State Fragility and Implications for Aid Allocation: An Empirical Analysis

Abstract

In recent years, state fragility has gained importance as a result of the perceived links between poverty, conflict, and global terrorism. In this paper, we examine the relationship between state fragility and aid by evaluating the literature and research programs currently extant. We bring conceptual clarity to the issue by developing and testing an alternative theoretical framework using CIFP's fragility index (articulated around the concepts of authority, legitimacy, and capacity [ALC]) and by using data collected for the period 1999–2005 to identify the empirical determinants of fragility. We then examine the effects of state fragility on aid allocation, using the ALC framework as defined. Our results indicate that aid allocation is directed toward states on the basis of their capacity and authority scores and not on the basis of their legitimacy scores. Finally, we assess the theoretical and policy implications of these findings and specify directions for future research.

Keywords:

• fragile states
• foreign aid
• failed states
• risk assessment
• Country Indicators for Foreign Policy

The authors thank the journal's anonymous reviewers for their helpful comments. We are indebted to Harvey Starr and Glenn Palmer for their suggestions and for giving us the opportunity to be included in this special issue. This research has been made possible through the financial support of the Canadian International Development Agency's Policy Branch. Special thanks to Michael Koros for his courage and foresight.

Notes

Kingwell, 2007. What fragility can teach us. The Globe and Mail (F7, October 27).

Since 1997, the Country Indicators for Foreign Policy (CIFP) project (together with the Canadian government, private sector and nongovernmental organizations) has collected and analyzed statistical information on a range of issues related to the political, economic, social, and cultural environments of countries around the world (see www.carleton.ca/cifp).

Compare the disparate rankings for Pakistan in the 2005 and 2006 editions of the Fund for Peace Failed States Index and then consider our ranking in Carment et al. (2006). In 2005, FFP ranked Pakistan 34th on their list, while in 2006, Pakistan ranked 12th (Fund For Peace, 2005, 2006). While the FFP list may provide some information regarding the relative level of violence in the state, it is unclear what else could drive such a shift in rankings. Moreover, by focusing on failure as its dependent variable, such an index provides little basis for any sort of forward-leaning policymaking, as it provides more information about where crises are currently occurring than when and where they may emerge.

Such an emphasis, exemplified in U.S. support for allies such as Pakistan in the global war on terror (GWOT), leads to potential distortions in both the selection of aid recipients, and the types of aid provided. Aid is provided to states with limited authority and capacity, regardless of the relative legitimacy of the regime resulting in deeply unpopular regimes heavily dependent on external aid that can be unstable over the long term. U.S. aid to Pakistan is predominately military aid.

It should be noted that the identified research programs within these streams do not themselves always use the term fragility. We use the term fragility here as a concept that contains elements that encompass all three streams.

The OECD defines fragile states as “countries where there is a lack of political commitment and insufficient capacity to develop and implement pro-poor policies” (Morcos, 2005). USAID uses the term fragile states to refer to a broad range of failing, failed, and recovering states, admitting that the distinction among them is not always clear in practice; they also identify states that are vulnerable from those that are in crisis, and report that about a third of the world's population now lives in areas that are fragile. The World Bank, on the other hand, focuses its attention on 30 low-income countries under stress (LICUS) that are characterized by a combination of weak governance, policies, and institutions, and ranked among the lowest on the Country Policy and Institutional Assessment (CPIA); the 30 LICUS countries comprise about half a billion inhabitants.

According to the original State Failure Task Force a failed state is one that is “utterly incapable of sustaining itself as a member of the international community” (Esty et al., 1995, 1). Narrowly defined however, “state failures consist of instances in which central state authority collapses for several years” (Esty et al., 1995, 1); emphasis. However, since fewer than 20 such episodes have occurred during the last 40 years, it is difficult to conduct statistical analysis. Therefore, the initial task force broadened the concept of state failure to include a wider range of civil conflicts, political crises, and massive violations of human rights that are typically associated with state breakdown. In line with such a broad definition, the task force isolated four kinds of state failure: (1) revolutionary wars, (2) ethnic wars, (3) mass killings, and (4) adverse or disruptive regime change.

To be sure, there are a variety of other elements posited to have an effect on conflict, state stability, or both. For instance, Cramer (2005) summarizes the relationship between inequality and conflict, concluding that most authors think there is a relationship, but virtually no consensus regarding its nature. To cite a limited number of examples, Muller (1985), invoking Aristotle and de Tocqueville and presenting empirical evidence in support, suggests that individual inequality is positively correlated with political violence. Nairn (1977) finds a link between class-based inequality and violence, while Holsti (1995) and Gurr (1993) highlight the effects of intercommunal inequality. Urdal (2004) finds that the “youth bulge” has a significant effect on the incidence of conflict in states. A number of writers, including Collier et al. (2003) and Hegre et al. (2001), highlight the significant relationship between the prior experience of instability and conflict, and their likely reoccurrence in the future.

In their systems dynamics model of state stability, Choucri et al. (2006) define state stability such that “a state is stable to the extent that its resilience (capabilities) is greater than the load (or pressures) exerted upon it” (Choucri et al., 2006, 2). Choucri et al. operationalize “loads” as the relative pressures exerted by insurgent activities, echoing the PITF. Some authors focus on the relationship between political systems and relative stability, focusing on the link between state legitimacy and stability. Hegre et al. (2001) are representative of a common school of thought, also supported by the Political Instability Task Force, that holds that inconsistent regimes experience greater instability, as measured by the relative incidence of civil war, than either democratic or autocratic regimes. Notwithstanding policy statements to the effect that increased democracy will necessarily decrease the risk of both intrastate and interstate conflict globally, there is little evidence supporting such a linear relationship; there is a growing consensus around the more nuanced “inverted-U” hypothesis articulated by Hegre et al. (2001) at the intrastate level, and Mansfield and Snyder (1995) at the interstate level.

A variety of other studies contribute to our understanding of legitimacy as well. For instance, Fein (1995) suggests a link between the level of political and civil rights in a state and the likelihood of political stability, finding an inverted “U” relationship similar to that found for regime type. Luckham et al. (2001) posit a link between the level of poverty of sub-Saharan Africa and the endemic nature of interstate conflict experienced there, suggesting that the region is particularly vulnerable to the development of conflict-sustaining political economies. Stedman (1996) and Van de Walle (2001) find a link between negative economic growth and instability consistent with Gurr's deprivation theory (1970).

A separate line of argument maintains that fragile states are the product of largely structural factors, including geography, access to the sea (or other international trade routes), the presence of sufficient population, and economic capacity to sustain centralized government. Diamond (1997), Collier (2007), and Collier and Hoeffler (1998) have contributed to this school of thought, although each identifies different processes contributing to the current state of radical inequality found in the world. While Diamond emphasizes the role that physical geography has played on the accumulation of human capital in different regions of the world over the very long term, Collier and Hoeffler (1998) focus on more proximate approach, suggesting that the so-called resource-curse plays an important role in the relative vulnerability of states to violent conflict. According to this argument, countries that are highly resource dependent are at an increased risk of conflict, as such resources can be used to fuel either government intransigence or rebellion. This argument has subsequently been refined to focus particularly on lootable resources, including drugs, secondary diamonds, timber, rubber, and a limited number of precious metals, which tend to be associated with elevated levels of conflict.

For a complete exposition of the 80 indicators that are included within each of the six clusters of our ALC framework see Carment et al. (2006).

The development of political capacity, legitimacy, and authority, all essential features of state-building, is not a linear process. This is especially relevant to explaining state failure, since changing environmental conditions can reverse (in very short periods of time, e.g., months and years) some or all of these essential features. For example, changes in system structure can reverse state-building in at least two non—mutually exclusive ways: through the creation of highly dependent weak states (compounded by the subsequent withdrawal of powerful patron states) on the one hand; and through processes of economic development, and the strengthening of international norms of self-determination, on the other.

Within the CIFP Index, states are scored from 1 to 9 on each indicator included, with 9 being the most fragile (see Table 2A in the Appendix for the fragility index scoring scale). Relevant scores are then averaged to provide overall scores for A, L, and C, as well as for overall fragility. For a full description see Carment et al. (2006).

∗Single measures.

∗p < 0.1.

∗∗p < 0.05.

∗p < 0.1

∗∗p < 0.05

All of these terms lead to the same econometric problem, namely that the independent variables are contemporaneously correlated with the error term.

We have also tested for endogeneity in our analysis in order to make sure that our estimates are not biased.

Aid allocation research answers the questions of who receives aid and why. Aid effectiveness research means large cross-country studies that have looked at the macroeconomic impact of aid on growth. The impact of aid has also been examined at the micro (project) level and in single-country case studies, not within the scope of this study.

More recent studies (for example, Hansen and Tarp [2000] and Clemens, Radelet, and Bhavnani [2004] are more positive and argue that aid works, regardless of the policy environment, effectively nullifying the micro-macro paradox of aid. Other than the lack of an impact on growth, critics of aid have also referred to fungibility, lack of donor coordination, too much tying of aid, lack of absorptive capacity, and the failure of conditionality to buy reforms, among others.

∗p < 0.1.

∗∗p < 0.05.

Aid allocation equations typically control for self-interest, recipient needs, and recipient merits. Here, our concern is determining which of the three main components of state fragility are more likely to influence aid allocation, recognizing that each category contains a number of underlying factors.

∗p < 0.1.

∗∗p < 0.05.

Compared to other extensive studies of aid effectiveness, which tend to rely on panel data, this is a preliminary step analyzing the issue of aid effectiveness using cross-sectional data, given that we have data on fragility for the past five years until 2005.

These findings clearly deserve further examination, which we for now leave for future research. We have in mind a model using panel data, with additional controls and a policy index a la Burnside-Dollar that will hopefully clarify the aid effectiveness debate for fragile states.