Shame on You: The Impact of Human Rights Criticism on Political Repression in Latin America

Abstract

The most commonly used weapon in the arsenal of human rights proponents is shaming the violating government through public criticism. But does this really affect the behavior of the violator? This study examines how governments that are targeted for human rights criticism respond to subsequent contentious challenges. Analyzing 873 challenges in seven Latin American countries between 1981 and 1995, it is found that human rights criticism does lead governments to reduce repression of subsequent challenges in cases where there are relatively strong economic ties to other countries. However, the duration of this impact is relatively short—less than 6 months. Examination of the source of human rights criticism shows that criticism by NGOs, religious groups, and foreign governments was more effective than criticism from inter-governmental organizations.

Since the 1970s, there has been an explosion of activity on the topic of human rights. Governments, intergovernmental organizations (IGOs), and nongovernmental organizations (NGOs) have increasingly seen human rights abuses as a cause for concern outside of the country in which they take place. These actors became part of an increasingly active human rights transnational advocacy network that has challenged traditional notions of sovereignty (Sikkink 1993; Keck and Sikkink 1998). These actors have worked to put pressure on human rights violators largely by gathering and sharing information, issuing public condemnations, and in the cases of governments, even occasionally applying material sanctions. An important question is to what extent these actors have been effective at pressuring governments to improve their human rights records. This research examines one form of influence, analyzing whether public condemnations of human rights abuses cause governments to be less repressive of subsequent political challenges, controlling for the domestic situation. The analysis examines 873 challenges that occurred between 1981 and 1995 in Argentina, Brazil, Chile, Guatemala, Mexico, Nicaragua, and Venezuela. Results show that human rights criticism does lead governments to reduce repression of subsequent challenges in cases where there are relatively strong economic ties to other countries. This effect is based on criticism that occurred in the month prior to challenges, and it weakens substantially when we consider criticism over a 6-month period. Examination of the source of human rights criticism shows that criticism by NGOs, religious groups, and governments was more effective than criticism made by inter-governmental organizations.

The Rise of Human Rights Activism

While many governments have long guaranteed rights to their citizens, in a world of sovereign states, human rights were traditionally seen as a matter of domestic jurisdiction. The prevailing doctrine was that foreign governments or other foreign organizations had no legitimate right to interfere with a state's treatment of its citizens. The idea of universal human rights gained ground with the formation of the United Nations and were explicitly declared in the Universal Declaration of Human Rights in 1948. However, during the Cold War, security concerns overwhelmed concerns of human rights among states and within the UN, and when human rights concerns were cited, they often became a tool in the ideological conflict between the superpowers (Donnelly 1998; Forsythe 1989; Schoultz 1981).

Major changes occurred in the 1970s, though, making human rights an important issue in world politics. One important change was the growth of NGOs dedicated to the issue of human rights. The most prominent group is Amnesty International, but a multitude of other groups have been formed. These groups have played an important role in gathering information, framing it in a way that can gain international attention, pressuring governments and IGOs to take action, and lobbying for the creation of additional agencies and institutions to enforce human rights standards (Cleary 1997; Keck and Sikkink 1998; Korey 1998). Intergovernmental organizations also became more actively engaged in human rights during the 1970s. The UN Commission on Human Rights (UNCHR) was given the authority to investigate human rights violations in particular countries in 1970, making the UN an important forum for international human rights concerns. Likewise, regional intergovernmental institutions, such as the Inter-American Commission on Human Rights (IACHR), connected with the Organization of American States (OAS), became increasingly active on the issue of human rights (Donnelly 1998; Keck and Sikkink 1998). Furthermore, human rights became a more prominent concern to governments starting in the 1970s. Human rights were initially put on the U.S. foreign policy agenda by members of Congress and this trend was reinforced by the election of Jimmy Carter as president. Human rights was only one of many concerns for U.S. foreign policy, but debates over foreign aid showed that it was an issue that could not simply be dismissed (Schoultz 1981; Forsythe 1989).

These developments led to the creation of what Keck and Sikkink (1998) call a transnational advocacy network dedicated to human rights. This network consists of parts of international and regional intergovernmental organizations, international NGOs (i.e. outside of the country of human rights concern), domestic NGOs (within the country where human rights abuses are taking place), private foundations, and parts of some governments.

The Impact of Transnational Human Rights Activism

The Case for Human Rights Pressures

Hawkins (2002) defines human rights pressures as “nonviolent activities carried out by transnational networks and states with the primary purpose of improving individual rights by creating economic and political costs for a repressive government” (20). Many scholars argue for the importance of transnational pressure in reducing repression by a foreign government (Brysk 1993; Hawkins 2002; Risse and Sikkink 1999; Sikkink 2004). Most prominently, Keck and Sikkink (1998) describe a boomerang pattern of influence for transnational advocacy networks. In this model, domestic human rights NGOs, facing a recalcitrant government, appeal to international NGOs, which gather information and promote the human rights concerns with other governments and with intergovernmental organizations, which then pressure the offending government. Private foundations play an important indirect role by funding NGOs. The boomerang model makes it clear that a variety of actors play an important role in applying pressure against repressive governments, but the way the model is depicted (in Figure 1 from Keck and Sikkink) suggests that NGOs work indirectly through states and intergovernmental organizations, with states applying the greatest degree of pressure.

Therefore, there are theoretical reasons to expect that transnational human rights pressure will have an impact on the target government. There are a variety of impacts, ranging from setting the agenda to influencing state behavior (Keck and Sikkink 1998). This study looks at this highest level of influence—stopping or at least reducing repression—since this is the ultimate aim of the human rights network (Brysk 1993).

Previous Research on Transnational Factors

Many studies address the connection between U.S. foreign aid and human rights records of recipient countries. However, this literature largely focuses on whether human rights records affect U.S. foreign aid outlays. A major goal of this literature was to determine whether Congressional directives discouraging giving foreign aid to human rights abusers was being followed in practice.¹

The concern of this study, though, is the opposite—whether outside actors’ pressure can cause a government to reduce subsequent repression. This issue has received less attention, though there has been a recent upsurge. One potential international influence is a set of international human rights agreements, including Article 55 of the UN Charter, and a set of international human rights treaties, such as the International Covenant on Civil and Political Rights. Almost all states in the world have become party to at least one of these agreements (Hafner-Burton 2005a). However, systematic studies show that ratification of these treaties fails to significantly improve a country's human rights record (Keith 1999; Hafner-Burton 2005a; Neumayer 2005). This is likely due to the weak enforcement mechanisms within these treaties (Donnelly 1998; Hafner-Burton 2005a). Hafner-Burton (2005a) finds, alternatively, that countries that are parties to preferential trade agreements that provide enforceable human rights standards tend to have fewer human rights abuses. However, a potential shortcoming of this study is that it does not control for a key factor in repression: threat associated with internal conflict.

Another potential influence on human rights involves states manipulating foreign aid to punish human rights violators. Regan (1995) examined whether changes in U.S. economic aid led to changes in political repression in 32 developing countries from 1977 through 1988. He found that changes in economic aid had no significant impact on changes in political repression. While this study fills a gap in the literature and does consider some domestic factors in repression, Regan, like Hafner-Burton (2005a), did not include any indicators of the level of political challenges or threats, which is perhaps the most important factor explaining political repression (Davenport 2000; Poe 2002). His study also only considered economic aid, not examining military aid which has been most connected to human rights policy, and the study did not consider whether changes in aid were connected to human rights considerations. In a similar vein, Sikkink (2004) compares U.S. military and economic aid to the incidence of political killings and/or disappearances in Argentina, Chile, Guatemala, and El Salvador. This simple visual comparison shows no clear relationship in the patterns of aid and human rights abuses.

Public Human Rights Criticism

While material sanctions made through foreign aid disbursements or preferential trade agreements are a potential source of influence on human rights violating governments, a great deal of activity by the human rights network involves public condemnations of a government's human rights abuses (Donnelly 1998; Keck and Sikkink 1998; Risse and Sikkink 1999). Indeed, Donnelly (1998), in his review of the international politics of human rights, concludes that

[i]n a world still organized around sovereign states, the international contribution to implementing human rights rests on persuasive diplomacy, which itself rests considerably on the power of shame that lies at the heart of investigatory and reporting mechanisms (85).

Sikkink (2004), likewise, argues that studying the messages and cues sent in regard to human rights (and how they are interpreted by officials) is more important than focusing mainly on foreign aid disbursements. Thus, the focus here is on public human rights criticism, which following Hawkins' (2002) concept of intangible human rights pressures, can be defined as “efforts to shame repressive governments by gathering and distributing information on the nature of repression, and to persuade repressive governments to alter their behavior by pointing out their hypocrisy” (20). Hawkins (2002) argues that governments respond to human rights pressures because of concerns over their domestic and international legitimacy. Risse and Sikkink (1999) see shaming as a tactic of persuasion that defines target states as part of an out-group separate from the community of civilized nations. Shaming seeks to convince leaders “that their behavior is inconsistent with an identity to which they aspire” (15). Such public shaming may also prod foreign governments to implement material sanctions on the offending country (Hawkins 2002; Risse and Sikkink 1999); thus, intangible human rights pressures may prompt more tangible pressures. This study, then, will examine this most available tool of actors, public human rights criticism, and whether it affects the future behavior of target governments.

Previous Qualitative Studies

There are several qualitative studies that examine the effect of criticism by the transnational human rights network on particular cases, yet I am not aware of systematic quantitative analyses. It is beyond the scope of this project to assess all the qualitative studies, but it is instructive to focus on three countries—Argentina, Chile, and Guatemala. These countries have been the subject of prominent studies and they are included in the sample of countries that I will analyze below. The strongest case for the impact of human rights pressure on actual practice concerns Argentina's military junta in power from 1976 to 1983. Martin and Sikkink (1993) and later Keck and Sikkink (1998) presented case studies of the activities of the human rights transnational advocacy network in opposition to the campaign of torture and disappearances of the military junta in Argentina during the late 1970s. Upon taking power in 1976, the military regime engaged in a systematic campaign of disappearances against suspected subversives.² They showed that disappearances dropped off dramatically after 1978, when international pressure from human rights NGOs, the UN, the IACHR, and the U.S. government was sufficient to force the Argentine government to allow the IACHR to conduct an investigation of human rights. This pressure included economic sanctions by the U.S. government, but Keck and Sikkink argue that it was the combination of criticism and sanctions from a variety of actors that was crucial. Brysk (1993) argues that the impact of human rights pressure from “above” and “below” in Argentina was mixed. On the negative side, the pressure seems to have encouraged a speeding up of secret executions. On the other hand, the pressure saved lives and may have encouraged the decline in repression after 1979, as well as protecting the human rights movement and even contributing to the military junta's miscalculated invasion of the Malvinas/Falkland Islands. Furthermore, Flood (1986) who was part of the U.S. State Department at the time, argues that U.S. foreign policy played a major role in human rights improvements in Argentina.

However, Escudé (1991) and Falcoff (1989) are more skeptical of the impact of Carter's human rights policies on repression in Argentina. Vicuña (1986) takes the more general position that internal factors are likely most important in explaining a country's human rights records. He suggests that human rights may worsen under a newly established authoritarian regime and then improve once the regime is consolidated. Indeed, Munck's (1998) detailed analysis of the military regime in Argentina asserts that key military leaders by 1977 recognized the need to move from eliminating threats to establishing a broader basis for authority, which presumably would involve reducing repression. Navarro (2001), likewise, quotes junta leader General Videla as stating in 1977 that the guerrillas no longer presented a threat to national security. Keck and Sikkink (1998) admit that an alternative explanation of the decline in human rights abuses is that the military in Argentina had already killed all the people they thought they needed to kill, but they counter this argument by asserting that there were hardliners within the military leadership who wanted to go much further with their “dirty war” against subversion.

While there is compelling (though disputed) evidence of human rights pressures reducing repression in Argentina, most authors seem to agree that human rights pressures had a mixed impact on repression in Chile and Guatemala. The military took power in Chile in a 1973 coup, seeking to eradicate Marxism. The greatest violence occurred shortly after taking power, though repression continued throughout the next 16 years of the regime (Constable and Valenzuela 1991). The transnational human rights advocacy network sprang into action in response to the repression in Chile, but with very little effect during this early period (Hawkins 2002; Muñoz 1986; Ropp and Sikkink 1999). Hawkins (2002) does consider the disbanding of the secret police organization, DINA, in 1977 to be, at least in part, a response to human rights pressures from within Chile and from the United States. Some authors also argue that human rights pressures—foreign and domestic—helped create space for the domestic political opposition and encouraged officials to allow a relatively fair plebiscite in 1988, which cleared the way for democratic transition (Fruhling 1992; Lowden 1996; Hawkins 2002).

Martin and Sikkink (1993) argue that human rights pressures in the 1970s had very little impact on human rights practices in Guatemala, which in sheer number of victims, was the worst human rights abuser in the hemisphere over the post-World War II era. They attribute the weak effect partially to external factors (less consistent pressure being wielded by the U.S.) and internal factors (lack of a military faction that was moderate on human rights, the weakness of domestic human rights groups, and the upsurge in guerrilla activity). The 1980–1983 period saw the worst phase of repression in Guatemala (Jonas 1991). Repression continued but the magnitude declined somewhat after a 1983 coup that brought a different military faction to power. This new government cleared the way for an election in 1985 that resulted in the first civilian president since 1970. Jonas (1991) sees Guatemala's isolation due to human rights abuses and the desire to increase international legitimacy as factors in these moves by the military, a point supported by Ropp and Sikkink (1999). However, Jonas cites several domestic factors in the 1983 coup, such as military rivalries, religious divisions, and economic grievances by the bourgeoisie. Furthermore, she maintains that this opening was simply a new phase in the military's counterinsurgency campaign, made possible by the tactical defeat of the rebels through brutal repression.

The debates over these cases exemplifies the difficulty of assessing the impact of human rights pressures on repression of human rights, because of the variety of potential external and internal factors. These studies cite several cases in which repression declined following human rights pressures. However, there were also domestic factors that could explain the decrease in repression, such as declining threat due to effective repression of opposition, and divisions among the regime's previous supporters. Assessing the impact of human rights criticism on repression of human rights requires a model of domestic factors as well, a point emphasized in Hawkins' (2002) study of human rights pressures in Chile.

Another potential problem facing qualitative studies on this topic is the attempt to assess changes in repression following an extreme campaign of repression, such as Argentina's Dirty War, Guatemala's massacres of the early 1980s, and disappearances and executions in Chile following the 1973 coup. These are certainly important cases that deserve detailed analysis. Yet, such egregious violations of human rights are likely to generate a great deal of activity by the human rights advocacy network, and levels of repression are almost certainly bound to decrease following these extremes. This is the statistical problem of regression toward the mean: the decrease in repression could just reflect the fact that “extreme experiences tend to be balanced by less extreme experiences” (Shaughnessy and Zechmeister 1990). We can avoid this potential problem by analyzing multiple cases that represent a wide variety of levels of repression and contexts.

Research Design

Based on the discussion above, I argue that a more definitive examination of this issue requires (1) controls for domestic factors that could also explain the patterns in repression, (2) systematic measurement of human rights criticism, (3) an examination of various levels of repression in multiple contexts, and (4) measurement of human rights criticism and repression that allows analysis of shorter time periods than years. This study offers this more definitive analysis by taking a different approach from previous studies on the topic. I will quantitatively analyze a large number of cases, but the unit of analysis is the individual contentious challenge rather than the country-year (as in most quantitative studies of repression). Contentious political challenges are defined here as collective, unconventional actions taken by inhabitants of a country to express opposition against their government, its policies or personnel, or the political regime itself. Collective acts involve at least two people. Unconventional acts take place outside of the usual institutionalized methods of conflict resolution, such as elections, the courts, contacting representatives, etc.

I recorded as much information as possible for all contentious political challenges mentioned between 1981 and 1995 for each of seven randomly chosen countries: Argentina, Brazil, Chile, Guatemala, Mexico, Nicaragua, and Venezuela. The sources for this information included full text news wire reports from dozens of wire services indexed on Lexis Nexis, and the CD-ROM versions of Keesing's Record of World Events and Facts on File news archives.³ This involved the reading of thousands of news articles in order to identify contentious political challenges and to collect information about each one. This process resulted in a sample of 1318 challenges in these seven countries that were relatively well reported.

The use of a random sample of countries was based on two considerations. First, the amount of information to code was too great to realistically allow a study of all Latin American countries, much less a global sample. Indeed, previous studies examining government responses to particular challenges only analyze one or two countries (Kowalewski 1980; 1987; Kowalewski and Schumaker 1981; O'Keefe and Schumaker 1983; Shin 1983), so the analysis of seven countries goes well beyond the norm. The second consideration was to avoid the practice in many more focused studies of repression and contention of concentrating only on highly contentious and/or highly repressive countries. It is certainly important to analyze these relatively extreme cases, but a more general understanding of repression requires a broader sample. At the outset of this project, a random sample was taken (using the Stata statistical package [StataCorp, College Station, Texas]) from the population of all countries with more than one million inhabitants. Due to the enormity of the data collection task and taking into account my own interests and expertise, I have chosen to study the seven Latin American countries that were randomly chosen in this process.⁴ The resulting sample includes countries that gained notoriety for being highly repressive during this time period, such as Chile and Guatemala as well as countries such as Mexico and Venezuela that were less repressive. The sample ranges from two of the poorest countries in the region (Guatemala and Nicaragua) to two of the most developed (Argentina and Chile). The sample includes various magnitudes and types of contentious challenges and government responses and also allows us to compare both democracies and nondemocracies. All in all, the sample should offer a fairly representative picture of repression in Latin America.

Justification of the Research Design

As mentioned above, this study examines the magnitude of repression utilized by governments in response to individual contentious challenges. Political repression is defined here as the use of coercion by political authorities of a country against inhabitants of that country. Coercion includes both violent acts meant to do physical harm to the targets and nonviolent acts that restrain a target's freedom of action or impose economic penalties.

Why not focus on whether governments make improvements in the aspect of repression that was criticized? For example, when a government is criticized for disappearances, we could examine whether there is a subsequent reduction in disappearances. I focus on repressive responses to political challenges for several reasons. First, distinguishing particular aspects of repression would require a wide variety of dependent variables which would make comparison problematic. Second, many human rights criticisms that occur are either very broad or multifaceted, so that it would be difficult to single out which aspects of repression to measure. Most importantly, though, there are theoretical reasons to expect that human rights criticism will have a more general impact on government decisions to use repression. Gartner and Regan (1996) argue that governmental leaders deciding how to respond to an opposition group will consider the possibility that repression will be deemed too harsh, inviting important domestic and international actors to impose costly sanctions on the government. Another way of stating it is that repression carries a backlash potential—the possibility that repression will lead to declining support for the government among citizens or international actors. This declining support can be manifested in future domestic challenges and in international condemnation or sanctions.⁵ How do leaders estimate this backlash potential? There are undoubtedly many ways, but one important guide is recent human rights criticism; this indicates that a government's repressive behavior is being scrutinized, increasing the likelihood that subsequent use of repression will result in further criticism with resulting loss of legitimacy and perhaps material sanctions (Hawkins 2002). Thus, I propose that any type of public human rights criticism can act as a constraint on subsequent use of repression.

Examining repression utilized in response to particular contentious challenges is also consistent with prior theory that posits threats or costs imposed by contentious challenges as a critical factor explaining repression (Gartner and Regan 1996; Gurr 1986; Poe 2002). Indeed, Davenport (2000) sums up the literature on repression by asserting that the tendency of authorities to use repression as a response to domestic threats is one of the few propositions to achieve law-like status in political science. Finally, there are a number of previous studies of government repression and accommodation that use contentious challenges as the unit of analysis (Kowalewski 1980; 1987; Kowalewski and Schumaker 1981; O'Keefe and Schumaker 1983; Shin 1983), though none of these have studied human rights criticism or Latin American countries.

Thus, there is a theoretical and empirical basis for examining repression in response to contentious challenges, and this research design has a number of advantages for analyzing the effect of human rights criticism. First, it allows for greater precision in measurement of human rights criticism and repression than would be possible in an aggregate country-year type of study. Rather than examining the effect of aggregate criticism in 1 year on the country's aggregate level of repression during the next year, it is possible to analyze human rights criticism during the weeks or months preceding a contentious challenge and then measure the magnitude of repression in response. Second, this research design allows me to carefully control for an important alternative factor—the level of threat. Thus, for each contentious challenge, the types of demands, the duration, number of participants, and tactics are measured. Third, other important controls can be applied, such as regime type and the amount of support for the government, at the time of the contentious challenge.⁶

The trade-off for analyzing repression in this way is that we are limited to studying reactive repression, as opposed to proactive repression that is not tied to any particular challenge. This is problematic in regard to the civil wars in Nicaragua and Guatemala. In these situations, governments cannot respond individually to the hundreds of challenges, and thus offer broader responses, such as large military offensives in Nicaragua against contra insurgents and a brutal scorched earth campaign of civilian massacres in Guatemala, designed to destroy active support for leftist guerrillas (see Schirmer 1998). Integrating these campaigns into this study would require questionable assumptions, such as assuming that the campaign of massacres in Guatemala was a reaction to the previous 1 or 10 or 100 guerrilla attacks. Thus, we do not have a full accounting of the level of repression connected to the guerrilla insurgencies in Nicaragua and Guatemala, so these challenges were omitted.⁷ However, for the 878 remaining challenges, acts of repression mentioned in news stories are more directly reactive in nature and, with a few reasonable rules, can be linked with particular challenges. These measures do not provide an all-encompassing indicator of repression, but it is doubtful that a single indicator or research design could fully measure the complex and varied nature of repression. This study does provide detailed information on over 300 instances of repression that resulted in well over 600 deaths and 75,000 arrests.⁸ Most of these individual instances of repression are too “small” to be the subject of detailed qualitative studies, yet this research design makes use of more detailed information on repression than is possible in large-N summary indicators of repression.

In sum, the specific issue being examined is whether the occurrence of critical statements relating to human rights caused the offending government to repress subsequent political challenges any less harshly, controlling for the type of challenge and the political context. This study examines, in a sense, critical test cases of a government's behavior in regard to repression of human rights. It is easy enough to profess a new-found dedication to human rights as long as political threats are nonexistent. However, the critical issue is whether governments will actually resist harshly repressing political challenges after being criticized for past human rights abuses.

Theory and Indicators

The primary theoretical question being examined is whether critical statements made against a government's human rights record actually reduce the offending government's subsequent repressive behavior. As discussed above, the general theory of the impact of human rights criticism was proposed by Hawkins (2002) and Risse and Sikkink (1999). In addition, I argued above that the impact of such criticism should not be limited to the aspect of human rights mentioned in the criticism but rather will affect repression more generally. This suggests the primary proposition to be tested.

Proposition 1:Governments that were recently publicly criticized for human rights violations will use a lower magnitude of repression in response to contentious political challenges.

Measuring Human Rights Criticism

The primary explanatory variable suggested by the first proposition is what I call human rights criticism, defined as public statements that cite or condemn human rights abuses in a particular country. This was measured from all public statements or reports concerning a country's human rights record that earned mention in the news archives Facts on File or Keesing's Record of World Events.⁹ In practice, this included human rights reports issued by NGOs, religious groups, intergovernmental organizations, and governments; it also included statements made by officials with these various organizations. Most of these instances of criticism emerge from actors outside the country, but keeping with Keck and Sikkink's (1998) boomerang model, domestic NGOs were also a valid source of criticism. Many of these statements or report findings were critical of a government's human rights record, but some defended or praised the government's record. A general listing of the number of instances of human rights criticism recorded in these news archives, broken down by target country and source of criticism, is shown in Table 1.

1

Instances of Human Rights Criticism in Media Archives, 1981–1995

	Source of Human Rights Criticism
Country	NGOs	Religious	IGOs	Governments	Total
Argentina	2	1	2	1 (4)	6
Brazil	8	0	0	1 (1)	9
Chile	15	9	15	11	50
Guatemala	17	2	9	12 (7)	40
Mexico	8	0	1	5 (1)	14
Nicaragua	12	0	4	10	26
Venezuela	6	1	0	0	7

Note: Numbers in parentheses indicate statements or reports that minimize human rights concerns or note an overall improvement in human rights.

As we can see, Chile was the country that was most criticized for this time period. This is consistent with the assessment of other authors. Ropp and Sikkink (1999) assert that “international network pressures ... were more comprehensive and forceful toward Chile than toward any other Latin American country” (175). Likewise, Tarrow (2005) quotes a Human Rights Watch staff member as identifying Pinochet as the “ ‘poster child’ of the human rights movement” (144). Chile also stands out in the activism of religious groups (especially the Catholic Church's Vicariate of Solidarity) and of international organizations (both OAS and UN organizations), which has also been noted by other authors (Donnelly 1998; Fruhling 1992; Lowden 1996; Ropp and Sikkink 1999). Guatemala is widely recognized as the worst violator of human rights during this period, connected especially with a systematic campaign of disappearances by death squads and wholesale massacres of villages as part of a scorched-earth anti-insurgency campaign (Archdiocese of Guatemala, Office of Human Rights 1999; Ropp and Sikkink 1999; Schirmer 1998). However, as Ropp and Sikkink (1999) note, Guatemala did not receive as much international attention. The data from news archives show that Guatemala actually did receive more criticism from NGOs than any other country in the sample, but religious groups and international organizations did not play as much of a role as in Chile. The extreme levels of repression in Guatemala did not allow space for a Vicariate of Solidarity-type organization. Furthermore, Guatemala was seen as more central to the Cold War by the Reagan Administration (Sikkink 2004), which defended Guatemala from human rights condemnation in the UN (Forsythe 1989: 116). Furthermore, the table shows that while Guatemala received a great deal of criticism of human rights abuses from various officials within the U.S. government, it led the way in comments that minimized human rights issues or emphasized improvements in human rights. These apologies for Guatemalan leaders by the Reagan Administration, many of which occurred during the most intense period of repression, are emphasized by Sikkink (2004) as the low point of U.S. human rights policy in the region.

The data shown in Table 1 for Argentina and Mexico are also consistent with the assessments of other experts in the field. One might be surprised that Argentina ranks lowest in terms of human rights criticism, but one must remember that the starting point for this study is 1981, and Keck and Sikkink (1998) noted that the worst abuses decreased after 1980. Furthermore, the Reagan administration, which took over in 1981, wanted to improve relations with anti-communist countries, and, moreover, Brysk (1993) asserted that once Argentina became democratic, international attention turned elsewhere. Finally, Keck and Sikkink (1998) argued that Mexico was off the agenda of the international human rights network until the mid to late 1980s. The data gathered from news archives show that the first prominent human rights criticism occurred in 1986, and then criticism picked up substantially in the early 1990s.

The use of media sources undoubtedly results in the omission or simplification of particular statements or report findings, but it is neither necessary nor feasible to have a census of all statements made on human rights. Instead, the goal is to assess whether violating countries truly made it onto the international human rights agenda, relative to other violators. I believe that using the media archives is appropriate for measuring this, particularly since a major tactic of the human rights network is to publicize human rights abuses, hoping to shame the offending government into improvements. Both Keck and Sikkink (1998) and Brysk (1993) emphasize the importance of the media in this task. Bob (2005) also argues that movements within repressive countries compete to gain the attention of the international media. In addition, Ron, Ramos, and Rodgers (2005) argue that Amnesty International has a strategy to use the media to achieve its goals, and they find evidence of strong reciprocal links between Amnesty International press releases and media coverage. Furthermore, the discussion above shows that the incidence of criticism measured from media archives is quite consistent with the assessments of the leading experts in the field.¹⁰

Since the unit of analysis for this study is the contentious political challenge, human rights criticism was measured for periods of six months and one month prior to each challenge. Human rights communications were initially coded separately for four sources (1) NGOs without an explicitly religious basis (domestic or international), (2) religious organizations, (3) intergovernmental organizations, and (4) governments. For each of these categories, all the human rights communications mentioned were assessed over the full 1- and 6-month periods using the scheme in Table 2. Each of these variables are analyzed below, and in addition these four variables were summed to create a general indicator of human rights criticism. For the present sample, this variable ranges from 0 to 3 for the month prior to challenges and from 0 to 4 for the 6 months prior. Of the 878 contentious political challenges studied here, 143 (16.3%) had human rights criticism that occurred in the previous month, and 358 (40.8%) followed human rights criticism that occurred over the previous 6 months.

2

Coding Scheme for Human Rights Criticism

Code	Description
0	No human rights communications or human rights communications are on balance neutral or favorable in regard to the country's human rights record
1	Human rights communications are on balance critical of the government in regard to its human rights record

Note: These codes are applied to periods of 1 and 6 months prior to each contentious political challenge in the sample.

Measuring Political Repression

The dependent variable being explained is the magnitude of political repression. The measurement of this concept began by distinguishing two types of political repression—violent repression and nonviolent repression. Violent repression refers to actions by authorities meant to do physical harm to challengers. The magnitude of violent repression was measured for each challenge, based on the number of deaths caused by violent repression. Nonviolent repression refers to arrests, detentions, dismissals from public sector jobs, and cases of forced exile utilized against suspected challengers. The coding procedures for both of these variables are described in Table 3.

3

Coding Procedures for Repression Indicators

Magnitude of Nonviolent Repression		Magnitude of Violent Repression
Code	Number of arrests, detentions, or job dismissals	Code	Description
0	None	0	No violence
1	1–19	1	Use of violence, but no one is killed
2	20–49	2	1–4 killed
3	50–99	3	5–19 killed
4	100–499	4	20–49 killed
5	500–999	5	50–99 killed
6	1000 or more	6	100 or more

The next task is to combine the two indicators described above into a single measure of the magnitude of political repression. Since taking someone's life is more severe than arresting him or her, the magnitude of violent repression is multiplied by two before the two indicators are summed. This is an unavoidably arbitrary weighting scheme, but it is less arbitrary than assuming that both components are of equal weight.¹¹ This produces a variable that varies from 0 to 18. This variable could be used for analysis, but it is problematic because while it resembles a continuous variable, its values have no precise meaning (except more or less repression). This variable is ordinal, so a method of analysis that assumes an ordinal dependent variable is preferable, and the interpretation of such an analysis is more straightforward if the dependent variable has fewer categories. Therefore, a final variable for the magnitude of political repression was created in which “0” indicates no repression, “1” indicates low repression (based on values 1–3 of the additive index), “2” indicates intermediate repression (values 4–8 of the index), and “3” indicates severe repression (values 9 through 18 on the index).¹² Political repression is measured from news archives and full-text news wire reports.

Vulnerability to Transnational Pressure

Keck and Sikkink suggested another important consideration in the effectiveness of transnational advocacy networks in achieving their goals—the vulnerability of the target government to material incentives or sanctions. Risse and Sikkink (1999) echo this argument, hypothesizing that “countries receiving large military and economic aid flows will be more vulnerable to human rights pressures than those not receiving such flows” (24). With the growth of international investment, we should also consider the role of foreign direct investment. Even though governments cannot easily pull the plug on investment as a punishment for human rights abuses, countries with greater investment may show greater concern for international opinion. Therefore, perhaps it is a combination of human rights criticism and foreign economic ties that result in changes of behavior for governments that have abused human rights.

Proposition 2:Governments that were recently publicly criticized for human rights violations and are more dependent on foreign aid and investment will use a lower magnitude of repression in response to contentious political challenges.

Foreign capital dependency combines two variables: foreign development aid as a percentage of the recipient country's gross national income, and foreign direct investment as a percentage of the recipient country's GDP. Both variables are from the World Bank and the two are combined in an additive index. In order to measure the interaction effect suggested in Proposition 2, human rights criticism variables were multiplied by the foreign capital dependency variable.

Strength of the Human Rights Network

Risse and Sikkink (1999) argued that the strength of human rights norms, institutions, and networks expanded greatly over time, with 1985 being the watershed to a period of “norms cascade.” In this new era, international human rights norms became widespread, human rights NGOs became more widespread and increasingly active, and most Western governments developed explicit human rights policies. Therefore, during this post-1985 period, they propose that countries will reform their human rights practices more quickly. We should also note that this general period also was a watershed in terms of democratization (Huntington 1991), increased globalization, and the winding down of the Cold War, which Cingranelli and Richards (1999) show, coincided with a decline in political imprisonment.

Proposition 3:Contentious challenges occurring after 1985 will receive lower magnitudes of political repression than challenges occurring before or during 1985.

Controlling for Domestic Factors in Political Repression

As mentioned above, it is critical to assess the impact of human rights criticism while also controlling for domestic factors. It is to be expected that authorities’ decisions on whether to repress contentious challenges and to what degree will depend on aspects of those challenges as well as the current political context within the country. Therefore, several additional factors will be assessed based on prior theory.

The growing literature on repression of human rights examines a wide variety of factors but Davenport (2000) and Poe (2002) emphasize threats as the central concept in explanations of repression. Gurr (1986) presents this argument most clearly, hypothesizing that the greater the threat associated with political challenges to political leaders, the greater the likelihood that the state will respond with violence. Several studies have found empirical support for this logic (Apodaca 2001; Davenport 1995; 1999; Davis and Ward 1990; Earl, Soule, and McCarthy 2003; Franklin 1997; Gartner and Regan 1996; Poe and Tate 1994; Poe, Tate, and Keith 1999; Poe, Tate, Keith, and Lanier 2000; Shin 1983). Thus, the concept of threat is the basis of this domestic model of political repression.

Four variables are included to indicate the level of threat. First, Gartner and Regan (1996) emphasize the importance of challenger demands in explaining repression. Revolutionary demands measures whether challengers seek to forcefully oust and replace governmental authorities or not, and was coded as a dummy variable. Second, Gurr (1986) argued that the level of violence is an important aspect of threat. The magnitude of challenger violence is measured with an indicator that is equivalent to the magnitude of violent political repression indicator described in Table 3, except, of course, that it measures violence inflicted by challengers rather than authorities. The third domestic variable is the level of participation in contentious challenges, which Gurr (1986) and Lichbach and Gurr (1981) posit as an important factor. Estimates of participation in media reports are often rough estimates and sometimes they use nonspecific terms like “hundreds” or “thousands.” Therefore, participation is measured using an ordinal scale that varies from 1 to 8, which is described in Table 4. Threat should also increase with the duration of particular challenges (Lichbach and Gurr 1981). Therefore, the fourth threat variable is the duration of each challenge, coded in days.

4

Coding Guidelines for Participation Variable

Code	Description
1	Less than 20 participants
2	20–99 participants, or “dozens”
3	100–499 participants
4	500–999 participants, or “hundreds”
5	1,000–4,999 participants
6	5,000–9,999 participants or “thousands”
7	10,000–99,999 participants
8	100,000 or more participants

Furthermore, two variables are included to indicate the political context. A variety of studies show that democracy tends to decrease political repression (Davenport 1995; 1999; Henderson 1991; Mitchell and McCormick 1988; Poe and Tate 1994; Poe et al. 1999; Rummel 1995). Democracy is measured here using the polity variable coded by the Polity IV project. This project measures degrees of political competition, executive recruitment, and constraints on the chief executive to create indicators of democracy and autocracy. The polity variable subtracts autocracy from democracy to create a score ranging from −10 (absolutely authoritarian) to 10 (absolutely democratic). Second, several authors have considered the strength or vulnerability of political leaders to be an important factor in their decision to use repression (Gurr 1986; Krain 2000). Executive support measures the breadth of presidents’ apparent support based, first, on how presidents came to office, coded as “1” for elected presidents or revolutionary leaders, “−1” for explicitly interim leaders, and “0” otherwise. This initial score is then adjusted based on events that indicate a narrowing of the president's support base. Such events include coup attempts, military mutinies, a party or faction that had previously supported the president moving into opposition, impeachment proceedings, or a presidential election that elects someone else as president. Each of these events subtracts one from the president's executive support score.

Finally, several studies show that leaders who relied on repression in the past are more likely to use it in the future (Apodaca 2001; Davenport 1995; Poe and Tate 1994). Prior repression is measured as the magnitude of repression for the previous challenge.¹³

Results

Once again, the primary question being examined is whether critical public statements made against a government's human rights record by various types of NGOs, intergovernmental organizations, or foreign governments actually reduce the offending government's subsequent repressive behavior. Three primary propositions were offered above, along with several controls for domestic factors. Since the dependent variable, the magnitude of political repression, is ordinal, ordered probit regression is used to assess the impact of the independent variables.¹⁴ This statistical analysis is testing which factors influenced the magnitude of repression meted out in response to 873 challenges that occurred in seven Latin American countries between 1981 and 1995.¹⁵ The first set of results shown in Tables 5 and 6 utilize human rights criticism that occurred no more than 1 month prior to each challenge. These two tables include five models, each of which includes a different source of human rights, while including the same set of control variables.

5

Impact of Human Rights Criticism (1 Month Prior) on the Magnitude of Political Repression

Variable	Model 1	Model 2
Human rights criticism [summary variable]	0.44** (0.05)
Human rights criticism by NGOs		0.49** (0.15)
Foreign capital dependency	−0.00 (0.00)	−0.00 (0.00)
Human rights × capital dependencya	−0.09** (0.02)	−0.25** (0.05)
Revolutionary demands	−1.34** (0.62)	−1.28** (0.62)
Magnitude of challenger violence	1.39** (0.17)	1.39** (0.17)
Participation	−0.04 (0.03)	−0.05* (0.03)
Duration of challenge	0.01 (0.00)	0.00 (0.00)
Executive support	−0.11** (0.04)	−0.13** (0.04)
Democracy	−0.03 (0.28)	−0.04 (0.03)
Prior repression	0.07** (0.02)	0.07** (0.02)
Post-1985	0.09 (0.24)	0.10 (0.23)
% Correctly predicted by model	70.68	70.79
% Error reduction	13.51	13.85
Number of observations	873	873

Note: Human rights criticism variables were calculated for 1 month prior to each contentious challenge. The number on the left is the unstandardized ordered probit regression coefficient, and the number in parentheses is the corresponding robust standard error.

^aThis variable refers to a combination of foreign capital dependency with the human rights variable that was included for that model.

*p < .05, one-tailed, **p < .01, one-tailed.

6

Impact of Human Rights Criticism (1 Month Prior) on the Magnitude of Political Repression (cont.)

Variable	Model 3	Model 4	Model 5
Human rights criticism by religious groups	1.18** (0.25)
Human rights criticism by intergovernmental organizations		0.68** (0.30)
Human rights criticism by governments			0.53** (0.13)
Foreign capital dependency	−0.00 (0.00)	−0.00 (0.00)	−0.00 (0.00)
Human rights × capital dependencya	−0.60** (0.14)	−0.06 (0.26)	−0.05* (0.03)
Revolutionary demands	−1.22** (0.61)	−1.38** (0.62)	−1.35** (0.65)
Magnitude of challenger violence	1.40** (0.18)	1.39** (0.17)	1.38** (0.18)
Participation	−0.05* (0.03)	−0.04 (0.03)	−0.04 (0.03)
Duration of challenge	0.00 (0.00)	0.01 (0.00)	0.01 (0.00)
Executive support	−0.12** (0.04)	−0.12** (0.04)	−0.13** (0.04)
Democracy	−0.04 (0.03)	−0.03 (0.03)	−0.04 (0.04)
Prior repression	0.07** (0.02)	0.07** (0.02)	0.07** (0.02)
Post-1985	−0.08 (0.22)	−0.08 (0.22)	−0.05 (0.23)
% Correctly predicted by model	71.02	70.79	70.56
% Error reduction	14.53	13.85	13.18
Number of observations	873	873	873

Note: Human rights criticism variables were calculated for one month prior to each contentious challenge. The number on the left is the unstandardized ordered probit regression coefficient, and the number in parentheses is the corresponding robust standard error.

^aInteraction between foreign capital dependency and the human rights variable that was included for that model.

*p < .05, one-tailed, **p < .01, one-tailed.

One prominent and consistent result shown in these tables is that all sources of human rights criticism are positively and statistically significantly related to the subsequent repression against contentious political challenges. This is the opposite impact from what was hypothesized in Proposition 1. Does this mean that criticizing governments’ human rights practices causes them to respond more harshly to protest? Several political leaders have responded to human rights criticism verbally with angry defiance, but I doubt that authorities deliberately and systematically increase their repressive responses because of criticism. The more plausible explanation is simply that human rights criticism tends to target the worst abusers of human rights, which are likely to be repressive in the future. The model includes a measure of prior repression based on how authorities responded to the previous contentious challenge, but this is not a complete indicator of past repressiveness. Indeed, I would argue that this result shows that media coverage of human rights criticism does, in a general way, measure the actual amount of repression (at least for Latin America during this period).

We certainly should not expect to see human rights criticism work against all governments. Some governments will simply ignore the condemnation and focus on quickly eliminating domestic threats, perhaps hoping to resuscitate their international image once this has been accomplished. Some authors have proposed the logical addendum that human rights criticism will be most effective against governments that are more dependent on their foreign relationships. Thus, Proposition 2 posits that the combination of recent human rights criticism with dependence on foreign aid and investment is crucial in causing governments to lessen repression. The results for Model 1 in Table 5 shows that this is indeed the case; the combination of human rights criticism (from all sources) with foreign capital dependency has a negative, statistically significant relationship with repression, signifying that human rights criticism significantly reduces repression against subsequent contentious challenges in countries that have greater reliance on foreign aid and investment.

Mexico in the 1994–1995 period illustrates this general finding. During this period there were many contentious challenges, including demonstrations, occupations, and blockades, that were repressed either relatively lightly, or in many cases, not at all. This was also a period in which Mexican officials were being criticized for abuses against participants in the Zapatista uprising that emerged at the beginning of 1994. Mexican officials were also under foreign scrutiny concerning the fairness of elections as opposition parties were mounting major electoral challenges across the country. This was a period that saw increasing economic ties between Mexico and other countries, particularly the NAFTA agreement with the United States and Canada, which went into effect January 1, 1994. World Bank data show that Mexico's foreign direct investment increased 140 percent from 1993 to 1994. In addition, a financial crisis that struck Mexico in December 1994 resulted in an emergency $20 billion loan from the United States (Preston and Dillon 2004). During NAFTA negotiations, President Clinton and Congressional Democrats had warned President Salinas about the importance of human rights improvements. Therefore, the combination of human rights criticism with foreign economic ties created pressures that appeared to encourage greater tolerance of dissent. This result also supports my proposal that the effect of human rights criticism is not limited to the exact aspect of human rights that is being criticized. Governments that were recently criticized for various aspects of their human rights record (and which depend more on foreign aid and investment) appear to be more reluctant to use repression in general.¹⁶

Human rights criticism comes from many sources, so an important question is whether any of the actors in the international human rights network are more effective than the others. One might expect that governments will carry the most clout with their criticism, since they have greater potential to carry out material sanctions than the other actors. Indeed, examining the results in models 2 through 5 in Tables 5 and 6 for the interaction of human rights criticism from the different actors with foreign capital dependence, we see that human rights criticism from governments significantly reduces repression against contentious challenges in countries that have greater reliance on foreign aid and investment. However, the results for NGOs and religious groups are even more highly significant. In contrast, IGOs have only a weak impact on repression.¹⁷

Another way to assess the results is to consider the impact the human rights criticism*foreign capital dependency variable has on the probability of subsequent repression, controlling for the other variables in the model.¹⁸ Countries that combine human rights criticism from NGOs with the highest observed levels of capital dependence are 33 percent less likely to use repression than countries with no criticism or no dependence, controlling for all the other factors in the model. The corresponding reduction in the probability of repression is 33 percent for criticism from religious groups, 10 percent for IGOs, and 31 percent for governments. As a point of comparison, countries that combine the highest levels of human rights criticism from all sources with the highest observed levels of capital dependence are 34 percent less likely to use repression than countries with no criticism or no dependence.

Thus, human rights criticism by NGOs and religious groups has the greatest impact on governments’ repressive behavior. This impact is only slightly stronger than the impact of governmental human rights criticism, but given the greater power of governments to apply tangible sanctions, this is somewhat surprising. A factor that likely weakened the impact of governmental criticism in this study was conflict among U.S. officials over the human rights situation in Latin America and over what the U.S. should do about it. This was particularly the case during the early 1980s when the Reagan Administration elevated Cold War concerns while congressional Democrats were holding firm on human rights concerns. For example, in February of 1982 the human rights report issued by the U.S. State Department cited growing political killings by groups tied to the Guatemalan government. Two months later, the U.S. Ambassador to Guatemala praised human rights improvements under General Rios Montt and called for a resumption of military aid, while the month after that saw the visit by a U.S. congressional fact-finding team that condemned continued killings of civilians by the Guatemalan military. The following months saw continued contradictory assessments of human rights in Guatemala and of whether the U.S. should renew military aid. This situation was not peculiar to Guatemala, as shown in Sikkink (2004) analysis of U.S. human rights policy in Latin America appropriately titled Mixed Signals. Therefore, the impact of U.S. criticism of human rights was likely reduced because U.S. policy was so muddled and contradictory.

Why does criticism from IGOs have such a weak impact? International power relations have limited the scope of IGOs, especially the UNCHR (Donnelly 1998) and likely make them slower to act on human rights. Thus, some countries are able to avoid human rights scrutiny in the UN either through their own power or through that of allies. Furthermore, IGOs do not have the capability to impose material sanctions that individual states do.

A full explanation for the greater efficacy of human rights criticism by NGOs and religious groups requires further analysis, but we can examine some possible factors. One advantage that NGOs and religious groups have in common when issuing human rights criticism is greater dedication to the human rights agenda with less internal dissension, compared to governments and IGOs. Catholic officials were not always active on the issue of human rights in Latin America, but where they were (particularly in Chile) they brought great zeal to the issue and such Catholic groups have special legitimacy in Latin America. Therefore, lay NGOs and religious groups offer the clearest cues in regard to human rights.

Another possible factor at play is that perhaps there are two groups of political leaders—those sensitive about their international image and those who have little concern (Hawkins 2002). Keck and Sikkink's (1998) boomerang model suggests that NGOs will first condemn human rights abuses and then work to get governments and IGOs to join them. Thus, the sensitive leaders may quickly respond to being put on the human rights agenda by NGOs by reducing repression, whereas more thick-skinned leaders continue to use repression even as states and IGOs join in the condemnation.

One final international factor, presented in Proposition 3, concerns the international environment. Risse and Sikkink (1999) proposed that human rights entered a new era of respect and institutionalization in 1985, making human rights concerns more prominent and human rights criticism more effective. This is tested with a post-1985 dummy variable, which is only very weakly related to repression. Thus, this proposition is not supported, but to some extent this factor is measured by foreign capital dependency, which increased after 1985 for most of the countries studied here as they opened their economies.

Analyzing the domestic control factors, we can see that four factors are consistently important in explaining repression. The most important factor by far is the level of violence utilized by challengers. Challenges using the highest level of violence were almost 100 percent more likely to receive severe repression than nonviolent challenges. Challenges with revolutionary goals received significantly lower magnitudes of repression controlling for other factors. This seems odd at first glance, but it surely represents the stealthy hit-and-run guerrilla attacks and bombings frequently used by revolutionary groups. Authorities cannot carry out repressive responses against challengers when they cannot find them. A third consistent factor is the level of executive support. Presidents with broader popular support (measured by popular election) utilize significantly lower magnitudes of repression. However, presidents who were never popularly elected or who have lost popular support are more repressive. Finally, challenges are repressed more harshly when authorities repressed the previous challenge.

One somewhat surprising result is that democracy does not significantly lessen repression, controlling for other factors. While many studies of large numbers of countries have found that democracies are less repressive (Davenport 1995; Henderson 1991; Poe and Tate 1994), other authors note that democratization does not bring about an immediate improvement in human rights behavior (Donnelly 1998; Fein 1995; Sikkink 2004). Indeed, most of the countries studied here were in a process of democratization during the period being studied, and establishing electoral democracy is not equivalent to establishing a rights-protective regime (Donnelly 1998). This is also consistent with Davenport and Armstrong (2004), who find that democracy, at low to moderate levels, has no impact on repression while high levels of democracy constrains repression. Bueno de Mesquita, Downs, Smith, and Marie Cherif (2005) support this threshold effect, arguing that the threshold for improving human rights is high levels of executive constraint and competition. The result for executive support does indicate a benefit of democracy, since elected leaders start out with higher values of executive support and are hence less likely to use repression. However, democratically elected leaders in desperate situations due to declining public support can lash out quite repressively against challenges.¹⁹

A final consideration is how long the effect of human rights criticism lasts. The results reported in Tables 5 and 6 measure human rights criticism for the month prior to each contentious political challenge in the sample. Table 7 shows the results for human rights criticism measured for 6 months prior to each contentious political challenge. Except for this change, the models shown in this table included all the same variables as the models analyzed above. However, since the purpose is to compare the effect of the 6-month criticism variables to the 1-month criticism variables, only the results concerning these variables are shown. The results for the control variables are quite similar to those shown in Tables 5 and 6.

7

Impact of Human Rights Criticism (6 Months Prior) on the Magnitude of Political Repression, Abbreviated Table of Results

Variable	Model 1	Model 2	Model 3	Model 4	Model 5
Human rights criticism [summary variable]	0.30** (0.12)
Human rights criticism by NGOs		0.33** (0.14)
Human rights criticism by religious groups			0.40** (0.24)
Human rights criticism by intergovernmental organizations				0.70** (0.26)
Human rights criticism by governments					0.46* (0.25)
Foreign capital dependency	−0.00 (0.00)	−0.00 (0.00)	−0.00 (0.00)	−0.00 (0.00)	−0.00 (0.00)
Human rights × capital dependencya	−0.01 (0.01)	−0.01 (0.01)	−0.06 (0.10)	−0.06 (0.12)	−0.03 (0.02)
Number of observations	873	873	873	873	873

Note: Human rights criticism variables here were calculated for the 6-month period preceding each contentious challenge. This is an abbreviated table intended for comparison with Tables 5 and 6. The coefficients were calculated using the same controls shown in Tables 5 and 6, but in the interest of parsimony, only the results for the human rights variables and foreign capital dependency are shown. The results for the other variables not shown here are quite similar to the results shown in Tables 5 and 6. The number on the left is the unstandardized ordered probit regression coefficient, and the number in parentheses is the corresponding robust standard error.

^aInteraction between foreign capital dependency and the human rights variable that was included for that model.

*p < .05, one-tailed, **p < .01, one-tailed.

The results here are clear. When we analyze human rights criticism for a full 6 months prior to challenges, the effect of the human rights criticism × capital dependency variables weaken substantially, compared to criticism measured only 1 month prior to challenges. Therefore, human rights criticism appears to be effective in countries that are more closely tied to the outside world, but only for a limited time—a period less than 6 months. Leaders may be wary of using repression immediately after being criticized for human rights abuses for fear of a domestic or international backlash, but as time passes, they may decide that the backlash potential is not as great.

Conclusions

Gone are the days when the world would routinely ignore human rights violations. A transnational advocacy network has developed that is active in researching, publicizing and working to end human rights abuses around the world. A natural question is whether foreign actors actually influence repressive practices by governments. This is also a methodologically thorny question because a decline in repression could be due to domestic factors, such as a change in regime or a decline in the amount of domestic threat facing governments. Therefore, this study attempts to ascertain the impact of foreign human rights criticism by examining the level of repression governments used in response to 873 contentious challenges that occurred in Argentina, Brazil, Chile, Guatemala, Mexico, Nicaragua, and Venezuela between 1981 and 1995. By examining governmental responses to contentious challenges, we can test whether human rights criticism causes a reduction of repression even when potential political threats are present. This analysis considers whether these governments had been criticized for human rights abuses prior to the challenges and their amount of dependency on foreign aid and investment. These effects are assessed in the context of a host of domestic factors.

The overall amount of human rights criticism was found to have a strong positive relationship with the magnitude of political repression. This most likely reflects the fact that human rights criticism tends to target governments that have used repression in the past, which are then more likely to use repression in the future. However, human rights criticism tends to significantly decrease subsequent repression for countries that receive greater amounts of foreign aid and foreign investment. Thus, countries with greater ties to the outside world, and hence more to lose, tend to be more sensitive to their international human rights reputation. This finding holds up in the face of multiple controls for domestic factors and despite the fact that human rights criticism does not necessarily target the reactive repression studied here. This result suggests that human rights criticism will become more effective as the current trend in economic globalization continues. Indeed, a number of recent studies have found that certain aspects of economic globalization are associated with improvements in human rights behavior (Apodaca 2001; Hafner-Burton 2005b; Richards, Gelleny, and Sacko 2001). This finding regarding the interaction of human rights criticism and foreign capital flows suggests an explanation for the globalization-human rights findings, and suggests that we examine globalization more broadly.

Furthermore, human rights criticism was broken down by the type of actor issuing the criticism and interesting and somewhat surprising results emerged. Human rights criticism emanating from NGOs (both domestic and international) and religious groups was slightly more effective at reducing subsequent repression than criticism emerging from foreign governments, while criticism from inter-governmental organizations was ineffective. This seems counter-intuitive since governments have the greatest ability to sanction human rights violators. A full explanation requires further analysis, but two possibilities were mentioned above. First, the U.S. government's position on human rights in Latin America has often been muddled and contradictory. NGOs and religious groups, in contrast, offered much clearer assessments on human rights. Secondly, Keck and Sikkink's (1998) boomerang model posits that information and criticism emanates from NGOs and then spreads to foreign states and intergovernmental organizations. Perhaps, then, offending governments sensitive about their international image will reduce repression after first being placed on the agenda by NGOs, but leaders not concerned about their image will carry on with abuses even as governments and IGOs join in the condemnation.

Again, more research is needed into this issue, but it does suggest one revision to Keck and Sikkink's prominent model. The boomerang model suggests that NGOs have an indirect impact on human rights violators, mediated through foreign states and IGOs; however, these results suggest that the effect of human rights NGOs (and religious officials or organizations) can actually be more direct, bypassing both states and IGOs.

For the above findings, human rights criticism is measured for the month prior to each contentious challenge. An additional analysis examines human rights criticism measured for the full 6 months prior to challenges. In this case, the effect of human rights criticism is much weaker. Therefore, human rights criticism does appear to be effective, but only in countries with greater ties to the outside world and only for a limited amount of time (i.e. less than 6 months). This suggests the importance of keeping human rights violators on the international agenda.

Finally, we must remember that this study is based only on Latin American countries. While it suggests broader relationships, these require further testing in other regions. Certainly, Latin America's pattern of authoritarianism followed by political and economic opening in the 1980s and beyond is not unique. Furthermore, the human rights advocacy network is global. However, human rights principles have a long history in Latin America (Ropp and Sikkink 1999) and strong human rights movements have developed in many Latin American countries. Thus, it is possible that the influence of human rights criticism in Latin America is greater than in other developing regions. This underlines the need for further empirical testing.

Footnotes

1

See Poe (1990) for a thoughtful review of this literature.

2

This continued and intensified repression that was being carried out by paramilitary groups prior to the coup (Roniger and Sznajder 1999).

3

Brockett (2005) advocates using domestic news sources on contentious events, but such sources are more vulnerable to government intimidation and repression. Indeed, one of Brockett's main newspaper sources for Guatemala declined and then ceased publication altogether in the early 1980s due to repression. The international news wires used offer regular reports from each of the countries studied and are a more reliable source during periods of repression and authoritarian rule. Davenport and Ball (2002) argue that human rights organizations offer more comprehensive coverage of repression in rural areas than news sources. However, news media provide more comprehensive coverage of contentious challenges.

4

While this is a subsample it is still random in that each Latin American country had an equal chance of being selected.

5

In similar veins, Francisco (2005) examines the backlash against harsh repression and Martin (2006) discusses the possible backfire from repression.

6

Shellman (2004) shows that aggregation at different temporal levels can have substantial impacts on studies of event data, and he recommends sequential analysis such as that used by Moore (1998, 2000). My method is also sequential, but while Moore's approach examined “turns” of actions by dissidents versus turns of actions by the government, my approach makes a more concerted effort through the data collection process to determine whether repressive events not only occur after a particular challenge, but also appear to be an actual response to that challenge.

7

The analysis was also conducted with the full sample. The basic result on the impact of human rights criticism is the same with the larger sample, but there are a few more specific differences, as described below.

8

The numbers of deaths and arrests are quite conservative estimates, as they only represent cases in which specific estimates were available.

9

These sources employ teams of editors who produce summaries of important world events from information in the world press, regional publications, wire services, and local broadcasts. Muñoz (1986) previously used Keesing's Record of World Events to measure UN condemnations of Chile's human rights record. Stack (1987) previously used Facts on File as a measure of media attention to executions.

10

The use of media sources ignores “quiet diplomacy,” which was claimed to be a major aspect of human rights policy under the Nixon and Reagan administrations. However, most scholars of U.S. human rights policy are skeptical of this approach (Schoultz 1981; Forsythe 1989; Donnelly 1998; Sikkink 2004), in the absence of more public and active statements and actions.

11

Several previous measures of repression distinguish between violent and nonviolent forms of repression and implicitly or explicitly weight violent acts of repression more heavily (see Davenport 2004; Krain 2000; Moore 1998; Poe and Tate 1994; Shin 1983). There is no consensus, though, on the exact weights to use. Krain (2000) and Moore (1998) use weights based on surveys of social scientists, but the categories they use do not translate directly to the context of this study.

12

The ranges of index values for these categories were chosen such that no cases of “low repression” involved deadly violence, and all cases of “high repression” involved deadly violence. Thus, the value of 3 (the highest index value for the “low repression” category) is the highest value on the index in which no deaths resulted from repression. Likewise, 9 (the lowest index value in the “high repression” category) is the lowest index value in which repression by definition resulted in death. “Intermediate repression” is merely the range between these two more absolute boundaries.

13

A number of additional variables, such as GDP per capita, inflation, international conflict, military expenditures, and the number of recent challenges were also included in earlier analyses but found to be largely unrelated to repression.

14

Standard errors are adjusted for country-level “clustering.” The challenges within a country may not be independent observations, since challenges may respond to previous challenges as well as to unobserved factors at play within that country. The clustering adjustment corrects for this and provides robust standard errors.

15

Adding the lagged repression variable results in the loss of five challenges in which there is no information on prior repression, which explains why the analysis uses 873 cases.

16

This significant negative impact of the combination of human rights criticism and foreign capital dependency on repression also holds for the full sample of 1,311 challenges (including the civil war challenges in Guatemala and Nicaragua that were omitted from this analysis). Furthermore, in another study (Franklin 2007), I found that human rights criticism, combined with foreign capital dependency, also increases the likelihood of government concessions in response to contentious challenges.

17

Testing these results against the full sample (including civil war challenges) the interaction of foreign capital dependency and human rights criticism by NGOs has a less significant impact on repression, falling short of the 5 percent level of significance by a tenth of a percent. The result for the government criticism interaction is slightly more significant than in the smaller sample results reported in Table 6. The results for religious groups and IGOs are quite similar.

18

These percentages were calculated using the prchange command introduced by Long and Freese (2006).

19

This also brings up the possibility that inter-correlation between democracy and executive support explains the weaker than expected impact of democracy on repression. However, omitting the executive support variable from the ordered probit regression does not appreciably affect the result for the democracy variable.

Author notes

Author's note: An earlier version of this article was presented at the 2003 annual meeting of the American Political Science Association. I am grateful to Whitney Franklin, Sidney Tarrow, the editors at ISQ, and three anonymous reviewers for their helpful comments on earlier drafts of this article. Replication materials may be found on the International Studies Quarterly web site at http://www.isanet.org/data_archive.html and at http://dvn.iq.harvard.edu/dvn/dv/isq.