Global Economic Prospects 2006 - ISBN: 0821363441

GlobalEconomicProspects34320Economic Implications ofRemittances and Migration2006

GlobalEconomicProspectsEconomic Implications ofRemittances and Migration2006

© 2006 The International Bank for Reconstruction and Development / The World Bank1818 H Street, NWWashington, DC 20433Telephone: 202-473-1000Internet: www.worldbank.orgE-mail: feedback@worldbank.org1 2 3 4 09 08 07 06This volume is a product of the staff of the World Bank. The findings, interpretations, andconclusions expressed herein do not necessarily reflect the views of the Board of ExecutiveDirectors of the World Bank or the governments they represent.The World Bank does not guarantee the accuracy of the data included in this work. Theboundaries, colors, denominations, and other information shown on any map in this work donot imply any judgment on the part of the World Bank concerning the legal status of anyterritory or the endorsement or acceptance of such boundaries.Rights and PermissionsThe material in this work is copyrighted. Copying and/or transmitting portions or all of thiswork without permission may be a violation of applicable law. The World Bank encouragesdissemination of its work and will normally grant permission promptly.For permission to photocopy or reprint any part of this work, please send a request withcomplete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers,MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, www.copyright.com.All other queries on rights and licenses, including subsidiary rights, should be addressed tothe Office of the Publisher, World Bank, 1818 H Street NW, Washington, DC 20433, USA,fax 202-522-2422, e-mail pubrights@worldbank.org.ISBN: 08213-6344-1E-ISBN: 0-8213-6345-X13-digit E-ISBN: 978-0-8213-6345-4DOI: 10.1596/978-0-8213-6344-7EAN: 978-0-8213-6344-7ISSN: 1014-8906Library of Congress Cataloging-in-Publication data has been applied for.Cover design: Naylor DesignCover photo: Panos

ContentsForewordviiAcknowledgmentsixOverviewxiAbbreviationsxviiChapter 1 Prospects for the Global Economy 1Global growth 3Long-term prospects and poverty forecast 8International finance 10Commodity markets 14World trade 16Risks and uncertainties 18Notes 22References 23Chapter 2 The Potential Gains from International Migration 25International migration trends 26The demographic challenge 29Migration and its development impact 31Returns to households 34Returns to factors of production 41Caveats—what the model leaves out 48Notes 51References 53Chapter 3 The Policy Challenges of Migration: The Origin Countries’ Perspective 57The migration decision and its impact on migrants and their families 59Low-skilled migration 64High-skilled emigration 66Diasporas 70The return of expatriates can benefit development 70Temporary migration and international agreements 72Notes 76References 78iii

C O N T E N T SChapter 4 Trends, Determinants, and Macroeconomic Effects of Remittances 85Remittance data and trends 86Factors affecting remittance flows 92Macroeconomic effects of remittances 99Annex 4A.1 World Bank data on remittances 105Annex 4A.2 A model-based estimation of informal remittance flows 108Notes 110References 112Chapter 5 Remittances, Households, and Poverty 117Remittances, poverty, and inequality 118Remittances and household consumption smoothing 122Remittances and indirect effects on household income 123Remittances, savings, and investment 125Annex 5A.1 Poverty simulation model: description and results 127Notes 129References 131Chapter 6 Reducing Remittance Fees 135Remittance fees and costs 136Factors underpinning high remittance fees 144Policies to reduce remittance costs 147Remittances and financial institutions 149Annex 6A.1 A stylized remittance transaction—structure, players, instruments 151Annex 6A.2 Licensing and registration requirements for remittance serviceproviders 152Annex 6A.3 A brief history of some remittance service providers 153Notes 154References 156Figures1.1 Industrial production 31.2 A sharp slowdown 51.3 Regional growth 61.4 Dollar-euro interest rate differentials 101.5 Financing of the U.S. current account deficit 111.6 Emerging market spreads 111.7 Real long-term interest rates in G-7 countries 121.8 World savings rate 121.9 Inflation rates 131.10 Cumulative real increase in housing prices, 2005 141.11 Commodity prices 141.12 Levels of spare oil capacity 151.13 World trade volumes 161.14 Change in textile exports to the developed world, first half of 2005 171.15 Estimated change in textile exports as share of total merchandise exports 181.16 Some countries are particularly at risk 202.1 International migrants as a share of destination countries’ population 27iv

C O N T E N T S2.2 Share of females in international migration 282.3 Immigration to selected countries, reasons for admittance, 2001 282.4 Labor force and dependency rates 302.5 Factor returns and migration 422.6 Source of gains for native workers 452.7 Factor returns and migration in high-income countries, 2005 463.1 Median wage levels for workers in the same occupation, relative to high-incomeeconomies, 1988–92 593.2 Major developing country diasporas in developed countries 623.3 Estimates of stock of irregular migration 623.4 Emigration rates for low-skilled workers 643.5 Emigration rates for those with a tertiary education, 2000 673.6 Number of temporary workers admitted under skill-based programs 724.1 Top 20 remittance-recipient countries, 2004 904.2 Estimated remittance payment, by country group, 2004 914.3 Remittances as percent of private consumption, two years before and two years afternatural disasters 1004.4 Remittances as a share of personal consumption, two years before and two years afterconflict 1004.5 Remittances as a share of personal consumption, two years before and two years afterfinancial crises 1014.6 Indebtedness classification including and excluding remittances, 2003 1014.7 Remittance securitization structure 1024.8 Securitization of remittances, 1994–2004 1035.1 Sri Lankan migration 1226.1 Remittance costs are high and regressive 1386.2 Remittance fees in the United States–Mexico corridor 1386.3 Fees and foreign exchange spreads for $200 in Western Union transfers from NewYork City 1446.4 Exclusive arrangements with post offices skew competition 1456.5 Barriers perceived by remittance service providers 146Tables1.1 The global outlook in summary 41.2 Long-term prospects 81.3 Regional breakdown of poverty in developing countries 91.4 Terms-of-trade impacts of commodity price changes 161.5 Impact of a 2 million bpd negative supply shock 191.6 Interest rate scenarios 202.1 Growth in international migration by destination, 1970–2000 272.2 Labor force structure in the base case and after increases in migrants 332.3 Change in real income across households in 2025 relative to baseline 342.4 Real income impacts across developing regions 382.5 Impact of different assumptions on the consumption of public goods and services byselected groups in 2025 393.1 Fees charged by recruitment agencies 603.2 Emigration rates of skilled workers, 2000 684.1 Workers’ remittances to developing countries, 1990–2005 88v

C O N T E N T S4.2 Recorded remittances have grown faster than private capital flows and ODA 884.3 Choice of remittance channel in selected countries 914.4 Estimated increase in formal remittances if transaction costs were reduced to2 to 5 percent and dual exchange rates were eliminated 924.5 Impact of remittances on country credit rating and sovereign spread 1024A.1.1 Countries with alternative estimates in 2004 1064A.2.1 Regression results: determinants of worker remittances 1094A.2.2 Regression results: determinants of transaction costs 1094A.2.3 Panel regression results: determinants of remittances 1105.1 Simulated impact of eliminating remittances on poverty rate 1205A.1 Effect of removing remittances on the poverty headcount rate 1286.1 Approximate cost of remitting $200 1376.2 Operating profits of major MTOs 1406.3 Estimating the cost of a remittance transaction 1406.4 Remittances are more cost-elastic when costs are higher 1436.5 Policies to reduce costs, regulate informal providers, and provide remittance-linkedfinancial services 148Boxes2.1 The model used in this study 322.2 Calculating and interpreting global welfare gains from migration 362.3 The impact of immigrants on fiscal balances 402.4 Empirical studies of the impact of immigration on wages 432.5 Increased migration and its impact on wages 473.1 Internal versus international migration 653.2 Mode 4 and international migration 754.1 International working group on improving data on remittances 874.2 The recent surge in remittance flows to India 894.3 Collective remittances through hometown associations and matching schemes 954.4 Forced remittances 974.5 Unlike oil windfalls, remittance inflows do not weaken institutional capacity 1055.1 Estimating a cross-country poverty change model 1196.1 Decline in remittance costs in the United States–Mexico corridor 1396.2 Estimating remittance industry costs 1416.3 Even charitable donations are sensitive to cost 1426.4 United States–Mexico FedACH 1486.5 The World Bank/CPSS task force on general principles for international remittancesystems 1496.6 Smart’s phone-based remittance system in the Philippines 150vi

ForewordFor millennia people have migrated insearch of economic opportunity. In thenineteenth and early twentieth centuries,technological advances and untapped naturalresources drove movements of population fromEurope and Asia to the Americas. Internationalmigration generated enormous improvementsin people’s lives. Immigrants enjoyed higherwages, countries of destination profited fromincreased supply of labor, and countries of originsaw labor market pressures ease.Current trends indicate that pressures formigration from the south to the north are setto rise again. This movement is driven largelyby income gaps and the rising number ofyoung adults in developing countries seekingbetter opportunities abroad. The economic,social, and political implications that comewith the movement of people differ from themovement of goods or money. As a result, thetopic of international migration has promptedmuch political debate in the internationalcommunity today.The prospects for migration flows are criticalfor development. Developing countriesbenefit through the money that migrants sendhome to their families (remittances), throughreduced labor market pressures, and throughcontacts with international markets and accessto technology.But migration is not always beneficial. Migrantscan be subject to exploitation andabuse, and the loss of highly skilled personnelthrough migration has hindered developmentin some countries.The World Bank’s research department, inpartnership with others, has launched a programto expand knowledge in an area that deservesgreater attention. The program addressesthe issues surrounding remittances;migration of high-skilled workers; the determinantsof migration; temporary movementsof persons; social protection and governance;and the links among trade, foreign direct investment,and migration.An integral part of this program, GlobalEconomic Prospects 2006 focuses on policiesto improve the developmental impact of remittances.It documents the high level of transactionscosts facing migrants sending small remittancesto their families, and it outlines theregulatory issues and market imperfectionsthat keep costs high.Fewer barriers to remittance flows andgreater competition among remittance serviceproviders could substantially reduce costs andboost remittance flows to developing countries.Global Economic Prospects 2006 showshow sound domestic policies and an investment-friendlyclimate can significantly increasethe contribution of remittances and migrationto improved living conditions back home.Migration remains an important force forfighting poverty, the key mission of the WorldBank, and it is our hope that this report willcontribute to this important debate.Paul WolfowitzPresidentWorld BankNovember 2005vii

AcknowledgmentsTHIS REPORT WAS prepared by the Development Prospects Group (DECPG). The lead authorsof this report were Dilip Ratha and William Shaw, with direction by Uri Dadush. Theprincipal authors of the chapters were Andrew Burns (chapter 1), Dominique van derMensbrugghe (chapter 2), William Shaw (chapter 3), and Dilip Ratha (chapters 4, 5, and 6). Thereport was prepared under the general guidance of François Bourguignon, chief economist andsenior vice president of the World Bank.The main macroeconomic forecasts in chapter 1 were prepared by the Global Trends Team ofDECPG led by Hans Timmer and including John Baffes, Andrew Burns, Maurizio Bussolo,Annette de Kleine, Betty Dow, Himmat Kalsi, Fernando Martel Garcia, Donald Mitchell, GaureshShailesh Rajadhyaksha, Mick Riordan, Cristina Savescu, Shane Streifel, and Shuo Tan. The outlookfor the East Asia and Pacific region was carried out with the cooperation of Milan Brahmbhattand Louis Kuijs. The team also benefitted from in-depth consultations and comments fromthe regional chief economists and their staff, as well as country economists. The long-term growthand poverty forecasts were prepared by Dominique van der Mensbrugghe, Shaohua Chen, andMartin Ravallion. The companion Prospects for the Global Economy web site was prepared byAndrew Burns, Sarah Crow, and Cristina Savescu, in collaboration with Reza Farivari, SaurabhGupta, David Hobbs, Shahin Outadi, Raja Reddy Komati Reddy, Malarvizhi Veerappan, andCherin Verghese.Maddalena Honorati and Prabal De provided research assistance. Chapter 2 benefitted fromcollaboration with Hans Timmer and from comments received from seminar participants, notablyLindsay Lowell and Susan Martin. Special thanks are due for the background materialprovided by Riccardo Faini, Robert Lucas, Julia Nielson, Kathleen Newland, and Irena Omelaniukfor chapter 3; Swaminathan S. Aiyar, Ralph Chami, Neil Fantom, Caroline Freund, GaryMcMahon, Irena Omelaniuk, Serdar Sayan, Nikola Spatafora, and K. M. Vijayalakshmi forchapter 4; John McHale for chapter 5; John Gibson, David McKenzie, George Kalan, DilekAykut, Nikos Passas, and Jan Riedberg for chapter 6. Ole Andreassen, Jose de Luna Martinez,Raul E. Hernandez-Coss, Massimo Cirasino, and Roger Ballard also contributed backgroundnotes for chapter 6. Thanks also to colleagues in the International Organization for Migrationwho helped collect information on remittance-related government policies (for chapter 4) usingtheir extensive international network, and Bernd Balkenhol of the International Labour Organizationfor preparing a background paper on forced remittances.ix

A C K N O W L E D G M E N T SMany colleagues provided excellent comments at various stages of the report’s preparation.L. Alan Winters provided comments on the report and guidance throughout its preparation. LucaBarbone, Kevin Barnes, Augusto de la Torre, Shantayanan Devarajan, Mustapha Nabli, JohnPage, Bryan Roberts, John Whalley, and Dean Yang were peer reviewers at the Bankwide review.Richard Adams, William Easterly, Isaku Endo, Jose Maria Fanelli, Shahrokh Fardoust, IanGoldin, Daria Goldstein, Yevgeny Kuznetsov, Ali Mansoor, Phil Martin, Maria Soledad MartinezPeria, Fernando Montes-Negret, Nayantara Mukerji, Latifah Osman Merican, Christopher Parsons,Guillermo Perry, Sonia Plaza, S. Ramachandran, and Terrie Walmsley also provided usefulcomments. Johan Mistiaen and Romeo Matsas provided excellent help in designing and implementinga survey of migrant remitters from Congo, Nigeria, and Senegal residing in Belgium.Maria Amparo Gamboa, Araceli Jimeno, Katherine Rollins, Sarah Crow, and Michael Paul providedinvaluable administrative support, including the collection of remittance fee data from allover the world.The report team held consultations in July 2005 in Accra, Brussels, Geneva, London, andParis. Thanks are due to Haleh Bridi, Barbara Genevaz, Carlos Braga, Sonia Plaza, MichelleBailly, and other colleagues in these country offices for efficiently and enthusiastically arrangingconsultations with several international, academic, financial, and non-governmental institutions.Thanks are also due to the International Organization for Migration for assistance with organizingconsultations in Geneva and to the International Labour Organization, the Global Commissionon International Migration, the European Commission, and the Commonwealth Secretariatfor participating in consultations and providing useful feedback.This report also benefitted from the comments of the Bank’s executive directors made at aninformal board meeting on October 20, 2005.Marilou Uy, Alan Gelb, Jeff Lewis, Amar Bhattacharya, Shaida Badiee, Robert Keppler, andMisha Belkindas provided guidance and encouragement to the team at various stages. Dorota A.Nowak managed production and dissemination activities by DECPG. Steven Kennedy’s contributionas an editor is gratefully acknowledged. Book design, editing, and production were coordinatedby the World Bank Office of the Publisher.x

OverviewTHE THEMES OF this year’s Global EconomicProspects are international remittancesand migration, their economicconsequences, and how policies canincrease their role in reducing poverty. Internationalmigration can generate substantial welfaregains for migrants and their families andfor the countries involved (countries of originand destination). The money that migrantssend home—remittances—is an importantsource of extra income for migrants’ familiesand for developing countries: in aggregate, remittancesare more than twice as the size of internationalaid flows. However, migrationshould not be viewed as a substitute for economicdevelopment in the origin country—development ultimately depends on sounddomestic economic policies.Over the past two decades, barriers tocross-border trade and financial transactionshave fallen significantly, while barriers to thecross-border movement of people remainhigh. Despite its economic benefits, migrationremains controversial and, for some people,threatening. In part, this is because migration,like trade and capital movements, has distributionalconsequences, whereby net gains forsociety may mask important losses for someindividuals and groups. But migration alsosparks resistance because the movement ofpeople has economic, psychological, social,and political implications that the movementof goods or money do not.This publication has two goals. The first isto explore the gains and losses from internationalmigration from the perspective of developingcountries, with special attention to themoney that migrants send home. The secondgoal is to consider policy initiatives that couldimprove the developmental impact of migration,again with particular attention to remittances.Our focus (for economic purposes) ison international migration from developingcountries to high-income countries. Despitetheir importance, internal migration, migrationamong developing countries, and the politicaland social impacts of migration arebeyond the scope of this work.It is important to keep in mind three basicprinciples. First, migration is a diverse phenomenon,and its economic impact in onelocation or another depends heavily on the particularcircumstances involved. Second, basicdata on migration and remittances are lacking,so predicting the impact of policy changes canbe problematic. This underlines the need forbetter data and more research. Third, migrationhas social and political implications thatmay be just as important as the economicanalysis provided here. These are ably andcomprehensively discussed in the recent reportof the United Nations’ Global Commission onInternational Migration. For all of these reasons,the analysis and policy recommendationsfor migration must remain qualified. Thisreport draws conclusions where they can bexi

O V E R V I E Wsupported by adequate data and points to anagenda for research where they cannot.Global economic prospectsThe slowdown among industrial economiesthat began in the second half of 2004 continuedin 2005, with GDP growth expected tocome in at 2.5 percent, down from 3.1 percentthe year before. The pace of the expansion inthe high-income countries is forecast to increaseslightly over the next two years, with accelerationin Europe offsetting a modest slowingin Japan and stable growth in the UnitedStates. Economic activity also slowed in developingeconomies during 2005. Higher oilprices, domestic capacity constraints, andslower demand for their exports brought GDPgrowth down from a very strong 6.8 percent in2004 to an estimated 5.9 percent this year.While GDP growth has remained robust,higher oil prices have sharply slowed real incomegrowth among oil importers from 6.4 to3.7 percent. Looking forward, continued highoil prices, coupled with inflationary pressures,are expected to restrain growth in most developingcountries over the next two years. Nevertheless,GDP in these economies shouldexpand by around 5.5 percent—much morequickly than during the past two decades.This relatively positive outlook is subject toimportant downside risks. The outlook for oilprices is particularly uncertain. Low excess capacityhas introduced a risk premium into oilprices and will make it more difficult to containthe impact of a future supply shock,should one arise. As a result, a significant supplydisruption could slow global growth, withlarge negative consequences for global economicprospects. The future path of interestrates, which despite recent increases are stilllow, is another source of uncertainty. Persistentglobal imbalances, signs of rising inflation,and concerns about the sustainability ofgovernment finances in industrialized countriesare all factors that could push rateshigher and possibly provoke a much more seriousslowdown.The impact of remittancesand migrationThe impact on migrantsThe bulk of the economic gains from migrationaccrue to migrants and their families, andthese gains are often large. Wage levels (adjustedfor purchasing power) in high-incomecountries are approximately five times those oflow-income countries for similar occupations,generating an enormous incentive to emigrate.Moreover, to the extent that migrants devotea portion of their income to remittances, thegains are even greater. Essentially migrants canearn salaries that reflect industrial-countryprices and spend the money in developingcountries, where the prices of nontraded goodsare much lower. Migrants, however, incur substantialcosts, including psychological costs,and immigrants (particularly irregular migrants)sometimes run high risks; many sufferfrom exploitation and abuse. The decision tomigrate is often made with inaccurate information.Given the high costs of migration—including the risks of exploitation and the exorbitantfees paid to traffickers—the netbenefit in some cases may be low or even negative.There are costs, too, for family membersleft behind—particularly children—althoughthese costs must be balanced against the benefitsof the extra income that migrants sendback home to their families.The impact on destination countriesDestination countries can enjoy significant economicgains from migration. The increasedavailability of labor boosts returns to capitaland reduces the cost of production. A modelbasedsimulation performed for this study indicatesthat a rise in migration from developingcountries sufficient to raise the labor force ofhigh-income countries by 3 percent could boostincomes of natives in high-income countries by0.4 percent. In addition, high-income countriesmay benefit from increased labor-market flexibility,an increased labor force due to lowerprices for services such as child care, and perhapseconomies of scale and increased diversity.xii

O V E R V I E WNevertheless, there are losers within destinationcountries. Some workers may see an erosionof wages or employment, although this effectis found to be small in most empiricalstudies. In the model-based simulation of theimpact of increased migration, earlier migrantssuffer significant income losses, while the impacton natives’ wages is small. (The differentialimpact is reduced if foreign-born workers areviewed as closer substitutes for natives.) Easingrules that limit labor-market flexibility, andstrengthening institutions that provide educationand training, will help workers displacedby immigration (both natives and resident migrants)to find work. Note that the simulationresults are not intended to incorporate all of theeconomic impacts of migration, nor do theycapture important social and political implications.The goal is not to forecast the overall impactof increased migration, but rather to giveus insights into the economic gains that mightbe expected from changes in policy or circumstances,as well as insights into the channelsthrough which migration affects welfare.The impact on origin countriesMigration also generates economic benefits fororigin countries, the largest being remittances.International remittances received by developingcountries—expected to reach $167 billionin 2005—have doubled in the past five years asa result of (a) the increased scrutiny of flowssince the terrorist attacks of September 2001,(b) changes in the industry that support remittances(lower costs, expanding networks), (c)improvements in data recording, (d) the depreciationof the dollar (which raises the dollarvalue of remittances denominated in other currencies),and (e) growth in the migrant stockand incomes. However, records still underestimatethe full scale of remittances, because paymentsmade through informal, unrecordedchannels are not captured. Econometric analysisand available household surveys suggest thatunrecorded flows through informal channelsmay conservatively add 50 percent (or more) ofrecorded flows. Several countries with significantmigrant populations do not report data onremittances at all, even those sent through formalchannels, or they report remittances underother balance of payments entries.Despite the prominence given to remittancesfrom developed countries, South-Southremittance flows make up between 30 and 45percent of total remittances received by developingcountries, reflecting the fact that overhalf of migrants from developing countriesmigrate to other developing countries.While the impact of remittances on growthis unclear, remittances do play an importantrole in reducing the incidence and severity ofpoverty (with no significant effect on incomeinequality). Remittances directly increase theincome of the recipient and can help smoothhousehold consumption, especially in responseto adverse events, such as crop failure or ahealth crisis. In addition to bringing the directbenefit of higher wages earned abroad, migrationhelps households diversify their sources ofincome (and thus reduce their vulnerability torisks) while providing a much needed source ofsavings and capital for investment. Remittancesappear to be associated with increasedhousehold investments in education, entrepreneurship,and health—all of which have a highsocial return in most circumstances.Measuring the poverty impact of remittancesis difficult: data are scarce, and calculatingthe income gains from remittances requiresassumptions concerning what migrantswould have earned if they had stayed at home.Careful analyses of the available householdsurvey data indicate that remittances havebeen associated with declines in the povertyheadcount ratio in several low-incomecountries—by 11 percentage points inUganda, 6 in Bangladesh, and 5 in Ghana, forexample. In Guatemala, remittances may havereduced the severity of poverty by 20 percent.Cross-country regressions and simulationsalso indicate that increases in remittances helpto reduce the incidence of poverty.By generating a steady stream of foreignexchange earnings, remittances can improve acountry's creditworthiness for external borrowingand, through innovative financingxiii

O V E R V I E Wmechanisms (such as securitization of remittanceflows), they can expand access to capitaland lower borrowing costs. While large andsustained remittance inflows can contribute tocurrency appreciation, this outcome may beless severe than it is in the case of natural resourceearnings, because remittances are distributedmore widely and may avoid exacerbatingstrains on institutional capacity that areoften associated with natural resource booms.Migration has economic implications for origincountries beyond remittances. The small sizeof migration flows relative to the labor force suggeststhat the effects of South–North migrationon working conditions for low-skilled workersin the developing world as a whole must be smallas well. However, in some countries low-skilledemigration can raise demand for the remaininglow-skilled workers (including poor workers) atthe margin, leading to some combination ofhigher wages, lower unemployment, less underemployment,and greater labor force participation.Thus low-skilled emigration can offer avaluable safety valve for insufficient employmentat home. In the long run, however, developingcountry policies should aim to generateadequate employment and rapid growth, ratherthan relying on migration as an alternative to developmentopportunities.High-skilled emigration has more compleximplications. Like low-skilled migration, itcan greatly benefit migrants and their familiesand help relieve labor market pressures. However,a well-educated diaspora can improve accessto capital, technology, information, foreignexchange, and business contacts for firmsin the country of origin. The return of expatriatesand the maintenance of close contactswith high-skilled emigrants have played animportant role in the transfer of knowledge toorigin countries. At the same time, large outflowsof high-skilled workers can reducegrowth in the origin country for these reasons:(a) the productivity of colleagues, employees,and other workers may suffer because theylose the opportunity for training and mutuallybeneficial exchanges of ideas; (b) the provisionof key public services with positive externalities,such as education and health (particularlyfor the control of transmissible diseases), maybe impaired; (c) opportunities to achieveeconomies of scale in skill-intensive activitiesmay be reduced; (d) society loses its return onhigh-skilled workers trained at public expense;and (e) the price of technical servicesmay rise. Highly educated citizens, if theystayed in their countries, could help to improvegovernance, improve the quality of debateon public issues, encourage education ofchildren, and strengthen the administrativecapacity of the state—contributions thatwould be lost through high-skilled emigration.It is impossible to reliably estimate the netbenefit, or cost, to origin countries of highskilledemigration because data are limitedand a myriad of individual country circumstancesenter into the calculus of that benefitor loss. We can only offer two rough observations,which reflect the wide variation in highskilledemigration rates among countries:• Very high rates of high-skilled emigrationare found in countries that representa small share of the population of the developingworld. Many of those countrieshave poor investment climates that likelylimit the productive employment of highskilledworkers. Of course, the loss ofhigh-skilled workers may aggravate thepoor investment climate and limit thepotential benefits of economic reform.• Some countries find it difficult to provideproductive employment for manyhigh-skilled workers because of theirsmall economic scale or because misguidededucational policies have resultedin a large supply of university graduatesfor whom no suitable jobs exist.Policies to improve thedevelopmental impact ofremittances and migrationMigration policiesGreater emigration of low-skilled emigrantsfrom developing to industrial countries couldxiv

O V E R V I E Wmake a significant contribution to poverty reduction.The most feasible means of increasingsuch emigration would be to promotemanaged migration programs between originand destination countries that combine temporarymigration of low-skilled workers withincentives for return. Temporary programshave several advantages, and some disadvantages,relative to permanent migration. Fromthe perspective of the destination country,managed, temporary migration programs easesocial tensions by limiting permanent settlement;they limit the potential burden on publicexpenditures because immigrants are guaranteeda job and are less likely to bringdependents; and they allow for controlledvariation of the number of immigrants in responseto changes in labor-market conditions,thus limiting adverse effects on low-skilled nativeworkers. However, temporary migrationcan be less efficient than permanent migrationfor firms in destination countries because ofhigh training costs. From the origin-countryperspective, managed, temporary migrationmay be the only means of securing deliberateincreases in low-skilled emigration and mayraise remittances and improve the skills of returningworkers. On the other hand, managedmigration programs do not guarantee futureaccess to labor markets (and thus to remittances),because it is easier for destinationcountries to suspend temporary programsthan to expel immigrants. Overall, however,such programs do represent a feasible approachto capturing the efficiency gains fromlabor migration.Origin countries that are adversely affectedby high-skilled emigration face challenges inmanaging it better. Service requirements foraccess to publicly financed education can beevaded and are likely to discourage return;and proposals for the taxation of emigrants tothe benefit of the origin country have made littleprogress. Origin countries can help to retainkey workers by improving working conditionsin public employment and by investingin research and development. Origin countriescan also take steps to encourage educated emigrantsto return by identifying job opportunitiesfor them, cooperating with destinationcountries that have programs to promote return,permitting dual nationality, and helpingto facilitate the portability of social insurancebenefits.By providing authoritative information onmigration opportunities and risks, governmentscould help avoid unfortunate, costly-toreversemigration decisions and limit the abuseof vulnerable migrants. Labor recruiters canplay a valuable role in promoting migration,but emigrants’ lack of information often enablesrecruiters to capture the lion’s share of therents generated by constraints on immigrationand imperfect information. Origin countrieswith effective public sector institutions mightconsider the regulation of recruitment agentsto limit rents and improve transparency.Remittance policiesGovernments in destination and origin countriescan sharpen the developmental impact ofremittances through the application of appropriatepolicies. Access of poor migrants andtheir families to formal financial services forsending and receiving remittances could beimproved through public policies that encourageexpansion of banking networks, allowdomestic banks from origin countries to operateoverseas, provide identification cards tomigrants, and facilitate the participation ofmicrofinance institutions and credit unions inproviding low-cost remittance services. Remittances,in turn, can be used to support financialproducts—housing and consumer loansand insurance—for poor people.A second set of promising policies couldimprove competition in the remittance transfermarket and thereby lower fees. The priceof remittance transactions is often unnecessarilyhigh for the small transfers typically madeby poor migrants. The cost of such transactionsis often well below the fees paid by customers.Reducing transaction charges increasesthe disposable income of poormigrants and increases their incentives toremit, because the net receipts of recipientsxv

O V E R V I E Wincrease. The overall result would be strongerremittance flows to developing countries.Competition among providers of remittanceservices could be increased by lowering capitalrequirements on remittance services and openingup postal, banking, and retail networks tononexclusive partnerships with remittanceagencies. Disseminating data on remittancefees in important remittance corridors and establishinga voluntary code of conduct for deliveringfair-value transfers would improvetransparency and reduce prices for remittancetransactions. Governments could help reducecosts by supporting the introduction of moderntechnology in payment systems. Alleviatingliquidity constraints by providing a credit lineeither to the sender or the recipient, based onpast remittance activity, would enable sendersto take advantage of the lower fee rates availableonly for larger remittances. Reducing exchange-ratedistortions could also lower thecost of remittance transactions. Finally, regulatoryregimes need to strike a better balance betweenpreventing financial abuse and facilitatingthe flow of funds through formal channels.Several origin countries have attempted toimprove the developmental impact of remittancesby introducing incentives to increaseflows and to channel them to more productiveuses. Such policies are more problematic thanefforts to expand access to financial servicesor reduce transaction costs, because they poseclear risks. Tax incentives to attract remittanceinflows, for example, may also encourage taxevasion, while matching-fund programs to attractremittances from migrant associationsmay divert funds from other local funding priorities.Efforts to channel remittances to investment,meanwhile, have met with little success.Fundamentally, remittances are privatefunds that should be treated like other sourcesof household income. Efforts to increase savingsand improve the allocation of expendituresshould be accomplished through improvementsin the overall investment climate,rather than by targeting remittances. Similarly,because remittances are private funds, theyshould not be viewed as a substitute for officialdevelopment aid.Organization of this studyAs is customary in this report, chapter 1reviews recent developments in andprospects for the global economy and theirimplications for developing countries. Chapter2 uses a model-based simulation to evaluatethe potential global welfare gains and distributionalimpact from a hypothetical increase of 3percent in high-income countries’ labor forcecaused by migration from developing countries.Chapter 3 surveys the economic literatureon the benefits and costs of migration for migrantsand their countries of origin, focusingon economically motivated migration from developingto high-income countries. We thenturn to remittances, the main theme of the report.Chapter 4 investigates the size of remittanceflows to developing countries, the use offormal and informal channels, the role of governmentpolicies in improving the developmentimpact of remittances, and, for certain countries,their macroeconomic impact. Chapter 5addresses the impact of remittances at thehousehold level, in particular their role in reducingpoverty, smoothing consumption, providingworking capital for small-scale enterprises,and increasing household expendituresin areas considered to have a high social value.The last chapter investigates policy measuresthat could lower the cost of remittance transactionsfor poor households and measures tostrengthen the financial infrastructure supportingremittances.xvi

AbbreviationsADBAMLAML/CFTARTEPATCBISBoABOPCECEICEMLACEPALCES-IFOCGAPCPICPSSDDPDFIDDHSDPRsDPSECAECOSOCEUEVFATFAsian Development BankAnti-money launderingAnti-money laundering/countering financing of terrorismAsian Regional Team for Employment Protection (ILO)Agreement on Textiles and ClothingBank for International SettlementsBank of AmericaBalance of paymentComité des établissements de crédit et des enterprises, d’investissementCenter for Latin American Monetary Studies (Centro de EstudiosMonetarios Latino Americanos)Economic Commission for Latin America and the Caribbean (ComisiónEconómico para América Latina y el Caribe)Center for Economic Studies at the University of Michigan—Institutefor Economic ResearchConsultative Group to Assist the PoorestConsumer price indexCommittee on Payment and Settlement SystemsDevelopment data platformDepartment for International DevelopmentDepartment for Homeland SecurityDiversified payment rightsDeferred payment schemeEurope and Central AsiaUN Economic and Social CouncilEuropean UnionEquivalent variationFinancial Action Task Force against Money Launderingxvii

A B B R E V I A T I O N SFDCFDIFFMCFinCENFISDLGATSGDPGEGTAPHIV/AIDSHSBCHTAsIADBIBMICICIICMCIFADILOIMFIOMIRCAIRDIRnetLACLDCsMCICMDGMIDSAMFIsMFNMICRMIDAMIDWAMPGMSBMTOsNAFTAFirst Data CorporationForeign direct investmentFirst Financial Management CorporationFinancial Crimes Enforcement NetworkSocial Investment Fund for Local Development (Fondo de InversiónSocial para el Desarrollo)General Agreement on Trade in ServicesGross domestic productGeneral equilibriumGlobal Trade Analysis ProjectHuman immunodeficiency virus/acquired immune deficiency syndromeHong Kong Shanghai Bank CorporationHometown associationsInter-American Development BankInternational Business MachinesIndustrial Credit and Investment Corporation of IndiaInternational Catholic Migration CommissionInternational Fund for Agriculture and DevelopmentInternational Labour OrganizationInternational Monetary FundInternational Organization for MigrationImmigation Reform and Control ActInstitut de recherche pour le développementInternational Remittance NetworkLatin American and the CaribbeanLeast-developed countriesMatricula Consular identification cardMillennium Development GoalMigration for Development in Southern AfricaMicrofinance institutionsMost-favored nationMagnetic ink character recognitionMigration for Development in AfricaMigration for Development in Western AfricaMigration Policy GroupMoney service businessMoney transfer operatorsNorth American Free Trade Agreementxviii

A B B R E V I A T I O N SNBERNGOsODAOECDOFACOPECOLSPADFPPPPRIOPROFECOR&DSADCSPVTDSTSGUAEUCSDUNUNDPUNESCAPUNESCO–ICSUUSAIDUSDWEOWOCCUWTONational Bureau of Economic ResearchNon-governmental organizationOfficial development aidOrganisation for Economic Co-operation and DevelopmentOffice of Foreign Assets ControlOrganization of the Petroleum Exporting CountriesOrdinary least squaresPan-American Development FoundationPurchasing power parityInternational Peace Institute, OsloProcuraduría Federal del ConsumidorResearch and developmentSouthern African Development CommunitySpecial purpose vehicleTelephone and data systemsTechnical Sub-Group on the Movement of PersonsUnited Arab EmiratesUniversity of California, San DiegoUnited NationsUnited Nations Development ProgrammeUnited Nations Economic and Social Commission for Asia and thePacificUnited Nations Educational, Scientific and CulturalOrganization–International Council of Scientific UnionsUnited States Agency for International DevelopmentUS dollarWorld Economic OutlookWorld Council of Credit UnionsWorld Trade Organizationxix

1Prospects for the Global EconomyFollowing very strong growth, the worldeconomy slowed in late 2004 and into 2005 asoutput began to push against capacity constraints.High oil prices cut into the incomes ofoil importers, but the expansion remainedstrong, partly because of favorable conditionsin financial markets, including still low inflation,interest rates, and interest-rate spreads.Tightness in the oil market, the threat of evenhigher fuel prices, and the possibility that interestrates may rise pose major threats to theexpansion.Slower but still strong growthWorld GDP is estimated to have increased by3.2 percent in 2005, down from 3.8 in 2004.Growth is projected to be stable in 2006, beforestrengthening somewhat in 2007. Theslowdown that began in the second half of2004 was experienced throughout the industrializedworld, with growth in Europe stillunderperforming its potential. In contrast, theeconomies of the United States and Japan, despitehaving slowed, are expanding at close totheir maximum sustainable rates.Among large developing economies, GDPin 2005 continued to expand rapidly in Chinaand India (in excess of 9 percent and about7 percent, respectively), but slowed in Russiaas growth in oil production weakened. Highoil prices, in combination with domestic capacityconstraints and slower import demandfrom high-income countries, are estimated tohave reduced growth among oil-importing developingcountries from 6.9 percent to 6.1percent. In terms of real incomes, the slowdownwas much sharper—from 6.4 percent to3.7 percent. Despite still growing oil revenues,reduced opportunities to expand productionin the petroleum sector meant that outputgrowth in oil-exporting developing countriesalso eased, from 6.6 percent to 5.6 percent.During 2006 the expansion among highincomecountries is projected to be stable, atabout 2.5 percent, before picking up a bit in2007. This reflects a combination of improvedperformance in Europe and stable growth inthe United States and Japan. In the UnitedStates, higher oil prices and tighter monetarypolicy are expected to offset the positive stimulusto growth from past depreciations. Theprojected pickup in Europe occurs despite asignificant drag on growth from high oil priceswhose effects are expected to be more thanoffset by low interest rates, pent up investmentdemand, and a dissipation of most of the negativeconsequences following the euro’s realeffectiveappreciation. In Japan, strengtheningdomestic demand and supportive macroeconomicpolicies should enable growth to remainclose to potential, despite high oil prices.Growth in developing economies is projectedto slow modestly from an estimated5.9 percent in 2005 to 5.5 percent by 2007.In East and South Asia, the expansion isprojected to moderate somewhat but remain1

GLOBAL ECONOMIC PROSPECTS 2006very strong, particularly in China and India.In the Middle East and in both North andSub-Saharan Africa, strong oil revenuesshould buoy internal demand among oil exportersand partially offset capacity constraintsthat will slow production growth. Theprojected easing of growth in Latin Americaand the Caribbean reflects weaker non-oilcommodity prices as well as a return to trendgrowth in several countries that reboundedvery strongly in 2004. In Europe and CentralAsia, the waning of the growth bonus followingEU accession and capacity constraints inoil-producing countries are expected to contributeto a modest slowing of the expansion.Tight commodity marketsWeaker global growth should reduce thestrain in non-oil commodity markets. Alreadythere are signs of stabilization, and even ofdecline, in the prices of agricultural products,where supply has responded to high prices.Metals and shipping prices also show signs ofeasing, although to a lesser extent.In oil markets, the projected slowdown isnot expected to be sufficient to generate a substantialeasing of prices. While crude oil supplyis growing marginally faster than demand,supply conditions are expected to remain tight.As a result, crude oil prices, which currentlyembody a large risk premium, are not expectedto fall rapidly. The baseline assumes that nomajor supply disruptions occur and that therewill be a gradual decline in oil prices toward$40 per barrel by 2010. This implies an averageprice of $56 for a barrel of oil in 2006 and$52 in 2007.Future spikes in oil prices form a potentialrisk to global prospects. A price hike generatedby a sustained negative supply shockwould be particularly disruptive, because outputwould be constrained directly by the reducedavailability of oil and petroleum-basedinputs. This would be in contrast to the recentpast, when prices rose in the context of rapidlygrowing supply. A supply shock that reducedoil deliveries by 2 million barrels per daycould push prices to more than $90 a barrelfor more than a year, resulting in a 1.5 percentreduction in global growth by the second yearfollowing the shock. The terms-of-trade impactfor low-income oil-importing economieswould reduce incomes in these countries bymore than 4 percent of their GDP (much morethan for high-income countries) because theireconomies are relatively oil intensive, and becausea supply shock–induced increase in oilprices is unlikely to be accompanied by highernon-oil commodity prices.Global imbalances remain an issueGlobal current account imbalances and theU.S. current account deficit (which is expectedto exceed $750 billion in 2005) remainimportant medium-term problems. Duringlate 2004 and early 2005 tensions easedsomewhat. Rising interest rate differentialsrelative to European short- and long-term assetsmade private sector purchases of dollar–denominated assets more attractive. As a result,the dollar appreciated some 2.5 percentin real-effective terms during the first sevenmonths of 2005, and reserve accumulation byforeign central banks became less important inthe financing of the current account deficit.This respite appears to have been shortlived.To some extent, the increased privateflows represented a one-off portfolio adjustmenttoward U.S. assets by investors. Beginningin the second quarter of 2005, the flowsdiminished, and the dollar faced reneweddownward pressure. As a result, foreign reserveaccumulation once again became a criticalcomponent in the financing of the U.S. currentaccount deficit, restoring the risk that achange in behavior on the part of foreign centralbankers could prove destabilizing. Recentdecisions by China and Malaysia to widen therange of currencies to which their own currenciesare pegged could help ease future pressures,especially if the scope for appreciationincluded in the regime is exercised in practice.Globally, policy should continue to focus onincreasing public and private savings in deficit2

PROSPECTS FOR THE GLOBAL ECONOMYcountries and increasing spending (notably oninvestment goods) in surplus countries.Low interest rates are a sourceof uncertaintyThe future path of long-term interest rates andspreads, which have been at historically lowlevels for an extended period, is an importantuncertainty. A number of factors have helpedmaintain interest rates at low levels, includingseveral years of very loose monetary policythroughout the developed world; increasedaging-related savings in Europe; balance-sheetconsolidation in the United States andAsia; and a low inflationary environment—thanks, in part, to increased competition followingthe entry into global markets of Chinaand members of the former Soviet bloc. Mostof these factors are temporary and are expectedto gradually abate, resulting in a steady rise inlong-term rates in the baseline. Indeed, yieldson 10-year U.S. Treasuries have risen 50 basispoints since September.However, these temporary factors couldcontinue to hold sway, reversing or bringingto a halt the recent increase in long-term rates(as they have in the past). This would promptstronger-than-projected demand, but also exacerbatecapacity constraints. As a result, oilprices could get pushed higher, which wouldprovoke a more brutal inflationary cycle, andultimately, a recession.Alternatively, these forces could dissipatemore rapidly, causing long-term interest ratesto rise more quickly toward long-term equilibriumlevels, which would provoke a more pronouncedslowdown. While not the most likelyscenario, the recent rise in long-term yieldsand inflation suggest that a higher interestratescenario is a real possibility.Finally, this environment of slowinggrowth and global imbalances raises the riskof rising protectionism. In this regard, policymakersneed to make a concerted effort to ensurethat the Doha round reaches a successfulconclusion so that developing countries specializingin the export of agricultural productscan benefit from trade liberalization in thesame way that other countries have profitedfrom freer trade in the manufacturing and rawmaterials sectors.Global growthThe global economy slowed markedly in2005, but still continued to expand at anestimated 3.2 percent pace, compared with3.8 percent in 2004 (table 1.1). The slowdownwas widespread, reaching virtually every economicregion. It was precipitated by higher oilprices, resource-sector capacity constraints,tightening monetary policy in the UnitedStates, and in some countries, the maturationof the investment cycle following a year ofvery fast growth.Outturns and prospects inhigh-income countriesGrowth among industrialized economies in2005 is estimated at 2.5 percent, substantiallylower than the 3.1 percent recorded the yearbefore. Industrial production and trade flowsamong high-income countries were particularlyweak. Growth rates of the former declinedfrom over 5 percent in mid-2004 to lessthan 1.5 percent in the middle of 2005 (figure1.1). High oil prices, rising short-term interestrates, and an unusually disruptive hurricaneseason 1 slowed growth in the United States toFigure 1.1 Industrial production% change, monthly, year over year151050Source: World Bank.Developing countriesDeveloping5Worldcountries(excluding China)(excluding China)101997 1998 1999 2000 2001 2002 2003 2004 20053

GLOBAL ECONOMIC PROSPECTS 2006Table 1.1 The global outlook in summaryPercentage change from previous year, except interest rates and oil price2003 2004 2005e 2006f 2007fGlobal ConditionsWorld trade volume 5.9 10.2 6.2 7.0 7.3Consumer pricesG-7 countries a,b 1.5 1.7 2.2 2.0 1.7United States 2.3 2.7 3.4 3.0 2.4Commodity prices (USD terms)Non-oil commodities 10.2 17.5 11.9 5.9 6.3Oil price (US$ per barrel) c 28.9 37.7 53.6 56.0 51.5Oil price (percent change) 15.9 30.6 42.1 4.5 8.0Manufactures unit export value d 7.5 6.9 2.4 2.4 2.1Interest rates$, 6-month (percent) 1.2 1.7 3.8 5.0 5.2€, 6-month (percent) 2.3 2.1 2.2 2.1 2.8Real GDP growth eWorld 2.5 3.8 3.2 3.2 3.3Memo item: world (PPP weights) f 3.9 5.0 4.4 4.3 4.4High income 1.8 3.1 2.5 2.5 2.7OECD countries 1.8 3.0 2.4 2.5 2.7Euro area 0.7 1.7 1.1 1.4 2.0Japan 1.4 2.6 2.3 1.8 1.7United States 2.7 4.2 3.5 3.5 3.6Non-OECD countries 3.7 6.3 4.3 4.2 4.0Developing countries in 5.5 6.8 5.9 5.7 5.5East Asia and Pacific 8.1 8.3 7.8 7.6 7.4Europe and Central Asia 6.1 7.2 5.3 5.2 5.0Latin America and Caribbean 2.1 5.8 4.5 3.9 3.6Middle East and N. Africa 5.2 4.9 4.8 5.4 5.2South Asia 7.9 6.8 6.9 6.4 6.3Sub-Saharan Africa 3.6 4.5 4.6 4.7 4.5Memorandum itemsDeveloping countriesexcluding transition countries 5.3 6.8 6.1 5.8 5.6excluding China and India 4.1 6.0 4.9 4.7 4.6Note: PPP purchasing power parity; e estimate; f forecast.a. Canada, France, Germany, Italy, Japan, the UK, and the United States.b. In local currency, aggregated using 1995 GDP weights.c. Simple average of Dubai, Brent, and West Texas Intermediate.d. Unit value index of manufactured exports from major economies, expressed in U.S. dollars.e. GDP in 1995 constant dollars;1995 prices and market exchange rates.f. GDP measured at 1995 PPP weights.an estimated 3.5 percent, compared with4.2 percent the year before. The slowdownwas not as marked as it could have been,because low long-term interest rates boosteddomestic demand, and the cumulative effectof past dollar depreciations improved netexports.In Europe, the growth slowdown was lesspronounced, but the expansion, at an estimated1.2 percent (1.1 percent in the euro zone), wasmuch weaker. The relatively low oil-intensityof European economies and relaxed macroeconomicpolicy stance help explain why theslowdown in Europe was not more pronounced.In Japan, GDP is estimated to haveincreased 2.3 percent. Rising domestic demandand household incomes, as a result oftighter labor market conditions and reduced4

PROSPECTS FOR THE GLOBAL ECONOMYindustrial restructuring, compensated formuch slower Chinese import demand.Looking forward, the increase in oil pricesobserved in 2005 is expected to slow globalgrowth by about one quarter of a percentagepoint in 2006, compared with what it wouldhave been had prices remained stable. In theUnited States, the pace of the expansion isprojected to remain broadly stable, becausethe negative effects of further expected increasesin interest rates and high oil prices willbe partly offset by a deficit-financed pickup inpost-hurricane investment and additional increasesin the contribution of the external sectorto growth. In Europe, economic activity isprojected to accelerate despite a significantdrag on growth from high oil prices, becauseof low interest rates, pent up investmentdemand, and a dissipation of most of thenegative consequences following the euro’sreal-effective appreciation. Meanwhile, inJapan, the negative consequences of higher oilprices are expected to be substantially offsetby strengthening domestic demand and continuedsupportive macroeconomic policies.Developing economy outturnsand prospectsDespite a slowdown of almost a full percentagepoint, growth in developing economies remainedvery robust, at an estimated 5.9 percentin 2005 (figure 1.2). In part this reflectsthe strong performance of China and India,where output continued to expand at rapidrates (in excess of 9 percent and about 7 percent,respectively). The slowdown among theoil-importing countries (excluding China andIndia) was sharper, from 5.6 percent to4.3 percent. 2 At the same time, dwindlingspare capacity in the petroleum sector causedgrowth in oil-exporting developing countriesto ease from 6.6 percent to 5.7 percent, eventhough oil revenues continued to rise.High oil prices, rising interest rates, andbuilding inflationary pressures are expected torestrain growth in most developing regions in2006 and 2007 (figure 1.3). As a group, however,low- and middle-income countries shouldFigure 1.2 A sharp slowdown% change, GDP76543211997Developing countriesSource: World Bank.again outperform high-income economies by awide margin through 2007.Regional outlooksWorld19981999200020012002Projection20032004200520062007Detailed descriptions of economic developmentsin developing regions can be foundin the Regional Outlooks section of http://www.worldbank.org/globaloutlook.The economies of the East Asia andPacific region continued to expand rapidlyin 2005. Regional GDP is estimated to haveincreased by 7.8 percent, down from 8.3 percentin 2004. Growth in China remainedvery strong—despite a substantial slowing inboth private consumption and investmentdemand—because exports continued togrow rapidly, and import growth declined byhalf. China appears to have been a majorbeneficiary of the expiration of quotas ontextiles (see the global trade discussionbelow), which contributed to rapid exportgrowth in the first half of the year. Sincethen, the re-imposition of quotas by theUnited States and the European Union (EU)have attenuated this positive force. For othercountries in the region, the slowdown inChinese imports, weak global high-tech demand,and elevated oil prices have translatedinto reduced export growth, rapidly rising5

GLOBAL ECONOMIC PROSPECTS 2006Figure 1.3 Regional growth% change, GDP9876200620052004 2007200354321East Asia andthe PacificEurope andCentral AsiaLatin Americaand theCaribbeanMiddle Eastand NorthAfricaSouth AsiaSub-SaharanAfricaOil importersOil importers,except ChinaSource: World Bank.producer prices, and a deterioration of currentaccount balances.Even higher oil prices on average in 2006, 3the longer-term implications of reduced investmentlevels of China, and a tightening ofmonetary policy are expected to slow regionalgrowth to 7.6 and 7.4 percent in 2006 and2007, respectively. The changes in the currencyregimes of China and Malaysia are notexpected to have a major impact on growth.Nevertheless, as discussed below, theseregimes should improve financial stabilityboth domestically and internationally.Economic activity in the Europe and CentralAsia region decelerated sharply in 2005,with GDP growing by an estimated 5.3 percent,down from 7.2 percent in 2004. Slowerincreases in oil production, a peaking of theinvestment cycle (especially among economiesthat recently joined the EU), and less robustworld demand for the region’s exports contributedto the slowdown, which was particularlyintense in a number of the largereconomies of the region. Russia deceleratedfrom 7.2 percent to 6.0 percent; Ukraine from12.1 percent to 4.4 percent; Polandfrom 5.4 percent to 3.5 percent; and Turkeyfrom 8.9 percent to 4.8 percent.Higher oil prices constrained domesticdemand in oil-importing countries, but oilrevenues in oil-exporting countries helped offsetmuch slower growth in the oil sector itself.Reflecting these capacity constraints and thevery strong growth recorded last year, inflationarypressures have built up in many countriesin the region, notably Russia. Turkey,where improved macroeconomic policy haspushed inflation below 10 percent, representsan important exception. The expected accelerationof demand in Europe, continued high oilprices—which for many countries in the regionare a positive factor—and additionalgains in European market share, suggest thatgrowth for the region as a whole should remainrelatively stable—at about 5 percent in2006 and 2007, which is close to the region’spotential growth rate.Economic activity in Latin America and theCaribbean is estimated to have increased bysome 4.5 percent during 2005, substantiallyslower than the 5.8 percent recorded in 2004but much faster than the region’s 0.4 percentaverage growth rate during the precedingthree years. Supply constraints and tight monetarypolicy are estimated to have slowed GDPgrowth in Brazil to some 3.8 percent (downfrom 4.9 percent in 2004), while in Mexicofive fewer working days in 2005 than in 2004are expected to contribute to a significantslowing. 4 Excluding these countries, regional6

PROSPECTS FOR THE GLOBAL ECONOMYgrowth in 2005 is estimated at a robust5.9 percent, boosted by both strong worlddemand for the region’s exports (particularlyoil, coffee, and copper, which account for 65percent of the regions’ commodity exports)and low interest rates. Domestic factors thatcontributed to the strong performance includepast efforts to open the region up to internationaltrade, more responsible budget policy,the introduction of more flexible exchangerate regimes, and lower inflation.Slower global growth is already easing tensionsin the non-oil commodity markets thathave driven the recovery in the Latin Americaand Caribbean region, and this trend is expectedto continue. Moreover, while manycountries in the region benefit from high oilprices, many others, particularly those in theCaribbean, are heavily oil dependent and facesubstantial income losses. 5 As a result, regionalGDP growth is projected to decline to3.6 percent by 2007.High oil prices and strong oil demand continueto be key drivers for the economies ofthe Middle East and North Africa, where GDPis estimated to have increased by 4.8 percentin 2005. Very high oil revenues generateddouble-digit advances in public spending,which have helped to increase GDP in oilproducingeconomies by an estimated 5.4 percent.Strong demand from these economiesspilled over to the labor-abundant economiesof the region through higher remittances andincreased intraregional tourism flows. However,weak growth in Europe, high oil bills,and a one-off negative effect from the removalof quotas under the Agreement on Textilesand Clothing (ATC) reduced growth of regionaloil-importing countries from 4.6 percentin 2004 to about 4.0 percent in 2005.Looking forward, high oil prices are expectedto continue feeding demand in oilproducingcountries, whose economies shouldexpand by 5.4 percent in 2006 and 5.1 percentin 2007. In the oil-importing economies,growth is expected to accelerate to about thesame level, supported by stronger Europeangrowth and a weaker negative effect fromthe ATC. The region’s strong performance reflects,in part, past reforms, such as steps toimprove transparency in the oil sector inAlgeria, as well as banking-sector reform, reductionsin customs duties, privatization, andregulatory reform in other Maghreb countries.These efforts, and in particular, the substantialreforms underway in Egypt, help to raise theregion’s growth potential by improving bothinfrastructure and the overall investment climate.While heartening, the pace of reformoutside of Egypt appears to have waned, perhapsbecause high oil prices have reducedthe sense of urgency attached to reform in oilexportingcountries.In contrast to the slowdown elsewhere inthe world economy, growth in South Asia isestimated to have picked up a bit in 2005,coming in at 6.9 percent, compared with6.8 percent in 2004. This mainly reflects improvedperformance in Pakistan, where GDPis estimated to have increased 8.4 percent (upfrom 6.6 percent in 2004), thanks to a broadbasedacceleration in the manufacturing andagricultural sectors. Like Pakistan, othercountries in the region have enjoyed verystrong export performance, in part because ofthe recent removal of ATC quotas. However,the sharp rise in oil prices and solid regionalgrowth over the past several years havecontributed to an acceleration of inflation.Addressing this issue will require a furthertightening of monetary policy, which, in combinationwith rising oil bills, is expected toresult in a modest deceleration of economicactivity to about 6.3 percent by 2007.GDP in Sub-Saharan Africa is estimated tohave increased 4.6 percent in 2005, bolsteredby very strong growth among resource-richcountries. Output in South Africa, the region’slargest economy, is estimated to have acceleratedto 4.2 percent, lifted by high metal prices,strong confidence, low nominal interestrates and the rand’s recent depreciation.The economies of oil-exporting countries,including Nigeria (the region’s second largesteconomy), grew an estimated 5.5 percent in2005, reflecting rapid increases in petroleum7

GLOBAL ECONOMIC PROSPECTS 2006production and investment inflows. Growthin some oil-exporting countries may exceed25 percent in 2006 and 2007, as new oil fieldscome on stream. However, the pace of theexpansion will taper off in other countries asthey reach capacity constraints.In West Africa, strong commodity prices in2005, improved rainfall, and more vigoroususe of insecticides are expected to lift regionalgrowth. In East and Southern Africa the expansionis projected to slow somewhat, partlybecause the removal of quotas under the ATCwill continue to put textile exports under pressure.Political strife and insecurity in Côted’Ivoire and the Great Lakes region are likelyto impact growth there. Countries are increasinglypassing higher crude-oil prices throughto consumers with the aim of containing budgetdeficits but will cut into consumer demandand add to inflationary pressures.The balance of payments and economicconsequences of higher oil prices are expectedto intensify over the next year as other commodityprices, which have attenuated theterms-of-trade impact of high oil prices, ease.Despite higher oil prices and increased passthrough,inflation is expected to remain in thesingle digits as a result of lower food pricesand prudent monetary policies. Recent economicreforms, and increased donor support—asmore countries reach the Heavily IndebtedPoor Country (HIPC) completionpoint—will also help support growth, which isprojected to be at or above 4.5 percent overthe medium term.Long-term prospects and povertyforecastThe recent strong economic performance ofdeveloping economies and the relativelyrapid growth projected for these economiesover the medium term owe much to the economicreforms undertaken over the past severalyears. Improved macroeconomic policies,reflected in lower inflation, trade liberalization(average tariffs have fallen from 30 percent toless than 10 percent since the 1980s), moreflexible exchange rate regimes, and lower fiscaldeficits have reduced uncertainty and improvedthe overall investment environment.More microeconomic structural reforms, suchas privatization and regulatory reform initiatives,have also played a key role.Table 1.2 Long-term prospectsReal GDP per capita, annual average percentage changeForecastMedium-term Long-term1980s 1990s 2001–06 2006–15World Total 1.3 1.2 1.5 2.1High-income countries 2.5 1.8 1.6 2.4OECD 2.5 1.8 1.6 2.4United States 2.3 2.0 1.8 2.5Japan 3.4 1.1 1.1 1.9European Union 2.1 1.8 1.4 2.3Non-OECD 3.5 4.0 2.0 3.5Developing economies 0.7 1.5 3.7 3.5East Asia & Pacific 5.8 6.3 6.4 5.3Europe & Central Asia 0.9 1.8 5.0 3.5Latin America & Caribbean 0.9 1.6 1.2 2.3Middle East & North Africa 1.1 1.0 2.5 2.6South Asia 3.3 3.2 4.5 4.2Sub-Saharan Africa 1.1 0.5 1.8 1.6Source: World Bank.8

PROSPECTS FOR THE GLOBAL ECONOMYThese factors are expected to contributeto better long-term growth performanceas compared with past decades (table 1.2).Consistent with recent improvements in economicperformance, per capita incomes in developingcountries are projected to growsome 3.5 percent a year, more than twice asfast as the 1.5 percent growth rates recordedduring the 1990s. Projected future growthrates are higher than during the 1980s and1990s in every developing region except EastAsia, where they are expected to declinesomewhat due to an aging population.Table 1.3 reports poverty projectionsbased on these real per capita income growthrates and the (re)distribution of incomewithin the population. The table indicatesthat over the next 15 years the share of thepopulation living in extreme poverty is expectedto decline in all developing regions. 6With the exception of Sub-Saharan Africa,all regions are expected to achieve theirMillennium Development Goal of reducingpoverty by 50 percent from its 1990 level. InEast Asia, the target has already beenachieved. Moreover, based on the currentlong-term forecast, extreme poverty would bealmost eliminated by 2015 in both the EastAsia and Pacific and the Europe and CentralAsia regions. Overall, the number of peopleliving on $1 a day or less will fall to around620 million, from 1.2 billion in 1990 and anestimated 1.0 billion in 2002.Despite these heartening prospects, there isno room for complacency. The percent of thepopulation in developing economies living atTable 1.3 Regional breakdown of poverty in developing countriesMillions of people living onless than $1 per dayless than $2 per dayRegion 1990 2002 2015 1990 2002 2015East Asia and the Pacific 472 214 14 1,116 748 260China 375 180 11 825 533 181Rest of East Asia and the Pacific 97 34 2 292 215 78Europe and Central Asia 2 10 4 23 76 39Latin America and the Caribbean 49 42 29 125 119 106Middle East and North Africa 6 5 3 51 61 40South Asia 462 437 232 958 1,091 955Sub-Saharan Africa 227 303 336 382 516 592Total 1,218 1,011 617 2,654 2,611 1,993Excluding China 844 831 606 1,829 2,078 1,811Percent of population living onless than $1 per dayless than $2 per dayRegion 1990 2002 2015 1990 2002 2015East Asia and the Pacific 29.6 14.9 0.9 69.9 40.7 12.7China 33.0 16.6 1.2 72.6 41.6 13.1Rest of East Asia and the Pacific 21.1 10.8 0.4 63.2 38.6 11.9Europe and Central Asia 0.5 3.6 0.4 4.9 16.1 8.2Latin America and the Caribbean 11.3 9.5 6.9 28.4 22.6 17.2Middle East and North Africa 2.3 2.4 0.9 21.4 19.8 10.4South Asia 41.3 31.3 12.8 85.5 77.8 56.7Sub-Saharan Africa 44.6 46.4 38.4 75.0 74.9 67.1Total 27.9 21.1 10.2 60.8 49.9 32.8Excluding China 26.1 22.5 12.9 56.6 52.6 38.6Source: World Bank.9

GLOBAL ECONOMIC PROSPECTS 2006or below $2 per day is projected to remain disturbinglyhigh. Moreover, notwithstandingthat inroads have been made recently, the incidenceof extreme poverty in Sub-SaharanAfrica in 2002 was actually higher than in1990. While current projections suggest 8 percentof the subcontinent’s population will belifted above the extreme poverty line by 2015,some 38 percent of Africans will still be livingin extreme poverty. Worse, the absolute numberof Africans living at or below the $1-a-daylevel is projected to increase. And, because percapita incomes elsewhere are projected togrow faster, the continent will continue to fallfarther behind the rest of the world—unlesssteps are taken to greatly improve economicgrowth in Africa.International financeThe significant adjustments of internationalexchange rates over the past several yearspaused in 2005. In particular, notwithstandingthe persistence of the U.S. current accountdeficit (expected to exceed $750 billion thisyear), the dollar’s trend decline with respect tomajor currencies came to an end. Initially, thecurrency appreciated against its trading partnersby some 3.5 percent in real-effective termsas of July 2005. It then lost value in Augustand September, before showing signs ofstrengthening in October.The strengthening of the dollar during thefirst seven months of 2005 is partly explainedby rising U.S. short-term interest rates (as theFederal Reserve Bank continued its policy ofgradual tightening) and falling long-term ratesin Europe (possibly in response to the continent’srelatively weaker economic performance).By July, these developments hadgenerated a 300 basis-point swing in the differencebetween U.S. and European shortrates, along with a 75 basis-point gap in favorof long-term U.S. bonds (figure 1.4).These growing interest rate differentials increasedthe financial incentive to hold dollarversuseuro-denominated assets, temporarilyproducing stronger net private sector capitalFigure 1.4 Dollar-euro interest ratedifferentialsPercentage points2.01.51.00.500.51.01.52.0Jan. 2003Source: World Bank.Long-termdifferenceShort-termdifferenceApr. 2003July 2003Oct. 2003Jan. 2004Apr. 2004July 2004Oct. 2004Jan. 2005Apr. 2005July 2005inflows in the first quarter of 2005 as investorsadjusted their portfolios. Not only didthese inflows help strengthen the dollar, theyalso financed a large share of the U.S. currentaccount deficit (figure 1.5). As a result, thedollar was much less reliant on the accumulationof reserves by foreign central banks (foreignofficial asset purchases) than in 2004.In the second quarter of 2005, however,private inflows eased, and foreign centralbanks once again assumed a large role in thefinancing of the dollar. Moreover, toward theend of July the dollar came under reneweddownward pressure and depreciated some1.7 percent in real-effective terms duringAugust and September. The dollar began toappreciate again only after the long-terminterest rate started to rise again. By October2005, the long-term interest rate differentialhad widened to about 120 basis points.The apparent sensitivity of the dollar andthe financing of the U.S. current accountdeficit to interest-rate differentials highlightthe problems posed by the large financingrequirements of the U.S. current accountdeficit.Until global imbalances are resolved, thedollar is likely to continue to come under10

PROSPECTS FOR THE GLOBAL ECONOMYFigure 1.5 Financing of the U.S. currentaccount deficit$ billions, annualized90060030003002004 2005Q12005Q2Figure 1.6 Emerging market spreadsBasis points1,000800EMBI global bond spreads600400200July 2002Jan. 2003July 2003Jan. 2004July 2004Jan. 2005July 2005Other private flowsOfficial assets(reserves)Private sectortreasury billsNet equity and FDISource: World Bank; Datastream.Source: U.S. Department of Commerce; BEA; U.S.Treasury Bulletin.downward pressure, unless foreign centralbanks accumulate substantial quantities ofdollars or interest-rate differentials widen further.In the baseline, interest rate differentialsare projected to widen further, and the dollaris projected to decline gradually, falling byabout 5 percent per year. Should central bankscease to be willing to accumulate reserves atcurrent rates, there could be a disruptive hikein interest rates or a more precipitous fall inthe dollar (World Bank 2005).The recent decision of the Chinese andMalaysian authorities to move from an exchangerate regime linked to the dollar aloneto one focusing on a basket of currencies representsa major and welcome move toward amore flexible currency regime. While it willnot resolve current account imbalances, itshould increase the stability of the renminbiand ringitt with respect to the currencies oftheir trading partners (other than the UnitedStates) and reduce the amount by which thedollar would have to depreciate relative toother currencies to achieve a given level of adjustment.7 How effective the new regimes willbe, depends importantly on how they aremanaged. While technically the announcedrules could allow the renminbi to depreciate asmuch as 9 percent per month, similar possibilitiesfor flexibility existed under the formerregime but were not exercised.Interest rates and spreads remain lowThe recent period of very low real interestrates has been particularly beneficial to developingeconomies. Together with narrowerrisk premia (figure 1.6), low rates have alloweddeveloping countries to reduce theirfinancing costs, restructure their debt, andpursue strong investment growth. Early repaymentof Paris Club debt has alreadyreached $22 billion in 2005, and amongemerging-market economies, virtually allfinancing requirements for this year had beenmet by August. 8Short-term rates have been rising, andthey can be expected to continue to rise asmonetary policy tightens, initially in theUnited States, but eventually in Europe aswell. In contrast and notwithstanding recentincreases, longer-term interest rateshave remained low longer than expected(figure 1.7), while spreads on more riskyemerging market and corporate assets havefallen even further.11

GLOBAL ECONOMIC PROSPECTS 2006Figure 1.7 Real long-term interest rates inG-7 countriesPercent864202461970Periodaverage1975Source: World Bank.GDP-weighted averageof the 10-year a governmentbond yields1980 1985 19901995 2000 2005Note: For data prior to 1972, exclude Italy, a 10-year ornearest government bond yields.Figure 1.8 World savings rate% of nominal GDP a25242322212019811986Averagesavings rateGlobalsavings rate1991 1996 2001 2005Source: World Bank.a. Sum of national savings divided by the sum of nationalGDP expressed in U.S. dollars at market exchange rates.Many reasons for these low interest rateshave been proposed (see IMF 2005 for a recentoverview), including the following:• Excess liquidity stemming from an extendedperiod of very low short-terminterest rates in almost all developedeconomies.• A low inflation environment, thanks toimproved credibility of monetary policy,and the disinflationary impact of increasedcompetition following the entryinto global markets of China and membersof the former Soviet bloc.• An increase in global savings, due to– increased savings in Europe followingheightened recognition of the need toprepare for the impending retirementof the baby-boom generation; and– increased corporate savings in dynamicEast Asia (caused by corporaterestructuring following the currencycrisis) and in the United States (followingthe stock market decline in 2000).However, while global savings have increasedrecently, this follows a period wherethey declined substantially, making it difficultto argue that the world savings rate is currentlytoo high (figure 1.8). Rather, investmentactivity, principally in the developed world, hasfailed to keep pace with savings as they havereturned to historical levels (see IMF 2005).Most of these explanations for lower longtermrates involve temporary factors, implyingthat long-term rates will eventually risetoward their long-run equilibrium level 9 (frequentlydefined as the long-run potentialgrowth rate of the economy). In this context,the question is not so much why long-termrates are low, but how much longer they willremain so. In the baseline, increased investmentin Europe and tighter monetary policyresult in a gradual rise in interest rates, whichwill nevertheless remain below recent estimatesof the long-term growth potential of theU.S. economy. The final section of this chapterexplores some of the economic implicationsshould interest rates stay low for an extendedperiod of time or, alternatively, should theyrise more quickly than anticipated.Signs of rising inflationLow interest rates have contributed directly tothe strong economic performance of recentyears. Growth has, in turn, provoked a pickupin inflation in many developing countries. Thelargest hikes have been in commodity prices(see below). However, producer price inflation12

PROSPECTS FOR THE GLOBAL ECONOMYhas jumped by more than 4 percentage pointsin some regions and exceeds 5 percent inevery developing region except Sub-SaharanAfrica. 10Consumer price inflation has also beenrising (if less spectacularly). Weighted byGDP, aggregate inflation among developingeconomies increased from 4.0 percent in thefourth quarter of 2003 to 5.4 percent by Julyof 2005. It has since eased somewhat. Regionally,inflation has picked up strongly in SouthAsia, Sub-Saharan Africa, and East Asia (figure1.9). Inflation in developing countriesis projected to continue rising in 2005, asgrowth remains at or above trend rates, andthe pass-through from high oil prices continuesto exert upward pressure on prices.In high-income countries, there are onlylimited signs of rising inflation. In the UnitedStates, where output is close to potential, inflationhas been rising steadily. It jumped to4.7 percent in September 2005, but the increaseis not expected to be permanent, becauseit reflects very high gasoline prices thatmonth, which have since declined. Nevertheless,data pointing to rising wages and lowerproductivity growth suggest that core inflation,which has been more stable, may beginto rise soon. In Europe, high oil prices havelimited disinflation despite significant slackand the appreciation of the euro.These same factors should continue to limitprice inflation in Europe. However, in theUnited States, high oil prices plus the projectedfurther depreciation of the dollar areexpected to generate additional upward pricepressure.Low interest rates have resulted in higherprices of interest-sensitive assets in marketswith strong financial intermediation—notablyin the United States and some Europeancountries—contributing to strong consumerdemand (World Bank 2005, IMF 2005). As interestrates rise, housing prices are expected toplateau and even decline, which has alreadybegun in the United Kingdom. As they do so,the rate at which household wealth increaseswill moderate and its contribution to consumerdemand should abate. 11Data indicate that house prices have alsobeen rising rapidly in a number of middleincomecountries, such as Bulgaria, India,Figure 1.9 Inflation ratesAnnual % change9876200320022004Most recent available data a543210East Asiaand PacificEurope andCentral AsiaLatin America Middle East andand the Caribbean North AfricaSouth AsiaSub-SaharanAfricaOECDSource: World Bank.a. Data concern August 2005, except for East Asia (September), the Middle East and North Africa (June), and Sub-Saharan Africa(March).13

GLOBAL ECONOMIC PROSPECTS 2006Figure 1.10 Cumulative real increase inhousing prices, 2005Percent100806040200India(Mumbai)South AfricaBulgariaSource: World Bank; BIS.a. Data for Russia and China reflect increases between 2000and 2004; data for other countries reflect increases between2000 and 2005.Indonesia, Malaysia, and South Africa (figure1.10). While fast economic growth andchanges in the regulatory environment havecertainly played a role in these countries, sohave low interest rates. Unfortunately, datalimitations prevent a thorough analysis of thecauses and consequences of rising housingprices in low- and middle-income economies. 12Commodity marketsRussia aThailand(Bangkok)United StatesShanghaiChina aAfter several years of rising commodityprices, there are indications of a stabilizationand even reversal of gains in the marketsfor agricultural products and for metals andminerals (figure 1.11).Agricultural prices have been decliningmost of this year and are down 5 percent sinceMarch 2005. However, prices of agriculturalraw materials are rising, partly because ofhigher prices for commodities that are closesubstitutes for crude oil-based products (forexample, natural rubber prices are up 41 percentbecause of increases in synthetic rubbercosts).Figure 1.11 Commodity pricesIndex, 1990 100300250200150100502000EnergySource: World Bank.AgriculturalproductsMetals andminerals2001 2002 2003 2004 2005 2006Although metals and minerals prices roseduring the first months of the year, they havesince stabilized, and in October 2005, theywere at the same level as in March 2005.Conditions in some metals and minerals marketsremain tight, due to low inventories. Inthe case of copper and aluminum, prices remainelevated (partly reflecting higher energycontent in the production of these goods).Demand has weakened markedly for lead, tin,and zinc.Analysis of past non-oil commodity cyclessuggests that this one may have run its course.Already it distinguishes itself from previousepisodes by having lasted longer, in part,because energy prices have also been high,which was not always the case during previousepisodes. In so far as high fuel prices increaseproduction costs in both agricultureand metals and minerals, they may have reducedthe supply response, keeping priceshigher longer.In line with the projected slowdown inglobal growth and increased supply, prices ofagricultural products and metals and mineralsare projected to decline somewhat in 2006.Limited spare capacity to keepoil prices highIn contrast with other commodity prices,oil prices continued to strengthen during the14

PROSPECTS FOR THE GLOBAL ECONOMYfirst nine months of 2005. During this period,they averaged some $52 per barrel, a 38 percentincrease compared with the average for2004. These increases occurred despite an easingof conditions in the oil market. Demandgrowth slowed from more than 3.5 percent in2004 (the highest growth since the late 1970s)to a 1.4 percent annualized rate during thefirst three quarters of 2005. As a result, supplyis actually increasing faster than demand, 13and inventories have begun to accumulate,although they remain low. 14Rising prices over the first eight months ofthe year reflected the market’s concern thatexisting spare capacity was insufficient to dealwith a major disruption to supply or an increasein demand (figure 1.12). In some sense,hurricane Katrina was the kind of seriousshock the market feared. Although oil pricesspiked briefly to more than $70 a barrel, theyare back below $60, following the release ofsome 29 million barrels of crude oil from thestockpiles of the International Energy Agencyand the U.S. government. Moreover, gasolineprices in the United States have returned to thelevels observed before hurricane Katrina, andmarket concerns have switched from a focuson inadequate oil supply to insufficientFigure 1.12 Levels of spare oil capacityMillion barrels per day876543210Jan.2002July2002Jan.2003Source: World Bank.July2003Jan.2004July2004Jan.2005July2005refining capacity, particularly of lower-qualitycrude oil.Spare production capacity is now about2 million barrels per day (mbpd), comparedwith almost 6 mbpd three years ago, and capacitywill remain tight over the near term.This reflects long lags in bringing significantnew quantities of oil into production 15 andshorter lags before demand substitution canhave an effect. 16 Moreover, approximatelyhalf of the expected new capacity is being producedby OPEC, suggesting that the organizationwill continue to exercise significant marketpower over the near term.In this environment, prices are likely to remainvolatile, as small events or even minorchanges in expectations may provoke significantprice swings. As a result, the World Bankhas adopted a technical assumption for thefuture path of oil prices based on a slow declinetoward $40 per barrel by 2010. Thisimplies an average price of $56 in 2006 and$52 in 2007, which is somewhat higher thanthe current consensus forecast. The economicconsequences of alternative scenarios, notablya sharp negative supply shock, are discussed inthe final section of this chapter.The impact of oil prices ondeveloping economiesThe world economy in general and developingcountries in particular have shown considerableresilience to higher oil prices. This reflectsincreases in non-oil commodity prices and avery robust global economy, which have, untilrecently, muted the impact of higher oil prices.The first round of oil price hikes (1999–2000) adversely affected low- and middleincomecountries. The price increase was verylarge in percentage terms (rising from justunder $12 to almost $30 per barrel between thefirst quarter of 1999 and the end of 2000) butsmaller than the most recent increases in dollarterms (table 1.4). Current account deficitsamong low-income oil-importing Africancountries increased by 0.5 percent of GDP onaverage. 17 Moreover, government deficits inthose countries that did not pass on the price15

GLOBAL ECONOMIC PROSPECTS 2006Table 1.4 Terms-of-trade impactsof commodity price changes1999–00 2001–03 2004–05Cumulative price changeOil 120.3 18.9 88.0Agricultural products 0.3 15.7 8.9Metals and minerals 25.0 10.2 47.9Manufactures 5.0 3.0 10.4Total terms-of-trade effect (% of GDP)Oil ImportersLow and Middle income 1.8 0.1 0.9Low income 3.8 0.9 2.9Sub-Saharan Africa 2.5 1.4 1.2South Asia 3.9 1.5 2.7Highly indebted 4.3 1.5 3.3poor countriesSource: World Bank.Note: Periods Jan. 1999–Dec. 2000, Dec. 2000–Dec. 2003,Dec. 2003–July 2005.hikes rose by about the same amount. 18 Amongthose countries with limited access to internationalfinance, non-oil imports fell by 3.8 percentin 2000, partly because insufficient foreignexchange was available to finance imports atprevious levels.During the second bout of (more gradual)price increases (2001–3) the same countriesperformed much better. Current account positionsactually improved as a percent of GDP,and there were no discernible slowdown innon-oil imports. Part of this improved performanceis explained by real exchange ratemovements (which diminished the domesticcurrency cost of the oil-price hike) and by anumber of non-oil commodity prices that alsoincreased rapidly during that time, thus providingthe necessary foreign currency to meetthe additional oil burden without cutting intonon-oil imports. On average, high non-oilcommodity prices reduced the negative termsof-tradeshock from higher energy costs bymore than half.Drawing from this experience, the impactof the latest hike in oil prices on poor oil importersis a concern, principally because it hasnot been accompanied by as much strength innon-oil commodity prices. Indeed, the estimatedterms-of-trade shock from price movementssince January 2004 is more than threetimes as large for various groups oflow-income countries than the cumulativeshock over the preceding three years. As a result,non-oil imports from poor, currentaccount–constrained countries are expected tocome under pressure in the coming months.Moreover, the impact on oil-importing poorcountries could be significantly aggravated ifoil prices remain at or close to current levelsand if non-energy commodity prices return topre-shock levels.World tradeThe expansion of world trade slowed significantlyduring 2005 (figure 1.13).While merchandise exports were growing at a16 percent or more annualized pace in themiddle of 2004, they subsequently slowed andwere expanding at an 8.5 percent pace duringthe third quarter of 2005.Figure 1.13 World trade volumesAnnual growth rates, 6-month moving average4036322824201612840Jan.2003July2003Source: World Bank.Jan.2004WorldChinaHigh-incomecountriesJuly2004Other developingcountriesJan.2005July200516

PROSPECTS FOR THE GLOBAL ECONOMYMost of the deceleration concerned the exportsof high-income economies, volumes ofwhich grew by less than 4 percent (annualized)in the first quarter of 2005, before strengtheningmore recently. Merchandise export volumesof developing countries (excluding China) wererelatively robust, increasing at an estimated12 percent pace toward the middle of 2005.Chinese exports, boosted by the removal ofATC quotas, grew at a 24 percent pace.Commodity markets have significantlyshaped developments in world trade. The merchandiseexports of oil-exporting countries,which had been rising at some 5.8 percent in2004, increased by an estimated 5.2 percentin 2005. The deceleration among developingoil importers (excluding China) was steeper inpercentage terms (from 15 percent to 11 percent),but growth rates remained much higher.Growth in the production of non-oil commoditiesin general is moderating, both becausedemand is easing and because of supplyconstraints.As a reflection of the slowdown already observed,international trade is forecast to slowdown relative to 2004 as a whole. Merchandisetrade volumes are expected to increaseby around 7.7 percent. The goods and servicetrade is expected to increase 6.2 percent in2005 before strengthening somewhat in2006–7.Exports of developing economies continueto be heavily influenced by developments in thevolatile high-tech market. After falling sharplyin the third quarter of 2004, global sales ofsemiconductors and other high-tech productspicked up before weakening once again in thesecond quarter of 2005. This volatility is apparentin East Asian export volumes (high-techproducts represent as much as two-thirds of theexports of some economies in this region). 19High-frequency data suggest a strengthening ofdemand for these products, implying a pick upin export flows from the region.Trade growth for some countries washeavily influenced by the removal of quotasunder the ATC of the Multifiber agreement inFigure 1.14 Change in textile exports tothe developed world, first half of 2005TajikistanMyanmarNepalKenya aPhilippinesMexicoMongoliaMauritiusSouth AfricaVietnamTunisiaLesotho aMadagascarMoldovaAlbaniaMacedonia, FYRPakistanMoroccoThailandSwaziland bIndonesiaRomaniaBulgariaLao PDREgyptBangladeshCambodiaPeru aJordanTurkeySri LankaIndiaChina b60 40 20 0 20 40 60% change, year over year, $ valuesSource: World Bank, IMF, U.S. Department of Commerce,and EuroStat.a. Data refer to the first five months of each year only.b. Data for China include Hong Kong and Macao.January 2005 (figure 1.14). While the removalof quotas was done in phases, and ten years ofadjustment time provided, backloading of theremoval of quotas meant that they were stillbinding when they were finally removed. As aresult, there were significant changes in patternsof trade among affected goods in 2005,most notably in the form of increased exportsfrom China (and other countries, whose marketshare had been artificially held back underthe old quota scheme) to the detriment of17

GLOBAL ECONOMIC PROSPECTS 2006Figure 1.15 Estimated change in textileexports as share of total merchandiseexportsBangladeshCambodiaSri LankaTurkeyJordanIndiaLao PDRPakistanRomaniaChina aBulgariaMoroccoEgypt, Arab Rep.Macedonia, FYRMadagascarPeru bTunisiaAlbaniaMoldovaIndonesiaVietnamThailandSouth AfricaMexicoMauritiusPhilippinesKenya bMongoliaMyanmarNepalTajikistan15 10 5 0 5 10 15PercentSource: World Bank, IMF, U.S. Department of Commerce,and EuroStat.Note: Data reflect the change in textile and clothing exportsbetween 2004 and 2005 as a percent of total merchandiseexports.a. Data for China include Hong Kong and Macao.b. Data refer to the first five months of each year only.those exporters that had benefited most fromthe old quota system.Using U.S. and European imports of textilesas a proxy for developments in the worldas a whole, China, 20 India, Jordan, Peru,Sri Lanka, and Turkey saw the dollar value oftheir exports increase between the first half of2004 and the same period in 2005 by morethan the 20 percent average increase in highincomeimports over the same period. The textilesectors in Kenya, Myanmar, Nepal, thePhilippines, and Tajikistan, on the other hand,saw the dollar value of their exports decline by4 or more percent.However, many of these countries are notlarge exporters of textiles. As a result, thesefigures may exaggerate the overall economicimpact of the relaxation of textile quotas.Expressed as a percent of these countries’total merchandise exports, the biggest gainswere experienced by Bangladesh, Cambodia,Jordan, India, Pakistan, Sri Lanka, andTurkey, while the largest losses were those ofKenya, Nepal, Myanmar, Mongolia, andTajikistan (figure 1.15).Risks and uncertaintiesLow interest rates, moderate inflation, andthe robust growth projected in the baselineconstitute a relatively benign scenario. However,the outlook is dominated by downside risks.An oil-market supply shock could causeserious disruptionThe most important potential risk comesfrom the oil market. As global demand andsupply are projected to increase broadly instep, excess capacity (currently estimated at1.9 million barrels per day) will remain veryconstrained. In this context, the market can beexpected to continue reacting to events in arelatively volatile manner. Rather than graduallydeclining, as in the baseline scenario,prices could remain at current levels, rise further,or even fall.More fundamentally, with spare productioncapacity so low, the market is particularlyvulnerable to a supply shock. Because nocountry can easily ramp up production, if outputin another producing country were to fallsignificantly, world supply would fall, provokinga decline in economic activity to the extentthat the global economy could not quicklyadopt an alternative energy source. 21Table 1.5 presents results from a simulationof the impact of a 2-million-barrels-per-day18

PROSPECTS FOR THE GLOBAL ECONOMYTable 1.5 Impact of a 2 million bpdnegative supply shock2006 2007 2008 2009Price of oil 90 70 44 40(Change from base line) 34 28 3 0Change in GDP % of baselineWorld 1.0 1.5 1.1 0.2High income 0.7 1.3 1.3 0.3Middle income 1.6 1.6 0.1 1.4Large low income 1.7 2.8 1.8 0.7Impact on inflation rateWorld 2.6 0.6 0.9 0.2High income 1.4 0.0 1.0 0.4Middle income 5.8 2.0 0.9 0.5Large low income 2.8 0.9 0.7 0.2Impact on real interest rates (levels)World 1.0 0.2 0.1 0.1High income 1.0 0.1 0.2 0.0Middle income 1.1 0.7 0.2 0.2Large low income 0.5 0.1 0.1 0.4Impact on current account balance (% of GDP)World 1.1 0.5 0.1 0.1High income 1.1 0.7 0.2 0.2Middle income 0.9 0.2 0.5 0.3Large low income 1.9 0.2 1.7 1.0Impacts on low-income current account–constrained countries (1)Terms of trade 4.1 .. ..GDP 0.3 0.1 0.0Domestic demand 2.7 1.1 0.0Current account balance 1.2 0.9 0Source: World Bank.Note: Impacts on low-income, current account–constrainedeconomies were estimates based on the terms-of-trade impactusing a purpose-built VAR model. Other estimates weresimulated using the World Bank’s macroeconomic simulationmodel.negative supply shock. 22 The disruption isassumed to last throughout the projectionperiod, causing prices to rise to $120 for aninitial period of three months before easingto $80 for three quarters. Thereafter, supplyand demand adjustments result in a gradualdecline in oil prices toward $40.Global output responds to the initial shockby contracting, as compared with the baseline,by 1.5 percent of GDP after two years, whileinflation picks up rapidly. On average, thecurrent account position of oil-importingcountries deteriorates by about 1.1 percent ofGDP. The impact is more severe in large lowincomeand middle-income countries, both becauseof higher energy intensities and a greaterinflationary impact, which requires a largercontraction to eliminate.While the impact in terms of GDP for currentaccount–constrained low-income countriesis smaller, it is more severe in terms ofdomestic consumption and investment. Suchcountries have limited access to internationalcapital markets, and their capacity to payhigher oil prices is limited by their export revenues.If these revenues are stable, they areforced to reduce domestic demand and nonoilimports in order to pay their higher oilbill. As a consequence, when oil prices rise,oil consumption remains relatively constantin volume terms (being generally inelastic inthe short run), but the oil bill rises. To compensate,non-oil imports and domestic demandtend to decline in unison—leaving GDPrelatively unchanged. For these countries, theterms-of-trade shock of the initial increase inoil prices is estimated at 4.1 percent of theirGDP, which would translate into a 2.7 percentdecline in domestic demand, with potentiallyserious impacts on poverty.The future path of interest rates representsan additional source of uncertaintyPersistent global imbalances continue to be aserious source of uncertainty. The current accountdeficit of the United States and its financingrequirements are very large, and thewillingness of investors to finance it is sensitiveto both interest-rate differentials andexchange-rate expectations. As net foreign liabilitiesaccumulate, markets will become increasinglysensitive to adverse shocks orchanges in sentiment, and the dollar is likelyto come under downward pressure onceagain, which would put upward pressure oninterest rates.Table 1.6 explores the possible implicationsof higher interest rates. In this scenario, facedwith sustained downward pressure on the dollar,investors demand higher returns on U.S.-denominated assets to offset further expecteddepreciations. This, combined with concerns19

GLOBAL ECONOMIC PROSPECTS 2006Table 1.6 Interest rate scenarios2005 2006 2007 2008 2009A. A 200 basis-point increase in interest rates and in spreadsInterest rates (change of Q4 level from baseline)World 1.8 1.4 0.6 0.2 1.3High income 1.7 1.1 1.1 0.3 0.9Low and middleincome 2.0 2.7 1.9 2.6 3.1GDP (% change from baseline)World 0.1 1.7 2.9 1.9 0.6High income 0.0 1.5 2.7 2.5 1.0Low and middleincome 0.2 2.4 3.5 3.0 1.5Inflation (change in inflation rate)World 0.0 0.3 1.1 1.1 0.3High income 0.0 0.3 1.5 1.6 0.5Low and middleincome 0.0 0.3 0.7 1.2 0.9B. Persistently low interest ratesInterest rates (change of Q4 level from baseline)World 0.7 0.5 0.6 0.1 0.6High income 0.7 0.5 0.8 0.1 0.6Low and middleincome 0.8 0.8 0.1 0.1 0.7GDP (% change from baseline)World 0.0 0.8 1.4 0.3 0.5High income 0.0 0.8 1.4 0.4 0.5Low and middleincome 0.0 1.0 1.4 0.1 0.7Inflation (change in inflation rate)World 0.0 0.2 0.6 0.6 0.1High income 0.0 0.2 0.8 0.8 0.2Low and middleincome 0.0 0.1 0.2 0.5 0.3Source: World Bank.about rising debt and pension liabilities inindustrialized countries, and a more rapid dissipationof the temporary factors depressinglong-term interest rates, causes them to increaseby some 200 basis points in highincomecountries. Risk premia in developingeconomies increase by an additional 200 basispoints as investors’ appetites for risk decline.The world economy reacts to the substantialtightening of monetary conditions by reducingglobal growth by half for a periodof two years, as higher interest rates cut intoinvestment and consumption demand, boththrough classic transmission mechanisms andvia the impact of interest rates on housingprices and consumer wealth. Slower growtheases inflationary pressure and global tensions,including in the oil market. As monetarypolicy loosens in response to increasingoutput gaps, growth starts to pick up again,bringing output back to the levels in the baselineby the end of the simulation period.Higher interest rates would also affect financingconditions for developing countriesby increasing future borrowing costs. Formany countries this will not pose a seriousshort-term risk, because they have taken advantageof low rates to reduce the share ofshort-term debt relative to their overall debtand to prefinance some of their future borrowingneeds. For others, particularly thosewith large debt-to-GDP ratios or those thathave accumulated large short-term debt positions(figure 1.16), a rapid rise in interest ratesFigure 1.16 Some countries are particularlyat riskArgentinaBelizeLebanonPapua New GuineaUruguayIndonesiaEcuadorTurkeyPhilippinesJamaicaBoliviaParaguayBrazilPakistanR.B. de VenezuelaUkraineDominican Rep.GuatemalaSource: World Bank.00Debt/GNI ratio50 100 1505 10 15 20 25 30Short-term debt/Total debt2003520

PROSPECTS FOR THE GLOBAL ECONOMYcould pose a real threat—particularly if therise in base interest rates also provokes a returnof spreads to more usual levels.Alternatively, if excess liquidity and globalsavings prevent long-term interest rates fromrising as quickly as in the baseline scenario,the resulting higher levels of demand could increasetensions in commodity markets, includingthe oil market. In addition, lower interestrates could cause a number of economies, includingthe United States, to overheat, generatingadditional inflation that forces a furthertightening of monetary policy. As a result,while growth would be initially higher, thesubsequent tightening of policy could provokea stronger-than-projected slowdown.The depth of the cycle would depend importantlyon the extent of the wealth effectgenerated by low interest rates on housingprices. The more pronounced, the deeper thecycle. Moreover, because asset prices becomeeven more out of step with their long-runlevels, an even longer period of slow growthcould be required to re-establish equilibrium.Policy challengesPolicy can help reduce both the economicseverity of such unfavorable outturns and thelikelihood that they will materialize.High oil prices will naturally induce substitutiontoward alternative energy sources andconservation. In the current context, wherethe increase in oil prices is expected to endure,countries that have not passed on recent pricehikes to consumers (and industry) may wishto revise their policies. Not only are the budgetarycosts of such subsidies likely to be difficultto support, but these policies also impedeadjustment.Moreover, countries with restrictive rulesconcerning the exploitation of oil reservesmight wish to re-examine them. Such policiesmay deny these countries access to technicalexpertise and financial capital, thereby preventingthem from investing in new productionto the extent that they might otherwise.This may slow the aggregate supply responseand encourage greater conservation and substitutiontoward alternative energy sources—to the ultimate detriment of oil-producingcountries.In the developed world, efforts to increaseenergy efficiency by developing more fuelefficienttechnologies, such as hybrid cars,could well pay important dividends. Thesetechnologies are already economic in somecountries, where gasoline is heavily taxed, andcould generate substantial savings in overallfuel demand. 23 In addition to the ecologicalbenefits, making such technology available indeveloping economies, where the increase intransportation-related energy demand is highest,would be particularly effective in limitingoverall demand.Finally, efforts to improve cooperation betweenusers and suppliers concerning the qualityand transparency of oil market data couldhelp reduce unwarranted volatility and perhapscontribute to lower prices by reducingthe oil-price risk premia.To further dissipate the risk from globalimbalances, policies need to promote bothpublic and private savings in countries withlarge current account deficits. Recent measuresto tighten fiscal policy in the UnitedStates are headed in this direction, but moretightening is required. Tighter monetary policyis helping. Higher interest rates in theUnited States promote private sector financingof the deficit but also promote private sectorsaving. In Europe, policymakers should seekto maintain low interest rates in an effort tostimulate demand. As output picks up, fiscalpolicy (rather than monetary policy) should beused to restrict demand, if necessary. Indeed,given unfunded public pension liabilities inthese countries, such a fiscal tightening is necessaryin its own right.Developing economies should react flexibly,seeking to maintain real effective exchangerates in line with their fundamentals, ratherthan a particular alignment with any one currency.In this regard, recent steps by somecountries to adopt an exchange rate regimethat reflects their overall trade patterns arepositive and could be emulated by other21

GLOBAL ECONOMIC PROSPECTS 2006countries. Petro-dollars could help reduce thelikelihood of a disruptive resolution of globalimbalances if they are recycled into the globaleconomy in a way that reduces tensions. Inparticular, the financing of investment expendituresboth domestically and in other developingcountries would help stimulate demandoutside of the United States, reducing thatcountry’s current account deficit. To the extentthat these funds are invested in U.S. financialsecurities, they could also help finance the U.S.current account deficit.Finally, global weakness could trigger increasedprotectionism or a slowing of tradeliberalization (which has been the basis formuch of developing countries’ recent success).The supply disruptions in Europe provoked bythe re-imposition of quotas on Chinese importsof textiles, following their liberalizationat the beginning of this year, is a good illustrationof how trade restrictions work to thedetriment of both exporting and importingcountries. Not only should countries resist thetemptation to intervene in already liberalizeddomains, concerted efforts need to be made toachieve meaningful liberalization in the agriculturaland service sectors in the Dohaprocess. To date, liberalization has largelyomitted the politically sensitive agriculturalsector, depriving many developing countriesof the benefits from trade liberalization thatmore manufacturing-oriented economies haveenjoyed.Notes1. The Congressional Budget Office (2005) estimatesthat hurricane Katrina reduced growth in theUnited States by 0.4 and 0.9 percent (annual rates) inthe third and fourth quarters.2. The importance of China to aggregate statisticsis also visible in the industrial production data. Growthrates for all developing countries showed little slowing,but excluding China, annualized growth rates declinedfrom about 7.5 percent in mid 2004 to less than 5 percenta year later.3. Although oil prices are projected to decline during2006, they will be higher, on average, than in 2005.4. The number of working days each year varies,generally because certain holidays do or do not fall onweekends. Occasionally, these fluctuations can have animportant impact on annual GDP growth. While fivefewer days corresponds to roughly a 2.5 percent reductionin working time, in general, the actual reductionin production is less pronounced.5. These losses are estimated at more than 5 percentof GDP for Antigua and Barbuda, Belize, Guyana,Honduras, Nicaragua, and Jamaica.6. This year’s projections differ somewhat from2004’s partly because of a shift in the base year for calculationsfrom 2001 to 2002, which reduces thepoverty level of the starting year in all regions thathave experienced positive per capita growth. In addition,new survey data was employed for a large numberof countries (more than half in the Europe andCentral Asia region), including important new householdsurveys in a number of Latin American countriesin the place of labor force survey data used in the past.Finally, revisions to national income estimates ofGDP, inflation, and consumption play a role. For moreinformation concerning the changes to the povertyforecast, please visit the Long-term Prospects andPoverty Forecasts section of http://www.worldbank.org/globaloutlook.7. The extent to which the new regime will contributeto increased stability in world markets will alsodepend on the extent to which other Asian currenciesfollow suit, and how much flexibility is permitted inpractice.8. As of August 2005, overall financing for emergingmarket sovereign debt was already 74 percentfunded; this share reached 93 percent for emergingEurope and Turkey, and 100 percent for Latin America.9. Long-term interest rates tend to be determinedby the long-term growth potential of the economy andexpected inflation. Historically, temporary factors havecaused them to deviate from this measure, sometimesfor extended periods of time. However, they have alwaystended to return to this level.10. In the first half of 2005, producer price inflationexceeded 15 percent in the Europe and CentralAsia region, was more than 9 percent in Latin Americaand the Caribbean, about 7 percent in the MiddleEast and North Africa, and around 6 percent in bothEast and South Asia.11. The stock of housing in the United States is estimatedby the Federal Reserve Bank to be equal to$15.2 trillion, or about 138 percent of GDP. A 10 percentchange in the value of that stock would represent13.8 percent of GDP, or 19 percent of consumption.Econometric estimates suggest that the long-term marginalpropensity to consume from housing wealth is0.05 (see, for example Catte and others 2004 andBenjamin, Chinloy, and Jud 2004), implying a reductionin consumption of 1.35 percent.22

PROSPECTS FOR THE GLOBAL ECONOMY12. Very few low- and middle-income countrieshave housing data similar to the data available in highincomecountries; and what does exist tends to belimited to wealthy neighborhoods in single cities.Moreover, there is little information on home ownershipratios, and mortgage-market completeness—allcritical components in determining housing-marketwealth effects.13. The U.S. Department of Energy estimates thatboth oil and gasoline consumption fell by 2.5 or morepercent in September 2005.14. In the second quarter, stocks equaled 54 daysworth of consumption versus an average of more than58 days during the first half of the 1990s.15. Some oil fields can be brought into productionwithin 1 year, but others would take as long as 10 years.Three to five years would be needed to bring in two millionbarrels per day over and above expected increasesin demand.16. Opportunities for demand substitution may beless plentiful than in the 1970s because of the substantialconservation steps undertaken then. Nevertheless,use of fuel-efficient cars and less intensive use ofexisting cars can have substantial impacts on overalldemand.17. Simple average of 32 oil-importing Sub-Saharan economies.18. Among countries that did not pass pricesthrough fully, fuel subsidy spending rose substantially,for instance in the Central African Republic, GuineaBissau, Malawi, and the Seychelles.19. In 2003, high-tech products represented 13 percentof Thailand’s exports, but more than 50 percent ofTaiwanese, Malaysian, and Philippine exports.20. China here is taken as the sum of Hong Kong,Macao, and China, based on the assumption that priorto liberalization some of the exports from Hong Kongand Macao had actually originated in China, andtherefore, the changes in their market share reported inofficial statistics are exaggerated.21. During the 1980s (and before), OPEC acted asa swing producer, stepping up production in responseto shortfalls elsewhere. With its spare capacity nowmeasured at less than 2 million barrels per day, itscapacity to act as a swing producer is limited.22. Beccue and Huntington (2005) estimate thatthe probability of such a disruption occurring duringthe next 10 years is high (70 percent for one lasting6 months and 35 percent for one of 18 months).23. In the United States, for example, hybrid carscurrently offer approximately an 80 percent improvementin fuel efficiency. Were these vehicles to gain a10 percent share of new car sales, new energy demandwould be reduced by about 12 percent (c. 0.3 percentagepoints) per year for about seven years.ReferencesBeccue, Phillip C., and Hillard G. Huntington. 2005.“An Assessment of Oil Market DisruptionRisks.” In Final Report EMF SR 8. Energy ModelingForum. October.Benjamin, J.D., P. Chinloy, and G.D. Jud. 2004. “RealEstate versus Financial Wealth in Consumption.”Journal of Real Estate Finance and Economics29(3): 341–54.Catte, P., N. Girouard, R. Price, and C. André. 2004.“Housing Markets, Wealth and the BusinessCycle.” Economics Department Working Paper394. OECD, Paris.Congressional Budget Office. 2005. “Macroeconomicand Budgetary Effects of Hurricane Katrina.”Report. September 6. Washington, DC.Huntington, Hillard G. 2005. “The Economic Consequencesof Higher Crude Oil Prices.” In FinalReport EMF SR 9. Energy Modeling Forum,October.International Energy Agency. 2004. World EnergyOutlook. Paris.IMF (International Monetary Fund). 2005. WorldEconomic Outlook. Washington, DC.OECD (Organisation for Economic Co-operation andDevelopment). 2004. Economic Outlook, no. 74.Paris.World Bank. 2005. Global Development Finance2005. Washington, DC.23

2The Potential Gains fromInternational MigrationInternational migration can generate substantialwelfare gains for migrants, their countriesof origin, and the countries to which they migrate.The main focus of this report is on gainsfrom the remittances that migrants send home(discussed in chapters 4–6); chapters 2 and 3address the economic costs and benefits ofmigration and the impact of migration onpoverty. In this chapter, we use an economicmodel to estimate the size of the welfare gainsresulting from migration from developing tohigh-income countries. 1 It must be recognizedat the outset that the model fails to capturesome known costs and benefits of migration;that the results are dependent on the specificationof the model and its key parameters;and that the model cannot incorporate socialor political considerations. 2 The results of thissimulation do not provide a precise forecast ofthe likely impact of migration; instead, theyprovide a consistent framework that offers insightsinto (a) the economic gains that can beexpected from changes in policy or circumstances,and (b) the channels through whichmigration affects welfare—and both are difficultto measure in reality. The conclusionsdrawn from the model are supported by severalempirical studies, and they hold up wellunder various alternative assumptions formodel specification and parameters.In chapter 3, we complement this modelbasedapproach to measuring the gains frommigration with a review of the economic literature,which covers the implications formigrants and for their origin countries. Herewe can refer to a broader range of economicissues than are captured by the model, althoughwithout the ability to quantify that themodel-based simulation provides.Starting from the base-case forecast of economicactivity described in chapter 1, we introducean additional increase in migrationfrom developing to high-income countries sufficientto raise the labor force of high-incomecountries by 3 percent over the period2001–25. The assumed increase, roughly oneeighthof a percentage point a year, is close tothat observed over the 1970–2000 period. Weimply no judgment concerning whether suchan increase is likely or politically feasible, butrather view the rise in migration as an exogenousshock. As discussed in chapter 3, pressuresto migrate are likely to rise over the nextfew decades, but the actual size of the migrantflows will depend heavily on political decisionsin destination countries. This exercisepresents us with the following key findings.The expected decline in the labor force inhigh-income countries will increase dependencyratios, which could add to the benefitsfrom migration. However, such increases inmigration are unlikely to be large enough tohave a significant impact on dependency ratiosin high-income countries.Under the assumptions adopted in this modelingexercise, the rise in migration—small25

GLOBAL ECONOMIC PROSPECTS 2006relative to the labor force of high-income countries,but large relative to the existing stock ofmigrants—would generate large increases inglobal welfare. Migrants, natives in destinationcountries, and households in origin countrieswould experience gains in income, although migrantsalready living in high-income countrieswould see a decline in wages relative to the basecase. Estimates of these gains and losses are particularlysensitive to assumptions about the degreeof differentiation among workers (betweennatives and migrants and between old and newmigrants), the impact of migrants on fiscal balances,and the extent of remittances.Empirical studies of the impact of migrationon natives’ wages have had mixed results. Inthis simulation exercise, the rise in migrationleads to a small decline in average wages inhigh-income countries relative to the baseline,which one would anticipate from a labor supplyshock. But the decline has a barely perceptibleimpact on the long-term growth ratefor wages.Native households in high-income countriesenjoy a rise in income, on average, as returnsto capital increase, offsetting the mild declinein wages. The impact on developing countriesis nearly the reverse, with wage income risingas labor-market conditions for workers improve,while returns on capital decline withthe smaller supply of workers. In developingcountries the gain from increased remittancesgreatly exceeds that from changes in factorreturns.The economic benefits for high-incomeeconomies could be even larger than those predictedby the model, due to several factors: themodel excludes the increased productivity ofmigrants (and the benefits to their offspring)over time; investment levels could increasesubstantially in response to higher returns tocapital; labor-force participation could riseamong natives with the greater availability ofmigrant labor (for household help, for example);the labor market would become moreflexibile, and diversity would increase.The costs of adjusting to increased migrationand the gains from migration depend, in part,on the investment climate. Adjustment costsas a result of migration will be lower if moreflexible labor markets and more efficient capitalmarkets in high-income economies reducetransitional unemployment and the cost of replacingcapital as economies adjust to the risein immigration. Similarly, developing countrieswith strong investment climates will beable to use increased remittances more efficiently,and enable workers who do not migrateto respond to improved labor marketconditions. The cost of adjustment may alsobe lower if migration is spread over timerather than concentrated in spurts.A principal conclusion from this exercise isthat migration can generate significant economicgains for migrants, origin countries,and destination countries—but migration alsocan have important political and social consequences.For example, natives in destinationcountries may become concerned about maintainingcultural identity in the middle of agrowing diversity, which also has implicationsrelative to minority languages and other issuessurrounding the integration of migrants. Tosome extent, opposition to migration is drivenby these concerns, and not by an economiccalculation of the gains and losses.We begin with a discussion of recent trendsand discuss how migration to high-incomecountries has grown over the past 30 years.We then turn to the prospects for migration,including the intense pressures generated bydemographic changes. We describe the basecasescenario for migration and the modelbasedanalysis of the welfare gains from increasedmigration. We conclude with issuesthat the model does not consider.International migration trendsMigration to high-income countrieshas acceleratedThe United Nations (UN) estimates that migrantsaccount for some 3 percent of the26

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONTable 2.1 Growth in international migration by destination, 1970–2000Percent change per year in stock of migrants1970–80 1982–90 1990–2000World 2.0 4.4 1.3High-income countries 2.4 2.9 3.1Developing countries 1.8 5.5 –0.1Excluding former USSR 1.9 2.1 0.0Former USSR 0.5 25.0 –0.3Source: United Nations.Figure 2.1 International migrants as ashare of destination countries’ populationPercentage9876543210WorldSource: United Nations.DevelopedcountriesDevelopingcountriesNote: a. Excluding countries of the former USSR.19702000Developingcountries aworld’s population, or about 175 millionpersons. 3 The stock of immigrants to highincomecountries increased at about 3 percentper year from 1980 to 2000, up from the2.4 percent pace in the 1970s (table 2.1). Atthat rate of growth, the share of migrants inhigh-income countries’ population almostdoubled over the 30-year period, and populationgrowth (excluding migration) fell from0.7 percent per year in the 1970s to 0.5 percentin the 1990s. Immigration has had a particularimpact on population growth in severalhigh-income countries. For example, withoutimmigration Germany, Italy, and Swedenwould have experienced a decline in populationin the past few decades (OECD 2005;IOM 2005). By contrast, migration to developingcountries rose by only 1.3 percent peryear from 1970 to 2000. With rapid populationgrowth, the share of migrants in developingcountries’ population (excluding the formerSoviet Union) fell (figure 2.1). 4Most high-income countries saw immigrationrise by at least 2 percent per year from1980 to 2000. 5 This increase reflected, inpart, increased demand for services accompanyingrising incomes, global competitionfor highly educated workers as technologicaladvances boosted the premium for skills, thegrowth of networks of immigrants in highincomecountries that facilitated new immigration,and increased refugee movements.Almost 70 percent of the increase in immigrationis accounted for by the United Statesand Germany, which together make up lessthan 40 percent of the population of thehigh-income countries. In the United States,the Immigration Reform and Control Act(IRCA) of 1986, which provided permanentstatus to 2.7 million migrants, facilitated furtherimmigration through rules governingfamily reunification and may have encouragedfurther irregular immigration (Passel2005) by encouraging expectations of futureamnesties. 6 Germany saw a large inflow ofethnic Germans following the breakup of theSoviet Union (Dustmann and Glitz 2005), aswell as an increase in temporary migrationunder bilateral agreements.Though the stock of migrants has acceleratedsharply relative to the population in theindustrial countries, in some respects the compositionand patterns of international migrationhave exhibited continuity over the pastfew decades. The share of female migrantshas remained almost unchanged (47 percent27

GLOBAL ECONOMIC PROSPECTS 2006Figure 2.2 Share of females ininternational migrationFigure 2.3 Immigration to selectedcountries, reasons for admittance, 2001Percentage515049484719702000AustraliaUnitedStatesFrance46454443WorldDevelopedcountriesDevelopingcountriesPercent80United70 Kingdom6050403020100WorkersReunitingfamiliesRefugee/asylumSource: United Nations.Note: Stock data, reported by countries of destination.Source: United Nations.Note: Stock data, reported by destination countries.of global migrant populations in 1970, comparedwith 49 percent in 2000—figure 2.2),although women are the great majority ofmigrants from some countries. More womentoday are migrating as independent wageearners, rather than to accompany their husbands(IOM 2005). Migration continues to beheavily determined by geographic proximity(from Mexico to the United States, fromNorth Africa to Southern Europe, and fromEastern to Western Europe), as well as bycolonial ties (from Latin America to Spain andfrom a number of Sub-Saharan African countriesto Belgium, France, Portugal, and theUnited Kingdom—OECD 2005). The majorcountries of destination continue to admit thelargest share of permanent immigrants forfamily reunification (or, in the case of the EUcountries, for humanitarian or refugee resettlement),although some countries are refocusingtheir migration policy toward economic(largely skilled) immigration (figure 2.3). 7 Butinternational migration is also changing, particularlyin the direction of flows. For example,more Asians are today seeking work inother Asian countries rather than in the MiddleEast (Wickramasekera 2002; OECD 2005;IOM 2005), while more Latin Americans areturning to Europe for work opportunities, inaddition to North America.It should be emphasized that the migrationdata on which these judgments are basedtend to be unreliable and incomplete. Manycountries and international agencies do notdistinguish between regular and irregularmigration or among types of temporary migration.Some record migrants’ country of birth;others their nationality (OECD 2005).National estimates of the number of migrantscan be vastly different depending on whether“migrant” is defined as foreign born or of foreignnationality.Migration is set to increaseIt is likely that the number of people who wishto migrate from developing to high-incomecountries will rise over the next two decades.About 31 percent of developing countries’population is below the age of 14, comparedwith 18 percent in high-income countries. Wecan thus anticipate a large influx in the agecategories most suitable for emigration, aslifetime earnings from migration tend to belargest for those emigrating early in theirworking life. The surge in immigration sincethe 1980s has established large diasporas inhigh-income countries, which help to reducethe costs and risks of migration (see chapter3). The demand for immigrant services inhigh-income countries will also rise as the28

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONaging of the population shrinks the workforceand increases demand for services thatimmigrants can supply (such as nursing care).As income standards rise, the demand forother services that employ migrants (such ashousehold and restaurant help) should growrapidly. The intensifying competition forskilled workers may also draw migrants, especiallyfrom countries with strong systems ofhigher education.Policies in destination countriescan affect migrationForecasts of migration flows remain problematic.But with the underlying demand for andsupply of migrants likely to increase in comingdecades, the number of migrants will dependon policy decisions governing admittanceand the effectiveness of efforts to policeborders and enforce workplace rules. Oppositionto immigration may grow as the numberof migrants increases, as it did in majorcountries of destination before World War I.But it is likely that the main policy issue willbe how best to manage and live with increasedmigration. In the simulations that follow,we explore the impact of an increase inmigration to 2025 in line with recent historicalexperience.The demographic challengeThe labor force in the high-incomecountries is set to declineA key driver in the demand for internationalmigrants over the next 20 years will be slowinggrowth, and then decline, of the laborforce in high-income countries. The age groupthat supplies the bulk of the labor force(15–65 years old) is expected to peak near500 million in 2010, and then fall to around475 million by 2025 (figure 2.4). In Japan thisage group has already begun to shrink, whilein Europe the peak will be reached in2007–08. In the other high-income countries,the peak will occur later—around 2020 forthe United States and 2015 for the rest. Assumingno change in labor-force participationrates, the high-income countries may loseabout 20 million workers by 2025, relative topeak employment. 8The expected decline in the labor force isaccompanied by a rise in the overall dependencyratio, defined as the ratio of nonworkersto workers. For the high-income countries as agroup, this ratio is forecast to remain at justunder one through 2009. However, by 2025,100 workers will be supporting 111 dependents,largely reflecting the increased numberof the elderly (also, in most countries the numberof children under 15 will fall). The largestrise in the dependency ratio will be in Europe.If we focus more narrowly on the number ofelderly per worker, every 100 European workersnow support 36 elderly people; by 2025they will have to support 52. In Japan 100workers will support 60 elderly in 2025.In the developing countries the laborforce will expandDeveloping countries show considerable diversityin demographic trends, but overall thebulge of youths born over the last two decadesis now entering the labor force, the number ofelderly is as yet still rising slowly, and thenumber of births is falling rapidly. Thusdeveloping countries are forecast to addnearly one billion workers to the world’s laborforce by 2025, again assuming no change inthe labor-force participation rate, and dependencyratios are expected to fall.The expected expansion of the labor forcein developing countries, coupled with largewage premiums in high-income countries,means that migration could help reduce dependencyratios in high-income countries.However, increases in immigration sufficientto have a noticeable impact on dependency ratioswould have to be very large. The scenariodiscussed below envisions an increase in thelabor force in high-income countries of 3 percentthrough migration, or a hike of nearly50 percent in working migrants in high-incomecountries. Even if migrants come with noelderly, the dependency ratio in the host coun-29

GLOBAL ECONOMIC PROSPECTS 2006Figure 2.4 Labor force and dependency ratesEuropeJapanLabor force, millions200Dependency rate1.3Labor force, millions70Dependency rate1.31951.2651.21901851801.11.06055501.11.01750.9450.91702001200420072010201320162019202220250.8402001200420072010201320162019202220250.8United StatesOther high-income countriesLabor force, millions170Dependency rate1.3Labor force, millions80Dependency rate1.31651.2751.21601551501.11.07065601.11.01450.9550.91402001200420072010201320162019202220250.8502001200420072010201320162019202220250.8Total, High-income countriesTotal, Developing countriesLabor force, millions500Dependency rate1.3Labor force, millions3,600Dependency rate1.34951.23,4001.24904854801.11.03,2003,0002,8001.11.04750.92,6000.94702001200420072010201320162019202220250.82,4002001200420072010201320162019202220250.8Labor forceDependency rateSource: World Bank databanks (DDP).Note: The dependency rate is measured as the ratio of the nonworking population to working population.30

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONtries would fall by only about 3 percent undersuch a scenario. In the case of Japan, it wouldlower the number of elderly dependents in2025 from 60 per 100 workers to 59 per 100workers—barely a dent. Nevertheless, asdiscussed in more detail below, selectivemigration—for example, of experienced andskilled workers—can help mitigate the transitionalcosts of financing pension benefits forrapidly aging populations in high-incomecountries.Migration and its developmentimpactTo illustrate the potential gains fromincreased migration, we compare the basecaseforecast for output and consumptionin chapter 1 with an alternative scenario, inwhich the stock of migrant workers is allowedto increase in the high-income countries soas to raise the overall stock of workers by 3 percent(a movement of 14.2 million workersfrom developing countries to high-incomecountries by the year 2025). A firstapproximation of the global gains from such ascenario is simply to calculate the income gainsaccruing to the new migrant workers—thiswill reflect the gains to the global economy,because it approximates the increase in globalproductivity derived from equipping the migrantswith more and improved capital andtechnology. This back-of-the-envelope calculationyields an increase in gross wage income of$772 billion in 2025. 9 As we will see later,when corrected for differences in prices thatmigrants face in high-income versus developingcountries, and taking into account otherimpacts of migration (on prices, for example)as calculated by the model, global gains fall to$356 billion—an 0.6 percent increase in globalincome. The scenario is particularly beneficialto developing countries relative to high-incomecountries. The aggregate percentage gain todeveloping countries (including the new migrants)is 1.8 percent, whereas the gains tonatives in high-income countries amount to0.4 percent relative to baseline income. Thesenumbers hold up well as an approximation ofthe gains to global output, regardless of variousassumptions made about taxes, non-wageincome distribution, key model parameters,and other factors.Our modeling exercise uses a global generalequilibrium model to measure the impactof migration (box 2.1). 10 One of the purposesof the global model is to verify the basic intuitiondescribed above—that migration producesa sizeable global gain. But it also is apowerful tool to evaluate distributionalimpacts—between skilled and unskilled workers,between native- and foreign-born workers,between capital and labor, and acrossregions—and to show how these distributionalimpacts vary with policy choices andparameters (for example, the role of fiscalpolicies or the propensity to remit).The assumption is that migrants as ashare of population remain constantin the baseline scenarioWe begin with a base case for global economicactivity (outlined in chapter 1), demographictrends (described at the outset of thischapter), and for migration. For the base case,the proportion of migrants in each region remainsthe same over time—somewhat contraryto the trends of the last two decades.This does not imply that gross migration isstagnant, or even declining. The stock of migrantsin any year will equal the previousstock of migrants, plus new migrants, less theattrition through death and return migration.We chose a relatively neutral assumption becauseof the difficulty in forecasting thesecomplex processes. For some countries—forexample Japan and those in Europe—the assumptionresults in an absolute decline in thestock of migrant workers. This decline parallelsthe overall decline in the European andJapanese labor forces. 11 For the high-incomecountries as a group, the stock of migrantworkers would increase by some 760,000between 2001 and 2025, just a small incrementfrom the estimated 27.8 million in 2001. Themain issue, however, is not the base case,31

GLOBAL ECONOMIC PROSPECTS 2006Box 2.1The model used in this studyThe underlying analytical framework used in thischapter is the World Bank’s standard globalgeneral equilibrium model—LINKAGE—which hasbeen used in previous reports for trade policy analysis.It has been modified to differentiate between migrantand native workers and to incorporate remittances.The model is based on release 6.0 of theGTAP database (base year 2001), developed by theGlobal Trade Analysis Project (www.gtap.org), aglobal network of researchers and policymakers engagedin the quantitative analysis of internationalpolicy issues. It is supplemented for use in ourmodel with a new database developed jointly byGTAP and the University of Sussex (Parsons andothers 2005). That database contains a comprehensiveestimate of bilateral stocks of migrants for 226countries and territories.While the new migration database is undergoingconstant improvements as new data become availableand obvious errors are corrected, its developershave done a remarkable amount of detective work,largely in national data sources. The GTAP centerhas used this underlying migration database to builda bilateral migration database for the 87-region levelof aggregation of the main GTAP database—including estimates of population and the stock ofworkers, both skilled and unskilled (Walmsley,Ahmed, and Parsons 2005). a World Bank data (describedin more detail in chapter 4 of this report)was used to provide the total level of remittances,and the bilateral stock of migrants was used to estimatethe bilateral remittance flows subject to theoverall total flows.The standard horizon for the LINKAGE model hasbeen 2015. For the work described here, the modelhorizon has been extended to 2025, in part becausedemographic dynamics play a more important roleover the longer-term horizon, and in part to allowfor more time to phase in the increase in migration.a The 87 regions of GTAP have been aggregated into 21 regionsfor the purposes of this study. Six of these are high-incomeregions using World Bank definitions—the EuropeanUnion and the European Free Trade Area, Canada, the UnitedStates, Japan, Australia/New Zealand, and the newly-industrializingeconomies. The fifteen developing countries/regions includeChina, the Philippines, India, Russia, Turkey, SouthAfrica, and Mexico as individual countries, plus 6 regions thatrepresent the remaining countries in each geographical area.but rather the impact of deviations fromit—although significantly different base assumptionscould affect the deviations aswell.According to the base-case scenario, migrantworkers would make up about 6 percentof the labor force of high-income countries in2025, though with sharp differences acrossregions and skills (table 2.2). The vast majorityof migrant workers are unskilled—some25.3 million migrant workers out of a projectedtotal of 28.5 million, or 7.8 percent ofhigh-income countries’ labor force. Skilled migrants,on the other hand, represent just 2.2 percentof the total skilled workforce on average.There are welfare implications ifmigration rises significantlyThe alternative scenario involves a rise inmigration sufficient to increase the labor forceof high-income countries by 3 percent, phasedin from 2010 through 2020. 12 As migrantsmake up about 6 percent of high-incomecountries’ labor force, a 3 percent rise in thelabor force (through migration) implies a50 percent increase in the number of migrantworkers. This may seem like a large change,but the resulting stock of migrants in Europe,Japan, and the United States would remain afar smaller share of population than currentlevels in some high-migration countries. (InAustralia, for example, about a quarter of thepopulation are migrants, in Canada 19 percent,in Kuwait 50 percent). The percentageincrease in migrants is large in Japan (as thebaseline share of migrants is relatively low),and lower in the United States. The increasecorresponds to an annual growth rate ofabout 1.9 percent, somewhat slower than the32

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONTable 2.2 Labor force structure in the base case and after increases in migrantsIn millions except where notedBaselineMigration shock2001 2025 Change in millions 2001–25 Change in percent 2001–25High-income countriesTotal labor force 480.8 474.0 14.2 3.0Developing-country migrant workers 27.8 28.5 14.2 49.9Unskilled 24.6 25.3 9.8 38.6Skilled 3.1 3.2 4.5 137.9Developing-country migrant workersas share of total labor force, percent a 5.8 6.0 8.8Unskilled, percent 7.4 7.8 10.5Skilled, percent 2.1 2.2 5.0Developing countriesTotal labor force 2,596.2 3,561.0 –14.2 –0.4Unskilled 2,395.9 3,294.3 –9.8 –0.3Skilled 200.4 266.7 –4.5 –1.7Source: Initial 2001 data from migration database under development by GTAP/University of Sussex (Parsons and others 2005and Walmsley, Ahmed, and Parsons 2005). Scenarios based on World Bank assumptions.Note: a. The percentage of migrant workers as a share of the total labor force is assumed to be the same for each individualregion of the model throughout 2001–25, but the share averaged across all developed regions will change through aggregationeffects.average increase over the period 1980–2000.Moreover, the growth rate is unbalanced, withan annual increase of only 1.5 percent in unskilledworkers, but 3.8 percent in skilledworkers. A number of additional assumptionsare critical to the results.First, the high-income countries’ laborforce of both skilled and unskilled workersincreases by 3 percent. 13 As the share of skilledworkers among migrants is much smaller thanthe share of skilled workers among highincomecountry natives, the shock results in amuch larger percentage increase for skilledmigrants. The number of unskilled migrantworkers increases by 39 percent, while thenumber of skilled migrant workers rises by138 percent. 14Second, the share of migrants by region oforigin remains constant; in other words, thenew migrants reflect the same allocation byregion of origin as existing ones. Thus ifMexicans constitute 30 percent of foreign migrantsin the United States in the base case,they maintain the same share after the increasein migration. This assumption is made to simplifythe analysis, although it does fail toreflect the likely migration pressures impliedby large differences in demographic trends insending regions (for example, Sub-SaharanAfrica versus Latin America).Third, foreign workers are assumed tobring family members in proportion to thedependency ratio in their home country. As aresult, the total number of migrants in highincomecountries increases from 65 million(6.5 percent of high-income countries’ population)in the baseline for 2025, to 93 million(9 percent of population) after the shock. Thisassumption can change the average dependencyratio of the host country. It canalso have other implications not modeledexplicitly—including fiscal impacts, becausethe families of new migrants may require additionalpublic services (such as schooling),not fully compensated by the taxes paid by thenew migrants.Fourth, remittances are assumed to be afixed proportion of migrants’ labor income,equal to the level in the base year. The averagefor developing countries is 17 percent, althoughthe level varies with the migrant’s origin anddestination countries. New migrants are33

GLOBAL ECONOMIC PROSPECTS 2006assumed to send remittances to their homecountry at the same rate (relative to income)as existing migrants. 15Returns to householdsThe gains from increased migrationare largeWith the labor force moving, it is best toassess the effects on real income in terms ofhouseholds as opposed to the national level(as is typically done in analyses of trade reform).Households are broken down into fourgroups. First are the native households inhigh-income countries. 16 Second are previousmigrants from developing countries now livingin high-income countries, that is, thosewho were in place in the baseline scenario.Third are native households in developingcountries—households that do not migrate. 17And finally, we have the households of thenew migrants. Each household’s welfare isbroken down between the change in privateconsumption and the change in the consumptionof public services.Natives in high-income countries gain$139 billion in real income, or 0.4 percent ofthe baseline, as a result of the rise in migration(table 2.3). Nonmigrating households in developingcountries see a rise in real income ofnearly 0.9 percent from baseline levels. 18 Asignificant portion of the increase is due to theremittances from the new migrants, with someimprovement in labor-market conditions forremaining workers. Those who are likely tolose—in the absence of any compensatorymechanism—are the existing migrants inhigh-income countries, who are relativelyclose substitutes for the new migrants. Theirprivate consumption would decline by over9 percent and overall consumption (includingpublic services) by 6 percent compared tobaseline levels.New migrants and their countriesof origin reap benefits(through remittances)The main gains come from the higher incomesthe new migrant workers can earn in the destinationcountry relative to what they wouldhave earned in their country of origin. Newmigrants earn $481 billion in real (after-tax)income in 2025 over the base case. However,the dollar increase in income overestimates thewelfare gains for migrants. Essentially, anadditional $1 spent in the high-income countriesdoes not provide the same amount ofwelfare as an additional $1 spent in the homecountry, because prices are higher in highincomecountries. Whereas the prices ofTable 2.3 Change in real income across households in 2025 relative to baselineReal incomeReal income adjusted forcost of livingPrivate Public Total Private Public TotalChange, $ billionsChange, $ billionsNatives in high-income countries 139 –1 139 139 –1 139Old migrants in high-income countries –88 0 –88 –88 0 –88Natives in developing countries 131 12 143 131 12 143New migrants 372 109 481 126 36 162World total 554 120 674 308 48 356Change, % Change, %Natives in high-income countries 0.44 –0.01 0.36 0.44 –0.01 0.36Old migrants in high-income countries –9.41 –0.02 –6.02 –9.41 –0.02 –6.02Natives in developing countries 0.94 0.44 0.86 0.94 0.44 0.86New migrants 584 607 589 198 203 199World total 1.20 1.15 1.19 0.67 0.45 0.63Source: World Bank model simulations.34

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONtraded goods (for example, cars and electronics)are the same worldwide, at least in principle,the prices of nontraded goods and services(for example, housing and haircuts) are muchhigher in high-income countries.A simple example may clarify the idea.Take a household of two persons living intheir home country. One works and earns$200. The other does not work. Each spends$100, half on tradable goods (each priced at$1) and half on nontradable goods (likewisepriced at $1). Now the worker moves to ahigh-income country and earns $700. Assumethat spending patterns do not change. Theworker remits $200 back to the home country,so the income (and welfare, in money terms)of the other doubles. The new migrant buysthe same goods—50 units of tradable goodsand 50 units of nontradable, 19 but the price ofthe latter is now $9 and not $1. The migrantthus spends $500, but welfare is unchanged,because the basket of purchased goods isidentical.Welfare evaluations are of course morecomplex than this simple example illustrates.For one thing, new migrants will have to adjusttheir spending patterns to deal with their newenvironment. Heating oil and warm clothesare necessities that will not boost a migrant’swelfare above what it was in the home country.For another, the decision to migrate is nottaken for simply static reasons; there are significantdynamic reasons for migrating—forexample, better opportunities for one’s childrenthat are not captured in this simple framework.Nonetheless, the difference in purchasingpower illustrated in the example is a strongmotivation for migrating, even on a temporarybasis. The more wage income earned in highincomecountries that can be spent in lowerincomecountries, the greater will be the welfarebenefits. Box 2.2 provides additionaldetail on the computation and interpretationof global welfare gains from migration.To account for the change in prices facedby the new migrants, their “new” consumptionin the destination country is adjusted toaccount for differences in the cost of living,using purchasing power parity (PPP) exchangerates from the World Bank’s database. 20 Thusinstead of an increase of $481 billion, the risein welfare for new migrants is $162 billion. 21Table 2.3 shows the change in these componentsfor the four household groups and theworld. Measured in national accountingterms, that is, with no adjustments for the differencein the cost of living for the new migrants,global real income rises under themodel by 1.2 percent relative to the baseline,or 0.6 percent with the cost-of-living adjustment.Global private consumption increases inreal terms by $308 billion in 2025 (with thecost-of-living adjustment), with real governmentexpenditures increasing by an additional$48 billion. The total real gain—with equalweight for high-income—and developingcountrygains—is $356 billion, with justunder half accruing to the new migrants,though natives in both high-income and developingcountries also are better off. In percentageterms—where relative weights betweenhigh-income and developing countries areirrelevant—the scenario clearly indicates thatthe relative gains are much higher fordeveloping-country households than highincomecountry households, rivaling gainsfrom global reform of merchandise trade.Obviously, global income and global gainswould also be larger if expressed in PPP terms.As the percentage increase in welfare for migrantsliving (originally) in developing countriesis larger than the percentage increase forthose living in high-income countries, a switchto PPP measures would also increase theglobal gains as a percentage of global income.If in the migration scenario presented here thegains are PPP-adjusted, the global gains wouldamount to 0.9 percent of global income in thebaseline, instead of 0.6 percent using the EVaggregation. This scenario illustrates that migrantsliving (originally) in developing countriesgain the most from migration in percentageterms.The impact of higher migration on prices ismild in aggregate in high-income countries,with a small decline in the average price of35

GLOBAL ECONOMIC PROSPECTS 2006Box 2.2 Calculating and interpreting global welfaregains from migrationTwo sets of issues arise with respect to the socalledglobal gains from a policy shock. First,how should the gains of specific groups be evaluatedand how do the gains compare with traditional measures,such as GDP or national accounting standards?Second, how should the gains be aggregatedover groups and countries, and how should theaggregated gains be interpreted?Evaluation of the welfare gains of specificgroups. In standard applications of general equilibrium(GE) models, the welfare impacts of specificgroups are evaluated using a concept from welfaretheory called equivalent variation (EV). The conceptis relatively straightforward. Welfare changesas a result of changes in nominal income andchanges in prices. EV calculations summarize thiswelfare change in terms of an equivalent changein income alone, showing by how much incomeat original prices would have to change toachieve the same change in welfare as observed ina simulation. aFor most households, the standard notion of thechange in real income, that is, the difference in nominalincome adjusted by the change in the CPI, is agood approximation of EV. bThis is not the case for new migrants, however.There is no standard price index that can be used asa deflator for the change in the nominal gains for thenew migrants, since the prices they face in their newhost country have no linkages to the prices they paidin their home countries. GE and macro models typicallycalibrate base-year prices in each region to one(or unit value) by choosing corresponding volumeunits. cThis approach does not allow one to take into accountthe price increases that new migrants face asa result of their migration. In the simulations, themacro PPP exchange rate (as an approximation ofthe rise in prices faced by migrants from developingto high-income countries) has been used to adjust thegains to the migrants—although this is just an approximationof the true welfare gains. dBecause of the cost-of-living adjustment to thewelfare gain of new migrants, the real gain reportedis no longer equal to real income gains ofcountries—and real output gains—measured usingnational accounting standards. However, the standardreal income measure is still a good approximationof the welfare gains for the other households inthe model.To the extent that new migrants remit part oftheir income to their country of origin and thatincome is spent in that country of origin, the increasein the cost of living that new migrants face is notrelevant. Therefore, the EV measure of remittances islarger than the same nominal income spent by thenew migrant in the host country. This differenceillustrates the incentive for new migrants to remitincome home.Aggregation. The second issue relates to the interpretationof the “global” gains. Typically, to deriveaggregate or global gains, EV (expressed in a commoncurrency, typically the U.S. dollar) is summedacross all households. For individual persons or homogeneousgroups this EV aggregation, expressed asa percentage of original income, is a good approximationof the change in welfare (or more precisely, itis a good indication of the change in welfare). eHowever, no clear link exists between global welfareand the aggregation of EV across heterogeneousgroups, because we do not know how to weigh individualwelfare across heterogeneous groups (a particularlydifficult issue in aggregating across countriesat very different stages of development, as is donehere). For example, while most groups gain from migrationin the scenario discussed in the text, somelose. The fact that the change in global welfare (expressedas the aggregation of EV across groups) ispositive does not mean that the welfare gains of thewinners are considered more important than the welfarelosses of the losers. Thus, global gains as expressedin aggregate EV should not be interpreted asa value judgment on how to weigh individual orlocal welfare gains.The aggregation of EV across groups does,however, have a useful interpretation, which islinked to the notion of compensation and Paretooptimality. As long as the global gains are positive—using the standard practice of adding up EVs acrosshouseholds—then it is possible through redistributionto compensate households that lose (so that noone is worse off relative to the baseline scenario),36

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONBox 2.2(continued)when some households are better off. In that sensethe global gain can be compared with an equal risein global output plus redistribution.In this report we maintain this standard practiceof reporting EV aggregates, making the gains comparableto global gains in many other studies.An alternative approach to calculating globalgains would be to add up changes in income measuredin PPP terms. The rationale for that alternativeis that because prices of nontraded goods arelower in developing countries, the addition of a dollarto a developing country would enable the purchaseof a larger amount of goods and services thanin an high-income country. In that case, both baseincome and gains for new migrants and for thosewho remain in developing countries would beroughly three times as large as reported here. Thisis true for all gains, whether they come from migrationitself, from remittances, or from changes inwages and prices in developing countries. As a result,the share of those who live (originally) indeveloping countries in global aggregates would increasein the measurement of both global incomelevels and global welfare gains. However, the percentageincrease in income for developing countrieswould not be affected.a One of the advantages of the EV measure is that it transformsthe ordinal concept of welfare into a cardinal concept ofincome. While it is impossible to measure how much one welfarelevel differs from another (one can only conclude that one level ispreferred to another), the corresponding increase in income canbe measured, and the size of the increase has a clear meaning.b For example, in trade-reform scenarios, the change in theprice index is a relatively good approximation of the welfareimpact, since the new price is approximately the old price lessthe tariff.c There are exceptions. For example, in the case of climatechangemodels, it is necessary to know the relative prices of thedifferent fuels to accurately determine the carbon tax.d See Timmer and van der Mensbrugghe (2005) for moredetails.e The size of the change in individual welfare is undetermined,since welfare is an ordinal concept.absorption (private consumption, private investment,and government spending) of 0.1percent. However, prices of some key nontradablesdecline by larger amounts—0.8 percenton average for public services (includinghealth-related services) and 0.2 percent for constructionand recreational services. These pricedeclines will be even sharper for specific subsectorswhere migrant workers are concentrated(for example, household help), for whichwe currently have no comprehensive data.The allocation of the gains across developingcountries depends on various factors, includingthe skill loss and the resulting impacton production, the locations to which migrantsmove and the relative wage differential,and the propensity to remit. By developing region,the gains to households under the modelvary from 0.6 percent for Europe and CentralAsia to 1.1 percent for South Asia and LatinAmerica and the Caribbean (table 2.4).For the new migrants, the real incomegains—cost-of-living adjusted—increase bynearly 200 percent. There are large differencesacross regions, with the highest gains (in percentageterms) accruing to migrants from Sub-Saharan Africa (619 percent) and the lowestto migrants from the Middle East and NorthAfrica and Europe and Central Asia. The mainreason for the disparity is the relative differentialbetween wages in origin and destinationcountries. Variations in wages paid to migrantsfrom different regions in destinationcountries are minor, whereas there are verywide variations in wages in countries of origin.For example, the average wage for a migrantin Europe in the base year is about$16,500—with only minor variation acrossmigrants. However, the average wage in Sub-Saharan Africa is only $470, whereas in theMiddle East and North Africa it is $2,700.Thus, the migrant from Sub-Saharan Africa37

GLOBAL ECONOMIC PROSPECTS 2006Table 2.4 Real income impacts across developing regionsChange in 2025 relative to the baseline, adjusted for differences in cost of livingNatives in regionNew migrants from region$ billions Percent $ billions PercentTotal developing 143 0.9 162 199East Asia and Pacific 37 0.7 32 215South Asia 21 1.1 2 175Europe and Central Asia 14 0.6 25 138Middle East and North Africa 18 0.9 11 134Sub-Saharan Africa 7 0.9 7 619Latin America and the Caribbean 47 1.1 85 224Source: World Bank model simulations.will gain much more in both absolute and percentageterms than one from the Middle Eastand North Africa.The impact of migration on tradewould be mildWhether migration and trade are substitutesfor each other is an old debate. For example,in the discussions leading up to the signingof NAFTA—the free trade agreementamong the United States, Canada, andMexico—one of the key arguments was thattrade would replace migration and reducethe pressure for Mexicans to migrate to theNorth. Likewise, allowing for increasedmigration—for example of unskilledworkers—could reduce trade, because itwould enable the high-income countries tocontinue producing low-skill-intensive productsat competitive cost.Evidence of the link between trade andchanging the comparative advantage emergesin the migration scenario described here. Forexample, the largest gains in export revenuefor high-income countries come in agriculture,clothing, other manufacturing, recreationalservices, and public services—all laborintensivesectors, the first four being relativelyintensive in unskilled workers and the last inskilled workers.Change in comparative advantage has onlya mild impact on trade flows in this scenario,however, as migration affects trade throughseveral channels, some of which increase, andothers that decrease, trade flows:• First, the rise in incomes due to migrationproduces a small rise in global tradeflows, with regional differentiation (becauseincome gains differ considerablyamong regions). In addition to higher incomes,the rise in migration changes thesize of regional economies, with implicationsfor their demand for imports andability to export.• Second, the nature of the shock assumedin our model differs from the standarddebate over trade and migration. Theshare of skilled workers in total migrantsis larger in the shock than in actual migrationover the recent past. A largeproportion of skilled workers will findemployment in nontraded sectors—forexample, as doctors and nurses—ratherthan in producing traded goods. Thiswill have general equilibrium effects tothe extent that the price of nontradedgoods will decline by more than the priceof traded goods. Thus there will be a relativeshift to nontraded goods and apotential reduction in demand for importsof traded goods. Overall, the largershare of skilled versus unskilled workersdoes tend to reduce trade flows.• Third, the increase in remittances providesan opportunity for developing38

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONcountries to import more and export less,as their current-account balance will increaseby the size of the remittances ($98billion in net terms). The model resultsshow that total imports into developingcountries would increase by $58 billionin 2025 (1.1 percent relative to the baseline),as aggregate exports decline by$40 billion (0.7 percent). 22 The changein remittances leads to an appreciation ofthe real exchange rate and therefore aloss in relative export competitiveness. 23For instance, the output price index indeveloping countries rises by 0.6 percenton average, whereas it declines by 0.1 percentfor high-income countries.In summary, the scenario provides evidencethat changes in comparative advantage due tomigration do influence trade flows. However,overall migration and trade are not substitutesfor each other, because migration has manyother economic effects that have more powerto stimulate or reduce trade. One implicationof this finding is that migration policies shouldnot be pursued because of their specific impacton trade flows. Likewise, in trade policies theimpact on migration should not be a mainfocus. 24 Trade and migration policies shouldbe evaluated on their own merits.Migrants’ impact on government fiscalaccounts is broadly neutralThe assumption concerning the level ofconsumption of public goods and services bynew migrants has important implications forindividual gains, and global gains, under themodeled scenario. We assume that the newmigrants’ level of consumption of publicgoods and services equals the amount they payin taxes, that is, their impact on the publicbudget is revenue-neutral. This is broadly consistentwith the available evidence (box 2.3).To provide some sense of how different approacheswould affect the scenario results, wepresent two alternative assumptions regardingthe distribution of public goods and servicesto the new migrants (table 2.5). The defaultassumption had a largely neutral impact forexisting residents in the host country. Underanother assumption—new migrants pay taxesbut receive no benefits from public goods andservices—existing residents, native and migrant,enjoy a rise in real incomes of $126 billion($117 billion for natives and $9 billionfor existing migrant households). Note thatthe global welfare gains increase as well, sincethe income accruing to natives (and existingmigrants) is not adjusted for the differences inthe cost of living between developing andhigh-income countries. 25 A second extremeTable 2.5 Impact of different assumptions on the consumption of public goodsand services by selected groups in 2025Change in cost-of-living-adjusted real income in 2025; billions of dollarsDefaultassumption—“New” migrants“New” migrantsreceive benefits receive no public “New” migrantsequal to their benefits but pay receive per capitataxes taxes average benefitPrivate Public Total Public Total Public TotalNatives in high-income countries 139 1 139 117 256 85 54Old migrants in high-income countries 88 0 88 9 79 6 94Natives in developing countries 131 12 143 12 143 12 143New migrants 126 36 162 18 108 75 201World total 308 48 356 120 428 4 304Source: World Bank model simulations.39

GLOBAL ECONOMIC PROSPECTS 2006Box 2.3The impact of immigrants on fiscal balancesImmigrants’ net contribution to fiscal revenues isusually considered to be small. The net fiscal impactof immigration on the United States has beenminimal (Coppel, Dumont, and Visco 2001;Auerbach and Oreopoulos 1999). The U.S.Binational Study on Migration (1997) found that irregularmigrants did impose a significant fiscal burdenon state and local government. However, schoolexpenses accounted for the bulk of these costs, and(as the authors note) education is an investmentthat may readily be recovered in greater future productivity.Moreover, Lee and Miller (2000) foundthat the overall fiscal consequences of altering thevolume of immigration to the United States wouldbe quite small. Gott and Johnston (2002) andSriskandarajah, Cooley, and Reed (2005) estimatedthat immigrants made a positive contribution topublic finances in the United Kingdom. Gustafssonand Osterberg (2001) found that new immigrants toSweden generated a net fiscal cost, but this turnedinto a positive contribution after a few years. Nanaand Williams (1999) found that immigrants to NewZealand had a positive fiscal impact. Bonin,Raffelhuschen, and Walliser (2000) found that thenet fiscal contribution of immigrants to Germanycould be significant if the government selects forskills.Calculations of the net fiscal cost of immigrationare fraught with difficulties, for several reasons.First, the computation at any point in timedepends heavily on the methodology used, whatexpenditures and revenues are included, which publicservices should be regarded as pure public goods (andthe extent of economies of scale in expenditures), andwhether households or individuals are considered.Second, static calculations of the current net fiscalimpact fail to take into account the age structure ofthe immigrant population. Smith and Edmonston(1997) found that immigrants arriving between theages of 10 and 25 years produced fiscal benefits undermost scenarios, while immigrants arriving in their late60s generally imposed a long-term burden. Studiesthat follow immigrants over time generally concludethat in net-present-value terms, immigrants and theirdescendants tend to contribute more in terms of taxrevenues than they absorb in expenditures, but the ordersof magnitude are typically small (OECD 1997).Intergenerational models are sensitive to the discountrate used and assumptions concerning the allocationof the fiscal burden over future generations.Third, the computation will depend on the levelof skills, experience, education, and fertility of immigrants.Rowthorn (2004) calculates that skilled migrantsto the United States typically make a largepositive contribution to the fiscal balance, whereasunskilled immigrants cost more on average than thetaxes they pay. Storesletten (2000) calculates that thenet-present-value contribution of the average highskilledimmigrant to the U.S. budget is $96,000; themedium-skilled immigrant’s contribution is $2,000;and low-skilled immigrant’s contribution is$36,000.The results may change over time, as migrantcharacteristics and government policies change. Theprobability that an immigrant to the United Stateswill receive public benefits has risen since the 1970s,probably due to an increasing share of immigrantsfrom poorer countries (Gustmann and Steinmeier1998).An issue of particular concern has been the impactof migration on government-financed pensions.Likely increases in immigration can make only asmall net contribution to strengthening the financingof pensions in the United States (Fehr, Jokisch, andKotlikoff 2004), although selecting immigrants forworking age and high skill levels could improve thepicture (Storesletten 2000). By contrast, increases inimmigration could make a significant contributionto financing pensions in Germany (Bonin, Raffelhuschen,and Walliser 2000) and Spain (Collado,Iturbe-Ormaetxe, and Valera 2004).assumption is that new migrants receive thesame amount in public benefits as the averagehousehold in the destination country. Thiswould imply a net positive transfer to the newmigrant households, since they would receivemore in public benefits than they paid intaxes. 26 In this case natives in high-incomecountries would lose $85 billion in aggregatepublic goods and services, although thisamount would not translate one-for-one into abenefit for new migrants due to the cost-oflivingadjustment. These simulations underline40

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONthe effect of public policy on the distributionof gains from migration.Additional gains from migrationcan be substantialThe gains for migrants from this scenarioessentially provide the same message as earlierestimates. In their seminal paper, Walmsleyand Winters (2003) estimate that a relaxationon the movement of temporary workerson the same order as that modeledhere—that is, 3 percent of the labor force ofthe high-income countries—would yieldglobal income gains of $150 billion (using a1997-based comparative static model). Theresult from our scenario that is roughlycomparable to their figures (that is, globalgains before adjustment for cost of livingand measured relative to 2001, rather than2025) are more than double their results. 27However, our figures are comparable withthe more recent work done by Walmsleyand her colleagues. 28 One of the key reasonsfor the increase in the global welfare impactis a reevaluation of the assumed wage differentialbetween the home and host country.In their initial work, Walmsley andWinters had assumed that new migrantsmade up 50 percent of the difference betweenthe home and host country’s wages.Their new assumption (used in our model aswell) is 75 percent, based in part on the factthat the migrants are permanent rather thantemporary. Hamilton and Whalley (1984)and Moses and Letnes (2004) have shownthat removing all restrictions on labormovement, admittedly not a realistic scenario,would yield a huge increase in worldoutput. Overall, these papers suggest thatlabor-market restrictions are imposing amuch larger burden on the global economythan are trade restrictions. The World Bank’strade model suggests that removing all remainingmerchandise trade barriers wouldyield $287 billion in global real income gainsin 2015. For the purpose of comparison, whenthe gains from the two different scenarios—those from an increase in migration, and thosefrom global trade reform—are scaled to thesame reference year, 2001, the gains fromtrade reforms are $155 billion versus $175 billionfrom the migration scenario. 29 Thisleaves little doubt that easing restrictions onthe movement of labor could provide a significantboost to the global economy. Moreover,in comparison with the most recentwork on global merchandise trade reform,the gains from an increase in migration aremore balanced toward income increases fordeveloping countries relative to developedcountries. In a study by Anderson, Martin,and van der Mensbrugghe (2005), the gainsto high-income and developing countries are0.6 and 0.8 percent, respectively, relative tobaseline income. In the scenario modeledhere, the income increases are 0.4 percentfor native households in high-income countriesand 1.8 percent for developing countries(including the new migrants).Returns to factors of productionFour critical factors determine the distributionof gains from migration among skilledworkers, unskilled workers, and owners ofcapital: (a) the size of the increase in migration;(b) the distribution of nonwage income(profits); (c) the degree of substitution betweenworkers by region of origin; and (d) the degreeof substitution or complementarity betweenworkers and capital. We have already positedthat the increase in migration is large, with anaverage increase in the migrant labor force ofaround 50 percent over a 20-year period, andcomparable (if somewhat less) to the rise in theshare of migrants in high-income country populationover 1970–2000. In the absence of anyspecific data on the source of migrant income,we assume that migrants—both existing andnew—receive no nonwage income. In essence,their real income will be driven by changes inwages. The effects of this simple assumptionon the distribution of gains are significant, andthe implications of relaxing it are discussedbelow.41

GLOBAL ECONOMIC PROSPECTS 2006Substituting between migrant and nativeworkers determines who gainsThe key issue of who reaps the benefits involvesthe degree of substitution among differentworkers. The allocation of demand forworkers assumes differentiation among workersfrom different regions. This is done in twosteps. First, “similar” workers are bundled togetherinto “native” workers and “foreignborn”workers. 30 In the second step, thesetwo bundles are decomposed into labor demandby region of origin. We assume thatthere is more differentiation between a nativeand a foreign-born worker (that is, a lowersubstitution elasticity) than between twoworkers from different countries of originwithin each of the two aggregate bundles. Forexample, in the case of the United States, employerssee a greater difference between a U.S.worker and a generic immigrant from a developingcountry than between a Mexican and aSalvadoran worker. The implication is that arise in the supply of migrants has a greater impacton old migrants than on native workers,which plays a key role in the distributionaloutcomes of the increase in migration. Theassumption of labor demand differentiationoperates for both skilled and unskilled laborin the model.In the default case, we assume that wagesare flexible, with a substitution elasticity betweenunskilled migrants and natives that isroughly comparable to that implied in theconclusions of the meta analysis in Longhi,Nijkamp, and Poot (2004); they concludethat a “one percentage point increase in theproportion of immigrants in the labor forcelowers wages across the investigated studiesby only 0.119 percent.” (See box 2.4 for a reviewof empirical studies of the impact of migrationon wages.) In the scenario describedhere, the 50 percent increase in the stock ofmigrants raises their proportion of the laborforce by about 3 percentage points, producinga 0.5 percent decline in the wages of natives.31 We also assume perfect substitutionbetween new and old migrants (the large majorityof both categories being unskilled). Theempirical evidence of the extent to which migrantsare substitutes for natives or for existingmigrants is sparse. Thus in addition toexploring the implications of the assumptionsmade, we also devote attention to alternativeassumptions.Finally, in a departure from previouswork but in line with a developing consensus,we assume that skilled workers are nearcomplements with capital (meaning that theyare more productive, and thus earn higherreturns, when used together with capital),whereas unskilled workers are substitutesfor capital and skilled labor. 32 This specificationhas important consequences for thedistributional impacts of increased migration.Whereas investment rises with increasedincome, the overall increase in thestock of capital is modest, so that the rise inthe supply of skilled workers is not matchedby an equivalent increase in capital. Thus themarginal productivity of additional skilledworkers declines, provoking a decline in thewage of skilled workers (by more than thefall in the wages of unskilled workers).Increased migration can generatesubstantial changes in income distributionamong workers and owners of capitalThe change in factor returns is depicted infigure 2.5. In the high-income countries onlyFigure 2.5 Factor returns and migration% change in factor returns, 2025, relative to baseline1050510152025CapitalDeveloping countriesUnskillednativeworkersHigh-incomecountriesSource: World Bank simulations.UnskilledforeignbornworkersSkillednativeworkersSkilledforeignbornworkers42

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONBox 2.4 Empirical studies of the impactof immigration on wagesMost cross-sectional studies find that immigrantshave no impact, or a very limited impact, onthe wages or employment of natives (LaLonde andTopel 1997 and Borjas, Freeman, and Katz 1997 forthe United States; Pischke and Velling 1994 forGermany). However, cross-sectional approaches relatewage differences across local labor markets tothe share of immigrants in each market. If immigrantsare attracted to high-wage areas, which islikely, it is difficult to identify the exogenous impactof immigrants on wages. a Studies of sudden, politicallydriven inflows of immigrants likewise fail to detecta significant impact on natives’ wages or employmentin affected areas. b However, native workersmay adjust to large, sudden inflows of immigrants bymoving to other areas (or through reduced inflowsfrom other areas), again obscuring the relationshipbetween immigration and labor-market outcomes. cThis problem has encouraged the use of paneltechniques that can discern the combined effects oftime and cross-sectional effects. Some panel studieshave found a significant impact on the wages of unskillednatives, who in addition have suffered declinesin wages due to skill-biased technical changeand increased trade. Borjas (2003a), analyzed the impactof immigration in the United States across differentlevels of skills and experience and estimatesthat immigration reduced the wages of native highschooldropouts in the United States by 8.9 percentfrom 1980 to 2000. Jaeger (1996) finds that immigrationlowered the real wage of U.S. high-schooldropouts by as much as 3.6 percent in the 1980s.DeNew and Zimmermann (1994) find that a onepercentage point rise in the share of migrants in thelabor force reduces the wages of blue-collar workersby almost 6 percent. By contrast, Dustmann and others(2003) find that immigration has little impact onnative wages (or employment) in the UnitedKingdom, including for the low-skilled.Where wages are relatively inflexible, an inflow ofmigrants may affect employment levels rather thanwages. Angrist and Kugler (2002) find that increasedimmigration in Europe is associated with a significantdecline in native employment, particularly forthe low-skilled. Hunt (1992) finds that a one percentagepoint rise in the share of immigrants in theFrench labor force (following Algerian independence)increased the unemployment rate by 0.2 percentagepoints. This shows more adjustment through employmentthan through earnings compared to U.S.studies, which may be due to French rules governingwages (Dustmann and Glitz 2005). The comparisonunderlines the importance of the investment climate,and in particular, labor-market flexibility, in the efficientabsorption of migrants. Some studies show thatimmigration reduces native unemployment in thelong term (Poot and Cochrane 2004), presumablybecause increased consumption demand from immigrantsraises the demand for labor. dThus some articles support the view that unskilledimmigrants are relatively close substitutes for nativeworkers (without attempting to distinguish betweenthe effects on native workers and old migrants). However,the share of low-skilled native workers in destinationcountries is falling. The share of U.S. adultswith less than a high-school education declined from47 percent in 1970 to 22 percent in 1998 (Massey2000). About 90 percent of new native entrants to theU.S. labor force in 2004 had completed high school(U.S. Labor Survey 2005). By contrast, the averagemigrant from rural Mexico has six years of educationand does not speak English (Mora and Taylor 2005).Many low-skill immigrants may have such limited educationand language skills that they do not competewith native low-skilled workers at all, but insteadtake jobs that natives are unwilling to do. In this view,the rise in immigration in high-income economiessince 1980 has been accompanied by increases in nativeeducational levels; essentially natives moved outof certain kinds of jobs, creating a demand for immigrantlabor.a The researchers do attempt to correct for endogeneity byusing instrumental variables.b Examples include the Mariel boatlift from Cuba (Card1989), the repatriation of Algerians of European origin to France(Hunt 1992), the inflow of workers to Austria after the breakdownof the communist regimes (Winter-Ebmer and Zweimuller1999), and the return of Portuguese from Africa in the 1970s(Carrington and de Lima 1996).c Still, Card (2001) finds no evidence that immigration intoan area leads to offsetting net outflows of workers.d See Gross (1999) for this result for France.43

GLOBAL ECONOMIC PROSPECTS 2006capital enjoys an increase in returns under themodel—with wages declining for all laborcategories, skilled and unskilled, native andforeign-born. With essentially only a laborshock, the scarcity value of capital increases.The negative impact on unskilled native wagesis small, at around 0.3 percent, depending onthe assumed elasticity of substitution betweenmigrant and native workers. 33 The greaterimpact is felt by existing, unskilled migrants,whose wages decline by more than 10 percent.34 At least two factors mitigate thatdecline. First, labor markets are not completelysegmented, so that part of the adjustmentfalls on native workers. Second, othergeneral equilibrium effects are at work, suchas a relative shift in the demand for unskilledworkers as the price of capital (combined withskilled labor) rises and a relative shift in demandtoward goods that use unskilled laborintensively, raising the relative demand for unskilledworkers.The impact of the shock on the wages ofskilled workers is greater than for unskilled.Wages decline by 1.1 percent on average forskilled natives, significantly more than forunskilled natives. Old skilled migrants suffer awage decline of 20 percent, which is doublethat of old unskilled migrants. The impact onskilled workers is larger than for unskilledbecause skilled workers are assumed to be nearcomplements with capital; with capital increasingonly slightly, this would tend to drivedown skilled wages. And the impact is largeston old skilled migrants because the rise inmigration of skilled workers is large relative tothe stock of old skilled migrants, and the newmigrants are assumed to be closer substitutesfor skilled migrants than are native workers.The greater impact of migration on skilledthan unskilled wages is not at first sight consistentwith the limited evidence available. Inthose studies that find any significant impactof migration on native wages, the largest impacttends to be on unskilled wages (seeabove). Our seemingly contrary result arisesfor three reasons. First, unskilled immigrationto high-income countries has been muchlarger than skilled migration, so that the wageimpact of unskilled migration is easier to detectin empirical work. Second, the shockmodeled here represents a one-time increase inskilled migration that is larger than the existingstock, which has built up over time. Andthird, the model assumes little change in thecapital stock, while increased investment in responseto migration would dampen the fall inskilled wages, a point to which we return inthe conclusion to this chapter.The impact in developing countries isnearly the reverse. Capital returns suffer andlabor returns improve, with larger improvementsfor skilled workers than for unskilledworkers. The magnitudes differ because therelative size of the shock differs. For example,the decline in unskilled workers in developingcountries is only 0.3 percent, versus 1.7 percentfor skilled workers.Assuming that all capital income accruesto native households, native households inhigh-income countries are on aggregate betteroff after the shock, with real incomes increasingby 0.4 percent. That is, the increase incapital income more than offsets the loss inwage income. Part of the old migrants’ 6 percentdecline in real income is due to the assumptionthat they own no capital, so enjoyno nonwage income. 35 An alternative, extremeassumption is that on a per capita basis,old migrants receive the same amount of nonwageincome as natives. This alternativewould reduce old migrants’ loss to 3.4 percentof base real income.To summarize, the new migrants areclearly the large winners, particularly in percentageterms. Under the assumptions of themodel, existing migrants are likely to belosers—though the extent of their loss willdepend on their degree of substitutabilitywith native workers and their share ofnonwage income. Native households in bothhigh-income and developing countries arebetter off. The sources of their gains, though,are very different (figure 2.6). In the highincomecountries the gains are generated byhigher returns to capital—somewhat offset by44

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONFigure 2.6 Source of gains for native workersChange in real income in 2025, $ billions350Native workers in high-income regionsChange in real income in 2025, $ billions200Native workers in developing regions3002501502001501001005050050010050150200CapitalLaborPublicRemittancesTotal100CapitalLaborPublicRemittancesTotalSource: World Bank simulations.lower wages. The gains for natives in highincomecountries would be lower if we assumeda more even distribution of capital (towardmigrants) and a greater degree of laborsubstitutability. In developing countries, thegains to natives essentially are generated byhigher wage income and higher remittances—somewhat offset by lower returns to capital.These gains would be lower should thepropensity of the new migrants to remit belower than average.If migrants are viewed the same asnatives, then increased migrationreduces natives’ wagesThe degree of labor-market differentiationplays a critical role in determining the effect ofthe increase in migration on native and foreignhouseholds. An alternative scenario—maintaining the same increase in migration—assumes that employers are perfectlyindifferent to hiring native workers versus foreign-bornworkers. 36 This empirical issue islinked to real-world dynamics since, over thelong run, differences in labor characteristicscould fade as migrants adjust to their newenvironment and as employers cease to seethem as different. 37The impacts of the alternative scenario onfactor returns are shown in figure 2.7. Themost notable impact is that native wages (forskilled and unskilled workers) decline by morewhen natives and migrants are viewed as perfectsubstitutes for each other, while the wagesof the foreign-born decline by significantlyless. 38 For skilled workers, the average declinebecomes negligible; the burden of adjustmentis spread out more evenly between native andforeign-born workers. For the given shock,and depending on the assumed elasticity ofsubstitution between foreign and native workers,the impact on native wages ranges froma slight increase to a decline of 1 percent(box 2.5).Because increasing migration constitutes aclear labor-supply shock, one would expect it45

GLOBAL ECONOMIC PROSPECTS 2006Figure 2.7 Factor returns and migrationin high-income countries, 2005% change relative to baseline50510152025CapitalUnskillednativeworkersStandardSource: World Bank simulations.UnskilledforeignbornworkersPerfectsubstitutionSkillednativeworkersSkilledforeignbornworkersto affect wages (or employment, where movementin wages is constrained). But it is importantto view these changes in dynamic terms.First, assuming differentiation between nativeand foreign-born workers, the impact on nativeworkers’ wages in high-income countriesis slight even in absolute terms (–0.04 percentfor unskilled workers and –0.4 percent forskilled workers). More important, in dynamicterms, these changes alter the rate of growthof wages over the next two decades onlyslightly. In the base case, nominal wages willincrease by 3.6 and 4.7 percent, respectively(average annual growth between 2001 and2025), for unskilled and skilled workers inhigh-income countries. With an increase in migration,the growth rate is unchanged for unskilledworkers and drops to 4.6 percent forskilled workers. Even in the worst-case scenariofor native workers, where foreign-bornworkers are assumed to be perfect substitutesfor natives, the dynamic trends are almost exactlythe same as in the baseline scenario.The effects of the modeled shock are obviouslylarger for existing migrants. With labordifferentiation, their long-term wage growthtrend drops to 3.1 percent for unskilled workersand 3.9 percent for skilled workers fromtheir baseline trend of 3.6 and 4.9 percent—still positive but significantly lower. In themore optimistic scenario for existing migrants,where native workers bear a largerpart of the burden, the trend growth in wagesis virtually identical to the baseline, decliningto 3.4 and 4.8 percent, respectively, for unskilledand skilled workers.The assumption of perfect substitutabilityof native and migrant labor has a small impacton the global income gain ($379 billion insteadof $356 billion), but significant distributionaleffects. First, in high-income countries perfectsubstitution implies a more pronounced procapitalbias, as native labor suffers a largerloss. However, the negative impact on existingmigrant households is much smaller: less than1 percent, compared with the 6 percent sufferedwith labor differentiation. This has apositive impact for developing countries, becausetheir loss in remittances from existingmigrants drops dramatically. Second, with perfectsubstitution the new migrants benefit froma larger wage differential and thus a higher incomegain. And again, for developing countries,this translates into higher remittances.Overall, the change in real net remittances is$129 billion under perfect substitutability, asopposed to $88 billion. The bottom-line is thatthe real income of new migrants increases by250 percent, as opposed to 200 percent, andthe gains for natives in developing countriesrise to 1.2 percent, instead of 0.9 percent.Several other parameters affect the relativeimpact of increased migration on wages andreturns to capital. For example, as shown inbox 2.5, the relative impact will depend on thesubstitution between capital and labor. Thelesser the degree of substitution (or the lessflexible the economy), the greater will be thenegative impact on wages. Most econometricevidence suggests a capital–labor substitutionelasticity of around 1, somewhat higher thanthat used in the model. 39 Another crucial assumptionin the standard model is that skilledworkers and capital are near complements. Analternative would be to assume that unskilledand skilled workers were both substituteswith capital. This would moderate the decline46

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONBox 2.5Increased migration and its impact on wagesThe impact of increased migration on wages canbe summarized by a few simple formulas. Thekey parameters are the share of foreign workers inthe economy, the capital-labor elasticity of substitution,and the substitution between native and foreignworkers. These relations are summarized in the tablebelow. a The values refer to an economy where foreignlabor is 5 percent of the labor force (s l f). Thecapital-labor substitution elasticity ( v ) is 0.9; andthe substitution ( l ) between native and foreignworkers is 4. b The point elasticity is given in thethird column of the table. The estimated impact of a50 percent increase in the stock of foreign workers issimply the point elasticity multiplied by 50. Theactual impact comes from a calibrated numericalmodel. The two are relatively close, despite the sizeof the shock. This is because the results are largelydriven by the low share of foreign workers in theeconomy; therefore most of the economy stays nearits initial equilibrium. The aggregate wage falls byonly 1 percent. This is allocated across the differentworkers depending on their substitutability. With anelasticity of 4 (used in the default scenario), migrantworkers see a 10 percent decline in wages; domesticworkers see only a 0.5 percent decline. If migrantsSummary of elasticities with respect to an increase in foreign workersPoint Estimated ActualExpression Description elasticity impact impactsf l s kW, F Aggregate wage 0.0222 1.1 1.0 vsnl lsf l FW, F s kWage of foreign workers 0.2597 13.0 10.1s l fsf l NW, F s kWage of native workers 0.0097 0.5 0.5 l v vWage impact from a 50 percent increase inforeign workers, under various substitutionassumptions% change in wages505101520253035Substitution elasticity between native and foreign workersSource: World Bank staff calculation.Nativeworkers’wagesForeignworkers’wages1.5 2.5 3 3.5 4 8 16 32 64and native workers are perfect substitutes, thenwages of native workers would decline by 1 percent,that is, by the same amount as the aggregate wage.The figure shows the impact on wages—of bothforeign and native workers—using different assumptionsabout their substitutability. The shock isa 50 percent increase in the stock of foreign workerswith the same assumptions used for the numericalexample in the table. The impacts on wagesconverge only at very high levels of substitutability.In actual econometric work, it might be hard to es-timate the substitution elasticity with great precision,particularly since the shocks are unlikelyto be as considerable as those modeled here.a These relations are for a small single-sector closed economybut line up relatively well with the impacts from theglobal model. Because it is closed and single-sector, the relationsmay not hold exactly because of other general equilibriumeffects. See van der Mensbrugghe 2005b.b The other parameters are the capital share (s k ), the shareof foreign labor in output (s f s l .s f l ), and the share of nativeworkers in the labor force (s n l ).47

GLOBAL ECONOMIC PROSPECTS 2006of skilled wages and the rise in the returns tocapital, essentially because there would bemore (assumed) flexibility in the economy.When capital and skilled workers are complements,a sharp increase in skilled workers,without a concomitant increase in capital,raises the scarcity value of capital. If we makecapital and skilled labor more substitutable,the decline in the wages of skilled workers willcreate more demand for them and dampen thenegative impact on their wages.Caveats—what the modelleaves outWhile the scenarios discussed here providea wealth of insights, they do not addressmany important aspects of the impact ofmigration. Using side calculations, it is possibleto get a sense of two such aspects.The model does not account for changesin migrant characteristics over timeThe model assumes that migrant characteristicsdo not change over time. This is useful inhighlighting the immediate impact of migration,but less realistic over the medium term.As migrants remain in the destination country,they tend to take on the characteristics ofnative workers. For example, they learn orimprove their fluency in the language, theybetter understand (and may tend to adopt)the social mores of the destination country,and they may become more educated. Employersare likely to view a migrant that haslived in the country for 20 years as beingmore similar to a native worker than a migrantwho arrived yesterday. This issue hasseveral implications for the calculation of thegains from migration. Migrants who havespent a longer time in the destination countrywill be less perfect substitutes for new migrants,mitigating the drop in their wagespredicted by the model. Similarly, as thedegree of labor differentiation declines, theimpact on native wages will rise, thus reducingnative households’ gains. On the otherhand, increased productivity as migrants improvetheir education may generate largergains to owners of capital and could benefitnative workers through spillover effects suchas training. Remittances may decline as migrantsbecome more removed from the origincountry.The process of catch-up in productivity isnot captured in our current model, but wehave done some side calculations to see howthe results could be affected. The catch-up rateand workers’ length of stay are two factorsthat affect catch-up. How long does it take theaverage migrant to achieve the level of productivityof native workers? Borjas (2003b)provides mixed evidence on this point formigrants in the United States. First, he showsthat the catch-up rate depends very much onwhen the migrants arrived, with earlier migrantsdoing better than later migrants. Second,he finds no absolute convergence, withmigrants’ wages remaining below those of nativeworkers.The second factor relates to workers’length of stay. The longer workers stay, thebetter placed they are to improve their skillsand adapt to local work practices, includinglanguage skills (if necessary). At one extreme,all migrants may be assumed to be temporaryworkers staying for a short period to returnhome permanently. Or they may be assumedto be permanent workers arriving young andwith high educational attainment or acquiredskills.Under the most optimistic scenario, wherecatch-up occurs within a year, our simple,calibrated model predicts gains in the outputof high-income countries that are about25 percent higher than in the case of no catchup.40 Under a more plausible scenario, wherethe process of catch-up takes 10 years andannual attrition 41 is around 10 percent, theoutput gain in high-income countries is about12 percent higher than with no catch-up. Itappears, therefore, that the catch-up phenomenoncould boost the gains from migrationsubstantially.48

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONThe model does not account for thepotential of migration to spur higherinvestmentThe model generates only a modest rise in thecapital stock as a result of the increase in migration.Increased returns to capital, and thusinterest rates, do increase savings. However,this effect is marginal, in keeping with empiricalestimates of the responsiveness of savingsto changes in interest rates. In reality, higherinvestment in response to the higher returns tocapital may be financed by capital inflows,which do not change in the model. There issome evidence, however, that large immigrationcan attract capital flows. For example,Davis and Weinstein (2002) argue that skilledlabor, unskilled labor, and capital are all attractedto the United States, owing to U.S.technological superiority. The mass migrationfrom Europe to the new world before WorldWar I encouraged large inflows of capital.Higher investment would lessen the declinein wages suffered by skilled workers, dampenthe rise in the return to capital, and increasethe demand for unskilled workers in the highincomecountries—but it could have the oppositeeffects in developing countries. To verifythis intuition, we simulated the same migrationshock and added to the shock a 0.4 percentdecline in the level of investment indeveloping countries, with a concomitanttransfer of these resources to high-incomecountries. 42 Those assumptions indeed have apositive impact for the high-income countriesand dampen the capital-income gains andlabor-income losses. Overall, the gain fornative households in high-income countriesimproves by 4.5 percent (from $139 billion to$145 billion). But it comes at the expense ofnatives in developing countries, whose incomegain drops from $143 billion to $125 billion,a drop of 12.5 percent. The global gains fall to$345 billion, a 3 percent fall. This suggests—in the absence of an increase in savings—thatthe potential reallocation of global savingstoward high-income countries is negative atthe global level and that capital is moreproductive in developing countries.Other factors not covered by the modelare more difficult to quantifyThere are various costs associated with migrationthat the model does not take intoaccount. One issue concerns adjustment costs,as changes in the technology of productionand in the mix of goods imply transitional unemploymentand changes in the pattern ofinvestment. The magnitude of these costs dependsin part on the structure of labor-marketinstitutions (such as constraints on hiring andfiring and minimum-wage legislation) and onthe efficiency of capital markets. Countrieswith more flexible labor markets and soundbanking, stock, and bond markets are likely toexperience lower adjustment costs, underliningthe importance of the investment climatefor realizing the potential gains from migration.The size of adjustment costs will alsodepend on whether migration is concentratedor spread over time. This point also has policyimplications. If a country anticipates needingmigrants in the future or recognizes that migrationpressures are bound to rise due to demographicchanges, it would be better toloosen constraints on migration earlier andmore gradually than to be confronted with asharp rise later on. Migration also involvesdirect costs, including transportation andtransitional expenses, as well as the noneconomiccosts suffered by migrants separatedfrom their families (for example, the impacton children raised without one or both parents—seechapter 3). In general these are eithershort-term costs that should not greatlychange the calculation of benefits from permanentmigration or problems that declineover time as migrants and families adjust topermanent changes or take steps to reunite.Several other issues that may affect thegains from migration are impossible to quantify.First, our model does not distinguish betweenirregular and regular migrants. If migrantsare irregular, they may be paid lowerwages (see chapter 3), which would reducetheir welfare gains (and remittances) relativeto the model results, while it increases thegains of natives. However, irregular migrants49

GLOBAL ECONOMIC PROSPECTS 2006also may impose costs on destinationcountries—among them the costs of enforcement(as governments seek to limit what somemay view as undesirable changes in the country’sculture and demographic characteristics);a possible burden on public spending, whichmay be higher for irregular migrants (see box2.3); and the potential for other forms ofillegality generated by the presence of a large,undocumented group of foreign workers.Second, immigrants may improve the efficiencyof employment from the perspective offirms by providing a source of labor that caneasily be employed in new geographic locations,and hired or fired in response to changesin cyclical conditions. Piore (1986) describeshow many migrants (at least initially) tend toview their stay as temporary, filling jobs withlower salaries and less stability than those ofnatives. 43 Large numbers of immigrants workin construction, which facilitates new developmentsin areas that require a mobile laborforce. However, over time, migrants will becomemore permanent and demand jobs similarto those held by natives.Third, our model does not reflect the socialor economic implications of increasing diversityin the destination country. The social impactlies outside our present scope, but diversityhas potential economic costs and benefitsthat should be considered. Some writers arguethat increased diversity has an economicvalue. Glasser, Kolko, and Saiz (2001) emphasizethe role of a rich variety of services andconsumer goods in enhancing the attractivenessof cities. Florida (2002) relates an indexof diversity to a concentration on high-technologyindustries. Ottaviano and Peri (2004)find that cultural diversity has a net positiveeffect on the productivity of U.S.-born citizens.By contrast, Schiff (1998) uses a theoreticalmodel to underline how a society’s sharedvalues can reduce the cost of transacting business,owing to higher trust and easier enforceabilityof sanctions. Thus immigration, whichincreases diversity, may lower productivity byraising transaction costs. Finally, Alesina andFerrara (2004) find that increases in ethnicdiversity are associated with lower growthrates, holding all else equal. However, diversitymay be more beneficial to growth at higher incomelevels. 44 Clearly much will depend onthe kinds of diversity involved: immigrantswho rely on national affinities to cement loyaltyto violent gangs presumably have a verydifferent impact on growth and welfare thanimmigrants who open ethnic restaurants.Fourth, the model may not fully capturethe beneficial effect of immigration on increasingthe supply of labor in the service sector.Although reductions in the prices of servicesare captured, the resulting expansion inthe supply of native labor (as more parentscan afford child care and workers have moretime to devote to their jobs) is not.Fifth, the model does not reflect the possibilitythat skilled migration may lowergrowth in origin countries, for example, becauseof positive externalities from the presenceof skilled workers or increases in theprice of services that require technical skills(see chapter 3).Finally, the model assumes constant returnsto scale, while immigration may bemore beneficial if significant sectors enjoyincreasing returns to scale. Increasing returnsmay be derived, for example, from fixed productioncosts, network effects (the unit priceof providing telephone service falls as the customerbase grows), reduced transport andcommunications costs (as the local market expands),or increased productivity due to interactionsamong highly skilled workers. In theirrole as consumers and workers, immigrantsmay facilitate an expansion of the market,thereby raising productivity by increasing returns.On the other hand, large inflows of immigrantsmay induce congestion, strainingpublic transportation systems, for example, orbidding up the price of land. Such effects areparticular to the sector and geographic areainvolved, so it is difficult to draw broad conclusions.However, skilled immigrants havemade significant contributions to high-technologysectors that are subject to increasingreturns to scale.50

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONThese qualifications to the scenario resultsillustrate how model exercises must abstractfrom reality to provide quantitative measuresof the impact of migration. Some of the issuesthat the model does not consider wouldlikely be small in the medium term (adjustmentcosts, transportation costs) or wouldtend to increase the economic benefits of migration(improved productivity of migrantsover time, greater labor-market flexibility andsupply of labor). Other issues would increasebenefits to destination countries, while potentiallyharming origin countries (higher investment,economies of scale). Still others mayhave both economic and social effects, withoutlending themselves to determinations oftheir direction and size (diversity, irregularmigration).Notes1. In keeping with the overall thrust of this report,we focus here on South-North migration, although it isimportant to recognize that a large portion of migrantsfrom developing countries move to other developingcountries.2. Some readers may also find the chapter too technical,as it necessarily deals with detailed specificationissues—for example, the degree of differentiation betweennative and migrant workers, the fiscal impact ofmigrants, and how to take into account the change inprices between developed and developing countrieswhen evaluating gains to migrants.3. Data on the stocks of migrants are generallytaken from census reports in countries of destination andthus include both regular and irregular migrants. However,irregular migrants tend to be less likely to reporttheir immigrant status, so the estimate of total migrantsis probably low.4. The breakup of the Soviet Union and emergenceof 15 new independent countries in 1991 created newpopulations of “international” migrants without migrationhaving taken place.5. The exceptions were Belgium, France, Ireland,Portugal, and the United Kingdom.6. In their regression equation explaining immigrationto the United States, Hatton and Williamson (2002)calculate that IRCA doubled the Mexican immigrationrate from 1989 to 1991.7. Germany, Ireland, and the Czech Republic are inthe process of establishing new immigration regimes,with a major focus on economic migration. The EU isalso discussing the Green Paper on an EU Approach toManaging Economic Migration (EU 2005).8. These numbers will be moderated to the extentthat labor-force participation rates in the 65 cohortare positive, if small. Moreover, labor-force participationrates for the elderly are likely to increase as pensionsand benefits stagnate or decline with fiscal pressuresand as life expectancy rates continue to increase.We may also witness an increase in labor-force participationrates among people of working age.9. These global gains are comparable to the recentfindings by Walmsley and Winters (2003), when adjustedfor the size of the economy in 2001 relative tothe projected size of 2025.10. The model’s specification is described in vander Mensbrugghe (2005a).11. “Migrants” refers to migrants from developingcountries unless otherwise stated.12. The phase-in period is somewhat arbitrary.Because of its 10-year implementation, it minimizesadjustment costs to some extent. The five-year periodbetween 2020 and 2025 enables an assessment of longrunsteady-state impacts.13. This is by design. An alternative would be toincrease the stock of migrants in proportion to theircurrent structure—by host region and skill level. In thiscase, the largest proportional increase would be for unskilledworkers in the United States.14. A switch of 14 million workers from developingto high-income countries has only a small impact(a decline of 0.4 percent) on aggregate employment indeveloping countries, albeit with potentially greaterconsequences among the relatively more scarce skilledworkers.15. Many factors determine the level of remittances.For example, new migrants may leave many dependentsin their home countries, which would tend to raise remittances.On the other hand, at least in the short-term,moving and start-up costs could lower remittances.16. For simplicity, migrants from other highincomecountries are added to the true natives.17. Again, for simplicity, all migrants in developingcountries—both from rich and developing countries—are lumped together for the purposes of the aggregateanalysis.18. The impact on households other than the“new” migrants is not affected by cost-of-living adjustments.Since these households do not move, they face thesame system of prices, and thus their change in realincome simply depends on the standard real incomemeasure.19. This assumes a perhaps implausible Leontiefutility function but the purpose is simply to illustratethe point that corrections need to be made for differencesin the cost of living.51

GLOBAL ECONOMIC PROSPECTS 200620. Were the true prices available, one could do astandard equivalent variation calculation that wouldtake into consideration the change in prices.21. In the high-income countries, new migrants’total real consumption is $562 billion, compared with$80 billion in the baseline, hence the real increase of$481 billion. When the $562 billion is adjusted for thedifference in the cost of living, real consumption, asperceived from the point of view of the new migrant, isonly $244 billion, taking the change in real incomedown to $162 billion. The cost-of-living adjustmentaverages 2.3, lower than the 3.1 GDP-weighted averagePPP of developing countries. This occurs becausemiddle-income countries (with a relatively low PPP adjustment)have a higher weight in migration than in developingcountries’ GDP.22. High-income countries, on the other hand, seea substantial rise in exports, $211 billion (2.2 percent),and a more modest $113 billion rise in net imports(1.2 percent), with imports from developing countriesdeclining by $23 billion. This implies that although alarge part of the increase in high-income exports can beattributed to the increase in remittances, a significantportion is also coming through intraregional tradeamong high-income countries driven in part by changingcomparative advantage.23. A standard “Dutch disease” effect of foreigninflows.24. Studies of the impact of trade reform indeveloping and industrial countries tend to show thatwages in developing countries rise relatively more—particularly for unskilled workers—than in industrialcountries, but those changes are relatively minor comparedto the initial gap in wages. For example, unreportedresults from Anderson, Martin, and van derMensbrugghe (2005) show that full merchandise tradereform would increase unskilled real wages in developingcountries by 3.7 percent (unweighted average), butby only 0.7 percent in industrial countries. This couldinduce a small reduction in the incentive to migrate,but it would not substantially alter the significant wagemultiple of 4 to 5 (taking into account cost-of-livingdifferentials).25. The cost-of-living adjustment for the new migrantstreats their consumption of public goods andservices the same as their private consumption—that is,it is adjusted by the same PPP factor.26. The assumption is that the new migrant householdscome with the dependency ratio of their homecountry.27. There will be compositional impacts in translatinggains from 2025 to 2001, since developing countriesare growing on average more rapidly than highincomecountries.28. Results presented at the eighth annual conferenceon Global Economic Analysis held in Lübeck,Germany. See http://www.gtap.agecon.purdue.edu/events/conferences/2005/program_day3.htm.29. The full merchandise trade reform scenario iswith a standard model and ignores any beneficial impactthrough higher trade-induced productivity orscale economy effects. Note that the gains from reformof services trade could be multiples of merchandisetrade reform. See Anderson, Martin, and van derMensbrugghe (2005).30. In the case of high-income countries, ‘similar’workers would be migrant workers from other highincomecountries. Other migrant workers are bundledtogether in a so-called ‘foreign-born’ aggregate.31. One would expect the elasticity to increase asthe proportion of migrants in the population increases(box 2.5).32. See, for example, Bchir and others (2002).33. Box 2.5 shows how wages—native andforeign—are related to an increase in the stock of migrants.Two parameters are crucial—the substitutionbetween native and foreign workers and the share offoreign workers in the labor force.34. The general equilibrium elasticity is only 0.27for unskilled workers; for roughly a 40 percent increasein supply, wages decline by around 10 percent.35. Migrants from other high-income countries(also assumed to have no nonwage income) see only asmall change in their real incomes, as their wages areclosely linked to the wages of native workers.36. Observed wage differentials can arise from acombination of two effects—differences in productivityand differentiated labor demand. If labor isperfectly substitutable, then the equilibrating conditionis the equality of efficiency wages, that is,productivity-adjusted nominal wages. If labor is differentiated,efficiency wages are no longer necessarilyequalized, and the equilibrium wage will be determinedby supply and demand conditions for the differentiatedlabor.37. The empirical evidence on “catch-up” is limited.In the case of migrants to the United States, Borjas(2003b) shows that migrants who arrived in the1960s almost caught up with natives within a 10–15year period. Those who arrived in the 1970s made lessprogress in closing the gap with natives. However, thewages of migrants arriving in the 1980s actually fellfurther behind those of natives after a 10-year period.The scenarios described in this chapter assume nochange in the relative productivity of migrants. Such anassumption would require a more elaborate specificationof migrants to capture their changing compositionover time, similar to modeling capital vintages. By52

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONignoring the catch-up process, our results may underestimatethe longer-term gains from migration.38. The changes in wages by region of origin areidentical for all workers in each high-income region,but due to aggregation effects, this will not necessarilybe true when averaging across regions.39. The model has a vintage structure with a lowersubstitution elasticity for “old” or installed capital anda higher substitution elasticity for “new” capital. Theactual substitution will be a weighted average of theold and the new vintages, with a higher average forcountries with relatively high rates of investment.40. This follows directly from the assumption thatthe productivity level of migrants is initially 75 percentthat of natives.41. The attrition rate will be a combination offactors—return migration, retirement, and death. Thefirst factor is probably most important the first year,whereas the other two factors will depend on the ageof the migrant.42. The value of 0.4 percent was chosen because itcorresponds to the change in the number of workers indeveloping countries—though it should be noted thatthe change in workers represents a change in the stocklevel, whereas the change in investment is a change inflows.43. He notes that this trend may be changing, astechnology and globalization encourages smaller-scaleproduction and more permanent immigration.44. This may occur because “the productivity benefitsof skill complementarities are realized only whenthe production process is sufficiently diversified,” orbecause high-income economies are able to develop institutionsthat help them cope better with the potentialfor conflict inherent in ethnic diversity (Alesina andFerrara 2004).ReferencesAlesina, Alberto, and Eliana La Ferrara. 2004. “EthnicDiversity and Economic Performance.” CentroStudi Luca D’Agliano Development StudiesWorking Papers 193. Milan and Turin. December.Anderson, Kym, Will Martin, and Dominique van derMensbrugghe. 2005. “Market and Welfare Implicationsof Doha Reform Scenarios.” In AgriculturalTrade Reform and the Doha DevelopmentAgenda, ed. Kym Anderson and Will Martin.New York: Palgrave Macmillan.Angrist, Joshua, and Adriana Kugler. 2002. “Protectiveor Counter-Productive? European Labour MarketInstitutions and the Effect of Immigrants on EUNatives.” Centre for Economic Policy ResearchDiscussion Paper 3196. London.Auerbach, Alan J., and Philip Oreopoulos. 1999. “Analyzingthe Fiscal Impact of U.S. Immigration.”American Economic Review Papers and Proceedings89(2): 176–80.Bchir, Mohamed Hedi, Yvan Decreux, Jean-LouisGuérin, and Sébastien Jean. 2002. “MIRAGE, aComputable General Equilibrium Model forTrade Policy Analysis.” CEPII Working Paper2002-17. Paris. December.Bonin, Holger, Bernd Raffelhuschen, and Jan Walliser.2000. “Can Immigration Alleviate the DemographicBurden?” Finanz Archiv 57(1): 1–21.Borjas, George J. 1994. “The Economics of Immigration”,Journal of Economic Literature 32December (1994): 1667–1717.———. 2003a. “The Labor Demand Curve IS DownwardSloping: Reexamining the Impact of Immigrationon the Labor Market.” Quarterly Journalof Economics. November.———. 2003b. “The Economic Integration ofImmigrants in the United States: Lessons for Policy.”WIDER Discussion Paper DP2003/78.World Institute for Development EconomicsResearch, United Nations University, Helskinki.December.Borjas, George, Richard B. Freeman, and LawrenceKatz. 1997. “How Much Do Immigration andTrade Affect Labor Market Outcomes?” BrookingsPapers on Economic Activity 1: 1–90.Butcher, Kristin F. and David Card. 1991. “Immigrationand Wages—Evidence from the 1980s.”American Economic Review 81(2): 292–6.Card, David. 1989. “The Impact of the Mariel BoatLift on the Miami Labor Market.” NationalBureau of Economic Research Working Paper3069. Cambridge, MA.———. 2001. “Immigrant Inflows, Native Outflows,and the Local Labor Market Impacts of HigherImmigration.” Journal of Labor Economics 19:22–64.Carrington, W., and P. de Lima. 1996. “The Impact of1970s Repatriates from Africa on the PortugueseLabor Market.” Industrial and Labor RelationsReview 49(2): 330–47.Collado, M. Dolores, Inigo Iturbe-Ormaetxe, andGuadalupe Valera. 2004. “Quantifying the Impactof Immigration on the Spanish WelfareState.” International Tax and Public Finance 11:335–53.Coppel, Jonathan, Jean-Christophe Dumont, andIgnazio Visco. 2001. “Trends in Immigration andEconomic Consequences.” Economics DepartmentWorking Paper 284. Organisation for EconomicCo-operation and Development, Paris.53

GLOBAL ECONOMIC PROSPECTS 2006Davis, Donald R., and David E. Weinstein. 2002.“Technological Superiority and the Losses fromMigration.” National Bureau of EconomicResearch Working Paper 8971. Cambridge,MA.DeNew, John P., and Klaus F. Zimmermann. 1994.Native Wage Impacts of Foreign Labor. Journal ofPopulation Economics 7: 177–92.Dustmann, Christian, and Albrecht Glitz. 2005. “Immigration,Jobs and Wages: Theory, Evidence andOpinion.” Centre for Research and Analysis ofMigration, Department of Economics, UniversityCollege London.Dustmann, Christian, Francesca Fabbri, Ian Preston,and Jonathan Wadsworth. 2003. “The LocalLabour Market Effects of Immigration in theUK.” Home Office Online Report 06/03.London.EU (European Union). 2005. “Green Paper on an EUApproach to Managing Economic Migration.”COM(2004)811 final, Commission of the EuropeanCommunities, Brussels. November.Fehr, Hans, Sabine Jokisch, and Laurence Kotlikoff.2004. “The Role of Immigration in Dealing withthe Developed World’s Demographic Transition.”National Bureau of Economic Research WorkingPaper 10512. Cambridge, MA.Florida, Richard. 2002. “Bohemia and Economic Geography.”Journal of Economic Geography 2:55–71.Glasser, Edward L., Jed Kolko, and Albert Saiz. 2001.“Consumer City.” Journal of Economic Geography1: 27–50.Gott, Ceri, and Karl Johnston. 2002. “The MigrantPopulation in the UK: Fiscal Effects.” RDS OccasionalPaper 77. Home Office, London.Gross, Dominique. 1999. “Three Million Foreigners,Three Million Unemployed: Immigration andthe French Labor Market.” IMF Staff WorkingPaper 99/124. International Monetary Fund,Washington, DC.Gustafsson, B., and T. Osterberg. 2001. “Immigrantsand the Public Sector Budget: Accounting Exercisesfor Sweden.” Journal of Population Economics14(4): 689–708.Gustmann, Alan L., and Thomas L. Steinmeier. 1998.“Social Security Benefits of Immigrants and U.S.Born.” National Bureau of Economic ResearchWorking Paper 6478. Cambridge, MA.Hamilton, Bob, and John Whalley. 1984. “Efficiencyand Distributional Implications of Global Restrictionson Labour Mobility.” Journal of DevelopmentEconomics 14: 61–75.Hunt, Jennifer. 1992. “The Impact of the 1962 Repatriatesfrom Algeria on the French LaborMarket.” Industrial and Labor Relations Review45: 556–72. April.IOM (International Organization for Migration).2005. World Migration 2005. Geneva.Jaeger, David A. 1996. “Local Labor Markets, AdmissionCategories, and Immigrant LocationChoice.” Unpublished paper. Hunter College,New York. June.LaLonde, Robert J., and Robert H. Topel. 1997. “EconomicImpact of International Migration and theEconomic Performance of Migrants.” In Handbookof Population and Family Economics,ed. Mark Rosenzweig and Oded Stark. Amsterdam:North-Holland.Lee, Ronald, and Timothy Miller. 2000. “Immigration,Social Security, and Broader Fiscal Impacts.”American Economic Review Papers and Proceedings90(2): 350–54.Longhi, Simonetta, Peter Nijkamp, and Jacques Poot.2004. “A Meta-Analytic Assessment of the Effectof Immigration on Wages.” Population StudiesCentre Discussion Papers 47. University ofWaikato, Hamilton, New Zealand. December.Massey, Douglas. 2000. “Higher Education and SocialMobility in the United States, 1940–1998.” Paperdelivered at Association of American UniversitiesCentennial Meeting, April 17, Washington, DC.Mora, Jorge, and J. Edward Taylor. 2005. “Determinantsof Migration, Destination and SectorChoice: Disentangling Individual, Household andCommunity Effects.” In International Migration,Remittances, and Development, ed. CaglarOzden and Maurice Schiff, Washington, DC:World Bank.Moses, Jonathon W., and Bjørn Letnes. 2004. “TheEconomic Costs to International Labor Restrictions:Revisiting the Empirical Discussion.”World Development 32(10): 1609–26.Nana, Ganesh, and Julian Williams. 1999. “Fiscal Impactsof Migrants to New Zealand.” Report tothe New Zealand Immigration Service, Auckland.Ottaviano, Gianmarco, and Giovanni Peri. 2004. “TheEconomic Value of Cultural Diversity.” NationalBureau of Economic Research Working Paper10904. Cambridge, MA.OECD (Organisation for Economic Co-operation andDevelopment). 1997. Trends in InternationalMigration. Paris.———. 2005. Trends in International Migration. Paris.Parsons, Christopher R., Ronald Skeldon, Terrie L.Walmsley, and L. Alan Winters. 2005. “Quantifyingthe International Bilateral Movements ofMigrants.” Paper presented at the Eighth AnnualConference on Global Economic Analysis, June9–11, Lübeck, Germany.54

THE POTENTIAL GAINS FROM INTERNATIONAL MIGRATIONPassel, Jeffrey S. 2005. “Estimates of the Size andCharacteristics of the Undocumented Population.”Pew Hispanic Center, Washington, DC.Piore, Michael J. 1986. “Perspectives on Labor MarketFlexibility.” Industrial Relations 25: 146–66.Spring.Pischke, Jorn-Steffen, and Johannes Velling. 1994.“Wage and Employment Effects of Immigrationto Germany: An Analysis Based on Local LaborMarkets.” Centre for Economic Policy ResearchDiscussion Paper 935. London. March.Poot, Jacques, and Bill Cochrane. 2004. “Measuringthe Economic Impact of Immigration: A ScopingPaper.” Immigration Research Programme,New Zealand Immigration Service, Auckland.December.Rowthorn, Robert. 2004. “The Economic Impact ofImmigration.” Civitas Online Report. Civitas:The Institute for the Study of Civil Society,London.Schiff, Maurice. 1998. “Trade, Migration and Welfare:The Impact of Social Capital.” In H. Singer, N.Hatti, and R. Tandon (eds.) Globalization, Technology,and Trade in the 21st Century. Vol. 19,New World Order, Delhi: B. R. Publishing.Smith, James P., and Barry Edmonston. 1997. TheNew Americans: Economic, Demographic andFiscal Effects of Immigration. Washington, DC:National Academy Press.Sriskandarajah, Dhananjayan, Laurence Cooley, andHoward Reed. 2005. “Paying Their Way: TheFiscal Contribution of Immigrants in the UK.”Institute for Public Policy Research, London.Storesletten, Kjetil. 2000. “Sustaining Fiscal Policythrough Immigration.” Journal of Political Economy108(21).Timmer, Hans, and Dominique van der Mensbrugghe.2005. “Migration, PPP and the Money Metric ofWelfare Gains.” Development Economics Department,World Bank, Washington, DC.U.S. Binational Study on Migration. 1997. BinationalStudy: Migration Between Mexico and the UnitedStates. www.utexas.edu/lbj/uscir/binational.html.U.S. Labor Survey. 2005. Bureau of Labor Statistics.http://bls.gov/cps/home.htm.van der Mensbrugghe, Dominique. 2005a. “LINKAGETechnical Reference Document: Version 6.0.” Unpublishedpaper. World Bank, Washington, DC.———. 2005b. “Derivation of Output and Wage ElasticitiesRelative to an Increase in Migrants.”Development Economics Department, WorldBank, Washington, DC.Walmsley, Terrie L., S. Amer Ahmed, and ChristopherR. Parsons. 2005. “The GMig2 Data Base: AData Base of Bilateral Labor Migration, Wagesand Remittances.” GTAP Research MemorandumNo. 6. Center for Global Trade Analysis,Purdue University. September.Walmsley, Terrie Louise, and L. Alan Winters. 2003.“Relaxing the Restrictions on the TemporaryMovements of Natural Persons: A SimulationAnalysis.” Centre for Economic Policy ResearchDiscussion Paper Series 3719. London. January.Wickramasekera, Piyasiri. 2002. “Asian Labor Migration:Issues and Challenges in an Era of Globalization.”International Labor Office, InternationalMigration Program. Geneva.Winter-Ebmer, Rudolf, and Josef Zweimuller. 1999.“Do Immigrants Displace Young Native Workers:The Austrian Experience.” Journal of PopulationEconomics 12: 327–40.55

3The Policy Challengesof Migration: The OriginCountries’ PerspectiveIn evaluating the impact of remittances, themain subject of this report, it is important totake into account the implications of theinitial decision to emigrate. This chapter willanalyze the implications for migrants and origincountries of migration for economic gainfrom developing to high-income countries. 1Focusing on this form of migration canhighlight some key policy dilemmas that governmentsface in improving the developmentalimpact of migration.Migration is an extremely diverse phenomenon.Its economic impact on each origincountry, and the impact of policy, will dependon many circumstances—among them theskills and former employment of migrants, thehistory of migration (the existence and locationof a large diaspora), the sectors affected,patterns of trade and production, the investmentclimate, and the size and geographicallocation of the country. For example, migrationpolicies appropriate for a large developingcountry with substantial low-skilled emigrationand effective institutions will differfrom the policies for a small island economywith substantial high-skilled emigration andweak institutions.Migration is as complex as it is diverse, sopredicting the impact of policy changes will beproblematic until more research is done andbetter data obtained. In particular, the genderimplications of migration are poorly understoodand require more research. Migrationalso has important social and political implications,that may be as important as the economicanalysis provided here. For all of thesereasons, the analysis and policy recommendationsin this chapter must remain heavily qualified.Our purpose is to signal to policymakersin developing countries, and to the developmentcommunity in general, the elements thatshould be considered in formulating migrationpolicy.International migration often generates greatbenefits for migrants and their families,although at some risk. Migration can greatlyincrease incomes of both migrants and theirfamilies and has helped countless householdsescape poverty. While most workers gain greatlyfrom migration, the decision to migrate issometimes made with inadequate informationand at high risk and cost, particularly if themigration is irregular. By providing informationon migration opportunities and risks,governments could help avoid unfortunate,costly migration decisions. Governments shouldalso consider means to prevent and prosecutetrafficking and other abuse of migrants, and tostrengthen migration-related partnershipsbetween origin and destination countries.Increasing the emigration of low-skilledworkers would significantly reduce poverty indeveloping countries. In addition to enablingemigrants to escape poverty and to reducingpoverty in the country of origin throughremittances (discussed in chapter 5),low-skilled emigration can increase wages and57

GLOBAL ECONOMIC PROSPECTS 2006reduce unemployment and underemploymentof poor workers in the country of origin.Many of the poorest lack the financialresources or the skills required for successfulemigration to high-income countries, butavailable data indicate that a significantnumber of them do emigrate, although atlower rates than the non-poor. Reducing therestrictions on low-skill emigration, whileremaining sensitive to concerns in destinationcountries over social tensions, jobopportunities for low-skilled natives and thepotential burden on public expenditures, maybest be achieved through managed migrationprograms designed jointly by origin anddestination countries. Such programsshould provide for temporary, low-skilledmigration—with incentives for return.Emigration of high-skilled workers mayreduce living standards of those left behind andimpair growth, but can also be beneficial. Likelow-skilled migration, high-skilled migrationcan greatly benefit migrants and their familiesand can help relieve labor market pressures.In addition, a well-educated diaspora canimprove access to capital, technology,information, foreign exchange, and businesscontacts for firms in the country of origin. Atthe same time, high-skilled emigration mayreduce growth in the origin country because(a) other workers lose the opportunity fortraining and mutually beneficial exchanges ofideas; (b) opportunities to achieve economiesof scale in skill-intensive activities may bereduced; (c) society loses its return on highskilledworkers trained at public expense; (d)the price of technical services (where thepotential for substitution of low-skill workersis limited) may rise. Highly educated citizensmay also help to improve governance, improvethe quality of debate on public issues,encourage the education of children, andstrengthen the administrative capacity of thestate—all of which may be reduced throughemigration of the highly skilled.Because of the lack of data and the myriadof individual country circumstances that caninfluence the impact, it is impossible to reliablyestimate the net benefit, or cost, to origincountries of high-skilled emigration. We canonly offer some rough observations that reflectthe wide variation in high-skilled emigrationrates among countries:• Very high rates of high-skilled emigrationaffect a small share of developingcountries’ population, and many countrieswith high rates of high-skilled emigrationhave poor investment climatesthat likely limit the productive employmentof high-skilled workers. On theother hand, the lack of high-skilledworkers may contribute to the poorinvestment climate and limit the supplyresponse to economic reform.• Some countries also find it difficult toproductively employ all high-skilledworkers because of small economic scaleor misguided educational policies that resultin a large supply of college graduatesfor whom no suitable jobs exist.• High-skilled emigration has had a severeimpact on public services with positiveexternalities. The loss of skills throughhigh-skilled emigration has particularlyimpaired health services in several developingcountries.Origin countries harmed by high-skilledemigration face difficulties in managing theproblem. Service requirements for access topublicly financed education can be evaded andare likely to discourage return, and proposalsfor the taxation of emigrants to the benefit ofthe origin country have made little progress.Improved working conditions in public employmentand investments in the infrastructurefor research and development may beeffective in retaining key workers. Workingconditions can also be improved by strengtheninggovernance, which may require politicalwill rather than money. Origin countries canalso encourage educated emigrants to returnby identifying job opportunities, cooperatingwith destination countries that have programsto promote return, permitting dual nationality,58

THE POLICY CHALLENGES OF MIGRATIONand helping to facilitate the portability of socialinsurance benefits.The migration decision and itsimpact on migrants and theirfamiliesMaking the costly and sometimes riskydecision to move to another countrygenerally involves the expectation of largeincreases (or lower variability) in income,described by economists as the net presentvalue of lifetime earnings. 2 The migrant’s expectedincome gain from emigration also reflectshis or her employment prospects at homeand the likelihood of employment overseas.Better economic prospects drive migrationMigrants from developing to high-incomecountries generally enjoy large increases inearnings. 3 A dataset compiled by the InternationalLabour Organization (ILO) shows thatworkers in high-income countries earn a medianwage that is almost five times the levelof that of workers in low-income countries,Figure 3.1 Median wage levels for workersin the same occupation, relative to highincomeeconomies, 1988–92 aPercent454035302520151050Upper-middleincomecountriesLower-middleincomecountriesSource: Freeman and Oostendorp (2000).Low-incomecountriesNote: Chart reports the median wage in each country/skillgroup relative to the highest wage for that skill group, withthe ratio in high-income economies as the numeraire. Thusthe median wage in low-income countries (averaged acrossskill groups) is 20 percent of the median wage in high-incomeeconomies.a. Adjusted for purchasing power parity.adjusted for differences in purchasing power(Freeman and Oostendorp 2000) (figure 3.1).These data may overstate the wages thatmigrants expect, because their earnings, at leastinitially, tend to be lower than those of natives(Lucas 2004a). Moreover, many poor workerswho lack local language skills and have minimaleducation may find limited employmentprospects in high-income country job markets.On the other hand, these data may understatethe benefit of migration from the perspectiveof the household. In measuring differences inwelfare between migrants and those who donot migrate, migrants’ earnings in highincomecountries are reduced to reflect thehigher cost of living in high-income countries—or purchasing power parity (PPP). To the extentthat migrants send earnings back home inthe form of remittances, however, this adjustmentis not relevant, so household gains mayexceed the PPP-adjusted rise in earnings. 4 Furthermore,the data on income differences mayinfluence expectations of future earnings formigrants and their children, and would undoubtedlygenerate much larger migration, inthe absence of controls. Evidence of substantialmigration pressure includes long queues ofapplicants for immigration to high-incomecountries, the rise in irregular immigration,the increase in asylum seekers (Hatton andWilliamson 2002), and the high fees paid tosmugglers who help migrants cross bordersillegally (Cornelius 2001).The expectation of higher earnings is notthe only economic incentive for migration.Households may decide to send some membersabroad to diversify the family’s source ofincome and thus reduce risk, as shocks affectingthe level of wages and the probability ofemployment in the destination country maynot be correlated with the shocks affectingdomestic workers (Daveri and Faini 1999). 5Migration involves considerable costsDespite clear gains for many, migration involvescosts and risks that, together with restrictionson migration, help explain why most peopleprefer to stay at home. Migration can entail59

GLOBAL ECONOMIC PROSPECTS 2006substantial up-front costs—transportation, feescharged by recruitment agencies, fees to obtaina visa and work permit, maintenance whilesearching for work, forgone earnings (if themigrant was or could be employed at home),the reduction in value of location-specific skills(for example, knowing one’s native language),and the pain of being separated from family andfamiliar surroundings. Obviously these costswill vary enormously among migrants.Lack of data makes it difficult to directlytest the relationship between costs and the decisionto migrate. However, distance can beused as a proxy for costs, representing notonly transport costs, but also migrants’ limitedfamiliarity with countries of destination.Adams and Page (2003) find that distance isa significant determinant of the directionof migration from developing countries tothe United States, European members of theOrganisation for Economic Co-operation andDevelopment (OECD), and the Arab Gulf.Long, Tucker, and Urton (1988) found thatdistance was an important constraint on internalmigration in the United States and theUnited Kingdom; what holds for internal migration(without political barriers or languagedifferences, and with limited cultural differences)probably holds even more strongly forinternational migration. Household surveysalso provide indirect evidence that distancehas an important impact on migration: whilesome of the poorest do migrate internationally,they are more likely to remain at home orto migrate internally (see below).Among the largest quantifiable costs tomigrants are fees paid to private recruitmentagencies, whose role in the international labormarket has increased substantially, in part becauseof the rise of temporary labor migrationprograms. 6 A major conduit of job opportunitiesfor migrant workers, private recruitmentagencies often are instrumental in seeking outnew markets for job opportunities abroad.They also provide services such as languagetraining and assistance with settlement (seeXiang 2003 in the case of China). However,they also can be a source of abuse (ILOTable 3.1 Fees charged by recruitmentagenciesCountry US$ YearMigration to the Middle EastBangladesh 1,727 1995India 900 1995Pakistan 768 1995Sri Lanka 689 1995Sri Lanka 893 1986Migration to JapanThailand 8,000 1996Migration to MalaysiaThailand 666–1,000 1991Sources: Abella 2004; Lucas 2004a; Eelens and Speckmann1990; Spaan 1994.2003b). Because many migrants lack informationon foreign job markets, and some agenciesmay have considerable market power,recruitment agencies have captured a substantialshare of the rents generated by limits onimmigration (Lucas 2004a). Available data(mostly from past years) indicate that recruitmentcharges range from $689 (in 1995) forSri Lankan immigrants to the Middle East to$8,000 (in 1996) for Thais seeking work inJapan (table 3.1).Recruitment agencies’ potential for earningrents raises the issue of whether governmentsshould regulate their fees. As constraints onmigration generate significant excess profits,efforts to regulate recruitment fees wouldappear to have merit. However, limitations onthe terms of mutually agreed contracts can bedifficult to enforce, and excessive regulationcan drive recruiters underground or lead themto switch to other countries. A few governmentshave attempted to regulate fees paid torecruiters, and require registration and minimumcapital requirements or financial guaranteesto limit abuses (ILO 2003a). In general,successful regulation of recruitment agenciesinvolves a large pool of potential migrants (toreduce the likelihood of recruitment agenciesswitching to alternative sources), effectivegovernment institutions, and regulations thatfocus on the most egregious abuses (insteadof just reducing the market rate). Even thePhilippines, considered to have a model60

THE POLICY CHALLENGES OF MIGRATIONprogram, experiences difficulties in enforcingfee limits, while most recruiters, including theUnion of Filipino Overseas Contract Workers,argue that stringent regulation of recruitersimpairs the competitive position of Philippinemigrants relative to workers from other countries(Martin 2005). Destination countriesare probably in a stronger position to regulaterecruitment agencies effectively.Decisions to migrate are often madewith inadequate informationThe distance and differences in language andculture between countries of origin and destinationimply that migration is particularlyaffected by inadequate information. Migrantsmay have a distorted notion of the possibilitiesof employment and the likely wage in countriesof destination, as well as insufficient informationon the costs and potential risks ofthe trip. Smugglers, recruitment agencies, andothers with a financial stake in encouragingmigration may present a biased picture of themigration experience, and poor informationincreases the potential for migrants to sufferfrom fraud and abuse. 7Some origin-country governments haveattempted to protect their emigrant workersby regulating the terms of labor contracts. Forexample, Indonesia, the Philippines, and SriLanka have drafted model contracts for variousoccupations and host countries that detailworking terms and conditions to be specifiedin advance (Abella 1997). Such efforts can beuseful in articulating standards and informingmigrants of their rights. However, enforcementof such contracts in destination countries canbe problematic (IOM 2003). The Philippines,India, and Bangladesh review contracts priorto migrants’ departure, and Bangladesh verifiesthe genuineness of overseas work visas.Some countries have entered into bilateralagreements that require destination countriesto issue visas only if the contract is approvedby the origin country. Restrictive policies ofthis kind run the risk of encouraging irregularmigration if workers cannot secure approvalof contracts in advance. Some countries use financialincentives, such as special exemptionsfrom travel taxes for those who clear contractswith government agencies, to facilitate governmentreview of contracts. Some countries withlarge numbers of emigrants offer a comprehensiveset of services, including predeparturetraining, information on labor markets in destinationcountries, legal services, 8 reintegrationsupport, and welfare funds financed fromfees paid to origin-country governments bydeparting workers. 9A diaspora can reduce the costsfacing migrantsThe stock of emigrants in countries of destinationcan reduce the costs facing new migrantsfrom the same origin country. The majorcountries of origin with significant diasporasin high-income countries (figure 3.2) are forthe United States: China, Cuba, El Salvador,India, Mexico, Philippines, and Vietnam; andin other countries: Turkey (for Germany);Serbia and Montenegro, Morocco, and Algeria(for France); and China (for Japan andCanada). As migrant networks spread, privateinstitutions and voluntary associations emergeto provide a range of services, including counseling,social services, and legal advice; lodging,credit and job search assistance; and themeans to reduce the cost of undocumented migration,including smuggling and transport,counterfeit documents, and arranged marriages(Massey and others 1993). The migrantdiaspora can also reduce the likelihood of, andfears concerning, abuse (see Gunatilleke 1998for the example of Sri Lanka).Evidence of how the diaspora encouragesmigration can be seen in the grouping ofimmigrants from the same country or localregion in countries of destination. Bartel(1989) and Jaeger (2000) find that U.S. immigrantstend to move near former immigrants ofthe same ethnicity. Munshi (2003) finds that anindividual is more likely to be employed and tohold a higher paying nonagricultural job whena large number of migrants from his homecommunity are in the United States. 10 Studiesof the Asia Pacific region have shown that61

GLOBAL ECONOMIC PROSPECTS 2006Figure 3.2 Major developing country diasporas in developed countriesMajor diasporas in the United StatesOther major diasporasMillions of people12108642Millions of people21.510.50MexicoPhilippinesIndiaChinaEl SalvadorCubaVietnam0Turks inGermanySerbiansin FranceMoroccansin FranceAlgeriansin FranceChinese inJapanChinese inCanadaSource: OECD 2004.Note: Data from the United States are based on country of birth, while most European migration data are based on nationality.networks raise migration rates once a few initialmigrants are established in destinationcountries (Massey and others 1998).Irregular migration is often subjectto substantial costs and risksIrregular migration appears to have increasedsignificantly in major countries of destination,although the estimates are unreliable. Theestimates are based typically on the differencesbetween census reports and other immigrantregistries, arrests at the border or internally,and regularization programs (Jandl 2004). 11Irregular migration may have doubled in theUnited States between 1990 and 2000, andmay now account for some nine million people(Passel, Capps, and Fix 2004)—about25 percent of the total stock of migrants. Thescale of irregular migration in Europe mayalso be high. Just under 700,000 irregular migrantsapplied for regularization under the recentamnesty drive in Spain. Mid-range estimatesprovided by Jandl (2003) indicate thatirregular migrants range from less than 10percent of the total reported stock of migrantsin France to 60 percent in Greece (figure 3.3).Most irregular migrants are low-skilledbecause (a) immigration laws in high-incomeFigure 3.3 Estimates of stock of irregularmigration% of reported migrant stock706050403020100FranceGermanyGreeceSource: Medium estimates from Jandl 2003.Note: See note 11 at the end of this chapter for a definition of“irregular” in this context.countries provide high-skilled workers withgreater opportunities for legal entry andresidence, and (b) it is more difficult for highskilledworkers to practice their professionswithout adequate documentation, such as educationalcredentials (Chiswick 2000). Also,skilled migrants often have more to lose athome and may be less inclined to run the risksinvolved in irregular migration. IrregularItalySwitzerlandUnitedKingdom62

THE POLICY CHALLENGES OF MIGRATIONmigrants also tend to be temporary, rather thanpermanent, immigrants (Carter 1999). 12Irregular migration imposes substantialcosts on migrants, compared with permanentmigration. It can be more expensive: the averageprice in 1991 for smuggling an illegalmigrant from China to the United States wasestimated at $30,000 and from $3,750 to$12,000 for migrants smuggled to Lithuania(Salt and Stein 1997). Irregular migrants canalso be paid low wages, have poor workingconditions, and be subject to violations of theprotections afforded under industrial-countrylabor laws (Vayrynen 2003). Employers maybe able to pay irregular migrants less thanlegal migrants and natives because only certainemployers will hire irregular migrants, orbecause the migrants are reluctant to moveaway from support networks. Lower pay andhigher costs of migration also make irregularmigration less desirable for the origin country,because they cut into remittances. Remittancescan be reduced by the relatively expensivemoney transfer operations used by irregularmigrants who lack access to bank accounts(see chapter 6).Irregular migrants can also be exposedto physical danger. Since 1994 an estimated2,600 undocumented migrants have diedcrossing the United States–Mexico border(Meek 2003). Entrapment into prostitutionis a danger for women and children (Wickramasekera2002). Trafficking in persons isestimated globally to involve some 600,000 to800,000 men, women, and children each year(U.S. Department of State 2004). Differentnational policies toward migration controlmake it difficult to combat trafficking andsmuggling, although the international protocolsagainst these activities provide a commoninstrument to criminalize them. 13 Severalgovernments, notably in Southeast Asia, haveinstituted restrictions on the emigration ofwomen, fearing their exploitation. Unskilledand semi-skilled women are allowed to emigratefrom Bangladesh only when accompaniedby a male partner (Siddiqui 2003). Suchoutright bans, in addition to being limitationson what is generally viewed as a basic right toemigrate, are likely to be counterproductive:they may compel many women to move asundocumented migrants, thus increasing theirvulnerability (Misra and Rosenberg 2003).More comprehensive, cooperative policies bygovernments are likely to have a more positiveeffect, including the dissemination of informationon the risks of migration, strengthenedprotection for women in destination countries,and stepped-up identification and prosecutionof traffickers. Migration agreementsbetween countries of origin, transit, and destinationcan help achieve such policy coherence(as in the bilateral agreements between someEU states and Morocco and Tunisia, forexample).There are costs for those left behindFinally, migration may impose costs on familymembers left behind, particularly children.For example, Battistella and Conaco (1996)find that the children of migrant parents fromLuzon, Philippines, performed worse in schooland tended to be less socially adjusted (particularlyif the mother had emigrated) than childrenwith both parents at home. On the otherhand, Bryant (2005) found that the improvementin the children’s health and schooling (financedby remittances), coupled with stronginvolvement of the extended family, tended tomitigate the social costs of a parent’s migration.In general, emigration does impose hardshipson family members left behind, but italso improves household income and improvesfamilies’ ability to make compensatingadjustments that mitigate those hardships. 14The impact of international migrationon countries of origin variesThe impact of migration on countries of originvaries greatly, depending on the size of emigrantflows, the kinds of migrants, and laborand product market conditions in the country.In describing these effects, it is useful to distinguishbetween skill levels, given the differencesin the labor markets for low- and highskilledworkers.63

GLOBAL ECONOMIC PROSPECTS 2006Low-skilled migrationLow-skilled migration can improve labormarket conditions for other poor workersThe stock of low-skilled emigrants who movedfrom developing to industrial countries in 2000averaged about 0.8 percent of developingcountries’ low-skilled, working-age residents—about the same as in 1990. The regions withcountries close to the major destination countrieshad relatively high rates of low-skilledemigration (figure 3.4). 15The effects of South–North migration onworking conditions for low-skilled workers inthe developing world as a whole must besmall. In individual countries, however, largescaleemigration can place increased pressureon wages or reduce unemployment of lowskilledworkers at the margin. For example,real wages in Pakistan’s construction sectorand the Philippines’ manufacturing sectorclosely trace the deployment of overseasworkers (Majid 2000; Gazdar 2003). 16 Lowskillemigration also may reduce underemploymentor raise labor-market participationwithout significant wage increases. Wagetrends in Albania, Bangladesh, and Sri Lanka,for example, display no obvious signs ofFigure 3.4 Emigration rates for low-skilledworkers% of low-skilled workers in home region43.532.521.510.50Europe andCentral AsiaLatin Americaand CaribbeanSource: Docquier and Marfouk 2004.Middle East andNorth AfricaEast Asia andPacificSouth AsiaSub-SaharanAfricaimprovement, despite massive emigration(Lucas 2004a). The wage response to emigrationdepends on the institutional setting inthe labor market in the home country (such asthe role of unions, public sector employment,and minimum wage laws); the extent of emigrationrelative to the domestic labor force;and the degree to which emigrants were productivelyemployed before migrating.The impact of international migration maydiffer considerably among regions withincountries of origin, depending on the degree ofgeographic concentration of emigration andthe links with other regions through internalmigration. People in regions lying close to acommon border with the destination countryor with easier access to overseas markets (suchas metropolitan centers or coastal areas) havea tendency to migrate (Long, Tucker, andUrton 1988; Malmberg 1997). These effectscan be greatly magnified through the influenceof migrant networks, once initiated from aspecific location (Gunatilleke 1998; Shah1998). Internal migration to areas of high departurecan be quite important to the trickledownbenefits of international migration, andinternal migration also can have an importantpoverty-reducing impact (box 3.1).Migration of low-skilled workersis usually beneficialWhether emigration results in reduced underemployment,increased labor-market participation,or higher wages, low-skilled workersin the home labor market gain, either directlyor indirectly, from additional remittancespending. Emigration of low-skilled workersthus can act as a safety valve for the failure tocreate appropriate employment at home.There have been cases, however, usually in thecontext of South–South migration, wherelarge outflows of temporary or irregular workershave resulted in massive return flows dueto economic or political shocks in destinationcountries. 17 Also, reliance on large-scale emigrationmay retard efforts to address the issueof employment expansion over the longterm, as a result of either the remittance-driven64

THE POLICY CHALLENGES OF MIGRATIONBox 3.1Internal versus international migrationAlthough internal migration is much larger thaninternational migration, they are in some respectssimilar phenomena. Both are largely driven byeconomic disparities among regions, although conflictand natural disasters (not discussed here) canalso catalyze large movements of people. Both internaland international migration can be permanent ortemporary, voluntary or forced. Both can subject migrantsto substantial risks. Both can result in improvedlabor conditions in the regions of origin.Internal migration in Bangladesh, for example,makes more rural land available for tenancy(Afsar 2003). While the risks are often higherwith international migration (particularly forirregular migrants), internal migration can also beplagued by trafficking of persons, particularly ofwomen and children, as evident in parts of Sub-Saharan Africa (Black 2004) and in India and China(Lee 2005).Internal and international migration also havetheir differences. Wage differentials are often lowerwithin countries than between countries, reflectingsmaller differences in economic conditions withincountries than between countries.Internal migration may have a larger role inreducing poverty than does international migration.While the expected wage gain is lower in internalmigration, poor workers may have a better chanceof finding work domestically than in economies withhigher wages, where workers’ lack of skills and languageability restrict job opportunities. Moreover,international migration tends to be more costly, sothe poorest workers may not be able to afford it.While data are sparse, household surveys in a fewcountries indicate the important impact that internalmigration has on poverty. In Sierra Leone, internalremittances helped reduce income inequality in poorareas in the 1980s (Black and others 2004). In theAsia Pacific region, international remittances accruedisproportionately to richer regions, while domestictransfers are directed mostly to poorer regions (UN-ESCAP 2003). In Ghana, internal remittances are estimatedto reduce the level of poverty by 14 percent,compared with only 5 percent for internationalremittances (Adams 2005).Internal migration can have large costs on receivingareas. The proliferation of HIV/AIDs in Ghana hasbeen linked to the movement of women from rural tourban areas, where unemployment and poverty oftenforce them into the unprotected sex trade (Black2004b). In some countries, internal migration to thecities has been so massive as to increase crowding andplace inordinate burdens on public services, whichlowers the quality of urban life. aInternal migration often responds to substantialemigration abroad. In the Philippines, there are indicationsof large movements from rural areas of thePhilippines into the Manila region from which mostoverseas workers are drawn, although this movementappears to have done little to help sustain wages inrural areas (Saith 1997). Bangladesh has seen rapid responsesof intra-village migration to replace departingworkers (Mahmud 1989). At the same time, as internalmigrants gain skills, resources, information, andnetwork contacts, they often emigrate internationally.For example, workers displaced by falling agriculturalprices in southern Mexico often moved to northerntowns to work in maquiladoras, later moving to theUnited States.The links between internal and international migrationare inadequately understood, in part becausebasic data are lacking. To gather more data, it hasbeen recommended that a migration module in demographicand health surveys, censuses, household incomeand expenditure surveys, and labor force surveysbe included (Afsar 2005).a In China, for example, rapid urbanization has been accompaniedby the emergence of urban enclaves of landless, unemployedmigrants from rural areas (Pan 2004).65

GLOBAL ECONOMIC PROSPECTS 2006exchange-rate appreciation or the reducedpressure for policy reform. In general, however,the opportunity to send low-skilledworkers abroad provides substantial benefitsto origin countries because of the impact onlabor markets and remittances.Low-skilled migration has contributedto poverty alleviationThe reduced supply of low-skilled workersmay help to alleviate poverty, if as a result ofemigration, poor people receive higher wagesor find new opportunities to work or receiveremittances (see chapter 5). Low-skilled emigrationalso alleviates poverty to the extentthat the people emigrating are poor. 18 It isunlikely, however, that a large proportion ofmigrants to industrial countries are pooraccording to the World Bank’s definitionof poverty as living on less than $2 a day—although certainly a very large share is poorcompared to even the poorest in high-incomecountries. Most migrants from Mexico to theUnited States come from households located atthe middle and upper-middle levels of the incomedistribution (Rivera 2005). Individualswith very low incomes are unlikely to be ableto obtain the financial resources necessary formigration (see, for example, Mahmud 1989for Bangladesh). Most of the world’s poor peoplelive in countries that are far away from industrialcountries (Bangladesh, Brazil, China,India, Indonesia, and most of the countries ofSub-Saharan Africa), so transportation is expensive.Moreover, many poor people lack therudimentary skills required to obtain a job inindustrial countries, as well as the social networksthat would facilitate migration and provideassistance once in the destination country.Nevertheless, the limited data indicate thatthe very poor do move abroad to some extent.In Sri Lanka, returns from household surveysshow that the share of households with a familymember abroad is approximately equalacross income groups. 19 Adams (2004) providesmodel-based estimates implying thatabout 5 percent of Guatemalan householdswith incomes of less than $2 a day receivedremittances from abroad (used here as a proxyfor having a household member who emigrated).Adams (2005) shows that less than8 percent of Ghanaian households that receivedinternational remittances had estimatedincomes (excluding remittances) that fellwithin the first to fourth deciles of householdsby per capita expenditures; 55 percent had expendituresin the top three deciles. Lucas(2004a) quotes studies of Kerala (India), Pakistan,the Philippines, and Thailand to supporta conclusion that most emigrants were notfrom the lowest income levels, although thepoorest did participate to some extent.High-skilled emigrationThere is a sharp increase in high-skilledmigrationThe emigration of high-skilled workers fromdeveloping countries has increased since the1970s. 20 By 1990, the stock of high-skilledSouth–North migrants in the United Statesalone was more than eight times the totalnumber of high-skilled migrants from developingto industrial countries over the 1961–72period, not counting foreign students(Docquier and Rapoport 2004). The numberof highly educated emigrants from developingcountries residing in OECD countries doubledfrom 1990 to 2000, compared to an approximate50 percent rise in the number of developing-countryemigrants with only a primaryeducation (Docquier and Marfouk 2004).Rates of high-skill emigration vary enormouslyamong developing countries, from lessthan 1 percent (Turkmenistan) to almost90 percent (Suriname) and by region, from15 percent for Sub-Saharan Africa to 5 percentfor Europe and Central Asia (figure 3.5).It is important to keep in mind this degree ofdiversity, as high-skill emigration can havevery different effects, depending on the sizeand economic conditions in origin countries.The increase in high-skilled migration ispartly due to the growing importance ofselective immigration policies first introduced66

THE POLICY CHALLENGES OF MIGRATIONFigure 3.5 Emigration rates for those witha tertiary education, 2000% of total tertiary-educated population1614121086420East Asiaand PacificEurope andCentral AsiaLatin Americaand CaribbeanMiddle East andNorth AfricaSouth AsiaSource: Docquier and Marfouk 2004.Sub-SaharanAfricain Australia and Canada in the 1980s andlater in other OECD countries. 21 Major recruitingcountries have increased their intakeof skilled migrants and relaxed their criteriarelating to labor-market testing and job offers.Some countries (for example, Germany,Norway, and the United Kingdom) have introducednew programs; others (such as Austria,Republic of Korea, the Netherlands, andSweden) offer fiscal incentives to attract talentto specific sectors (OECD 2005). These programs,and the migrants themselves, are respondingto rising skill premiums in industrialcountries that have tightened global competitionfor skilled workers.In some instances, high-skilled emigrationhas a negative impact on living standardsof those left behind and on growthThere are several reasons that migration ofhigh-skilled workers may decrease living standardsand growth. First, the total return toeducation may be greater than the privatereturn, because highly educated workers maybe more productive when interacting withsimilar workers, and they may help train otherworkers. One statistical measure of the beneficialimpact of high-skilled immigrants isthat in the United States, both internationalgraduate students and skilled immigrants werefound to be positively correlated with patentapplications (Chellaraj, Maskus, and Mattoo2005). Highly educated citizens may alsomake contributions to public goods—for example,in improving governance and strengtheningthe administrative capacity of thestate—which may be lost through high-skilledemigration (McMahon 1999).Second, the productivity of firms mayincrease with size. If large firms require networksof professionals with specialized skills,then overall productivity will be higher withmany professionals. For example, the value oftelephone networks increases with the numberof people connected. Expanding networksefficiently may require highly technical skills.Third, emigration of high-skilled workersmay impose a fiscal cost. In most developingcountries education is heavily subsidized bythe state, so that the permanent emigration ofeducated workers represents a loss of fiscalrevenues. 22Finally, emigration of high-skilled workerswill increase the price of services that requiretechnical skills. It is difficult to provide comparablelevels of service with low-skilled workers,and greater resources devoted to trainingmay be lost through further emigration.But high-skilled migration is oftenbeneficial for origin countriesThe costs of high-skilled emigration shouldbe evaluated against the beneficial effects ofmigration, skilled and unskilled: increased remittances,higher wages (for migrants andworkers who stay home), and benefits to destinationcountries (see chapter 2). Moreover,high-skilled emigration will have a limited impactif it is difficult for high-skilled workers tofind productive employment in the country oforigin. This may be the case for three reasons.First, the investment climate may be so poor,because of political instability or other reasons,that many high-skilled workers cannotpursue their professions. Even under such conditions,however, high-skilled emigration maybe harmful if it deprives the government of67

GLOBAL ECONOMIC PROSPECTS 2006competent administrators and limits theprospects for growth once the investmentclimate improves. Second, a significant proportionof high-skilled workers may not betrained in professions required by the economy,perhaps because of government subsidypolicies. And third, some of the smallest developingcountries lack the economic scale toproductively employ a large number of specializedprofessionals. 23 These issues serve tounderscore concerns over the appropriatenessof state subsidization of university educationin many countries. 24Some recent articles have claimed thathigh-skilled emigration, even of productivelyemployed workers, may benefit development.The opportunity to emigrate increases the returnsto education, leading more individuals toinvest in education with a view to emigrating.However, only some of the educated peoplewill actually emigrate. If the increase in humancapital of those unable to emigrate exceeds theloss from those who do emigrate, then society’shuman capital rises following the openingof emigration opportunities (a phenomenonknown as the “brain gain”). 25 The effect willbe largest in countries with large stocks ofemigrants (so that the probability of emigrationis high). These models have beenquestioned, however, because they assume thatforeign firms are not able to discriminateamong educated workers (otherwise theywould take the best qualified, and so destroyincentives for education by marginal candidates),and because these models do not applywhere family reunification programs, unrelatedto the skills, predominate (Schiff 2005).Findings on the impact of high-skilledemigration are mixedIt is difficult to generalize about the impact ofskilled migration. The dispute over gains andlosses has remained largely conjectural andhas not been settled by the available empiricalstudies. On balance, it is not possible at presentto provide an aggregate, reliable estimateof the true impact of high-skilled emigration.Some partial conclusions follow.Table 3.2 Emigration rates of skilledworkers, 2000Percentage of workers with tertiary education living abroadLess than 10% to 20% to More than10% 20% 30% 30%Number of countries 62 33 16 28Share of developing 75 19 3 3country population (%)Source: Docquier and Marfouk 2004.The available data indicate that high ratesof high-skill emigration affect only a smallshare of developing countries’ population. Adata set developed by Docquier and Marfouk(2004) indicates that the 77 countries withhigh-skilled emigration rates (to industrialcountries) in excess of 10 percent account foronly one-quarter of developing-country population(table 3.2). 26 Moreover, about half ofthese people live in countries with very poorinvestment climates (included in the bottom25 percent of developing countries, as measuredby the United Nations’ Human DevelopmentIndex), which may indicate that manyhigh-skilled workers face limited opportunitiesto practice their professions. It is important tonote that these data do not distinguish by profession(even though high emigration rates forliterature professors and physicians wouldhave different economic impacts) or by quality(the emigration of a Nobel laureate physicistwould represent a greater loss than the emigrationof an average university graduate).Some countries encourage skilled migration.China, Cuba, India, the Philippines, SriLanka, and Vietnam all have programs tofacilitate training for migration, suggestingthat some policymakers see the benefits ofskilled migration—among them remittances,relieving job market pressures, developmentof an extensive diaspora, and expectationsthat many migrants will eventually returnwith improved skills (as discussed below).Direct, cross-country tests of the relationshipbetween high-skilled emigration andgrowth have been mixed. The preponderanceof evidence supports the view that education68

THE POLICY CHALLENGES OF MIGRATIONmakes an important contribution to growth. 27Beine, Docquier, and Rapoport (2001) detecteda positive and significant impact onhuman capital formation from the opportunityfor emigration, whereas Faini (2003)found that a higher probability of migrationfor workers with secondary education had novisible impact on secondary educationalachievement in the home country.High-skilled emigration has had enormousimpact on some sectors, especially healthThe sectoral distribution of high-skilled emigrantsis important for assessing the implicationsfor countries of origin. Meyer andBrown (1999) estimate that about 12 percentof developing-country nationals trained in scienceand technology live in the United States.If accurate, these estimates suggest that highskilledemigration may be much more seriousfor production than shown by the data fromDocquier and Marfouk (2004) given above,where total high-skilled emigration to industrialcountries was estimated at about 8 percentof the stock of high-skilled developingcountrynationals.High-skilled emigration may have a particularlysevere impact on the health sector, andthe emigration of doctors and nurses may reducethe likelihood of some countries meetingthe Millennium Development Goals for reducingchild mortality, improving maternalhealth, and combating HIV/AIDS and tuberculosis.Chanda (2001) estimates that at least 12percent of the doctors trained in India live inthe United Kingdom, that Ethiopia lost half ofits pathology graduates from 1984 to 1996,that Pakistan loses half of its medical schoolgraduates every year; and that in Ghana onlyabout one-third of medical school graduatesremain in the country. Perhaps one-half of thegraduates of South African medical schoolsemigrate to high-income countries (Pang,Lansang, and Haines 2002), and Jamaicahad to train five doctors, and Grenada 22, tokeep just one (Stalker 1994). 28 Of course, theincentive for migration is often conditionednot only by the opportunity for higher earningsabroad, but also by poor working conditionsand public sector services in origin countries.Origin countries face considerabledifficulties in limiting high-skilledemigrationEven if high-skilled emigration were found tobe detrimental to living standards and growth,countries of origin would face serious obstaclesin reducing it. Some countries haverequired that graduates of publicly fundededucation work for a period of time in publicsector jobs. But such requirements can beevaded, and their existence is likely to discouragereturn of migrants to the country of origin.Several proposals have been made for internationalschemes to tax high-skilled emigrants,with the funds earmarked for developing countries.Such schemes have made no progress, asthey would be hard to enforce. Calculating thewelfare loss from high-skilled emigration andthus setting an appropriate level of tax wouldbe difficult. Moreover, the schemes wouldrequire the cooperation of migrants and countriesof destination—something not likely to beachieved. Bhaghwati (1976) advocates thatdeveloping countries should subject their nationalsworking abroad to local taxes, as doesthe United States. However, many developingcountries would find such a system of taxationdifficult to administer.Some governments encourage skilled workersto stay by improving working conditions,providing research facilities, and giving incentivesfor research (see the discussion of incentivesfor return, below). China has reporteda nine-fold increase from 1995 to 2003 inforeign programs offered in cooperation withlocal institutions, which has resulted in lowernumbers of students going abroad (Vincent-Lancrin 2004). Such programs may requiresubstantial resources, and poorer countrieswill face difficulties in creating the conditionsrequired to retain their most-skilled workers.In some cases improvements in governance,which may require political determinationrather than large expenditures, may help toretain workers.69

GLOBAL ECONOMIC PROSPECTS 2006Cooperation from destination countriesmay helpEffective means of limiting high-skilled emigration,or increasing its benefits, will requirethe cooperation of countries of destination.One example has been the United Kingdom’scode of practice for the international recruitmentof health-care workers, which restrictsthe active solicitation of health care professionalsfrom developing countries. However,implementation has been difficult because thecode does not apply to private sector recruiters(House of Commons 2004). A recent proposalto require countries of destination to provideCaribbean countries with subsidies for traininghealth professionals could increase thesupply of health workers in both origin anddestination countries (Commonwealth Secretariat2005). 29 More broadly, contributions bydestination countries to education in origincountries could both compensate origincountrygovernments for their training of emigrantsand improve the qualifications of workerscoming to destination countries. Anotherapproach, which would improve the coherenceof development policies, would be for destinationcountries to increase their investment insectors in which they lack skills, rather than“raiding” those skills from poor countries.DiasporasA large diaspora can expand marketaccess for origin countriesA potent benefit of high-skilled emigration isthe creation of a large, well-educated diaspora,which improves access to capital, information,and contacts for firms in countries of origin. Immigrantsplay a role in facilitating trade by providinginformation and helping to enforce contracts(Rauch and Trindade 1999) and by actingas intermediaries that can match buyers with reliablelocal suppliers (Yusuf 2001). Johnson andSedaca (2004) emphasize that diasporas can actas “first movers” who catalyze growth opportunitiesand make connections between marketsthat otherwise would not exist. Barré and others(2003) cite the importance of diasporas ingenerating possibilities for codevelopment betweenfirms in the countries of origin and destination,and expanding technical cooperation.With the growth of outsourcing of manufacturingcomponents and telecommunications andother services, diaspora networks may be of increasingimportance. However, despite thebroad agreement on the importance of diasporasand the many anecdotal comments on howthey have assisted development, it is difficult toquantify these benefits.The diaspora can be a significant source offoreign-exchange earnings (beyond remittances)for countries with sizeable emigration.Orozco (2003) documents diaspora-relatedincreases in exports and tourism receipts forCentral America. Gould (1994) and Headand Ries (1998) found that increased emigrationto the United States and Canada raisedexports from countries of origin.There is some evidence that the diasporaplays an important role in the transfer ofknowledge between destination and origincountries. Agrawal, Cockburn, and McHale(2003) find that patent applications are likelyto be filed in both the country of residence andthe country of origin. Meyer and Brown (1999)and Brown (2000) identify Internet-based expatriatenetworks of skilled professionals andstudents that facilitate the transfer of knowledge.However, the effectiveness of these networksis open to question: less than half of the61 Internet-based networks examined in 2004were updated regularly, and only 56 percentwere updated within the past year. 30 Origincountrygovernments can help maintain ties tothe diaspora by supporting professional networks,promoting dialogue with government,and funding educational, linguistic, and culturalprograms.The return of expatriates canbenefit developmentThe return of expatriates to their homecountry is widely perceived as benefittingdevelopment (Ellerman 2003). Expatriates may70

THE POLICY CHALLENGES OF MIGRATIONbe more effective than foreigners in transferringknowledge back home because of theirunderstanding of local culture. However, returneesmay also represent retirees, or the lessskilledof the skilled cohort of emigrants(Borjas and Bratsberg 1994; Lowell andFindlay 2001), or may have difficulties inreadapting to the home country (Faini 2003),or their skills may have deteriorated whileabroad (Ghosh 1996). Returnees may be thosedisappointed by the wages or working conditionsor may have more difficulty in finding orretaining jobs. 31A range of programs have been establishedto encourage return of highly educated nationalsliving abroad, with mixed results.Among developing-country governments, forexample, China, the Philippines, Taiwan(China), Thailand, and Tunisia have offered awide range of incentives, including researchfunding, access to foreign exchange, expandedreal estate investment options, and studyopportunities. 32The domestic policy environment is criticalto productive return. Cervantes and Guellec(2002) cite the favorable impact of returningexpatriates in the Republic of Korea, attractedby strong research and development (R&D)environments and infrastructure investments.Industrial parks helped to lure entrepreneursback to China. In Taiwan (China), the HsinchuIndustrial Park attracted more than 5,000returning scientists in 2000 alone (Saxenian2002). Conversely, a poor investment climatewill inhibit return. In Armenia, barriers toforeign direct investment (FDI) and inadequateenforcement of contracts have preventeda more active involvement of the Armenian diasporain local development (Gevorkyan andGrigorian 2003). Saxenian (2000) cites the reluctanceof Indian entrepreneurs to return becauseof government regulations that increasethe administrative cost of operating abusiness—although the Indian diaspora hascontributed to the development of informationtechnology in Bangalore. 33Both origin and destination countries canhelp facilitate return, on both a temporary andpermanent basis, through changes in regulaion.Both can allow dual citizenship, anincreasingly common practice (Aleinikoff andKlusmeyer 2002). Origin countries can eliminaterules that prevent emigrants from owningor investing in property back home. Permanentresidents can be protected from losing theirstatus if they leave for a relatively limited periodof time, as this discourages productivetemporary returns to the origin country.Destination countries can also allow returningmigrants to benefit from the rights they acquireduring their work abroad, such as pensions,health insurance, and disability programs(Holzmann, Koettl, and Chernetsky 2005). 34Such arrangements, however, require effectiveinstitutions in the origin country to providesuch services and are best achieved through negotiationsbetween origin and destinationcountries.Destination countries have provided variousincentives for the return of migrants. For example,France has provided loans and technicalassistance to migrants from Mali and Senegalto establish businesses in their home countries.However, few of the businesses appear to havebeen successful, either because of the inadequateinvestment climate in the recipient communities(Gubert 2005) or because participantshad worked in low-level jobs in France andlacked entrepreneurial skills (Magoni 2004).Many of these programs are quite small. 35International organizations, too, have managedprograms to promote return, althoughthey tend to cover few emigrants. The IOM’sReturn of Qualified African Nationals programsuccessfully attracted more than 2,000highly skilled persons back to 41 Africancountries from 1974 to 1990, and the programwas later expanded to the Migration forDevelopment in Africa program (MIDA). 36Similar programs have been run for LatinAmerican countries, Afghanistan, andBosnia and Herzegovina. 37 The United NationsDevelopment Programme’s TOKTENproject promotes temporary return (threeweekto three-month development assignments),which is often easier to achieve.71

GLOBAL ECONOMIC PROSPECTS 2006Temporary migration andinternational agreementsThe number of temporary workers admittedto high-income countries under skillbasedprograms rose substantially in the1990s, doubling in the United Kingdom andalmost quadrupling in the United States (figure3.6). Foreign student programs are alsoproliferating, in part due to competitive marketingand entry policies by Australia,Canada, Germany, and the United Kingdom.Inflows of unskilled, seasonal workers alsohave increased since the early 1990s in mosthigh-income countries that have such programs(UN 2004). Greater emphasis on temporarymigration may reflect opposition toexpanding permanent migration.Temporary labor migration schemes varygreatly from country to country. Skilled migrantstend to have more opportunities underunilateral visa programs (such as the U.S.H-1B visa and the temporary skilled migrationprograms of Australia and Canada), whileunskilled migrants often must rely on bilateralor regional agreements, such as the seasonalwork programs between Germany and countriesin central and eastern Europe.Figure 3.6 Number of temporary workersadmitted under skill-based programsThousands600500Australia400Canada300United States aFrance200United Kingdom10001992Source: United Nations.2000Note: Double counting occurs if a person enters the UnitedStates more than once during the year.a. Reflects the number of admissions under H-1B visas,not the number of persons.Temporary migration may facilitategreater migration flowsWhile temporary migration programs can generatebenefits similar to those of permanent migration(such as higher incomes for migrants,improved efficiency in destination countries,remittances for origin countries), in some respectsthose benefits differ between permanentand temporary migration. For the legal migrant,being admitted on a temporary basismay be less desirable than permanentresidence, which provides for free movementbetween destination and origin countries.However, the existence of temporary migrationprograms (reflecting resistance to permanentmigration) may facilitate larger legal flows.From the point of view of the destinationcountry, temporary migration offers the flexibilityto target required skills and to adjustentry in response to changes in labor demand.Temporary migration can reduce fiscal pressuresthat may be associated with low-skilledimmigration, in that temporary migrants tendnot to bring their dependents (who may requirepublic services). At the same time, programscan be made conditional on employment,thereby limiting social tensions fromimmigration. Provisions for temporary migrationalso can be viewed as offering an alternativeto irregular migration.The record of destination governments inensuring that temporary migration programsare indeed temporary has been mixed. Forexample, over 22 years, most workers in theMexico–U.S. guest workers program returnedat the end of their seasonal jobs, although somefound ways to obtain permanent status (Martin2003). Similarly, from 1960 to 1973, threequartersof the 18.5 million foreigners whocame to work in Germany left. But 25 percentremained. Coupled with rules that allowedmany to eventually bring their dependents, thisresulted in significant permanent settlement.There is some evidence that the recent guestworker programs in Europe, particularly thoseinvolving subcontracting for short-term projectsand some return incentives, have managedhigher return rates (Lucas 2004a).72

THE POLICY CHALLENGES OF MIGRATIONThere are advantages and disadvantagesto temporary migrationReliance on repeated, temporary migrationfor workers also has some economic drawbacks.Hiring a temporary immigrant maymean a shorter duration of employment comparedwith hiring a permanent immigrant, andthus higher costs for training. Temporary migrantsare also less likely than permanent onesto invest in skills specific to the destinationcountry (such as language proficiency andlicensing requirements), because the returnsare enjoyed over a shorter period of time(Chiswick 2000). Nevertheless, for emigrantsfrom developing countries, the wage differentialsare so large that they may justify substantialinvestments in acquiring such skills,even for temporary stays.For the origin country, remittances (andrepatriation of assets) may be higher withtemporary migration, because temporary migrantsare less likely to bring their dependentsand more likely to maintain close ties withthe home country. Perhaps most importantly,temporary migration programs can provide anopening to increase legal, unskilled migration,which generates the greatest developmentalimpact for origin countries, as already noted.On the other hand, temporary migration mayprovide a less reliable means of exportinglarge labor surpluses, as cancellation of futureaccess is easier for destination countries thanexpelling existing migrants. However, it is thisflexibility (coupled with less long-run populationpressure, fewer concerns over integration,and fewer pension commitments) that makestemporary migration desirable for destinationcountries, thus facilitating agreements forlarger unskilled migration (Winters 2005).Bilateral agreements can play animportant role in low-skilled, temporarymigrationBilateral labor agreements have become amajor vehicle for low-skilled, seasonal workersin agriculture, tourism, and construction, asevidenced by agreements between the UnitedStates and Mexico, Germany and central andeastern European states, and Saudi Arabia andEgypt and Libya. There are several hundredsuch agreements worldwide, including some168 signed in the last 50 years in Latin Americaalone, half in the past 10 years (IOM 2005a). 38Bilateral agreements could improve the benefitsof temporary migration for origin countriesthrough greater certainty of access and conditions.This may be particularly important inmarkets where increased competition fromother suppliers might lead to a reduction in access(as occurred in Saudi Arabia and the MiddleEast in the 1990s). Bilateral agreements canhelp build the confidence in both origin anddestination countries that a particular channelof migration will generate real benefits andminimize costs—for example, that migrantswill be treated well and will return at the end oftheir contract.Several factors impede the maximizationof gains from bilateral agreements, however.Some origin countries may lack sufficientlyreliable information on demand for theirworkers in destination countries, and in whichsectors, to negotiate appropriate agreements.Destination countries may likewise have difficultyreliably estimating labor shortages inparticular sectors. And origin countries mayface resource constraints in implementingobligations with regard to prescreening of migrantsor monitoring of their return, althoughthese may be covered by the destinationcountry (as in a nurses program betweenRomania and Italy). Origin countries may alsolack bargaining power to conclude termsfavorable to them or to conclude agreementsat all. For example, of 18 bilateral agreementsproposed by the Philippines with countriesin Africa, Asia, Europe, and the MiddleEast, five countries refused to enter intoagreements, and others have remained inactive(Go 2004).Nevertheless, there is scattered evidencethat countries like the Philippines have beenable to use bilateral agreements to gainfavorable employment conditions for theirmigrants—and in some cases to support theirreturn and reintegration (Lucas 2004b).73

GLOBAL ECONOMIC PROSPECTS 2006Bilateral agreements can help ensure thatorigin-country credentials are accepted indestination countries, for example. 39 They canhelp ensure that temporary migration is indeedtemporary, and that returning migrants arereintegrated, by supporting the transfer oftechnology and human resource developmentin the origin country, as under Spain’s agreementswith Colombia and Ecuador (IOM2005a). Bilateral agreements can also ensurethat the origin country cooperates in monitoringand managing migration, for example, byincorporating a readmission provision (as in the1997 agreement between Italy and Albania).They also can limit the effects of braindrain. For example, a pilot scheme betweenthe Dutch and Polish ministries of healthprepared Polish nurses for employment in theDutch health care system for a maximumperiod of two years and to facilitate theirsubsequent return and reintegration into thePolish health care system. 40 Other proposalstake a development-cooperation approach,under which destination countries fund thetraining (to their standards) of a given numberof nurses in excess of origin country demand,with the surplus nurses granted temporaryvisas to work in destination countries for aspecified period, with guaranteed return.Except in the EU, regional andinternational agreements have hadlittle impact on migrationAt the regional level, there has been someprogress on removing technical and administrativebarriers to the cross-border exchangeof skilled personnel for business purposes inAfrica, Europe, Latin America, and parts ofAsia. Also, several consultative processes onmigration have emerged at the regional andglobal levels. 41 However, with the major exceptionof the EU, most regional arrangementshave had little impact on the free movement ofless-skilled foreign workers or on permanentmigration (World Bank 2005).International treaties have had only limitedimpact on migration. Mode 4 of the GeneralAgreement on Trade in Services (GATS) hasthe potential to improve cooperation on laborservices between countries of origin and destination,but so far it has not facilitated a significantrise in cross-border labor movements(box 3.2). The ILO has pioneered the developmentof international instruments for protectingthe rights of migrant workers, and the UNGeneral Assembly adopted the InternationalConvention on the Protection of the Rights ofAll Migrant Workers and Members of TheirFamilies, which clearly defines the rights ofmigrant workers, including irregular workers(Wickramasekera 2002). The convention enteredinto force in 2004. However, none of themajor destination countries have ratified it yet,and its means of enforcement are limited.International agreements governingmigration contrast sharply with thosefor tradeA major impulse behind the General Agreementon Tariffs and Trade and its successor,the World Trade Organization, was that multilateralagreements that provide for nondiscriminationamong countries would maximizethe gains from trade. By contrast, there is littlesupport for multilateral, nondiscriminatoryapproaches to migration, at least in destinationcountries. In part this is because theeconomic implications of nondiscriminationdiffer between trade and migration. In trade,nondiscrimination maximizes economic efficiencyby allowing the lowest-cost supplier tocompete, thus reducing prices and forcinghigh-cost producers to improve efficiency orexit the market. But labor markets in highincomecountries are generally not permitted(through minimum-wage laws and socialinsuranceschemes) to adjust fully to thelowest-cost supplier. Thus the benefits ofnondiscrimination are weaker in migrationthan in trade. U.S. consumers benefit if Indianshirts are cheaper than Mexican shirts, butU.S. employers benefit little if Indians are willingto work for less than Mexicans—thedecline in wages is limited.74

THE POLICY CHALLENGES OF MIGRATIONBox 3.2The WTO General Agreement on Trade in Services(GATS) does not cover labor migration perse, but rather the narrower concept of movement ofpeople across borders as one of four modes of deliveringservices. Mode 4 covers the temporary movementof persons across borders for the purpose ofsupplying a service. “Temporary” movement is notdefined, but permanent migration is explicitlyexcluded, as are workers in most nonservice sectors,such as agriculture or manufacturing.Mode 4 service suppliers can be viewed in termsof both duration and purpose of stay: they enter acountry for a specific purpose (to fulfill a servicecontract), for a limited and (generally) specified periodof time, and are usually confined to one sectoror one job (they do not enter the labor market andare not free to search for employment). Mode 4 isnormally understood to include business visitors(persons who come for three months or less to negotiatea contract), intracorporate transferees (personstransferred within a company from one country toanother), and suppliers of contractual services (individualsor employees of foreign companies with acontract to supply a service to a client in the receivingcountry). While Mode 4 includes persons at allskill levels, to date market-opening commitments byWTO members have been limited to the highlyskilled.Relatively few market-opening commitments onMode 4 have been made by both developed and developingcountries. Those that have been made tendto be subject to restrictions on number, type, and durationof stay of service suppliers. Countries’ actualregimes for temporary entry of workers tend to bemore liberal than their GATS commitments, however,and considerable movement on service issues isoccurring in a range of sectors (such as health),notwithstanding the near-absence of relevant GATScommitments.Five issues arise regarding GATS Mode 4 as aninstrument to manage labor mobility. First, GATScommitments are fixed commitments of guaranteedMode 4 and international migrationtreatment, while migration regimes seek to retainflexibility to make adjustments in line with labormarketconditions. Second, GATS commitmentsfollow the most-favored-nation (MFN) principle—that is, treatment offered to one country must beextended to all WTO members—whereas migrationregimes can offer special treatment to countries withwhich regulatory trust or other special relationshipshave been built (through visa waiver programs, forexample). Third, GATS Mode 4 covers only a relativelylimited subset of the workers moving aroundthe globe. Agricultural workers, for example, are notgenerally viewed as falling under the GATS. Fourth,multilateral trade negotiations have a 50-year history,while migration has largely remained a national policyprerogative characterized by limited internationaldialogue. Finally, movement of people, especially thelower-skilled, raises a raft of issues related to socialand cultural integration, exploitation, impact onlocal labor markets, and, more recently, security thattrade agreements are ill-equipped to address.Against this background, what role might theGATS play in managing labor mobility? The GATSis a narrow, but sharp, instrument that can deliver apowerful guarantee of access, but only for certaintypes of workers. Beyond this, however, GATS negotiationscan be used to create a sense of urgency thatmay serve to bring migration authorities to the tableto discuss ways to manage mobility. Bilateral or regionalapproaches could, for example, include lowskilledworkers and develop creative, cooperative approachesto issues such as remittance transfer, braindrain, and loss of investment in education, prescreeningof temporary workers, and return. These agreementscould assist in building regulatory trust andimproving management schemes in receiving andsending countries. Over time, by creating a templateof basic requirements or criteria that could be appliedto all countries on a nondiscriminatory basis,they could be used to extend access to a wider groupof countries and so approach the MFN principle ofGATS.75

GLOBAL ECONOMIC PROSPECTS 2006But the largest reason that nondiscriminatoryapproaches are limited is that people arenot goods: migration has much broader implicationsfor society than does trade. Destinationcountries tend to be concerned thatimmigrants from countries with very differentcultures will not integrate easily into society,and high-income countries tend to limit lowincomemigrants for fear of overburdeningpublic services (see chapter 2). Thus eventhose countries that have immigration regimesthat do not discriminate by country tend todiscriminate by level of skill.A final important distinction between tradeand migration is that trade is subject torelatively effective regulation, while manycountries of destination face considerabledifficulties (and internal disagreements) in regulatingimmigration. The lack of effectiveregulation and incomplete efforts to controlimmigration encourages many low-skilled migrantsto run substantial risks that can lead toconditions akin to slavery, great physical danger,and even death. On the other hand, thesame lack of control works to the advantage ofmigrants by offering opportunities that mightnot otherwise exist and by benefiting groupswithin destination countries. The evidence inthis chapter suggests that cooperation betweenorigin and destination countries, throughagreements that provide for temporary, lowskilledmigration, and through enforcement oflaws protecting migrants from exploitationand abuse, can improve the impact of migrationfor countries and for migrants.Notes1. Of course, migration may arise out of a combinationof economic, political, and social goals. Also,migration among developing countries is an increasinglyimportant phenomenon, but given data limitationswe focus here on migration from developing toindustrial countries.2. Empirical work largely confirms the view thatincome differentials are important determinants ofmigration. Borjas (1987), Karemera, Oguledo, andDavis (2000), and Hatton and Williamson (2002)found that migration to the United States was negativelyrelated to source-country income per capita,among other variables. Solimano (2002) found thatreal per capita income differentials between Argentinaand source countries were the main determinant of netmigration flows in the twentieth century.3. See also chapter 2, which points out thatmigrants’ earnings (per worker) increase eleven-fold,before adjusting for differences in purchasing powerin high-income versus developing countries.4. The basic idea is that the opportunity to earnmoney based on developed country prices but spend it(through remittances) based on developing countryprices is a major benefit from migration. The sameadjustment from the perspective of the migrant isdiscussed in the modeling exercise in chapter 2.5. See the discussion of remittances and smoothingof household consumption in chapter 5.6. By the late 1990s, public employment servicesalready played an insignificant role in the recruitmentof foreign workers, except where migration wascovered by bilateral labor agreements (ILO 1997). Forexample, nine out of ten workers sent from Asiahave used private recruiters (Abella 1997); and forRomania, most jobs in countries with which the governmenthas not secured bilateral agreements are foundby private intermediaries (Diminescu 2004).7. Hugo (2004) describes how work contractorsare the primary source of information for potentialmigrants from Indonesia, and relates this to the highlevels of exploitation of Indonesian contract workerscompared with workers from other countries.8. Support services provided through Philippinelabor attachés have provided critical legal counselingand protection (Moreno-Fontes Chammartin 2005).9. The funds operated by the Philippines, Pakistanand Sri Lanka provide scholarships, legal aid indestination countries, insurance against death anddisability, and loans for predeparture costs, housing,and self-employment. The administration and deliverycan often be difficult, particularly on insurance (Tan2004), and some emigrants may resent the mandatorynature of the schemes (Abella 1997).10. Also regarding Mexico, Mora and Taylor(2005) find that the presence of a family member in theUnited States increases by 7 percent the probabilitythat an individual will migrate, while McKenzie (2005)shows that larger migration networks increase theprobability of other community members migrating.11. An irregular migrant in this context is definedas any person entering, residing, and working in a countrywithout proper documentation of their legal statusin that country, or any person who has committed acrime or breach of immigration law in that country andtherefore is not entitled to remain in that country.12. However, Cornelius (2001) notes that the shareof irregular migrants who settle permanently in the76

THE POLICY CHALLENGES OF MIGRATIONUnited States has increased—a trend accelerated bytighter border enforcement adopted in the mid-1990s.13. Protocol to Prevent, Suppress and Punish Traffickingin Persons, Especially Women and Children;and Protocol against the Smuggling of Migrants byLand, Sea and Air, 2000. These supplement the Conventionagainst Transnational Crime, 2000.14. Dedicated government offices such as thePhilippines Overseas Workers Administration, unionssuch as the seamen’s union in the Philippines, and nongovernmentalorganizations (NGOs) can help familiesand communities make adjustments when family membersmigrate.15. These data are described in Docquier andMarfouk (2004), which relies on census data (plusextensive estimations), and thus undercounts irregularmigrants, who are mostly low skilled. The data aretaken largely from industrial countries, so that lowskilledmigration to other developing countries, as wellas high-income countries in the Middle East and Asia,is not reflected (which, for example, reduces the ratioof low-skilled emigrants from South Asia).16. Similarly, mine labor recruiting in South Africaincreased wages in the plantation sectors in bothMalawi and Mozambique, which ultimately resulted inthe curtailed permission to recruit in Malawi in theearly 1970s (Lucas 1987).17. For example, the compulsory repatriation ofworkers to Kerala following the Gulf conflict in1990–91 threw Kerala into a fairly sharp recession(Lucas 2004b).18. This is not invariably true, for example, if thedeparture of one household member leaves his or herdependents impoverished. In general, family income islikely to rise with emigration, but cases of real hardshipcaused by emigration do exist.19. This calculation is based on income that includesremittances from the emigrant, so the number ofSri Lankan households with a family member abroadthat were poor prior to migration is probably larger.20. Much of the data on high-skilled migrationrefers to individuals who have some tertiary education,although other kinds of qualifications (electrician,plumber, ability to handle sophisticated machinery) areof economic interest.21. See Lowell (2001) for a list of programs toattract high-skilled workers.22. For example, forgone income tax revenuesassociated with Indian-born residents of the UnitedStates may be equal to one-third of current individualincome tax receipts in India (Desai, Kapur, andMcHale 2001), although this is a very low share oftotal government revenues.23. Such countries benefit highly from remittances.Remittances to countries with populations of less than1.5 million totaled about 6 percent of gross nationalincome, compared with an average of 1.7 percent forall developing countries.24. Available data do not distinguish émigréseducated at home from those educated abroad, an issueof growing importance as education is increasinglymarketed to the developing world by high-incomecountries.25. This theory is developed in Mountford (1997),Chau and Stark (1999), Stark (2003), and Drinkwaterand others (2002).26. These data do not include high-skilled emigrantsto other developing countries, which may be animportant issue for many developing countries.27. Microeconomic evidence tends to find thateducation is associated with higher earnings (Mincer1991). After some considerable debate, recent articlesfind that years of schooling have a positive impact onproductivity growth (de la Fuente and Domenech2002), and that the quality of education (as measured,for example, by pupil-student ratios or the dropoutrate) may matter more than the quantity (Barro andLee 2000).28. Clemens (2005) presents an alternative view,arguing that health systems in Africa are not greatlyweakened by emigration because the option to emigrateencourages entry into the medical field.29. Institutions exist in countries of origin thattrain workers for external labor markets, for examplesome nursing schools in the Philippines and a medicalschool in Budapest that teaches in German. However,these schools do not receive funds from potential countriesof destination (World Bank 2004).30. Data are based on a survey carried out for abackground paper, available on request.31. Workers who stayed in Albania had higherqualityskills than returnees (De Coulon and Piracha2002), and returnees from Sweden were found to beless successful economically than emigrants who stayed(Edin, LaLonde, and Åslund 2000). See also Hugo(2002) on re-emigration from Australia and Constantand Massey (2003) on Germany.32. See Pang, Lansing, and Haines 2002 onThailand, Lucas (2004a) on China, and the IOM officein Tunis on Tunisia.33. The High-Level Committee on the IndianDiaspora (2001) notes the role of expatriates in attractingR&D investments from Intel, Oracle, TexasInstruments, Sun Microsystems, and IBM.34. Recognizing benefits earned abroad mayreduce costs to the destination-country government, asmany such services are likely to be less expensive in developingcountries. The cost implications have someuncertainty, as some migrants will choose to returneven without portability of benefits.77

GLOBAL ECONOMIC PROSPECTS 200635. For example, a program between Germany andEritrea disbursed only 65 loans from 1993 to 1997,while a project in Italy’s Veneto region trained only 30Albanian immigrants in setting up companies orlaunching joint ventures with local companies in theircountry of origin.36. The Return of Qualified Africans program wasevaluated by the European Commission as contributingto development at micro levels (IOM 2005b).37. The Return for Qualified Afghans has beencriticized for offering low compensation packages andinvolving few individuals despite being expensive torun (Jazayery 2002).38. Australia, Argentina, Canada, and the UnitedStates entered into bilateral labor agreements withcountries of origin in the mid-twentieth century. Thebracero program admitted some five million Mexicanfarm workers to the United States between 1942 and1966. In Europe, Germany and France recruited guestworkers from southern Europe, Turkey, and NorthAfrica after the Second World War until the economicdownturn of the 1970s.39. An example is the agreement that provides forthe acceptance of Vietnamese information technologycredentials in Japan (Vietnam Trade 2005).40. The pilot ended in January 2005, and the outcomesare being evaluated by the Dutch government.41. Regional examples include the Regional Conferenceon Migration (Puebla Process) and Lima Processin the Americas; MIDSA and MIDWA in Africa; and theManila, APC, and Bali Processes in Asia. Inter-regionalprocesses include the “5 plus 5” (a migration dialogueestablished in 2002 between southern Europe—France,Italy, Malta, Portugal, Spain—and the Maghrebgroup—Algeria, Libya, Mauritania, Morocco, andTunisia. Global consultation forums include the UNGlobal Commission on International Migration, theBerne Initiative, IOM’s International Dialogue on Migration,and ILO’s International Labor Conference.ReferencesAbella, M. 1997. Sending Workers Abroad: A Manualfor Low- and Middle-Income Countries. Geneva:International Labour Office.———. 2004. “The Role of Recruiters in LaborMigration.” In International Migration:Prospects and Policies in a Global Market, ed.Douglas S. Massey and J. Edward Taylor. Oxford:Oxford University Press.Adams, Richard. 2004. “Remittances and Poverty inGuatemala.” Policy Research Working Paper3418. World Bank, Washington, DC.———. 2005. “Remittances and Poverty in Ghana.”Mimeograph. World Bank, Washington, DC.Adams, Richard, and John Page. 2003. “InternationalMigration, Remittances, and Poverty in DevelopingCountries.” Policy Research Working Paper3179. World Bank, Washington, DC.Afsar, Rita. 2003. “Internal Migration and the DevelopmentNexus: The Case of Bangladesh.”Bangladesh Institute of Development Studies,Dhaka.———. 2005. “Internal Migration and Pro-PoorPolicy.” In Migration, Development, and PovertyReduction in Asia, ed. Gervais Appave and FrankLaczko. Geneva: International Organization forMigration.Agrawal, Ajay, Iain Cockburn, and John McHale.2003. “Gone But Not Forgotten: Labor Flows,Knowledge Spillovers, and Enduring Social Capital.”Working Paper 9950. National Bureau ofEconomic Research, Cambridge, MA.Aleinikoff, T. A., and Klusmeyer, D. 2002. CitizenshipPolicies for an Age of Migration. Washington,DC: Carnegie Endowment for International Peace.Barré, Rémi, Valéria Hernandez, Jean-Baptiste Meyer,and Dominique Vinck. 2003. Diasporas scientifiques:Comment les pays en développementpeuvent-ils tirer parti de leurs chercheurs et deleurs ingénieurs? (Scientific diasporas: how candeveloping countries benefit from their expatriatescientists and engineers?) Paris: IRD Editions.Barro, R., and J. W. Lee. 2000. “International Data onEducational Attainments—Updates and Implications.”Working Paper 7911. National Bureau ofEconomic Research, Cambridge, MA.Bartel, A. P. 1989. “Where Do the New U.S. ImmigrantsLive?” Journal of Labor Economics 7(4):371–91.Battistella, Graziano, and Ma. Cecilia G. Conaco.1996. “Impact of Migration on the Children LeftBehind.” Asian Migrant 9(3): 86–91.Beine, Michel, Frederic Docquier, and Hillel Rapoport.2001. “Brain Drain and Economic Growth:Theory and Evidence.” Journal of DevelopmentEconomics 64: 275–89.Bhagwati, Jagdish. 1976. “The Brain Drain.”International Social Science Journal 28: 691–729.Black, R., 2004. “Migration and Pro-Poor Policy inAfrica.” Working Paper C6. Sussex Centre forMigration Research, University of Sussex.Black, Richard, Savinna Ammassari, Shannon Mouillesseaux,and Radha Rajkotia. 2004. “Migrationand Pro Poor Policy in West Africa.” WorkingPaper C8. Sussex Centre for Migration Research,University of Sussex.Borjas, George. J. 1987. “Self-Selection and Earningsof Immigrants.” American Economic Review77(4): 531–53.78

THE POLICY CHALLENGES OF MIGRATIONBorjas, George J., and Bernt Bratsberg. 1994. “WhoLeaves? The Outmigration of the Foreign Born.”National Bureau of Economic Research WorkingPaper 4913. Cambridge, MA.Brown, Mercy. 2000. “Using the Intellectual Diasporato Reverse the Brain Drain: Some Useful Examples.”Paper presented at the Regional Conferenceon Brain Drain and Capacity Building in Africa.Addis Ababa, February 22–24.Bryant, John. 2005. “Children of International Migrantsin Indonesia, Thailand, and the Philippines:A Review of Evidence and Policies.” WorkingPaper 2005-05. UNICEF Innocenti ResearchCenter, Florence.Carter, Thomas J. 1999. “Illegal Immigration in anEfficiency Wage Model.” Journal of InternationalEconomics 49: 385–401.Cervantes, M., and D. Guellec. 2002. “The BrainDrain: Old Myths, New Realities.” Observer,May 7. www.oecdobsever.org/.Chanda, Rupa. 2001. “Trade in Health Services.”Paper prepared for the Working Group on Healthand International Economy of the Commissionon Macroeconomics and Health, World HealthOrganization, Geneva.Chau, Nancy H., and Oded Stark. 1999. “Migrationunder Asymmetric Information and HumanCapital Formation.” Review of InternationalEconomics 7(3): 455–83.Chellaraj, Gnanaraj, Keith Maskus, and AadityaMattoo. 2005. “The Contribution of Skilled Immigrationand International Graduate Studentsto U.S. Innovation.” In International Migration,Remittances, and Development, ed. CaglarOzden and Maurice Schiff. Washington, DC:World Bank.Chiswick, Barry R. 1988. “Illegal Immigration andImmigration Control.” Journal of Economic Perspectives2(3): 101–15.———. 2000. “The Economics of Illegal Migrationfor the Host Economy.” Paper presented at theNational Association for Business EconomicsAnnual Meeting. September.Clemens, Michael. 2005. “Do No Harm—Is the Emigrationof Health Professionals Bad for Africa?”Prepared for the G-20 Workshop on DemographicChallenges and Migration. Sydney,Australia, August.Commonwealth Secretariat. 2005. “A ManagedMigration Program for Teachers and Nurses.”London.Constant, Amelie, and Douglas S. Massey. 2003. “Self-Selection, Earnings, and Out-Migration: A LongitudinalStudy of Immigrants to Germany.” PopulationEconomics 16: 631–53. November.Cornelius, Wayne A. 2001. “Death at the Border: TheEfficacy and ‘Unintended’ Consequences of U.S.Immigration Control Policy, 1993–2000.”Working Paper 27, Center for ComparativeImmigration Studies, University of California–San Diego.Daveri, Francesco, and Riccardo Faini. 1999. “WhereDo Migrants Go?” Oxford Economic Papers 51:595–622. oep.oxfordjournals.org/cgi/content/abstract/51/4/595.De Coulon, Augustin, and Matloob Piracha. 2002.“Self-Selection and the Performance of ReturnMigrants: The Case of Albania.” DiscussionPaper 0211. Department of Economics, Universityof Kent, Canterbury.De la Fuente, A., and R. Domenech. 2002. “HumanCapital in Growth Regressions: How Much DifferenceDoes Data Quality Make? An Update andFurther Results.” Unpublished paper. UniversidadAutonoma, Barcelona.Desai, Mihir A., Devesh Kapur, and John McHale.2001. “The Fiscal Impact of the Brain Drain:Indian Emigration to the U.S.” Prepared for theThird Annual NBER-NCAER conference, December17–18. Harvard University and NationalBureau for Economic Research, Cambridge,MA.Diminescu, Dana. 2004. “Assessment and Evaluationof Bilateral Labour Agreements Signed byRomania.” In Migration for Employment: BilateralAgreements at a Crossroads. Paris: Organisationfor Economic Co-operation and Development.Docquier, Frederic, and Abdeslam Marfouk. 2004.“Measuring the International Mobility of SkilledWorkers (1990–2000).” Policy Research WorkingPaper 3381. Development Research Group,World Bank, Washington, DC.Docquier, Frederic, and Hillel Rapoport. 2004.“Skilled Migration: The Perspective of DevelopingCountries.” Policy Research Working Paper3382. Development Research Group, WorldBank, Washington, DC.Drinkwater, Stephen, Paul Levine, Emanuela Lotti, andJoseph Pearlman. 2002. “The Economic Impactof Migration: A Survey.” Paper prepared forSecond Workshop of the Fifth Framework ProgrammeProject, “European Enlargement: TheImpact of East-West Migration on Growth andEmployment,” December 6–7, Vienna.Edin, Per-Anders, Robert J. LaLonde, and OlofÅslund. 2000. “Emigration of Immigrants andMeasures of Immigrant Assimilation: Evidencefrom Sweden.” Swedish Economic Policy Review7: 163–204. Fall.79

GLOBAL ECONOMIC PROSPECTS 2006Eelens, Frank, and J. D. Speckmann. 1990. “Recruitmentof Labor Migrants for the Middle East:The Sri Lankan Case.” International MigrationReview 24(2): 297–322.Ellerman, David. 2003. “Policy Research on Migrationand Development.” World Bank Policy ResearchWorking Paper 3117. Washington, DC.Faini, Riccardo. 2005. “Does the Brain Drain BoostGrowth?” Research Program on InternationalMigration and Development. DECRG. Mimeo.World Bank. Washington, DC.Freeman, Richard B., and Remco H. Oostendorp.2000. “Wages Around the World: Pay acrossOccupations and Countries.” Working Paper8058. National Bureau of Economic Research,Cambridge, MA.Gazdar, Haris. 2003, “A Review of Migration Issues inPakistan.” Paper presented at the Regional Conferenceon Migration, Development and Pro-PoorChoices in Asia, June 22–24, Dhaka.Gevorkyan, Alexandr V., and David A. Grigorian.2003. “Armenia and Its Diaspora: Is There Scopefor a Stronger Economic Link?” Armenian Forum3(2): 1–35.Ghosh, Bimal. 1996. “Economic Migration and theSending Countries.” In The Economics of LaborMigration, ed. Julien van den Broek. Cheltenham,UK: Edward Elgar.Go, S. 2004. “Fighting for the Rights of MigrantWorkers: The Case of the Philippines.” In Migrationfor Employment: Bilateral Agreeements at aCrossroads. Paris. Organisation of Economic Cooperationand Development.Gould, David. 1994. “Immigrant Links to the HomeCountry: Empirical Implications for U.S. BilateralTrade Flows.” Review of Economics and Statistics76: 302–16. May.Gubert, Flore. 2005. “Migrant Remittances and TheirImpact on the Economic Development of SendingCountries: The Case of Africa.” Paper presentedat OECD International Conference on Migration,Remittances, and the Economic Developmentof Sending Countries, February 23–25,Marrakech.Gunatilleke, Godfrey. 1998, “The Role of SocialNetworks and Community Structures in InternationalMigration from Sri Lanka.” In EmigrationDynamics in Developing Countries II: SouthAsia, ed. Reginald Appleyard. Aldershot,England: Ashgate.Hatton, Timothy J., and Jeffrey G. Williamson. 2002.“What Fundamentals Drive World Migration?”Discussion Paper 3559. Centre for EconomicPolicy Research, London.Head, K., and J. Ries. 1998. “Immigration and TradeCreation: Econometric Evidence from Canada.”Canadian Journal of Economics 31(1).High-Level Committee on the Indian Diaspora. 2001.“The Indian Diaspora.” Ministry of ExternalAffairs, Government of India. December.indiandiaspora.nic.in/contents.htm.Holzmann, Robert, Johannes Koettl, Taras Chernetsky.2005. “Portability Regimes of Pension andHealth Care Benefits for International Migrants:An Analysis of Issues and Good Practices.” SocialProtection Discussion Paper Series 0519. WorldBank, Washington, DC.House of Commons. 2004. “Migration and Development:How to Make Migration Work for PovertyReduction.” Sixth Report of Session 2003–04,HC 79-1. London. July.Hugo, Graeme J. 2002, “Migration Policies Designedto Facilitate the Recruitment of Skilled Workersin Australia.” In International Mobility of theHighly Skilled. Paris: Organisation for EconomicCo-operation and Development.———. 2004. “Information, Exploitation, and Empowerment:The Case of Indonesian ContractWorkers.” Unpublished paper. University ofAdelaide.ILO (International Labour Office). 1997. “Protectingthe Most Vulnerable of Today’s Workers.” InternationalMigration Branch. Geneva.———. 2003a. “Booklet 2: Decision-Making andPreparing for Employment Abroad.” In PreventingDiscrimination, Exploitation and Abuse ofWomen Migrant Workers: An InformationGuide. Geneva: International Labour Office.———. 2003b. “Booklet 4: Working and LivingAbroad.” In Preventing Discrimination, Exploitationand Abuse of Women Migrant Workers:An Information Guide. Geneva: InternationalLabour Office.IOM (International Organization for Migration).2003. Labour Migration in Asia. Geneva.———. 2005a. Migration, Development, and PovertyReduction in Asia. Geneva.———. 2005b. World Migration 2005. Geneva.Jandl, Michael. 2003. “Estimates on the Numbers ofIllegal and Smuggled Immigrants in Europe.” Presentationat the Eighth International MetropolisConference of the Inernational Centre for MigrationPolicy Development, September 17, Vienna.———. 2004. “The Estimation of Illegal Migration inEurope.” Studi Emigrazione/Migration Studies61(153): 141–55. March.Jaeger, David A. 2000. “Local Labor Markets, AdmissionCategories, and Immigrant Location80

THE POLICY CHALLENGES OF MIGRATIONChoice.” Unpublished paper. Hunter College,New York.Jazayery, Leila. 2002. “The Migration-DevelopmentNexus: Afghanistan Case Study.” InternationalMigration 40(5): 231–52.Johnson, Brett, and Santiago Sedaca. 2004. “Diasporas,Émigrés, and Development: EconomicLinkages and Programmatic Responses.” Studyconducted for the Trade Enhancement ServiceSector (TESS) Project under contract with the U.S.Agency for International Development, CaranaCorporation, Washington, DC. January.Karemera, David, Victor I. Oguledo, and Bobby Davis.2000. “A Gravity Model Analysis of InternationalMigration to North America.” AppliedEconomics 32(13): 1745–55.Kanbur, Ravi, and Hillel Rapoport. 2005. “MigrationSelectivity and the Evolution of SpatialInequality.” Journal of Economic Geography 5:1–15.Lee, J. 2005. “Human Trafficking in East Asia: CurrentTrends, Data Collection, and KnowledgeGap.” In Data and Research on Human Trafficking:A Global Survey. Geneva. IOM.Long, Larry, C. Jack Tucker, and William L. Urton. 1988.“Migration Distances: An International Comparison.”Demography 25: 633–60. November.Lowell, B. Lindsay. 2001. “Policy Responses to the InternationalMobility of Skilled Labour.” InternationalLabour Office, International MigrationBranch, Geneva.Lowell, B. Lindsay, and Allan Findlay. 2001. “Migrationof Highly Skilled Persons from DevelopingCountries: Impact and Policy Responses. SynthesisReport.” International Labour Office, InternationalMigration Branch, Geneva.Lucas, Robert E.B. 1987. “Emigration to SouthAfrica’s Mines.” American Economic Review 77:313–30. June.———. 2004a. International Migration Regimesand Economic Development. Report from theseminar of the Executive Group on DevelopmentIssues on International Migration Regimes andEconomic Development, May 13, Stockholm.www.egdi.gov.se/seminars6.htm.———. 2004b. “International Migration to the HighIncome Countries: Some Consequences for theSending Countries.” Unpublished paper. BostonUniversity. Economics Department.Magoni, Raphaele. 2004. “France.” In InternationalMigration and Relations with Third Countries:European and U.S. Approaches, eds. Jan Niessenand Yongmi Schibel. Brussels: Migration PolicyGroup.Mahmud, Wahiduddin. 1989. “The Impact of OverseasLabour Migration on the Bangladesh Economy:A Macro-Economic Perspective.” In To theGulf and Back: Studies on the Economic Impactof Asian Labour Migration, ed. Rashid Amjad.New Delhi: ILO-ARTEP.Majid, Nomaan. 2000, “Pakistan: Employment, Outputand Productivity.” Issues in DevelopmentDiscussion Paper 33. International Labour Office,Geneva.Malmberg, Gunnar. 1997, “Time and Space in InternationalMigration.” In International Migration, Immobilityand Development: Multidisciplinary Perspectives,ed. Tomas Hammar, Grete Brochmann,Kristof Tamas, and Thomas Faist. Oxford: Berg.Martin, Philip. 2003. “Managing Labor Migration:Temporary Worker Programs for the Twenty-FirstCentury.” International Institute for LabourStudies, Geneva.———. 2005. “Merchants of Labor: Agents of EvolvingMigration Infrastructure.” Discussion Paper158/2005. International Institute for LabourStudies, Geneva.Massey, Douglas S., Joaquin Arango, Graeme Hugo,Ali Kouaouci, Adela Pellegrino, and J. EdwardTaylor. 1993. “Theories of International Migration:A Review and Appraisal.” Population andDevelopment Review 19(3): 431–66.———. 1998. Worlds in Motion: Understanding InternationalMigration at the End of the Millennium.Oxford: Clarendon.McKenzie, David J. 2005. “Beyond Remittances: TheEffects of Migration on Mexican Households.” InInternational Migration, Remittances, and Development,ed. Caglar Ozden and Maurice Schiff.New York: Palgrave Macmillan.McMahon, Walter M. 1999. Education and Development:Measuring the Social Benefits. Oxford:Oxford University Press.Meek, Miki. 2003. “Life and Death on the SouthwestBorder.” National Geographic. November.Meyer, Jean-Baptiste, and Mercy Brown. 1999. “ScientificDiasporas: A New Approach to the BrainDrain.” Paper prepared for the UNESCO-ICSUWorld Conference on Science, June 26–July 1,Budapest.Mincer, Jacob. 1991. “Human Capital, Technology,and the Wage Structure: What Do Time SeriesShow?” Working Paper 3581. National Bureau ofEconomic Research, Cambridge, MA.Misra, Neha, and Ruth Rosenberg. 2003. “MigrantWorkers.” In Trafficking of Women and Childrenin Indonesia, ed. Ruth Rosenberg. Jakarta: InternationalCatholic Migration Commission81

GLOBAL ECONOMIC PROSPECTS 2006(ICMC) and American Center for InternationalLabor Solidarity (Solidarity Center).Mora, Jorge, and J. Edward Taylor. 2005. “Determinantsof Migration, Destination and SectorChoice: Disentangling Individual, Household, andCommunity Effects.” In International Migration,Remittances, and Development, ed. Caglar Ozdenand Maurice Schiff. New York: PalgraveMacmillan.Moreno-Fontes Chammartin, G. 2005. “DomesticWorkers: Little Protection for the Underpaid.”Migration Information Source. April. www.migrationinformation.org/.Mountford, Andrew. 1997. “Can a Brain Drain BeGood for Growth in the Source Economy?”Journal of Development Economics 53(2):287–303.Munshi, Kaivan. 2003. “Networks in the ModernEconomy: Mexican Migrants in the United StatesLabor Market.” Quarterly Journal of Economics118(2):OECD (Organisation for Economic Co-operation andDevelopment). 2004. Trends in InternationalMigration. Paris.———. 2005. Trends in International Migration.Paris.Orozco, Manuel. 2003. “Worker Remittances,Transnationalism, and Development.” Inter-American Dialogue, Washington, DC.Pan, W. 2004. “Yikao nongmin gaosu tuijing chengshihua”(Making farmers accelerate urbanization).Zhanlue Yu Guanli (Strategy and Management)6(2).Pang, Tikki, Mary Ann Lansang, and Andy Haines.2002. “Brain Drain and Health Professionals.”British Medical Journal 324(7336): 499–500.March 2.Passel, J. S., R. Capps, and M. E. Fix. 2004. UndocumentedImmigrants: Facts and Figures. UrbanInstitute, Washington, DC. www.urban.org/url.cfm?ID1000587.Rauch, James E., and Victor Trindade. 1999. “EthnicChinese Networks in International Trade.”NBER Working Paper 7189, National Bureau ofEconomic Research, Cambridge, MA. June.Rivera, Jose Jorge Mora. 2005. “The Impact of Migrationand Remittances on Distribution and Sourcesof Income: The Mexican Rural Case.” Paperprepared for the meeting of the United NationsExpert Group on International Migration andDevelopment, July 6–8, New York.Saith, Ashwani. 1997. “Emigration Pressures andStructural Change: Case of the Philippines.” InternationalMigration Papers 19. InternationalLabour Office, Geneva.Salt, John, and Jeremy Stein. 1997. “Migration as aBusiness: The Case of Trafficking.” InternationalMigration 35(4).Saxenian, AnnaLee. 2000. “Brain Drain or Brain Circulation?The Silicon Valley–Asia Connection.”Lecture in the Modern Asia Series, HarvardUniversity Asia Center, September 29.———. 2002. Local and Global Networks of ImmigrantProfessionals in Silicon Valley. San Francisco:Public Policy Institute of California.Schiff, Maurice. 2005. “Brain Gain: Its Size and Impacton Welfare and Growth Have Been GreatlyExaggerated.” In Caglar Ozden and MauriceSchiff (eds.) International Migration, Remittancesand the Brain Drain. New York: PalgraveMacmillan.Shah, Nasra M. 1998, “The Role of Social Networksamong South Asian Male Migrants in Kuwait.”In Emigration Dynamics in Developing CountriesII: South Asia, ed. Reginald Appleyard. Aldershot,England: Ashgate.Siddiqui, T. 2003. “Migration as a Livelihood Strategyof the Poor: The Bangladesh Case.” Paper presentedat the Regional Conference on Migration,Development, and Pro-Poor Policy Choices inAsia, June 22–24, Dhaka, Bangladesh.Solimano, Andres. 2002. “International Migration,Globalization and Development: Main Issues.”Unpublished paper. UN Economic Commission forLatin America and the Caribbean, Santiago, Chile.Spaan, Ernst. 1994. “Taikongs and Calos: The Role ofMiddlemen and Brokers in Javanese InternationalMigration.” International Migration Review28(1): 93–113.Stalker. 1994. “The Work of Strangers: A Survey ofInternational Labour Migration.” InternationalLabour Office, Geneva.Stark, Oded. 2003. “Rethinking the Brain Drain.”World Development. 32(1): 15–22.Tan, E. 2004. “Welfare Funds for Migrant Workers:A Comparative Study of Pakistan, Philippines,and Sri Lanka.” International Organization forMigration, Geneva.UNESCAP (UN Economic and Social Commission forAsia and the Pacific). 2003. “Migration Patternsand Policies in the Asian and Pacific Region.”Asian Population Studies Series 160. United Nations,New York.UN (United Nations). 2004. World Economic andSocial Survey. New York.U.S. Department of State. 2004. Trafficking in PersonsReport. Washington, DC.Vayrynen, Raimo. 2003. “Illegal Immigration, HumanTrafficking, and Organized Crime.” DiscussionPaper 2003/72. World Institute for Development82

THE POLICY CHALLENGES OF MIGRATIONEconomics Research, United Nations University,Helsinki.Vietnam Trade. 2005. “Japan Accepts Credentials ofVietnam’s IT Engineers.” Vietnam Trade. January20. www.info.vn/index.php?news1&id9&nid5783&langen&start\20&archive.Vincent-Lancrin, S. 2004. “Building Capacity throughCross-Border Tertiary Education.” Paper preparedfor the conference of the UNESCO/OECDAustralia Forum on Trade in Educational Services,“Bridging the Divide: Building Capacity forPost-Secondary Education through Cross-BorderProvision,” October 11–12, Sydney. www.oecd.org/dataoecd/43/25/33784331.pdf.Wickramasekera, Piyasiri. 2002. “Asian LabourMigration: Issues and Challenges in an Era ofGlobalization.” International Labour Office, InternationalMigration Program, Geneva.Winters, Alan. 2005. “Demographic Transition andthe Temporary Mobility of Labor.” Paper preparedfor the G-20 Workshop on DemographicChallenges and Migration, August, Sydney.World Bank. 2005. Global Economic Prospects.Washington, DC.Xiang, B. 2003. “Emigration from China: A SendingCountry Perspective.” International Organizationfor Migration, Geneva.Yusuf, Shahid. 2001. “Globalization and the Challengefor Developing Countries.” Background paper forWorld Development Report 1999/2000. WorldBank, Washington, DC.83

4Trends, Determinants, andMacroeconomic Effectsof RemittancesChapter 3 reviewed the trends, opportunities,and policy challenges associated with internationalmigration. It also introduced theeconomic importance of the funds that internationalmigrants send back to their countryof origin. In recent years, those funds haveemerged as a major source of external financingin developing countries. Although there isno universal agreement yet on how to measureinternational migrants’ remittances to developingcountries, a comprehensive measure ofcertain officially recorded flows—workers’remittances, compensation of employees, andmigrant transfers—produced an estimate of$167 billion for 2005, up from $160 billion in2004. Given measurement uncertainties,notably the unknown extent of unrecordedflows through formal and informal channels,the true size of remittance flows may be muchhigher—perhaps 50 percent or more. Becauseof their volume and their potential to reducepoverty, remittances are attracting growingattention from policymakers at the highestlevels in both developed and developingcountries. 1This chapter and chapters 5 and 6 considerremittances from several angles. The organizingframework is driven by three items on theinternational policy agenda: (1) understandingthe true size and trends in remittance flows todeveloping countries, as well as their macroeconomicimpact; (2) evaluating the impact ofremittances on the households that receivethem; and (3) designing policies to reduce thetransaction costs of remittances, strengthenthe formal financial infrastructure supportingremittances, and leverage remittances to improveaccess to financial services in recipienteconomies.Officially recorded remittance estimatesmay significantly underestimate the real magnitudeof remittances. Model-based estimatesand household surveys suggest that informalflows could add at least 50 percent to the officialestimate, with significant regional andcountry variation. The true size of remittanceflows could be even larger, in view of substantialunderrecording of flows through formalchannels.Despite the prominence given to remittancesfrom developed countries, South–Southremittance flows make up 30–45 percent oftotal remittances received by developing countries,reflecting the fact that over half ofmigrants from developing countries migrate toother developing countries. Remittance flowsto poor countries originate largely in themiddle-income developing countries.Recorded remittance flows have surged inrecent years, driven by a combination offactors—among them better data collection,reflecting greater awareness of the developmentpotential of remittances, as well as concernsabout money laundering and terroristfinancing; lower costs and wider networks inthe industry that supports remittance; and85

GLOBAL ECONOMIC PROSPECTS 2006growth in the number of migrants and theirincomes. Government policies to improvebanking access and the technology of moneytransfers have also helped increase the flow ofremittances and promote their transferthrough formal channels.Efforts to encourage remittances, however,sometimes generate unwanted effects. Taxincentives may attract remittance inflows, forexample, but they also create opportunitiesfor tax evasion. Likewise, matching-fundprograms for migrant associations may channelcollective remittances to developmentprojects, but in so doing they may divert fundsfrom other local funding priorities.For some recipient countries, remittancesare large enough to have broader macroeconomicimplications. By generating a steadystream of foreign-exchange earnings, theycan improve a country’s creditworthiness forexternal borrowing, and through innovativefinancing mechanisms (such as securitization),they can expand access to capital and lowerborrowing costs. While large and sustainedremittance inflows can contribute to currencyappreciation and so affect the production ofcost-sensitive tradables (such as laborintensivemanufactures), this outcome maybe less severe than it is in the case of natural-resourceearnings (since remittances are distributedmore widely and may avoid exacerbatingthe strains on institutional capacity that areoften associated with natural-resourcebooms). Furthermore, the “Dutch disease”effects of remittances are of relatively minorconcern insofar as remittances grow graduallyover long periods. Remittances have a largepositive effect on national income in many developingcountries, and there is compellingevidence that they contribute significantly topoverty reduction (see chapter 5). Althoughthe evidence on the effect of remittances onlong-term growth remains inconclusive, ineconomies where the financial system is underdeveloped,remittances appear to alleviatecredit constraints and may stimulate economicgrowth.The plan of this chapter is as follows. In thenext section, trends in remittance flows to developingcountries are presented along with arange of estimates for their true size—that is,with informal flows included. We identify themajor sending and receiving countries, includingthose in the South. In the following section,we examine the factors affecting remittanceflows, including the prospects for futureremittance growth, and policies and regulationsin source and destination countries thataffect the cost of remittances. In the final section,we consider the macroeconomic effectsof remittances, including the effects on stability,country creditworthiness, internationalcapital-market access, the real exchange rate,and competitiveness.Remittance data and trendsThe quality and coverage of data on remittancesleave much to be desired. First, thereis no consensus on the boundaries of the phenomenonunder study. Should only workers’ remittancesbe counted, or should we includecompensation of employees and migrant transfers?(See annex 4A.1 for more details on thesenomenclatural disputes.) Second, in severalcountries, many types of formal remittanceflows go unrecorded, due to weaknesses in datacollection (related to both definitions and coverage).2 Reporting of “small” remittance transactionsmade through formal channels is notmandatory in most countries, 3 and remittancessent through post offices, exchange bureaus,and other agents of money transfer operators(MTOs) are often not reflected in official statistics(de Luna Martinez 2005). Third, flowsthrough informal channels (such as unregulatedmoney transfer firms or family and friends whocarry remittances) are rarely captured. Finally,remittances are often misclassified as exportrevenue, tourism receipts, nonresident deposits,or even foreign direct investment (FDI). Improvingthe quality of remittance statistics is thefocus of ongoing cooperative internationalefforts (see box 4.1).86

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESBox 4.1 International working group on improvingdata on remittancesAt its meeting in Sea Island in April 2004, the G-8called upon the international financial institutions(IFIs) to lead a global effort to improve remittancestatistics. In January 2005, the World Bankand International Monetary Fund (IMF) held aninternational meeting of data users and compilerswho agreed that balance-of-payments statistics werethe appropriate framework for collecting, reporting,and improving official statistics on remittances; thatbalance-of-payments concepts and definitions relatingto remittances should be reviewed; and thatimproved guidance for collecting and compilingremittance statistics, including through the use ofhousehold surveys, was needed. Participants at theinternational meeting also agreed that improvementsto relevant statistical concepts and definitions shouldbe discussed in a Technical Sub-Group on the Movementof Persons (TSG), chaired by the UN StatisticsDivision with membership from central banks andnational and international statistical agencies.The TSG recommended that the “workers’ remittances”item in the balance of payments be replacedby “personal transfers.” The new term would coverall current transfers in cash or in kind made orreceived by resident households to or from othernonresident households. It went on to recommendthat a new aggregate, “personal remittances,” bereported in the standard balance-of-payments presentationas a memorandum item. It was proposed thatpersonal remittances comprise current and capitaltransfers in cash or in kind, made or received, byresident households to or from nonresident households,and “net” compensation of employees frompersons working abroad for short periods of time(less than one year).The TSG also recommended that institutionalremittances—those involving government, corporations,and nonprofit institutions servinghouseholds—should also be reported as a new memorandumitem in the standard presentation ofbalance-of-payments statistics. That item would leadto a further memorandum item, “total remittances,”the sum of personal and institutional remittances.Because the concepts of personal transfers andremittances are based on the concept of residencerather than migration status, the TSG recommendedthat the concept of “migrant” be replaced by theconcept of “residence” in the balance-of paymentsframework.Reporting of bilateral remittance flowsis not currently required in the balance of payments,but the recommendation of the TSG is that flows toand from major partner countries be identified.The TSG is expected to produce a final report inspring 2006. aa A concurrent effort to improve remittance statistics isunder way at the Center for Latin America Monetary Studies(CEMLA) with support from the Multilateral Investment Fundof the Inter-American Development Bank and technical advicefrom an international advisory council that includes the IMFand World Bank.Officially recorded remittance flowsare surgingIn this report (as in past editions of the WorldBank’s annual Global Development Financeand the IMF’s 2005 World Economic Outlook),migrant remittances are calculated asthe sum of workers’ remittances, compensationof employees, and migrant transfers (seeannex 4A.1). Thus defined, remittances receivedby developing countries, estimatedusing officially recorded data, rose to$167 billion in 2005, up 73 percent from2001 (table 4.1). More than half of that increaseoccurred in China, India, and Mexico.Low-income countries, led by India, registeredan increase of $18 billion during this period(box 4.2). Of 34 developing countries that receivedremittances in excess of $1 billion in2004, 26 countries registered more than30 percent growth during 2001–4: Algeria87

GLOBAL ECONOMIC PROSPECTS 2006Table 4.1 Workers’ remittances to developing countries, 1990–2005$ billionsChange (%)1990 1995 2000 2001 2002 2003 2004e 2005e 2005–2001Developing countries 31.2 57.8 85.6 96.5 113.4 142.1 160.4 166.9 73Lower middle income 13.9 30.0 42.6 47.4 57.3 72.5 83.5 88.0 86Upper middle income 9.1 14.5 20.0 22.3 23.0 27.8 33.0 33.8 52Low income 8.1 13.3 22.8 26.8 33.1 41.8 43.9 45 68Latin America and the 5.8 13.4 20.1 24.4 28.1 34.8 40.7 42.4 74CaribbeanSouth Asia 5.6 10.0 17.2 19.2 24.2 31.1 31.4 32.0 67East Asia and the 3.3 9.7 16.7 20.1 27.2 35.8 40.9 43.1 114PacificMiddle East and North 11.4 13.4 13.2 15.1 15.6 18.6 20.3 21.3 41AfricaEurope and Central 3.2 8.1 13.4 13.0 13.3 15.1 19.4 19.9 53AsiaSub-Saharan Africa 1.9 3.2 4.9 4.7 5.2 6.8 7.7 8.1 72World (developing & 68.6 101.6 131.5 147.1 166.2 200.2 225.8 232.3 58industrial)Outward remittances from 6.1 12.5 12.1 14.3 18.7 20.2 24.1 – –developing countriesOutward remittances from 11.2 16.6 15.4 15.1 15.9 14.8 13.6 – –Saudi ArabiaSource: World Bank staff estimates based on IMF BoP Yearbook 2004 and country sources.Note: Remittances are defined as the sum of workers’ remittances, compensation of employees, and migrant transfers (seeannex 4A.1). e estimate.– Data not available.and Guatemala reported more than a tripling ofremittance inflows; Brazil, China, Honduras,Nigeria, Pakistan, and Serbia and Montenegroreported growth in the range of 101–170 percent.(Also, five high-income countries—Austria, Australia, Belgium, Germany, andSpain—reported 45–79 percent growth inremittance inflows during 2001–4.)The growing importance of remittances asa source of foreign exchange is reflected in thefact that remittance growth has outpacedprivate capital flows and official developmentassistance (ODA) over the last decade(table 4.2). Recorded remittance receipts wereequivalent to about 6.7 percent of developingcountries’ imports and 7.5 percent of domesticinvestment. They also were larger than officialflows and private equity (non-FDI) flowsin 2004. Remittances were larger than publicand private capital inflows in 36 developingcountries in 2004 and larger than totalmerchandise exports in Albania, Bosnia andTable 4.2 Recorded remittances havegrown faster than private capital flowsand ODA$ billions1995 2004Workers’ remittances 58 160Foreign direct investment 107 166Private debt and portfolio equity 170 136Official development assistance 59 79Source: World Bank (2005).Herzegovina, Cape Verde, Gaza, Haiti,Jamaica, Kiribati, Lebanon, Nepal, Samoa,Serbia and Montenegro, and Tonga. In another28 countries, they were larger than theearnings from the most important commodityexport; for example, in Mexico, remittancesare larger than FDI; in Sri Lanka, they arelarger than tea exports; and in Morocco, theyare larger than tourism receipts.88

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESBox 4.2India has reported a spectacular increase in remittanceinflows—from $13 billion in 2001 to morethan $20 billion in 2003 (see figure). Several factorsaccount for this remarkable increase. First, the numberof migrants has grown sharply. During the oilboom in the 1970s and 1980s, thousands of lowskilledIndian workers migrated to the Persian Gulfcountries. In the 1990s, migration to Australia,Canada, and the United States, increased significantly,particularly among information technology(IT) workers on temporary work permits. aSecond, the swelling of migrants’ ranks coincidedwith (a) better incentives to send and invest money$ billions25201510The recent surge in remittance flows to IndiaRemittance flows to Indiain India’s growing economy and (b) an easing of theregulations and controls, more flexible exchangerates, and gradual opening of the capital account.The elimination of the black-market premium on therupee and convenient remittance services provided byIndian and international banks have no doubt shiftedsome remittance flows from informal hawala channelsto formal channels.Third, nonresident Indians have also respondedto several attractive deposit schemes and bondsoffered by the government of India. These offerattractive interest rates and an appreciating rupee.While nonresident deposits are conceptually differentfrom remittances (they are a liability item inthe capital account), evidence suggests that a largepart of such deposits is converted to local currency.For example, for the Resurgent India Bond thatmatured in 2003, most of the redemption valuestayed in India to meet various local currencyneeds of the nonresident depositors and their families.Nevertheless, remittances in the form of foreign-currencydeposits can be speculative and mayreverse in the event of deterioration in theinvestment sentiment.5019871988198919901991199219931994199519961997199819992000200120022003a In particular, migration to United States doubled duringthe 1990s. Remittances from United States as a share of totalremittances to India grew from 37 percent in 1997 to 51 percentin 2003.Figure 4.1 identifies the top 20 remittancerecipients in 2004. Among developing countries,China, India, Mexico, and the Philippineswere among the top recipients. Several industrialcountries appear in this list as well.When remittances are calculated in percapita terms or as a share of GDP, a differentpicture emerges. The top 20 recipients inshares of GDP are all developing countries;all receive more than 10 percent of GDP asremittance flows (figure 4.1). Small countries(Bosnia and Herzegovina, Haiti, Lesotho,Moldova, and Tonga) are among the most dependenton remittances.High-income countries are the dominantsource of global remittance flows (figure 4.2).The United States was the largest sourcecountry with nearly $39 billion in outwardremittances in 2004. However, outwardremittances from developing countriesamounted to $24 billion in the same year. 4When expressed in terms of GDP shares, outwardremittances play the largest role in theupper-middle-income developing countries(0.7 percent of GDP in these countriescompared to 0.2–0.4 percent in other developingcountries and in high-income countries;figure 4.2).89

GLOBAL ECONOMIC PROSPECTS 2006Figure 4.1 Top 20 remittance-recipient countries, 2004Billions of dollarsShare of GDPIndiaChinaMexicoFrancePhilippinesSpainBelgiumGermanyUnited KingdomMoroccoSerbiaPakistanBrazilBangladeshEgypt, Arab Rep.PortugalVietnamColombiaUnited StatesNigeria6.96.86.56.44.24.13.93.63.43.33.23.23.23.02.812.711.618.121.721.3TongaMoldovaLesothoHaitiBosnia and HerzegovinaJordanJamaicaSerbia and MontenegroEI SalvadorHondurasPhilippinesDominican Rep.LebanonSamoaTajikistanNicaraguaAlbaniaNepalKiribatiYemen, Rep.31.127.125.824.822.520.417.417.216.215.513.513.212.412.412.111.911.711.711.310.00 5 10 15 20 25 0 15 30 45Source: IMF BoP Yearbook, 2004, and World Bank staff estimates.South–South remittance flowsare considerableOfficial data show that several developingcountries (China, Malaysia, and the RussianFederation) are among the top 20 sources ofremittances. Anecdotally, outward remittancesfrom India and South Africa are also believedto be large, although this is not reflected in theofficial data (Genesis Analytics 2005). TheWorld Bank (2005a) points out a strong associationbetween remittance receipts and thelength of the border shared with more prosperousneighbors. Harrison and others (2004)also report that most remittance flows occurwithin the same region.These factors all point to the conclusionthat South–South remittance flows are substantial.But placing more precision on theseflows is hard to do. First, relatively little isknown about bilateral migration flows—thatis, about how many migrants (or what share)in each receiving country come from eachsending country. Comprehensive global dataare not available, 5 but estimates are that inpoor countries of East Asia, South Asia, andSub-Saharan Africa, more than two-thirds ofemigrants migrate to a country in the same region.In South Asia and Sub-Saharan Africa,most migrate to another developing country.Second, even less is known about how bilateralremittance patterns differ. We do notknow, for example, how much, in total, is sentfrom one country to another, or how remittancepropensities differ across sending andreceiving countries. But by making plausibleassumptions about these flows (in particular,that bilateral remittances are a function of thestock of migrants in the sending country), itis possible to estimate bilateral remittanceflows and to calculate what proportioncomes through South–South links. 6 Using thismethod, we estimate that nearly 30 percent oftotal remittance flows to developing countriesoriginate in other developing countries. This90

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESFigure 4.2 Estimated remittance payment,by country group, 20043.0Low-incomecountries0.3Low-incomecountriesBillions of dollars5.7LowermiddleincomecountriesShare of GDP0.2Lowermiddleincomecountries15.4UppermiddleincomecountriesUppermiddleincomecountries125.3High-incomecountriesHigh-incomecountriesestimate is consistent with the fact that nearlyhalf of the migrant stock from the South migrateto another country in the South. 7One of the challenges of understandingremittance flows is that their characteristics,costs, and channels can vary widely from onebilateral corridor to another (and also widelyfrom different locations within each country).Understanding how remittance corridorsdiffer in the kinds of migrants they serve andtheir means of transferring money is useful forproviders of remittance services as well as policymakers(Hernandez-Coss 2004; Terry2005; and chapter 6 of this volume). Some ofthe major remittance corridors are those thatconnect Canada and the United States to LatinAmerica and Asia; the European Union to0.70.4Eastern Europe, Turkey, and North Africa;and the Persian Gulf to South and SoutheastAsia.Informal remittances are largeRemittances transferred through informaloperators or hand carried by travelers areunlikely to be captured in official statistics,although they may represent a substantialaddition to remittances sent through officialchannels. While it is extremely difficult to estimatethe flows through informal channels,they appear to be large. First, the fact thatrecorded remittances to several countriesthrough formal channels doubled, tripled, orquadrupled between 2001 and 2003 suggeststhat a significant part of the increase is likelyto reflect a shift from informal to formal channelsin response to the tightened regulatoryscrutiny that has occurred since September 11,2001.Second, evidence from household surveyssuggests widespread use of informal remittancechannels (table 4.3). 8 Household surveysalso help identify factors affecting the useof remittance channels. In the presence of awell-developed formal sector, regular remittersand large remitters are unlikely to use theinformal sector. Trust in the financial system isan important prerequisite for a growing bankpresence in the (formal) remittance market.High remittance costs and the presence ofdual exchange rates are two key factorsaffecting the choice of informal remittanceTable 4.3 Choice of remittance channelin selected countries% remittancesFormalInformalDominican Republic 96 4Guatemala 95 5El Salvador 85 15Armenia 62 38Moldova 53 47Bangladesh 46 54Uganda 20 80Source: World Bank staff estimates based on householdsurveys. See also Freund and Spatafora 2005.91

GLOBAL ECONOMIC PROSPECTS 2006channels. 9 If there were no cost advantages tousing informal channels, there would be littleincentive to use them, and remittances couldarguably shift entirely to formal channels.Thus if the costs of formal transfers werereduced to the range reported in the informalsector (2–5 percent), and if official and parallelexchange rates were unified, the resultantincrease in recorded remittance flows couldbe interpreted as an estimate of the size ofinformal flows.Table 4.4 reports the results of an exerciseto estimate the size of the informal remittancesector (see annex 4A.2 for a fuller explanation).Cross-country regression analysisshows that reported remittances are lower,and informal flows higher, in corridors whereremittance costs are higher and where thereare significant black-market premiums overthe official exchange rate. Using the estimatedcoefficients from these regressions, thepredicted increase in officially recorded remittancesis calculated in response to a 2–5 percentdecline in remittance costs and eliminationof the exchange-rate premium. Thesecalculations suggest that the informal remittancesector is at least 50 percent of theTable 4.4 Estimated increase in formalremittances if transaction costs werereduced to 2 to 5 percent and dualexchange rates were eliminatedPercentCross-sectional PanelRegion estimates estimatesAll developing countries 69 54Sub-Saharan Africa 201 122Eastern Europe and Central Asia 151 73East Asia and the Pacific 56 ..South Asia 25 55Middle East and North Africa 165 ..Latin America and the Caribbean 51 99Source: Freund and Spatafora 2005.Note: Results averaged over 1995–2003. See annex 4A.2 fora fuller explanation of the procedures used. In column 3,a reduced form equation is estimated on the basis of theexplanatory variables used in the cost regression reported intable 4A.2.2... Negligible.official sector. 10 They also show significantregional variation. Informal remittances appearto be larger in Sub-Saharan Africa, theMiddle East and North Africa, and Europeand Central Asia than in other regions. 11While the magnitude of the regional estimatesvaries across methods, the relative ranking ofregional effects is more robust.Factors affecting remittance flowsThe surge in remittance flows over the pastfew years reflects a mix of factors, asnoted. In some areas, there have been significantreductions in remittance costs—60 percentin the United States–Mexico corridorsince 1999. On the measurement side, the sizeabledepreciation of the dollar against mostother major currencies (the euro in particular)since 2002 has increased the dollar value ofnondollar remittances over time. 12 Improvementsin data recording by central banks—inresponse to growing recognition of the importanceof remittances by national authorities,and as a result of broader efforts to improvedata quality—have generated sharp increasesin remittance flows in some cases. In addition,heightened security and scrutiny by immigrationand finance authorities in many highincomecountries may have encouraged outwardsurges in remittances, as undocumentedmigrants responded to increased uncertaintyand risk of deportation or other legal action byremitting a larger share of their savings orincome. This factor has reportedly been importantin Pakistan, which recorded a triplingof remittance receipts from 2001 to 2003.The surge in remittances is likely tocontinue in the medium termIn addition to these special factors, powerfuleconomic factors also influence the growthof remittances. Increases in the number ofmigrants will have the greatest and most directimpact, of course, along with compositionalfeatures, such as the mix between temporaryand permanent workers (temporary workers92

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESare believed to remit a larger share of theirincome) and the skill mix (low-skilled workerstend to send a higher proportion of their lowerincomes). Employment opportunities in thehost country affect income, and thereforeremittances, while changes in the cost of livingmay affect the size of the surplus that remittersare able to send.The complex interplay of these factorsmakes assessing the future growth potential ofremittance flows quite difficult. It is plausiblethat in the coming years, official remittanceflows will continue to rise at the 7–8 percentannual rate seen during the 1990s. With boththe supply and demand for migrants growing,migration flows—especially temporarymigration—are likely to continue to be strong.Growing income levels in source countries andrising costs of living in receiving countries,together with the falling costs of remittances,would also imply larger remittances, especiallythrough recorded channels.It is unlikely, however, that the surge inremittance flows seen in some countries since2001 will continue much longer. The shift inflows from informal to formal channels, to theextent that it occurred in response to tightenedscrutiny, is likely to dwindle. (In Pakistan,for example, remittance flows have flattenedsince 2003.) In the more mature UnitedStates–Mexico corridor, where remittancecosts have already fallen drastically (by 60 percentsince 1999), the effect of further costreduction will not be as large as it was fiveyears ago.Some analysts argue that in the more maturemarkets, “remittance decay” may set in, especiallyif temporary or undocumented workersare allowed permanent and legal residence.While it is true that the marginal propensity toremit tends to decline with the length of a migrant’sstay in a host country, and ties with thehome country weaken over time, there is noempirical evidence that the dollar amount ofremittances actually declines in these circumstances.13 On the contrary, the effect of risingincomes of the migrant sender may show up asan increase in remittances over time.Government policies can affectremittance flowsMany sending and receiving governmentsare only now beginning to think about policiesto increase remittance flows and promotetransfers through formal channels. In theremittance-receiving countries, these policiesinclude tax exemptions for remittance income;improved access to banking services byrecipients; incentives to attract investmentsby the diaspora; access to foreign exchangeor lower duties on imports; support for theprojects of migrant associations; and help formigrants in accessing financial systems. In theremittance-source countries, they include policiesaffecting access to banks, access to foreignexchange, support to migrant groups, types ofimmigration regimes, and cooperation withreceiving countries. 14Policies in remittance-receiving countriesTaxes on incoming remittances. Mostremittance-receiving countries today do notimpose taxes on incoming remittances. Theremay be some implicit tax on remittances, however,in the form of a general financial servicestax 15 or on remittances in kind (for example,food, clothing, electronic items, or vehicles).When Vietnam removed its 5 percent tax onremittances in 1997, it found that the flow ofremittances through formal channels increased.Such tax exemptions may well increaseremittance inflows, 16 but they also raisethe possibility of misuse for tax evasion.Travel and customs privileges for returns andimported goods. Many remittance-receivingcountries give preferential treatment tomigrants sending home or bringing with themgoods and equipment. For example, once ayear Tunisians are entitled to import goodsand/or services up to a customs value ofTD1,000 without paying tax, and a privatevehicle, home equipment, and furniture aretax free when they return; Guatemala permitsa once-a-year tax-free remittance of any commodityvalued up to $500. Pakistan, Turkey,Vietnam, and many other countries also offersuch import privileges. 1793

GLOBAL ECONOMIC PROSPECTS 2006Relaxation of exchange and capital controls.Unification of exchange rates and allowingmore banks and financial institutions to undertakeforeign exchange transactions havebeen among the most successful ways of attractingremittances to formal channels andexpanding remittance services in many countries.Also, allowing residents to hold foreigncurrency deposits using remittances fromabroad is believed to have resulted in a largeincrease in formal remittances in many countriesin South Asia and Africa (Siddiqui 2004).India’s liberalization of the exchange rate in1991 has been linked to a decrease in the useof illegal transfer channels to the state ofKerala; and the Philippines found that byabolishing exchange controls it quadrupled itsformal inward remittances in the same year(Buencamino and Gorbunov 2002). Allowingthe market to decide exchange rates in 2002also helped the Bangladesh Bank to curb theinformal hundi business significantly (Siddiqui2004). In 2004, an increase in foreign currencyreserves in Zimbabwe was ascribed, inpart, to the introduction of a new moneytransfer system (Homelink) set up by the governmentto facilitate formal transfers.Allowing domestic banks to operate overseas.Governments have allowed more of theirdomestic financial institutions (includingmicrofinance institutions in some countries) toopen branches and provide services to theirmigrants working in other countries. Thesedomestic banks bring trust and offer remittanceservices at competitive prices. For example,the Groupe Banques Populaires haspicked up 66 percent of total remittances toMorocco by offering low fees, simple procedures,and other nonfinancial services toMoroccans abroad (Amin and Freund 2005).Two small Armenian banks specializing in remittancetransfers, Anelik and Unibank, havecome to dominate the formal transfer systemfor Armenians in parts of Europe; andFonkoze in Haiti has expanded its U.S.–basedclientele in partnership with the City NationalBank of New Jersey. In Bangladesh the dramaticincrease in formal remittances since2001 is, in part, the result of the improved servicesof the banking sector (Siddiqui 2004).ID cards for migrants. Providing identificationcards to migrants (regardless of their legal migrationstatus) to access banking facilities hasalso opened up more opportunities for formalremittance transfer. Mexican immigrants, forexample, can obtain a photo-identificationcard in the form of a matricula consular fromthe Mexican consulates abroad. This card iswidely accepted by commercial banks in theUnited States to open bank accounts (and inmany states, for issuing driving licenses, seebox 6.1). Other Latin American governmentsare discussing similar arrangements for theirnationals in the United States. Most sendingcountries require legal documentation for anybank transaction. Some receiving countriesissue ID cards to expedite domestic servicesfor their emigrants, for example, the Tunisiancarte consulaire for special customs clearance,reduced airfares, and foreign currency bankaccounts in Tunisia. 18Support to hometown associations (HTAs)and matching grants. Providing funds tosupplement or match collective remittancesmade by emigrant groups is another means toengage migrants in the development of homecommunities. With enhanced institutionalcapacities, HTAs could be valuable developmentpartners for governments, the privatesector, and communities, but importantly as acomplement to, not a substitute for, strengthenedfinancial and investment systems on theground (Gubert 2005). A careful evaluation ofsupport to HTAs through matching grantschemes and other means is yet to be undertaken(see box 4.3).Loans/pension schemes and bonds targeted atthe diasporas. These measures can expandopportunities for investment and provide incentivesfor the formal transfer of money fromabroad (see also chapter 6). While investments94

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESin the form of nonresident deposits or diasporabonds are not, strictly speaking, remittances(because they involve the purchase ofassets, rather than transfers to households),they may indirectly encourage remittances.Many countries have successfully issued premiumbonds to their diaspora (forBangladesh, China, Eritrea, India, Israel,Lebanon, Pakistan and the Philippines, seeCarling 2005). Even when investments inthese bonds are in foreign currency terms,after maturity some portion is likely to remainin the country. Such schemes were a major factorbehind the doubling of remittance flows toIndia between 2002 and 2003 (box 4.2).Active policies and institutional arrangementsto support the diaspora. Countries likeMexico and the Philippines with more successfulremittance programs tend to have wellestablished institutional frameworks to train,support, and ensure the welfare of their expatriatesabroad. There is also a broad range ofoutreach activities to assist migrant welfareBox 4.3 Collective remittances through hometownassociations and matching schemesMany migrants are increasingly pooling theirresources and investing collectively indevelopment-related activities in home communities,either through hometown associations (HTAs) orother migrant group schemes. a HTAs are the mostprominent, because of their proliferation among theLatin American and Caribbean diaspora in Canadaand the United States since the late 1990s. Similarassociations exist in France (some 1,000 organizationsde solidarite internationale issues de migrationsor OSIMs), the United Kingdom, and Africa. b Theactivities of HTAs are mixed and poorly documented,but they range from diaspora support inthe host country to community investment projectsin villages in the home country.Collective remittances via HTAs currently accountfor only 1 percent of all remittances in CentralAmerica, but it is estimated that they could rise to3–5 percent in ten years if their management andinstitutional capacity improves (IFAD 2005).Most HTAs tend to be small scale and philanthropicin orientation, and they invest in projects ofno more than $10,000. They have traditionallyfocused on infrastructure and social projects(schools, churches, recreational parks, medical outreachclinics, and household support) and on channelingpost-disaster humanitarian aid (for example,in El Salvador). In Africa, there is evidence that themore sustainable projects tend to facilitate householddistribution of consumer goods (as in general storesor grain banks) or the purchase of farming equipment(Gubert 2005). In Latin America, it is observedthat when at least 30 percent of households in atown receive remittances, HTAs can help improvethe quality of life of households (IOM 2005). But thefocus of HTAs is expanding to include more investmentin economic infrastructure and incomegeneratingprojects managed by the community andlocal NGOs or banks (Orozco 2003).Governments have, on occasion, offeredmatching grants for remittances from diasporagroups or HTAs to attract funding for specificcommunity projects. c The best known of thesematching schemes is Mexico’s 3-for-1 program,started in 1997, under which the local, state, andfederal governments all contribute $1 for every $1of remittances sent to a community for a designateddevelopment project. By 2002, the 3-for-1 programhad established projects totaling $43.5 million, twothirdsof which benefited labor-intensive agriculturaleconomies in four high emigration states (IOM2005). In the period 2002–4, more than 3,000 suchprojects benefited some 1 million inhabitants in23 Mexican states. dEvidence from Mexico suggests, however, thatHTAs have not been very successful. But in somecases (for example, Zacatecas) where HTAs haveexchanged or debated project ideas and investment95

GLOBAL ECONOMIC PROSPECTS 2006Box 4.3(continued)climate issues with the local and state governments,they are believed to have been successful. eOn the positive side, HTA involvement in projectsis argued to ensure that programs are focused oncommunity needs. Resources have gone primarily torural areas, where they have increased the supply ofessential services (health, education, roads, and electricity).Donations by HTAs are often as much as ormore than the municipal budget for public works,particularly in towns with small populations (Orozco2003). HTAs can promote higher standards of transparencyand accountability among local authorities,and higher labor standards.There are obviously limitations on the potentialfor HTAs to serve as conduits for broader developmentprojects. They may not have the best informationabout the needs of the local community, or theymay have different priorities. The capacity of HTAsto scale up or form partnerships is limited by the factthat their members are volunteers, and theirfundraising ability finite. They can also becomedivided and weaken their own advocacy potential(Newland and Patrick 2004). When matching fundscome from fiscally constrained governments, there isalso the problem that they may be diverted fromother—perhaps higher priority—developmentprojects, or from other regions with a greater needfor assistance.a HTAs are grassroots migrant organizations, usually formedaround the interests and needs of a mutual hometown. Theterm has been coined in the United States, where many thousandsof Latin American and Caribbean HTAs have sprung upin the past 15 years or so (Orozco and Welle 2004).b Migrant associations exist in many countries, but aremostly concerned with the conditions of the diaspora and networkingabroad. Some, like the Sierra Leonean Women’s Forumin the United Kingdom, are concerned with immediate survivalneeds (food, clothing) back home (Black and others 2004).c In addition to Mexico, the Salvadoran government partnerswith HTAs in rural development projects in El Salvador. In2001, the federation of HTAs (COMUNIDADES) and the NationalCorporation of Municipalities created the Social Investmentfor Local Development Fund (FISDL) to providematching project funding. In France, the Osims can alsoreceive institutional and financial subsidies from the governmentfor social and economic development projects back home(Magoni 2004).d See “3 por 1. Proyectos Compartidos,” prepared for theseminar “Migracion, remesas y el Programa 3 por 1 para Migrantes,”Secretaria de Desarollo Social, Mexico and IADB,Washington DC, June 2005.e See Gubert 2005, Iskander 2005, and Orozco 2004. Theliterature is not clear on what “success” means in these cases(beyond mere survival of the HTAs).and promote remittances and investment inthe home country, from pre-migration informationand orientation (Philippines), IDs forcustoms and other purposes (Colombia,Tunisia), finance for study (Tunisia), support inlegal and administrative disputes (Morocco),fairs and re-orientation visits for émigrés andtheir families (Colombia, Tunisia), shortenedmilitary service (and payment of fee in lieu,Turkey), hotline for migrant investors(Tunisia), and a diaspora trust fund (Nigeria).Some countries like Bangladesh, Egypt,Eritrea, Pakistan, Philippines, and Thailand(and Mexico and Turkey in the 1960s) havetried to impose mandatory remittance requirementson their émigrés, but with little success.Also, restrictive emigration policies have drivenmigrants into using clandestine remittancechannels. 19Policies in remittance-source countriesOnly a handful of remittance-sending countrieshave proactive remittance-supportingpolicies. Most are noninterventionist or havehad little engagement to date, but this ischanging with the growing appreciation of thesignificance of remittances for development incountries such as Australia, Canada, theUnited States, and most West European states(Ellerman 2003, Carling 2005). USAID has96

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESundertaken extensive research on remittances,as has the United Kingdom’s Department forInternational Development (DFID) and theNorwegian International Peace Institute(PRIO). All propose ways forward for moreproactive policies by sending countries—forexample, to support migrant associations, facilitatelow cost, reduce bureaucratic remittancetransfer, greater competition in the remittancemarket, and inform decision makingby migrants and affected communities.Immigration policies. Policies that affect thesize, type, and tenure of migration flows alsoaffect remittance patterns. A larger migrationstock would in general imply larger remittanceflows to the country of origin. Given the migrationstock, a larger share of temporarymigrants is likely to lead to larger remittances.Also, as discussed above, the ties of migrantsto their home country weaken with the passageof time, causing remittances to decline.Given the personalized nature of remittances,governments are unlikely to have muchsuccess in using remittance policies to steermigration differently. Some countries, likeCanada, France, and Germany, have tried todirect remittance flows to investments in thehome country to encourage return migration,but these efforts have met with little success.There are also some examples of “forced” remittancetransfer programs between sendingand receiving countries, although these raisevexing legal issues and do not appear to be effectiveeither in encouraging migrant return ormobilizing resources (box 4.4).Box 4.4 Forced remittancesWhile it is generally assumed that migrantworkers are free to choose how much, when,and to whom to send money, there have been caseswhen sending or receiving governments, banks inthe home country, or employers have decided to retaina certain proportion of pay for remittances. Therationale for such “forced” remittances is to ensurethat temporary migrant workers do not stay on, butreturn home after the end of their contract. Sometimes,the objective of such measures is to steer theuse of remittances to investment in the country oforigin.For example, from 1942 to 1964 the “Braceroprogram” regulated migration of 4.6 million farmworkers between Mexico and the United States.From 1942–9, a tenth of the wages earned by thesebraceros was deducted from their pay by the U.S.employers and paid into accounts held by the Bankof Mexico at two commercial banks in San Francisco.From there it was transferred to the Bank ofMexico and then on to the Banco de Credito Agricola.Alternatively, the employers gave the worker acheck for the deducted amount at the end of the contractto be cashed back home in Mexico. A 1946 reportby the Mexican government claimed that$8 million in forced savings had been paid out to exbracerosand only $6 million was unaccounted for;but the LA Times reported (on March 30, 2001) thata total of $34 million in forced savings was collectedduring 1942–6. The loss of the money was explainedby successive bank consolidations and restructuring,and as a result, records of accounts had disappeared(Migration News, http://migration.ucdavis.edu/mn).The braceros were mostly poorly educated peasants,who did not even know about the deductions andwho later were intimidated by the forms and correspondenceneeded to claim their money (LA Times).op.cit.). In March 2001, a class action suit was filedon behalf of former braceros at a San Francisco districtcourt claiming $30 million–$50 million in savingsnot returned and additional punitive damages.This claim was rejected because of the statute of limitations(San Francisco Chronicle, August 29, 2002).In 2003, the Mexican government agreed to reimburse,within six months, an upfront sum of $150per person and then monthly rates of pesos 200 forup to pesos 60,000, provided the ex-braceros couldproduce identification (the Bracero Net program).Forced remittances may also be used by a governmentto encourage the use of remittances for97

GLOBAL ECONOMIC PROSPECTS 2006Box 4.4(continued)investment in the domestic economy. Under theDeferred Pay Scheme, mine workers from Lesothohave 30 percent (initially 60 percent, until 1990) oftheir pay deposited at a Lesotho bank and the balanceinto a savings account at TEBA (The EmploymentBureau of Africa) Bank. The certificate confirmingthe identity of the account holder is handedout by the TEBA Bank at the end of the contract,before the mine worker goes home to collect thebalance from the deferred pay scheme.A similar arrangement is foreseen in the memorandumof understanding between the Governmentsof Thailand and Laos on employment cooperation.All Laotian guest workers are obliged to pay 15 percentof their earnings into a “deportation fund” setup by the host country, Thailand. Workers who wishto return home can claim their contribution in fullwith interest. The request must be filed three monthsbefore the return date, and the money is to be paidto the workers within 45 days after the last day ofemployment (articles 11 and 12).A milder form of induced remittances has beenintroduced for temporary Mexican farm workers inthe United States and Canada. Before their departurevisas and work permits are issued, the temporaryfarm workers register with the Ministry of Labor inMexico. After the papers are delivered, migrantsopen a savings account with the subsidiary or anassociated institution of a North American bank inMexico. Once they arrive in the United States orCanada, the temporary workers either make theremittance transaction themselves or arrange withthe farmer-employer to pay directly into their savingsaccount via payroll deduction.Forced savings of this type raise legal issues inthat they violate an accepted principle of wage protection,that is, the idea that “wages shall be paid directlyto the worker concerned” (article 5 ILO Convention95 of 1949). The only exception providedfor is that the “worker concerned has agreed to thecontrary.” It is not clear whether that has been thecase with the braceros or with the other examplescited here. Convention 95 states that “employersshall be prohibited from limiting in any manner thefreedom of the worker to dispose of his/her wages.”Article 8.2 further spells out that “workers shall beinformed of the conditions under which suchdeductions may be made.” (Mexico ratified thisconvention in 1955. The United States has notratified it.)Forced remittances are also probably not the mosteffective measure to ensure that temporary migrantworkers return home. If they return, it is likely notdriven by their desire to reclaim their savings. Whenoffered a choice, migrants avoid such systems. InSouth Africa a considerable number of mine workersfrom Lesotho did not participate in the deferred payscheme, often in complicity with the mining companies(Sparreboom 1996, p. 13). If the Lesotho deferredpay scheme was voluntary, then the volume ofsavings would drop to a level of the voluntaryschemes of workers from Botswana and Swaziland,namely 1 percent of the levels of the obligatoryscheme (TEBA 1995).Banking and financial markets. Greater relaxationand competition in money transfer marketsleads to reduced prices and more moneyreaching the beneficiaries. This process is facilitatedfurther by improving access of remittanceservice providers to national paymentand settlement systems. This seems to haveworked well within framed agreements suchas the United States–Mexican Partnership forProsperity program of 2001, involving thematricula consular to improve banking accessof Mexican immigrants in the United Statesand low-cost electronic transfers through theFederal Reserve Bank’s automated clearinghousesystem for Mexico (see chapter 6).Spain has initiated agreements betweenSpanish and Latin American financial institutionsto reduce transfer fees and foster theentry of new agents into the financial market,particularly in rural areas. In the past,Germany worked closely with Turkey to encourageremittances into formal channels98

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCES(UN 2005). In some remittance-source countries,outward remittance flows are affected byexchange controls. For example, SouthAfrica’s policy of limiting foreign exchangedealings only to banks has prompted (unbanked)remitters to use informal channels—only 5 percent of remittances to other SouthernAfrican Development Community (SADC)countries are being sent via formal channels,according to Genesis Analytics (2005). 20ID arrangements for migrants. The U.S. facilitationof banking for both regular and irregularmigrants from Mexico through the matriculaconsular mechanism has been highlysuccessful in drawing more migrants intosafer and cheaper remittance modes. The FederalDeposit Insurance Corporation (FDIC)through its New Alliance Task Force initiative,in collaboration with the Mexican consulatesand commercial banks, has been successful inimproving banking access as well as the financialliteracy of immigrants. 21Support to HTAs or migrant associations.HTAs and similar entities receive some supportfrom host governments in the United States,France, and parts of Africa in recognition oftheir development assistance potential. WhileHTAs could potentially play a useful role incommunity infrastructure and other collectivelyfunded projects, their ability to effectively channellarge amounts of aid remains untested. 22Macroeconomic effectsof remittancesUntil recently most of the discussion andresearch on remittances was focused onthe (microeconomic) end use by the recipienthouseholds, including the effects on poverty(see chapter 5). But as outlined earlier in thischapter, the large size of remittances relative toother external flows and to the GDP in manycountries suggests that the macroeconomiceffects of remittances may be of critical importancein many countries (recall that the top 19remittance recipients receive more than 10 percentof their GDP in remittances).High levels (or large increases) in remittanceflows can be expected to have directrepercussions on foreign exchange rates, domesticinterest rates, and the balance of payments,and indirect repercussions on macrovariables.Because of their relative stabilityand targeting (directly to households), theymay bring some additional benefits. However,as the experience with and analysis of naturalresource booms have shown, large inflows canalso have some undesirable side effects (seealso box 4.5). And to the extent that remittanceflows may naturally just go to countriesthat are doing poorly or respond anticyclically(increase during downturns, due to a drought,for example), it may be hard to disentanglehow remittances affect macro-performance. Inthis section, we consider some of the macroeconomicchannels through which remittancesaffect recipient countries.Remittances are stable andmay be countercyclicalRemittances may move countercyclically relativeto the economic cycle of the recipientcountry. Remittances may rise when the recipienteconomy suffers a downturn in activity ormacroeconomic shocks due to financial crisis,natural disaster, or political conflict, becausemigrants may send more funds during hardtimes to help their families and friends. Remittancesmay thus smooth consumption andcontribute to the stability of recipient economiesby compensating for foreign exchangelosses due to macroeconomic shocks.Many authors have observed an increase inremittance inflows following a natural disaster(Clarke and Wallsten 2004) or an economicdownturn (Kapur 2003). Yang (2004)showed that remittance receipts by Filipinohouseholds increased following the 1997 financialcrisis. A 10 percent appreciation of amigrant’s currency against the Philippine pesoled to increases in household remittance receiptsand a 0.6 percentage point decline in thepoverty rate in migrant households. He alsofound evidence of positive spillover effects onhouseholds without migrant members due to99

GLOBAL ECONOMIC PROSPECTS 2006Figure 4.3 Remittances as percent of private consumption, two years before and two yearsafter natural disastersPercent201816141210864203.8 4.2 4.4 4.7 5.1 7.9 8.9 10.4 11.3 10.82 1 0 1 2BangladeshNote: The bar marked 0 is the year of the natural disaster.5.78.4 9.62 1 0 1 2 2 1 0 1 2 2 1 0 1 2Dominican Republic Haiti Honduras11.315.55.2 5.5 5.6 7.7 8.4increases in remittance-driven economic activityas well as by direct transfers from the migrant’sorigin household. Mishra (2005) findsthat a 1 percent decrease in real GDP was associatedwith a 3 percent increase in remittancesafter a two-year lag in 13 Caribbeancountries during 1980–2002. To the extent thatremittances are used for investment purposes,however, they may behave procyclically just asother investment flows do. In Turkey and thePhilippines, remittances were more volatile andprocyclical in the 1990s than in the 1980s. 23Remittance flows (as a share of personalconsumption) continued to rise after naturaldisasters in Bangladesh, Dominican Republic,Haiti, and Honduras (figure 4.3). In Albania,after an initial disruption in remittanceinflows (as a share of personal consumption)in the year of conflict, remittance flows recoveredquickly (figure 4.4). In Sierra Leone,remittances increased in the year of the conflict.24 Remittances as a share of personalconsumption rose in response to the financialcrisis in Mexico in 1995 and in Indonesia andThailand in 1997 (figure 4.5).Yang (2005) found that the increase in remittancesmakes up for 13 percent of incomelosses in the current year and 28 percentwithin four years of a hurricane. In contrast,increases in ODA and FDI make up forroughly 26 and 21 percent, respectively,within four years.Remittances can improve countrycreditworthinessRemittances can improve a country’s creditworthinessand thereby enhance its access tointernational capital markets. The ratio ofFigure 4.4 Remittances as a share ofpersonal consumption, two years beforeand two years after conflictPercent2015105016.9 16.912.72 1 0 1 2 2 1 0 1 2Albania17.012.32.9 3.20.71.0 0.8Sierra LeoneNote: The bar marked 0 is the year of the conflict.100

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESFigure 4.5 Remittances as a share ofpersonal consumption, two years beforeand two years after financial crisesPercent2.521.510.501.2 1.12.0 2.1 1.9Mexico0.5 0.52 1 0 1 2 2 1 0 1 2 2 1 0 1 2IndonesiaThailanddebt to exports of goods and services, a keyindebtedness indicator, would increase significantlyif remittances were excluded fromthe denominator (figure 4.6). Country creditratings by major international rating agenciesoften fail to account for remittances. 25 Modelbasedcalculations using debt-to-exportratios that include remittances in the denominatorindicate that including remittances increditworthiness assessments would improve1.42.0 2.01.81.71.51.0 1.1Note: The bar marked 0 is the year of the financial crisis.credit ratings for Lebanon and Haiti by twonotches; these would result in implied sovereignspread reductions ranging from 130 to334 basis points (table 4.5). 26Remittance securitization can helpcountries raise external financingAnother way in which remittances affectinternational capital market access is throughthe use of structured finance techniques.Several banks in developing countries (forinstance, Brazil) have been able to raise relativelycheap and long-term financing frominternational capital markets via securitizationof future remittance flows.Remittance securitization typically involvesthe borrowing entity (such as a bank) pledgingits future remittance receivables to an offshorespecial purpose vehicle (SPV). The SPV issuesthe debt (figure 4.7). Designated correspondentbanks are directed to channel remittanceflows of the borrowing bank through an offshorecollection account managed by atrustee. The collection agent makes principaland interest payments to the investors andsends excess collections to the borrowingbank. Since remittances do not enter the issuer’shome country, the rating agenciesFigure 4.6 Indebtedness classification including and excluding remittances, 2003Debt as percent of exports a800732700Excluding remittances600500400355Including remittances300200100257208245193151 127167121190145 114 126178116 124820LebanonEcuadorPakistan Philippines Jamaica Morocco Jordan El SalvadorGuatemalaSource: World Development Indicators, and World Bank staff calculations.Note: a. Present value of external debt as percent of exports of goods and services, and remittances.101

GLOBAL ECONOMIC PROSPECTS 2006Table 4.5 Impact of remittances on country credit rating and sovereign spreadRemittances as % Rating excluding Rating including Spread savingof GDP, 2004 remittances remittances a (basis points)Serbia and Montenegro 7 B BB 150Lebanon 14 B B 130Haiti a 28 CCC B 334Nicaragua a 11 CCC B 209Uganda a 5 B B 161Sources: Standard and Poors and World Bank staff calculations.Note: a. Calculated using a model similar to Cantor and Packer (1995); see Ratha and De (2005).Figure 4.7 Remittance securitization structureRemittance sendersBeneficiary’saccountRemittancepayments(foreign currency)Issuing bank creditsbeneficiary’s accountin domestic currencyCorrespondent banksMessageIssuing bankRemittance flows (foreign currency)Trustee collateralaccountExcess cash(foreign currency)Issuing bank’saccount atcorrespondent bankDebt service paymentInternational investorsOffshoreDomesticbelieve that the structure mitigates the usualsovereign transfer and convertibility risks.Such transactions also often resort to excesscoverage to mitigate the risk of volatility andseasonality in remittances.By mitigating currency convertibility risk,a key component of sovereign risk, the futureflow securitization structure allows securitiesto be rated better than the sovereign credit rating.These securities are typically structured toobtain an investment grade rating. In the caseof El Salvador, for example, the remittancebackedsecurities were rated investment grade,two to four notches above the sub-investmentgrade sovereign rating. Investment-grade ratingmakes these transactions attractive to awider range of “buy-and-hold” investors (forexample, insurance companies) that face limitationson buying sub-investment grade. As aresult, the issuer can access international capitalmarkets at a lower interest rate spread andlonger maturity. Moreover, by establishing acredit history for the borrower, these dealsenhance the ability and reduce the costs ofaccessing capital markets in the future.The first major securitization deal involvinginternational migrant remittancesoccurred in 1994 in Mexico. The volume ofremittance securitization has grown rapidlysince then (figure 4.8a). Using this instrument,102

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESFigure 4.8 Securitization of remittances,1994–2004a. Millions of dollars65 206 350 1151994b. By countryMexico24%Panama1%19951996199719981999El Salvador6%5401,1751,170Turkey35%1,600 1,690 2,837Kazakhstan1%Note: Applies to the first half of 2004 only.1,05520002001200220032004Brazil 31%Peru2%Mexico, El Salvador, and Turkey raised about$2.3 billion during 1994–2000. As electronictransfers became more widespread, it waseasier to track complex transactions, andremittances securitization gave way tosecuritization of diversified payment rights(DPRs), including migrant remittances, butalso payments related to exports and FDI.During 2000–4, a total of $10.4 billion wasraised through securitization of DPRs byBrazil ($5.3 billion), Turkey ($4.1 billion), ElSalvador, Kazakhstan, Mexico, and Peru (figure4.8b). Following a sharp increase in borrowingcosts in 2002 (in part because ofelection-year uncertainties), Brazil has raisedover $4 billion by issuing bonds backed by diversifiedpayment rights. These bonds resultedin a spread saving of more than 700 basispoints compared to Brazil’s sovereign spread.As experience with this instrument broadens,and investors become more comfortablewith its characteristics, it is possible that itcould be used by a wider range of countries (includingpoor countries) and for a broader rangeof external flows (remittances, tourism receipts,and commodity earnings). It is not easyto estimate the potential size of such futureflowsecuritization. But preliminary calculations,assuming an over-collateralization ratioof 5:1 and using migrant remittance figures for2003, show that developing countries couldpotentially issue nearly $9 billion and lowincomecountries could raise up to $3 billionannually from international capital markets.Several policy hurdles need to be crossedbefore securitization deals can proceed. Highfixed costs of legal, investment banking, andcredit-rating services and long lead times canpose difficulties for developing countries withfew large entities and high borrowing needs. Amaster trust arrangement can permit issuers tostructure a large deal but to tap the market inseveral tranches. Pooling receivables of severalbranches (or even several borrowers) couldalso help increase the deal size to justify largefixed costs. While the absence of an appropriatelegal infrastructure can also constrainissuance, this need not require an overhaulof the entire legal system. A more focusedapproach that concentrates on bankruptcylaw may suffice, by making sure that pledgedassets remain pledged in the event of default.So far, only the top-rated (in local currencyterms) financial institutions have issued futureremittance-backed bonds in an effort to piercethe sovereign foreign currency rating ceiling(that is, to obtain a higher rating for thesebonds than the sovereign foreign currency rating).The securitization transactions typicallydo not affect financial institutions’ ability todeliver remittances to the ultimate beneficiaries.Loosely speaking, the financial institutionsthat undertake a securitization transactionare pledging their rights to foreign103

GLOBAL ECONOMIC PROSPECTS 2006currency, but not their obligations to deliverremittances (typically in local currency terms).Potential issuers should be reminded, however,of significant risks—currency devaluationand, in the case of flexible rate debt, unexpectedincreases in interest rates—that areassociated with market-based foreign currencydebt. Moreover, securitized debt is inflexibledebt. Securitization of remittances (and otherfuture flows) by public sector entities reducesthe government’s flexibility in managing itsexternal payments and can conflict with thenegative pledge provision included in multilateralagencies’ loan and guarantee agreements,which prohibit the establishment of a priorityfor other debts over the multilateral debts.Large remittance inflows can leadto exchange rate appreciation andlower export competitivenessLarge and sustained remittance inflows cancause an appreciation of the real exchangerate and make the production of cost-sensitivetradables, including cash crops and manufacturingless profitable. Although empirical evidenceon the adverse effect of large inflows offoreign exchange in terms of trade and growthis limited, 27 it is plausible that this effect existsand is significant for some small economieswhere remittances are very high. Amuedo-Dorantes and Pozo (2004) found that a doublingof workers’ remittances resulted in realexchange rate appreciation of about 22 percentin a panel of 13 LAC countries (see alsoWinters and Martins 2004). Rajan andSubramanian (2005), however, did not findany evidence that remittance flows slow downgrowth by affecting competitiveness. 28 Moreover,as remittances tend to be relatively stableand persistent over long periods, the “Dutchdisease” effects of remittances are less of aconcern than similar effects of natural resourcewindfalls and other cyclical flows, andthe real exchange rate level achieved throughsensible policies may be sustainable (IMF2005). Governments in countries receivinglarge remittances can mitigate the effects ofreal exchange rate appreciation by allocating alarger portion of government expenditures oninfrastructure and also practicing more liberaltrade policies; both these measures wouldtend to increase exports and also contributeto improved labor productivity andcompetitiveness.A related concern is whether reliance on unearnedincome in the form of remittances hasadverse effects on the incentives to work, aswell as on the quality of economic policies andgovernance, similar to the well-documentedeffects of windfall gains from natural resourcessuch as oil. While oil exports are almostalways found to have a strong negativeimpact on various governance indicators, suchas control of corruption and rule of law, preliminarycross-country analysis suggests thatremittance flows may not have such negativeeffects (box 4.5). 29The evidence on the effect of remittanceson long-term growth is inconclusiveTo the extent that they finance education andhealth and increase investment, remittancescould have a positive effect on economicgrowth. Remittances may relieve credit constraintsin the recipient community and spurentrepreneurial activity (Funkhouser 1992,Yang 2004, Woodruff and Zenteno 2004).Faini (2002) finds that the impact of remittanceson growth is positive. He argues thatremittances overcome capital market imperfectionsand allow migrant households to accumulatepositive assets, as claimed by Starkand Lucas (1988) and Taylor (1994). Mishra(2005) found that a 1 percentage point increasein remittance inflows in 13 Caribbeancountries increased private investment by0.6 percentage point (all measured relative toGDP). To the extent that they increaseconsumption, remittances may increase percapita income levels and reduce poverty andincome inequality, even if they do not directlyimpact growth (see chapter 5).On the other hand, large outflow of workers,especially skilled workers, can reducegrowth in labor-sending countries. Remittancesmay also indirectly affect labor supply,104

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESBox 4.5 Unlike oil windfalls, remittance inflows donot weaken institutional capacityThe economic performance of most mineral exporters,in particular oil exporters, has been farless impressive than that of resource-poor countries(Gelb and others 1988; Auty 2001; Gelb, Eifert, andTallroth 2002). To a large extent this outcome seemsdriven by mismanagement of the economy and weakinstitutions. Sala-i-Martin and Subramanian (2003)show empirically that concentration of resourceflows has deleterious effects on the institutionalframework and capacity of a country. Natural resourcewindfalls—oil rents, for example—often fosterweak institutions because they allow the authoritiesto pursue arbitrary, costly, and inefficient policies(Ross 2001). States that control such resources mayresist secular modernization pressures because theycreate alternative sources of power (Isham andothers 2003). These rents also perpetuate economicinequality, which results in nepotism and a weakcivil society. Resource rents are also believed to beassociated with civil conflict (Collier and Hoeffler2002).In contrast, remittances are widely dispersed, thegreat bulk of them is allocated in small amounts, andfor the most part, remittances avoid the government“middleman.” Hence the expectation is that they canavoid the negative effects of natural resource windfallson poverty, growth, and institutional capacity.This is similar to an argument by Birdsall and Subramanian(2004) that countries would be better off ifthey distributed the bulk of the returns from resourceflows to the general population, who would use thefunds more effectively than a highly centralized government,and also greatly reduce the incentives forcorruption.by encouraging some remittance-recipienthouseholds to choose more leisure thanlabor. Chami, Fullenkamp, and Jahjah (2005)argue that remittances may slow down growthby reducing work efforts by remittancerecipients. 30One recent study of the impact of remittanceson growth over an extended period(1970–2003) for 101 developing countriesfound no significant link between remittancesand per capita output growth, or betweenremittances and other variables such as educationor investment rates (IMF 2005). Thisstudy, however, attributed this inconclusive resultto measurement difficulties arising fromthe fact that remittances may behave countercyclicallywith respect to growth. 31 Also, empiricallyit is difficult to measure the effects ofremittances on human capital formation,which may occur over a very long period oftime.Remittances, like aid, may be more effectivein a good policy environment. Forinstance, a good investment climate with welldevelopedfinancial systems and sound institutionsis likely to imply that a higher share ofremittances is invested in physical and humancapital (IMF 2005). Indeed, Giuliano andRuiz-Arranz (2005) show that in theeconomies where the financial system is underdeveloped,remittances alleviate creditconstraints and work as a substitute for financialdevelopment, improving the allocation ofcapital and therefore accelerating economicgrowth. Recent research also shows thatremittances may promote financial development(Aggarwal and others 2005), which inturn can enhance growth and reduce poverty(Beck and others 2004).Annex 4A.1 World Bank dataon remittancesUsing the definition in chapter 7 of GlobalDevelopment Finance 2003, migrant remittancesare considered the sum of workers’remittances, compensation of employees, andmigrants’ transfers. Data for these variables105

GLOBAL ECONOMIC PROSPECTS 2006Table 4A.1.1 Countries with alternativeestimates in 2004$ millionsAlgeria 2,460China 21,283Gambia 8Iran 1,032Kenya 464Lebanon 2,700Malaysia 987Mauritious 215Nigeria 2,751Serbia and Montenegro 4,129Vietnam 3,200Total $39,259are taken mostly from the balance of payments(BoP) data file of the IMF (see also Ratha2003). However, many countries do not reportdata on remittances in the IMF BoP statistics,even though it is known that emigration fromthose countries took place (see table 4A.1.1 fora list of these countries). In 2003 about 87countries did not report any remittances’ data.Further, there was no consistency in reportingthe data. For example, only 28 countries reportworkers’ remittances, compensation ofemployees, and migrants’ transfers. Forty-fivecountries report both workers’ remittancesand compensation of employees; 11 countriesreport compensation of employees and migrants’transfers; and 3 countries report workers’remittances and migrants’ transfers. Thereare 14 countries that report only workers’ remittancesand 19 countries that report onlycompensation of employees.Reported data for developing countriesshow only $113.4 billion in total remittancesfor the year 2003 (workers’ remittances$97.3 billion, compensation of employees$14.8 billion, and migrants’ transfers $1.3 billion),and 83.8 billion in 2004 (workers’ remittances$68.7 billion, compensation of employees$13.5 billion, and migrants’ transfers$1.5 billion). By filling in gaps for some developingcountries for which remittance datawere missing, we arrived at an estimate of$142 billion in 2003, and $160 billion in2004 (the latest year for which BoP data arecurrently available). The gap-filling methodsfollowed, and the reasons for making theadjustments are documented below.Workers’ remittances, as defined in theIMF Balance of Payments manual, publishedin 1993 (fifth edition), are current privatetransfers from migrant workers who are consideredresidents of the host country to recipientsin their country of origin. If the migrantslive in the host country for a year or longer,they are considered residents, regardless oftheir immigration status. If the migrants havelived in the host country for less than a year,their entire income in the host country shouldbe classified as compensation of employees.Workers’ remittances are transfers, whereascompensation of employees is considered factorincome. In the earlier, fourth edition of theBoP manual, compensation of employees wascalled labor income and was classified as nonfactorservices (referred to just as services inthe fifth edition).Although the residence guideline in themanual is clear, this rule is often not followedfor various reasons. Many countries compiledata based on the citizenship of the migrantworker rather than on their residency status.Further, data are shown entirely as eithercompensation of employees or as worker remittances,although they should be split betweenthe two categories if the guidelineswere correctly followed; for example, SaudiArabia and Israel record only compensationof employees. India shows very little compensationof employees, but large workers’remittances, although it is well known thatIndia supplies a large number of temporaryIT workers to the United States and Europeancountries. On the other hand, thePhilippines shows large compensation ofemployees and very few migrants’ transfers.The distinction between these two categoriesappears to be entirely arbitrary, dependingon country preference, convenience, and taxlaws or data availability. This fact has beenrecognized at the World Bank since the1980s, and worker remittances have been106

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCEStreated as part of labor income and added toexports of goods and services in calculatingdebt service ratios.Though small in comparison to compensationof employees and workers’ remittances,migrants’ transfers have become anothersource of confusion. Migrants’ transfers arethe net worth of migrants that are transferredfrom one country to another at the time ofmigration (for a period of at least one year).Migrants’ transfers are considered capitaltransfers in the BoP fifth edition manual, althoughthey were considered current privatetransfers in the fourth edition. As the numberof temporary workers increases, the importanceof migrants’ transfers may increase.Therefore, in order to get a complete pictureof the resource flow, one has to consider thesethree items together.There are four main reasons for gaps inremittance data: vintage, missing data, datarecorded under other than the three categoriesmentioned above, and data collection practices.VintageThe Balance of Payments Yearbook publishesdata with a one-year lag. That is, the yearbookpublished in December of the current yearshould have data up to December of the previousyear. However, this is not true for anumber of developing countries, for which thelatest data available are two or even morethan three years old. For about 28 countries in2003, and 59 countries in 2004, data havebeen obtained from World Bank country desksor extrapolated on the basis of earlier trends.In addition, two countries, Algeria andNigeria, have not reported data to the IMF fora number of years. For Algeria the IMF datastop in 1991, and for Nigeria data, stop in1999. However, data for these countries areavailable from the country and reported in thecountry databases of the World Bank andIMF.Missing dataSeveral developing countries (for example,Lebanon) do not report to the IMF. 32 Datafrom the country desks are used for The Gambia,Iran, and Serbia and Montenegro; anddata from central banks were used forLebanon and Vietnam. Some high-incomecountries (notably Canada, Singapore, UnitedArab Emirates) also do not report remittancedata.Classification under other categoriesDue to the difficulty in classifications, countrieshave often classified workers’ remittanceseither as other current transfers or as transfersfrom other sectors. For example, in the case ofHaiti, before 1989 and after 1997, data wererecorded as workers’ remittances, but during1990–7, they were recorded as transfers fromother sectors. Kenya and Malaysia data havesimilar difficulties. For these countries, dataunder “other sectors” from the IMF aretreated as worker remittances. In China alarge proportion of workers’ remittances areclassified as other private transfers in the IMFBoP file. Therefore, instead of the IMF’s workers’remittances, we have used workers’ remittancesdata from the country desk. It is notjust the developing countries that follow thispractice, many high-income OECD countries(for example, the United Kingdom) do thesame.There are also other problems in the data,such as the difficulty in separating travel expenditurefrom remittances, which have notbeen addressed here. The increased acceptanceof credit and debit cards in developing countriesfurther complicates the issue. In somecountries, notably China, remittances mayhave been misclassified as FDI. The OECDdefinition of FDI (including the purchase ofholiday or second homes by nonresidents)may be counted as FDI—a likely case inChina. In the case of India and many othercountries, remittances may have been classifiedas nonresident deposits, especially thosein local currency terms.Data collection practicesA survey of central banks, based on responsesfrom 40 central banks, reveals widespread107

GLOBAL ECONOMIC PROSPECTS 2006problems with remittance data collectionmethodology (de Luna Martinez 2005). Mostof the central banks use remittance data reportedby commercial banks, but leave outflows through money transfer operators andinformal personal channels.Even when data are available and properlyclassified, in many cases they are not based onactual exchange records. In a number of cases,the preferred methodology of estimating theworkers’ remittances is based on taking thenumber of emigrants, and multiplying by anaverage amount sent. The sources for thesedata are migration records, surveys ofexchange and financial houses, and householdsurveys. However, these data are often weakor out of date. Also the methodology forpreparing estimates is not the same in all countries,and it is not always described in thecountry notes in the publicly available balanceof-paymentsdata. It is hoped that the increasedawareness about the importance of remittancesand the shortcomings in both the remittanceand migrant workers’ data will result in effortsto improve the data transmission.Table 4A.1.1 shows the countries where wehave used alternative estimates of workers’ remittances’using either country desk or thecentral bank data.Perhaps the most difficult aspect of remittancedata is estimating informal flows. Inannex 4A.2, we discuss different ways of estimatinginformal flows. One way to estimatethe true size of remittances is to undertakesurveys of remittance senders and recipients.Unless new, adequately randomized and representativesurveys of recipients and sendersare carried out, evidence from existing householdsurveys would only be indicative ratherthan comprehensive.Annex 4A.2 A model-basedestimation of informal remittanceflowsEstimating the size of unrecorded flows isalmost impossible. In what follows, wemake an effort to arrive at some crudeestimates using a set of variables that are notedin the literature to affect the choice of the remittancechannel. Empirically, this involvesfirst estimating officially recorded remittancesas a function of fee, exchange commission, andthe presence of a dual exchange rate (and othervariables shown in the equation below). Next,using the estimated coefficients on thesevariables, we predict what remittances wouldbe if the values of these variables becomecloser to those prevailing in regions whereinformal flows are small. We then interpret thedifference between these predicted remittancesand the actual remittances as an estimate of informalflows.For this purpose, we propose the followingmodel of remittances:REMIT 0 1Host 2Home 3Migrant 4Fee200 5Spread200 6Dualwhere REMIT is the log of remittances (orremittance per migrant or per capita); Host isthe log of the host-country per capita output(trade or migration weighted across hosts);Home is the log of home-country per-capitaoutput; Migrant is the log of the stock ofmigrant workers in OECD countries; Fee200is the fixed fee for sending $200 from theUnited States to the source country; Spread200is the exchange commission for sending $200;and Dual is a dummy variable for dual exchangerates. The last three variables are likelyto have large impacts on the extent to whichmoney is sent via formal channels. The data onremittances are available on a panel basis; dataon transaction costs and on the number of migrantworkers are only available for a crosssection.In table 4A.2.1 below, we report regressionresults for a cross-section of countries(using average figures for 1995–2003 for remittancesand other time series variables). Intable 4A.2.2, we show results of remittancecost functions estimated using cross-countrydata. These equations estimate panel data onremittance costs for use in panel data regressionsreported in table 4A.2.3. Reduced form108

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESTable 4A.2.1 Regression results: determinants of worker remittancesDependent variable: Dependent variable: DependentDependent variable: Ln (Remittances) Ln (Remittances per variable:Ln (Remittances) IV a emigrant) Ln (RemittanceExplanatory variables (1) (2) (3) per capita) (4)Dual exchange rate 0.42 0.18 0.31 0.74**(1.32) (0.46) (0.90) (2.04)Service fee 0.06** 0.12* 0.07** 0.04(2.32) (1.94) (2.47) (1.00)Exchange-rate spread 0.04 0.02 0.05(0.50) (0.26) (0.52)Stock of migrant 0.73** 0.64 0.22**workers (7.66) (5.41) (2.45)Main host per capita 0.10 0.05 0.22* 0.11income (0.77) (0.31) (1.75) (0.69)Home per capita 0.15 0.17 0.06 0.72**income (1.03) (1.00) (0.40) (4.35)Income 0.31** 0.31(4.05) (3.75)Number of 104 85 104 104observationsR 2 0.70 0.69 0.08 0.35Source: Freund and Spatafora (2005).Note: Robust t-statistics appear in parentheses.a. Instruments include financial development and dollarization. Hansen’s J-statistic is 2.49 (p-value 0.12), and the Shea partialR-squared of the instruments is 0.37.**significant at the 5 percent level; *significant at the 10 percent level.Table 4A.2.2 Regression results: determinants of transaction costsDependent variable: remittance costExplanatory variables (1) (2) (3) (4) (5)Bank concentration 0.05** 0.03 0.03(2.29) (1.10) (1.35)Financial development 0.05** 0.05** 0.05** 0.06** 0.05**(2.41) (2.38) (2.42) (2.54) (2.53)Financial risk 0.04 0.04 0.03 0.04 0.01(0.32) (0.26) (0.24) (0.31) (0.02)Dollarization 4.12** 3.92** 3.92** 4.10** 4.00**(4.47) (3.87) (4.20) (4.14) (4.33)Domestic output 0.22 0.56 0.32 0.75 0.46(0.46) (1.02) (0.66) (1.47) (0.98)Remittances 0.30 0.42*(1.18)(1.85)Emigrant stock 0.34 0.59**(1.18)(2.40)R 2 0.52 0.56 0.53 0.52 0.52Number of observations 76 69 76 70 78Source: Freund and Spatafora (2005).Note: The dependent variable is the percentage cost of remitting $200. Robust t-statistics appear in parentheses.**significant at the 5 percent level; *significant at the 10 percent level.109

GLOBAL ECONOMIC PROSPECTS 2006Table 4A.2.3 Panel regression results: determinants of remittancesLn (RemittancesLn (RemittancesLn (Remittances) Ln (Remittances) per capita) per capita)Explanatory variables (1) (2) (3) (4)Dual exchange rates 0.28** 0.22** 0.26** 0.21**(2.69) (2.18) (2.48) (2.03)Fitted cost 0.08* 0.08*(1.70)(1.74)Bank concentration 0.00 0.00(1.00)(1.03)Financial development 0.03** 0.02**(5.13) (4.30)Financial risk 0.00 0.01(0.02) (1.12)Dollarization 0.62* 0.58(1.73) (1.60)Net errors and omissions 0.02** 0.02** 0.02** 0.02*(2.07) (2.03) (2.12) (1.97)Home per capita income 2.10* 3.64** 0.43** 0.63**(1.71) (2.92) (2.37) (3.31)Host per capita income 1.37 1.28 1.33 1.41(1.39) (1.34) (1.33) (1.46)Home income 1.71 3.13**(1.38)(2.46)Number of observations 295 295 295 295R 2 0.27 0.35 0.21 0.29Source: Freund and Spatafora (2005).Note: All regressions include country and year fixed effects. Robust t-statistics appear in parentheses.**significant at 5 percent level; *significant at 10 percent level.equations for remittances using statisticallysignificant variables in the cost regressions arealso shown in table 4A.2.3.Notes1. In March 2004, the G-7 finance ministers indicatedtheir intention to “continue to work on initiativesto reduce barriers that raise the cost of sending remittancesand to integrate remittance services in theformal financial sector” and their commitment to“work with governments, the private sector, and multilateraldevelopment banks to broaden the access forfamilies and entrepreneurs to financial services.” At theSea Island Summit in June 2004, the G-8 heads of statecalled for “better coherence and coordination of internationalorganizations working to enhance remittanceservices and heighten the developmental impact of remittancereceipts.” They indicated that “G-8 countrieswill work with the World Bank, IMF, and other bodiesto improve data on remittance flows and to developstandards for data collection in both sending and receivingcountries.”2. One market study estimates that global remittancesare about 2.5 times the size reported in the IMFbalance of payments (Aite Group 2005). The recentupward revision of China’s remittances to $21 billionin 2004, from an earlier estimate of $4.6 billion, lendssome support to this notion, although there is nostrong evidence that systematic misreporting is solarge. The discrepancy for China is reportedly due tothe fact that the Chinese figures include compensationonly for state employees. Some authors believe that aportion of China’s FDI attributed to overseas Chinesemay actually be a misclassification of migrant remittances.Some also believe that the recent surge in remittancesto China in part reflected speculative inflowsin anticipation of a revaluation of the yuan.3. For example, the reporting threshold (typicallyper person per day) is $10,000 in the United States,12,500 euros in western European countries (on average),and 3 million yen in Japan.4. Saudi Arabia, the second largest source of remittancesat $13.6 billion (or 5.4 percent of GDP) in2004, is now classified as a high-income country. SaudiArabia’s per capita income level has risen in response110

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESto the current high oil prices. The authorities preferthat Saudia Arabia be treated as a developing country.5. Efforts are under way through the GTAP consortiumto compile and estimate a comprehensive set ofbilateral migrant stocks, which are used here. SeeWalmsley, Ahmed, and Parsons 2005.6. More precisely, bilateral remittance flows arecalculated by allocating reported remittance inflows ineach country according to weights constructed asfollows:R(i,j) [Remittance flows to country j] *[M(j,i)/[sum over i of the nominator]]R(i,j) remittance flows from country i to countryj and M(j,i) stock of migrants from country j incountry i. Data on migrant stocks M(.) are taken fromthe GTAP database (Walmsley, Ahmed, and Parsons2005).7. Including Saudi Arabia as a developing countrywould raise South–South remittances to 45 percent andSouth–South migration stock to 60 percent.8. A World Bank survey of the African diaspora inBelgium conducted in spring 2005 revealed that42 percent of remittances from Belgium to Senegal,and 55 percent to Congo and Nigeria, go through informalchannels. Anecodotal evidence suggests thatnearly 70 percent of remittances in the France–Malicorridor take place through informal channels. Handcarriageis a popular yet informal channel of remittancesin many countries. In the Philippines, 40 percentof total flows are estimated to be remittances broughthome by migrants in person. Nearly 42 percent of outwardremittances from South Africa are believed tomove through informal channels (Genesis Analytics2005).9. In a calibration model, El Qorchi, Maimbo, andWilson (2003) argue that the black-market premium isthe key factor determining informal flows. Other factorsaffecting the choice of the channel are trust in theintermediary and anonymity and convenience factors,such as location, hours of operation, and language.10. Results will underestimate the size of informalflows to the extent that they are affected by other factors,such as a lack of legal documentation of migrantsand high tax rates. To the extent that there would stillbe some informal flows even at this lower remittancecost level, the estimates are actually lower bounds onthe true size of informal remittances. However, it ispossible that the increases estimated in table 4.4 representnew remittance flows, and not just the shift fromthe informal to the formal sector, in which case theseestimates would overstate informality.11. Page and Plaza (2005) use a similar methodologyand find that the share of unrecorded remittancesrelative to the total remittances averages 48 percentworldwide (and 73 percent in Sub-Saharan Africa).12. Between 2001 and 2004, the euro appreciatedby 28 percent relative to the U.S. dollar. During thisperiod, outward remittances from France andGermany actually declined by 5 percent in euroterms. Remittances from Italy and Spain increasednearly 40 percent in euro terms and 93 percent inU.S. dollar terms.13. A survey of Congolese, Senegalese, and Nigeriandiasporas in Belgium did not reveal any significant relationshipbetween the propensity to remit and thenumber of years a migrant has lived in Belgium. On thecontrary, several migrants who had been in Belgiumfor more than two decades continued to send significantamounts of remittances. Evidence from the PacificIslands also do not support remittance decay (Connelland Brown 2005; Simati and Gibson 2001). Grieco(2003), however, reported evidence of remittancedecay in the case of Micronesian migrants in Guamand Hawaii, caused by family reunification or death ofthe beneficiaries.14. The information presented here derives from asurvey of IOM country missions on the policies of theirhost countries, as well as studies by ADB, ECOSOC,USAID, DFID, and the World Bank.15. For example, Colombia has a 0.4 percent taxon transactions through money exchange bureaux andbanks, a temporary arrangement in effect until 2007.Belarus also taxes remittances from nonimmediatefamily members.16. For example, in Tajikistan the removal of thestate tax on cross-border bank transactions in 2003reportedly helped raise remittances from $78 million in2002 to $256 million in 2003 (Olimova and Bosc2003).17. For example, nonresident Pakistanis remittingover $10,000 through banking channels can importany personal item valued up to $1,200 duty-free perannum (World Bank 2005b).18. Extension of voting franchise to migrants overseasand other policies of political inclusion may alsocatalyze remittances and other financial flows to thecountry of origin (Carey 2003; Yang 2003).19. For example, Pakistan does not permit womenunder 35 to emigrate as domestic workers and Vietnambans females from working overseas in the entertainmentsector. Bangladesh recently abandoned similar restrictionsrecognizing that although such restrictionsmay protect migrants from exploitation, they may alsoencourage more irregular migration, rendering themeven more vulnerable.20. This is in part responsible for Western Union’swithdrawal from the South African market (GenesisAnalytics 2005).21. Such activities complement government objectivesto improve banking access in poor neighborhoods111

GLOBAL ECONOMIC PROSPECTS 2006(e.g., through the Community Reinvestment Act; seeFrias 2004).22. USAID also established an 18-month pilot programin 2004 with the PanAmerican DevelopmentFoundation to strengthen the capacity of U.S.-basedHTAs. The UK government is also looking at the possibilityof using HTAs as a conduit for developmentaid.23. See also Global Development Finance 2003,chapter 7, and Sayan (2004). A separate study bySayan (2005) finds that in a sample of 12 low-incomeand lower-middle-income countries during 1976–2003,real remittances responded to a fall in real GDP with aone-year lag. He also found evidence of countercyclicalitydue to consumption smoothing in India andBangladesh and procyclicality due to a stronger investmentmotive in Jordan and Morocco.24. Black (2004, p. 12) reports that remittances remainedsubstantial during the civil war in Côted’Ivoire.25. This is likely to be the case in countries (suchas the Philippines or Lebanon) where the headlineworker remittance variable has underestimated ormissing data.26. Sovereign spread rises exponentially as creditratings worsen along the rating scale. A one-notch improvementin credit ratings, therefore, results in higherspread saving for countries at the bottom of the ratingscale.27. See McMahon (1997) for a review of empiricalstudies on the so-called Dutch disease, a term coined byThe Economist in 1977.28. They argue that migrants may lose interest inremitting money and prefer to send goods instead, ifthe currency in the remittance recipient country is overvalued.Thus controlling overvaluation throughprudent macroeconomic policies can help attractremittances.29. Note that this result applies to cross-countrycomparison. It would be extremely difficult to empiricallyestimate the effect of remittances on institutionalcapacity over time in a given country, since institutionalchanges take place over a very long time. Alsosuch an exercise would require controlling for reversecausality: remittances may respond to cyclical orabrupt changes in economic growth and governance. Apriori, the effect of institutions on remittances can runeither way: On the one hand, better institutionalcapacity may attract remittances meant for investmentpurposes. On the other hand, better institutionalcapacity (if they also mean better performance) maymean less emigration and dependence on remittances.30. However, reduced work effort by some individualsmay not reduce the aggregate work effort in atypical developing country with a large pool ofunemployed.31. It is difficult to disentangle the reverse-causalityproblem (that growth also affects remittances) whilemeasuring the effect of remittances on growth. Someresearchers argue that the empirical results showinga negative association between remittances and growthmay largely reflect the fact that remittances tend torise when growth is weak in the remittance-recipientcountry.32. The list of countries that do not report remittancedata also includes the following 29 countries:Afghanistan, Angola, Bahamas, Bahrain, Bhutan,Burundi, Canada, Central African Republic, Chad,Congo Democratic Republic, Equatorial Guinea, Iraq,Kuwait, Liberia, Singapore, Somalia, Taiwan (China),Turkmenistan, United Arab Emirates, Uzbekistan,Zambia, and Zimbabwe.ReferencesAggarwal, Reena, Asli Demirguc-Kunt, and MariaSoledad Martinez Peria. 2005. “Do RemittancesPromote Financial Development? Evidence froma Panel of Developing Countries.” Unpublishedpaper, World Bank, Washington, DC.Aite Group. 2005. “Consumer Money Transfers: PoweringGlobal Remittances.” Unpublished paper.January. www.aitegroup.com.Alesina, Alberto, Arnaud Devleeschauwer, WilliamEasterly, Sergio Kurlat, and Romain Wacziarg.2003. Fractionalization. Journal of EconomicGrowth 8(2): 155–94.Amin, Mohammad, and Caroline Freund. 2005. “Migrationand Remittances in ESA Countries.”Washington, DC: The World Bank.Amuedo-Dorantes, Catalina, and Susan Pozo. 2004.“Workers’ Remittances and the Real ExchangeRate: A Paradox of Gifts.” World Development32(8): 1407–17.Annual Finance & Accounting InternationalConference—Managing Securitization forLebanon and the MENA Region, December 3-4,2004. Lebanese American University, School ofBusiness, Beirut.Auty, Richard. 2001. “Introduction and Overview.” InResource Abundance and Economic Development,ed. R. M. Auy. New York: Oxford UniversityPress.Beck, Thorsten, Asli Demirguc-Kunt, and Ross Levine.2004. “Finance, Inequality and Poverty: Cross-Country Evidence.” NBER Working Paper10979. National Bureau of Economic Research,Cambridge, MA.112

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESBirdsall, Nancy, and Arvind Subramanian. 2004. “SavingIraq from Its Oil.” Foreign Affairs (July/August): 77–89.Black, Richard, Savina Ammassari, Shannon Mouillesseaux,and Radha Rajkotia. 2004. “Migrationand Pro-Poor Policy in West Africa.” WorkingPaper C8. Sussex Centre for Migration Research,University of Sussex. November.Buencamino, Leonidas, and Sergei Gorbunov. 2002.“Informal Money Transfer Systems: Opportunitiesand Challenges for Development Finance.”DESA Discussion Paper 26.Cantor, Richard, and Frank Packer. 1995. “SovereignCredit Ratings.” Current Issues in Economics andFinance 1(3): 1–6. June.Carey, John M. 2003. “Political Institutions in ElSalvador: Proposals for Reform to ImproveElections, Transparency, and Accountability.”Unpublished paper. Dartmouth College.www.dartmouth.edu/~jcarey/el percent20salvador.pdf.Carling, Jorgen. 2005. “Migrant Remittances and DevelopmentCooperation.” PRIO Report 1/2005, Oslo.Chami, Ralph, Connel Fullenkamp, and Samir Jahjah.2005. “Are Immigrant Remittance Flows aSource of Capital for Development?” IMF StaffPapers 52(1).Clarke, George, and Scott Wallsten. 2004. “Do RemittancesProtect Households in Developing Countriesagainst Shocks? Evidence from a NaturalDisaster in Jamaica.” Unpublished paper. WorldBank, Washington, DC.Collier, Paul, and Anke Hoeffler. 2002. “Greed andGrievance in Civil War.” In Annual Bank Conferenceon Development Economics. Washington,DC: World Bank.Collyer, Michael. 2004, “The Development Impact ofTemporary International Labour Migration onSouthern Mediterranean Sending Countries: ContrastingExamples of Morocco and Egypt,”Working Paper T6. Development Research Centreon Migration, Globalisation and Poverty, Universityof Sussex, Brighton, UK.Connell, John, and Richard P. C. Brown. 2005. “Remittancesin the Pacific: An Overview.” AsianDevelopment Bank, Manila. March. www.adb.org/Documents/Reports/Remittances-Pacific/default.asp.de Luna Martinez, Jose. 2005. “Workers’ Remittancesto Developing Countries: A Survey with CentralBanks on Selected Public Policy Issues.” PolicyResearch Working Paper 3638. World Bank.El Qorchi, M., S. Maimbo, and J. Wilson. 2003. “InformalFunds Transfer Systems: An Analysis ofthe Informal Hawala System.” IMF OccasionalPaper 222. International Monetary Fund,Washington, DC.Ellerman, David. 2003. “Policy Research on Migrationand Development.” World Bank Policy ResearchWorking Paper 3117. Washington, DC.Faini, Ricardo. 2002. “Migration, Remittances, andGrowth.” Unpublished paper. University of Brescia.www.wider.unu.edu/conference/conference-2002-3/ conference%20papers/faini.pdf.Freund, Caroline, and Nikola Spatafora. 2005. “Remittances:Costs, Determinants, and Informality.”Background paper prepared for this report.World Bank.Frias, Michael. 2004. “Linking International RemittanceFlows to Financial Services: Tapping theLatino Immigrant Market.” Supervisory Insights.FDIC. Winter 2004.Funkhouser, E. 1992. “Migration from Nicaragua:Some Recent Evidence.” World Development20(8): 1209–18.Gelb, Alan, and others. 1988. Oil Windfalls: Blessingsor Curse? New York: Oxford University Press.Gelb, Alan, Benn Eifert, and Nils Borje Tallroth. 2002.“The Political Economy of Fiscal Policy and EconomicManagement in Oil Exporting Countries.”World Bank Policy Research Working Paper2899. Washington, DC.Genesis Analytics. 2005. “Supporting Remittances inSouthern Africa: Estimating Market Potential andAssessing Regulatory Obstacles.” Johannesburg,South Africa.Giuliano, Paola, and Marta Ruiz-Arranz. 2005. “Remittances,Financial Development, and Growth.”IMF Working Paper. Washington, DC.Grieco, Elizabeth. 2003. The Remittance Behavior ofImmigrant Households: Micronesians in Hawaiiand Guam. New York: LFB Scholarly Publishing.Gubert, Flore. 2005. “Migrant Remittances andTheir Impact on the Economic Development ofSending Countries: The Case of Africa.” Paper presentedat the OECD International Conference onMigration, Remittances, and the Economic Developmentof Sending Countries, February 23–25,Marrakech.Harrison, Anne, assisted by Tolani Britton and AnnikaSwanson. 2004. “Working Abroad—The BenefitsFlowing from Nationals Working in OtherCountries.” Paper prepared for the OECD RoundTable on Sustainable Development, Paris. September.www.worldbank.org/data/Remittances/3aHarrison.pdf.Hernandez-Coss, Raul E. 2004. “The U.S.-MexicoRemittance Corridor: Lessons on Shifting from113

GLOBAL ECONOMIC PROSPECTS 2006Informal to Formal Remittance Systems.” FinancialSector, World Bank.IFAD (International Fund for Agricultural Development).2004. “Remittances and Rural Development.”Discussion paper for the Governing Council,27th session, Rome, February 18–19, 2004.IMF (International Monetary Fund). 2005. WorldEconomic Outlook: Globalization and ExternalImbalances (chapter 2). Washington, DC. April.IOM. 2003b. World Migration 2003: Managing Migration—Challengesand Responses for People onthe Move, Geneva.IOM (International Organization for Migration).2005. World Migration Report. Geneva.Isham, Jonathan, Michael Woolcock, Lant Pritchett,and Gwen Busby. Forthcoming. “The Varieties ofResource Experience: Natural Resource ExportStructures and the Political Economy of EconomicGrowth.” In World Bank EconomicReview.Iskander, Natasha. 2005. “Social Learning as a ProductiveProject (Zacatecas and Guanajuato’s CautionaryTales).” Paper presented at OECD InternationalConference on Migration, Remittances,and the Economic Development of Sending Countries,February 23–25, Marrakech.Kaufmann, Dani, Art Kraay, and M. Mastruzzi. 2005.“Governance Matters IV: Governance Indicatorsfor 1996–2004.” World Bank Institute. Draft,May 9, 2005.Ketkar, Suhas and Dilip Ratha. 2001. “Securitizationof Future Flow Receivables: A Useful Tool for DevelopingCountries.” Finance and Development.International Monetary Fund. Washington, DC.Ketkar, Suhas and Dilip Ratha. 2004. “RecentAdvances in Future-Flow Securitization.” Paperpresented at the OECD International Conferenceon Migration, Remittances and the EconomicDevelopment of Sending Countries.” February23–25, Marrakech.La Porta, Rafael, Florencio Lopez-de-Silanes, AndreiShleifer, and Robert Vishny. 1999. “The Qualityof Government,” Journal of Law, Economics,and Organization 15(1): 222–79.Lam, Ricky and Leonard Wantchekon. 2003. PoliticalDutch Disease. April 10. New York University.www.nyu.edu/gsas/dept/politics/faculty/wantchekon/research/lr-04-10.pdf.Magoni, Raphaele. 2004. “France.” In InternationalMigration and Relations with Third Countries:European and U.S. Approaches, ed. Jan Niessenand Yongmi Schibel. Migration Policy Group.McMahon, Gary. 1997. “The Natural Resource Curse:Myth or Reality?” Unpublished paper. WorldBank Institute, Washington, DC.Mishra, Prachi. 2005. “Macroeconomic Impact ofRemittances in the Caribbean.” Unpublishedpaper. International Monetary Fund, Washington,DC.Newland, Kathleen, and Erin Patrick. 2004. “BeyondRemittances: The Role of Diaspora in Poverty Reductionin Their Countries of Origin.” MigrationPolicy Institute, Washington, DC.Olimova, Saodat, and Igor Bosc. 2003. Labour Migrationfrom Tajikistan. Dushanbe: Mission ofthe International Organization for Migration.www.untj.org/library/?mode=details&id=73.Orozco, Manuel. 2003. “Hometown Associations andTheir Present and Future Partnerships: New DevelopmentOpportunities? Report commissionedby the U.S. Agency for International Development.Orozco, Manuel, and Katherine Welle. 2004. “HometownAssociations and Development: A Look atOwnership, Sustainability, Correspondence, andReplicability.” Inter-American Dialogue. Unpublishedpaper.Page, John, and Sonia Plaza. 2005. “Migration, Remittances,and Development: A Review of GlobalEvidence.” Paper presented at the Plenary Sessionof the African Economic Research Consortium,May 29, Nairobi.Rajan, Raghuram, and Arvind Subramanian. 2005.“What Undermines Aid’s Impact on Growth?”NBER Working Paper 11657. October.Ratha, Dilip. 2003. “Workers’ Remittances: An Importantand Stable Source of External DevelopmentFinance.” In Global Development Finance2003, Chapter 7. World Bank. Updated in Remittancesand Development: Development Impactand Future Prospects, ed. Samuel Maimbo andDilip Ratha. Washington, DC: World Bank.Ratha, Dilip. 2004. “Understanding the Importance ofRemittance Flows.” Migration Policy Institute.Migrationinformationsource.org.Ratha, Dilip, and Prabal De. 2005. “Predicting SovereignRating: International Capital Market Accessfor Unrated Countries.” Background paper forthis report. World Bank.Ravallion, Martin. 2003. “Measuring AggregateWelfare in Developing Countries: How WellDo National Accounts and Surveys Agree?” TheReview of Economics and Statistics 85(3):645–52.Reuter, Peter, and Edwin M. Truman. 2004. ChasingDirty Money: The Fight against Money Laundering.Institute for International Economics.Washington.Ross, Michael. 2001. “Does Oil Hinder Democracy?”World Politics 53(3): 325–61.114

TRENDS, DETERMINANTS, AND MACROECONOMIC EFFECTS OF REMITTANCESSala-i-Martin, Xavier, and Arvind Subramanian. 2003.“Addressing the Natural Resource Curse: AnIllustration from Nigeria.” IMF Working PaperWP/03/139. International Monetary Fund,Washington, DC.Sayan, Serdar. 2004. “Guest Workers’ Remittances andOutput Fluctuations in Host and Home Countries.”Emerging Markets Finance and Trade40(6): 68–81.———. 2005. “Business Cycles and Workers’ Remittances:How Do Migrant Workers Respond toCyclical Movements of GDP at Home?” Unpublishedpaper. Bilkent University.Siddiqui, Tasneem. 2004. “Efficiency of MigrantWorkers’ Remittances.” Prepared for the AsianDevelopment Bank.Simati, A., and J. Gibson. 2001. “Do RemittancesDecay? Evidence from Tuvaluan Migrants in NewZealand.” Pacific Economic Bulletin 16(1):55–63.Sparreboom, P. 1996. “Migrant Worker Remittancesin Lesotho: A Review of the Deferred PayScheme.” Social Finance Programme WorkingPaper 16. International Labour Office, Geneva.Stark, Oded, and Robert E.B. Lucas. 1988. “Migration,Remittances, and the Family.” EconomicDevelopment and Cultural Change 36: 465–81.Taylor, Edward. 1994. “International Migration andEconomic Development: A Micro Economy-wideAnalysis.” In Development Strategy, Employmentand Migration, ed. J. Edward Taylor. Paris:Organisation for Economic Co-operation andDevelopment.TEBA (The Employment Bureau of Africa). 2005. “APresentation on the Existing and Potential Contributionto Rural Southern Africa.” MarshallTown, Johannesburg.Terry, Donald. 2005. “Sending Money Home: Remittancesas a Tool in Latin America and theCaribbean.” Development Bank, Manila, JointConference on Remittances, September 12–13,2005.UN (United Nations). 2005. Letter of February 23,2005, to the Secretary General from the PermanentRepresentatives of Brazil, Chile, France,Germany, and Spain (A/59/719-E/2005/12) TechnicalGroup on Innovative Financing Mechanisms.Economic and Social Council. New York.Walmsley, Terrie, Syud Amer Ahmed, and ChristopherParsons. 2005. “The GMig2 Data Base: A DataBase of Bilateral Labor Migration, Wages andRemittances.” GTAP Research Memorandum 6,Center for Global Trade Analysis, University ofSussex. www.gtap.agecon.purdue.edu/resources/download/2338.pdf.Winters, L.A., and Pedro M.G. Martins. 2004. “WhenComparative Advantage Is Not Enough: BusinessCosts in Small Remote Economies.” World TradeReview 3(3): 1–37.Woodruff, Christopher, and Rene Zenteno. (2004).“Remittances and Microenterprises in Mexico.”IR/PS working paper. Graduate School of InternationalRelations and Pacific Studies, Universityof California–San Diego.World Bank. 2005a. Global Development Finance2005. Washington, DC.World Bank. 2005b. “Migrant Labor Remittances inthe South Asia Region.” Finance and Private SectorDevelopment Unit/South Asia Region. ReportNo. 31577. February 2005.Yang, Dean. 2003. “Salvadorans Overseas: The Foundationof a Pro-Poor Growth Strategy.” Unpublishedpaper. Gerald R. Ford School of Public Policy,University of Michigan. www-personal.umich.edu/~deanyang/yang_cv.pdf.———. 2004.———. 2005. “Coping with Disaster: The Impact ofHurricanes on International Financial Flows,1970–2001.” Research Program on InternationalMigration and Development. DECRG. WorldBank.Yang, Dean, and Claudia Martinez. 2005. “Remittancesand Poverty in Migrants’ Home Areas:Evidence from the Philippines.” In InternationalMigration, Remittances, and the Brain Drain, ed.Caglar Ozden and Maurice Schiff. New York:Palgrave Macmillan.115

5Remittances, Households,and PovertyChapter 4 presented evidence on the macroeconomicdimensions of remittance flows—their overall size, determinants of their composition(formal versus informal), the role ofgovernment policies in determining theirmagnitude and use, and their macroeconomicimpacts—to developing countries. But as previouslynoted, these aggregate flows are comprisedof millions of individual remittancetransfers among private households, all undertakenby senders and receivers striving toimprove household welfare. This chapter considersthe impact of remittance flows at themicro-level, in particular on the welfare andopportunities of the recipient households andtheir members.Evaluating household impact depends ondata and analysis carried out at the householdlevel, often through household surveys.Surveys are available for many countries andperiods, and many of these have common orcomparable structures, but substantial differencesin coverage and circumstances complicatetheir interpretation. Such caveatsnotwithstanding, the evidence presented inthis chapter suggests that remittances can:• Reduce poverty, even where they appearto have little impact on measured inequality;• Help smooth household consumption byresponding positively to adverse shocks(for example, crop failure, job loss, or ahealth crisis);• Ease working capital constraints onfarms and small-scale entrepreneurs;• Lead to increased household expendituresin areas considered to be importantfor development, particularly education,entrepreneurship, and health. 1Our evaluation of the empirical analysis onremittances and development is structured asfollows. In the next section, we consider the effectsof remittances on poverty and inequality.We then explore how remittances can alleviatethe difficulties that households face in smoothingconsumption. The next section considersthe indirect effects of remittances on householdbudgets in terms of induced labor supplyeffects, increased access to working capital,and multiplier effects. We then examine howhouseholds allocate remittances to variouscategories of spending, with a particular emphasison evidence of remittance-funded investmentsin human capital, micro-enterprises,and property.Before continuing, two broad observationson the scope and interpretation of the availableanalysis help to put the results in perspective.First, in evaluating the impact ofremittances, it is important to consider thealternative (or counterfactual) situation thatserves as a comparison. If a household117

GLOBAL ECONOMIC PROSPECTS 2006member migrates and sends back remittances,one could evaluate the net change in the migrant’scontribution—that is, adding the remittancesand subtracting the income of themigrant had he or she stayed and worked athome. 2 That approach is the appropriatefocus when the goal of migration is to generateremittances, or when we are interested inthe overall effect of migration on remaininghousehold members. The alternative is to ignorethe lost domestic contribution—so thatthe counterfactual is now simply no remittances.This approach measures the narrowimpact of remittances, which seems appropriatewhen migration is being treated as exogenouslygiven, and our interest is simply in theremittance flows generated by the existing migrationstock. The second approach meritsclose attention because the existing migrantstock is large, and also because not all remittancesare received from migrant relativesabroad; third-party remittances are common. 3Second, in evaluating the benefits of remittances,we also need to weigh the welfare ofthe migrants themselves. To take a concreteexample, imagine the migration decision facinga young husband and father in a countrywith a long-established migration history thatborders on a much richer country. Economicopportunity (and possibly social pressure)may make remittance-motivated migrationirresistible. However, separated from familyand community support, this young mancould end up living a quite miserable existence.Clearly, a simple tracking of cash providesan inadequate guide to the welfare implicationsof the move.Remittances, poverty,and inequalityRemittances directly affect poverty by increasingthe income of the recipient. Theyalso indirectly affect poverty in the recipientcountry through their effects on growth, inflation,exchange rates, and access to capital.Measuring the impact of remittances is complex(in part because of the difficulties ofaccounting for the counterfactual loss ofincome from migration, as just mentioned).But a growing body of evidence from povertysimulation models, cross-country regressions,and analysis of household survey data showsthat remittances, in fact, do reduce poverty—although the evidence of their effect on inequalityis mixed.Remittances reduce povertyIn what follows, we present evidence on thepoverty effects of remittances, based on threesources: a poverty simulation model, a crosscountryregression analysis, and householdsurvey data from selected countries. The illustrativepoverty simulation model asks astraightforward question: how would povertyrates change in our sample of developingcountries if remittances were to disappearcompletely? Because this model is easy to implementfor most countries, it can providesome sense of the effect of remittances acrosscountries. However, the model is relativelycrude and cannot account for the fact thatwhile remittances affect poverty, the level ofpoverty also affects the volume of remittances.In comparison, cross-country regressionanalysis requires more data and is harder toimplement, but it is better able to control forreverse causality between remittances andpoverty. Household surveys are most likely toprovide the data required for a rigorous analysisof the relationship between remittancesand poverty. The surveys also allow one to analyzethe counterfactual loss of income due tomigration. It is difficult, however, to generalizeacross countries on the basis of householddata, particularly because most availablehousehold surveys do not have usabledata on remittances, especially internationalremittances.To understand how remittance flows mightaffect measures of poverty, we start with an illustrativepoverty simulation model and ask astraightforward (although unrealistic) question.The model relates the change in povertyto income growth and inequality change. Itis estimated using cross-country data for118

REMITTANCES, HOUSEHOLDS, AND POVERTYBox 5.1 Estimating a cross-country povertychange modelThe basic idea behind a poverty change modelis that a particular measure of poverty (say thefraction of the population with incomes below$1 per day) is a function of descriptive parametersof the income distribution, such as the mean and theGini coefficient. Building on Ravallion (1997), weposit a conditional constant elasticity specification,in which there is a constant growth elasticity ofpoverty reduction that varies with the initial levelof inequality. After reformulation, the basicrelationship (see annex 5A.1) relates the rate ofpoverty change to a measure of inequality-adjustedincome growth and an income-adjusted change ininequality.To estimate the relationship, we use the datasetassembled by Adams and Page (2005), in which theobservations relate to the period between comparablenationally representative household surveys. Foreach country, we have data from surveys for the initialand final values of poverty, inequality, and percapita income. Below is the estimated poverty changeequation for the headcount measure of poverty.Using survey income and consumption data as theincome variable (last column), results from themodel are robust, with statistically significant impactsfrom both income growth and inequalitychange on the headcount rate.Estimated poverty change modelIncome variable Income variable Survey mean income or consumptionDependent variable—proportionateGDP per capitachange in the headcount rateIntercept 0.39 0.33(2.21) (2.42)Inequality-adjusted growth 4.93 5.60(2.57)(2.27)Income-adjusted inequality change 0.60 1.11(1.57) (2.82)R 2 0.08 0.30Number of observations 81 81Source: World Bank staff estimates.Note: Standard errors are robust to country-level clustering; t statistics are in parentheses. The model is estimated using firstdifference of variables. The poverty line for the headcount poverty calculation is $1.08 in purchasing-power-parity-adjusted1993 dollars (equivalent to $1 in 1985 dollars). The per capita income measure is mean survey income or consumption(depending on availability from the survey).81 countries. The methodology and results arepresented in box 5.1 and in annex 5A.1. Thepremise behind the analysis is that the incrementalincome from remittances can be analyzedin the same way as incremental incomefrom economic growth—so we can simulatethe impact of eliminating remittances by modelingan income decline equal to the originalremittance level. For the sake of simplicity, weassume that nothing else changes—that thereare no offsetting increases in domestic sourcesof income or other adjustments to spendingbehavior or labor supply. 4Results from the simulation are summarizedin table 5.1 (see also annex 5A.1). We reportaveraged results for different groups ofcountries—first, by distinguishing betweenhigher-remittance recipients (greater than119

GLOBAL ECONOMIC PROSPECTS 2006Table 5.1 Simulated impact of eliminating remittances on poverty rateChange in headcount rate, noCountry group No. of countries Remittances/GDP (%) Poverty headcount rate Gini changeLow remittances 23 2.2 25.6 5.0Low headcount rate 12 2.0 11.8 1.2High headcount rate 11 2.5 40.6 9.1High remittances 14 11.0 24.8 12.2Low headcount rate 7 8.0 10.7 4.1High headcount rate 7 14.1 38.9 20.3Source: World Bank staff estimates.Note: Low (high) remittances refer to the remittance share in GDP less (greater) than 4 percent. Low (high) headcount rate refersto a rate less (greater) than 20 percent. Allowing inequality to change 2 Gini points has minimal effects on the change in theheadcount rate. See annex 5A.1 for detailed results.4 percent of GDP) and lower (less than 4 percentbut greater than 1 percent), and second, bythe extent of poverty (headcount above 20 percentor below 20 percent).These results show that the impact of eliminatingremittances depends on how large theyare to begin with (higher initial levels meansteeper income declines), the initial extent ofpoverty, and the degree of inequality. For example,the average increase in the headcountratio for higher-remittance countries (12.2 percentagepoints) is more than twice that ofthe lower-remittance countries (5 percentagepoints). Similarly, with each of these twogroups, the impact is much greater for thosecountries with higher headcount ratios to startwith. The estimated impact of inequality—an assumed 2 point worsening in the Ginicoefficient—has only a small marginal impacton the estimated change in the poverty rate.This simple analysis has significant limitations.First, the simulated effects depend onaccurate country-level measures of remittances,which, as emphasized in chapter 4,are of variable reliability. Second, many of thecountry simulations are made outside the sampleused for the regression analysis and aretherefore subject to the standard out-ofsampleprediction problems. Third, the analysisassumes that remittances are included inhousehold income when calculating the measuresof poverty and inequality from householdsurveys. In reality, there is variationacross surveys in how remittances are accountedfor in the household surveys. 5The results just described provide an indicationof the role that remittances can play inreducing poverty, but because of the simplicityof the model and other limitations, the resultsare not conclusive. More rigorous analyticwork has been undertaken to investigate thelink between remittances and poverty basedon careful analysis of cross-country data.In a model that relates national povertylevels to mean income and the Gini measureof inequality for 71 developing countries, a10 percent increase in per capita official internationalremittances leads to a 3.5 percent declinein the share of people living in poverty(Adams and Page 2005). 6 Other recent studieshave broadly confirmed these findings, includingIMF (2005) (see chapter 2), which uses asample of 101 countries for the period1970–2003.Although the available evidence is still relativelylimited, growing evidence from householdsurvey data complements the findings ofthe model that international remittances havereduced the incidence and severity of povertyin several low-income countries. Accordingto that evidence, remittances are believed tohave reduced the poverty headcount ratio by11 percentage points in Uganda, 6 percentagepoints in Bangladesh, and 5 percentage pointsin Ghana (Adams 2005b). Completely removingremittances for Lesotho would raise the120

REMITTANCES, HOUSEHOLDS, AND POVERTYheadcount poverty ratio (with a poverty lineequal to 60 percent of mean household expenditure)from 52 to 63 percent (Gustafsson andMakonnen 1993).While remittances had only a limited rolein reducing the number of poor people inGuatemala, they did significantly reduce thedepth and severity of poverty (Adams2004a). 7 International remittances accountedfor 60 percent of income for households in thelowest income decile, but were not very largefor households located near the poverty line(roughly the fifth income decile). As a result,international remittances had more impact onreducing the depth of poverty than on thepoverty headcount; in other words, they werereally helpful for the poorest of the poor.Wodon and others (2002) conclude that inGuerrero and Oaxaca, two southern Mexicanstates with significant international emigrationand remittance inflows, the share of thepopulation living in poverty is lower by 2 percentagepoints due to remittance income. Theyargue that this poverty effect is similar in magnitudeto that of many government programsin poverty reduction, education, health, andnutrition. Taylor, Mora, and Adams (2005),using data from a 2003 survey, find that internationalremittances account for 15 percent ofper capita household income in rural Mexico.They conclude that an increase in internationalremittances would reduce both thepoverty headcount and the poverty gap.Yang and Martinez (2005) studied the impactof variations in the exchange rate on remittancessent by Filipino workers and the ultimateimpact of remittances on poverty in therecipient regions. Using a large dataset fromthe Overseas Filipino Survey, they found thatan appreciation of the Philippine peso led toan increase in remittance flows, which contributedto the reduction in poverty. Interestingly,increased remittances not only reducedpoverty in the migrant families, they alsohad spillover effects on nonmigrant families.(We will have more to say on multiplier effectslater in this chapter.)The effect of remittances on inequalityis unclearIn contrast to the link between remittancesand poverty, no strong conclusion is found inhousehold studies of the relationship betweenremittances and inequality: remittances sometimesgo disproportionately to better-offhouseholds and so widen disparities, but inother cases they appear to target the less welloff, causing disparities to shrink. Some studiessuggest that the remittances from new migrationmay raise inequality in the short term,but the effect on inequality is small over thelong term. 8 Calculations that impute incomesfor the migrant had he stayed and worked athome generally show an increase in inequalityfrom the combined effect of migration and remittances.For example, inequality was foundto have increased in Bluefields, Nicaragua,when an imputation was made for the lost domesticincome of migrants, but it fell when thedomestic income of migrants was ignored(Barham and Boucher 1998).Two recent studies, however, did not findan increase in inequality even after controllingfor the counterfactual income loss frommigration: Adams (2005a) found that inGhana, the inclusion of international remittancesin household expenditures led to only aslight increase in income inequality, but thatthe Gini coefficient remained relatively stable,between 0.38 and 0.40. 9 De and Ratha (2005)found that in Sri Lanka, the Gini coefficientdrops from 0.46 to 0.40 as a result of remittancereceipt.Differences in findings on the impact ofremittances on inequality also stem from varyinggeographic and historic circumstances,such as the distance from high-incomedestination countries and the prevalence ofnetworks of earlier migrants. Both proximityto high-income countries and establishednetworks will tend to reduce the cost ofmigration, making migration an option forpoorer (and often credit-constrained) households.10 For example, remittances to a Mexicanvillage with a well-established history of121

GLOBAL ECONOMIC PROSPECTS 2006international migration had an equalizing effect,whereas remittances to another Mexicanvillage for which international migration wasa relatively new phenomenon tended to makethe distribution of income more unequal(Stark, Taylor, and Yitzhaki 1986). For a largenumber of Mexican communities, the overallimpact of migration and remittances is estimatedto reduce inequality for communitieswith relatively high levels of past migration(McKenzie and Rapoport 2004). 11One reason for the inconclusive empiricalevidence on inequality effects is that the Ginicoefficient does not adequately capture incomemobility; it remains unchanged, for example,if one person moves up and another movesdown the income ladder. Using household surveydata from Sri Lanka, De and Ratha (2005)show that the poor income deciles also havesignificant overseas migration and that remittancerecipients in the middle income decilesmove up the income ladder (figure 5.1). 12Beyond the contradictory or inconclusiveresults, some scholars question whether thelink between remittances and inequality is allthat important. Inequality matters when it interfereswith the functioning of the economy(for example, when credit constraints bindmore households 13 ) or the political system(for example, when growing inequality increasessupport for governments that pursuedamaging populist policies 14 ). Greater inequalitymay also be considered bad becauseof its impact on social welfare (see Sen 1973for a discussion). But it should be kept in mindthat in the context of remittances, inequalityrelates to income differences among groupsthat would all be viewed as relatively poor inan industrial-country context. The rich in developingcountries probably receive little inthe way of remittances; the rich who migratetend to take their families with them.Remittances and householdconsumption smoothingRemittances may play a significant role insmoothing consumption. Poor householdsFigure 5.1 Sri Lankan migrationa Percent of Sri Lankan households withmigrants abroadPercent12108642012 3 4 5Source: De and Ratha 2005.Income decile6 7 8 9 10b Percent of Sri Lankan households that moved upto a higher income decile after receiving remittancesPercent252015105012 3 4 5Source: De and Ratha 2005.Income decile6 7 8 9 10that lack access to insurance and creditmarkets are vulnerable to severe declines inincome from adverse shocks, and they maybe forced to forgo income-generating—butrisky—strategies (Morduch 1994). Informalcommunity institutions generally play a limitedrole in mitigating risk (see, for example, Coateand Ravallion 1993 and Fafchamps 2004), especiallyin the face of adverse events such as acommunity-wide crop failure. One strategy toreduce risk is for households to send familymembers to other regions or countries, where122

REMITTANCES, HOUSEHOLDS, AND POVERTYthey are not likely to face the same incomeshocks as those found in the domesticmarket. 15 Migration patterns and policies thatencourage migrants to travel unaccompaniedby family members encourage this form of risksharing.There is some evidence that remittancesfrom internal migration provide insurance.Remittances to Botswana increased with theextent of drought in the migrant’s home region,and the responsiveness of remittance levelsto drought was greater for households withmore drought-sensitive assets such as cattle(Lucas and Stark 1985). 16 The anticipation ofinsurance may allow the household to pursuea more risky asset accumulation strategy—although it is also possible that householdswith more to lose from drought (whatever thereason) are simply more likely to receive remittances.The likelihood that Thai internalmigrants move to Bangkok is reduced themore closely income in Bangkok aligns withincome in the province of origin (Paulson2000). 17 The effect is particularly strong forremittances to rural households, which arelikely to be poorer and have less access to formalinsurance products to mitigate weatherrelatedrisks. The more volatile a household’sincome (and the more restricted its ability toself-insure), the greater the distance thathouseholds in rural India tend to send theirdaughters to marry (Rosenzweig and Stark1989). Greater distance means that the covarianceof income shocks with the home regionwill be smaller, facilitating consumptionsmoothingtransfers between these relatedhouseholds.Studies of how remittances respond to adversehousehold shocks generally support theview that remittances provide some insurance.However, interpreting these correlations iscomplicated by the likelihood of reversecausality (remittances can influence householdoutcomes as well as be influenced by them)and omitted variable bias (certain hard-tomeasurehousehold characteristics may affecta household’s susceptibility to risks as wellas the likelihood of receiving remittances).Consider, for example:• Migrants responded to the cost of hurricanedamage borne by Jamaican households,with each additional dollar ofhurricane damage leading to $0.25 inadditional remittances (Clarke andWallsten 2004. The authors use paneldata to control for the household-levelrisk aversion and vulnerability effectsthat potentially bias the estimates. 18 )• Remittances are estimated to have replaced60 percent of income loss due toweather-related shocks in a sample ofFilipino households (Yang and Choi2005. 19 Rainfall is used as an instrumentfor income to avoid reverse causality;panel data are used to control for the tendencyfor risk-averse households to locatein places where incomes are more stableand to send migrants to manage risk.)• In cross-country data, a dollar’s worth ofhurricane damage leads to roughly $0.13in additional remittances in the year ofthe hurricane and $0.28 cents over fiveyears (Yang 2005). (Yang uses meteorologicaldata to instrument for reporteddisaster damage, because damage reportsmay be affected by the anticipation offinancial flows.)Remittances and indirect effectson household incomeRemittances may indirectly affect householdincome through changes to the laborsupply of those remaining behind; relaxationof working capital constraints that expand incomefrom entrepreneurial or farming activities;and multiplier effects on household income.Unfortunately, the evidence on each ofthese channels is quite limited, so we are constrainedhere to identifying important areas foradditional research.Remittances may affect labor supplyRemittances may tend to reduce the supply oflabor provided by remaining householdmembers, who may take a portion of the123

GLOBAL ECONOMIC PROSPECTS 2006remittance gain as leisure. This income effect isgenerally not a concern, because it representspart of the welfare gain from remittances. Bycontrast, remittances may change the return tosupplying labor, for example, if the migrantconditions the remittance on low householdincome. 20 Such a substitution effect will reducethe welfare gain from remittances bydistorting household labor decisions.However, it is difficult to separate incomeand substitution effects of remittances on thelabor supply of those remaining behind. Lookingat the overall effect, a rise in remittancesreduced labor force participation in Managua,Nicaragua, but increased self-employment(Funkhouser 1992). Remittances were estimatedto reduce the participation rates ofremaining household heads in a number ofCaribbean countries, although the directionof causality was hard to establish (Itzigsohn1995). Yang (2004) points to more encouraginglabor-supply effects than the standardmodel when he determined that remittancesreduce the supply of child labor but increasethat of adult labor.Remittances provide working capitalThere is some evidence that remittances provideworking capital to households that lackaccess to credit markets. For example, migrationto South Africa’s mines initially reducedagricultural production in countries of origin,because labor was removed from the farm(Lucas 1987). However, over time productionrose with migration, perhaps due to remittance-fundedcapital investment and a greaterwillingness to take risks with agricultural production,owing to the more diversified sourcesof family income. Remittances had a smallnegative effect on household income for Mexicoin 1982, but a large positive effect for 1988(Taylor 1992). One possible explanation isthat over time the development of migrant networksallowed migration from poorer householdsthat are more likely to be credit constrained(see the discussion of inequality,above). The effect of remittances on householdincome depends on both the liquidity ofhousehold assets (which determines their valueas collateral) and on the availability of inputsthat complement entrepreneurial activity (Taylorand Wyatt 1996). The role of remittancesin relaxing household credit constraints inrural cropping income in China dominated thedirect loss of productive labor from migration,so that internal migration increased per capitahousehold income (excluding remittances) by14 to 30 percent (de Brauw, Taylor, andRozelle 2001). Mishra (2005) found that a 1percentage point increase in remittance inflowsin 13 Caribbean countries increased privateinvestment by 0.6 percentage point (allmeasured relative to GDP).Remittances may ease credit constraints becausea stable stream of remittance incomemay make households more creditworthy inthe eyes of formal sector financial institutions.Remittance receipts that increase when thehousehold receives an adverse shock may beeven more important in relaxing credit constraints,since they increase the lender’s confidencethat they will be repaid even if thingsturn out badly for the household. This creditworthinesseffect deserves careful empiricalinvestigation, given the increasing interest inchanneling remittances through formal financialchannels.Remittances may have multiplier effectsSome studies have found that remittanceshave a multiplier effect, whereby the increasein domestic income is some multiple of the remittanceincome. For example, each dollarsent by Mexican migrants to the United Stateswas estimated to boost Mexican GDP by$2.90 (Adelman and Taylor 1992). Such multiplierswill occur if output is constrained byinsufficient demand. However, in many developingcountries sustained underemployment islikely to have supply-side causes, for example,government policies that increase the cost ofhiring and firing workers, so that increaseddemand will ultimately result in higher inflationrather than increased output. 21Nevertheless, there may be greater scopefor sustained multiplier effects at the regional124

REMITTANCES, HOUSEHOLDS, AND POVERTYlevel. The local spending of remittance incomewill generate further income for other localhouseholds, which in turn is likely to causelocal inflation for nontraded goods and possiblya small increase in national inflation. A nationalgovernment with a formal or informalinflation target is likely to respond to any increasein the national inflation rate by tighteningmonetary policy, thereby leading to an offsettingeffect on national aggregate demand.The net effect would be multiplier effects atthe local level but not at the national level. Indeed,the local gains come partly at the expenseof the regions that do not receive the remittancesbut are forced to suffer the tightermonetary policy.Remittances also may have multiplier effectsin the context of increasing returns, typicallyas the expansion of one sector increasesthe optimal size of other sectors. 22 Althoughsuch income-expanding feedback loops couldbe present at the national level, they are againmore likely to be relevant at the regional level,because expanding regions attract labor andcapital from elsewhere in the economy. Thebottom line is that remittance-induced multipliereffects cannot be ruled out—especially atthe regional level—but our current empiricalunderstanding of their importance is quitelimited.Remittances, savings,and investmentDoes it matter how households allocate remittanceincome between consumptionand saving? Allocations to the latter may boosthousehold investment or national investmentthrough allocation to financial assets. But froma welfare perspective, an extra dollar of investmentis only better than an extra dollar of presentconsumption if the marginal social valueof investment is greater than its marginal privatevalue. 23 Although a number of factors candrive a wedge between social and private values(such as capital income taxes, monopolypowers, and credit constraints), one prominentreason raised in the development context is thepossible existence of positive externalities frominvestment expenditure. 24 Thus the way thatremittances are allocated by households mayaffect the social value of a given remittanceflow.The rate of investment of remittance incomewill be high when:• Remittance flows are viewed by thehousehold as transitory rather than permanentand thus should be saved (andinvested) rather than spent.• The sender conditions the remittance onit being spent for particular purposes,which are more likely to involve investmentthan current consumption. Examplesinclude education or the purchase ofnew farm machinery.• The remittance is targeted (or “tagged”)to household members more likely touse the funds for investment purposes(women rather than men). 25• Households practice a form of mentalaccounting with their overall budget,with remittances being disproportionatelyput in accounts set aside for investmentpurposes. 26On the other hand, some of the literature alreadyreviewed suggests reasons to expect thatthe marginal propensity to invest remittanceincome will be low when (a) remittances aretargeted to poor households that are strugglingto meet subsistence needs and (b) they are targetedto credit-constrained households that areexperiencing adverse consumption shocks.The empirical challenge in identifying thecausal effect of remittances on investment isthat remittances are likely to be correlatedwith the extent of opportunities for investment,thereby biasing the estimated remittanceeffect. That correlation could be positive ornegative. When more enterprising householdsare the ones sending migrants and the oneswith substantial investment opportunities,high remittances will be wrongly associatedwith high investment. On the other hand, tothe extent that households send migrants125

GLOBAL ECONOMIC PROSPECTS 2006when investment opportunities are absent athome, then high remittances will be wronglyassociated with low investment. The empiricalsolution is to find a source of variation in remittancesthat is plausibly unrelated to householdinvestment opportunities.Measuring the impact of remittances on investment—eitherin physical or in human capital—isnot easy. Household budget surveysare best suited for this purpose, but most ofthe existing surveys either do not record dataon international remittances or are poorlydesigned. Since remittances are fungible, it isdifficult to isolate their effects from those ofother sources of income. Simply asking howremittances are spent is unlikely to reveal thetrue marginal effect on spending, because remittances,even when used for investment purposes,may free up the marginal dollars forconsumption spending. 27Remittances can lead to investmentsin education and healthSome of the clearest evidence for remittanceinducedinvestments comes from work onhuman capital. The dramatic depreciation ofthe Philippine exchange rate during the Asianfinancial crisis increased remittances fromFilipino migrants (because from the migrants’perspective, exchange-rate depreciation raisedthe relative price of their own consumptionin the destination country compared withconsumption by household members backhome), leading to greater child schooling, reducedchild labor, and increased educationalexpenditure in origin households (Yang2004). 28 In El Salvador, remittances are estimatedto reduce the probability of childrenleaving school by 10 times the effect of othersources of income in urban areas and by 2.6times in rural areas (Cox Edwards and Ureta2003). 29 They speculate that remittances havea disproportionate influence on schooling expendituresbecause the migrant has made it acondition for the financial support. Mexicanchildren in households with migrants completedsignificantly more schooling, with thelargest impact (an additional 0.89 years ofschooling) for girls in households where themother has a low level of education (Hansonand Woodruff 2003).Health status is both an important componentof human capital and a central element ofwell-being in its own right. Unfortunately,the effect of migration on the health offamily members remaining behind—notablychildren—is poorly understood. Migrationfrom Mexico is associated with lower (by3 percent) infant mortality and higher birthweights of children left behind (Hildebrandtand McKenzie 2005). The positive health effectscome through increased access to healthrelatedknowledge as well as through increasedhousehold wealth. Notwithstandingthese encouraging outcomes, the authors cautionthat the impact of migration on childhealth is quite nuanced, with migration associatedwith lower measures of preventivehealth care such as breast-feeding and vaccinations.30 De and Ratha (2005) find that inSri Lanka, remittance income has a positiveand significant impact on the weight of childrenunder five; this result is especially strongfor female-headed households. However, thehealth impact of absenteeism of one of theparents is negative.Remittances can encourageentrepreneurshipThere has been a marked shift from the beliefthat migrants are unlikely to establish newbusiness enterprises in their countries of origin(either upon return or through remittancefinancing) to the view that migration encouragesentrepreneurship. Large receipts ofremittances from the United States are associatedwith a greater likelihood of productiveinvestment in Mexico (Massey and Parrado1998). 31 A survey of 6,000 small firms in 44urban areas in Mexico shows that remittancesare responsible for almost 20 percent of thetotal capital in urban micro-enterprises(Woodruff and Zenteno 2001). The share risesto one-third for the 10 states with the highestrates of United States–bound migration.126

REMITTANCES, HOUSEHOLDS, AND POVERTYRemittances also appear to ease credit constraintson new business formation in thePhilippines (Yang 2004). The effect of exogenousincreases in remittance income on theprobability of entering into entrepreneurshipis larger for low- to middle-income households,which are the ones most likely to facecredit constraints. Policies that facilitate easyexit and reentry for migrants may encourageincreased involvement in remittance-fundedinvestments or enterprises.Remittances are often investedby recipient householdsContrary to the conventional wisdom thatremittances tend to be “frittered away” by recipienthouseholds, recent work has estimatedthat a large proportion of remittance incomeis saved. Only 12 percent of net incrementsto expenditure by rural Egyptian householdswere allocated to consumption, with largepropensities to invest in the construction andrepair of houses, and in agricultural or buildingland (Adams 1991). This relatively highpropensity to invest is assumed to result fromhouseholds treating remittance receipts astemporary income flows, which forwardlookinghouseholds save (and invest) ratherthan consume. These findings are largely confirmedin a later study of Pakistani households(Adams 1998). 32 In Guatemala, remittancereceivinghouseholds are found to have lowermarginal propensities to consume and ahigher propensity to invest in education,health, and housing than other households(Adams 2005c).It should be noted that some survey resultsfor a number of Latin American countriespoint to much higher propensities to consumeremittance income (see, for example, IADB-MIF 2004). The percentages of remittancesspent on household expenditures are 78 percentin Mexico, 77 percent in Central America,and 61 percent in Ecuador, while spendingon real estate and education is low. However,surveys of how income from a particularsource is spent tend to be unreliable, becausemonies from different sources are consideredperfect substitutes by the household. In contrast,studies such as Yang (2004) econometricallyestimate expenditure propensities givenexogenous changes in remittance income, sothat the estimates should be less susceptible tothe fungibility problem. A second explanationfor the different results is that the econometricstudies measure marginal propensities,whereas the direct surveys measure averagepropensities. It is the marginal propensity thatis of interest when we consider the expenditureeffects of policies that increase remittanceflows. 33The role of remittances in funding investmenthas recently been questioned in a macroeconomicpaper by Chami, Fullenkamp, andJahjah (2005), who find that remittances tendto be negatively associated with economicgrowth. This countercyclical behavior of remittancesis consistent with the evidence discussedabove that remittances respond to adversehousehold shocks. But the observationthat remittances tend to move countercyclicallydoes not necessarily obviate their role infunding investment. The micro studies we reviewedpoint to remittances as both smoothingconsumption and providing funds for investment.Moreover, the increased flow ofremittances in the face of adverse shocks mayallow households to sustain funding for keyinvestments in areas such as business workingcapital, education, and health care.The evidence reviewed in this chapter suggeststhat remittances play multifaceted rolesin poverty reduction, consumption smoothing,and investment, with the balance of rolesvarying by time and place.Annex 5A.1 Poverty simulationmodel: description and resultsThe poverty change model assumes that aparticular measure of poverty (say thefraction of the population with incomes below$1 per day) is a function of descriptive parametersof the income distribution, such as themean and the Gini coefficient. Building on127

GLOBAL ECONOMIC PROSPECTS 2006Ravallion (1997), we assume that there is aconstant growth elasticity of poverty reduction,but we allow that elasticity to vary with theinitial level of inequality. 34 We call this aconditional constant elasticity specification.Specifically, the poverty measure, P, is given by,P AY (1 I) ,where Y is per capita income (measured asmean survey income or consumption), I is themeasure of inequality (which we take to be theGini coefficient), and A and are parameters.Differentiating the poverty equation and writingit in rate-of-change form yields our basicpoverty change model,dPP(1 I)dYY lnY dIThis equation can be interpreted as sayingthat the rate of poverty change depends on aTable 5A.1 Effect of removing remittances on the poverty headcount ratePoverty headcount rate Remittances as shareCountry Survey year ($1 a day, PPP, $1993) of GDP (%) Change in headcount rate (Gini 0)Using per Using surveycapita GDP mean incomeArmenia 1998 13 5 2 4Azerbaijan 2001 4 2 0 0Bangladesh 2000 36 4 5 13Bolivia 1999 14 1 0 1Botswana 1993 31 2 1 3Burkina Faso 1998 45 3 3 4Colombia 1999 8 2 0 0Côte d’Ivoire 1998 16 1 0 1Ecuador 1998 18 3 2 3Egypt, Arab Rep. 2000 3 3 0 1El Salvador 2000 31 13 10 41Gambia, The 1998 54 7 9 12Georgia 2001 3 6 0 0Guatemala 2000 16 3 1 2Honduras 1999 21 6 3 4India 2000 35 3 3 23Kenya 1997 23 3 2 2Lao PDR 1997 26 2 2 4Lesotho 1995 36 44 29 34Mali 1994 72 6 10 17Mexico 2000 10 1 0 1Moldova 2002 22 19 13 21Mozambique 1997 38 2 2 2Nepal 1996 39 1 1 2Nicaragua 2001 45 2 3 13Pakistan 1999 13 2 1 2Panama 2000 7 2 0 0Paraguay 1999 15 4 1 2Peru 2000 18 1 1 2Philippines 2000 16 8 3 10Senegal 1995 22 3 2 3Sierra Leone 1989 57 3 3 4Sri Lanka 1996 7 6 1 3Swaziland 1996 8 6 1 1Tajikistan 1999 14 6 3 3Uganda 1999 85 4 9 40Yemen, Rep. 1998 16 19 10 7Source: World Bank staff estimates.Note: Remittance share is 1995 for Sierra Leone. Estimates are based on poverty reduction model.128

REMITTANCES, HOUSEHOLDS, AND POVERTYmeasure of inequality-adjusted growth and anincome-adjusted change in inequality. 35To estimate the relationship, we utilize thedataset assembled by Adams and Page (2005).The observations relate to spells between comparablenationally representative householdsurveys. For a given spell, we have data for theinitial and final values of poverty, inequality,and per capita income. 36 The estimated equationfor the headcount measure of poverty isreported in the table in box 5.1 in the maintext. Similar results were found for thepoverty gap measure.The next step is to simulate the effect ofremoving remittances on the poverty measureunder various assumptions about how remittancesaffect inequality. The proportionate increasein per capita income due to remittancesis given simply by the share of remittances inGDP multiplied by the ratio of per capita GDPto mean survey income/consumption. 37 It isimportant to emphasize that the simulatedpoverty-increasing effect of removing remittancesapplies to the latest year for which asurvey is available for that country, and thusdifferent years are being used for differentcountries. Care should thus be taken in makingcomparisons about the importance of remittancesin reducing poverty across differentcountries. Where the headcount rate is below2 percent we do not attempt to estimate thepoverty change effect of remittances.Table 5A.1 shows results on the povertyheadcount when remittances are removed(assuming there is no impact of remittances oninequality) for 37 countries where remittancesare above 1 percent of GDP and where thepoverty headcount rate is greater than 2 percentat the outset.Notes1. This represents a significant shift from the traditionalearlier pessimism about the role of remittancesin development. For example, Papademetriou andMartin (1991) emphasize how migration increases thedependence of emigration countries that are unable “toregulate or channel remittances,” while Jacobs (1984)states that remittances “did nothing to convert stagnationto development” in abandoned regions.2. Since data on what the migrant was earning beforeleaving are typically unavailable, the lost domesticincome is estimated or imputed based on observedcharacteristics of the migrant and on knowledge ofhow those characteristics are rewarded in the domesticeconomy.3. In the Sri Lanka Integrated Survey 1999–2000,nearly a third of households receiving remittances didnot report having a migrant member overseas. It is possiblethat those households received remittances fromtheir extended family; it is also possible that they receivedremittances from friends (De and Ratha 2005).The literature also notes third-party remittances inother countries (see, for example, Yang 2004 for thePhilippines). A survey of African diasporas in Belgiumfound that more than one member in a typical migranthousehold sent regular remittances and that each remittermight send remittances to different recipients.4. This situation is akin to the second counterfactual(a decline in remittances but no change in migrationand hence household income) discussed at the outset.To the extent that a decline in remittance income mayencourage households to devote more labor hours todomestic income-generating activities, the total declinein household income and the consequent poverty effectmay be smaller than assumed in the simulation, a pointmade in Adams (2004a).5. If we adopt the other extreme assumption—thatremittances are not included in household income—theresults can be interpreted with a simple reversal of thesign, which gives the reduction in poverty that wouldresult if remittances were included.6. As higher poverty increases the incentive to migrate,the ordinary least squares (OLS) estimates ofthe impact of remittances and stock of migrants arebiased downward. On the other hand, when creditconstrainedpoor families do not have the resources tosend migrants, the OLS coefficients are biased upwardas poorer countries send fewer migrants. To deal withthe bias, Adams and Page (2005) allow for remittancesand emigrant stock variables in their regressions, usingmeasures of international distance, government stability,and levels of education.7. The poverty depth is the average shortfall belowthe poverty line expressed as a fraction of the povertyline (or simply the poverty gap ratio); and povertyseverity is the squared poverty gap ratio. A key featureof this severity measure is that it is sensitive to the distributionof income among the poor (Foster, Greer, andThorbecke 1984).8. Stark, Taylor, and Yitzhaki (1986) found that a1 percent increase in international remittances leads to129

GLOBAL ECONOMIC PROSPECTS 2006a 0.14 percent increase in the Gini coefficient in thecase of Mexican villages with a short migration history;but in other villages with a long migration history,the Gini coefficient actually declines by 0.01 percent.Taylor (1992), McKenzie and Rapoport (2004),and Taylor, Mora, and Adams (2005) also find negativeeffects on the Gini in the case of Mexico.9. Nearly 40 percent of households in a representativesample in Ghana receive remittances, of those remittances,4 times as many households receive internalthan international remittances (Adams 2005b). Householdsreceiving internal remittances are disproportionatelypoor, indicating the importance of internal remittancesin reducing poverty. Households surveys inEurope and Central Asia also show that in a number ofcountries (Albania, Kosovo, Moldova, and Tajikistan,for example) a large number of households receive remittances,many in rural areas (World Bank 2005).10. Research has also shown that the inclusion ofremittance-induced, indirect effects on income—suchas income-induced reductions in nonmigrant laborsupply or increases in entrepreneurial income due tothe relaxation of credit constraints—can change thedirection of the inequality effect.11. The authors experiment with various instrumentsfor location-specific migration levels in their regressions.The instruments include the historic (1924)state-level migration rate and unemployment rates forthe U.S. state that includes the city that is the likelydestination for migrants from a particular Mexicanlocation.12. This study imputes the counterfactual incomeby calculating the income from equivalent activities athome. In the bottom two deciles, remittance income isoffset by the counterfactual loss of income frommigration, whereas in the top two deciles, remittanceincome falls short of the counterfactual loss of income.13. See, for example, Banerjee and Newman (1993).14. See, for example, Alesina and Rodrik (1994)and Persson and Tabellini (1994).15. The New Economics of Labor Migration(NELM) emphasizes that (a) migration is often betterviewed as a family rather than an individual decision;(b) risk management and provision of credit are seen toplay a central role in migration and remittance decisions;and (c) migration is often seen as a response tothe failure of markets for insurance and credit (Taylor1999). Rosenzweig (1988, p. 1167) highlights theinformational problems that undermine crucial marketsand emphasizes how ties of common experience,altruism, and heritage “enable families to transcendsome of the informational problems barring the developmentof impersonal markets.”16. See also Stark and Lucas (1988).17. Rainfall is used as an instrument for provincialincome in establishing the covariance pattern.18. Simple cross-sectional estimates of how remittancesrespond to hurricane damage will be biased downwardif more risk-averse households are more likely tosend migrants as a general insurance strategy and aremore likely to take actions to reduce the risk of costlydamage to the home. They will be biased upward ifhouseholds with more vulnerable dwellings are morelikely to send migrants and more likely to suffer hurricanedamage. An additional complication is moral hazard,where insured (remittance receiving) households haveless incentive to avoid risky behavior. Clarke andWallsten (2004) deal with the potential endogeneityproblem by using the average damage done in the neighborhoodas an instrument for household-specific damage.19. They cannot reject the null hypothesis that allof an exogenous decline in income is matched by anincrease in remittances.20. There is a disincentive to work if remittancesare conditioned on low income. Conversely, if remittancesare conditioned on domestic labor supply—“Iwill help you if you help yourself”—there is an addedincentive to work. From the migrant’s perspective,there is some similarity with the challenges faced bygovernments in providing social assistance without creatingpoverty traps and dependency. A traditional“welfare” model conditions remittances on householdincome, whereas the modern “workfare” model attemptsto condition remittances on household effort inan attempt to avoid putting the household in a trapwhere working makes little economic sense. However,from the strict welfare perspective of the standardmodel, any distortion to the labor supply of the remainingmembers—negative or positive—reduces thewelfare gain from the remittance.21. See Layard, Nickell, and Jackson (1991) for adiscussion of supply-side constraints on employment.22. See Cooper (1999) for an introduction to multipliereffects under increasing returns where the outputsin different sectors are “strategic complements.”23. These values are assumed to be expressed inunits of current consumption.24. As examples, consider that the social returnfrom human capital investments is greater than theprivate return due to knowledge spillovers (Moretti2003); the social return from investments in vaccinationis greater than the private return due to thespillover of reduced disease contagion (Miguel andKremer 2001); and that the social return from entrepreneurshipis greater than the private return due toexternalities from demonstration effects about where acountry’s comparative advantage might actually lie(Hausmann and Rodrik 2002).130

REMITTANCES, HOUSEHOLDS, AND POVERTY25. See Bourguignon and Chiappori (1992) andBrowning and others (1994) for treatments of collectivedecision making in the “nonunitary household.”Duflo (2003) demonstrates that pensions received bywomen in South Africa have a larger impact on theweight and height of girls than of boys living in thehousehold. In other words, how that income is spentdepends on who receives it. Using data from the Côted’Ivoire, Duflo and Udry (2004) show that where husbandsand wives farm different plots of land, the effectof rainfall shocks that differentially affect the plots hasimplications for the composition of household expenditure.They consequently reject the hypothesis of “fullinsurance” within the household.26. See Thaler (1990) for a discussion of mentalaccounting.27. Adams 2005a. See also Swaroop and Devarajan(2000) for a discussion of the fungibility problem in thecontext of evaluating the impact of official flows.28. This is a useful test of the allocation ofremittances between consumption and savings if thedepreciation-induced change in remittances is independentof household investment opportunities.29. This finding is subject to the problem of identifyingwhether remittances increase schooling orwhether households with migrants are more likely touse additional income for schooling. The authors arguethat remittances are closer to a randomly assigned transfer,particularly for political exiles whose migration isless likely to be correlated with household factors thataffect the likelihood of human-capital investment.30. Again, it is difficult to separately identify theimpact of migration on health outcomes. Individualsfrom households with poor health status may not bewell enough to make a difficult border crossing; themost prosperous and healthy households may findthat local opportunities outweigh those yielded by arisky illegal move; or adverse shocks may affect bothmigration decisions and health status. Hildebrandtand McKenzie’s (2005) empirical solution to thisidentification problem is to instrument for migrationusing the historic migration rate for the migrant’scommunity.31. Remittances sent during a household head’sabsence do not affect the likelihood of starting a newbusiness; rather the resources accumulate and areavailable as seed capital after an adjustment periodfollowing the migrant’s return. This delay may explainwhy contemporaneous surveys miss the business fundingeffect.32. Likewise, Yang (2004) finds no evidence thataggregate household consumption expenditures wereaffected at all by the remittance-inducing exchangerateshocks he studies, which contrasts with the significantpositive effects he finds for education spending,adult labor supply, and capital investments.33. On the other hand, the exogenous changes inremittance income that are used to identify the expenditurepropensities in studies such as Yang (2004) arelikely to be viewed by the household as temporary,leading the forward-looking households to investrather than consume. These estimates would then providea poor guide to the expenditure effects of policiesthat led to more sustained increases in remittanceflows—for example, policies that permanently lowerthe cost of sending remittances.34. Bourguignon (2003) uses a highly flexible functionalform with multiple interactions between the keyvariables to estimate the relationship between povertyreduction, growth, and changes in inequality. Here wemake stronger assumptions about the functional formof the relationship as a first pass in estimating thepoverty-reducing effect in a relatively simple povertyreductionmodel.35. We actually take a slightly less restrictive versionof this equation to the data by allowing for an interceptand allowing the coefficients on the two explanatoryvariables to differ. We also allow for the change in inequalityto enter in separately, but the coefficient onthis variable is insignificantly different from zero.36. Observations where the initial and/or finalpoverty measure for the interval is zero are excluded.We also exclude observations from the Eastern Europeand Central Asian region due to concerns aboutcomparable measurement during post-communisttransition.37. The growth in per capita income is given by∆Y/Y. Denoting per capita remittances as R, and lettingthe absolute change in per capita income equal the levelof per capita remittances (that is, ∆Y R), then thegrowth rate of income due to remittances is simplyR/Y. Since we use mean survey income/consumption asour measure of per capita income, R/Y is convenientlycalculated as (R/Y GDP ) (Y GDP /Y), where the firstterm is equal to remittances as a share of GDP, and thesecond term is the ratio of per capita GDP to meansurvey income.ReferencesAdams, Richard. 1989. “Worker Remittances and Inequalityin Rural Egypt.” Economic Developmentand Cultural Change 38(1): 45–71.———. 1991. “The Economic Uses and Impact of InternationalRemittances in Rural Egypt.” EconomicDevelopment and Cultural Change 39(4):695–722.131

GLOBAL ECONOMIC PROSPECTS 2006———. 1998. “Remittances, Investment, and RuralAsset Accumulation in Pakistan. Economic Developmentand Cultural Change 41(1): 155–73.———. 2004a. “Remittances and Poverty inGuatemala.” Policy Research Working Paper3418. World Bank, Washington, DC.———. 2004b. “Economic Growth, Inequality andPoverty: Estimating the Growth Elasticity ofPoverty.” World Development 32(12): 1989–2014.———. 2005a. “International Remittances and theHousehold: Analysis and Review of Global Evidence.”Paper presented at the Plenary Sessionof the African Economic Research Consortium,May 29, Nairobi.———. 2005b. “Remittances and Poverty in Ghana.”Unpublished paper. World Bank, Washington, DC.———. 2005c. “Remittances, Household Expenditureand Investment in Guatemala. In InternationalMigration, Remittances, and the Brain Drain, ed.Caglar Ozden and Maurice Schiff. Washington,DC: World Bank.Adams, Richard, and John Page. 2005. “Do InternationalMigration and Remittances ReducePoverty in Developing Countries?” World Development33(10): 1645–69.Adelman, I., and J. E. Taylor. 1992. “Is StructuralAdjustment with a Human Face Possible?” Journalof Development Studies 26: 387–407.Alesina, Alberto, and Dani Rodrik. 1994. “DistributivePolitics and Economic Growth.” QuarterlyJournal of Economics 109(2): 465–90.Banerjee, Abhijit, and Andrew Newman. 1993.“Occupational Choice and the Process of EconomicDevelopment.” Journal of Political Economy101(2): 274–98.Barham, Bradford, and Stephan Boucher. 1998. “Migration,Remittances, and Inequality: Estimatingthe Net Effects of Migration on Income Distribution.”Journal of Development Economics 55:307–31.Bourguignon, Francois. 2003. “The Growth Elasticityof Poverty Reduction: Explaining the Heterogeneityacross Countries and Time Periods.” In Inequalityand Growth: Theory and Policy Implications,ed. Theo Eicher and Steven Turnovsky.CES-IFO Seminar Series. Cambridge, MA: MITPress.Bourguignon, Francois, and Pierre-Andre Chiappori.1992. “Collective Models of Household Behavior:An Introduction.” European EconomicReview 36: 355–64.Browning, Martin, Francois Bourguignon, Pierre-Andre Chiappori, and Valerie Lechene. 1994.Journal of Political Economy 102(6): 1067–96.Chami, Ralph, Connel Fullenkamp, and Samir Jahjah.2005. “Are Immigrant Remittance Flows aSource of Capital for Development?” IMF StaffPapers 52(1): 55–81.Clarke, George, and Scott Wallsten. 2004. “Do RemittancesProtect Households in Developing CountriesAgainst Shocks? Evidence From a NaturalDisaster in Jamaica.” Unpublished paper. WorldBank, Washington, DC. November.Coate, Stephen, and Martin Ravallion. 1993. “Reciprocitywithout Commitment: Characterizationand Performance of Informal Insurance Arrangements.”Journal of Development Economics 40:1–24.Cooper, Russell. 1999. Coordination Games.Cambridge: Cambridge University Press.Cox Edwards, Alejandro, and Manuelita Ureta. 2003.“International Migration, Remittances, andSchooling: Evidence from El Salvador.” Journal ofDevelopment Economics 72(2): 429–61.De, Prabal, and Dilip Ratha. 2005. “Remittance Incomeand Household Welfare: Evidence from SriLanka Integrated Household Survey.” Unpublishedpaper. Development Research Group,World Bank, Washington, DC.de Brauw, Alan, J. Edward Taylor, and Scott Rozelle.2001. “Migration and Incomes in Source Communities:A New Economics of MigrationPerspective from China.” Unpublished paper.Department of Agricultural and Resource Economics,University of California-Davis. October.www.agecon.ucdavis.edu/aredepart/facultydocs/Taylor/migration_income.pdf.Duflo, Esther. 2003. “Grandmothers and Granddaughters:Old-Age Pensions and Intrahousehold Allocationin South Africa.” World Bank EconomicReview 17(1): 1–25.Duflo, Ester, and Christopher Udry. 2004. “IntrahouseholdResource Allocation in Côte D’Ivoire:Social Norms, Separate Accounts, and ConsumptionChoices.” NBER Working Paper 10498.National Bureau of Economic Research,Cambridge, MA.Ellerman, David. 2003. “Policy Research on Migrationand Development.” World Bank Policy ResearchWorking Paper 3117. August.Fafchamps, Marcel. 2004. Market Institutions inSub-Saharan Africa: Theory and Evidence.Cambridge, MA: MIT Press.Foster, James, Joel Greer, and Erik Thorbecke. 1984.“A Class of Decomposable Poverty Measures.”Econometrica 52(3): 761–66.Funkhouser, Edward. 1992. “Migration fromNicaragua: Some Recent Evidence.” World Development20(3): 1209–18.132

REMITTANCES, HOUSEHOLDS, AND POVERTYGustafsson, Bjorn, and Negatu Makonnen. 1993.“Poverty and Remittances in Lesotho.” Journal ofAfrican Economies 2(2): 49–73.Hanson, Gordon, and Christopher Woodruff. 2003.“Emigration and Educational Attainment inMexico.” Working Paper. University of California-San Diego.Hausmann, Ricardo, and Dani Rodrik. 2002. “EconomicDevelopment as Self Discovery.” NBERWorking Paper 8952. National Bureau of EconomicResearch, Cambridge, MA.Hildebrandt, Nicole, and David McKenzie. 2005.“The Effects of Migration on Child Health inMexico.” Policy Research Working Paper 3573.World Bank, Washington, DC.IADB-MIF (Inter-American Development Bank, MultilateralInvestment Fund). 2004. “Sending MoneyHome: Remittance to Latin America and theCaribbean.” Washington, DC. May.IMF (International Monetary Fund). 1995. “MigrantRemittances, Labor Markets, and HouseholdStrategies: A Comparative Analysis of Low-Income Household Strategies in the CaribbeanBasin.” Social Forces 74(2): 633–55.IMF. 2005. World Economic Outlook.Jacobs, Jane. 1984. Cities and the Wealth of Nations:Principles of Economic Life. New York: RandomHouse.Layard, Richard, Stephan Nickell, and RichardJackman. 1991. Unemployment: MacroeconomicPerformance and the Labour Market. Oxford:Oxford University Press.Lucas, Robert. 1987. “Emigration to South Africa’sMines.” American Economic Review 77(3):313–30.Lucas, R.E.B. 2004. “International Migration to HighIncome Countries: Some Consequences for EconomicDevelopment in the Sending Countries.”Unpublished paper. Boston University.Lucas, Robert, and Oded Stark. 1985. “Motivationsto Remit: Evidence from Botswana.” Journal ofPolitical Economy 93: 901–18.Martin, Philip, and Thomas Straubhaar 2002. “BestPractices to Reduce Migration Pressures.” InternationalMigration 40(3): 5–23.Massey, Douglas, and Emilio Parrado. 1998. “InternationalMigration and Business Formation inMexico.” Social Science Quarterly 79(1):1–20.McKenzie, David, and Hillel Rapoport. 2004. “NetworkEffects and the Dynamics of Migration andInequality: Theory and Evidence from Mexico.”BREAD Working Paper 063. Harvard University,Cambridge, MA. www.cid.harvard.edu/bread/papers2.htm.Miguel, Edward, and Michael Kremer. 2001. “Worms:Education and Health Externalities in Kenya.”NBER Working Paper 8481. National Bureau ofEconomic Research, Cambridge, MA.Papademetriou, Demetrious, and Philip Martin. 1991.“Introduction.” In The Unsettled Relationship:Labor Migration and Economic Development,ed. Demetrious Papademetriou and Philip Martin.New York: Greenwood.Morduch, Jonathan. 1994. “Poverty and Vulnerability.”American Economic Review 84(2): 221–25.Moretti, Enrico. 2003. “Human Capital Externalitiesin Cities.” NBER Working Paper 9641.National Bureau of Economic Research, Cambridge,MA.Paulson, Anna. 2000. “Insurance Motives for Migration:Evidence from Thailand.” Unpublishedpaper. Northwestern University.Persson, Torsten, and Guido Tabellini. 1994. “Is InequalityHarmful for Growth.” American EconomicReview 84(3): 600–21.Ravallion, Martin. 1997. “Can High-Inequality DevelopingCountries Escape Absolute Poverty?” EconomicLetters 56: 51–57.Rosenzweig, Mark R. 1988. “Risk, Implicit Contracts,and the Family in Rural Areas of LowIncome Countries.” Economic Journal 98(393):1148–70.Rosenzweig, Mark R., and Oden Stark. 1989. “ConsumptionSmoothing, Migration, and Marriage:Evidence from Rural India.” Journal of PoliticalEconomy 97(4): 906–26.Seabright, Paul. 2004. The Company of Strangers: ANatural History of Economic Life. Princeton:Princeton University Press.Sen, Amartya. 1973 (rev. 1997). On Economic Inequality.Oxford: Clarendon.———. 1976. “Real National Income.” Review ofEconomic Studies 43(1): 19–39.Stark, Oded, and Robert Lucas. 1988. “Migration,Remittances, and the Family.” Economic Developmentand Cultural Change 36(3): 465–81.Stark, Oded, J. Edward Taylor, and Shlomo Yitzhaki.1986. “Remittances and Inequality.” EconomicJournal 96(383): 722–40.Swaroop, Vinaya, and Shantayanan Devarajan. 2000.“The Implications of Foreign Aid Fungibility forDevelopment Assistance.” In The World Bank:Structure and Policies, ed. Christopher Gilbertand David Vines. Cambridge: Cambridge UniversityPress.Taylor, J. Edward. 1992. “Remittances and InequalityReconsidered: Direct, Indirect, and IntertemporalEffects.” Journal of Policy Modeling 14(2):187–208.133

GLOBAL ECONOMIC PROSPECTS 2006———. 1999. “The New Economics of Labor Migrationand the Role of Remittances.” InternationalMigration 37(1): 63–86.Taylor, J. Edward, and Philip Martin. 2001. “HumanCapital: Migration and Rural PopulationChange.” In Handbook of Agricultural Economics,ed. Bruce L. Gardner and Gordan C. RausserNew York: Elsevier Science.Taylor, J. Edward, Jorge Mora, and Richard Adams.2005. “Remittances, Inequality, and Poverty: Evidencefrom Rural Mexico.” Research Programon International Migration and Development.DECRG. Mimeo. World Bank.Taylor, J. Edward, and T. J. Wyatt. 1996. “The ShadowValue of Migrant Remittances, Income, and Inequalityin a Household-Farm Economy.” Journalof Development Studies 32(6): 899–912.Thaler, Richard. 1990. “Anomalies: Saving, Fungibility,and Mental Accounts.” Journal of EconomicPerspectives 4(1): 193–205.Wodon, Quentin, Diego Angel-Urdinola, Gabriel Gonzalez-Konig,Diana Ojeda Revah, and CorinneSiaens. 2002. “Migration and Poverty in Mexico’sSouthern States.” Regional Studies Program, Officeof the Chief Economist for Latin America andthe Caribbean, World Bank, Washington, DC.Woodruff, Christopher, and Rene Zenteno. 2001.“Remittances and Microenterprises in Mexico.”Unpublished paper. Graduate School of InternationalRelations and Pacific Studies, University ofCalifornia–San Diego.World Bank. 2005. “Remittances in Europe and CentralAsia: Size, Distribution, and Impact.” ECAMigration Study, Europe and Central Asia Region,World Bank, Washington, DC.Yang, Dean. 2004. “International Migration, HumanCapital, and Entrepreneurship: Evidence fromPhilippine Migrant’s Exchange Rate Shocks.” ResearchProgram on International Migration andDevelopment. DECRG. Policy Research WorkingPaper 3578. World Bank.———. 2005. “Coping with Disaster: The Impact ofHurricanes on International Financial Flows,1970–2001.” Unpublished paper. Gerald R. FordSchool of Public Policy, University of Michigan.Yang, Dean, and HwaJung Choi. 2005. “Are RemittancesInsurance? Evidence from Rainfall Shocksin the Philippines.” Research Program on InternationalMigration and Development. DECRG.Mimeo. World Bank.Yang, Dean, and Claudia Martinez. 2005. “Remittancesand Poverty in Migrants’ Home Areas:Evidence from the Philippines.” In InternationalMigration, Remittances, and the Brain Drain, ed.Caglar Ozden and Maurice Schiff. Washington,DC: World Bank.134

6Reducing Remittance FeesFollowing the discussion in chapters 4 and 5,on the macro- and microeconomic importanceof remittances, chapter 6 focuses on a specificpolicy issue: reducing remittance costs andstrengthening the financial infrastructure thatsupports remittances. Reducing the cost ofpersonal remittances is the most promisingarea of policy intervention for several reasons.First, it will stanch a drain on the resources ofpoor migrants and their families back home.Second, it will increase flows through formalchannels, especially banks. Third, it willimprove financial access for the poor in developingcountries.Cost is usually not an issue in large remittances(made for the purpose of trade, investment,or aid), because, as a percentage of theprincipal amount, it tends to be small. But forsmall, personal transfers, remittance costs arehigh—unnecessarily so. Providers of remittanceservices in the formal sector typicallycharge a fee of 10–15 percent of the principalamount to handle the small remittances typicallymade by poor migrants. 1 High fees placea financial burden on the migrant remittersand on the recipients of the remittances, whoreceive a smaller amount of the much-neededfunds sent by their family members.Major international banks tend to focus onlarge-value remittance services rather than onservices for migrants. Poor migrants may feeluneasy about using a major bank for remittanceservices; they tend to prefer smaller financialinstitutions, money transfer operators(MTOs), or informal channels, such as afriends, family members, export-import firms,and transport companies.The main messages of this chapter are asfollows:There is significant scope to reduce the fees onremittance services, especially for the smalltransfers typically made by poor migrants.Remittance transaction costs are often significantlylower than the fees that most customerspay. Reducing transaction fees will increasethe disposable income of poor migrants andincrease the incentives to remit (as the net receiptsof beneficiaries increase). It may alsosignificantly increase annual remittance flowsto developing countries. Cross-border paymentsfor retail trade, investment, and pensionbenefits (typically defined in foreign currency)would also increase in response to a reductionin remittance fees.A weak competitive environment in the remittancemarket, lack of access to technologysupportingpayment and settlement systems,and burdensome regulatory and compliancerequirements all tend to keep fees high. Competitionin the remittance market could be increasedby lowering capital requirements onremittance services and opening up postal,banking, and retail networks to nonexclusivepartnerships with remittance agencies. Disseminatingdata on remittance fees and establishinga voluntary code of conduct for fairtransfers would improve transparency in135

GLOBAL ECONOMIC PROSPECTS 2006remittance transactions. In countries withexchange controls, efforts to align the officialand the market exchange rates would reducethe foreign-exchange spread in remittancetransactions.Reducing remittance costs and improvingaccess to the financial system for migrants andtheir families will do more to encourage theuse of formal channels than will regulation ofso-called informal services. While regulation isnecessary for curbing money laundering andterrorist financing, overregulation of informalservices may conflict with the objective ofreducing remittance costs.Expanding migrants’ access to bankingservices may enable remitters to bundle remittancesand thereby take advantage of thelower fees available for larger remittances.This would require expanding the bankingnetwork, allowing domestic banks from origincountries to operate overseas, providing identificationcards to migrants, and facilitatingthe participation of savings banks, creditunions, and microfinance institutions inproviding low-cost remittance services. Remittances,in turn, can be used to support financialproducts—such as deposits, loans, andinsurance—for poor people, and to contributeto the financial development of the recipienteconomy.The plan of this chapter is as follows. Thefirst section describes remittance fees andcosts in various remittance channels. It shows(a) that remittance fees paid by customers arehigh for smaller transactions, especially inlow-volume corridors; (b) that the cost ofproviding remittance services need not be sohigh; and (c) that remittance flows to developingcountries would increase if remittancecosts were reduced. The next section examinesthe factors underpinning remittancefees—market competition, regulations, paymentinfrastructure, and technology—andsuggests policies for reducing costs and fees.This section also briefly discusses the recommendationsof the international FinancialAction Task Force (FATF) to prevent misuseof remittance systems for criminal purposes.The last section discusses complementaritiesbetween remittances and other financialproducts such as loans, deposits, and insurance.Finally, an annex to the chapter brieflydescribes the historical evolution of threemajor remittance service providers (WesternUnion, MoneyGram, and Bank of America)to provide a perspective on the remittancemarket.Remittance fees and costsThe remittance industry consists of formaland informal fund transfer agents. Majorcompetitors include a few large global players,such as the major money transfer operators(MTOs) and banks, as well as hundreds ofsmaller participants that serve niche markets inspecific geographic remittance corridors. Theinformal fund transfer agents include friends,family, and unregistered MTOs such as hawaladealers and trading companies.The price of a remittance transactionincludes a fee charged by the sending agent(typically paid by the sender when initiatingthe remittance transaction) and a currencyconversionfee for delivery of local currency tothe beneficiary in another country. (A stylizedremittance transaction is presented inannex 6A.1.) Some smaller MTOs require thebeneficiary to pay a fee to collect remittances,presumably to account for unexpectedexchange-rate movements. In addition, remittanceagents (especially banks) may earn anindirect fee in the form of interest (or “float”)by investing funds before delivering them tothe beneficiary. The float can be significant incountries where overnight interest rates arehigh. 2 Many recipients spend considerabletime and travel considerable distances tocollect remittances. These costs typically arenot included in the price.Remittance fees are high, regressive,and nontransparentRemittance fee pricing is complex, and rarelyare senders informed about the full and136

REDUCING REMITTANCE FEESprecise price of a remittance transaction. Feesmay be as high as 20 percent of the principal,depending on the remittance amount, channel,corridor, and transaction type. The averageprice is reported to have been around12 percent of the principal in 2004 (Taylor2004; Kalan and Aykut 2005). Prices are believedto have declined recently but are stillvery high in low-volume corridors. Currencyconversioncharges are even less transparentthan remittance fees; they, too, vary dependingon the competitor, corridor, and channel,ranging from no charge in dollarizedeconomies to 6 percent or more in somecountries (Orozco 2004; Hernández-Coss2004; Kalan and Aykut 2005).Major MTOs such as Western Union andMoneyGram apparently charge higher remittancefees than banks and other financialinstitutions that offer remittance servicesto attract migrant customers (table 6.1).Informal channels such as hawala are reportedto be cheaper than formal services. Someheavily traveled remittance corridors, such asUnited States–Mexico and South Africa–Mozambique, are much cheaper than others.Urgent transactions delivered in minutes costmuch more than next-day transfers, and electronictransfers cost more than bank checks ordrafts, because they also clear much fasterthan the latter.The fee amount also depends on the remittanceamount. Average remittance fees, as apercentage of money sent, decline rapidly asthe transaction size increases, indicating scaleeconomies and the potential advantage ofbundling remittances—that is, the advantageof sending more funds, but less frequently.According to one firm’s fee schedule, the costof sending money from Belgium to Africadrops from 21 percent to below 4 percentas the transaction amount increases from40 euros to 900 euros (figure 6.1). Similarly,the cost of remittances from the United Statesto Mexico (through the major MTOs) is morethan 10 percent for $100, but less than 3 percentfor $500 (figure 6.2).In recent years, remittance fees have declinedin high-volume corridors in response toseveral factors. First, global and regionalMTOs have intensified their competitionin mature corridors (United States–LatinAmerica, for example), as new competitorshave been attracted by high and growingremittance volumes. In the United States–Mexico corridor, for example, remittance feeshave dropped nearly 60 percent since 1999(box 6.1). 3 Second, Bank of America andTable 6.1 Approximate cost of remitting $200Percent of principal amountMajor MTOs Banks Other MTOs HawalaBelgium to Nigeria* 12 6 9.8 —Belgium to Senegal* 10 — 6.4 —Hong Kong, China, to the Philippines 4.5 — — —New Zealand to Tonga ($300) 12 3 8.8 —Russia to Ukraine 4 3 2.5 1–2South Africa to Mozambique — 1 — —Saudi Arabia to Pakistan 3.6 0.4 — —United Arab Emirates to India 5.5 5.2 2.3 1–2United Kingdom to India 11 6 — —United Kingdom to the Philippines — 0.4–5.0 — —United States to Colombia — 17 10 —United States to Mexico 5 3 4.7 —United States to Philippines 1.2–2.0 0.4–1.8 — —Source: Brocklehurst 2004; Orozco 2004; Gibson, McKenzie, Rohorua 2005; Hernandez-Coss 2004; Ratha and Riedberg 2005;Kalan and Aykut 2005; Andreassen and others 2005.*World Bank survey of African diasporas in Belgium.Note: Figures do not include currency-conversion charge.— Data not available.137

GLOBAL ECONOMIC PROSPECTS 2006Figure 6.1 Remittance costs are high andregressive2140751310Percent of principal amount98150225300737565Principal amount, €4 4 4 435607459301120130514901860Source: Western Union branches in Brussels and Paris.policies to improve transparency in remittancetransactions (as in the United Kingdom), providefinancial training to migrants (as in thePhilippines), and establish bilateral initiatives(such as the Partnership for Progress betweenthe United States and Mexico) have helpedreduce remittance costs.These positive developments remain theexception. In most corridors, particularlythe low-volume corridors, remittance feescontinue to be very high. In the NewZealand–Tonga corridor, for example, feesare about three times as high as those in theUnited States–Mexico corridor. The wide gapbetween remittance fees and costs shows thatboth should be reduced.other banks in source countries are using minimaltransfer fees to attract migrant accounts,while a growing number of banks in recipientcountries (including ICICI and Bancomer) arecompeting for remittance customers. Third,the use of Internet-based technology for messagingand advanced clearing and settlementhas reduced the cost of remittance transactions.In some countries, new remittance toolshave emerged, based on cell phones (seebox 6.6) and smart cards. Finally, governmentThe cost of a remittance transactionappears to be far lower than the priceService providers’ remittance costs appear tobe much less than the fees charged to customers.Domestic transfer fees are only a fractionof the cross-border remittance fees (net ofthe currency-conversion charge). The cost of adomestic automated clearinghouse (ACH)payment in the United States is one-third ofa cent. Domestic transfers using Visanet cost2 cents per transaction, as opposed to 51 centsFigure 6.2 Remittance fees in the United States–Mexico corridorFees as % of remittance16141210864Western UnionMoneygramVigoDolex20100200 300 400 500 600 700 800 900 1,000Size of remittance, $Source: Kalan and Aykut 2005.138

REDUCING REMITTANCE FEESBox 6.1 Decline in remittance costs in the UnitedStates–Mexico corridorThe cost of sending $300 from the United Statesto Mexico declined nearly 60 percent between1999 and 2005—from $26 to $11 (according toPROFEC data, see figure). The decline can be tracedto greater competition. Prices generally remainedstagnant when one MTO dominated the transmissionservice through exclusive contracts with distributors.Practitioners inside the industry cite thebreakup of these exclusivity contracts and the entryof new competitors—especially banks—into thecorridor as key events leading to a steady decline inprices.Starting with Citibank’s acquisition of Banamexin 2001—a $12.5 billion deal reportedly motivatedby the attractiveness of Banamex’s remittance business,U.S. banks have increased their stake in theUnited States–Mexico remittance market in the pasttwo years. Being able to use the matricula consularidentification card to establish identity when openingan account has helped this process. The card isUnited States–Mexico, remittance fee per $300Dollars282522191613101999Cost(left scale)2000 2001Billions of dollarsRemittancesto Mexico(right scale)2002 2003 2004 2005accepted as a valid identity document in 32 U.S.states, more than 1,000 police stations, 409 cities,125 counties, and 280 banking institutions.2015105per transaction for international transfers(Brocklehurst 2004). In some corridors, feesfor international remittances are as low as$1.80 per transaction (London-Manila),which hints at a falling lower bound for thecost of remittances. The fact that some bankshave been offering free remittance services asloss-leaders to attract new business suggeststhat the actual cost of remittances is modest.Courier services that offer remittances alsocharge small fees for this additional service.Finally, industry cost estimates as well asother calculations presented below suggestthat remittance costs are not very high.The cost of providing remittance servicesvaries with the business model used by the serviceprovider. Western Union, MoneyGram,and Vigo use agents who pay all operatingcosts in exchange for their franchise and acommission on sales. In the “branch” modelused by Dolex and many of the smallerregional MTOs, the fixed and operating costsassociated with each branch are paid by theMTO. By leveraging existing businesses on acommission basis, the agency model is muchless capital-intensive than the branch modeland can be expanded rapidly through partnerships,but it has higher variable costs. 4 In bothmodels, relatively high fixed costs are associatedwith transaction-processing operations,compliance with regulatory requirements,marketing, and administration. 5Data on MTOs’ costs of providing remittanceservices are hard to obtain. However, ananalysis of profitability of the market leaders 6using publicly available financial statementssuggests that remittance costs are significantlylower than the fees charged to customers.Western Union has sustained operatingmargins that are at least 50 percent higherthan other MTOs and industry peers in thepayments and electronic processing market139

GLOBAL ECONOMIC PROSPECTS 2006(table 6.2). 7 Its operating profit per remittancetransaction may have averaged $8 to $9 in2004. This is consistent with an earlier annualreport (Western Union 2000) that put thecompany’s operating profit at $684 million(or 30 percent of its $2.3 billion revenue). Theoperating profitability of the other major marketplayers (MoneyGram and Dolex) has beenin the range of 15–20 percent (table 6.2). Avery simple model for Western Union (whichassumes that agency commission costs are 35percent of revenues after deduction of fixedcosts and that all other costs are fixed costs)suggests that average transaction fees could bereduced by as much as one-third while maintainingoperating margins within the samerange as those of other major MTOs and peers.Reducing these operating profits to zerowould provide a rough estimate of the breakevencost for these firms. Such an exercise revealsthat the break-even fee for WesternUnion is probably around $9 per transactionand would fall below $5 if the volume oftransactions were to double (box 6.2). Althoughit would be unreasonable to suggestthat any company reduce its prices to cost, thissimple model does appear to indicate that thereis considerable latitude for reductions in transactionfees within the higher-priced corridors.A more direct way of estimating the cost ofa remittance transaction in a hypotheticalMTO is to add up plausible cost components,Table 6.2 Operating profits of major MTOsPercentage of revenue2004First Data Corporation, Western Unionmoney transfer operations a 32MoneyGram money transfer operations 15Global Payments money transfer operations, 20including Dolex bPeer group average b 18Source: Yahoo Finance Database financial summaries; Kalanand Aykut 2005; Piper Jaffray Equity Research; MoneyGramInternational.a. 90 percent Western Union; 50–55 percent non-UnitedStates/Canada consumer-to-consumer money transfers.b. Includes American Express, Total System Services, DSTSystems, Sunguard Data Systems, and Fiserv. These companiesare not directly comparable to the MTOs as they are notnecessarily in the money transfer business.such as staff to process the transaction andprovide security, rental of the premises, fixedcosts (including franchise licensing), the costof network and technology, and administrativecosts for regulatory compliance. 8 Thismethodology yields a cost estimate of $5.50for the first remittance transaction (table 6.3).Because most remittance transactions tendto be repetitive—the same amount is remittedfrom the same location to the samebeneficiary—the cost for subsequent transactionsdrops to $3.60 (less staff time is required).It drops to under $3 per transaction ifelectronic processing is used.Admittedly, the calculations in table 6.3are based on a theoretical model of a basicTable 6.3 Estimating the cost of a remittance transactionCost in dollarsFirst Subsequent Electronictransaction transaction processing ExplanationSending staff 2.50 0.83 0.50 10 minutes of staff time at $15 per hourReceiving staff 0.17 0.17 0.17 10 minutes of staff time at $1 per hourFixed costs 0.27 0.27 0.27 $40 million system cost recovered over10 years; 2,000 branches with20 transactions per dayIT, telecommunications 0.60 0.60 0.60 1 minute international phone callRent 1.50 1.50 1.50 $30 rent per day; 20 transactions per dayAdministrative costs 0.50 0.50 0.50 Compliance, general overheadTotal costs 5.54 3.60 2.94Source: Ratha and Riedberg 2005.140

REDUCING REMITTANCE FEESBox 6.2Estimating remittance industry costsRemittance industry costs are difficult to obtain.Isolating the cost of remittance services is difficultin the case of financial institutions that provideother services as well. Estimating costs is not easyeven in the case of dedicated remittance serviceproviders because of the differences in the qualityand reliability of remittance services (only someproviders give customers legal redress). In Remittanceindustry costs, therefore, we have used publiclyavailable information on Western Union, the largestMTO that is also a publicly listed company.We used a simple model to estimate a break-evenfee for Western Union’s international money transferoperations. The model suggests that for WesternUnion’s operating margins on its international moneytransfers to drop to the peer group average of 17.8percent (table 6.2), the average transaction fee wouldhave to be lowered from $22.90 to $15.30 (column 2of the table below)—very close to the company’s currentfee in several U.S. corridors. The model also indicatesthat the break-even fee at which the operatingprofit becomes zero is $9.30 (column 3). This price isin the same range as MoneyGram’s standard flatprice in the U.S. corridors. A sensitivity analysis usingthis model suggests that the break-even fee would be$6.50–$7.00 if agency commissions were 25 percent,and around $11 if commissions were 45 percent.Dollars108642075Western Union: Operating profitbreak-even price vs. volume100 125 150 175 200Source: Kalan and Aykut 2005.Transaction volume (millions)225 250 275 300The figure illustrates how the break-even feeshown in the table decreases as the number of transactionsincreases. If transaction volume doubledfrom the current 76 million to 150 million, thelowest fee at which the international operationwould remain profitable would be $4.74.Calculation assuming Calculationpeer group marginassuming2004 data of 18% break-even marginOperating margin (operating profit 30 18 0over revenue) (%)Operating profit per transaction 8.8 3.9 0.0(revenue minus costs) ($)Costs ($) 20.4 17.7 15.7Agency commission, 35% of fee 8.0 5.3 3.3Fixed costs 12.4 12.4 12.4Revenue ($) 29.3 21.6 15.7Foreign-exchange commission 6.4 6.4 6.4Fee 22.9 15.3 9.3Source: Western Union financial statement for 2004.Note: Reflects 76 million transactions in 2004. Fixed costs include marketing, administration, depreciation, and amortization,agency start-up, and other unidentified costs. Figures may not add up due to rounding errors.141

GLOBAL ECONOMIC PROSPECTS 2006remittance transaction that does not capturethe global network and diversified servicesprovided by major MTOs. Moreover, themodel’s assumptions are subject to considerableuncertainties, the greatest of which is thataverage costs would be higher if the number oftransactions were smaller. It is worth noting,however, that many independent agents provideremittances as a side business: for them,fixed and variable costs could be significantlylower than for dedicated remittance serviceproviders. Indeed, there may be a case for providingfree remittance services in order todraw customers for other products and services,as practiced by certain banks.Remittance costs should continue to fallunder the influence of increased competitionand better technology. Large MTOs mayhave considerable latitude to reduce feeswhile maintaining reasonable profit margins.In corridors where costs have already fallensignificantly, further decline may be modest;but elsewhere there is scope for significant decline,especially with the volume of transactionsrising rapidly.Reducing remittance fees will increaseremittance flows to developing countriesReducing remittance fees would increase thedisposable income of remitters, encouragingthem to remit more. It also might encouragesmaller and more frequent remittances.And lower prices in a particular channelmight encourage remitters to shift from otherchannels—notably informal ones.The degree to which a fee reduction wouldresult in an increase in flows depends on thepurpose of the remittance. At one extreme,where the purpose is to meet a specific need—payment for tuition, a medical emergency, asocial ceremony, or the purchase of a giftitem—the amount of remittance may not besensitive to the remittance fee. At the otherextreme, remittances by a poor, cash-strappedremitter may be highly cost elastic. Similarly,remittances meant for investment are likely tobe cost elastic. In reality, most remittancetransactions fall between these two extremes.Even when remittances are driven by altruism,they will tend to be cost elastic, as evidencedby the literature on charity, which shows thatpeople tend to donate more as the cost ofdonating declines (box 6.3).In a recent survey of Senegalese migrants inBelgium, two-thirds of the migrants said theywould send more if the cost of sending wentdown. In a survey of Tongan migrants in NewZealand, 30 percent of remitters said theywould increase the amount of remittances by0.74 percent (on average) if costs fell by 1 percent(Gibson, McKenzie, and Rohorua 2005).That survey found the overall cost-elasticity ofBox 6.3to costEven charitable donations are sensitiveCharitable donations and bequests, like altruisticremittances, increase when the costs of such actionsdecline. In one of the best-known early studiesof the responsiveness of gift-giving to tax deductions,Feldstein and Taylor (1976) estimated a price elasticityof the amount given of 1.3. a Although this findinghas been challenged on the ground that giftgivingresponds more to temporary changes in thecost of giving (Glenday, Gupta, and Pawlak 1986),the general agreement is that people give more whenit costs less to do so (Cordes 2001). The literature oncharitable bequests reaches a similar conclusion.Bakija, Gale, and Slemrod (2003) estimate a priceelasticity of 2.14 for charitable bequests.a In the United States, taxpayers can deduct the amount ofcharitable contributions from their income for tax purposes.Thus, Feldstein and Taylor (1976) view an increase in the incometax rate as a decline in the price of charitable donations.142

REDUCING REMITTANCE FEESremittances with respect to the fee (averagingthe elasticity over those who would increaseremittances and those who would not) to be–0.22. Based on this estimate, Gibson and others(2005) calculate that lowering the fixedcost of sending money through banks andMTOs from New Zealand and Tonga to competitivelevels in the world market would resultin a 28 percent increase in remittancesfrom existing remitters. It might also inducesome nonremitters to start remitting. 9If the cost elasticity (–0.22) of the NewZealand–Tonga study were applicable to alldeveloping countries, a reduction inremittance cost from 12 percent to (say) 6percent could result in an 11 percent increasein annual remittance flows to developingcountries. One caveat to this calculation isthat the cost elasticity applies only to highcostcorridors, which also tend to have lowvolumes. In corridors where the remittancecost is already low, further decreases may notincrease flows. For example, a fee reductionby a major MTO may not produce much effectif a major part of the flows is alreadymoving through low-cost informal channels.This is confirmed by the World Bank surveyof Senegalese migrants in Belgium; half of therespondents who paid remittance fees of 20percent or more said they would send more ifcosts were halved; not even one-fourth ofthose who paid less than 10 percent said theywould send more (table 6.4). Almost 75 percentof the Senegalese migrants who sendmoney through the large MTOs said thatTable 6.4 Remittances are more costelasticwhen costs are higher% of respondents who wouldremit moreCost(% of principal) Senegal Nigeria1–9 23 6410–19 50 6720 and above 50 83Source: World Bank Survey of Senegalese and Nigeriandiasporas in Belgium.they would send more if the costs werelowered, a result confirmed by findings froma World Bank survey of the Nigerian diasporain Belgium.An indirect implication for cost elasticitymay be drawn from Yang’s (2004) finding ofan elasticity of 0.6 for remittance receiptsdenominated in Filipino pesos with respectto the peso–dollar exchange rate. Applyingthis elasticity to a remittance transaction of$150, if the remittance fee were halved from(say) 12 percent to 6 percent, remittance receiptswould rise by 3.6 percent, or $5.4,while the remittance fee would decline from$18 to $9.31. 10 If the same elasticity wereto apply to the entire flow of remittances todeveloping countries, remittance receipts, inresponse to a halving of costs would increasesignificantly, by more than $5 billionusing only recorded flows, and more than$8 billion using both recorded and unrecordedflows.Reductions in remittance fees would also belikely to increase other cross-border retailflows such as transfers from public and privateinstitutions to individual beneficiaries (pensions,child-care payments), small-value paymentsin exchange for goods and services, acquisitionsof assets, and debt servicing. 11 Inmore developed countries, migrant remittancesare only a small share of retail payments,which, in turn, are a fraction of wholesale payments.But in developing countries, especiallyin smaller and poorer countries, remittancesare a significant source of funding in relationto the size of the economy and, therefore, ofthe retail payment system. A reform of the retailpayment system to facilitate remittanceswould probably benefit other (not easily quantifiable)components of retail payments.Based on the evidence presented above,notably the finding that the cost elasticity ofremittances is negative, policies that aim tolower remittance costs by increasing accessto banking services, promoting competition,and disseminating information have thepotential to provoke sizeable increases inremittance flows to developing countries.143

GLOBAL ECONOMIC PROSPECTS 2006Factors underpinning highremittance feesFigure 6.3 Fees and foreign exchangespreads for $200 in Western Uniontransfers from New York City (percent)7.010.6Transfer feeLow-volumecorridorsHigh-volumecorridors2.91.8Foreign exchangespreadSource: Kalan and Aykut 2005.Total costWhat accounts for high transaction costsin the remittance business? Are they relatedto large sunk costs, regulatory measuresthat restrict competition, or the lack of accessto low-cost public infrastructure (such as paymentssystems)? Do exchange controls, countryrisks, or other specific factors keep the costof cross-border transfers higher than those ofdomestic transfers?Before answering these questions, it isworth noting some general findings from across-country regression analysis of remittancefees (Freund and Spatafora 2005). Remittancefees tend to be higher in corridors in whichbank concentration is high and competitionlow (as reported in chapter 4, table 4A.2.2).They tend to be lower in more developed financialsectors (proxied by the ratio of depositsto GDP) in the recipient country and in dollarizedeconomies and other economies that presentlow exchange-rate risk. Greater creditrisk 12 reduces the willingness of agents to provideremittance services. Finally, high wages atthe recipient end (as proxied by domestic output)are associated with more costly remittanceservices. These results, which should be treatedas indicative rather than conclusive, suggestthat measures to increase competition amongremittance-service providers and to reduce financialrisk and exchange-rate volatility arelikely to reduce the transaction costs associatedwith remittances. 13Several factors related to conditions in thecorridor and in the sending and receivingcountries have significant impacts on remittancepricing. Two corridor-related factorsthat have a significant impact on price are the(potential) level of competition in the corridorand special arrangements with postal systemsto handle distribution.A high level of competition may considerablyreduce remittance prices in a corridor.The level of competition in a corridorcan be proxied by remittance volume, sincehigh-volume corridors attract more competitors,particularly small niche players thatcompete primarily on price. The relativelylower prices in high-volume corridors, suchas United States–Mexico and Saudi Arabia–India, can be ascribed in part to the presenceof regional and smaller players, in addition tothe major MTOs (and in part to scale economies).The remittance prices of the global marketleader, Western Union, are significantlylower in the high-volume U.S. corridors thanelsewhere (figure 6.3). 14High volume in a corridor does not alwaysguarantee high competition, however. Exclusiveaccess to an extensive distributionalnetwork (such as a post office network) maydistort competition in the corridor. Using postoffices as money transfer agencies can give anMTO a significant advantage, because thepostal system almost always offers the most extensivedistribution networks in both sendingand receiving countries, particularly in ruralareas. Exclusive arrangements with postal systemshave been employed by the two largestMTOs, Western Union and MoneyGram.Exclusive arrangements can block orbar entry by small competitors and maythus allow the company that enjoys the arrangementto maintain a high price premium. Moreover,exclusive arrangements with the post office,typically a trusted and ubiquitouspresence, may facilitate price leadership and8.813.5144

REDUCING REMITTANCE FEESFigure 6.4 Exclusive arrangements withpost offices skew competitiona. Fees for $200 remittance, France–Morocco corridorPercent121086420Price premium, $11.7Western Union4.9MoneyGramb. Fees and foreign exchange spread for $200remittance, United States–Vietnam corridorPercent6543210c. Western Union price premium over MoneyGram for$200 remittance when Western Union has post office link16128405.0AllcorridorsHigh-volumecorridorsSource countryWestern Union5.0Transfer feeWith post office linkSource: Kalan and Aykut 2005.MoneyGram1.0Allcorridors3.0FX spreadHigh-volumecorridorsRecipient countryWithout post office linkprice signaling—thereby raising the fee structureacross the board. For example, arrangementswith the postal systems in France andMorocco may help explain Western Union’sability to charge significantly higher fees inthat corridor than its major competitor,MoneyGram (figure 6.4a). However, inthe United States–Vietnam corridor, whereMoneyGram has such an arrangement, WesternUnion lowers its price to the same level asMoneyGram’s, whereas MoneyGram chargeshigher foreign-exchange commissions (figure6.4b).At the global level, Western Union’s pricepremium over MoneyGram’s fees, even inhigh-volume corridors, appears to be significantlyhigher when the company has a linkwith the postal system in either the sending orreceiving country (figure 6.4c). 15 On the otherhand, when MoneyGram, which almostalways offers lower prices than WesternUnion, has the agency relationship with thepost office, its strong distribution advantageforces Western Union to lower its prices tocompete in the corridor.Other factors that appear to have animpact on corridor pricing include activeparticipation of banks, credit unions, or othernonbank financial institutions in the remittancemarket; cultural and geographic commonalitywith a group of countries thatincludes one highly competitive, high-volumecorridor with lower prices; the size of informaltransfer network in the corridor; and governmentpolicy initiatives within the corridor.Regulatory and policy decisions have asignificant effect on remittance costs. In a recentsurvey of providers of remittance servicesin the United States, 40 percent of those surveyedcited the process of getting a license asthe chief barrier to their operation, followed bybuilding a compliance system (figure 6.5). Inthe United States, remittance service providersare supervised by state departments of consumeraffairs and banking. Not all states havespecific regulations on remittances. And registrationrequirements vary widely from state tostate (annex 6A.2). To set up a money transfer145

GLOBAL ECONOMIC PROSPECTS 2006Figure 6.5 Barriers perceived by remittance service providers% of firms504040323027 2723 23201018.21614117 70Getting U.S.licenseCreating agentnetworkBuilding compliancesystemsGetting bank accountGetting U.S. bondFunding workingcaptialGetting licenseabroadBuilding consumertrustBarriers to entry cited by firmsAcquiring know-howTechnologyEstablishing commercialcontactsAccessing U.S.financial infrastructureSource: Andreassen 2005.business with offices in all 50 states would requirenet worth and bonds of more than $5million (Ratha and Riedberg 2005). Althoughbond and capital requirements protect consumersand deter fraudulent practices, the widevariation in requirements from state to state,mirrored in wide variances among countries,can be confusing and costly, thereby discouragingcompetition from new, smaller players.Many countries, including France, Italy, andthe Russian Federation, require a provider ofremittance services to be a fully licensed bankor financial institution. Only recently did Germanyallow remittances to be conducted undera financial institution license instead of underbanking regulations. Costly and stringent licensingrequirements, like bond and capital requirements,discourage the entry of smallerplayers that could provide effective competitionin many remittance corridors.Regulating informal remittancesmay raise costsSince the terrorist attacks of September 11,2001, authorities in many countries haveadopted more stringent regulations and steppedup enforcement of existing rules governing thetransfer of foreign exchange. 16 An increasingnumber of countries are requiring MTOs to registerwith the authorities and to report transactionson a regular basis. These regulatory requirementshave raised the cost of fundtransfers to the remittance service providers,which tend to pass them on to customers.National requirements center on the registrationof transfer businesses, application ofknow-your-customer procedures, detailedrecord-keeping, and frequent reporting.“Money service businesses” in the UnitedStates must maintain a list of their agents andmake the list available to the Financial CrimesEnforcement Network (FinCEN) upon request.Operating such a business without registeringit is a crime. The introduction of theUSA Patriot Act in late 2001 tightened theknow-your-client requirements for fund transfers.In addition, U.S. financial institutions arerequired to comply with the recommendationsof the international Financial Action TaskForce to Prevent Money Laundering (FATF146

REDUCING REMITTANCE FEES2005), which are incorporated into U.S. regulations,and to comply with the sanctions listmaintained by the Treasury Department’sOffice of Foreign Assets Control.Since early 2005, correspondent bank accountsof hundreds of money service businessesin the United States have been closed by banksfor fear that they may be targeted by authoritiesfor servicing customers regarded as “highrisk.”The wave of closures can be traced to aJune 2004 notice from the Office of the Comptrollerof Currency that “[s]ome nationalbanks also provide banking services to foreign[money service businesses], a line of businessthat can carry significant money launderingrisks.” 17 Clear guidance on how to assess risksand spot suspicious activity is lacking.Some argue that the users of informalremittance channels face a great risk of fraudand default. Requirements for bonds, capitalization,auditing, reporting, and disclosurecan shield consumers from excessive fees,fraud, or other losses. At the same time, trustand self-regulation, characteristics of informalremittance networks such as hawala,hundi, padala, fei chien, and others, haveproven effective in protecting customersagainst losses, although they are by no meansimmune to fraud. Moreover, their low cost,speed, reach, and convenience of informaldoor-to-door remittance services remainextremely competitive compared with the inefficienciesof formal operators (Ballard2005; El Qorchi, Maimbo, and Wilson 2003;Maimbo and Passas 2005). Law enforcementcases from all continents show that formaland informal remittance channels are bothsusceptible to criminal abuse. 18The regulatory regime governing remittancesmust strike a balance between curbingmoney laundering, terrorist financing, andgeneral financial abuse, and facilitating theflow of funds through efficient formal channels.Policies that encourage formal operatorsto imitate the best practices of informaltransfer systems will benefit poor migrants.Strengthening the formal remittance infrastructureby offering the advantages of lowcost, flexible hours, expanded reach and language,and increasing efforts to identify andregulate the unregulated sector, would effectivelyfacilitate remittance flows while preservingtheir integrity.Policies to reduce remittance costsMeasures to reduce remittance costsshould aim to improve the efficiency ofremittance transactions by (a) enhancing marketcompetition to reduce high profit margins;(b) helping remittance service providers’ accessto new payments technology; and (c) devisingways to encourage remitters to send largeramounts (table 6.5). As a way to enhancecompetition, governments can encouragepostal systems and other state-owned distributionalternatives to open their networks tomultiple MTO partnerships on a nonexclusivebasis. In addition, they should avoid overregulation,excessive monitoring, or reporting requirementsthat could drive out smaller competitorsthat lack the economies of scale toabsorb the cost of compliance.Developing a shared network would be apowerful way to increase competition. Cooperationon infrastructure and competitionin service provision would allow networkbenefits to accrue to the consumer. 19 Thetechnology required to set up a paymentprocessinginfrastructure with large capacityis no longer an expensive proposition. Afunctioning payment infrastructure could beextended to a new country at a minimal costand in a matter of weeks. 20 There have beensome attempts to set up shared networks inthe remittance-source countries (for example,the United States–Mexico FedACHsystem, box 6.4). Also some governmentsin remittance-receiving countries havefacilitated the establishment of payment networksthat are shared by savings banks,credit unions, and microfinance institutionsoperating in poor and remote areas (forexample, BANSEFI in Mexico 21 and ApexLink in Ghana). 22147

GLOBAL ECONOMIC PROSPECTS 2006Table 6.5 Policies to reduce costs, regulateinformal providers, and provide remittancelinkedfinancial servicesSourcecountryRecipientcountryReducing costsIncrease competition X XAvoid exclusive arrangements X XHarmonize regulation and capital Xrequirements (same policy forall players)Introduce and harmonize electronic Xpayment systems (card-basedproducts)Improve data on corridors X XVoluntary code of conduct X XBundling of transactions X XRegulating informal providersMake formal sector operations X Xmore convenient anduser friendlyImprove banking access X XLeveraging remittancesImprove banking access X XEncourage microfinance institutions X Xand credit unions to provideremittance servicesAnother way to address the issue of high feesin the remittance industry would be to developbest-practice guidelines for remittance serviceproviders. Several such guidelines have been issuedby Credit Union National Association,Inter-American Development Bank, and WorldSavings Bank Institute, which urge serviceproviderstodisclosefees,exchangerates,andthetime of delivery. At the end of 2004, the WorldBank and the Bank for Committee on Paymentand Settlement Systems (CPSS) set up a taskforce, with participation from the IMF, to developvoluntary principles for remittance serviceproviders, regulators, and supervisors for improvingtransparency in the market (box 6.5). 23Such guidelines would have to be voluntary.Central banks generally are not willing to imposesuch guidelines or to cap remittance feesand foreign-exchange commissions. A recentsurvey (de Luna Martinez 2005) revealed thatin only 9 of 40 countries—Brazil, Bulgaria, Indonesia,Pakistan, Philippines, Russian Federation,Thailand, Tunisia, and República Bolivarianade Venezuela 24 —did central bankseven have the legal power to do so. All 40 centralbanks indicated that even if they had thepower to limit fees, they would not do so, preferringto leave fee-setting to financial institutionsin response to market competition. 25Raising consumer awareness throughfinancial literacy efforts and publicizing informationon costs (as Mexican authorities havedone through the PROFECO initiative) willstrengthen competition among remittanceservice providers. In April 2005, Britain’sDepartment for International DevelopmentBox 6.4United States–Mexico FedACHUnited States–Mexico automated clearinghousewas created as part of the U.S.-Mexico Partnershipfor Prosperity to reduce the cost of sending remittancesbetween the two countries. In 2002, thecentral banks of the United States and Mexico undertooka cooperative effort to link their automatedclearinghouse (FedACH) systems. The cross-borderservice began operating in February 2004. Today, approximately23,000 payments are sent from theUnited States to recipients in Mexico through thischannel each month. The cost of a remittance transactionis just $0.67; the exchange-rate spread is only0.21 percent. Since July 2005, the cross-borderservice has made funds available to a recipient on thefirst business day after a payment is originated.But FedACH has not been as popular with thebanking community as was expected. It sufferedfrom some technical weaknesses—for example, insufficientcoding flexibility for certain remittances. Also,major international banks that earn significant remittancefees from their own proprietary payment systemsand from foreign-exchange commissions havebeen slow to join FedACH. And other cross-borderACHs between the United States and Canada andbetween the United States and Europe have had similardifficulty attracting participation from banks.148

REDUCING REMITTANCE FEESBox 6.5 The World Bank/CPSS task force on generalprinciples for international remittance systemsAt the end of 2004, the World Bank and the Bankfor International Settlements’ Committee forPayment and Settlement Systems (CPSS) convened atask force, with members from central banks ofsending and receiving countries, international financialinstitutions, and development banks, to addressthe need for international policy coordination in remittancesystems. The task force is expected to issuegeneral principles for international remittance systemsin the first half of 2006. The purpose of theprinciples is to promote a sound, efficient, andcompetitive market in remittance services. Therecommendations of the task force are expected tocover market environment, consumer protectionand transparency, market infrastructure, and publicpolicy.(DFID) launched a website that provides informationon remittance costs and options inseveral countries. 26Assisting remittance service providers toadopt new payment systems technology andinstruments would help lower their service costs.Some technologically advanced methods ofsending transfers already exist. Card-based instruments,such as stored value cards (similar tophone cards), credit cards, and debit cards, arenow frequently used to send remittances tourban locations that have access to card-processingmachines. Systems such as iKobo.comuse the Internet to make remittances. PayPal andother services move money between virtual accounts,although they do not (yet) focus on immigrants’transfers. Similar technology has beenadapted by an operator in the Philippines tosend fast—and reportedly cheap—remittancesusing a cell phone (box 6.6).Migrant workers need easier accessto the formal financial systemImproving migrant workers’ access to bankingservices could reduce transaction costs, and atthe same time, help to develop the financialsystem in the countries where remittances arereceived. Sending and receiving countries alikecould support migrants’ access to banking byproviding them with the means to establishtheir identity.In receiving countries, the factor that exertsthe greatest effect on remittance costs (and onthe choice of remittance channel) is the reach ofthe remittance agent’s distribution network. Recipientsin rural areas underserved by banks mayhave to pay high costs for receiving remittances,especially through formal channels. Partnershipsbetween remittance operators and institutionsthat have wide networks in rural areas (such aspost offices) would help reduce such costs. 27 Incountries where residents are allowed to holdforeign currency deposits, permission to deliverremittances in U.S. dollars (or the same foreigncurrency sent by the remitter) would significantlyreduce (if not eliminate) the exchange ratespread on remittances (as seen in the case of dollarizedeconomies such as El Salvador).Remittances and financialinstitutionsBanks and smaller financial institutions,such as credit unions and microfinance institutions,can deliver convenient and possiblylow-cost, remittance services in developingcountries, especially in rural areas. In contrastto cash transactions, remittances channeledthrough bank accounts may encourage savingsand enable a better match for savings and investmentin the economy. Remittances, in turn,can be used to support business and consumerloans, insurance, and other financial productsfor remittance recipients. 28 Some institutionsare exploring ways to target remittances tospecific uses such as school fees or medical149

GLOBAL ECONOMIC PROSPECTS 2006Box 6.6 Smart’s phone-based remittance systemin the PhilippinesThe largest mobile phone company in the Philippines,Smart Communications, has developed aninnovative remittance system based on cell-phonetext messaging. Cell phones are widespread in thePhilippines, in use by at least 30 percent of the 84million Filipinos. A standard Smart remittance workslike this: A Filipino in Hong Kong, China, depositsmoney to be remitted with one of Smart’s remittancepartners, which then sends a text message to the beneficiaryin the Philippines, informing him or her ofthe transfer. The remittance is credited into a SmartMoney “electronic wallet” account by any Smartmobile customer. The money can be withdrawn froman ATM using the Smart Money cash card, whichcan also be used as a debit card for purchases.Smart’s partners in the Philippines—among themMcDonald’s, SM malls, SeaOil gas stations, 7-Elevenstores, and Tambunting pawn shops—will also payout cash to Smart customers.Smart has already formed remittance partnershipswith Travelex Money Transfer; Forex InternationalHong Kong; Dollar America Exchange in California;CBN Grupo in Greece, Ireland, Japan, Spain, and theUnited Kingdom; New York Bay Remittance; andBanco de Oro Bank in Hong Kong, China.The system’s simplicity keeps fees down. Fees atorigination vary from country to country. In HongKong, China, it is about $2. In the Philippines, it is1 percent plus the cost of the text message.The Smart system also appears to be secure. Theuse of different PINs for the cell phone and theSmart account make it difficult for a thief to accessthe funds. An ID is required when collecting cash.bills. Others are exploring insurance products,for example, to ensure a stable flow of incometo the remittance beneficiary in the event thatthe sender suffers an income shock.Credit unions in El Salvador, Guatemala,Honduras, Nicaragua, Mexico, and Jamaicathat are members of the World Council ofCredit Unions (WOCCU) encouragedWOCCU to establish the International RemittanceNetwork (IRnet) in July 1999 to facilitateremittance flows from the United States toLatin America. That initiative has lowered remittancecosts by raising customer awarenessof remittance fees and by generating somecompetition in the remittance market. Tosend up to $1,000, IRnet charges a flat $10—much less than the fees charged by majorMTOs. Besides fee income, IRnet institutionshope to use remittances to build relationshipswith customers. It is reported that 14–28 percentof nonmembers who visited WOCCUaffiliatedcredit unions to transfer funds eventuallyopened an account; 37 percent of creditunion members saved a part of theirremittance receipts (Grace 2005).Smaller nonbank financial institutions facechallenges in entering the remittance marketbecause of regulatory constraints—such aslicenses for transactions involving foreignexchange and access to national paymentsystems. For prudential reasons, access to paymentand settlement systems is typically restrictedto well-capitalized and well-establishedbanking institutions. Microfinance institutionsand smaller institutions generally mustenter into corresponding banking relationshipswith commercial banks 29 and with internationalremittance providers (such as theIRnet or the major MTOs).A survey of central banks found that 35 of 40developing-country authorities were not enthusiasticabout allowing small financial institutionsto have access to clearing and settlementsystems (de Luna Martinez 2005). Centralbanks appear to believe that most nonbank financialinstitutions in developing countries lackthe technological infrastructure required to participatedirectly in clearing and settlement systems.Also, central banks believe that givingnonbank financial institutions direct access to150

REDUCING REMITTANCE FEEScentral banks’ clearing and settlement systemsmay not help reduce the remittance fees chargedby those institutions. According to the survey,only five countries—Azerbaijan, Belarus, Bolivia,Philippines, and Thailand—are contemplatinggranting access for clearing and settlementsystems to a few large nonbank financialinstitutions (mostly post offices).Even if microfinance institutions were tooffer remittance services, they might face restrictionson taking deposits and offering loanand insurance services, again for prudentialreasons. Given these constraints, sources offunds for such institutions tend to be expensiveand their capacity to offer financial productslimited.Annex 6A.1 A stylized remittancetransaction—structure, players,instrumentsAtypical remittance transaction takes placein three steps: (1) initiation of remittancesby a migrant sender using a sending agent,(2) exchange of information and settlement offunds, and (3) delivery of remittances to thebeneficiary. In step 1, the migrant sender paysthe principal amount of the remittance to thesending agent using cash, check, money order,credit card, debit card, or a debit instructionsent by e-mail, phone, or Internet. In step 2,the sending agency—which may be an MTO,bank, or other financial institution, moneychanger, or merchant (gas station, grocerystore)—then instructs its agent in the recipient’scountry to deliver the remittance. Instep 3, the paying agent makes the payment tothe beneficiary. In most cases, there is no realtimefund transfer; instead, the balance owedby the sending agent to the paying agent issettled periodically according to a mutuallyagreed schedule. Settlement usually occursthrough commercial banks acting through thenational clearing and settlement system. A portionof informal remittances is settled throughgoods trade.Step 1 Step 2 Step 3SenderInformationby phone or mailBeneficiaryCashCheck, money orderBank transferCredit, debit, prepaid cardTransfer order in person,by phone, or by internetCashCheck, money orderBank transferCredit / debit cardInternetGoodsSending agentMTOs, banks, creditunions, microfinanceinstitutions, merchants,money exchanges,friends, relatives,missionaries, cell-phonecompaniesInformation forwarded bye-mail, SWIFT, proprietarynetworks, fax, phone, ACHSettlement in cash, via banktransfer, netting against othertradePaying agentMTOs, banks, creditunions, microfinanceinstitutions, merchants,money exchanges,friends, relatives,missionaries, cell-phonecompanies151

GLOBAL ECONOMIC PROSPECTS 2006Annex 6A.2 Licensing and registration requirements for remittanceservice providersAuditedfinancialsCountry Net worth($) required? Bond CommentUnited States 27CaliforniaFloridaIllinoisMassachusettsNew JerseyNew YorkPennsylvaniaTexasVirginiaWisconsinCanadaFranceGermany(federallegislation)Min. $500,000 in equity100,000 plus $50,000 perlocation up to $500,000Depending on locations:1 $35,00025 $500,000None(1) Min. $100,000 plus$25,000 per location (oragent) in NJ up to$1,000,000(2) $50,000 for foreignmoney transmitter plus$10,000 per location (oragent) up to $400,000Liquidity equivalent tooutstanding payments$500,000$25,000 per location up to$1,000,000$100,000–$1,000,000 asdetermined by thecommission“Suitable to conductbusiness”Should not be lower than$10,000NoneMin. €2,400,000plus capital to cover firstyear’s expenses€125,000 capitalNet worth must besufficient to coverexposuresIf availableYesYesNoYes,2 yearsYesNoNoYes,3 yearsYesDiscretionary depending on sizeof business. Min. $200,000.1% of annual turnover, max$250,000; can be set at$500,000 in exceptionalcircumstances; may be waivedupon requestGreater of $100,000 or theaverage daily outstanding for12 months, maximum$2,000,000$50,000 (or 2x amount ofoutstanding transactions)(1) Not less than $100,000 andnot more than $1,000,000(2) Foreign remitters: dependingon business volume, $25,000to $100,000; commissionermay require up to $900,000In general: investments not lessthan outstanding paymentinstruments; this can bewaived by the commissioner$500,000 unless thesuperintendent lowers theamount$1,000,000$100,000 for first location,$50,000 for each additional,max. $400,000$25,000–$1,000,000 asdetermined by the commission$10,000 for 1st location $5,000 for each additionalMax. $300,000NoneNoneNoneFee $5,000 plus $50 peragentApplication fee $500 plus$50 per agent; renewal$1,000 plus $50 peragent up to $20,000.Fee $100Licensing $100$10 per location; $100renewalFee $250Application fee $1,000Licensing fee up to$4,000Biennial fee $25 perlocation up to max. of$5,000Fee $500Licensing $1,000investigation.Application fee $1,000Renewal fee $300Fee $500 licensing $2,500 investigationfeeLicensing fee $500Renewal fee $750Fee $500 license (annual) $300 investigation $5 per location(annual)Reporting threshold:Can$3,000STR and CTR aboveCan$10,000Full bank license;the ownershipstructure must beadequateAML proceduresscrutinizedReporting threshold:€2,500STR; AML laws must befollowed; 2 managingdirectors must havesuitable backgrounds152

REDUCING REMITTANCE FEESAnnex 6A.2 (continued)AuditedfinancialsCountry Net worth($) required? Bond CommentItalyUnited Kingdom€750,000NoneYesNoNoneNoneReporting threshold:€12,500STR; the license is onlyrequired by the serviceprovider, not by hisagentsRegister normal business;Moneys may not be heldfor more than 3 days, asa bank license (deposits)would be required inthis eventSource: For United States, www.rubinsanchez.com; Canadian Bankers Association; French Central Bank, Banque de France, Comitédes Etablissements de Crédit et des Entreprises, d’Investissement (CECEI), Committee for Credit Institutions and Investment Companies;German Financial Supervisory Board, Bundesanstalt für Finanzdienstleistungsaufsicht; Italian Law 106; Bank of England.Note: Licensing and registering approaches may differ. See FATF Typologies Report (FATF 2005). STR suspicious transactionreport; CTR currency transaction report; SAR suspicious activity report; AML anti-money laundering.Annex 6A.3 A brief history ofsome remittance service providersThis annex describes the historical roleof three key providers of remittanceservices—Western Union, MoneyGram, andBank of America. The intention is to shed lighton their business strategies.Western UnionWestern Union descended from the New Yorkand Mississippi Valley Printing TelegraphCompany, originally formed by a group ofbusinessmen in Rochester, New York, in1851. The Western Union Telegraph Companywas subsequently formed in 1956 followingthe acquisition of several competingtelegraph systems. Having completed the firsttranscontinental telegraph line by 1861, thetelegram network started providing the WesternUnion Money Transfer service nationallyby 1871.Historically, Western Union has beeninvolved in a wide range of telecom and otherproducts. These included introducing theNew York Stock Exchange stock ticker in1866, offering a nationwide standard time, introducingteletypewriters in 1923, offering thefirst intercity fax service, and launching thetelex. It also launched the first domestic communicationssatellite, Westar I in 1974. 30Today, Western Union is primarily a remittancecompany.Western Union has always followed thestrategy of developing its own proprietaryproducts. For money transfers, WesternUnion has developed its own software andnetwork of exclusive agents. It has been ableto develop this network by being the firsttruly global remittance company. Currentlyits network comprises slightly more than220,000 locations in about 195 31 countries(including the network of its subsidiary OrlandiValuti). The growth rates in WesternUnion’s international remittance transactions(excluding Mexico) were 32 percent, 25 percentand 24 percent for the years 2002, 2003,and 2004, respectively.MoneyGramThe MoneyGram money transfer service wasstarted in 1988 by Integrated PaymentService, a U.S.–based division of First DataCorporation (FDC), a data processing companyowned by American Express at the time153

GLOBAL ECONOMIC PROSPECTS 2006of inception. 32 FDC divested its MoneyGramoperation in December 1996 through an initialpublic offering of its common stock tocomply with the company’s agreement withthe Federal Trade Commission as part of amerger with FFMC.In 1998, Travelers Express, a division ofViad Corp. acquired MoneyGram. On June30, 2004, the Travelers Express business wasspun off from Viad Corp., 33 and became anindependently traded company called Money-Gram International, Inc. Travelers Expressand MoneyGram Payment Systems, Inc.continue as operating companies under thisnew corporate umbrella.Whereas Western Union insists on a strategyof exclusive partnership with its agents,MoneyGram allows its agents to representother remittance companies as well, as shownby its partnership with the World Councilof Credit Unions and Bancomer TransferServices.Bank of AmericaBank of America (BoA) was established in1929 as an outgrowth of the merger betweenthe Bank of Italy and the Bank of America,Los Angeles. California became the fastestgrowingstate after World War II, with thehighest use of checking accounts. To cope withthe transaction volume, the bank investedheavily in information technology and isgenerally credited, together with GE and SRI,with inventing modern centralized bank operations;BoA has a number of financial transactionprocessing technologies, such as automaticcheck processing, account numbers, andMagnetic Ink Character Recognition (MICR),and, based on these technologies, credit cardslinked directly to individual bank accounts.Because of the efficiency of these technologies,BoA had significantly lower administrativecosts than other banks and was able to expandfurther, until it was the world’s largest bank bythe early 1970s.In 1959, BoA invented the bank creditcard, the BankAmericard, which changed itsname to VISA in 1975. A consortium ofother California banks founded MasterCharge (now MasterCard) in order to competewith the BankAmericard.BoA offers remittance services along withregular savings and loan products to its customers.It has a banking relationship with44 percent of all Hispanic households; itopened more than 1 million checking accountsfor Hispanic customers in 2004. The bank offersremittances under the SafeSend brand to itscustomers wishing to send money to Mexico. Inearly 2005, it announced that it would eliminatethe $10 transfer fee for all checking accountholders for remittances from the UnitedStates to Mexico to attract new business frommigrant customers.Notes1. Western Union reports an average fee of6–8 percent and an additional foreign exchange spreadof 2 percent on its global remittance services. The averagesize of a Western Union remittance (covering personalremittances as well as small business-to-businessremittances), however, is around $700, much higherthan the average transaction size (under $200 reportedin household surveys of migrants by, for example, aPew Hispanic Center study on Mexican migrants in theUnited States and a Genesis Analytics study on SouthAfrica). A World Bank survey of African diaspora inBelgium found that the average monthly remittanceswere 154 euros in the case of Senegalese migrants, 126euros for Nigerian migrants, and only 78 euros forCongolese migrants.2. For example, if a Brazilian bank received remittanceson October 11, 2004, delayed payment to thebeneficiary for two weeks, and invested the funds inthe overnight money market, it would earn a float of2.85 percent (IOM 2005). Humphrey, Keppler andMontes-Negret (1997) note widespread use of floats bybanks, especially in the Russian Federation.3. Some banks have announced even more aggressiveprice cuts recently. Bank of America, for example,eliminated fees for United States–Mexico remittances,to attract customers from the Mexican migrant community.Banks have also been providing free remittanceservices in some other corridors as well. It is worthnoting, however, that there are hidden fees—accountmaintenance fees, minimum balance requirements,taxes on interest income—involved in such transactionsbesides the cross-selling of loan and deposit products.4. According to the Piper Jaffray Global MoneyTransfer Report (2005), Western Union has said that it154

REDUCING REMITTANCE FEEScan add a new agency for $1,000–$1,500 (the cost ofsetup, terminal, software, and training), while Dolexrequires $12,000–$15,000 to establish a new branch.5. In the case of banking institutions (also gasstations and grocery stores), remittance services may becross-subsidized by other product lines, which rendersremittance costs hard to determine.6. The leading MTOs are Western Union, with areported 13 percent market share, and MoneyGram,with a 3 percent market share. Vigo and Dolex are thethird and fourth largest MTOs, respectively, withabout a 2 percent combined market share. See Aite(2005) for their U.S. market shares.7. Western Union has sustained high margins bytaking the initiative to enter underserved markets,building a strong distribution network, and leveragingits brand name. Overall, the company has provided remittanceservices to millions of individuals in previouslyunderserved markets.8. This methodology is similar to that suggestedby Humphrey, Keppler, and Montes-Negret (1997) forpricing payment services. These calculations do not includethe costs of advertising and security.9. This analysis is less conclusive with respect tothe sensitivity of remitters to exchange-rate commissions.Migrants do have fairly accurate knowledge ofthe exchange rate used by their remittance operator,but they may be less informed about the premiuminvolved in this rate. The estimate of cost-elasticityfrom the New Zealand study is based on responsesfrom a small random sample of new Tongan migrantsto questions about how they would react to a potentialchange in costs. It was not based on actual reactions.10. This example assumes no change in theexchange rate.11. There is no standard definition of a retailpayment. Here it is defined as a transaction originatedby, or payable to, an individual, the counterparty beingan individual, a firm, or a government agency. It alsoincludes frequent, small-value business-to-businesspayments. See also BIS (1999).12. As measured by the International CountryRisk Guide, credit risk is based on foreign debt as apercentage of GDP, foreign debt service as a percentageof exports of goods and services, current accountas a percentage of goods and services, the import coverof international reserves, and exchange-rate stability.13. The regression that generated these results isbased on remittance fees from a single large MTO, so theresults are not representative of costs in the remittanceindustry as a whole. Also the low R 2 suggests that theregression does not fully explain the cost structure.14. The remittance price estimates provided byWestern Union and MoneyGram on their Web sitesoften differ from the actual transfer fees. For this study,we gathered remittance price data by visiting WesternUnion and MoneyGram agents in Washington, NewYork, Brussels, Paris, London, and Singapore and bycalling agents in other cities in various parts of theworld. At various points, seven individuals were collectingremittance fee data in various corridors (for example,North America to Latin America and Asia, theEU to Africa and South Asia, the Gulf to South Asia,Eastern Europe to Central Asia, and East Asia to SoutheastAsia). We collected daily foreign exchange data (forWestern Union transfers to a broad range of countriesfrom the United States and the United Kingdom and forMoneygram transfers from the United States to thesame countries) on four consecutive business days inearly June and calculated the FX spread by comparingthese data to the exchange rates quoted on Bloomberg.15. Lack of competition has been a persistent problemin this industry. In December 1996, MoneyGramwas spun off from First Data Corporation, the holdingcompany of Western Union, as part of an agreementwith the U.S. Fair Trade Commission. See the annex fora history of remittance service providers.16. In July 2005, the European Commissionproposed that banks in the European Union be requiredto register the name, address, and bank accountof anyone making an international money transfer. Therequirements, which the commission hopes will comeinto force in January 2007, are the latest EU responseto terrorism following the bombings in London on July7, 2005.17. OCC Advisory Letter 2004-7, www.occ.treas.gov/ftp/advisory/2004-7.doc.18. Abuses include money laundering, the transferof corrupt payments, payment of human smugglingfees, tax evasion, customs offenses, violations ofcurrency controls, subsidy frauds, smuggling, illegalarms sales, and funding of terrorism. Internationaltrade is subject to many of the same abuses, but it iswidely recognized that efforts to curb them must notinterfere unduly with vital trade.19. This is easier said than done, however, becausemajor remittance service providers that have investedin their own proprietary networks and used them toexpand their market share may not willingly sharethem. Furthermore, even if a shared network weredeveloped with public funding, it may not easily gainparticipation by key banks and financial institutions(box 6.4). Federal Reserve Bank (2004, pp. 33–37) listsmajor proprietary payment networks.20. Visa reported that in 2004 it set up a Visanetsystem in Iraq within eight weeks and for less than$200,000 (Brocklehurst 2004).21. BANSEFI has a commercial alliance (L@Redde la Gente) with 62 regulated saving banks and MFIsoperating mostly in areas where commercial banks155

GLOBAL ECONOMIC PROSPECTS 2006have no presence. This network provides a commonplatform for collecting and distributing financial products;for example, it facilitates migrant remittances aswell as government transfers for pension and socialand education programs, and it offers savings accountsand mortgage and consumer loans, small businesslending, and health insurance products to the poor. ItsIT network is aimed at offering the advantage of scaleeconomies to its members.22. Apex Bank in Ghana was set up (with supportfrom the government and the World Bank) to providebanking services in rural areas. The Apex Bank has developedApex Link, a domestic funds-transfer schemeamong 75 rural and community banks. The Apex Bankis also collaborating with some financial institutionsfor the payment of foreign inwards remittancesthrough the rural and community banks to beneficiariesin the rural areas.23. The International Remittance Protection Actproposed by U.S. senator Paul Sarbanes in September2004 marks an effort to improve disclosure of fees andexchange-rate commissions in remittance transactions.24. In the European Union, the fees that financialinstitutions charge their customers for money transfersbetween EU countries cannot be higher than the feescharged for domestic money transfers.25. In some cases, consumer rights legislation hasenabled customers to challenge price gouging. In 2002,Western Union paid $30 million to settle two classactionsuits stemming from its use of different exchangerates for converting remittances than the rates itreceived in the international money market (Aite 2005).26. The related websites are www.profeco.gob.mxand www.sendmoneyhome.org.27. For example, BANSEFI and Apex Link (asmentioned earlier).28. This correspondent banking relationship betweenMTOs and some commercial banks, which hadbeen working smoothly for a long time, came underpressure recently because of a misunderstanding ofthe know-your-customer rules. More than 300 smallMTOs that collected remittances and then wired themthrough a correspondent bank were told in February2005 that such transfers were no longer permitted. OnMarch 30, 2005, FinCEN, FDIC, the Federal Reserve,and the Office of the Comptroller of Currency issued ajoint statement that such transactions were indeedlegitimate.29. Source: www.rubinsanchez.com, representingthe legislations for the different states of the UnitedStates.30. Historical data on Western Union fromwww.westernunionalumni.com/history.htm.31. PR Newswire, February 1, 2005.32. See www.sec.gov/divisions/investment/noaction/firstdata011304.htm. In 1992, AmEx andFirst Data completed an initial public offering that resultedin approximately 40 percent of the commonshares of First Data being held by the public. Over thenext five years, AmEx sold its remaining shares to thirdparties, and by 1997 AmEx had no reportable ownershipinterest in First Data.33. Source: www.MoneyGram.com.ReferencesAite Group. 2005. “Consumer Money Transfers: PoweringGlobal Remittances.” www.aitegroup.com.Andreassen, Ole. 2005. “Remittance Service Providersin the United States: How Remittance Firms Operateand How They Perceive Their BusinessEnvironment?” Unpublished paper. Washington,DC: World Bank.Bakija, Jon M., William G. Gale, and Joel B. Slemrod.2003. “Charitable Bequests and Taxes onInheritances and Estates: Aggregate Evidence fromacross States and Time.” American EconomicReview Papers and Proceedings 93(2): 366–70.Ballard, Roger. 2005. “System-Security in Hawala Networks:An Analysis of the Operational Dynamicsof Contemporary Developments.” Unpublishedpaper. Centre for Applied South Asian Studies,University of Manchester.BIS (Bank for International Settlements). 1999. “RetailPayments in Selected Countries: A ComparativeStudy.” Committee on Payment and SettlementSystems, Basel.Brocklehurst, Stuart. 2004. “Sending It Home: How aCard Payment System Can Facilitate Remittancesto Help Developing Economies.” Presentation atBarcelona Forum 2004–HMI World Congress,September 1–5, Barcelona. www.mhicongress.org/.Cordes, Joseph. 2001. “The Cost of Giving: How DoChanges in Tax Deductions Affect CharitableGiving?” Emerging Issues in PhilanthropySeminar Series, Urban Institute, Washington, DC.www. urban.org/UploadedPDF/philanthropy_2.pdf.de Luna Martinez, Jose. 2005. “Workers’ Remittancesto Developing Countries: A Survey with CentralBanks on Selected Public Policy Issues.” PolicyResearch Working Paper 3638. World Bank.DFID (Department for International Development).2005. “Sending Money Home? A Survey ofRemittance Products and Services in the UnitedKingdom.” London. www.sendmoneyhome.org.El Qorchi, M., S. Maimbo, and J. Wilson. 2003. “InformalFunds Transfer Systems: An Analysis of156

REDUCING REMITTANCE FEESthe Informal Hawala System.” IMF OccasionalPaper 222. International Monetary Fund, Washington,DC.FATF (Financial Action Task Force Against MoneyLaundering). 2005. “Money Laundering andTerrorist Financing Typologies, 2004–2005.”Paris. June. www.fatf-gafi.org.Federal Reserve Bank of Chicago. 2004. “Global ElectronicPayments.” April.Feldstein, Martin, and Amy Taylor. 1976. “The IncomeTax and Charitable Contributions.” Econometrica44(6): 1201–22.Freund, Caroline, and Nikola Spatafora. 2005. “Remittances:Costs, Determinants, and Informality.”Background paper prepared for this report.World Bank.Gibson, John, David J. McKenzie, and HalahinganoRohorua. 2005. “How Cost-Elastic Are Remittances?Estimates from Tongan Migrants in NewZealand.” Background paper prepared for this report.World Bank, Washington, DC.Glenday, Graham., Anil K. Gupta, and Henry Pawlak.1986. “Tax Incentives for Personal CharitableContributions.” Review of Economics and Statistics68(4): 688–93.Grace, David. 2005. “Exploring the Credit Union Experiencewith Remittances in the Latin AmericanMarket.” In Remittances and Development:Development Impact and Future Prospects, ed.Samuel Maimbo and Dilip Ratha. Washington,DC: World Bank.Hernández-Coss, Raúl. 2004. The U.S.-Mexico RemittanceCorridor: Lessons on Shifting from Informalto Formal Transfer Systems. Washington,DC: World Bank.Humphrey, David B., Robert Keppler, and FernandoMontes-Negret. 1997. “Cost Recovery and Pricingof Payment Services: Theory, Methods, andExperience.” World Bank Working Paper 1833.Washington, DC.IMF. 2005a. Approaches to a Regulatory Frameworkfor Formal and Informal Remittance Systems:Experiences and Lessons. Washington, DC.———. 2005b. World Economic Outlook: Globalizationand External Imbalances, April 2005,Washington, DC.IOM (International Organization for Migration).2005. World Migration 2005. Geneva.Kalan, George, and Dilek Aykut. 2005. “Assessment ofRemittance Fee Pricing.” Background paper preparedfor this report. World Bank, Washington,DC.Orozco, Manuel. 2004. “The Remittance Marketplace.”Unpublished paper. Pew Hispanic Center. June.Maimbo, Samuel M., and Nikos Passas. 2005. “TheRegulation and Supervision of Informal FundsTransfer Systems.” In Remittances and Development:Development Impact and Future Prospects,ed. Samuel Maimbo and Dilip Ratha. Washington,DC: World Bank.Passas, Nikos. 1999. “Informal Value Transfer Systemsand Criminal Organizations: A Study intoSo-Called Underground Banking Networks.” TheHague, Ministry of Finance.Piper Jaffray. 2005. “Global Money Transfer Report.”Prahalad, C. K. 2005. The Fortune at the Bottom ofthe Pyramid: Eradicating Poverty throughProfits. Philadelphia: Wharton SchoolPublishing.Ratha, Dilip, and Jan Riedberg. 2005. “On ReducingRemittance Costs.” Unpublished paper. DevelopmentResearch Group, World Bank, Washington,DC.Taylor, John B. 2004. “Remittance Corridors and EconomicDevelopment: A Progress Report on a BushAdministration Initiative.” Remarks presented atthe Payments in the Americas Conference, FederalReserve Bank of Atlanta, October 8.Western Union. 2000. Annual Report. GreenwoodVillage, CO.Yang, Dean. 2004. “International Migration, HumanCapital, and Entrepreneurship: Evidence fromPhilippine Migrants’ Exchange Rate Shocks.” ResearchProgram on International Migration andDevelopment. DECRG. Policy Research WorkingPaper 3578. World Bank.157

International migration createsopportunities for developingcountries but also poseschallenges. Migration canmake migrants and their familiesbetter off, while the remittancesmigrants send home boostdeveloping countries’ foreignexchange earnings and helpalleviate poverty. But in manycases, the size of flows and theirrole in poverty reduction islimited—especially for poormigrants—by high transactioncosts, restricted access to financialservices, and inappropriateregulations or policies. GlobalEconomic Prospects 2006 analyzesthe impact of remittancesand migration on migranthouseholds and migrant-sendingcountries, and considers policiesto increase the developmentimpact of remittances.—François BourguignonSenior Vice President andChief EconomistGlobal Economic Prospects 2006: Economic Implications of Remittancesand Migration explores the gains and losses from internationalmigration and policies to improve the developmental impactof migration, with particular attention to remittances.Recorded remittances sent home to developing countries by internationalmigrants are expected to reach $167 billion in 2005. Unrecordedremittances through formal and informal channels are estimated to be atleast half as large as recorded flows, making remittances the largest sourceof external financing in developing countries. Remittances substantiallyreduce the incidence of poverty and help support household consumptionin response to adverse events.International migration generates significant economic gains formigrants, their countries of origin, and their countries of destination.The benefits to the countries of origin are especially large in the case ofmigration of low-skilled workers.The most feasible means of increasingsuch migration would be to promote joint origin-destination countryprograms that combine temporary migration of low-skilled workers withincentives for return.The price of remittance transactions is often unnecessarily high for thesmall transfers typically made by poor migrants. Governments could lowerfees and expand remittances by improving the access of poor migrants andtheir families to formal financial services, and by promoting competitionin the remittance transfer market (for example, by lowering capital requirementson remittance services, improving transparency, and opening up retailfinancial networks to nonexclusive partnerships with remittance agencies).Increasing migration of low-skilled workers and reducing remittance feescould significantly help alleviate poverty.Chapter 1 of this study reviews recent developments in and prospectsfor the global economy, and their implications for developing countries.Chapters 2 and 3 evaluate the costs and benefits of migration, throughmodel-based simulations and a review of the economic literature. Subsequentchapters address remittances, including their size and macroeconomicimpact (chapter 4), their impact on households (chapter 5), and policymeasures to lower the cost of remittance transactions for poor migrants(chapter 6).For additional information, please visit www.worldbank.org/prospects.To access Prospects for the Global Economy, an online companionpublication, please visit www.worldbank.org/globaloutlook.ISBN 0-8213-6344-1