Conflict Prevention and Reconstruction Unit April 2002 Dissemination Notes Aid, Policy and Growth in Post-Conflict Countries Number 2 Aid is considerably more effective in augmenting growth in post-conflict situations than in other situations, but the pattern of aid matters. Key post-conflict priorities should be social policies first, followed by sectoral policies and macro policies last. Introduction Since growth and poverty reduction tend to reduce the risk of conflict (Collier and Hoeffler, 2001 ),2 Since the needs of post-conflict situations compete the aid allocation warranted on the criterion of poverty for the same pool of resources devoted to aid for reduction will also contribute to peacebuilding. development, it is useful to benchmark the efficacy of aid The second purpose of the paper is to investigate in post-conflict situations relative to development whether and how priorities for the reform of policies, assistance more generally. governance and institutions might differ in post-conflict The general effectiveness of aid in reducing poverty societies from those in other developing countries. At the can be benchmarked using the analysis of Collier and most obvious level, priorities might differ because some Dollar (2001).1 However, post-conflict situations are not problems are atypically severe. For example, if inflation explicitly considered in the Collier-Dollar analysis: were to be atypically high in post-conflict situations, then 'poverty-efficient" aid for post-conflict countries makes improved macroeconomic management would be no allowance for their special circumstances other than atypically important. Less obviously, some reforms might what is already reflected in the Country Policy and be atypically important not because the attained level of Institutional Assessment (CPIA) rating. The economic performance is worse than in other societies, but circumstances of post-conflict societies are distinctive in because the economy is atypically sensitive to them. We several respects, and this is likely to change the use a new data set on disaggregated ratings of different relationship between aid, policy and growth. Typically, aspects of policy, governance and institutions to test for post-conflict societies face substantial financial needs for differential priorities. reconstruction, implying that they might be able productively to absorb more aid than in peaceful The Pattern of Post-Conflict Recovery societies. However, they also tend to have high levels of opportunism, making the management of aid more We first investigate whether there are exogenous difficult. Hence, a priori, aid might be more or less forces for economic recovery from conflict, in the sense preflctiveing post-conflicthe rsuiednee for fthat growth is more rapid during the post-conflict phase, Reflecting the presumed need for differential cotoln.o olce n oi inlw.. treatment, the IDA allocation formula has recently been controlling for policies and for aid Inflows. treatment, the ID allocation.formulahasrecent For the growth relationship we rely upon the analysis revised to allow post-conflict countries to receive and database used in the Collier-Dollar growth additional temporary resources. However, at present, regression (Collier and Dollar, 2001). This analyzes the this Is based on Judgment rather than quantitative analysis. Thebprimarypurposetofathiper is than quantite per capita growth rate over each four-year period from analysis. The primary purpose of this paper is to bring 1974 to 1997 for 62 countries. We introduce conflicts post-conflict situations explicitly into the poverty- into this analysis using the database of Collier and efficiency framework of aid allocation. The resulting Hoeffler (2001), which provides a comprehensive listing benchmarks are all the more necessary given the highly and dating of civil wars for 1960-99. They use a politicized nature of aid allocation In post-conflict..... situations. Of course, aid in post-conflict situations has definition of civil war that is conventional in the academic legitimate objectives other than poverty reduction. There literature: namely, an internal conflict between a is a considerable risk that conflict will resume and aid government and an identifiable rebel organization that misht conideretle redu that confict willyond ume an d e aid results in at least 1,000 combat-related deaths, of which might directly reduce this risk beyond any bects ia at least 5% must be Incurred on each side. Most of these growth and poverty reduction. However, a benchmark in 73 wars did not end within the period analyzed by Collier terms of the objective of poverty reduction can at least and Dollar, or, due to data limitations, were in countries provide guidance to donors as to the lower bound for other than the 64 included in their analysis. appropriate response. 1 Collier, Paul and David Dollar. Forthcoming. "Aid Allocation and 2Collier, Paul and Anke Hoeffler. 2001. "Greed and Grievance in Civii Poverty Reduction," European Economic Review. War." We have investigated the time-profile of post- Our results also suggest that the increased scope conflict growth through two approaches. The first defines for effective aid absorption does not occur immediately. three four-year episodes: peace onset and the two There is neither a supra-normal growth effect nor a subsequent periods. The second allows us to date the supra-normal effect of aid in the peace onset period. onset of peace more precisely but encounters some Since the moment of peace onset is randomly distributed difficulties of interpretation. Both approaches reveal an across the four-year peace-onset episode, its average inverted-U pattern. On the first approach supra-normal length is two years. We have also found that there growth peaks in the second period. This period will appears to be no supra-normal growth effect during the range from being virtually the first four years of peace to first three years of peace, or beyond the seventh year of virtually the second four years of peace. The second peace. Hence, there is some presumption that the key approach dates the peak growth phase more precisely period during which aid absorption is exceptionally high as being from the fourth to the seventh year of peace. is approximately between the fourth and the seventh Such a pattern of recovery is not a priori surprising. In year of peace. From the perspective of effective use of the immediate aftermath of conflict there are probably aid for economic recovery, aid volumes should gradually many uncertainties, and basic functions of government build up during the first few years of peace, and have yet to be re-established. If peace is maintained gradually revert to normal levels after around a decade. there is then a phase of catch-up, but this peters out and We should stress that these results are tentative: as data the economy reverts to its long-run growth rate. for the episode 1998-2001 become available, the sample of post-conflict countries will increase and the results Aid During Recovery should be re-assessed. An implication of the above analysis is that donors It does not necessarily follow from this that post- have not responded appropriately to post-conflict conflict countries should get more aid than other situations. The initial response during the peace-onset countries with similar levels of poverty. Allowance must period - typically the first two years - has indeed be made for the unsurprising fact that policies and restored lending, perhaps (on the narrow criterion of institutions tend to be less satisfactory in post-conflict poverty reduction) even excessively, but thereafter aid situations. Hence, the greater absorptive capacity should have continued to taper in whereas it has tended conditional upon policy, is qualitatively offset by worse to taper out. The recent experiences of Afghanistan and policy. To quantify this, for the 1990s we compare the East Timor suggest that donor behavior may be average CPIA score for all countries with that for those changing: the volume of aid allocated to post-conflict countries in their first full period of post-conflict peace. situations may well have increased substantially, which The former is 3.00 and the latter is 2.88. The typical on our analysis would be appropriate. However, the country in its first full four-year period of post-conflict timing of the inflow may not be appropriate, arriving too peace thus has a saturation point around 2.36 times that soon and tapering out too early. For example, the for the typical country in other circumstances. present rules concerning IDA allocation to post-conflict During the first full period of post-conflict, the typical countries allow for supra-normal aid allocations only in country experiences a temporary growth spurt of around the first three years of peace, which on our analysis is two percentage points per year in excess of normal too early for effective absorption. Of course, the growth. This growth spurt is largely, or entirely, benchmarks provided by regression analysis, especially dependent upon aid: for given policies aid is more than on such a small sample of episodes, can provide only twice as productive in post-conflict circumstances, and limited guidance. Each situation will appropriately be so at normal levels of aid, growth is higher. In the assessed using much richer country-specific information. absence of aid there would be no growth spurt. However, given the highly politicized context of aid We next consider whether these effects of aid are allocations in post-conflict situations, it would not be distinct to the first full period of post-conflict peace, or surprising if historically they have not been appropriately whether they apply both earlier and later. The end of a aligned with the opportunities for reinforcing the civil war creates a temporary phase during which aid is opportunities for economic recovery. particularly effective in the growth process. Our results suggest that during the first full peace period the Policy Priorities During Recovery absorptive capacity for aid is around double its normal level and that some such effect extends at least We now turn to the question of whether policy qualitatively into the second full peace period. As with priorities for growth should be distinctive in post-conflict aid in more normal circumstances, absorptive capacity societies. The previous analysis has already implicitly depends upon policy, but, conditional upon policy, aid is answered this at the aggregate level of policy captured considerably more effective. Although policy is worse in by the overall CPIA rating, which is an average over post-conflict societies than in most other societies, this is ratings of twenty different particular policies. Since the insufficient to offset the greater absorptive capacity, so effect of aid is dependent upon policy, as measured by that post-conflict societies constitute an important the CPIA, the better is policy, the larger is the growth exception to the proposition that for given levels of spurt. Thus, when policy reform is coordinated with aid poverty, aid should be lower in societies with worse flows, it is atypically effective in promoting growth in policies. post-conflict situations. The previous analysis has also found that the interaction term between policy and the three post-conflict variables is insignificant. Hence, other disaggregation into the three distinct post-conflict than through its effects on enhancing aid absorption, episodes analyzed above so that the post-conflict policy is neither more nor less important for growth in dummy refers to all three episodes. We then add the post-conflict situations than in other situations. Policy interaction term for governance. The interaction term is matters more in post-conflict situations because it insignificant, so that governance is approximately as differentially augments the effectiveness of aid. important for growth in post-conflict situations as in other contexts. We now investigate whether particular policies are Macro, structural and social policy scores on the differentially important in post-conflict situations. For CPIA are too highly correlated with the overall CPIA to this, we disaggregate the overall policy rating into four be entered together in the same regression. To components: macro, structural, social and governance. overcome this problem, we measure each relative to the This disaggregation is dictated by the availability of data, average CPIA score. Thus, we retain the overall CPIA but it corresponds to important broad categories of policy score in the regression and add variables showing how and so is potentially useful. The data on macro, the components deviate from the overall score. structural and social policies are from the components of Evidently, since the overall score is simply the average the CPIA ratings and are available since 1990. Prior to of its three components, once the deviation of any two of this the CPIA was only an aggregate indicator of policy. the components from the average is specified, the Within the CPIA each of these three components is deviation of the third component is also determined. scored on a scale of 1-5. Thus, if one component has a Hence, only two of the deviations in the components can lower score than another this has no intrinsic meaning. be entered together in the regression. We introduce the However, in practice, the mean values of the three deviations for the macro and social components of the components are all very similar: during the 1990s the CPIA as additional variables, and also the interactions of average rating for macro policies was 4% higher than for these terms with the post-conflict dummy, so that structural policies, and 6% higher than for social policies. structural policies are the excluded term and so the This suggests that each component of the scale was benchmark. approximately ordinal, with a country that was average Since the regression is run for all episodes since for macroeconomic policies, getting approximately the 1974, but the disaggregated CPIA data is only available same rating on these policies as the rating for social for the 1990s, we initially include both a dummy variable policies for a country that was average for those policies. for the 1990s and an interaction of this dummy with the In addition to the CPIA, which is measured on a common overall CPIA score. This allows both that exogenous basis by World Bank staff, the ICRGE is used to growth might have been different during this period and measure governance. Again, this is a subjective that the effect of policy might have been different. assessment on a scale of 1-6. Without these terms the CPIA component terms might Unsurprisingly, post-conflict societies have worse be spuriously picking up such effects. In the event, CPIA scores than other societies. The scores by post- neither term is significant and they are dropped from the conflict episode reveal a steady improvement as long as regression. peace is maintained. In the peace-onset period the CPIA The regression includes the direct effect of the is only 2.50, in the first full period of peace it is 2.88, and macro and social components of the CPIA (i.e. without in the second full peace period it has risen to 3.05 and being interacted with the post-conflict dummy). These so is in effect back to normal. Hence, the phase of terms are necessary in order to determine how the distinctively problematic policy is the first 4-8 years of effects of these components of policy are distinctive in post-conflict. However, although the level of overall post-conflict situations, through their interaction with the policy- is distinctively poor during this period, there post-conflict dummy. However, the temptation to appears to be no systematic difference between policies. interpret these direct effects as showing which Macro, structural and social policy scores are all equally components of policy are most important for growth discounted in post-conflict countries and show fairly outside the context of post-conflict should be resisted. All uniform improvements during the three post-conflict they show is what would happen if the three component periods. Governance, as measured by the ICRGE, is parts of the CPIA score were to be varied in such a way markedly worse during the peace onset period, than as to keep the aggregate score constant. At the most, other developing countries. However, in contrast to the these results will tell us that the reforms represented by CPIA components, it actually appears to deteriorate over a one point increase in one component are more the ensuing decade. valuable for growth than the reforms represented by a one point increase in another component. They are We now return to the regression analysis, therefore a comment not upon the relative importance of introducing terms which interact each of the four macro, structural and social policies, but upon the components of policy with a post-conflict dummy scoring systems for them. There is no reason why a one variable. This tests whether any of these policies is point change in one component should be in any sense differentially important for growth in post-conflict commensurate with a one point increase in another situations and hence provides some guidance as to component. priorities for policy improvement. Since our analysis can While the direct effects of the three policy components only be conducted for the period since 1990, the number must therefore be dismissed, the variables generated by of post-conflict observations is too small to permit interacting the policy components with the 2 post-conflict dummy are readily interpretable. They test Conclusion for whether policy priorities should be distinctive in post- conflict situations when compared to other Countries coming out of conflict are in atypical need circumstances. of both financial resources and policy advice. Their Both the interaction terms are significant, with societies are often extremely fragile and so it is macro negative and social positive. Further, the important that the response of the international coefficients on both terms are large. Given the objective development community should be as appropriate as of promoting growth, consider priorities as between possible. Although such situations are becoming more macro, structural and social reforms in two societies with common, they still constitute a small minority of identical CPIA scores on each component, one society development experience, and so there is a danger that being post-conflict and the other having no history of they will receive both finance and advice that largely conflict. The results tell us that the post-conflict society ignores their special characteristics. Most donors now should pay more attention to improvements in social have units specially dedicated to post-conflict, but to policy than the other society, and less attention to date the learning process has largely been highly improvements in macro policy. This formulation of the context-specific. Indeed, a common assessment from result is not only meaningful, it is pertinent. Post-conflict policy practitioners is that each situation is so distinctive situations constitute only a small minority of the that there are no general lessons. In this paper we have situations on which IFI experience is based. Hence, in investigated post-conflict economic recovery statistically, the absence of such knowledge, IFI staff are likely to using all episodes for which data are available. Since the advise for post-conflict situations those policy priorities number of such observations is quite limited, the degree that are effective in normal circumstances. While the of confidence in the results must be correspondingly broad direction of such advice might be correct, our discounted. The basis for our analysis has been to results suggest that priorities based on general incorporate post-conflict situations explicitly into the experience are likely to be misplaced. For example, existing analysis of the relationship between aid, policy suppose that a post-conflict society starts with each and growth as undertaken in Collier and Dollar (2001). component - macro, structural and social - rated at 2.5, and the matter for judgment is whether a small Two general patterns have emerged from this improvement in social policies at the expense of a small analysis. deterioration in macroeconomic policies would be advisable from the perspective of growth. To be specific, First, we find that aid is considerably more effective let these small changes in policy amount to an in augmenting growth in post-conflict situations than in improvement in social policies to 2.6 and a deterioration other situations. For "poverty efficiency," aid volumes in macroeconomic policies to 2.4. The coefficients on the should be approximately double those in other direct effects of these components of policy suggest that situations. The pattern of aid disbursements should were the situation not post-conflict, such a change would probably gradually rise during the first four years, and reduce growth. By contrast, in a post-conflict situation gradually taper back to normal levels by the end of the growth would be increased by around one percentage first post-conflict decade. Actual aid practice has not, point. historically, followed this pattern. We can also distinguish to an extent between the two possible ways by which policy priorities might be Second, we find that among policies the key distinctive - differential severity of policy problems and priorities for improvement, relative to an otherwise differential effects of policies. Recall that the three similar society without a history of recent conflict, should components of the CPIA are all equally poor in post- be social policies first, sectoral policies second, broadly conflict situations. There does not appear to be with the same priority as in other contexts, and macro differential deterioration. Hence, it appears that the policies last. Again, actual improvements in policies differential importance of social policy is not because during the first decade of peace do not appear to reflect social policies are differentially bad in post-conflict these priorities: all policies other than governance situations, but rather that they are differentially appear to improve more or less in tandem. important. This is indeed consistent with much of the practical policy work in post-conflict situations which CPR Unit tends to prioritize social issues. However, recall that the policy ratings tend to improve through the various ofic poinst-cnflito inmtandem thourganlysi suggests This note was prepared by Paul Collier, of the Development Research phases of post-conflict in tandem. Our analysis suggests Group (DECRG) This note senes is intended to disseminate good that it would be desirable if social policy could improve at practice and key findings on conflict prevention and reconstruction. a faster rate than structural policy, which in turn should The series is edited by the Conflict Prevention and Reconstruction Unit improve at a faster rate than macro policy. This is in the Social Development Department of the Environmentally and decidedly not to say that macro does not matter. In post- SSocially Sustainable Development Network of the World Bank. The views expressed in these notes are those of the authors and do not conflict, as elsewhere, everything matters. But the necessarily reflect the views of the World Bank Group CPR practical process of reform is always a matter of Dissemination Notes are distnbuted widely to Bank staff and are also priorities. available on the CPR website