The mechanisms by which societies come to be governed in a democratic manner is an important and accordingly oft-studied topic for scholars of comparative democratization. Democratic states, however, do not always remain democratic, and the causes and consequences of transition from democratic to nondemocratic rule are much less studied, though important academically as well as practically. In his book, Jay Ulfelder addresses the interesting question of when and why democracies fail. In it, he argues that existing theories of democratization—be they structural, process oriented, or game theoretic—offer little to explain a transition from democracy to nondemocracy, and develops a model to explain transitions away from democracy. In brief, the main argument is that different domestic actors—an incumbent political actor, an opposition group, and the military—all interact in a strategic environment with different amounts of information to pursue their own material interests.

This book has many strengths, starting with the topic. For scholars of comparative democratization as well as policymakers interested in promoting democracy, understanding not only the origins of democratic rule but also the conditions under which democracies fail is important. Methodologically, Ulfelder employs a broad array of methods in his analysis, including game theory, large-n statistical studies, and historical cases (“analytic narratives”). Additionally, he employs an impressive data set of democratic episodes from 1955 to 2007, including when and how democratic states transition to nondemocracies. The book is composed of one chapter presenting a formal model, three empirical chapters, and a chapter devoted to the implications of the approach for organizations promoting democratic policies.

The end of democratic rule is conceptualized as occurring in one of three ways: by executive coup in which the incumbent party instigates nondemocratic practices, by rebellion in which the opposition group attempts to seize power, or by military coup. With this taxonomy of transitions to nondemocracy, the main logic of the argument is presented in Chapter 2 as “the democracy game” in which three domestic actors—the incumbent party, an opposition party, and the military—each may challenge the status quo of democracy by “defecting” or not. Democracies fall when one or more actors successfully challenge the status quo. Certainly, the behavior of these groups will depend not only on their preferences for “democracy” or not but also on who is doing the ruling. Accordingly, a variety of preferences are examined. For example, one interesting finding is that even if all parties prefer democracy to nondemocracy, but there is uncertainty regarding what others will do—if incumbent and opposition take actions simultaneously—democracy can fail. The reason for this perhaps surprising finding is risk: If incumbents think that the opposition is going to rebel, incumbents would prefer to stage an executive coup, because they prefer an autocratic state with incumbents in charge to an autocratic state with the opposition in charge. It is interesting that in the converse—when all actors prefer to be in power at the expense of democratic rule—the analysis shows that democracy can be a sustainable outcome, provided that the costs of punishing “defectors” is sufficiently high. But here the presentation could be clearer: Sustaining punishment is itself a difficulty and, while acknowledged, is left unmolded.

Chapter 3 employs event history analysis to study the relationship between duration of democratic rule and a variety of covariates, such as GDP, region, the presence of international observers, and the Cold War. The findings are interesting, and are tied to the literature nicely, though not tied to the game-theoretic model developed in Chapter 2 as directly as could have been.

Chapters 4 and 5 nicely show examples of the scenarios modeled in Chapter 2, relating equilibrium of the game-theoretic model to historical episodes of democratic failure, and to nonfailure. Six cases are discussed in detail, illustrating the types of transitions conceptualized: breakdown by military coup (Fiji 2001–6 and Thailand 1992–2006); breakdown by executive coup (Ukraine 1994–99 and Venezuela 1958–2005); breakdown by rebellion (Cyprus 1960–63) and no breakdown (Spain since 1997). The use of analytic narrative is impressive, randomly selecting cases (by strata), investigating the historical details in depth, and interpreting the facts in terms of the model.

Chapter 6 address policy prescriptions, and Ulfelder argues that organizations trying to promote democracy should appreciate the strategic concerns of domestic actors; that current strategies place too much emphasis on “technocratic” approaches to promoting and consolidating; and that instead, the incentives of actors—and especially uncertainty of others' actions—play prominently in democratic failures and, hence, should be addressed.

Here again, the specifics could be articulated more clearly, however. For example, uncertainty can allow a transition from democratic rule to undemocratic rule, even if all involved prefer democracy, but lack of uncertainty can give additional advantage to the “final mover,” that is, the military. So how is one to evaluate the usefulness of policies that reduce uncertainty in “fragile” democratic states?

Yet the model and approach developed do offer useful criticisms of existing practices. For example, strengthening opposition groups might not enhance democracy for strategic reasons; stronger civil groups might be viewed as threats to military actors, prompting less civic societies.

The main theoretical workhorse of the book is used to yield interesting comparisons and explanations for different types of democratic failure. A key feature of the states and political actors under consideration, however, is that they repeatedly interact. A main difficulty in strategic interactions that is repeated is that of credibility—promises of democratic practices today are not necessarily indicative of future behavior. The difficulty actors have in overcoming credible commitments is discussed briefly in Chapter 2, but the game-theoretic tools employed to do so were surprisingly coarse. For example, analyzing dynamic interactions directly, as in a reputed game, and examining resulting (perfect) equilibria could help elucidate the commitment problem. Likewise, some of the claims could be supported much better regarding the role of international actors. The “game” set up is purely domestic, yet the conclusions are made about international aid and development. This is certainly possible, such as by raising the costs of coups on domestic actors, and so on, but the connection to the main theoretical model was at times less than transparent.

This is not to detract from the contribution of the work. From a policy perspective, if promoting democracy is desirable, then democracy, once instilled, must continue to be employed, and Ulfelder both calls our attention to this and offers testable mechanisms for the failure of democracy. Academically, this is a nice contribution, employing formal, large-n, and historical narrative approaches to study when and why democracies fail. In summary, this is a readable, well-written book that asks a very interesting question and takes a fruitful, multimethod approach to advancing the study of democratization. In particular, the synthesis with existing literature, with an emphasis on practical implications and engagement with existing policy prescriptions, is especially welcome and well done.