Abstract
This article argues that coup-proofing does not necessarily reduce the general propensity among military officers to stage coups d’état against authoritarian incumbents. Sophisticated coup-proofing terminates coup cascades and buys incumbent rulers time in office, but general coup risk remains high even if they maintain power for long periods of time. The article uses a new data set on coups in the Middle East and North Africa covering the period 1950–2013. The data reveal that the number of coup instances has decreased over time, but the risk for incumbents of falling to a coup during their extended time in office remains constant, even in coup-proofed autocracies. When autocrats apply coup-proofing measures, timing and agency become increasingly important. Plotters attempt coups either immediately after power turnovers or when incumbents turn into lame ducks after excessively long periods of rule. Success rates increase when coup plotters are supported by central military leaderships.
Coups d’état are a classic indicator for measuring a military’s involvement in politics. For Comparative Politics in general, regime change and democratic transition inspire the focus on civil–military relations in nondemocratic regimes. The underlying puzzle asks when and why armies would return to the barracks and allow for the rise of elected, accountable political officials. With political regimes—including autocracies and new democracies—increasingly institutionalizing throughout the late twentieth and early twenty-first centuries, coups d’état seem to have become a phenomenon of the past. Conventional wisdom has it that the consolidation of those regimes includes the successful application of coup-proofing by civilian incumbents, designed to provide disincentives for power-hungry officers.
This article finds that coup-proofing in authoritarian regimes is only partially successful in that it reduces the number of coups and coup attempts. Yet, general coup risk remains high as long as authoritarian rule persists. In the absence of regular, competitive, and constitutional succession of incumbents, coups remain attractive for politically ambitious elites and officers as a vehicle for assuming power. 1 Coup-proofing works to buy incumbents time in office but does not reduce the general risk of falling to coups.
One prominent argument in the more recent debate maintains that coup-proofing can diminish the triggering causes of coups, while leaving structural underpinnings for a military’s involvement unaltered. 2 The argument is intuitively compelling, but difficult to trace empirically. How can we account for general coup risk at a time when the number of coup instances decreases owing to the successful application of coup-proofing measures? A simple example illustrates the puzzle: in the period 1960–2010, no single coup attempt was reported in either Egypt or Germany; yet, most scholars would agree intuitively that general coup risk was higher in the former country than in the latter, even before the Egyptian military’s interventions in politics in February 2011 and July 2013.
Coups d’état are “illegal and overt attempts by the military or other elites within the state apparatus to unseat the sitting executive.” 3 Coup risk does not necessarily indicate the weakness of authoritarian regimes, but rather the likelihood of regime incumbents to experience coup attempts. The likelihood for coups increases when opportunities exist and/or when military or political elites breed increasing motivations to oust the incumbent. Coup-proofing contributes to the consolidation of authoritarian regimes. Yet, it reduces coup risk for incumbents only insofar as opportunities for coup plotters taper off, while incentives and motivations to stage coups remain high.
The aim of this article is two-fold. I use a new data set on coups in the Middle East and North Africa (MENA) to show that coup-proofing does not reduce general coup risk for regimes and incumbents. Using data from the MENA region allows for a cross-time comparison of coup-proofed autocracies (CPAs) with authoritarian regimes vulnerable to coup cascades. A test of authoritarian coup-proofing is facilitated by the fact that authoritarianism has remained a persistent phenomenon, while in all other world regions the absence of coups might as well be the result of democratic consolidation. Second, I discuss the conditions under which incumbents in CPAs become vulnerable to coups. Their successful execution becomes increasingly conditional on two factors: timing and the organizational resources of coup agents. Coup-proofing helps to institutionalize militaries and hence incentivizes coup plotters: first, to introduce centralized military leaderships among their ranks and, second, to execute coups shortly after the accession to power of a new autocrat or when he turns into a lame duck after a long period of rule.
Theorizing Coup Risk in Authoritarian Regimes
A theoretical point of departure is to study CPAs in the context of authoritarian power maintenance. Scholars suggested that authoritarian regimes consolidate as a result of their adaptive capacities. Autocrats would arrange their coercive capacities, 4 craft ruling coalitions and regime parties, 5 establish political institutions, 6 and frame contentious relations with domestic opponents 7 in a way conducive to regime resilience. The assumed consolidation of authoritarian regimes remains a contested subject, mainly since some of the seemingly most durable autocracies have collapsed, such as during the breakdown of the Soviet bloc or the Arab Spring. I therefore refrain from studying authoritarian regime consolidation. I am more interested here in the strength and weakness of incumbencies, that is, the ruling coalitions in authoritarian regimes rather than their institutional infrastructures.
In mature autocracies, sophisticated coup-proofing is part of a larger menu of strategies applied to keep incumbents in power in that it helps to balance intra-elite interests and thus replace “contested” with “established dictatorships.” 8 Coup-proofing includes the establishment of loyalties between officers and incumbents through ethnic, religious, and personal bonds; the recruitment of military personnel from among privileged minorities and mercenary soldiers; the counterbalancing of divided security apparatuses; the frequent rotation of officers to avoid the emergence of alternative power centers; and buying off the officer corps through economic privileges and opportunities for self-enrichment. 9 The crafting of alliances with international powers, including the stationing of foreign troops, also helps to avoid coups since plotters would have to assume that status quo oriented foreign powers would stand by their allies. 10
Autocrats have an even stronger interest to adhere to power than their democratic counterparts, mainly since their own personal fate is often grim when they fail. 11 Among a whole variety of different challenges, coups d’état are the most eminent threat to an autocrat who will develop a preference for coup-proofing over any other power-saving strategy. Yet, coup-proofing remains a double-edged sword and carries serious risks. The preference of incumbents to avoid coups d’état at any cost often leads to the establishment of identity-based militaries, ethnic exclusion, and consequently increasing risk of civil war. 12 Authoritarian rulers’ anticipation of the dire consequences of losing office through coups might lead them to employ extraordinary measures, such as diversionary wars or violent repression. 13 Coup-proofing can also reduce military effectiveness, with detrimental effects on combat capacities 14 and on the military’s deployment against domestic unrest. 15
There is also reason to assume that coup-proofing, irrespective of its sophistication and duration, does not guarantee the complete absence of coups d’état. Mehran Kamrava reminds us that military “professionalization enhances the autonomy of the military and, if politically unchecked, can increase its tendency to intervene in the affairs of the state.” 16 Coup-proofing—along with other power-saving strategies—serves to reduce a regime’s imminent vulnerability. 17 Using the terminology of Samuel Finer, successful coup-proofing reduces the opportunities of coup plotters, but not necessarily the general disposition of military personnel to intervene in politics. 18 Hence, the absence of coup attempts indicates the successful application of coup-proofing in so far as opportunities are reduced—declining imminent coup risk—but not coup risk as a general propensity of officers to target authoritarian incumbents.
It is rational for military elites in authoritarian regimes to maintain a credible coup threat. At times, officers even resort to formulating explicit warnings in order to deter incumbents from diverting resources pledged for the military. 19 In the absence of electoral contestation, the threat of insubordination has a disciplining effect on political elites who are encouraged to maintain established distributive practices. 20 General coup risk therefore remains a constant danger for regimes as long as nondemocratic rule persists, and irrespective of the extent of coup-proofing measures. This contention turns the puzzle expressed in Zoltan Barany’s seminal study on its head: 21 rather than wondering how a military force would be crafted to allow for democratic transition, it is consolidated democracy and the accountability of incumbents to their citizens which will ultimately keep armies in their barracks. 22
Measuring coup risk is difficult when coup instances are rare or absent over an extended period of time. Moreover, the literature on coups d’état lacks agreement on the relative weight of potential causal variables, including endogenous factors (e.g., regime type, military grievances, and organizational cohesion) and exogenous factors (e.g., economic crises, civil society activism, and prior instances of coups). In order to account for general coup risk, I propose to depart from simply counting the number of incidents of coups over time. Rather, I measure the likelihood for executive incumbents to experience a coup attempt during their terms in office.
I define coup cascades as rapid sequences of multiple coups d’état, which allows for limited periods of tenure between turnovers of up to three years. Incumbents will usually have stabilized ruling coalitions within three years of takeover. Coup cascades make a good indicator for weak ruling coalitions. New incumbents remain vulnerable to challenges from army officers and elite segments, particularly in the early stages of their ascent to power. Succession constitutes the moment in time when ambitious members of the political and military elites feel sidelined and might act to take power themselves. Coup-proofing diminishes coup cascades. Coups do not disappear but remain isolated incidents.
Scholars of civil–military relations have largely ignored the effects that coup-proofing has on the political capacities of militaries—beyond the implicit assumption that it will keep them out of politics. While authoritarian regimes acquire adaptive capacities in their quest for power maintenance, potential coup plotters do too, and they learn from past errors. Upon thorough consideration, coup-proofing endows militaries with additional organizational resources. 23
It leads to the growth and institutionalization of armies which enhances hierarchical chain of command, task cohesion, and functional separation of units in the coercive forces, including armies, praetorian guards, domestic security forces, and intelligence services. Hence, superior coordination and organizational capacities are required for coups to succeed. Coup plotters refrain from attempting coups when they perceive that success is unlikely. Coup-proofing therefore incentivizes plotters to include the military leaderships among their ranks, because it is the upper echelon in hierarchical apparatuses that commands over sufficient information and superior organizational capacities.
Coup-proofing measures can have paradoxical effects: while they reduce the number of opportunities for coups d’état, an investment is made in the organizational capacities of potential coup plotters. The result should indicate a positive effect on the likelihood of success whenever a coup is executed.
Authoritarian incumbents’ strength and popularity are difficult to measure—let alone the military officers’ perceptions of incumbent strength. I employ the duration of incumbents’ tenure as a proxy variable. 24 Timing is crucial for coup plotters and “captures variables that in practice we cannot measure independently, such as leaders’ ability to build political networks and to acquire and use information.” 25 Scholars have provided overwhelming evidence that authoritarian incumbents are most vulnerable in the immediate aftermath of their takeover of power. 26 The reason is that they have not yet reconfigured elite coalitions. Scholars also found that ruling coalitions enjoy greater stability over the medium term than newcomers: “initially an autocrat’s coalition is relatively unstable since members fear exclusion. However, as the learning process continues, it becomes increasingly unlikely that supporters will be replaced, and so their fear of exclusion diminishes and the loyalty norm strengthens.” 27 Accordingly, “as a rule the longer each chief executive stays in power, the more power will be concentrated in his or her hands.” 28 This explains why the rate of coup-induced turnover declines with an autocrat’s extended term in office. 29
Yet, while autocrats enjoy considerable mid-career job security, most existing studies fail to recognize that incumbents remain vulnerable to decreasing popular and elite support in the case of an excessively long period of rule. Such leaders turn into lame ducks and are likely to face popular opposition for having accumulated a track record of broken promises, political and economic crises, and repressive treatment of political rivals and ordinary people. Lame duck autocrats also, more often than not, face political elites engaging in power bargains, possibly in opposition to the chosen successor. Military officers in particular will find their corporate loyalty for a lame duck autocrat put to a stress test when such a succession crisis correlates with social and political upheaval which would necessitate an active engagement of armed forces to save the incumbent. 30
General coup risk will have decreased significantly only when the following two conditions are met: (1) officers remain in their barracks, even (2) if incumbents are weak. If, first, militaries have been exposed to sophisticated coup-proofing, and, second, general coup risk remains high, officers analyze the strength and support base of incumbents. They refrain from challenging rulers at a time when those have a solid grip on power. Beyond the immediate aftermath of the takeover of power, coup plotters will be patient to wait for the moment to come when incumbents turn into lame ducks. In order to test a coup-proofed military’s general propensity to stage coups d’état, I therefore reemphasize the assumption of Bienen and van de Walle “that there would be some threshold, say two or three years, after which risk [of the ousting of incumbents] decreased only to rise again later, thus producing a pattern that could be graphed as a U-shaped curve.” 31
The authors of this study did not find proof in their empirical material to verify their hypothesis. Rather, they found that the risk of losing office declines with age, which could be depicted in an L-shaped curve. 32 Yet, their study covers the period 1801–1987; that is, a large part of their empirical observations of coups d’état occurred in new and instable regimes, most of which had to deal with difficult state-building processes. The effect is the selection of a universe of cases the majority of which witnesses rapid sequential turnovers, with very few observations of long-serving autocrats. Within the whole universe of autocracies—and among a sample as large as the one, for instance, by Jennifer Gandhi comprising 558 observations 33 —the number of CPAs is certainly underrepresented. I assume that the results of a sample of observations isolating coups d’état in consolidated autocracies will confirm the Bienen/van de Walle study’s original hypothesis and hence depart significantly from its empirical findings.
Jonathan Powell provides evidence that organizational capacities of coup plotters—namely the size of military apparatuses—are decisive for coup outcome. 34 I support this contention and introduce an explanatory variable showing how coup plotters adapt to coup-proofing: my agency variable determines whether coups d’état are executed by the central military leaderships or factions of military apparatuses, most often consisting of low-ranking officers. Significant differences prevail between military leaderships and low-ranking officers concerning the respective agents’ control over economic resources, the degree of corporate grievances, and organizational cohesion. Given the hierarchical organization of military apparatuses, it is also compelling to assume that military leaderships are more successful than army factions in the performance of coups. If militaries are vulnerable to coup-proofing, I expect to see an increasing number of coup attempts executed by the core military leaderships, rather than subordinate units.
Scope Conditions, Data, and Method
With its proven track record of authoritarian regime endurance, the MENA region provides a valuable field for the comparison of coup-proofed versus unconsolidated autocracies. Scholars of MENA politics have analyzed the successful application of coup-proofing strategies which—since the late 1970s—contributed to the consolidation of authoritarian incumbencies. 35 Middle Eastern rulers seemed to have successfully solved the major puzzle of authoritarian power maintenance: 36 the “principal-agent problem,” that is, the paradox of establishing coercive power designed to support the incumbency, while avoiding that this power within the state turns into a very threat for the incumbency. Scholars of Egyptian politics, for instance, maintained that coup-proofing led to civilian control, firmly established over a military that “refrains from intervening in politics, has re-professionalized, has allowed for the civilianization of the regime, and has succumbed to the control of civilians.” 37 It was assumed that “Mubarak’s close association with the military has rendered Egypt ‘coup-proof.’” 38 The ousting of Hosni Mubarak through a military coup on February 11, 2011, and the coup against Mohamed Morsi on July 3, 2013, begs for new research on contemporary civil–military relations in the MENA region that has only just begun to surface. 39
Concerning the identification of coup instances, I do not emphasize differences among singular coup instances concerning the agents (higher vs. lower officers; regime outsiders vs. insiders, and sometimes incumbents’ family members), duration (some coups take only hours, others several days or even lead to a protracted standoff), and the intrinsic motivations of coup plotters. As to coup trajectories, I follow Powell and Thyne who coded successful coups d’état when the successor assumed executive office for at least one calendar week. 40 In a few instances—such as in Egypt 1952, Turkey 1960, and Tunisia 2011—successful coup plotters have used their powers to invite a new executive office holder from outside of their own ranks.
My data set on military coups d’état in the MENA region (n[total] = 89) covers the time period 1950–2013 and is exhaustive on the number of successful coups d’état (n[suc] = 44) (see Appendix B). Observations of unsuccessful coup attempts (n[fail] = 45) use data by Powell/Thyne 41 and the Center for Systemic Peace (CSP). 42 I excluded a number of cases coded in the Powell/Thyne and CSP data sets whenever I could not find sufficient information on the identity of the coup plotters; in turn, some of the CSP observations did not match the employed definition and could not be clearly distinguished from other forms of military insubordination, such as assassinations and mutinies. I included a small number of observations when I found sufficient empirical evidence that a coup attempt may have taken place (e.g., Jordan 1972, Iraq 1996, and Tunisia 2011).
In order to study the effects outlined in Hypothesis 1, I compiled a list of all executive office holders in the MENA region for the period 1950–2013 (n = 137) (see Appendix A). The list includes presidents, prime ministers, junta leaders, kings, emirs, and revolutionary leaders. The unit of analysis here is the universe of individual incumbents, which—when paired with observed coup instances—allows for an assessment of the likelihood of executive office holders to becoming the target of coup attempts. To respond to Hypothesis 2, I coded two dichotomous variables: first, the success of coups d’état; and second, agency. As to coup agents, I distinguish between those coups planned by the central military leadership, and others executed by subordinate military units. Centrally planned coup attempts happen when the ministers of defense, chief of staff, or a minimum of two junta members (e.g., “Revolutionary Command Councils” in postcolonial military regimes) are among the coup plotters.
The main method employed in the empirical discussion is a cross-time comparison of two historical periods, which has particular value for detecting changes in the broader patterns of the phenomenon under investigation. 43 The approach helps to compare CPAs with those authoritarian regimes where incumbents and ruling coalitions have remained structurally unstable. In my comparison, the first period (P1), 1950–1979, covers the postindependence, state-building era. Concerning the relationship between incumbents and officers, the 1950–1979 period was the “golden age of coups in the Arab world” whereas, in the post-1980 period, “governments did have a great deal of success in preventing their armies from intervening in politics.” 44
While the phenomenon under investigation here shows a gradual decline of coup incidents since the second half of the 1970s, the selection of 1980 as a cutoff year is useful for methodological purposes and adopts the suggestion of James Quinlivan. 45 In most MENA countries, coup-proofing became effective from the mid-1970s onward owing to substantial amounts of financial resources available to authoritarian incumbents. Almost all coup-proofing strategies require a material investment; and investments were facilitated by increasing oil rents. They started to flow into the region from 1973 onward. Rent income was not monopolized by the oil-rich states on the Arabian Peninsula and in North Africa but was rather distributed across the Arab world mainly through the diffusion of budgetary assistance, political rents, and workers’ remittances. 46 While the diffusion of material resources was unequal, every country in the region had increased resources at the disposal to undertake significant coup-proofing measures by the turn of the decade. I also do not find evidence that the end of the Cold War had a significant impact on the likelihood of coup attempts. MENA data reveal that the number of coup incidents had already declined by 1990 and stayed relatively constant ever since.
Measuring Coup Risk in Middle Eastern Autocracies
It is obvious that the number of coups d’état in the MENA region has decreased significantly over the past forty-five years. Figure 1 shows that the period from the late 1950s to the mid-1970s saw a peak in coup incidents; and their steady decline since 1965 lends support to widespread assumptions on the effects of coup-proofing: the decline of observable incidents. In 1950–1979 (P1), sixty-two coups were observed, twenty-nine of which were successful; 1980–2013 (P2) saw a total of twenty-seven coup attempts, fifteen of which triggered the replacement of incumbents.
Coups d’état in the Middle East and North Africa, 1950–2013.
Yet, some indicators support the assumption that general coup risk in the MENA region has remained higher than implied in longitudinal data. In P1, a total of eighteen countries experienced coup attempts: Algeria, Egypt, Iran, Iraq, Jordan, Lebanon, Libya, Mauritania, Morocco, Oman, Qatar, Saudi Arabia, Sudan, Syria, Turkey, United Arab Emirates, Yemen Arab Republic, and the People’s Democratic Republic of Yemen. In P2, the number of countries affected by coups d’état remains high at thirteen: Algeria, Bahrain, Egypt, Iraq, Libya, Mauritania, Qatar, Sudan, Syria, Tunisia, Turkey, United Arab Emirates, and the People’s Democratic Republic in Yemen. While seven countries counted in P1 have not reported a coup attempt in P2 (Iran, Jordan, Lebanon, Morocco, Oman, Saudi Arabia, and Yemen Arab Republic), two countries experienced a coup attempt in P2, but not in P1 (Bahrain and Tunisia).
The decreasing number of observations in the MENA region does not imply the ultimate subordination to civilian autocrats, but rather the successful containment of coup cascades in the majority of the concerned countries. In Iraq, eight coup attempts characterized the years 1963–1973; in Sudan, four coups happened in 1955–1959 alone in addition to six coups in 1966–1977; Syria witnessed seven military interventions between 1961 and 1970; and the two Yemeni states were struck by a combined six coups during 1967–1978. All these postcolonial regimes were ruled by juntas that had come into being through military coups. This explains why they are challenged primarily from within their military apparatuses, 47 but it does not explain why the same countries continued to be affected by coup incidents once ruling coalitions coup-proofed.
Most countries in the MENA region experienced coups as isolated incidents during P2: seven of the thirteen countries witnessed only one coup attempt (Algeria, Bahrain, Iraq, Libya, People’s Democratic Republic of Yemen [PDRY], Syria, and United Arab Emirates); Egypt, Tunisia, Turkey, and Qatar reported two incidents each. The effects of coup-proofing were most significant in Iraq, Syria, and Yemen. In P1, these three countries alone account for a total of twenty-three coups (37.1 percent of all coups in P1), but only three attempts in the post-1980 era (11.1 percent of all coups in P2). Mauritania (eight coups) is somewhat exceptional because several coups have been observed throughout the greater part of P2. Sudan saw a belated consolidation, with four coups between 1983 and 1990, but the ruling coalition has somewhat stabilized since the takeover of Omar al-Bashir.
There are other empirical indicators supporting Hypothesis 1: authoritarian consolidation led to the containment of coup cascades, but this did not completely immunize incumbents. The containment of coup cascades facilitated the extension of incumbents’ tenure, but did not remarkably reduce the average likelihood of each individual incumbent to experience at least one coup attempt during their time in office.
Figure 2 visualizes the number of coup attempts in P1 in relation to the number of years of incumbents’ tenure. The data reveal that the vast majority of coup attempts happened during the first three years of a new ruler’s ascent to power. Throughout the 1950s, 1960s, and 1970s, the vast majority of executive incumbents in Iraq, Sudan, Syria, Turkey, and Yemen did not come to enjoy a particularly long period of time in office without facing a coup attempt—and they often fell over their military rivals’ political ambitions.
Coups d’état and incumbent tenure, 1950–1979.
In fact, many coups in P1 were directed against incumbents who had held office for a significant amount of time prior to the observed period, often as puppets of colonial powers: Egypt in 1952 (ousting King Farouq I, ruled for 16.2 years), Iraq in 1958 (replacing King Faisal II, in power for 19.2 years), and Oman in 1970 (replacing Emir Said Bin Taimur, in charge for 38.4 years). Only few incumbents have avoided coup attempts for long during P1: Libya’s King Idris al-Senoussi ruled for 17.8 years before being ousted by Colonel Mu’ammar Qadhdhafi in 1969; Mauritania’s Moktar Ould Daddah presided over Mauritania for 17.7 years before falling to a junta in 1978; Egypt’s Gamal Abdel Nasser came to power in 1954 by sidelining his rival Mohammed Naguib and passed away in 1970 without any reported coup attempt; Bahrain’s Salman Bin Hamad Al Khalifa served in 1942–1961; the Emirs of Kuwait, Abu Dhabi, and Dubai have never reported a coup attempt; and the Saudi Kings Faisal and Khalid ruled during the 1960s and 1970s. All other executive office holders in the pre-1980 period have either experienced a coup attempt or had to abandon power after a short period for other reasons, such as assassinations, wars, natural death, or voluntary resignation.
The immediate consequence of coup-proofing was the extension of incumbents’ tenure in office. During P1, the average tenure of executive office holders, when they experienced coups d’état, amounts to 5.2 years; the average leadership duration for the same group of individuals in P2 increased to 11.2 years. While, in P1, not many executive office holders came to enjoy a status as elder statesmen, P2 saw an increasing number of elderly rulers.
Figure 3 visualizes the effects of coup-proofing in authoritarian regimes. The early stage of an incumbent’s takeover is still a moment of opportunity for coup plotters. But coup-proofing leads to a significant degree of job security once personal rule and ruling coalitions are stabilized. In P2, no executive office holder ever became the target in a coup attempt during his seventh to twelfth year in office. This observation is in line with the findings of studies of authoritarian leadership consolidation. 48 Yet, MENA data isolating CPAs reveal a substantial increase in the number of targeted rulers during the later stages of their political careers. Several long-serving autocrats were targeted by ambitious military personnel, whose actions were often catalyzed through political or economic crises.
Coups d’état and incumbent tenure, 1980–2013.
Among long-serving rulers, Khalifa Bin Hamad Al Thani was replaced by his son in 1995, after running affairs in Qatar for 23.3 years. Chadli Benjedid was replaced by a military junta in 1992, after serving 12.9 years as Algeria’s president. Despite the obvious differences between the two countries, both rulers—in Algeria and Bahrain—were sidelined when they proved unable to solve protracted social and economic instability. Mauritania’s Maaouya Ould Taya seemed to have stabilized the country to some extent, only to fall in 2005 after serving 20.7 years as president. Iraq’s Saddam Hussayn survived two Gulf wars, several uprisings, assassination attempts, and a coup attempt in 1996 (after 16.9 years in power). Tunisia’s Bourguiba was sidelined by Ben Ali in 1987 after 30.3 years in power and amid social unrest following economic reforms in the 1980s. Ben Ali himself was forced out of the country by his military during the Arab spring in 2011, 23.2 years later. Hosni Mubarak, ruling over Egypt for 29.2 years, met Ben Ali’s fate only three weeks later. In Libya, Mu’ammar Qadhdhafi survived a coup attempt in 1993—24.1 years after his takeover—and various attempts on his life only to succumb to Libya’s uprising in 2011. The sidelining of Yemen’s Ali Abdullah Saleh in November 2011 and the desperate struggle of Bashar al-Assad throughout Syria’s uprising were not coded as coups d’état (and thus not included in the data), but these events also reveal the eminent threat that military officers’ insubordination imposes on long-serving autocrats.
The comparison of the two periods P1 and P2 reveals intriguing insights in individual incumbents’ likelihood of becoming the target of coup attempts, whether successful or not. Of the eighty-three rulers serving during P1, forty-four have experienced coup attempts (53.0 percent). During P2, forty-nine executive office holders enjoyed conclusive terms, twenty-two of which experienced a coup attempt (44.9 percent). 49 The difference between the two periods becomes all but insignificant when comparing success rates. In P1, twenty-six of eighty-three incumbents fell to coup plotters (31.3 percent); during P2, the rate among incumbents with conclusive terms remained equally high at 30.6 percent (fifteen of the forty-nine).
The high personal coup risk for post-1980 incumbents in the MENA region stands in remarkable contrast to the results of a global study on the ways of losing power, which maintains that 11.6 percent of rulers worldwide lose power through coups d’état. 50 Yet, the MENA data on P1 correlate with the global data underlying the study of Bienen and van de Walle, the results of which were informed by an underrepresentation of CPAs. The initial hypothesis of Bienen and van de Walle, predicting a U-shaped curve, receives empirical support from the MENA region only for the post-1980 period, largely characterized by CPAs (with Mauritania as exception). Coup plotters strike when they believe that incumbents command over limited elite or popular support. This explains the high number of coup cascades in P1, the limited average duration of tenure, and the few examples of long-term rulers in that period. While turnover rates have decreased in the CPAs during P2, the general propensity of coup plotters to strike when an opportunity arises has remained high. This explains the rising number of coups attempted against long-serving autocrats.
Who Are the Coup Plotters in CPAs?
Coup plotters adapt to coup-proofing. Not only will they wait for the right moment in time to challenge a weakened incumbent. In coup-proofed authoritarian regimes, coup plotters will also increasingly make use of superior organizational resources. In the CPAs of the MENA region, state apparatuses have grown and institutionalized. Military apparatuses have also undergone processes of institutionalization, leading to the establishment of large standing armies, an infrastructure consisting of modern equipment and communication tools, and a more effective chain of command than in the early years of state building throughout the 1950s and 1960s. In the wake of this modernization process, it has become much more challenging for peripheral elite segments and fragmented army units to successfully stage coups d’état and usurp power. The “Young Officers,” leading postcolonial independency movements in Egypt, Syria, Yemen, and Libya, consisted of small groups of middle-rank military officers and would most likely be unsuccessful today. In order to stage a coup d’état in CPAs, agency is as decisive as timing. Hence, I assume that adaptation to sophisticated coup-proofing will be indicated by an increasing number of centrally planned coups, corresponding with increasing success rates.
A simple cross-tabulation constructs a relationship between coup agency and coup success (see Tables 1 and 2). The data show that the number of coups undertaken by the leadership of military apparatuses (n[cen]) increased significantly from 27.4 percent in P1 to 40.7 percent in P2. The increase in the success rate of coups d’état (n[suc]) is modest but noteworthy, from 46.8 percent in P1 to 55.6 percent in P2. 51 Yet, the data indicate a strong relationship between coup agency and coup success in CPAs. In P2, all eleven coups d’état attempted by central military leaderships have been successful. These coups make 40.7 percent of n[total] in P2—up from only 21.1 percent of successful, centralized coups in P1. The drop in the number of false positives is also significant: in P1, 55.2 percent of all successful coups were undertaken by fragmented groups within the military, down to only 26.7 percent in P2.
Agency and Outcome of Coups d’état, 1950–1979.
Agency and Outcome of Coups d’état, 1980–2013.
That fragmented coup plotters have been quite successful in P1—compared to their fellow officers in P2—can be explained with the particular situation of state building in the 1950s and early 1960s. The data in P1 include those coups directed against political elites, often backed by colonial powers—such as in Egypt 1952, Iraq 1958, Syria 1954, Turkey 1960, and Yemen Arab Republic 1962. Most of these “founding coups” have been successful even though the majority was launched by younger officers. When left-censoring them—a total of eighteen coups, including eleven successful and seven failed coups—the success rate (n[suc]) in P1 drops to 40.9 percent. With state building in the MENA region a bygone era, longitudinal data therefore show a significant increase of coups executed by military leaderships, correlating with higher success rates.
One limitation of this study, mainly owing to its reliance on existing data sets, is a possible bias on failed coup attempts. It is likely that various incidents of failed and aborted coup attempts are not reported in the data. Successful coups d’état are relatively easy to detect; failed and aborted coup attempts are not. Information on the latter is often based on rumors or the claims of governments to have uncovered plots. Such claims are difficult to verify because independent information on the events in an authoritarian “black box” is scarce. The authorities’ announcements to have uncovered a coup, the exiling of alleged conspirators, and mass executions of military personnel are viable indicators. Yet, autocrats may employ such measures for a variety of alternative purposes: the elimination of potential rivals in elite coalitions, preemptive measures to forestall imagined coups, and uprisings not necessarily aimed at overthrowing incumbencies. Spreading the rumor of a coup plot might then be invoked to legitimize exaggerated coercive action.
On other occasions, incumbents may have an interest to conceal coup attempts, for example, in order to sustain the picture of cohesive elite coalitions. Hereditary monarchies, where elite configurations are based on strong family bonds, are usually interested not only in avoiding coups but also in hiding them from public scrutiny, for coup plots might indicate fissures within ruling families. Rigorous qualitative and archival research will help uncover failed and aborted coup incidents, and hence improve the quality of large-n and medium-n data collections.
Concluding Remarks: Guidelines for Coup Plotters, and Autocrats
Table 3 lists all coups d’état executed in the post-1980 period in the MENA region, analyzed as a historical period of CPAs. This table helps to summarize my argument. First, coups d’état are not a phenomenon of the past. Second, coup plotters in CPAs are successful when they are executed by military leaderships, and they are promised to fail when coup plotters are organized in peripheral military units.
Agency and Incumbent Tenure in Coups d’état, 1980–2013.
Note: PDRY = People’s Democratic Republic of Yemen.
Clever coup plotters consider timing to be an important factor for success. Centralized and successful coup plotters have acted against incumbents shortly after their takeover, as in Turkey 1980 and 1997, or in Egypt 2013. Once this moment of opportunity passes, coup plotters will have to be prepared to wait patiently—exceeding the period of power maintenance to which most democratic leaders aspire in their political careers. Yet, coup plotters should not be discouraged by authoritarian coup-proofing. Lame duck autocrats become vulnerable and, more often than not, do not retain the means to fend off their militaries’ ambitions.
Opportunities for the takeover of power through coups arise in domestic politics: as a response to severe economic crises, such as in Tunisia 1987 and Sudan 1989; to contain domestic opposition, as in Algeria 1992 and Turkey 1997; triggered by popular mass uprisings, such as in Egypt 2011 and 2013 and in Tunisia 2011; and when rulers are temporarily incapacitated, such as in early 1983 in Syria. International factors might aggravate such domestic crises or more directly create incentives for coup plotters by weakening incumbents. Western support to oust the rulers in Libya and Iraq during the 1990s proved unsuccessful at the time, yet it may well have encouraged coup attempts in 1993 and 1996, respectively.
It is more difficult to offer suggestions to autocrats as to how to avoid coups d’état. The simplest trick is to embark on a path toward democratic consolidation, a strategy with strong disincentives for authoritarian incumbents who would surrender to elected officials, rather than their officers. Autocrats want to refrain from excessive personal resource extraction and rather broaden ruling coalitions if they have the necessary financial resources at their disposal. Excluded parts of society or the political establishment will breed grievances and likely engage in opposition to incumbents in crisis. The most effective strategy to avoid coups d’état is the sophisticated management of political succession. Timing is decisive for all new incumbents. The sooner they reestablish ruling coalitions, the more likely it is that militaries are prevented from removing them early on.
Yet, there are also significant differences between subtypes of autocracies. MENA data reveal that hereditary monarchies are able to reduce general coup risk better than presidential republics. Four of the twenty-seven coups happened in those countries, with only one successful case in the observed period: Qatar in 1995. The financial capacities of the oil-rich Gulf states are one reason. Yet, more generally, institutionalizing succession without allowing democratic procedures to take root is more challenging for incumbents in presidential republics. The strategy of quasi-hereditary turnover, in which power would be transferred from the father to son, was attempted but triggered the Arab uprisings, 52 disastrous events for incumbents in the Arab republics. A more promising approach seems to lie in the establishment of term limits. The institutionalized retrenchment of a single individual’s tenure in executive office might not necessarily threaten the core mechanisms of authoritarian rule—as can be seen in Iran, China, or Putin’s Russia. But it might prevent personalized authoritarian regimes from slipping into severe crises when the mighty ruler turns into a lame duck. With institutionalized mechanisms in place to rejuvenate ruling coalitions, democratic or not, coups d’état might well become a phenomenon of the past.
Footnotes
Appendix A
Appendix B
Acknowledgment
I would like to thank Eva Bellin, Dina Bishara, Matt Buehler, Elliott Colla, Maria Glenna, Steven Heydemann, Michelle Jurkovich, Hazem Kandil, Kevin Koehler, Jean Lachapelle, Ellen Lust, Joshua Stacher, Lucan Way, and three anonymous reviewers for helpful comments.
Authors’ Note
A first version of this article was presented at the 4th Annual Conference of the Project on Middle East Politics (POMEPS), Washington DC, May 23–24, 2013. All errors and omissions are mine.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Much of the research for this article was done when the author was a Jennings Randolph fellow at the United States Institute of Peace (USIP), in 2012 and 2013. The views expressed in this article are those of the author and do not necessarily reflect the views of the United States Institute of Peace.