Determinants of US foreign policy in multilateral development banks

Abstract

The purpose of this study is to identify and explain recent US human rights policy in the Multilateral Development Banks (MDBs). Foreign aid, whether distributed directly (bilateral aid) or indirectly through multilateral institutions such as the MDBs, is one of several tools through which the USA furthers its human rights policy. Several studies show the conflicted human rights policies the USA pursues with bilateral aid, but very few examine the role of human rights in US multilateral aid policy. It is this deficit that the present study addresses by examining what factors determine how the USA votes on proposals before the Executive Board in the various MDBs. Using a multinomial logistic regression model I test whether US votes in the MDBs are conditioned on recipient countries’ respect for human rights along with other strategic interest variables. The study finds that respect for political rights has an important role in determining US votes in the MDBs although much of the impact is accounted for by votes specifically against China. Conversely, respect for rights of personal integrity do not have an impact on US voting. Level of economic development and whether the country receives military aid from the USA are also important determinants of US votes, as well as the level of trade between the USA and the recipient country.

Keywords

human rights multilateral development banks US foreign policy

Introduction

Are US votes in the multilateral development banks (MDBs) conditioned on recipient countries’ respect for human rights? In other words does the USA sanction countries with poor human rights records by voting to deny them loans in the MDBs? Human rights have become an important fixture of US foreign policy over the last several decades. However, the specific content of US human rights policy, as well as its place among foreign policy priorities more generally, has varied considerably over the years (Forsythe, 2002; Mertus, 2004; Apodaca, 2006; Blanton & Cingranelli, 2010). Despite the establishment of a stable body of international law on human rights from which to draw, there has been significant variability in how the USA promotes human rights in its foreign policy, what rights it promotes, and what countries are targeted over others. The most common explanations for this inconsistency is that US human rights policy is simply a mask for pursuing core national security or economic interests, or in other words the liberalism of human rights is trumped by the realpolitik of self-interest.

Foreign aid, whether distributed directly through bilateral aid or indirectly through the MDBs, is one of the economic tools through which the USA furthers its human rights policy (Apodaca, 2006). Using foreign aid to influence another country’s human rights policies involves either using assistance as a positive reinforcement – providing aid to those countries that have made improvements in the protection of human rights – or as a negative sanction – withholding aid to countries that have an overall awful human rights record or are currently engaged in massive violations of human rights. Several studies highlight the inconsistency with which the USA promotes human rights with bilateral aid – conditioning aid on the human rights practices of some states but not others. However, except for a brief flurry of research in the 1980s, very few studies examine the role of human rights in US multilateral aid policy. The USA votes on whether it will support or oppose every loan or grant proposal that comes before the MDBs. Therefore, the purpose of this research is to assess the question of whether US votes in the MDBs are conditioned on recipient countries’ respect for human rights.

Literature review

MDB background and lending

Multilateral development banks are international institutions comprised of member states and dedicated to the alleviation of poverty through the financing of development projects in poor countries. Towards that end the MDBs generally engage in three types of activities. First, their traditional activity is providing financing for large-scale investment projects that are deemed important for development. Second, the MDBs also provide loans for policy reforms. These are non-project loans with conditions attached that provide governments with financing in exchange for specific domestic policy reforms. Finally, the MDBs also conduct and disseminate research on development. This is especially true for the World Bank, which produces a large quantity of scholarly research on development every year (Babb, 2009).

There are five major development banks operating today. The World Bank, founded in 1944, is by far the largest lender, distributing billions of dollars a year in low-interest loans and grants to the poorest countries in the world. In addition to the World Bank there are four regional banks – the Inter-American Development Bank (IDB), the Asian Development Bank (AsDB), the African Development Bank (AfDB), and the European Bank for Reconstruction and Development (EBRD) – which target their loans to countries within their respective geographical regions of the world. According to Babb (2009: 23), the MDBs are ‘members of a common family and share family traits’. In other words, the MDBs are separate organizations but they share many common characteristics especially when it comes to their organization and decisionmaking structure.

The MDBs are staffed and run by international civil servants but the main supervisory bodies of the MDBs – the Boards of Governors and the Boards of Executive Directors – represent member state governments. The Executive Board is where the majority of decisionmaking occurs within the MDBs including decisions on lending and policy proposals. Lending decisions made by the Executive Boards are based on majority consensus with formal votes only taken on rare occasions. Despite the fact that formal votes are rarely taken, the decisions on loan projects take place in what Lyne, Nielson & Tierney (2009: 413) describe as, ‘the shadow of the formal majority rule’. In other words, projects that are unlikely to garner majority support of the Executive Board are not brought before the board for approval.

This brief overview on the MDBs outlines their background and purpose as well as their Executive Board decisionmaking process. An additional question to consider with regards to the USA and the MDBs is for what benefits, or to serve what ends, does the USA choose to channel a measure of its foreign aid through multilateral institutions? The question is relevant because as the hegemonic state the USA is giving up a significant measure of autonomy to allow its resources to be distributed by a multilateral body. Ikenberry (2001) argues that the USA pursued a strategy of building multilateral institutions after World War II in an effort to ‘lock in’ other states to an international order that is inherently favorable to the USA. The USA was willing to give up some of its sovereignty to multilateral institutions to create an institutionalized international order that serves its long-term interests.

An additional benefit of channeling aid through multilateral institutions such as the MDBs is burden-sharing. Milner & Tingley (2010) suggest that support for multilateral aid from the US public derives from notions of burden-sharing and cost control. Americans tend to support multilateralism, to the extent that they do, because they think it ‘will be cheaper than unilateralism in the long run’ (Lake, 1999: 220). Finally, another way of looking at why the USA would cede some of its sovereignty to join international organizations is because in practice it does not actually give up much sovereignty. Since the USA exercises considerable influence in international organizations it can influence the policies that IOs pursue. Both quantitative (Fleck & Kilby, 2006; Kilby, 2006) and qualitative studies (Wade, 2002; McKeown, 2009) have shown that the decisions made by the MDBs are often well matched with US policy

In summary, the USA operates in multilateral organizations such as the MDBs because they help achieve common goals, which the USA either cannot achieve on its own, or cannot financially afford to. The major downside of multilateralism for individual states is that it restricts autonomy and range of action, but the USA through its exercise of informal influence is largely able to make up for those restrictions in the MDBs. The next issue to consider is exactly what US preferences are regarding foreign aid and human rights.

Human rights and US multilateral aid policy

There is considerable research on the question of whether human rights are a significant determinant of MDB loans (Neumayer, 2003a,b; Lebovic & Voeten, 2009) and whether MDB loans and structural adjustment programs are detrimental to human rights (Abouhard & Cingranelli, 2006, 2008; Eriksen & de Soysa, 2009). Research on whether human rights influence US policy in the MDBs however has not been as prevalent. With the general interest in human rights developing in both Congress and the Carter administration in the 1970s, and with the passage of the International Financial Institutions Act in 1977 (IFI Act, 1977), there was an initial surge of interest in the question of human rights in US multilateral aid policy. The IFI act, specifically section 701, instructed the US Executive Directors (USEDs) in the MDBs to use their ‘voice and vote’ to direct assistance away from countries which systematically violate individual human rights unless the assistance would directly contribute to the basic human needs of the recipient countries citizens. Despite this early interest in human rights and multilateral aid the question has not received sustained attention since the 1980s. The original research into the question generally concluded that human rights took a back seat to geopolitical concerns during the Cold War, and the ideology of anti-communism was a significant motivator of US voting behavior in the MDBs.

According to Spiro (1977) US human rights policy regarding multilateral aid consists of using bilateral tactics within a multilateral environment. Two specific tactics stand out: first, trying to create a human rights coalition among other like-minded countries willing to deny loans for human rights reasons, and, second, pursuing a human rights policy that also serves the economic and geopolitical interests of the United States. Sanford (1982) points out, however, that from the period of 1977 (when the IFI act went into effect) to 1981 there is no discernible evidence that abstentions and ‘no’ votes by the USA for human rights reasons affected lending. He does note, however, that as a signaling mechanism, US ‘no’ votes can have an impact. In addition, he points out that the votes painted a somewhat inconsistent picture as the decision to vote against certain loan proposals stemmed not so much from the human rights record of the recipient country as from its ideological orientation (anti-communist vs. pro-communist) (Sanford, 1982).

Schoultz (1982) also examined US votes in the MDBs during the same time period and found that human rights were one of the main reasons the USA voted ‘no’ on particular loans (although ‘no’ votes overall were quite rare). But he also points out how the ‘no’ votes cast ostensibly for human rights reasons also had a distinct ideological dimension, in that the ‘no’ votes were more likely to be cast against a communist state. In addition, Schoultz (1982) points out the overwhelming presence of Latin American countries that were targeted by the USA for their poor human rights records, with Argentina receiving one-fourth of all opposing votes.

In a survey that spanned the Carter administration as well as the first Reagan administration, Forsythe (1988) argued that although the Carter administration was not enthusiastic about section 701 of the IFI act, viewing it as a constraint on executive authority, it did vote against numerous loans in the MDBs on human rights grounds. During the first Reagan administration however a double standard emerged from using the provision primarily against certain left-wing or communist governments while voting for loans to right-wing or authoritarian governments (Forsythe, 1988). Mower (1987) finds a similar pattern in both the Carter and Reagan administrations. In summary, this literature points out that US votes against loans for human rights reasons had a distinct Cold War bias in that the recipient state’s ideology, specifically if the state was communist, as well as its relationship with the USA, appears to be the primary reason for the ‘no’ votes, with human rights a secondary concern.

Human rights and US bilateral aid

One limitation of this literature review is that all of the works under review examine the time period from 1977 to the late 1980s and primarily rely on descriptive statistics without testing the full array of variables that may determine US votes in the MDBs. In contrast to the thin literature on US human rights policy and multilateral aid, there is an extensive literature on the place of human rights in US bilateral aid. An overview of this literature focusing specifically on the post-Cold War era supplies additional variables that enable a more fully specified overall test of the place of human rights in recent US foreign policy in the MDBs.

Meernik, Krueger & Poe’s (1998) expansive study suggests that three broad factors govern the distribution of foreign aid. The first includes systemic factors that primarily encompass national security concerns. The second includes societal factors that include liberal economic and free market policies driven by domestic business interests. The third category includes ideological concerns regarding promotion of democracy, human rights, and development. Meernik, Krueger & Poe conclude that since the end of the Cold War, security factors have declined in importance while ideological concerns have increased. Specifically they found that human rights were significant in determining whether a country would receive aid from the USA. However, of those countries that did receive aid, those countries with worse human rights records tended to receive a greater amount of aid than countries with better human rights records. They speculate that once a country passes some basic minimum standard of human rights protection they are not punished for the less severe violations they may commit (Meernik, Krueger & Poe, 1998).

Apodaca & Stohl (1999) expanded the scope of previous works on US aid and human rights. They found that at the gate-keeping stage (the stage at which it is decided whether a country receives aid or not) human rights impacted whether a country received US aid or not, except during the Clinton administration. In addition, human rights were important in determining the amount of aid a country received – countries with awful human rights records received less aid. They also found that national security interests played a significant role, with countries viewed as important to US national security receiving aid despite their record on human rights (Apodaca & Stohl, 1999).

More recent work from Lai (2003) focuses on the gate-keeping stage of bilateral aid by looking at the initial disbursement of aid to a country, as opposed to which countries receive aid each year. He found human rights only predicted the initial amount of aid a state was allocated and this was only significant after the Cold War (Lai, 2003). The most recent scholarship posits that the influence of human rights on bilateral aid decisions is dependent on both the regime type of the recipient country and its level of economic development. Demirel-Pegg & Moskowitz (2009) find that human rights are not an important determinant of US aid for countries transitioning to democracy or established democracies, nor are they an important determinant for economically underdeveloped countries. They did find that human rights are an important determinant for aid going to autocracies, with autocracies that have poor human rights records receiving less aid (Demirel-Pegg & Moskowitz, 2009).

From the literature review one can see that human rights have been an important but uneven component of US foreign aid policy, becoming more of a factor in the post-Cold War environment. More specifically, these studies show that economic and security concerns, or strategic interests, often take priority over human rights norms as a determinant of aid, although the shift in the international security environment away from the bipolarity of the Cold War has somewhat changed the content of those priorities as well as their importance.

Research design

Hypotheses

From the literature review one can derive two types of hypotheses: those that test human rights claims and those that test strategic interest claims. Regarding human rights claims the first hypothesis is:

H1: The USA is more likely to vote against countries in the MDBs with poor human rights records.

Turning to strategic interests that may influence US votes in the MDBs, I test two hypotheses that touch on both dimensions of strategic interests: supporting allies and promoting domestic economic prosperity. The first hypothesis tests the economic connection between the recipient country and the USA through the volume of trade between the two states:

H2: The USA is more likely to vote against countries in the MDBs with which it has a lower volume of trade.

The alliance component of the strategic interests dimension will be tested using two different measures – UN General Assembly voting similarity and military aid:

H3: The USA is more likely to vote against countries in the MDBs if it has a low voting similarity with those countries in the UN.

H4: The USA is more likely to vote against countries in the MDBs to which it does not give military aid.

Finally, I also hypothesize that the USA will support loans to poorer countries. This may happen regardless of their human rights record or strategic interest to the USA. Since the purpose of multilateral aid is to channel money to the poorest states and the USA is operating in these institutions, one can assume that the USA ultimately supports the raison d’être of the MDBs.

H5: The USA is less likely to vote against poorer countries in the MDBs.

Data description

The main data source used in this study is the number of US votes in the MDBs by recipient country for each year under study for a total of 13,107 observations. The temporal dimension of the data ranges from 2004 to 2011 and the spatial dimension includes 166 countries.¹ The dataset is unbalanced since not every country seeks approval for projects in the MDBs every year. The data are skewed toward the ‘yes’ votes, with the USA expressing approval for particular MDB decisions 87.10% of the time while voting ‘no’ 4.42% and abstaining from voting 8.48% of the time. However, since the number of observations overall is large, there is enough variation in the dependent variable to enable meaningful inferences. Also because such a small percentage of loan projects are rejected by the USA, those that are rejected enhance the signal the USA is sending. Since rejection is used sparingly, recipient states (and other states as well) must take notice.

After the initial swell of research on human rights and US voting in the MDBs in the late 1970s and early 1980s, interest among scholars tapered off. This is partially because of data limitations on US votes. The IFI Act required the Treasury Department to report to Congress, on a quarterly basis, its votes in the MDBs. This practice was discontinued in the early 1990s as the Treasury Department claimed that it had insufficient staff to meet the requirement of monitoring US participation in the MDBs (Sanford, 1997). The Consolidated Appropriations Act of 2004, specifically section 581 of Division D, required the Treasury Department to make public US voting decisions in the MDBs by publishing them on its website (2012). The voting record published by the Treasury Department contains a brief description of the projects voted on, which Bank and Lending Window it was brought before, the date of the vote, the amount of the loan, and how the USA voted on the proposal (‘yes’, ‘no’, ‘abstain’).

The main limitation in utilizing the voting record of the USA in the MDBs is that the USA only votes on projects that have come before the Executive Boards for formal approval. One cannot discern from the voting record the US position on projects that do not make it to the Executive Board for final approval, and there is evidence to suggest that the USA has considerable informal influence at the stages before a loan proposal makes it to the Executive Board. It has been suggested that proposals that the USA would seriously disapprove of are stopped well before they get to the Executive Board stage (Wade, 2002; Woods, 2003). In addition, several studies have shown that the USA has been particularly successful in affecting MDB activities through informal channels involving staffing decisions and choice of leadership (Babb, 2009; Andersen, Hansen & Markussen, 2006; Woods, 2003; Wade, 2002). Ultimately this creates potential bias in the data as the projects from countries the USA favors may be more likely to make it to the Executive Board stage to be voted on than projects from countries the USA does not support. Nevertheless, despite this limitation it is still useful to use US votes to analyze US foreign policy in the MDBs since they still provide a picture of what US priorities are in the MDBs.

Dependent and independent variables

The main dependent variable used in this study is votes in the MDBs (USVOTE) by country and year. The voting takes place on the Executive Board of each MDB and US votes have been made publicly available by the Treasury Department on its website since 2004 (2012). The range of vote choices available are ‘yes’, ‘no’, or ‘abstain’. A ‘yes’ vote expresses approval for the particular project under consideration, while a ‘no’ vote expresses disapproval of the project. By abstaining from voting the USED is signaling a lesser means of disapproval than if they voted ‘no’ outright. According to Rebecca M Nelson & Martin A Weiss, ‘generally, the United States abstains in order to show opposition, though it may cast a negative vote for emphasis in particular situations’ (2013: 9). Therefore the dependent variable for US votes on MDB lending decisions is ‘yes’ = 2, ‘abstain’ = 1, and ‘no’ = 0.

The main independent variables are: human rights, supporting allies, promoting domestic economic prosperity, and level of development. Additional control variables included in the model are ideology, conflict, and population. The human rights category is broken down into two components: political rights (POL) and personal integrity rights (PI). Political rights are measured using the Freedom House survey of government respect for political rights (Freedom House, 2012). The variable POL is the country’s score on Freedom House’s measure of a country’s respect for political rights. The organization Freedom House releases an annual Freedom in the World survey, which rates countries on their respect for political rights on a 7-point scale with 1 indicating the highest level of freedom and 7 representing the lowest level of freedom. Under Freedom House’s designation the category of political rights includes free and fair elections, ability to form political parties, and popular sovereignty (Freedom House, 2006). To ease interpretation the original scale is inverted so 1 represents the lowest level of political freedom and 7 represents the highest level of political freedom. Many studies on human rights and US foreign aid policy often include a measure of regime type in their models. Political rights are closely related to democracy and the Freedom House political rights measure used here is strongly correlated with the common measures of democracy including Polity IV. The Pearson correlation between the two measures is 0.85 and in robustness checks substituting the Polity IV score of regime type with Freedom House political rights measure did not alter the results substantially.²

Personal Integrity (PI) rights are operationalized using the Political Terror Scale (PTS), which uses US State Department and Amnesty International country reports to create a 5-point scale of repression of personal integrity rights (Gibney, Cornett & Wood, 2010). The personal integrity rights violations encompassed by the PTS include extrajudicial killing, torture, disappearances, and political imprisonment (Gibney, Cornett & Wood, 2010). A PTS score of 5 indicates large-scale and indiscriminate repression while a 1 describes countries with little to no threats to individual personal integrity. To ease interpretation the original PTS scale is inverted so 5 represents countries with the greatest respect for individual personal integrity while 1 represents countries with the least amount of respect for personal integrity rights.

Conceptually there is overlap between the measures of political rights and personal integrity rights. However, in terms of the scales used in this study there is actually very little overlap. Previous studies that have used both Freedom House and the Political Terror Scale have also found little overlap in the measures (Hafner-Burton & Ron, 2009). The Pearson correlation between the two scales used in this study is 0.433, indicating that the two scales are measuring different aspects of human rights. Additionally there is support in the literature suggesting a diversion in support and promotion for political rights and rights of personal integrity (Lebovic & Voeten, 2009; Neumayer, 2003a). Finally, robustness checks using alternative measures of personal integrity rights yield similar results.

Turning to strategic interests, I have hypothesized that the USA will be more supportive of proposals for countries that it considers allies (either formal or informal). ‘Allies’ therefore is measured in two ways. First is whether a country receives military aid (MILAID) from the USA or not. Whether a country receives military aid from the USA or not is a good indicator of how important an ally the USA perceives that country to be, or wants that country to be. The reasons why the USA chooses to give military aid to recipient countries are often complex and multifaceted. Usually military aid is given by the USA to exert some measure of influence in the recipient country, primarily in the name of national security. Receipt of military aid from the USA indicates a security relationship between the USA and the recipient country and therefore signals the country as a military ally (Rosenblum & Salehyan, 2004). As with previous studies which use US military aid (Rosenblum & Salehyan, 2004), I created a dummy variable coded 1 for countries that receive military aid from the USA and coded 0 for countries that do not. The justification for coding military aid in this way is to accommodate the wide range in the amount of aid the USA distributes to various countries. Certain countries receive large amounts of aid because of their importance to the USA and also because of their size and capacity, while smaller states will not require as much aid regardless of their strategic importance to the USA. Military aid data are gathered from the US Overseas Loans and Grants (USAID, 2012) document, otherwise known as the ‘Greenbook’, published annually by USAID.

In addition to military aid, strategic interests are also operationalized using a country’s voting consistency with the USA in the UN General Assembly (UNVOTE). Voting similarity in the UN General Assembly is a common measure to assess how close countries are politically (Kim & Russett, 1996; Alesina & Dollar, 1998; Gartzke, 1998; Voeten, 2004). I measure a country’s voting similarity with the USA as a country’s alignment with the USA on important UN votes. The US State Department (2012) annually produces the United States Participation in the United Nations document, which identifies those votes in which direct US interests are at stake, and votes in which the USA lobbied extensively to persuade other countries to support its position. Narrowing the focus of UN General Assembly votes to only ‘important’ votes provides a better measure of congruency between the USA and other countries since these votes are by definition ones that the USA feels the most strongly about, perceiving its interests to be at risk. A country’s vote is coded 1 if it is congruent with the US vote (either both ‘yes’ or both ‘no’), 0.5 if the country abstained from voting, and 0 if the country’s vote is the opposite of the USA’s (voted ‘no’ when the USA voted ‘yes’ or voted ‘yes’ when the USA voted ‘no’).

Economic importance is measured by the volume of trade (TRADE) a country has with the USA, which is operationalized as the sum of US exports to and imports from a recipient country in a given year. In the model a natural log scale is used for level of trade. Volume of trade is an appropriate measure of how important a country is economically to the USA. The greater the trade volume between countries, the greater the opportunity for a slight variation in the volume of trade to influence the economic performance of each country, therefore making the countries highly dependent on one another. The export and import data come from the US Census Bureau on foreign trade statistics (2012a).

Finally, level of economic development (DEV) is measured using GDP per capita in US dollars. As with level of trade, level of economic development is also transformed into a natural log scale for the model. GDP per capita provides a measure of both a country’s economic output and its level of economic development. GDP per capita (or level of economic development) also serves as a proxy for recipient countries’ needs (Demirel-Pegg & Moskowitz, 2009). Stated another way, GDP per capita represents the demand a country has for development projects. Higher GDP per capita numbers represent less need or less demand, and lower GDP per capita represents more need or greater demand. GDP per capita information comes from the IMF World Economic Outlook database (2012).

The control variables include whether the country is involved in an ongoing conflict (CONFLICT), its ideology (IDEOLOGY), and its population (POPULATION). Whether the country is involved in an ongoing conflict is a dichotomous variable coded 0 if the country is not involved in a conflict for that year and 1 if the country is involved in a conflict for that year. The conflict variable is included to account for the fact that human rights abuses and the presence of an ongoing conflict in a country tend to be correlated. Also the presence of an ongoing conflict may inhibit the desirability for approving development projects. The conflict data are taken from the Center for Systematic Peace Major Episodes of Political Violence 1946–2012 dataset (2012), which defines conflict as a major episode of political violence that involves more than 500 directly related fatalities.

The ideological variable (IDEOLOGY) is coded 1 if a country currently has a communist government or is a post-communist country. All other countries are coded 0. The inclusion of the ideological variable is to determine if the ideological bias in US voting during the Cold War is still evident in the post-Cold War era. Finally, POPULATION is a continuous variable that measures a country’s population for the given year. In the model a natural log scale is used for population. The inclusion of population is intended to account for any bias against countries with large populations, since those countries have greater capacity for MDB projects and hence will account for more votes. The population data are taken from the US Census Bureau (2012b). Additionally, in keeping with the scholarly practice, all independent variables are lagged one year to reflect the information available to decisionmakers when voting decisions were made (Meernik, Krueger & Poe, 1998; Apodaca & Stohl, 1999; Lai, 2003; Demirel-Pegg & Moskowitz, 2009).

Estimation

This article uses multinomial logistic regression to test the hypothesis under investigation. Although the categories of the dependent variable are partially ordered in that the options for US voting in the MDBs are ‘no’, ‘abstain’, and ‘yes’, multinomial logistic regression is utilized over ordered logistic regression because the parallel regression assumption is violated (Long, 1997). The categories are partially ordered in that a ‘yes’ vote is giving approval while ‘no’ and ‘abstain’ votes are withholding approval but voting ‘no’ sends a stronger signal of disapproval (Nelson & Weiss, 2013).

Additionally, when dealing with cross-national data it is likely that the errors will not be independent as situations that occur in one country often have spillover effects that influence other countries. In the area of human rights one can imagine that two countries involved in conflict with each other might both see their record on human rights deteriorate due to the conflict. To account for potential heteroskedasticity I employ robust standard errors clustered around country-years. Finally, due to the unbalanced nature of the panel using fixed effects would help to mitigate any potential bias. However, since many of the independent variables vary considerably across countries but do not vary considerably over time, the fixed effects model dropped too many observations for meaningful inferences to be made. Therefore that method was not utilized. As Allison (2009: 3) notes, ‘If predictor variables vary greatly across individuals but have little variation over time for each individual, then fixed effects estimates will be very imprecise’.³

Table I.

Multinomial logistic estimates: Likelihood of US ‘no’ votes and abstentions in MDBs, 2004–11

	‘No’ vote vs. ‘yes’ vote		‘Abstain’ vs. ‘yes’ vote
Variable	Coefficient	Change in probability (‘no’)	Coefficient	Change in probability (‘abstain’)
Political rights	–0.084 (0.069)	–0.01	–0.202*** (0.043)	–0.09
Personal integrity rights	–0.411 (0.228)	–0.06	0.062 (0.082)	0.02
Military aid	–1.25*** (0.275)	–0.04	–0.656*** (0.166)	–0.05
UN vote similarity	1.70** (0.652)	0.04	–0.616 (0.417)	–0.03
Level of trade	–0.588*** (0.091)	–0.70	–0.005 (0.048)	0.05
Level of econ. development	0.921*** (0.145)	0.25	0.072 (0.075)	0.01
Ideology	0.869** (0.291)	0.02	0.419*** (0.126)	0.03
Conflict	–0.217 (0.286)	–0.004	–0.464** (0.166)	–0.03
Population	0.452*** (0.126)	0.12	0.165* (0.068)	0.09
Constant
Number of obs.	12,180
Wald chi2(18)	259.06
Prob > chi2	0.000
LL	–4,979.591

*p < .05, **p < .01, ***p < .001. Robust standard errors clustered around country years included in parentheses.

Results

The results are found in Table I, which presents both the multinomial logistic regression coefficient and the change in predicted probability of a US ‘no’ vote compared to a US ‘yes’ vote and of an abstention by the USA compared to a ‘yes’ vote by the USA. The results show that respect for political rights is significant for the abstention compared to ‘yes’ vote category only. Respect for rights of personal integrity was not significant in either category. This shows that human rights do play some role in determining US voting in the MDBs. Specifically, respect for political rights has an impact in that the worse a country’s record is on political rights the greater the likelihood the USA will abstain from voting on a proposal from that country as opposed to supporting that country’s proposal. Looking at the other categories, whether the country receives military aid from the USA is significant for both categories. Countries that receive US military aid are less likely to receive a ‘no’ vote or an abstention than countries that do not receive US military aid. Voting similarity with the USA in the UNGA is also significant in the ‘no’ vote compared to ‘yes’ vote category. However, the sign is in the opposite direction than what was hypothesized. The results show that the greater the voting similarity with the USA, the greater the likelihood the USA will vote ‘no’ as opposed to ‘yes’ for proposals from that country. Both level of trade and level of economic development are significant in the ‘no’ vote compared to ‘yes’ vote category. The higher the level of trade between the country in question and the USA, the less likely the USA will vote ‘no’ as opposed to ‘yes’ for proposals brought by that country. Conversely, the greater the level of economic development the greater the likelihood that the USA will vote ‘no’ as opposed to ‘yes’ for proposals from that country.

For the control variables, whether the country has a communist government or is a post-communist country is significant for both categories and indicates that communist and post-communist countries are more likely to receive a ‘no’ vote or abstention than a ‘yes’ vote from the USA. Presence of conflict is only significant in the abstention compared to ‘yes’ vote category, with countries involved in a conflict less likely to receive an abstention from the USA compared to a ‘yes’ vote. Finally, population level is significant for both categories, with countries that have a larger population more likely to receive a ‘no’ compared to a ‘yes’ vote and also more likely to receive an abstention compared to a ‘yes’ vote.

Table I also contains the change in predicted probability of a US ‘no’ vote for a change from the minimum to maximum value for each independent variable and the change in predicted probability of an abstention for a change from the minimum to the maximum value for each independent variable. Regarding the political rights variable, one can see a decrease in the predicted probability of an abstention when moving from countries that do not respect political rights to countries that do. There is a 9% decrease in the predicted probability of the

Table II.

Probability of US ‘no’ votes and abstentions in the MDBs, 2004–11, given the value (highest and lowest) of political rights, military aid, level of trade, and level of economic development

Dependent variable	Independent variable	Political rights
		1		7
		Probability	(95% CI)	Probability	(95% CI)
‘No’ vote	Military aid
	No	0.073	(0.035–0.112)	0.053	(0.023–0.083)
	Yes	0.025	(0.011–0.038)	0.016	(0.010–0.023)
	Level of trade
	Low	0.413	(0.166–0.659)	0.319	(0.101–0.537)
	High	0.0014	(−0.0002–0.003)	0.001	(−0.0001–0.002)
	Level of econ. dev.
	Low	0.0038	(0.0008–0.007)	0.0025	(0.0004–0.005)
	High	0.191	(0.060–0.323)	0.138	(0.054–0.221)
‘Abstain’	Military aid
	No	0.196	(0.137–0.255)	0.070	(0.045–0.095)
	Yes	0.119	(0.089–0.150)	0.039	(0.029–0.049)
	Level of trade
	Low	0.080	(0.029–0.132)	0.031	(0.012–0.049)
	High	0.131	(0.064–0.197)	0.043	(0.020–0.065)
	Level of econ. dev.
	Low	0.115	(0.072–0.159)	0.037	(0.022–0.053)
	High	0.124	(0.074–0.175)	0.044	(0.027–0.061)

The predicted probabilities are computed based on the values of the political rights variable (values in the Independent variable row) and values of three additional independent variables (military aid, level of trade, and level of economic development, which are values in the Independent variable column). High and low levels of Level of Trade are defined as two standard deviations above and below the mean. High and low levels of Level of Economic Development are defined as two standard deviations above and below the mean.

USA abstaining in the MDBs when comparing countries that severely violate political rights to those that fundamentally respect political rights. If a country receives US military aid, the probability of a ‘no’ vote decreases by 4% and the probability of an abstention decreases by 5%. The probability of a US ‘no’ vote also decreases by 70% for countries with a high volume of trade with the USA as opposed to those countries with a low level of trade. Conversely, the probability of a US ‘no’ vote increases by 25% for wealthy countries. Being a communist country or a post-communist country increases the probability of a US ‘no’ vote by only 2% and increases the probability of an abstention by 3%. Being engaged in a conflict decreases the probability of an abstention by the USA by 3%. Finally, the probability of a US ‘no’ vote increases by 12% and the probability of an abstention increases by 9% for countries with large populations.

The main focus of this study is on human rights, and respect for political rights is significant in determining US abstentions in the MDBs. Therefore it is important to examine how influential political rights are in different contexts such as when a country receives military aid from the USA, or whether a country has a low level of economic development or a relatively high level of trade with the USA. Table II displays the predicted probabilities of ‘no’ votes and abstentions from the USA by comparing states that severely violate political rights (POL = 1) and states that respect political rights (POL = 7)⁴ and different contexts such as whether a country receives military aid or not, whether it has a high or low volume of trade with the USA, and whether it has a high level of economic development or not.

The probabilities displayed in Table II show that both political rights and strategic interests – such as whether a country receives US military aid or if it has a high volume of trade with the USA – condition votes in the MDBs. For example, a country with low respect for political rights that does not receive US military aid has a 7.3% probability of receiving a ‘no’ vote from the USA. Conversely, a country that has high respect for political rights and does receive US military aid only has a 1.6% probability of receiving a US ‘no’ vote. There is a much starker difference when examining countries that vary on their level of trade with the USA and respect for political rights. Countries that have a low level of respect for political rights and a low volume of trade with the USA have a 41% probability of the USA voting ‘no’ on their proposals in the MDBs, while countries with a high level of respect for political rights and a high volume of trade with the USA only have a 0.1% probability of receiving a US ‘no’ vote in the MDBs. Additionally one can also see that a country’s level of economic development also conditions US votes even in conjunction with its respect for political rights. Countries that have a low level of economic development and low respect for political rights have a 0.4% probability of receiving a US ‘no’ vote, but countries with a high level of economic development and a low level of respect for political rights have a 19% probability of receiving a US ‘no’ vote.

The abstention category yields somewhat different results from the ‘no’ vote category in Table II, especially with regards to level of trade and political rights. For countries that have a low level of respect for political rights and a low level of trade with the USA, the probability of a US abstention is 8%; however for countries that also have a low level of respect for political rights but have a high volume of trade with the USA, the probability of a US abstention is 13%. This is the reverse of what was seen in the ‘no’ vote category, with countries with low levels of both respect for political rights and volume of trade with the USA much more likely to receive a US ‘no’ vote (43%) than those with an equal level of respect for political rights but a high level of trade with the USA (0.14%). This might suggest that respect for political rights is a greater driver of determining US abstentions than is level of trade with the USA. This is reinforced when comparing the probability of a US abstention for two countries that both have high levels of trade with the USA but one country does not respect political rights (13% probability of an abstention) and the other does (4.3% probability of an abstention).

The anomaly presented in Table II with regards to abstentions for countries with high levels of trade with the USA and lack of respect for political rights makes sense when one considers the number of abstentions from the USA for proposals from China. China has the fifth most votes over the period under investigation with 423 in total. What makes China stand out, however, is the number of ‘no’ and ‘abstain’ votes the USA cast against proposals from China, particularly the abstentions. Of the 423 votes on proposals from China, the USA voted ‘no’ on ten of those proposals and abstained on 233 of those proposals. Taken together the USA voted ‘no’ or abstained from voting on 57% on the proposals from China.⁵ China also represents a significant percentage of the worst violators of political rights observations in the study (POL = 1). Twenty-four countries in the dataset (at least for one year) had a score on the political rights variable of 1 indicating the least amount of respect for political rights. In total there were 1,142 observations of the worst violators of political rights but 423 of those were China, which accounts for 37%.⁶ Finally, China’s human rights record was given by the USA as one of the primary reasons for voting ‘no’ on those proposals.⁷

As can be seen above, the USA abstains or votes ‘no’ on most of the proposals brought by China. This is because the USA considers China to be an automatic basic human needs vote, meaning that the USA will only support proposals from China if they support basic human needs.⁸ This is part of a broader strategy of economic sanctions that the USA has leveled on China since the Tiananmen Square massacre and that have been maintained to varying degrees since (Rennack, 2005). Therefore, one important thing to investigate is whether the significance of the political rights variable is driven just by China.

The results for the model (Model 2) without the China observations are presented in Table III. The political rights variable is still significant for the ‘abstain’ vs. ‘yes’ vote category, as it was in the full model. However, the change in predicted probability is much less when compared with the full model (4% vs. 9%). This suggests not only that the worst violators (POL = 1) of human rights drive a lot of the significance of the political rights variable but also that votes against China specifically drive a lot of the significance as well. As with the full model, military aid is significant for both categories, as is ideology. In contrast to the full model, however, voting similarity with the USA in the UN, level of trade, and level of economic development are now significant for both categories. Population level was significant in

Table III.

Multinomial logistic estimates: Likelihood of US ‘no’ votes and abstentions in MDBs, absent China (Model 2), 2004–11

	‘No’ vote vs. ‘yes’ vote		‘Abstain’ vs. ‘yes’ vote
Variable	Coefficient	Change in probability (‘no’)	Coefficient	Change in probability (‘abstain’)
Political rights	–0.079 (0.071)	–0.01	–0.116** (0.040)	–0.04
Personal integrity rights	–0.410 (0.230)	–0.06	0.025 (0.079)	0.01
Military aid	–1.23*** (0.286)	–0.04	–0.404** (0.153)	–0.02
UN vote similarity	1.68** (0.653)	0.04	–1.00** (0.388)	–0.05
Level of trade	–0.592*** (0.096)	–0.66	–0.104* (0.049)	–0.01
Level of econ. development	0.926*** (0.148)	0.24	0.169* (0.081)	0.04
Ideology	0.858** (0.293)	0.02	0.256* (0.131)	0.01
Conflict	–0.201 (0.293)	–0.004	0.006 (0.166)	0.0006
Population	0.450*** (0.126)	0.12	0.05 (0.063)	0.02
Constant
Number of obs.	11,757
Wald chi2(18)	179.42
Prob > chi2	0.000
LL	–4,579.341

*p < .05, **p < .01, ***p < .001. Robust standard errors clustered around country years included in parentheses.

both categories in the full model but is only significant in the ‘no’ vs. ‘yes’ category in the reduced model. Finally, the presence of conflict is the only variable that was significant in the full model but not significant in the reduced model. As was mentioned above, China had the fifth most votes in the data analyzed and a majority of those were ‘no’ or ‘abstain’ votes. However, China is also one of the USA’s largest trading partners. When China is removed from the data the level of trade variable becomes even more prominent and is significant for both categories presented in Table III as opposed to only the ‘no’ vote vs. ‘yes’ vote category in Table I. The implication is that in general the USA is more likely to support countries in the MDBs if they have a high level of trade with the USA; however, China is a major exception.

Conclusion

In conclusion, the hypothesis that human rights influence US votes in the MDBs receives mixed support. Recipient country respect for political rights was a significant factor in determining US abstentions in the MDBs but respect for rights of personal integrity was not significant. In other words, countries that abused the political rights of their citizens were more likely to face a sanction from the USA in the form of an abstention on proposals they brought forth in the MDBs. However, countries that abused the personal integrity rights of their citizens faced no sanction. Previous studies have also found a similar pattern (Lebovic & Voeten, 2009). A potential explanation for this distinction between support for countries that respect political rights and not for countries that respect rights of personal integrity is that support for political rights is often intertwined with the practice of democracy. Democracy promotion is an important component of US foreign policy overall and of the Bush administration in particular, which was creating policy for the majority of the time period under investigation in this study. More broadly, the US approach to foreign aid has been to emphasize performance-based lending, which stipulates that recipient countries should meet certain policy requirements before receiving aid, and should also be rewarded with aid for meeting those requirements (Babb, 2009).

Moving to strategic interests, whether a country receives military aid from the USA has an impact on whether the USA supports loans to that country in the MDBs. Voting similarity with the USA in the UN was significant in the ‘no’ vs. ‘yes’ vote category but the sign was not in the hypothesized direction. This means that a greater degree of voting similarity between the country in question and the USA in the UN leads to a greater chance the USA would vote ‘no’ on proposals for that country in the MDBs. Complicating the issue further is that, although it was not significant, in the abstention vs. ‘yes’ vote category the sign for UN voting similarity was in the hypothesized direction. Additionally, in Model 2 (as presented in Table III) voting similarity with the USA in the UN was significant for both categories but the signs were in the opposite direction. It is hard to tease out the exact implications of this finding, but possible explanations are that UN voting similarity is not a good measure of US strategic interests or that other factors such as level of trade with the USA and support for political rights are more important.

The volume of trade a country had with the USA was significant for one category in the full model and for both categories in the model with the China observations removed. This indicates that trade is a significant driver of US votes in the MDBs and also shows the outside influence abstentions against China have. The level of economic development of the recipient country was also a significant factor and in the hypothesized direction in determining US ‘no’ votes compared to ‘yes’ votes in the full model and was also significant and in the hypothesized direction in Model 2 for both categories. In other words, the USA was less likely to vote ‘no’ on loans going to poorer countries than relatively wealthier countries. When proposals from China were removed the USA was less likely to abstain on loans going to poorer countries than relatively wealthier countries. Additionally, the main control variables of ideology, conflict, and population were all significant in one or more of the categories in both models with the exception of the conflict variable in Model 2.

In summary, the general findings of this study are in line with findings from studies focusing on the post-Cold War distribution of bilateral aid. Those studies found that strategic goals, such as supporting allies with military aid, and normative goals, such as promoting political rights, are both foreign policy goals the USA pursues with bilateral aid. Those are also policies the USA pursues in the MDBs. On top of that the USA did also align itself with the mission of the MDBs by supporting proposals for the poorer countries in the MDBs.

In broader terms, what emerged from this study is a notion that there are multiple factors that influence US foreign policy and human rights are an important but limited one. This study fits with previous works on human rights and US foreign policy in that it shows that human rights have a variable role in US foreign policy. Or in other words, human rights in US foreign policy are context specific. What was seen in this study is that human rights, particularly political rights, influence US foreign policy but they have to be situated in the broader context of strategic interests. This can also be taken as evidence that strategic interests and normative values are not altogether separate issues and are instead perhaps reinforcing and not competing.

Footnotes

Replication data

The dataset, codebook, and do-file used for the empirical analysis in this article, as well as the online appendix, can be found at .

Acknowledgements

The author would like to acknowledge David Rapkin, David Forsythe, Ross Miller, Danielle Shea, and three anonymous reviewers for their helpful comments and suggestions. Any errors or omissions are solely the responsibility of the author.

Notes

References

Abouhard

M Rodwan

Cingranelli

David L

(2006) The human rights effects of World Bank structural adjustment, 1981–2000. International Studies Quarterly 50(2): 233–262.

Abouhard

M Rodwan

Cingranelli

David L

(2008) Human Rights and Structural Adjustment. Cambridge: Cambridge University Press.

Alesina

Alberto

Dollar

David

(1998) Who gives foreign aid to whom and why? Journal of Economic Growth 5(1): 33–64.

Allison

Paul D

(2009) Fixed Effects Regression Models. Thousand Oaks, CA: Sage.

Andersen

Thomas Barnebeck

Hansen

Henrik

Markussen

Thomas

(2006) US politics and World Bank IDA-lending. Journal of Development Studies 42(5): 772–794.

Apodaca

Clair

(2006) Understanding US Human Rights Policy. New York: Routledge.

Apodaca

Clair

Stohl

Michael

(1999) United States human rights policy and foreign assistance. International Studies Quarterly 43(1): 185–198.

Babb

Sarah

(2009) Behind the Development Banks. Chicago, IL: University of Chicago Press.

Blanton

Shannon L

Cingranelli

David L

(2010) Human rights and foreign policy analysis. In: Denemark

Robert A

(ed.) The International Studies Encyclopedia. Blackwell Reference Online (http://www.blackwellreference.com/public/tocnode?id=g9781444336597_yr2011_chunk_g978144433659710_ss1-12).

10.

Center for Systemic Peace (2012) Major Episodes of Political Violence 1946–2012 (http://www.systemicpeace.org/warlist.htm).

11.

Demirel-Pegg

Tijen

Moskowitz

James

(2009) US aid allocation: The nexus of human rights, democracy, and development. Journal of Peace Research 46(2): 181–198.

12.

Eriksen

Silja

Soysa

Indra de

(2009) A fate worse than debt? International financial institutions and human rights, 1981–2003. Journal of Peace Research 46(4): 485–503.

13.

Fleck

Robert

Kilby

Christopher

(2006) World Bank independence: A model and statistical analysis of US influence. Review of Development Economics 10(2): 224–240.

14.

Forsythe

David P

(1988) Human Rights and US Foreign Policy: Congress Reconsidered. Gainesville, FL: University of Florida Press.

15.

Forsythe

David P

(2002) US foreign policy and human rights. Journal of Human Rights 1(4): 501–521.

16.

Freedom House (2006) Methodology (http://www.freedomhouse.org/report/freedom-world-2006/methodology#.UzoSdxbY8lI).

17.

Freedom House (2012) Freedom in the World 2012 (http://www.freedomhouse.org/report/freedom-world/freedom-world-2012#.UzoTZRbY8lI).

18.

Gartzke

Erik

(1998) Kant we all just get along? Opportunity, willingness, and the origins of the democratic peace. American Journal of Political Science 42(1): 1–27.

19.

Gibney

Mark

Cornett

Linda

Wood

Reed

(2010) Political Terror Scale 1976–2008 (www.politicalterrorscale.org).

20.

Hafner-Burton

Emilie M

Ron

James

(2009) Seeing double: Human rights impact through qualitative and quantitative eyes. World Politics 61(2): 360–401.

21.

Ikenberry

G John

(2001) After Victory: Institutions, Strategic Restraint and the Rebuilding of Order after Major War. Princeton, NJ: Princeton University Press

22.

International Financial Institutions Act (1977) Public Law 95–118.

23.

International Monetary Fund (2012) World Economic Outlook Database (https://www.imf.org/external/pubs/ft/weo/2012/02/weodata/index.aspx).

24.

Kilby

Christopher

(2006) Donor influence in multilateral development banks: The case of the Asian Development Bank. Review of International Organizations 1(2): 173–195.

25.

Kim

Soo Yeon

Russett

Bruce

(1996) The new politics of voting alignments in the United Nations General Assembly. International Organization 50(4): 629–652.

26.

Lai

Brian

(2003) Examining the goals of US foreign assistance in the post-Cold War period, 1991–96. Journal of Peace Research 40(1): 103–128.

27.

Lake

David A

(1999) Entangling Relations: American Foreign Policy in Its Century. Princeton, NJ: Princeton University Press.

28.

Lebovic

James

Voeten

Eric

(2009) The cost of shame: International organizations and foreign aid in the punishing of human rights violators. Journal of Peace Research 46(1): 79–97.

29.

Long

J Scott

(1997) Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage.

30.

Lyne

Mona M

Nielson

Daniel L

Tierney

Michael J

(2009) Controlling coalitions: Social lending at the multilateral development banks. Review of International Organization 4(4): 407–433.

31.

McKeown

Timothy

(2009) How US decision-makers assessed their control of multilateral organizations, 1957–1982. Review of International Organizations 4(3): 269–291.

32.

Meernik

James

Krueger

Eric L

Poe

Steven C

(1998) Testing models of US foreign policy: Foreign aid during and after the Cold War. Journal of Politics 60(1): 63–85.

33.

Mertus

Julie

(2004) Bait and Switch: Human Rights and US Foreign Policy. New York: Routledge.

34.

Milner

Helen

Tingley

Dustin

(2010) The political economy of US foreign aid: American legislators and the domestic politics of aid. Economics and Politics 22(2): 200–232.

35.

Mower

A Glenn

Jr (1987) Human Rights and American Foreign Policy. Westport, CN: Greenwood.

36.

Nelson

Rebecca M

Weiss

Martin A

(2013) Multilateral development banks: How the United States makes and implements policy. CRS Report for Congress. Washington, DC: Congressional Research Service.

37.

Neumayer

Eric

(2003a) Is respect for human rights rewarded? An analysis of total bilateral and multilateral aid flows. Human Rights Quarterly 25(2): 510–527.

38.

Neumayer

Eric

(2003b) The determinants of aid allocation by regional multilateral development banks and United Nations agencies. International Studies Quarterly 47(1): 101–122.

39.

Rennack

Dianne E

(2005) China: Economic sanctions. CRS Report for Congress. Washington, DC: Congressional Research Service

40.

Rosenblum

Marc R

Salehyan

Idean

(2004) Norms and interests in US asylum enforcement. Journal of Peace Research 41(6): 677–697.

41.

Sanford

Jonathan

(1982) US Foreign Policy and Multilateral Development Banks. Boulder, CO: Westview.

42.

Sanford

Jonathan

(1997) China and the Multilateral Development Banks. CRS Report for Congress. Washington, DC: Congressional Research Service.

43.

Schoultz

Lars

(1982) Politics, economics, and US participation in multilateral development banks. International Organization 36(3): 537–574.

44.

Spiro

Elizabeth

(1977) Front door or back stairs: US human rights policy in the international financial institutions. In: Rubin

Barry M

Spiro

Elizabeth P

(eds) Human Rights and US Foreign Policy. Boulder, CO: Westview, 133–147.

45.

United States Agency for International Development (2012) US Overseas Loans and Grants (http://gbk.eads.usaidallnet.gov/).

46.

United States Census Bureau (2012a) Foreign Trade Statistics (http://www.census.gov/foreign-trade/statistics/index.html).

47.

United States Census Bureau (2012b) International Database (http://www.census.gov/population/international/data/idb/informationGateway.php).

48.

United States Department of State (2012) United States Participation in the United Nations (http://www.state.gov/p/io/rls/rpt/c47366.htm).

49.

United States Department of Treasury (2012) Loan Review Votes (http://www.treasury.gov/resource-center/international/development-banks/Pages/data.aspx).

50.

Wade

Robert

(2002) US hegemony and the World Bank: The fight over people and ideas. Review of International Political Economy 9(2): 201–229.

51.

Woods

Ngaire

(2003) The United States and the international financial institutions. In: Foot

Rosemary

Neil MacFarlane

Mastanduno

Michael

(eds) US Hegemony and International Organizations. Oxford: Oxford University Press, 92–114.

52.

Voeten

Eric

(2004) Resisting the lonely superpower. Journal of Politics 66(3): 729–754.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB