cop44301 351..373


1 Review Article What Do We Know About Authoritarianism After Ten Years? David Art Political Institutions under Dictatorship , New York, Cambridge Uni- Jennifer Gandhi, versity Press, 2008. Steven Levitsky and Lucan Way, Competitive Authoritarianism: Hybrid Regimes After the Cold War , New York, Cambridge University Press, 2010. Beatriz Magaloni, Voting for Autocracy: Hegemonic Party Survival and its Demise in Mexico , New York, Cambridge University Press, 2006. Pablo Policzer, The Rise and Fall of Repression in Chile , Notre Dame, University of Notre Dame Press, 2009. Dylan Riley, The Civic Foundations of Fascism in Europe: Italy, Spain, and Romania, 1870 – 1945 , Baltimore, Johns Hopkins University Press, 2010. Dan Slater, Ordering Power: Contentious Politics and Authoritarian Leviathans in Southeast Asia , New York, Cambridge University Press, 2010. After two decades in which democratization was pretty much the only game in town, the study of authoritarian regimes has recently become one of the hottest sub fi elds in comparative politics. The “ transitology ” paradigm, which conceived of authoritarian regimes as theoretically interesting insofar as they told us something important about their democratic successors, now has the taste of ashes. Even by the standards of a faddish discipline, the magnitude of the switch in scholarly focus from democratization to authoritarianism has been remarkable. As always, real-world developments were primarily responsible for this shift. The third wave of democracy had runs its course by the turn of the twenty- fi rst century, 351

2 Comparative Politics April 2012 leaving a host of regimes that did not fi t neatly into existing categories. Of course, 1990s, and Fareed the signs of democratic slippage were already apparent by the mid – Foreign Affairs “ The Rise of Illiberal Democracy, ” ’ , titled Zakaria s 1997 piece in 1 underscored what many specialists already knew. Two years later, Barbara Geddes ’ s attempt to summarize what scholars had learned about democratization over the last twenty years succeeded instead in raising a host of research questions about authori- 2 tarian regime types and their durability. Some of these issues had been highlighted in previous work on authoritarianism, much of which focused on either military regimes 3 or totalitarianism. Yet the rise of regimes that were neither purely democratic nor authoritarian led to many new questions. After several years devoted to constructing typologies, a young cohort of scholars turned to the issue of regime trajectories. To paraphrase one representative work in this strain, what explained the persistence of 4 authoritarianism in an age of democratization? “ ” After a decade of work in this vein, it is worth pausing and considering what we have learned. The books that comprise this review provide a good overview of the existing research 5 programs on authoritarianism. Two are particularly concerned with the origins of authori- ’ s Ordering Power focuses on the construction of “ authoritarian tarian regimes. Dan Slater ” in Southeast Asia; Dylan Riley ’ s Civic Foundations of Fascism in Europe leviathans offers a new theoretical perspective on the development of fascism in Italy, Romania, and Spain. The other four books concentrate on the durability of authoritarian regimes, 6 though in so doing they also speak to the issue of democratization. As the title suggests, Political Institutions under Dictatorship argues that democratic-looking ’ s Jennifer Gandhi institutions, particularly legislatures, are not merely window-dressing ” but essential “ s arsenal. Pablo Policzer ’ s The Rise and Fall of Repression in Chile weapons in a dictator ’ is one of the few existing studies of coercive institutions, and shows that repressing a society is a much more complicated organizational endeavor than is usually assumed. Competitive Authoritarianism , by Steven Levitsky and Lucan Way, combines interna- tional and domestic organizational variables in a theoretical framework to explain the fi ve nondemocratic regimes over the past fi fteen years. Beatriz trajectories of thirty- ’ Voting for Autocracy not only demonstrates that hegemonic parties can suc- s Magaloni cessfully buy elections, but also shows how running up huge electoral majorities pre- serves authoritarian durability by dissuading defections by hegemonic party elites and demonstrating to both voters and the opposition the futility of challenging the system. Taken together, these books use very different methodologies to make a number of important points. I highlight four in this review article. First, by taking the institutions of authoritarian regimes seriously, several scholars are able to gain real traction on the question of durability. Rather than pointing to exogenous shocks, they are able to locate the reasons for authoritarian stability or breakdown in longstanding patterns of behavior, both formal and informal. Second, several of the books demonstrate the continued utility of comparative historical analysis for explaining both past and contemporary regime types. Third, with the exception of Magaloni, and to a lesser extent Gandhi, the books move away from patronage-based accounts of authoritarian durability, pointing to other, noneconomic factors, which allow or prevent elite cohesion. Fourth, Slater and Riley in 352

3 David Art particular seek to explain differences between authoritarian regimes, such as their degree of infrastructural power and pattern of state-society relations. fi ne future research on These are all important contributions, and will help de authoritarianism. But the books should appeal to a broader readership because each, s goal ’ in its own way, latches onto a more general concept in comparative politics. Slater fi is to demonstrate that contentious politics “ non-routine political — which he de nes as (p. 5) — should be treated as an events involving considerable popular mobilization ” independent variable, as well as a dependent one, in theory building. Riley engages the literature on civil society. Levitsky and Way bring international variables back into the study of regime change. Policzer demonstrates the utility of a principal-agent frame- races the new institutionalism and formal work in organizing coercion. Gandhi emb modeling. Magaloni synthesizes decades of insights about elite and voter behavior to craft new models of both authoritarian durability and regime transition. After analyzing what these books contribute to our understanding of the origins and persistence of authoritarian regimes, the latter half of this review essay raises three issues that emerge from my reading of them, and which I believe future work on rst is the relatively weak authoritarianism should think about more explicitly. The fi theory of institutions and organizations that characterizes several of them, and other works in this vein. It is not enough to iden tify institutions or organizations as key independent variables in explaining the origins or persistence of authoritarianism. The causal mechanisms linking these variables to outcomes need to be better spelled out. Given the wealth of theorizing about institutions and organizations in several sub- elds of political science, including those that would seem far removed thematically fi from the authoritarianism, such as the political economy of advanced industrial socie- ties, future work would do well to borrow from these insights. Second, we need better evidence to test our hypotheses. While work on authoritarian regimes should obviously not be held to the same evidentiary standards as work on advanced industrial democracies, I am not convinced that the situation is as hopeless rst blush. An entire generation of Sovietologists overcame hurdles in as it appears at fi collecting evidence, and contemporary researchers have worked fruitfully in places, 7 fi eld work. and on topics, that would seem ill-suited for Third, while the last decade of research on democratic-looking institutions such as legislatures and political parties in authoritarian regimes has been productive, there are many other compelling issues to explore. Coercion, for example, remains the core feature of authoritarian regimes, and several of the books in this review refocus our attention on the role of fear, violence, intimidation, and surveillance in the origins and maintenance of authoritarianism. Yet our understanding of the coercive institu- tions of modern authoritarian and hybrid regimes is pretty thin, especially when com- pared to studies of totalitarianism in which organizations like the Gestapo, KGB, and Stasi fi gure prominently. We need to know much more than we currently do about — militaries, paramilitaries, presidential guards, secret the myriad of security forces — that comprise the coercive apparatus of contemporary police, militias, armed thugs authoritarian regimes. 353

4 Comparative Politics April 2012 Origins of Authoritarianism rst place, and what types of legacies How are authoritarian regimes constructed in the fi do these foundations leave? Slater and Riley both adopt comparative historical analysis (CHA) to get at these questions. This research tradition has enjoyed something of a rebirth in comparative politics over the last decade. Of course, one could claim that a mode of social scienti c inquiry that traces its lineage from Barrington Moore, back to fi Karl Marx and Max Weber, and ultimately back to Aristotle, was never likely to disap- pear entirely from view. Still, self-conscious efforts to de fi ne and defend best practices eld of CHA have appeared prominently in some of the most visible publication in the fi 8 outlets. Both Slater and Riley continue in this vein. Both draw on the work of Gregory one of the most in fl uential recent exemplars of CHA – although Slater does Luebbert — 9 so more explicitly than Riley. Both also make contentious politics the starting point of their analysis. “ contentious politics receives far more attention as a Slater laments the fact that product rather than a producer of political phenomenon ” (italics in original, p. 275). There is a massive literature on the causes of r evolutions, riots, civil wars, strikes, and protests. Yet there are far fewer works, according to Slater, that treat these phe- nomena as independent variables. His central argument is that different patterns of con- tentious politics explain regime outcomes in Southeast Asia. By outcomes, he means different types of authoritarian regimes, and is explicitly not concerned with explain- ing the presence or absence of democracy. “ Since democracy broke down everywhere ” Slater argues, “ the region lacks the requisite variation to in postwar Southeast Asia, ” support any systematic explanation for democratic breakdown (fn.33, p. 12). What do vary, however, are the strength of state infrastructures in Southeast Asia and the durability of the regimes that control them. Following Luebbert, who linked the outcomes of liberalism, fascism, and social democracy in Western Europe to the responses to mass unrest in the wake of WWI, Slater begins his analysis with the conten tious politics following the end of WWII. Although the seven cases (Burma, Indonesia, Malaysia, Philippines, Singapore, South t a general pattern of communist insurrection Vietnam, and Thailand) would appear to fi es three distinct patterns of contentious fi amid colonial retreat, Slater instead identi politics that had long-term consequences for state building. The fi rst leads down a pathway of domination, illustrated by the cases of Malaysia ’ s central insight — one that draws from Hobbes and scholars of and Singapore. Slater — is that elites will only act collectively to maintain an Hobbes like Ioannis Evrigenis 10 authoritarian, extractive state if they view it as less threatening than the alternative. To get a sense of that alternative, elites gauge the extent of the threat to their lives and livelihoods by analyzing current and past forms of contentious politics. Speci fi cally, endemic or episodic , and whether it is they need to decide whether mass unrest is or unmanageable under existing institutional arrangements. Elites are manageable most threatened when they conceive of contentious politics as endemic and unmanage- able. Although there are many potential reasons for reaching this conclusion, Slater 354

5 David Art argues that elites are most threatened when (1) contentious politics takes the form of demands for radical redistribution; (2) it touches urban as opposed to rural areas; and (3) it sparks communal in addition to class tensions. These conditions pertained in Malaysia and Singapore, where communist movements ignited urban race riots and — — led elites to band together in what Slater calls a both ethnic Malay and Chinese “ protection pact. ” Elite collective action resulted in a strong authoritarian state under- pinned by durable political parties, bureaucracies, and coercive institutions. s emphasis on noneconomic incentives for elite collective action is a wel- Slater ’ come correction to a literature that has focused overwhelmingly on patronage as the central mechanism in authoritarian durability. While some authors, such as Magaloni, have produced theoretically sophisticated and nuanced elaborations of this basic argu- t speci fi ment that fi c cases, the ability to buy off sectors of the population is only one part of the story. The distribution of resources alone cannot explain patterns of break- down and persistence, since the number of authoritarian regimes that have weathered severe economic crisis is much higher than an exclusive focus on patronage would predict. Slater rightly points out that elites are not only afraid of losing licenses and contracts, but also are concerned about a return to a Hobbesian state of nature. Although he includes leftist political demands as a central variable in elite threat perception, his theory is much more grounded in security studies than in political economy. But the argument is not simply that we should move beyond patronage-based accounts. In fact, Slater claims that authoritarian regimes can indeed be founded on provision pacts patronage (what he terms elite ), but that such an arrangement leads “ ” — fragmentation. When elites do not face a form of contentious down a second pathway politics that is leftist, urban, and communa l, they are not motivated to form a strong “ Flimsy coalitions produced fl imsy institutions ” in coalition versus a common enemy. the Philippines, South Vietnam, and Thailand (p. 23). Authoritarianism was less durable, and state strength was lower, in these cases. militarization with Indonesia and Burma. Slater illustrates his third pathway — — The key point here is that contentious politics took the form of regional rebellions that threatened the territorial integrity of the state more so than they threatened the elites. The army took responsibility for squashing these rebellions and, much like Brandenburg, Prussia, after the military revolution of the sixt eenth century, created a military-bureaucratic 11 state. In contrast to Prussia, however, elites in Burma and Indonesia were never incor- porated into state institutions, which meant that their interests were never perfectly aligned with those of the ruling military. Slater admittedly has more dif fi culty fi tting these cases into his theoretical framework, yet they do support the broader point about fl icts — as opposed to the external ones that Tilly and others normally how internal con 12 — fl uence state building. in highlight This attention to con fl ict and state building is one fi eld boundaries, and I would not be surprised of the many ways that Slater crosses sub if selections of it appeared on graduate syllabi in international relations, in addition to comparative politics. Although Riley ’ s story commences a half century earlier and on a different con- tinent, there are clear overlaps with Ordering Power . Like Slater, Riley begins with 355

6 Comparative Politics April 2012 contentious politics — in the form of associational activity that challenges the political to explain both the development of authoritarianism as well as three par- status quo – ticular subtypes of it. As the title suggests, Riley is clearly taking on the Tocquevillian thesis that posits a causal relationship between associational life and liberal democracy. ’ s article on civil society in This would seem to be well-trodden territory; Sheri Berman Weimar Germany, among other critiques, led the prominent neo-Tocquevillian Robert Putnam to acknowledge over a decade ago that there was a “ dark side of social capi- 13 tal. ” ’ s analysis is novel in several respects. First, the choice of cases demon- Yet Riley strates that Weimar Germany was not unique in its degree of associational development. An explosion of mutual aid societies, credit associations, and cooperatives at the turn of particularly rural civil society — the twentieth century transformed civil society — in three of the least developed states in Europe. arguing that the type of associational Second, and more provocative, Riley is not the notion that fascism arose out of a patho- “ activity led to fascism. He explicitly rejects (p. 11). Rather, he views fascist movements as “ a twisted logical form of civil society ” (p. 2). The fascists rejected the institutions of ” and distorted form of democratization parliamentary democracy, while claiming to represent the legitimate democratic aspira- tions of the people. This reframing of fascism is less audacious than it sounds: Peter Fritzsche made a similar point a decade ago in emphasizing the populist aspects of 14 National Socialism. s formulation is virtually guaranteed to provoke Still, Riley ’ storms of protest from scholars of fascism on several grounds, not the least being that fi — the de fi ning features of fascism for many paramilitarism and glori cation of violence 15 scholars nition of democracy. — are hardly compatible with any reasonable de fi Poten- tially even more severe is the charge that Riley overstates the intellectual coherence of fascist movements. Even the pioneer of fascist ideology, Benito Mussolini, waited until 1932 to write down its essential elements, and there is some debate about whether the 16 components of Italian fascism ever held together, either in theory or in practice. These points will be hashed out among specialists. More important for com- parativists is the causal chain of the argument. Here Riley leans heavily on Antonio Gramsci, particularly on the concepts of hegemony and organic crisis. As in most works of CHA, sequencing plays a central role. When associational development preceded the development of strong political organizations, the political crises that followed WWI overwhelmed liberal democracies and produced fascism. When political organizations preceded associations, liberal democracy withstood these challenges. The reader might legitimately ask how this explanation is different from Luebbert. Although Riley sets up his argument as a challenge to neo-Tocquevillians, neo-Weberians like Michael Mann, and Barrington Moore, the similarities and differences with Luebbert are acknowledged only in passing. Luebbert claimed that division among liberals produced weak liberal ’ s political parties in states like Italy and Germany (weak intra-class hegemony in Riley 17 terms). He noted that this contributed to the failure of a cross-class coalition between liberal and labor parties (weak inter-class hegemony according to Riley). World War I led to an explosion of political participation, redistributive claim-making, and a variety of other pressures on liberal democracies (an organic crisis). Political scientists will also 356

7 David Art not ’ s claim that “ the rise of a fascist movement in each country was connected fi nd Riley (p. 200) particularly with the tendency of democratic forces to split and fragment ” 18 novel, as Linz and Stepan are the best-known proponents of this line of thinking. s tripartite distinction among fascist regimes does, however, break with pre- ’ Riley ’ s book, the attempt not only to explain an authoritarian outcome vious work. Like Slater of authoritarian regime gives The but also to say something interesting about the type Civic Foundations of Fascism in Europe added value. Riley argues that the particular types of associational activity under fascism produced different patterns of state-society relations. When associational life developed autonomously from the state, as in Italy, the result was party fascism. When civil society was fostered by the landed elite and the Church, as in the Spain, the result was traditional fascism. Statist fascism, as in Romania, emerged because a statist liberal party played the dominant role in fostering fi cult to say whether one of these fascist subtypes the development of civil society. It is dif would have proven more durable than the others. Putting aside the question of whether uence in the Spain was ever a fascist regime, and whether the brief period of fascist in fl Romanian government counts as a fascist outcome, there seems no reason to believe that 19 one pattern of state-society relations would have been more stable than the others. Slater has more to say on the question of durability. In his view, protection pacts are clearly the strongest foundation upon which to build an authoritarian order. A com- bination of historical memories of violence and institutional reproduction prevented Mahatir from falling like Suharto, whose leviathan was far less extractive and organized, fi nancial crisis hit. Still, it is fair to say that both Slater and Riley are when the Asian primarily concerned with the origins of authoritarianism, and only secondarily with its trajectory. To understand authoritarian maintenance and breakdown, we need to turn to the other four books. Durable Authoritarianism In Competitive Authoritarianism: Hybrid Regimes After the Cold War , Steven Levitsky and Lucan Way identify a new type of regime that, like fascism, is the product of a distinct historical epoch. The title is potentially misleading, for the authors are clear that competitive authoritarian regimes and hybrid regimes are not the same thing. In fact, the former is a subset of the latter. According to their de fi nition, a competitive authoritarian “ in which formal democratic institutions exist and are widely regime is a civilian regime ’ abuse of the viewed as the primary means of gaining power, but in which incumbents state places them at a signi fi cant advantage vis-à-vis their opponents ” (p. 5). The book is not primarily about the origins of these regimes. The authors do claim that the rise of competitive authoritarianism is a post-Cold War phenomenon, driven by a combination of a shift in Western foreign policy, political conditionality, and transnational activism. These factors combined to raise the external costs of authoritarianism, leading elites to adopt the trappings of democracy while abusing those same institutions to preserve their power. 357

8 Comparative Politics April 2012 This explanation for origins is certainly plausible, if only an initial step in explain- ing the rise of this new regime type. Still, because of its precise conceptualization and ’ s book will likely draw wide acceptance mass of empirical evidence, Levitsky and Way of the notion of competitive authoritarianism among specialists. Levitsky and Way iden- ve of these regimes between 1990 and 2008. Compare this with fascist tify thirty- fi regimes, of which there was a handful at most, depending on how one counts. Given the mass of theories that purport to explain the origins of fascism, we can probably expect a slew of studies devoted to assessing the impact of U.S. foreign policy, political conditionality, transnational activism, and potentially a range of other causal variables to provide a convincing theory of the rise of competitive authoritarianism. This is of course a good thing. To fault Levitsky and Way for failing to adequately explain the origins of a new regime type is a bit like criticizing Jared Diamond for taking the size and shape of continents as given, rather than beginning Guns, Germs, and Steel 20 a hundred million years earlier. is one of those rare Competitive Authoritarianism books that no student of comparative politics or international relations can afford to ignore. It is written so that, with a little guidance, it can be used in both introductory and upper-level courses in comparative politics at the undergraduate level. It is worth reading for the case studies alone, which serve as thumbnail sketches of the political ve countries between 1990 and 2008. histories of thirty- fi The puzzle that Levitsky and Way pose is this: what explains the diverse trajec- tories of competitive authoritarian regimes over the last two decades? Why did some democratize? Why did some develop into stable authoritarianism? And why did a third group become unstable author itarian, meaning that autocrats lost out to other non- democratic successors? They make a three-step argument that begins with international — factors, turns to the domestic politics, and brings international power in some cases — relations back in to solve the puzzle. Simply put, the subset of competitive authoritarian regimes that were tightly linked with the West, meaning that they possessed a high degree of economic, social, communication, and intergovernmental ties, by and large democratized. These regimes were clustered in Central and Eastern Europe and in Latin America. In those cases in which links with the West were low, as in Africa and Cen- tral Asia, the organizational power of incumbents — a product of coercive capacity, elite — determined the outcomes. cohesion, party strength, and state control of the economy egimes became stable authoritarian. If When organizational power was high, these r — — determined Western leverage organizational power was not high, then a third factor whether they would become stable or unstable authoritarian. States that were weak economically or politically, or that counted on the support of a powerful external ally for whom democratization was not a foreign policy consideration (Russia in the past, and possibly China in the future), became unstable authoritarian. The authors spend a chapter outlining the intervening variables connecting their three main independent variables (linkage, leverage, and organizational power) to their dependent variable (regime outcomes). This is a remarkable effort in theoretical syn- thesis, and it is impossible to do it justice in this review. It is a big-picture book that offers a structural explanation for regime trajectories (Luebbert is cited approvingly here 358

9 David Art as well), and therefore an alternative to contingency- and leadership-based accounts. As for Slater and Riley, historical processes leave enduring legacies that cannot be overcome easily by simply changing the formal rules of the game. The book is also a challenge to a modernization theory in that levels of economic development do not explain the patterns of democratization, stable authoritarianism, and unstable authori- tarianism emerging from competitive authoritarian regimes. One of the central lessons is that institutions determine regime trajectories, even if the institutions that really matter are sticky and therefore not fundamentally altered through formal constitutional design. ’ s Political Institutions under Dictatorship offers a somewhat dif- Jennifer Gandhi ferent perspective on institutions, one that draws more from a rational choice as opposed ’ s to a historical-institutionalism perspective. Like ,Gandhi Competitive Authoritarianism book is broad in scope and ambition. Her dichotomous conceptualization of regimes as either democracies or dictatorships means that her universe of cases is much larger than s, which allows her to run a number of statistical tests. She defends her Levitsky and Way ’ regimes in which rulers acquire power by means minimalist conception of dictatorship as “ (p. 7) in a couple of paragraphs, and thereby chooses not other than competitive elections ” to engage at length with the massive literature on hybrid regimes. This will strike some students as wise, and others as misguided. Indeed, it is dif fi cult to see how both Gandhi and Levitsky and Way can be right. If competitive authoritarianism represents a distinct and widespread regime type, then Gandhi fi - ’ s dichotomous categorization is clearly de cient, and the basic assumption of unit homogeneity is violated in her statistical tests. If Gandhi (and Geddes) are correct, and the useful distinction among dictatorships is whether they are monarchies, juntas, or one-party states, then introducing yet another concept like fi eld, as do Levitsky and Way, is a mistake. Since competitive authoritarianism into the sub this debate is likely to involve larger arguments about the pros and cons of both universal and middle-range theorizing, I will merely fl ashpoint. fl ag it here as a potential — Although the universe of cases is larger, Gandhi picks a narrower range of institutions speci er extent political parties — to focus upon than do Levitsky cally legislatures and to a less fi and Way. These institutions are not coercive, and they are formal rather than informal. The overarching goal of her book is to demonstrate that these democratic-looking institutions “ window dressing, ” but that they play a central role in the construction, are not simply policymaking, economic performance, and durability of authoritarian regimes. She is 21 extending research on formal institutions under dictatorship, and her book is convincing in two respects: rst, that the conventional wisdom that authoritarian institutions are empty fi shells and therefore not deserving of serious scholarly inquiry is wrong; and, second, that we still do not know nearly enough about them. Her argument is simple and parsimonious. She assumes that dictators want to stay in power, and that they must deal with threats when they arise. If those threats are not severe, they do not require anything beyond the normal instruments of coercion. But, fi cant enough, dictators need to make conces- Gandhi argues, if those threats are signi sions to outside groups. Since these concessions take the form of policy compromises, institutions are needed in order to signal preferences and to forge agreements. “ For the potential opposition, assemblies and parties provide an institutionalized channel through 359

10 Comparative Politics April 2012 which they can affect decision-making even if in limited policy realms. For incumbents, these institutions are a way in which opp osition demands can be contained and ” answered without appearing weak (p. xviii). In short, institutions are instruments of s main theoretical contribution. This account co-optation; and this point is the book ’ of origins is clearly inspired by rational c hoice and functionalism, and Gandhi uses evidence from the case studies of Kuwait, Ecuador, and Morocco in her second chapter to illustrate that her deductions are empirically supportable. Turning to the effects of institutions, Gandhi argues that institutionalized dictator- ships produce outcomes different from those of noninstitutionalized ones. Her measure of institutionalization combines political parties and legislatures as follows: a dictator- ure with multiple political parties, a 1 if the ship scores a 2 if it possesses a legislat regime controls all the seats within the leg islature, and a 0 if either political parties are not allowed in the legislature or the legislature is closed (p. 191). Using institu- tionalization as her independent variable, Gandhi then explores its effects on the pro- nds that vision of public goods, economic performance, and regime survival. She fi institutionalized dictatorships provide more public goods than noninstitutionalized ones, which supports her argument about policy concessions. Economic performance is higher under institutionalized dictatorships, and Gandhi suggests that this relationship supports her claim that institutionalization forces autocrats to enter into policy compro- mises that increase the provision of public goods and limits their ability to extract rents. ndings on regime survival, which ultimately is her most important dependent Her fi variable, are not as clear. Her statistical tests yield no relationship between institution- alization and regime survival, and she is forced to concede that this result could consti- tute a major strike against her claim that institutions preserve authoritarianism. Yet Gandhi offers another interpretation. Since ultimately the degree of threat determines whether or not dictators will construct institutions to co-opt other elites, lack of insti- tutionalization could simply signal a lack of threat, while strong institutions suggest a “ Because observed institutions re fl ect a best response on the part strong potential threat. ... of she writes, “ these institutions do not advantage or disadvantage any of dictators, ” (p. 178). In this sense, she would not disagree with a line of ” them in terms of power scholarship that claims that strong political parties and other types of quasidemocratic institutions, in addition to the ones pointed out by Slater and Levitsky and Way, aid 22 regime survival. It is worth re fl ecting at this point on how the works reviewed thus far compare to ’ s Voting for Autocracy . To be clear, Magaloni is concerned with a Beatriz Magaloni smaller subset of authoritarian regimes than either Gandhi or Levitsky and Way. She includes only those in which one political party remains in of fi ce uninterruptedly “ under semi-authoritarian conditions whil e holding regular multi-party elections ” (p. 32). She terms these cases hegemonic-party autocracies, and asks two basic questions about them: why they persist, and why they go to such effort to create supermajorities through elections when minimal winning coalitions would be enough to ensure their dominance. Magaloni ingeniously answers the fi rst question by way of the second. In short, by run- ning up huge electoral margins of the victory, the hegemonic party creates an image of 360

11 David Art invincibility that keeps potential elite defectors bound to it, demonstrates to voters that they have no other viable choice, and exacerbates the coordination dilemmas that opposition forces must face. ’ Since the book appeared in 2006, and since Magaloni s arguments already have uenced much of the work on authoritarianism, I will focus on two points and leave in fl aside some of the book s subsidiary arguments. The fi rst is her causal mechanism link- ’ ing patronage to authoritarian stability. Like Slater, Magaloni focuses on maintaining nition of elite cohesion as the central issue in authoritarian durability, and her de fi hegemonic parties as “ collusive pacts among ruling party politicians to divide the spoils ce among themselves ” (p. 79) is essentially the same as Slater ’ of of “ provision fi s ” “ protection pacts ” in creating pacts. Yet while Slater views these pacts as inferior to durable authoritarianism, Magaloni marshals original empirical evidence to show how the Mexican PRI (her book is primarily a study of Mexico) distributed payments to its voters and punished opposition voters by starving their constituencies of public funds. The primary contribution here is empirical, for the notion that a hegemonic party buys at least some of its support is, of course, not new. What Magaloni effectively demon- strates is that the PRI was remarkably adept at targeting particular constituencies where they were electorally vulnerable, primarily through a government poverty relief pro- gram (PRONASOL), and that this strategy worked. Still, Magaloni is forced to address the limits of this strategy in the second half of the book. Leaving aside the econometric and game theoretic models that purport to explain Mexico ’ s democratic transition, ’ s conclusion when she writes that “ eco- Magaloni actually comes quite close to Slater nomic recession translates into a decrease in economic resources available for patron- age and pork, making the hegemonic party more vulnerable to elite divisions, voter (p. 263). ” defection, and opposition entry ’ s conception of when and why institutions are impor- The second point is Magaloni tant. On fi rst blush, her account appears to be deeply institutional and therefore in line with the other books included in this review. The PRI, after all, is an institution that seeks to ensure its own survival, and she speci fi nes electoral institutions as a “ means fi cally de ” (p. 19). Yet, to regularize payments to their supporters and punishment to their enemies in fact, Magaloni departs from treating institutions as causal forces in their own right “ endogenous to the electoral game ” (p. 260, italics in and instead conceives of them as original). It is the ability of hegemonic elites to win supermajorities that matters, for that is what ultimately allows them to shape electoral rules in their favor. So power comes rst, fi and institutions are constructed to preserve it. This may seem like an obvious point, but it is central to keep in mind when dealing with authoritarian regimes. There is a danger in granting too much causal power to constitutions that can be rewritten and legislatures than can be summarily dismissed. Indeed, it is the asymmetric attention that quasidemocratic institutions have ’ s The Rise and Fall of received thus far in the literature that renders Pablo Policzer such a useful corrective. He writes that “ the goal of this book is Repression in Chile to understand how authoritarian regimes organize and in some cases reorganize their coercive institutions ” (p. 15). He correctly notes that there is a dearth of theoretical 361

12 Comparative Politics April 2012 treatments of organizations such as the police, the military, and other organizations that monitor their societies and engage in violence against perceived enemies of the regime. Although Slater, and to a lesser extent Levitsky and Way, deal with coercive tions are not the theoretical focus of their institutions to some degree, these institu work. In this sense, they share the general assumption that coercion is not a problem as long as rulers maintain the support of those who are doing the coercing. ’ s central insight is that organizing coercion is a problematic undertaking, Policzer even under the most auspicious conditions for dictators. There are several different “ the coercion problem. ” manifestations of what Policzer identi fi es as Dictators must craft coercive institutions that can deal with threats without undermining support for the regime. They must also not allow these institutions to become alternative power 23 centers. As Geddes and others have pointed out, most dictators do not fall as a result of mass opposition movements, but rather are replaced by small groups of elites, often military elites. The organization of coerc ion thus becomes a classic principal-agent problem that demands a form of monitoring by the principal (the dictator) to keep tabs on the agent (the coercive institution). But monitoring is no easier in dictatorships than in democracies, and the common types of monitoring introduce their own trade-offs. The dictator can pursue a strategy of what Policzer terms internal monitoring, ” whereby the dictator sets up his own institu- “ tions to keep tabs on the coercive institution. Alternatively, he can rely on a system of “ whereby the dictator relies on groups outside of his control — external monitoring, ” such as opposition forces within the country, foreign governments, or international NGOs — to keep track of the agent. The former strategy is analogous to what students “ police patrol, ” while the latter is the equivalent of a of American politics refer to as a 24 fi “ re alarm. ” Both strategies have costs. Internal monitoring can be extremely expen- sive, create alternative centers of power, and introduce a new principal-agent relation- ship ( ” ). External monitoring can limit the ruler ’ s autonomy, who monitors the monitor? “ and entails at least a partial liberalization whose dynamics the ruler may not be able to control. fi rst Policzer applies this framework to Chile under Pinochet, focusing on the fi ve years of the dictatorship (1973 – 1978). Using archival and secondary sources, he explains how Pinochet experimented with different forms of monitoring until an equi- librium was reached. He documents how the blind coercion in the immediate aftermath of the coup introduced principal-agent issues that led Pinochet to construct an internal security service, the DINA, which was charged with monitoring the armed forces. This type of internal monitoring, however, did not solve the problem as the DINA began to engage in its own high pro fi le acts of violence — including the assassination of Chilean fi gures on foreign (including US) soil — that damaged Pinochet interna- opposition eep accurate tabs on the four branches of tionally. In addition, the DINA failed to k the military, and thereby failed in its basic task of monitoring. Pinochet eventually dismantled the DINA and replaced it with the CNI, which was not under his direct control. He also began to rely on a form of external monitoring for information. In short, Policzer uses a parsimonious explanatory framework that helps us understand 362

13 David Art why Pinochet — would replace his own internal secu- — and indeed why any dictator rity service with another, and why he would embrace the seemingly counterintuitive de actors for information about his own strategy of relying on opposition and outsi coercive institutions. Institutions and Organizations All six books emphasize institutional and organizational factors. This is unmistakable in the works of Gandhi and Policzer. The other authors, Magaloni in particular, give structural factors a more prominent role. Linkage and leverage, according to Levitsky and Way, are primarily determined by geographic and material factors. Riley ’ s notion of hegemonic politics appears to be deeply social-structural. Slater focuses on geographic and class variables to explain patterns of e lite collective action. Yet institutions and s frame- organizations are clearly central to their accounts as well. In Levitsky and Way ’ “ organizational power determines regime outcomes when linkage work, the degree of ” s key transmission mechanism from contentious politics to fascism is ’ is weak. Riley voluntary associations. Slater relies on institutions to explain how provision pacts per- sist after the violence that produced them has subsided. Thus, for all of these authors, institutions or organizations are critical causal variables. Each book largely succeeds in demonstrating that institutions and organizations under dictatorship are not just window dressing. ” Legislatures and political parties “ — although they are certainly these things are not merely rubber stamps for autocrats — as well but forums for policy compromises and tools of cooptation. State capacity and party strength help determine whether i nternal opposition or exogenous shocks lead to authoritarian breakdown. Administering terror involves a host of organizational decisions and processes. To the extent that these points were ignored in previous studies of authoritarianism, these books represent an important correction. However, given the enormous body of scholarship devoted to institutions and organizations, one might claim that each of these works adopts an overly simplistic view of how institutions and organizations matter. To varying degrees, the authors all use institutions and organizations as convenient placeholders for a variety of different causal fi ed. This is true whether the inspiration is rational mechanisms that are left underspeci choice institutionalism, as it is for Gandhi and Policzer, or historical-institutionalism, as it clearly is for Slater and arguably is for Levitsky and Way and Riley. fi rst Gandhi ’ s central claim that institutions To pick a couple of examples, consider are required for policy concessions in dictatorships. Why? Because once a dictator is threatened enough to concede something to the opposition, institutions are required to lower transaction costs by identifying reliable bargaining partners and revealing infor- mation. Although certainly in line with ratio nal choice assumptions, this leaves two issues unanswered. First, why are institutions the only means of making policy com- ’ t ad hoc bargaining between potential threats work equally well, promises? Why can 25 and why might this not be the preferred option of dictators? Put another way, why 363

14 Comparative Politics April 2012 do we need the intervening variable of ” between the independent variable “ institutions ” and the dependent variable policies ” ? Second, even if institutions are in fact “ threat “ doing much of the work, is it possible to say more about them than that they are venues t this depend on the particular type of legislature or ’ for policy compromise? Doesn political party? The decision to classify regimes on the basis of whether they have mul- tiple, single, or no institutions does not allow us to pick up institutional variation among dictatorships, much less to see how institutions shape behavior at the micro level. ’ Slater s treatment of institutions also fails to unpack the causal mechanisms linking institutions to outcomes. To be sure, part of this has to do with the fact that Slater is concerned with institutions largely as a dependent rather than an independent variable. His causal arrows run from contentious politics, to elite coalitions, to institutions. Yet he does claim that protection pacts are reproduced through two types of mechanisms — rst is attitudinal and the second is institutional. fi The more powerful of the two is the “ Slater claims, for merely by organizing actors in particular ways at the ” “ institutional, outset of a new political dispensation, leaders create structures that assume a movement of their own. Expectations converge, relationships are forged, and interests adapt to (p. 18). This is, of course, the language of path prevailing institutional frameworks ” dependence, and one cannot fault Slater for relying on the logic of path dependence to support a part of his argument. Yet these mechanisms strike me as too general to provide a convincing account of institutional reproduction. It is unclear what the increas- ing returns of institutional arrangements are, how loyalty to existing institutions is sus- 26 tained, and where endogenous institutional change might come from. Although Slater Ordering Power do an has dealt with the latter issue in other places, the case studies in excellent job of describing and measuring the institutions of the authoritarian leviathans, but are less geared toward demonstrating how institutions shape the calculations of politi- cal actors and make alternative outcomes less likely when the world around the actors 27 changes dramatically. s book also fails to specify exactly how organizations in this case voluntary ’ — Riley “ associations provided — associations are linked to fascist outcomes. He notes that (p. 68) and that the “ fascists had entrenched organizational resources to the fascists ” ” (60 – 61). The causal mechanism would thus appear to themselves in civil society be the colonization of preexisting groups by fascists, and the use of these groups to fl uence in society and, ultimately, in government. This is a familiar story, extend their in although it has certainly been told more in the German context than in the Italian, 28 Romanian, or Spanish cases. But an alternative account would be that voluntary associations — particularly those with leftist goals — provoked a counter reaction from fl uence that elites, and that it was not associations themselves but the threat of their in fi cult to tell which of these causal mecha- led to fascism. Looking at the evidence, it is dif nisms is more compelling. Since Riley does not spend much time teasing out the micro mechanisms of his theory, which is not surprising given his reliance on Gramscian con- cepts, the reader is left wondering about the precise role that associations play. To some extent, each of the authors can be forgiven for privileging macro over micro factors. With the exception of Policzer, their concern is demonstrating that institutional 364

15 David Art variables provide a more convincing account than the alternatives. Given the number of both countries and years that each author attempts to cover, there is scarcely room for “ in institutions matter a micro-level account of institutions. Yet having discovered that ” understanding authoritarian regimes, the next step for scholars is to unpack exactly how they do. Finding the Evidence One potential problem with pursuing riche es of institutional r and more nuanced typ analysis is that fi nding the evidence to support such accounts might be challenging at fi best, impossible at worst. The problems of conducting eldworkinauthoritarianor hybrid regimes are obvious. Access is often denied, and fi eld work can be dangerous to both the researcher and his/her interview subjects. To provide one illustration, Lisa s book on symbolic power in Syria was based on two years of fi eld work, yet Wedeen ’ fi she was unable to include much in the way of interview material with government of - 29 cials or ordinary citizens. While there may be strategies for overcoming some of these problems, it would be foolish to hold students of authoritarian or hybrid regimes to the 30 same standards of data quality as students of advanced industrial democracies. Given this limitation, how well do the six books deal with issues of evidence? Each uses a very different approach to theory testing, and taken together the books outline the menu of methodological options. As one would expect, each approach has strengths and weaknesses. Policzer ’ s diachronic comparison of the Chilean case, not surprisingly, provides ’ s exhaustive treatment the greatest level of detail of the six books (though Magaloni of budget cycles under PRI hegemony as well as her analysis of PRONASOL funding patterns stand out as well). Policzer pieces together evidence from many different sources, including: minutes of government meetings under the Junta, testimony col- fi ed lected by the Chilean National Committee on Truth and Reconciliation, declassi CIA documents, contemporary newspaper reports, and interviews conducted in the 1990s. Given the theoretic l dynamics and elite deci- – mid al focus on organizationa sion making, Policzer needs every bit of this material to make his account compelling and to demonstrate the utility of his principal-agent framework. The result is a study shine a light ” on “ the darkest spaces of politics ” “ that largely succeeds in its aim to fi nd a wide (p. 15). The historical narrative also makes for a good read, one that should readership among students of Latin American politics and that can be used in under- graduate teaching. Yet the book also has the feel of a mystery that is never entirely resolved. The problem is that Policzer often lacks the type of “ smoking gun ” evidence needed to support his theory. In the absence of the minutes of some key meetings (some of which fi ed), written correspondence between Augusto Pinochet may eventually be declassi key participants (which may or may not and Manuel Contreras, or interviews with be reliable), Policzer is necessarily tenta tive about his conclusions. Even his central 365

16 Comparative Politics April 2012 contention that the purpose of the DINA was to monitor other coercive institutions rests on a combination of logical deductions and counterfactual reasoning. The holes in the evidence give a speculative aura to what is otherwise a meticulously researched study. It also stands as a warning to other scholars looking to undertake qualitative research on the workings of coercive institutions. If an exhaustive examination of all the existing evidence of a single case by a country expert still leaves the reader with some unanswered “ dark spaces questions, then what are the chances that a less intensive case study of the of politics, such as those in article-length treatments or as parts of a multiple case com- ” eld needs more parison, can surmount a reasonable evidentiary hurdle? Although the fi ’ work like Policzer s, this is an important issue to consider. Gandhi adopts a methodological strategy very different from that of Policzer. “ provide some intuitions ” to inform Although she draws upon three case studies to her arguments, these ” really do not succeed in demonstrating “ qualitative snapshots exactly how institutions work in authoritarian settings (p 44). The reliance on secondary ne-grained detail render her narratives about authoritarian sources and the absence of fi institution-building in Kuwait, Morocco, and Ecuador plausibility probes rather than rigorous case studies. Gandhi is up front about this limitation, and the empirical heart of the book is comprised of a statistical test of the observable implications of her theory (which she lays out formally in chapter three). Chapter four tests her predictions that institutionalized dictatorships spend more on public goods and provide their subjects with more liberties and freedoms since they allow for policy compromises with the population. ve assesses her claim that institutions have a positive impact on economic Chapter fi growth since they provide for political stability, greater information, and policy predict- ability. Chapter six analyzes the relationship between institutions and regime survival. Here Gandhi predicts a statistically insigni fi cant relationship according to the following “ because dictators formulate their institutional strategies as a best response to the logic: conditions they face, those rulers who choose to rule with institutions should not survive signi ” fi cantly longer than those who govern without them (xxiv). It is striking that Gandhi runs into evidentiary challenges similar to those of Policzer. “ Lacking good data on public spending in dictatorships, she is forced to rely on the crude of military spending (p. 111). She fi nds a positive relationship here. She also proxy nds ” fi that civil liberties are more expansive in institutionalized as opposed to noninstitutionalized dictatorships. Yet when she analyzes the effect of institutions on social spending — which — she fi nds no rela- is arguably the best test of her argument about policy compromise fi nding, Gandhi suggests that the issue is tionship. Puzzling through this unexpected either that the data are poor (only 17 percent of the sample included data on social spend- ing) or that her causal mechanism linking institutions to social spending is incorrect. Her fi ndings in chapter fi ve linking institutions to economic growth are more persuasive, although her data do not allow her to tease out the causal mechanism linking the two fi nding on the relationship variables. As noted earlier, chapter six concludes with a null ’ s expla- between institutions and regime survival, which is either consistent with Gandhi nation that dictators adjust their strategies given the degree of threat, or is evidence against her claim that institutions are useful instruments in co-opting potential opposition. 366

17 David Art What all this means is that Gandhi ’ s. She ’ s conclusions are as tentative as Policzer not wholly unreasonable is forced into phrases like “ it is possible to con- ” “ (p. 133), “ ” (p. 179), and (p. 161) when defending some clude ” indirectly supports the notion fi of her proxies and interpreting some of her ndings. Yet rather than criticizing her for signaling weakness, although some readers will take issue with her degree of hedging, Gandhi is refreshingly transparent about the limitations of her data and research ndings, rather than design. Perhaps using case studies to elaborate on her statistical fi - to merely establish the plausibility of her argument, would have led to greater con fi ndings. dence in her fi If both a diachronic case study and a large cross-national comparison have trouble mustering enough evidence to support their claims, then perhaps a medium N strategy is the way to go? Levitsky and Way argue that their medium N analysis allows for greater measurement validity and better illust rations of causal mechanisms than a large N approach. And with thirty- fi ve cases worldwide, their research design allows for sig- fi cant variation on both the dependant variable (regime outcome) and potential explan- ni atory factors. The case study evidence certainly supports the general argument. Yet some fore- fi t thirty- fi ve cases into a book of less than seeable problems emerge when trying to 400 pages. With a little more than six pages devoted to each case on average, many of the studies are actually justi fi cations of the coding (how much linkage, leverage, and organizational power) rather than an exploration of causal mechanisms. Some of the less-studied cases, such as Guyana and Mali, rely on the views of a couple of experts. There is no room to seriously entertain alternative explanations. Moreover, the brevity opens space for the charge that the authors are only marshaling evidence consistent with their explanation. Would it have made more sense to include one longer case study (or possibly two) in each region, and bundle the other cases into an appendix? The answer depends on taste. The thirty- fi ve cases demonstrate that the argument seems to travel well across countries and regions, and this will count as excellent evidence for some compara- tivists. Others will probably prefer more detail and better treatment of causal processes. ’ s Country specialists will most likely agree with the basic outlines of Levitsky and Way case studies, where they have done a heroic job of summarizing. But country experts may also dispute the codings and readings of particular cases. Such a reaction is less likely in Riley and Slater ’ s cases, as both use a small N case study approach. Riley looks at three countries, while Slater analyzes three “ congruent countries in depth and includes a chapter in which he looks at four other ” to extend the argument. While neither book is based on extensive archival cases materials or fi eld work (although Riley has some of the former, and Slater some of the latter), practitioners of comparative historical analysis generally have not 31 viewed relying on secondary sources as a problem. fi ve Riley draws on sources in for not using Romanian sources), and it is different languages (one cannot fault him dif fi cult to imagine anyone contesting Slater ’ s grasp of Indonesian, Malaysian, or Philippine history. 367

18 Comparative Politics April 2012 As one would imagine, Riley and Slater ’ s accounts leave more room to tease out the key causal mechanisms and marshal evidence in support of them. Yet, even here, ’ s central argument is that asso- one could take issue with some of the empirics. Riley ciations led to fascism under conditions on nonhegemonic politics. Putting aside the thorny issues of whether the three outcomes were in fact fascist, and whether one can come up with a coding scheme that would differentiate hegemonic from nonhegemonic politics, the evidence linking voluntary organizations to fascist development is pretty “ fascists had entrenched themselves in civil society, ” thin. He asserts that, in Italy, yet he lacks the type of micro evidence to demonstrate his case (p. 60). With regard to Spain, he demonstrates a correlation between fascist groups and associational den- that the fascists did not cap- — sity, though this leaves room for an alternative reading ture preexisting organizations but rather emerged to contest left-leaning voluntary associations. As for Romania, Riley provides data on the absolute number of rural associations, but there is no way of telling how large they were, how important they s defense, ’ nding were, and whether the fascists succeeded in capturing them. In Riley fi good evidence of fascist takeover of existing voluntary organizations may be tricky. But a more detailed case study of at least one organization in each of the three countries would have helped his argument signi cantly. fi s evidence is the most persuasive (with Magaloni ’ Of the six books, Slater ’ sa close second). The combination of the three central cases and four congruent ones allows both for empirical richness and a wider scope. Yet one problem that recurs throughout the case material is in demonstrating empirically the degree of threat that elites perceived. Slater is aware this problem, and tries to head it off in the introductory “ since human perceptions are not directly observable, this book ’ s chapter, noting that strategy is to lay out a deductively grounded framework as to what types of conten- tious politics should be considered most challenging and threatening to the widest (p. 13). It certainly seems that elites perceived events as Slater s ” ’ range of elite actors theory predicts, and one could claim that verifying elite views is either impossible or unnecessary. Yet while perceptions may not be observable, students of international security have long wrestled w ith the problem of measuring elite threat perception, 32 and of nding evidence to demonstrate its intensity. fi Given the centrality of elite threat to his argument, more empirical evidence would have made Slater ’ s argument even more convincing. Slater is certainly not alone in this regard. Gandhi takes elite threat perceptions as a given, although this strategy is not a major problem for her argument. More broadly, contemporary work on democratization has taken the strategy of deducing elite threat 33 from levels of inequality and the mobility of elite assets. This leaves no room for the possibility of misperceptions, deliberate threat in fl ation, or a whole host of deviations 34 from a bare-bones, rational threat model. Indeed, there are good reasons to believe that dictators in particular will perceive threats differently than one would expect given objective material factors, such as the type of contentious politics or the strength of the opposition. As seen clearly in the case of Saddam Hussein, dictators may receive limited information from their henchmen and, over time, may form ideas that are 368

19 David Art wildly at odds with reality. They may fail to take note of some threats, and deliberately ate others. They may manufacture threats to achieve certain goals. in fl Although getting inside the head of a dictator is of course impossible, the impor- tance of elite ideas should not be ignored. A generation of Soviet scholars relied on a host of creative practices to get inside the heads of decision makers in the Kremlin as best they could. It would be dif cult to argue that any regime — with the possible fi exception of North Korea is as impervious to would-be researchers as the former — Soviet Union. The problems may be more formidable with historical cases, but even here contemporary journalistic accounts and archival material, where available, can or even elites in a relatively help us make reasonable suppositions about how dictators — perceived threats to their rule. I do not suggest open society like Wilhelmine Germany — that scholars abandon deductive logic, formal modeling, statistical inference, or any of the standard practices in contemporary comparative politics in favor of an interpretive practice like Kreminology. Yet given the massive technological changes that now allow a researcher to download all sorts of primary source materials, not making the effort to do so will, I believe, impoverish our work. Future Work Since criticizing outstanding work in our fi eld is an easy enterprise, let me make clear fi ed diner looking for extra that the points raised in this review are akin to those of a satis helpings. More attention to institutions and organizations would be welcome, but let us fi rst to identify institutions and organiza- not forget that these authors were among the tions as important variables in the study of authoritarianism. More micro-level evidence and attention to causal mechanisms is desirable, but a reader can still learn a great deal about the workings of actual authoritarian regimes by reading any of these books. To single out one issue that deserves some rethinking, it is the tendency to focus on the democratic-looking features of authoritarian regimes at the expense of, ironically, their authoritarian ones. Slater, Policzer, and Levitsky and Way have already started to reverse a trend toward downplaying the coercive aspects of authoritarian regimes. dictators use institutions fi rst As Slater writes in a chapter on institutional change, “ and foremost to craft collective compliance, and only secondarily to solicit policy advice 35 fl uence in exchange for support. ” or to offer in One could argue that a research agenda ’ ’ s book would misdirect our attention s or Magaloni based too closely on either Gandhi away from the core features of authoritarian institutions. This is not to deny the value in demonstrating that political parties and legislatures in authoritarian regimes matter. Yet just as Gandhi succeeded in showing that institutions are far more consequential than the ’ s observation that coercion does not fol- conventional wisdom would have it, Policzer low seamlessly from the wishes of the dictator can potentially generate the types of puzzles that make for a productive research agenda. I list four possible puzzles below. Puzzle number one: why do some dictators succeed in building effective coercive institutions while others fail? To take one historical example, we know that Benito 369

20 Comparative Politics April 2012 Mussolini aspired to create a totalitarian society — it was he who popularized the term ” in the rst place — in which coercive institutions would enforce complete “ totalitarian fi compliance and intrude into the private lives of all citizens. To this end, he founded the OVRA, which Hitler looked to as a model for the Gestapo. Yet while the Nazi secret ’ s expectations, if not to its post-facto reputation for police generally lived up to Hitler omnipotence, OVRA was largely a failure. Indeed, there are most likely many failed cases of coercive institution building that have escaped our notice. Obviously, looking and at variations among “ successful ” coercive institutions — will help us better at them — understand how violence is institutionalized in authoritarian regimes. Puzzle number two: how do authoritarians prevent coercive organizations from undermining them? A dictator need not to have read Geddes to realize that the most and their life will come from within the ruling class than likely threat to their rule — — from the opposition. Military, paramilitary, and secret police elites are likely candi- dates for coup attempts. What strategies do dictators use to avoid this fate while main- taining a grip on the coercive institutions that they need to survive? Is there a logic to the ts of strong purges that characterize many authoritarian regimes? If the costs and bene fi coercive institutions are well known to dictators, why do they so often miscalculate? Puzzle number three: why are some coercive institutions more violent than others? If Policzer is correct that Pinochet — and by extension other dictators — would have pre- ferred calibrated to brute violence, then why do we observe very different patterns fi nding the “ optimum ” level of violence across countries? What prevents dictators from that keeps potential opposition at bay without completely undermining domestic and international legitimacy? And a separate but related question: what explains variation in the same coercive organization across time? “ ” that many contemporary dictators Puzzle number four: who exactly are the thugs seem to rely on to suppress the opposition? The large literature devoted to the ques- who were the fascists ” tion of fi le and organizational “ taught a lot about the social pro dynamics of fascist movements, and also provided a window into their basis of support, at a time when electoral data were either nonexistent or unavailable at the individual 36 level. fi cult Although researching this question is likely to be some combination of dif out paramilitary organizations (or street and dangerous, asking similar questions ab gangs) in authoritarian and hybrid regimes is necessary in order to really understand the nature of contemporary coercive institutions. The puzzles are many, and the theoretical and empirical challenges to solving them should not be underestimated. But these issues are likely to become even more impor- tant to the fi eld of comparative politics over the next decade. Rather than having crested, there are good reasons to believe that research on authoritarianism will continue apace. It is too soon to tell whether the Arab Spring will lead to democratic consolidation or the reemergence of authoritarian leviathans. At any rate, the sudden collapse of what were once thought to be among the most durable of authoritarian regimes in the Middle 37 East has already forced specialists of the region to rethink some of their theories. And, if Levitsky and Way are correct that the rise of competitive authoritarianism was the result of particular historical epoch in which the geopolitical environment allowed the 370

21 David Art West the luxury of concentrating on democracy promotion at the expense of traditional security goals, the implications for democ ratization are not very rosy. The unipolar moment, the end of history, or whatever else one calls the 1990s produced a wave awed democracies, many of which had reverted back to authoritarianism a decade of fl later. If this is the best case scenario for democracy, what can we expect as international fl relations are increasingly in uenced by other powers that have little inherent interest in democracy promotion, and may even seek to work against it? NOTES I thank Ben Ansell, Michael Bernhard, Jane Gingrich, Oxana Shevel, Victor Shih, Daniel Ziblatt, and two anonymous reviewers for their helpful comments on earlier versions of this article. ” Foreign A ff 1. Fareed Zakaria, , 76 (November/December 1997). “ The Rise of Illiberal Democracy, airs What Do We Know About Democratization After Twenty Years? ” Annual Review of “ 2. Barbara Geddes, Political Science , 2 (1999): 115 – 44. I thank Barbara Geddes for not objecting to my use of a similar title for the present review. 3. The literature on both totalitarianism and military regimes is massive. Classic works on the former Totalitarian Dictatorship and Autocracy include Carl J. Friedrich and Zbigniew Brzezinski, (Cambridge: (New York: Meridian, Harvard University Press, 1956); and Hannah Arendt, The Origins of Totalitarianism Soldiers in Politics: Military Coups and Governments 1951). On military regimes, see Eric Nordlinger, ff s, NJ: Prentice Hall, 1977); and Alfred Stepan, (Englewood Cli Rethinking Military Politics: Brazil and the Southern Cone (Princeton, NJ: Princeton University Press, 1988). (New York: Cambridge University 4. Jason Brownlee, Authoritarianism in an Age of Democratization Press, 2007). 5. This review is not meant, however, to be exhaustive. One of the several important research programs it ect of economic crises on does not cover is the economic performance of authoritarian regimes, and the e ff e of the latter, see Thomas Pepinsky, authoritarian survival. For an exampl Economic Crises and the Breakdown of Authoritarian Regimes (New York: Cambridge University Press, 2009). 6. Given space constraints, I do not consider the transitions from authoritarianism to democracy in this review. It should be noted that Levitsky and Way, Magaloni, and to a lesser extent Slater deal explicitly with democratization. 7. Two recent examples are Jeremy Weinstein, Inside Rebellion: The Politics of Insurgent Violence (New York: Cambridge University Press, 2007); and Scott Strauss, The Order of Genocide: Race, Power, and War in Rwanda (Ithaca: Cornell University Press, 2006). 8. James Mahoney and Dietrich Rueschmeyer, Comparative Historical Analysis in the Social Sciences (New York: Cambridge University Press, 2003). See, also, the recent special issue edited by Daniel Ziblatt and Giovanni Capoccia, “ The Historical Turn in Democratization Studies: A New Research Agenda for Comparative Political Studies , 43 (August/September 2010): 931 – 68. Europe and Beyond, ” 9. Gregory Luebbert, Liberalism, Fascism, or Social Democracy: Social Classes and the Political Origins of Regimes in Interwar Europe (New York: Oxford University Press, 1991). (New York: Cambridge University Fear of Enemies and Collective Action 10. Ioannis Evrigenis, Press, 2008). The Military Revolution and Political Change: Origins of Democracy and Autocracy in 11. Brian Downing, Early Moderns Europe (Princeton: Princeton University Press, 1992). “ War Making and State Making as Organized Crime, ” in Peter 12. The classic statement is Charles Tilly, Bringing the State Back In (New York: Cambridge Evans, Dietrich Rueschmeyer, and Theda Skocpol, eds., University Press, 1985), pp. 169 – 87. See, also, Thomas Ertman, Birth of the Leviathan: Building States and Regimes in Medieval and Early Modern Europe (New York: Cambridge University Press, 1997). “ Civil Society and the Collapse of the Weimar Republic, ” World Politics , 49 (April 1997): 13. Sheri Berman, – 29; Robert Putnam, Bowling Alone: The Collapse and Revival of American Community (New York: Simon 401 and Schuster, 2000). 14. Peter Fritzsche, Germans Into Nazis (Cambridge: Harvard University Press, 1998). 371

22 Comparative Politics April 2012 15. Scholars of fascism that use this de Fascists (New York: Cambridge fi nition include Michael Mann, (New York: Knopf, 2004). On the University Press, 2004); and Robert Paxton, The Anatomy of Fascism A History of Fascism: 1914 – 1945 (Madison: University of Wisconsin de nitional debate, see Stanley Payne, fi Press, 1995). TheDoctrineofFascism, Fascist 16. Benito Mussolini and Giovanni Gentile, ” in Benito Mussolini, “ 42. (Rome: Ardita Publishers, 1935), pp. 7 Doctrine and Institutions – Social Origins of Dictatorship and Democracy: Lord and Peasant 17. Mann, Fascists . Barrington Moore, in the Making of the Modern World (Boston: Beacon Press, 1966). The Breakdown of Democratic Regimes: Europe (Baltimore: Johns 18. Juan Linz and Alfred Stepan, eds. Hopkins University Press, 1996). cation of Spain and Romania fi 19. It is odd that Riley does not spend much time defending his classi — albeit as fascist regimes as the consensus among historians is that both were conservative dictatorships Fascism in Spain , — with some of the trappings of fascism as opposed to fascist ones. See Stanley Payne, 1977 (Madison: University of Wisconsin Press, 1999). In Romania, the fascists were violently repressed 1923 – rst by King Carol and later by General Antonescu. fi (New York: W.W. Norton, 1997). 20. Jared Diamond, Guns, Germs, and Steel: The Fates of Human Societies fl 21. See, for example, Ellen Lust-Okar, ict in the Arab World: Incumbents, Opponents, and Structuring Con Institutions (New York: Cambridge University Press, 2005). “ Life of the Party: The Origins of Regime Breakdown 22. See Geddes; Brownlee; and Benjamin Smith, , 57 (Spring 2005): 421 – and Persistence under Single-Party Rule, ” World Politics 51. 23. Geddes. “ Congressional Oversight Overlooked: Police Patrols versus 24. Matthew McCubbins and Barry Weingast, American Journal of Political Science Fire Alarm, – 79. ” , 28 (1984): 165 25. Students of international relations may be reminded of a similar debate between those who believe that, cient outcomes with or without institutions. drawing on Ronald Coase, bilateral bargaining can reach pareto e ffi 44; and the The Problem of Social Cost, Journal of Law and Economics “ – ” See Coase, , 3 (Autumn 1960): 1 “ International Organizations and the Theory of application of Coasian bargaining in John A.G. Conybeare, 44. International Organization Property Rights, ” , 34 (Summer 1980): 307 – 26. James Mahoney and Kathleen Thelen, eds., Explaining Institutional Change: Ambiguity, Agency, and (New York: Cambridge University Press, 2010). Power ” 27. Dan Slater, “ in Streek and Thelen, Institutional Complexity and Autocratic Agency in Indonesia, pp. 132 – 67. The Nazi Seizure of Power: The Experience of a Single 28. See Berman; and William Sheridan Allen, (New York: Franklin Watts, 1965). 1945 German Town 1922 – 29. Lisa Wedeen, Ambiguities of Domination: Politics, Rhetoric, and Symbols in Contemporary Syria (Chicago: University of Chicago Press, 1999). Rede fi 30. See J. Paul Goode, ” Perspectives “ ning Russia: Hybrid Regimes, Fieldwork, and Russian Politics, , 8 (December 2010): 1055 75. on Politics – American Political Science Review 31. But perhaps it is time that it should be. As a recent debate in the revealed, simply relying on secondary accounts for coding data is not likely to produce consensus. There are numerous reasons for this, the most important being that scholars can choose a historical account that er from best supports their theory. Tests of causal arguments that use primary source material do not su ff “ this problem. On the debate, see Marcus Kreuzer, Historical Knowledge and Quantitative Analysis: The Case of the Origins of Proportional Representation, ” American Political Science Review , 104 (May 2010): – 92; Thomas Cusask, Torben Iversen, and David Soskice, “ Coevolution of Capitalism and Political 369 ” Representation: The Choice of Electoral Systems, American Political Science Review , 104 (May 2010): 403; Carles Boix, Electoral Markets, Party Strategies, and Proportional Representation, ” American – 393 “ – 13. For an example of testing established arguments using Political Science Review , 104 (May 2010): 404 Shaping Democratic Practices and the Causes of Electoral Fraud: primary source material, see Daniel Ziblatt, “ The Case of Nineteenth Century Germany, ” American Political Science Review , 103 (2009): 1 – 22. See also “ Does Landholding Block Democratization? ATest of the ‘ Bread and Democracy ’ Thesis and Daniel Ziblatt, ” 60 (2008): 610 – 41. World Politics, the Case of Prussia, 32. For a recent contribution to this literature, see A. Trevor Thall and Jane K. Cramer, eds., American Foreign Policy and the Politics of Fear (New York: Routledge, 2009). Economic Origins of Dictatorship and Democracy (New York: 33. Daron Acemoglu and James Robinson, Democracy and Redistribution (New York: Cambridge Cambridge University Press, 2006); Charles Boix, University Press, 2003). 372

23 David Art 34. On this point, see Nancy Bermeo, ” “ Interests, Inequality, and Illusion in the Choice for Fair Elections, – , 43 (2010): 931 Comparative Political Studies 68. 35. Slater, “ Institutional Complexity, ” p. 135. (Oslo: Universitetsforlaget, 36. S. Larsen et al., eds., Who Were the Fascists? Social Roots of European Fascism – The Logic of Evil: The Social Origins of the Nazi Party, 1925 1980); See, also, Mann; and William Brustein, 1933 (New Haven: Yale University Press, 1996). Debating Arab 37. On the durability of authoritarianism in the Arab world, see Oliver Schlumberger, ed., rability in Nondemo cratic Regimes (Palo Alto: Stanford University Authoritarianism: Dynamics and Du “ The Robustness of Authoritarianism in the Middle East: Exceptionalism in Press, 2007); Eva Bellin, ” Comparative Perspective, – 57; Brownlee; Steven Heydemann, Comparative Politics , 36 (January 2004): 139 fl , 1946 – 1970 (Ithaca: Cornell University Press, 1999). Authoritarianism in Syria: Institutions and Social Con ict fl ection on the failure of academics to predict the Arab Spring, see Gregory Gause, “ Why For a critical re g: The Myth of Authoritarian Stability, ” Foreign A ff airs , Middle East Studies Missed the Arab Sprin 90 (July/August 2011). 373

Related documents