Oppressive Governments, US Closeness, and Anti-US Terrorism

econstor A Service of zbw Make Your Publications Visible. Leibniz-Informationszentrum Wirtschaft Leibniz Information Centre for Economics Redlin, Margarete; Gries, Thomas; Meierrieks, Daniel Conference Paper Oppressive Governments, US Closeness, and AntiUS Terrorism Beiträge zur Jahrestagung des Vereins für Socialpolitik 2014: Evidenzbasierte Wirtschaftspolitik - Session: Causes of Islamistic and Anti-American Terrorism, No. G08-V1 Provided in Cooperation with: Verein für Socialpolitik / German Economic Association Suggested Citation: Redlin, Margarete; Gries, Thomas; Meierrieks, Daniel (2014) : Oppressive Governments, US Closeness, and Anti-US Terrorism, Beiträge zur Jahrestagung des Vereins für Socialpolitik 2014: Evidenzbasierte Wirtschaftspolitik - Session: Causes of Islamistic and Anti-American Terrorism, No. G08-V1, ZBW - Deutsche Zentralbibliothek für Wirtschaftswissenschaften, Leibniz-Informationszentrum Wirtschaft, Kiel und Hamburg This Version is available at: http://hdl.handle.net/10419/100588 Standard-Nutzungsbedingungen: Terms of use: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Documents in EconStor may be saved and copied for your personal and scholarly purposes. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. www.econstor.eu If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. Oppressive Governments, US Closeness and Anti-US Terrorism (Draft Version, do not cite: 2014-02-28) Thomas Gries1, Daniel Meierrieks2 and Margarete Redlin3 Abstract Since the 9/11 attacks more attention has been given to the question why the United States is a major target for transnational terrorism. What conditions motivate these terrorist activities? Are there specific characteristics in the terrorists’ home countries that provide a breeding ground for anti-US terrorism? In particular, we ask whether oppressive and bad governments in these countries and/or close connections with the US encourage attacks against the US. Oppressive and bad government behavior, such as human rights violations or poor governance, may provoke resistance against the authorities, including violent attacks. Attacking the United States instead of one’s own government may be a promising option, especially if the government’s capacity seems dependent on US support. In a count data approach we use panel data for 149 countries from 1981 to 2005. We measure governmental “oppressiveness” using the physical integrity rights index, and measure a government’s closeness to the US with a range of measures. Controlling for a variety of variables, our findings indicate that both oppressiveness and US closeness are important determinants of anti-US terrorism. Furthermore, both effects do not seem completely independent. Sorting into groups, US closeness seems to relate to more attacks the greater the oppressiveness of one’s home government’s. Keywords: anti-US terrorism, international aid, human rights JEL Classification: F35, F51, F52 1 University of Paderborn, Department of Economics, Center for International Economics. Germany. Ph.: +49-(0)5251-60-3823, fax: +49-(0)5251-60-3540, e-mail: [email protected]. 2 University of Freiburg, Wilfried Guth Endowed Chair for Constitutional Political Economy and Competition Policy, Germany, Ph.: +49-761 203-67651, fax: +49-761 203-67653, e-mail: [email protected] 3 University of Paderborn, Department of Economics, Center for International Economics. Germany. Ph.: +49-(0)5251-60-3823, fax: +49-(0)5251-60-3540, e-mail: [email protected]. 1 1 Introduction The United States are a major target of terrorist groups and the attacks of 9/11 led to worldwide attention on transnational anti-U.S-terrorism. Even if there is no simple explanation for this phenomenon, single aspects might be possible to study. Anti-U.S. terrorism is not equally distributed with respect to terrorists’ country origin. Attacks are prevalently done by persons from certain countries? Hence, an interesting question is, if specific country characteristics provide a breeding ground for terrorism directed towards the United States. Furthermore, do certain economic and political conditions in these countries abet the emergence of anti-U.S. terrorism? Specifically, oppressive state behavior and violations of basic human rights may lead to political frustration and in turn may involve violent attacks and terrorism against the ruling authority. In fact, the US government seems to be aware of this reasoning, as President Obama argued “when the United States stands up for human rights, by example at home and by effort abroad … we also strengthen our security and well being, because the abuse of human rights can feed many of the global dangers that we confront -- from armed conflict and humanitarian crises, to corruption and the spread of ideologies that promote hatred and violence.” (Barak Obama 10th December, 2008). Moreover, when governments enjoy international support the international supporter is often regarded as ally. Hence, if the local authority were a repressive violent regime and actively supported by the US, attacking not the own government but the United States might be an auspicious option, in particular when the disliked home regime is stabilized by the US. Following the assumption that the supporters of the hostile home government are considered as enemies, US support may turn hate against local rulers into anti-US terrorism. In this paper we empirically examine the effects of state repression and US closeness on anti-US terror in a panel of 149 2 countries for the period from 1981 to 2005. The results of the negative binomial regressions indicate that abuses of the subset of human rights known as physical integrity rights as well as dependence from the United States measured by economic and military aid are crucial factors driving the extent of anti-US terrorism. In addition, we reveal that a combination of both conditions increases the probability of anti-US terrorism over proportionately. The next section reviews the relevant literature and theoretically motivates the formulation of the hypotheses to test, section three illustrates the empirical design including the data and the methodology and presents the results and section four concludes. 2 Theoretical Foundations and Related Literature Rational theories assume terrorists and terrorist groups to follow an economic calculus (see, e.g., Sandler and Enders, 2004; Caplan, 2006; Anderton and Carter, 2005). Individual terrorist behavior is regarded as a strategic weighing up of costs and benefits of achieving idealistic and political goals (e.g. regime change, territorial change, political control). Terrorist attacks are seen as a radical instrument to follow these goals by rebelling against the antagonistic home government. Radical behavior is often seen as the only possibility to push political and ideological values since first, the home country oftentimes does not allow peaceful political participation – for example in autocratic countries – and second, the differences in power between the home government and the radical group are too pronounced to permit a change in the unaccepted status quo. Accordingly, for groups who are dissatisfied with the current economic situation, and not in a position to bring drastic political and institutional changes, it can be rational to engage in terrorism (Blomberg et al. 2004). Considering the crucial determinants that trigger the participation in violent conflict, recent literature has focused mainly on economic and political factors. In this vein Blomberg et al. 3 (2004), Freytag et al. (2011) and Gries and Meierrieks, (2013) suggest that poor economic conditions promote terrorism and a wide body of literature focused on the nexus between political institutions and terrorism analyzing if terrorism is more likely in non-democratic (e.g., Eyerman, 1998; Li, 2005, Piazza, 2008a) and politically unstable countries (e.g., KurrildKlitgaard et al., 2006; Hegre et al., 2001). However, coming from the assumption that terrorist behavior is a reaction on the discontent relation with the government, we should give more attention to government behavior and government effectiveness as potential drivers of terrorist activity. Only few studies investigate the possible effect of governance and the government manner in this context. Presumably, a malfunctioning political system and oppressive government activity in terms of e.g. political violence and human rights violations may have a strongly provoking effect on the emergence of oppositions against the government. In this context Walsh and Piazza (2010) show that terrorist activity in a country is positively related to local human rights violations. They suggest that states that violate the overall physical integrity rights of their citizens are actually more frequently targeted by terrorists than those characterized by a higher respect for such rights. This is also supported by Krueger and Maleckova (2003) and Krueger and Laitin (2003) who find that repressive states are typical source countries of terror. In search of an explanation for why the unaccepted status quo in the home country is also projected on transnational targets – in particular the U.S. -, we follow a recent developed rational theory of international terrorism, which argues that from a strategic perspective it makes sense for terrorists to attack transnational goals, even if they ultimately intend to induce political change in their home (Neumayer and Plümper, 2011). Often, repressive state behavior of the home government is buttressed by foreign powers. In this case, the terrorists can have a strategic interest in attacking nationals of these foreign countries. For instance, on the question why the Taliban are fighting America in his “Letter to America”, Bin Laden gives the answer that “under 4 your supervision, consent and orders, the governments of our countries which act as your agents, attack us on daily basis” (Bin Laden, 2002). So, it becomes clear, that though the goals are of domestic nature, the discontent at home is not only directed against the home government but also against alliances and supporters. Further, the higher the economic and political dependence on the US, the more the dissatisfaction within the country is projected on the United States. In summary, assuming that human rights violations and repressive governments might induce more terrorism not only directed at the government but also politically close transnational targets, we arrive at our first hypothesis to test: Hypothesis 1a: Anti-U.S. terrorists are more likely to come from countries with highly repressive governments and violations of human rights. Further, recent research provides insights that the extent of transnational terrorism is affected by the quality of the relation of the home government and the target government. The domestic political discontent is the stronger projected on America, the more the disliked home political system is supported by the American government. According to the motto “the friend of my enemy is my enemy”, Plümper and Neumayer (2010) argue that for terrorist groups which seek to overthrow their home country’s political system it might be attractive to target foreign allies which stabilize and support the home country. For the case of the US, Neumayer and Plümper (2011) find that more anti-American terrorism emanates from countries that have a higher military closeness measured by US military aid, arms transfers and stationed American military personnel. In this context, aid flows should not be seen as unconditional gifts, but as political strategies. For example Dreher et al. (2008) show that US aid flows go in line with voting compliance of recipients in the UN General Assembly and Alessina and Dollar (2000) as well as 5 Sullivan et al. (2011) find evidence that foreign aid is determined by political and strategic considerations and affects the foreign policy behavior. The aid flows, which are seen as support of the local government, in turn might lead to increased transnational terrorism directed towards the sending country. Considering the relationship between aid and terrorism, Boutton and Carter (2013) demonstrate that anti US terrorism from a country is strongly correlated with US aid flows entering that country and Bapat (2011) and Boutton (2013) show that US aid as counterterrorism strategy is counterproductive and increases the extent of transnational terrorism against the US. This leads us to our next hypothesis to test: Hypothesis 1b: Anti-U.S. terrorists are more likely to come from countries with a high military (military aid) and economic (economic aid) dependence on the US. Finally, based on the preposition of hypotheses 1a and 1b, we assume that the combination of both terrorism driving factors, namely state repression and US support, might generate an overproportional effect on terrorism. The terrorism generating effects of aid flows might be higher in countries with higher human rights violations, and the other way around, repression might boost anti-US terrorism over proportionally in countries with high aid flows from the US. This leads to the final hypothesis to test: Hypothesis 2: The effects of government repression and military and economic dependence on the US interact, so that countries with both, high repression and high US dependence are even more likely to be the origin of anti-US terrorism. 6 3 Empirical Design In the analytical part of the paper an empirical design is established with the purpose of testing the hypotheses developed in Chapter 2. The hypotheses are subjected to a formation of empirical tests in the form of negative binomial statistical regression models using annual data from 149 countries over the period 1981-2005. In this section we describe the data – including the dependent variable, the main explanatory constructs and the set of control variables – and econometric methods. The summary statistics of all variables are reported in Table 1. Table1: Summary statistics Variable Obs. Mean S.D. Min Max US victims 2639 0.0966 0.9178 0 26 Us attacks 2639 0.3899 2.2489 0 88 Killings in anti US attacks 2639 0.4252 6.2839 0 263 Attacks with US victims 2639 0.3384 2.1772 0 90 Human rights violations 2560 3.1746 2.3098 0 8 Economic aid 2639 0.0029 0.0122 0 0.2875 Military aid 2639 0.0003 0.0016 0 0.0305 GDP p.c. (ln) 2639 8.6324 1.1628 5.0315 11.1971 GDP p.c. growth 2630 1.5897 6.9265 -64.3602 88.7483 Population (ln) 2639 9.3245 1.4510 5.7821 14.0779 Distance to the US (ln) 2639 8.3357 1.1698 0 9.0993 Regime stability 2639 27.5316 26.1525 1 136 State failure 2639 0.61709 1.6298 0 13.5 Trade openness 2639 72.5688 45.6322 1.0860 446.064 Government size 2639 17.6295 8.3704 3.05556 62.4492 7 3.1 Data Dependent Variable: Anti-US terrorism Since the intention is to test for possible sources of anti-American terrorism, we use bilateral transnational terrorism data to compile anti-US terrorism – measured by US victims, US attacks, killings in anti US attacks and attacks with US victims based on the International Terrorism: Attributes of Terrorist Events (ITERATE) database.4 It contains data on the characteristics of transnational terrorist groups, their activities with international impact, and the environment in which they operate (Mickolus et al. 2007). We define transnational anti-US terrorism by incidents carried out by basically autonomous non-state non-US actors on US citizens (irrespective of the location of the incident), so that the country classification of the terrorism source is defined by the nationality of the terrorist actor. The variable US victims is defined by the sum of US citizens killed in terrorist attacks, the variable US attacks is defined by the number of attacks that primarily victimized US citizens, the variable killings in anti US attacks quantifies the number of killed people in these attacks, and the variable attacks with US victims is the number of all attacks that victimized US citizens. Independent Variables Our key independent constructs to test the hypotheses developed in Chapter 2 involve the repression of a terrorist’s home country government and the closeness of this government to the US government. Further we assume a combined effect of both constructs captured by interaction effects of both variables. 4 Terrorism is characterized as transnational when a terrorist incident in one country involves victims, targets, institutions, governments, or citizens of another country. 8 Repression and Human Rights Violations We measure the repressive behavior of a countries’ government using human rights violations based on the physical integrity rights index from the Cingranelli and Richards (2010) (CIRI) Human Rights data set. It is an additive measure consisting of four individual components: (1) Torture – defined as the inhuman and degrading inflicting of extreme pain by purpose by government officials or by private individuals at the instigation of government officials. (2) Extrajudicial killings – defined as killings by government officials without due process of law. (3) Disappearances - characterized as cases in which people have disappeared, the victims have not been found and political motivation appears likely. (4) Political imprisonment – defined as the incarceration of people by government officials because of: their speech; their non-violent opposition to government policies or leaders; their religious attitudes; their non-violent religious practices including proselytizing; or their associations in certain groups (ethnic, racial etc.). Each of these sub-indices is built as ordinal scale with values ranging from 0 (practiced frequently) over 1 (practiced occasionally) to 2 (did not occur), so the cumulative physical integrity index ranges from 0 to 8. We rescale the variable, so that higher values go in line with higher human rights violations and following Hypothesis 1a we expect a positive effect on antiUS terrorism. 9 Military and Economic Closeness To measure the closeness of the home government to the US government we use military and economic US support and dependence generated by military and economic aid from the US. According to recent empirical studies investigating the relationship between aid flows and political support of the receiving country, such as Alesina and Dolar (2000), Kuziemko and Werker (2006), Dreher et al. (2008), Mesquita and Smith (2009) and Fleck and Kilby (2001, 2006) and Milner and Tingley (2010), assistance may not serve as an unselfish giveaway or charity of donors, but as an investment with self-interest that may also be used to buy political support from the recipients of aid. The data for economic and military aid flows is provided by the USAID Economic Analysis and Data Services (EADS). This database offers information on aid flows in constant dollars from the US to all receiving countries classified into assistance programs since 1946.5 To control for scale effects we divide by national GDP. Controls To avoid the problem of spurious relationships between the dependent and the independent variables, our baseline model controls for a number of variables commonly identified as potential determinants of terrorism. In particular, we control for the impact of economic development, population size, distance to the US, regime stability, and state failure. 5 We experimented also with alternative measures of military dependence like US military personal measured as the ratio of US to domestic military personal stationed in the country and US arms exports (divided by domestic military expenditures). The results support the general effect of military aid but are not presented here due to space limitations (available upon request). 10 Economic Development First, we account for the effect of economic development. Theoretically, the opportunity costs reflected in a country’s economic development may influence the cost-benefit expectations of terrorists and particularly their supporters (Freytag et al. 2011; Caplan, 2006, Sandler and Enders, 2004; Blomberg et al., 2004). The current empirical literature is highly suggestive that poor economic development is a root cause of terrorism and increases the probability of terrorist attack from that country (Blomberg et al., 2004; Freytag et al., 2011 and Gries and Meierrieks, 2013). To control for the economic development of a country we use (logged) GDP per capita extracted from the Penn World Table. Population Further, we account for the effect of population size on terrorism. From the theoretical perspective countries with larger populations might have a larger potential for terrorism by producing more potential groups and individuals willing to engage in terrorism. Further, more populous countries might have higher monitoring costs making counter terrorism more difficult. This perception is confirmed by numerous empirical studies which find that population size is positively correlated with terrorism (e.g., Piazza, 2006; Li and Schaub, 2004; Burgoon, 2006). The variable (logged) population size is also extracted from the Penn World Table. Distance Following the assumption that a higher distance between origin and target country is associated with higher planning effort and higher costs for terrorist attacks we control for the distance using the (logged) distance between Washington, D.C., and the respective foreign country’s capital as reportet in the CEPII GEODist Database (Mayer and Zignago, 2011). 11 Regime stability A broad field of literature has already investigated the influence of political institutions on the extent of terrorism with mixed results. On the one hand, results of Piazza (2006), Eyerman (1998), and Krueger and Laitin (2008) suggest that democracy can reduce terrorism. On the other hand, studies of Li (2005), Blomberg and Hess (2008a, 2008b), Braithwaite and Li (2007) and Burgoon (2006) present opposite results suggesting that liberal democracies are thought to provide a favorable environment for terrorism by giving terrorists freedom of association and movement. Abadie (2006) and Kurrild-Klitgaard et al. (2006) indicate a non-liner relationship between political liberty and terrorism and suggest that democratic regimes as well as the repressive practices commonly adopted by autocratic regimes may help keeping terrorism at bay. They provide support that it is not primarily the regime type but regime stability has an effect on the emergence of terrorism and that especially in times of regime changes and political transitions political instabilities provide incentives and opportunities for terrorist groups to strike. We control for regime stability using the regime durability variable from the Polity IV Project operationalized by the number of years since the most recent regime change and expect a positive effect on terrorism. State failure Finally, we control for state failure and assume it to have a positive effect on terrorism for the same reasons as regime stability. Involving political instability, failed states might provide conditions under which certain types of terrorist groups can operate. Results of Piazza (2007, 2008a, 2008b), Campos and Gassebner (2013) and Newman (2007) show that states that experience more incidents of state failures are more likely to experience transnational terrorism. We use the variable state failure from the State Failure Task Force database which measures the 12 intensity of revolutionary wars, ethnic wars, adverse regime changes, and genocides and politicides. While the former variables enter our baseline model, we run a battery of specifications including further controls to assess the robustness of our findings. Specifically, to account for the auto correlation structure of the dependent variable, we expand the specification by considering a lag of the dependent variable as a regressor. Further, we analyze the influence of economic growth (as an alternative to the level of development), the impact of trade openness and government size.6 3.2 Econometric Methodology and Results Since the set of our dependent variables is classified as count data variables, which assume only discrete, non-negative values, a sporadic and rare nature of incidents with a high number of zeroobservations, an uneven distribution and a possible dependence between the incidents, we have to apply a count-data regression technique accounting for the special structure of the data. Furthermore, preliminary tests suggested overdispersion of the dependent variables;7 therefore we estimate pooled negative-binomial regression models on our panel dataset.8 6 Further robustness checks control also for the effects of democracy, ethnic and religious tension and the cold war period. The results are available upon request. 7 In the statistical sense overdispersion defines the situation when the variance of a variable is higher than its mean. 8 Alternative estimation techniques like random effect negative-binomial regression and Poison regression do not change the quality of our results To account for a possible problem of zero inflation we run also zero inflated negative binomial regression as a robustness check which also support our main finding. The results are available upon request. 13 3.2.1 Human Rights Violations and US Dependence In a first step we test Hypotheses 1a and 1b and investigate the impact of human rights violations and economic and military assistance on anti-US terrorism estimating a set of equations of the following form: terrorism i,t = α + β1 hr_violations i,t-1 + β2 aid i,t-1 + γ Z i,t-1 + ηi + µt + εi,t where terrorism i,t (1) is the dependent variable measured alternatively by US victims, US attacks, killings in anti US attacks, or attacks with US victims in country i, hr_violations i,t-1 measures the abuses in human rights quantified by the violations in physical integrity rights in country i and aid i,t-1 represents alternatively the US economic or US military aid flows into country i. Z i,t-1 is a set of control variables presented above, ηi and µt are region and time dummies and εi,t is the disturbance term. To counteract the problem of a possible endogeneity between human rights violations and terrorism and aid flows and terrorism and because we assume the independent variable not to have an immediate but a deferred effect on the independent variable we let all independent variables enter the model with a lag of one year. 9 The results are presented in Table 9 In the context of reverse causation Dreher et al. (2010) investigate the effect of counter terrorism on human rights violations with the result that counter terrorism strategies, as an outcome of appearance of terrorism have a negative impact on human rights. Piazza and Walsh (2009) investigate the effect of terrorist attacks on human rights behavior of governments and find significant effects for a couple of human rights categories. These results suggest a possibly endogenous relationship between terrorism and human rights abuses in a country. Further, Azam and Thelen (2010) discuss a possible endogeneity problem between aid and terrorism. To account for this possible endogeneity and to check the robustness of the negative binomial regressions we additionally reverse the regression and account for a possible effect of (lagged) terrorism on aid flows and (lagged) terrorism on human rights violations using the same set of control variables. Our results indicate that a reverse causality is not present. 14 2, whereat the first four specifications present the results for economic aid and the last four for military aid.10 We find a highly significant robust correlation between human rights violations and anti-US terrorism from the respective country. Greater respect for physical integrity rights consistently reduces anti-US terrorism regardless of the type of the dependent variable. This confirms Hypothesis 1a and is in line with the prediction that anti-US terrorism is more likely to emerge in countries with high repression measured by violations in physical integrity rights. This result is similar to the findings of Walsh and Piazza (2010) who show that terrorist activity in a country is positively related to local human rights violations. However, our results go even further and suggest that government repression seems to provoke more terrorism not only directed on the hostile government but also on politically close transnational targets. Political violence and human rights violations seem to favor oppositions directed also towards countries supporting the domestic power structure. The coefficients of economic and military aid are positive and significant through all specifications. This result confirms Hypothesis 1b and reveals that aid flows from the US increase the probability of transnational US terrorism from that country so that anti-US terrorism is more likely to have its genesis in countries with a high military and economic dependence on the US. Our results confirm the findings of Neymayer and Plümper (2011) who show that attacking foreign countries is the more attractive the more the home government depends on support from the foreign country. In line with studies of Boutton and Carter (2013), Boutton (2013), Sullivan et al. (2011), Findley (2012), and Young and Findley (2011) we show that aid flows, which are 10 Since economic and military aid are highly correlated we do not include both variables simultaneously in one specification to avoid multicollinearity problems. 15 often addressed as instruments of counter terrorism strategies (Bandyopadhyay et al., 2011, Azam and Delacroix, 2006) seem to have a counterproductive effect and rather drive terrorism directed on US interests. As argued by Lizardo (2006) economic dependence and the influence of the US seem to trigger fear of a cultural globalization and lead to an increase in anti-US terror. Considering the set of control variables population size shows a consistently positive effect on anti-US terrorism. It can be assumed that the pure size effect of the country and the associated greater risk of the emergence of terrorism as well as higher monitoring costs are crucial. As expected, state failure positively affects anti-US terror, indicating that countries with high political instability may provide conditions under which terrorists and terrorist groups can operate. This is also the case for regime stability, where the results show a negative effect of regime durability on terror measured by US attacks and attacks with US victims. This is in line with the studies of Abadie (2006) and Kurrild-Klitgaard et al. (2006) who argue that political liberty has a non-linear effect on terrorism and not the regime type but the regime durability is essential for poverty reduction.11 Finally, the distance to the US coefficient is positive albeit not consistently significant. Surprisingly, this result is contrary to our assumption that transnational terrorism decreases with increasing distance. However, one should bear in mind that the most part of anti-US terrorism is not within America, so that the distance becomes less important. To check the robustness of our main findings we implement a set of additional controls, namely GDP per capita growth, trade openness and government size and include also a lag of the dependent variable as a regressor. The results presented in Table 3 show that the identified 11 Alternatively to regime stability, we also account for the effect of democracy, however the results show no significant effects. 16 positive effects of human rights violations and economic and military aid on anti-US terrorism are robust and stay highly significant with the inclusion of alternative controls. The positive effect of the lagged terrorism variable is in line with Freytag et al. (2011), Enders and Sandler (2005) and Lai (2007) and suggests path dependence. For instance, longer terrorist campaigns generate economies of scale by reducing per-incident costs and generating a higher media attention. Considering the coefficients of economic growth, trade openness and government size the results show no significant effects. 3.2.2 Interaction of Human Rights Violations and US Dependence In a second step we test Hypothesis 2 and investigate the impact of a possible interaction effect of human rights violations and military and economic dependence. We assume that the terrorism generating effects of aid flows are bigger in countries with higher human rights violations, and the other way around, that repression boosts anti-US terrorism over proportionally in countries with high aid flows from the US. To account for this effect we augment model specification (1) by the product of both predictors resulting in the following model: terrorism i,t = α + β1 hr_violations i,t-1 + β2 aid i,t-1 + β3 (hr_violations* aid) i,t-1 + γ Z i,t-1 + ηi + µt + εi,t (2) The interpretation of interaction terms in count models in not trivial. As shown in Ai and Norton (2003) the magnitude of the interaction effect in nonlinear models does not equal the marginal effect of the interaction term. A comprehensive analysis of interaction terms in negative binomial models can be found in Hilbe (2011). Following this methodology we approach the analysis 1) by binary interaction terms and 2) by continuous interaction terms. 17 Table 2: Human Rights Violations US Aid and anti US Terrorism – Baseline Regression GDP p.c. (ln) t-1 Population (ln) t-1 HR violations t-1 Economic aid t-1 (1) US victims (2) US attacks 1.035*** (0.290) 0.636*** (0.171) 0.391*** (0.096) 94.927*** (27.256) 0.494** (0.209) 0.450*** (0.118) 0.319*** (0.064) 46.733** (19.564) (3) Killings in anti US attacks 0.739** (0.318) 0.857*** (0.203) 0.488*** (0.115) 135.987* (71.749) (4) Attacks with US victims (5) US victims (6) US attacks 0.380* (0.208) 0.431*** (0.114) 0.339*** (0.069) 45.705** (22.925) 0.963*** (0.307) 0.538*** (0.165) 0.410*** (0.095) Military aid t-1 Distance (ln) Regime stability t-1 State failure t-1 Mean VIF Wald χ2 (Prob. > χ2) LogPseudolikelihood Observations 0.690 (0.428) -0.003 (0.010) 0.274*** (0.085) 2.14 1455.81 0.00 -359.08 0.309*** (0.062) -0.020*** (0.007) 0.141** (0.067) 2.14 457.96 0.00 -1224.13 1.661* (1.001) -0.015 (0.011) 0.359* (0.207) 2.14 725.53 0.00 -541.90 0.313*** (0.068) -0.020*** (0.007) 0.167** (0.067) 2.14 553.25 0.00 -1130.56 232.518*** (75.401) 0.575** (0.258) -0.005 (0.009) 0.264*** (0.080) 2.13 32724.10 0.00 -359.61 2420 2420 2420 2420 2420 (8) Attacks with US victims 0.428** (0.208) 0.397*** (0.113) 0.312*** (0.060) (7) Killings in anti US attacks 0.673** (0.315) 0.723*** (0.188) 0.529*** (0.110) 169.674*** (47.990) 0.290*** (0.060) -0.020*** (0.007) 0.140** (0.065) 2.13 444.73 0.00 -1219.51 325.097*** (120.394) 1.107 (0.752) -0.015 (0.011) 0.317 (0.202) 2.13 783.77 0.00 -542.66 190.129*** (50.270) 0.294*** (0.072) -0.021*** (0.006) 0.166** (0.065) 2.13 601.74 0.00 -1122.39 2420 2420 2420 0.328 (0.212) 0.381*** (0.107) 0.335*** (0.065) Notes: Constant not reported. All models include year and regional dummies (not reported). Robust standard errors clustered over countries in parentheses. * p < 0.10, ** p < 0.05, *** p < 0 18 Table 3: Human Rights Violations US Aid and anti US Terrorism (US victim) – Additional Controls US victims t-1 GDP p.c. (ln) t-1 (1) US victims 0.359** (0.156) 0.965*** (0.268) GDP p.c. growth t-1 Population (ln) t-1 HR violations t-1 Economic aid t-1 0.600*** (0.163) 0.380*** (0.092) 89.102*** (25.543) (2) US victims -0.036 (0.024) 0.500*** (0.169) 0.327*** (0.092) 58.223*** (20.380) (3) US victims (4) US victims 1.032*** (0.294) 1.107*** (0.299) 0.662*** (0.222) 0.397*** (0.110) 95.138*** (26.374) 0.658*** (0.183) 0.394*** (0.097) 95.191*** (26.673) Military aid t-1 Distance (ln) Regime stability t-1 State failure t-1 0.685 (0.437) -0.004 (0.010) 0.273*** (0.083) 0.935 (0.735) 0.007 (0.011) 0.276*** (0.084) Trade openness t-1 0.699 (0.425) -0.004 (0.011) 0.272*** (0.087) 0.001 (0.011) Government size t-1 Mean VIF Wald χ2 (Prob. > χ2) Log Pseudolikelihood Observations 0.670* (0.387) -0.005 (0.010) 0.263*** (0.092) 2.12 1659.99 0.00 -357.63 2.04 690.99 0.00 -364.78 2.16 3974.09 0.00 -359.05 0.012 (0.018) 2.17 1757.10 0.00 -358.92 2420 2420 2420 2420 (5) US victims 0.420*** (0.142) 0.901*** (0.282) (6) US victims 0.505*** (0.153) 0.397*** (0.089) -0.036 (0.024) 0.450*** (0.165) 0.340*** (0.091) 225.779*** (68.452) 0.566** (0.263) -0.005 (0.009) 0.261*** (0.078) 156.551** (66.975) 0.727 (0.504) 0.006 (0.010) 0.270*** (0.080) (7) US victims (8) US victims 0.962*** (0.310) 1.017*** (0.330) 0.540** (0.239) 0.410*** (0.103) 0.550*** (0.177) 0.413*** (0.095) 232.512*** (75.559) 0.575** (0.258) -0.005 (0.010) 0.263*** (0.082) 0.000 (0.011) 232.374*** (74.485) 0.569** (0.250) -0.005 (0.010) 0.256*** (0.088) 2.11 11769.01 0.00 -357.61 2.03 14448.14 0.00 -364.73 2.15 40118.74 0.00 -359.61 0.009 (0.019) 2.16 33010.88 0.00 -359.54 2420 2420 2420 2420 Notes: Constant not reported. All models include year and regional dummies (not reported). Robust standard errors clustered over countries in parentheses. * p < 0.10, ** p < 0.05, *** p < 0 19 Binary interaction Computing binary interactions allows us to account for different interaction coefficients and their significance levels subjected to the value of the binary variables. The approach requires both variables composing the interaction term to be binary. Therefore, we transform the variables hr_violations and aid (economic and military) into dummy variables, where 1 denotes high human rights violations and high aid flows and 0 respectively low values.12 Given that β1 is the estimated coefficient for physical integrity rights, β2 is the estimated coefficient for economic or military dependence and β3 is the estimated coefficient for their product, the risk ratios of the interaction effects are calculated by: exp[β1+ β3* aid] and (3) exp[β2+ β3* hr_violations]. (4) Table 3 presents the results for all combinations of binary interaction including the coefficients, the standard errors, and the incident rate ratios (IRR).13 The corresponding regression results are presented in the Appendix in Table A1. 14 12 We construct the dummies according to the mean value of the variables, so that the dummy is 1 when the value is higher than the mean and 0 when the value is lower than the mean. Alternatively, we also run regressions with a dummy based on the median of the variables, however, our main findings do not change. 13 The standard errors differ for each interaction combination. They are computed as the square root of the variance of the interaction coefficient, which is V(β2)+ hr_violations2V(β3) + 2 * Cov (β2, β3) for equation (3) and V(β2)+ aid2V(β3) + 2 * Cov (β2, β3) for equation (4). 14 We present only results for the dependent variable US victim. However, results for alternative dependent variables like US attack, Killings in anti US incidents, and Incidents with US victims support our main findings. Results are available upon request. 20 Table 4: Binary Interaction Effect of Human Rights Violations and Economic/Military Aid Dummy relation - high HR violations and high aid : high HR violations and low aid - high HR violations and low aid : low HR violations and low aid - high HR violations and high aid : low HR violations and high aid - low HR violations and high aid : low HR violations and low aid Economic Aid Coef. Std. Err. 1.0472 0.6586 2.8497 Military Aid Coef. Std. Err. 1.3385 0.6586 3.3677*** 1.1008 29.0117*** 1.5838* 1.1008 4.8734* 1.7146*** 0.5768 5.5545*** 1.160** 0.5768 3.1899** 4.0352*** 1.0824 56.5542*** 1.4054* 1.0824 4.0772* IRR IRR 3.8133 Considering countries that receive high aid our results show that the incident rate ratios of high human rights violations to low human rights violations are 5.5545 for economic aid and 3.1899 for military aid. These effects are significant and reveal that in countries with a high US dependence terrorism is 555% and 319% more likely when human rights are violated. Further, in countries with low human rights violations the emergence of anti US-terrorism is higher when the aid flows from the US are high (5655% for economic and 408% for military aid) and focusing on low aid countries we can identify a higher probability (2901% for economic and 487% for military aid) for anti-US terrorism with predominant government repression. The results reveal that the terrorism driving effects of repression and US dependence mutually depend on each other. Anti US-terrorism is not only boosted by the isolated factors but rather by their combination, here a repressive government structure plus a high economic and military dependence on the US seem to create favoring conditions for its emergence. Continuous interaction In a next step we implement a continuous interaction term that provides more detailed information on the impact of a variable in dependence on the value of the other variable. It allows 21 for a more precise analysis of the interdependencies of human rights violations and US dependence on anti-US terrorism. Adding an interaction computed as the product of two continuous variables to a model can result in a host of coefficients. Assuming n values of human rights violations and m values of aid the slope results in n * m separate coefficients. To present the results more clearly Hilbe (2011) suggests to group continuous variables and to present the results for selected values. So we compute the interaction coefficient holding one variable constant and varying the other whereat we consider nine grades.15 The results for US victims as the dependent variable are presented in Table 5 and 6, where Table 5 shows the effect of aid dependent on varying values of human rights violations (according to equation (3)) and Table 6 shows the effect of human rights violations for varying values of aid (according to equation (4)). The corresponding regression results are presented in the Appendix in Table A2. The human rights violations value ranges from 0 (no violations) to 8 (very high violations). The results reveal that the effect of economic aid and military aid on US terrorism increases with increasing human rights violations. While for countries with no human rights violations (HR violations =0) the IRR of economic aid is 6.53E+36, rising continuously with increasing repression arriving at a value of 1.74+E44 for the highest repression (HR violations = 8). For military aid the IRR rises continuously from 1.13+E72 for very low repression to 9.29+E132 for high repression with the effect being significant for a repression value of 2 upwards. 15 We group the ranges of the variables into nine categories. For human rights violations we consider all integers from 0 to 8, which is exactly the range of the variable. For economic and military aid we consider values ranging from the minimum to the maximum value of the variable, however in addition we weight the intervals according to the distribution within the range. 22 Table 5: Effect of Economic Aid Dependent on the Value of HR Violations HR violations value 0 1 2 3 4 5 6 7 8 Economic aid Coef. 84.78** 86.92*** 89.06*** 91.20*** 93.33*** 95.47*** 97.61*** 99.75*** 101.89*** Std. Err IRR 37.16 32.71 29.20 26.99 26.42 27.58 30.27 34.14 38.83 6.53E+36 5.55E+37 4.72+E38 3.97+E39 3.38+E40 2.87+E41 2.44+E42 2.07+E43 1.74+E44 HR violations value 0 1 2 3 4 5 6 7 8 Military aid Coef. 165.91 183.44 200.98* 218.51*** 236.04*** 253.58*** 271.11** 288.64** 306.18* Std. Err IRR 170.93 136.05 104.44 80.08 70.87 82.08 107.51 139.59 174.69 1.13+E72 4.64+E79 1.93+E87 7.90+E94 3.24+E102 1.33+E110 5.52+E117 2.26+E125 9.29+E132 Considering the effect of human rights violations as a function of aid (Table 6) we can identify a stronger effect with increasing aid flows. While for countries not receiving aid from the US a one unit increase in human rights violation results in a 47% (economic aid) and 50% (military aid) increase in US terrorism (given the other variables are held constant in the model). In countries which are highly dependent on US aid flows a one unit increase in human rights violations results in a 60% (economic aid) and 64% (military aid) increase in anti-US terrorism. Table 6: Effect of Human Rights Violations Dependent on the Value of Economic Aid Economic aid value 0 0.0001 0.0005 0.001 0.005 0.01 0.02 0.03 0.04 HR violations Coef. 0.3835*** 0.3837*** 0.3845*** 0.3856*** 0.3942*** 0.4049*** 0.4262*** 0.4476** 0.4690* Std. Err. IRR Military aid value 0.1046 0.1043 0.1031 0.1017 0.0939 0.0950 0.1282 0.1821 0.2434 1.4673 1.4676 1.4689 1.4705 1.4845 1.4990 1.5384 1.5646 1.5982 0 0.000001 0.000005 0.00001 0.00005 0.0001 0.0005 0.001 0.005 HR violations Coef. 0.40402*** 0.40404*** 0.40411*** 0.40420*** 0.40490*** 0.40578*** 0.41279*** 0.42156*** 0.49169** Std. Err. IRR 0.0952 0.0952 0.0952 0.0952 0.0949 0.0946 0.0939 0.0966 0.2037 1.4978 1.4979 1.4980 1.4981 1.4981 1.5005 1.5110 1.5243 1.6351 23 So both, the results for the binary interaction as well as for the continuous interaction reveal that the effects of home repression and US dependence should not be considered independently of each other but, what is remarkable, that they influence each other. In summary, we are able to confirm hypotheses 1a and 1b as well as hypothesis 2. Both, repressive state behavior – measured by human rights violations – as well as a close relationship of the home regime to the US – measured by economic and military aid flows from the US – seem to increase anti US terrorism in each case. Further, interaction effects of these variables show that a combination of state repression and state support by the US leads to an over proportional effect. The discontent with the domestic regime is more strongly projected on the US the more repression takes place in the country and the more the domestic political situation is stabilized and supported by the US. Conclusion We investigate possible determinants of anti-US terrorism. In particular, we ask whether oppressive governments and/or close connections with the US encourage attacks against the US. We argue that oppressive state behavior, such as human rights violations, may provoke resistance against the authorities, including violent attacks. Further, for terrorist groups which seek to overthrow their home country’s political system it might be attractive to target the US as foreign allies especially if the government’s capacity seems dependent on US support. Arriving at our test hypotheses we assume that anti-U.S. terrorism is more likely to emerge in countries with high repression (hypothesis 1a) and in countries with a high dependence to the US (hypothesis 1b). Further, a combination of both conditions may drive anti-US terrorism to an over proportional degree (hypothesis 2). The negative binomial regressions for a panel of 149 countries from 1981 to 2005 indicate that both oppressiveness and US closeness are important 24 determinants of anti-US terrorism. In summary, we are able to confirm hypotheses 1a and 1b as well as hypothesis 2. Both, repressive state behavior – measured by human rights violations – as well as a close relationship of the home regime to the US – measured by economic and military aid flows from the US – seem to increase anti-US terrorism in each case. Further, both effects do not seem completely independent. Interaction effects of these variables show that a combination of state repression and state support by the US leads to an over proportional effect. The discontent with the domestic regime is more strongly projected on the US, the more repression takes place in the country, and the more the domestic political situation is stabilized and supported by the US. 25 References Abadie, A. (2006). Poverty, political freedom, and the roots of terrorism. American Economic Review, 96(2): 50–56. Ai, C.; Norton, E.C. (2003). Interaction terms in logit and probit models. Economics Letters, 80:123-129. Alesina, A., Dollar, D. (2000). Who gives foreign aid to whom and why? Journal of Economic Growth, 5(1): 33–63. Anderton, C.H.; Carter, J.R. (2005). On Rational Choice Theory and the Study of Terrorism, Defence and Peace Economics, 16(4): 275-282 Azam, J.P.; Delacroix, A. (2006).Aid and the Delegated Fight Against Terrorism, Review of Development Economics, 10(2), 330–344. Azam, J.P.; Thelen, V. (2010). Foreign aid versus military intervention in the war on terror. Journal of Conflict Resolution, 54(2): 237-261. Bandyopadhyay, S.; Sandler, T.; Younas, J. (2011). Foreign aid as counterterrorism policy, Oxford Economic Papers, 63(3): 423–447. Bapat, N.A. (2011). Transnational terrorism, US military aid, and the incentive to misrepresent, Journal of Peace Research, 48(3) 303–318. Bates, R.H.; Goldstone, J.A.; Gurr, T.R.;, Harff, B.; Levy, M.A.; Marshall, M.G.; Epstein, D.L.; Kahl, C.H.; Woodward, M.R.; Surko, P.T.(2006). State Failure Task Force Report: Phase IV. McLean, VA: Science Applications International Corp.; 2006. Bin Laden, O. (2002). Letter to America, The Guardian, published online: http://www.theguardian.com/world/2002/nov/24/theobserver. 26 Blomberg, S.B., Hess, G.D., Weerapana, A., (2004). Economic conditions and terrorism. European Journal of Political Economy 20(2), 463–478. Blomberg, S. B., Hess, G. D. (2008a). The Lexus and the olive branch: globalization, democratization and terrorism. In Keefer, P., & Loayza, N. (Eds.) Terrorism, economic development, and political openness (pp. 116–147). New York: Cambridge University Press. Blomberg, S. B., Hess, G. D. (2008b). From (no) butter to guns? Understanding the economic role in transnational terrorism. In Keefer, P., & Loayza, N. (Eds.) Terrorism, economic development, and political openness (pp. 83–115). New York: Cambridge University Press. Boutton, A. (2013). Interstate Rivalry, U.S. Foreign Aid, & Incentives for Counterterrorism Cooperation, Journal of Peace Research, forthcoming. Boutton, A.; Carter, D.B. (2013). Fair-Weather Allies? Terrorism and the Allocation of US Foreign Aid, Journal of Conflict Resolution, forthcoming. Braithwaite, A., Li, Q. (2007). Transnational terrorism hot spots: identification and impact evaluation. Conflict Management and Peace Science, 24(4): 281–296. Burgoon, B. (2006). On welfare and terror: social welfare policies and political-economic roots of terrorism. Journal of Conflict Resolution, 50(2): 176–203. Campos, N.F.; Gassebner, M. (2013). International terrorism, domestic political instability, and the escalation effect, Economics and Politics, 25(1): 27–47. Caplan, B. (2006). Terrorism: the relevance of the rational choice model. Public Choice 128(1-2): 91–107. Cingranelli, D.L.; Richards, D.L. (2010). The Cingranelli and Richards (CIRI) Human Rights Data Project, Human Rights Quarterly 32(2): 401-424. Dreher A.; Nunnenkamp, P.; Thiele, R. (2008). Does US aid buy UN general assembly votes? A disaggregated analysis, Public Choice 136(1-2): 139-164. 27 Dreher, A.; Gassebner, M.; Siemers, L.H. (2010). Does Terrorism Threaten Human Rights? Evidence from Panel Data, Journal of Law and Economics, 53(1): 65-93. Enders, W.; Sandler, T. (2005). Transnational terrorism 1968–2000: Thresholds, persistence, and forecasts, Southern Economic Journal, 71(3), 467–482. Eyerman, J. (1998). Terrorism and democratic states: soft targets or accessible systems? International Interactions, 24(2): 151–170. Findley, M.G.; Piazza, J.A.; Young, J.K. (2012). Games Rivals Play: Terrorism in International Rivalries, The Journal of Politics, 74(1): 235-248. Fleck, R.; Kilby, C. (2001). Foreign aid and domestic politics: voting in Congress and the allocation of USAID contracts across congressional districts, Southern Economic Journal 67, 598–617. Fleck, R.K.; Kilby, C. (2006). How do political changes influence U.S. bilateral aid allocations? Evidence from panel data, Review of Development Economics, 10(2): 210–223. Freytag, A.; Krüger, J.; Schneider, F.; Meierrieks, D. (2011), The origins of terrorism: Crosscountry estimates of socio-economic determinants of terrorism, European Journal of Political Economy. 27(1): 5-16. Gries, T., Meierrieks, D. (2013), Causality between terrorism and economic growth, Journal of Peace Research. 50(1): 91-104. Hegre, H.; Ellingsen, T.; Gates, S.; Gleditsch, N.P. (2001). Toward a Democratic Civil Peace? Democracy, Political Change, and Civil War, 1816–1992, American Political Science Review, 95(1): 33-48. Heston, A.; Summers, R.; Aten, B. (2012). Penn World Table Version 7.1, Center for International Comparisons of Production, Income and Prices at the University of Pennsylvania, Nov 2012. 28 Hilbe, J.M. (2011), Negative Binomial Regression. Cambridge University Press: New York. Krueger, A.B.; Laitin, D.D. (2008). Kto kogo?: a cross-country study of the origins and targets of terrorism. In Keefer, P., & Loayza, N. (Eds.) Terrorism, economic development, and political openness (pp. 148–173). New York: Cambridge University Press. Krueger, A.B.; Malečková, J. (2003). Education, poverty and terrorism: Is there a causal connection? The Journal of Economic Perspectives, 17(4): 119-144. Kurrild-Klitgaard, P.; Justesen, M.K.; Klemmensen, P. (2006). The political economy of freedom, democracy and transnational terrorism. Public Choice, 128(1-2): 289–315. Kuziemko, I., Werker, E. (2006). How much is a seat on the Security Council worth? Foreign aid and bribery at the United Nations. Journal of Political Economy, 114(5): 905–930. Lai, B. (2007). Draining the swamp: An empirical examination of the production of international terrorism, 1968–1998, Conflict Management and Peace Science, 24(4): 297– 310. Li, Q. (2005). Does democracy promote or reduce transnational terrorist incidents? Journal of Conflict Resolution, 49(2): 278–297. Li, Q.; Schaub, D. (2004). Economic globalization and transnational terrorism: A pooled time-series analysis, Journal of Conflict Resolution, 48(2): 230-258. Lizardo, O. (2006). The effect of economic and cultural globalization on anti-U.S. transnational terrorism: 1971-2000, Journal of World Systems Research, 12: 149-186. Marshall, M.G.; Jaggers, K. (2012). Polity IV project: Political regime characteristics and transitions, 1800-2002. Mayer, T.; and Zignago, S. (2011). Notes on CEPII’s distances measures: The GeoDist Database, CEPII Working Paper 2011-25. Paris: Centre d’Etudes Prospectives et d’Informations Internationales. 29 Mesquita, B.B. De; Smith, A. (2009). A political economy of aid, International Organization, 63(2): 309-340. Mickolus, E.F.; Sandler, T.; Murdock, J.M.; Flemming, P. (2007). International Terrorism: Attributes of Terrorist Events, 1968-2011. Dunn Loring, VA: Vinyard Software. Milner, H.V.; and Tingley, D.H. (2010). The political economy of US foreign aid: American legislators and the domestic politics of aid, Economics and Politics, 22(2): 200–232. Neumayer, E.; Plumper, T. (2011). Foreign Terror on Americans, Journal of Peace Research, 48(1): 3–17. Newman, E. (2007). Weak states, state failure, and terrorism, Terrorism and Political Violence, 19(4): 463-488. Obama, B. (2008). Statement of President-elect Obama on Human Rights Day, The Office of the President-elect, published online: http://change.gov/newsroom/entry/statement_of_president_elect_obama_on_human_rights_d ay/ Piazza, J.A. (2006). Rooted in poverty? Terrorism, poor economic development, and social cleavages. Terrorism and Political Violence, 18(1): 159–177. Piazza, J. A. (2007). Draining the swamp: democracy promotion, state failure, and terrorism in 19 Middle Eastern countries. Studies in Conflict, Terrorism, 30(6): 521–539. Piazza, J. A. (2008a). Do democracy and free markets protect us from terrorism? International Politics, 45: 72–91. Piazza, J. A. (2008b). Incubators of terror: Do failed and failing states promote transnational terrorism? International Studies Quarterly, 52(3): 469–488. Piazza, J. A.; Walsh, J.I. (2009). Transnational Terror and Human Rights, International Studies Quarterly, 53, 125–148. 30 Plumper, T.; Neumayer, E. (2010). The friend of my enemy is my enemy: International alliances and international terrorism, European Journal of Political Research, 49: 75–96. Sandler, T.; Enders, W. (2004). An economic perspective on transnational terrorism, European Journal of Political Economy, 20(2): 301–316. Sullivan, P.; Tessman, B.F.; Li, X. (2011). US military aid and recipient State cooperation, Foreign Policy Analysis, 7(3): 275–294. US Agency for International Development (USAID) Economic Analysis and Data Services. 2013. ‘‘U.S. Overseas Loans and Grants, Greenbook.’’ Technical report, United States Agency for International Development, Washington, DC. Walsh, J.I.; Piazza, J.A. (2010), Why Respecting Physical Integrity Rights Reduces Terrorism, Comparative Political Studies, 43 (5), 551-577. Young, J.K.; Findley, M.G. (2011). Can Peace be Purchased? A Sectoral-Level Analysis of Aids Influence on Transnational Terrorism, Public Choice, 49 (3/4): 365–81. 31 Appendix Table A1: Human Rights Violations US Aid and anti US Terrorism (US victim) with Binary Interaction (1) (2) US victims US victims GDP p.c. (ln) t-1 1.018*** 0.481* (0.301) (0.262) Population (ln) t-1 0.358** 0.364** (0.170) (0.171) Dummy (1) HR 3.368*** 1.584* violations t-1 (1.101) (0.852) Dummy (2.1) 4.035*** Economic aid t-1 (1.082) Dummy (2.2) 1.405* Military aid t-1 (0.725) Interaction -2.321** (1)*(2.1) (1.136) Interaction -0.245 (1)*(2.2) (0.796) Distance (ln) 1.159 0.608 (0.917) (0.448) Regime stability t-1 -0.011 -0.004 (0.009) (0.009) State failure t-1 0.522*** 0.432*** (0.118) (0.110) Mean VIF 2.36 2.21 Wald χ2 5667.54 6769.22 (Prob. > χ2) 0.00 0.00 Log-404.24 -409.80 Pseudolikelihood Observations 2513 2513 Robust standard errors clustered on countries in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01 32 Table A2: Human Rights Violations US Aid and anti US Terrorism (US victim) with Continuous Interaction (1) (2) US victims US victims GDP p.c. (ln) t-1 Population (ln) t-1 (1) HR violations t-1 (2.1) Economic aid t-1 1.030*** (0.289) 0.638*** (0.174) 0.383*** (0.105) 84.771** (37.162) (2.2) Military aid t-1 0.960*** (0.306) 0.538*** (0.164) 0.404*** (0.095) 165.911 (170.926) Interaction (1)*(2.1) Interaction (1)*(2.2) 2.138 (6.832) Distance (ln) 0.694 (0.440) -0.003 (0.010) 0.274*** (0.085) -29.087*** (5.932) 2.16 1885.51 0.00 -359.06 17.533 (39.405) 0.579** (0.264) -0.005 (0.009) 0.260*** (0.079) -26.558*** (5.176) 2.70 25414.75 0.00 -359.57 2420 2420 Regime stability t-1 State failure t-1 Constant Mean VIF Wald χ2 (Prob. > χ2) LogPseudolikelihood Observations Robust standard errors clustered on countries in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01 33