Arunatilake, Nisha, Sisura Jayasuriya and Saman Kelegama, 2001, “The Economic Cost of the War in Sri Lanka,” World Development, Vol. 29, No. 9, pp. 1483.
Catao, Luis, and Marco Terrones, 2001, “Fiscal Deficits and Inflation: A new Look at Emerging Markets,” IMF Working Paper 01/174 (Washington: International Monetary Fund).
Collier, Paul, V.L. Elliot, Håvard Hegre, Anke Hoeffler, Marta Reynal-Querol, Nicholas Sambanis, May 2003, “Breaking the Coflict Trap: Civil War and Development Policy,” (Washington: World Bank).
Davoodi, Hamid, Benedict Clemens, Jerald Schiff and Peter Debaere, 1999, “Military Spending, the Peace Dividend, and Fiscal Adjustment,” IMF Working Paper 99/87 (Washington: International Monetary Fund).
de Melo M., C. Denizer and A. Gelb: “From Plan to Market,” 1996, World Bank Policy Research Working Paper 1564 (Washington: World Bank).
Demekas, Dimitri, James McHugh and Theodora Kosmax, 2002 “The Economics of Post-Conflict Aid,” IMF Working Paper 02/198 (Washington: International Monetary Fund).
Elbadawi, Ibrahim, and Nicholas Sambanis, December 2000, “Why Are There So Many Civil Wars in Africa? Understanding and Preventing Violent Conflict,”Journal of African Economies, Vol. 9, No. 3 pp. 244-269.
Elbadawi, Ibrahim, and Nicholas Sambanis , January 2002, “How Much War Will We See? Explaining the Prevalence of Civil War,” (Washington: World Bank).
Fallon, Peter, Nicholas Staines and others, 2004, “Review of Recent IMF Experience in Post-Conflict Countries,” IMF Occasional Paper No.____, forthcoming (Washington: International Monetary Fund).
Fearon, James and David Laitin, 2003, “Ethnicity, Insurgency and Civil Wars,”American Political Science Review; Vol. 97, No 1, pp.75-90.
Fischer, Stanley; Ratna Sahay, Carlos Vegh, 2002, “Modern Hyper and High Inflation,” IMF Working Paper 02/197 (Washington: International Monetary Fund).
Gerson, Philip, 1998, “The Impact of Fiscal Policy Variables on Output Growth,” IMF Working Paper 98/1 (Washington: International Monetary Fund).
Ghosh, Atish and Steven Philips, 1998, “Inflation, Disinflation and Growth,” IMF Working Paper 98/68 (Washington: International Monetary Fund).
Gupta, Sanjeev, Benedict Clemens, Emanuele Baldacci, and Carlos Mulas-Granados, 2002a, “Expenditure Composition, Fiscal Adjustment and Growth in Low-Income Countries,” IMF Working Paper 02/77 (Washington: International Monetary Fund).
Gupta, Sanjeev; Benedict Clemens, Rina Bhattacharya, and Shamit Chakravarti, 2002b, “Fiscal Consequences of Armed Conflict and Terrorism in Low- and Middle-Income Countries,” IMF Working Paper 02/142, (Washington: International Monetary Fund).
Hemming, Richard, Michael Kell and Selma Mahfouz, 2002, “The Effectiveness of Fiscal Policy in Stimulating Economic Activity—A Review of the Literature,” IMF Working Paper 02/208 (Washington: International Monetary Fund).
Humphreys, Macartan, August 2002, “Economics and Violent Conflict,” Harvard University, Internet Essay for the Program on Humanitarian Policy and Conflict Research, Conflict Prevention Initiative (HPCR-CPI), www.hsph.harvard.edu/hpcr/index.htm.
Imai, Kosuke and J.M. Weinstein, June 2000, “Measuring the Economic Impact of Civil War,” Working Paper No 51, Centre for International Development, Harvard University.
Khan, Mohsin, and Abdelhak Senhadji, 2000, “Threshold Effects in the Relationship Between Inflation and Growth,” IMF Working Paper 00/110 (Washington: International Monetary Fund).
Knight M., N. Loayza, and D. Villanueva, 1996, “The Peace Dividend: Military Spending Cuts and Economic Growth,” Staff Papers, International Monetary Fund, Vol. 43, pp.1-37.
Michailof, Serge, Markus Kostner and Xavier Devictor, April 2002, “Post-Conflict Recovery in Africa: An Agenda for the Africa Region,” World Bank Africa Region Working Paper Series, No. 30 (Washington: World Bank).
Nafziger, E. Wayne, and Juha Auvinen, 2002, “Economic Development, Inequality, War, and State Violence,” World Development Vol. 30, No. 2, pp. 153–163.
Ranveig, Gissinger, Petter Gleditsch Nils, and Håvard Hegre, 2002, “Globalization and Internal Conflict” in Globalization and Conflict, Boulder, Gerald Schneider, Katherine Barbieri and Nils Petter Gleditsch, eds, Rowman and Littlefield.
Ross, Michael L, June 2002, “Oil, Drugs, and Diamonds: How Do Natural Resources Vary in their Impact on Civil War?” International Peace Academy.
Sambanis, Nicholas, November 2001, “A Review of Recent Advances and Future Directions in the Quantitative Literature on Civil War,” Yale University.
Smaldone, Joe, 2003, “War and Peace in Sub-Saharan Africa, 1989-1999: Does Military Spending Matter?,” paper presented to 46th Annual Meeting of the African Studies Association.
This paper has benefited from extensive discussions with Peter Fallon and Nancy Happe and from comments by Donal Donovan, Mark Plant, Louis Valdivieso, Ragnar Gudmundsson, Anton Op de Beke, and Sonali Jain-Chandra. Any errors are the sole responsibility of the author.
Throughout the paper, the policy stance is evaluated solely in terms of its estimated impact on growth and not relative to some benchmark.
The IMF’s experience in post-conflict countries since the introduction of its emergency post-conflict assistance policy in 1995 and especially in the six countries that received this financial assistance is discussed in Fallon, Staines and others (2004)
Direct budgetary assistance could be provided while meeting concerns over governance, through donor trust funds to help pay, for example, external debt service or the civil service wage bill.
While dividing the sample in this way creates a risk of sample selection bias, as lengthy conflicts are more likely to be excluded from the 1990s subgroups, this problem does not arise here as all conflicts that began in the 1990s had ended by 2000. Conflicts excluded from the sample because they had outlasted the 1990s, i.e., Angola and Sudan, were lengthy confrontations that commenced in earlier decades.
SIPRI defines a country to be in conflict in any single year if there are more than 1,000 casualties with at least 5 percent of casualties on either side.
World Economic Outlook, Winter 2003.
Caplan (2001) looked at conflicts over 1950–1992 and concluded that while civil conflicts have on average reduced annual real GDP growth by 2 percent, international conflicts have increased growth by 2 percent.
The conflicts are here arranged by starting date. By way of comparison, Fearon (2002) looked at the 122 civil wars from 1945 to the mid-1990s and estimated that the average length of civil conflict was 9 years but with a high variance so that about half in equal numbers were either less than 2 years or more than 12 years. He also noted the shortening of conflicts in the 1990s but concluded that the average length of outstanding civil conflict has been steadily increasing since 1945 and reached 15 years by the mid-1990s.
For the 14 post-1990 conflicts, the contractions in real GDP and real GDP per capita ended before the end of conflict in 5 and 3 countries, respectively, and both ended after conflict in 3 countries.
Output in the DFEs fell to 49 percent of its pre-conflict level during conflict and further after conflict but much of this reflected external factors that were unrelated to conflict and also possibly related to changes in national accounting methodology. Once these factors are accounted for, the depth of the contractions in the DFEs and the post-1990 conflicts were broadly similar. By way of comparison, output for the former Soviet block countries (CIS) as a whole contracted to 62 percent of its pre-conflict level, suggesting that the additional impact of conflict was moderate. Also, de Melo and others (1996) investigated the effect of civil wars in the transition economies of Eastern Europe on the average growth rate over the period 1989-94 and found that civil conflict reduced the annual average growth rate during the five years by 9 percent.
Contractions continued after conflict in Azerbaijan, Georgia and the DRC and was particularly strong in Azerbaijan.
The return of real growth to ‘normal’ is an alternative candidate employed, for example, by Collier (1999).
By way of comparison, Collier and others (2003) estimated that the cost of conflict to be about seven months’ output.
Revenues include grants since data on tax revenues is only available for about half the countries.
Important exceptions were the DFEs whose fiscal balances were stronger at the end of conflict than at the start. This was because these countries entered conflict with very little administrative capacity to raise revenues and this capacity had to be mobilized quickly once conflict started.
Caplan (2001) looked at 66 conflicts from 1953 to 1992 and found that during civil conflicts in low-income countries higher military spending tended to crowd out other spending with little overall effect on total government spending. He also found that tax revenues tended to remain unchanged or declined as a share of GDP. Gupta and others (2002b) looked at 22 conflicts from 1985 to 1999 and also found evidence that military spending crowded out other spending. Smaldone (2003) looked at military spending in 42 sub-Saharan African countries in the 1990s. He found no clear difference in real military spending levels between countries that remained at peace and those affected by conflict but found that the military’s share of total government spending tended to be higher in the latter. These findings are echoed here: in the 14 countries (10 of which are post-1990) which reported data to the IMF, average military spending increased during conflict as a share of both GDP and total fiscal expenditures but declined in real terms.
Domestic financing is estimated by the financing balance after taking into account all external assistance. IMF net financing is included although it is typically routed through the Central Bank as balance of payments support. This is because IMF financing also provides the resources to permit the sterilization of the monetary expansion that accompanies domestic fiscal financing and is therefore implicitly budgetary financing.
The linkage between domestic financing and CPI inflation is not direct but is likely to be more closely related in countries affected by conflict where access to indirect monetary instruments is often heavily curtailed.
For 66 conflicts over 1953-92, Caplan (2001) found that civil conflicts had no discernible impact on inflation. Gupta (2002b) found that 22 episodes of armed conflict in the 1990s had a discernible impact on inflation.
The World Bank commodity price index increased in the 1970s but dropped in the 1980s and 1990s. There were important conflicting tendencies on import unit prices. In the earlier conflicts, unit prices increased (volumes dropped while expenditures rose) probably reflected increased defense related imports. In the later conflicts, unit prices dropped (volumes increased while expenditures declined), probably reflecting a shift in the composition of imports towards basic staples.
Even allowing for an average grant element of 65 percent (comparable to IDA terms), this implies an average NPV-of-debt to current exports ratio at the end of conflict of about 260 percent, well above the 150 percent threshold normally considered sustainable.
The contractionary and in-conflict recovery periods are here defined in terms of real GDP per capita. Also, the contraction period only includes contractions during conflict and not after conflict.
These indicators provide imperfect proxies for the policy stance. For example, assessing the fiscal policy stance would involve measurement of the structural fiscal balance for which, unfortunately, adequate data is lacking. CPI inflation is included instead of, for example, money growth since most of the literature focuses on the impact of inflation. Including money growth instead does not change the results significantly.
There are non-linearities at work, especially for inflation, that are not here explored.
The shift in the earlier conflicts is difficult to explain but, for example, the reduced impact of government spending arguably reflects the shift in the composition of government spending towards unproductive and more import intensive military spending observed by Gupta and others (2002b). The only significant change in the later conflicts was for the impact of a change in the fiscal balance whose sign was reversed. This shift could also reflect the endogeneity of the policy variables. However, it is also consistent with the presence of non-linearities on the impact of the policy variables. Estimates including the square of both the change in the fiscal balance and inflation indicate that these non-linearities were significant, especially for inflation.
The results provide no support for the presence of convergence effects. The effect of changes in the terms of trade was generally positive outside the conflict cycle, but negative during the conflict cycle. This result could reflect a shift in the structure of trade (especially imports) but is also consistent with the observation that, once conflict started, rents from natural resource exports tended to prolong and arguably accentuate the depth of conflict, Collier and others (2000b and 2001)
The contribution of a variable in any period is the actual value of the variable multiplied by the coefficient from the GLS regression for equation (1) in Table 8, taking serial correlation into account where relevant.
This is done using the same coefficients from Table 8 as for the set of post-1990 conflicts as a whole.
With respect to inflation, Fischer (1993) has argued that causality is more likely to run from inflation to growth and studies of the impact of inflation on growth have tend to downplay the endogeneity of inflation: Cateo and Terrones (2001), Fischer (2002), Ghosh and Philips, (1998), Khan and others (2000). However, in countries affected by conflict, the large swings in output could be expected to have a more discernible impact on inflation. The endogeneity of government expenditures is also difficult to ignore because of the impact of conflict on the tax revenue base. However, looking at fiscal adjustment and growth in a selection of low-income countries (not necessarily affected by conflict), Gupta and others (2002a) found little evidence of significant reverse causality between output and government expenditure growth.
The set of i country equations are estimated as a set of system equations with the coefficients constrained to be the same across the country equations. No single equation statistics can therefore be calculated.
Net official external resource inflows are equal to official transfers plus loans less debt service paid (after debt relief and the accumulation of arrears).
The link between net resource inflows and inflation is not direct. Unless sterilized, external aid inflows can lead to higher monetary growth and consequently inflation. However, by providing a source of foreign currency, the provision of external assistance also makes it easier for the central bank to sterilize domestic credit to the central government, especially when the availability of other monetary instruments has been curtailed by conflict.
It is difficult to distinguish whether assistance was increased at the end of the conflicts in the 1990s in anticipation of the end of conflict or whether this increased involvement itself helped bring about an earlier termination of conflict than might have otherwise occurred. However, the experience in some countries suggests that earlier and well-timed involvement of the international community can help provide the impetus for parties in conflict to reach a resolution.
These six countries are: Albania, Republic of Congo, Guinea-Bissau, Rwanda, Sierra Leone, and Tajikistan. Tajikistan is also a DFE. Three other countries have also received EPCA from the IMF and are not included here—Burundi, the Federal Republic of Yugoslavia and Bosnia-Herzegovina. Burundi emerged from conflict in 2000 and only received EPCA in late 2002; Bosnia-Herzegovina is a new country so that no data is available before and during conflict; and the FRY is not a low-income country.
The IMF introduced a new policy to assist post-conflict countries in 1995 and one of the policy changes was to enable the IMF to provide financial assistance to post-conflict countries soon after the end of conflict and before they are ready to move to a comprehensive IMF program that could garner broader support from donors.
Seven of these also received financial assistance from the IMF through vehicles other than EPCA, five of them soon after conflict.
Average growth is skewed down by the particularly poor performance in Azerbaijan, Georgia and the DRC, where output continued to contract after conflict and where inflation remained in triple digits. The median real GDP per capita growth of about 2 percent is perhaps more indicative.
The IMF arguably helped play a role in this virtuous cycle. On the one hand, the main role of the IMF was to help formulate a macroeconomic strategy for recovery and rebuild administrative capacity for its implementation. On the other hand, the EPCA-supported programs played an important catalytic role in mobilizing this donor support at an early stage after conflict by signaling that the strategy was being implemented.