Africa's Missing Billions - Oxfam International

More documents

Recommendations

Info

Appendix: Methodology for cost toAfrica calculationThe methodology used is very similar to that used by Stewart andFitzgerald of the University of Oxford in War and Underdevelopment(2001). 129 We have also taken advice from academics and experts,including Anke Hoeffler of the Centre for the Study of AfricanEconomies at Oxford University and Graham Harrison of theUniversity of Sheffield.The calculation was done as follows:1. Determining where and when conflicts happened in Africa from1990. The basic source of data used was the Heidelberg Institute forInternational Conflict Research (HIIK)’s register of conflicts (newmethodology, not yet published) which does not rely on a narrowdefinition of combat deaths. 130 Of HIIK’s five levels of violence, weused a ‘severe crisis’ (level 4) or a ‘war’ (level 5). These results werecross-checked against the Uppsala Conflict Data Program 131 andother sources such as the BBC, The Economist, and the UN’s IRIN.2. Predicting what GDP should have been (the counterfactual). Eachconflict country was assigned to a broad economic grouping usingWorld Bank categories: low income, low-middle income or uppermiddleincome. 132 The GDPs of all countries (constant year 2000 $)were taken from the World Development Indicators. 133 It should benoted that data may not be totally reliable and in some cases isdependent on estimates. For each grouping, the annual growth ratewas calculated for each year, 1990–2005.3. Calculating a measure of the cost of conflict = the cumulative lossin GDP. Starting with the GDP at the beginning of the conflict, thegrowth rate of the economic grouping was applied for every year thatthe country was in conflict. The difference between the predictedGDP and the actual GDP was calculated for each conflict year andsummed.Limitations to this methodologyThere is no perfectly accurate way of developing the counterfactual;however, our method is reasonable and reflects the typicalperformance of a country of that economic level. The large number ofcountries in each economic group (around 45) smoothes peculiaritiesdue to particular countries. As a form of comparison and verification,we have used other counterfactuals (for example, predicting a28Africa’s missing billions, IANSA, Oxfam, and Saferworld, October2007
country’s GDP according to its previous non-conflict growth), andthese suggest that our estimate is satisfactory. The comparison withgrowth projected from past performance was not used to generate thetotal cost figure because it was impossible to generate a non-conflictgrowth rate for some countries, due to the length of the conflict or itscyclic nature.For most of the conflicts considered, conflict had an overwhelmingeffect on the economy and therefore the fall in GDP can be used as areasonable proxy for the cost of the conflict. However, if the conflict isgeographically limited and situated away from main economic zones,the impact on the national economy may not be so marked. Forexample, the conflict in South Africa in the 1990s was geographicallyconfined and so only had a small impact on the economy. In a fewsituations, other factors also have an impact on GDP: e.g. majorchanges in world commodity prices or drought in an agriculturaleconomy (the impact of which would obviously be made much worseby conflict).Our figure is likely to be an under-estimate because it does notinclude the effects on neighbouring countries (of reduced trade andremittances, refugee movements, etc.) or the social costs of conflict(propensity to increase crime, loss of social capital), and it only looksat the period of conflict, not the war overhang.All countries showed a loss due to conflict, apart from Uganda. Themajor reason for this peculiar and unconvincing result is the boomingUgandan economy. 134 Our methodology compares actual growthwith predicted growth, but the Ugandan economy has grown muchfaster than average.In addition, the conflict is confined to the north of the country andalthough this has had some economic impacts at national level(cotton production has suffered), other economic drivers (includingcoffee, the country’s main export) have not been affected.A 2002 study, updated in 2006, used an accounting methodology (i.e.counting individual costs, rather than modelling) to estimate theeconomic cost of the war in the north of Uganda as $1.7bn over 20years, or $85m per year. 135 This is clearly significant and has majoropportunity costs, particularly in the north. $85m represents 1.1 percent of GDP in 2005.For more details on the methodology, please email:africaattreport@controlarms.org29Africa’s missing billions, IANSA, Oxfam, and Saferworld, October2007
Page 4 and 5: abuse human rights and ignore the r
Page 6 and 7: conflict is one important factor in
Page 8 and 9: Using the definitions from the Heid
Page 10 and 11: Table 1: Selected country results f
Page 12 and 13: Table 2: Key costs to the national
Page 14 and 15: Direct costs, with a focus on medic
Page 16 and 17: example, tend to be major losers in
Page 18 and 19: Faso increased military investments
Page 20 and 21: combatants and have brought their w
Page 22 and 23: Exactly quantifying African product
Page 24 and 25: principles of customary internation
Page 26 and 27: Diversion from stockpiles: As state
Page 30 and 31: Notes1 International Action Network
Page 32 and 33: Education in Africa: $5bn/year, fro
Page 34 and 35: 65http://www.oxfam.org.uk/applicati
Page 36 and 37: 104 Amnesty International, IANSA, a
Page 38: © The International Action Network

Africa's Missing Billions - Oxfam International

Create successful ePaper yourself

Delete template?

Save as template?