In the early 1990s, Samuel Huntington famously noted that political transitions tended to cluster historically into ‘waves’: broad, global movements in which the number of global democracies either expanded or retreated during reverse waves (Huntington Reference Huntington1991). The third wave, beginning in the 1970s, toppled autocracies in a number of regions: southern Europe, Latin America, East Asia, the crumbling Soviet bloc and sub-Saharan Africa, before largely cresting in the mid-1990s. However, scholars soon began to identify serious deficiencies in many transitional democracies. Nominally democratic elections were held for national offices, but incumbents failed to protect the civil liberties of their citizens, harassed and intimidated opposition parties and manipulated or rigged elections. These regimes were at first conceptualized as ‘democracy with adjectives’ (Collier and Levitsky Reference Collier and Levitsky1997: 430–51), including as ‘illiberal’ (Zakaria Reference Zakaria1997), ‘pseudo’, ‘partial’ or ‘electoral’ democracies (Diamond Reference Diamond1996). Thereafter, many researchers began to question the assumption that these regimes could be designated democracies at all. Instead, they argued such cases were better understood as a new, novel and often stable form of dictatorship emerging in the post-Cold War era – ‘competitive’ authoritarianism (Howard and Roessler Reference Howard and Roessler2006: 367; Levitsky and Way Reference Levitsky and Way2002).
Even as many emerging autocrats manipulated formally democratic institutions to maintain their grip on power, they began to face serious challenges from below. Beginning in the late 1990s with the ‘colour revolutions’ of Eastern Europe and Eurasia, well-organized opposition forces began contesting rigged elections and then challenging the fraudulent results on the streets, often bringing about the collapse of incumbents (Bunce and Wolchik Reference Bunce and Wolchik2006: 5–18). In the Arab world, largely untouched by the third wave of democracy and populated by long-surviving autocracies, sudden protests emerged in late 2010.
These two most recent regional waves of anti-regime unrest have invited reflection on what we have learned about the sources of authoritarian resilience and vulnerability. Drawing on the general literature on authoritarianism and contentious politics, this study assesses the power of various explanations for the breakdown or survival of autocratic regimes in new sites of contention, where large-scale protests spread from neighbouring initiating sites. These explanations are formulated into alternative hypotheses centring on the causal role of economic crises, regime subtype, international linkages, leadership turnovers and rentierism in contributing to regime vulnerability in the face of popular challenges.
The five hypotheses under investigation are assessed through a cross-national comparison focusing on 17 protest years in which non-democratic regimes were challenged by large-scale popular unrest inspired by events in neighbouring countries, which then diffused into their own states. In regional protest waves such as the post-communist colour revolutions and the Arab revolutions, novel forms of large-scale contentious collective action were innovated in initiating sites – Slovakia (1998) for the former and Tunisia (2011) for the latter – and produced political change, revealing vulnerabilities in these regimes. Observers and activists in neighbouring states drew inspiration from the successful outcomes of these protests and emulated these forms of protest in their own states. In these new sites, regime challengers were often (but not always) successful in organizing large protests. But even successfully organized large protests enjoyed mixed success in producing regime breakdown and compelling autocrats to exit from power. In the 17 cases of large-scale protests examined here, seven produced regime breakdown, whereas in the remaining 10, regimes maintained cohesion in the face of protests and successfully retained their grip on power. By focusing on a relatively small sample of rare moments in which large-scale protests emerged in the context of spatially and temporally bounded regionalized waves, this study observes tests of the durability of a regime when faced with sudden, acute and often unanticipated adversity, rather than longevity – the length of a regime’s tenure in power. It asks why certain regimes collapse and others perish when faced with sudden, externally inspired large-scale political contention.
Notably, some alternative explanations in the broader literature on authoritarianism and democratization are excluded from the analysis. These include the relationship between nominally competitive elections, inequality and the robustness of security forces in contributing to authoritarian resilience or vulnerability (Bunce and Wolchik Reference Bunce and Wolchik2010: 49–50; Brownlee, Reference Brownlee2007: 32; Gandhi Reference Gandhi2008; Gandhi and Przeworski Reference Gandhi and Przeworski2007). In this article, elections and measures of electoral competitiveness are excluded because of their negligible role in the Arab uprisings. Unlike the demonstrations organized against rigged elections typical of post-communist colour revolutions, protests in the Arab world did not precede or follow elections and were not in reaction to the stealing of elections.
Second, Acemoglu and Robinson (Reference Acemoglu and Robinson2006) and Houle (Reference Houle2009) suggest that inequality plays an important causal role in initiating political transitions in authoritarian countries (but also is harmful to consolidation). As noted by Acemoglu and Robinson (Reference Acemoglu and Robinson2006), high inequality contributes to large protests since unequal dictatorships are more likely to see large protests when citizens place redistributive pressure on the regime. In this study, large protests are treated as a constant, and because inequality is largely conceptualized as a driver of unrest rather than breakdown in the face of unrest, it is excluded as an independent variable seeking to explain regime survival amidst large protests. All regimes under examination experienced large-scale unrest.
Thirdly, scholars also suggest that the robustness of security forces is a pivotal factor in determining whether or not an autocrat survives moments of large-scale unrest. In an important 2004 article, Eva Bellin (Reference Bellin2004: 144–6) argues that the unique resilience of authoritarianism in the Middle East is largely linked to the ‘robustness of the coercive apparatus’. This variable is determined by the fiscal health of a regime, its ability to apply repression without the loss of international support, the degree of institutionalization and patrimonialism in the security forces and, finally, the occurrence of large-scale popular mobilization, which increases the cost of applying repression (Bellin Reference Bellin2004: 144–6; Reference Bellin2012: 127–49). This article considers international linkages, economic conditions, regime type and political institutionalization, and rentierism as causal explanations in their own right and in case selection assumes the presence of large-scale unrest for various regimes. Thus, because of the inclusion of these variables and their high degree of interrelatedness with other independent variables, a measure for the robustness of the coercive apparatus, as envisioned by Bellin (Reference Bellin2004, Reference Bellin2012), is not included.
Of the five hypotheses explored in this exercise, leadership turnovers and the absence of resource wealth were confirmed as the most powerful predictors of regime vulnerability in the face of regional waves of unrest. With the exception of one outlier (Serbia), all cases of regime breakdown occurred in the absence of an economic crisis. Economic crises were certainly bad news, but they were not necessary ingredients for regime breakdown in the post-communist colour revolutions or Arab uprisings. Monarchies were particularly resilient regime types in the face of regionalized unrest relative to hybrid and personalist regimes. Social and economic linkages to the outside world did not have an observable effect on regime vulnerability, although political linkages such as participation in international organizations and treaties had a mild destabilizing effect. In short, regimes strengthened by structural factors such as resource wealth, monarchical political organization and isolation from international political pressures were resilient to unrest. Many otherwise resilient regimes were caught at a moment of vulnerability – during leadership transitions – when exogenous protests suddenly spread across their regions.
The Diffusion of Protest in Regional Waves
The cases in this study are protest years associated with two temporally and spatially defined waves of unrest: the post-communist colour revolutions of the late 1990s through the 2000s and the 2010–11 Arab uprisings.Footnote 1 During these highly interrelated waves (Beissinger Reference Beissinger2009: 74–7), new repertoires of contention were developed in initiating sites such as Slovakia (1998) and Tunisia (2010) to challenge autocrats successfully, revealing vulnerabilities in the regimes. Domestic activists introduced new forms of contention – often borrowing tactics from outside the region, including the Philippine People Power Revolution (1986) or the strategies circulated by Gene Sharp and the Albert Einstein Institute – but adapted them to local contexts. Once successfully implemented in initiating sites, protest tactics and strategies were then transferred to new sites within the region via demonstration and/or brokerage effects, resulting in the sudden appearance of similar modes of popular contention in neighbouring countries (Beissinger Reference Beissinger2007: 259–60; Patel and Bunce Reference Patel and Bunce2012: 10–11).
The focus of this study is the survival of autocrats in new sites of contention. In initiating sites such as Slovakia (1998) or Tunisia (2010) activists introduced new and novel forms of popular contention to their regions, and the triggers that sparked unrest were domestic. Conversely, in new sites such as Serbia (2000) or Egypt (2011) large-scale unrest emerged in reaction to exogenous events – namely the protests and regime breakdown observed in initiating sites. Externally originated unrest in new sites presented a unique challenge: it emerged suddenly, applied new and novel repertoires of contention, mobilized and united previously unconnected groups and challenged strategies for political control that generally functioned well under normal conditions. In other words, exogenously motivated unrest overwhelmed the methods for preventing and managing popular contention of many regimes in new sites which had functioned well under normal conditions. As observed in this study, these rare moments of sudden, exogenously inspired protest resulted in regime breakdown and the exit of leaders in seven of 17 cases, including Hosni Mubarak (Egypt) and Slobodan Milošević (Serbia), whereas 10 others, including Alexander Lukashenko (Belarus) and Bashar al-Assad (Syria), proved capable of maintaining regime cohesion and their grip on power. This study seeks to bring greater clarity to these divergent outcomes.
In the two waves examined – the colour revolutions and Arab uprisings – the specific modes of resistance and diffusion differed, following region-specific patterns. In the colour revolutions, opposition activists organized their efforts around regular ‘rigged electoral rituals’. They formed united opposition movements, conducted extensive ‘get-out-the-vote’ efforts and made use of election monitoring and other support provided by international organizations and the donor community. When elections were ultimately rigged, the opposition placed pressure on the regime with prearranged campaigns of non-violent resistance (Beissinger Reference Beissinger2007: 261; Bunce and Wolchik Reference Bunce and Wolchik2006: 6). The diffusion of this mode of protest relied not only on demonstration effects – the observed success of early cases such as Slovakia and Croatia – but also brokerage, in which graduates of early colour revolutions travelled to new sites and, with the support of democracy-promoting organizations, shared their experiences and worked to adapt earlier revolutions to local conditions (Bunce and Wolchik Reference Bunce and Wolchik2006: 11–12).
In the Arab uprisings, demonstration effects outweighed brokerage. In some early cases, particularly Egypt, non-violent street demonstrations were in part coordinated by youth activists who had trained with Serbian protest veterans and received assistance from the democracy-promotion community (Rosenberg Reference Rosenberg2011). Unlike the colour revolutions, which spanned the course of a decade and involved organized oppositions conducting careful preparation and coordination in advance of elections, protest actions diffused almost immediately across the Arab world. Participants observed events on social media and satellite television and quickly adopted similar tactics, organizing ‘Days of Rage’ in central squares involving diverse actors such as youths, secular oppositionists, labour unions and Islamists (Patel and Bunce Reference Patel and Bunce2012: 12–13). Thus, the form of protest activities and the speed of diffusion varied between regions.
In both the Arab uprisings and post-communist colour revolutions, waves of unrest drew international attention but were ultimately bounded by physical and cultural geography; supporting studies suggest that geographic proximity plays a powerful causal role in driving political transitions (Elkink Reference Elkink2011: 1651–74; Gleditsch and Ward Reference Gleditsch and Ward2006: 911–33). These limits were related to several factors. First, at the onset of unrest, both regions were defined by region-specific structural factors which created unique opportunities for particular modes of contention across the affected sites. Countries experiencing attempted colour revolutions shared important legacies associated with the communist/Soviet era: they had long experiences with highly fraudulent elections, high levels of education and near-universal adult literacy and, finally, long traditions of civilian rule over the military (Bunce and Wolchik Reference Bunce and Wolchik2006: 7–9).
In the Arab world, experience with elections varied and was largely non-existent in some venues, literacy rates ranged from over 90 per cent in the Gulf States to as low as 54 per cent and 64 per cent in Morocco and Yemen (World Bank 2013), national identities competed with overarching Islamic forms of identity, and civilian control over the military was often very suspect. These regional variations in part explained why post-communist regions saw rigged elections emerge as the focal point for anti-regime popular contention whereas Arab protests involved more spontaneous mass demonstrations organized without reference to the electoral calendar.
Secondly, as noted by Tarrow (Reference Tarrow2005), the diffusion of popular activism is a highly ‘relational’ process that is made easier when populations are connected through social bonds and personal ties. This social connectedness encourages participants to attribute similarity to their overseas counterparts, to see themselves in the struggles of similar peoples in neighbouring lands, and ultimately to emulate their repertoires of contention (Tarrow Reference Tarrow2005: 103–4). Peoples in Eastern Europe and Eurasia had long-standing transnational ties from the communist and post-communist eras, common and nearly simultaneous experiences with the difficult transition away from state socialism and a set of cultural symbols associated with the time period.
Similarly, populations in the Arab world spoke a common language, shared a cross-national traditional and popular culture and had dense transnational interpersonal networks. Importantly, they relied on international media outlets that broadcast real-time news across the region – with al-Jazeera emerging as particularly important during 2011. Geographic proximity and dense region-specific social bonds meant the early colour revolutions and the symbols and strategies had greater cultural resonance in sites within the post-communist region than beyond it. Meanwhile, unrest within the Arab world had much stronger appeal and more powerful demonstration effects in other Arab states than in countries outside the region, where organized efforts such as the 2011 Chinese ‘Jasmine Revolution’ attracted relatively little popular attention or interest from their respective publics.
Waves of unrest were also bounded by time.Footnote 2 Scholars note that explosions of popular unrest and the regime changes that often follow them are extraordinarily difficult to anticipate (Kuran Reference Kuran1989). In many respects, the possibility of large-scale unrest is dependent on this element of surprise. Autocrats have a massive resource advantage over their opposition challengers and, if forewarned of impending social unrest, can take preventive measures, arresting activists and dispersing small crowds before they grow too large to control. Of course, large-scale unrest can begin with small, seemingly inconsequential sparks and in the case of regional waves be triggered by developments outside the country and the autocrat’s territorially limited zone of control. As discussed in studies of popular mobilization, individuals under normal conditions are deterred from protesting in the public square (Easley and Kleinberg Reference Easley and Kleinberg2010: 483–508; Tucker Reference Tucker2007: 535–6). Since the regime is perceived as strong and omnipresent, collective actions taken in isolation are personally risky and unlikely to be effective. However, as larger groups of individuals begin to gather, the decisions of individual bystanders change. They make inferences based on the choices of observed others and judge that risks are lower and benefits higher. As crowds grow, more bystanders determine the regime to be vulnerable, raising their expectations of success and lowering their assessment of the personal dangers associated with participation. Such ‘information cascades’ may also transcend national borders, as individuals make decisions to organize demonstrations in their own countries based on observed patterns in foreign lands. When unrest in neighbouring countries is successful in producing regime change, audiences in connected sites are more likely to act in a similar fashion.
Of course, as suggested by Weyland (Reference Weyland2012: 917–34), when activists and demonstrators use information from events occurring abroad to inspire collective action at home, it may lead to miscalculation. Using incomplete information, protesters may underestimate the strength of the regime at home and overestimate their likelihood of success. As unrest spreads cross-nationally, some regimes may collapse while others suppress popular challenges. Cases of regime resilience subsequently embolden surviving autocrats and discourage demonstrators, while autocrats may also observe events abroad and learn, taking preventive measures and fine-tuning their responses to challenges from below (Heydemann and Leenders Reference Heydemann and Leenders2011: 647–53). Thus, as the protesters’ tactical advantage dwindles, participants are discouraged and surviving regimes maintain their grip on power. The regional wave of unrest subsides.
Research Design
The cases included in this study are protest years in which national sites were connected to two spatially and temporally bounded regional waves of contention: the 1998–2010 post-communist colour revolutions and 2010–11 Arab uprisings. Outbreaks of unrest outside the post-communist region – Eastern Europe and the former Soviet Union – for the colour revolutions, and outside of Middle East and North Africa for the Arab uprisings, were not included, nor were outbreaks of unrest beyond these time periods.
Only protests coded as ‘large scale’ are included in the sample. In the literature on authoritarianism and regime change, there is no single definition for what constitutes a high- or regime-threatening level of unrest. In this study, the author compiled media accounts of large protest events associated with colour revolutions and the Arab uprisings. Sources included the New York Times, BBC News, Al Jazeera, Associated Press, The Guardian, Swarthmore Global Nonviolent Action Database and International Crisis Group. National populations for each country site during the protest year (World Bank 2013) were then divided by estimates of peak protest size (the most participants involved in a single protest action in a given day during the protest year) provided in media sources to give a rough estimate of the proportion of citizens involved in protest activities. Protests were considered as large scale if peak participation rates in a single protest event equalled or exceeded 0.1 per cent of the national population during that year. Setting the threshold at a high magnitude reduced the number of overall cases to a relatively small sample but ensured that demonstrations were sufficiently large to threaten regimes’ hold on power, thus providing a more serious test of their durability in the face of serious popular challenges. Cases such as Azerbaijan (2011), Russia (2011), Oman (2011), Saudi Arabia (2011), UAE (2011) and Algeria (2011) were excluded because of low participation rates. Colour revolution protest years above the threshold included Serbia (2000), Armenia (2003), Azerbaijan (2003), Georgia (2003), Ukraine (2004), Azerbaijan (2005), Kyrgyzstan (2005), Belarus (2006), Armenia (2008) and Kyrgyzstan (2010). Arab uprising protest years included Egypt (2011), Jordan (2011), Libya (2011), Morocco (2011), Syria (2011), Bahrain (2011) and Yemen (2011).
Protest years for these 17 cases are based on the date of the inception of protests. They are considered to be incidents of regime breakdown if they resulted in an incumbent leader exiting from power within one calendar year of the initiation of protests. Notably, Syria (2011) and Libya (2011) are coded as cases of regime survival despite their subsequent descents into civil war, both leaders having maintained regime cohesion in the face of initial protests. In the case of Libya, Muammar Gaddafi’s regime resisted popular protests but later collapsed in the face of an extensive NATO-led military intervention, culminating in Gaddafi’s death. The author endorsed the counterfactual assumption that without this extensive international intervention the regime would have survived the year, explaining Libya’s treatment as a case of regime survival. Consequently, seven cases are coded as incidents of regime breakdown: Serbia (2000), Georgia (2003), Ukraine (2004), Kyrgyzstan (2005), Kyrgyzstan (2010), Egypt (2011) and Yemen (2011), whereas the remaining 10 are treated as cases of regime survival: Armenia (2003), Azerbaijan (2003), Azerbaijan (2005), Belarus (2006), Armenia (2008), Jordan (2011), Libya (2011), Morocco (2011), Syria (2011) and Bahrain (2011).
Economic Crises
In the midst of acute short-term economic crises, regimes are more likely to be vulnerable to popular challengers. In his observations of the third wave, Samuel Huntington (Reference Huntington1991: 13) suggests that non-democratic regimes are highly dependent on ‘performance legitimacy’. By distributing rewards to the public, autocrats ensure that most citizens remain on the sidelines if and when opposition forces confront the regime. Strong economies also enable leaders to distribute benefits to regime insiders, buy off leading members of the opposition and lavishly finance their internal security apparatuses (Bellin Reference Bellin2004; Skocpol Reference Skocpol1979).
On the other hand, as demonstrated by the longevity of economically stagnant autocracies in the former Soviet bloc and much of the Middle East, non-democratic regimes can survive long and gradual economic declines. Moreover, as revealed by the eclipsing of authoritarian regimes such as South Korea and Taiwan, long-term economic achievement is not a lasting cure for authoritarian vulnerability. As noted by Gurr (Reference Gurr1968: 1110), rising economic standards create rising expectations that regimes need to achieve with regularity. If leaders fail to provide these benefits, rent-seeking regime insiders may calculate that their interests are better served by alternative leadership. These might be found among internal rivals or even the opposition.
Supporting these views, more recent research suggests that long-term economic declines are much less destabilizing than sudden, short-term economic crises. Geddes’ (1999) large-N empirical study reveals that among comparatively fragile military regimes, even mediocre economic news – an average 0.4 per cent increase in annual per capita income growth – typically precedes breakdown. For personalist and single-party regimes, much more serious short-term economic shocks are necessary. On average, collapse in these autocracies respectively occurs after declines of 0.5 per cent and 4 per cent in per capita income in the preceding year (Geddes 1999: 135–6). Recent research has confirmed that while long-term economic declines are largely manageable, short-term economic crises contribute strongly to regime breakdown (Ulfelder Reference Ulfelder2005: 324). In short, when large protests occur in the context of sharp economic decline, autocrats are more likely to see supporters divide and defect in the face of powerful grassroots opposition, contributing to regime collapse. In this study, GDP per capita growth rate in the year prior to the protest year is utilized as a measure for assessing the short-term economic conditions of regimes facing popular challenges (World Bank 2013).
Regime Type
Scholarship indicates that the type of authoritarian regime plays an important role in the resilience of autocracies, both during crises and in normal times. One strand of the literature on authoritarianism has focused on hybrid regimes – those political regimes that hold nominally competitive multiparty elections, which are then manipulated to advantage the incumbent party (Gandhi and Przeworski Reference Gandhi and Przeworski2007: 1279–301; Levitsky and Way Reference Levitsky and Way2010a: 3). Scholars have suggested that dividing electoral authoritarian regimes into subtypes based on the degree of competitiveness in these contests is a strong predictor of their vulnerability to political challenges. ‘Hegemonic’ authoritarian regimes where incumbents consistently won by large margins and elections involved almost no uncertainty in their outcomes have been remarkably stable in the post-Cold War era. On the other hand, ‘competitive’ autocracies, while operating on an ‘uneven playing field’ advantaging the incumbent (Levitsky and Way Reference Levitsky and Way2010b: 57–68), have involved a much greater degree of opposition contestation and uncertainty. They have been much more fragile to popular challenges, particularly when facing a well-organized and unified opposition coalition applying the electoral model developed during many colour revolutions (Roessler and Howard 2009: 108–24).
Students of authoritarianism have also noted that non-democracies vary significantly in the way in which regime insiders are recruited, promoted and rewarded for their service. This variation has a critical impact on elites’ perceived interests, the nature of their loyalty to the regime and their likelihood of defecting when popular opposition forces challenge the state. In recent decades, generations of scholars of authoritarianism have divided up non-democratic regimes into discrete subtypes (Friedrich and Brzezinski Reference Friedrich and Brzezinski1965; Linz, [1975] Reference Linz2000; Linz and Stepan Reference Linz and Stepan1996: 41–2; O’Donnell Reference O’Donnell1979). Applying a military, personalist and single-party categorization scheme, Geddes (1999: 131–2) and Brownlee (Reference Brownlee2007: 32) have found single-party regimes to be the most durable and long-lived of autocracies. In non-democracies where power is channelled through a hegemonic, highly institutionalized political party, this political organization functions as the only avenue for accessing rents, political offices and economic resources (Magaloni Reference Magaloni2008: 725; Nathan Reference Nathan2003: 9–11; Way Reference Way2008: 55–69). When the regime is threatened by popular challenges from below, the tendency of cadres is to close ranks rather than defect to the opposition. In personalist regimes, political power is oriented around the person of the leader. Important political offices are occupied by individuals with informal connections to the ruler – his kin, cronies and clients, and these elites manipulate their political positions to gain personal access to state resources and wealth. Like their single-party counterparts, personalist regimes are quite resilient under normal conditions (Bratton and van de Walle Reference Bratton and van de Walle1997: 86). However, they can quickly become fragile with the death or illness of the leader, the regime fragmenting as aspiring successors compete for the highest office. In military regimes, the government is dominated by officers who have left the barracks to lead the government (Cheibub et al. Reference Cheibub, Gandhi and Vreeland2010: 85–6; Geddes 1999: 125). Their prevailing concern is with preserving the military as an institution – maintaining its unity, cohesiveness and readiness. Consequently, when minority factions in the government challenge the leadership, lower-ranking officers tend to remain on the sidelines, fearing that intervention would result in disorder within the military or even civil war between factions (Geddes 1999: 125–9). For these reasons, single-party autocracies are typically more resilient than personalist and military regimes.
In more recent formulations, monarchies are treated as a fourth distinct subtype (Geddes et al. Reference Geddes, Wright and Frantz2014). The resilience of this subtype has been more controversial. Whereas the general literature on authoritarianism has treated the monarchy as an anachronism – a holdover from a pre-modern era ill-suited for the challenges posed by the contemporary political world, many Middle East scholars argue that these traditional political regimes are quite resilient. They have developed effective formal and informal institutions for resolving internal divisions, maintaining elite cohesion and managing dissent (Cheibub et al. Reference Cheibub, Gandhi and Vreeland2010; Herb Reference Herb1999: 1–19; Menaldo Reference Menaldo2012: 1–16; Yom and Gause Reference Yom and Gause2012: 76–7). Consequently, this study expects that single-party regimes and monarchies are likely to be more resilient than their personalist counterparts. The regime types of cases in this study are assigned according to Geddes et al.’s (2014) recently updated data set.
International Linkages
Many scholars note that non-democracies with dense linkages to powerful democracies are more likely to experience breakdown than those regimes with weaker ties. When autocracies have dense economic, organizational and social links to powerful democracy-promoting countries, they face stronger popular oppositions and are under greater pressure to restrain their use of repression when faced with challenges from below. Economic linkages involve trade, investment and the extension of loans (Kopstein and Reilly Reference Kopstein and Reilly2003: 120–54; Levitsky and Way Reference Levitsky and Way2010a: 43). Organizational ties include participation in regional and international organizations such as the European Union (EU) or the Organization for Security and Cooperation in Europe (OSCE). Social links include growing interpersonal contact between a country’s citizens and their counterparts in powerful democracies.
The role of the EU in influencing opposition protests in the post-communist region is a case in point. Milada Vachudova (Reference Vachudova2005: 3–7) has suggested that the EU provided both ‘passive’ and ‘active’ leverage that influenced the democratizing pathways of post-communist states in Central and Eastern Europe. In terms of passive leverage, the EU influenced neighbouring countries simply by existing. Membership of the EU offered the prospect of higher living standards, access to European markets and the symbolic benefits of ‘joining Europe’, whereas not entering the EU meant being excluded from these benefits and suffering the fairly harsh terms the organization extended to non-members. Active leverage involved the rigorous entry requirements associated with accession to the EU, which compelled aspiring members to implement economic and political reforms driving them in the direction of freer markets and more democratic politics. For incumbent rulers in states such as Romania, Bulgaria, Slovakia and Croatia, the reforms associated with the EU accession process threatened their grip on power. However, when these rulers baulked at unwanted reforms, the general public rallied behind opposition forces, compelling these leaders reluctantly to accept economic and political changes that undermined their hold on power (Levitsky and Way Reference Levitsky and Way2010a: 114–18; Vachudova Reference Vachudova2005: 3–7). As suggested by Gleditsch and Ward (Reference Gleditsch and Ward2006: 919–26), both the presence and direct action of powerful international actors such as the EU can tip the balance between incumbent parties and pro-democracy opposition forces in favour of the latter, even leading to revolutionary cascades across regions. Thus, autocracies with strong economic, organizational and social links to democratic states are more likely to experience breakdown than their more insulated counterparts, who have a comparatively freer hand to use repression against their own citizens.
The measure used in this study to approximate a regime’s degree of international linkages to influential democratic countries is the KOF (Konjunkturforschungsstelle) Globalization Index for the year preceding large-scale protests. The index for economic globalization combines measures of trade flows and foreign investment with those assessing the degree of trade restrictions a country has in place. The social globalization index combines data such as transnational telecom traffic, cross-national job transfers, a national population’s exposure to tourism, and measures for cultural proximity – specifically the consumption of Western products. Finally, the political globalization index measure is derived from data on the number of international organizations a country was a member of, the number of embassies within the country and the number of treaties it has signed with foreign countries (KOF Globalization Index 2012).
One potential shortcoming for using this measure to approximate a regime’s linkages to democratic countries is that globalization might well include greater interconnectedness with powerful authoritarian allies such as Russia or China, which might act as counterweights to external democracy promoters (Levitsky and Way Reference Levitsky and Way2010a: 41). However, based on the large economic weight of Western democracies and the dense linkages between them, the most globalized countries during the period under examination are either democracies or countries closely linked to them, as clearly reflected in KOF’s rankings and map visualizations of the index. In this study, countries with high scores on the globalization index are expected to have stronger links to influential democracies and thus be more vulnerable to popular contention than those regimes more insulated from outside pressures.
Leadership Turnovers
Leadership turnovers are expected to make regimes more vulnerable to challenges from below. During normal times, a stable, well-positioned leader heads an authoritarian system. The regime exhibits a high level of cohesion as high-ranking individuals within the system have received their positions and the rewards associated with them in exchange for their loyalty and connections to the leader. Moreover, long-tenured dictators are typically masterful strategists. They balance rival factions against one another, rotate elites to prevent them from establishing autonomous bases of economic or political power and eliminate, intimidate, marginalize or co-opt potential rivals. As a result, the stable presence of the ruler minimizes the corrosive impact of unbridled factionalism, while also providing the cohesive focal point that binds together the power elite (Geddes 1999: 132–3; Howard and Roessler Reference Howard and Roessler2006: 372). Consequently, even when popular opponents do challenge the regime, party insiders, business elites and other influential powerbrokers in society are reluctant to defect to the opposition or make their own bids for power.
However, this changes dramatically when various signals suggest that the leader is leaving office. After he declares his intention to leave office and retire, becomes old or begins to demonstrate visible signs of serious illness, the perceived interests of regime loyalists suddenly alter. Economic and political elites, in the interest of maintaining their privileged positions, attempt to predict who will succeed the outgoing leader so that they can curry favour with a likely successor. Even if the exiting leader has announced a successor, other rivals with comparable bases of support might emerge as challengers (Hale Reference Hale2005: 138–41). During this period of transition and uncertainty, the control of the exiting lame duck leader weakens. Meanwhile, competition between rival factions within the regime intensifies as potential successors jockey for power. This leaves the regime fragile when challenged from below, constraining its ability to suppress or co-opt leading opposition figures. Perceiving a window of opportunity, opposition activists redouble their efforts to topple the regime, attract greater numbers of supporters and appeal to insiders to defect. In the midst of mounting unrest, the regime loses cohesion and breaks down. In this vein, research suggests that during periods of leadership transition or uncertainty, autocracies are much more prone to collapse (Hale Reference Hale2005; Howard and Roessler Reference Howard and Roessler2006: 372; Magaloni and Kricheli Reference Magaloni and Kricheli2010: 127).
In this research, a leadership turnover is coded as being present in situations first, in which a leader has died in the year preceding a major protest challenge. Secondly, a leadership turnover is present when a national election is contested by a successor with the intention that he/she will replace an incumbent leader. This was the situation in which Ilham Aliyev replaced Heydar Aliyev in Azerbaijan in 2003, Viktor Yanukovych succeeded Leonid Kuchma in Ukraine in 2004, and Serzh Sargsyan followed Robert Kocharyan in Armenia in 2008. Finally, leadership turnovers are considered to take place when an incumbent ruler is at or above the age of 70 or gravely ill during the protest year in question. Advanced age is associated with leadership turnovers in Georgia in 2003 (Eduard Shevardnadze, aged 75), Tunisia in 2010 (Zine El Abidine Ben Ali, aged 75) and Egypt in 2011 (Hosni Mubarak, aged 83). During leadership turnovers, autocracies are expected to be significantly more vulnerable to popular challenges than during other times.
Rentierism
Scholars have long noted that authoritarian regimes in countries with large stocks of easily extractable mineral resources, particularly hydrocarbons, are uniquely resilient to political challengers (Belawi and Luciani Reference Belawi and Luciani1987). These resources enable autocrats to accumulate large volumes of wealth from their countries without the participation of most of the general population. They thus avoid the need to extract taxes from the population, which may sow dissent and require the regime to extend political concessions to the citizenry (Tilly Reference Tilly1985: 183–4). In addition, the availability of easily extractable mineral wealth enables the autocrat to establish ‘discretionary control over the economy’ (Way Reference Way2008: 60), in which funds can be strategically distributed as a reward to loyalists or withheld from potential challengers.
Under such conditions, a rentier state can emerge, in which informal connections to the political elite become the primary pathways to personal advancement and material wealth. The state, moreover, becomes a mechanism for rent-seeking, as ambitious individuals abandon opportunities available in the normal economy, instead currying favour with political insiders to win access to mineral wealth and employment positions. Because the political status quo provides high social standing and generous incomes to regime officials, authoritarian regimes in countries with large stocks of mineral resources are expected to be less likely than those in resource-poor countries to fragment and see defections in the face of large protests; insiders have little to gain and everything to lose from doing so under such conditions. To measure the degree of rentierism in a given case, this study utilizes the World Bank’s World Development Indicator for Fuel Exports as a percentage of all merchandise exports (World Bank 2013). The measure is comparable to a number of well-known studies on oil-dependent states (Ross Reference Ross2001: 325–61). To capture national conditions at the onset of a protest year, statistics are derived from the year preceding large-scale unrest.
Results
A summary of the results is shown in Table 1. For the 17 cases examined, economic crises provided a limited explanation for regime vulnerability. The 10 autocrats who survived large-scale protests were experiencing solid economic growth rates in their countries, with the average rate of annual per capita GDP growth in these regimes reaching a robust 5.6 per cent. For the seven leaders who fell from power in the face of popular unrest, the rate of annual per capita GDP growth occurred at the lower rate of 3.0 per cent. However, this figure is skewed downward by the weight of Serbia’s economic collapse in 1999, in which the GDP per capita declined by 10.9 per cent in a single year. If this outlier is excluded, per capita growth rates in failed autocracies were 5.3 per cent, nearly identical to the 5.6 per cent observed in surviving regimes. Of course, no other case observed in this sample was suffering from an economic crisis of the magnitude of Milosevic’s Serbia – a regime that ultimately did experience breakdown. This suggests that, as expected in much literature (Geddes 1999: 132), sharp, short-term economic crises can play an important role as drivers of regime breakdown, although these crises need to be extremely severe to have an observable destabilizing impact. However, acute economic crises are not a necessary precondition for breakdown, as six regimes collapsed even as their economies encountered positive growth, ranging from the slower 1.7 per cent rate in Kyrgyzstan (2010) to the impressive 10.3 per cent in Ukraine (2004). In short, strong economic performances are certainly positive news for autocrats but do not inoculate them from challenges from below. Regime breakdowns in these two regional waves occurred amidst growing economies and in the absence of sharp economic downturns.
Regime type also provides only a partial explanation for regime survival. This sample included no pure single-party regimes, and hybrid and personalist regimes demonstrated no significant pattern of resilience or vulnerability. However, monarchies proved unusually resilient in the face of protests. Eleven pure personalist regimes were observed – five (46 per cent) of which experienced breakdown during protest years. Three regimes were classified by Geddes et al. (Reference Geddes, Wright and Frantz2014) as authoritarian hybrids: Serbia (2000) as personalist/single-party, Egypt (2011) as personalist/single-party/military and Syria (2011) as personalist/single-party/military. In these regimes, ‘control over policy, leadership selection, and the security apparatus’ was not firmly controlled by one single leader, party or military officer corps, but shared by some combination of these poles of power (Geddes et al. 2014: 318). Of these hybrids, Serbia (2000) and Egypt (2011) experienced breakdown, whereas Syria (2011) survived popular challenges. In the former two cases, the efforts by Milosevic and Mubarak to bolster their personalist regimes by establishing institutionalized political parties did not appear to improve significantly their resilience to exogenously inspired protests relative to their pure personalist neighbours. Of course, neither Egypt nor Serbia constituted a pure single-party regime. Perhaps regimes with higher levels of institutionalization around a political party might have demonstrated greater unity and resolve in the face of demonstrations. This seemed particularly likely in Egypt, where regime fragmentation appeared as military elites intervened against Hosni Mubarak amidst popular demonstrations. In short, personalist and hybrid autocracies were not particularly resilient or fragile, suggesting that their collapse or survival in the face of large-scale, exogenously inspired protests was driven primarily by other factors.
Supporting the findings of a number of Middle East specialists and challenging the assumption in the general literature that these regimes were anachronisms, monarchies proved the most resilient of subtypes. Bahrain (2011), Jordan (2011) and Morocco (2011) all survived their popular challenges. As suggested by Yom and Gause (Reference Yom and Gause2012: 77), ‘The stability of monarchies during the Arab Spring has so strikingly defied theoretical expectations that many analysts have reversed the decades-long consensus and now contend that inherent cultural and institutional forces make such regimes more durable than their republican peers.’ Thus, the Arab revolutions suggested that the informal institutions and relationships cultivated under monarchies probably provided a significant, if underappreciated, boost to regime resilience in the face of crises.
Concerning international linkages, political links had a more observable impact than social or economic links on increasing the fragility of authoritarian regimes during the two regional waves. Among regimes that collapsed, the average total KOF Globalization Index was 54.3, whereas survivors were only marginally more closed, with an average index of 53.3. When regimes were analysed according to political, social and economic indexes in isolation, this revealed that, counter to expectations, failed regimes were less socially globalized (44.0) and economically globalized (53.1) than surviving regimes, which averaged 47.4 on the social index and 59.1 on the economic index, respectively. Meanwhile, political linkages had a stronger impact. Autocracies that broke down had more dense political links (60.6) to the outside world than surviving autocracies (57.3). When ranking these regimes by levels of political globalization, the trend was more striking. Of the eight most politically globalized regimes, five collapsed. Of the eight least politically globalized, only two – Serbia (2000) and Georgia (2003) – collapsed. These findings suggested that participation in international organizations and deeper diplomatic ties with the outside world made autocracies more likely to collapse in the face of large-scale regionalized unrest, whereas dense social and economic bonds made less of an impact.
As expected, leadership turnovers greatly increased the likelihood that regimes would collapse during regional waves of protest. Eleven regimes had continuity in their leadership during protest years; only three (27 per cent) experienced breakdown. Six regimes were undergoing leadership transitions during protest years; four (67 per cent) collapsed. Supporting this pattern, the leaders of collapsed regimes were on average significantly older than survivors. The average age of incumbent autocratsFootnote 3 included in the sample was 60 years. Among those incumbents who were driven from power during protest years, the average age at the time of large-scale protests was 68 years, whereas the age of incumbents who survived protest years was lower: 52 years old. In short, regimes with leaders who were old, ill or had announced a plan to leave office were much more vulnerable to collapse than regimes with younger, healthier and more stable incumbents. When large protests appeared during times in which the regime’s future leadership was in question, powerful insiders were more likely to compete with one another for power or defect to the opposition, resulting in regime fragmentation and collapse.
These conditions were evident in cases such as Egypt, where the succession to the ageing and ailing Hosni Mubarak was shrouded in secrecy and intrigue. A leaked 2009 US embassy cable reported: ‘Despite incessant whispered discussions, no one in Egypt has any certainty about who will eventually succeed Mubarak or under what circumstances’ (The Guardian 2010). Mubarak’s son, Gamal Mubarak, head of the ruling National Democratic Party (NDP), was identified as the mostly likely candidate, although intelligence chief Omar Suleiman, Field Marshall Mohamed Hussein Tantawi and Arab League Secretary General Amr Moussa were also mentioned as alternatives (The Guardian 2010). As Gamal rose within the National Democratic Party and developed growing influence within the business community and over economic policymaking, he struggled to establish power and influence over the defence forces, as he had weak contacts and no armed forces service record. There was visible resentment within the military officer corps, and rumours circulated that the military might intervene if the elder Mubarak were to die and be succeeded by his son (The Guardian 2011). Ultimately, competition between factions aspiring for the post-Mubarak leadership weakened the regime’s cohesion amidst the sudden appearance of protests inspired by events in Tunisia in late 2010 and early 2011. Important regime insiders, such as Mohamed Hussein Tantawi in Egypt, reached out to popular demonstrators as the old regime crumbled. Thus, when leaders were old, sick or otherwise expected to leave office, regimes were much more vulnerable to large-scale unrest than regimes with healthy, young and stable individuals likely to lead for future decades.
Rentierism had a powerful impact on the resilience of autocracies in the face of regional waves of protest. Among surviving autocracies, fuel exports were an average of 43.5 per cent of all merchandise exports, compared with 23.7 per cent in failed regimes. In fact, among the seven regimes in which over one-third of all exports were fuel exports, only one autocracy – Yemen (91.2 per cent) – experienced breakdown during a protest year. All others, including Belarus (34.6 per cent), Syria (49.9 per cent), Bahrain (74.3 per cent), Azerbaijan 2005 (82.2 per cent), Azerbaijan 2003 (88.9 per cent) and Libya (97.7 per cent) maintained regime cohesion in the face of large-scale protests. As suggested in much academic literature, access to easily extractable resource wealth provides an advantage to autocrats threatened by large-scale outbreaks of popular unrest. With the single exception of Ali Abdullah Saleh in Yemen, autocrats in resource-rich states were able to buy supporters and buy off potential opponents, maintain regime cohesion and ride out outbursts of exogenously inspired protests.
For resource-poor regimes – with fuel exports making up less than one-third of all merchandise exports – the appearance of large-scale unrest was much more threatening. Six of 10 resource-poor autocracies – Serbia (2000), Georgia (2003), Ukraine (2004), Kyrgyzstan (2005 and 2010) and Egypt (2011) – collapsed during protest years. As suggested by Levitsky and Way (Reference Levitsky and Way2010a: 67), regimes without discretionary control over the economy were less able to ‘starve oppositions of resources’. Thus, relative to states with abundant resource wealth, these regimes were more likely to see outside opponents win support among civil society and private economic interests and attract insiders to defect, simultaneously strengthening popular challengers and weakening the cohesion of the regime. In short, large stocks of easily extractable natural resources substantially improved survival rates for autocrats threatened by large-scale unrest.
Implications
Recent decades witnessed the deepening of transnational bonds between national societies, including the appearance of new forms of transnational activism and the diffusion of contention across borders (Bunce and Wolchik Reference Bunce and Wolchik2010; Levitsky and Way Reference Levitsky and Way2010a; Tarrow Reference Tarrow2005). These developments have complicated the efforts of autocrats to maintain political control and have also emboldened domestic dissidents in their struggle against authoritarian regimes. Of course, the transnationalization of contention has not entirely eroded the relevance of physical and cultural space. Following earlier patterns observed during the third wave and even in waves of contention observed in 1848 and 1917–19 (Weyland Reference Weyland2014), powerful outbreaks of unrest over the last several decades have clustered temporally and spatially, constituting regional waves bounded by cultural and physical geography. In the 1970s, popular opposition movements challenged authoritarian regimes that transcended physical geography but were predominantly Catholic and classified as military regimes. This subsection of the third wave began in Portugal’s Carnation Revolution, followed by Spain, six Latin American cases, Mexico and Chile (Huntington Reference Huntington1991; Threlfall Reference Threlfall2008). A smaller regional wave took place in East Asia with the Philippines People Power Movement of 1986, South Korea’s June Democracy Movement in 1987 and Taiwan’s acceptance of a formal opposition party and lifting of martial law in 1986–7. The year 1989 began a cascade of revolutions among communist regimes that began with China’s failed Tiananmen protests and Poland’s successful Solidarity-led revolution and spread rapidly into mass uprisings in locations such as Hungary, East Germany, Czechoslovakia, Bulgaria, Romania and the Soviet Union itself.
As discussed in this article, in the more recent post-communist colour revolutions of the 2000s and the Arab uprisings of 2011, regional waves of unrest were triggered by successful popular challenges in initiating sites such as Slovakia (1998) and Tunisia (2010), which then diffused to regional neighbours, generating large-scale unrest in new sites. As these waves subsided, it became clear that unrest had largely concentrated within culturally similar and geographic regions, having less impact further afield. While the rapid diffusion of popular contention often surprised and overwhelmed the repressive capacities of autocrats, as suggested by Weyland (Reference Weyland2014: 8), popular claimants in neighbouring regimes also often ‘overestimate[d] the evidentiary value of this foreign success and jump[ed] to the conclusion that a challenge to their own government [was] feasible and promising as well’. As such, in both colour revolutions and Arab uprisings, the rapid diffusion of large-scale unrest in new sites had uneven outcomes within their respective regions. Some regimes collapsed in the face of sudden outbreaks of protest and others closed ranks, containing challenges and holding onto power. By looking at new sites of contention connected to regional waves, this article has provided a glimpse at the resilience of authoritarian regimes when suddenly faced with exogenously inspired large-scale unrest.
This exercise has suggested that in recent iterations of regionalized protest waves, autocratic survival has followed several distinct patterns. First, regimes with stable leaderships associated with younger, healthier incumbents were more resilient than those with older and less healthy leaders as well as those undergoing leadership transitions. Second, rentier states with large stocks of readily extractable natural resources were more durable than their resource-poor counterparts. Third, in the Arab uprisings, monarchies proved more resilient than their personalist counterparts. Finally, more politically isolated regimes were less susceptible to popular challenges than those tightly enmeshed with the international community via treaties and international organizations. These findings have provided some clarity to patterns of resilience and breakdown in the midst of regionalized protest waves. Regimes that collapsed generally exhibited certain structural weaknesses – they lacked easily extractable natural resources, had less stable forms of political organization and had dense political linkages to the international community. Regimes containing a number of these structural characteristics, specifically Kyrgyzstan in 2005 and 2010, proved vulnerable to regional unrest even in the absence of leadership transitions.
Other regimes that collapsed contained one or more of these destabilizing structural variables but were also victims of circumstance – collapsing amidst a double coincidence of two largely unrelated events. Regimes such as Georgia (2003), Ukraine (2004), Yemen (2011) and Egypt (2011) were undergoing leadership transitions at the same moment in which a modular form of unrest had been developed in a foreign initiating site and was now spreading across their region. During normal times, an autocrat might be able carefully to choreograph a leadership succession: publicly tapping a successor, building up a supportive coalition within the regime for the anointed successor and distributing threats and rewards to deter internal and external challengers from exploiting a temporary moment of vulnerability. Successful transitions, such as from Syrian President Hafez al-Assad to his son Bashar in 2000, occurred in the absence of regional waves of unrest, meaning these potential windows of opportunity were not exploited by potential challengers backed by large, effective protests. For other less fortunate leaders, exogenous unrest suddenly appeared during unfolding leadership transitions. In Ukraine, an exiting Leonid Kuchma was handing power over to Viktor Yanukovych precisely as opposition activists imported a tested electoral model from abroad and emboldened demonstrators joined the nascent Orange Revolution. In Georgia, Egypt and Yemen, succession questions surrounding ageing and ailing leaders contributed to divisions among regime insiders, ultimately resulting in fragmentation and collapse in the face of new, novel and powerful forms of popular contention imported from abroad. If these exogenous protests had not appeared amidst leadership transitions, demonstrations may have struggled in the face of more cohesive and formidable autocracies, resulting in stronger prospects for regime survival.
These findings suggest that, in line with recent research in contentious politics and authoritarianism, such as Levitsky and Way (Reference Levitsky and Way2010a), Bunce and Wolchik (Reference Bunce and Wolchik2010) and Weyland (Reference Weyland2014), explanations for the outbreak as well as success or failure of popular movements against authoritarian regimes need to integrate both international and domestic levels of analysis. On the international level, when an autocratic regime collapses in the face of protests, this example resonates strongly within neighbouring societies that share cultural similarities but is less likely to do so in venues further afield. Opponents of regimes in these new sites may draw inspiration from these initiating sites and import their tactics and symbols. But structural factors on the domestic level will ultimately determine success. While efforts inspired from abroad may result in large demonstrations, they fail in the face of strong and united autocratic regimes bolstered by stable leaderships, resource wealth and effective forms of political organization. Opposition activists under such circumstances overestimate their chances of affecting regime change, presuming that their own rulers are vulnerable to the same tactics successfully applied abroad. Of course, authoritarian neighbours who lack structural advantages or are caught during a leadership transition may find themselves vulnerable to large protests. The result is elite fragmentation, defections to the opposition and ultimately regime collapse. Consequently, authoritarian survival amidst a regional wave of unrest requires a combination of structural strength as well as fortunate timing.
Table 1 Summarized Results
Notes: Change in GDP per capita is the percentage change in GDP per capita in the year before the protest year. Linkages measure according to the KOF Globalization Index, Oil exports are the fuel exports as a percentage of all merchandise exports. P/SP refers to personalist/single-party hybrid regimes. P/SP/Military refers to personalist/single-party/military hybrid regimes.
