direct investment, and portfolio investment). Harrelson-Stephens and Calloway (2003) examine three measures of trade openness over a significantly longer period than Apodaca (1976–1996): trade openness (the sum of exports and imports divided by GDP, which captures the level of trade but not the symmetrical nature of trade between states), level of exports relative to GDP, and a 4-point trade liberalization index. They find that, when controlling for the standard human rights model, each of their trade openness measures does demonstrate a decrease in the likelihood of state repression in three separate models, supporting liberal expectations. However, they do caution that the coefficients are rather small. But given the effect over time through the lagged dependent variable, they calculate .5 decrease in the 5-point personal integrity abuse scale in ten years’ time. This is roughly comparable to the impact that Poe, Tate, and Keith (1999) found for the presence of international war and about half the effect they found for economic development, which, interestingly, fails to achieve statistical significance in Harrelson-Stephens and Calloway’s model. Apodaca’s (2007) study, which covers a longer period (1989–2002) than her first study, confirms her earlier finding even though she employs a different measure of trade openness (exports plus imports as a percentage of GDP). Thus much of the initial empirical evidence does support the liberal perspective that open trade has a positive effect on at least one dimension of human rights, the right to personal integrity.
The evidence examining the effect of foreign investment has been much more inconsistent and in one study has varied depending upon the category of repression. Apodaca (2001) found that foreign direct investment (FDI) did lead to moderate decreases in repression of personal integrity rights for the period 1990–1996; however, she also found that the portfolio investment did not have a statistically significant effect. She had in fact hypothesized that this type of investment, which tends to be “short-ventures, highly mobile, and subject to capital flight,” places more restrictions on the regime’s taxing and spending policy options, and in particular restricts spending on social welfare programs while doing little to enhance economic growth (595). In her 2007 study of the years 1989–2002 the effect of FDI disappears, although here her model is more focused on the media and thus controls for an additional set of measures than in her 2001 study. She does not include the portfolio measure here either, so the results are somewhat difficult to compare. Richards, Gelleny, and Sacko’s (2001) analysis of both categories of political repression in 1981–1995 provides some additional insight into these mixed findings. They also find that FDI reduces the likelihood of political repression, but only in regard to the civil liberties restriction category, not the personal integrity category. This finding contradicts Apodaca’s earlier study of the seven-year period but fits with her longer study. However, contrary to Apodaca, they find that portfolio investment decreases the odds of repression of personal integrity rights but has no effect on civil liberties restrictions. They also find that foreign debt increases the repression of civil liberties restrictions but has no effect on personal integrity rights. Ultimately, these mixed results suggest that it is still premature to draw firm conclusions about the effect of foreign investment.
International Treaties
Most empirical studies of state compliance with international human rights treaty obligations focus on the state’s provision or protection of the guaranteed rights embedded in the document. The bulk of the empirical compliance literature, while grounded in rich theoretical debate concerning the influence of human rights treaties, has largely been limited to tests of whether being a state party produces a positive or negative effect, when controlling for a variety of factors. It does not enable us to determine which of the mechanisms that are hypothesized to be at work are actually operative. Realist theory views treaty commitments as cheap talk and therefore predicts no effect or perhaps even a negative effect. Rationalist theory does not provide as clear an expectation, but because of the lack of direct benefits from compliance and the weak enforcement mechanisms that make noncompliance relatively costless (except, possibly, in terms of reputation), treaties would be expected to have a weak effect at best. The decoupling effect posited by the world society approach would predict that we would observe no direct effect and a negative effect over time. Thus, three theories predict little or no effect or even a negative effect. The expectations of the domestic institutions approach would be conditional, expecting better human rights, but only in democratic regimes or regimes with effective legal institutions. The normative perspectives, however, would predict compliance due to norm diffusion and the effect of transnational human rights networks, but perhaps would make the expectations contingent upon the strength of the networks (Neumayer 2005).
Empirical evidence of the influence of participation in an international human rights treaty has been mixed. For example, Keith (1999) finds no effect from the International Covenant on Civil and Political Rights (ICCPR) on personal integrity rights or civil rights and liberties, unless controlling for state derogations. Hathaway (2002) finds that most treaties within the human rights regime do not significantly affect human rights behavior, and that participation in some of the treaties, such as the Genocide Convention and the Convention Against Torture (CAT), produces negative effects, a result that is confirmed by Hafner-Burton and Tsutsui (2005) in regard to a wide range of treaties. On the other hand, the two most exhaustive compliance studies (Landman 2005 and Simmons 2009) found consistent evidence of an association between rights behavior and state commitment to a variety of treaties within the international human rights regime. Landman finds that commitment to the ICCPR and the CAT, even while controlling for the level of reservations, decreased state repression of personal integrity rights, torture, and civil and political rights. Simmons extends Landman’s work and finds that participation in the ICCPR, the Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW), and the Children’s Convention are associated with the provision of several related rights. In the case of the CAT and the ICCPR in regard to fair trials she finds that the influence is limited to a narrow set of circumstances. For example, the impact of the CAT is conditioned on the strong rule of law. In most empirical studies of compliance, the theoretical mechanisms (for example, cost/benefit calculations and power relationships) are merely assumed rather than directly tested in the models. Kelley’s (2007) work on bilateral nonsurrender agreements represents a significant exception and takes advantage of a unique opportunity to test particular costs that a state could not have anticipated when it became a party to the Rome Statute. Contrary to realist expectations, she finds that approximately half of the countries pressured by the United States went against their own self-interest and resisted the pressure, incurring diplomatic and sometimes economic costs. She does find some evidence to support realist assumptions in that states that have Generalized System of Preferences (GSP) status with the United States, poor states, and members of the “Iraq coalition of the willing” were more likely to sign nonsurrender agreements.
We find another notable exception within the asylum literature. A growing body of research has examined the United States’ obligations under the Geneva Convention Relating to the Status of Refugees and under the CAT to respect the norm of nonrefoulement.1 Most of these empirical studies have demonstrated that U.S. foreign policy interests (security and trade relationships) and domestic policy interests (economic and national security concerns) influence decisions on who receives a grant of asylum in the United States (Rosenblum and Salehyan 2004; Salehyan and Rosenblum 2008; Rottman, Fariss, and Poe 2009; Keith and Holmes 2009). Overall, the weak association of human rights with treaty commitment supports realist expectations. However, most of these studies do not specifically operationalize and test the underlying assumptions that would predict no effect; thus our confidence in this conclusion is rather weak. Moreover, Kelley’s findings and those in the asylum literature, which account more directly for realist and rationalist assumptions, suggest that interests may not matter as much as the realists think; and, indeed, their analyses demonstrate that norms and domestic political contexts also matter.
Assumptions of the domestic institutions approach have received much more specific empirical attention than those of realist and rationalist theory; presumably because scholars can more readily observe and more directly measure domestic institutional contexts than cost/benefit calculations or the diffusion of norms. Several studies have examined the conditional or interactive effect of democratic regimes on treaty compliance. When Hathaway (2002) limited her analysis to democratic regimes she continued to find no effect of the ICCPR and negative effects of the CAT on human rights behavior;