Introduction

Dictatorships are typically viewed as brutally repressive regimes where power lies in the hands of a single individual. Some dictatorships fit this portrait well, such as Iraq under Saddam Hussein and Libya under Muammar Qaddafi. But many do not, including Singapore under the People’s Action Party and Botswana under the Botswana Democratic Party, both of which on most days appear more like modern democracies than the stereotypical dictatorship.

Misconceptions about dictatorships abound at least partially because political scientists have historically paid less attention to them than they have to democracies (Payne, 1996, p. 1187). As political scientist Adam Przeworski noted in an interview in 2007, “dictatorships are by far the most understudied area in comparative politics. We need to start thinking about them” (Munck & Snyder, 2007, p. 473). Because the bulk of comparative research has focused on democratic systems, the literature on dictatorships has lagged behind, only beginning to gain traction in the second decade of the 21st century (Pepinksy, 2014).

One reason for this is that it is significantly more challenging to study dictatorships than democracies. As Paul H. Lewis (1978) writes, “it is more difficult to study dictatorships than democracies because the internal politics of the former are deliberately hidden from the public view. There is no free press, no free public opinion, no open lobbying or party competition” (p. 622). Gathering reliable information in authoritarian contexts is no easy task. Because of the nature of dictatorship, where censorship is common and government-sponsored propaganda pervasive, accurate and reliable data are hard to come by (Magee & Doces, 2014). Though formal rules may exist, informal processes often guide authoritarian politics and major decisions are typically made behind closed doors. In other words, dictatorships are particularly challenging to study precisely because they are dictatorships.

The example of Iran helps illustrate this. Since 1979, the official leader of Iran is the supreme leader. Ayatollah Ruhollah Khomeini held this position until his death in 1989; his successor, Ayatollah Ali Khamenei, has held it ever since. In addition to the supreme leadership post, however, the Iranian regime also features a president who is elected by the people (albeit in contests that fall short of free and fair).¹ During Khomeini’s time in office, power clearly lay in the hands of the supreme leader, but since his death the lines of power have become blurred, with observers of Iranian politics occasionally unsure of who the de facto leader of Iran is: the supreme leader or the president.² At various junctures (such as during Mahmoud Ahmadinejad’s presidency from 2005 to 2013), the president seemed more powerful than the supreme leader (Sahimi, 2011). The Iranian regime is not the only dictatorship where the lines of authority have been unclear. In Russia, though many observers viewed Prime Minister Vladimir Putin to be that country’s leader during Dmitry Medvedev’s presidency from 2008 to 2012, others contested this assertion, citing Medvedev’s efforts to distance himself from Putin while in office (Reuters, 2009).

Unlike in democracies, where decision making procedures are clear and the rules for delegating power are respected, politics in dictatorships usually occurs via informal bargaining. While in democracies the de facto leader of the regime is generally fairly obvious, in dictatorships such basic features of the regime can be difficult to ascertain.

Despite the challenges that autocracies pose for researchers interested in studying them, recent scholarship on dictatorships has illuminated many features of authoritarian regimes, improving our understanding of them. We know now, for example, that dictatorships are not one and the same (Geddes, 2003; Gandhi, 2008a), as the stark contrast between places such as Mexico under the Institutional Revolutionary Party (PRI) and Nicaragua under the Somoza family illustrate. We also know that dictatorships are not synonymous with the leaders who rule them (Geddes et al., 2014; Frantz & Ezrow, 2011). Though in some instances the leader and the regime are indistinguishable, in many the regime lasts well beyond the tenure of any single leader, as in North Korea under the Kim family or in China since Mao. This helps to explain why international pressures on dictatorships to democratize or change their behavior that focus on the leader as the unit of analysis may fail to bring about the intended effects (Escribà-Folch & Wright, 2010). In addition, we know that dictatorships often adopt the same institutions that scholars have historically viewed as quintessential hallmarks of democracies—including elections, parties, and legislatures—even if they have no intention of using them for democratic purposes (Gandhi & Lust-Okar, 2009; Gandhi, 2008a). As a consequence, elections in dictatorships do not necessarily suggest an increase in the democratic nature of the regimes governing them. After all, Saddam Hussein frequently won elections in Iraq with 100% of the vote (BBC News, 2002).

The boom in the study of dictatorships in the 2010s has dramatically improved our understanding of these regimes, but many questions remain and contentious debates persist. There is still a very real need, in other words, for scholars to focus their attentions on authoritarian politics. This is particularly true given that as of 2021 dictatorships governed about 40% of the world’s countries. Though this marks a drop from the Cold War era, when dictatorships often outnumbered democracies, there are indications that authoritarianism is on the rise. According to the watchdog organization Freedom House’s 2020 report assessing the state of political rights and civil liberties worldwide, as of that year democracy had suffered 15 consecutive years of decline, the longest ever since it started measuring these trends more than four decades earlier (Freedom House, 2020). Dictatorships do not appear to be going away anytime soon; if anything, they seem to be experiencing a revival (Ibrahim, 2019, p. 5). Improving our understanding of how politics works within them therefore remains as important a task as ever.

Defining Dictatorship

Scholars have defined dictatorships in a variety of ways. Juan Linz (1964), for example, writes that “authoritarian regimes are political systems with limited, not responsible, political pluralism, without elaborate and guiding ideology (but with distinctive mentalities); without intensive nor extensive political mobilization (except at some point in their development), and in which a leader (or occasionally a small group) exercises power within formally ill-defined limits but actually quite predictable ones” (p. 255). From this perspective, dictatorships are regimes where power is concentrated, political pluralism is restricted, and the masses are not mobilized, a depiction similar to those offered by Samuel P. Huntington and Clement H. Moore (1970), Charles W. Anderson (1970), and Phillipe C. Schmitter (1971), to name a few. Scholars have also emphasized other related factors in their definitions of dictatorships, including lack of representation (Acemoglu & Robinson, 2006, p. 17) and the absence of an electoral process (Brooker, 2000, p. 3). Though definitions of dictatorship range in their focal points and complexity, the operational definition underlying the discussion to follow will be Przeworski et al.’s (2000) simple description of democracies as systems in which “those who govern are selected through contested elections” (p. 15) and dictatorships as “not democracies” (p. 18). Dictatorships, in other words, are regimes in which rulers assume power through means other than free and fair elections (Gandhi, 2008a, p. 7). This discussion will also use the terms authoritarian regime, dictatorship, nondemocracy, and autocracy interchangeably, unless otherwise indicated.

The Evolution of Dictatorships

In many ways, definitions of dictatorship have evolved in line with changes in the nature of dictatorship over time. For example, Przeworski et al.’s (2000) conceptualization of dictatorship pits dictatorships against democracies, yet the latter type of political system only became widely used in the past few hundred years. Throughout history there has been enormous variation in the structure and scope of dictatorships—from the pharaohs of Egypt to the monarchs of Europe—and in the ways in which political observers have theorized about them—from Aristotle’s (2000) Politics to Machiavelli’s (2012) The Prince. While a handful of monarchic dictatorships still exist today, for example, they look quite different than their predecessors did centuries earlier. Today’s monarchies, such as Jordan and Qatar, may still use hereditary forms of leadership succession procedures, but they are also apt to incorporate institutions such as legislatures and elections. As Paul Brooker (2000) points out, modern dictatorships differ substantially from earlier forms of authoritarian rule because monarchs and chiefs are no longer the sole individuals in power. This section will discuss the evolution of the nature of dictatorships and the scholarship devoted to them since the 20th century.

At the turn of the 20th century, stilted attempts at democratic governance around the globe prompted many scholars to reflect on whether democracy was an ideal form of government. The key studies at this time were largely normative in nature, often promoting the merits of oligarchic rule and questioning the feasibility of liberal democracy. The ideas proposed drew heavily from the elite theorists of this era, including Gaetano Mosca, Robert Michels, and Vilfredo Pareto, who believed that all social and political organization was destined to become oligarchical. In this view, concerted political action required rule by the elite given both the intellectual superiority of elites and the disorganization of the masses. Carl Schmitt (1921/2013), for example, picked up on these themes in his seminal 1921 book Dictatorship, which advocated dictatorship as a necessary institution due to the need for governments to have extraordinary powers during times of crisis. In this conceptualization, liberal democracy is a contradiction and therefore impossible to implement. Other scholars of the time took a slightly more nuanced approach, for example Emilio Rabasa, whose 1912 analysis of Mexican politics put forth that authoritarian rule is often a necessary prelude to a truly liberal regime. According to Rabasa (1912/1976), the dictatorships of Mexican leaders such as Benito Juarez and Porfirio Diaz arose due to a poorly conceived constitution that contained unrealistic limitations on executive authority.

Following World War II, international exposure to regimes such as Adolf Hitler’s Nazi Germany and the Soviet Union under Josef Stalin led scholars of authoritarian politics to focus on a new concept: totalitarianism. Totalitarian regimes are led by an individual at the helm of a single political party with a highly developed ideology and powerful secret police (Huntington, 1991). The government maintains its control over citizens through its reliance on propaganda, which it uses to try to create the illusion of an ideal society. In totalitarian systems, there are no limitations to the use of state power. Hannah Arendt’s (1951) seminal work on totalitarianism views totalitarian regimes as extreme forms of dictatorship exercising control over “atomized, isolated individuals” (p. 323). Arendt emphasizes the role of ideology in driving the totalitarian system, where the goal is to fundamentally transform society in line with the government’s ideological leanings through the use of terror. Carl Friedrich and Zbigniew Brzezinski (1956) pick up on these themes and identify six features of totalitarian regimes: the implementation of an official ideology, a single political party, party control over mass communications, party control over the military, a centrally controlled economy, and a secret police.

The concept of totalitarianism accurately describes a number of dictatorships that dominated news headlines in the 20th century, including the Communist regimes in Eastern Europe and Asia. Other dictatorships soon grabbed scholars’ attention, however, where ideology was not the central method of control and societal transformation was not the end goal, such as in Spain under Francisco Franco. Linz (2000) labels these regimes “authoritarian” and argues that, in stark contrast to totalitarian systems, they seek to demobilize and depoliticize their citizens. Because there appeared to be many additional factors that differentiated dictatorships, scholars slowly moved even further away from the focus on ideology and in turn the use of the term totalitarianism (Snyder & Mahoney, 1999).

For example, a wave of independence movements in the 1950s and 1960s brought to power one-party dictatorships in places ranging from Kenya under the Kenya African National Union to Singapore under the People’s Action Party. Regimes such as these featured a dominant political party but looked very different from Nazi Germany or the Soviet Union. Building on these developments, work by Huntington and Moore (1970) disaggregates dominant-party dictatorships based on the type of political party in power. For example, in strong one-party states, the party is supreme. In weak one-party states, by contrast, other actors such as the leader, the military, or the police often eclipse the role of the party. Huntington and Moore assert that the strength of the regime party can be assessed by analyzing the intensity and duration of its struggle to assume power. They further disaggregate strong one-party states by their level of institutionalization, as well as the types of survival strategies the party employs. Revolutionary one-party systems, for example, are characterized by “social dynamism, autocratic and charismatic leadership, disciplined party, highly developed ideology, stress on propaganda and mass mobilization, combined with coercion and terror” (Huntington & Moore, 1970, p. 23). Established one-party systems, by contrast, are more bureaucratic in nature. Power is less personalized, and there is less emphasis placed on mobilizing the masses.

The 1970s saw a spike in the number of dictatorships led by the military, such as those in Brazil (1964–1985) and Chile (1973–1989), largely due to Cold War geopolitics (Geddes et al., 2014).³ In line with this development, a number of scholars focused their attentions on the particular features of military rule (O’Donnell, 1973). Amos Perlmutter (1969, 1977), for example, identifies two types of military dictatorships: rulers and arbitrators. The ruler type of military dictatorship seeks to maximize its power and sees civilian politicians as a danger to stability. The arbitrator type, by contrast, is not interested in ruling for long periods of time but rather wants to restore to the country order and stability.

Concurrently, the emergence of strongman dictators in a number of places, such as Mobutu Sese Seko in the Democratic Republic of the Congo and Jean-Bédel Bokassa in the Central African Republic, prompted scholars to delve into the dynamics of one-man rule, or personalist dictatorship. Like the Latin American “caudillos” who dominated that region’s political landscape decades earlier, these regimes feature a single, often charismatic, man at the helm of the political system.⁴ As Samuel Decalo (1985) puts it, personalist dictatorship is “an authoritarian system of social repression set by an individual—civilian or military—in which, whether social or political structures are pro forma retained or not, all policy dictates derive from him and all of society is viewed as his personal fief” (p. 212). Unlike in other dictatorships where a party or military may wield political influence, in personalist dictatorships the leader rules unchecked by other actors. The system is structured by politicians rather than institutions (Bratton & van de Walle, 1997, p. 62; Acemoglu et al., 2004, p. 167).

Greater diversity in the manifestations of dictatorship, coupled with greater awareness of the full scope of the authoritarian regimes in power, sparked more comprehensive efforts to categorize them. The section that follows therefore turns to a discussion of the major typologies of dictatorship.

Typologies of Dictatorship

Since the late 20th century a number of typologies have emerged to capture differences in autocratic rule considered theoretically relevant by their authors. These typologies fall into two categories: continuous and categorical. Because a number of them promote concepts that scholars can measure cross-nationally, their development has facilitated comprehensive analyses of the consequences of different types of dictatorship for a range of political outcomes, including conflict behavior, economic performance, and regime transitions.

Continuous typologies of dictatorship disaggregate regimes according to how “authoritarian” they are. They focus on the different gradients of dictatorship, the idea being that political systems can be placed along an autocratic–democratic scale. A number of scholars claim, for example, that many dictatorships lie in the middle of this continuum, using formally democratic institutions that conceal a “reality of authoritarian domination” (Diamond, 2002, p. 24). Scholars have referred to these regimes as grey zone (Diamond, 2002), competitive authoritarian (Levitsky & Way, 2010), and electoral authoritarian (Schedler, 2006). Dictatorships such as these are neither fully authoritarian nor fully democratic. Though they often hold elections in which opposition parties are allowed to compete and occasionally win legislative seats, electoral contests are not fully competitive because the electoral playing field is not even. Incumbents have access to a bevy of state resources to help secure their own victories. They control the media and can deny the opposition access to it, they can use the security apparatus to harass and intimidate members of the opposition, and they can manipulate the electoral rules and outcomes in ways that tilt outcomes in their favor.

A number of data sets have emerged in recent years that enable scholars to place dictatorships on a continuous scale with democracies, including the widely used Polity data set (Polity, 2020), the Varieties of Democracy data set (Coppedge et al., 2021), and Freedom House’s (2020) measures of political rights and civil liberties, to name a few.

Unlike continuous typologies of dictatorship, categorical typologies differentiate regimes based on specific dimensions, enabling scholars to avoid making any assumptions about the linearity of the path from dictatorship to democracy. Though a few scholars have developed categorical typologies of dictatorship that reflect differences in the strategies of the dictator (see, e.g., Haber, 2006; Wintrobe, 1998), most categorical typologies focus on differences in the structure of the regime.

The typology developed by Barbara Geddes (1999, 2003), for example, categorizes dictatorships based on differences in the identity of the group that controls leadership selection and policy choices. Geddes differentiates regimes according to whether this group consists of a single individual and his supporters (i.e., personalist dictatorships), military officers (i.e., military dictatorships), or a dominant party (i.e., single-party dictatorships). She argues that political actors in these different institutional environments behave differently, leading to differences in their regimes’ survival rates and propensities for democratization.⁵ A number of studies have used the Geddes typology to explain variations in policy outcomes across dictatorships, including their propensity for conflict (Peceny et al., 2002), reliance on repression (Davenport, 2007), and use of foreign aid (Wright, 2008).

Some scholars note that because many dictatorships exhibit some level of personalism, personalism should be viewed as a feature of dictatorships, rather than a unique category (Hadenius & Teorell, 2007). The typology proposed by Hadenius and Teorell (2007) and updated by Hadenius et al. (2012) seeks to address this concern. It categorizes dictatorships as monarchic, no-party, military, one-party, or multiparty, relying on formal institutions to distinguish regimes, especially those governing the number of legal political parties. Geddes et al. (2018) build off of these ideas in their personalism index, which captures different levels of personalism across dictatorships and within them over time.

An additional widely used typology, first introduced by Mike Alvarez et al. (1996) and further developed by Jose Cheibub et al. (2010), categorizes dictatorships based on the type of leader in power, specifically whether the leader is a monarch, a member of the military, or a civilian. Whereas civilian dictators do not rely on a preexisting organization to govern (and in turn typically turn to a political party), monarchs rely on the royal family and military dictators rely on the armed forces. The motivation underlying this typology is that differences in the nature of the leader bear on the incentives leaders face and in turn their strategies and prospects for survival in office. Scholars have used these data to explain patterns across dictatorships in conflict initiation and the posttenure fates of leaders (Debs & Goemans, 2010), their exchange rate policies (Steinberg & Malhotra, 2014), and their economic performance (Gandhi, 2008b).

The development and expansion of the typologies put forth to capture differences across dictatorships have improved the ability of scholars to answer key questions about authoritarian politics. They have also triggered debates among scholars regarding which among them offers the most accurate representation of authoritarian politics (Cheibub et al., 2010; Kailitz, 2013; Lidén, 2014; Wahman et al., 2013). The critical message that has emerged from these discussions is that scholars should pay close attention to ensure that the theoretical concepts that they are interested in are reflected in the typology they choose. Because there are advantages and disadvantages to each, the appropriate typology to use is context dependent.

The proliferation of autocratic typologies reflects an expansion in the ways in which scholars conceptualize authoritarianism. Dictatorships are no longer viewed as monolithic political systems, marking a major advance in our understanding of the authoritarian world. With this in mind, the sections that follow address contemporary themes of authoritarian politics.

The Causes of Dictatorship

Early research on the causes of dictatorship focused on the origins of totalitarianism, in line with this literature’s emphasis on this particular form of autocratic rule. Arendt (1951), for example, argues that totalitarianism emerges in places where the rigors of extreme levels of individualism in capitalist societies draw people to a totalitarian ideology. Linz (2000) also sees totalitarianism as a product of class and ideological conflicts. Totalitarian movements spread in response to the forces of capitalism, which has failed to incorporate the masses into the political process.

Following World War II, researchers turned their attentions not to the causes of dictatorship but to the conditions favorable to democratization. Huntington (1968), for example, identifies institutional conditions as particularly important. In his view, modernization and mass mobilization, in conjunction with underdeveloped political institutions, are barriers to democratization. Governments that are poorly institutionalized—as evidenced by high levels of corruption, lack of civil associations, and a blurring of private and public spheres in government—make democracy unfeasible. In a similar vein, George Kahin et al. (1955) highlight a set of features that are roadblocks for democracy, including charismatic leadership and “a tendency for unorganized and generally inarticulate segments of society, such as peasants and urban masses, to involve themselves in politics in a discontinuous, sudden, erratic and often violent way” (p. 1026).

Other researchers emphasize economic and demographic factors as critical to understanding where democracy is likely to emerge. Seymour Martin Lipset (1959), for example, argues that urbanization, industrialization, education, and a healthy population together lead to pressures for democratization. This “modernization” of society leads to a change in the mindset of citizens, paving the way for democracy. As an indicator of this, a number of studies have established that there is a positive and robust relationship between democracy and economic development: as countries grow richer they are more likely to be democratic (Bollen, 1979; Burkhart & Lewis-Beck, 1994; Jackman, 1973). Relatedly, the most important predictor of a new dictatorship sprouting is poverty (Londregan & Poole, 1990, 1996). Przeworski et al. (2000) find, however, that although the proportion of democracies is higher in rich countries than in poor ones, there is little evidence of a causal relationship. In other words, democracy is correlated with wealth (or what Lipset refers to as “modernization”), but wealth does not cause democratization. The implication is that political stability—whether in the form of long-lasting democracy or long-lasting dictatorship—is more likely in rich countries. Countries that are poor may democratize, but their experience with democracy is likely to be short-lived. Przeworski et al.’s (2000) findings improve our understanding of the experiences of a number of poor countries, such as Haiti and Guinea-Bissau, that have been led by a series of short-lived regimes (both democratic and authoritarian), as well as the experiences of a number of rich countries, such as Qatar and Bahrain, that have been governed by remarkably robust dictatorships for decades.

Researchers have suggested a number of arguments to explain the causes of dictatorship and the roadblocks to democracy more specifically. The major message to emerge from this literature is that long-lasting democracy is more difficult to achieve absent economic prosperity.

Future research is needed, however, to better grasp the conditions that foster the emergence of one type of dictatorship over another. For example, a number of recent studies have shown that personalist rule is associated with a host of negative outcomes. Compared with dictatorships where leaders face constraints on their rule, personalist dictatorships are more likely to become embroiled in military disputes with democracies (Peceny et al., 2002; Reiter & Stam, 2003), initiate wars (Peceny & Butler, 2004; Weeks, 2012), and pursue a risky and aggressive foreign policy agenda (Frantz & Ezrow, 2011). They are also the least likely form of dictatorship to engage in cooperative behavior internationally (Mattes & Rodriguez, 2014). They are the most corrupt of all dictatorships (Golden & Chang, 2010), some of the most repressive (Davenport, 2007), and some of the most prone to civil war (Fjelde, 2010). At the same time, the evidence indicates that they are increasingly common (Geddes et al., 2014). Not only are new personalist dictatorships sprouting around the globe in countries as varied as Turkey, Benin, and Nicaragua, but we are also seeing increases in personalism in existing authoritarian regimes, such as in China under Xi Jinping (Frantz, 2018). The rise in personalist rule globally, coupled with its pernicious consequences, highlights the importance of identifying methods for preventing it. Research suggests that more fractured and less institutionalized seizure groups facilitate the emergence of personalist rule after the seizure of power (Geddes et al., 2018), but future research is needed to better understand these dynamics, as well as the conditions that steer once-collegial autocracies toward a path of strongman rule. Gaining insight into vulnerabilities to personalism remains a critical task.

Authoritarian Survival

Every political leader confronts the challenge of remaining in office.⁶ This is particularly true for dictators, who cannot rely on the legitimacy conferred by free and fair electoral victories to defend their rule (Gandhi & Przeworski, 2006). Political survival in dictatorships depends not on maintaining the favor of voters, as in democracies, but on securing the backing of a considerably smaller coalition of supporters (Svolik, 2012).

Bueno de Mesquita et al.’s (2003) selectorate theory, for example, identifies two key groups that influence leaders’ political survival: the selectorate (a subset of the population that has a say in the selection of the leader) and the winning coalition (a subset of the selectorate large enough to maintain a leader in power). The leader’s position is the most secure when the selectorate is large and the winning coalition is small, partly because the costs of defection are high in such situations, but also because members of the winning coalition can easily be replaced by members of the selectorate. A key insight of this theory is that not all citizens have influence over the leader. Leaders will cater their policy choices to meet the needs of the coalition of individuals whose support they require to stay in office.

The composition and scope of this coalition varies from one dictatorship to the next. In some, it is comprised of the dictator’s friends and family, while in others it may be made up of members of the party or military upper echelons (Geddes, 2003).

Politics in dictatorship is the product of bargaining between leaders and this coalition of elite supporters (Frantz & Ezrow, 2011). Yet the absence of a reliable third party to enforce compromises among these players means that power-sharing deals lack credibility, and the threat of forced ouster is omnipresent (Svolik, 2012). Uncertainty is therefore a pervasive feature of authoritarian politics.

Modern autocrats respond to this uncertain environment in a variety of ways. They use traditional tactics such as repression to maintain control, but reliance on coercive methods can be costly to maintain. For this reason, they often incorporate cooptation as well to protect themselves from overthrow (Wintrobe, 1998), meaning the intentional extension of benefits to potential challengers in return for their loyalty (Corntassel, 2007). In addition to patronage, dictators use institutions as tools of cooptation, often in the form of political parties, legislatures, elections, and other institutions typically associated with democracies. Though pseudo-democratic institutions such as these create positive levels of risk for dictatorships (Cox, 2010), the survival benefits they confer appear to outweigh them.

Elections, for example, are widely viewed as a means through which dictatorships maintain power (Gandhi & Lust-Okar, 2009). Most dictatorships hold them, though they vary in terms of the regularity with which they are held, their competitiveness, and their scope. Elections provide incumbents the opportunity to mobilize their supporters, sending a message to potential challengers that attempts to unseat them are unlikely to be successful, thereby lowering the risk of coup (Cox, 2010; Geddes et al., 2018).

Support parties fulfill a similar purpose in dictatorships. Beyond serving as vehicles to mobilize popular support, support parties give their members a vested interest in the regime’s survival, making them less likely to join subversive coalitions or back efforts to seek the regime’s overthrow (Magaloni, 2006). Support parties may also allow dictators to credibly commit to sharing power and the perks of office with those who belong to them (Magaloni, 2008).

Likewise, it is argued that legislatures confer survival benefits for dictatorships, specifically partisan legislatures in which members of opposition parties are allowed to hold seats. Partisan legislatures prolong regime survival by creating a forum for incorporating potential opposition forces into the system, broadening the support base of the regime apparatus and lowering the chance of coups (Gandhi & Przeworski, 2007; Boix & Svolik, 2007). Partisan legislatures provide an arena through which dictators can offer potential rivals policy concessions and bargain over the terms of these deals. Rivals, in turn, use legislatures as a tool for monitoring the dictator and ensuring he is upholding his end of the bargain (Boix & Svolik, 2007).

Despite the façade of democracy, these institutions are key components of most autocrats’ survival strategies; those that incorporate them last longer in power than those that do not (Gandhi, 2008a; Gandhi & Lust-Okar, 2009). Dictatorships appear to have caught wind of this message: since the end of the Cold War, most dictatorships have featured regular elections, multiple political parties, and a legislature (Kendall-Taylor & Frantz, 2014).

Scholars have yet to fully uncover, however, the specific vulnerabilities of these heavily institutionalized regimes.⁷ Researchers have identified a number of factors that shape the risk of autocratic regime collapse, including the internal balance of power (Moore, 1966), economic performance (Haggard & Kaufman, 1995; Przeworski & Limongi, 1997), income distribution (Boix, 2003), the identity of the autocratic ruling group (Geddes, 2003), and the regime type of neighboring countries (Gleditsch & Ward, 2006). But how do these factors interact with dictatorships’ increasing reliance on institutions to ensure regime survival? What conditions make even heavily institutionalized regimes vulnerable to collapse, as occurred in Malaysia in 2018 following many decades of authoritarianism? Moreover, how can we better unpack the category of authoritarian institutions to move beyond the current emphasis on parties, elections, and legislatures? Importantly, has the widespread use of institutions created new risks for dictatorships? Has it altered the ways in which they use their coercive apparatuses? Are we seeing autocratic repression evolve in line with this institutional evolution? Future research is needed to address important questions such as these.

The Road Ahead

In ancient Rome, the challenges posed by military emergencies created incentives for unity of command. The Romans’ solution for this problem was to appoint a dictator in place of the consuls. The utility of dictatorship is reflected in its persistence as a form of government. Dictatorships have dotted the worlds’ political landscape for thousands of years, a trend that appears likely to continue in the years to come.

Despite the pervasiveness of dictatorships, scholars know relatively less about authoritarian regimes than they do about their democratic counterparts. The literature on authoritarian politics has expanded greatly in the 2010s, but there is still a lot of catching up to do. A number of key questions remain unanswered. Among them are: What sets of factors lend themselves to the formation of strongman as opposed to collegial autocratic rule? Why do some institutions in dictatorships work to constrain leadership choices to prevent consolidation of power into the hands of a single individual, while others do not? What are the specific vulnerabilities of today’s heavily institutionalized dictatorships to regime collapse? How is the institutional and coercive toolkit in autocracies evolving? Answering questions such as these requires further theoretical attention to the underpinnings of authoritarian politics, including who the key political actors are, the nuances of the institutional environments they operate in, the tools they use to enhance political survival, and how the incentives they face may change.

At the same time, evaluating theories of authoritarian politics requires systematic data. Though scholars have developed a number of new data sets in recent years to enable this, high-quality cross-national data capturing a number of characteristics of dictatorships—including the nature and scope of opposition networks and the size and composition of elite support groups—are not yet publicly available.

Developing a deeper theoretical and empirical grasp of the key features of dictatorships is a challenging task given the black box of decision making and the restricted political space that characterizes most of them, but one that will offer insight into many of the contemporary puzzles that currently exist in the study of authoritarian politics.