer as the foreign (dollar)price of wheat is kept artificially low by EU/US farm subsidies. That is, we have. Here is the exogenous dollar price in the absence of a subsidy,and is the subsidy in foreign currency. This implies that is set accordingto. Using this expression we canformalise Hobson’s (2006, p. 14) statement that South African wheat farmers are in essence
FIGURE 3: WHEAT PRODUCER PRICE VERSUS THE R/$ EXCHANGE RATE, JANUARY 1999 - MAY 2004*79<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>R/$ exchange rateSAFEX price (R/ton)Source: Hobson (2006, p.11)As can be seen from equation (3.12), in case of IPP the (nominal) exchangerate becomes a major driver of the rand producer price of home-producedintermediate goods (in fact the correlation between and becomes1 or 100%!). Figure 3 [from Hobson (2006)] illustrates the strong correlationbetween the producer price of wheat and the R/$ exchange rate fromJanuary 1999 to May 2004.Thus as the exchange rate appreciates, the producer price goes down andvice versa. Hobson (2006, pp. 13-14) reports that – in turn – the retailprice of bread ( in the notation of this study) is closely related to theSAFEX wheat price (or ), although there is a lag of approximately fourmonths.In case of IPP for intermediate goods we get equivalence between theproducer price of final goods, , and the rand price of imported intermediategoods, as can be seen from equation (3.13):being squeezed (low value of ) by the impact of the strengthening rand on the onehand (low or negative value of ), and the heavily subsidised world price of wheat (negativevalue of ) which allows millers to import at a price lower than the South Africancost of production.(3.13)Table 7 illustrates how things might work under IPP for home-producedintermediate goods. Note that in row 1 we have a 1% tariff, and this isdropped by the authorities in row 2. Therefore, by comparing the numbersfor the CPI in rows 2 and 1, we see the immediate beneficial effect onlower tariffs for the (poor) consumer. The CPI drops from 1.99 to 1.09 (a90 percentage points fall). Compare this to the case where home producersof intermediate goods do not engage in IPP – then the fall in the CPIis a mere 27 percentage points. Thus, the benefits of cutting tariffs onintermediate inputs for the poor consumer are much larger if domesticproducers of intermediate goods engage in IPP; here the benefits are morethan three times as large! Note that this result is similar to the one wefound with respect to IPP of imported final goods (although there thequantitative differences were bigger). 3333This result (benefits for the consumer of cutting the tariff on final goods under IPP by homeproducers of final goods is 10 times the same benefit when they do not engage in IPP) isdriven by the large share in the CPI of home-produced final goods, namely 90%. So, therelevant ratio of 10 follows from , where . Similarly, when we discuss thecase of IPP in the context of intermediate inputs, the share of imported intermediate inputs( ) is 30%, so the relevant ‘benefit ratio’ is now .<strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty ReductionTABLE 7: FOREIGN PRODUCERS ENGAGE IN PCP, HOME PRODUCERS ENGAGE IN IPP: ***1.1 0 1 2.1 1 2.1 0 2.1 -1.1 1.991.1 0 0 1.1 1 1.1 0 1.1 -0.1 1.091.1 0.5 0 1.6 1.5 1.6 0 1.6 -0.1 1.591.1 0 0 1.1 1.5 1.1 0 1.1 0.4 1.1480<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>* All numbers are computed assuming = 0.30 and = 0.10.** We assume here that is exogenous, so terms of trade movements are not driven by rand prices of imported final goods. Inaddition, in this case we are not interested in changes in , or .Further, note that the reason why domestic producers engage in IPP in thefirst place, is the presence of tariffs, which raises above (the priceof the home-produced intermediate input in the absence of IPP). This canbe formalised as follows. It makes sense for domestic producers to setaccording to the rational pricing rule:(3.14)where is the (autonomous) price of the locally produced intermediategood not linked to the price of the imported intermediate good andis the operator that returns the largest value in a set of values. So,the presence of the tariff raises above and it therefore becomeslucrative – according to rule (3.14) – to price according to IPP. 34In row 3 we analyse the effects of a nominal depreciation of the rand. Themost important thing that we see here is that the CPI goes up by a lotmore than in the absence of IPP. For here the CPI jumps from 1.09 to 1.59,an increase of 50 basis points (the increase in the case of autonomouspricing for intermediate inputs was only 18.5 basis points). The reason theeffects on inflation of a nominal depreciation are much larger now is thatthe increase in the rand price of the imported intermediate input (goes up from 1.1 to 1.6; the same increase as under AP) is now followedby an increase in the rand price of home-produced intermediate inputs aswell. The latter price goes up from 1.1 to 1.6 as well! As a consequence,the CPI’s increase is much larger (2.7 times as big as under AP). This rowis interesting as it has a real-life counterpart in the form of the inflationaryeffects of the 2001 depreciation of the rand. Then inflation spiked asthe price level jumped upwards. For example, the price of bread went up,following the depreciation of the rand, which is not hard to understand,given the fact that wheat is priced according to IPP.We are now ready to state the conclusions from this Section:The first conclusion that can be drawn from the case of IPP for imported intermediategoods is that the nominal exchange rate becomes a major driverof the rand producer price of home-produced intermediate goods.The second result is that the benefits of cutting tariffs on intermediate inputsfor the poor consumer are much larger if domestic producers ofintermediate goods engage in IPP; here the benefits are more than three timesas large as in the case of AP.<strong>Trade</strong> and competition policy will work in parallel – in the sense that tradepolicy will achieve some of the objectives of competition policy, namely a lowerlevel of . The reason is that given the level of prices for home produced intermediategoods (), lower rand pricesfor imported intermediate goods as a result of the lower tariff will decrease thecompetitiveness of domestic producers with respect to supplying home-producedintermediate inputs as the relative price between foreign and home intermediateproducts will deteriorate. This may then put pressure on domesticproducers to also lower their prices ( ) in order to regain some of thelost ground.Finally, the inflationary effects of a rand depreciation are also muchlarger than under AP of intermediate inputs. Again, we find the effects to beabout three times as bad as under AP.4. Final goods: case studies4.1 Small household appliances(Chaponda and Stern)Before we proceed with the implications of Chaponda and Stern’s (2006)case study of small household appliances for the price transmission mechanismin this sector, we discuss a table similar to the decomposition of theCPI used in Section 1 – Table 8.34Note that the effects of a weaker rand (higher number for ) are completely analogous, soin this sense a weaker exchange rate has the same effect as a tariff.Based on Table 8 we see that the consumption price of the basket of(home and foreign produced) appliances, , consists for 80% of the price
TABLE 8: PRICE TRANSMISSION IN THE SMALL HOUSEHOLD APPLIANCES SECTOR*0.20 x (prices of home-produced appliances) = 0.20 x+ 0.80 x (prices of imported appliances)= 0.80 x=0.5 x (price of home-producedintermediate goods) ( )**+ 0.5 x (price of importedintermediate goods) ( )* Abstracting from margins (mark-ups) and VAT.=0.10 x (price of home-producedintermediate goods)+ 0.10 x (price of importedintermediate goods)+ 0.80 x (prices of importedfinal goods)81of the imported final good, , and for 20% of the price of the locallyproduced appliances, . The latter, in turn, can be split into two equalshares allocated to the prices of home- and foreign-produced intermediategoods aggregates (mostly material costs such as plastics, steel, aluminium,switches and elements). 35The implication of all this is that the consumption price of appliances canbe split into a 80% imported foreign final price component (includingthe tariff) and two 10% shares allocated to the prices of home- and foreign-producedintermediate goods, respectively. We will use this decompositionlater on in the proposed sequencing of recommended trade andcompetition policy measures for this sector.Here are our conclusions for price transmission (based on the insightsfrom our conceptual framework and those of the CS (2006) case study):** This share is assumed to be 0.5, since there is no information in the case study to help calibrate it.Foreign competitors appear to engage in PCP (perhaps not for 100%, but thereappears to be significant exchange rate pass-through from the exchange rate torand denominated import prices)In response to a deteriorating terms of trade (a lower number for ), the domesticindustry has been restructured (away from manufacturing towards brandmanagement, overseas manufacturing and the associated importing). Existing suppliershave become leaner and much more profitable.The relevant companies (Nu-World and AMAP) – that supply about 80% of all smallhouseholds appliances in South Africa – have adapted very well to the ‘threat’ ofinternational competition and have captured most of the gains.The gains to consumers in terms of product variety and quality have also beensubstantial.But input prices (prices of intermediate inputs) and output prices (final goodsprices) remain distorted by unnecessary tariffs, and South African consumers continueto pay a high price for basic home appliances. This conclusion is in line withour analysis in Sections 2.1 and 3.1, which shows how tariffs on imported finalgoods and intermediate inputs 36 both push up the prices of final goods via a directand indirect channel respectively.35The labour cost component of home-produced intermediate goods is very small, just 2% oftotal production costs.36For example, the surveyed companies have to pay import duties on specialised importedcomponents (that are not locally available) such as switches, elements, valves and pipes.The poor who are least able to bear the cost of tariff duties are affected most bythem.As the share of imported goods such as irons, kettles, toasters, vacuum cleanersand irons makes up about 80% of total domestic consumption of small householdappliances, the elimination of the tariff on goods of this type would contribute toa substantial reduction in the price of these appliances in South Africa [CS (2006,p. 11)].The manufacturing plants of the relevant companies source inputs from a rangeof local and overseas suppliers. Plastics, steel and aluminium are largely sourceddomestically, while specialist component parts are imported.One of the surveyed companies identifies the high cost of South African metals andplastics as its main competitive disadvantage relative to foreign (Chinese) producers.Again, this is in line with the results of Section 3.2, which shows how the practiceof IPP for intermediate inputs pushes up the producer price level ( ) and given theprice if imported final goods ( ) of domestic (South African) final goodsproducers. 37Surprisingly, labour costs are not a significant constraint to competitiveness. Labourcosts contribute just 2% of total production costs.If we cast this result in terms of the algebra of Section 3.1, we find that the unitlabour cost componentof the price of the home-produced intermediateinput aggregate is small relative to the ‘other’ component (materialcosts and profits).In terms of the value chain associated with the domestic price of home-producedsmall household appliances, we can conclude the following: Chinese competitionhas resulted in deteriorating terms of trade (lower number for ( ).Although the domestic industry appears to have restructured successfully, thereremain a number of problems. First, by far the largest component of the domesticcost base ( ) is cost of material rather than unit labour costs. Since domesticsupplies of the relevant intermediate inputs (plastics, steel) tend to engage in IPP, and hence is pushed up, this puts further pressure on the competitivenessof local final goods producers (who were already under threat from the Chinesein the first place). Second, with respect to imported intermediate inputs, such asswitches and elements, producers of home final goods are charged import duties bySARS, which further threatens their competitiveness (which was already under pressurefrom (i) the Chinese and (ii) the presence of IPP by local intermediate producerssuch as Mittal Steel and Sasol).37Material costs can reach between 60% and 70% of production costs, with steel the mostimportant input. While manufacturers of white goods enjoy a rebate from ISCOR, this doesnot apply to manufacturers of smaller appliances.<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong><strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty Reduction82<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong> Of course, the fact that domestic intermediate goods producers do engage in thepractice of IPP in the first place is largely caused <strong>38</strong> by the presence of the tariff onimported intermediate inputs ( ) in the first place (which raises above )– that is, by government policy. 39 We now repeat CS’s (2006, p. 12) main conclusions. They say that the real loss inthe small consumer appliances sector is one of economic waste. Domestic producersof final goods are supported by very high levels of effective protection 40(due to the high level of which pushes up and thus improves the terms oftrade and decreases competitiveness). As a consequence, they do notexport. The full cost of this protection is borne by mostly poor consumers of mostlyimported appliances (if they buy imported appliances the SARS also benefits; ifthey buy local they pay the same due to IPP, but then local producers, not theSARS, benefit). 41 If South African producers are internationally competitive, they willsurvive under lower or zero tariffs. It will certainly require the removal of tariffs onall inputs and some attention to IPP in the steel, aluminium and plastic sectors. Allof these actions will contribute to lower domestic prices in this and a much widerrange of sectors. It will also raise the incentive to export.CS’s last conclusion can be further sharpened by drawing on the formalanalysis in this study.4.2 The wheat-flour-bread value chain(Simon Hobson)Before we proceed with the implications of Hobson’s (2006) case study ofthe wheat-flour-bread value chain for the price transmission mechanismin this sector, we want to discuss a table similar to the decomposition ofthe price of small household appliances in Section 4.1 – Table 9.Based on Table 9 we see that the consumption price of the basket ofhome-produced white bread, , consists for 76% of the price of locallyproduced white bread, (we assume that importing large quantitiesof fresh white bread is not feasible), plus the retailer’s margin(12%) and VAT (also 12%)., in turn can be split into two shares, and allocatedto the prices of home- and foreign-produced flour respectively, isthe mark-up (margin) of the baker. Based on Hobson’s Figure 2, we cancompute this to be about 20%. 42Based on Table 8, we know that 80% of small household appliances are imported.We also know that the ‘producer price level’ – or the price index of home-producedappliances - in turn consists of, say, 10% imported intermediate components (steel,plastics, aluminum, elements and switches) and 10% home-produced intermediategoods aggregates. Taking this into account it is clear that the first priority withrespect to policy reform should be doing away with the tariff on the relevant finalgoods (the output tariffs). This will have the largest beneficial impact (‘biggest bangper buck’) on the consumption price, namely -0.80. That will automatically make itless attractive for domestic producers to stick to their policy of marking-up prices tothe duty-inclusive price for (equivalent) products sold in South Africa – to continueengaging in IPP strategies for final goods.The second priority should be doing away with the tariff on the relevant intermediateinputs (the input tariff on steel, plastics, etc.). The reason is that here we areonly talking about 20% of the total consumption basket price. Again, the scrappingof the input tariff will then make it automatically less attractive for domesticproducers of intermediate goods (plastics, steel, aluminium) to persist in followingthe IPP model.Home-produced flour, in turn, can be milled using either home-producedor imported wheat, with shares and respectively.Here the miller’s margin is symbolised by . Therefore, the price ofhome-produced flour can be split into three components. Home-producedwheat (with share ), imported wheat (with share ) andthe profit or mark-up component (with share ). Based on the NationalAgricultural Marketing Council’s (NAMC’s) (2004) estimate of the impactof the wheat tariff on the price of bread at 0.14 – and assuming local producersof wheat price according to IPP – we compute this mark-up to be54% (!) of the price of flour. We will obviously return to this issue below.We do not have any information on the share. Therefore we simplyassume the remaining 80% of the baker’s price to consist of 40% of theprice of home-produced flour and 40% of the price of imported flour (thatis, we assume that). That means that the equation forwill read:(4.1)<strong>38</strong>Another possibility is a positive spike in the exchange rate like the one that happened inSouth Africa after the rapid depreciation of the rand in 2001.39A rational pricing rule for those intermediate goods producers would be to set their priceaccording to where is the(exogenous) price of the locally produced intermediate input not linked to the price of theimported intermediate input and is the operator that returns the largest value in aset of values.40The effective rate of protection (ERP) is a standard indicator of protection used in internationaltrade policy analysis that accounts for tariffs on outputs ( ) as well as those oninputs ( ).41Remember that 80% of those small appliances are imported.42From this figure we know that this margin is 15% of the retail price of bread.Since the baker’s price is 76% of the retail price (12% is VAT and another12% is the retailer’s margin), we can compute the baker’s margin in terms ofthe baker’s price to be 0.197 or about 20%. Formally, the argument goes as follows:from and we can infer that.Here the superscript ^ indicates an absolute value.
TABLE 9: PRICE TRANSMISSION IN THE WHEAT-FLOUR-BREAD VALUE CHAIN*0.76 x (price of home-produced white bread) = 0.76 xx (price ofhome produced intermediategoods) = 0.40 x(price of importedintermediate goods) = 0.40 x* Based on Figure 2 of Hobson (2006) and NAMC (2004)83FIGURE 4: THE COST OF WHITE BREAD<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>We know that there is a tariff on imported flour. Assuming foreign producersof flour engage in PCP, 43 the expression for can be expandedinto:(4.2)by more than 3% (and vice versa). So, it seems that – based on the interestof the poor consumer – policy priority number one would be to cut orrepeal the tariff on flour.We know that wheat is priced according to IPP. So, the last row of Table8 will read:where is the tariff (duty) on imported flour. Combining this expressionwith the one for the retail price (first row of Table 9) we get:(4.3)From this expression we see that the tariff on flour has a significant impacton the retail price of bread;! Thatmeans that if the relevant tariff is lifted by 10%, the retail price goes upAssuming foreign wheat producers engage in PCP and incorporating atariff on wheat we get:(4.4)Combining equation (4.4) with the retail price (first row of Table 9), weget the final expression for , namely:43This assumption is consistent with the fact that according to Hobson (2006), there appearsto be significant pass-through from the exchange rate of the rand to rand-denominatedimport prices as South African consumers seem to have benefited from the appreciatingrand.(4.5)<strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty Reduction84<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>From equation (4.5) we see that the tariff on wheat also has a significant impacton the retail price of bread; . 44This elasticity has been estimated by NAMC (2004) to be equal to 0.14,which is substantially lower than our calibrated impact of a tariff on flour(0.304). This means that if the tariff is lifted by 10%, the retail price ofbread goes up by 1.4% and vice versa.From the final expression for the retail price we can deduce a very importantinsight with respect to exchange rate pass-through. It can be seenthat there are two channels here.The first channel is what we call the ‘flour channel’. If the exchange rateappreciates by 10%, say, the rand price of flour goes up proportionally.This affects the retail price of bread by 0.76 x 0.40 x 10%, being theproduct of the share of the baker’s price in the retail price of bread(0.76) and the share of imported flour in the baker’s price (0.40). 45 Thus,via the first channel a 10% nominal depreciation will cause the retail price ofbread to increase by about 3% (). Of course, the size ofthis effect is identical to that of the impact of a tariff on flour.The second channel could be labelled the ‘wheat channel’. A nominal depreciationwill cause the rand price of imported wheat to increase. Then,given IPP, local producers will also increase their rand selling price. As aconsequence, a 10% nominal depreciation will cause the retail price ofbread to increase by 0.76 x 0.40 x x 10%, being the product ofthe share of the baker’s price in the retail price of bread (0.76), theshare of home-produced flour in the baker’s price (0.40), and the share ofthe rand price of wheat in the price of flour . NAMC (2004) hasestimated this elasticity to be equal to 0.14. This means that a 10% randdepreciation via the ‘wheat channel’ will cause the retail price of bread toincrease by 1.4%. Of course, the size of this effect is identical to that ofthe impact of a tariff on wheat.Here the impacts of the first and second channel are different, namely0.304 and 0.14 respectively. Note that the first effect is calibrated and thesecond estimated by NAMC (2004). 4644This elasticity has been estimated by NAMC (2004) to be equal to 0.14. Assumingthis reflects the truth – and ceteris paribus – then we can calibratefrom, which implies thatequals 0.54, suggesting a huge mark-up at the milller’s stage of the value chain.To be more specific, the relevant pricing equation under IPP would then read.45Of course, this is based on an assumption as we have no empirical information to helpcalibrate this share.46Thus, the size of the first effect is based on our assumption regarding the shareof home-produced flour in the producer price of bread. As this share goes up(meaning that goes down), the impact of the nominal exchange rate on and thuson will become lower.But, all in all, the consumer will be faced with the sum effect of bothchannels, which here equals 0.304 + 0.14 = 0.44! This is a big number.It indicates that a 10% nominal depreciation of the rand will – by virtueof both the flour and wheat channels – push up the retail price of breadby almost 4.5%.Here are our conclusions for price transmission (based on drawing on theinsights from our conceptual framework and those of the Hobson (2006)case study):Bread is an increasingly important component in the poor consumers’ basket. 47Foreign wheat producers appear to engage in PCP (perhaps not for 100%, butthere appears to be significant pass-through from the exchange rate to rand denominatedimport prices) as South African consumers seem to have benefited fromthe appreciating rand.Wheat is mainly produced for human consumption in South Africa and there istherefore a fairly direct value chain from producer to consumer. According to Hobson(2006, p. 6) the essence of the value chain runs from the producer to the millersto the bakers to the retail sector and on to final consumers. Figure 5 [taken fromHobson (2006, p. 6)] illustrates.In the current deregulated market the producer price of wheat is determinedthrough the South African Futures Exchange (SAFEX). This means that grain producers,traders and processors now trade in a ‘free’ market and must respond toworld-wide supply and demand forces in setting prices.These forces include weather conditions, consumer preferences, government policy,trade agreements, exchange rate fluctuations, changes in living standards and technology.Given that South Africa is a relatively small player on the world market and a netimporter of wheat, producers are subjected to global pricing pressures and generallyfollow import (or export) parity pricing calculations to determine prices. There is,for example, little incentive for wheat grain millers to pay South African producersmore than what they can import wheat for (including the cost of transport, insurance,the tariff and the exchange rate). This is called an import parity price andeffectively becomes the maximum producer price.Relatively detailed quantitative information is available on the impact of changingthe tariff level and structure is available from the Bureau for Food and AgriculturalPolicy Research (BFAP) and the Provincial Decision-Making Enabling (PROVIDE)Project who did an economy-wide computable general equilibrium model (CGE).The Food Price <strong>Monitor</strong>ing Committee [NAMC (2004)] also undertook a modelingexercise to determine the value adding and price transmission at the five levelsof the value chain as set in Figure 5. This was updated in the 2005 BFAP report.Hobson (2006, p. 13) finds that the cost of raw material flour has the largest sharein the retail price.The cost of the raw material flour is followed closely by the cost of baking. 48 Figure4 (the pie chart) illustrates.47According to Nzeni and Flatters (2006), expenditure on food makes up about half of allexpenditure by very low and low expenditure households Very low expenditure householdsspend up to R8,071 and low expenditure households spend between R8,071 and R12,263[Nzeni and Flatters (2006, p. 6)].48However, if one takes into account both the cost of baking and the baker’s margin, then themaking operation in the supply chain has the largest share in the retail price of bread.
FIGURE 5: THE WHEAT-FLOUR-BREAD VALUE CHAINProducerMillers85BakersRetailers<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>ConsumersWith regard to price transmission further up the value chain, it is generally recognisedthat changes in farm and wholesale prices are not evenly transmitted toconsumer prices. A common feature here is that retailers may quickly pass on priceincreases, while price decreases have a longer lag and/or are not fully passed on tothe consumer [Hobson(2006, p. 13)]. 49 The extent of this will largely be determinedby factors such as the relative level of concentration and competition in the particularsegment of the value chain. 50The same model was also used to estimate the impact of an import tariff on theretail price of bread ( ). In essence, an input tariff ( ) of say R300per ton of wheat (that is, a 23% increase from R1,300 per ton to R1,600 per ton)will only result in an increase of 3.2% in the price of white bread (in terms of thenotation of this paper we have). Hobson(2006, p. 14) says that the evidence suggests that there are no parties in the valuechain that extract ‘super profits’.49The BFAP report used a time-series model to test for asymmetric price transmission andconcluded that the retail price of bread ( ) is indeed closely related to the SAFEX wheatprice ( in the notation of this study) although there is a lag of approximately fourmonths.50Studies on bread have shown that since the abolition of price control in 1991, the retailsector share of the actual bread price has been increasing; it grew from 3% up to 12% in1998/1999 [Hobson (2006, p. 13)]. Importantly, the BFAP (2005) report concludes that the SAFEX price formation processfor wheat is efficient and is a true reflection of prices for the domestic wheatmarket [Hobson (2006, p. 14)]. At the producer level South African wheat farmers are in essence being squeezedby the impact of the strengthening rand on the one hand and the heavily subsidisedworld price of wheat, which allows millers to import at a price lower than the SouthAfrican cost of production. 51 However, our distance from the rest of the world provides a buffer against foreigncompetition. This is especially true for farmers in the interior, who are also closer tothe most important consumer markets. According to Hobson (2006, p. 18), commercial producers (of wheat) are generallyworse off due to the decrease in protection, increased competition from importsand decreasing real prices for wheat. While millers are negatively impacted by the protection on wheat which has theeffect of increasing their input prices, they are positively affected by the duties onimports of flour. The miller’s effective protection is positive given that the ad val-51In terms of the notation of this paper this can be formalised as follows:. Here is the exogenous dollar price in the absence of a subsidy,and is the subsidy in foreign currency.<strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty Reduction86<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>orem equivalent rate on flour is higher than on wheat. In this regard Erasmus andFlatters (2003) argue that the milling industry is the main beneficiary of protectionfor the wheat and flour industries.It appears as though the milling industry has had sufficient flexibility and power toat least protect its margins, while the baking and retail parts of the value chain havesecured a greater portion of the value added over time [Hobson (2006, p. 18)].The available evidence indicates that the benefits from deregulation and trade liberalizationhave generally been passed down through the chain with the result thatconsumers as a whole have benefited due to the decreasing real price of bread. Thisis benefiting more poor consumers as well.The above positive outcome has however played out in more recent times withinthe environment of a strengthening rand. Consumers will also be exposed to apass-through of rapidly increasing prices should the rand weaken substantially.Hobson’s conclusions can be further sharpened by drawing on the formalanalysis in this Section. Here are our conclusions:The tariff on flour has a significant impact on the retail price of bread;! That means that if the relevanttariff is lifted by 10%, the retail price goes up by more than 3% (and vice versa).So – based on the interest of the poor consumer - policy priority number 1should be to cut or repeal the tariff on flour. This recommendation willbe supported below by the finding – contrary to Hobson (2006) – thatthere is a party in the value chain that extracts ‘super profits’, namelythe millers. Thus, the presence of the tariff on flour (on the importedintermediate input) allows local producers of flour (millers) to extractan extraordinary profit margin, which is ultimately at the expense ofthe poor consumer.The tariff on wheat also has a significant impact on the retail price of bread;(this elasticity has been estimated by NAMC (2004)).This means that if the tariff is lifted by 10%, the retail price of bread goes up by1.4% and vice versa.Based on this NAMC (2004) estimate we can then compute the share ofthe miller’s mark-up (margin) in the price of home-produced flour to beequal to 54%! This suggests a huge mark-up at the milller’s stage of thevalue chain. 52 Note that the finding of super profits at the miller’s stage of thevalue chain is consistent with (i) the fact that millers are protected from foreign competitionby the tariff on imported flour which allows them to raise , and(ii) that they face low cost of locally produced wheat because local wheat farmersare being squeezed by global pricing pressures (and a strong rand) which keep the(rand) wheat price, , low. 53From the final expression for the retail price we can deduce a very important insight52To be more specific, the relevant pricing equation under IPP would then read.53Formally, the miller’s profit margin (in absolute terms) is given by. From thiswith respect to exchange rate pass-through. It can be seen that there are twochannels here.The first channel is what we call the ‘flour channel’. If the exchange rate depreciatesby 10%, say, the rand price of flour goes up proportionally. Via the first channel a10% nominal depreciation will cause the retail price of bread to increase by about3% ( ). Of course, the size of this effect is identical to that ofthe impact of a tariff on flour.The second channel could be labelled the ‘wheat channel’. A 10% rand depreciationvia the ‘wheat channel’ will cause the retail price of bread to increase by 1.4%.The size of this effect is identical to that of the impact of a tariff on wheat.But, all in all, the consumer will be faced with the sum effect of both channels,which here equals 0.304 + 0.14 = 0.44! This is a big number. It indicates that a10% nominal depreciation of the rand will – by virtue of both the flour and wheatchannels – push up the retail price of bread by almost 4.5%.4.3 The clothing industry (Van derWesthuizen)Before we proceed with the implications of Van der Westhuizen’s (2006)(hereafter VW) case study of the clothing industry for the price transmissionmechanism in this sector, we want to discuss a table similar to thedecomposition of the price of small household appliances in Section 4.1Table 10.First of all it should be stressed that Table 10 is highly speculative as it isbased on one interview quoted by VW (2006. p. 25). If this interview isrepresentative of what’s going on in the industry than the table is not toofar off either. VW (2006, p. 25) mentions that in the clothing industry thevalue chain is (retail) buyer driven without much space for the manufacturerto manoeuvre. 54Therefore, small household-based manufacturers or CMTs are forced toaccept the prices offered by retailers. Competition in this sector is thus ofthe ‘race to the bottom’ variety, with CMTs engaged in a ‘price war’ witheach other (and the Chinese presumably). VW (2006, p. 25) mentions anexample of the small cut that a CMT may receive of the final price. Ofa woman’s shirt retailing at R249, a KwaZulu-Natal CMT would receiveR18.50. This would be about 7.4% of the selling price. Given VAT, thatwould leave a margin for the retailer of 78.6% of the retail price! Thenumbers in Table 10 have been calculated with that example in mind(since there was no other quantitative information in the case study todraw on). We assume that the pricing structure would be similar if thegarment was sourced from China. Further, we do not know the share ofexpression we see that the margingoes up, the lower the inputcost, , given the selling price , or the higher the selling price giventhe input cost. Here the first effect is supported by the low world price of wheat and IPP,and the second by the tariff in imported flour.54Interestingly, in the case of the motor trade we will find exactly the opposite. There theindustry is manufacturer driven, without much leverage of the dealer or ‘car-retailer’. SeeSection 4.4 below.
imports (and therefore of home-produced) garments in the consumptionbasket of total clothing, and therefore set this share to 50%. 55Here are our conclusions for price transmission (drawing on the insightsfrom our conceptual framework and those of the VW (2006) case study: Foreign (Chinese) competitors appear to engage in PCP. We base this on VW (2006,p. 18) who states that benefits that accrue to Chinese producers are – inter alia– a fixed exchange rate undervaluing the Chinese currency by up to 40% (onlychanged in the second half of 2005) and lower labor costs in the absence of ahuman rights regime. As a consequence of PCP and the stronger rand (during and after the recoveryfrom the 2001 low) imports from China jumped from 85-million units in 2001 to335-million in 2004. Cheap imports from (mainly) China caused a price deflation within the industry. Thisis evident from Table 11, which shows that the rate of change of prices of consumerprices of clothing and footwear has been negative since 2001. As elsewhere thishas forced adjustments in the South African industry, especially by manufacturers,as buyers now had the choice of imported goods at lower prices, and manufacturerswere therefore competing with such goods. As happened also in the small household appliances sector [see CS (2006)], oneresponse by domestic manufacturers to a deteriorating terms of trade (a lowernumber for ) has been to shed or reduce the manufacturing functionand thereby the labour component partially or wholly, 56 leading to some manufacturersbecoming design houses or importers. This particular adjustment (‘downsizing’through retrenchments and casualisation) has led to the assembly of apparel(garments) being externalised (‘outsourced’), including to the household, either asa cost initiative by large manufacturers or as a livelihood strategy by retrenchedworkers [VW (2006, p. 21)]. The second response strategy, which was supported with government incentives,was to reorient production from the domestic to the export market. Another important aspect of the clothing industry – which is completely the oppositeof the corresponding feature of the small household appliances sector – is thatit appears to be very labour intensive [VW (2006, p. 37]. In terms of the notationof this study this means that the unit labour cost componentof theprice of the home-produced intermediate input aggregate is potentially large relativeto the ‘other’ component (material costs and profits). However, in practice it appears to be the case that wages are concerned to bea residual, rather than an independent driver of the producer price of clothing.The reason is that as the value chain is buyer driven, without much space for themanufacturer to manoeuvre. Suppliers of clothing to retailers – often CTMs or smallhousehold-based clothing manufacturers – are forced to accept the prices offered55To be more precise, if the garment is home produced we assume it to be priced accordingto:If it was sourced from China we assume the price is set as:Now, assuming50% of clothing is home-produced, we take the average of these pricing equations andthen get:which is the equation that is reported in Table 10 .56VW (2006, p. 29) reports that the strong rand also played a part in making producing forthe domestic market unsustainable.by retailers. 57 This means that – using the notation of this study – they have nosay over , so is basically an exogenous variable from theirperspective. Given the fact that this price depends on a unit labour cost componentand a margin (mark-up) component, and the fact that about 700 small operatorsmay be prosecuted for not paying the minimum wage as agreed in the clothingbargaining council (BC), 58 it appears that wages are considered a residual (afterprofits have been set by CMTs given the selling price to the retailer).As the value chain is buyer-driven, the retailers are the most important actors inthe clothing value chain. Power is located at the end of the pipeline as the retailersdetermine the orders and the prices, causing surpluses (mark-ups) not to be distributedthrough the pipeline but to be located at the retail end.A plausible theory on what now determines the price , is (again) the practice ofIPP. If Truworths can import clothing from China why would it pay a KwaZulu-NatalCMT more than what they import those goods for (including the cost of transport,insurance, the tariff – if any – and the exchange rate)?The market in South Africa is dominated by a handful of large companies on whichmanufacturers have been dependent, and which have shown large increases insales (between 20% and 30%) between 2004 and 2005. The largest of the retailers,Edcon, improved its headline earnings with 76% between 2004 and 2005, in acontext of lower interest rates and a consequent boom in retailing since 2003.We have no quantitative information on the share of imported final pieces of clothingin the retail price index of clothing. The relevant information that is mentionedin VW (2006) points in different (and contradictory) directions. On the one hand,she quotes Tony Taylor, Deputy Managing Director of (the large retailer) Truworthssaying that ‘Importing is regarded as a headache, which is why Truworths importsonly what cannot be produced locally.’[VW (2006, p. 28)]. On the other, she reportson units of clothing imported from China which shows a very clear upward trend(up from 85-million units in 2001 to 335-million in 2004). We have more confidencein the numbers than in the statement of (the representative of) an interestgroup. Therefore, in the remainder of this Section we assume that this share is large;definitely bigger than 50%. This means that we assume that the majority of clothingis imported and not locally produced 59 .This assumption is also consistent with the sharp drops observed in consumer pricesof clothing and footwear (see Table 11). Remember that based on the definition ofthe (macroeconomic) CPI we have(2.2), where therand price of the imported final good (clothing) is set according to PCP. Thus, in orderto observe price deflation in this sector concurring with an undervalued Chinesecurrency and a strong rand, 60 the share of imported clothing in the total clothing‘basket’ has to be substantial. 6157According to VW (2006, p. 23) CMTs are the product of two phenomena associated withthe increased competition in the clothing market: first, manufacturers downsizing to ‘core’value-adding functions such as design and marketing, and secondly, the closures of manufacturersunable to cope with the concomitant price pressures. In the first case, retrenchedworkers are set up with equipment in the household by erstwhile employers, thus creatingCMTs. In the second case, retrenched workers would start CMTs as a survival strategy to‘put food on the table’.58See VW (2006, p. 24).59In line with the facts on small household appliances where 80% is imported, and not inline with the wheat-flour-bread value chain where we assume that most – if not all – freshbread is locally produced.60In fact, what we are seeing here is a substantial real appreciation (increase in) as the nominal exchange rate of the rand appreciates ( low or negative)driven by an undervalued Chinese currency and a recovering (appreciating) randfollowing the 2001 low.61Note that the ‘price deflation effect’ of cheap imports on the consumer price of clothing87<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong><strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty ReductionTABLE 10: PRICE TRANSMISSION IN THE CLOTHING INDUSTRY*0.035 x (prices of homeproducedapparel) = 0.035 x+ 0.035 x (prices of importedapparel) = 0.035 x+ 0.79 x +0.14 x VATTABLE 11: CONSUMER PRICES, CLOTHING AND FOOTWEAR: METROPOLITAN AREAS*881995 1996 1997 1998 1999 2000 2001 2002 2003 20044.9% 3.9% 3.2% 1.0% 1.0% 0.9% -0.5% -1.3% 1.6% -3.8%<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>VW’s conclusions can be further sharpened by drawing on the formalanalysis in this Section. Here is our final conclusion:Based on Table 10 – that is, based on the VW (2006, p. 25) example and the assumptionof a 50-50 split between home- and foreign-produced garments – we getthe result of an enormous retail margin: about 79% (of the retail price). Thissuggests that the big winners of trade liberalisation have been the retailers,and perhaps to some extent also the consumers as the strengtheningrand and an undervalued Chinese currency have resulted in lowerretail prices for clothing and footwear.4.4 The motor trade industry (Nzeni andFlatters)Before we proceed with the implications of Nzeni and Flatter´s (2006)case study of the motor trade for the price transmission mechanism in thissector, we want to discuss a table similar to the ones we have used in theother case studies – Table 12. Based on Table 12 (which abstracts fromimported final goods) we see that the most important driver of consumerprices of home-produced cars is the manufacturers´ selling price, . Thisis consistent with the result we find in this section that the value chain inthe motor trade is manufacturer driven.Here are our conclusions for price transmission (drawing on the insightsfrom our conceptual framework and those of the Nzeni and Flatters (hereafterNZ) (2006) case study:*Source: VW (2006, p. 16).The motor trade appears to provide relatively well-paying employment.According to the Competition Commission, it was clear that manufacturers notonly monitored dealer’s prices carefully but also imposed fines on them if they soldbelow agreed prices, or granted higher discounts than allowed. That means thatthe value chain is manufacturer driven where – using the notation of this studyand footwear would become even more pronounced if was set by the retailers in linewith IPP. Why would they pay South African producers (CMTs, say) more than whatthey can import clothing for from China! In the case of IPP by retailers the relevant priceindex (in line with equation (2.2) and abstracting from VAT and margins) would read(where ).– the manufacturer makes sure that , where is the rand retail price(dealer’s) price of a new vehicle. 62The Commission also found that in the large merger case between UnitransMotors and the motor division of Senwes, manufacturers dictated the marginsmade on their product, the number of cars sold and even how the carswere sold. Using the notation of this study that would suggest that we have, whereis the mark-up (margin) of the dealer (set by the manufacturer) here expressedas a share of the selling price .Note that given this share and the VAT percentage, the shareis also determined. For example, assume the marginis 12%. Then, , the share of the manufacturer’s selling price (in terms ofthe dealer’s price) is 0.74. Therefore, given a level set for by setting the dealer’smargin the manufacturer actually fixes his own margin.There is little doubt that the actions of vehicle manufacturers stifled competitionin the motor trade.The position of dealers is worsened by the loss of sales to manufacturers. The percentageof new vehicle sales that pass through dealers fell from 81.9% of the marketin 1999 to about 79.2% in 2002, while there were increases in direct sales bymanufacturers to government and rental firms. Manufacturers are also increasinglyopting to integrate vertically by distributing their own cars directly.From this we infer that it is likely that (absolute values of) margins of dealers aredecreasing in the degree of vertical integration (or dealer disintermediation) bymanufacturers. Thus, the value chain is (increasingly) manufacturer driven (evenmore so over time, as is evident from the 1999-2002 developments), with manufacturerstrying to increase their share of the (final selling price ‘pie’) at the expenseof dealers.Second-hand vehicles in South Africa are over-priced relative to other markets.Moreover, South Africa only permits the import of second-hand vehicles undervery special circumstances like returning diplomats’ vehicles, expatriates’ vehicles,hearses and the like. 6362It is estimated that in the US it takes a household 20 working weeks to cover the price of anew car (including finance costs), compared to 70 working weeks that a high-income SouthAfrican family needs to purchase and finance a similar car.63In terms of our earlier discussions this could be modelled as a tariff on imported secondhandvehicles (final goods), say , where makes the imports prohibitivelyexpensive. Moreover, vehicle ownership density (ownership per 1,000 people) in SouthAfrica remains 108 compared to 450 in mature markets. Clearly, an opportunity to grow
TABLE 12: PRICE TRANSMISSION IN THE MOTOR TRADE *= =xhome-manufactured cars)(price index of+ x + 0.14 x VAT* Abstracting from imported carsThe high cost of second-hand vehicles constitutes a bulwark in the assault on poverty.Often vehicles the poor can afford are old, barely roadworthy, costly to maintain andgenerally unsafe. In such communities opportunities to generate income throughtransporting people and goods are available. However, aspirant entrepreneurs areconstrained by high prices of second-hand vehicles.According to the Stats SA expenditure survey of 2000, the bulk of low-income andvery low-income consumption services goes to minibus taxis and hired transport.The National Household Travel Survey has revealed that almost half of the householdswith a monthly income of R500 or less spend more than a fifth of theirincome on transport!It is safe to assume that vehicles owned by the poor are not new. This impliesthat the cost maintaining such vehicles are key. Historically, imports of componentsdestined for the after market faced lower duties compared to original equipment.However, the aftermarket components and accessories faced a much slower tariffphase down. Nzeni and Flatters (2006, p. 10) point out that vehicle parts have notfaced tariff reduction of any significance. Indeed, the phase down of tariffs is setto decline only to about 20% in 2012 under the Motor <strong>Industry</strong> Development Plan(MIDP). This places independent distributors of such parts at a disadvantage relativeto domestic vehicle assemblers. 64 The latter can import the parts duty free usingIRCC while the former do not qualify for IRCCs because they are not exporters.There is no evidence to suggest that assemblers pass on the benefits to consumers.Generally, the prices of independent distributors are lower than those of vehicleassemblers [NZ (2006. p. 11)].NZ´s conclusions can be further sharpened by drawing on the formalanalysis in this section. Here is our final conclusion:Based on Table 12 – that is, based on the finding that the value chain in the motortrade is manufacturer driven – in some cases to the extent that manufacturersdictate the margins of their products – we find that the big winners of the MIDP (ofprotectionist trade policies vis-à-vis the domestic industry) have been the domesticmanufacturers (producers) and the big losers the South African consumers in generaland poor consumers in particular.the market exists, particularly for cheaper (home-produced) second-hand vehicles [NZ(2006, p. 7)].64Imported cars generally have a higher-than-locally-produced-cars parts to price ratio.This is the result of the aforementioned tariffs. For example, the Tata Indica’s parts basketconstitutes 45.1% of its price [NZ (2006, p. 11)].5. Conclusions and policyrecommendationsThe master conclusion that can be drawn from this study is that protectionistpolicies – the presence of tariffs on imported goods – are very badnews for the consumer in general, and the poor consumer in particular.First, a tariff directly increases the rand price of imported goods, muchlike a weaker exchange rate – following a nominal depreciation – woulddo. Here we can distinguish between tariffs on imported final goods, theso-called output tariffs, and tariffs on imported intermediate inputs suchas chemicals and steel, called input tariffs.Obviously, the adverse price impact of a tariff is larger – again in line withthe effect of a nominal exchange rate depreciation – the more importantthat good is (or the larger its share in the (poor) consumer’s overallbasket. For example, food prices are very important in their potential effectson poverty (relief) as expenditure on food makes up about half ofall expenditure by very low and low expenditure households (householdsthat spend up to R8,071 and between R8,071 and R12,263 respectively).Another example of an important final good (or rather service) in thepoor consumer’s basket is transport. The National Transport Survey hasrevealed that almost half of the households who earn R500 or less permonth spend more than a fifth of their income on transport. We will returnto food prices and transport services when we present our policy recommendations.We have shown that a tariff is more problematic for price transmission themore important that particular price – or its induced effect, given certainpricing strategies such as import parity pricing – for the real income ofthe consumer. Similarly, for the South African economy as a whole a tariffhas a higher impact on the overall level of the consumer price index thelarger the share of that final good or service in the CPI. Now, if we lookat the macroeconomic architecture of the South African CPI, it followsthat only 10% of all final goods and services are imported. Therefore,the presence of tariffs on outputs or final goods, say, is not the majorthreat to an orderly price transmission in general, or to poverty reductionin particular. Rather, the crux of international trade in South Africa is trade89<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong><strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty Reduction90<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>in intermediate inputs, such as chemicals and steel. Here, we find that ona macroeconomic level about 30% of such inputs are imported. Therefore,the potential adverse impact on poverty via higher rand prices of importedintermediate inputs is much larger than that of more expensive foreignfinal goods (as the latter effectively constitutes only 10% of the CPI).Of course, in the end the effects of imported final and intermediate goodsprice inflation interact, as illustrated by Table 13, which shows that thecombination of 10% imported final goods (row 1) and 30% importedintermediate goods (row 2) together implies that the largest part of theCPI (row 1) is driven for 63 (0.7 x 0.9 x 100)% by the prices of homeproducedintermediate goods (row 3). The remaining 37% comes fromabroad – is ‘imported inflation’ – and is split between 27% importedintermediate goods inflation (row 3) – and 10% imported final goodsinflation (row 3).5.1 General conclusions and policyrecommendationsSo, using this decomposition as a guideline it is clear that scrapping alltariffs on imported final goods will have quite a small impact on the CPI,and therefore on poverty. Much more can be achieved by repealing tariffson imported intermediate inputs where the impact on poverty relief willbe much more substantial. 65So, the first policy recommendation is that more gains are to be made for the domesticconsumer by lowering tariffs on intermediate goods than on final goods. Thenumbers in Table 13 suggest that those gains are almost three times as large as anequivalent drop on tariffs on final goods. 66Another very important result we found in this study is that the presenceof a tariff – either on final or intermediate goods – lifts the rand price ofthe imported good above the rand price of the home-produced good. Itbecomes a rational pricing strategy of the domestic (South African) producernot to charge below the cost (to the consumer) of imported goodson which duties are charged. This is called IPP. Then the only options availableto South African consumers (or to South African downstream producers)are to pay the duty-inclusive price on local goods or purchase importsand pay the duty anyway.IPP is unfortunately a well-established business practice in South Africa.For example, the case study on the wheat-flour-bread value chain [Hobson65Of course, there are exceptions. For some goods, for example, small household appliances,the home-produced component is quite small (in fact about 20% of the basket). So, in thatcase the policy priorities are reversed: first repeal the tariff on imported final goods and thenon imported intermediate inputs.66For the supporting analysis see Section 3.1.(2006)] finds that wheat is priced according to IPP as millers have littleincentive to pay South African farmers more than what they can importwheat for. Similarly, in our assessment of the case study on the clothingindustry [Van der Westhuizen (2006)], we find that it is plausible that inthis (retail) buyer-driven value chain, retailers also engage in IPP vis-à-vistheir suppliers (manufacturers). The case study on small household appliances[Chaponda and Stern (2006)] finds that South African producers ofplastics, steel and aluminium also follow the IPP model. Finally, the sharprise in the prices of South African-produced cars following the rapid depreciationof the rand in 2001 is also completely consistent with the logicof an IPP strategy. The reason is that it does not matter what drives thehigher imported rand price; a hike of the tariff or a weaker currency havea similar effect in that they both push up the rand price of the importedfinal good. In this study we show that the presence of import parity pricing increases the adverseprice impact of tariffs – either on intermediate or final goods – (or of a weakerrand) on the CPI and therefore on poverty.Consider the important case of IPP for imported intermediate inputs. WithoutIPP the higher tariff would increase the CPI by 27% of the change inthe tariff (being the share of imported intermediate inputs in the consumptionbasket, indicated in row 3 of Table 13). But now the higher randprice of imported intermediate inputs is followed by a parallel increasein the rand price of home-produced intermediate goods (which are 63%of the CPI). When those goods are priced to parity with the prices of importedintermediate inputs, the effects of tariffs on those inputs on the CPIwould increase from 0.27 (in the absence of IPP) to 0.9. This means that with import parity pricing the adverse effect of input tariffs on theCPI is now more than three times as large (or 0.9/0.27 = 3.33)!IPP has another extremely negative influence on the South African pricetransmission mechanism: it substantially increases the adverse effect of anexchange rate depreciation on the CPI and hence on poverty. This effectis important as it has a real-life counterpart in the form of the 2001 depreciationof the rand. As a consequence of IPP of imported intermediateinputs the adverse effects of a rand depreciation on the CPI are almostthree times as large as in the absence of IPP.Further, note that the reason why South African producers engage in IPPin the first place is the presence of tariffs on intermediate inputs, whichraises the rand price of those inputs above the autonomous price of thehome produced input (the price in the absence of IPP). So, the second policy recommendation is that tariffs on intermediate inputs need tobe scrapped as soon as possible as they invite domestic producers to engage in IPP,which not only increases the adverse price effects of tariffs on the CPI (more than
TABLE 13: MACROECONOMIC PRICE TRANSMISSION (FOR THE AGGREGATE SOUTH AFRICAN ECONOMY)CPI = 0.90 x (prices of home-produced final goods) = 0.90 x PPI + 0.10 x (prices of imported finalgoods)PPI =0.7 x (price of home produced intermediategoods)+ 0.3 x (price of imported intermediategoods)CPI =0.63 x (price of home produced intermediategoods)+ 0.27 x (price of imported intermediategoods)+ 0.10 x (prices of imported finalgoods)91three times as large as in the absence of IPP), but also of a nominal rand depreciation(almost three times as large as in the absence of IPP). 675.2 Conclusions and policyrecommendations for sectors basedon case studiesThe above conclusions can be further sharpened by drawing on the casestudies on small household appliances, the motor trade, the wheat-flourbreadvalue chain and clothing and textiles.5.2.2 The wheat-flour-bread value chainFrom the case study on the wheat-flour-bread value chain we have learntthat here we have the opposite situation of the small household appliancessector in the sense that there is no imported component associatedwith (fresh) white bread. So all that matters for the retail price are localproducer prices and those of imported intermediate inputs.We find that the tariff on flour has a significant impact on the retail priceof bread, namely about 30% (30% of a tariff increase is passed on to theretail price).<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>5.2.1 Small household appliancesFrom the case study on small household appliances [Chaponda and Stern(2006)], we know that South African consumers continue to pay too higha price for small household appliances and that 80% of those appliancesare imported. Therefore it is clear that the first policy recommendation for the small householdappliances sector is to do away with the tariffs on the relevant final goods.That will make it automatically less attractive for domestic producers tostick to their policy of marking-up their prices to the duty-inclusive pricefor (equivalent) products sold in South Africa – to continue engaging inIPP strategies for final goods. The second policy recommendation for the small household appliances sector is toscrap the tariffs on the relevant intermediate inputs (steel, plastics, etc.).Again, the repealing of the input tariff will then make it automaticallyless attractive for domestic producers of intermediate goods to persist infollowing the IPP model. So, the first policy recommendation for the wheat-flour-bread value chain is to cutor repeal the tariff on flour.This recommendation can be supported by our finding that there is a partyin the value chain that extracts ´super profits´, namely the millers. 68 Anotherway of saying this is that the wheat-flour-bread value chain is drivenby the millers or by the ‘middle men’ (who are positioned between thefarmers and bakers/retailers). Thus, the presence of the tariff on flour (onthe imported intermediate input) allows local producers of flour to extractan extraordinary profit margin, which is ultimately at the expense of thepoor consumer.The tariff on wheat also has a significant impact on the retail price ofbread. There the corresponding number is 14% (14% of a tariff increaseis passed on to the retail price).As the effects of a nominal exchange rate depreciation are similar to thoseof an increase in tariffs, we can conclude that a weaker exchange rate willsaddle the poor consumer with a substantially higher retail price of bread.In the end, 44% (!) of the depreciation will work its way to the price ofbread (30% via the ´flour channel´ and 14% via the ´wheat channel´).67To be more precise the impact of a rand depreciation in the absence of IPP is 0.37 (or 0.27+ 0.10), and under IPP this increases to 1(!). The reason is that the remaining 63% ofprices (namely those of home-produced goods) also respond in a one-for-one manner to anominal depreciation.68In Section 4.2 we compute the share of the miller’s mark-up (margin) in the price ofhome-produced flour to be equal to 54%! This suggests a huge mark-up at the milller’s stageof the value chain. Note that the finding of super profits at the miller’s stage of the valuechain is consistent with (i) the fact that millers are protected from foreign competition bythe tariff on imported flour which allows them to raise the local price of flour, and (ii) thatthey face low cost of locally produced wheat because local wheat farmers are being squeezedby global pricing pressures (and a strong rand) which keep the (rand) wheat price low.<strong>Trade</strong> and Poverty in South Africa: The Roles of the Exchange Rate, Tariffs and Import Parity Pricing for CPI Inflation
Development & Poverty Reduction92<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>5.2.3 The clothing industryIn the clothing industry the value chain is (retail) buyer driven withoutmuch space for the manufacturer to manoeuvre. Therefore, small household-basedmanufacturers or CMTs are forced to accept the prices offeredby retailers.Cheap imports from China have caused a price deflation within the industry.As elsewhere this has forced adjustments in the South African industry,especially by manufacturers, as buyers now had the choice of importedgoods at lower prices, and manufacturers were therefore competing withsuch goods.As the value chain is a buyer-driven chain, the retailers are the most importantactors in the clothing industry. Power is located at the end ofthe pipeline as the retailers determine the orders and the prices, causingsurpluses not to be distributed through the pipeline but to be located as(very substantial) mark-ups at the retail end.A plausible theory on what determines these prices offered to manufacturersis (again) the practice of IPP. If Truworths can import clothing fromChina why would it pay a KwaZulu-Natal CMT more than what they canimport those goods for (including the cost of transport, insurance and thetariff – if any – and the exchange rate?Based on this case study – and some simple (but plausible) assumptions– we get the result of an enormous retail margin, namely about 79% ofthe retail price.goods (flour) producers (millers) in the wheat-flour-bread value chain anddomestic manufacturers of final goods (cars) in the motor industry. Interestingly,these rent-extracting opportunities seem to be made possible bythe practice of IPP by clothing retailers (due to trade liberalisation andcheap imported Chinese final goods) on the one hand and tariffs on flour(due to protection)/IPP by millers on the other.Second-hand vehicles in South Africa are over-priced relative to othermarkets.Based on the finding that the value chain in the motor trade is manufacturerdriven – in some cases to the extent that manufacturers dictate themargins on their products – we find that the big winners of the MIDP (ofprotectionist trade policies vis-à-vis the domestic industry) have been thedomestic manufacturers (producers) and the big losers the South Africanconsumer in general and the poor consumer (and aspiring entrepeneur)in particular.ReferencesAgénor, P. (2004) ‘The Economics of Adjustment and Growth’, HarvardUniversity Press.Chaponda, T. and M. Stern (2006) ‘<strong>Trade</strong> and Poverty Case Study: SmallHousehold Appliances’, January.Edwards, L. (2005) Not in refs Chaponda and Stern (2006)This suggests that the big winners of trade liberalisation here have beenthe retailers, and to some extent also the consumers as the strengtheningrand and an undervalued Chinese currency have resulted in lower retailprices for clothing and footwear.5.2.4 The motor tradeThe most important driver of consumer prices of South African-producedcars is the manufacturer’s selling price. This is consistent with the resultwe find in this study that the value chain in the motor trade is manufacturerdriven.Compare this result with the other case studies. For the clothing industrywe found that the retailers where ‘top dogs’ (the clothing value chain is‘buyer driven’), whereas in the wheat-flour-bread value chain the millers –who are in the ‘middle’ of the value chain – realised the highest margins.So, it appears we now have an example of each stage being dominant ina different industry – the retailers in the clothing industry, intermediateErasmus E. and Flatters F (2003) ‘Rent-Seeking in SADC <strong>Trade</strong> Liberalization:Rules of Origin and Other Barriers to <strong>Trade</strong> in Wheat Products’,April.Hobson, S. (2006) ‘The Effects of <strong>Trade</strong> Liberalisation on the Wheat-Flour-Bread Value Chain in South Africa’, <strong>Trade</strong> and Poverty Case Study, March.NAMC (2004) ‘Final Report of the Food Pricing <strong>Monitor</strong>ing Committee’,National Agricultural Marketing Council, Pretoria.Nzeni and Flatters, F. (2005) ‘The Motor <strong>Trade</strong> in South Africa’, <strong>Trade</strong> andPoverty Case Study, December.Van der Westhuizen, C. (2006) ‘<strong>Trade</strong> and Poverty: A Case Study of the SAClothing <strong>Industry</strong>’, January.
Southern African Development ResearchFunded by the IDRC93Current and previous donor initiatives on economicpolicy in Southern Africa have generally focused onproviding direct technical assistance in specific areas,or targeted capacity-building often focused on individualsrather than institutions. Whilst some of these initiatives havebeen successful in meeting short-term policy needs, they are notsustainable solutions to Southern Africa’s relatively weak policyand research communities.TIPS has undertaken a substantive needs analysis of both thepolicy and research communities. In addition, through ourAusAID-funded SADC <strong>Trade</strong> Development Programme – which isnow in its second year – we believe we have a reliable pictureof the research and policy environment in SADC. Whilst generalisationsshould be treated cautiously, policy and researchcommunities in SADC are characterised by: Poor integration with one another; Policy agendas which are large, growing and complex; Policy processes that are not always based on sound evidenceor research; Research communities that are under-resourced and notproducing a pool of technically skilled researchers; and Researchers who are not sufficiently skilled in new, emergingpolicy themes.We propose establishing a broad-based policy and researchnetwork, the Southern African Development Research Network(SADRN), which attempts to fill some of these institutional andskill gaps. The general objectives of SADRN will be to fill theinstitutional, skill and knowledge gaps conspiring against afruitful integration of SADC countries in the global economyand their capacity to promote inclusive growth.More specifically, SADRN aims to:i. Increase the supply of policy-relevant research in the SADCregion by creating a pool of suitably skilled researchersbased in institutions in SADC.ii. Improve the policy relevance of research through strengtheningthe capacity of policy-makers to be discerning research‘users’.iii. Develop an appreciation for evidence-based policy-makingby engaging policy-makers in the design, specification,implementation and review of research projects.iv. Build institutional capacity in key policy research institutesin SADC via the creation of Working Groups in focused thematicareas of research.However, the Network needs to balance two different priorities: The need to deliver, in the short run, policy-relevant researchwhich is useful to policy-makers in the national capitalsof the SADC countries; and The need to build, in the medium term, institutional capacitywithin policy research institutes and universities in theregion.Ideally, we would wish to sequence our activities such that capacityin the research community is first strengthened beforeattempting to draw policy-makers into the process. However,with policy-makers’ agendas already substantial and increasinglycomplex, there simply is not enough time to do this, andthe long-term credibility of the Network could be damaged ifit cannot demonstrate its value to the policy community at theoutset.We therefore propose an evolving role for TIPS in the Network.In the first 12 to 15 months, we will co-ordinate the delivery ofshort-run policy-relevant research. This will ensure that policymakerssee value in SADRN, thereby establishing its credibilityand their continued support. To establish strong accountabilityto policy-makers’ needs, we will co-ordinate an acceleratedround of stakeholder consultation through workshops andseminars with the objective of launching substantial researchprojects in three key areas.TIPS believes, based on the consultation already undertaken,that there is strong demand for research in the three thematicareas of: Regional industrial policy; <strong>Trade</strong> policy and pro-poor growth; and Service sector development for poverty reduction.These themes are discussed in detail in the proposal. Nonetheless,we view a further round of consultation as essential toestablish the modus operandi for how themes are chosen by theNetwork. This will also serve as an opportunity to draw the researchcommunity more firmly into the ongoing policy processes.Moreover, if this round of consultation confirms the relevance<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The Impact of the Asian Drivers Announcementson Sub-Saharan Africa’s Wooden Furniture <strong>Industry</strong>
Announcements94<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>of the three themes we propose, the workshops can be used todebate the sub-themes and their order of priority. These workshopswill be carried out in the first three months of the projectand will have as their outcome an agreed-upon formal workplanfor each research theme.However, in the long run, this is not a sustainable solution. TIPSdoes not have the capacity to undertake research for the entireSADC region and there will almost certainly be large questionsaround the credibility of a regional Network dominated by aSouth African institution. It will therefore be essential that TIPS,in parallel with co-ordinating short-term research, undertakesa programme of institutional strengthening in the region. Wepropose that we set up a programme of institutional assessmentsusing external expertise as well as self assessments forkey policy research institutions wishing to be part of SADRN.These assessments should culminate in a plan for institutionalcapacity-building, agreed between TIPS and the institution, andbased around ‘Outcome Mapping’ to ensure that this can bemonitored properly. As part of the agreed institutional strengtheningplan for each institution, we envisage staff exchanges,mentorships – not only with South African-based technical staff– training courses and the like. A key incentive for participatinginstitutions is the commitment to devolve selected researchprojects to the institutions as part of the strengthening programme.We envisage this happening via the establishment ofThematic Working Groups (TWGs). Alternatively, where capacityalready exists, an institution may wish to bid to launch a newTWG rather than take over an existing one.The TWGs will in time be responsible for policy/research co-ordinationand become the main mechanism for building capacityboth throughout the Network as well as within TWG host institutions.We would argue that this modality will increase thecredibility of SADRN, improve communication amongst researchpartners and key policy-making institutions (in part throughgeographical proximity) and become an effective channel fortraining and dissemination activities for the Network.We are mindful that SADC policy-makers require research supportnow and that a Network which promises much in the longrun but delivers little in the short run is a luxury that SADCcannot afford. We therefore propose a number of arrangementssuch as formal mentorships between established researchersand younger researchers to increase delivery potential in theshort term. In addition, SADRN will simply have to find a balancebetween building sustainable organisations in the mediumterm whilst dealing with immediate policy research needs. Thiswill not be an easy task but we believe that through a varietyof tools such as ‘wikis’ and online learning, an appropriate balancecan be found.The institutional assessments and agreed Plan, as well as otherpreparatory activities, will provide a solid foundation for implementingSADRN’s activities and should ensure that fragileinstitutions are not overwhelmed with the sudden demandsassociated with implementing large projects. Rather, we envisagea sequenced devolution of responsibility to TWGs as agreedmilestones are met and their capacity grows. In addition, SADRNwill assess the data needs and available resources that are relevantto the TWGs and establish online resource pages so thateach TWG is well-resourced at the outset.Over the first one to three years of the Network we expect to seethe following outputs. An increase in the quantity of policy-relevant research availableto, and undertaken by, researchers and policy-makersin SADC. An increase in the number of seminars, summer schools anddiscussion fora for policy-makers and researchers to interactand exchange views. An increase in the availability of (relevant) research resourcessuch as online databases, technical publications,policy briefs and methodological tools in SADC.Over the first one to five years of the Network we expect to beable to see the following outcomes being achieved. Increasingly integrated research and policy processes. Enhanced involvement of local researchers and research institutesin key policy studies. More effective policies based on ex-post and ex-ante evaluations.Governance of the Network and efficient ‘back-stopping’ supportfrom TIPS will be crucial to the success of SADRN. We believethat TIPS has the necessary experience and expertise tomanage these functions effectively and efficiently. We are alsoof the view that the design of the Network – for example theAdvisory Board and Board of Directors structures – contributesto the ease of managing these functions whilst ensuring that thecredibility of the initiative is high and that it is well integratedwith the policy community.The modality for SADRN is based on an appreciation of the experienceof implementing SATRN. One of the key balances thatSATRN sought to manage was the need for high-quality and relevantresearch in the short term whilst the local research communitywas heavily under-resourced and lacked capacity. Whilstwe are cognizant and concur with IDRC’s desire for the Networkto have a single ‘home’, we believe that our proposal – whichaccepts this requirement in the short term but is based on amodality which places great importance on capacity-buildingof policy research institutes in SADC through the TWGs – willimportantly be a sustainable and effective model to meet theprimary objective of strengthening SADC’s policy and researchcommunities.SADRN was launched on 31 August 2007.
industrialpolicy95THE INDUSTRIALCHALLENGE<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>FACING AFRICA IN THE GLOBALTRADING SYSTEM*UNITED NATIONS INDUSTRIAL DEVELOPMENT1. IntroductionORGANISATION (UNIDO)Most African countries have yet to make significant progress towards achieving the MDGs, mostparticularly the target of halving world poverty by 2015. If the productive sectors are to benefit fromnew market opportunities stemming from globalisation and so contribute to poverty reduction, itis essential that both productive and trade capacity be built up. Microeconomic efficiency gains inmanufacturing are potential sources of poverty reduction.Decision-makers should be made aware of the new features of the industrialisation process thathave to be included in any shared growth strategy focused on wealth creation in Africa. The technologygap has prevented most African countries from integrating into the global economy and usingmanufacturing as a dynamic force to reduce poverty. In the ultimate analysis, poverty can only bealleviated through the creation of wealth that, in turn, stems from improved productive capacitiesand adaptive capabilities. Essential to developing an internationally competitive economy is thecapacity to absorb appropriate technologies and produce competitive products that meet marketdemand and comply with international quality standards.* This paper was prepared by UNIDO for the 17 th Conference of the African Ministers of <strong>Industry</strong> (CAMI), 19-21 June2006, Cairo, Egypt.The first half of this paper describes the weakposition held by Africa in the global tradeof goods and identifies key determinants ofthe region’s industrial performance. It goeson to outline shifts in the overall structure ofinternational trade and production, together withthe challenges they pose for Africa. It analysesthe reasons for low productivity rates, points tomissed export opportunities and highlights thekey determinants of competitiveness. The seondhalf of the paper describes the role of trade instimulating growth and fighting poverty. For allthe potential benefits of trade, most developingcountries have failed to reap significant gainsfrom trading opportunities in rapidly expandingliberalised markets. The paper further describesthe main reasons for this failure and focuseson the constraints on supply-side capacities,most particularly as they relate to standardscompliance and conformity assessment issues.Finally, the paper highlights a series of issuesthat the experts might wish to take up in theirdeliberations.
Industrial Policy96<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Africa has missed many market opportunities because of the obstacles ithas encountered in building up technological capacity and introducing anintegrated knowledge-sharing system.Many African countries are often unable to demonstrate compliance withinternational standards and compete on an equal footing in internationalmarkets. Linking productive capacity and trade is of fundamental importanceto global market penetration. If the countries in the region are toderive any benefit from the competitive environment, they will need topromote the diffusion of technology and compliance with internationalstandards in a comprehensive manner.In the component on sustainable industrial development contained inthe New Partnership for Africa’s Development (NEPAD) these new industrialrealities are acknowledged. With the support of UNIDO, the regionaleconomic communities in Africa, governments and the private sectorare adopting a new approach to enhancing supply-side capacities andstrengthening productive capabilities so as to meet the requirements ofthe global market, inter alia, in terms of standards and conformity assessment.Against this background, effective participation in the many segmentsof the global production system takes on particular importance.The document comprises two distinct sections.The first part focuses on Africa’s weak participation in the global productionnetwork; it highlights the challenges facing Africa as the individualcountries in the region try to improve productivity and increase valueaddedat the local, national, regional and global levels. It points to theneed to build on cumulative advantages in the region and identifies keydeterminants of competitiveness in the private sector within the contextof developing public private partnerships. A proactive strategic approachto promoting both productive and trade capacities is described, as arenew regional programmes within the framework of NEPAD that are focusedon sectoral or cross-cutting issues (see Annexes 1 and 2).The second part describes trade as a potential source of poverty reduction,highlighting Africa’s major impediments to trade and the need for compliancewith market requirements. An overview of trade capacity-relatedinterventions is followed by an explanation of the main multilateral initiatives,with particular emphasis on the strategies and initiatives launchedby UNIDO. This part also presents the determinants of success in tradecapacity building, indicating the need to combat supply-side constraintsand strengthening standardisation, certification, accreditation, metrology,testing and quality assurance capacities at institutional and enterprise levels.A list of issues in the final section is intended to stimulate informeddiscussion on the challenge of addressing those issues.2. Building productive capacity:competing in the context ofglobal production 12.1. The share of Africa in global trade inmanufactured goods2.1.1 The region’s weaknessesIn 2004 the volume of world merchandise trade expanded by 9% 2 . Thiswas primarily due to the dynamic performance of trade in manufacturesthat grew by 10% in the same year. Of the total volume of global merchandiseexports valued at US$8,907bn, manufactures accounted forsome US$6,570bn as against fuels and mining products (US$1,281bn)and agricultural products (US$783bn). In terms of value, Africa’s shareamounted to some US$212bn. In structural terms Africa’s merchandiseexports in 2004 were skewed towards fuels and mining products (59.1%),followed by manufactures (25.1%), agricultural products (12.1%) and unspecifiedproducts (3.7%).African decisions-makers in the productive sector need to be aware of thedirect correlation between their position in the different segments of theglobal production system and the positive impact on economic and industrialgovernance at the local, national, regional and continental levels.They should gradually adjust their approaches to industrial development,if there is to be any chance of reducing Africa’s marginalisation in theglobal production system. Greater knowledge of the individual countries’specific industrial capacities and capabilities across the whole value chainis becoming a key tool in modernising the region’s industries.In 2004 Africa exported goods to the value of US$232bn, equivalent to2.6% of global trade; it imported goods to the value of US$212bn equivalentto 2.3% of global trade 3 . Of its exports, manufactures accounted forUS$58.1bn, agricultural products US$28bn and fuels and mining productsUS$137bn. Taking fuels, mining products and agricultural productstogether, it can be seen that unprocessed goods accounted for more than71.% of Africa’s total merchandise exports in 2004. Only 10% of thegoods exported were traded within Africa (intra-trade); the larger shareswent to Europe (42.8%), Asia (16.8% [5.8% to China alone]) and Southand Central America (2.9%).1This part of the paper has been prepared by Yves Ekoué Amaïzo and Anders Isaksson (1.3only), UNIDO, Vienna (Austria).2World <strong>Trade</strong> Organisation, International <strong>Trade</strong> Statistics 2005.3Ibid, p. 80.
FIGURE 1: AFRICA: SHARE IN WORLD TRADE IN MANUFACTURES, 2004 (%)World trade in manufactures = US$6,570bn (73.8%)Africa’s exportAfrica’s import97Source: WTO, International <strong>Trade</strong> Statistics 2995, pp 124<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Africa may well face difficulties in terms of long-term sustainable development,if it fails to promote its capacity and capability to processgoods. Productive capacity is an essential complement to trade capacityin Africa. According to UNIDO, the share of Africa in global manufacturingvalue-added in 2005 stood at 0.7% as against 5.1% in Latin Americaand 18.4% in South-East Asia 4 (see Figure 1). Global concern is growingover the manner in which effective operational support can be lent tostakeholders in development as they endeavour to attain the targets setin the MDGs. Poverty reduction needs to be tackled from both the macroeconomicand the micro-economic levels.2.1.2 Shifts in the structure of tradeThe changing structure 5 of global trade is directly related to the globalproduction system, which basically works along the lines of a highly segmentednetwork. This bears many implications for productive capacity inindustrially less developed countries. Upstream activities such as policies,business environment and reduction of barriers to competition are of crucialimportance. Downstream activities such as support to institutions orthe enterprises themselves are fundamental given the ripple effects theyhave on job creation and a country’s overall capacity to upgrade production.In any wealth generation process, such factors as productive capacity,productivity and competitiveness have to be closely linked.4UNIDO, International Yearbook of Industrial Statistics, 2006, p. 34 (percentage is inconstant 1995 price).5William Milberg, The changing structure of international trade linked to global productionsystems: what are the policy implications? World Commission on the Social Dimensionof Globalisation, International Labour Office, Geneva, Working paper No. 33, May 2004.That linkage is one of the prerequisites for sustainable supply capacityand regional integration. It is predicated on the pro-active participationand complementary initiatives of the private sector, the government andsupport institutions, as well as learning and innovation centres.UNIDO has identified some of the key determinants of industrial performance:productivity enhancement and the crucial role of institutions andlinkages in promoting industrialisation. It actively supported the AfricaProductive Capacity Initiative, which ultimately emerged as the sustainableindustrial component of the New Economic Partnership for Africa’sDevelopment (NEPAD). One of the clear objectives set by NEPAD in thecoming years is to increase Africa’s current share in global trade 6 . This underscoresthe need to support all production related initiatives designedto increase Africa’s share of manufactures in the total merchandise trade.In 2004, no less than 73.8% of global trade was in processed goods (seeGraph 2 for manufactures) 7 . In that year, manufactured goods accountedfor 25.1% of the region’s exports and 71% of its imports. This asymmetryneeds to be corrected over time through programmes and projects focusedon productivity enhancement in Africa.2.1.3 The strategic approach of UNIDOUNIDO has adopted a methodology based on the synergies between thepromotion and gradual upgrading of local technological content and upgradingthe value-chain process.62.6% of global exports and 2.3 % of global imports7World <strong>Trade</strong> Organisation, International <strong>Trade</strong> Statistics 2005., p. 124.The industrial challenge facing Africa in the Global Trading System
Industrial PolicyFIGURE 2: WORLD TRADE IN MANUFACTURES, 2004Manufactures versus non-manufactures: Share in world merchandise trade, in %98<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: WTO, International <strong>Trade</strong> Statistics 2995, pp 124This is clear acknowledgment that: (a) competition takes place in globalproduction networks and segmented systems; and (b) the private sectorderives major benefits from participation in the global market. Partnershipand cooperation should thus take place at both the enterprise and institutionallevels with the objective of promoting clusters. Special attentionshould be paid to the priority sectors 8 and crosscutting issues 9 alreadyidentified in the NEPAD African Productive Capacity Initiative with the aimof increasing Africa’s participation in both local and global production systems.To that end, UNIDO proposes a three-step approach: 3Ls - Linkages,Leverage and Learning (see section 2.6.1 below).2.2. Global production networks andproductive capacity2.2.1 Shifts in the structure of productionThe overall structure of international trade has shifted with the emergenceof more intermediates and the increase in outsourcing activities in specificsegments of the value chain. As a consequence, industrial production haschanged and the location of production became a less important con-8Priority sectors in the value-chain approach in NEPAD/APCI: (i) food processing; (ii)textiles and garments; (iii) leather and leather products; (iv) mineral and metal products(processing of); (v) wood and wood products; (vi) automobile equipment and assembly/spare parts; (vii) pharmaceuticals; and (viii) building materials.9Crosscutting issues in NEPAD/APCI:(i) harmonisation of industrial strategies and policies(including statistical data); (ii) improving quality infrastructure, investment promotionand supply chains (trade capacity building); (iii) promoting energy supplies and efficiency,especially in rural areas; (iv) developing information and communication technologies as amean to reduce transaction costs in productive activities; (v) focusing on technology diffusion,clean production and productivity (vi) promoting conducive regulatory and businessenvironment (support systems and capability building); and (vii) upgrading skills throughlearning and innovation processes.sideration given the mobility of capital, goods and human resources. Aparadigmatic shift towards “externalisation” of production and expansionof manufacturing export activities has occurred. In such a setting,increasing the level of value added in Africa poses a challenge because ofthe region’s dependence on the export of unprocessed goods. Commodity-basedand mineral-based exports are hardly conducive to increasingvalue-added. Industrial policy and strategy in Africa should thus focus onthe specialisation of production processes.The segmentation approach favours major transnational corporations(TNCs) which control both the entire production process and the completesupply chain without any direct involvement in the segmentationof the production process. Most major corporations operate via a systemof global production networks, relocating or outsourcing lower valueaddedactivities with lower technology content. Investors and TNCs alikeare adapting their activities to this new development pattern; they selectcountries that present the lowest risks in terms of barriers to competition,availability of skilled human resources, overall flexibility and linkages atthe level of both enterprises and support institutions. Africa, however, maynot necessarily meet those requirements. Although improving, the level offoreign direct investment in sub-Saharan Africa only reached US$11.3bnin 2004 as compared to US$4.1bn in North Africa (self-financing investment)and the Middle East, US$42.4bn in Latin America and the Caribbeanand US$63.6bn in East Asia and the Pacific 10 . For the transnationalcorporations, the main objective in developing productive capacity throughglobal production networks is to control that part of the value-chain yield-10World Bank, Global Development Finance: Mobilising Finance and Managing Vulnerability,2005.
FIGURE 3: SHARES OF MANUFACTURED PRODUCTS IN WORLD EXPORTS BY TECHNOLOGY (%)99<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: Sanjala Lall in UNIDO, IDR 2004ing the highest value-added; that is invariably their “core” competency.The downward trend displayed by non-processed goods (see Figure 3 forresource-based products) in global exports overall is a clear indication tothe industrially less developed countries (ILDCs) that they should reconsidertheir industrial and competition policies.With ever increasing demands in terms of technological content 11 , skillsrequirements, efficiencies and linkages between firms, barriers becomealmost insurmountable. Many ILDCs often find themselves forced tocompete for outsourcing activities in specific segments of the value chainthat are easier to enter. For some countries, it often appears to be oneof the best ways of entering the production system. In such a dynamicand changing environment, however, the distribution of wages and theirgradual increase may well give rise to new problems. Whereas the revenuegenerated might well improve the overall standard of living, it is notalways immediately available for reinvestment. Profits accruing to foreignenterprises are often repatriated by virtue of the liberal investment codesin Africa. Rents from productive activities are highly dependent on theoverall business and investment climate, as well as the general infrastruc-ture, including connectivity, perceived levels of risk for businesses andpolitical stability.Countries will need to address the implications that these developmentsbear for industrial policies as part of their comprehensive competitionpolicies in the context of other related sectoral and crosscutting policies.They cannot be de-linked from the global production network where oneneeds to understand and master the ability to break down the productionprocess into segments. Each of those segments can be developedindependently or in a comprehensive manner. Each of the segments canbe developed across national boundaries. Selected parts of each segmentcan be produced in one of many countries, assembled in one of manyother countries and sold on the global market. The role of technical assistanceand foreign direct investment should always be viewed in thelight of this new reality: manufacturing within a production system that isdisintegrated, segmented and fragmented. Mastering productive capacityunder these circumstances becomes a crucial issue for countries whoseoptions in terms of integrating and competing in the global productionsystem are limited.11Sanjaya Lall, Investment and technology: Policies for competitiveness: Review of successfulcountry experiences, UNCTAD, Technology for Development Series, 2003, p. 23.The industrial challenge facing Africa in the Global Trading System
Industrial Policy100<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>2.2.2 The implications of those shiftsOne of the consequences of the shifts in the structure of global productionand international trade is to be seen in the international division oflabour and the specialisation in the production of sub-components as partof the international value chain. Productivity becomes crucially importantin a competitive setting. Actors better suited to dealing with competitionshould also be given more opportunities to lead the process. Enhancingproductive capacity, defined in the report on Africa Productive CapacityInitiative (APCI) as the “ability, first to produce goods that meet the qualityrequirements of present markets and second, to upgrade in order totap future markets” is one of the best approaches to Africa’s sustainableparticipation in global production networks and segmented systems. Theefficient allocation of resources in specialised production sites or agglomeratesaround the world has become an important governance issue. Themovement of investment and human capital is often more closely relatedto specific and interlinked segments of the value-chain extending fromproduction right up to the sale of the final product. Value-added needsto be generated at each stage of the segmented production process. Inthe context of this paper, two segments in the value chain should behighlighted: Downstream segments: these are usually based on comparative and competitiveadvantages directly related to incremental value-added in each segment/stage ofproduction. The value-added and technology content is low, as is the level of skillsrequired. Upstream segments: these are usually based on cumulative advantages. Their focuslies on coordination, knowledge-sharing and brand-ownership in the globalproduction systems.Producers in the least developed countries (LDCs) need to identify thosesegments/stages of the production and trading process in which they canparticipate. The traditional determinants of competitiveness such as land,cheap labour and capital are not necessarily the most important. Equallyimportant are support institutions, which can bring about improvementsin productivity and efficiency. Compared to the high productivity achievedin industrialised countries, it is becoming increasingly obvious that thelower wages paid in Industrially less developed countries do not alwayswarrant entering the development process. Simply promoting industrialisationon the basis of industries with a low technology content can wellprove counter-productive unless the country’s overall competition policyprovides for resilient learning and upgrading processes . The asymmetricknowledge-based differences between the ILDC and industrialised countriesare reflected in their trade imbalances in terms of both quantity andquality. Improving productive and trade capacity in the ILDCs is predicatedon securing similar improvements in their industrial infrastructure and infostructure.Less than 5% of the Fortune 500 companies are based in low-incomecountries 12 . The challenge of attracting high calibre multinational corporationsto Africa is directly linked to the ability of African governmentsand regional economic communities to reduce substantially the currentbarriers to competition in the region. At present, FDI flows primarily tothe extractive sectors. Manufacturing in Africa is characterised by a broadspread of production. Regrettably, enterprise agglomerates in Africa areoften only to be found in the low value-added segments of the valuechainbecause they cannot meet the challenge of maintaining sustainabilityin niche-markets. Measured in terms of the Herfindahl-Hirschmanindex, a standard measure of industry concentration, it would appear thatproduction is more widespread and diversified in Africa than elsewhere;production, moreover, is mainly related to low-tech manufacturing withlow value-added 13 . Moreover, when assessing the long-term profitabilityof private firms in developing countries as compared to industrialisedcountries, it was found that both the short- and long-term profit rateswere lower in industrially less developed countries than industrialisedcountries 14 . In this regard, Africa is no exception, especially when it comesto average profit rates in manufacturing as compared to mining or services.Despite increasing global competition and the overall trend towardsglobalisation of production, the irony is that concentration of industriesseems to encourage or even force the ILDCs to enter global productionfrom the lower end of the value chain.THIS SITUATION MAY WELL PERSIST FOR A NUMBER OF REASONS:The existence of barriers to entry at the high end of the value chain: this is reinforcedby the inability to achieve economies of scale;The asymmetric technological divide: this will call for a collective effort on the partof the ILDCs at both the national and regional levels to reorient their industrial andcompetitive policies towards a process focused more on learning and knowledgesharing,with especial emphasis on innovative culture and approaches;The difficulty of penetrating markets owing to “brand loyalties”: over time thesehave become “brand penalties” at the upper end of the value chain, thus making itessential that the ILDCs create and benefit from their own brands;Competition between plant location in low-wage countries and movement of capital:a major factor is the ability of enterprises and foreign direct investors to movearound the world, typified by dynamic capital mobility. Unfortunately for the ILDCs,the speed with which subsidiaries of multinational companies relocate from onecountry to another can have a destabilising effect on local economies, especiallywhen regional economic policy cannot be enforced.Tariff and non-tariff barriers: although tariffs are being lowered in Africa, exports12Fortune (2000) and Financial Times (2000).13Mayer, J. A. Butkevicius and A. Kadri. 2002. Dynamic Products in World Exports,UNCTAD, Discussion Paper, No. 159, Geneva.14J. Glen, K. Lee and A. Singh, Corporate Profitability and the Dynamics of Competition inEmerging Markets: A Time Series Analysis, ESRC Centre for Business Research, Universityof Cambridge, Working Paper, No. 248, December 2002, p. 1
of industrial goods from Africa, other than those that are resource-based and havea low technology content or value-added, are facing tariff and non-tariff barriersin industrialised countries. This has skewed production towards low value-addedsectors, at the lower end of the value chain. Wage increases: wage increases might ultimately pose a problem, unless enterprisesprovide for relocation as part of the modernisation process, failing whichtechnology would have to be upgraded within the value chain.2.2.3 Industrial policy responseIn the light of the above, the industrial policy response has to be changed.Decision-makers will have to decide on the manner in which productionmoves up the value chain. They will also have to determine the strategiesthat governments, private sector enterprises and support institutionsshould adopt to upgrade their industries in a competitive environment.Their comparative advantages will have to be converted into dynamiccompetitive advantages. For industrial development policy to be sustainablein the long term, decisionmakers in Africa may seriously considermoving towards operationalising the concept of “cumulative advantages”.Dynamic efficiency and the maximisation of sustainable productivegrowth in an asymmetric market structure dominated by powerful playersare the new development factors; they determine productive capacity andwealth creation.Producers in the ILDCs find it difficult to add value in the production process.Their low absorptive capacities related to learning 15 and knowledge isdirectly correlated to the difficulties they face in enhancing their productivecapacities and capabilities, increasing efficiency and productivity, breakingdown barriers to competition through the creation of an enabling businessenvironment: all of which cannot be taken in isolation. In addition,weak support institutions and generally inadequate infrastructure arenot conducive to lowering transaction costs. The business environment,institutional upgrading, public private partnership approaches to enterprisepromotion and policies pertaining to the building of both productiveand trade capacities should be embedded in a culture of innovation andgreater value-added. The involvement of the private sector should focuson changing the general trading culture of many industrialists. A specialglobal forum might be held to raise awareness of these issues.2.3. Low productivity and missed exportopportunities2.3.1 Productivity and exportIn Africa, differences in labour productivity across firms of differing sizeare substantial measured in terms of value-added per employee 16 . In micro-enterprises,value-added per employee in dollar terms is three to fourtimes higher than in larger enterprises. Variation across sectors is alsosubstantial. The range of physical capital per employee across firms ofdiffering size is even larger than that for labour productivity. Differencesin the physical capital endowments across the firms are a much moreimportant determinant of differences in labour productivity than differencesin human capital measured in years of education and tenure. Thereis no clear evidence that large firms have higher labour productivity thansmall enterprises. The fact that large enterprises have more capital peremployee (capital intensity) is not necessarily an indication of higher labourproductivity, although greater capital intensity usually implies higherlabour productivity.The correlation between productivity growth and propensity to export isworth examining. A marked propensity to export despite relatively lowproductivity is often explained by the absence of an adequate enablingbusiness environment. The rates of return on physical and human capitaland productivity at the enterprise level are clearly correlated. The ratesof return on physical capital are far higher than those on human capital.The rates of return on capital could be affected by increased investment.If investment costs were low, high returns on physical capital would implyhigh investment rates. In Africa, investment in firms in the manufacturingsector is often low. From a sectoral perspective, it is worth highlightingthe “productivity” of the sector through the use of a benchmarking approachor peer reviews (comparing data within the sector and within asub-region).The issue of export becomes fundamental to understanding the specialisationprocess. Relatively few firms in Africa export; most of them arenot specialised in exporting. It is therefore difficult to sustain productivityand upgrade in terms of increasing the technology content. Masteringthe segmentation factor previously mentioned is essential to reducingmissed export opportunities. The differences in productivity are largely theoutcome of the firms’ operational efficiency rather than the amounts ofphysical capital or even technology transferred. Exports are important to101<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>15UNIDO, Industrial Development Report 2002/2003: Competing through Innovation andLearning, Vienna, 2002.16Measuring or comparing productivity is still a difficult task. Difficulties in collectingappropriate data and availability of appropriate statistical data are some of the difficultiesfaced when trying to discuss the level of African enterprises productivity.The industrial challenge facing Africa in the Global Trading System
Industrial Policy102<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>firms for several reasons . For example, the increased scale of operationsand access to foreign exchange are significant factors. The latter facilitatesthe acquisition of capital goods from abroad; that, in turn, can boostproductivity.Many economists argue that one of the main obstacles to entering exportmarkets is the cost involved 17 . To overcome this, firms have to produce atcompetitive costs; the only way to achieve this is by being productive (becausegreater productivity reduces the average cost of production). Thisexplains the oft observed statistical correlation between productivity andexport-participation 18 .2.3.2. Increasing productivityIt is now accepted that relatively high productivity tends to go hand inhand with exporting; firms in small open economies that fail to achievesufficiently high levels of productivity are confined to the domestic market19 . The issue thus becomes one of increasing productivity.Enhanced competition is often put forward as a key remedy to low productivity.It is likely that among both formal and informal small and medium-sizedfirms, competition is fairly stiff, whereas among large firms, ofwhich there are but few in Africa, real domestic competition hardly everoccurs. Interestingly enough, competition among small firms can only takethem so far because they are too small to enjoy reasonable economies toscale. Large firms, on the other hand, exploit economies of scale but tendto be shielded from competition as long as they operate in a “closed” or“protected” market.A good example are the former large state-owned enterprises (SOEs) inAfrica that tended to be protected against competition and operate atlow productivity levels. The question is whether on being privatised thoseenterprises maintained their productive capacity i, increased their productivityor merely continued operating under a different ownership structure.For want of competition and given their sheer size, some of those enterpriseshave managed to stay in business. In the longer term, however,17See for example Clerides, S., Lach, S. and J. Tybout (1998), “Is ‘Learning-by-Exporting’Important? Micro- Dynamic Evidence from Colombia, Mexico and Morocco”, QuarterlyJournal of Economics, <strong>Vol</strong>. 113. and Granér and Isaksson (2002), “Export Performancein the Kenyan Manufacturing Sector”, in A. Bigsten and P. Kimuyu (Eds.) Structure andPerformance of Manufacturing in Kenya, Palgrave and Granér and Isaksson (2006),“Firm Efficiency and the Destination of Exports: Evidence from Kenyan Plant-Level Data”,mimeo, UNIDO.18See for example Aw, B.Y., S. Chung, and M. Roberts (1998), “Productivity and the Decisionto Export: Micro Evidence from Taiwan and South Korea”, NBER Working Paper6558 and Roberts, M. and J. Tybout (1997), “The Decision to Export in Colombia: AnEmpirical Model of Entry with Sunk Costs”, American Economic Review, <strong>Vol</strong>. 87.19Exceptions to this include firms in relatively large (for the region) economies such as Nigeriaand South Africa.smaller domestic enterprises or foreign companies will seize whateverbusiness opportunities there are, whereupon the former SOEs will eitherhave to become more efficient or close down.2.3.3 Increasing competitionOne way of increasing competition among large firms propagated by majorinternational financial institutions is to open them up to internationalcompetitors. While in theory the benefits seem to abound, the extent towhich the presence of foreign firms yields any net benefits remains unclear.Three factors speak against positive net benefits. First, foreign firmsoften operate in extractive industries and repatriate their profits. Secondly,if scarce skilled labour tends to work in foreign firms rather than domesticfirms, the latter’s productivity will be reduced. Thirdly, since African industriestend to be weak, permitting transnational corporations to competeon local markets can in principle crowd out domestic production. In theultimate analysis, competition can have a negative impact on domesticprivate sector enterprises.It is claimed that the major benefit to be derived from foreign firms is thetransfer and diffusion of technology. Much of this occurs at the managerialand shop floor levels. The extent to which this newly acquired knowledgespills over to domestic firms is debatable. Moreover, it is questionablewhether the knowledge is useful or the technology appropriate. Foreignfirms also create job opportunities; here again, however, the openingsare mainly for skilled labour. If, in the process, domestic firms have toshut down, the net employment generated could be negative. Industrialpolicies should take this into consideration when drafting competitionpolicies. For technology transfer to be beneficial, the absorptive capacityof the economy needs to be strong. It cannot take place without strongsupport institutions. Otherwise, many of the positive effects of foreignpresence will bypass the host country. The same holds true for technologytransfer and diffusion through international trade; however, the differencein this case being that much of technology is embodied in capitalgoods and can thus be more readily applied to production. Furthermore,transaction costs in Africa are higher than elsewhere. Key to productivityperformance in Africa is the development of the region’s human capital: aconcept that relates to both education and health. Slight improvements inthe quantity and quality of food can have an enormous impact on workerproductivity. A broadly educated population is also important; emphasisshould thus be placed on primary and secondary (in particular technical)schooling rather than tertiary education. Allowing women to completetheir schooling will undoubtedly contribute to the creation of a skilledworkforce. It is also likely that, as more people enjoy a better educationand wages rise accordingly, the risk of a brain drain will diminish.
2.3.4 The need for supportDoes the African private sector need support? Like any other private sector,it does and whatever helps to improve its productivity will increaseits competitiveness. The public sector could lend support in the form ofimproved infrastructure, provision of adequate education and healthsystems, as well as general institutional support to increase domesticsavings and investment. It could thus help to create a healthy businessenvironment (the cost of doing business appears to be higher in Africathan elsewhere). Administrative and legal barriers to the entry and exitof firms should be reduced in order to enable the more productive firmsto survive and grow, even if it be to the detriment of relatively unproductivecompanies. If best-practice technologies cannot be implemented, forwhatever reason, governments should endeavour to remove or at leastlower barriers to their introduction. The debate on the type of businessenvironment to be created should be conducted within the context of apublic-private partnership (PPP).External support may also be needed; in particular when official fundsare so constrained that a government can no longer assist private sectordevelopment. That support could include debt relief and official developmentaid to key sectors such as education and health.It could also take the form of help to ensure that the goods produced forthe export market are of adequate quality, meet international standardsand norms and are of relevance to the global segmented market. Africanfirms need to increase their productive capacity to the point that they cancompete internationally in terms of prices. The African private sector andgovernments should complement each other in their endeavours to buildup trade capacity.2.4. Key determinants of competitiveness2.4.1 The need for a comprehensive approachMany issues such as productivity, value chain, enabling competitive environment,technology content, infrastructure and support institutions areconsidered key determinants of competitiveness. In global production networksand their segmented systems, however, it is important to concentrateon a holistic approach. UNIDO has promoted the national learningprocess or system supporting interaction between factors and actors inthe interest of collective efficiency: another key determinant. <strong>Trade</strong>, industrialpolicies, macroeconomic conditions, location and resource endowments,human capital and technological endeavours, as well as the natureof factor markets and institutions are all related factors; they should beduly considered within the framework of public-private partnerships. Oneoutcome of the regional NEPAD/APCI sub-regional conferences has beenthe establishment or strengthening of those partnerships with the objectiveof forging stronger links between innovation and productive capacities/capabilities.Many countries have been advised to focus their industrialdevelopment performance on a technology-driven economy which, inturn, should ensure high growth rates. In fact, as mentioned by SanjayaLall, it is “Skill development, industrial specialisation, enterprise learningand institutional change create cumulative, self-reinforcing processes thatpromote or retard further learning. Countries set on a pattern with a lowtechnology,low-skill and low-learning specialisation find it increasinglydifficult to change course without a concerted shift in a large numberof interacting markets and institutions” 20 . A comprehensive approach istherefore required when designing programmes and projects in supportof productivity enhancement.The one-way solution directed exclusively towards “economic liberalisation”does not contribute to the development of a comprehensive approachnor does it engender cumulative advantages. All too often endowmentssuch as natural resources and cheap unskilled labour per se arenot conducive to creating the dynamic and innovative culture needed tosustain growth and structural change. Marginalisation and the inability toparticipate meaningfully in the global production networks and segmentedsystem thus become almost inevitable. Most ILDCs fall into the trap ofmaintaining local production cultures based on “static” approaches andentrenched in “declining markets” while still pursuing obsolete importsubstitutionpolicies. In other words, they continue to import technologies,services and products without mastering the requisite production, servicingand maintenance skills. Productive capacity initiatives must abandonthis long outdated approach.They should learn the lessons of agglomeration (the cluster approach)that open up opportunities for industrial “leapfrogging”. The acquisitionof technology from abroad is embedded in local centres of excellenceor learning centres; they provide optimal access to the latest technologiesand know-how. Given the pressure emanating from profit-orientedtransnational companies, this comprehensive leaning process should bestepped up.However, if not adequately backed by government, a conflict with externaltechnologies and know-how might a occur, much to the detriment of theILDCs. New capabilities need to be set up as part of national or regional20Sanjaya Lall, “Investment and Technology: Policies for competitiveness: review of successfulcountry”, UNCTAD, Technology for Development Series, 2003, 79 p.103<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The industrial challenge facing Africa in the Global Trading System
Industrial Policy104<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>learning processes. It would even be more effective, were it to be focusedon a particular sector (such as leather and leather products) or a crosscuttingissue (such as information and communication technologies).With a major part of the ILDC economies based on low-wage productivecapacities, resource mobility becomes another key determinant of competitiveness.It is universally accepted that technical change erodes thecompetitive advantage of cheap and unskilled labour. Countries cannotbuild their industrial performance strategy on static factors. The quality oflocal capabilities and institutions becomes another key determinant of theability to attract and use foreign resources. The global production networkis based on interactive specialisation.Leading players in each value chain rely increasingly on independentsuppliers of valueadded inputs, services and innovations. Africa shouldrecognise as a given the fact that the overall objective in a competitiveenvironment today is to structure the national/regional economy towardsa knowledge-led economy based on gradual mastery of technological“niches” at the local level. Global and local (“glocal”) visions are of paramountimportance: a point emphasised during the NEPAD/APCI regionalconferences. As a consequence, it was requested that an “observatory” beestablished to maintain a watch on developments related to competitivenessand employment creation so as to alert productive actors to changesin the global market and their impact on national and regional markets.The Africa Commission of the UK Government recognised the Africa ProductiveCapacity Initiative as one means of diversifying and reducing dependenceon volatile commodity and mineral sectors. Coherence needs tobe strengthened between national and regional industrial developmentstrategies on the one hand and global economic dynamic changes onthe other. This would bring about the creation of additional wealth derivedfrom sound productive capacity and secure greater gains from trade.Forging industry’s links with both agriculture and trade is central to allstakeholders in the process, including UNIDO, NEPAD, CAMI and the AfricanUnion.2. 5. Improving competitiveness in theprivate sector in Africa2.5.1 Measuring competitivenessIn its Industrial Development Report 2002/2003, UNIDO highlighted thefact that some of the main structural “drivers” (enabling factors) of industrialcompetitiveness are FDI, domestic R&D, skills, licensing and physicalinfrastructure. Even using the UNIDO Competitive Industrial DevelopmentIndex , it is clear that competitive industrial performance cannot be assessedin a readily comprehensible form for want of reliable statistics.Statistics and disaggregation are not enough. They should be complementedby a peer review system. At the end of the day, it is the quality ofthe work performed by the private sector actors themselves that offers agood indicator of industrial performance. It is therefore crucial to supportthe African private sector as enterprises gradually adapt to the new globalindustrial segmentation and shift away from a culture of assistance andnon-productive rent-seeking approach to an innovative culture of buildingproductive capacities and capabilities. If they are to be successful, theywill have to stop working in isolation.2.5.2 Private sector considerationsIt is generally agreed that the creation of wealth and reduction of povertyare closely correlated to economic growth, which in itself is directly linkedto the volume of sustainable investment entering a country or region. Itshould be recalled that compared to other developing regions, Africa, especiallysub-Saharan Africa, does not receive excessive amounts of privateinvestment. Support to the private sector in Africa presupposes the pursuanceof clear policies with predictable incentives designed to promotesustainable private investment and build support institutions. Three keyenabling factors that help to improve competitiveness are: Direct support to small and medium enterprises using technical assistance fundsto provide business advisory services, upgrade overall productive capacity buildingand introduce training activities related to entrepreneurial development. A onestop-shopapproach that not only provides solutions to the SMEs but also helps toreduce costs associated with bureaucracy is called for. Advisory services including the maintenance of a watch on developments are fundamentalto success in a segmented production system based on value chains. SMEsin Africa should be “embedded” in a system where support is assured throughoutthe operational cycle; A predictable investment and business climate is necessary to attract private investmentthat does not need to rely on public sector guarantees. Providing an enablingbusiness environment for enterprises and promoting public-private sector dialoguethrough peer review mechanisms are key to raising collective awareness and securingproactive responses from both the public and private sectors.Innovative approaches are very much part of cultural change. The privatesector should be more adequately represented in the group of donorslending support to the private sector. For example, it is important thatprivate sector representatives be an integral part of the poverty reductionstrategy programme process at the country level. Consultations are notenough. Jointly preparing such programmes could secure the incorporationof private sector views at the very outset of the process, thus contributingto a change in the donor-driven approach to supporting the privatesector as though it were the public sector.
2. 6. A strategic approach to promoteproductive capacity2.6.1. The three-step processAs a follow-up to the recommendations of the Third United Nations Conferenceon Least Developed countries on the role of industrial growthin alleviating poverty in those countries 21 and after NEPAD had beenlaunched as Africa’s new approach to improving political and economicgovernance 22 , UNIDO reoriented its approach towards Africa and shiftedgradually to the promotion of integrated programmes/projects. Not onlydid the African Ministers of <strong>Industry</strong>, NEPAD and African Union and theAfrican Heads of States of approve the NEPAD component on sustainableindustrial development based on the African Productive Capacity Initiativedeveloped by UNIDO, but the African and international communityalso recognised that any new efforts directed towards industrialisation inAfrica should come primarily from within Africa itself. This is very much inline with the MDGs and represents a comprehensive attempt to reach acommon African position on mobilising resources and expertise in supportof the region’s industrialisation. Joint efforts and partnerships constitutethe very basis of this newfound spirit of cooperation. This particular approachshould also be embodied in the design and implementation ofprojects and programmes related to productive and trade capacities inAfrica. Earlier approaches based on sectoral or crosscutting entry pointsneed to be integrated if they are to compete successfully in a global setting.The overall objective is to transform comparative and competitiveadvantages into cumulative advantages for Africa. Three key steps are essentialto enhancing productivity at the local, regional and global levels:Linkages: improving internal and external linkages in building competitive productivecapacities and capabilities;Leverage: leveraging growth by reducing barriers to competition at all levels; andLearning: promoting a learning and innovative culture through clustering and networkingapproaches.These three steps make for the emergence of an innovative culture andprocesses, while generating value-added in each segment of the productionnetwork. Productivity enhancement is the ultimate goal. The majorchange over previous “stand-alone” approaches is the manner in whichdevelopment stakeholders, including the private sector, organise themselveswhen promoting comprehensive programmes. UNIDO will focus onNEPAD priorities, special consideration being given to regional integrationand co-operation (see Figure 4). The way forward might be to ensurepredictable joint co-operation between African Development stakeholders21UNIDO, Building Productive Capacity for Poverty Alleviation in Least Developed Countries(LDCs): The Role of <strong>Industry</strong>, May 2001 (Conference took place in Brussels).22NEPAD, www.nepad.orgin the preparation of comprehensive programmes, with NEPAD coveringboth sectoral and crosscutting issues. Selected examples of specific actionsare provided in the annexes.2.7. Supporting the industrialisation ofAfrica: regional programmes withNEPAD2.7.1 A common visionNEPAD 23 is a common vision based on a shared conviction that Africa’sleaders are duty bound to eradicate poverty and set their countries, individuallyand collectively, on a path towards sustainable growth and development.To that effect, they should participate actively in the worldeconomy. The private sector should take on a proactive role. Moreover,NEPAD has set up the NEPAD Business Group to ensure proper co-ordinationbetween the public and private sectors in keeping with the NEPAD vision.The programme is based on the determination of Africans to assumeresponsibility and ownership and extricate the region from the viciouscircle of poverty and transform it into a virtuous circle of wealth creation.THE NEPAD OBJECTIVES IN RESPECT OF MANUFACTURING, PRODUCTIVEAND TRADE CAPACITIES CAN BE SUMMED UP IN THREE KEY POINTS:To increase production in, and improve the competitiveness and diversification of,the domestic private sector, especially in the agro-industrial, mining and manufacturingsub-sectors, with potential for exports and employment creation;To establish national standards bodies throughout Africa;To harmonise national technical regulatory frameworks.FOR THE PRIVATE SECTOR, THE MAIN OBJECTIVE IS TO ENSURE A SOUNDAND ENABLING ENVIRONMENT FOR PRIVATE SECTOR ACTIVITIES, WITHPARTICULAR EMPHASIS ON DOMESTIC ENTREPRENEURS. TWO KEYOBJECTIVES SHOULD BE ADDRESSED WHEN BUILDING UP PRODUCTIVECAPACITY:To promote foreign direct investment and trade, with particular emphasis on exports;To develop micro, small and medium-scale enterprises, including the informalsector.2.7.2 Specific actionsNEPAD suggested the following specific actions; they will need to be revisitedin the context of global production networks and the segmentedproduction system.23The New Economic Partnership for Africa’s Development, 2001. See: http://www.uneca.org/eca_resources/Conference_Reports_and_Other_Documents/nepad/NEPAD.htm105<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The industrial challenge facing Africa in the Global Trading System
Industrial PolicyFIGURE 4: SUPPORTING AU/NEPAD INDUSTRIALISATION PROCESS: AFRICAN PRODUCTIVE CAPACITY INITIATIVECompeting in a global segmented production systemLinkages Leverage Learning106<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>LOCALSectoral entrypointBuilding competitiveproductive capacitiesand capabilitiesCross-cuttingentry pointSectoral entrypointEducing barriers tocompetition at alllevels(policy, institution,enterprise)Cross-cuttingentry pointSectoral entrypointClustering and networkingapproaches(stopping working inisolation)Cross-cuttingentry pointGLOBALComparative advantages Competitive advantages Cumulative advantagesComprehensive value chain approach for regional programmes with NEPADAT THE AFRICAN LEVEL: Undertake measures to enhance the entrepreneurial, managerial and technicalcapacities of the private sector by supporting technology acquisition, productionimprovements, and training and skills development; Strengthen chambers of commerce, trade and professional associations, and theirregional networks; Organise dialogue between the government and the private sector to develop ashared vision of economic development strategy and remove constraints on privatesector development; Strengthen and encourage the growth of micro, small and medium-scale industriesthrough appropriate technical support from service institutions and civil society, andimprove access to capital by strengthening micro-financing schemes, with particularattention to women entrepreneurs.AT THE INTERNATIONAL LEVEL:Given the fundamental changes in the way global and local actors approachproduction, UNIDO needs to be innovative and demonstrate itscommitment to the MDGs. With donors moving increasingly from basketfunding (multi-donor contributions to programmes in the field) to generalbudget support to governments to implement their own programmes, it isimportant to establish a core group for resource mobilisation composed ofmajor stakeholders in the NEPAD initiative on sustainable industrial development.A partnership comprising NEPAD, the African Union, private sectorrepresentatives (for example, the NEPAD Business Group 24 ), regionaleconomic communities and government representatives, should ensurethat all regional programmes take account of the value chain approachand the segmentation of the production system and enjoy counterpartsupport.Promote entrepreneurial development programmes for training managers of Africanfirms;Provide technical assistance in relation to the development of an appropriate regulatoryenvironment, promotion of small, medium and micro-enterprises and, establishmicro-financing schemes for the African private sector.24NEPAD Business Group is composed of: ABR: African Business Roundtable; ICC: InternationalChamber of Commerce; CBC: Commonwealth Business Council; FFA: ForumFrancophone des Affaires; BHF: Business Humanitarian Forum; CIAN: Conseil Françaisdes Investissements en Afrique; CCA: Corporate Council on Africa; IBLF: InternationalBusiness Leaders Forum; BCEAM: Business Council Europe Africa Mediterranean; PEF:Pan-African Employers; CCA: Canadian Council on Africa.
It is important to analyse the cross-border activities of firms. The newapproach to industrial development will gradually replace the obsoletestate-centred form of analysis. Within the context of a global productionnetwork framework, it is vitally important that industrial leadersunderstand and master the dynamics of global organisation of productionand the subsequent division of labour. Employment generation andthe creation of decent sustainable jobs are directly related to decisionson productive capacity building. This calls for a regular updating of localand global regulations, competition policies and economic integration.Industrial upgrading must be firmly bedded in a learning-led approachto industrial development, with special emphasis on flexible small andmedium enterprises.Overall private sector objectives in Africa should focus on sustainingshared economic growth, creating wealth and increasing security throughpoverty reduction. Flexibility and adaptability in all forms, products, management,processes, systems and policies are required. Innovation mustbe an on-going process based on learning from mistakes, capitalising onsuccesses and drawing on best practices. With aid funding for productivecapacity activity on the decline, it is essential that private sector capacityand capabilities be strengthened.AS PART OF A NEPAD BUSINESS GROUP PROGRAMME, DONORS ANDSTAKEHOLDERS SHOULD FOCUS ON THREE MAIN FEATURES OF ACOMPREHENSIVE PRODUCTIVE CAPACITY STRATEGY: Improving the on-going process of creating an enabling policy environment withspecial focus on reducing barriers to competition. This could be achieved by:(a) Initiating a public-private policy dialogue and setting up global peer reviewfora for improved business decisions;(b) Improving legal and regulatory frameworks in order to attract investment innew innovative culture. Networking with investment and financial bodies in order to leverage or accesslarger investment flows to SMEs; Develop businesses and products that take account of the segmentation of globalproduction, while focusing on selected key sectors and crosscutting issues (see Annexes1 and 2).3. Building trade capacity: theneed for compliance andconformity 253.1. IntroductionThe evolution from protectionism to liberalisation in the global tradingsystem and the emphasis on development in the Doha Round offers opportunitiesfor the advancement of trade and industry in Africa. However,most countries in the region have failed to reap significant benefit fromtrading opportunities in expanding markets and concessionary schemessuch as the African Growth and Opportunities Act (AGOA) and EverythingBut Arms (EBA). This paper argues that the reasons for Africa’s failure tobenefit from these opportunities are not primarily related to tariff barriers.They relate to: the lack of productive capacity needed to ensure necessaryquantity and quality of supply; inability to prove compliance of potentialexport products with international standards and the problems with integrationinto the multilateral trading system.African enterprises need to develop regional value chains and link withglobal supply chains to market their products internationally. SMEs, whichpredominate in African economies, have inherent difficulties with accessto capital, productive capacity, technology and servicing because of resourcelimitations. If these enterprises are to trade on global markets,they need to build capabilities to increase their supply capacity, quality,competitiveness and conformity with importer-mandated product standards.This involves both increased investment at the enterprise level andgovernment backing in introducing and improving support to productivityand technology extension services, training, export consortia and clusterdevelopment. Clearly, in the present global setting the role of such intermediariesbecomes even more important. They can provide support forbuilding capabilities to use new technologies, adapt and improve processesand products and move up the value chain into more sophisticatedproduction activities.Standards and technical regulations drawn up by individual countriesto protect health and the environment, as well as to ensure quality andsafety, can also act as technical barriers to trade. In addition to technicalregulations and product standards, African exporters face the morestringent private standards of developed country retailers. Increasingly,107<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>25This part of the paper has been drafted by Mohamed Lamine Dhaoui and Rishi Choprawith inputs from Lalith Goonatilake, Olga Memedovic, Gerardo Pataconni and FabioRusso, UNIDO, Vienna (Austria).The industrial challenge facing Africa in the Global Trading System
Industrial Policyinternational buyers require effective application and recognised proof ofenterprise system management standards such as ISO 9000 for qualitymanagement, HACCP and ISO 22000 for food safety, ISO 14000 for environmentalmanagement and SA 8000 for social accountability.rently holds the largest trade capacity building (TCB) portfolio among allthe multilateral agencies listed in the WTO/OECD database. Implementationreached US$39.8m in 2004, having increased rapidly from US$7.6min 2002.108<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Enacted in 1995, the WTO agreements pertaining to technical barriersto trade (TBT) and sanitary and phyto-sanitary measures (SPS) in effectlegalised the use of standards and technical regulations in global trade.The agreements define the non-discriminatory use of those standards andregulations, while recognising that developing countries need to have thenecessary infrastructure for standards and conformity in place in orderto fulfil their commitments. The two agreements specifically mention thedifficulties developing countries have in this area and include provisionsfor technical assistance although, to a large extent, this has still not happened.Increasing supply capacity is essential to, but not sufficient for, gainingentry into world markets. Proving conformity with standards and technicalregulations requires establishing efficient testing, certification and accreditationmechanisms that conform to the requirements of the SPS andTBT agreements and enjoy international recognition. Testing, calibrationand certification facilities thus take on extreme importance for Africancountries wanting to benefit from trade opportunities.To date, the trade facilitation support through existing multilateral programmessuch as the Integrated Framework for Technical Assistance toLeast Developed Countries (IF) and the Joint Integrated Technical AssistanceProgramme (JITAP), as well as individual agencies such as the WorldBank, United Nations Conference on <strong>Trade</strong> and Development (UNCTAD),and the International <strong>Trade</strong> Centre (ITC) have largely focused on traderelatedpolicies, cross-border facilitation issues such as improving customsprocedures, efficiency, transparency, transport and port development, governanceand export promotion. Although the initial IF needs assessmentincluded supply capacity and conformity, subsequent IF activities concentratedmore on the individual mandates of the participating agencies.Clearly the Aid for <strong>Trade</strong> initiative should build on the experience of the IFto date and aim to fill such a gap by expanding the areas of interventionto supply capacity and conformity.UNIDO has a long track record of developing supply capacity (technologytransfer activities in such areas as agro-processing, cluster development,cleaner production, hygiene and quality management and export consortia)and building standards and conformity infrastructure (standards bodies,product testing laboratories, inspection services, certification bodiesand accreditation systems) over the past 30 years. The Organisation cur-The growing realisation of the need for this type of technical assistancehas also led UNIDO to formulate a more holistic approach to trade capacitybuilding, drawing into the initiative many established and successfulprogrammes such as competitiveness analysis, cluster development andthe promotion of export consortia. The resulting approach is problemorientedand not limited to the mandate of UNIDO. It actively integratesthe areas of expertise of other bi- and multi-lateral development partnersand can be summarised around the three key areas of intervention, the‘three Cs’: supply Competitiveness, demand Conformity, market Connectivity.While promoting supply competitiveness and demand conformityare the main areas of the Organisation’s competence, market connectivityprovides the link to other agencies such as WTO, UNCTAD and ITC.The trade capacity building approach adopted by UNIDO is in line withthe capacity building needs identified by The New Partnership for Africa’sDevelopment (NEPAD) Market Access initiative. The approach has beenrecognised by the UN General Assembly. It is also reflected in major donorinitiatives for trade/export development, such as the Norwegian Agencyfor Development Cooperation (NORAD) and the Swedish InternationalDevelopment Cooperation Agency (SIDA) strategy for export developmentin Africa. This approach also constitutes the basis for the cooperation betweenUNIDO and WTO in trade-related technical assistance.The technical know-how and experience that UNIDO has acquired inprogramme conceptualisation, implementation and fund-raising in thefield of building supply capacity and conformity infrastructure, togetherwith the Organisation’s extensive field presence, are valuable assets inthe context of trade-related technical assistance. UNIDO stands ready tocontribute to the efforts of the African Union (AU) and other developmentpartners to make the Aid for <strong>Trade</strong> initiative a success for Africa in generaland for the least developed countries (LDCs) in particular.3.2. <strong>Trade</strong>, industry and povertyalleviation3.2.1 The global contextThe achievement of the UN MDGs, especially poverty alleviation, is still anelusive target. Since 1980 the number of people living in poverty (earn-
ing less than US$2 per day) has risen by 50% to 2.8-billion 26 . Between1981 and 2001, the number living in absolute poverty (less than a dollara day) worldwide fell from 40% to 21%, but in the Sub- Saharan regionthe figure rose from 42% to 47% 27 . The divide between the rich and thepoor, between nations and within nations, continues to widen. Urgentremedial action by the international community is called for; failure is notan option.The advance of the global trading system from protectionism to liberalisation,with the Doha Development Agenda as a landmark still tobe achieved, offers major opportunities. World trade has risen fromUS$1.0-trillion in 1970 to the current (estimated) level ofUS$14.4-trillion. But the development is lopsided. The share of developingcountries is still only around 25% 28 .While the newly industrialised countries (NICs), China, India, Chile andsome others have gained, many other developing countries in Africa andthe LDCs have fallen back in the highly competitive and regulated marketsthat trade liberalisation has created.Historically, the natural progression of countries in export trade has beenfrom commodities to labour-intensive manufactures to higher technologygoods to capital goods and, finally, to the services sector. WTO negotiationsfor the liberalisation of the service sector have yet to come. Mostdeveloping countries and LDCs, however, are either in labour-intensivemanufacturing or drawing up plans to enter that area. Many are activein agro-processing, which requires less investment and technology, whilea small number are to be found in high technology or services (besidestourism).3.2.2 <strong>Trade</strong>, industry and poverty reduction“I am proposing that industry-related trade issues shouldbe accorded greater importance since the industry-tradedevelopment-povertynexus is of major concern to developingcountries in the light of the Doha Round of trade negotiationsand regional initiatives such as NEPAD.” Dr. K. Yumkella, UNIDODirector-General,(Towards Pro-Poor Sustainable Industrial Development: A Shared Visionfor UNIDO, 2005)The role of an open trade regime in stimulating growth and fighting povertyhas been widely researched and debated by academic economists.According to one study, an increase of one percentage point in the shareof trade in GDP raises income by at least one-half of 1% 29 . The same studysuggests that trade increases income by spurring investment in humanand physical capital, as well as by stimulating productivity.Empirical analysis of country-specific data also confirms the existenceof a link between economic growth and poverty reduction. For example,the experience of India, over the past two decades, shows a correlationbetween growth and poverty reduction on the one hand and increasedopenness to trade on the other.<strong>Trade</strong> is inextricably linked with growth and poverty alleviation in developingcountries. The developing countries’ low internal purchasing capacitycannot support industrial expansion: it requires the expansion of trade toachieve sustainable development and create employment.The World Bank estimates that a successful Doha round could lift 140-millionpeople out of extreme poverty 30 . The impact on Africa alone would bean additional income of US$70bn, about five times the current aid flowto the continent 31 .The poverty-reducing effects of an open trade regime are likely to bestronger in the case of developing countries because trade in goods is amore important source of income than in the developed world. However,the relatively large share of trade in developing country’ GDPs also meansthat their economies are more exposed to volatility in demand. The LDCsare extremely vulnerable; their export earnings tend to depend on one ortwo commodities, for which current world prices are inordinately low.Greater participation in international trade has replaced import-substitutionindustrialisation as the key development policy of developingcountries. In the liberalised and competitive global markets, developingcountries have to: (i) enhance their industrial supply capacity; (ii) conformto an increasingly complex and ever-expanding rules-based internationaltrading system; and (iii) ensure cleaner and sustainable production.109<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>26World Bank Official Website: http://econ.worldbank.org.27UNIDO (2004): Industrial Development Report 2004; UNIDO, Vienna (Austria).28UNDP (2005): Human Development Report, UNDP, New York (USA). UNCTAD(2005): <strong>Trade</strong> and Development Report 2005; UNCTAD, New York (USA) and Geneva(Switzerland).29Frankel, J. and Romer,D. (1999): Does <strong>Trade</strong> Cause Growth?; American EconomicReview, 89 (3): pp379-99.30World Bank (2003): Press Release (September 12th), World Bank Announces Stepped-Up<strong>Trade</strong> Assistance Programme; Washington, DC (USA).31House of Commons-International Development Committee (2002-2003).The industrial challenge facing Africa in the Global Trading System
Industrial PolicyFIGURE 5: WORLD TRADE DEVELOPMENT, 1970-2003110<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: UNCTAD3.2.3 The position of African countries in global tradeOver the past 20 years, Africa’s share of world exports has fallen from4.5% to 2.4%, while the number of people living in absolute povertyhas increased from 42% to 47% 32 . In terms of manufactured exports,Africa lags behind every other region; a mere 0.9% of its total exportsare in manufactured goods 33 (see Figure 5). This contrasts sharply withdeveloping countries as a whole, which currently account for roughly onethird of global manufactured exports, up from 18% in 1980 34 . Moreover,while South-South trade has increased to account for some 40 per cent ofthe developing countries’ total exports, the share of Africa in South-Southexports of merchandised goods has declined from 5.4% in 1970 to 2.3%in 2003 35 .A recent study by the Economic Commission for Africa (ECA) on the incidenceof non-tariff barriers shows that 48% of agricultural and fish exportsfrom LDCs to developed economies face non-tariff barriers such as productstandards, phyto-sanitary and environmental controls and are morelikely to be subject to anti-dumping measures 36 . This is significant sinceAfrica’s greatest comparative advantage lies in agro-based industries 37 .In addition, the large number of TBTs that exist between the developingcountries themselves merely serves to make matters worse. It is estimatedthat 70% of the benefits from a successful Doha Round would accrue todeveloping countries as a result of increased South-South trade <strong>38</strong> .3.3. Failure to benefit from trade and thecontributing factorsThe multilateral trade regime has created an environment in which themost competitive countries benefit most from trade liberalisation. Withoutefficient supply-side capabilities, developing countries will not be able toparticipate meaningfully in international trade nor will they benefit fromit. Developing countries must improve the competitiveness of their productsand services through efficient upgrading of their industrial productivecapacity at both the institutional and enterprise levels 39 . In Africa, in36UNECA Study (2005).32UNIDO (2004): Industrial Development Report 2004; UNIDO, Vienna (Austria).33WTO (2004): International <strong>Trade</strong> Statistics 2004; WTO, Geneva (Switzerland). www.wto.org/english/res-estatis-e/its2004_e.pdf34UNDESA (2006): Trends in Sustainable Development; New York, May 2006, FourteenthSession of the Commission on Sustainable Development.35Awori, M. (2006): Official Opening Speech of the 4th ordinary session of the AfricanUnion <strong>Trade</strong> Ministers Conference on April 14th, 2006 in Nairobi (Kenya).37None the less, the stark truth is that only a small proportion of produce is being processed.Taking cotton production as an example, market prices rise along the value chain fromraw cotton to carded/combed to yarn. Yet 96% of Africa’s cotton exports are in raw cotton,whereas 90% of East and South Asian exports are in the form of processed yarn (UNIDO,2005).<strong>38</strong>World Bank: http://econ.worldbank.org.39Dhaoui, M.L. (2002): Guide for Restructuring, upgrading and industrial competitiveness;UNIDO, Vienna (Austria).
FIGURE 6: SHARE OF WORLD MANUFACTURED EXPORTS, 19980-2000111<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: UNIDO scoreboard database, Industrial Development Report 2004common with most developing countries, a number of factors at bothlevels hinder effective participation in global trade. These factors includemacro-economic instability, lack of human and physical capital, weak infrastructureand economic governance institutions, as well as a poorlydeveloped private sector 40 .3.3.1 The major impediments to trade in AfricaLack of supply capacityThe foremost impediment is a lack of capacity to produce a surplus ofexportable goods of sufficient quantity, stable quality and required standardthat can be traded internationally. As the United Nations Secretary-General, Kofi Annan, said “LDCs have neither the surplus of exportableproducts nor the production capacity to take immediate advantage of newtrade opportunities. They will need substantial investment and technicalassistance in order to expand their production.” On the production side,major constraints include: a labour force that is hampered by poor healthand insufficient nutrition; a labour force suffering from a lack of educationand training; an infrastructure unable to cope with adverse geography, thelandlocked situation and erratic energy supplies; and capital equipment40South African Department of <strong>Trade</strong> & <strong>Industry</strong> (2001): “The Millennium Partnership forthe African Recovery Programme: A Market Action Plan for Africa”, South Africa. www.nepad.org/2005/files/documents/24.pdfthat is often old and cannot be easily replaced for want of funds (limitedaccess to finance). These constraints result in production at relatively lowutilisation rates and equally low productivity levels.At the enterprise level, African countries suffer from a lack of: (i) capabilityto assess and address their technical needs; (ii) sufficient technology torespond to rapidly evolving markets; (iii) adequately trained skilled workers(human capital); (4) effective industrial productive capacity (physicalcapital); and (v) access to finance.All this underscores the need for financial resources and technical assistancein the area of productive capacity upgrading. Not only do productivesectors with high export potential have to be identified through a comprehensivemarket and product analysis, but in the interests of both socioeconomicand environmental sustainability, scarce resources have to beefficiently and effectively allocated to the sectors so identified. Diagnosticand upgrading programmes have to be formulated for the pilot enterprisesselected and matching investment programmes implemented.In the final analysis, upgrading productive capacity yields improved competitivenesswhich, in turn, allows developing countries to: (i) producehigher value-added products; (ii) enter/capture more markets; iii) experiencetrade growth; and (iv) enjoy the benefits of employment creation andpoverty alleviation.The industrial challenge facing Africa in the Global Trading System
Industrial Policy112<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>These and other aspects are covered extensively in the paper on improvingproductive capacity.Lack of comparative analysisGiven that 75% of global trade today is in manufactures, attaining andmaintaining competitiveness in the manufacturing sector calls for continuousmonitoring and assessment of global and regional trends in bothmanufacturing and trade. Identifying sectors or subsectors that have thegreatest competitive potential and are sustainable in the medium andlong term takes on particular importance. Starting from an overview ofcompetitive performance of the industrial sector as a whole, it is possibleto work down to an assessment of sectoral performance and of performanceat the product level and see which stages in the production processyield the greatest share of manufacturing value-added or at which stagecompetition from peers is weakest or strongest. Competitive analysesbecome essential tools in ensuring that scarce resources are allocatedefficiently and are focused on a sector or sub-sector that stands to benefitmost from industrial upgrading, productivity enhancement or enterpriserestructuring.Tariff and non-tariff barriers to international tradeA World Bank study carried out in 2002 established that low-incomecountries face tariffs on global markets that are more than double thosefaced by non-poor producers. To make matters even worse, these marketaccess barriers are particularly high in sectors where developing countrieshave a comparative advantage such as agriculture and labourintensivemanufacture. Non-tariff measures, which include agricultural subsidies,antidumping duties and threats, as well as rules of origin also affect thetrade performances of developing countries, especially in the manufacturedgoods sector 41 .labelling and testing, and certification requirements 43 . Meeting those marketrequirements call for upgrading skills and capabilities, mastering newtechnologies, and enhancing old and establishing new institutions (e.g.accreditation bodies, metrology, standardisation and technical supportfacilities). The costs implications can be very high relative to the value ofexports, they can thus pose a formidable barrier to exports.Poor infrastructure at the institutional levelThe problems related to poor export and infrastructure capacities includea low skill base, low level of technology adoption, absence/weaknessof institutions to support private sector development (industry associations,export agencies, productivity centres, technology information centres,marketing boards, R&D laboratories and cluster development programmes)and weak capacity for the mobilisation of domestic and foreigninvestment.Lack of effective market integrationEstablishing a successful presence in foreign markets is considerablymore difficult than in domestic markets. SMEs, which constitute the largemajority of African enterprises, are often deterred from exporting by thecomplexities of the export business and the high risks involved. Withoutsignificant knowledge and preparation, attempts to export are doomedto failure and may even jeopardise the financial stability of the enterpriseas whole. Evidence suggests that particularly during the early stagesof exporting, failure rates are relatively high. To develop export markets,significant know-how, effort and financial resources are needed. Foreignmarkets have their particularities and stringent requirements. By cooperatingwithin an export consortium that combines the expertise and financialresources of several firms, the enterprises can reduce their risks, improvetheir profitability, secure efficiency gains and accumulate know-how.Inability to meet market standards, conformity and certificationrequirementsRecent enterprise surveys 42 in both developed and developing countriesshow that the highest non-tariff barriers include performance standards,product quality standards, technical measures, product requirementsrelating to standards and technical regulations, conformity assessment,41Hoekman,B. and Olarreaga,M. (2002): : “Tariff Peaks in the Quad and Least DevelopedCountry Exports”; World Bank Economic Review, <strong>Vol</strong>. 16, p1-22.42Including those carried out in Africa under the UNIDO-WTO joint programme.3.3.2 The need for compliance and conformityIn today’s global market, meeting standards or other technical requirements(certification, measurement, testing) has become essential to tradesuccess. The International Development Research Centre in Canada foundthat countries that cannot meet standards and regulations in developedcountry markets are effectively barred from trading with those markets.43World Bank: Press Release (September 12th, 2003), World Bank Announces Stepped-Up<strong>Trade</strong> Assistance Programme; Washington, DC (USA).
International standards make an important contribution to the globaleconomy since they improve the efficiency of production and trade andreduce consumer costs. As a result, compliance with standards has becomea requisite for the expansion of inter-regional and internationaltrade. However, standards and technical regulations drawn up by individualcountries to protect health and the environment, as well as to ensurequality and safety, can also act as technical barriers to trade. Increasingly,international buyers require effective application and recognised proof ofenterprise system management standards such as ISO 9000 for qualitymanagement, HACCP and ISO 22000 for food safety, ISO 14000 for environmentalmanagement and SA 8000 for social accountability, as wellas private-sectorled standards such as EUREPGAP. ‘The situation is madeall the more complex with the emergence of an ever-increasing number ofproduct-specific standards.For the most part, these ’market requirements’ can be classified as TBTor SPS measures. Despite the WTO TBT and SPS agreements 44 reached inthe context of the Uruguay Round and designed to ’ensure that technicalregulations and standards do not create unnecessary obstacles to trade’,such market requirements can still constitute non-tariff barriers that obstructaccess to markets. The obstruction is all the larger since thousandsof new standards and private product specifications, as well as relatedconformity assessment procedures, are being introduced each year.For most developing countries, standards have become impediments totrade. Products often fail to comply with international standards and certificationrequirements. The technical regulations and standards applied, includingpackaging, marking and labelling requirements, are often incompleteor incompatible with international standards. Furthermore, Africancountries often do not have the necessary laboratory facilities to test andcertify goods for developed markets, while the tests their laboratories performand the certificates they issue are not recognised in export markets.As a result, African countries need to enhance (or establish) institutionalinfrastructure to ensure that they produce goods that are of a standardthat can be sold on the global market.They also need to upgrade standardisation agencies and services alike,as well as develop – where appropriate – national testing, certification,inspection, laboratory and accreditation infrastructure that conforms tothe requirements of the WTO SPS and TBT agreements and thus enjoysinternational recognition. Finally, they need to combat the lack of aware-44Discussions are taking place at the WTO on how to deal with private-sector standards andtheir relation to the TBT and SPS agreements.ness on standards and regulations and the lack of investment in creatinga solid knowledge base at both the institutional (R&D) and human capital(training) levels.African producers must be able to prove the reliability of their test dataand inspection procedures, as well as the conformity of their productsto international standards and/or those applied in the recipient country.Testing, calibration and certification facilities are therefore of extremeimportance. In the industrialised countries this infrastructure is taken forgranted; for the most part, it passes unnoticed. The developing countries,however, particularly the LDCs, lack even the most rudimentary elementsof the complex infrastructure. If such facilities are not recognised internationally,a country’s trade potential is seriously hampered and the pricesits products can command on global markets are correspondingly low.Moreover, local metrological and testing capabilities reduce the associatedcosts for products that would otherwise either have to be testedabroad or locally by international certification companies at very highrates. Effective assessment also provides domestic companies with objectiveresults that are essential to improving designs and technologies andassuring quality.In addition, weak quality infrastructure (standards, conformity assessment,import inspection, markets surveillance) has a marked negative effect onAfrican economies and their citizens.The weakness also prevents countries from controlling the quality, quantityand safety of imported and locally produced goods. African countriesthus become the ‘dumping ground” for low quality/sub-standard productsthat can cause dramatic health problems for the local consumers andreduce further the ability of domestic producers to compete in terms ofprice with imported goods.If African countries are to meet standards and conformity requirements,major investments are needed to establish and upgrade the standardisation,metrology and conformity assessment infrastructure. For example,UNIDO estimates that the added cost to the developing countries of complyingwith food and safety regulations for the export of shrimps to theEU market is as follows: (i) 2.8% cost increase for testing and complianceactivities; (ii) 5% cost increase for setting up quality and safety activitiesat the enterprise level; (iii) US$5m - US$25m for setting up quality andtesting infrastructure at the national level. In another instance, the WorldBank has estimated that the Ugandan honey industry would require up toUS$300m to construct the processing facilities and purchase the equipmentneeded in order to be able to conform to ISO food safety standards,113<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The industrial challenge facing Africa in the Global Trading System
Industrial Policy114<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>while the Ugandan coffee producers would, on average, see their productioncosts go up by 200%, if compliance costs for good quality coffeewere included 45 .3.3.3 Beyond complianceIn addition to their being hampered by supply-side and technical constraints,the developing countries’ failure to exploit the export opportunitiesoffered by the liberalised global trading system is compounded byshortcomings in such essential trade-related requirements and proceduresas registration and documentation, customs valuation, import regulations,licensing and similar issues. Strengthening customs capacity, issuing import/exportlicences more effectively, extending computerisation, improvingpre-shipment inspection methods, increasing the transparency of therules of origin more strictly and simplifying import/export procedureswould facilitate the cross-border movement of goods as well as heightencredibility in export markets. Similar benefits stand to be gained by improvingtransportation facilities at ports and airports.The effective dissemination of international trade rules, particularly thoseagreed within the context of the WTO, as well as regulations governingfinancing, direct investment and exchange measures in various countriesreduce the scope for misunderstandings and result in fuller integrationwith the international trade framework. This information and effectivemarket intelligence enable the developing countries to keep pace withshifts in global market trends. If the small and medium enterprises in theleast developed countries in particular are to make use of the vast amountof information required to improve productivity, they need that information;however, they need operational solutions as well.One such operational solution is the creation of export consortia, whichare voluntary alliances of firms with the objective of promoting the exportof goods and services of their members through joint actions. The twomain types of consortia are promotional and sales consortia. Promotionalconsortia are created to explore specific export markets by sharing promotionaland logistical costs among participating firms. Actual sales, however,are the responsibility of the individual firms. Sales consortia, on theother hand, also perform business promotion activities, but they handlethe sale of member firms’ products as well.45World Bank (2003): “Standards and Global <strong>Trade</strong>: A Voice for Africa”; World Bank,Washington, DC (USA).3.4. Main trade capacity buildinginitiatives3.4.1 Definition of termsOne concept, many definitionsDonors and agencies working with TRTA/CB usually see TCB as “one (outof many) means to promote economic development and reduce poverty,by supporting developing countries’ greater participation in the multilateraltrading system and their integration into the world economy” 46 .UNIDO defines `trade capacity’ as “the ability of domestic providers ofgoods and services to penetrate a related market in a foreign country” 47 .According to the 2005 Joint WTO/OECD Report, TRTA and CB can bedefined as activities that intend to enhance the ability of the recipientcountry to:Formulate and implement a trade development strategy and create an enablingenvironment for increasing the volume and value-added of exports, diversifyingexport products and markets and increasing foreign investment to generate jobsand trade;Stimulate trade by domestic firms and encourage investment in trade-oriented industriesor;Participate in and benefit from the institutions, negotiations, and processes thatshape the national trade policy and the rules and practices of international commerce.”48Delivery of TRTA/CBActivities in the field of TRTA/CB can generally be implemented via: (a)technical assistance or (b) financial assistance. Technical assistance includessuch activities as training sessions, workshops, studies, and consultancieswith the main objectives of either the provision of technicalknowledge or of strategic advice on policy issues.Financial assistance is disbursed through loans or grants and has the keyobjective of overcoming a lack of capital 49 . TRTA/CB focuses on aiding the46OECD (2005): 2004 Overview of Donor and Agency Policies in TRTA/CB; DCD (2005)10/REV1; April 2005, Paris (France).47UNIDO (2002): “Enabling Developing Countries to Participate in International <strong>Trade</strong>-Strengthening the Supply Capacity. A UNIDO Strategy for Capacity Building” during theWTO Financing for Development Conference held from 18-22 March 2002 in Monterrey(Mexico).48WTO/OECD (2005): Joint Report on <strong>Trade</strong>-Related Technical Assistance and CapacityBuilding.49Miller, E.T.(2005): “Achievements and Challenges of TCB: A Practitioner’s Analysis of theCAFTA Process and its Lessons for the Multilateral System- Occasional Paper 32; ITAL,Buenos Aires (Argentina)
ecipient countries in such a way that will subsequently allow them tocarry out their trade-related activities in a sustainable and independentmanner.In that particular context, the proportion of ODA provided to help buildtrade capacity is an indicator used in connection with one of the targetsset under the MDGs: more specifically in relation to making debt sustainablein the long term [Target 15]. It is quite conceivable that by buildingeffective trade capacity and so increasing the potential for greater exportearnings, trade capacity building can be seen as one of the principalsources for sustaining debt. In the ultimate analysis, it might well ensuethat trade capacity indicators may be prescribed as a conditionality forODA flows.3.4.2 Progress to date 50OverviewOver the past few years, aid for TRTA/CB has experienced a remarkableboost in the wake of increased awareness regarding the importance oftechnical barriers to trade. Both bilateral and multilateral donors havebecome much more active. The total for TRTA/CBrelated assistance grewfrom just over US$2bn in 2001 to almost US$3bn in 2004. A similar upsurgecan be observed in terms of the number of activities recorded; theseincreased from 4,339 in 2001 to 8,195 in 2004.After a major increase in 2003, TRTA/CB remained stable overall in 2004,whereas commitments to trade policy and regulations decreased fromUS$934m in 2003 to US$811m in 2004. At the same time the volumeof aid allocated to trade development rose by US$140m to reach nearlyUS$2.2bn in 2004. In addition, donors committed US$9.3bn to supporteconomic infrastructure: transport, energy and telecommunications.Focus on AfricaIn 2004, the total amount of TRTA/CB committed to Africa stood atUS$980.7m – nearly twice the amount in 2001 US$599.2m. This substantialincrease reflects a rise in aid to both trade policy/regulations andtrade development. The number of activities recorded in Africa also increasedsignificantly over the same span of time. In 2004, trade developmentaccounted for 83% of total TRTA/CB expenditure in Africa and tradepolicy/regulation for the remaining 17%.In terms of the distribution of TRTA/CB-related aid between North Africaand sub-Saharan Africa, it can be seen that the southern part of the regionattracts most assistance. In both 2001 and 2004, aid to sub-SaharanAfrica was significantly greater than aid to North Africa. In 2001, totalassistance in the area of TRTA/CB allocated to North Africa amountedto US$154.8m, while sub-Saharan Africa was allocated US$417.8m.Similar patterns can be found in 2004, when North Africa received 34%(US$288.4m) of total TRTA/CB-related aid, and sub-Saharan Africa 66%(US$567.3m). The only exception is trade policy/regulations aid to NorthAfrica in 2001, which exceeded aid to sub-Saharan Africa in the same categoryby some US$5m. In both regions, assistance to trade developmentconsistently outstripped that to trade policy/regulations.In 2004, the main bilateral donors of TRTA/CB-related aid to Africa werethe United States of America (US$163 million), the UK (US$4.4m) andFrance (US$24.9m). In 2004, the EU delivered more than US$169m ofTRTA/CB-related aid to Africa, US$110m of which went to trade development.In 2004 the largest multilateral donor of TRTA/CB-related aid toAfrica was the International Development Association (part of the WorldBank); it gave US$80m to trade development and US$10m to trade policy/ regulations.3.4.3 Major agencies and initiativesMain agenciesAs indicated in a joint WTO/OECD report covering the period 2001-2004 51 , UNIDO is the most active multilateral agency in the area of TRTA/CB. UNIDO TRTA/CB projects reached US$39.8m in 2004. They went onto increase to US$64.6m in 2005 and projects to the value of an additionalUS$90.4m are currently under negotiation with donors and recipientcountries. The other major multilateral donors include ITC, whichimplemented projects worth US$29.9m in 2004, followed closely by FAO(US$25.4m) and WTO (US$19.4m).Major InitiativesTHE INTEGRATED FRAMEWORK FOR TRTA TO LDCS (IF)The IF, a multi-agency and multi-donor programme, was first set up at the WTO MinisterialConference in Singapore in December 1996 and launched in October 1997. Theparticipating agencies are the IMF, ITC, UNCTAD, UNDP, the World Bank and WTO. In115<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>50The figures and analysis in this section are taken direct from the 2005 WTO/OECDReport on TRTA/CB.51Joint WTO/OECD Report on <strong>Trade</strong>-related Technical Assistance and Capacity Building,Geneva/Paris, December 2005The industrial challenge facing Africa in the Global Trading System
Industrial PolicyFIGURE 7: TRTA/CB IN NORTH AND SUB-SAHARAN AFRICA, 2001-2004<strong>Trade</strong> policy and regulations ($ million)<strong>Trade</strong> development ($ million)116<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: WTO/OECD data cubeJuly 2000, the agency heads met to review the programme and recommended severalchanges in order to enhance its effectiveness. The number of countries benefiting fromthe IF has grown steadily; at the end of March 2005, 28 countries were involved, all ofwhich were at different stages of the IF process.The IF has two main objectives: (i) to mainstream and integrate trade fully into nationaldevelopment plans; and (ii) to assist in the coordination and delivery of trade-relatedassistance provided by each of the core agencies in their specific field of competenceand by other development partners (see also www.integratedframework.org). The IF wasoriginally conceived to provide assistance in all areas related to supply development, aswell as in the areas of standards, testing, quality and conformity. However, even it its ‘revamped’form, the focus remains on customs and export marketing, with the result thatthe benefits accruing to the least developed countries have been relatively limited.“Aid for <strong>Trade</strong> should aim to help developing countries,particularly LDCs to build supply-side capacity and trade-relatedinfrastructure that they need to assist them to implement andbenefit from WTO agreements and more broadly expand theirtrade. It can be a valuable complement to the Doha DevelopmentAgenda, particularly on market access” 52 .This explicit recognition of a WTO interest in and responsibility for aid has raised manyexpectations; however, the recently designated Task Force on Aid for <strong>Trade</strong>, together withthe WTO Director-General, still has to clarify the terms of the initiative in order to fulfilthose expectations.THE GLOBAL TRUST FUND AND DOHA DEVELOPMENT AGENDA TRUSTFUNDBoth the Global Trust Fund and the Doha Development Agenda Trust Fund, establishedin 1999 and 2001 respectively, provide a vehicle through which WTO member statescan make extra-budgetary contributions to ensure: (a) the participation of all developingmember countries in WTO negotiations; and (b) financial and technical assistance fortrade-related capacity building.THE AID FOR TRADE INITIATIVE (AFT)The WTO defined the scope and objectives of the AFT in December 2005 during the WTOMinisterial Conference in Hong Kong as:The AFT is also re-introducing components that though initially envisaged in the IF weresomewhat neglected subsequently. The AFT initiative calls for further emphasis on providingTRTA in the field of strengthening/upgrading supply-side capacities and conformityassessment infrastructures: areas that correspond exactly to the areas of competenceof UNIDO.3.4.4 The trade capacity building initiative of UNIDOUNIDO launched its trade capacity building initiative at the InternationalConference on Financing for Development in Monterrey, Mexico, in March2002. It has also placed trade capacity building at the top of its corporatestrategy and development agenda. UNIDO has thus realigned the focus of52WTO (2005): Mandate- WTO Hong Kong Ministerial on Aid for <strong>Trade</strong>, paragraph 57,2005.
its technical assistance. In March 2006, it established a branch specificallydedicated to trade capacity building and has since entered into strategicpartnerships with other key organisations.The main goal of the UNIDO initiative 53 is to promote export-orientedproduction, within the framework of poverty reduction. It focuses on enhancingsupply capacities and securing for the developing countries anincrease in both their manufactured exports and value-added, thus makingfor a greater share in global trade. The initiative provides developingcountries with technical assistance for the upgrading of their institutions,support infrastructures and competitive productive capacities in order to:facilitate their integration into the world market, overcome both tariff andnon-tariff barriers to trade and, ultimately, create a favourable trade environment.The strategic approach of UNIDOThe reasons for the developing countries’ failure to benefit from the opportunitiesoffered by the rapidly changing global markets are not relatedsolely to tariffs and quotas; other factors also play a role. As the developingcountries themselves recognise, they lack:a) Effective industrial productive capacity needed to ensure optimisationof production and product diversificationb) Ability to comply with international standards needed to exploitthe opportunities offered by the liberalised global trading systemc) Equitable integration into the multilateral trading systemUNIDO responds to these problems through its three-pronged trade capacitybuilding approach: the 3Cs: Developing competitive manufacturing capability; Developing and promoting conformity with market requirements; Enhancing connectivity to markets.In its approach UNIDO focuses on upgrading standards, metrology, testingand quality infrastructure and services in relation to TBT and SPS, as wellas on market analysis and development more closely related to supplysidecapacity and access to export markets.However, complementary to those operational activities at the plant andinstitutional levels, UNIDO actively supports the developing countriesin the creation of export consortia in different sectors, training national53In his recent book “Fair <strong>Trade</strong> for All”, Joseph Stiglitz refers to the UNIDO TCB initiativeas presented in Monterrey in 2002.promoters of such consortia in both the public and private spheres andpromoting a favourable institutional and regulatory environment for theirdevelopment. In terms of regional focus, Africa is the main recipient in ofUNIDO SMTQ/TCB projects: 34 projects over the period 2001-2005 (seeFigure 9).The strategic partnerships of UNIDOUNIDO signed a cooperation agreement with the WTO at Cancun in 2003,under the terms of which assistance is provided to the developing countriesin removing supply-side constraints, developing systems to prove conformitywith market requirements and enabling better integration into themultilateral trading system. In addition to providing for close involvementin each other’s work, the partnership currently covers nine pilot countriesand the joint WTO/UNIDO cotton initiative for African countries.UNIDO actively pursues a policy of close operational partnership withother UN agencies such as UNCTAD, FAO and UNEP, as well as with technicalorganisations such as ISO, ILAC and IAF. UNIDO currently acts assecretariat to the Joint Committee on Coordination of Assistance to DevelopingCountries in Metrology, Accreditation and Standardisation (JCD-CMAS). Established in 2002, the joint committee seeks to bring togetherall specialist organisations that operate at a global level and are activein promoting metrology, accreditation and standardisation as a tool forsustainable economic development.3.4.5 Principles governing assistance to trade capacitybuildingIn an OECD report assessing assistance related to trade capacity building,success was said to hinge on four key principles 54 .1. COORDINATION OF EFFORTSThe scope of successful TRTA/CB in developing countries is beyond the means of anysingle multilateral or bilateral donor. Consequently, donor coordination is essential. Requiredpolicy frameworks cannot function effectively if the institution and arrangementsconstituting it are assembled and/or strengthened independently. Furthermore, in orderto save resources, increase effectiveness and respond to regional needs, efficient donorcoordination is primordial.54The following 4 paragraphs are taken direct from the 2001 OECD report: The DACGuidelines, Strengthening <strong>Trade</strong> Capacity for Development; OECD, Paris (France).117<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The industrial challenge facing Africa in the Global Trading System
Industrial Policy2. COMPREHENSIVE AND INTEGRATED ACTIVITIES5. FOCUS ON SUPPLY-SIDE CONSTRAINTSComprehensive and integrated efforts in development cooperation are critical in achievingviable trade policy frameworks. This requires action in multiple areas and the inclusionof a wide range of key stakeholders in the design, implementation and evaluationof projects.Developing countries, especially African countries, need assistance in:a. Removing supply-side constraints in order to become more productive and competitivevia upgrading programmes at the institutional and enterprise level.118<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>3. PROMOTE LOCAL OWNERSHIP AND PARTICIPATIONCivil society and local business people are the key stakeholders of any TRTA/CB activities,since they are most acutely affected by trade capacity constraints and trade developmentchallenges. Hence, local participation and consultation are defining features of aneffective trade framework. This will ensure that the trade development process is locallyowned, demand driven and will foster local ownership.4. FOLLOW SUSTAINABLE APPROACHESb. Strengthening/ establishing internationally credible standardisation, metrology,testing and quality (SMTQ) frameworks and infrastructures in order to meet themultiplicity of technical requirements and standards set by the markets.c. Enhancing their ability to analyse market trends and opportunities and so improvetheir capacity for formulating effective export strategies and approaches.Consequently, the international development community, as well as multilateral and bilateraldonors, should increase their allocations to the supply-side and SMTQ aspects oftrade capacity building.Socio-economic sustainability should be the key objective of any TRTA/CB activity and willensue, if processes are truly participatory and inclusive. None the less, explicit attentionhas to be paid to encourage sustainability when designing and implementing projects,for example through a greater use of national experts (ibid). Environmental sustainability,should, also be an integral element of project designing and implementation.A fifth principle can be added to the list:FIGURE 8: UNIDO TRADE CAPACITY BUILDING3.4.6 In a broader settingWith the overall aim of securing greater developing country participationin global trade, TRTA/ T CB activities can be seen to be a step in the rightdirection that will ultimately lead to achieving two major MDGs: povertyreduction and a global partnership for development.The significance of industry and trade and their close correlation to povertyalleviation are undisputed. The correlation between a global partnershipfor development and trade capacity building is equally significant asachievement of that partnership hinges on an open trading and financialPProduct to marketMRADUCDevelop competitivesupply-sidecapacitiesProve conformitywith marketrequirementsConnect to themarketRKETTSCompete Conform ConnectS...by upgrading supply-side capacity and quality infrastructure
FIGURE 9: UNIDO SMTQ PROJECTS, IMPLEMENTED, 2001-2005119<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: UNIDO datasystem that is rules-based, predictable and non-discriminatory, while takingaccount of the special needs of the LDCs, including quota-free accessfor their exports. It can thus be seen that in its efforts to surmount thechallenges that the developing countries face in their endeavours to producea surplus of exportable goods, the UNIDO trade capacity buildingprogramme complements the targets set in the MDGs. In a world of interconnectedthreats and opportunities, it is in each country’s self-interestto address the challenges they face: something that can only be achievedthrough collective action at the national, regional and global levels.3.5. Issues to be discussedIn view of the foregoing, the following concerns have been identified asthe main issues to be addressed: What are the main reasons for the failure of developing countries in general andAfrican countries in particular to benefit from the opportunities offered by the globalmarket? How can developing countries, in general, and African countries, in particular, gainappreciable benefits in terms of economic growth and wealth creation from greatermarket penetration and fuller participation in global trade? What are the supply-side constraints hindering the African countries’ trade performance? How are the productive sectors with high export potential best identified and assessed?How can the productive sectors selected be assisted to enhance their competitivenessand increase their access to export markets?What is the critical mass of support, institutional and otherwise, that African industriesand enterprises require to upgrade and improve competitiveness?What is the critical mass of infrastructural/institutional support required at the nationallevel?What is the critical mass of infrastructural/institutional support needed at the regionallevel to foster exports and reduce the inflow and production of hazardousproducts that pose a high risk to African consumers?Can developing countries, in general, and LDCs, in particular, afford to invest in thatquality infrastructure?How can the provision of cost-effective, sustained and internationally recognisedservices in testing, conformity assessment and metrology be enhanced through cooperationat the national, regional and international level?How can the constraints that SMEs encounter in their endeavours to secure entryinto foreign markets be best overcome?Can trade capacity targets be measured in terms of the indicators used in theMDGs?The industrial challenge facing Africa in the Global Trading System
© Greatstock/X Collection120<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>
south africa’sgrowthtrajectory121AN ASSESSMENT OF THEACCELERATED & SHARED GROWTHINITIATIVE<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>BY ROB DAVIESUNIVERSITY OF ZIMBABWEAND DIRK ERNST VAN SEVENTERTIPS1. IntroductionWHILE IT IS DIFFICULT TO pin down with any great accuracy what the Accelerated& Shared Growth Initiative for South Africa (ASGISA) entails in terms offiscal expenditures, it nevertheless suggests that there is a renewed urgencyamongst South Africa’s policy makers to redress the neglect of public sectorinfrastructure expenditure that occurred during the heyday of the Growth, Employment and Redistribution(GEAR) initiatives of the late 1990s and early years of the following decade. A number ofadditional concerns have been raised in this debate around financing of the initiative, skill shortages,productivity impacts and reduction of import content.Policy makers are now keen to understand the economy-wide impact of the ASGISA and an assessmenthas been called for. The central part of ASGISA seems to be investment. This will impacton the economy through its demand and through its supply effects. Macoeconomists would usesome form of macro multiplier analysis to look at the demand impact but would point out thatthis rise in demand will not necessarily translate into a rise in income. It assumes that there are noUsing a comparative static economy-wide policymodel, the authors evaluate an infrastructureinvestment program for its impact on the SouthAfrican economy. The infrastructure investmentprogram is loosely based on the Accelerated& Shared Growth Initiative for South Africa(ASGISA). In addition to the usual demand sideinjections, particular attention is paid to supply sideimpact of the investment program, productivityenhancement and skilled labour shortages. Thenational accounting consistency requirements ofthe modeled economy dampen the final impactof the demand injection considerably. Althoughproductivity enhancement will have an additionalpositive impact this will be reduced by skilledlabour shortages.
South Africa’s Growth Trajectory122<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>macroeconomic crowding-out effects. Indeed, some styles of macroeconomicswould argue that raising demand in this way will have no effecton output, since they believe there is a completely inelastic aggregatesupply curve. ASGISA from this perspective would simply cause inflationand change the composition of the economy towards the public sector.Clearly output in South Africa is not as constrained as this – there are unemployedresources, particularly labour, that could be brought into play toexpand production. But equally clearly, aggregate supply is not perfectlyelastic, able to expand to meet any demand increase. Finite existing productioncapacity, import capacity and skills (amongst other) will constrainthe economy’s ability to grow.The architects of ASGISA recognise this when they emphasise the capacitycreating nature of the investment. It is not simply any investment but istargeted at specific sectors, particularly infrastructure, in order to try toease bottlenecks.Another simple analytical framework that can be considered for such anassessment is a SAM multiplier model. In the same way as the macromultipliermodel, this framework typically multiplies the initial expenditureassociated with ASGISA at an industry/commodity level up. This can potentiallybe counterbalanced by negative impact of various financing optionssuch as higher taxes or higher private savings. Such an approach hasbeen taken by McCord & van Seventer (2005) to evaluate the impact ofexpenditure switching of Public Works Programs from a machine based toa labour based approach. Their model is best suited for the evaluation ofexpenditure programs of limited size. However, ASGISA is more substantialand important potential broader impacts of ASGISA, such as raisingproductivity and reducing skill shortages, cannot be considered in such asimple first generation SAM based multiplier framework.Policy makers are all too aware of unintentional counterbalancing effectsof expansionary fiscal policy and would like to account for such eventualitiesin their design. We therefore employ Computable General Equilibrium(CGE) framework, which in its standard format accounts for a richer setof feedback effects that the framework used by McCord & van Seventer(2005) due to flexible prices and substitution effects while in addition it allowsfor further assumptions around productivity effects and skill shortages.We use a simple comparative static framework similar to the standardIFPRI CGE model for South Africa (Thurlow & van Seventer, 2002, whichwas updated and applied recently by (Davies & van Seventer, 2005). TheCGE framework is characterized by standard neoclassical attributes butallows for structural features relevant to the South African economy. A fulldescription of this type of model is found in Lofgren (2000).2. DataIn order to evaluate the economy-wide impact of ASGISA, the first issuethat needs to be tackled is how this initiative differs from a business asusual scenario. We will be very modest in terms of what can reasonablybe associated with the ASGISA program and are only considering the“economic infrastructure priorities” and ignore other priorities identified,such as “meeting skills needs”, “social security, health and education”,“housing and neighbourhood development”, “industrial development”,“justice”, “international relations, peace and security” and “public administrationreform”, all important areas but beyond the scope of ouranalysis here. The 2006 Budget documentation offer limited specific numberson “economic infrastructure priorities” such broad categories suchas road, rail, Gautrain and Industrial Development Zones. The total “additional”amount allocated under this chapter is about R13 billion over 3years. For more detail we then turn to earlier Medium Term ExpenditureFramework (MTEF) capital expenditure estimates of the November 2005Medium Terms Budget Policy Statement (MTBPS). This shows that capitalexpenditure of the public sector, broadly defined here as the three tiers ofgovernment, development finance institutions (IDC, DBSA and LandBank)and State Owned Enterprises and Public Private Partnerships managedin the Treasury, is expected to jump by about 1% of GDP from a (recent)historical average of about 5.5%. For reasons of convenience we associatethe 1% of GDP one-off increase in infrastructure expenditure withthe ASGISA program. The proportion of capital expenditure to GDP in theMTEF is expected to remain constant in the next two years (2007/08 and2008/09).We employ a range of additional, secondary data sources to express thecapital expenditure program in terms of commodity groups used in ourmodeling framework. The results of these calculations show that the threetiers of government only invest in the asset type building and constructionworks, the largest share of the expenditure is on building construction andcivil engineering goods and services, with small shares on metal productsand business services. Public Enterprises are investing in machinery andtransport equipment as well as building and construction according to theproportions identified in the underlying Quantec SAM data. The distributionof this expenditure is across a wider range of goods and serviceswith machinery and equipment and motor vehicles featuring more thanbuilding construction. The combined and consolidated expenditure patternsare then applied to the overall benchmark value of 1% of GDP (in2003 values as this is the latest year for which the underlying SAM is estimated)to arrive at a final set of exogenous investment injections, readyfor model input.
3. Economy-wide modelingassumptionsA number of critical assumptions need to be considered in order to get arealistic view on the economy-wide feedback of the expenditure programdiscussed in the previous section.3.1 Demand side effects: representationof public investmentFirstly, in the underlying SAM database, investment by the public sectoris added to that of the private sector. While we can simulate the demandside of government investment, we cannot simulate alternative financingarrangements directly. The model focuses on the real resource flows behindthe investment, not on financial implications. The additional real resourcesfor the investment have to be matched by increased real savings.This is the standard Keynesian story: a rise in investment must be matchedby a rise in savings. The additional real savings can come from one or acombination of private, public and foreign savings because the agentsconcerned save a higher proportion of a given income or because incomesrise. The financing arrangements might influence who does the saving, notthe amount required. It is best to think of the model as giving a real storythat can be consistent with a number of different financing options.Although part of the expenditure program is accounted for by the threetiers of government and it is tempting to consider increases in direct and/ or indirect taxes or increases in the budget deficit as the most likelyfinancing options, the underlying SAM reports on the deficit on the currentaccount of the public sector and not on the public sector borrowingrequirement (PSBR). Therefore, in our modeled economy, capital expendituresby the public sector cannot (and should not) be financed by meansof higher taxation. The only other finance options available is by issuinggovernment bonds, which implies that private savings have to be raised.In reality this typically means that enterprise savings will increase and thedistribution of dividend income to households, with its typical distributioncharacteristics will decline accordingly.3.2 Supply side effects: capital expansionExogenous injections of investment expand productive capacity as wellas raise demand. By adding to the stock of capital, in particular infrastructure,the production function is shifted outwards, increasing supply.We can simulate this in a comparative static framework by exogenouslyincreasing capital stocks in the targeted sectors. In doing so, we assumethat the supply impacts work themselves out over the modeling period of2 – 3 years. Although there will be depreciation of capital stock over theperiod, we assume that there will also be ‘normal’ replacement investmentso that the ASGISA figures represent a net addition to the sectoralcapital stocks. Any crowding-in of private sector investment due to theASGISA push are ignored here due to lack of evidence. If we had informationon this we could model it as additional exogenous increases in capitalstocks. Further research into these private responses to public investmentwill be important if we wish to understand the full effects of ASGISA.3.3 Supply side effects: productivitygrowthSince it is the purpose of additional capital expenditure associated withASGISA to make the South African economy more competitive, it makessense to consider efficiency increases. Under constant returns to scale,efficiency increases in an industry’s average cost function is equivalentto increasing the productivity scale factor of the production functionsemployed in our modeling framework. Most of the capital expenditurehas an infrastructure theme and we therefore evaluate the impact of aproductivity increase in the transport services sector. Without reference toparticular studies in this regard we assume that this sector will experiencea productivity increase of 5%.3.4 Skills shortagesPolicy makers are aware of critical shortages of skilled labour to drive thisprogram. In general, South Africa’s growth prospects are seen to be heldback by skill shortages. A number of initiatives are currently underway toaddress these problems and it makes sense to evaluate the impact of addressingthe skill shortage on the likely outcome of the capital expenditureprogram. We do this by assuming one scenario with and one without fixedhigh skilled labour supply. The latter typically drives up wages in thosesectors most impacted by the demand and may well result in other sectorshaving to release skilled labour, thereby reducing growth elsewhere. Unconstrainedskilled labour would therefore typically have a greater morepositive impact from the injections such as those evaluated here.3.5 Macroeconomic adjustmentsWe assume that the government budget deficit and real government currentconsumption are both fixed. To ensure this, revenue is adjusted bya uniform percentage change to income tax rates across all households123<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>An assessment of the Accelerated & Shared Growth Initiative
South Africa’s Growth Trajectory124<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>and enterprises. The current account on the balance of payments is fixedand adjustment is effected through a flexible exchange rate. Investmentis fixed, with savings adjusting where necessary though a change in themarginal propensity to save of enterprises.The fixed budget deficit could be interpreted as assuming that ASGISAis not financed through a larger budget deficit. This is appropriate, sincecapital expenditure is not part of current spending. The model does notaddress the source of finance for the ASGISA. Assuming a fixed currentaccount deficit effectively means that ASGISA is not based on an inflowof foreign resources. The assumption of investment driven savings is inpart necessitated by the desire to impose the ASGISA investment on ourmodeled economy (if we did not do this, ASGISA would simply crowdout current investment as the level of investment adjusted to match theexogenous level of savings).Taken as a package, these three macroeconomic adjustment rules (fiscal,foreign, savings=investment) essentially place the burden of releasing theresources for ASGISA on the private sector. This does not mean that theprivate sector finances or undertakes the investment. Rather, we are referringhere to the real resource flows needed to accommodate ASGISA inthe economy. This is an important issue as ASGISA will lay claim to a largeamount of resources. These have to come from somewhere. If we thinkabout the standard national income balance, when we raise investmenteither savings or income has to rise. Our assumptions imply that it is savings,and more specifically private sector savings, that goes up.How does this tally with the view that the resources for ASGISA will comefrom growth? We deliberately do not model this because we think thatthe growth dividend will come with a lag, and that the resources haveto be released before the dividend is realised. The growth dividend providesthe rationale for ASGISA by holding out the hope that diversion ofresources away from current uses now – and most importantly currentconsumption – will lead to a better future. But it would be wrong, andpolitically would create hard-to-manage short-run expectations, to suggestthat it has no real immediate cost.We assume that there are no constraints on labour supplies of all skilltypes. In the last simulation we consider the effects of a skills shortage.However, we prefer to assume perfectly elastic supplies of labour becauseit gives an upper bound on possible employment creation. Those who donot like the assumption should interpret our results as showing what theimpact will be on labour demand rather than employment. We also assumethat capital is sector specific and operates at full capacity.4. ResultsWe present results for a sequence of four scenarios. They are discussedin turn below.Scenario 1: Supply expansion, without skill constraintWe start with a supply expansion without skill constraint. ASGISA injectscapital into three targeted sectors, “electricity, gas and water”, “transport”and “government services”. One of the main motivations for thisis to remove some bottlenecks or constraints on growth. This is a supplyside effect. Our first simulation therefore attempts to look at pure supplyeffects. It does so by exogenously increasing the capital stock in thetargeted sectors.Additional capital will increase labour demand. With the assumption ofperfectly elastic supplies of labour, employment rises. Output of targettedsectors will rise and the price of output should fall. At the macro level,GDP must rise since more resources are being drawn into use. Importdemand increases and exports may rise on the back of expanded supply.Because of fixed foreign savings, imports may have to rise further dependingon the net effect. Sectors supplying the target industries are likely tosee an increase in demand for their product, such as rubber.The capital stock in each of the target sector rises according to the shockwe have imposed.In our standard set-up this would give rise to a paradoxical result (certainlyagainst the intuition explained above). There would be a fall in thedemand for labour in the three targeted sectors as the increase in capitalstock leads to a large fall in the price of capital in those sectors (Row 3).This has to happen against the neoclassical background of the model,since capital is factor specific and fully employed. The price of labour isfixed, in accordance with the assumption of perfectly elastic supply. Thereis therefore a substitution away from labour towards capital.This goes against all beliefs that sensible people might hold. How is it possiblethat a large injection of capital into a sector could not lead to a risein employment? Those beliefs are, however, based on a view that capitaland labour are complementary. We see this often in arguments like “Eachjob in the formal sector costs R300 000 to create. Therefore if we investedR3bn we would create 10 000 jobs.” Constant capital / labour ratiosmight be plausible in the short run. In the long run however, the investmentaffects the relative prices of capital and labour. Firms will adjust to
TABLE 1: IMPACT ON TARGET SECTORS (% CHANGE)Description Model abbreviation Electricity Transport Government services1 Sector capital stock QF(CAP,A) 2.1 0.7 1.32 Sector labour demand QF(LAB,A) 0 0 03 Sector price of capital WFA(CAP,A) -19.7 -1.5 -23.94 Activity level QA(A) 1.2 0.4 0.25 Activity price PA(A) -10.3 -0.2 -15.16 Exports QE(C) 41.7 1.6 0.4Source: Model resultsTABLE 2: MACRO VARIABLES, % CHANGE IN CONSTANT PRICESBASERbnScen 1PINVCAP% changeScen 2PINV2% changeScen 3PINVPROD% changeScen 4PINVSK% change1 Total Absorption ABSORP 1,231 0.07 0.33 0.65 0.562 Private consumption PRVCON 786 0.11 -1.22 -0.71 -0.863 Fixed investment FIXINV 200 6.80 6.80 6.804 Change in inventories DSTOCK 55 Government consumption GOVCON 2396 Exports EXPORTS 340 0.05 0.31 0.60 0.537 Imports IMPORTS -319 0.06 0.33 0.63 0.568 GDP at market prices GDPMP 1,251 0.07 0.32 0.64 0.559 Exchange rate (LCU per FCU) EXRXP 1.00 0.01 0.57 0.35 0.33Source: Model results, LCU is local currency unit, FCU is foreign currency unit125<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>these new price ratios and change their decisions about capital intensity.In its standard set-up, the model would capture these long run effects 1 .Output in all three target sectors rises, but by relatively small amounts.This is not surprising: the capital injection is small relative to their initialcapital stocks, and therefore has insignificant impacts on production. Pricesfall (Row 6). This is one of the main microeconomic channels throughwhich the ASGISA investment will impact on other sectors. The fact thatthe falls are small should lead us to anticipate negligible economy-wideeffects through this channel.1However, we can simulate that ASGISA is trying to promote labour intensity, and preventshifts in factor intensity in favour of capital. It is possible to fix employment within onesector (or group of sectors), while still allowing it to vary in all other sectors. Consequently,we have to allow the wage in the fixed sectors to vary. Thus, if we add to the capital stock inthe target sectors, wages in these sectors go down. The logic is obvious, there is an injectionof capital, which cheapens capital and firms would like to substitute away from labour.But we do not let them. However, they continue to maximise profit and will employ thesame amount of labour as before only if it is cheaper. This may makes sense within thecontext of public works type programmes. Essentially there the decision about employmentis not being made on business grounds , i.e., not profit maximising. If you were doing it viaprofit maximising firms, you would have to subsidise wages to get the firms to take on theadditional workers. The fall in wages that the model generates shows how much we wouldhave to subsidise by.Column 2 of Table 2 shows the main macroeconomic effects (with thefirst column showing the base levels in 2003 Rbillion). GDP market pricesis expected to rise by 0.07%. This is largely driven by rising consumptionwhich rises by 0.11% due to the real household income effect followingthe price decreases discussed above. The reason for this increase can thusbe traced back in part to higher real household incomes. However, in addition,it can also be traced back to reduced tax rates (with GDP rising,the tax base rises, but with fixed government consumption and savings,revenue can only remain the same if tax rates come down. 2 Finally, recallthat in this simulation the demand effects of the additional investment arestill excluded, as this is part of the second scenario.Targeted sectors may impact on other sectors through either forward orbackward linkages. The results are shown in column 2 of Table 3 (withthe first column showing the base levels in 2003 Rbillion). The forwardlinkage channel operates through reduced prices of the output of targeted2We considered a flexible budget deficit and keep tax rates fixed. This may be more in linewith current fiscal practices. But it doesn’t effect the scenarios directly as the main adjustmentgoes through enterprise savings. Sensitivity analysis revealed little or no impact.An assessment of the Accelerated & Shared Growth Initiative
South Africa’s Growth TrajectoryTABLE 3: IMPACT ON REAL OUTPUT OF ALL SECTORS (% CHANGE IN QAXP LEVEL OF DOM ACTIVITY)126<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>BASE1Scen 1PINVCAP2Scen 2PINV23Scen 3PINVPROD4Scen 4PINVSK51. Agriculture, forestry & fishing 81.5 0.00 0.10 0.10 0.002. Coal mining 30.8 0.00 0.10 0.20 0.203. Gold & uranium ore mining 35.5 0.00 0.10 0.10 0.104. Other mining 86.7 0.00 0.20 0.20 0.205. Food 108.2 0.00 -0.20 -0.10 -0.206. Beverages & tobacco 37.6 0.00 -0.10 -0.10 -0.107. Textiles 19.9 0.00 0.00 0.00 -0.108. Wearing apparel 16.9 0.10 -0.50 -0.30 -0.509. Leather & leather products 4.6 0.00 0.50 0.30 0.2010. Footwear 3.7 0.00 0.10 -0.10 -0.2011. Wood & wood products 18.5 0.00 1.10 1.10 1.1012. Paper & paper products <strong>38</strong>.8 0.00 0.30 0.30 0.1013. Printing, publishing & recorded media 18 0.10 -0.20 -0.10 -0.3014. Coke & refined petroleum products 65.2 0.00 0.10 0.30 0.3015. Basic chemicals 55.7 0.00 0.50 0.50 0.4016. Other chemicals & man-made fibres 68.7 0.10 -0.20 -0.10 -0.3017. Rubber products 9.3 0.20 -0.10 0.40 0.3018. Plastic products 24.5 0.00 0.90 0.90 0.7019. Glass & glass products 4.7 0.00 0.60 0.60 0.5020. Non-metallic minerals 18.4 0.00 4.90 5.10 5.1021. Basic iron & steel 74.8 0.00 0.10 0.10 0.1022. Basic non-ferrous metals 28.9 0.00 0.10 0.00 0.0023. Metal products excluding machinery 46.1 0.10 2.60 2.50 2.4024. Machinery & equipment 40.5 0.00 0.20 0.20 0.1025. Electrical machinery 26.9 0.00 2.50 2.60 2.5026. TV, radio & communication equipment 6.6 0.00 0.30 0.30 0.1027. Professional & scientific equipment 3.8 0.00 -0.10 -0.20 -0.4028. Motor vehicles, parts & accessories 111.1 0.00 0.20 0.20 0.1029. Other transport equipment 10 0.10 0.50 0.50 0.2030. Furniture 10.9 0.10 -0.30 -0.30 -0.4031. Other industries <strong>38</strong>.6 0.00 -0.20 -0.10 -0.2032. Electricity, gas & steam 41.2 0.30 0.30 0.50 0.4033. Water supply 14 0.10 -0.10 0.20 0.1034. Building construction & civil engineering 110.7 0.00 11.10 11.20 11.2035. Wholesale & retail trade 278.9 0.10 0.20 0.40 0.3036. Catering & accommodation services 29.8 0.10 -0.20 -0.20 -0.3037. Transport & storage 162.2 0.40 0.30 3.50 3.40<strong>38</strong>. Communication 106.2 0.10 -0.10 0.10 0.0039. Finance & insurance 160.6 0.10 0.00 0.20 0.1040. Business services 249.3 0.10 -0.10 0.10 0.0041. Medical and other services 61.8 0.10 -0.90 -0.60 -0.8042. Other services 74 0.10 -0.80 -0.40 -0.7043. Government services 262.1 0.00 0.00 0.00 0.00Source: Model results
sectors leading to reduced costs on intermediate inputs in purchasing sectors.This is a price effect and will stimulate output on those sectors onlyin so far as the reduced costs allow a fall in the price of their output asthis may stimulate demand. Both “Electricity” and “Transport” are widelyused as intermediates, so there are numerous forward linkages. However,since, as we have seen, the reduction in the input prices is small, thiseffect is small. The largest increase in marketed output is 0.4%, while inmost sectors it is not much different from zero.The backward linkage channel operates because the targeted sectors increasetheir purchases of intermediate inputs from other sectors. Againthe effects are small. Without making any adjustments to the model setupthat allow for special employment arrangements in the governmentexpenditure program as discussed above, in construction as well as thetargeted sectors (electricity, transport services and government services),we proceed by presenting the results for employment in the next table.As before, results are shown for all scenarios, here we’re only interestedin the second column as it represents the first scenario, focusing on thesupply expansion only.Note that we assume that labour of all skills is plentiful available, even inthe targeted sectors. For good measure, we report in the first row on thepercentage increase in the demand for capital. It can be seen that this isfixed at 0.49% for all scenarios, i.e., an increase in the capital stock of 1%of GDP translates into a 0.5% increase.In terms of the first scenario, the increase in semi-skilled labour is 0.07%,which is roughly equal to about 3,000 full time equivalents (FTE) workers.The total increase associated with the supply increase only is expectedto be about 4,000 FTEs. However, if the expenditure would not be main-tained the next period, employment would revert back to the base levelsshown in the first column.Scenario 2: Supply plus demand expansion, without skillconstraintNext, we add the demand side stimulus to the supply side expansion ofthe first scenario described above. The demand-side stimulus is typicallythe one that is most often referred to in impact analyses that employinput-output or SAM based multiplier models. A useful benchmark is thatthe multiplier of a typical demand side injection is round about 1.2 – 1.3.However, as can be seen in Table 2, the increase in GDP rises from 0.07%(only accounting for the supply stimulus) to 0.32% (accounting for thesupply and demand stimulus) of GDP when comparing scenario 1 and 2,while the initial increase in investment demand is a full 1% of GDP. Ourchallenge here is trying to find an explanation for this somewhat disappointingresult.We described earlier that that the increase in investment demand is assumedto be financed out of enterprise savings. Not shown here is thatthe enterprise savings rate increases from 42% to about 46%. What is leftfor distribution to households by enterprises after savings are accountedfor will clearly be less. Consequently, household income and thus householdexpenditure will be lower. This is confirmed in the second entry ofTable 2 where it can be seen that household consumption declines by1.2%. Since the decline in consumption affects those households that arerecipients of unearned income (from enterprises) more than householdsthat receive income from labour, this will impact negatively on those sectorsthat supply goods and services that are typically consumed by these127<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>TABLE 4: IMPACT ON FACTORS OF PRODUCTION (% CHANGE IN QFSP LEVEL OF QUANTITY OF FACTOR SUPPLY)% increaseBASE1Scen 1PINVCAP2Scen 2PINV23Scen 3PINVPROD4Scen 4PINVSK51. Capital 0.49 0.49 0.49 0.492. Low skilled 0.07 1.63 1.83 1.773. Semi skilled 0.05 0.31 0.57 0.544. Highly skilled -0.02 0.07 0.31Absolute increase (‘000)5. Low skilled 2,566 3 57 65 636. Semi skilled 3,537 2 11 20 197. Highly skilled 3,416 0 1 4 08. Total 8,377 4 69 89 81Source: Model resultsAn assessment of the Accelerated & Shared Growth Initiative
South Africa’s Growth Trajectory128<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>households. In particular this applies to demand for goods and servicesfrom food, beverage, clothing, furniture, and some of the services, as canbe seen in Table 3.Nevertheless, the initial investment demand stimulus benefits those industriesthat produce directly and indirectly the relevant investmentgoods. The main recipients are, amongst others, construction and relatedindustries such as metal products and wood products, but also transportequipment, plastic and glass products and electrical machinery.Investment draws imports in (see row 7 of Table 2) and, given the assumedfixed current account on the balance of payment, exports need torise in order to achieve this. As a result the exchange rate has to depreciate.The model suggests (see row 9 of Table 2) that the exchange ratedepreciation is about 0.57% in the second scenario, up from 0.01% inthe first scenario.The construction bias in the initial expenditure has an impact on thedemand for low skilled labour, as can be seen in Table 4, with demandfor more skilled labour increasing to a lesser extent. Employment in constructionis reported (but not shown here) to increase by more than 20%across all skill levels but along with changes in output, employment levelsare lower than the base for a number of other sectors including, as before,food, beverage, clothing, furniture, and some of the services.The impact on household income is less clear. In Table 5 we report onhousehold income for the 10 income deciles, with the top decile brokendown into further detail, one of 5% and 4 of 1.25%. With higher enterprisesavings, unearned income distributed to households will decline andas a result consumption will be lower. Interestingly, due to the distributionof unearned income, top income earning households are more exposedto this negative impact.Scenario 3: Supply plus demand expansion plus productivityincrease, without skill constraintIntroducing productivity gains is the good news story of this paper. Rememberthat we raised multifactor productivity by 5% and that only forthe transport sector. The impact on GDP is significant , almost doublecompared to the previous scenario, i.e., without the productivity increase.Nevertheless, the accumulated impact is still only 0.6% of GDP. The negativeimpact on consumption is now also less given the higher incomes.The productivity effect works mainly through exports, as producers arenow more competitive. The exchange rate does not depreciate as muchas in the previous scenario but increased income still draws imports in.The sectors that will benefit most from the increase in productivity in thetransport sector are those that produce for the export and consumptionmarkets, not those that produce for investment demand (see Table 3). Thelatter typically remain stationary relative to the previous scenario whileTABLE 5: IMPACT ON HOUSEHOLD INCOME (% CHANGE IN YIXP)BASE1Scen 1PINVCAP2Scen 2PINV23Scen 3PINVPROD4Scen 4PINVSK51. HHD1 15.84 0.02 0.00 0.13 0.082. HHD2 21.98 0.02 -0.04 0.07 0.013. HHD3 29.31 0.03 -0.07 0.07 0.014. HHD4 <strong>38</strong>.79 0.03 -0.36 -0.18 -0.275. HHD5 50.35 0.03 -0.51 -0.28 -0.366. HHD6 71.57 0.03 -0.62 -0.37 -0.437. HHD7 103.7 0.03 -0.82 -0.53 -0.588. HHD8 160.98 0.03 -1.05 -0.72 -0.779. HHD91 127.48 0.02 -0.90 -0.59 -0.5610. HHD921 42.77 0.02 -0.96 -0.64 -0.5711. HHD922 50.01 0.00 -0.44 -0.21 -0.0512. HHD923 57.83 0.03 -1.41 -1.00 -0.9813. HHD924 114.03 0.04 -1.84 -1.36 -1.40Source: Model results
food, beverage, clothing, furniture, and some of the services typically expand.Employment gains of productivity increases are distributed differentlywith demand for semi and highly skilled labour increasing relativelymore (see Table 4), but in terms of income distribution, the gains are moreevenly shared (see Table 5), with some low income households not improvingtheir income position, while high income households still 3 sufferalbeit to a lesser extent.Scenario 4: Supply plus demand expansion plus productivityincrease under skill constraintWe end our analysis with a scenarios that builds on the previous onein which we combined supply, demand and productivity increases underunlimited supply of labour. Imposing a constraint on the supply of highlyskilled labour will change the impacts considerably, as can be seen in thelast column of Table 2. The impact on GDP reduces from 0.64% to 0.55%.At the macro level, the additional impacts of the skill constraint is distributedfairly evenly, without any particular variable bearing the brunt. At thesector level, it would appear from Table 3 that services sectors are mostlyaffected by the constraint on high skilled labour. A number of channels areat work here: not only is a sector that uses highly skilled labour held backmore in its supply of output, their prices will rise, making their goods andservices lose market share vis a vis goods and services of other sectorsand households more likely receiving less income from high skilled labour(see the last column of Table 5) typically demand goods and especiallyservices that are produced by these sectors. The impact on the demand forlabour is shown in the last column of Table 4. As expected, employment ofhighly skilled labour is fixed by assumption, but other skilled labour is alsonegatively impacted. The supply constraint of highly skilled labour costsabout 3 000 full time equivalents of other semi and low skilled workers,even though these are available for work.ConclusionsIt is important to note that the results discussed here are not forecastsof what may or may not happen to the South African economy. At best,the results give a sense of direction off a comparative static benchmark.Moreover, we have been very modest in terms of what can reasonably beassociated with the ASGISA program so as to avoid raising unwarrantedexpectations. At best we are considering the “economic infrastructure priorities”of ASGISA and ignore other priorities.In terms of our quantitative analysis, the innovation here is that, comparedto traditional impact analysis with Input-Output and SAM basedmultiplier models, we have accounted for a number of channels throughwhich an economic infrastructure investment program such as the oneassociated with ASGISA may play itself out in the South African economy.In particular we account for:1) a combination of economy-wide price, quantity and substitutioneffects while standard analysis typically only considers quantityeffects2) specific finance decisions, in particular we consider a consistenteconomy-wide accounting framework that forces us to thinkabout the inevitability that somewhere someone will have to payfor the outlays. Although we made it clear that since ASGISAis about investment expenditure and the public sector’s currentaccount is not impacted (at least not directly), the savings-investmentbalance cannot be ignored. If foreign savings are not anoption, the only remaining adjustment mechanism is domesticsavings.3) Although we use a comparative static framework, we incorporatesupply as well as demand considerations, while traditionalimpact analysis only considers the latter.4) Productivity increases are incorporated which shifts the productionfrontiers out for those sectors that are to benefit most fromthe investment. In this case we assume a hypothetical productivityincrease in the transport services sector.5) Much has been said about shortages of highly skilled labour.While we ignore this problem initially, we returned to it in ourlast scenario. It does appear to make a difference, in particularfor those activities that are know for their high skilled labourrequirements such as medical services.On balance, we find that the impact of ASGISA is therefore not likely tobe as large as policy makers would have hoped for. The reasons are bynow clear: simple multiplier analysis ignores the hard realities of nationalaccounting consistency. In the end, our modeled impact of the investmentprogram is then not more than an exercise in compositional shifts. Whatdoes get the economy going is increases in productivity. We’ve been fairlymodest by assuming productivity increases in one sector only. Whetherour assumption is on target remains to be seen. Only bottom-up sectorspecific analysis of the impact of investment programs on productivity willbe able to shed light on this, but the pay-off may well be worth it.129<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>3 Remember that our scenarios accumulate impacts.An assessment of the Accelerated & Shared Growth Initiative
South Africa’s Growth Trajectory130<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>ReferencesDavies, RJ & van Seventer, DEN, 2005: The Economy-Wide Effects Of PriceReducing Reforms In Infrastructure Services In South Africa, TIPS Forum,www.<strong>tips</strong>.org.za/events/forum2005/programme.asp.Lofgren, H, Harris, G & Robinson, S (2001) A Standard computable generalequilibrium (CGE) model in GAMS, International Food Policy Research Institute,<strong>Trade</strong> and Macroeconomic Division, Discussion Paper no 75, http://www.ifpri.org/divs/tmd/dp/tmdp75.htm.McCord, A. and Van Seventer, D.E.N. (2004): “The economy-wide impactsof the labour intensification of infrastructure expenditure in South Africa”,Centre of Social Science, University of Cape Town, CSSR Working Paper, no93, http://www.cssr.uct.ac.za.Thurlow, J & van Seventer, DEN, (2002) A Standard Computable GeneralEquilibrium Model for South Africa, <strong>Trade</strong> and Macroeconomics DiscussionPaper No. 100, International Food Policy Research Institute, Washington,D.C. and Working Paper, <strong>Trade</strong> and Industrial Policy Strategies, Johannesburg,http://www.<strong>tips</strong>.org.za/research/item.asp?ID=606.National Treasury, 2006: Medium Terms Budget Policy Statement, http://www.treasury.gov.za/
sectorstrategies131THE SA AUTOMOTIVEINDUSTRY IN AGLOBALISING WORLD<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>WHAT HAS HAPPENED TO IMPORTS?BY ANTHONY BLACKSCHOOL OF ECONOMICS, UNIVERSITY OF CAPE TOWNANDSIPHO BHANISIDEPARTMENT OF AGRICULTURE (WESTERN CAPE) ANDSCHOOL OF ECONOMICS, UNIVERSITY OF CAPE TOWN1. IntroductionSINCE THE EARLY 1990S, THE SA automotive industry has been through a rapid processof adjustment as protection has been reduced. The globalisation of the industry hasreceived much attention. 1 Much of the focus has been on the rapid increase in automotiveexports, and the Motor <strong>Industry</strong> Development Programme (MIDP) – introduced in1995 – is generally credited with having played a very significant role in promoting competitivenessand export expansion. Imports constitute the other side of the adjustment coin and have receivedmuch less attention. In assessing the impact of globalisation, the supply response of firms to therealignment of incentives towards global rather than domestic markets is a fundamental determinant.In any process of liberalisation, import expansion would be anticipated but successful adjustmentmay require new investment and growing efficiencies to at least partly offset the impact ofdeclining protection.This paper examines the impact of globalisationand automotive policy on imports of automotiveproducts since the early 1990s. A series ofimportant policy interventions, Phase VI of thelocal content programme (1989) and the Motor<strong>Industry</strong> Development Programme (1995), haveliberalised imports and encouraged exports.The industry is now structurally stronger andmore competitive. Investment and exports haveincreased but vulnerabilities remain. Importshave grown rapidly and the industry has still notachieved a sufficient volume of production torealise full economies of scale. Another constraintis South Africa’s location, remote from majorautomotive markets. After a long period of heavyprotection followed by liberalisation and exportsupport, it is now time for the sector to movetowards a balanced growth path on the basis ofpolicies which impose a more neutral incentivestructure. This would involve some ongoingprotection and assistance for production at lowto moderate levels. Under such a scenario, boththe domestic market and exports could providethe basis for sustained future growth.1See for example, Barnes and Kaplinsky (2000); Black (2001); and Barnes, Kaplinsky & Morris (2004).
Sector Strategies132<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The overall picture in the SA automotive sector is that exports have increasedvery rapidly but imports have expanded rapidly as well. Thegrowth of vehicle imports is clearly visible on our roads. Consumers havebenefited from a rapidly expanding range of makes and models, many ofwhich are imported. Less visible and more difficult to measure is whatlies beneath the bonnets of the half million vehicles being assembled annuallyin SA. Each vehicle contains some 10,000 components, which reallyconstitute the heart of the industry. A significant percentage of thesecomponents are imported.Total automotive imports increased from R18.0bn in 1996 to R73.3bnin 2005, while exports increased from R5.6bn to R45.6bn over the sameperiod. The value of imported vehicles has increased even more sharply;from only R2.8bn in 1996 to R28.3bn in 2005, accounting for nearly 40%of the domestic market. However, imports of components are considerablylarger, although the growth rate has been less dramatic because theystarted from a higher base. Imports of original equipment and aftermarketcomponents rose from R15.2bn in 1996 to R45.0bn in 2005. Automotiveproducts therefore constitute a large and growing segment of SA’s totalimports (see Figure 1). Vehicle ownership is highly income elastic and withGDP growth in excess of 4%, further rapid increases in vehicle sales arelikely. This means that automotive imports are likely to remain an importantcomponent of the current account.The objective of this paper is to analyse the import side of the globalisationequation. There are a number of factors at play here. Clearly importsare a function of factors such as the growth in the internal market and theexchange rate. Our focus is primarily on determinants which are more specificto the automotive industry. These include the position of the industryin relation to global trends in the location of automotive production, thepolicy regime (tariff reductions and the ability to rebate import duties byexporting), the strategies of vehicle manufacturers, changes in the structureof production, economies of scale and the role of foreign ownership.Section two of the paper briefly considers key global trends and theirimplications for the SA industry. Early automotive policy developmentsand their impact are discussed in section three. Section four outlines theparameters and objectives of the MIDP, while section five assesses thelink between the policy regime and the expansion of automotive imports.2. Global trends and theirimpact on the SA automotiveindustryInternationally, global integration is occurring very rapidly in the automotiveindustry, driven by lower trade barriers, including the formation andextension of regional trading blocs 2 as well as by the global strategies ofmajor international firms. Global trade in automotive products grew by anaverage annual rate of over 8% from 1980 to reach US$914bn by 2005,accounting for 9% of total merchandise trade and 12.5% of total tradein manufactures.Table 1 indicates the most important global importers of automotive productsfor the period 1980 to 2005. The bulk of world trade takes placewithin the EU and NAFTA, and the importance of the regional dimensionto global automotive trade has led some to argue that this is a moreimportant force than globalisation. 3 SA’s share of global automotive imports,while small, had been growing very rapidly.The share of developing countries in global production and exports hasalso increased very rapidly, driven by rapidly expanding markets in thesecountries but also by global automotive firms’ desire to seek out cheaperlocations. Automotive exports from the 20 major developing country exportersincreased to US$137.4bn in 2005 from US$12.3bn in 1980, representingan increase from 3.9% of global automotive exports to 15%.In this fast-changing milieu, the outcome of liberalisation is not predetermined.Rapid liberalisation in the automotive industry has had verydifferent outcomes in, say Chile and New Zealand, compared to Mexicoand Poland. In considering the prospects for the growth of the automotiveindustry in the developing world, four types of industry locations can beidentified. 4 Big emerging markets (BEMs) such as China and India haveclear advantages in that they constitute large existing markets with hugepotential. A second category includes countries which are part of regionaltrading blocs, which collectively can constitute viable ‘automotive spaces’.These countries include Brazil and Argentina (in Mercosur) and Thailandand Malaysia (in ASEAN). Neither grouping is yet fully effective and Brazil,of course, on its own constitutes a large market.The paper argues that the rapid expansion of automotive imports hasimportant implications for policy, which are discussed in the concludingsection.2NAFTA and the enlarged EU have been of particular importance. Other regional formationssuch as Mercosur (Laplane & Sarti, 2004) and ASEAN (Shimokawa, 2004) havebeen much less significant, although this could change.3See, for example, Van Tulder and Audet (2004).4We draw here on the typologies used by Sturgeon & Florida (1999) and Humphrey &Oeter (2000).
FIGURE 1: AUTOMOTIVE IMPORTS AS A SHARE OF TOTAL IMPORTS20%18%16%14%12%10%8%6%4%2%0%Source: TIPS data1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005Total auto imports / Total imports133<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Countries on the periphery of large emerging markets (PLEMAs) constitutea third category. 5 This group includes countries such as Mexico andalso new members of the EU such as the Czech Republic and Poland.These countries are attracting very large-scale inward investment in productioncapacity to supply the large adjacent market area. Mexico and thenew EU member states have the advantages of much lower labour coststhan the US and western Europe respectively, high productivity in modernplants and close proximity to major markets. The fourth category includescountries with independent strategies such as Korea.SA manifestly lacks the attributes of a big emerging market, is not partof a significant trading bloc 6 and is not on the periphery of a major market.It also does not have home-grown firms or governmental strategycapable of driving a successful, Korean-style independent strategy. Theseattributes, or rather the lack of them, have implications for how SA isperceived by the major global decision-makers and how they choose toposition SA within their global networks. As we shall see, this affects theorientation of international firms to the SA market and the types of investmentsthat have been undertaken, which in turn impacts on SA’s tradeprofile and development prospects.5This is a term used by Sturgeon and Florida (1999).6As far as the automotive industry is concerned, SADC is insignificant, although this maywell change in the long term (Black & Muradzikwa, 2004).3. Early policy developments3.1. Tariffs and the local content programmeIn promoting the development of the automotive industry, SA initiallyfollowed a programme of import substitution similar to that adopted inother developing countries, especially in Latin America. High tariffs wereplaced on built-up vehicles which, when combined with a rapidly growingmarket, acted as a magnet to a large number of (initially foreign) companieswhich established assembly plants in the country. These operations,although in many cases highly profitable, were very small in internationalterms, with correspondingly high unit costs. Production was aimed solelyat the domestic market and the SA assembly plants were kept isolatedfrom the global production networks of the parent companies, except asmarkets for completely knocked down (CKD) packs.Ford and General Motors were the first to establish a production presencein SA. They were granted protection and established assembly plants inPort Elizabeth in the 1920s. The domestic market expanded rapidly in thepost-war period, reaching 120,000 vehicles in 1960, and a large numberof assembly plants were established. The level of local content at thisstage was only 20%. The adverse impact on the balance of paymentsled to increasing government support for greater usage of domesticallyproduced components. As a result, the first in a series of local contentprogrammes was introduced in 1961. Domestic sourcing of 11 peripheralitems such as tyres, batteries and trim was required and higher localThe SA Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports?
Sector StrategiesTABLE 1: LEADING IMPORTERS OF AUTOMOTIVE PRODUCTS, 1980 - 2005Value (US$bn) Share in world imports (%) Annual % change2005 1980 1990 2000 2005 2000-05Importers134<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>European Union (25) 407.54 - - 41.9 44.1 11extra-EU (25) imports 54.91 - - 5.5 5.9 11US 205.45 20.3 24.7 28.9 22.2 4Canada d 57.61 8.7 7.7 7.9 6.2 4Mexico a, d 25.14 1.8 1.6 3.4 2.7 5Australia d 15.19 1.3 1.2 1.5 1.6 12China a 13.55 0.6 0.6 0.6 1.5 29Japan 13.17 0.5 2.3 1.7 1.4 6Russian Federation b 12.90 - - 0.4 1.4 40Turkey 11.96 ... 0.4 1.0 1.3 15Saudi Arabia b 11.64 2.7 0.9 0.6 1.3 25South Africa d 9.27 ... ... 0.4 1.0 31Switzerland 8.52 1.8 1.9 1.1 0.9 6UAE b, c 6.05 0.4 0.3 0.5 0.7 ...Norway 5.25 0.6 0.4 0.4 0.6 15Brazil 4.51 0.3 0.2 0.7 0.5 1Above 15 807.72 - - 90.7 87.5Notes: a Includes significant shipments through processing zones bIncludes Secretariat estimatesc2004 instead of 2005 dImports are valued Free On Board (f.o.b.)Source: WTO World <strong>Trade</strong> Statisticscontent levels were rewarded with additional import permits (Duncan,1991). Net local content rose rapidly, reaching approximately 52% bymass by 1971, which marked the end of Phase II of the programme. Rapidgrowth was accompanied by a proliferation of assemblers and also by thedevelopment of a low-volume components industry oriented towards theproduction of heavier components such as body pressings (due to localcontent being measured on a mass basis).Under Phase III, local content (on a mass basis) was to reach 66% by1977 in the case of ‘manufactured’ vehicles. Phase IV was a consolidationperiod with no additional requirements and Phase V, which was introducedin 1980, applied a local content requirement of 50% to light commercialvehicles, rising to 66% in 1982.The automotive industry has always been import intensive and in all thesedevelopments the main motivating factor for increasing local contentwas the desire to save foreign exchange. The import penetration ratio in1965 was 37.1%. After two decades of prohibitive protection on built-upvehicles, combined with moderate local content requirements, this hadreduced to 30% by 1985, still the highest of any industrial sector butmachinery (Kahn, 1987:248).3.2. The Phase VI ProgrammeThe problems inherent in the above approach to the promotion of localcontent were aggravated by the severe slump which followed the goldboom of the early 1980s. By late 1986, there were seven assemblersproducing over 20 basic model variants for a market of 172,000 passengercars. These low volumes meant that the industry was uncompetitive.Exports were minimal, and with the increased introduction of highlysophisticated components, it had become increasingly easy to meet massbasedlocal content requirements while increasing the value of importedcomponents.According to the Board of <strong>Trade</strong> and <strong>Industry</strong> (BTI) (1988), the local contentprogramme up to and including Phase V had two main deficiencies.It led to:“a tendency to produce low-cost, low-technology componentswhich were unremunerative to export and were produced inuneconomic volumes, so locking the industry into a lowvolume,high-cost production structure; and...
a very high import bill as source companies tended to load theprice of components they supplied to local producers. As theywere supplying largely high-technology components which thelocal industry did not produce, this too tended to raise prices,as there was no incentive to produce low-mass, high-costcomponents locally”.The new Phase VI local content programme, introduced in 1989, wastherefore aimed at promoting investment, improving productivity, minimisingprice increases and maintaining competition (BTI, 1989). It markeda significant change in direction. Local content was to be measured byvalue rather than mass. Most importantly, local content was to be measurednot just by the value of domestically produced components fitted tolocally assembled vehicles but on a net foreign exchange usage basis. Inother words, exports by an assembler counted as local content and enabledit to reduce actual local content in domestically produced vehicles.The system operated through the imposition of an excise duty of 37.5%on all locally assembled vehicles. However, this duty was rebatable to theextent of 50% of the local content value, so that if the local content target(75%) was achieved, no duty was payable. A minimum average level of50% actual local content (i.e. irrespective of exports) had to be maintainedacross the model range but local content was defined very broadlyas the ex-works price less foreign exchange used. It therefore includedprofit margins and overheads.In addition to the protection added by the local content programme, theindustry also received tariff protection on sub-sectors which fell outsidethe ambit of Phase VI, such as completely built up (CBU) vehicles, spareparts and accessories. Built-up vehicles received tariff protection of 100%and were subject to a 15% surcharge. As a result, imports were minimal.The calculation of the effective rate of protection on built-up vehicles iscomplicated by local content arrangements, but because of reduced protectionon components, it increased sharply under Phase VI.Exports under Phase VI received a substantial effective subsidy in the formof a rebate of excise duty of 50 cents in the rand. All exports were channeledvia vehicle producers and component exporters had to negotiatethe extent of the ‘subsidy’ that they received. Component producers usuallyreceived 60% to 70% of the rebate or 30 cents to 35 cents per rand(of local content value) of exports. However, there was pressure to reducethis as assemblers approached their required local content levels.The impact of Phase VIUnder Phase VI, exports grew rapidly from negligible volumes in the mid-1980s to approximately R2,245m in 1994. Many component suppliersand all the assemblers instituted significant export drives. Assemblers developedinternational marketing channels, frequently via their overseasprincipals, and identified the types of components where local producershad a competitive advantage. The position of assemblers in the auto industry’sproducer-driven value chain proved critical. As we will show withregard to the response to the MIDP, this has been a key factor in explainingthe strong supply response to the changes to the incentive regime.The assembly industry remained highly protected. As a result of prohibitivetariffs, vehicle imports remained at minimal levels, consisting mainly ofexotic vehicles which were not being assembled domestically.The main impact as far as imports were concerned was on the componentsector. Rapidly rising exports gave assemblers considerably greater flexibilityin their sourcing arrangements. By the end of Phase V, local contentin terms of mass (which was the measurement used) had reached 66%,but was lower in value terms. Under Phase VI, actual local content neededonly be a minimum of 50% averaged across the model range, as longas the rest of the total 75% ‘local content’ requirement was made upthrough exports. It is clear, therefore, that with the growth in exports,vehicle producers were able to reduce the local content in domesticallyassembled vehicles significantly.Imported components were mainly brought into the country in the formof CKD packs. If a vehicle producer opted to use a local component, theforeign supplier would remove this from the pack and subtract a ‘deletionallowance’ from the cost of the pack. Deletion allowances werewidely held to be below competitive international prices. The result wasthat a domestic component manufacturer was competing not with ‘competitive’international prices, but with a lower deletion allowance. Whilethe component industry was previously excessively protected comparedto the other intermediate goods sectors, under Phase VI this situation hadchanged dramatically.THE SHORT-TERM IMPACT OF PHASE VI ON THE DOMESTIC COMPONENTINDUSTRY WAS FELT IN THREE MAIN AREAS:1. The switch from mass to value had a highly differentiated effect on the componentsector. Vehicle producers began looking at ways of increasing localcontent by value rather than mass. Heavy components such as body pressingswere no longer required and came under increasing pressure. Because of hightooling costs and short production runs, this was one of the most vulnerable135<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The SA Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports?
Sector Strategies136<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>sectors, especially as it had enjoyed exceptionally high protection in terms ofthe mass-based scheme.2. Components which formed part of sub-assemblies also came under threatbecause it became cheaper to import these in a semi-assembled form thussimplifying assembly and limiting the problems of local re-engineering, qualityand supply complexities.3. Components with high tooling costs in relation to the cost of the componentwere also vulnerable (for example, plastic moulded components). Again, lowvolumeproduction for the domestic market made these uneconomic.However, for models introduced under Phase V, manufacturers tended tomaintain their sourcing arrangements due to sunk investment in toolingand contractual obligations. Also, it took time to build up large exportvolumes. Thus the increased flexibility to source additional componentsabroad was most apparent with new model introductions and started tohave a major impact from 1992. Estimating the impact of Phase VI is complicatedby the difficulties of disaggregating the impact of recession fromreduced domestic sourcing. Also, reduced domestic sourcing was to someextent compensated for by exports, although much of this was in ‘non-traditional’components such as catalytic converters. There was, however, asignificant fall in employment in the component industry in the first threeyears following the introduction of Phase VI.As explained above, the growth in exports greatly increased the flexibilityof component sourcing, allowing assemblers to take advantage ofcheaper foreign components. This led to a substantial reduction in costs,especially as new models were introduced. Component suppliers, whowere used to prices being determined on a ‘cost plus’ basis, were forcedto become more efficient and reduce their margins as they faced ultimatumsto reduce prices in real terms or lose contracts.A major defect of Phase VI is that it did not address the major factorimpacting on the scale of production in the components sector – proliferationof makes and models in the domestic market. Increased flexibility incomponent sourcing increased the effective rate of protection on built-upvehicles, and the predictable result was an increase in the variety of modelsand makes being assembled locally.The series of local content programmes introduced in SA was seriouslyflawed. They were directly responsible for the development of a fragmentedand non-competitive industry. Phase VI was an attempt to address thissituation. It encouraged exports but at the same time drastically reducedprotection of the components sector while doing nothing to reduce proliferationof models being assembled domestically, which was one of themajor reasons for the component sector being uncompetitive. A furtherproblem was that Phase VI was introduced at a time of great politicaland economic uncertainty and a generalised lack of investor confidence.This provided an inappropriate environment for a programme of structuraladjustment, the success of which is contingent on a positive supplyresponse.4. The MIDP4.1. The introduction of the MIDPPhase VI came in for heavy criticism, with frequent changes adding tothe atmosphere of uncertainty. In particular, there was pressure from thecomponent producer federation, NAACAM, and in late 1992 the Motor<strong>Industry</strong> Task Group (MITG) was appointed to re-examine the programmeand advise government on a future development policy for the industry.The MITG was a tripartite forum representing industry, trade unions andgovernment. Government made it clear that tariffs had to be reduced inline with its obligations to the World <strong>Trade</strong> Organisation (WTO).The eventual outcome was the MIDP, which was introduced in 1995. Therecommendations of the MITG were only partly accepted. Most notably,the contentious proposal to encourage higher model volumes and force adegree of rationalisation was not accepted, as a result of strong oppositionfrom the vehicle producers federation, NAAMSA. The MIDP continuedthe direction taken by Phase VI and entrenched the principle of importexportcomplementation. However, it went a step further by abolishing localcontent requirements and introducing a tariff phase-down at a steeperrate than required by the terms of SA’s offer to the WTO.THE MAIN ELEMENTS OF THE MIDP WERE THE FOLLOWING:a) The excise duty-based local content system was changed to a tariff-driven programme.b) There was no minimum local content requirement.c) Tariffs were to be phased down to 40% for light vehicles and 30% for componentsby 2002.d) Manufacturers of light vehicles were entitled to a duty-free allowance in termsof which components to the value of 27% of the wholesale price of the vehiclecould be imported duty free.e) Import duties on components and vehicles could be offset by import rebatecredits derived from the export of vehicles and components.f) Provision was made for a Small Vehicle Incentive (SVI) in the form of a higherduty-free allowance for low-cost vehicles.While nominal duties on imported vehicles were set to remain high, evenuntil 2002, the ability to rebate import duties by exporting enabled importersto bring in vehicles at lower effective rates of duty. Import-exportcomplementation also enabled assemblers to use import credits to source
components at close to international prices. Thus declining nominal protectionon vehicles was to some extent being offset by reduced protectionfor components.4.2. Mid-term review and 2003 reviewIn response to the need to assess the impact of the MIDP as well as toprovide long-term policy certainty to the industry, the Department of <strong>Trade</strong>and <strong>Industry</strong> (the dti) conducted a Mid-Term Review in 1998, the resultsof which were published in 1999. 7 A further Review was conducted in2002/2003. Both policy reviews extended the MIDP with minor adjustmentsand with a continued gradual decline in assistance to the industry(see Table 2).In the Mid-Term Review, the MIDP was extended to 2007. One significantadjustment was that while import-export complementation provisionswere extended, this was on a phasing down basis. The qualifying valueof eligible export performance was scheduled to decline from 2003 (seeTable 2). This meant that while exports of components (or vehicles) witha local content value of R100 would allow the exporter to import R100of components (or vehicles) on a duty-free basis in 2002, from 2003 agradually declining value of components could be imported duty free.Again, extensive discussions regarding the imposition of direct industrialpolicy measures to rationalise the industry took place, but these were notadopted. An important late change introduced into this process as a resultof concerted pressure on government by vehicle manufacturers who wereplanning major export programmes, was the introduction of a ProductiveAsset Allowance (PAA). In terms of the PAA, firms making qualifying investmentsreceive import duty credits equal to 20% of the value of theseinvestments, spread over five years.A third review was held in 2002 to provide clarity on policy until 2012.The brief was to maintain the basic architecture of the MIDP. Tariffs wereset to decline to 25% and 20% for built-up vehicles and componentsrespectively and there were other minor adjustments. 84.3. The objectives of the MIDPThe orthodox rationale for tariff reductions is to realign relative prices,reduce input costs and correct anti-export bias. The expected result wouldbe a shrinking of the sector concerned, with benefits being felt in othersectors of the economy.The initial objectives of the MIDP were to provide high-quality, affordablevehicles, provide sustainable employment and through increased productioncontribute to economic growth (the dti, 1997). These, of course, aregeneric objectives, which are important to all sectors. More specifically,the MIDP was devised as a trade facilitating measure with very particularindustry policy objectives. As a result of protection, the industry structurehad historically been very fragmented and the resultant failure to achieveeconomies of scale not only made the assembly industry inefficient, butimposed major negative externalities on the component sector. With theproliferation of makes and models being produced in low volumes in SA,component firms were in turn required to produce at way below minimumefficient scale. An objective of the MIDP was therefore to increase thevolume and scale of production though a greater level of specialisation interms of both vehicle models and components.Essentially what was required was a transition from CKD assembly, whichtypically has been characteristic of vehicle production in protected developingcountry markets through a transition stage to full manufacturing(see Table 3). CKD assembly involves relatively light investments, in spiteof the fact that the need for precision welding and advanced paintingprocesses in modern CKD plants increasingly require larger capital outlays(Sturgeon & Florida, 1999). Under CKD assembly, production costs areusually quite high, especially if a high level of localisation is stipulated bygovernment policy.High local content requirements would necessarily require much higherlevels of investment and would tend to encourage rationalisation. In theCKD assembly stage, quality is likely to be below international standardsand assemblers would be likely to introduce their own adaptations, usuallywith the purpose of extending model life. As a result, in many protected,emerging economy markets, models have continued in productionlong after they have been phased out in advanced country markets. In SA,the VW Citigolf and Toyota Tazz are examples of this.In the transition and full manufacturing stages where exports may becomesubstantial, both quality standards and the number of derivativesoffered need to be in line with international practice. <strong>Vol</strong>umes per modelalso increase in the transition stage, and under full manufacturing wouldapproach world scale. Because firms are exporting, they would need accessto components at world prices, so in spite of higher volumes in thetransition stage, local content levels may not increase. In the full manufacturingstage, much higher volumes would normally be attained, allowingvehicle makers to localise components on an economic basis.137<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>7See Republic of SA (1999) for further detail on the proposals.8For the policy recommendations, see Barnes & Black (2003).The SA Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports?
Sector StrategiesTABLE 2: THE MIDP AS AMENDED IN THE MID TERM-REVIEW AND THE 2003 REVIEWYear Import duty Value of exportperformanceQualifying PGM*contentRatio of exports against imports1<strong>38</strong><strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Built-up light vehiclesOriginal equipmentcomponentsBuilt-up vehicles andcomponents (excludingtooling)Catalytic convertersexportedComponents,heavy-duty vehicles& tooling exported:CBU light vehiclesimportedComponents, vehiclesand tooling exported:components,heavy vehicles andtooling imported1999 50.5% 37.5% 100% 90% 100:75 100:1002000 47% 35% 100% 80% 100:70 100:1002001 43.5% 32.5% 100% 60% 100:70 100:1002002 40% 30% 100% 50% 100:65 100:1002003 <strong>38</strong>% 20% 94% 40% 100:60 100:1002004 36% 28% 90% 40% 100:60 100:1002005 34% 27% 86% 40% 100:60 100:1002006 30% 26% 82% 40% 100:60 100:1002007 30% 25% 78% 40% 100:60 100:1002008 29% 24% 74% 40% 100:60 100:1002009 28% 23% 70% 40% 100:60 100:1002010 27% 22% 70% 40% 100:60 100:1002011 26% 21% 70% 40% 100:60 100:1002012 25% 20% 70% 40% 100:60 100:100Notes: The Duty Free Allowance of 27% remained unchanged during this period. The Productive Asset Allowance (PAA) was put in place until 2007, to be reviewed later.* PGM: Platinum Group MetalBuilt-up light vehiclesexported: builtuplight vehiclesimportedSource: Adapted from Black & Barnes (2003) and NAAMSA (2005)The main instruments of the MIDP have been falling nominal duties combinedwith export assistance derived from the ability to offset import duties.While nominal duties on imported vehicles have remained fairly high,at least in the early stages of the MIDP, the ability to rebate import dutiesby exporting has enabled importers to bring in vehicles at lower effectiverates of duty. Import-export complementation has also enabled assemblersto use import credits to source components at close to internationalprices, so declining nominal protection on vehicles has to some extentbeen offset by reduced protection for components. This has meant thatthere has still been a significant incentive to assemble locally.The MIDP seeks to provide support for the automotive industry on a graduallydeclining basis. This requires that it meet a number of objectives,including: Some protection of assembly; Some protection for the component sector; and Some export support.The support mechanism is principally via rebates on import duties whichcan be gained by exporting. 9 The various components of the system aretherefore interdependent. If credits can be generated too ‘easily’, importprotection is effectively removed, while if it is too difficult to earn thesecredits, the industry becomes more protected. The latter outcome wouldlead to rising car prices and higher vehicle production costs. So the volumeof credits being generated is a key policy issue, as it affects the ‘balance’of the programme.5. The impact of the MIDPInternational competition in the SA automotive industry has increasedsubstantially as a result of the MIDP. Vehicle manufacturers faced theprospect of the domestic market being eroded by imports, as tariffs werereduced from prohibitive levels. The component sector, which had only just9Since 2003, firms undertaking ‘qualifying investments’ could earn duty rebates under theProductive Asset Allowance, but the level of this is relatively low.
egun the transition from low-volume, flexible production faced furtherrestructuring and consolidation.The impact of the changes was awaited with a degree of trepidation bypolicy-makers, the industry and unions. Clearly, the outcome of this shifttowards more open markets depended not only on the level of importpenetration, but also on the supply response in terms of investment andexport expansion. A survey of component firms undertaken in 1995 justbefore the introduction of the MIDP showed that firms were well awareof the changes that would have to be made in response to the new programme(Black, 1995).In spite of the fact that two-thirds of firms expected competition in theirproduct line to increase dramatically as opposed to one-third who expectedthe increase in competition to be slight or negligible, firms were generallyadopting a positive approach. Firms planned to upgrade productivityby improving production efficiencies, expanding exports and increasing investment.There was much greater emphasis on a positive supply response(expanding exports and new investment) than on reducing employment,curtailing operations or sourcing sub-components internationally.The restructuring process, therefore, looked likely to be centered aroundefforts to improve in-house productivity, including work organisation, byattempts to expand production volumes in a more focused range of productsthrough exporting and to upgrade plant and equipment, includingincreased use of automation. It should be recognised, however, that theindustry was experiencing boom conditions at the time of the survey, withsales growing rapidly and this is likely to have influenced expectationsabout the outlook for investment and employment.The respondents also proved to be remarkably accurate in forecastingthe impact of the MIDP on exports, investment and employment. Firmsanticipated a massive increase in exports, moderate increases in investmentand roughly stable employment, with two clear categories of firmsemerging - those who were linked into export markets and expected toincrease employment and another group of more traditional componentsuppliers who expected employment to decline.5.1. Imports, the trade balance andautomotive policySince the introduction of the MIDP, automotive imports have grown rapidlyin absolute terms. Even more striking has been the growth of automotiveimports as a share of manufactured imports, which increased froma low point of 12% in 1998 to 23% in 2005 (see Figure 2).As a result of growing component and vehicle imports, the overall automotivetrade deficit widened dramatically from under R5.1bn in 1992 (ayear of weak demand) to R14.1bn in 1996 before declining as a resultof falling domestic vehicle production (requiring fewer imported components)and growing exports. In 1999, it had declined to only R8.0bn asimports increased moderately while exports continued to grow rapidly.139<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>TABLE 3: STAGES IN THE DEVELOPMENT OF VEHICLE PRODUCTION IN SACKD assembly Transition Full manufacturingTarget market Domestic Domestic and export Domestic and exportLevel of integration with parentcompanyLow; import of CKD packs Medium HighModel line up Many models One or two One or twoDerivatives Limited to reduce costs Full range to supply export market Full range to supply export marketLocal contentGenerally low but may be quite high as aresult of local content requirementModerate, based primarily on cost factorsMedium to highQuality Below source plant Equal to source plant Equal to source plantProduction costHighDomestic design Local adaptations NoneSource: InterviewsMedium; penalties incurred by highlogistics costsLowNone - may do worldwide R&D inniche areasThe SA Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports?
Sector Strategies140<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The rand value of imports has since grown rapidly, initially as a result ofthe weaker rand in 2001/2002. Since then the rand has strengthened, butbooming domestic demand for imported vehicles as well as componentsto supply the increase in domestic assembly has led to a further rapidincrease in automotive imports and a negative automotive trade balanceof R18.8bn in 2004 and R27.7bn in 2005.A number of factors have a bearing on imports. These factors include marketdemand, the exchange rate and the policy provisions of the MIDP. Ourfocus is on the impact of automotive policy, including tariffs, the Duty FreeAllowance (DFA), import-export complementation and the PAA.As tariffs are reduced, imports can be expected to gain a larger share ofthe domestic market and rapid import expansion can threaten the viabilityof local producers by eroding their domestic market share. Nominal tariffsare declining gradually and do not on their own explain the rapid increasein automotive imports.A key strategy of the car-makers operating in SA is to expand marketshare. They seek to achieve this via a combination of local productionand vehicle imports. Importing vehicles and components incurs importduties and much of the strategic behaviour of firms is therefore directedat optimising their duty position.Minimising duty payments can be achieved in a number of ways. First,firms can limit vehicle imports. Secondly, local content in domesticallyproduced vehicles can be adjusted. Thirdly, vehicle producers can expandexports of either vehicles or components. As exports have increased, sohas the ability to import automotive products without paying duty. Independentimporters, such as Renault and Hyundai, which do not producevehicles in SA are also trying to expand market share and are able tooffset import duties by facilitating exports into their global networks. Inaddition, car-makers undertaking specified investments which qualify underthe PAA also receive import credits. These are currently at a low levelcompared to the credits earned via exporting.These considerations have had a decisive effect on the strategic choicesmade by vehicle producers. The structure of the MIDP has been such thatit clearly has been easier to generate exports than develop high local contentin domestically assembled vehicles. In the early stages of the MIDP,the strategy adopted by vehicle manufacturers was to develop componentexports. As Figure 4 indicates, component exports increased rapidly until2002 but a large share was taken by just two types of components – automotiveleather and catalytic converters. These types of ‘peripheral’ componentsoffered the opportunity of generating large export volumes withlimited investment. Global demand for catalytic converters was expandingrapidly due to environmental legislation. Also, due to their platinumcontent, catalytic converters are high-value products. Automotive leatheris a labour-intensive and footloose industry. This kind of strategy contributedlittle to the overall development of the industry, as it co-existed withlow-volume CKD assembly and did not contribute to reducing the cost ofdomestic vehicle production by reducing the cost of components beingsupplied to the domestic industry or by bringing down assembly costs.From the late 1990s, vehicle exports expanded rapidly. This requiredmuch greater investment by vehicle manufacturers and assisted in raisingvolumes – which at least helped component suppliers to become competitive.Exports, initially of components but later including vehicles, weretherefore the main strategic choice adopted to minimise duty payments inthe face of increasing imports.The key to understanding the impact of exports on the ability to importis the value of Import Rebate Credit Certificates (IRCCs) (see Figure 5).Exports grew rapidly until 2002. The certificate value has, however, grownless rapidly over the period because of lower local content in total exportsas a result of the growth in the relative importance of vehicle exports,which have much lower local content than components. Additional factorsare the phasing down in the qualifying percentage of platinum incatalytic converter exports from 1999 and the phased reduction in thequalifying percentage of all exports from 2003. In spite of the phasingdown of export assistance, the incentive structure continues to encourageexport growth.5.2. Vehicle importsUntil the early 1990s, prohibitive tariff levels resulted in negligible importsof vehicles into SA. Vehicle prices were significantly above internationallevels. The opening up of the economy and the phasing down of tariffshave led to an increased level of light vehicle imports from under 2%of the market in 1990 to 13.9% 10 in 1997 and nearly 40% in 2005. Untilthe surge in imports during 2004 to 2005, increases were roughly in linewith the expectations of policy-makers, and as indicated in Figure 6, thenumbers of vehicles exported in some years exceeded imports.Domestic vehicle producers, especially those firms which have already establishedlarge-scale vehicle export projects, account for the major shareof vehicle imports. In interviews conducted in 2002, all assemblers wereplanning to expand imports of models which they did not produce locally.10This includes imports of semi-knocked down vehicles imported under a temporary concession.
FIGURE 2: AUTOMOTIVE IMPORTS AS A SHARE OF MANUFACTURED IMPORTS, 1996 - 200525%20%15%14110%5%0%Source: TIPS data1996 1997 1998 1999 2000 2001 2002 2003 2004 2005Total auto imports / Manufacturing importsPassenger vehicle imports / manufacturing imports<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>FIGURE 3: AUTOMOTIVE TRADE BALANCE, 1996 - 2005 (RBN, CONSTANT 2000 PRICES)706050403020100-101996 1997 1998 1999 2000 2001 2002 2003 2004 2005-20-30ImportsExports<strong>Trade</strong> balanceSource: Compiled from the dti dataThis was clearly related to plans to rationalise production in the domesticmarket to a reduced number of platforms, raise production per model andexport part of the expanded output. This strategy generally required anexport allocation by the parent company, which in turn was seeking toexpand market share (including the sale of imported models) in SA. Independentimporters have also increased their market share significantlysince 1995.The rise in vehicle imports has been due to four main factors. First, decliningtariffs have played a role (see Figure 7). Import duties have declinedsteadily from the 65% level at the start of the MIDP in 1995. Nevertheless,the normal import duty on light vehicles, at 34%, remained quitehigh in 2005. The second important factor has been the ability to rebateimport duties by exporting. Figure 8 shows the close relationship at thefirm level between vehicle exports and imports. In 1996, no firm exportedsignificant numbers of vehicles and imports were minimal. By 2001, threeThe SA Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports?
Sector Strategies142firms, all German-based, had implemented vehicle export strategies andwere generating nearly 50% of their IRCCs from vehicle exports, whichenabled them to make a greater contribution to the overall group by raisingmarket share in SA. During this period, other car-makers such as Fordand Nissan continued to pursue multi-model strategies (in some caseswith low local content levels). They were able to do this by generatinglarge-scale exports of components, which allowed them to offset componentimports. This option was not sustainable in the medium to longterm, and by 2005, both Toyota and Ford had started to implement exportprogrammes, with the other vehicle producers announcing plans as well.High-volume vehicle exports have allowed vehicle manufacturers to importsignificant volumes of both vehicles and components duty free.<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>FIGURE 4: EXPORTS OF CATALYTIC CONVERTERS, AUTOMOTIVE LEATHER AND OTHER COMPONENTS, 1995 - 2004R-million25,00020,00015,00010,0005,00001995 1996 1997 1998 1999 2000 2001 2002 2003 2004Other componentsStitched leather componentsCatalytic convertersSource: Compiled from the dti dataFIGURE 5: VALUE OF IRCCS GENERATED BY EXPORTING45,00040,00035,00030,000R-million25,00020,00015,00010,0005,00001997 1998 1999 2000 2001 2002 2003 2004 2005Source: Compiled from the dti dataExports Eligible exports (local content) Certificate value
FIGURE 6: NUMBERS OF VEHICLES EXPORTED AND IMPORTED, 1996 - 2005250,000200,000150,000143100,00050,00001996 1997 1998 1999 2000 2001 2002 2003 2004 2005ExportsImports<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: NAAMSA, the dtiBy 2001, vehicle exports accounted for over 30% of the IRCCs beinggenerated, up from only 10% in 1996, and this percentage has continuedto increase. Independent importers have also been able to offset dutieson vehicle imports by facilitating component exports into their global networksbut have less capacity to do this compared to the vehicle producerswith established facilities in SA.A third factor which has had an impact since 2003 has been the recoveryof the rand from very depressed levels. While this reduced costs for localassemblers by lowering the costs of imported components, the impact onimported vehicles was correspondingly greater.Fourthly, the boom in the domestic market, spurred by the stronger currency,has contributed to growing imports since 2004. To some extent thishas been due to capacity constraints among local assemblers, which hasencouraged more aggressive marketing of imported vehicles.5.3. Components and local contentA key policy issue in the development of the automotive sector, both inSA and other developing countries, has been the level of local contentin domestically assembled vehicles. All too frequently vehicle assemblyhas been characterised by low volumes, which has led to low levels ofdomestic content.Figure 9 shows that, in real terms, the value of components imported pervehicle assembled domestically has increased significantly. Measuring thevolume and value of vehicle and component imports is simple enough,but the level of local content in the SA industry remains a major issue ofdebate between government and the various industry federations.Local content is notoriously difficult to measure. As Table 4 indicates, itcan be defined in a number of ways and there are significant differencesbetween the various measures. Each definition is also subject to measurementdifficulties and to the vagaries of the exchange rate.For example, the ‘official’ definition of local content as vehicle wholesaleprice (value of production) less import content (Measure D) leaves muchroom for changes in vehicle prices, assembly costs and profit margins.With no change in the actual sourcing of components, higher prices andprofits would mean a ‘higher’ level of local content. In fact, this is whathappened recently in SA. According to Measure D, local content hasranged between 50% and 60% and shows a slight upward trend (seeTable 4). But this reflects a stronger rand from 2003, as well as risingassembly industry profitability. The latter dimension is reflected in the assemblycontribution and hence higher ‘local content’. The data based onother definitions show a stable or declining position, which may in factindicate that ‘actual local content’, based on the number of domesticallyproduced parts incorporated in locally assembled vehicles, is declining.The SA Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports?
Sector Strategies144<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>What is also evident is the low level of local content (between 32% and26%) in total component purchases (Measure B). If one considers localcontent in component purchases as a percentage of total vehicle revenue(Measure C), this dropped to as low as 16% in 2004, a year in which boththe rand and assembly industry profitability were strong.Explaining changes in local contentThe introduction of a series of local content programmes from 1961 led toincreased levels of local content in SA assembled vehicles. With the introductionof Phase VI, the component sector came under increasing pressurefrom imports. This continued under the MIDP, with annual price increasessignificantly below inflation levels, especially in the early stages.Under the MIDP, protection of the component sector has been reduced.Local content requirements were abolished and duties have continued todecline, albeit gradually. Conventional trade theory would predict a declinein local content. Apart from declining protection, there are, however,a number of further considerations. One of the policy objectives has beento increase model volumes, which could be expected to impact on localcontent levels. Changes in local content usually take place when newmodels are introduced and it would therefore be important to be ableto measure whether new models have significantly lower levels of localcontent than the vehicles they are replacing. There has also been significantforeign investment by first-tier suppliers and a further question iswhether these firms are mainly engaged in assembly of imported parts ordraw on the domestic supply base of second-tier firms.It is clear that there has been some decline in local content but this hasnot been dramatic. Of major concern in the first few years following theintroduction of the MIDP was the tendency to introduce new modelswith low local content levels. As Table 5 shows, the local content level insome new models introduced was very low, especially as the measurementof local content in this table includes assembly and profit marginsapart from actual local content. 11 On this measure, a local content level ofunder 40% is very low in terms of actual local components and translatesinto less than 30% of components fitted being locally sourced. Thesewould comprise mainly peripheral components such as wheels, exhausts,certain trim components, body panels, batteries and glass.Given the fundamental objective of the MIDP, government has beenanxious to see higher levels of localisation. In the course of interviewsconducted in 2002, assemblers voiced their concerns regarding the do-11Some caution should be exercised in the interpretation of the data. The data on new modellocal content levels are not weighted according to volume. Lower local content modelswould tend to be lower volume vehicles.FIGURE 7: TARIFFS AND THE SHARE OF THE LIGHT VEHICLE MARKET TAKEN BY IMPORTS90807060Percentage504030201001994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008Vehicle tariffImports as % of domestic salesSource: Compiled from the dti data
FIGURE 8: LIGHT VEHICLE IMPORTS AND EXPORTS BY FIRM, 1996 - 2005199640,00035,00030,000Units25,00020,00014515,00010,0005,0000Ford Delta/GM Fiat/Nissan Toyota DCSA BMW VWImports Exports2001<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>40,00035,00030,000Units25,00020,00015,00010,0005,0000Ford Delta/GM Fiat/Nissan Toyota DCSA BMW VWImportsExports200540,00035,00030,000Units25,00020,00015,00010,0005,0000Ford Delta/GM Fiat/Nissan Toyota DCSA BMW VWImportsExportsSource: Compiled from NAAMSA dataThe SA Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports?
Sector StrategiesFIGURE 9: MOTOR VEHICLE PRODUCTION AND ORIGINAL EQUIPMENT COMPONENT IMPORTS (CONSTANT 2000 PRICES)30,000600146<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Component Imports (R-million)25,00020,00015,00010,0005,00001996 1997 1998 1999 2000 2001 2002 2003 2004 2005Total domestic productionOEC imports5004003002001000Motor Vehicle Production (‘000 units)Source: Compiled from NAAMSA datamestic components industry. They argued that in many cases it did nothave the required technology to supply components for the advanced vehiclesbeing exported. A particular problem related to the fact that insufficientinvestment had occurred to upgrade the technology in this sector. Todeal with this constraint would require large-scale investment, including asubstantial expansion in foreign investment.Achieving higher levels of local content is not easy. Given the very largeinvestments involved, even model volumes of 40,000 to 50,000 units perannum do not justify the investments required to raise local content muchabove a level of about 60% 12 One solution would be for component firmswhich supply domestic car-makers to achieve minimum efficient scale byexporting, say, half of their output. The presence in SA of the three majorGerman car firms, all with significant vehicle export activities, should beattractive to German component firms. Indeed, the German-based carmakershave co-operated to attract investments by first-tier suppliers. Tosome extent they have been successful, but real constraints remain. As anintegral part of the global production capacity of the parent companies,SA-based assemblers would normally be expected to use exactly the samesupplier as the parent company, a practice known as lead sourcing. Thesesuppliers may be different for VW, BMW and DaimlerChrysler. Increas-ing output by supplementing production in order to reduce unit costs isconstrained by the fact that this might make inroads into the establishedcapacity of existing foreign suppliers.Part of the trend indicated above reflects changes in the level of local contentof components themselves. While there has been investment, mainlyby foreign firms (or joint ventures) in high technology, first-tier componentsto supply large-volume vehicle projects, in many instances thesefirms operate as just-in-time sub-assemblers of imported components usingtechnologically advanced assembly jigs and testing equipment. Theyare not, however, responsible for any materials conversion processes andas such cannot be considered true manufacturers. The advanced materialsconversion (and the associated tooling and technology investment) takesplace outside of SA and local content and local value adding, even onlarge-scale vehicle export projects, has remained low (Barnes & Black,2003). The latter characteristic was clearly borne out in the course of the2002 interviews with vehicle producers (Barnes & Black, 2003). It is furthersupported by recent data drawn from the 70 firms which belong tothe SA Automotive Benchmarking Club, which shows striking differencesin the purchasing patterns of local and foreign owned component firms(see Figure 10).12As explained above, the definition of local content is problematic. Here we are using the‘official’ definition of wholesale value less forex (Measure D in Table 4). This thereforeincludes local assembly. Under narrower definitions (excluding assembly) the local contentvalue would clearly be significantly lower.The reliance on foreign inputs partly reflects the assembly or ‘system integrator’character of many foreign-owned supplier operations, and is in
TABLE 4: THE PERCENTAGE LEVEL OF LOCAL CONTENT (LC) ACCORDING TO VARIOUS DEFINITIONSLocal content measure 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005A. Component purchases by OEMs 39 <strong>38</strong> 40 43 34 35 <strong>38</strong> 40 na naB. Component purchases less all imports na na na na 32 27 27 26 27 31C. Component purchases less all imports as % of revenue na na na na 21 19 20 16 16 20D. Average value of production less average forex 56 58 56 52 51 52 50 56 60 59E. LC in exported passenger cars na 50 49 na 42 <strong>38</strong> 39 35 36 37F. LC in exported light commercial vehicles na 51 50 na na na na 39 36 35G. LC in exported medium commercial vehicles na 45 50 na na na na 42 39 39Notes: A. Local component purchases by vehicle producers based on the dti survey data B. Local component purchases less all import content as % of all component purchases.C. Local component purchases less all import content as % of vehicle revenue. D. Average value of production less average forex (from OEM aggregate data).E, F, G. F.o.b. export value less forex.Sources: IRCC data, MIDP Customs Account and survey data supplied by the dti, NAAMSA and NAACAM.TABLE 5: LOCAL CONTENT LEVEL OF NEW MODELS INTRODUCED FROM SEPTEMBER 1995 - 1998147<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>New model Local content (%)A 61B 60C 58D 51E 41F 39G 37Average (new models - unweighted) 49.6Weighted average for the industry 57.5Note: Local content measured by ex-works price less foreign content.Source: the dti, unpublished data.part a global trend. 13 This lack of embeddedness may partly result fromthe limited time that these firms have been operating in SA. But it alsoreflects the fact that vehicles are being produced in volumes of 50,000units per annum or less, which does not justify heavy investment in componentproduction.This is further demonstrated by the fact that component exporters tendto have higher levels of local content than domestic market suppliers.Exports are mainly of less complex components and tend to be in highvolumes, with concomitant high levels of local content (see Table 6).An interesting development noted by certain vehicle producers and anumber of component producers, was that some suppliers had becomereluctant to supply assemblers, even where volumes were fairly large.This was seen to be because of demanding price and quality certificationrequirements on the one hand and the fact that some component suppliershad access to more lucrative opportunities in international markets,especially the less technologically demanding and price sensitive aftermarket.1413See, for instance, Humphrey & Salerno (2000).14Interviews.The SA Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports?
Sector StrategiesFIGURE 10: LOCAL CONTENT IN COMPONENTS PRODUCED BY VARIOUS CATEGORIES OF COMPONENT FIRMS1001489080<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Local content % weighted706050403020100MNCs SA firms Non-exporters ExportersSource: Derived from data supplied by B&M AnalystsTABLE 6: LOCAL CONTENT IN MAJOR COMPONENT EXPORTS (%)Component type 2003 2004 2005Catalytic converters (including PGM reduction) 81 82 naSeats and components (including leather covers) 77 78 79Engines and engine parts 82 65 72Tyres and tubes 67 65 68Wheels, brakes and parts 89 89 87Body parts 79 72 76Electrical equipment 62 57 62Other parts 72 72 80Silencers & exhaust systems 84 84 79Glass 75 78 89Suspension parts 84 79 70Radiators 82 86 85Tooling 84 78 71Gearboxes, clutches & parts 73 74 75Axles: drive & non-drive 69 74 71Source: International <strong>Trade</strong> Administration Commission (ITAC)
The case of Toyota’s suppliersThe implications of the changing nature of the supplier system for localcontent and domestically owned firms are illustrated by the case of ToyotaSouth Africa (TSA). 15 Having previously been a licensed operation,the company is now a subsidiary of Toyota Motor Corporation (TMC)and has been converted into a global production hub. It is the onlySA assembler which has attained world-scale volumes, and productionis set to increase to 220,000 units by 2008. This includes two highvolumemodels, the Corolla and International Multipurpose Vehicle(IMV). Furthermore, according to a recent poll of suppliers, Toyota isthe firm which is regarded as most supportive of building the supplynetwork.According to TSA president Johan van Zyl,“unless you have a strong supplier base, there is no reasonfor your existence... it’s in everyone’s interest for us to go toour suppliers and help them with quality and logistics, andlink them into our system”. 16But at the same time, the number of domestic suppliers has been sharplyreduced. In 2002, Toyota had seven platforms with 160 local suppliers(see Table 7). By early 2007, there were only two primary platformsand 75 suppliers. While local content was hardly changed, the positionof suppliers had been transformed. For the Corolla model that waslaunched in 2002, 41% of locally produced parts (by value) were fromglobal suppliers. 17 For the model being launched in 2007, this figure was82%, the large majority of which were foreign owned.In line with global practice, Toyota no longer issued drawings requestingquotations from suppliers. Suppliers in the global design centres inEurope, Japan and America were involved from the concept stage. Thismeant that access to the design centres was crucial for local suppliers,and from 2002, Toyota SA had spent much effort trying to partner localfirms with major foreign suppliers. For example, Toyota had helped tofacilitate two joint ventures by firms within the large local Metair Groupand a third was in the pipeline.Toyota’s preference was to deal with foreign-owned suppliers or jointventures. Even a 25% foreign-owned joint venture would effectively berun by the foreign firm and in most cases the foreign partner would increaseits stake over time. For this reason, local firms were frequently notsupportive of this initiative as they were reluctant to cede control. Manyasked Toyota to help them to set up technical agreements, but whilethis route remained a possibility, it was not Toyota’s favoured option, aslicence payments were cost raising. In dealing with locally owned firmsreluctant to relinquish control, TSA cited itself as an example of a firmwhich had only been able to survive and expand because the previousowners, the Wessels family, had seen the necessity of the TMC taking acontrolling stake.According to TSA, the parent company strategy was to localise in thecountry of assembly and it was also important to achieve certain levelsof local content in order to meet EU rules of origin for duty-free marketaccess. In spite of the high volumes being achieved and the fact thatToyota was keen to increase local content, this had not materialised butremained at about 40% for both the Corolla and the IMV. To move to60% local content would require a locally produced engine and therewere no plans to do this. Engines for the IMV are built in Thailand, theglobal hub of the project, while transmissions are produced in Indonesia.Previously, a large number of components had been produced by inhousefirms such as Toyota Automotive Components (TAC) and theToyota Stamping Division (TSD). Soft trim and plastic parts, wheel andtyre assembly and small stampings have now been outsourced. Plasticpainting, regarded as a core competence, has been brought in-house.TSD continues to produce major stampings.The position of local suppliers excluded from this network was extremelyprecarious. This group consisted of a multitude of smaller firms withcapabilities in areas such as tooling and press parts. However, withoutclose links to the design and technology of global suppliers they werelikely to be marginalised. Ironically, while the low model volumes in theSA assembly industry has always been a major complaint of the componentindustry, these smaller firms now found it difficult to cope withmodel volumes in excess of 100,000 units per year that Toyota wasproducing.149<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>15This section is based on interviews conducted in early 2007. ‘16‘Fortuner favours the brave’ (Financial Mail Special Report: The Motor <strong>Industry</strong>,15 September 2006).17‘Global supplier’ in this context refers to a foreign-owned firm or joint venture operatingin SA or a domestically owned firm with a technical agreement with a global supplier.The second tier of suppliers was also regarded as weak, with limited skilllevels, and Toyota was pessimistic about their prospects. While Toyotaencouraged the new global suppliers to expand their own local content,these firms were importing from countries such as China and local firmsThe SA Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports?
Sector StrategiesTABLE 7: CHANGES AT TOYOTA FROM 2002 - 20072002 2007150<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>No of platforms 7 2No of suppliers 160 75Percentage local content (component purchases only) 40-45 40-45Old Corolla (2002 model)New Corolla (2007 model)% of local components produced by global suppliers (Corolla) 41 82(46%: global firms, non-Japanese; 23%: Japanese firms;13%: technical agreements with global company)Source: Interviewscould not compete. Even large raw material suppliers appeared to be losingground. Only 30% of the steel used in the new Corolla was to belocally produced, compared to 90% in the previous model.Supplier development depended on the cycle of new model introductions.Following the launch of the IMV project in April, 2005, Toyota hadidentified 17 high-risk suppliers and appointed a team of mainly retiredtechnical staff, dubbed ‘grey berets’, to work with these firms in areasranging from kaizen plans to press loading and HR development. Thesefirms had been assisted through to late 2006. At the time of the interviewsin January 2007, production preparation was in progress for thenew Corolla launch in mid-2007. After this, product development staffwould again be deployed to assist 15 key suppliers to the Corolla andIMV. These firms were smaller, domestically and foreign-owned suppliers.The larger foreign-owned suppliers received back-up from their parentcompanies. However, foreign ownership did not necessarily signify parentcompany support and Toyota regarded some foreign-owned suppliers asbeing quite weak in some respects.Toyota’s evolving relationship with its suppliers is fairly typical of assembler-supplierrelations in the context of globalisation. The supplier basehad been consolidated and the level of foreign ownership much increased.Toyota now dealt with global follow-source firms which had either establishedoperations in SA through wholly owned firms or joint ventures.Much effort had gone into facilitating these arrangements. Toyota continuedto work with local suppliers, but the number of these was muchreduced. Some domestic firms which were not part of this new supplychain would become second-tier suppliers to the new global first tier andothers would seek out opportunities in the after-market. The remainderfaced very bleak prospects.6. Conclusions: the implicationsfor policyThe automotive industry is widely regarded as a major success story ofpost-1994 SA, mainly on account of the well-documented expansion inexports. Since the introduction of the MIDP in 1995, rapidly increasingvolumes of components – and more recently of vehicles – have beensupplied, mainly to first-world markets. Although export expansion hasbeen accompanied by greater efficiencies and a more rational industrystructure, key vulnerabilities remain.This paper has focused on the rapid rise in imports, which have risen particularlysharply in recent years in line with the boom in domestic vehiclesales. But the behaviour of many firms is revealing about their long-termstrategy. The investment in export capacity has been driven to a largeextent by the objective of earning import rebate credits. Firms have beenexporting in order to import. This is an important conclusion because itprovides pointers as to how firms would respond to a more open tradeenvironment and must be of concern to policy-makers. If the real target isthe domestic market, where does this leave the industry when assistanceis phased down still further?This is where critics of industrial policy may argue that what policy hasdone is to simply create an unsustainable export edifice in the same waythat the import substitution phase created an inefficient, inward-lookingindustry. While it is undoubtedly true that the automotive industryhas been subject to excessive government support – first high protectionand then overly generous export support – the counter to this is that thestructure of the industry is much sounder than it was. Investments havetaken place in reasonably high-volume production and achieved significanteconomies of scale. The industry is operating under a far less protectiveregime than was previously the case. Although nominal tariffs remainquite significant, the ability to rebate import duties effectively reducesprotection. With the sharp decline in export support since 1995, the incentivestructure is also tending towards a more neutral stance.
With quite high rates of growth in vehicle sales likely to be sustained asvehicle ownership increases, automotive imports are likely to loom largeas a major policy issue. Already in 2005, vehicle imports accounted fornearly 40% of the market in unit terms, and given the large number ofluxury vehicles being imported, the proportion is higher in value terms. Itis also significantly higher than expected at the time of the 2003 Review.While the strong rand, boom in domestic sales and related capacity constraintshave contributed to growing imports, policy-makers will be lookingat the policy parameters of the MIDP itself to assess whether the paceof liberalisation is appropriate.The hoped-for increases in local content have not fully materialised becausethe bulk of the industry is still stuck in a ‘transition phase’ (see Table3) and, with the possible exception of Toyota, is not yet advancing to thefull manufacturing stage with sufficient volume to justify high local contentlevels. But as the case of Toyota illustrates, even high volumes do notguarantee higher local content in a situation where the domestic supplybase is weak. Tariffs are continuing to decline, but as we have indicatedabove, it is also the ability to rebate import duties which has had a majorliberalising impact. With export growth leveling off, the result has been ahuge increase in the trade deficit in the sector to a record R27bn in 2005.Two questions emerge. The first is whether the growth path of the past10 years can be maintained. If we take this to mean an export-led growthpath, the answer is probably not. The export base is now very large andgrowth has already slowed. Expansion could continue but at a much slowerpace than the exponential growth which has occurred since 1995.The second question is whether automotive policy should proceed to fullyliberalise the domestic market. The reality is that, as we indicated earlier,SA is not part of a viable ‘automotive space’. But there are goodreasons to believe that it could become one. The domestic market couldeasily reach one million units by 2015. Assuming that economic recoverycontinues in the Southern African Development Community (SADC), thecombined region will begin to constitute a significant regional market.The conclusion, therefore, is that further substantive liberalisation nowwould be premature.So from a strategic point of view, where does this leave the industry andautomotive policy? The industry has been through a period of heavy protectionand then of large-scale export support and export expansion. Theincentive regime still favours exports but to a much lesser extent andprotection is effectively quite low, which has led to rapid import growth. Itis now time for the industry to move to a more balanced growth path onthe basis of policy which imposes a more neutral incentive structure. Thiswould involve some ongoing protection and assistance for production atlow to moderate levels. Under such a scenario, both the domestic marketand exports could provide the basis for sustained future growth.ReferencesBarnes, J. and Black, A. (2003) Motor <strong>Industry</strong> Development Programme:Review Report, Department of <strong>Trade</strong> and <strong>Industry</strong>.Barnes, J. and Kaplinsky, R. (2000) Globalisation and trade policy reform:Whither the automobile components sector in South Africa, Competitionand Change, <strong>Vol</strong>. 4, pp. 211-243.Barnes, J., Kaplinsky, R. and Morris, M. (2004) Industrial policy in developingeconomies: Developing dynamic comparative advantage in theSouth African automobile sector, Competition and Change. <strong>Vol</strong>. 8, No. 2,pp. 153-172.Black, A. (1995) Reducing protection in the motor industry: Likely impacton the component sector, <strong>Trade</strong> <strong>Monitor</strong>, <strong>Vol</strong>. 11, December.Black, A. (2001) Globalisation and restructuring in the South Africanautomotive industry, Journal of International Development, <strong>Vol</strong>. 13, pp.779-796.Black, A. and Muradzikwa, S. (2004) The limits to regionalism: The automotiveindustry in the Southern African Development Community. InCarrillo, J., Lung, Y., and van Tulder, R. (eds.) Cars, Carriers of Regionalism?Hampshire: Palgrave Macmillan.Board of <strong>Trade</strong> and <strong>Industry</strong> (1988) Investigation into the <strong>Industry</strong> ManufacturingPassenger Cars and Light Commercial Vehicles, Report No. 2627.Pretoria, Government Printer.Board of <strong>Trade</strong> and <strong>Industry</strong> (1989) Investigation into a Structural AdjustmentProgramme for the Industries Manufacturing Motor Vehicles andAutomotive Components: Phase VI of the Local Content Programme, ReportNo. 2767. Pretoria, Government Printer.Department of <strong>Trade</strong> and <strong>Industry</strong> (1997) Current Developments in theAutomotive <strong>Industry</strong>, Pretoria, Department of <strong>Trade</strong> and <strong>Industry</strong>.Duncan, D. (1997) We are Motor Men, Scotland: Whittles Publishing.Humphrey, J., and Oeter, A. (2000) Motor industry policies in emergingmarkets: Globalisation and the promotion of domestic industry. In Hum-151<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>The SA Automotive <strong>Industry</strong> in a Globalising World: What has Happened to Imports?
Sector Strategies152<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>phrey, J., Lecler, Y. and Salerno, M. (eds.) Global strategies and local realities:The auto industry in emerging markets, London: MacMillan.Humphrey, J. and M. Salerno (2000) Globalisation and Assember-SupplierRelations: Brazil and India. . In Humphrey, J., Lecler, Y. and M. Salerno(eds.) Global strategies and local realities: The auto industry in emergingmarkets, London, MacMillan.Kahn, S.B. (1987) Import penetration and import demands in the SouthAfrican economy, South African Journal of Economics, <strong>Vol</strong>. 55, No. 33, pp2<strong>38</strong>-248.Laplane, M. and Sarti, F. (2004) Mercosur: Interaction between governmentsand producers and the sustainability of the regional automotiveindustry. In Carrillo, J., Lung, Y. and van Tulder, R. (eds.) Cars, Carriers ofRegionalism? Hampshire: Palgrave Macmillan.NAAMSA (2005) Annual Report 2005. Pretoria: National Association ofAutomobile Manufacturers of South Africa.Republic of South Africa (1999) Board on Tariffs and <strong>Trade</strong>: Mid term Reviewproposals for the Motor <strong>Industry</strong> Development Programme, GovernmentGazette, Pretoria; 19 March.Shimokawa, K. (2004) ASEAN: Developing a division of labour in a developingregion. In Carrillo, J., Lung, Y. and R. van Tulder (eds.) Cars, Carriersof Regionalism? Hampshire: Palgrave Macmillan.South African Automotive Benchmarking Club (2006) Does ownershipmatter in the SA auto components industry: SA vs MNC-owned firm performancelevels, South African Benchmarking Club Newsletter, <strong>Vol</strong>. 9, No. 2.Sturgeon, T. and Florida, R. (1999) The World that Changed the Machine:Globalization and Jobs in the Automotive <strong>Industry</strong>, Final report to the AlfredP. Sloan Foundation, International Motor Vehicle Program, MassachusettsInstitute of Technology.Van Tulder, R. and Audet, D. (2004) The faster lane of regionalism. In Carrillo,J., Lung, Y. and van Tulder, R. (eds.) Cars, Carriers of Regionalism?Hampshire: Palgrave Macmillan.
sectorstrategies153CASSAVA TRADEINFORMATION BRIEFBY ALISON GOLDSTUCKTRADE AND INDUSTRIAL POLICY STRATEGIES (TIPS)1. IntroductionThe aim of the <strong>Trade</strong> <strong>Industry</strong> Brief (TIB) is to highlight potential export markets to SADC producerswho may not have the financial resources to engage in preliminary market research activities. TheTIB is not a detailed market intelligence report but rather highlights potential lucrative businessopportunities in a market. A TIB should not be used to determine whether one enters a particularmarket but rather to ask questions about a market and stimulate further research. A series of TIBshas been produced that covers a range of product clusters. These clusters represent an existing keyset of export products with potential for expansion, or a relatively new set, where an indicationof a competitive advantage for the region is apparent. This TIB showcases opportunities for SADCproducers in the cassava industry.Cassava is known as a poor man’s crop. It is predominately grown by subsistence farmers, as astaple crop, in developing countries that have a temperate climate. This has two important marketimplications. The amount of cassava traded compared to global production is miniscule; and thelargest exporters of cassava are not necessarily the largest producers. <strong>Trade</strong> patterns illustrate thatParticipation in international trade has becomeone of the most important factors in increasingthe prosperity of countries. Yet for manydeveloping countries, perhaps particularly forthose in SSA, trade is viewed primarily froma defensive perspective, with a focus on thedisruptive effects of imports rather than on theopportunities presented by increased access toworld markets. A key reason is the existence ofinformation market gaps that are often associatedwith trade facilitation and development indeveloping countries – information on theexport performance and potential of manydeveloping countries remains incomplete. TIPS’<strong>Trade</strong> Information Service series of marketbriefs, as part of its AusAID-funded SouthernAfrican <strong>Trade</strong> Development Programme, aimsto contribute to bridging this information gapfor existing producers in the SADC who maynot have the financial resources to generate afully fledged market research process. The briefsare not in-tended to act as the detailed exportmarket intelligence that successful exportingrequires, but rather as a basic first-cut analysisof export prospects, to allow enterprises to makethe decision on whether to initiate further marketresearch. This <strong>Trade</strong> Information Brief analysestrade in cassava.<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>
Sector Strategies154<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>import/export activity is concentrated between South East Asia and EastAsia. If regional trade is broken down it becomes apparent that China,Thailand and Vietnam are responsible for driving world trade in Cassava.Cassava is a versatile crop. It has a multitude of applications cutting acrossvarious industries and is used in a variety of products: Flour, food, animalfeed, paper, textiles, sweeteners, convenience meals, and bio-degradableplastics. To produce these products cassava is processed in numerousways. The simplest being the preparation of food for human consumption,such as flour, that involves peeling, grinding and drying cassava. While themost complicated process involves the creation of modified starches. Toensure that this TIB is concise, its primary focus is on “fresh, chilled, frozenor dried cassava, whether or not in the form of pellets made either frompieces of the root or from its flour, meal or powder. This product categoryfalls under HS 0714.10 Manioc (cassava).Cassava products fall into these broad categories, human consumption,animal consumption and industrial applications. These categories havedifferent supply and demand side drivers; as developing a generic agriculturaland industrial strategy for generic cassava products is not a usefulexercise. This TIB implicitly proposes that growers should target a particularmarket segment, on a global basis, or a particular region. Basedon trade data, the market for starches seems to offer the most promisingprospects as it provides the raw material base for an array of processedproducts. With respect to geographical markets, Africa’s demand for cassavapellets to feed its livestock offers potential for intra-African trade.The international starch market is extremely competitive and is dominatedby corn, maize and potato starch products. These crops have benefitedfrom substantial scientific research and thus have a technological advantagecompared to cassava. Cassava’s future prospects are rooted in improvingits supply side, with respect to increasing productivity by adoptingimproved varieties, which are more resistant to pests and have higherstarch content, and improve processing technologies. Developing countries’progress in reforming agricultural processes through distributingbetter planting material and implementing intensive production methodsfor small-scale farmers has been inconsistent. To address this failure, byproviding growers access to biotechnology and extension services, thecontent of national research programmes should be revisited, but moreimportantly, the manner they feed into international and regional agriculturalresearch programmes must be investigated.characteristics are discussed which provides the knowledge to develop aseries of value chains for cassava’s product clusters. The second section isa market study that describes the consumption, production and trade patternsbetween regions and countries. This information is used to establisha market’s size, with respect to its value, shape and growth patterns. Thisknowledge is used to identify where prospective export opportunities liefor SADC’s farmers. The last part of this analysis is to investigate pricetrends to gauge, at a simplistic level, whether an opportunity is economical.The last section provides exporters with information about gainingmarket access and placing their product into a market. This section highlightsimportant tariffs and non-tariffs barriers and also proves informationabout marketing and distribution channels.2. Rationale behind selectingcassavaBased on the following reasons, which will be explained in greater detailin this TIB, cassava was selected as a potential export crop for SADC’sfarmers:Cassava can be grown in difficult environmental conditions characterised by low orextreme rainfall and infertile, poor, sandy soil (ITC, 2003: 2);Cassava is a simple crop to maintain as it has no definite maturation point andthus can be left in the ground from 7 months to 2 years after planting and thenharvested as needed, in addition it can recover from pest damage and diseases;Cassava provides an opportunity to improve rural dwellers’ income by opening upmarginal lands under cultivation;Cassava provides farmers with the flexibility to opt for more capital intensive,efficient production processes as they develop as “production practices may becompletely manual, partially mechanized, or animal-powered, especially for landpreparation (Howeler, 2003 :5);Cassava is a labour intensive crop to harvest, and as a result it will provide employmentunskilled labour in rural areas;Cassava is a highly perishable, bulky crop and thus must be processed before it istransported, which opens up numerous opportunities for small-scale farmers to getinvolved in producing simple, value added products;Cassava has a wide range of applications ranging from food products to industrialstarches. The processes required to produce these products vary in complexitywhich gives different parties the flexibility to pursue markets that suit their skill andresource base; andCassava’s supply chain has an hour-class shape, which makes it simpler for smallscale farmers to be absorbed into the cultivation stage of cassava’s value chain.This TIB is divided into four broad sections. The first section defines theproduct and establishes its market. In this section cassava’s physical
3. Product definitionCassava is known by various names (see Table 1). For the purpose of thisTIB, the commodity’s common name is used and trade data is discussedat the HS 6 digit level under the classification 071410, Manioc (cassava).Other trade classifications are provided in the table, but are not discussedin the TIB, as a reference point to make it easier for interested parties toaccess leading importers/exporters trade data.4. Cassava’s characteristicsCassava is a perennial, woody shrub that grows between one to fourmetres in height. The root can grow up to 15cm in diameter and reach120cm in length to weigh between one to eight kilograms (ITC, 2003:2).The roots of a 1 to 1.5 year old cassava plant have a starch content thatranges from 20% to 32%, which is good compared to other starch foodcrops. Cassava is an excellent source of carbohydrates but is an inferiorsource of protein, fat and vitamins (ITC, 2003: 2).Cultivating cassava requires one to perform the following activities, inchronological order: select a site, prepare the land, prepare planting materials,plant, apply fertiliser, weed and cultivate, harvest, dry roots, grindroots and store. Processing a commodity to create a final product is themost complex stage in the value chain. This step will be dealt with indetail in the following section as processing activities are tied to productmarkets. Other stages of the value chain, such as packaging, marketing,distribution and transportation will be discussed in section 14.Cassava has been selected as a potential cash crop for SADC’s farmers asthe cultivation stage of its value chain, outlined in the above paragraph,is relatively simple. As a consequence these activities can be performedat the farm gate by a small-scale farmer or at the village or local level.Non-agricultural activities, so-called secondary activities, such as marketing,processing and packaging products are performed by fewer, largescale units. The unique feature of cassava’s supply chain is its hour-glassshape because it provides opportunities for numerous small-scale farmersto be involved in cultivating, harvesting and rudimentary processactivities compared to other activities along the value chain. The supplychain “begins with small-scale production units, followed by small-scaleprocessing units for the drying and/or milling of cassava” (ARC, 2006).The structure of cassava’s value chain provides potential contact pointsfor small scale farmers to participate in a larger market. This suggests thatthe growth and development of cassava product markets should benefita large number of resource-poor farmers located on poor lands and localprocessing units. Reaping the pro-poor benefits associated with cultivatingcassava hinges on developing distributed, simple micro technology forfarmers to process cassava into a transportable product that feeds intoindustrialists’ downstream processing activities.Cassava is propagated vegetatively from stem cuttings. This has both apositive and negative implication. The negative implication is that the rateof multiplication of new improved varieties is slow as cuttings do not storewell, and they are costly to cut and handle (FAO, 2000). The positive implicationis that it is easy to share good genetic material. This is importantas cassava’s yields are slightly less than other starch crops’ yields. This isdue to a dearth of research being allocated to cassava as in the past ithad an image of being a poor man’s crop. However interest in cassavahas been growing due to its use as a feedstock, at the same time its usein other industrial applications has also become more widespread. Thisinterest has sparked research in cassava to create better cultivars. Given155<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>TABLE 1: CASSAVA’S NAMING CONVENTIONSDescriptionCommon NameBotanical NameHarmonised Customs ClassificationEU: Combined Nomenclature of the EUChinese CustomsSource: International <strong>Trade</strong> Centre, 2003: 1-2NameCassava (Africa & Thailand) , Manioc (Brazil), Tapioca (India) ,Yucca (South America)Europe & USA: Cassava (roots) andtapioca ( products such as starch, pellets or dried chips)Manihot EsculentaHS 0714.10 Fresh or dried maniocsCN 07.14-1010 Pellets of flour and meal CN 07.14-1091 Human consumption, fresh and whole or without skin andfrozen, whether or not sliced and packagedCN 07.14-1099 Other07.14-1010 Fresh manioc 07.14-1020 Dried manioc 07.14-1030 Chilled or frozen maniocCassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies156<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>that cassava is vegetatively propagated, good genetic material can becopied, and thus a new strain can be quickly introduced into mainstreamagricultural systems.Although cassava can grow in dire circumstances, the best conditions are150 inches of rainfall, temperature between 25-30C, an altitude below2000 meters and fertile, sandy-clay soil with a ph range of 5.5-6.5. Cassavacannot survive flooding or freezing conditions. SADC has ideal climaticconditions to grow Cassava, especially Mozambique, Swaziland, Lesotho,Malawi, Tanzania Zimbabwe, Zambia, Madagascar and the DemocraticRepublic Congo (DRC). Furthermore even countries that are not endowedwith good arable land, such as Angola and Namibia, could replace theirmore “fragile” crops with cassava.Cassava does not have a mature stage. This allows the crop to be harvestedat a farmer’s discretion. A plant can be harvested when its rootsare sufficiently developed to meet a consumer’s requirement, or “it canbegin six to eight months after planting (ITC, 2003: 2) or delayed till thenext growing season. This feature makes cassava an ideal secondary cropfor small-scale farmers in SADC to grow, as farmers can stagger their harvestingactivity to ensure that resources are not thinly stretched betweencrops. In addition this feature allows farmers to influence the market’ssupply by delaying harvesting if the market is over supplied and to takeadvantage of price swings.One of the benefits of growing cassava is that farmers can generally decidewhen they prefer to harvest the crop, however due to cassava’s physicalattributes post-harvesting activities must follow a strict, short-timeframe.Therefore a farmer’s ability to devote resources to post-harvestingactivities will affect when cassava should be harvested. The issue is that afarmer’s activities are subject to time constraints and the flexibility gainedduring the pre-harvesting stage should be weighed up against post-harvestingactivities.Raw cassava roots comprise 70% water and are highly perishable. Oneto three days after harvesting the roots start to deteriorate. If the roots donot receive special treatment, they must be processed within two or threedays after they have been harvested (ITC, 2003: 2). Given the perishablenature of the crop’s roots and time delay between processing the crop dueto inadequate processing machinery at the farm gate, storing it in woodencrates, trenches or moist mulch to increase its shelf life is important.The high water content of cassava’s roots not only shortens its shelf life,but also increases the cost of transporting the product, as it tends tobe heavy and bulky. These factors suggest that transporting raw cassavalarge distances is uneconomical and logistically difficult. SADC’s smallscale farmers tend to be in rural areas, where access to roads and infrastructureis poor. This will complicate transportation issues. For cassava tobe a viable crop cash earning crop, cassava’s first stage of processing tocreate a transportable product must be simple and done at the farm gate.This opens up an opportunity for small-scale farmers to get involved increating a value-added product, albeit a simple one, and could serve asan initial entry point for them to participate in supplying other processedproducts. One potential issue that would however inhibit this step, andreduce the ability for farmers to become integrated into the value chain istheir limited’ access to infrastructure (and other-technology. Finance, etc)that would allow them to process the cassava into a storable product.Nigerian engineers in response to this problem are currently trying to developprocessing equipment that can be used by farmers in the rural areasto process cassava into an easily transportable product. SADC’s engineerscould potentially collaborate with their Nigerian counterparts to developthis technology, which has the additional benefit of not only creating newtechnology but also fostering regional co-operation.The cost incurred to produce cassava is location specific and time boundand thus a generic cost schedule cannot be provided. A farmer’s costsare dependant on climatic factors that affect a plant’s growth pattern,which is tied to the time of planting (Lamchaiyaphum et al 2006:1). Eventhough farmers’ cost structures are not identical, they share a similar profile.The bulk of a farmer’s production costs are made up of three maincomponents which are in descending order, labour, then land and finallymaterials (Lamchaiyaphum et al, 2006). Compared to the other regionsthat would be SADC’s main competitors, predominately Asia and SouthAmerica, SADC has something of a competitive advantage in unskilledlabour and land (although in Asia unskilled labour is abundant, whilst inSouth America land is abundant). Typically, variable costs comprise 60%of total costs, while fixed costs account for the remainder. A higher ratioof variable to fixed costs makes it easier for SADC’s farmers to becomecassava growers. First, their initial capital outlay is lower, which reducestheir bank loan. The structure of the banking industry makes it difficult forSADC’s small scale farmers to rise capital from traditional financial institutions,as a result they are often forced to borrow money from micro-lendersthat charge exorbitant interest rates. Second, lower fixed costs reducesa farmer’s potential downside and thus reduces his/her risk profile. Third,a smaller fixed cost ratio gives a farmer more flexibility to manage his/hercash flow.
FIGURE 1: PRODUCTS DERIVED FROM CASSAVA’S ROOTCassava rootCassava starchCassava chips & pellets Direct consumption Peels & pulp157Direct consumptionSago PearlsNoodlesTraditional dessertsModified starchAnimal feedAlcoholic: fuelCitric acidAcetylated: sauces, frozen food, instant soup, pastries, glueCrosslinked: salad dressing, canned food, sauces, paper textilesOxidized: candies, instant soup, salad dressing, paper textilesCationic: paper, textilesAlpha: animal feed, mosquito coil, sauces, dessertsBoiling, roastingDrying: flourAnimal feedCompostMushroomsAlcoholEthanol: liquor,industrial and medicalalcoholOrganic acid<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>SweetenersGlucose/Dextrose: candies, beverages, canned food, medicine, creamersFructose/high fructose syrup: beverages, pastries, dessert, candies, saucesSorbitol: toothpaste, cosmetics, vitamin CCitric acidAmino acid & derivativesMonosodium glutamateLysine: amimal feedSource: Howeler, 2003:135. Demand- and supply-sidevariablesThe cassava shrub contains a root and leaves, which can both be processedto make various products. More products can be made from cassava’sroot than its leaves. These products require more complex valueaddedactivities and have a greater value. As a result this TIB exclusivelydiscusses products made from cassava’s root (see Figure 1). Cassava’sproducts fall into three broad categories: Food for human consumption,animal feed and industrial products. These categories must be discussedindividually as the manner in which cassava is processed, distributed andmarketed is different for each category.5.1. Human consumptionBefore the root of the bitter cassava variety can be eaten it must be processedto eliminate potentially toxic concentrations of cyanogenetic glucosides.Processing can take the form of soaking the root in water, crushingor heating it. Countries have developed various traditional methods toprepare cassava, which include peeling, boiling, baking, frying and gratingit to extract starch. The refined product is then dried over a fire or leftin the sun to dry for 2-3 days. It is then added to soups and stews as athickener, or fermented and cooked. Extracted starch can be used to makebreads, crackers or pasta. The leaves of the cassava plant are edible andprovide a rich source of protein and vitamins A and B. They are eatenas a green vegetable and prepared in a similar manner to spinach (ITC,2003:3).Processing cassava and selling it as a product for human consumption todeveloped countries’ specialised food markets is a potentially lucrativemarket. Increased awareness of wheat allegories among consumers hascreated a market for a substitute product. Cassava’s dried roots provideanother source of carbohydrates for people who have wheat, corn or riceallergies. In addition cassava products could be marketed to consumerswho have a taste for exotic foods and health foods that have a lower fatand sugar content. Cassava absorbs less fat when it is fried than otherstarches, as a result cassava can be used as a healthier alternative toproduce snack and convenience foods. Over the past five years LatinAmerica’s snack and convenience food industry has created a range ofCassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies158<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>cassava products and successfully marketed them in the United States(US), European and Japanese market” (ITC, 2003: 15).Cassava has the potential to become a lucrative speciality food product.However turning that potential into trade flows requires substantial resources.Marketing cassava would be an expensive undertaking as one’smarketing strategy would involve educating consumers about an unknownproduct and then creating an appetite for the product. The market’s incumbentshave an interest in hindering the spread of cassava productsas they have invested in technology that favours potato based products.To introduce cassava products into this type of market would require accessto financial resources and a strategy that builds on present demand,explores alternative distribution channels and emphasises the health benefitsof consuming cassava products compared to potato or maize basedalternatives. It might be argued that pursuing this market is a lengthy andexpensive process, and given SADC farmers’ limited resources it might notbe viable. However this market has the potential to be very profitable andthus this opportunity could be marketed to venture capitalists or boutiquefood processors. To gain a foothold into this market, a starting point couldbe to target developed countries that have a large immigrant populationand health food stores. In an effort to make the product more attractiveto consumers, at the onset of the marketing campaign, the cassava basedproduct would probably be priced below the traditional alternative. In Brazilpre-cooked, deep-frozen cassava fingers are priced 10-15% below theprice of deep-frozen potato chips (ITC, 2003:31). However as the productgains popularity and consumers’ perceive it to be a superior productbecause of its health benefits, it could probably be sold at a premiumcompared top its maize or potatoes based product. In addition fresh cassava’sshort shelf life and bulky nature complicates logistics and increasestransportation costs, and thus introducing processed cassava foods intothe market is a better strategy than supplying fresh cassava.The International <strong>Trade</strong> Centre of the UNCTAD/WTO publishes marketwholesale import prices for cassava destined to be used for human consumption.Although these prices cover Costa Rica’s exports of cassava tothe European Market they give one a sense of the market’s volatility andvalue. This information can be found at http://www.intracen.org/mns (ITC,2003:23).5.2. Animal feedCassava animal feed is used to feed cattle, sheep and poultry. Feed ismade from processing the plant’s roots into either pellets or chips. Cassava’sroots are an excellent source of carbohydrates but its protein andvitamin content is poor. As a result cassava feed must be supplementedwith soymeal or leaves from the cassava plant (ITC, 2003).Cassava chips are more widely produced than pellets. They are producedin Thailand, Malaysia and Nigeria. Processing cassava into chips involvesslicing them into pieces, not longer than 5cm to ensure they can be storedin silos, and drying them in the sun for 2-3 days or until their moisturecontent is between 13%-15% (ITC, 2003). During the drying phase, thechips must be regularly turned over. Slicing the roots can either be performedmanually or mechanically. The mechanised option is more efficientas it takes one to complete a task that would require three days manuallabour (International Starch Institute, 2007). In SADC the need for efficiencymust be weighed against surplus unskilled labour and a farmer’sability to access finance.The diesel/electric powered machine required to slice roots is not complicatedor high-tech, it comprises “a rotating notched cutting disk or knifeblades mounted on a wooden frame equipped with a chopper” (InternationalStarch Institute, 2007). Roots can be trimmed, peeled and washed,before processing, to create a superior quality product (ITC, 2003). Ingeneral, 2kg-2.5kg of fresh roots is required to produce 1kg of chips(ITC, 2003:3), which can be translated into recovery rate of roughly about20%-40% (International Starch Institute, 2007). The by-product from thisprocess is used to make cassava meal, which is categorised as an inferiorproduct compared to cassava chips, pellets and broken roots because ofits lower starch content, higher impurity content and it is more difficultto transport.Producing cassava chips is a fairly simple process that does not requirelarge capital investment. It provides farmers and small scale businesseswith an opportunity to invest in a chipping factory to get a foot hold intothe value-added product market. As processing must be done within closeproximity of the growing areas due to the perishable and bulky natureof cassava, it ensures the benefits arising from value-added activities aretrapped in communities where cassava is grown.Cassava chips are used as the starting point to produce cassava pellets.When chips are dry, they are transported to a pellet processing factory.To make pellets, chips are mixed with palm oil, grind, steamed, dyed andcooled into a cylindrical shape. The industry standard for these cylindersis roughly 2-3 cm long and about 0.4-0.8 cm in diameter, and their appearanceand texture should be uniform (ITC, 2003). Compared to chips,cassava pellets are regarded as a superior value-added product. This isdue to the following reasons. First, pellets’ product quality is more uni-
FIGURE 2: PRICE TRENDS OF CASSAVA FEED AND OTHER COMPETING PRODUCTS300300250250159US$/tonne2001501005001990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004Year200150100500<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>barley Spainsoybean mealbarley Germanycassava+soybean mix cassava pellets maize USA #2Source: Howeler, 2003:29form. Second, pellets are more compact, occupying 25%-30% less spacethan chips. This reduces transportation, handling chargers for off-loadingproducts and storage costs. Third, pellets are a more stable, sturdy productand reach their destination with considerable less damage than chips. Onaverage, one ton of fresh roots produce 450 kg of chips or 440 kg of hardpellets (Howeler, 2003:15).Generally the demand for cassava chips and pellets is driven by a population’sconsumption of livestock products. Generally wealthier consumersinclude more complex proteins in their diet. Therefore as a country develops,reflected in rising per capita income levels, its population improvesthe quality of its diet resulting the consumption of livestock products toincrease. Second, the demand for cassava is driven by its relative pricecompared to substitute products. Third, the price of complementary products,in this case protein-rich meals, affects the demand for cassava pellets/chips. According to the International <strong>Trade</strong> Centre (ITC) the industrystandard for cassava feed comprises 80% cassava pellets and 20% soybeanmeal. As a result soybean meal prices affect the competitiveness anddemand for cassava feed compared to its substitute products, ultimatelyaffecting cassava’s price. Indirect factors that affect the demand for cassavaare exchange rates, especially the Euro/US$ exchange rate; countries’agricultural polices and climatic conditions.Prices for animal feed are subject to fluctuations (see Figure 2) as the marketis influenced by the interaction of various factors. The market for cassavafeed is affected by countervailing forces, or the knock on effect frommovements in grain markets, and thus is relatively unstable compared toits substitute grain products. For instance, the EU’s grain policy increasedsoymeal prices, which in turn dampened the demand for cassava feed,however the Euros strength compared to the US$ increased the cost ofimporting wheat from Eastern Europe, thereby increasing the demand forcassava feed. Given the impact of various countervailing forces, it is difficultto judge whether this market should be targeted by SADC farmers asan export product. Even though the same set of factors holds for variousmarkets, there are slight nuisances. As a result SADC’s farmers shouldnot base an export strategy on generalisations. For example in the EUthe demand for cassava is influenced by domestic grain prices, especiallybarley, and manufacturers’ ability to source cheap wheat from EasternEurope. Alternatively in China demand is affected by the price of sweetpotatoes..Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies160<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Based on experts’ opinion the market for cassava feed is entering intoits consolidation phase, which is characterised by demand growing at asteady rate and demand side driven pressure to reduce supply side costs.Since the 1980s, the demand for cassava feed has remained stable causingtrade levels to stabilise. Although Asia and Africa‘s demand for feedhas grown it has barely managed to offset the EU’s 1970-1980 demandlevels (ITC; 2003). It is predicted that South Korea’s demand for cassavafeed should decline over the medium term as the growth of its livestockindustry decreases due to greater imports of livestock products. In essenceKorea is shifting its exports from feed to the finished good. The demand forcassava feed in Viet Nam, Indonesia, China, and Thailand should continueto grow as their populations’ per capita income increases, stimulating thedemand for livestock products. Apart from China, the other countries cansatisfy domestic demand through domestic production. In Latin America,in particular Brazil and Columbia, the demand for cassava feed shouldincrease. An interesting trend is emerging. Although the consumption ofcassava feed should increase in Asia and Latin America, it will be met bylocal production; as a result export opportunities represent a fraction ofthe total market.Another factor that SADC’s farmers should consider before building capacityto supply this market is its rivals, such as Thailand, Vietnam andIndonesia’s pre-existing level of investment. Generally the greater thelevel of investment, the more likely a country will defend its markets. Thisis especially the case for Thailand, whose government has a history ofproviding its farmers support throughout the value chain. Thailand’s 200pellet factories have a combined capacity to produce roughly 10 milliontonnes of pellets per year, but the EU’s quota is only 5 million tonnes p.a(Lamchaiyaphum et al, 12). Excess capacity could lead to a scenario wherefactories might operate at 50% of their capacity if they rely solely on theEU’s demand for pellets. As a result the Thai government and the its cassvaassociation are motivated to aggressively capture other markets, Thisanecdote raises the issue whether SADC should compete against Thailandor rather explore other markets. The demand for cassava feed has potentialin Sub-Saharan Africa. This market does not “exist” in an establishedform due to institutional and supply side factors. Perhaps a possible strategyfor SADC’s farmers to explore is to put effort into creating a marketrather than trying to break into established markets, such as China, whichare highly competitive.5.3. Industrial usesStarches are used in various markets, such as the adhesives, explosive,paper, construction, metals, textiles, cosmetic, pharmaceutical, miningand food industries, and applied to a host of applications within markets.The food industry uses starch to produce monosodium glutamate (MSG),lysine, high fructose, glucose syrup, dextrose monohydrate, dextrose anhydrousand sorbitol. Given the widespread use of starches, this TIB does notprovide an exhaustive list of applications for cassava starches. Cassavacan be used to produce a native or modified starch. These starches canbe used as a finished product or either as a raw material to create a substancethat is used in a manufacturing process. An example of a finishedproduct is MSG, and an example of a raw material is organic acids andamino-acids that are used by to produce food, plastics, synthetic resins,rubber products etc.As a native starch, cassava has high amylose content, and as a result ithas a neutral taste, is odourless and has the smoothness and transparencyof a gel. Its unique properties are its viscosity and resistance to shear(ARC, 2007). These properties make cassava starch an ideal product forthe food processing industry. In addition cassava starch can with standacidic conditions and is stable in freezing conditions but breaks downwhen it is heated.Modified starches are produced by manipulating a native starch’s intrinsicphysical, chemical or micro-biological processes to meet a user’s requirementsfor his/her specific application. For example cassava starchwould need to be modified to produce bio-gradable plastics or any applicationthat requires properties associated with a low amylase content.This process uses advanced technologies that are rapidly evolving. Nativeand modified starches are not perfect substitute products, even thoughthey are used in cross-over markets. These starches are used to producesweeteners (maltose, glucose syrup, glucose and fructose), hydrogenatedsweeteners (sorbitol, mannitol and maltol) and MSG. In certain markets,where consumers are against genetically modified products, such as babyfood, a native starch would be a preferable option (Howeler, 2003). Generallymodified starch is used in “heavy” manufacturing applications: Paperindustry, textile industry (warp sizing, cloth finishing and printing),construction materials, medicines, etc.On a global basis the market for starch is growing as economies continueto industrialise and consumerism spreads into peri-urban and rural areas,changing people’s cultural preferences and values, altering their lifestylesand what they consume. These demand and supply factors have increasedthe level of consumption, and also, changed the type of products demandedby end-users. Demand for processed foods, paper products, biodegradableplastics and cosmetics continue to rise. These products areproduced using starches. Although the market for starches is growing, the
pertinent question is whether is the market for cassava starch is growing?The answer to this question lies in exploring what type of products isdemanded, whether cassava starch has the properties to cater to this market,and does cassava starch face competition from substitute products.As mentioned previously, native cassava starch has ideal properties to beused by the food industry to produce processed foods and sweeteners.However cassava starch would need to be modified to produce plasticsor any product that requires a “waxy” compared to a gel-like substance.Substitutes for cassava starch are maize, potato and wheat. These productsare entrenched in developed countries markets the US’ prefer maizestarch and Europeans prefer potato and wheat starches. These starchesdominant market position is due to historical usage patterns, the continueddevelopment of products that require these starches’ properties andthe fact that the producers of these starches reside in developed countriesand thus have the resources to conduct scientific research to create newapplication for these starches. For cassava starch to gain a sizable marketposition, research is required into its properties and the development ofmodified starches “with specific properties that make them preferable forcertain industries” (ITC, 2003:32). As mentioned previously to compete inthis market requires substantial scientific resources, which SADC does nothave access too.The market’s growth potential is impressive because the demand forstarch based applications in the food industry and industrial sector is increasing,and industry is searching for a cheaper substitute, as a resultmarket timing to introduce a new starch alternative is excellent. Howeverthis has no consequence if SADC’s farmers do not have the ability totap into this market due to their technological constraints. In totality thestarch value chain is technologically advanced, however within the chainexists relatively simple components. Over the short to medium term thereexists an opportunity for SADC’s small scale farmers to produce wet starchthat could be sold to factories to produce, higher quality dry starch. Althoughthis option provides an entry point for small scale farmers to enterinto the cassava starch value-chain, it reduces the overall quality of thefinal produced product. Factories’ quality of starch and the efficiency of itsconversion process are optimal when roots are used.An emerging market for cassava starch is to produce biodegradableproducts, such as packaging material and kitchenware. Discarded plasticproducts have the potential to cause environmental pollution, and as aresult discarding these products places a burden on municipalities’ wastemanagement system (ITC, 2003). Studies show that consumers and industryparticipants are interested in buying and supplying bio-degradableplastic products. This market’s annual growth is estimated to be 30% inEurope and the US, products provided these product’s physical propertiesmeet industry standards and they can be placed o the compost heap (ITC,2003:32). This could represent an FDI opportunity for SADC which hasthe land, labour and climatic conditions to grow cassava, but requires atechnology partner and capital to build factories.6. Countries’ productionpatternsAccording to the Food Agricultural Organisation (2002) cassava is grownin 101 countries. These countries are not evenly dispersed among regions,“in 2003 about 54% of cassava in the world was produced in Africa,29% in Asia, and only 14% in Latin America and the Caribbean” (Howeler,2003: 2). Furthermore the demand drivers stimulating productionamong regions are different. In Asia, Latin America and the Caribbean,cassava is primarily produced for the domestic feed, while in Africa cassavais produced for human consumption. Although Thailand and Chinaproduce cassava to make animal feed, it is not their primary market. Chinaproduces cassava for industrial applications, in particular raw material forstarch production (MSG, sweeteners), and Thailand produces cassava asmainly an export crop.From 2000-2004 the global production of cassava grew at a modest rateof 5% p.a calculated on an average annual basis (see Table 2). The top10 global producers of cassava grew their production by 3% from 2000-2005, while other producers achieved 5% growth. This indicates thatemerging producers have the potential to move into the top 10. Theseemerging producers include Vietnam, Paraguay, Malawi, Madagascar,Peru, Zambia, Rwanda, Senegal, Cambodia and Costa Rica.The world’s production of cassava is geographically concentrated in Africaand Asia. This is confirmed by the fact that nine out of the world’s top 10producers are located in the above regions, and that the world’s top 10producers comprise 76% of the world’s production. Although the world’sproduction of cassava chiefly resides in 10 counties, these countries’ shareof global production is relatively small, excluding Nigeria. From 1990 to2004 countries’ positions within the top 10 changed. In 2004 Nigeriabecome the world’s largest producer of cassava, relegating Brazil to secondposition. Eight of the top 10 producers’ share of global productiondeclined in 2004 compared to their 1990 level. The biggest losers wereBrazil, Thailand and the Democratic Republic of Congo (DRC). Nigeria andGhana managed to increase their share of global production during 2004compared to 1990.161<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies162<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>A notable feature is the presence of six African countries among the top10 producers. Furthermore, five of these countries’ average annual growthrate was positive and greater than their counterparts on the list, excludingIndonesia. Africa has the distinction of having the largest producer,Nigeria, and the fastest growing producer, Angola. Africa’s position as theworld’s dominant producer of cassava can be attributed to governmentpolicies to improve food security, introducing new higher yielding, diseaseresistant cultivars and favourable climatic conditions (ITC, 2003). Cassavahas the potential to be an attractive export crop as it builds on Africa’sexisting strong productive capacity.In 2004 the world’s largest producers of cassava were Nigeria, Brazil andThailand, whose share of global production was 19%, 12% and 11%,respectively. Over the 2000-2004 period, both Brazil and Thailand’s averageannual growth rate was significantly lower than Nigeria’s. Cassavaproduction in Nigeria grew, on an average annual basis, by 4.50% p.ato become the third fastest growing top 10 producer. Nigeria’s growthis impressive as it is off a larger base than other top 10 producing countries.Nigeria’s success is due to the government’s initiative to improve theinteraction between the industry’s supply and demand side capabilities.The remarkable feature about the Nigerian government’s approach wasthe manner in which policies were sequenced and the ability to drawon international resources by forming partnerships with internationalagencies. Initially the government’s policies focused on improving farmers’yields and product quality. The second stage was to create a stablesource of demand for a relatively simple, value-added product that couldbe processed at the farm gate. The government legislated that bread mustcontain a certain percentage of cassava flour. Once the government hadstimulated demand for cassava, its next initiative was to build the industry’ssupply-side to produce sophisticated value-added products. SADC’sfarmers could learn from Nigeria’s experience to build a regional industryas SADC’s farmers face similar constraints. Furthermore encouragingTABLE 2: MAJOR PRODUCERS OF CASSAVA (‘000TONS)Year Average annual growth Percentage of total1990 1995 2004 1990-2004 2000-2004 1990 2004Nigeria 19 043 31 404 <strong>38</strong> 179 4.09 4.50 12.54 18.82Brazil 24 322 25 423 23 927 -0.12 0.63 16.02 11.79Thailand 20 701 16 217 21 440 0.25 2.98 13.63 10.57Indonesia 15 830 15 441 19 425 1.47 4.82 1042 9.57DRC 18 715 16 870 14 951 -1.59 -1.62 12.32 7.37Ghana 2 717 6 611 9 739 9.55 4.69 1.79 4.80Tanzania 7 792 5 969 6 890 -0.87 -0.82 5.13 3.40India 4 962 5 85 6 700 2.17 2.74 3.27 3.30Angola 1 600 2 550 6 650 10.71 10.67 1.05 3.28Mozambique 4 590 4 178 6 413 2.42 4.58 3.02 3.16Vietnam 2 276 2 212 5 573 6,61 29,42 1.50 3.28Paraguay 3 550 3 054 5 500 3.18 19.25 2.34 2.71Uganda 3 420 2 224 5 500 3.45 2.59 2.25 2.71China 3 216 3 517 4 216 1.95 2.48 2.12 2.08Benin 937 12<strong>38</strong> 2 955 8.55 5.89 0.62 1.46Malawi 145 328 2 559 22.77 -1.84 0.10 1.26Madagascar 2 292 2 400 2 191 -0.32 -2.88 1.51 1.08Colombia 1 939 1 801 1 934 0.02 2.04 1.28 0.96Philippines 1 854 1 906 1 641 -0.87 -1.82 1.22 0.81Coted’Ivoire 1 393 1 608 1 500 0.53 -2.95 0.92 0.74Top 20 producers 141 293 150 809 187 891 2.06 3.49 93.04 92.61Other producers 10 571 10 993 14 988 2.53 1.65 6.96 7.39Total production 151 865 161 802 202 879 2.09 3.35 100.00 100.00Source: Food and Agriculture Organisation Statistics (FAOSTAT)
intra-regional knowledge could be the first step toward establishing anAfrican cassava hub that gives Nigeria and SADC access to supply-sideresources and a demand base to build a lucrative industry.SADC’s relative position is dropping. It needs to be borne in mind thatthese statistics might be conservative as a large percentage of cassavagrain in SADC is not traded and consumed as a subsistence crop.The interaction between cultivar type, planting season and soil type determineyields. If high yielding cultivar varieties are planted combined withgood management practices, cassava yields can reach 20-25 tonnes perhectare (IFAD, 2000:9). Productivity levels, based on yields per hectare,are higher in Asia, Latin America and the Caribbean compared to Africa.However Africa’s yields have reported the fastest growth, albeit off a lowbase, while Latin America and the Caribbean’s yields have stagnated(IFAD, 2000:9).Over the past decade the area allocated to cassava production in Asiahas decreased but yields have markedly increased and as a result productionhas steadily increased. Improved productivity levels stems from therespective governments’ “effort to distribute widely the new high-yieldingand high-starch varieties, as well as the adoption of improved culturalpractices, such as more balanced fertiliser use and soil conservation measures”(Howeler et al, 2004). Thailand and Vietnam have aggressively reformedtheir cassava sector. In Thailand new cassava varieties are plantedin roughly 100% of its farmlands and 70-80% of farmers apply chemicalfertilisers (Howeler et al, 2004). In Vietnam new cassava varieties areplanted in about 50% of its cassava growing area and about 80% offarmers apply chemical and/or organic manures (Howeler et al, 2004).This has two implications for SADC farmers’ ability to reduce Asia’s dominanceof the cassava market. Thailand has access to a growing domesticand international market for its cassava products. However Thailand’sability to service this demand could be potentially strained in the mediumterm, as it does not have any more land available for cassava cultivationand it has exploited productivity gains associated with planting new cultivarsand crop management. Implicitly Thailand is reaching its productiveceiling, yet demand in the region and domestically is increasing. Africahas access to the factors of production and has already established itspresence as a large producer, which can be built upon to create the momentumto improve its productivity, required to capture potential surplusdemand in the Asian market.In 2004 SADC’s production comprised 20% of global supply (refer to Table3), which is slightly larger than the world’s largest producer, Nigeria.This comparison illustrates that SADC’s productive capacity is significant.From 2000-2004 SADC’s production grew by 1.08%, which is lower thanthe global average of 3.35%. This is a troubling trend as it indicates thatThe decline in SADC’s production could be easily reversed as the regionhas the climatic conditions, access to land and abundant labour to improveits performance. SADC has access to the factors of production toproduce cassava but not trade it. Simple processing technology that canbe used at the farm gate or in the village to create an easily transportableproduct does not exist. A general lack of infrastructure exacerbates theproblem of transporting a product which by its very nature is difficult todistribute unless it is processed. As a result trade in cassava is constrainedchiefly by two bottlenecks, access to simple, cheap micro technology asfarmers access to capital is limited and general infrastructure. On the demandside SADC’s farmers and its industries are not taking advantage ofcassava’s various applications, as it is rigidity regarded as a staple crop.This illustrates that there is an underlying marketing problem and alsothat industries’ supply chains act in isolation. For example, although cassavais an agricultural product its value chain could interact with livestockproducers’ value chain or South Africa’s energy value chain, as cassavacan be processed in animal feed or bio-fuels.SADC has the potential to increase its production, and more importantly,use cassava as a crop to bring marginal subsistence farmers into the casheconomy, and based on Nigeria’s example (refer to the appendix) it isan achievable task. The region also has the opportunity to learn fromNigeria’s experience with respect to moving the production of cassavaaway from subsistence farming to inclusive commercial farming. This is avaluable source of intangible capital that SADC’s farmers can tap into, andif used properly, should reduce the potential hurdles that SADC’s farmerscould face when they establish a cassava supply chain.7. Countries’ consumptionpatternsData suggests that a region’s economic development influences thetype of value-added cassava products it demands and consumes. Generallyleast developed regions consume cassava as a staple food, whiledeveloped regions use cassava as a raw material to produce starches.In Africa cassava is predominately consumed as a staple crop for humanconsumption. Africa’s consumption level is tied in theory to its productioncapacity. A miniscule share of Africa’s total consumption is used as animalfeed. This should change over the medium term as government and163<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector StrategiesTABLE 3: SADC’S PRODUCTION OF CASSAVA (‘000TONS)Year Average Annual Growth Percentage of Total164<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>1990 1995 2004 1990-2004 2000-2004 1990 2004Democratic Republic of Congo 18,715 16,870 14,951 -1.59% -1.62% 12.32% 7.37%Tanzania 7,792 5,969 6,890 -0.87% -0.82% 5.13% 3.40%Angola 1,600 2,550 6,650 10.71% 10.67% 1.05% 3.28%Mozambique 4,590 4,178 6,413 2.42% 4.58% 3.02% 3.16%Malawi 145 328 2,559 22.77% -1.84% 0.10% 1.26%Madagascar 2,292 2,400 2,191 -0.32% -2.88% 1.51% 1.08%Zambia 640 744 957 2.92% 4.09% 0.42% 0.47%Zimbabwe 95 150 190 5.08% 2.08% 0.06% 0.09%Seychelles 0 0 0 0.00% 0.00% 0.00% 0.00%Mauritius 0 0 0 -2.67% -3.51% 0.00% 0.00%Total SADC Production 17,154 16,320 25,851 0.92% 1.08%Other Producers 134,710 145,483 177,028 2.42% 3.96% 88.70% 87.26%Total Production 151,865 161,802 202,879 2.09% 3.35%Source: FAOSTATinternational agencies’ initiatives to build a livestock feed industry gainsmomentum. In Latin America and the Caribbean approximately 60% ofcassava is consumed by the traditional food sector, while the remainder isprocessed into animal feed and used by industry to produce starch (ITC,2003: 15). In Asia cassava is predominately used as animal feed, in theform of pellets, or industrial applications to produce starches. An exceptionto this generalisation is Indonesia, India and Vietnam; where cassavais utilised in human consumption. This region is also experimenting withproducing ethanol from cassava. In the European Union (EU) cassava ismostly consumed by the livestock industry as an animal feed for its porkindustry. The EU’s consumption of cassava feed is falling and the slack isbeing absorbed by the demand for industrial starches.The top 10 consumers of cassava are located in Asia and Africa; as aresult it is fair to say that the consumption of cassava has a geographicaldimension. Based on data, this trend should not change as the emergingconsumers of cassava are Thailand, China, Guinea, Rwanda, Peru, Kenyaand Vietnam.The ten largest consumers of cassava, with respect to volume and notvalue, comprised 73% of global consumption in 2004. In totality the top10 consumers’ market share remained relatively stable from 1990-2004,as it moved within a 1% range. Countries relative ranking within the top10 from 1990-2004 also remained relatively unchanged, barring Indonesiaand Nigeria. With respect to market share over the period, countries’fortunes have changed: the biggest loser was the DRC and Indonesia andNigeria were the biggest gainers.From 2000-2004 growth in global consumption was negligible, reachingonly 0.15% (refer to Table 4). The top 10 consumers’ demand for cassavadeclined by 0.35% from 1990-2004. Nine of the top 10 countries usecassava as a staple food, and thus it is not surprising that the market’sgrowth in demand is insignificant. The market’s historical low growthrate should not deter investors’ interest as cassava has a dual market.The tradable market is dominated by the Asian exporters that supply cassavapellets and chips to the world, and the staple food market, mostlyin Africa countries. The consumption data reflected in Table Four gives aconservative picture of cassava’s trade prospects as it is skewed towardpoorer countries that use cassava as a staple food. Growth prospects forcassava are prevalent in middle-income developing countries that requirean alternative source of fuel and raw material feedstock to support the industrialisationof their economies. Therefore growth prospects for cassavaexist for its use as an industrial feedstock to produce starch and bio-fuels.Even though these markets are in their developmental stage, on a volumebasis, they have outpaced the consumption of cassava as food and feed(refer to Figure 3).An interesting observation is that the largest producers of cassava tendto be the largest consumers. Nine countries are among the ten largestconsumers and producers of cassava. The only two countries to buck this
TABLE 4: MAJOR CONSUMERS OF CASSAVA (‘000TONS) Table 4:Year Average Annual Growth Percentage of Total1990 1995 2004 1990-2004 2000-2004 1990 2004Democratic Republic of Congo 15,464 5,463 14,122 -0.65% -1.77% 19.18% 14.23%Indonesia 8,155 5,730 12,425 3.05% 1.25% 10.11% 12.52%Nigeria 8,236 3,327 12,3<strong>38</strong> 2.93% -2.80% 10.22% 12.43%Brazil 8,058 3,757 6,771 -1.24% -3.77% 9.99% 6.82%India 4,649 2,431 5,722 1.50% 0.02% 5.77% 5.77%Tanzania 5,886 1,463 5,122 -0.99% -2.60% 7.30% 5.16%Mozambique 3,598 1,803 4,758 2.02% 5.21% 4.46% 4.80%Ghana 1,949 551 4,528 6.21% 1.87% 2.42% 4.56%Angola 1,520 1,630 3,559 6.27% 5.06% 1.89% 3.59%Uganda 2,251 1,766 3,098 2.31% 7.24% 2.79% 3.12%Madagascar 1,726 1,313 2,005 1.07% -2.21% 2.14% 2.02%Thailand 513 1,448 1,989 10.16% 35.72% 0.64% 2.00%China 1,263 541 1,941 3.12% 3.97% 1.57% 1.96%Philippines 1,650 860 1,551 -0.44% -2.22% 2.05% 1.56%Colombia 1,245 300 1,546 1.56% 1.79% 1.54% 1.56%Côte d’Ivoire 1,254 244 1,330 0.42% -3.37% 1.56% 1.34%Guinea 334 707 1,202 9.58% 9.67% 0.41% 1.21%Benin 675 1,337 1,135 3.78% 3.41% 0.84% 1.14%Malawi 132 7<strong>38</strong> 1,095 16.34% -3.76% 0.16% 1.10%Rwanda 258 <strong>38</strong>5 1,002 10.19% 5.90% 0.32% 1.01%Top 20 Consumers 68,816 35,792 87,2<strong>38</strong> 1.71% 0.22% 85.35% 87.93%165<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Other Consumers 11812 54914 11980 0.10% -0.37% 14.65% 12.07%Total Consumption 80,628 90,706 99,218 1.49% 0.15% 100.00% 100.00%Source: FAOSTATtrend are Thailand and Uganda that only appear on the producers andconsumers list, respectively. Also most countries produce more cassavathan they consume. The important factor to establish is whether countries’production surplus is exported, which would create competition for SADCfarmers’ product. To answer this question, trade flows are analysed in thenext section.In 2004 SADC’s share of global consumption was 32%, a fall of 4%from its 1990 level of 36% (refer to Table 5). From 2000-2004 SADC’sconsumption of cassava declined by 0.26%, managing to fall below theworld’s annual average growth rate of 0.15%. SADC’s consumption profilewith respect to its absolute value and composition has remained relativelystatic over the period. This reflects cassava’s status as a substancecrop that is gown on marginal land. SADC’s poor performance should notbe viewed as negative but as an opportunity as it reflects a dearth of investmentand interest in an industry where SADC’s competitive advantagewith respect to land and labour has not been harnessed.8. Regional trade<strong>Trade</strong> in cassava comprises mostly pellets and chips for animal feed, whilethe remainder is for starch and flour for industrial use. <strong>Trade</strong> in fresh cassavais generally limited to exchanges between bordering countries dueto the product’s bulkiness and perishable nature. Although cold chainmanagement can improve a product’s shelf-life, it complicates logisticswhich increases transportation costs that cannot be passed onto the finalconsumer, unless the product is destined for a specialised marketCassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector StrategiesFIGURE 3: CONSUMPTION OF CASSAVA ON A PRODUCT BASIS (‘000 TONS)120,000100,0016680,000<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>60,00040,00020,000-1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004Food 80,628 87,370 89,527 89,340 89,625 90,706 91,9<strong>38</strong> 93,475 97,010 97,312 98.622 100,997 100,703 104,509 99,218Feed 57,557 53,023 53,282 50,549 48,806 40,233 37,144 <strong>38</strong>,264 37,344 45,173 46,606 52,573 47,186 50,931 50,044Other 24,710 30,929 33,059 33,748 34,062 <strong>38</strong>,483 37,245 28,823 37,511 39,632 43,123 40,252 49,072 52,700 50,291Source: FAOSTATTABLE 5: SADC’S CONSUMPTION OF CASSAVA (‘000TONS)Year Average Annual Growth Percentage of Total1990 1995 2004 1990-2004 2000-2004 1990 20.04DRC 15,464 5,463 14,122 -0.65% -1.77% 19.18% 14.23%Tanzania 5,886 5,730 5,122 -0.99% -2.60% 7.30% 5.16%Mozambique 3,598 3,327 4,758 2.02% 5.21% 4.46% 4.80%Angola 1,520 2,431 3,559 6.27% 5.06% 1.89% 3.59%Madagascar 1,726 1,803 2,005 1.07% -2.21% 2.14% 2.02%Malawi 132 300 1,095 16.34% -3.76% 0.16% 1.10%Zambia 608 707 902 2.85% 3.87% 0.75% 0.91%Zimbabwe 90 142 180 5.07% 2.05% 0.11% 0.18%South Africa - 3 0 98.95% 0.00% 0.00%Seychelles 0 0 0 2.48% 5.53% 0.00% 0.00%Mauritius 0 0 0 2.79% 4.46% 0.00% 0.00%SADC Consumption 29,025 29,826 31,744 0.64% -0.26% 36.00% 31.99%Other Consumers 51,604 60,880 67,474 1.93% 0.35% 64.00% 68.01%Total Consumption 80,628 90,706 99,218 1.49% 0.15% 100.00% 100.00%Source: FAOSTAT
From 1995-2005 trade in cassava was erratic. This is not surprising as tradein cassava is dominated by animal feed, whose prospects are affected bythe grain market’s behaviour. The grain market tends to be volatile as itis subject to government interventions. In addition cassava’s trade spikesare due to the fact that it is a thinly traded market, whose behaviour isdriven by five countries’ demand patterns and three countries’ supply capacity.For example, the major surge in 2003 is due to China’s increasingdemand, while the decline in 2001 is due to the EU’s falling demand.Regional trade in cassava has geographic and product specific dimensions.From Table 7 it becomes apparent that in 1995 the predominant regionalexporter was South East Asia with an 86% share of global exports andthe largest importer was East Asia with a 74% share of global imports.Another feature that is immediately apparent from glancing at the matrixis that trade occurs between trading blocks: East Asia and South East Asiaare trade partners, NAFTA and Central America are trade partners, andthe EU’s trade partners are South East Asia and Central America. <strong>Trade</strong>between the identified regional blocks has specific product dimensions,which is discussed in this section.8.1. Regional exportsIn 2005 South East Asia was the world’s dominant exporter of cassava;comprising 86% of the market (refer to Table 7). Although South EastAsia’s share of global exports has decreased since 1995, it is a marginaldecline of 3%. The drop in market share does not imply that the region’sproductive capacity is diminishing as it managed to grow its exports by3% from 2000-2005. This level of growth is below the global average andis the second lowest growth rate achieved by the top 5 exporting countries.Nevertheless South East Asia’s export growth is significant becauseit is off a large base.The region’s export growth was driven by Thailand, Vietnam and Indonesia.In 2005 intra-regional trade was negligible accounting for less than1% of the region’s exports. The region’s export market is geographicallyconcentrated and country specific. The majority of South East Asia’s exportsare destined for East Asia, in particular China and Korea. The region’stop three export markets comprise 90% of its trade. China is essentiallySouth East Asia’s export market, as it accounted for 79% of the region’sexports in 2005. In second and third place, respectively, were Korea andSpain that comprised 6% and 5% of South East Asia’s exports. These tradepatterns are not accidental. The Thai government has pursued a focusedexport strategy that spans the entire value chain, from selecting cultivarsthat have the best properties to produce a specific product to mappingthat product to a market. This case study illustrates that developing anexport strategy is more complicated than selecting a country / region toexport one’s product too; rather it involves taking activities throughout thevalue chain into consideration. SADC’s farmers can learn from Thailand’s167<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>FIGURE 4: TRADE IN CASSAVA 1995-2005 (US$’000)1,200,0001,000,000973,967800,000600,000400,000463,7<strong>38</strong>645,221400,<strong>38</strong>7321,524478,589398,335456,621335,219436,072623,635200,000-1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005Source: World Integrated <strong>Trade</strong> Solution (WITS)World <strong>Trade</strong>Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector StrategiesTABLE 6: REGIONAL TRADE MATRIX FOR 2005 (US$‘000):Exporting CountriesWorldImportsPercentage168<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Importing CountriesSouth East AsiaCentral AmericaEU25South AmericaNAFTASouth AsiaSADCMiddle EastEast Asia 459,244 - - 816 - - - - 460,070 73.8EU25 36,513 11,549 7,357 1,550 61 1 1 - 59,534 9.5NAFTA 645 43,228 2 3,480 13 1 - - 48,725 7.8Central America - 910 - - 2 - - - 913 0.1South America 1 - - 614 0 - - - 615 0.1South East Asia 324 - - 0 14 4 - - 445 0.1Middle East 1 - 3 - 0 0 0 8 14 0.0South Asia 0 - 0 - - 9 - 0 9 0.0SADC - - - - - - - - 3 0.0World Exports 533,926 66,173 7,853 7,790 201 18 11 8 623,635 100.0Percentage 85.62 10.61 1.26 1.25 0.03 0.00 0.00 0.00Source: World Integrated <strong>Trade</strong> Solution (Wits)experience as the country is also a developing country that faces similarconstraints with respect to small scale farmers’ access to resources. ThisTIB does not advocate that SADC’s farmers should copy their Thai counterpartsbut could use their experience to stimulate ideas about integratingactivities throughout a value chain to create an export strategy andmost importantly methods to include small scale farmers into this valuechain. Information about Thailand’s experience can be found at www.fao.org, under Global Cassava Strategy.Central America was the world’s second largest exporter of cassava in2005, managing to secure a 12% share of global exports, which is impressiveconsidering that in 1995 it had a 4% market share. Although theregion’s exports experienced strong growth over the decade its growthspurt occurred from 2000-2005, when the region’s exports experiencedphenomenal growth of 19%. The region’s export growth is driven byCosta Rica and Nicaragua. Intra-regional trade from 2000-2005 was minuscule,fluctuating between a low of 0.65% and a high of 1.65%. In2005 the region’s import partners were the US and the EU, in particularthe Netherlands and France. The US is the region’s single largest importer;it imports 64% of Central America’s exports. This region’s export successdemonstrates that exporting a specialised product, in this case fresh cassava,can be a profitable strategy. However if this strategy is pursued, aspectsof geography and importing into “cold-chain” hubs are important.It is not a coincidence that Costa Rica’s largest EU trading partner is theNetherlands as it has the infrastructure to distribute a perishable production,relatively quickly, throughout Europe.In 2005 the EU was the third largest exporter of cassava. Over the pastdecade the EU’s market share has declined from 7% in 1995 to 1% in2005. Growth rates indicate that the EU shed more of its market shareduring the second part of the decade from 2000-2005 as the EU’s exportscontracted by 23%. From 2000-2005 intra-regional trade ranged from94% to 99% of the region’s trade activity. In 2005 the region’s top sixexport destinations were Spain, the Netherlands, Belgium, Italy, Portugaland France, which comprised 90% of the region’s total exports. <strong>Trade</strong> activityis predominately concentrated within EU15 states. The EU predominatelyimports cassava in a pellet form and it is used as animal feed. Since2000 this market has followed a downward trend due to the BSE scare,falling domestic grain prices, strengthening of the Euro against the dollarand a change in the EU’s agricultural polices that made the relative priceof grain feed attractive. Spain’s imports of cassava reflect its growingdemand for industrial starch to support its food processing industry.An interesting observation is that Africa produces the majority of theworld’s cassava, but it is not classified as a major exporting region. Thisis due to the fact that cassava is grown as a subsistence crop for farmers’own usage as a staple food. In addition cassava’s physical attributes,especially the requirement to process the crop within days of post activity,
TABLE 7: REGIONAL EXPORTS OF CASSAVA (US$’000)Year Average Annual Growth Percentage of Total1995 2000 2005 1995-2005 2000-2005 1995 2005South East Asia 411,394 336,252 533,926 2.64% 9.69% 88.71% 85.62%Cental America 18,367 27,494 66,173 13.67% 19.20% 3.96% 10.61%EU25 32,517 28,988 7,853 -13.24% -22.99% 7.01% 1.26%South America 545 2,623 7,790 30.47% 24.32% 0.12% 1.25%NAFTA 82 434 201 9.41% -14.25% 0.02% 0.03%South Asia 3,212 126 18 -40.62% -32.62% 0.69% 0.00%SADC 1,874 139 11 -40.43% -40.26% 0.40% 0.00%Middle East 285 - 8 -29.65% 0.06% 0.00%WORLD 463,7<strong>38</strong> 398,335 623,635 3.01% 9.<strong>38</strong>% 100.00% 100.00%Source: Witsexacerbates Africa’s supply-chain bottlenecks. These supply side featuresinclude the availability of micro processing technology at the farm gate,farmers’ access to capital to purchase inputs and good quality transport.On the demand side, market’s for cassava products have not been developedas commercial interest in the product has been lacklustre due to itsimage as a poor man’s crop. Supply side bottlenecks coupled with limitedmarkets for the cassava based products created unfavourable conditionsfor the tradability of cassava products. Nigeria is an interesting case study(refer to the appendix) as one of the steps the government took to createEast Asia’s growth spurt increased its market share from 21% in 1995 to74% in 2005, toppling the EU from its 1995 dominant market position of88% in 1995 to 10% in 2005. Therefore East Asia’s growth spurt changedthe balance of power in the import market. East Asia’s demand for cassavais driven by China’s demand for livestock feed and starches. A relativelylarge importer in the region is Taiwan, but it is small compared toChina. Intra-regional trade is not significant. In 2005 86% of East Asia’simports were from South East Asia, in particular Thailand, Vietnam andIndonesia.a market was reducing “easy” supply side bottlenecks and then creatinga mass market for a simple processed product and then reinvestigatingthe industry’s supply side to address advanced issues.In 2005 the second largest import market for cassava was the EU witha 10% share of global exports. From 1995-2005 this market’s share ofimports has continued to decline, however the rate of its decline was more8.2. Regional importsFrom 1995 to 2005 the top three importers’ share of global imports andtheir relative ranking considerably changed (refer to Table 8). Furthermoreover this period, the top three regional importers increased their value ofimports, and also, the range of products that they imported. Over the pastdecade East Asia’s imports of cassava grew, however the majority of thisgrowth occurred after 2000. From 1995-2005 East Asia’s imports grewby 17% but between 2000-2005 imports increased by 54%. This growthspurt can be attributed to the interaction between the following factors:China’s rapid industrialisation, Thailand’s search for another export marketafter the collapse of its key export market (EU), and the impact of freetrade agreements, such as the ASEAN Free <strong>Trade</strong> Area and Thailand’s EarlyHarvest Agreement with China.pronounced during 2000-2005. Demand for cassava pellets to feed itslivestock industry has steadily decreased due to the EU’s agricultural policies,such as subsiding farmers’ cereal production, which made substitutegrain products more attractive and exchange rate movements. The EU’simport basket of cassava products can be divided into three submarkets.The animal pellet market is dominated by Thailand. The market for pelletsof flour and meal is imported from Coast Rica into the Netherlands(ITC, 2003). The Netherlands re-exports these products throughout Europe(ITC; 2003). Lastly, a growing market for food products made forhuman consumption which is dominated by Costa Rica (ITC, 2003). Inthe respective markets it would be very difficult for a country to challengethe market leaders’ position (ITC, 2003). As a result entering intodirect competition with the respective market leaders by selling a similarproduct at similar price could start a price war. SADC’s farmers/ producerswould probably not win this war as the do not have access to establishednetworks. Therefore SADC farmers/ producers’ ability to enter this mar-169<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies170<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>ket would be based on creating innovative, processed food products andmarketing them to distribution channels that serve specialised retailers,such as health stores, ethnic cuisine caterers, and food outlets catering toimmigrant populations.In 2005 the third largest import market was NAFTA with an 8% marketshare, which is considerably better than its 3% market share in 1995. Theregion’s primary importer is the US comprising 97% of the region’s imports.According to the FAO (2004), the majority of cassava imported intothe US is used for its livestock industry. The next largest user of importedcassava is for industrial applications in the form of starches and the reminderis consumed food. Although cassava used for human consumptionis the smallest market, it is the fastest growing sub- sector. This market’sgrowth rate in value of 15% was largely driven by the US’ demand for cassavafor human consumption to produce starches and ethnic cuisine forits immigrant population. The region’s preferential supplier is Costa Rica,which comprised 88% of its imports in 2005.9. Country trade9.1. Countries’ importsGlobal imports have grown at a steady rate of 9% per annum from 2000-2005. However this figure hides the fluctuation and variation in growthbetween import markets. This is an important point for potential exportersto realise, as an exporter’s ability to choose a “growing” import partnerwill determine his/her success. Even within the top 10 importing countriesa wide variation exists between markets’ prospects: China’s imports grewby 80% from 2000-2005, while The Netherlands’ imports declined by37%. In 2005 four countries (China, the US, Korea and Spain) comprised87% of global imports. Given these countries’ dominance of the market,they effectively are the market (refer to Figure 5.Compared to the export market, the four largest importers are moregeographically dispersed. On logical grounds this is to be expected as“creating artificial” climatic conditions to grow cassava is not economicaland thus production is tied to areas that have suitable climatic conditions.The geographic dispersion of large import partners provides morenodes for potential exporters to enter the cassava market. As mentionedin the supply chain section, the manner in which cassava is processedaffects its perishability and its weight, which has a host of transportationimplications. This in turn affects exporters’ logistical arrangements andalso shapes their decision regarding which product to supply and whichpartner to select. For example, Costa Rica supplies fresh cassava to theUS; however this strategy would probably not be an effective one forThailand to adopt.From 1995-2005 the top 7 importers’ share of global imports changedsubstantially. China’s share of the import market increased from 15% n1995 to 67% 2005, while Spain, The Netherlands and Portugal’s share ofthe import market fell 16%, 33% and 9%, respectively. The majority of theEU’ s imports were absorbed by China, as a result the geographic locationof cassava’s demand base shifted to East Asia. The European Marketpredominately imported cassava pellets for animal feed, while the Asiancountries basket of imported cassava products is more diverse, includingpellets and industrial starches.TABLE 8: REGIONAL IMPORTS OF CASSAVA (US$’000)8Year Average Annual Growth Percentage of Total1995 2000 2005 1995-2005 2000-2005 1995 2000East Asia 97,809 52,251 460,070 16.75% 54.51% 21.09% 73.77%EU25 408,328 305,709 59,534 -17.51% -27.91% 88.05% 9.55%NAFTA 15,807 23,7<strong>38</strong> 48,725 11.92% 15.47% 3.41% 7.81%Central America 34 297 913 <strong>38</strong>.92% 25.17% 0.01% 0.15%South America 307 2,060 615 7.19% -21.48% 0.07% 0.10%South East Asia 289 368 445 4.39% 3.84% 0.06% 0.07%Middle East 87 142 14 -16.87% -37.31% 0.02% 0.00%South Asia - 7 9 6.03% 0.00% 0.00%SADC 4 55 3 -2.54% -43.84% 0.00% 0.00%World 463,7<strong>38</strong> 398,335 623,635 3.01% 9.<strong>38</strong>% 100.00% 100.00%Source: Wits
FIGURE 5: TOP 10 IMPORTERS OF CASSAVA IN 2005 (US$’000)450,000400,000420,82680.33%100.00%80.00%350,00060.00%300,000250,000200,000150,000100,00050,000-171<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>ChinaUnited StatesKorea, Rep.SpainThe NetherlandsPortugalFranceJapanBelgiumUnited KingdomUS$ ‘00015.50%13.60%10.39%4.42%-14.40% -11.58%-21.94%47,406-37.01%34,785 28,764-43.63 %10,581 10,168 3,862 3,539 2,907 2,03640.00%20.00%0.00%-20.00%-40.00%-60.00%Average Annual Growth(%)Imports Average annual growth rate (%)Source: WITSIn 2005 China was the world’s dominant importer of cassava, comprising67% of global imports. More impressively, given China’s large import volumesfrom 2000-2005, it managed to grow its imports by 80%, on an averageannual basis. China’s surging demand for cassava products fuelledthe global import market’s 9% growth rate. A few factors contributed tothe growth in China’s demand for cassava products. First, China reducedits import duty from 30% to between 7% and 11.2% in preparation forits accession to the WTO in December 2001 (ITC.2003:11). Second, China’srapid industrialisation created demand for industrial feed-stocks toproduce ethanol and starches. Third, rising per capita income, urbanisationand the growth of the middle class increased the population’s consumptionof meat.China’s imports its cassava from Thailand, Vietnam and Indonesia. Giventhe low value of cassava products, it is uneconomical to transport theseproducts long distances and thus geography influences trade patterns.This is illustrated in China’s decision to import cassava products fromSouth East Asia instead of South America.From 1997-2000 both Thailand and Indonesia were vying to becomeChina’s dominant supplier, Thailand won the battle. Thailand was aggressivelylooking for new markets to reduce its dependency on the EU. In2000 the government launched a purchasing programme to support producerprices resulting in the Thai Public Warehouse Organisation holdingstockpiles of cassava. As a result Thailand had the capacity to meet China’sunexpected surge in demand. Vietnam’s cassava industry also grewon the back of China’s increased demand, Indonesia’s industry failed tocapture benefits from China’s growth phase. This scenario illustrates thatmarket timing is a crucial factor determining an exporter’s potential success.SADC’s entrepreneurs, producers, and policy-makers could apply thislesson to entering into the ethanol market, which has the potential to beextremely lucrative.China’s imports comprise primarily of dried chips and pellets, used foranimal feed, however its trade is weighted in favour of pellets (ITC; 2003).China imports about 60% of its chips to produce alcohol from Thailandand 11% from Vietnam (TTTA, 2004). China imports 40-50% of its starchto produce sweeteners and MSG from Thailand and 20-30% from Vietnam(TTTA, 2004).<strong>Trade</strong> data shows that China’s demand for cassava has grown exponentially.However this is past behaviour: Is China’s growth sustainableCassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector StrategiesTABLE 9: TOP IMPORTERS OF CASSAVA (US$’000)172<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong><strong>Trade</strong> (US$’000)AverageYearsAnnualGrowth Percentage of Total Uses (‘000Tons)1995 2000 2005 2000-2005 1995 2005 2000 2004China 67,680 22,065 420,826 80.33% 14.59% 67.48% Feed:42% Feed:62%United States 15,062 23,064 47,406 15.50% 3.25% 7.60% Feed:59% Feed:61%Korea, Rep. 27,261 28,015 34,785 4.42% 5.88% 5.58% Food:98% Feed:93%Spain 96,277 99,275 28,764 -21.94% 20.76% 4.61% Feed:100% Feed:99%The Netherlands 160,257 106,692 10,581 -37.01% 34.56% 1.70% Feed:100% Feed:100%Portugal 49,321 22,130 10,168 -14.40% 10.64% 1.63% Feed:100% Feed:99%France 11,029 7,146 3,862 -11.58% 2.<strong>38</strong>% 0.62% Feed:91% Starch:60%Japan 2,855 2,159 3,539 10.39% 0.62% 0.57% Food :100% Food :100%Belgium - 51,076 2,907 -43.63% 0.00% 0.47% Feed:100% Feed:100%United Kingdom 944 1,076 2,036 13.60% 0.20% 0.33% Starch:50% Starch:58%Canada 744 671 1,319 14.48% 0.16% 0.21% Starch:80% Starch:68%Taiwan, China - 12 920 1<strong>38</strong>.81% 0.00% 0.15%Italy 12,370 4,000 917 -25.51% 2.67% 0.15% Feed:98% Feed:96%Australia 329 445 732 10.45% 0.07% 0.12% Starch:90% Starch:88%Honduras - 83 654 50.98% 0.00% 0.10% Food:94% Food:75%New Zealand 43 296 528 12.29% 0.01% 0.08% Food:97% Starch:60%Switzerland 248 260 487 13.35% 0.05% 0.08% Starch:85% Feed:81%Colombia 288 1,519 <strong>38</strong>4 -24.03% 0.06% 0.06% Food:79% Food:79%Singapore 195 291 309 1.19% 0.04% 0.05%Iceland 1 2 220 156.<strong>38</strong>% 0.00% 0.04% Starch:95% Feed:92%Total Top 20 Imports 444,905 370,277 571,343 9.06% 95.94% 91.61%Other Importers 18,833 28,058 52,292 13.26% 4.06% 8.39%World Imports 463,7<strong>38</strong> 398,335 623,635 9.<strong>38</strong>% 100.00% 100.00%Source: Wits and FAOSTATover the long-term? Based on industry reports, the author asserts thatChina’s growth in demand is sustainable, provided demand is driven bya new sub-segment, which in this case is the demand for bio-fuels. China’sdemand for fuel is increasing at an increasing rate (refer toError!Reference source not found.Fuel is required to power its industrialisationdrive and rising per capita income has stimulated consumers’ demandto own a car. Over the past two decades, China was the fastest growingautomobile market in the world. From 1986-2004 growth in car ownershipwas11.8% and by 2004 the number of cars reached 26.94 million(refer to Figure 7) (Latner & O’Kray & Jiang, 2006:3). This growth shouldcontinue, and even increase, as China continues to industrialise. Over themedium term the demand for bio-fuels should grow as the economy’sdemand for fuel increases based on the government’s policy that 15%of China’s transportation energy needs must be supplied by bio-fuels by2020 (Latner & O’Kray & Jiang, 2006:3). The Chinese government tabledlegislation requiring consumers to use gasohol, which is 10-20% ethanolmixed with gasoline, as automotive fuel by 2008 (Howeler, 2003:21). Thispiece of regulation should increase the demand for cassava chips (Howeler,2003:21).The Ethanol industry’s future growth is dependant on government supportin the form of subsidies or obligatory usage rates as a result potential investor’sor suppliers to this industry are exposed to political risk. The onlyissue that could impact one’s exposure to political risk is the government’splan to change the manner in which it supports the ethanol industry. Subsidieswill be phased out, but state municipalities will be mandated to useethanol and municipalities will be given government grants to constructethanol production facilities (Latner & O’Kray & Jiang, 2006). This changeaffects the manner in which the government allocates funds to financeethanol plants, but it does affect the demand for ethanol. As a conse-
FIGURE 6: GASOLINE AND DIESEL CONSUMPTION IN CHINA FROM 1980-200414,00012,00011,59410,000Unit: 10,000 kl8,0006,0004,0002,00005,8683,0061,8592,1732,3431,7441,24580 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04173<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: Latner & O’Kray & Jiang; 2006:7Gasoline ConsumptionDiesel ConsumptionFIGURE 7: NUMBER OF VEHICLES IN CHINA AND THEIR RATE OF GROWTH FROM 1986-20043,00018.22018Unit: Car(s)2,5002,0001,5001,00050012.713.810.17.89.914.115.210.45.810.88.210.110.712.013.916.113.116141210864Unit: %2-86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04Number of Owned AutomobilesAnnual increaseSource: Latner & O’Kray & Jiang; 2006:70quence China’s production of ethanol should continue to increase passedits 2004 level of 3.7 billion litres of ethanol, making it the third largestglobal producer (Brant &Fang &Lin, 2006:5).China produces ethanol from corn and wheat; however its primary feedstockis corn, for example, in 2006 90% of China’s ethanol was producedfrom corn. Corn has become a “precious” commodity as China’s economicgrowth increases the demand for grain and simultaneously reduces supply.Industrialisation increases urbanisation which changes a population’sfood consumption patterns toward convenience food and increased consumptionof meat products. This led to the expansion of the livestock industrywhich uses corm as feed. In China corn is used as animal feed.However due to China’s industrialisation and urbanisation both land andlabour allocated to agricultural activities is falling. Thus a situation has de-Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies174<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>veloped where China’s consumption of corn is increasing at a faster ratethan its can increase it productive capacity, as a result from 2004 Chinabecome a net importer of grain (Brant et al, 2006). The Chinese governmentis concerned that being a net importer of grain could jeopardiseChina’s long term goal of food security and food self-sufficiency, placing itin a precarious situation where it is reliant on foreign resources to achievefood security (Brant et al, 2006). The government’ realises that its ethanolpolicy will increase the demand for ethanol to comprise a greater proportionof China’s energy mix, which is growing at an increasing rate. As aresult the consumption of ethanol is given a double boost which will haverepercussions on the consumption of gain. If China does not diversify itsfeedstock it is predicted that by 2010 corn consumption could increaseby 25% which would require China to import 10 million tons of corn ayear to meet demand (Latner & O’Kray & Jiang, 2006:17). Also highercorn prices will reduce the competitiveness of China’s growing food industry.These considerations caused the government to revisit the mannerin which its bio-fuels industry will be developed in the future.feedstock for China’s ethanol industry. First, cassava is a cheaper optionthan grain, producing ethanol from cassava costs approximately $500/MT(4,000 Yuan/MT compared to 563/MT (4,500 Yuan/MT) for stale grain, asa result cassava is a cheaper option (Latner & O’Kray & Jiang, 2006:9).Second, the waste pulp from the production of cassava starch is usedto make ethanol. It is more cost effective to use a by-product than todiscard it (Howeler, 2003:21). Third, cassava is a versatile crop, it can beprocessed in form of fresh roots during the harvest season or dried chipand extracted starch when fresh roots are out of season. Lastly, stricterpollution regulations make the use of molasses uneconomical resultingin energy companies switching their feedstock from molasses to cassava(Howeler, 2003:21). The government’s intent to use cassava as a feedstockto produce ethanol is demonstrated by the construction of a productionfacility capable of producing one million MT of fuel ethanol by 2010 in theGuangxi Zhuang Autonomous Region in southern China. This productionfacility is scheduled to begin operations in October 2007 at a productioncapacity of 110,000 MT per year (Latner & O’Kray & Jiang, 2006:20).According to Yang Jian, the director of the development planning departmentunder the Agriculture Ministry, the Government’s policy regardingbio-fuels is that it “should neither impact the people’s grain consumption,nor should it compete with grain crops for cultivated land (Pattern, 2006).A short-term solution to the pending problem led to China’s ministry ofadministration publishing a warning in December 2006 that only fourcompanies had permission to produce corn-based fuel (Pattern, 2006).This short-term solution is not viable over the long-term because Chinarequires fuel to support its industrialisation drive. Ethanol productionin 2005 was approximately 920,000 metric tons (MT), with a productioncapacity of 1,020,000 metric tons. The government predicts thatChina’s ethanol production capabilities should increase to 4 million MTp.a by 2010 and to 15 million MT (Pattern,2006 and Latner & O’Kray &Jiang,2006:3).If China is to meet the demand for ethanol without relegating its food securityneeds to second position, an alternative feedstock must be sought.This sentiment is echoed by Chinese officials, “the development of biofuelshouldn’t be at the expense of the expansion of farmland, since food isstill the priority of China. We should put more attention on sweet potatoand cassava that are rich in starch and suitable for planting in Chinabased on its terrain” (Pattern,2006). Sugarcane was considered but thenruled out: China’s agricultural conditions are not suitable to grow thiscrop, industrial demand for sugar is increasing due to China’s expandingfood-processing industry and the sugar price is volatile (Brant et al2006). After considering various options, cassava was selected as the newIt is expected that over the medium term a greater proportion of cassavawill be used to manufacture ethanol and the demand for ethanol will increase.China’s economic growth is energy intensive and it is expected toincrease and China’s energy policy is being reformulated to use bio-fuelsthat are not produced from grains. Both these factors should increase thedemand for cassava at a faster rate than China can domestically supplycassava, opening up a market for imports. Even if China grew cassavaon 2.471 million acres its barren land, it production capabilities couldincrease from 13.3 million MT in 2006 to 34.3 million MT p.a, and employedthe latest technology to increase its yields by 7 million MT pa,its production capacity would fall short of demand (Latner & O’Kray &Jiang, 2006:20). Despite these supply side measures China will need toimport cassava. Not only does this provide an opportunity for SADC’sfarmers to export cassava, but it also could expose SADC’s farmers to newforms of business arrangements that simplify their operations throughoutthe value-chain. For example, Henan’s Tian Guan Group entered into acontract with the government of Laos to lease 15 square km of land toproduce cassava (Latner & O’Kray & Jiang, 2006:20).In 2005 the US was the second largest importer of cassava with an 8%share of the market, which is minuscule compared to China. However an8% market share is impressive considering that in 1995 the US’ shareof the market was only 3%. This market has also experienced steady,strong growth. The growth rate of imports between 1995-2005 of 12%and 2000-2005 of 16% was within a tight range; however the marketperformed slightly better since 2002. It should be noted that on a volume
TABLE 10: TOP IMPORTERS’ LARGEST TRADING PARTNERS FOR 2005 (BASED ON A PERCENTAGE OF TOTAL IMPORTS)ExportersImporter First Second ThirdChina Thailand 81.09% Vietnam 11.97% Indonesia 6.94%United States Costa Rica 87.82% Ecuador 5.86% Ghana 1.41%Republic Korea Vietnam 51.91% Thailand 28.80% Indonesia 16.76%Spain Thailand 87.71% Costa Rica 5.03% The Netherlands 4.66%The Netherlands Costa Rica 43.65% Thailand 18.11% France 14.08%Portugal Thailand 91.64% France 5.30% Spain 1.80%France Costa Rica 52.88% Cameroon 18.21% Belgium 6.61%Japan Thailand 96.91% Brazil 0.00% Indonesia 0.96%Belgium Costa Rica 59.61% The Netherlands 34.66% Ghana 2.13%United Kingdom Costa Rica 72.35% Ecuador 12.93% Ghana 3.45%Source: Wits175<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>basis, the US’ primarily uses cassava in the form of feed, but with respectto value, the US’ largest market is the one for human consumption. Giventhat trade is discussed on a value basis the US’ imports are primarily fromCosta Rica, which specialises in supplying cassava for human consumption.Demand for cassava is not broadly based throughout the populationbut is driven by the Hispanic and Asian population (ITC, 2003). Givencassava’s characteristics this would not be a lucrative market for SADC’sfarmers. However SADC’s farmers could use this market as a case study.It shows that specialising and exporting a “niche” product to a targetmarket could be a more profitable strategy than competing against lowcost producers to supply a commodity based product to the Asian market.The question SADC’s farmers could ask is where does the next profitablegeographically accessible market for a niche product exist.In 2005 Korea was the third largest importer of cassava with a marketshare of 5%. The demand for cassava in Korea seems to have stagnated.Korea’s share of the global import market has remained unchanged since1995 compared to 2005, even though its economy has grown and thusthe demand for a feedstock should have increased. This lacklustre performanceis reflected in the rate of growth of imports of 5% from 2000-2005. Since 1996 Korea’s imports of pellets has followed a downwardtrend. The expansion of Korea’s livestock industry outpaced the productionof livestock feeds; as a result the shortfall was imported. Korea’s cassavaimports are limited to the import of dried cassava in the form of chipsand pellets. Chips are imported from Vietnam and pellets are importedfrom Thailand. These markets are not contested are but dominated byboth parties.Based on trade statistics the following countries are potential emergingimporters and could serve as a potential market: Canada, Australia, NewZealand, Switzerland, Iceland, Brazil (on a volume basis the feed industryuses 49% of total cassava) and Malaysia (on a volume basis 52% of totalcassava is used by industry to produce starches).9.2. Countries’ exportsThe export market for cassava experienced steady growth of 9.<strong>38</strong>% from2000-2005 (refer to Table 11). Most of this growth was driven by theTop 10 exporters (refer toFigure 8 Top 10 Exporters of Cassava in 2005(US$’000) Figure 8). In 2005 Thailand, Vietnam, Coast Rica and Indonesiacomprised 96% of global exports and thus these four countries effectivelyconstitute the global export market. A notable feature is that the exportmarket is dominated by Asian producers: In 2005 they comprised 85%of global exports and occupied three out positions in the top 4 list ofexporters.Over the past decade, from 1995 to 2005, the top 5 exporting countries’market share and their relative rankings changed. Thailand maintainedits position as the world’s dominant exporter but lost market share. From1995 to 2000 Vietnam and Costa Rica substantially increased their shareof the export market. It appears that these two countries absorbed Indonesia’sshare of the market, which significantly declined during 1995-2005. An interesting observation is that Vietnam and Costa Rica pursueda different export strategy. Costa Rica exports cassava for human consumptionto developed countries. Products are tailored towards nichemarkets, such as health stores and speciality food stores that sell ethnicCassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies176<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>food. In contrast Vietnam exports cassava pellets to China to be used asanimal feed for its livestock industry. These antidotes illustrate that to bea successful exporter, it is important to specialise in a product and mapit to a country’s demand profile instead of trying to export every productto every country. The above principle is an important point that SADC’sfarmers should note.In 2005 Thailand was the world’s dominant exporter of cassava, comprising69% of the export market. It managed to grow its exports from2000-2005 by 8%, which is impressive, even though it is substantiallybelow Vietnam and Costa Rica’s growth rate because Vietnam’s exportgrowth is off a large base. One of the reasons behind Thailand’s successis its ability to deliver better quality products, on a more consistent basis,than other producers, especially its African counterparts, who are plaguedby adverse weather conditions, various crop diseases and civil unrest. Thaifarmers have ample support from the government throughout the supplychain, ranging from technical to financial assistance. The government’srole in supporting its cassava industry ha contributed to Thailand beingthe preferred supplier to the top four global importing counties: China,Spain and Korea. Thailand’s trade partners import cassava primarily fortheir livestock industry and also for industrial applications; as a resultThailand’s largest export product is dry cassava products. Thailand wasthe world’s third largest producer in 2004 and exports its own production,and thus does not rely on re-exports to service demand.In 2005 Vietnam was the world’s second largest exporter, but it trails behindThailand, with 11% of the global export market. Vietnam’s exportsmaybe smaller than Thailand’s but Vietnam grew its export base at an exponentialrate, from 2000-2005 its exports grow by 22%. Vietnam’s tradingpartners are China, Korea and Australia. Similar to Thailand, Vietnam’sdominant trading partner is China; this is largely due to the benefits ofeconomic geography. It should be noted that Vietnam’s strategy to placereliance on a single importer could be a risky long term strategy.Costa Rica was the third largest exporter of cassava in 2005, with a marketshare of 10%. Costa Rica is an interesting case study because it iscompeting with the other top 4 exporters with respect to export volumesand growth rates but is follows a different export strategy with respect tothe product it exports and markets it pursues. It exports frozen or waxedroots for human consumption to the US and the EU, mainly to the Netherlandsand France.FIGURE 8: TRADE IN CASSAVA 1995-2005 (US$’000)500,00060.00%450,000428,86651.04%50.00%400,000350,000300,000250,000200,000150,000100,00050,000-ThailandVietnamCosta RicaIndonesiaEcuadorThe NetherlandsGhanaFranceCameroonBrazilUS$ ‘0008.22%22.41%18.87%11.44%15.<strong>38</strong>%40.41%5.42%<strong>38</strong>.80%40.00%30.00%20.00%10.00%0.00%Average Annual Growth(%)68,741 64,79935,126-23.26%4,<strong>38</strong>2 4,171 3,518 1,989 1,469 1,399-10.00%-20.00%-30.00%Source: WITSExports Average annual growth 2000-2005
TABLE 11: TOP EXPORTERS OF CASSAVA (US$’000)YearAverage AveragePercentage of Total1995 2000 2005Growth 00-051995 2005Thailand 330,703 288,988 428,866 8.22% 71.31% 68.77%Vietnam 14,155 25,008 68,741 22.41% 3.05% 11.02%Costa Rica 18,200 27,301 64,799 18.87% 3.92% 10.39%Indonesia 65,115 20,435 35,126 11.44% 14.04% 5.63%Ecuador 54 2,143 4,<strong>38</strong>2 15.<strong>38</strong>% 0.01% 0.70%The Netherlands 1,028 15,674 4,171 -23.26% 0.22% 0.67%Ghana 873 645 3,518 40.41% 0.19% 0.56%France 158 1,527 1,989 5.42% 0.03% 0.32%Cameroon 8 187 1,469 51.04% 0.00% 0.24%Brazil 46 272 1,399 <strong>38</strong>.80% 0.01% 0.22%Nicaragua 121 101 984 57.56% 0.03% 0.16%Fiji 271 450 900 14.86% 0.06% 0.14%Venezuela 266 - 816 0.06% 0.13%Belgium - 2,315 678 -21.78% 0.00% 0.11%Philippines 1,205 453 674 8.29% 0.26% 0.11%Suriname 142 3 565 186.50% 0.03% 0.09%Malaysia 202 371 448 3.84% 0.04% 0.07%Spain 13 452 373 -3.76% 0.00% 0.06%Colombia 22 139 364 21.22% 0.00% 0.06%Tonga 39 157 350 17.42% 0.01% 0.06%Top 20 Exporters’ Total 432,622 <strong>38</strong>6,621 620,611 9.93% 93.29% 99.52%Other Exporters 31,116 11,715 3,024 -23.73% 6.71% 0.48%Total Exports 463,7<strong>38</strong> 398,335 623,635 9.<strong>38</strong>% 100.00% 100.00%Source: Wits177<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>TABLE 12: TOP EXPORTERS’ LARGEST TRADING PARTNERS FOR 2005 (BASED ON A PERCENTAGE OF TOTAL EMPORTS)Import MarketsExporters First Second ThirdThailand China 79.57% Spain 5.88% Korea, Rep. 58.06%Vietnam China 73.28% Korea, Rep. 26.27% Australia 0.17%Costa Rica United States 64.25% The Netherlands 7.13% France 3.15%Indonesia China 83.13% Korea, Rep. 19.97% Romania 0.93%Ecador United States 63.35% Colombia 8.75% European Union 12.75%The Netherands Spain 32.14% Belgium 24.16% Italy 16.44%Ghana European Union <strong>38</strong>.96% The Netherlands 35.91% United States 19.03%France The Netherlands 74.89% Spain 9.82% Portugal 3.68%Cameroon European Union 48.91% France 47.86% Switzerland 1.46%Brazil European Union 32.27% The Netherlands 26.56% France 10.76%Source: WitsCassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector StrategiesTABLE 13: SADC’S USAGE OF CASSAVA IN 2004 (’000 TONS)178<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Domestic SupplyDomestic UtilisationProduce Import Export Feed & Seed Other FoodAngola 6,650 1 0 1,300 1,792 3,559Botswana - 0 - - 0 -Congo, Dem Republic of 14,951 2 - 359 471 14,122Madagascar 2,191 1 0 84 104 2,005Malawi 2,559 1 - <strong>38</strong>5 1,080 1,095Mauritius 0 2 0 - 2 0Mozambique 6,413 0 0 369 1,287 4,758Namibia - 1 - - 0 1Seychelles 0 0 - - - 0South Africa - 76 0 - 75 0Swaziland - 0 0 - 0 -Tanzania 6,890 4 2 58 1,711 5,122Zambia 957 0 1 - 54 902Zimbabwe 190 1 0 - 10 180Total SADC 40,801 88 4 2,555 6,586 31,744Total World 202,879 20,203 18,930 50,044 54,890 99,218SADC’s Share of Total 20.11% 0.44% 0.02% 5.11% 12.00% 31.99%TABLE 14: SADC’S TRADE WITH THE WORL FROM 2000-2004 (US$)Imports (US$)2000 2001 2002 2003 2004Botswana 3,076 27,259 1,958 19,560Lesotho 2,761 146 9,263 7Maurituis 5,748 7,5<strong>38</strong> 9,127 10,865Mozambique 161South Africa 371 188 162 <strong>38</strong>0,716 1,451Swaziland 370 186 17 256 3,875Tanzania 1,131Zambia 584 57 1,164 1,347Total SADC 4,247 9,401 44,239 394,359 37,098Exports (US$)2000 2001 2002 2003 2004Botswana 82,603Lesotho 16,374 30,155Maurituis 58 177South Africa 203 1,961 708 9,077 7,906Swaziland 180 581 46Tanzania 17,052 147,328 1,516 7,796 1,048Zambia 40 174 45Total SADC 33,809 149,968 115,333 16,919 8,999Source: SADC <strong>Trade</strong> Data Base, <strong>Trade</strong> and <strong>Industry</strong> Strategies
FIGURE 9: REGIONAL EXPORTS IN 2005 (US$’000)700,00030.00%US$ ‘000600,000500,000400,000300,000200,000100,000-533,926SouthEast Asia9.69%66,173CentralAmerica19.20%EU25-22.99%SouthAmerica24.32%NAFTA-14.25%SouthAsia-32.62%SADC-40.26%7,853 7,790 201 18 11 8MiddleEast623,6359.<strong>38</strong>%World20.00%10.00%0.00%-10.00%-20.00%-30.00%-40.00%-50.00%Average Annual Growth(%)179<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Source: WITS<strong>Trade</strong> Average annual growth 2000-2005Based on trade statistics the following countries are potential emergingexporters and thus could be SADC’s farmers’ competition: Nicaragua, Colombia,Paraguay and Nigeria.The largest exporters of cassava are not the largest producers. In 2005only Thailand, Indonesia and Ghana managed to occupy a place in thetop 10 producing countries and exporting countries. Considering Africancountries are large producers of cassava this indicates that Africa is nottaking advantage of its productive capacity.10. SADC trade10.1. <strong>Trade</strong> with the worldThree out of the world’s top ten producing countries are in the SADCregion. These countries are the DRC, Tanzania and Angola, and they areranked in fifth, seventh and eighth positions, respectively. SADC’s importsare negligible, which is understandable, given its productive capacity withrespect to climate, land and labour. An area for improvement is SADC’s exportingcapacity. SADC accounts for less than 1% of global exports, yet itsproduction comprises 20% of global output. This indicates that cassava’spotential as a cash crop is not being exploited due to supply and demandside bottlenecks. Given the crop’s perishable nature, subsistence farmerson marginal land grow an adequate amount to consume as a staple food,and thus their yields are low. Marginal areas tend to be underdeveloped;as a result farmers’ access to supply-side infrastructure and their resourcesto engage in marketing activities are limited. These supply-side rigiditiesmake it difficult for farmers to export their crops. Therefore improvingSADC’s ability to export cassava will require sequenced, supply-side anddemand-side initiatives.SADC is a net exporter of Cassava to the world (refer to Table 14). Overthe period under review, from 1995-2005, SADC’s participation in globalimport and export markets was poor. A worrying sign is that the region’sparticipation in global markets deteriorated from 2000-2005 when tradein cassava entered into its growth phase (refer to Figure 9). The world’strade in cassava grew by 9% from 2000-2005, while SADC’s exports declinedby 40% over the period. This suggests that SADC has been lockedout of trade and was unable to tap into growing markets. SADC countries’import partners only include two of the top 10 importing countries. Thesecountries are Portugal and France that are ranked in sixth and seventhposition, respectively, whose combined share of global exports is 4%. Malawiexported cassava to Portugal. France imported cassava from Madagascarand the DRC.Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies180<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>10.2. Intra-SADC tradeIntra-trade between SADC states is negligible and appears to be random.This is in line with expectations as SADC countries grow cassava as a subsistencecrop. Table 15 was included in the TIB for illustrative purposes, asit highlights that trade between member countries is sporadic.11. Key countries’ propensity totrade in cassavaThe differential between production and consumption is due to the factthat cassava does not have a set harvest period (refer to Figure 10). Astriking feature of this market is that only a small percentage of cassavaproduced is traded. Only one of the top 10 producing countries, Thailand,also appears on the list of top 10 exporters. This reflects the crop’s statusas a lowly subsistence crop grown on marginal land by small-scale farmersin lower-income developing countries. The low level of trade in cassavais not entirely a demand issue but also a supply issue. Raw cassava is notan export friendly product as it is bulky and perishable. The majority ofcassava is produced in poor countries, whose infrastructure is poor andaccess to resources to create a market is limited. Also the majority ofproduction is required to satisfy domestic consumption needs. Both thesefactors have made cassava a relatively obscure “tradable” product. Anotherissue that has also affected cassava’s tradability is a lack of scientificresearch into the plant’s properties and its industrial applications. This isnot surprising as the dominant producers of cassava are poor countriesthat do not have the resources to conduct research.The market for cassava is on the cusp of entering into a new phase. Overthe long-term, a four tiered market for cassava products should develop.The first market will be for domestic consumption as a staple productand cassava will be grown and consumed in lower-income developingcountries. International trade in this market will be thin, as the product iscomplicated to transport and has a low value.The second market will be for animal feed that will be consumed in middle-incomedeveloping countries. Economic development and urbanisationhas led to the growth of the middle class, whose diet comprises more livestockproducts than their rural counterparts. Changes in social-economicconditions have fuelled the livestock industry’s expansion, which increasesthe demand for feed. Another factor that will affect this market’s growth isthe price of complementary and substitute products. Grain prices reachedtheir highest level in seven years and this trend is expected to continueas the demand for gain increases due to the production of bio-fuels. Cassavafeed is cheaper than its grain based substitute products and as it isa commodity product relative prices should impact consumers’ behaviour.The popularity of cassava feed should increase and trade should growat a steady rate. Cassava animal feed will be grown and processed inmiddle-income developing countries, such as Thailand, Vietnam and Indonesia,as exporting cassava requires infrastructure, marketing and distributionchannels. Traditionally cassava feed is considered an entry levelvalue-added product, and as such would provide SADC’s farmers withan opportunity to produce value-added products. Also marketing cassavafeed in SADC would not be difficult as the livestock industry is expandingand the price of grain base products is becoming prohibitively expensivefor farmers to feed livestock. This product is simple to transport which isimportant as infrastructure in SADC is poor. As a result this product haspotential for intra-regional trade.The third market is the one for industrial applications, such as starchesand the production of ethanol. These products will be produced in uppermiddle-income countries, and consumed in upper middle income countriesand developed countries. <strong>Trade</strong> in this market is projected to growat an increasing rate. The implication is that SADC’s farmers should ultimatelyintegrate their operations into supply chains to create industrialproducts.The fourth market is for premium quality cassava for human consumptionin developed markets. This market’s degree of tradability is dependant onTABLE 15: INTRA SADC TRADE IN 2004 (US$)ImportersAngola Botswana Mozambique South Africa ZimbabweExportersBotswana 14DRCNamibia 179South Africa 7,178 16,139 728 7,Zambia 45Source: SADC <strong>Trade</strong> Data Base, <strong>Trade</strong> and <strong>Industry</strong> Strategies
FIGURE 10: MARKET TRENDS IN CASSAVA (‘000 TONS)250,000200,000150,000100,00050,000181<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>-1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004Source: FAOSTATTotal consumption Total production Total importwhether consumers’ demand fresh or processed cassava, in the form ofconvenience foods. <strong>Trade</strong> in fresh cassava has a geographical dimensionand requires complicated cold chain management. This is an area whereSADC countries could share infrastructure across countries and products.For example South Africa has developed chain management to exportcut flowers. This infrastructure could be used by cassava exports, especiallyconsidering that Rotterdam is a hub for fresh flower imports and alarge re-exporter of cassava throughout Europe. Processing cassava intoconvenience foods simplifies or eliminates logical issues and the SADCproducers could tap into South Africa’s sophisticated processed food sector.As a result exporting processed food presents a viable opportunity forthe region. This market is expected to experience strong growth; howeverit is off a low base.The previous paragraph established that over the long term trade in cassavaas a percentage of production should increase. To understand themarket’s unfolding dynamics it is useful to analyse the manner in whichthe industry’s leaders source and consume cassava. The countries representedin Table 16 have been selected as they are dominant producers,suppliers, importers and exporters.Based on industry trends over the medium term both Brazil and Indonesiawill continue to grow cassava for their domestic use. China, Korea, theNetherlands, Spain and the US will import cassava to satisfy domesticdemand. Experts predict that the nature of demand in the US, Korea, theNetherlands and Spain should remain relatively unchanged. Another factorthat will drive China’s demand for cassava is its bio-fuels industry;as a result “other uses” should substantially increase. Although Chinahas its own plans to grow cassava on marginal land, and thereby supplytheir bio-fuels industry, this should not significantly affect its imports inthe near term, as its increase in demand will be so great as to continueincreasing at a much faster rate than its productive capacity well into thefuture. Indeed given China’s incredible growth over the last 30 years andits massive population, which is rapidly becoming ever more affluent, thedemand for bio-fuels, and other products made from cassava (e.g MSG)are likely to result in large importations for my years to come. Vietnam’sdomestic usage of cassava should also increase as a wealthier populationconsumes more livestock products. At this stage however, there is littleevidence as to how much of the increase in demand Vietnam ad meetdomestically. Should they be able to expand production significantly, theywill not only meet their domestic demand, but will also, in most probability,capture a large amount of the Chinese market.Costs Rica’s strategy to export a specialised product to niche markets indeveloped countries does not seem likely to change. Nigeria’s participationin export markets is likely to improve due to the government’s successfulCassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector StrategiesTABLE 16: KEY COUNTRIES USAGE OF CASSAVA (‘000TONS):Supply (‘000 Tons)Utilisation (‘000 Tons)Produce Imports Export Feed Other FoodLeast DevelopedBangladesh 371 0 252 119Cambodia 362 3 24 140 202182<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Cameroon 2,093 0 3 1<strong>38</strong> 509 1,443Côte d’Ivoire 2,128 0 18 42 105 1,963Ecuador 89 14 19 -6 89Gabon 230 1 0 130 100Ghana 9,739 2 78 1,057 3,308 5,298Honduras 18 8 1 11 14Kenya 643 12 5 149 501Myanmar 139 8 - 34 113Nicaragua 87 2 16 44 -13 42Rwanda 766 1 -98 865Middle-Income DevelopingArgentina 170 11 2 105 4 70Brazil 23,927 180 1,<strong>38</strong>9 11,714 4,048 6,955China 4,216 11,305 426 9,077 4,088 1,931Colombia 1,919 45 157 78 208 1,523Costa Rica 295 3 174 -3 127Czech Republic 18 0 17 0India 6,700 12 8 437 6,268Korea, Republic of 1,044 43 994 8Malaysia 430 432 151 64 259 <strong>38</strong>8Paraguay 5,500 3 30 2,471 2,016 987Peru 971 36 2 2 269 735Philippines 1,641 199 3 45 230 1,562Thailand 20,209 4 15,604 2 2,164 2,443Viet Nam 5,821 1 2,731 2,328 364 400Developed CountriesArgentina 170 11 2 105 4 70Brazil 23,927 180 1,<strong>38</strong>9 11,714 4,048 6,955China 4,216 11,305 426 9,077 4,088 1,931Colombia 1,919 45 157 78 208 1,523Costa Rica 295 3 174 -3 127Czech Republic 18 0 17 0India 6,700 12 8 437 6,268Korea, Republic of 1,044 43 994 8Malaysia 430 432 151 64 259 <strong>38</strong>8Paraguay 5,500 3 30 2,471 2,016 987Peru 971 36 2 2 269 735Philippines 1,641 199 3 45 230 1,562Thailand 20,209 4 15,604 2 2,164 2,443Viet Nam 5,821 1 2,731 2,328 364 400Source: FAOSTAT
initiatives to build the industry’s supply and demand side. Furthermorethe government has alluded to the fact that it is building its domesticindustry to provide a base to create an export crop. Nigeria’s economydoes have the capability to build critical mass in a specific product asdomestic consumption is spread between various cassava products. Theinteresting question is whether Nigeria’s exporting ambitions will cometo fission. Perhaps this provides an opportunity for SADC’s farmers and itsNigerian counterparts to form an alliance. Although South Africa does notimport vast quantities of cassava starch it has the potential to and as ithas established food processing, paper and chemical industries that use awide range of industrial starches it might consider sourcing cassava fromNigeria in the future. This need not necessarily run contrary to the idea ofSADC countries exporting to the South African market, and Nigeria doesnot have to be a major competitor, or inhibit development of commercialgrowing of cassava in the SADC region. Rather if Nigeria, South Africa,or another SADC country were to develop an industry that processes cassava,and achieve some scale economies, this could be a positive developmentfor the region, as they would then gain a geographic advantageover the South East Asian and Latin American countries. Cassava has thepotential to be a truly African product.The interesting issue is whether Thailand will have the capacity to meetits growing domestic demand, while simultaneously satisfying China’s increasingdemand for imports. On the supply side, Thailand has plantedcassava crops on its marginal land. The only alternative to increase its productionis to increase productivity. However 100% of its crop comprisesnew, improved, high-yielding cultivars. Thailand’s production costs shouldincrease over the medium term as it faces land constraints, a shortage oflabour and fertilisers tend to be relatively expensive compared to its Asiancounterparts (Howeler, 2003). As a result Thailand’s productive capacityis approaching its limit. On the demand side, domestic demand for cassavashould increase to produce ethanol. Although ethanol is made fromsugarcane this could change. A study by Kasetsart University (KuakoonPiyachomkwan et al., 2002) concluded that in Thailand using dry cassavachips is the cheapest and must convenient way to produce automotivefuel, on a large-scale (Howeler,2003:21). Already a facility to produceethanol from cassava is being constructed in Khon Kaen.In addition, since 1990 Thailand has become a net importer of soybeansand maize, which are used to feed livestock. The growth of China’s livestockindustry has increased the demand for soybeans, pushing-up itsprice. This could cause Thailand to increase its domestic consumptionof cassava pellets, creating a gap for other exporters to cover Thailand’sexports to other Asian countries. A distinction must be made between divertingexisting export supply into the domestic economy and an increasein the Asian market’s ability to consume cassava feedstock. The region’slivestock industry has grown, which would imply greater demand for cassavapellets. However creating potential demand into actual demand isdependant on the price of substitutes and complementary products. “Inspite of recent price increases of all three crops, the cassava-soybean orcassava chips-leaf meal-soybean mixes are now considerably cheaperthan maize soybean mixes with the same crude protein contents” (Howeler,2003:25).Over the medium term Thailand’s ability to satisfy the demandrequirements of its trade partners might be constrained, which providesan opportunity for SADC’s farmers to broaden their export markets,provided they are low cost producers.12. Prices12.1. Producer pricesThis section’s discussion is based on price data obtained from FAO andcovers cassava in its fresh and dried form. The FAO’s price data is a goodstarting point to form a basic understanding of price trends and thus providesone with information to ask pertinent questions. One of the problemsassociated with using this data source is that the thinner a countries’trade, the more likely information will be inaccurate. The implication is thatprice data for SADC countries will be inaccurate.Producer prices for countries represented in Figure 11 followed a progressivedownward movement over the period, except for China and Nigeriathat experienced price swings. Nigeria’s price swings are due to the government’sinitiatives to invest in the industry and the policies it has implementedto increase demand. China’s prices are volatile because of theinteraction between constrained supply due to crop failures and increaseddemand. It is interesting to note that the dominant exporters of cassavashare a similar producer price structure, and that their producer pricestend to be low. An encouraging sign is that Malawi’s producer price isin line with the world’s largest exporter, Thailand, and the world’s fourthlargest exporter, Indonesia.183<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>12.2. Average import pricesThe average import prices of countries illustrated in Figure 12, excludingSouth Africa and US, seem to follow a general trend: From 1994-1995prices increased, then declined from 1995-1997, stabilised from 1997-2003 and from 2003 started to enter into an upward phase (refer toCassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector StrategiesFIGURE 11: COUNTRIES’ PRODUCER PRICES (US$/TON)900800700184600<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>US$/Ton5004003002001000China Costa Rica Indonesia Madagascar Malawi Mozambique Nigeria Thailand1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003Source: FAOSTATFigure 12). This brief trend line illustrates that prices tend to exhibit acyclical pattern, which could be linked to the business cycle of an industrythat uses a particular product application. The implication is that farmersshould be aware of the business cycle that affects their market, and theyshould supply more than one product to an industry to minimise theirexposure to business risk.The US’ average import price per ton is significantly higher than TheNetherlands, China and Korea. This price differential reflects a productdifference. The majority of the US’ imports are superior quality cassava forhuman consumption (ITC, 2003:13).12.3. Average export pricesCountries’ average export prices from 1993-2005 seem to be random butwhen the data is analysed there appears to be a tenuous link betweenexport prices, product markets and geographic markets (refer to Figure 13Countries’ Average Export Price (US$/Ton). The Netherlands and CostaRica’s average export prices are volatile, but not random as they movethrough high peaks and low troughs. Costa Rica exports cassava for humanconsumption, which is a more specialised product than animal feed.It is no coincidence that the Netherlands’ average export price is similar toCosta Rica’s as it re-exports Costa Rica’s product throughout Europe.Thailand and Vietnam’s prices tend to move in tandem and thus exhibitthe same trend, although off a different base. Both these countries exporta “commodity” based industrial product to China. This raises the questionwhether China’s position as the world’s dominant import market givesit the ability to negotiate prices with its suppliers. If this is the case thenSADC farmers’ ability to supply this market will be cost, and not necessarily,quality driven. As a result if SADC’s farmers wish to enter the Chinesemarket, they must be able to compete against Thailand and Vietnam’s lowaverage export price.12.4. Pellet and starch pricesThe FOB price for pellets and starch follow a similar trend over the period;although off different bases (refer to Figure 14). The average annualprice for starch is higher than the price for pellets. This differential reflectsthat starch is a higher value product than pellets and as such involves a
FIGURE 12: COUNTRIES’ AVERAGE IMPORT PRICE (US$/TON)350300250185US$/Ton20015010050<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>0United States ofAmericaSouth Africa Japan China Netherlands Korea, Republic of Spain1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004Source: FAOSTATFIGURE 13: COUNTRIES’AVERAGE EXPORT PRICE (US$/TON)250200US$/Ton150100500Source: FAOSTATCosta Rica Netherlands Indonesia Thailand Vietnam1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies186<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>more complicated process. The export price of pellets fell from 1996-2000due to competition from substitute products but this trend was broken in2001 due to the Thai government’s intervention and greater demand inEast Asia. Demand was largely driven by China’s consumption, which wasdue to cheaper cassava pellet prices and its poor sweet potato crop. China’sbuoyant economy increased the demand for pellets, lifting depressedprices. The recovery of pellet prices during 2004-2005 was due to a combinationof factors; China’s economic growth, product scarcity caused bydrought and the impact of the Thai government’s ethanol programme.Thailand’s investment in the cassava feed industry gives the governmentan incentive to use its resources to safeguard its investment by manipulatingprices. This could have a potentially negative effect for SADC’s farmers,as they are exposed to additional market risk.13. Market accessCountries use tariffs barriers and non-tariffs barriers to protect domesticfarmers from imported goods. Tariffs increase the price of imported goodscompared to domestic goods, thereby giving domestic producers a relativeprice advantage. Non-tariff barriers usually take the form of strict sanitaryand phytosanitary measures or adherence to certification measures, suchas 1SO 9000 standards. Non-tariff barriers increase a producer’s coststhroughout the supply chain due to the complexity of the processes thathe/she must adhere too and the bureaucratic cost of ensuring that proceduresare documented. As a result non-tariff barriers’ potential to hinderexporters’ ability to sell their products into foreign markets is greater thantariff barriers. Unlike tariffs, non-tariff barriers do not affect all producersequally. It is more onerous for farmers in developing countries to satisfynon-tariff barriers as their access to supply-side inputs is limited comparedto their developing country counterparts. However collective organisationand the pooling of resources among SADC’s farmers could be an effectivestrategy to reduce this burden.On average, countries place higher tariff rates on a good as its move upthe value chain. As a result the tariff rate applied to cassava starch productswill be greater than the one applied to raw cassava. Also a biggerdiscrepancy exists between counties’ tariffs for value–added goods comparedto commodities. For example, “tariffs on cassava starch in the mainimporter countries range from zero in Canada, Indonesia, Malaysia andthe United States to 480% in the Republic of Korea (IFAD, 2000: 25).When a farmer plans to export a value-added good, he/she should payspecial attention to investigate tariff rates and quotas, and also any discrepanciesthat might exist between countries’ rates.FIGURE 14: STARCH AND PELLET FOB PRICES (US$/MT)400350300250US$/MT2001501005001990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005StarchPelletsSource: The Tapioca <strong>Trade</strong> Association
TABLE 17: EUROPEAN UNION’S TRARIFF SCHEDULE07.14-1010 Pellets of flour andmeal07.14-1091 Cassava for humanconsumption9.50EUR/100kg07.14-1099 Pellets made of cassavachipsConventional rate of dutyPreference for WTO members (excl. TH*, ID*,6% for imports below a quota of 145,590 tonsCN*)Preference for countries, which are not members6% for imports below a quota of 6% for imports below a quota ofof the WTO2,000 tons30,000 tonsPreference for ACP countries 8.60EUR/100kg 0% 8.80 EUR/100kgPreference for OCT 0% 0% 0%Preference for least developed countries underGSP (excl. MM*)Preference for AL*, BA*, YU*, AD*, HR*, MK*,LB*, SM*Preference for ChinaPreference for IndonesiaPreference for Thailand0% 0% 0%0% 0% 0%6% for imports below a quota of 350,000 tons6% below a quota of 825, 000 tons6% for imports below a quota of 5.5 million tons within a maximum quantity of 21 million tons over eachfour-year period.* AD - Andorra, AL - Albaria, BA - Bosnie - Herzegovina, CH - China, ID - Indonesia, HR - Croatia, MK - Former U=Yugoslav Republic of Macedonia, MM - Myanmar, LB- Lebanon, SM - Sam Marino, TH- Thailand, YU- Yugoslavia (Serbia and Montenegro)Source: TARIC cited in ITC, 2003:16187<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>13.1. Tariffs13.1.1.EUA general import duty of 9.5EUR/100kg is levied on cassava productsthat fall into the following sub-categories: Pellets of flour and meal, cassavafor human consumption and pellets made of chips (ITC,2003). Thisgeneral tariff does not apply to countries that have negotiated a bilateraltrade agreement or qualify for a special provision. As a uniform tariff rateis not applied to countries imports, exporters should refer to macmap@intracen.og and TARIC for more in-depth tariff information. The table belowis provided to give farmers a broad sense of tariff rates applied tocountries’ imports. It is not an in-depth study of tariff rates applied toSADC’s exports. SADC’s farmers should investigate whether they qualifyfor preferential treatment as a least developed country. The fact thatThailand receives preferential treatment is a concern as it is the world’sdominant and cheapest exporter of cassava chips. Indonesia also receivespreferential treatment but its exports to the EU are of a smaller magnitudethan Thailand’s.According to the ITC (2003) before a product can freely circulate withinthe EU, an import certificate must be obtained in accordance with Reg.(EC) No 1291/2000 (OJL152). An import license must be obtained beforean importer can take advantage of quota arrangements (ITC; 2003). Togain an import licence, importers must satisfy the conditions stipulatedin EC 2449/1996 (0JL333). For more information refer to http://europa.eu.int/eur-lex.13.1.2.USThe general tariff rate levied on countries’ exports is not excessive. A hostof countries have preferential access to the US market, but they are notmajor exporters of cassava. SADC countries receive preference under GSP,and thus their products enter into the US at a lower price than productsfrom Asia’s dominant exporters. Preferential access could give SADC’sfarmers a price advantage, provided their initial cost base is competitivewith industry standards.Costa Rica has preferential access to this market under the CaribbeanBasin Economic Recovery Act. Given Costa Rica’s preferential access andits proximity to the US market, it is doubtful whether SADC’s farmers couldcompete against with respect to exporting fresh cassava for human consumption.Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector StrategiesTABLE 18: US’ TARIFF SCHEDULE07.14-1010 Frozen cassava 07.14-1020 Fresh, chilled or dried cassava188<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>General tariff 7.9% 11.3%Tariff for Cuba, Laos and North Korea 35% 50%Preference under GSP (excl. Costa Rica) 0% 0%Preference under Caribbean Basin Economic Recovery Act 0% 0%Preference under Andean <strong>Trade</strong> Preference Act 0% 0%Preference for Canada 0% 0%Preference for Israel 0% 0%Preference for Mexico 0% 0%Preference for Jordan 1.9% 4.5%Source: ITC, 2003;18TABLE 19: CHINA’S TARIFF SCHEDULE07.14-1010: Fresh ManiocMFN:10%07.14-1020:Dried ManiocMFN:7%07.14-1030 Chilled / Frozen ManiocMFN:11.2%Source: Mac Map cited in ITC, 2003:17Cassava can only be imported into the US once an import permit is obtained.This permit certifies that the product satisfies phytosanitary regulationsand the produce is pest and disease free. The first step of the processto obtain an import permit is to contact a national plant protection agencyin the exporting country, for a list of foreign contacts refer to www.aphis.usda.gov/ppq/permits/phytosanitary/contact.pdf13.1.3.ChinaChina’s tariff rates are in line with other large importers, notably the EUand the US. An important issue to consider is that Thailand’s exports toChina are subject to a zero tariff duty. This gives Thailand’s exports a relativecost advantage in this market. Based on average import prices it ismore apt to state that a zero tariff rate entrenches Thailand’s status as alow cost supplier to China. The issue facing Thailand is not market accessfor it product but whether its rate of production is sufficient to satisfyboth domestic and China’s demand for cassava. This could imply that asthe market for cassava becomes constrained, the advantage that preferentialtariff access gives an importer becomes less important. For exampleChina is entering into arrangements to secure agriculture products fromcountries that do not necessary have preferential access. The larger issuefor China is access to land and labour to provide a consistent source ofsupply.13.1.4.South KoreaThe tariff rate applied to cassava products is in line with other major importingcountries’ rates, provided exporters do not exceed import quotas.Import quotas for chips are 150,000 tons and 296,000 tons formanioc pellets. In 2003 duties were 10% for chips and 2% for pellets(ITC, 2003:17). Once an importer exceeds the quotas, an excessive tariffis applied to his/her goods that could be in the region of 907.1%(ITC: 2003:17). A tariff rate of 47.8% is applied to frozen cassava (ITC,2003:17).13.2. Non-tariffs barriersCassava’s products span a multitude of industries that are subject to differentrequirements. Regulation that covers food for human consumptionis probably more comprehensive than standards applied to animal feed orfeedstock for industrial applications, such as bio-fuels. This section doesnot attempt to provide an exhaustive list of non-tariff barriers for the variouscassava products. Instead this section provides examples of generalnon-tariff barriers to illustrate their breadth. After this discussion exportersshould realise that general research about this topic is insufficient, andthat an exporter’s research should be product / market specific.
Cassava should be prepared and handled in accordance with the appropriatesections of the Recommended International Code of Practice- General Principles of Food Hygiene (CAC/RCP 1-1969, Rev. 3-1997) andother relevant Codex texts, such as Codes of Hygienic Practice and Codesof Practice.Cassava should also comply with any microbiological criteria establishedin accordance with the Principles for the Establishment and Application ofMicrobiological Criteria for Foods (CAC/GL 21-1997).Although it is not mandatory, it is generally accepted industry practice thatsuppliers have Hazard Analysis and Critical Control Point (HACCP) qualitysystems in place. If a supplier’s systems are not certified it limits the marketabilityof his/ her goods, especially in developed countries.Cassava’s quality is based on its moisture, ash, crude fibre and starch content(ITC; 2003:28). Different countries have different quality standards.However a general “norm” does exist throughout the industry. Moisturecontent should range from “12% to 14 %, ash content and extraneousinorganic contaminants, such as sand and soil, should not exceed 3% andcrude fibre content is generally accepted at 14% and starch content at 74to 82%” (ITC; 2003:28).Qualify standards tend to be market and product specific. The EU’s qualitystandards for feed material are stipulated in Commission Directive 98/67/EC (OJ L 261). If the moisture content of cassava feed exceeds 14% ofthe weight of the feed material it must be declared (ITC, 2003). “For rootsof cassava, regardless of their presentation the maximum content of ashinsoluble in hydrochloric acid is 4.5% of dry matter” (ITC, 2003:18). Accordingto directive 02/32/EC (OJ L 140) a cassava product’s hydrocyanicacid content must be below 100 mg/kg, its “aflatoxin content must notexceed 0.05mg/kg if cassava is used as a complementary feeding stufffor cattle, sheep and goats and 0.03mg/kg for pigs and poultry. For dairyanimals and young animals the maximum content is 0.005mg/kg” (ITC,2003:18).14. Marketing activitiesDistribution channels tend to be product and market specific. As a resultdistribution channels will be different for animal feed, food ingredients,convenience foods and starch based products used to manufacture goods.Given the difference between distribution channels, on a product specificbasis, and also the nuisances exiting within product specific distributionchannels, this TIB does not cover this issue.Although the manner a product is packaged is largely determined by thebuyer, countries have minimum regulations. As a result an exporter shouldconsult an industry association in his/her country to ensure that a product’spackaging satisfies the importing countries’ regulations.15. Way forwardThe way forward comprises five stages. First, identify markets that havegrown over the past five years or are entering into their growth phase.This is not a simple task because a market’s economic stage of developmentand its government’s agricultural policies affects the type of cassavaproduct it demands. For example the market for cassava feed in Africais predicted to grow and be lucrative, whereas in Europe its growth istapering-off. Given this TIB’s scope, this section does discuss the prospectof exporting specific product to specific markets but rather highlightsemerging trends. Second, investigate whether the quality and volume ofcassava exists to satisfy a users’ requirements. Third, consider the impactof substitute products on the demand, price and properties of cassavabased products. Fourth, propose measures that could be used to improvecassava’s competitiveness in its target market compared to its substituteproducts. Lastly, develop strategies to integrate small-scale farmers intoestablished industry value chains, such as the processed food value-chainor the bio-fuel value chain.<strong>Trade</strong> patterns are influenced by the interaction of regional, country andproduct dimensions. Experts predict that cassava feed is a potentiallylucrative market in SADC. The region’s demand for livestock products isincreasing, which increases the consumption of feed. Importing maize andwheat to satisfy this demand is a relatively expensive option and it drainsforeign reserves. This policy is not preferable considering that on averageAfrica uses 20% of its productive capacity to produce feed. Given the lowvalue of cassava feed, transporting the product far distances is not profitable.These circumstances provide an opportunity for SADC’s farmers tosupply cassava feed to the region’s livestock industry. However this doesnot imply that a market does not exist for native and modified starchesto be used in SADC’s food, textile, paper and pharmaceutical industries.Alternatively, in Asia, it is touted that the next growth market for cassavawill be its usage in industrial applications as a modified starch and toproduce bio-fuels, especially in Thailand and China. Although the aboveantidotes illustrate that differences exist between countries’ demand driversand usage trends, if data is analysed on a generalised level, it becomesapparent that the market for starches or starch based products offers thegreatest potential for cassava, on a global basis.189<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies190<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Starches are a versatile product that can be used in an array of industriesand applications, such as processed food, paper, textiles; ethanol and biodegradableplastics. Even within an industry, starch has a multitude ofuses. According to the FAO, “the extent of specific functional propertiesof starches required by the food industry, alone, is almost unlimited, asno other ingredient provides texture to as many foods as starch does”.In addition, given the rise of consumerism and urbanisation, consumers’demand for processed goods continues to grow. Another important factoris that starch is used to manufacture a range of products that cut acrossvarious industries that have different business cycles and demand drivers.Therefore farmers’ exposure to volatile market movements, on average,should be reduced. Starch is the quintessential value-added cassava product.Generally, value-added products tend to be application specific and asa result are less prone to price swings than commodity products.Competition between cassava and other starch sources is not betweenphysical commodities; instead, it is based on the functional characteristicsof these commodities’ value-added products. As a result a starch shouldbe viewed as “a set of functional characteristics suited to a particularapplication”, and not a product per se (FAO,21). Native starches havebiological properties that make them better suited to certain applications.Native cassava starch is also very resistant to acid conditions, it is intermediatelyresistant to freezing but very unstable during heating (sterilization),making it suitable for some and unsuitable for other applications(Dufour et al., 2000 cited in Howeler; 2003:29). Furthermore, cassavastarch has a “neutral taste and odour, and the transparency, smoothnessand viscosity of a gel, making it particularly suitable for many processedfood items (Howeler; 2003:29).The starch market is lucrative but entrance is based on a product’s functionalityand its cost. Cassava might be a cheaper product but does it havethe desired properties to compete against other starches? To explore acrop’s functional properties requires research and development activities.Cassava is viewed as a poor man’s subsistence crop; as a consequenceresearch about cassava’s physical starch properties lags behinds its substitutestarches. This places cassava at a relative disadvantage compared toother crops that have benefited from the development of superior cultivarstrains and processing technologies that make it easier for them to fit intoindustrial activities. A comparative lack of interest in cassava has createdan unstable crop characterised by sporadic cultivation, poor processingmethods and lower quality products, which has marginalised cassava’s usagein industrial applications. Thus substitute starches have an unfair advantage,and in essence one is not comparing like with like. Therefore fortropical starches to tap into the lucrative industrial starch market researchis required to develop new products to satisfy users’ specifications.Initially this research should explore cassava’s physical, chemical and organicproperties to create a product that is easier to process and distribute.This implies that research to produce a superior starch begin with designingan improved cultivar type and end in the laboratory, where a starchis engineered. Based on this logic creating a better starch encompassesimproving activities throughout the value chain, as a consequence the followingactivities are proposed as part of the way forward. First, create amore competitive crop. This involves modifying cassava’s functional characteristicsto develop a cultivar that produces roots which have a higherstarch and nutritional content and thin, easy-to-peel skins (FAO, 21). Alsoresearch at this level should consider end users’ requirements and thendesign a starch with the functional requirements to satisfy this need. Thisresearch is not overly complicated as slight changes in amylose/amylopectinratios have a significant effect on functional characteristics (FAO, 21).Furthermore the costs incurred to produce tailor made starches could berecouped from end-users. <strong>Industry</strong> has an appetite to invest in developinga starch that is not grain or soybean based due to these commodities continuedprice hikes. This situation makes it easier for cassava starch to gainacceptance, provided it can compete against substitute starches in termsof its functional properties and consistent availability of quality supply ata relative price advantage.Developing countries access to resources to conduct scientific researchis limited. However this TIB argues that cassava starch can only competeagainst other starches if more value-added research is done on its functionalproperties. It seems that the starch market is potentially lucrativebut is inaccessible; however this is not the case. Developing countriesshould devise strategies to spread R&D costs between institutions andconsider forming partnerships with the private sector. This process couldstart with collaboration between various domestic institutions, regionalinstitutions (forming relationships between SADC and other African tradingblacks/ agencies) and tapping the resources of international institutions.These ventures should not solely focus on producing “outputs’(research) but rather foster cooperation throughout the value chain byencouraging “close and effective collaboration between national researchand international extension institutions that work with local and provincialgovernment officials” (Howeler, 2003:35). Encouraging participationat the local level is vitally important as it ensures that research is not anacademic activity but improves the function throughout the supply chain.As a consequence the full benefits derived from this process will not be
ealised unless “farmers become directly involved in testing, selecting anddisseminating new practices and technologies (Howeler, 2003:35).Asian counties have used the above approach to improve the profitabilityof their cassava industry. These countries’ national research institutionscollaborated with the Centro Internacional de Agricultura Tropical (CIAT)to develop a cassava cultivar that produces a 20%–40% higher yieldingcrop whose roots have a greater starch content (APBN,1998:<strong>38</strong>5).Research had social and financial benefits as cassava is grown predominatelyby small-scale farmers. Thailand used farmer participatory researchprogrammes to create opportunities for small-holder farmers in marginalareas on sloping and undulating land to participate in commercialisedagriculture. These farmers’ were involved in evaluating promising cassavagermplasm, developing effective soil conservation practices and investigatingbalanced fertilization and cropping systems. This project was financedby the Nippon Foundation from 1994 to 2003. By the end ofthe project, new high yielding and high starch varieties were adopted innearly 1 million ha (98% of cassava area) in Thailand, 100,000 ha (40%)in Vietnam and 36,000 ha (10%) in China, benefiting at least 800,000cassava farmers (Howeler R & Kawano K & Watananonta W & Ngoan T).The next area for research should be improving processing methods, especiallypost-harvesting techniques, to lower production costs. Given freshcassava’s bulky, perishable nature it is not economical viable to transportit long distances. This opens up opportunities for small scale farmers toget involved in rudimentary processing activities, provided they have accessto machinery, finance and skills. Small scale farmers tend to dry cassavain the sun. This processing technique is not ideal as cassava can becontaminated, which makes it unfit for consumption. Small scale farmerscannot afford to purchase a flash-dryer, and addition, this technology isdiesel/ fuel powered which pushes its operational cost above must farmers’means. Portable micro rotary dryers need to be developed to processcassava into a storable, high-quality product at the farm gate. It might beargued that incorporating small farmers into processing activities couldreduce the competitiveness of a country’s exports as scale economies arenot fully exploited. This is not the case. Thailand has structured its cassavaindustry in such a manner that export led growth benefits parties throughoutthe value chain. Their model integrates small scale farmers into thevalue chain by creating farmer associations that pool supply side resourcesto invest in machinery and creates nodes for scale farmers to tap intothe private sector’s marketing resources and distribution networks.The third stage of research should include developing new products andmarkets that exploit cassava’s unique starch characteristics, with emphasisplaced on the processed food industry, and applications to createbio-fuels. Developing new products tends to simulate demand for otherproducts based on the initial product. There is a reinforcing relationshipbetween market demand and product development: “’Market demanddrives product development, and sometimes, new products create newmarket opportunities’ (Howeler, 2003:34). This implies that “’for eitherto succeed, products and markets need to develop in coordination, andproduction, processing and marketing need to be fully integrated” (Howeler,2003:34).Although international starch markets have great potential, they alsohave drawbacks. They are complex and protected. Before starch producersventure into international markets, they build up critical mass byconcentrating on supplying their domestic markets. In SADC‘s case, dueto the market’s size, this approach could be extended to creating a regionalmarket. This may imply two policy outcomes: “Import substitutionfor competing products, especially if they are subsidised by the supplyingcountries, as is the case of potato starch from the EC, and further progressin reducing tariff and non-tariff barriers” (IFAD, 2000:7).16. ConclusionCassava was selected as an export crop for SADC’s farmers because of itshour-glass supply chain (ARC; 2007). Growing and harvesting cassava isa manual intensive activity and thus lends itself to small-scale productionunits. Post-harvesting activities involve milling and drying cassava and arenot capital intensive or complicated and thus they can be conducted atthe farm-gate or within the community or village. Other activities in thesupply chain, such as refining, extracting, marketing, and packaging; tendto be more capital and knowledge intensive and thus benefit from scaleeconomies. These activities are done by fewer larger-scale units, whichthen distribute their final product to a larger number of consumers. Theshape of the supply chain provides possibilities for small-scale farmers locatedon marginal lands to become involved in producing a cash crop andto participate in rudimentary value-added processing activities. Howeverfor these benefits to be realised new micro drying technology must bedeveloped and the supply chain must be analysed to provide small-scalefarmers with nodes to tap into commercial agriculture’s marketing anddistribution networks.Cassava can be processed into an array of products that can be used bynumerous industries. The complexity of processes required to make theseproducts vary. The simplest product is staple food and the most advanced191<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>Cassava <strong>Trade</strong> <strong>Industry</strong> Brief
Sector Strategies192<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>is a modified starch. A country’s demand profile for a specific cassavaproduct is linked to its stage of economic development. Therefore trade incassava products tends to be country and product specific. Despite productspecificity among countries, broad generalisations can be made. Overthe long-term, cassava’s best potential growth market is its applicationin starch and starch-based products. On a price basis cassava starch cancompete against its substitute products, but cassava’s ability to competeagainst them with respect to their functional properties is limited to specialisedmarkets. Cassava’s status as a poor man’s crop has resulted inminimal scientific interest in the crop. This has created the perception thatcassava lacks the “wide range of intrinsic starch characteristics found inthe gene pool of some competing crops like maize, wheat and potato”(Howeler,2003:1). As a result cassava starch’s application is limited comparedto its substitutes, but perhaps more importantly, it is locked out ofspecialised markets that are less volatile and more profitable than massmarkets (Howeler, 2002:1). Research into cassava’s properties and itsvalue-added applications has been made a priority by the FAO.The market for cassava feed is growing, but its growth rate has taperedoff since the 1980s. This market’s growth is driven by the expansion ofthe livestock industry in Africa, South America and Asia. Consumption ofcassava feed is growing in countries that produce cassava, while consumptionin non-producing countries, such as the EU, has stagnated dueto competition from substitute products, as a result trade flows shouldremain unchanged. Another issue to consider is that growth does notimply profitability, the demand for cassava feed is growing but averageexport prices are too low to cover production costs, unless a farmer canmatch Thailand’s average export price, which might not be feasible givenThai farmers direct and indirect government support throughout the valechain.The market for animal feed is unstable and thus relatively high risk comparedto other markets because the price of cassava feed is affected theprice of substitute and complementary products, which in turn is influencedby agricultural polices and climatic conditions. It might be arguedthat cassava starch faces the same problem, but this is incorrect asstarches complete on a basket of factors, of which one factor is price.This TIB suggests that SADC’s farmers should enter the feed market, asit represents the second stage in supplying a value-added product, butbe selective about which market it supplies. Competing against Thailand,the world’s low cost producer, will be difficult. The best strategy would beto supply a geographically close, uncontested growing market, such asSub-Saharan Africa.Data presented in this TIB illustrates that building on SADC’s productioncapacity to turn cassava into a cash crop has economic and social benefits.For these potential benefits to become tangible requires sequencedprogrammes that builds on the industry’s supply and demand side abilities,identifies bottlenecks and addresses constraints. On the supply sidethese measures should initially focus on improving yields and reducingprocessing costs, as these activities form the basis of developing valuechains for lucrative products, such as cassava feed for Africa, starchesfor developed markets ethanol for Nigeria, Ghana, Thailand and China).Access to resources is a constraining factor in SADC. Establishing a newvalue chains is a resource intensive activity, therefore to stretch scarceresources further, the development of a processing network that sharessimilar standards is important. On the demand side restructuring measurescould encompasses the following activities: stimulate the developmentof a domestic market that has the flexibility to “slot” into regionalmarkets and use local markets to test the popularity of innovative cassavabased products. For example, cassava can be used as a raw material toproduce bio-fuels. To explore lucrative opportunities in this field, the valuechain involved in producing cassava and bio-fuels should be mapped andareas of co-operation should be explored.
ReferencesAgricultural Research Council, (2007) CassavaIFAD (2000), The World of Cassava: Facts, Trends and OutlookAsia Pacific Biotech News (1998), Framers Cash in on Cassava Boom inSEA, <strong>Vol</strong> 1, Nos. 16&17IFAD, Cassava Processing and Marketing, Regional Initiative Workshop,20-22 March 2006Brant. S. and Fang. R. and Lin C.C (2006) “The Potential for Ethanol FuelIn China””CGPRT Flash (2003), “Good Prospects for Cassava Development”, <strong>Vol</strong> 1,No 2, ISSN 1693-4636CIAT, Cassava: A crop for Hard Times and Modern Times, In Focus CropCommitmentsDukec, Z (2003), The World Market for cassava. Manioc, Market Brief, International<strong>Trade</strong> Centre, UNCTAD, GenevaFAO (1988), Tropical Starch Misses Market, Agriculture 21International Starch Institute (2007), CassavaKamateng Kahoy, B, CassavaLamchaiyaphum B and Piyachomkwan and Sriroth K (2006), Present Situationand Future Potential of Cassava in ThailandLatner K and O’Kray C and Jiang J (2006), “People’s Republic of China:An Alternative Future for Agriculture”, USDA Foresign Agricultural Service,CH6049Liu Y (2006) China Embarks on a Million Ton Cassava Ethanol Base inGuangxi, WorldWatch Institute193<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>FAO (2000) a New Strategy for Cassava, Agriculture 21 MagazinePattern D (2006) China Clamps Down on Corn Use in BiofuelsFood and Beverage Development Europe, Cassava to Compete with MajorCrops in the Starch GameRoonaphai N (2007) “Pathways Out of Poverty through Cassava, Maizeand Soybean in Thailand, UNESCAP-CAPSA Newsletter, Palawija NewsHoweler R (2003), Cassava in Asia: Present Situation and its Future Potentialin Agro-<strong>Industry</strong>.Tewe P Strategies for Sustainable Development of Cassava Based Industriesin Osun State, Osun State Economic SummitHoweler. R. and Kawano. K and Watananonta. W and Ngoan. T .N, “Workingwith Farmers: Spreading New Cassava Varieties, Improved Practices……. and New Hope”, CIATCassava <strong>Trade</strong> <strong>Industry</strong> Brief
194<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>
195<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>
196<strong>Trade</strong> & <strong>Industry</strong> <strong>Monitor</strong>
TRADETRADE & INDUSTRY MONITOR@A GLANCE
Q4 2005 Q4 2006 Q3 2006 Q4 2006Rbn US$bn Rbn US$bn Rbn US$bn Rbn US$bnTotal Exports 83.96 12.86 112.71 15.41 103.80 14.52 112.71 15.41Total Imports 91.41 13.97 140.93 19.20 119.<strong>38</strong> 16.73 140.93 19.20<strong>Trade</strong> Balance -7.45 -1.11 -28.22 -3.80 -15.58 -2.21 -28.22 -3.80ProductsTotalExports (Rbn)% of TotalExportsProductsTotalImports (Rbn)% of TotalImportsPrecious metals 32.0 28.3 Mineral and fuel oils 30.02 21.3Iron and steel 12.7 11.2 Machinery and boilers 21.74 15.4Mineral and fuel oils 10.8 9.6 Electrical equipment 13.95 9.9Vehicles 9.9 8.8 Vehicles 13.11 9.3Machinery and boilers 9.0 8.0 Motor Vehicle parts 8.83 6.3Ores, slag and ash 7.2 6.4 Medical & surgical equipment 4.02 2.9Aluminium 3.6 3.2 Plastics 3.26 2.3Electrical and electronic equipment 2.1 1.8 Pharmaceutical products 2.39 1.7Inorgainic chemicals 1.9 1.7 Aircraft 2.28 1.6Articles of iron or steel 1.5 1.3 Iron and steel 2.19 1.5Total 90.74 80.5 Total 97.32 69.1TOP THREE NON-MINERAL EXPORTS FROM AND IMPORTS TO SA FROM REGIONS (HS4, Q4 2006)Region Exports ImportsProducts Value (Rbn) % share Products Value (Rbn) % shareCoal briquettes & ovoids 4.61 12.<strong>38</strong> Motor vehicles for transport of persons (except buses) 4.51 10.03EULiquid & gas centrifuges 4.00 10.73 Original equipment components 3.85 8.57Stainless steel sheets 2.06 5.53 Radio and TV transmitters, television cameras 1.53 3.41Motor vehicles for transport of persons (except buses) 2.26 10.43 Original equipment components 2.67 8.88East AsiaNAFTASADCMiddle EastSouth-East AsiaSouth AmericaFerro-alloys 2.18 10.07 Motor vehicles for transport of persons (except buses) 2.11 7.03Iron ores and concentrates 2.03 9.40 Radio and TV transmitters, television cameras 1.51 5.03Motor vehicles for transport of persons (except buses) 1.10 8.22 Turbo-jets, propellers and other gas turbines 0.80 6.21Ferro-alloys 0.64 4.83 Aircraft, spacecraft and satellites 0.77 6.02Stainless steel sheets 0.48 3.64 Original equipment components 0.57 4.48Motor vehicles for the transport of goods 0.39 4.5 Nickel mattes, nickel and other products of nickel metallurgy 0.37 6.78Structures of iron/steel 0.22 2.5 Nickel ores and concentrates 0.32 5.92Coal briquettes & ovoids 0.22 2.5 Copper wire 0.25 4.65Coal briquettes & ovoids 0.56 15.45 Mineral or chemical fertilisers, nitrogenous 0.37 2.93Liquid & gas centrifuges 0.20 5.45 Polymers of ethylene, in primary forms 0.18 1.39Hot-rolled products, iron/steel 0.16 4.45 Acyclic hydrocarbons 0.11 0.89Hot-rolled products, iron/steel 0.66 26.48 Original equipment components 0.86 12.39Chemical wood pulp 0.30 11.93 Rice 0.36 5.21Motor vehicles for transport of persons (except buses) 0.11 4.54 Automatic data processing machines (computers) 0.35 5.01Coal briquettes & ovoids 0.17 22.51 Original equipment components 0.77 13.77Ferro-alloys 0.12 15.44 Soya-bean oil 0.30 5.44Nickel plates, sheets, strip and foil 0.04 5.61 Meat and edible offal 0.30 5.41TRATRA@A G
Imports into South AfricaExports from South AfricaQ4 2005 – Q4 2006 (%) Q3 2006 – Q4 2006 (%)Total Exports 34.25 8.58Total Imports 54.18 18.05Note: Growth rates have been calculated on the Rand valuesTOP 10 EXPORT MARKETS AND IMPORT SOURCES (Q4 2006), ALL PRODUCTSExportsImportsCountry Value (Rbn) Share (%) Country Value (Rbn) Share (%)USA 12.00 10.7 Germany 16.02 11.<strong>38</strong>Japan 10.69 9.5 China 15.29 10.87UK 9.06 8.1 US 10.97 7.79Germany 8.00 7`.1 Sudi Arabia 9.<strong>38</strong> 6.67China 5.77 5.1 Japan 8.08 5.74Netherlands 5.30 4.7 Iran 6.79 4.83Switzerland 3.18 2.8 UK 6.46 4.59DEE & INDUSTRY MONITORLANCESpain 2.87 2.6 Nigeria 4.82 3.43Belgium 2.83 2.53 France 4.01 2.85Italy 2.76 2.47 Italy 3.99 2.84Total 62.46 55.8 Total 85.82 61.0SA TRADE BY REGION (RBN)Region Q4 2005 Q4 2006 Q3 2006 Q4 2006Exports Imports Exports Imports Exports Imports Exports ImportsEU 27.17 32.91 37.25 44.98 33.87 43.28 37.25 44.98East Asia 14.60 21.08 21.63 30.02 19.87 27.87 21.63 30.02NAFTA 9.10 8.47 13.32 12.84 11.79 10.93 13.32 12.84SADC 8.36 2.29 8.81 5.45 8.88 4.18 8.81 5.45Middle East 2.92 8.14 3.61 12.71 3.72 7.30 3.61 12.71South-East Asia 1.78 4.<strong>38</strong> 2.48 6.94 2.28 6.53 2.48 6.94South America 0.83 3.05 1.30 5.59 1.17 4.65 1.30 5.59Rest of Africa 3.78 1.05 4.76 5.99 3.91 3.53 4.76 5.99Rest of the World 15.42 10.03 19.55 16.41 18.30 11.12 19.55 16.41
<strong>Trade</strong> and Industrial Policy Strategies (TIPS)814 Church StreetArcadia 0083PO Box 11214Hatfield 0028South Africatfew+27(0)12 431 7900+27(0)12 431 7910info@<strong>tips</strong>.org.zawww.<strong>tips</strong>.org.za
Change language