2nd MULTI-conference2nd session: The UN system

Scott Gates

(Research Professor at PRIO & Associate Professor at Michigan State University),

Anke Hoeffler

(Research Officer at the Centre for the Study of African Economies,

Oxford University)

Paul Collier

(Director, Development Research Group, World Bank)

Aid, Policy, and Conflict:

The Multilateral System and Conflict Prevention

Paper prepared for presentation at the 2nd MULTI-seminar, Asker, Norway, 18-19 January 2001. The authors welcome comments. and .

Abstract

This paper is based on our research proposal to the NFR, Environment & Development MULTI-program. Multilateral development assistance agencies increasingly provide aid to conflict ridden areas. Unavoidably, whether before, during, or after civil violence, aid has an effect on conflict. Likewise, conflict has an effect on aid. The purpose of this project is to investigate the effects of economic policies and aid flows upon the risk of civil conflict and to consider how aid can best be allocated so as to enhance the prospects of peace. The policy implications for the multilateral system will be studied in close collaboration with the Development Research Group at the World Bank.

The project consists of two stages. The first stage features the patterns of aid allocations with regard to conflict. Comparisons between bilateral and multilateral assistance allocations (EU and World Bank) will be made. New data will be collected. Africa will receive special country-by-country attention. The second stage features more policy relevant analyses. We will concentrate on policy advice with regard to how multilateral aid can be used to prevent conflict.

Aid, Policy, and Conflict:

The Multilateral System and Conflict Prevention

1.0Project Description

Multilateral development assistance agencies increasingly provide aid to conflict ridden areas. Unavoidably, whether before, during, or after civil violence, aid has an effect on conflict. Likewise, conflict has an effect on aid. The purpose of this project is to investigate the effects of economic policies and aid flows upon the risk of civil conflict and to consider how aid can best be allocated so as to enhance the prospects of peace. The policy implications for the multilateral development aid system will be studied in close collaboration with the Development Research Group at the World Bank. The principal investigators include: Scott Gates (Research Professor at PRIO & Associate Professor at Michigan State University), Anke Hoeffler (PRIO Research Associate & Research Officer at the Centre for the Study of African Economies, Oxford University), and Paul Collier (Director, Development Research Group, World Bank).

The project will consist of two stages. The first stage will focus on the patterns of aid allocations with regard to conflict and develop a dataset on the distribution of aid within the recipient country. Three subprojects assess these patterns and linkages between aid and conflict. They are as follows:

  • Patterns of Aid & Conflict: Comparing Multilateral & Bilateral Assistance
  • Conflict & the Phases of Conflict
  • Political Stability, Aid & Conflict

The second stage features a more policy relevant analysis. We will concentrate our policy analysis on how multilateral aid can be used to prevent conflict. The central topic to consider is how aid can best be allocated so as to enhance the prospects of peace. We also plan to investigate the policy implications for the multilateral development system more generally. In turn we hope to be able to provide policy recommendations for shaping a better integrated policy for the South and for the Norwegian government with respect to the multilateral system.

1.1 Aid, policy and conflict -- a theoretical overview

Very little has been written on the relationship between aid and conflict (Anderson 1999). Much of what we propose here is exploratory. We intend to integrate several bodies of literature, ranging from theories of conflict, political stability, economic growth, and the political-economy of the multilateral system. The literature on the analysis of conflict can be broadly categorized into two parts. On the one hand there is the theoretical analysis of conflicts and on the other hand there is a large number of case studies. Most of these studies are generated by political scientists, anthropologists and development scholars. Recent work examining the role of political and social institutions in civil conflict is most relevant to our study (Fearon & Laitin 2000; Hegre et al. 2000; Moore 1998). Only a few papers analyse the causes of civil conflict within the framework of economic theory (Roemer 1985; Hirshleifer 1991; Grossman 1991) and there are very few studies that test these predictions empirically. Our research aims to make a contribution in terms of improved modelling of the economics and politics of conflict and in terms of testing these theoretical propositions empirically. We hope to make further contributions by integrating economic models based on individual choice and political science models of political institutions and organisations. The policy recommendations will be based on the results of the theoretical and empirical findings.

1.2 Bilateral and multilateral aid and conflict

Much of the aid literature has focused on the use of aid in poverty reduction and on aid as a catalyst of economic growth and economic reform. To our knowledge there is no systematic review of aid allocation and conflict. For this project we want to analyse to what extent conflicts determine aid allocations. One interesting question is whether bilateral and multilateral aid allocations are distinct and if this is the case whether multilateral aid allocations are preferable in terms of peace promotion.

The recent work by Alesina and Dollar (2000) examines which donor countries give aid to whom and what the main determinants of this bilateral aid allocation are. Their analysis shows that only a few donors, namely the Nordic countries, are motivated by poverty reduction, democratisation, and openness. Other countries (notably France) give aid to former colonies without much regard to other factors. The United States’ aid allocation is vastly influenced by that country’s interest in the Middle East. Countries such as Japan seem to give aid to political allies (defined by the similarity of UN voting patterns) and according to Japan’s economic interests. The study also presents some evidence that donors tend to reward countries that democratise (measured by the change in Gastil’s democracy index over time), however, they do not tend to cut aid drastically for countries that experience decreases in democracy. Collier and Dollar (2000) also find that aid increases with political reform. However, aid is reduced after the economic reform process, thus there is no long-term reward mechanism.

During the past decade the share of multilateral aid has increased. Analogous to the work by Alesina and Dollar (2000) we want to analyse the determining factors of bilateral as well as multilateral aid allocation. Distinguishing between the different multilateral aid agencies and individual donor countries we want to examine whether conflicts are a determinant of aid allocations. In addition, if multilateral aid allocations are less determined by political allegiances, colonial ties and geo-strategic interests, one policy implication could be that more aid should be given through multilateral agencies.

1.3 Aid and the phases of conflict

This aspect of the project aims to examine how aid has been used in conflict prevention and how the effectiveness of the use of aid can be increased. This analysis will have three main parts. First, we want to present an empirical analysis of the pattern of aid allocation before, during and after civil wars. Secondly, building on previous work we examine the causes and duration of war and thirdly, we make suggestions as to how aid could be used more effectively for conflict prevention.

Empirical studies show that countries that have a history of internal conflict are more likely to experience further conflict. Collier and Hoeffler (2000a) suggest that while conflict in countries without a history of internal war is mainly motivated by greed (loot-seeking), grievance created by the first conflict is likely to be a cause for renewed conflict. This suggests that conflict prevention strategies should distinguish between countries that have not experienced a previous civil war and countries that have.

Prelimary work on the duration of war and post-war peace (Collier, Hoeffler, and Söderbom, 1998) suggests that there are distinct patterns of duration. A large number of civil wars end within one year. However, a small number of wars continue for very long periods. Econometrically the war to peace transition probability is characterised by a downward sloping hazard function. In other words if internal conflicts are not resolved within a short time, the longer the conflict has lasted the more likely it is to continue. Conflicts become self-sustaining. The findings for post-war peace are similar. In many cases post-war peace is short lived and the fighting resumes within one year of the end of the last conflict. However, once peace has been sustained for some time (about five years) the risk of renewed conflict decreases and peace becomes self-sustaining. Hegre et al (2000) provide similar findings.

We propose to extend this work by examining how the interaction between aid flows and conflict changes from one phase of conflict to another. To do this we will need additional data. In particular we will need information regarding the distribution of aid resources within a recipient country. We also will need to gather data regarding the geographical location of the fighting, where the conflict is initiated and where the rebel camps are located. By analysing these data we can begin to ascertain how aid (particularly emergency aid) can shape conflict. In turn we can evaluate how conflict effects aid distribution at a micro level.

1.4 Political stability, conflict & aid

To better understand the relationship between conflict and aid we need to understand it in the political context of the recipient country. Civil war is a real danger in states having recently experienced political transition. Hegre et al. (2000) conclude that greater institutional changes are more hazardous than smaller ones and moves towards autocracy appear to be more dangerous than moves towards democracy. With this subproject, we propose to extend the analysis in Hegre et al. by examining all forms of regime transition. Rather than focusing exclusively on civil war we propose to examine civil war in the broader context of political stability and instability.

In studying political instability we feature polity change (or regime collapse) and polity duration.[1] The polity is a broader unit of analysis than the duration of political leaders or the hazards of governmental downfalls. Moreover, the polity features institutional characteristics rather than the personal qualities of a leader or the partisan characteristics of a government. Extending Weingast’s (1997) theory of the foundations of democracy, we posit that democratic institutional stability depends on a self-enforcing equilibrium while autocratic institutions exhibit asymmetric equilibria. Thereby it is in the interest of political officials themselves to work to maintain the institutional structure of a polity, whether through autocratic control or by the self-enforcing equilibria of democracy. Where such equilibria are lacking, instability will follow. Using this theoretical structure we hypothesize that strongly autocratic and strongly democratic regimes will exhibit the greatest stability. The in-between regimes (semi-democracies) will be least stable. Institutional maturation (the duration of a polity) only serves to reinforce these tendencies, such that relatively young polities are more likely to collapse.

Using a Cox proportional hazard model we can assess the relative hazards of regime collapse, accounting for different institutional structures and histories of political instability, including violent instability. By examining the underlying cofactors associated with regime duration and political instability, we can assess those institutional structures that lead to political instability. We can also assess those institutional characteristics at greatest risk of experiencing violent and nonviolent instability. In turn, we can then link this analysis to our study of the pattern of aid flows. This will allow us to see how recipient countries’ political institutions interact with aid with respect to conflict. This analysis will be integrated with our analysis of aid and the phases of conflict (section 1.3). More generally the connection between aid and political stability will also be examined. Does aid help prop up regimes that would otherwise have collapsed? To what extent does the multilateral system help promote (or encourage) different policies and political institutional structures?

1.5 Aid & conflict prevention – policy analysis

Our policy analysis will focus on how multilateral aid can be used to prevent conflict. How can aid be allocated so as to enhance the prospects of peace? What are the policy implications for the multilateral development aid system. In turn, we aim to provide policy recommendations for the Norwegian government with respect to the multilateral system. We also hope to be able to provide advice for the shaping of a integrated set of policies to reduce the chances of conflict in the South.

The notion that aid may be ‘fungible’ has recently received empirical support (Devarajan et al. 1999). If a donor gives aid for a project that the recipient government would have undertaken anyway, the aid finances expenditures other than the intended project. Donors should therefore be concerned about the quality of public spending programs in recipient countries. With regard to conflict, donors should be aware of how aid can be used unintentionally to finance conflict or the conditions that promote conflict.

Work by Collier and Hoeffler (2000a), Collier et al. (1998), and Hegre et al. (2000) suggests that conflict prevention should distinguish between countries that already experienced a civil war and those that have not. Since greed (or loot-seeking activity) can be seen to be the main motivating factor of conflict in countries without a history of war conflict prevention should concentrate on generating higher income and economic growth and reducing natural resource dependency. Conflicts create and increase existing grievances, thus, in post-conflict societies conflict prevention strategies should concentrate on addressing objective and perceived grievances. The timing of conflict intervention/prevention seems also crucial. Early intervention seems to have higher chances of success while post-war assistance needs to be long enough to generate a peace that will become self-sustaining. We will draw extensively from the research discussed in sections 1.1-1.5 to draw policy conclusions.

1.6 Summary and conclusion

Why should aid donors be worried about conflict? Why should we care about the connections between aid and conflict? Peace is a global public good. Conflict not only causes immense human suffering in the countries in which the fighting takes place, but a large number of externalities are created as well. Internal conflict often flows over into neighbouring countries. Refugees seek safe places beyond the national border as well as the rebel fighters who have often base camps in neighbouring countries. Ethnic, religious and political allegiances frequently draw governments or part of the population in neighbouring countries into the conflict, thus destabilising the entire region. Easterly and Levine (1998) show that the spillover effects from destabilising macro-economic policies are large and seriously hinder economic development in neighbouring countries. To our knowledge the spillover effects of conflict have not been studied in a systematic way, but they are likely to have a much larger impact on the neighbouring countries’ development than bad macro-economic policies. Conflicts cause a large number of displaced persons seeking refuge within and outside the region. International interventions are costly in terms of human lives and resources and are often unsuccessful in bringing lasting peace to conflict ridden countries. Terrorism can result as a by-product of conflict with devastating consequences reaching far beyond the national boundaries of the country at war.

Much of our work proposed here is exploratory in nature and oriented towards theory-building. We plan to publish most of our research in international peer-reviewed journals. We also hope to publish work that is useful and accessible for the policy-making community. Ultimately we hope to learn how to better use the multilateral system to prevent conflict.

2.0Methodology & Data

Our economic and political analysis is based on the principles of rational choice and in our empirical analysis we use econometric methods appropriate for the analysis of panel as well as duration data. The data sets we use are mainly based on publicly available data sources such as the economic indicators provided by the World Bank, aid data from OECD, policy and political institutional indicators from Polity IVd, and the conflict data from the Correlates of War Project, and the Department of Peace and Conflict Research, Uppsala University, and the US State Department. An important part of the project is to compile a global data set by pooling and updating information from the different data sources. We plan to participate in a conference on conflict data to be held in Uppsala (Identifying Wars: Systematic Conflict Research and Its Utility in Conflict Resolution and Prevention. Uppsala, April 2001. Co-sponsored by the EU, the World Bank, and the Department of Peace and Conflict Research, Uppsala University).