1 __________________________________________________________________________________
How interstate relations affect the likelihood of the
coup d’état in Africa, 1945-1999
___________________________________________________________________________
T.S. Leenstra
11262214
Master Thesis
MA Political Science: International Relations
University of Amsterdam
Graduate School of Social Sciences
June 23, 2017
Under the supervision of:
1
streader: Dr. A.A. Steele
2
ndreader: Dr. J. Krause
2
_________________________________________________________________________
Abstract
This thesis examines the effect of interstate signals on the likelihood of coups in Africa between 1945 and 1999. The purpose of this thesis is to improve our understanding how interstate relations affect the probability of coups. The dominant school of thought in the literature relies heavily on intrastate conditions (structural factors in society) to explain coup attempts. These explanations appear inadequate to explain considerable variation in coup attempts in Africa: why numerous African states with similar country characteristics faced several coups, while others did not. This thesis addresses the empirical puzzle by claiming that external actors play a key role in destabilizing unfavored governments in Africa. The general argument asserts that hostile signals send by external actors to a targeted state shape a favorable climate to launch a coup against the incumbent government. This argument is tested using measures for troop mobilizations and sanctions as costly signals in combination with events data as a measure for cheap signals. This thesis uses a nested analysis research design as outlined by Lieberman (2005). The first part entails a preliminary large sample of cases (LNA) which found no significant results of the expected relationships. The second part engages in an in-depth case study (SNA) of Libya to show a non-relationship more generally, and seeks alternative explanations to explain Libya’s non-coup outcome. The final section concludes that structural intrastate factors partly explain the variation of interest. The effective coup-proofing techniques implemented by Qaddafi offer a more satisfying explanation in relation to the non-coup outcome. The results demonstrate that hostile signals do not increase the probability of coups in Africa.
__________________________________________________________________________________
3
Table of contents
1. Introduction 4.
2. Literature review 8.
2.1 The relevance of International actors 9.
2.2 Coup attempts 10.
2.3 The coup d’état and its relevance 10.
2.4 The purpose of study 11.
3. Theoretical framework 13.
3.1 Rational choice theory 13.
3.2 Signaling theory 15.
3.3 Indicators of hostile signals: MID’s, sanctions and cheap hostile signals 18.
3.3.1 (MID’s) Military interstate disputes 18.
3.3.2 Sanctions 19.
3.3.3 Cheap hostile signals 20.
3.4 Control variables 21.
3.5 Interstate relations and coup risk 22.
4. Research design 25.
4.1 Data 25.
4.2 Nested analysis research design 26.
4.3 Variables of the theoretical model 27.
4.3.1 Dependent variable 27. 4.3.2 Independent variables 27. 4.4 Limitations 28. 5. Data analysis 29. 5.1 Descriptive statistics 29.. 5.2 Bivariate analysis 30. 5.3 Multivariate analysis 32.
5.4 SNA model building 34.
5.5 Case selection 34.
6. Case evidence of Libya & alternative explanations 36.
6.1 Libya and interstate relations 36.
6.1.1 US-Libya relations 37.
6.1.2 Critique on Thyne’s theory 38.
6.2 Alternative explanations: structural factors in Libyan society 39.
6.2.1 Economic performance 40.
6.2.2 Strength of civil society 40.
6.2.3 Regime legitimacy 41.
6.2.4 Recent coups 41.
6.3 Three forms of political violence: coups, revolutions, civil wars 42. 6.4 The final piece of the puzzle: Qaddafi’s efficient coup-proofing techniques 43.
7. Conclusion & discussion 46.
Bibliography 48.
4
1. Introduction
On the 24th of June 1960, Patrice Lumumba became the first democratically elected prime minster of the Democratic Republic of Congo. In the height of the Cold War, an anxious Washington presided by Dwight Eisenhower, feared a spread of communism in the heart of Africa. A country rich in minerals and of significant strategic importance to the United States (US) to restrain Soviet ambitions in the region. Due to Lumumba’s inclination towards the Soviet-Union in the Congo’s struggle for independence, the US, together with Belgium, deliberately set out to isolate the Congo from outside influence and simultaneously aimed to oust Lumumba from power to safeguard Western interests.1 As a result, the US exercised diplomatic pressure on Lumumba, ceased financial assistance and imposed (economic) sanctions on the Congo.2 Moreover, Lumumba further undermined Western interests by engaging in an intrastate military conflict with the pro-Western mineral rich Katanga province to curb its secessionist ambitions. The following quotation clearly illustrates Washington’s hostile position towards the Congo and its resolute stance for regime change.
“On 26 August 1960, the head of the CIA, Allen Dulles, sent a telegram to Lawrence Devlin, the CIA station chief in Leopoldville, saying: ‘In high quarters here it is the clear-cut conclusion that Lumumba’s removal must be an urgent and prime objective’. At a meeting of the US National Security Council (NSC) three days later, President Eisenhower, authorized the CIA to ‘eliminate’ Lumumba”.3
In 1961, already after seven months after his democratic victory, Lumumba was captured and assassinated through a Western backed military coup led by pro-Western army Colonel Mobutu Sese-Sokko. The coup d’état in the Congo is just one clear example where external actors played a key role in destabilizing a foreign government and fostered a favorable climate to stage a coup against an incumbent regime.4
Despite considerable involvement of the US in coup attempts in the Congo and elsewhere in the third world, substantial variation exists in coup attempts on a global scale. Uganda for instance, experienced seven coups between 1966 and 1986. Whereas, Botswana never experienced a coup. Considerable variation exists in the proliferation of coup attempts elsewhere in Africa. While Ghana and Sudan were frequently confronted with coups, Senegal and Eritrea faced none. Again, where countries like Burundi, Sierra Leone and Liberia were continuously plagued by coups, Namibia, Malawi and Mozambique witnessed none. Figure 1 reflects the dispersion of global coup activity and reports values for the total number of coup attempts experienced by a state between 1960 and 1999.5 These comparisons raise the following question: what explains the variation in coup attempts in Africa, diversified from practically zero to more than ten in a specific country between 1945 and 1999?
Since the end of the Second World War the world witnessed 616 coup attempts, where 245 occurred in Africa alone.6 Coups, irrespective to their outcome, “can alter political processes and social institutions as drastically as any classic revolution”.7 Despite this recognition among academics and policy makers of the immense consequences for society and the international community, the relevance of coups is seldom covered in the coup literature to claim generalizations about their causes.8 Since coups occur all over the world, they – to a certain extent – are a general phenomenon.9 This suggests there is room for a theory to explain the general causes of coups.
1 Foreign Affairs (2004).
2 Meredith (2006).
3 Meredith (2006) Chapter 6, pp.106. 4 De Witte (2002), Van Reybrouck (2010). 5 Powell (2012).
6 Luttwak (2010), Powell and Thyne (2011). 7 Luttwak (1979).
8 O’Kane (1984). 9 O’Kane (1984).
5 The coup literature offers two different approaches to address the causes of coups, which I prefer to categorize as traditional and progressive approaches. The traditional coup literature focusses primarily on internal characteristics of the state, such as economic performance, regime legitimacy, internal stability, military spending and whether a country experienced a coup in the past.10 These characteristics are framed into intrastate conditions, which on average can predict the probability of a coup in the near future.11 This line of research argues that these objective preconditions influence the stability of the incumbent regime and determine the underlying structural risk of a coup.12 More specifically, the general argument asserts that coups occur as a result of deep rooted causes in society, which in turn can shape a viable political climate to stage a coup. Accordingly, a coup, prerequisites the presence of these objective preconditions to be possible, these conditions simultaneously serve as reliable indicators to determine the likelihood whether a coup would be attempted.
Importantly, this assumption is widely shared in the coup literature, and various studies corroborated these findings over time.13 Hence, the relevance and impact of these structural factors is undisputed and I recognize its explanatory power. However, this line of reasoning raises an empirical puzzle: various countries in Africa, where structural factors which shape the underlying structural risk of a coup are present, were indeed plagued by a series of coups, while others seemed immune to the coup. In other words, following the traditional approach, we would expect a coup in each of these countries, because the level of coup risk is high due to presence of structural factors in society. Empirical evidence shows that this argument does not hold. According to the United Nations (UN), on average, both Ghana and Namibia reflect poor economic performance, inhibit a weak civil society and favor high military spending.14 However, both countries vary substantially in the number of coup attempts: Ghana experienced ten coups between 1966 and 1983, while Namibia faced none.15
In a similar view, domestic structural factors fail to explain why coups did not occur in Malawi, Zambia, or Eritrea. Because these general explanations would expect a coup in each country given their poor economic performance, illegitimate government, weak democratic institutions, high unemployment and low income rates. For example, Eritrea meets all structural factors as outlined by the traditional coup literature, but never experienced a coup. Despite Eritrea’s young period of independence, its president Isaias Afwerki, holds power since the country’s independence in 1993. Afwerki’s domestic policies foster economic backwardness and exhibits one of the highest poverty and unemployment rates in Africa. In addition, according to the UN, Eritrea lacks any form of civil society, seats an illegitimate government, enhances strong civic-military relations and structurally violates human rights on a large scale.16 These observations suggest that there is room for a theory which focusses specific on coup risk, rather than wider forms of political instability to explain the variation in coup attempts on the African continent.
The puzzle mentioned above suggests that existing explanations put forward in the traditional literature rely too heavily on intrastate conditions and appear inadequate to account for the variation in coup attempts in Africa. A further examination reveals that most African countries – on average – share similar macro level conditions.17 The variation within groups – between African countries – in terms of economic performance, strength of civic society, military expenditure and regime type does not deviate drastically from the continents average. It follows logically to treat these conditions as constant over time. The propensities of these macro level conditions make them partly inadequate to account for the variation in coup attempts and leaves room for alternative explanations. A wider
10 Belkin and Schofer (2003, 2005), Powell (2012), De Bruin (2017), Roessler (2011), O’Kane (1984), 11 Powell, Lasley and Schiel (2016), Agyeman-Duah (1990).
12 Belkin & Schofer (2003). 13 Luttwakk (1968),
14 Nambia: http://data.un.org/CountryProfile.aspx?crName=namibia Ghana: http://data.un.org/CountryProfile.aspx?crName=ghana 15 Luttwak (1968).
16 United Nations (2016) http://www.ohchr.org/EN/HRBodies/HRC/CoIEritrea/Pages/commissioninquiryonhrinEritrea.aspx 17 Luttwak (1968).
6 analysis which includes international factors might provide more satisfying answers to explain the causes of coups.
As mentioned earlier, theories focused on intrastate conditions treat coups as pure national phenomena and proclaim that the coup has been, and will be a domestic affair.18 As a result, international dynamics, such as the role of external actors on coups, remains largely untouched. Proponents of the progressive approach19 like Agyeman-Duah, Thyne and Powell, argue that the absence of a systematic examination of the possible linkage between interstate relations and coups in the literature is problematic.20 Because the traditional coup literature assumes that states exist in a vacuum and therefore neglects foreign influence on coup onset and its latter course of action.21 Since African states are interdependent political entities who are embedded in a highly interactive environment where policies of one country have implications for the other. Accordingly, we cannot treat coups and states in isolation to exclusively focus on macro level conditions of a single African country to predict the probability of a coup.
The empirical evidence suggests that coups are still relevant in Africa and we can expect that interstate relations continue to fulfill a pivotal role in these events in the future. These observations motivate to explore the direct relationship between hostile signals from external actors and coup risk. This thesis aims to improve our understanding how interstate relations affect the likelihood of coup attempts in Africa by combining qualitative and quantitative literature. Historical qualitative analysis show that the permanent member states of the UN Security Council (P5 countries)22: US, Russia, China, UK and France played an important role in (de)stabilizing African governments over time.23 Whereas quantitative accounts revealed the impact of intrastate conditions to explain coup activity in a specific country. This discrepancy in the literature leaves the role of interstate relations on the occurrence of coups underspecified. This gap in the literature is however identified by Clayton Thyne, who strives to bridge the quantitative research (intrastate conditions) and qualitative research (interstate relations) in relation to coups. The scientific relevance lies within joining both types of literature in order to build a theoretical model to predict potential coup attempts as a function of interstate relations measured by hostile signals. Following Thyne, the contribution to the literature depends whether evidence proposed by the case literature holds across time and space and enables us to compare the impact of interstate relations with intrastate conditions identified in the quantitative literature.24 This objective contributes to narrowing the existing gap in the coup literature.
The empirical evidence mentioned earlier shows that African states vary in their vulnerability to coups based on the ability of plotters to organize and execute a coup.25 This thesis attempts to explain the variation in coup attempts in Africa, as a function of interstate relations. It focusses on hostile signals an African country receives from relevant actors in the international system. I attempt to fill the gap in the coup literature by empirically assessing a variety of international signals that can impact the stability of an African regime and the ability of the plotters to organize and execute a coup. This approach therefore recognizes the importance of international actors as relevant indicators to
18 Agyeman-Duah (1990).
19 For the sake of clarity, I prefer to classify the literature which focusses on international dynamics in relation to coups as progressive. I belief this is a valid distinction since a growing number of scholars emphasize the importance of external actors to explain coups. This tendency emerged since the 2000’s and therefore deviates from previous ‘traditional research’ which focus exclusively on national dimensions.
20 Thyne (2011), Powell (1990), Powell and Thyne (2011) and Agyeman-Duah (1990).
21 Thyne (2006), Lutwakk (2016) and Agyeman-Duah (1990). With exception of Gleditsch (2007) and Cunningham and Gleditsch (2009).
22 From now on: permanent member states of the UN Security Council referred to as: P5. In a similar vies, United States is US and United Nations is UN.
23 See Agyeman-Duah (1990) for a detailed account of foreign involvement in African coups. Due to declassified reports, detailed research emerged revealing UK and US involvement in the military coup against the Ghanaian government of Kwame Nkrumah in 1966. Similar research conducted by Ludo de Witte (2000) and David van Reybrouck (2010) show the proactive role of the US and its CIA to overthrow the first democratically elected prime-minister Patrice Lumumba of DRC in 1961 and install Mobutu Sese Seko as president.
24 Thyne (2011), pp.450. 25 Powell (2012).
7 explain the causes of coups. The value of this probabilistic approach to the causes of coups lies within the existence of the conditions that promote coups with the focus on interstate conditions rather than intrastate conditions. More specifically, this thesis tests a general theory which predicts, when a country receives costly hostile signals from international actors over time, increases the likelihood of a coup in the near future. The underlying logic asserts that international signals give rise to internal instability, which in turn encourage coup plotters to stage a coup.
The purpose of this thesis is to expand the scope of inquiry by incorporating the role of international factors in the coup literature. My general argument asserts that signals from international actors should have considerable impact on the likelihood of coups in Africa. The answer to the puzzle I put forward is that international actors play a decisive role in coup attempts. In the example of Eritrea, the absence of costly hostile signals from foreign powers, might explain why the country never experienced a coup.
In other words, this thesis attempts to explain the covariation that exists as described above, as a function of a hypothesized causal relationship between hostile signals and coup attempts. In doing so, this thesis builds on two main approaches: signaling theory and rational choice, to examine how international actors influence the likelihood of coups in Africa. Therefore, we arrive at the following research question: how international relations affect the likelihood of African coups between 1945 and 1999?
This thesis proceeds as follows. First, I discuss the coup literature to support the claim of a shift from intrastate to interstate conditions to address the causes of coups. Then I develop a theoretical framework by drawing on rational choice and signaling theory to explain coup attempts. Next, I lay out the research design of this study by using a nested analysis research design which combines quantitative and qualitative analysis. Thereafter, I run a statistical analysis to test the effect of the hypotheses. Subsequently, I examine an in-depth case study of Libya as a logical next step to complement the results from the statistical analysis. Finally, I draw a general conclusion and discussion of the overall study of this thesis.
Source: Powell (2012).26
Figure 1. Global instances of coups from 1950 until 2010.
26 Powell (2012).
8
2. Literature review
A profound body of the coup literature devotes its interest to the factors that determine the likelihood of coups. In a popular study conducted by Belkin and Schofer, the authors developed a structural conceptualization of coup risk. The authors conceptualize coup risk as “a function of deep, structural attributes of government, society, political culture and state society relations”. Of course, these structural conditions appeared as important determinants influencing coup benefits, but partly fail to explain the variation in coup attempts. Because traditional coup literature attributes the patterns of coup attempts to intrastate conditions, such as economic performance, regime legitimacy, regime type and internal stability. In this regard, international factors have largely been neglected in this scope of inquiry to explain the likelihood of coups. The first scholar who stressed the importance to abandon the traditional understanding of coups as an internal, isolated event whose cause and effect are limited to the country where it takes place, was Steven David.27 In a normative analysis, David proclaims that a considerable amount of coups have been influenced by external actors who have much to lose or to gain in their outcome.28 Several scholars followed David’s recommendation to widen the scope of inquiry to assess the relevance of interstate relations on coup risk as presented by Rowe, Bienen and Luttwak.29 This proposition is consistent with findings proposed by Kristian Gleditsch, who stresses the importance to acknowledge transnational dimensions of social phenomena like civil wars and coups.30
Since the late 2000’s more progressive studies emerged where scholars found well-established findings in the realm of international politics of coup risk. These findings show an important role of external shocks, international war and international signals in shaping a favorable climate to stage a coup.31 Agyeman-Duah for instance, provides convincing evidence that exogenous shocks as international war can trigger coups in the countries involved.32 From a similar approach, Thyne and Sikkink emphasize the ability of foreign powers to indicate their support for a potential coup.33 More specifically, Thyne developed a framework in which external actors send encouraging signals to potential coup plotters in advance, actions that lower the perceived costs of coup attempts.34 Thyne presents a quantified theoretical model to assess the impact of international signals on the cost-benefit analysis of coup plotters. Thereby exhibits the degree of the signal send – cheap or costly – to an incumbent government, which in turn influences the probability of a coup attempt.35 Foreign governments can indicate their support for a potential coup through material supplies to coup plotters to strengthen their capacity to oust the incumbent regime from power. In a post-coup setting – once the coup is successful – the coup plotters regime relies on international recognition as the new legitimate government to enhance political survival. In sum, the central argument in this thesis departs from the general assumption that international actors shape the ‘pre’ and ‘post’ phase of coup politics.
Recent empirical research conducted by Thyne shows a causal mechanism between US hostile signals and the likelihood of coups of Latin-America.36 Hostile signals sent by the US to an incumbent government of a Latin-American country has destabilizing internal effects. Thyne’s findings imply that hostile signals from the US shape an appealing environment for potential coup plotters to execute a coup, knowing that the US will likely be inactive to assist the incumbent government when a coup is underway, therefore increasing the likelihood the coup will succeed. In other words, at least by implication, a coup attempt against the government might be a desired outcome of hostile signals.
27 David (1986) p.4.
28 David (186).
29 Rowe (1974), Luttwak (1979), Bienen (1969). 30 Gleditsch (2002).
31 Tansey (2016).
32 Belkin and Schofer (2003, 2005). 33 Thyne (2010), Sikkink (2004). 34 Thyne (2011), Tansey (2016). 35 Tansey (2016).
9 This argument is consistent with Luttwak’s reasoning that the level of foreign political influence is often sufficed to prevent, oppose or consolidate a coup.37 Similar evidence is found by Powell, Lasley and Schiel, who strongly emphasize the impact of international organizations on coup plotters decision to stage a coup.38 This is remarkable, since the dominant school of thought of comparative analysis primarily consider a role for external actors on democratic processes, but not on coups. The argument developed in this thesis attempts to bolster the overall argument that hostile signals increase the likelihood of a coup. Further, I build on prior literature on international dynamics and coups to hypothesize that hostile signals increase the probability of a coup. The following section illustrates the ability of foreign actors to influence coups in the developing world.
2.1 The relevance of international actors
Since the process of decolonization, the Western superpowers multilaterally advocated a new foreign policy of deliberate support of democratic institutions on a world wide scale, with the final objective to end dictatorships in the developing world.39 The permanent member states of the United Nations Security Council (P5 countries), to a certain extent, share the general belief to oppose authoritarian regimes – especially those who undermine their strategic interests. Yet, there exists considerable variation in state behavior among the P5 countries how to promote regime change for the sake of democracy. The P5 countries have a long historical record of foreign interventions to influence political regimes and implement democratic structures in the third world. Since, the method of foreign intervention became rather controversial because it provoked fierce international criticism inasmuch it is in violation with international law, P5 countries found more subtle means to advocate regime change abroad. Besides foreign interventions, nation states obtain a broad arsenal of policy options to exercise pressure on foreign governments. A state can impose sanctions, engage in military interstate disputes or sent other hostile signals (e.g., closure of an embassy) to a recipient country.
P5 countries obtain the capabilities to exercise significant influence unitarily or multilaterally on a single African country. In the past the US unilaterally imposed economic sanctions on Somalia in the 1990’s which targeted specific individuals and governments for leading unconstitutional usurpations of power.40 These policies eventually led to the overthrow of Siad Barre in 1991. The P5 countries are widely perceived as the major super powers in the international system. The US, China and Russia’s influence in Africa continues to accumulate and therefore their incentives to protect their interests in the region. To a lesser extent, as former colonial powers, the UK and France still obtain severe political influence in Africa.
In addition, African countries are embedded in a highly interactive community of African states: the African Union (AU). Which implies that, similar to P5 leaders, African leaders are significant stakeholders in African politics. The important role of P5 countries and other African countries in domestic political affairs has been widely covered in the literature, and are therefore evidently relevant actors in the region with a certain degree of political influence on other African states. In general, the case evidence shows that coups are just too important to leave its course of action to chance by simply rolling the dice. Indeed, all P5 member states have attempted to deter, organize, back and suppress coups in order to strengthen its position in the developing world.41 In sum, external actors can play a key role in destabilizing foreign governments and serve as a reliable indicator to explain the causes of coups in the developing world.
37 Luttwak (2016).
38 Powell, Lasley and Schiel (2016). 39 Escriba-Folch & Wright (2010).
40 US sanctions policy in Sub-Saharan Africa; congressional testimony. Lyman (2016). https://www.usip.org/publications/2016/06/us-sanctions-policy-sub-saharan-africa 41 David (1986).
10
2.2 Coup attempts
It is evident that external actors play a pivotal role in inciting coups in Africa. The next step is to determine which stage of the coup is mostly related to international involvement. The coup literature is divided into two subcategories: coup attempts and coup outcome – success or failure. Although the literature examining the effect of external actors on coups is scarce, it nevertheless implies that the role of external actors is largely related to coup attempts rather than outcomes. From recent research we can infer that once a coup is launched, it usually succeeds or fails within hours.42 Thus, the short time period leaves no room for the international community to influence the outcome of the coup. The outcome of the coup is regularly determinant before anyone outside the capital realized that a coup has been attempted. This particular feature distinguishes coups from civil wars. Coups are often led by members of the elite and use a minimum degree of violence to seize power. In contrast, civil wars are defined as an armed conflict that involves military action against the national government by at least one insurgency group and requires that the state violence exceeds a certain threshold of thousand related battle deaths per year.43 In sum, the involvement of external actors in coups is most likely related to the attempt stage of the coup than its outcome.
2.3 The coup d’état and its relevance
This section explains the notion of the coup d’état and why it is still relevant as a form of political violence today. A coup d’état is an attempt by elite actors from the state apparatus to overthrow the incumbent of a country in order to seize the executive power.44 According to Luttwak, “a coup consists of the infiltration of a small but critical segment of the state apparatus, which is then used to displace the government from its control of the remainder”.45 The expression coup d’etat implies a direct ‘stroke of state’.46 In contrast to a revolution or a rebellion, which prerequisites a substantial military operation before it poses a credible threat to the political entity in power, the coup, as a rapid surprise strike, constitutes a more acute and unpredictable danger to the regime which – when successful – can cause a ‘stroke of the state’.47 The immediate threat comes from those inside the government who obtain the knowledge and skills, and have access to utilize the state apparatus against the incumbent. This definition excludes forceful seizers of power through civil wars, revolutions, mass demonstrations, foreign interventions and coup rumors and coup plots. As Walter Laqueur remarks in the preface of Edward Luttwak’s practical handbook of the coup d’etat: “much has been written under which (objective) conditions revolutions and civil wars take place, but almost nothing on coups”.48 This is surprising, since revolutions and civil wars are much less common forms of political violence in a world were coups enjoyed the most popular mode of political change in the developing world.49 As remarked by Decalo, who claims that “coups became the most visible and recurrent characteristic of the African political experience”.50 In addition, the unpredictable feature of the coup in combination with its swift and lethal outcome – when carried out successfully – poses the greatest threat to vulnerable regimes. Besides the immediate threat against the survival of the incumbent regime, coups can have destabilizing effects on a larger scale. The impact of coups on the domestic and international political level can be disastrous. Coups can result into civil war, revolution, mass protests, crimes against humanity or jeopardize democratization processes.51 These undesirable outcomes evidently
42 Thyne (2011).
43 Sambanis (2004) pp.817.
44 De Bruin (2017), Powell (2012), Thyne (2010), Powell & Thyne (2011), Luttwak (2016). 45 Luttwak (2016), pp.12.
46 Oxford English Dictionary (2010). 47 Roessler (2012), pp. 308.
48 Laqueur (1979), In Luttwak’s preface of the Coup d’état (1979). 49 Luttwak (2016).
50 Decalo (1990). 51 Thyne (2014).
11 pose a significant threat to national, regional and international stability. These propensities of the coup d’etat make it a valuable social phenomenon to fully understand its logic in order to better predict future coups since they contain ever recurring patterns – “the same always different” – from its initial conspiracy to the ultimately seizure of power.52 It is the task of the coup plotters to successfully segregate the political leadership from the military command to effectively seize power.53
2.4 Purpose of study
The literature review stressed the importance to shift the focus from intrastate factors to interstate factors to assess the effect of the latter on coup attempts. In order to achieve this objective I enlarge the scope of inquiry to focus primarily on coup plotters expectations for staging a successful coup as a function of interstate signals to the targeted country. This approach focusses on the effect of signals sent from external actors on the likelihood of coups in Africa. As mentioned by Thyne, the role of external actors on coup activity has foremost been ignored in the coup literature while the accumulation of empirical evidence continues.54
This however, does not imply superiority of one approach over the other – interstate conditions over intrastate conditions. Rather, I attempt to expand the scope of inquiry to combine intrastate and interstate conditions as proxy indicators for coup onset. In other words, intrastate and interstate conditions are complementary, and are therefore not in contradiction. Nevertheless, to test the effect of interstate conditions, intrastate conditions require a secondary role and will be used as control variables in the analysis. If the coup literature – and this thesis attempts to contribute to this objective – is able to combine both types of conditions into one general framework to explain the causes of coups, provides valuable information for scholars, policy makers and statesmen to anticipate a coup more accurately and act accordingly.
Furthermore, it is not the aim of this thesis to engage in a foreign policy analysis of the P5 member states in Africa nor to analyse the process of decolonization and its implications for coups. Rather, to demonstrate that on average, external actors play an important role in the likelihood of coup attempts in Africa between 1960 and 1999. The following paragraph narrows down the primary objective of this thesis.
In general, the causal argument proposed by Thyne – and proposed here – entails that hostile signals empower the decision making process of coup plotters to stage a coup and therefore increases the probability of coup attempt.55 Although Thyne draws general conclusions based on his model, his analysis has several limitations which could undermine his generalizable results. First, US and Latin-American relations are a unique set of cases due to their historical background and anti-colonial struggles against European powers. Second, other major powers might fulfill a similar role as potential signalers in destabilizing governments to protect national interests in the region (e.g., UK or Russia). When we change the actors of the theoretical model we might discover different mechanism of the role external actors play in the decision making process to execute a coup. This observation raises the question whether Thyne’s results hold over time and space when we include other actors – Russia, China, UK and France – in the analysis to test the theory and change the context – Africa, instead of Latin-America. In other words, can we validate Thyne’s theory when we exclude a unique set of cases (US-Latin-American relations) in our analysis and change its original context?
In addition, as mentioned by Bienen, Africa is one of the least studied areas in terms of the relationship between national and international systems and especially the linkage between international signaling and coup activity.56 The absence of Africa in previous research, provides an additional argument to select the continent as the subject of analysis.
52 Laqueur (1978), pp.36-37. In: Luttwak (1979). 53 Luttwak (1979).
54 Thyne (2010). 55 Luttwak (2016). 56 Bienen (1969).
12 Finally, the aim of this thesis is to develop and test a theoretical model for general application, which allows causal inference and comparative analysis of their precision and relevance.57 As a result, we improve our understanding of the relationship between international actors and coup risk measured by MID’s, sanctions and cheap hostile signals.
13
3. Theoretical framework
The purpose of this section is to propose a theory how international actors affect the likelihood of coups. This theoretical framework builds on two theoretical approaches: signaling theory and rational choice theory. This section proceeds as follows. I start by laying out the general framework of analysis, which is derived from rational choice theory in the civil war literature. The concept of rational choice proclaims rational expectations, which in turn reflect clear expectations how different types of international signals should affect the likelihood of coups.58 Thereafter, I explain the content of signaling theory and its implications for coup attempts.
3.1 Rational choice
Since we narrowed down the stage of the coup to be the focal point of inquiry – coup attempts – we can specify the mechanisms at play during the decision making process of coup plotters. The decision making process whether to stage a coup finds it origin in rational choice theory. The relevance of rational choice has been recognized in several fields within social sciences (e.g., economics and psychology) and has proven to be a useful framework to explain micro and macro level decision making processes.59 Coup plotters will therefore more likely stage a coup when they belief it will succeed. This assumption implies that when coup plotters decide to execute a coup the benefits outweigh the costs. The nexus between the civil war and coup literature is the widely shared assumption of rational choice to explain why individuals join an insurgency during civil conflict and why individuals launch or join a coup.60 From the logic of rational choice one can infer that the decision to stage a coup is determinant by the anticipated benefits of staging a successful coup and the perceived probability of success.61 These conditions are conceptualized by Thyne and depicted in table 1. The theory predicts, that the likelihood of coups increase when the perceived probability of success increases and the anticipated benefits to stage a coup increase.62 The assessment of coup plotters to stage a coup will be discussed in the following section.
Following rational choice theory, this thesis treats coup plotters as rational actors who calculate the expected utility of action by weighing the benefits of success against the probability and impact of failure.63 Departing from a cost-benefit analysis, we expect coup plotters to execute a coup based on an analysis of the contemporary political situation within a country compared with the anticipated situation following a successful coup.64 In such a situation coup plotters calculate the expected pay-off of a particular decision by weighing the gains of success against the consequences of failure.65 A failed coup attempt is most likely to result in execution or exile, while success coincides with a government position and corresponding privileges. Following a rationalist approach, coup plotters are most likely to stage a coup when they are convinced the incumbent regime is weak and when they believe foreign actors will support a successful coup. Support in this context has several facets which are all beneficial to coup plotters. The level of support from foreign actors varies from inaction to fierce condemnation of the coup or the lack of support to the incumbent regime to repel the coup plotters attack. Albeit these actions lack any legitimate support for the coup plotters, in practical terms the course of action from other states immensely affects the likelihood of a coup and its outcome.
58 Thyne (2009).
59 For example, governments will align with other governments when they believe such a decision will benefit their national interest. A similar line of reasoning can be ascribed to individual decision making process when individuals believe a specific action will result in potential gains.
60 Cederman (2004), Macartan and Weinstein (2008), Thyne (2006, 2011), Powell (2012) and De Bruin (2017). 61 Thyne (2011).
62 Thyne (2011).
63Powell, Lasley and Schiel (2016), Thyne (2010), Powell and Thyne (2012), Powell (2012). 64 Thyne (2011) pp.
14 It follows logically to frame inaction or indifferent state behavior from other states as supportive to the coup, because this course of action does not obstruct or hinder the process of the coup. Inaction from influential actors is essential for the coup to succeed. In other words, the indifference from other states sends a signal of indirect support to the coup plotters which implies that a coup would be tolerated. Namely, the broadest range of the spectrum of political action favors coup plotters rather than the incumbent government as depicted in figure 2. Figure 2 captures three possible outcomes of foreign state action as a response to a coup attempt: hostile towards the coup plotters, neutral (inaction) or supportive of the coup plotters. Due to the swift nature of the coup, an incumbent government depends on immediate support from foreign states in order to repel the coup plotters attack. A government which is confronted with a coup only gains support from the far left side of the spectrum above the critical line of -0.5. All state actions from -0.5 onwards is beneficial to the coup plotters. Similar to zero in roulette benefits the casino, zero (inaction) favors coup plotters. Coup plotters obtain the luxury position that they often prefer an inactive position of the international community, because foreign involvement can jeopardize success. In contrast, in order to stay in power aid dependent regimes require outside sources in terms of troop mobilizations or international condemnation of the coup.66 The political survival of these regimes depends on the political will of other states to protect the regime against an ongoing coup.
Figure 2. Foreign state behavior in reaction to a coup.
-1. -0.5 0. 0.5 1.
Hostile to coup plotters Neutral (inaction) Supportive to coup plotters
Critical measure of Critical measure of hostile action supportive action
In sum, according to rational choice theory, coup plotters are more likely to stage a coup when the benefits outweigh the costs. The influence of external actors on this decision making process can be decisive in triggering a coup. The following section exemplifies the notion of signaling theory following a rationalist approach.
66 Powell et al. (2016).
15
3.2 Signaling theory
This section explains the content of signaling theory in the context of international politics and its implications on coup risk. International relations covers a large spectrum of political interaction. Interactions between states occur on a day-to-day basis; which varies from events such as frequent trade agreements to rare events like military disputes. These interactions influence the likelihood of coups. Most importantly, interstate interactions act as signals to the coup plotters, which affect their estimation of success, if a coup were attempted.67 Gartzke defines signals as, “actions or statements that potentially allow an actor to infer something about the unobservable, but salient, properties of another actor”.68 Signals are therefore a product of interstate interaction, which in turn provides actors with information to judge the nature of the relationship. The underlying logic of signaling theory asserts that actors – individuals, groups or states – can assess the internal instability of a country as result of hostile signals and simultaneously can anticipate future behavior from the signaler.
State interaction therefore coincides with public or private statements which connotes supportive or hostile interaction. These signals might play a pivotal role in relation to coups, because when external actors advocate regime change or a continuation of power in an African country serves as a useful indicator whether the coup plotters will enjoy support or face opposition from influential foreign actors.69 If we accept the claim that signals – as a product of interstate relations – affect coup activity, we must further specify the notion of signals to develop testable hypotheses.
Theoretical models of game theory as applied in the interstate conflict literature shows us that the perception of credibility of the signal sent from external actors is the main denominator whether a coup will be attempted.70 This line of reasoning suggests a bargaining framework to explain civil war onset or coup attempts.71 I believe that a bargaining framework is a rather inadequate framework to explain the onset of civil wars and coups. These intrastate events often involve multiple actors that are seldom acknowledged or recognized as legitimate actors by the state (e.g., rebel leaders or rebel groups) and therefore rarely obtain a seat at the negotiation table.72 For example, when two states are on the brink of war, communication runs through formal chains of command with special representatives from both sides. In such a setting, the bargaining framework accurately captures all possible outcomes, and predicts that both states strive towards their preferable outcome which creates a bargaining range of mutually acceptable outcomes.73 In contrast, coups are orchestrated in secrecy and lack any process of negotiations between coup plotters and the incumbent government. The absence of two negotiation players rules out any form of a bargaining framework to explain the occurrence of these events.74 Rather, we can infer that coup plotters make a decision on the basis of the credibility of the threat, not to strike a deal with the government to arrange a mutually acceptable agreement. This might occur in the aftermath of a coup when it fails, then a mutual acceptable agreement is very appealing for the coup plotters to negotiate a safe passage. The coup outcome, however, is not the subject of discussion in this thesis. In other words, intergovernmental relations affect interstate bargaining not intrastate bargaining. In sum, relevant for the theoretical framework of this thesis is that rational choice explains the importance of the credibility of the threat as main indicator whether a coup will be staged, not to determine the preferable outcome between both parties.
67 Thyne (2009).
68 Gartzke (2003). 69 Thyne (2011).
70 See for example Jervis (1988) or Fearon and Laitin (1996). 71 Morrow (1989), Thyne (2006) and Powell (1999). 72 Autesserre (2009).
73 Fearon and Laitin (2003).
74 A bargaining framework is inadequate to explain the onset of coups and to a lesser extent to civil war onset. If a coup fails, negotiations often take place to arrange a peaceful settlement between the coup plotters and the government. In this stage of the coup, a bargaining framework is useful to predict coup outcomes. This stage of the coup is however not the focal point of this thesis and is therefore neglected in the theoretical framework.
16 In relation to rational choice, the credibility of the signal shapes the expected effect of the signal on the coup plotters decision to stage a coup. According to rational expectations, coup plotters respond not only to contemporary circumstances but also to expectations for future events. In other words, coup plotters extrapolate the past and present course of state relations to the future to predict future behavior from external actors when a coup will be attempted. Rational expectations therefore enable coup plotters to infer current and future behavior of external actor’s. For example, the continuum of hostile signals from the US, the UK and Belgium to Lumumba’s government of the Congo, allowed Mobutu to realize that the Western governments opted for regime change due to Lumumba’s sympathizes with the Soviet Union. As a result, Mobutu rightfully anticipated that the Western governments would support a coup against Lumumba, to curb Soviet expansionism in central Africa. This example clearly illustrates the hostile positions of external actors which was signaled in a predictable manner prior to the coup. Mobutu’s final decision to stage a coup and to claim absolute power came as no surprise. The costly signals imposed on Congo – economic sanctions and troop mobilizations – created the perfect environment for Mobutu to stage a coup. One could argue that without the role of external actors it was rather unlikely that Mobutu would have been able to execute a coup. The effect of the cost-benefit analysis of costly signals resulted in rational behavior from Mobutu to stage a coup is outlined in table 1.
Interstate relations affect the probability of a possible coup because they signal information about an external actor’s likelihood of supporting either the coup plotters or the government to repel a coup when it is underway.75 From a coup plotters point of view, signals must enjoy a certain level of commitment from the signaler to convince the coup plotter to perceive the signal as credible before they launch a coup. A signal that lacks any credibility would not convince a coup plotter to execute a coup. To scrutinize the effect of a variety of signals from international actors on coup risk enables us to better understand how international signaling influences the likelihood of coup attempts.
The credibility of the signal is of vital importance to the coup plotters. To treat signals as an aggregate of all signals sent by external actors, leaves us unable to infer any conclusions regarding the degree of credibility of the signals. To improve our understanding of signals and coup risk, requires a disaggregated analyses of the various signals sent by external actors. In other words, a disaggregated approach results in a detailed analysis of how variations in hostile signals influence the likelihood of coups in Africa. In order to assess the credibility of a signal, Thyne distinguishes signals between ‘cheap’ and ‘costly’ signals.76
Cheap hostile signals can be defined as “day-to-day communications across borders that come with little costs to the signaling state”.77 These can be characterized as verbal criticism in terms of public statements of an incumbent government or the withdrawal of a high level diplomat. Other examples are the closure of an embassy or threats to sanction the recipient state.78 According to Fearon, “these signals are cheap because they do not require significant investments to reinforce a country position”.79 In this regard, cheap hostile signals lack the assurance of future support from foreign governments if a coup were attempted due to its uncertain signal if it sides with the government or potential coup plotters.80 This is the main difference between cheap and costly signals. Both signals have destabilizing effects on regime stability, but cheap signals are less ‘costly’ for the signaler to reflect its position.
Costly signals reflect the clear and direct position from the external actor and bear an expensive cost to initiate and preserve.81 In addition, costly signals are more transparent and stable over time as observable flows of information from one country to another. To illustrate, military disputes - which can be observed as troop mobilizations - and economic sanctions can be defined as 75 Thyne (2006, 2011). 76 Thyne (2006) pp.939-940. 77 Thyne (2009) pp.27. 78 Thyne (2009). 79 Fearon (1997). 80 Thyne (2009). 81 Thyne (2010).
17 costly signals, because these are financially expensive and transparent. Costly signals are largely perceived as credible signals because they have a stronger impact on the targeted country and reflect a higher level of commitment of the signaler due to the real costs it brings along.82 The following section will provide a more detailed description of each hostile signal – MID’s, sanctions and cheap hostile signals.
Table 1. Decision making process of staging a coup. Source: Thyne (2011).83 __________________________________________________________________________________
Perceived probability of coup success (P)
_______________________________________________________
Low High
___________________________________________________________________________________________ Anticipated benefits of Low (-) Potential gain: low (-) Potential gain: low
staging a successful coup (-) Interstate signals: supportive (+) Interstate signals: hostile (-) Prediction: low probability of (+/-) Prediction: moderate coup attempts probability of coup attempts ___________________________________________________________________________________________
High (+) Potential gain: high (+) Potential gain: high (+) Interstate signals: hostile (+) Interstate signals: hostile (+/-) Prediction: moderate (+) Prediction: high probability of probability of coup attempts coup attempts
Figure 3 reflects the expected effect of signaling theory in a rationalist framework on coup risk. The figure simultaneously shows different degrees of hostile signals.
Figure 3. Variation in coups
1. Hostile signals Credibility 0. 0. Coup risk 1.
82 For example, if the US imposes economic sanctions on Nigeria will result in economic loss for American investments in the region and are therefore classified as a costly hostile signal to the Nigerian government.
83 Thyne (2011) pp.450. Low (Cheap) Public statements of condemnation, closure of embassy High (Costly) Military interstate disputes (MID’s), sanctions Low probability of a coup Medium probability of a coup
18
3.3 Indicators of hostile signals: MID’s, sanctions and cheap hostile signals
The previous section outlined the general expectations how hostile signals from external actors impact the likelihood of coups in the recipient country. This section provides a detailed description to operationalize the indicators of hostile signals: MID’s, sanctions and cheap hostile signals. The descriptive analysis ends with a formulated hypotheses for each indicator. Before I discuss the different hostile signals, I turn to a methodological issue for clarification purposes.
A quantitative analysis to assess the effect of external actors on coups raises a methodological issue. The clandestine nature of coups and its violation of international law to illegally overthrow a foreign government, stimulates external actors to keep their involvement secret with the possibilities of false accusations of complicity by international condemnation.84 As a result, most interactions occur through obscure flows of information through cables or face to face meetings between foreign government officials and coup plotters. Such a dataset is of course not available to academics. Consequently, a quantitative analysis is restricted to observable transparent indicators of interstate relations on coup risk. These indicators are MID’s, sanctions and cheap hostile signals and will explained in the following section.
3.3.1 Military interstate disputes (MID’s)
The first indicator used by Thyne to operationalize costly signals are MID’s. Thyne defines a military interstate dispute (MID) as “an international interaction involving a threat, display, use of military force, or war”.85 Importantly, “these interactions are explicit, overt and authorized by the government and illuminate transparency to the international community”.86 Thyne uses two variations of MID indicators in his model. The first MID indicator contains the total number of MID’s a country is engaged in, by any state during each state-year.87 The second MID indicator is restricted to neighboring countries only.88 It seems plausible that when political instability rises in a country, neighboring countries become more involved than peripheral states to influence the political situation to a desired outcome. For example, following this argument, we expect Angola to be more actively involved than Kenya when political instability rises in South-Africa.
According to Thyne, MID’s accurately capture costly hostile signals because they compel the signaling state to engage in financial and political investments to credibly signal its stance.89 Recent events in Africa clearly illustrate this logic. Since the late 2000’s, Rwanda, Tanzania, Zambia and Angola hosted several conferences to discuss the political instability in the Democratic Republic of the Congo (DRC). Because Joseph Kabila’s ambitions to remain in power for a third term in office can result in destabilizing effects to neighboring countries. As a result, these African countries mobilized troops along the DRC’s border and threatened to intervene if the political situation further deteriorates. Following Thyne’s argument, a coup is mostly likely to occur if this situation results into a MID, when signals become costly to maintain stability within Congolese politics. The underlying logic entails that coup plotters judge these signals as credible and calculate potential foreign support if a coup were attempted against the Kabila government. This expectation results in the following two hypotheses:
H1a: Military interstate disputes with all countries increase the likelihood of a coup.
H1b: Military interstate disputes with neighboring countries increase the likelihood of a coup.
84 David (1986).
85 Thyne (2007) pp.94.
86 Thyne (2007, 2010), Gochman and Moaz (1984). 87 Thyne (2006, 2007).
88 Thyne (2010). 89 Thyne (2006, 2010).
19
3.3.2 Sanctions
The second indicator to operationalize hostile signals are sanctions. In contrast to MID’s, are sanctions a frequent used tool in international politics to stimulate political change in foreign countries.90 Over time, the international community witnessed a substantial increase of the scope and frequency of sanctions imposed on foreign governments.91 Sanctions are often imposed when a targeted regime violates international law, with a clear intent to destabilize the executive to revise or revoke certain policies and conform to international standards of government.92
The general assumption is that sanctions eventually can result in the overthrow of an incumbent regime. Indeed, as argued by Robert Pape, regime change can be the principal reason of those implementing the sanctions.93 This line of reasoning clearly puts forward that sanctions share a clear purpose, namely a deliberate instrument to push for regime change as a desirable outcome. Thus, as argued by Pape, sanctions can be an effective policy to remove a head of state of a specific country.94 This argument is consistent with Marinov’s conclusion that sanctions are a highly efficient method to undermine foreign governments.95 Other research corroborates this finding and proclaims that the effect of sanctions is most effective when implemented against internal instable regimes which lack democratic institutions and suffer economic backwardness.96
Escriba-Folch and Wright, for instance show that international sanctions have destabilizing effects on regimes in Africa. The authors proclaim a conditional effect of international sanctions mediated by regime type in the target country.97 Accordingly, there exists a causal relationship between international sanctions and the destabilization of authoritarian regimes.98 Because personal regimes are more susceptible to the deprivation of foreign assistance to fund their security units and their cronies who ensure the political survival of the regime. The majority of African regimes build supporting coalitions whose loyalty is largely dependent on financial compensation and political privileges provided by the incumbent regime. In order to maintain the loyalty of these groups requires a constant allocation of benefits to the coalition. Central in the literature is that a decrease in the government’s budget as a result of international sanctions increases the likelihood of a coup or an insurgency because the incumbent loses the backing of the coalition which protects him.99 If political developments take such a course of action the economic payoff to the coalition can be higher under an alternative regime, which triggers the potential support for a coup. More specifically, and particular in authoritarian regimes, political factions will remain loyal to the regime if the payoff from supporting the ruler exceeds the expected payoff of a risky coup plot.100 Political leaders of these regimes are more likely to be destabilized by sanctions than leaders of other regimes. Consequently, sanctions increase the probability of a legal or illegal course of regime change through coups or an insurgency. As a result, international sanctions shape the vulnerability of the regime and produce rational incentives for political actors to defect.
For example, western countries effectively destabilized the government of Idi Amin in Uganda by imposing economic sanctions through a boycott on coffee exports for obtaining foreign exchange. According to Escriba-Folch and Wright, Amin relied too heavily on foreign exchange to fund his patronage, who protected his regime. Eventually Amin was ousted in 1979 by Tanzanian forces. It would be an overstatement to claim that the economic sanctions resulted in Amin’s downfall. Nevertheless, one could argue that the economic sanctions brought Uganda in economic turmoil which
90 Escriba-Folch & Wright (2010). 91 Escriba-Folch & Wright (2010). 92 Escriba-Folch & Wright (2010). 93 Pape (1997).
94 Pape (1997), Powell et al. (2016). 95 Marinov (2005).
96 Lam (1990). 97 Lam (1990).
98 Escriba-Folch and Wright (2010). 99 O’Kane (1981).
20 made it impossible for Amin to finance the forces who protected him. Despite the fact that Amin was ousted through a foreign intervention, the economic sanctions clearly signaled the weakness of the regime, constrained its resources and maneuverability, subverted its support and empowered opponents of the regime.101
In sum, the dependence on external resources makes authoritarian regimes vulnerable to international sanctions because it undermines the conditional support of coalition forces to protect the regime. A reduction in the revenues of external resources decreases the maneuverability of the incumbent to maintain his clientelist network which safeguards the political survival of the regime and increases the expected utility from defection of the former. This expectation results in the following hypothesis:
H1c: Economic sanctions increase the likelihood of a coup. 3.3.3 Cheap hostile signals
The third indicator to operationalize hostile signals are cheap hostile signals. The literature is quite decisive that costly signals have a stronger effect on coup risk than cheap signals. Nevertheless, in a similar view, cheap hostile signals have destabilizing effects on foreign governments. These include day-to-day interactions and occur therefore more frequently than costly signals which exhibit more rare events. For this reason, a considerable amount of information is captured by cheap signals. Though these are rather unpredictable over time, and acquire limited explanatory power compared to costly signals, they nonetheless provide useful information.102 Cheap signals include statements of criticism and accusations or threats to another government.103
An African government which faces fierce criticism on its policies from the international community, may be an incentive for potential coup plotters to detect and expose the regime’s vulnerability. Regime vulnerability enables coup plotters to organize and execute a coup conspiracy.104 Potential coup leaders could utilize a political situation wherein the incumbent regime encounters criticism to alter the status quo in their advantage.105 External threats offer a window of opportunity for coup leaders, because they can rely on a potential alley if a coup were attempted. Cheap signals therefore may play in important role for coup plotters to estimate the probability of future support from external actors. At the very least, cheap signals will contribute to the coup plotters perception of staging a successful coup due to the increased support which resonates through public statements from external actors.106 In sum, although cheap signals bear less weight than costly signals in their credibility it makes a coup a likely prospect. This results in the fourth hypothesis:
H1d: Cheap hostile signals increase the likelihood of a coup.
The estimation of the relationship between hostile signals and coup attempts is captured by the following model:
Coup attempt i = xi + β1 MID’s with all countries + β2 MID’s with neighbors + β3 Sanctions + β4 Cheap hostile signals + βj xj + εij
101 Kirshner (1997), Escriba-Folch and Wright (2010). 102 Thyne (2006, 2009).
103 Thyne (2006). 104 Luttwak (2016). 105 Tansey (2016). 106 Thyne (2009).
21
3.3.4 Control variables
The traditional coup literature discussed earlier shows significant results of structural factors on coup risk. In order to control for the effect of these variables we must explicitly measure them in the analysis. To avoid omitted bias, I will include these variables as controls in the analysis. The literature found strong effects of regime instability, democracy, authoritarian regime, military regime and GDP per capita. These variables will be extensively discussed in the case study section, so I surpass them here.
Figure 4. Expected effect of hostile signals on coup probability (P) an African state
0 2 4 6 8 10 12 14 16 18 20 C o up risk Hostile signals
