Foreign Aid, Democracy and Political Stability in Post Conflict Societies

Foreign Aid, Democracy and Political Stability in Post Conflict Societies Marijke Breuning* and John Ishiyama** What is the relationship between development aid and political stability in post conflict societies? Although there has been a considerable amount of literature that empirically investigates the relationship between development aid and corruption (Tavares. 2003; Alessina and Weder, 2002; Knack, 2000; Rimmer, 2000; Svensson 1998; Ijaz, 1996), the quality of governance (Knack 2001) ethnic conflict (Esman and Herring, 2003; Herring, 2001) and post conflict economic growth (Collier and Hoeffler, 2004; 2002; Hamburg, 2002; Casella and Eichengreen, 1994) no study of which we are aware has examined the direct effects of both the quantity and timing of development aid on promoting political stability in post conflict societies. In this study we examine twenty-six post conflict countries across whose civil wars ended after 1980: Angola, Armenia, Azerbaijan, Burundi, Chad, Croatia, The Democratic Republic of the Congo, Republic of Congo, El Salvador, Ethiopia, Georgia, Guatemala, Indonesia, Iran, Morocco, Mozambique, Nicaragua, Nigeria, Peru, Philippines, Russia, Rwanda, Serbia and Montenegro, Somalia, Tajikistan and Uganda. The dependent variable is measured using Kauffman et al’s (2003) measure of political stability for 1996-2002. The principal independent variables are the amount of aid provided in the periods following the conflict settlement, the timing of aid as well as domestic political factors (such as the extent to which democracy has taken root) and ethnolinguistic homogeneity/heterogeneity. Several scholars have examined the impact of aid on post conflict societies. Many argue that there is good reason to believe that post conflict development aid may help promote political stability. As Hamburg (2001) notes the provision of massive aid will help populations to secure Alternatives: Turkish Journal of International Relations, Vol. 6, No.1&2, Spring & Summer 2007 82 minimum standards of living and restore key infrastructures. Technical assistance is needed for institutional development, adopt new legislation, enhance the quality of governance, and support civil society and democracy. On the other hand, several scholars suggest that development aid may be deleterious to political stability. As Esman and Herring (2003) note, development aid can exacerbate conflict and instability, particularly in ethnically divided societies. Whether or not aid exacerbates conflict depends heavily on the type of aid provided. Alesina and Weder (2002) contend that development aid may promote "wasteful public corruption" and in turn have a negative impact on economic growth and ultimately political stability. Svennson (1998) contends that large amounts of aid money is counterproductive for good public policies since it promotes a form of rentierism and ultimately corruption. Collier and Hoeffler (2002) have conducted some very influential work in which they contend that what is most important in the provision of development aid in post conflict societies, is not so much the amount or the type, but the timing of aid. The timing of aid is critical to achieve aid-growth efficiency. To this end, donors should phase in aid gradually during the first four years after the conflict, and then gradually taper back to normal levels by the end of the first post-conflict decade@ (p.14). in reality donors tend to do the opposite by providing large amounts of aid soon after the conflict when the situation commands world attention, and then reducing aid. Historically aid Ahas tapered out just when it should have been tapering in@ (p.14). On the other hand, many scholars have argued that the impact of external factors on post conflict settlement has been overstated, and that insufficient attention has been paid to intervening variables, such as political institutions, that assist in mitigating strife. This argument is made by Sadowski who contends that the “global chaos” theorists oversimplify the complex relationship between external international forces on conflict, and miss the importance of domestic economic, social and political factors that exert effects independent of external forces (Sadowski, 1998). Another argument is made by Crawford and Lipshutz (1998), who contends that although external economic forces may “trigger” cultural conflict, its effects are mitigated by other factors (Crawford and Lipshutz, 1998, pp. 4-5). In this regard state institutions play a key role. Those institutions define the rules of political membership, representation, and resource allocation. When these institutions structure membership, representation, and resource allocation according to previously established cultural criteria, “identity politics” dominates the political game (Sadowski, 1998, p. 517). Alternatives: Turkish Journal of International Relations, Vol. 6, No.1&2, Spring & Summer 2007 83 Indeed for Crawford and Lipschutz (1998), political institutions that afford representation and an outlet for the aspirations of politicized groups in post conflict scenarios is a key intervening variable. Democratization and openness helps to prevent the consistent denial of representation to important groups (Lijphart, 1985). Further, by securing representation for minority groups, openness serves to facilitate the integration of disaffected groups into the political system, which ultimately leads them to moderate their demands. However, Cohen (1997) argued that broad representation and openness has different effects on what he calls “low level” (non violent “legal” protest) and “high level” (violent, extra-institutional). This is because the broader the representation the more likely the politicized conflict groups feels bound to the existing system. This does not mean that grievances go away but conflicts are channeled away from extra institutional directions (Ishiyama, 2000). Design and Methodology The above studies on the role of foreign aid in post conflict societies, although valuable, have tended to focus on economic performance as opposed to the political performance of recipient countries. In this paper we are primarily interested in examining the impact and timing of development aid on political stability. To that end we measure political stability by an index calculated by the World Bank (2006) based on Kauffman et al. (2002) from 1996-2005. The measure is one of six indicators of the quality of governance, which includes: 1) Voice and Accountability; 2) Government Effectiveness; 3) Regulatory Quality; 4) Rule of Law 5) Control of Corruption. It is conceptually defined as “perceptions of the likelihood that the government will be destabilized or overthrown by unconstitutional or violent means, including domestic violence and terrorism” (World Bank, 2006, p. 2). The measure is based on several hundred underlying variables that reflect perceptions of a wide range of governance issues. The governance indicators are drawn from 31 separate data sets. The data consist of surveys of firms and individuals, as well as the assessments of commercial risk-rating agencies, nongovernmental organizations and think tanks, and multilateral aid agencies. Each of the many individual indicators becomes part of one of the six aggregate indicators A statistical methodology known as the “unobserved components model” is then used to construct aggregate indicators from the individual measures. The aggregate indicators are weighted averages of the underlying data. The Alternatives: Turkish Journal of International Relations, Vol. 6, No.1&2, Spring & Summer 2007 84 resulting political stability indicator is measured in units ranging from about -2.5 to 2.5, with higher values corresponding to better governance outcomes. The higher is the index, the more stable is the country. We take as our dependent variable the average score on the index from 1996-2005 (for further explanation see World Bank, 2006). Our principal external independent variable is the extent to which aid is infused in a country following the settlement of the internal conflict. In particular we are interested in two issues—the amount of aid and the timing. Indeed as Alesina and Weder (2002) and Svennson (1998) contend, the amount of aid should be inversely related to both the development of democracy and political stability, particularly because the infusion of large amounts of aid money is counterproductive for the development of good public policies. However, we would expect that if Collier and Hoeffler (2004; 2002) are correct, that massive infusion of aid provided in the first four years after the conflict (only to be followed by a period of less aid) would produce less politically stable results. To measure both the amount of aid and whether or not that aid was provided early on, we employ two independent variables in our model. The first is the amount of development assistance aid per capita received by the post conflict country in the decade prior to the measurement of political stability (1996-2005). We take the average amount of aid per capita received (in current US dollars) by a country from 1990-2004. The second measure is a dummy variable, taking into account whether the amount of development assistance aid increases or decreases after four years. If the country receives more in the second four year period (or 5 to 8 years following the conflict) as opposed to the first four years (years 1 through 4 after the conflict) then that country is scored a “1” for this variable. If a country receives more aid in the first four year period than the second four year period then the country receives a score of “0”. If Collier and Hoeffler are correct then we would expect that countries that receive more aid in the first four year period would be more prone to be more politically stable. The second independent variable has to do with the extent to which democratic inclusion affects the propensity for politically stable polities to emerge in the wake of a post conflict settlement. We use the combined polity score developed by the Polity IV group at the University of Maryland College Park to measure the degree to which a political system is autocratic or democratic. In part this value is based on the level of autocracy (or autonomy of the state authorities from constraints) and includes measures of the competitiveness of political Alternatives: Turkish Journal of International Relations, Vol. 6, No.1&2, Spring & Summer 2007 85 participation, the regulation of participation, the openness and competitiveness of executive recruitment, and constraints on the chief executive The polity score is computed by subtracting the Autocracy score from the Democracy score; the resulting unified polity scale ranges from +10 (strongly democratic) to 10 (strongly autocratic). We take the average combined polity score for each country in the period 1990-2002, as a lead up to the political stability scores for 1996-2005. If the internal factors are what really affects the extent to which a country develops a political stable outcome, then we would expect that this independent variable is also positively and significantly related to the dependent variable. In addition to these primary independent variables we also control for other independent variables that have been posited to impact on both democracy and political stability. The first is the extent to which a country is ethnically divided. To measure the degree of ethnic homogeneity and heterogeneity in a society, this paper employs the often used measure the proportion of the population made up of the largest ethnolinguistic group, where the higher the value the greater the degree of ethnic homogeneity. In addition, we also examine the number of years that have passed since the end of the conflict. Presumably the greater the number of years that have passed since the end of the post war settlement, the more likely a political stable outcome will have emerged. In the analysis we include 26 post conflict countries: Angola, Armenia, Azerbaijan, Burundi, Chad, Croatia, The Democratic Republic of the Congo, Republic of Congo, El Salvador, Ethiopia, Georgia, Guatemala, Indonesia, Iran, Morocco, Mozambique, Nicaragua, Nigeria, Peru, Philippines, Russia, Rwanda, Serbia and Montenegro, Somalia, Tajikistan and Uganda.. This list is taken from the list of countries whose civil wars ended after 1980 provided by Collier and Hoeffler (2002, p. 24). Table 1 lists the countries included in the analysis and also reports the values of the variables used in the model. Alternatives: Turkish Journal of International Relations, Vol. 6, No.1&2, Spring & Summer 2007 86 Table 1: Values of Variables Country Average Political Stability Score (19962005) Average Combined Polity Score (19902005) % largest Ethnic Group Years of Conflict Angola Armenia Azerbaijan Burundi Chad Congo, Rep. Congo, Dem Rep. Croatia El Salvador Ethiopia Georgia Guatemala Indonesia Iran Morocco Mozambique Nicaragua Nigeria Peru Philippines Russian Federation Rwanda Serbia and Monte negro Somalia Tajikistan Uganda -1.99 -.33 -.69 -1.68 -1.16 -2.39 -1.52 -2.62 3.67 -5.08 -2.31 -3.23 -1.31 -3.30 37.00 93.00 90.00 85.00 40.00 48.00 25.00 75-91 88-94 88-94 .48 .25 -.60 -1.08 -.79 -1.27 -1.24 -.06 -.16 -.10 -1.38 -.58 -.19 -.55 -.83 6.92 .15 .80 5.69 -2.54 -7.00 -6.76 2.23 7.24 -2.46 3.00 8.00 2.60 89.60 90.00 40.00 70.00 55.00 45.00 52.00 99.10 99.66 69.00 29.00 45.00 91.50 70.00 -1.26 -1.64 -5.77 -3.07 84.00 37.10 90-94 -1.50 -1.87 -1.21 -.54 -2.34 -4.69 85.00 64.90 17.00 88-92 92-97 80-88 88 80-88 91-95 96-97 97 79-92 74-91 91-93 78-84 75-82 81-82 75-89 76-92 82-90 80-84 82-96 72-96 94-96 91 Dummy Variable Aid Higher in Second Four Year Period after the Conflict than First Four Year Period 14.00 11.00 11.00 17.00 17.00 10.00 8.00 Aid Provided per capita in 2000 constant dollars in 10 years following conflict 47 60.41 14.51 220.74 168.32 1.59 219.46 8.00 13.00 14.00 12.00 21.00 23.00 23.00 16.00 13.00 15.00 21.00 9.00 9.00 9.00 16.15 17.43 60.11 79.72 15.26 4.13 0.14 10.96 150.88 115.15 3.72 3.82 4.18 2.32 1.00 .00 .00 1.00 1.00 1.00 .00 .00 .00 1.00 1.00 .00 .00 1.00 233.79 .00 1.00 Number of Years since the end of Civil War (up unit 2005) 11.00 14.00 13.00 8.00 17.00 . . 88.43 104.33 1.00 1.00 1.00 .00 .00 1.00 1.00 .00 1.00 1.00 Alternatives: Turkish Journal of International Relations, Vol. 6, No.1&2, Spring & Summer 2007 87 Results Table 2 reports the results of regressing the dependent variable political stability (1996-2002) against the independent variables, the average combined polity score for 1990-2002, the percentage of the population made up of the largest ethnic group, number of years since the end of the conflict (up until 2002), the average aid per capita received by the country (1990-2002), and the dummy variable for whether or not aid received in the second four year period after the end of the conflict exceeded the amount of aid received in the first four year period after the end of the conflict (see Table 2). As indicated in Table 2 below, the two variables that relate to both the amount and timing of foreign aid (“Average Aid per capita” and the “Dummy variable aid greater later or earlier”) are unrelated to the average political stability score for 1996-2002. Further, the sign of the dummy variable is opposite to what was anticipated by the work of Collier and Hoeffler (2002). It appears that more aid offered in the second four years when compared to the first four years after a conflict is inversely related to political stability. Thus, in terms of timing, providing aid later as opposed to earlier does not necessarily promote political stability. Table 2: Coefficient Estimates and Collinearity Diagnostics, Political Stability, Amount and Timing of Foreign Aid , and Internal Political and Social Variables Variable Coefficient VIF (standard error) Average Combined Polity .06* 1.15 Score (1990-2002) (.03) Percentage of Population .01* 1.26 Largest Ethnic Group (.005) Years Since Conflict End .004 1.08 (.02) Average Aid per capita .002 1.09 (1990-2002) (.004) Dummy variable aid greater -.25 1.17 later or earlier (.26) *p< .05 Adjusted R-square= .36 N= 26 On the other hand, the primary internal variable, the combined polity score was both statistically significant and the single most important variable in the model explaining political Alternatives: Turkish Journal of International Relations, Vol. 6, No.1&2, Spring & Summer 2007 88 stability (with the Beta coefficient at .44 as compared to both .09 and -.02 for the two aid variables). Also significant was the variable that measured the percentage of the population made up by the largest ethnic group. This indicates, not surprisingly, that countries that were more ethnically homogenous were more likely to be politically stable. Finally, the number of years since the end of the conflict was unrelated to the extent to which a political system was relatively more politically stable. In all the model was not plagued by collinearity problems (as indicated by the low variance inflation factor –VIF- scores). This also indicates that the independent variables were largely unrelated (i.e. that foreign aid was uncorrelated with the level of democracy/autocracy). Conclusions In sum, the results reported above are consistent with the literature that suggests internal political variables are far more important in explaining political stability in post conflict societies than are external variables (like the provision and timing of foreign aid). However the results do not support the notion that foreign aid, either the amount nor the timing, is related to the emergence of political stability in post conflict states. Ultimately, this also suggests that perhaps a relatively more effective means to promote political stability in post conflict societies than merely the provision of foreign financial assistance is the careful design, during the post settlement period, of political institutions that promote inclusiveness of political groups. This does not mean that international forces cannot impact on the political course in post conflict countries. However, it does suggest that the provision of technical assistance particularly via the design of political institutions may be more important than merely the provision of financial aid in the rebuilding of countries torn by civil war. Although these findings are somewhat preliminary, the above results suggest further questions for future investigation. First, what kinds of political institutions best promote political stability? Second, perhaps it is not the amount of aid that makes a difference in promoting political stability but the type. In other words, perhaps aid that is designed to promote civil society institutions may be more effective and providing voice to potentially disaffected populations and hence better promote political stability. Although the evidence thus far cannot yet answer these questions, the findings above indicate that further investigation into political Alternatives: Turkish Journal of International Relations, Vol. 6, No.1&2, Spring & Summer 2007 89 consequences of foreign aid in post conflict societies is warranted, and represents a promising avenue for future inquiry. *Truman State University, Division of Social Science, Kirksville Missouri, USA. An earlier version of this paper was presented at the 2005 Annual Meeting of the American Political Science Association, September 1-4 2005, Marriott Wardman Park, Washington DC. References Alesina, Alberto and Beatrice Weder (2002) "Do Corrupt Governments Receive Less Foreign Aid?" American Economic Review 92 (4): 1126–1137 Casella, Alessandra & Barry Eichengreen, (1994). "Can Foreign Aid Accelerate Stabilization” NBER Working Papers 4694, National Bureau of Economic Research, Inc Cohen, Frank S. (1997) AProportional Versus Majoritarian Ethnic Conflict Management in Democracies.@ Comparative Political Studies 30 (4): 607-630. Collier Paul and Anke Hoeffler (2002) “Aid, Policy and Growth in Post-Conflict Societies” World Bank Policy Research Working Paper 2902 Crawford Beverly & Ronnie D. Lipschutz eds. (1998). The Myth of Ethnic Conflict: Politics, Economics, and Cultural Violence. Berkeley: UC Berkeley International and Area Studies Press. Esman Milton J. & Ronald J. Herring (2003) Carrots, Sticks, and Ethnic Conflict : Rethinking Development Assistance Ann Arbor: University of Michigan Press Hamburg, David A. (2002) No More Killing Fields: Preventing Deadly Conflict. Lanham, MD, Rowman & Littlefield Publishers Herring, Ronald J. (2001) “Making ethnic conflict: The civil war in Sri Lanka.” In Carrots, sticks, and ethnic conflict: Rethinking development assistance edited by Milton J. Esman and Ronald J. Herring, 140-174. Ann Arbor, MI: The University of Michigan Press. Ijaz Mansoor (1996) "A Primer on the Perils of Foreign Aid," Wall Street Journal, October 2. Ishiyama, John (2000) AInstitutions and Ethnopolitical Conflict in Post-Communist Politics Nationalism and Ethnic Politics 6(1):51-67. Kaufmann Daniel A. Kraay, and M. Mastruzzi (2003) Governance Matters III: Governance Indicators for 1996–2002 World Bank Policy Research Working Paper 3106 Knack, Steven (2001). “Aid Dependence and the Quality of Governance: Cross-Country Empirical Tests”. Southern Economic Journal 68 (2): 310–329. Knack, Steven (2000). Does Foreign Aid Promote Democracy? World Bank, Washington, DC Lijphart, Arend (1985) Power Sharing in South Africa. Berkeley: University of California Press. Rimmer, Douglas (2000) "Review Article. Aid and Corruption", African Affairs, 99(392): 121-128. Sadowski, Yahya (1998) The Myth of Global Chaos. Washington, DC: Brookings Institution Press. Alternatives: Turkish Journal of International Relations, Vol. 6, No.1&2, Spring & Summer 2007 90 Svensson, Jakob 1998). “Foreign aid and rent-seeking,” World Bank, Development Economics Research Group, Policy research working paper 1880. Tavares, Jose (2003) “Does foreign aid corrupt?” Economics Letters 79 (2): 99–106 World Bank (2006) A Decade Measuring the Quality of Governance. Washington D.C.: The World Bank. Alternatives: Turkish Journal of International Relations, Vol. 6, No.1&2, Spring & Summer 2007 91