Authoritarian Regimes

This article does not intend to contribute to the very broad literature on authoritarian regimes.¹ Rather, its first section aims at clarifying some basics (What are authoritarian regimes? Why study them?), and the second section, the bulk of this article, is divided into eight subsections to present a structured overview of the field and a critical discussion of the main debates. This rough and cautiously commented on overview of the evolving field demonstrates the extent to which research has not only increased output dramatically but also broadened thematically since the early 2000s. On this basis, the conclusions take a step back and reflect on the state of the art. I argue that there is a need for some emerging challenges to be addressed by future scholarship, as we otherwise risk producing research output without necessarily deepening our understanding. Working to address such challenges will thus shape the future research agenda for students of authoritarian regimes—or at least it should.

In political science, authoritarian regimes, since roughly the mid-twentieth century and along with democratic and totalitarian regimes, have constituted one of three classical types of political rule. Authoritarianism, totalitarianism, and democracy have been referred to as “the classical triad” of systems of governance or political regimes.²

Unlike the essentially contested concept³ of “democracy,” the definition of authoritarian regimes advanced by Juan Linz more than half a century ago is still broadly accepted by the scholarly community. According to Linz, authoritarian regimes are characterized by four defining traits: (1) limited, nonresponsible, political pluralism; (2) the absence of an elaborate and guiding ideology, having instead “distinctive mentalities”; (3) the absence of both intensive and extensive political mobilization; and (4) the exercise of power within formally ill-defined, but actually quite predictable, limits by a leader or a small group (Linz 1964, 297; for a more elaborate discussion, see Linz 1975, 275–411). This characterization of authoritarian regimes captures them as one type of autocracy, a broader term that also includes other nondemocratic political regimes that are nonauthoritarian such as totalitarianism or, in Linz’s (and Chehabi’s) understanding, sultanism (Chehabi and Linz 1998).

Although some authors have suggested alternative definitions of authoritarianism, none has achieved the scholarly consensus that Linz’s classical definition has reached. It will probably take much more time for new and potentially more conceptually sound propositions to stand a chance of arriving at similar levels of agreement within the academic community.

Authoritarian regimes must be distinguished from “authoritarianism,” which is a significantly broader concept. The latter comprises not only a specific type of political regime, but also a broader understanding of “authoritarian” individual (human) and collective (social) traits, themes, situations, behaviors, modi, and practices that can be found across a range of academic disciplines beyond political science, such as psychology, business administration, management, sociology, media and mass communication, cultural studies, or anthropology.

Beetham (2015, 12) draws this distinction between authoritarian regimes and authoritarianism, in which the latter for him depicts a “mode of governing which is intolerant of public opposition and dissent.” If that is the case, an “authoritarian mode of governing is possible within a democratic system, though it only remains democratic so long as elections are genuinely ‘free and fair’, and formal civil and political rights are respected” (12). Whether or not this holds up against Linz’s definition need not be discussed here, as this article focuses on the study of political regimes only.

Contrary to the frame that dominated late twentieth-century research on political regimes—that is, the assumption that the world faced a democratic “end of history” (Fukuyama 1989, 1992; cf. also Fukuyama 2014)—authoritarian polities seem to be here to stay for the foreseeable future. Fukuyama’s hypothesis triggered a wave of almost immediate and often fundamental criticism,⁴ which suggests that his hypotheses have proven short lived, and that the “end-of-history” hypothesis marks not an end to, but only a by now long-passed point in, academic debates about political rule and how it is exerted. For the past ten years Freedom House’s analysts have decried every single year a further weakening of democracy and/or a backlash of authoritarian regimes. In a nutshell, thus, one compelling reason for studying authoritarian regimes is the sheer empirical necessity: we need to study authoritarian regimes because they exist.

A theoretical necessity is the second—and until now largely neglected—reason: While authoritarian regimes continue to exist in no small numbers and historically have been the rule rather than the exception, academic debates on the topic and scholarly interest have been overwhelmingly dominated by research agendas established (and by research conducted by scholars residing and publishing) within democracies.

Among the consequences of this fact is that while the systematic study of authoritarianism can be said to have started more than half a century ago with Linz’s seminal 1964 article, this older literature was dwarfed by the vast conceptual, theoretical, and empirical literature on democracy. Linz’s 1964 piece on Spain under Franco, when he first conceptualized the authoritarian regime, was followed by a number of works written mostly by specialists on Latin American politics (cf., e.g., O’Donnell 1973; Malloy 1977; Collier 1979; and Perlmutter 1981). Along with the four volumes edited and in part written by Linz and Stepan (1978), the Breakdown of Democratic Regimes, this older literature can be seen as the “formative” or “constitutive phase” of research on authoritarian rule. However, comparatively very little time and effort were invested in researching authoritarian rule after this formative period and through the beginning of the twenty-first century.

Another effect of the preoccupation of scholars with democracy rather than autocracy was that students of authoritarian regimes were mostly concerned with either the question of how to best get rid of them and transform them into normatively more desirable democracies (e.g., O’Donnell, Schmitter, and Whitehead 1986; Di Palma 1990; as well as large parts of the “transitology” literature) or the question of why certain countries and regions had not been reached by Huntington’s famous “waves of democratization” (cf. Huntington 1991); that is, what puzzled scholars was the fact that “yet they persist” (Brownlee 2002).

After the “constitutive phase” of research on authoritarianism (from the 1960s to the early 1980s), however, little effort was devoted to examining authoritarian regimes in their own right, to conceptualizing and analyzing such regimes as a distinct form of political rule and a specific mode of governance. The teleological lens of democracy as the ultimate point of reference seems to have been necessary to justify academic interest in nondemocratic regimes.

More recently, new research on authoritarianism has emerged. At least in major parts, this new research promises to overcome the persistent, albeit latent, democracy bias that has long existed even in studies on authoritarianism.⁵

Apart from the constitutive phase’s limited writings on authoritarian regimes, the new century saw the proliferation of a by now extremely broad and comprehensive body of literature. In fact, “the study of authoritarian regimes has recently become one of the hottest subfields in comparative politics” (Art 2012, 351). By 2017 it still was the fastest growing area in the subdiscipline.

What follows is a structured overview of some of the most important research themes of the past two decades: (1) typologies and regime characteristics; (2) institutionalist approaches to the study of authoritarianism, which have been for the most part focused on formal state institutions; (3) state-society relations beyond state institutions, including questions of legitimacy, civil society, and the media; (4) repression; (5) political-economic arguments for the survival or breakdown of authoritarian rule; (6) international dimensions of authoritarianism; (7) the performance of authoritarian regimes; and (8) the nexus between the authoritarian regime and the state.

In any new field of research, taxonomic efforts come first. In the case of authoritarian regimes, this has been particularly complicated because of a conceptual uneasiness with the empirical “grey zone” between democracies and nondemocracies. To analytically capture this grey zone of regimes that fit common definitions of neither democracy nor authoritarianism required extensive efforts at “disentangling political regimes” (Munck 1996). One attempt at sub-typologizing democracies that are found to be less than “fully” democratic is through the invention of “diminished subtypes” (conceptually, see Collier and Levitsky 1997; cf. also Collier and Levitsky 2009; Zakaria 1997; and O’Donnell 1994 as early contributions along this methodological approach; for an alternative conceptualization, see Merkel 2004; for a critique, see Møller and Skaaning 2010). A similar method of typologizing consists of classifying grey zone regimes as “authoritarianisms with adjectives” rather than as democracies (competitive, liberalized, semi, etc.; see Levitsky and Way 2002, 2010; Brumberg 2002; Ottaway 2003; and Schedler 2002, 2006).⁶ Diminished subtypes do not comprise all the definitional characteristics of the superordinate class. Rather, the adjective that is added in front of the regime type (or “root concept”) specifies what these democracies lack in comparison to “full,” “real,” or “liberal” democracies, or what makes such autocracies with adjectives stand out from other authoritarian regimes.

Rather than subsuming grey zone regimes under the root concept of either democracies or authoritarianisms, the concept of “hybrid regimes” inserts a fourth basic type of political rule into a previously tripartite regime typology (totalitarianism, authoritarianism, democracy). It locates such “hybrid regimes” between democracy and authoritarianism as a distinct regime type (Karl 1995; Rüb 2002; Zinecker 2004; Wigell 2008; Gilbert and Mohseni 2011). All these discussions take place on a macro level of the “root concepts” of political regimes (Sartori 1970) and deal with (a) the question of how to grasp the grey zone and (b) the question of where exactly the boundary between democracies and nondemocracies is to be drawn.

It was with Geddes’s (1999) distinction among different subtypes of authoritarian regimes (military, personalist, and single-party) that the search for typological clarity within the spectrum of nondemocratic regimes gained new momentum; the aim here was to find out about survival prospects for various types of authoritarian regimes.⁷ Geddes’s lean tripartite typology is easily applicable, in particular for macro-quantitative purposes and the creation of larger data sets. Brooker (2009, ch. 2) suggests essentially the same categories, while Ezrow and Frantz follow Geddes’s typology but add the category of monarchies “to give a fuller picture” (2011, xvi; cf. also Frantz and Ezrow 2011). It is unclear, however, why monarchies, if indeed they are not merely representative-democratic ones, would not be captured by the personalist subtype. Also, it is not self-evident that “personal authoritarian regimes” include all sorts of personalist forms of nondemocratic rule.⁸

Various data sets on authoritarian regimes have been constructed. The three most prominent are by Geddes et al. (2014b; cf. also 2014a); Hadenius et al. (2012; an update of which is by Wahman et al. 2013); and Cheibub et al. (2010). Although they resemble each other, they differ in their conceptualization of authoritarian regimes as well as in the subtypes they recognize (i.e., in the criteria they identify for definition, as well as in the number of subtypes and the definition itself). “Military,” “royal,” and “civilian” are the autocratic subtypes in Cheibub et al. (2010), whereas “military,” “monarchical,” “party-based,” and “personal” are those of Geddes et al. (2014a). Hadenius et al. (2012) use a different theoretical underpinning of subtypes by not following the identity of incumbents but rather types of underlying institutions, yet the subtypes they arrive at (“military,” “monarchy,” “one-party authoritarian,” “multi-party authoritarian,” and “no-party authoritarian,” plus “other”) come very close to those of the other data sets (cf. also Roller 2013). Such data sets are valuable because they are a precondition for engaging in diachronic studies and for running time series, as well as for testing hypotheses that view regime subtypes as an explanatory variable. Overall, while it seems that a consensus on authoritarian subtypes is emerging, a more critical question is whether these comply with the meta-methodological rules of class building.

Divergent origins lead to the formation of different types of authoritarian regimes. This is the argument advanced by a range of authors, although they do not agree about the exact patterns that make authoritarian regimes emerge or about the types that result from different modes of state formation. Slater’s (2010) comparative historical analysis of Southeast Asian cases sees various kinds of contentious politics as leading to either “provision pacts” or “protection pacts,” with the latter having greater prospects for durability (Singapore and Malaysia are the cases in which Slater sees such pacts emerging out of post–World War II mass mobilization). Bueno de Mesquita et al. (2003), by contrast, elaborate what has become widely known as “selectorate theory,” building on earlier works on coalition building in democratic polities, while Brownlee (2002) finds extensive neopatrimonial ties causal for regime survival and restabilization in fifteen cases. Brownlee thus also refers to domestic regime- and state-society-related variables in explaining regime outcomes.

Both Slater and Bueno de Mesquita build on literature that has been adapted from the study of democracy/democratization to the study of authoritarianism/authoritarianization.⁹ Alternatively, Solt (2012) offers a “relative power theory” to explain why prolonged and increasing social inequality leads to authoritarianism: Citizens then tend to regard hierarchical relations as natural and thus accept authority more readily. Yet this contradicts other authors who focus on various types of elite bargains or winning coalitions. Acemoglu and Robinson (2006) also look at strategic elite behavior, but in relation to class constellations. For them, dictatorship is a function of the interplay of wealth and inequality, with both very low and very high inequality making the emergence of authoritarianism more likely than medium levels.

Strikingly, a large share of the literature on origins, as well as on the question of durability of authoritarian regimes, focuses on questions of power sharing among elites and/or between an imagined ideal-type “leader” and elites (cf. Boix and Svolik 2013; Svolik 2012; Bueno de Mesquita et al. 2003; and ultimately also Slater 2010).¹⁰ In light of the colored revolutions in the former Soviet Union and Balkans, the upheavals of 1989–1990, and the Arab spring, it might not be so clear that threats from within are actually greater for the survival of dictators than those that originate from the masses; this is so also in light of the findings of works like Slater’s (2010), in which regime breakdown and re-formation starts with mass contention.

Contributions on elite politics that extend beyond the question of state or regime formation have also resulted in important intraregional or general insights (cf., e.g., Perthes 2004 on Arab countries and Bo 2007 on China). Depending on the nature of the regimes, elite conceptualizations other than “ruling” or “winning coalitions,” such as those of concentric circles, seem to be better able to empirically capture “politically relevant elites” (Perthes 2004) and to present a fuller picture of who exactly influences not only decisions but also agendas.

Without doubt, the largest share of the new research on authoritarian regimes has focused on the role of formal and often democratic-looking institutions, and this focus has been particularly (and strikingly) popular among US- and US-based scholars. Institutions examined include (competitive and noncompetitive) elections; parliaments; oppositions; and political parties, both single parties as well as parties in multiparty systems (to name just a few of the vast flood of literature, cf., e.g., Lust-Okar 2001, 2005, 2006; Lust-Okar et al. 2011; Gandhi 2008; Gandhi and Przeworski 2007; Gandhi and Lust-Okar 2009; Magaloni 2006, 2008; Magaloni and Kricheli 2010; Shambaugh 2008; Donno 2013; Brownlee 2007a, 2007b; and many others).

To this we might add other institutions such as courts or constitutions in authoritarian contexts (e.g., the edited volumes by Ginsburg and Moustafa 2008 and Ginsburg and Simpser 2013), about which one of the core research interests is the fine line between the rule of law and “rule by law.” Proof has yet to be furnished, however, that a politically independent judiciary can actually exist within nondemocratic regimes, or in other words, that the rule of law is compatible in principle with authoritarianism.

Although notoriously difficult to study, there is also a recently renewed interest in civil-military relations and the role of armies in autocracies, in particular in situations of coups and during revolutionary uprisings, along with questions of coup proofing. This research has regional origins in studies on South and Southeast Asian countries such as Myanmar, Pakistan, the Fijis, and Thailand on the one hand, and Middle Eastern cases such as Egypt, Syria, Bahrain, Yemen, and Tunisia on the other (e.g., Barany 2009, 2016; Croissant 2013; Bellin 2004; Springborg 2015; Albrecht and Ohl 2016). While scholarly access to key decision makers almost never exists, heuristic categorization of professional versus nonprofessional armies or deductive principal-agent modeling remains the most common approach to tackling these questions.

Any number of correlations between institutions and autocratic survival, durability, reproduction, plurality, and other variables have been assumed and detected. Oppositional forces, whether “loyal” or “anti-systemic,” have also been examined for their intentional and unintentional contributions to authoritarian survival (cf Albrecht 2006). The bottom line in all of them is that institutions that bear the same name as their democratic counterparts decisively fulfill different functions, and produce dissimilar output, than those institutions in democratic contexts. This gives credence to the assumption that dictators actually engage in institution building for a variety of reasons, but all of them aim at enhancing autocratic rule.

What makes those contributions valuable is that they shed light on a long-neglected or misread component of authoritarian regimes. Formal democratic-looking institutions had been neglected because either they were considered mere rubber stamps that were at best created to maintain some semblance of democracy for external (Western) audiences and thus facilitate the acquisition of aid, or this semblance of democratic institutions was taken at face value and not further questioned. Yet even if such institutions as parliaments, political parties, and (unfree) elections were mere window dressing, and when citizens would not believe in them, why would autocrats then invest in having them? Many possible answers have been suggested; scholars assume them to be anything from opinion barometers, to patronage machines, to facilitators of power-sharing arrangements, to instruments that enable regimes to acquire some form of nondemocratic responsiveness (on the latter, cf. Miller 2014).¹¹

What remains conspicuously absent from this voluminous research, however, are informal institutions. They seem to have been relegated to the back burner and largely omitted. Art (2012, 355) surprisingly considers this not a lack, but “a welcome correction to a literature that has focused overwhelmingly on patronage.” Have informal institutions thus lost their importance? Most probably not, but they are much harder to study.

In the 1990s there was a largely transitology-inspired euphoria about the anticipated role of the “third sector.” Civil society organizations in particular were thought to be a vehicle for democratization, and in fact the prodemocratic functions of voluntary intermediate organizations in established democracies along the lines of Putnam’s neo-Tocquevillean argument (civic traditions and institutions as promoters of social capital and of democratic practices and values such as engagement or toleration; cf. Putnam 1993) remain little contested. But the picture became more serious in the 2000s regarding civil society under authoritarian conditions; autocrats responded to the upsurge in civil societies through a strategy of (a) co-opting organizations that emerged outside of governmental control and (b) deliberately creating a mimicry of look-alike institutions through processes of “imitative institution-building” (Albrecht and Schlumberger 2004; see also Heydemann 2007). De facto, however, the public sphere was closely surveilled and controlled by regime-organized NGOs (GONGOs or government-organized nongovernmental organizations; cf. Hasmath et al. 2016 for a good conceptualization; see also Naím 2009 and Carapico 2000).¹² A third strategy pursued by regimes consists of publicly delegitimizing autonomous third-sector organizations as external agents that interfere with the sovereignty of authoritarian regimes. Also, a survey on Vietnamese NGOs finds that they can help embed the state and ruling party more deeply into local society (Wischermann et al. 2015).

In line with the polyvalent roles civil society organizations play in different types of political regimes, there is still a widespread belief that the new media constitute “liberation technologies” (Diamond 2010) that tend to undermine authoritarian rule (cf. also Lynch 2011 for a more nuanced discussion). Yet Rød and Weidmann (2015, 339) emphasize the Janus-faced nature of new media, emphasizing that they also “can serve evil purposes” and be a “repression technology.” More specifically, autocrats use the Internet for three distinct purposes: to (a) track dissidents, (b) send signals of surveillance and control that aim at deterrence, and (c) influence public opinion through the spread of propaganda and misinformation. The latter in particular has become part and parcel of authoritarian policies, either directly or through hired regime-friendly bloggers who flood the Web with pro-regime posts and “information,” often also delegitimizing counterarguments (see also Chestnut Greitens 2013, 263ff.). Kalathil and Boas (2001) demonstrate with the examples of Cuba and China how different types of authoritarian regimes make effective use of the Internet through different mixes of proactive and reactive strategies, respectively. Overall, autocrats seem to be able to gain net benefits from the new technology and effectively manage to enhance control, surveillance, and agnotology (conceptually on the latter, see Proctor and Schiebinger 2008 as well as Ahram and Goode 2016).

In this context the commercialization of media service provision does not help. Private suppliers of Internet access are compelled to abide by the rules and requests of local authorities. Likewise, Chinese TV privatization makes content appear more modern, but leaves censorship unchanged while support for governmental institutions is effectively raised (Stockmann and Gallagher 2011). Even when content originates from liberal democracies, its effects are difficult to predict. Kern and Hainmueller (2009, 378) find that on East Germans in the 1980s, “the net effect of West German television exposure was an increase in regime support.” In sum, therefore, what Rød and Weidmann (2015, 349) state about the Internet goes for the media as a whole: “If democratic governments know how to take advantage of it, it might be naïve to think that autocratic governments do not.”

What applies to the media also holds true in the broader politics of symbols in virtually every corner of daily life, as Wedeen (1999) demonstrated. Using the Syrian case, she showed how a personality cult that nobody believes in is at the same time obviously phony and effective because it drives citizens to act as if they believed in it, thereby generating obedience and rules of behavior that strengthen the authoritarian regime. An iconography that appears ridiculous at first sight thereby turns out to be an effective tool for control and regime maintenance that goes beyond what we have come to know as preference falsification (on this latter point, with reference to the Chinese case, see Jiang and Yang 2016; see Kalinin 2016, on Russia).

The effects of the media and symbol politics on political support for incumbents point to another dimension of authoritarianism that is important, but almost impossible to pin down and operationalize: legitimacy. There is today an overwhelming consensus in the social sciences to understand the concept along with its Weberian meaning as a reciprocal category. In this understanding, legitimacy is seen as the subjects’ belief in the legitimacy of the ruler(s) (Legitimitätsglaube), to be gained through either input procedures (that is: democratic elections) or through output or performance (Scharpf 1999). Strikingly, studying legitimacy in authoritarian contexts seems an almost exclusively European (and German in particular) endeavor (e.g., Holbig 2010; Holbig and Gilley 2010; Schlumberger 2010; Kailitz 2013; Grauvogel and von Soest 2013; von Soest and Grauvogel 2016; Kneuer 2013; Pickel 2013; etc.), almost entirely neglected in Anglo-American research.

While common wisdom has it that autocracies by definition suffer from an inherent legitimacy gap, recent studies have presented both theoretical (Schlumberger 2004) and empirical (Gilley 2009) counterarguments: Legitimacy originates from various sources and does not per se depend on regime type. Gilley, for example, finds autocratic China among the most “empirically legitimate” countries in his sample of seventy-two cases.¹³  Schlumberger (2010) distinguishes analytically between (a) the “who” (the addressees of claims to legitimacy); (b) the success or failure (whether addressees buy these claims); (c) the “what” (the content of legitimation strategies invoked); and (d) the “how” (How are claims to legitimacy delivered to their addressees?). Country-specific and regional studies such as those by Yakoutchyk (2016b; on Belarus); White (2005; on Southeast Asia); and Schlumberger (2010; on the Middle East and North Africa) have flourished recently. The database by von Soest and Grauvogel (2016) takes this approach one step further; in their data, they differentiate along the content of legitimation strategies (or the “what”: foundational myth, ideology, personalism, international engagement, procedural mechanisms, and performance) and establish different “legitimation profiles” of political regimes. If the profile is coherent, their argument holds, authoritarian resilience is enhanced. Yet this of course evades the most difficult question of how legitimacy (as opposed to legitimation strategies) could be reliably and validly measured.

The flip side of legitimacy is repression, the other mainstay of authoritarian rule.¹⁴ In comparison to “the asymmetric attention that quasidemocratic institutions have received thus far in the literature” (Art 2012, 361; see the subsection “Institutionalist Approaches” above), repression still remains understudied.¹⁵  Landman (2013, 4) correctly reminds us that we “must not lose sight of the study of repression and coercion.” He suggests that “the varied combinations of principals and agents in the face of diverse threats produce great variation in both the use and severity of repression, and a differentiation of regimes necessary for a comparative politics of authoritarianism” (Landman 2013, 24). Existing work on repression mostly centers on questions about why, when, how, and to what avail authoritarians engage in repressive activity. One core demand has been for “disaggregating repression” spatially, temporally, and with respect to its types, as well as for integrating the topic into wider political science research (cf. Davenport 2007, 18ff.). Part of this is practiced by Josua and Edel (2015), who model variation in repression as caused by (a) the setup of the authoritarian regime, (b) the state and its capacities, and (c) the nature of the challenge to the regime, while the decision for or against repression hinges upon the availability of alternatives—and alternatives, to them, “can be framed as the availability of legitimation strategies” (291).

This latter point in particular brings in threat perception, which influences the way incumbents perceive the need for repression (against mass protest or other contentious politics). Indirectly, this also brings in society, including the societal targets of repression (How targeted is repression?). Information from the “recipient’s side” (i.e., the victims) is usually limited because repression instills fear. However, sometimes it becomes possible to take on a bottom-up perspective, as is impressively demonstrated by Pearlman (2016), who in a case study on Syrian refugees in neighboring countries traces shifts in this notion of fear.

Two other important dimensions that demonstrate the need for deeper research are added to the emerging agenda by Landman (2013, 23ff.). First, he hints at possibilities of an internationalization/transnationalization of repression and second at its privatization. Other authors point to a path dependency in the emergence of repressive patterns that can be traced back to incumbents’ previous interactions with contenders (Boudreau 2004). Art, in his turn, asks about the nature and identity of regime-hired “thugs” (which could be seen in combination with an increasing privatization of repression¹⁶) and also about the origins of repressive institutions: “Why do some dictators succeed in building effective institutions while others fail?” (2012, 369ff.)

Ultimately, most of this literature takes repression as a dependent variable to be explained by regime features, state-society constellations, perceptions of the latter, or historical experiences. For obvious reasons, micro-level empirical studies are rare in this field, but they are all the more desirable, with respect to both the institutions and the effects of repression. In particular, further studies that do not take an isolated look at repressive action but view repression as part of the larger picture of how autocrats manage their societies are still much needed.

Economic factors are often seen as an independent variable that explains regime outcomes.¹⁷ In this regard, by far the most coherent argument links economic structures to nondemocratic political rule and is made by the rentier state approach (for a recent critical discussion, see Waldner and Smith 2015; for a related debate, see also Papyrakis 2016). It is quite firmly established that states that to a large extent rely on external revenues derived from rent income (such as the export of oil or natural gas) foster authoritarian rule and hardly democratize (Beblawi and Luciani 1987; Ross 2001, 2012). In a nutshell, this is because external revenues can be distributed domestically both to privilege strategically important elites and to aliment the population at large. In turn, extraction remains marginal, so that the shibboleth of the American War of Independence, “no taxation without representation,” seems effectively reversed: no representation without taxation. This argument has been further validated in quantitative studies by Smith (2004) and Ulfelder (2007), but has also been criticized (for instance, by Herb 1999 and Alahmad 2007).¹⁸ Further development and refinement of the core arguments have been made in various directions (Peters and Moore 2009; Smith 2005, 2006). Looking at revenues must be complemented by a view of disbursement patterns and allocative power per person (Richter 2012), because this is what engenders loyalty. Further confirmation was recently provided by Lucas and Richter (2016); richness in resources strengthens authoritarian regimes. Interestingly, however, the likelihood of authoritarian leaders being ousted by competing elites does not seem to be reduced by the availability of external rent income.

Counterintuitively, bust periods do not automatically lead to regime crises (Smith 2006). Rather, as Moore (2015) demonstrated for the 2011 Arab revolts, earlier changes within the general fiscal framework toward indirect rather than direct taxation (itself a consequence of neoliberal adjustment policies) were key in a longer term relative deprivation of Arab societies that ultimately led to protest.

Neoliberal reform and its consequences, in their turn, have empirically led to more social inequality and to a relative deprivation of broad social segments in otherwise different world regions. They therefore constitute in themselves a link to protest and repression (cf. subsection “Repression” above); to regime performance (see the subsection “Performance” below); to questions of legitimacy (subsection “State-Society Relations beyond State Institutions” above); and to the origins of authoritarian regimes (see subsection “Origins of Authoritarianism, Regime Building, and Elite Constellations” above).

Since rentier states allocate resources to society, incentives to comply with regime restrictions are high as long as individual benefits are realized from such distributional politics. Such benefits are usually extended through informal, pyramidal patron-client relations that are also at the heart of a growing political economy literature on cronyism in autocracies. In this context, an amalgamation of regime-business interests can constitute powerful economic underpinnings of authoritarianism. Haddad (2012), for example, shows in a single-case study on Syria how state agents orchestrated collusion with select business actors in the light of dwindling revenues such as oil or aid. Heydemann (2004) finds similar cross-regional results for the Middle East (cf. also Demmelhuber and Roll 2007 on Egypt) that strikingly correspond with the findings that Ledeneva (1998) reached on “Russia’s economy of favors.” With respect to Asia, Kang (2002) examines state-business relations with a similar focus, whereas Schlumberger’s (2008) “patrimonial capitalism” tries to conceptualize and theorize, but also to make empirically identifiable, cases in which this particular type of politically induced capitalism prevails.¹⁹ Both Haddad (2012) and Schlumberger (2008) conclude that this type of state-business relations has detrimental impacts on development prospects, but the latter sees in this a type of economic order that emerged out of selective neoliberal reform in nondemocratic developing countries more generally.

It took democratization studies almost four decades to take into account the fact that actors, processes, and structures outside the borders of individual states impact their political regimes within. In the recent research on authoritarianism, this subfield was included early on, so there is no doubt about the importance of international factors. But the range of possible influences from abroad and the conceptualizations of such influence are very broad indeed. They start with Levitsky’s and Way’s various studies on linkage and leverage²⁰ (2005, 2006, 2007, 2010; and many subsequent studies by various authors). They identify important transnational mechanisms that impact regime outcomes. This subfield continues with studies on an often questionable efficacy of democracy promotion policies by Western powers, which furthermore can have unintended opposite effects (e.g., Kienle 2007; Dandashly 2014). Moreover, Western actors do not usually accord democracy promotion the policy priority in actual commitment that exists in rhetoric (e.g., Youngs 2010; Schlumberger 2006). Rather, it can unintentionally and mostly unconsciously make the representatives of democracy-promoting agencies part of local power games (see Leininger 2010 on Mali). Foreign aid more generally cannot automatically be assumed to be regime neutral (cf., e.g., Bader and Faust 2014). The (erroneous) assumption of the purely positive effects of Western influence that shines through early studies like those by Levitsky and Way has thus given way to a more differentiated picture.

From the viewpoint of authoritarian regimes, diverse options exist to shape relations to donors and influence donor perceptions by emphasizing themes that are foreign policy priorities of those donors (order and stability)—a phenomenon that Jourde (2007) studies by building on Bayart (2000), with the help of the African cases of Gabun and Mauretania.

At the other end of the spectrum of international factors that affect regime outcomes, some autocracies have started an outright export of autocracy or aim at disrupting democratic regimes in their home environments. Not surprisingly, this “black knight” phenomenon is most prominently observed in the Russian case (see, for instance, Ambrosio 2009; Bader et al. 2010; Jackson 2010, Dannreuther 2014) and in other post-Soviet countries like Belarus (Yakoutchyk 2016a). There is far less agreement about whether China also represents a black knight, and if so, under what conditions (Bader 2015; cf. also Burnell 2010; for more skeptical accounts of autocracy promotion in general, see Tansey 2016 and Way 2016).

Finally, authoritarian learning, cooperation, and diffusion represent probably the most recent strand in the growing literature on international factors that impact authoritarian regimes. Apart from individual studies, Germany’s GIGA Institute Hamburg embarked on a global project on this topic (cf. Erdmann et al. 2013; Bank and Edel 2015; Lynch et al. 2016; cf. also Heydemann and Leenders 2011). The studies emphasize the adaptive capacities of authoritarian regimes in learning from each other’s experience as well as collaborating among themselves in order to preserve and uphold authoritarian rule.

The debate on whether autocracies or democracies fare better in creating growth and development dates back to the Cold War era but continues unabatedly, without consensus. Why is this question important? “A political order that does not perform well will ultimately be considered illegitimate no matter how democratic the policymaking process,” Risse and Kleine (2007, 74) argue. Whether and how regime type affects the (economic, social, cultural, etc.) performance of countries is thus not restricted to research on authoritarianism, but has been a field of intense and contradictory arguments (“authoritarian advantage” versus “democratic dividend”). Much of the current literature can still be attributed to this dichotomous view of “who is doing better” (cf., e.g., Halperin, Siegle, and Weinstein 2005; Faust 2007).

Art (2012, 365), however, rightly demands that “having discovered that ‘institutions matter’ in understanding authoritarian regimes, the next step for scholars is to unpack exactly how they do.” This process has started with respect to the specific performance of some institutions (and sometimes, in a next step, what that means for authoritarian survival, emergence, or durability).²¹ Topics include, among others, environmental sustainability (Wurster 2013), social performance (McGuire 2013), property rights (Knutsen and Fjelde 2013), and responses to economic crises (among others, see Tanneberg et al. 2013). Most of all, however, they deal with developmental performance. While some precursors do exist in the literature (e.g., King 1981), to date the lack of consensus on how to classify authoritarian regimes continues to have a negative impact on this research; taken together, clear directions of results are difficult to find (cf. also Croissant et al. 2014, 2015).

A case in point is the impact of elections on performance. Miller (2015) claims that when the electoral process in autocracies is competitive, this has positive effects on health, education, gender equality, and civil liberties. While the mere existence of legislatures and elections, according to Miller, does not produce heightened levels of human development, the boldly stated finding is: “Multiparty elections […] motivate regimes to promote citizen welfare” (2015, 26). By contrast, McGuire (2013) states that single-party autocracies perform better than multiparty authoritarian regimes, having lower infant mortality rates. Yet again, Little (2016) claims that elections benefit citizens even if they are not competitive—that is, under any conditions. If we trust all three, that leaves us with confusion about contradictory findings. Juxtaposing these results, we would need to state more generally that elections are always good, but single-party rule is also, and add the gradualist assumption that the freer elections are, the better for citizens; from a scholarly viewpoint, this is not exactly satisfying. Schedler (2013, 380ff.) concurs (“bad elections are better than no elections”), albeit on totally different, noneconomic grounds, whereas the various works by Lust-Okar (2006, 2009), Magaloni (2006, 2008), and others (see the subsection “Institutionalist Approaches” above) aim at demonstrating how elections foster and facilitate authoritarian regime maintenance. All in all, the research on regimes’ impact on performance cannot yet be considered to have produced unquestioningly consolidated new knowledge. One core question is: Are findings really generalizable on the basis of the quality, amount, and operationalization of the data used? And to add another challenge: because of the absence of any agreed upon typology of nondemocratic regimes, it seems very demanding to establish generalized claims on performance not about authoritarianism in comparison to democracy, but within the universe of authoritarian regimes in intra-class, subtype comparisons.

A final research arena is just about to open up, but, upon closer inspection, is not entirely new: the relationship between regime and state(-ness). Authors of the constitutive phase had either implicitly or explicitly assumed that “the autonomy and centrality of the state is the most important concept in modern politics” (Perlmutter 1981, 2). Against this backdrop, authoritarian regimes were seen as one specific way of organizing the state and its institutions, whereby authoritarian regimes would enhance state autonomy “at the expense of society and its development” (Perlmutter 1981, 4). State and authoritarian regime could thus be seen as corresponding, mutually reinforcing entities. Today authors are re-examining the relationship between state and democracy (Andersen et al. 2014) and between state and authoritarian regimes (Schlumberger 2016b). As Russia’s autocracy consolidates with the help of huge oil and gas revenues, Perlmutter’s statement seems to hold: authoritarian regimes might be able to enhance stateness. Then again, states are crumbling where authoritarian regimes fall that have largely been congruent to the very states they ran, as evidenced in North Africa and the Persian Gulf (e.g., Libya or Yemen; cf. Schlumberger 2016a). At any rate, the nexus between the two fundamental concepts is again under scrutiny, even though this research is still at its beginning.

Interestingly, work that focuses on the performance of authoritarian regimes is likely among the first within the vast new research on authoritarianism that takes the regimes’ side not as a dependent but as an independent variable, explaining by their institutional setup how well regimes fare on a number of performance criteria and in different areas. However, this is true only prima facie; if performance impacts regime legitimacy, and we know that legitimacy is heavily affecting the complex of stability and persistence, then the deeper research interest in performance ultimately lies, once again, in the survival, persistence, or breakdown of regimes (cf. Keremoglu-Waibler 2016). This is different with the emerging research on the nexus between state and (authoritarian) regime; here the survival and breakdown of political order as such, including the institutions and organizations typically associated with the state rather than with the regime, are taken as dependent variables, while political regimes are seen as a potential explanandum for the strengthening, weakening, or erosion and collapse of statehood.

“The field is in its infancy,” Gehlbach et al. (2016, 566) correctly write about the study of authoritarianism. Yet this is not the case regarding the size of output, the number of publications, the thematic breadth of the field, and the diversity of methods applied. In these respects, the field has been exploding. But it is less than mature regarding the certainty with which we can trust the many and at times contradictory results that are published. This hints at a number of new challenges that are waiting to be addressed by future research, some of which have become apparent in the preceding discussion of the field.

First, there is a terminological fuzziness that stems from the “foggy zone” (Schedler 2002) between authoritarianism and democracy. The blurred lines between the two, as well as the absence of a clear distinction between them, make the universe of cases unclear. At the heart of this problem lies the question of how to establish classes, and the inventors of “regimes with adjectives” in particular have ignored fundamental rules of categorization such as mutual exclusiveness and joint exhaustiveness. The same goes for sub-typologies of authoritarianism, as discussed above.

Second, while most contributions in the field are concerned with how autocracies emerge, how they survive, or what stabilizes them and makes them resilient, these variables are not all the same thing. Conceptual fuzziness is thus accompanied by a fuzziness in the dependent variable. Further clarification being absent, comparative checking across different studies then becomes difficult. We need to know what the precise explanatory purpose of a new study is and what similar or competing claims have already been advanced by others, on what grounds, data, and independent variables. This requires us to distinguish more sharply factors that lead to the emergence of authoritarianism from those that sustain it.

Third, dozens of arguments and hypotheses are on offer to explain authoritarian resilience, and their authors all too often claim that their argument “provides the best explanation” (Slater 2010, 230). But the claim is not enough: What arguments are central, which ones more ephemeral? Many authors borrow from earlier literatures, but few engage with competing claims by contemporaries. While this is not to deny multicausality, a science of politics should be able to adjudicate and both methodologically and intellectually discriminate between competing claims. This has hardly happened in the new literature on authoritarian regimes. Rather, it still seems to be caught in a phase of seeking to produce as much publishable output as quickly as possible, at the expense of creating as much scientifically sustainable knowledge as possible.

Fourth, and related to the third point, mere statistical testing will often not be enough in such a process of adjudicating various hypotheses and arguments. This is because autocracies are not only “information-shy” (Henry and Springborg 2010, 55, 110) but deliberately produce false data and misinformation, and thus engage in what is called “agnotology” (cf. Betancourt 2014). Regimes’ efforts at creating a post-factual world cannot be overestimated, and we “need to reconsider the implications of opacity for our own research” (Schedler 2013, 385). Yet very few publications address this problem with even a single word. No small part of the scholarly community sometimes seems to have replaced confidence in the degree of certainty in our results with confidence in the respectability of some methods as opposed to others. Thus, “one cannot expect to be able to study contemporary authoritarian regimes using the same kinds of methods for researching open regimes” (Ahram and Goode 2016, 838; see also Croissant 2014). If we still do, and do so unaware of the inherent methodological problems, we might even run into the danger of replicating the ignorance, misinformation, and false data that authoritarian regimes so characteristically produce themselves.

These points are certainly not exhaustive regarding the challenges the broad but adolescent research on authoritarian regimes faces. Overall, we see a very encompassing literature that is in various respects still unsystematic and offers lots of publications, but little consolidated knowledge. Questions of research design, concept specification, and methodological pitfalls abound; it is likely that results published today will face important qualifications or disconfirmation tomorrow as further research unfolds. Furthermore, some of today’s findings, such as the one discussed above (subsection “Performance”: “elections are always good [for development or other things], the freer the better”), might more be reflective, albeit subtly, of an age-old democracy bias that Western political research has suffered from for too long, rather than offering consolidated knowledge.

However, if scholars adhere to the maxims of conceptual clarity, thorough concept specification, critical (self-)reflection and comparative checking, and if we adhere to the principle of producing not just publications, but knowledge, then this field of investigation will doubtlessly mature over the next decade. What is needed most for this to materialize is less a further expansion of the field by new arguments and more a rigorous checking and testing of existing ones.