Political Repression. Linda Camp Keith

a significant component of U.S. foreign policy and has received considerable attention by scholars, commentators, and critics of the United States, especially during the Cold War. Most scholarly attention has sought to determine what factors actually affect U.S. aid allocation and has been especially concerned with whether a state’s human rights practices affect these decisions as mandated by Congress or whether geopolitical and economic interests truly drive these decisions (for example, Carleton and Stohl 1987; Cingranelli and Pasquarello 1985; Poe 1992; Poe et al. 1994; Poe and Sirirangsi 1994; Meernik, Krueger, and Poe 1998; Apodaca and Stohl 1999; Neumayer 2003a, 2003b, 2003c). Considerably less attention has been given to the question of whether U.S. aid influences human rights practices of recipient states. Regan’s (1995) study is the only empirical study I am aware of that examines this question specifically, beyond two studies supported by the U.S. Agency for International Development (USAID) itself (Finkel et al. 2006, 2008) and a study by Knack (2004), a senior research economist for the World Bank who examined the impact of western foreign assistance on democratization in recipient states. While there are strong theoretical links between aid and improved human rights practices, both direct and indirect, there is also some justification to expect that aid will not influence human rights or that it might have a harmful effect. While Knack agrees that foreign aid can potentially contribute to democratization by providing support for electoral processes, strengthening legislatures and independent judiciaries, promoting of civil society organizations, and improving education and per-capita income levels, he argues that it may also undermine some components of democratization such as accountability and stability. Not only are we left with mixed theoretical expectations, but also the small number of empirical studies examining this question have been somewhat mixed in their findings, though generally pessimistic that aid will reduce levels of political repression. Beginning with Shoultz’s (1981) early analysis of U.S. economic aid to Latin America, we see that aid was sometimes used to support dictatorial regimes and sometimes used to overthrow nonrepressive regimes for more repressive ones that supported key U.S. policy goals during the Cold War. Regan’s (1994) cross-national analysis of U.S. economic aid in 1979–1988 found that it accounted for little change in recipient states’ human rights behavior. In fact, he found that it would take a $250 million increase in economic aid to a state to produce a 1-point change in his 15-point repression scale. It is tempting to attribute these findings to the limited period or geographic region (in the case of Shoultz) under analysis, especially since it is the height of the Cold War and, in the case of Regan, to attribute the result to the possibility that his aggregate index may not appropriately capture all the potential variations in state repression.

      However, two comprehensive analyses of the USAID democracy and governance aid commissioned by USAID itself have consistently demonstrated that while spending in the other subsectors of aid (elections, civil society, and governance) all produce the desired effects, spending on the human rights component of the rule-of-law program has a strong negative effect on states’ human rights protection, regardless of numerous controls, including controls for endogeneity, and regardless of which measure of human rights is employed (Finkel et al. 2006, 2008). As the investigators note in the final report of their second study, their “effort to untangle the web of relationships that may underlie this distressing and presumptively anomalous relationships and to model them statistically has been largely unsuccessful in its basic purpose” (Finkel et al. 2008, 57). Apodaca (2001) has examined the influence of bilateral aid from OECD Development Assistance Committee countries during 1990–1996, and while she does find that the bilateral aid reduced the likelihood of repression, the size of its impact is rather small, and it achieves only marginal statistical significance. Apodaca is more optimistic about her findings than subsequent work supports. Knack (2004) also examined the influence of foreign aid from the OECD countries, but over a twenty-five-year period (1975–2000) that overlaps somewhat with the fifteen-year period of Finkel et al.’s 2008 study (1990–2004) and is inclusive of Apodaca’s entire period. In his more extensive analysis Knack found no effect from either of his measures of development assistance on either of his two measures of democracy. As Knack admits, his data do not allow for the disaggregation of aid intended to promote democracy from aid intended for other purposes, and thus urges some caution in interpreting his results. Thus, while scholars seem somewhat reluctant to accept these null or negative findings, we are left with the dilemma that aid may actually promote political repression, at least under some circumstances, as indeed Knack posited as a counterhypothesis. Recent empirical studies of the influence of multilateral aid programs have been just as pessimistic.

      Multilateral Aid/Structural Adjustment Programs. While some scholars might argue that there are direct links between structural adjustment programs and improved human rights through the program requirements of a reduced or limited state, most neoliberal arguments link structural adjustment programs indirectly to human rights practices through increased levels of economic growth or wealth, which have been weakly linked to increased human rights protection. Critics, however, argue that the programs have harmful effects on the economic well-being of citizens, especially women and the poor (for example, Buchmann 1996; Sadasivam 1997; Zack-Williams 2000; Fields 2003), and on workers’ rights (Abouharb and Cingranelli 2008); and other scholars have argued that regimes’ reduction in subsidies, social welfare programs, and public employment have led to mass protests to which recipient states have responded with repressive action (for example, Pion-Berlin 1984; Keith and Poe 2000; Abouharb and Cingranelli 2008). In addition, increasingly rigorous analyses have consistently demonstrated that the programs do not even lead to economic growth (for example, Pion-Berlin 1984; Harrigan and Mosley 1991; Rapley 1996; Prezworski and Vreeland 2000; van de Walle 2001; Vreeland 2003); thus even the hypothesized indirect benefits seem unlikely.

      These results bring into serious question the expectation that the loans will lead to improved human rights, since this influence is expected to come indirectly through economic growth and political stability. Interestingly, empirical studies have also demonstrated that the programs do not lead to increased political stability but rather increase the probability of civil conflict (for example, Sidell 1988; Di John 2005; Keen 2005; Abouharb and Cingranelli 2008). A growing number of empirical studies have addressed specifically the impact of these programs on human rights. The earliest empirical study to my knowledge is Pion-Berlin’s (1984) analysis of the impact of International Monetary Fund (IMF) loans on repression in Argentina during 1958–1980, which demonstrated that the stabilization programs increased repression against labor groups. Subsequent studies of personal integrity abuse have also demonstrated a link between these programs and increased state imprisonment, torture, killing, and disappearances (Franklin 1997; Keith and Poe 2000). The most comprehensive and rigorous study to date (Abouharb and Cingranelli 2008) demonstrates that not only do the structural adjustment programs (both IMF and World Bank) fail to deliver economic development, but also the agreements fail even to promote increased political stability. They demonstrate that instead, the deleterious economic and social effects of the agreements have destabilized the countries with increased levels of civil conflict—antigovernment demonstrations, riots, and rebellion in particular. The increased civil conflict in turn increases the state’s level of physical integrity abuse (torture, murder, disappearance, and political imprisonment). Indeed, the likelihood of physical integrity abuse increases the longer the state participates in a structural adjustment agreement, even when controlling for selection effects and most of the known factors that affect such abuse. And similarly, the longer the state participates in a structural adjustment agreement, the weaker the state’s protection is of worker rights. Thus, we are left with overwhelming evidence that multilateral aid is generally harmful to a broad range of rights, and that some components of bilateral aid have either no effect or a deleterious effect. The evidence in regard to trade openness and foreign economic penetration is mixed but somewhat more optimistic than that of the aid relationship.

      Trade Openness and Foreign Economic Penetration. Empirical studies that have tested the liberal perspective on the linkage between trade and human rights protection, though small in number, have consistently confirmed these expectations (Apodaca 2001, 2007; Harrelson-Stephens and Calloway 2003). Apodaca (2001, 2007) found that exports had the third-strongest impact in predicting personal integrity abuses in 1990–1996, with only democracy and conflict producing larger effects. Her model controls for several important domestic conditions (conflict, population, democracy, and education expenditures) and international factors (conflict,