Why do some governments—both in different countries and in regions within those countries—employ more workers than others? Existing theories focus on the level of economic development, political redistribution, and social insurance. But they raise additional puzzles and do not account for all evidence or for a global trend toward decentralization of public employment. The authors propose a new theory, inspired by Russia's recent experience, that locates one motive for subnational public employment growth in a political and fiscal game between central and subnational governments. In countries with weak legal systems, local and regional officials may deliberately set their employment levels beyond their fiscal capacity, prompting bailouts from the central government, which fears the political cost to it if wage arrears accumulate and provoke strikes. The authors model the logic of such brinkmanship, derive several propositions, and show that they—and the model's assumptions—fit empirical evidence from Russia in the 1990s. Deficiencies of that country's overstaffed, underequipped, irregularly paid, ineffective, and strikeprone public sector appear to result in part from a system of dysfunctional incentives created by the interaction of electoral pressures with the system of fiscal federalism. The authors suggest parallels with Latin American countries such as Argentina and Brazil.

This article focuses on new democracies in Eastern Europe and addresses two questions about the translation of social cleavages into political oppositions. The first question concerns the translation of preexisting cleavages: does the evolution of new party systems influence the politicization of social conflicts? The second question concerns the translation of new social cleavages, that is, cleavages that emerge once a party system freezes: can a new social cleavage be politicized? To answer these questions, the article integrates a formalization of social cleavage theory with a game-theoretic model of a new party system. The first result is that translation of preexisting cleavages depends on which parties survive the early rounds of electoral competition. In fact, depending on which parties survive, the axis of political conflict can shift by 90 degrees. This implies that party systems in new democracies should be seen as important founding moments, during which political actors determine the long-term axes of political conflict. The second result is that once a party system freezes, the politicization of a new social cleavage is difficult. Indeed, it is possible that a new social cleavage will remain politically dormant. In the context of Eastern Europe, this result suggests that political salience of class conflict is likely to be low because competitive elections and political parties predate the entrenchment of propertyowning classes.

The transition from communism in Europe and the former Soviet Union has only sometimes produced a transition to democracy. Since the crumbling of the Berlin Wall in 1989 and the collapse of the Soviet Union in 1991, most of the twenty-eight new states have abandoned communism, but only nine of these have entered the ranks of liberal democracies. The remaining majority of new postcommunist states are various shades of dictatorships or unconsolidated “transitional regimes.” This article seeks to explain why some states abandoned communism for democracy while others turned to authoritarian rule. In endorsing actorcentric approaches that have dominated analyses of the third wave of democratization, this argument nonetheless offers an alternative set of causal paths from ancien regime to new regime that can account for both democracy and dictatorship as outcomes. Situations of unequal distributions of power produced the quickest and most stable transitions from communist rule. In countries with asymmetrical balances of power, the regime to emerge depends almost entirely on the ideological orientation of the most powerful. In countries where democrats enjoyed a decisive power advantage, democracy emerged. Conversely, in countries in which dictators maintained a decisive power advantage, dictatorship emerged. In between these two extremes were countries in which the distribution of power between the old regime and its challengers was relatively equal. Rather than producing stalemate, compromise, and parted transitions to democracy, however, such situations in the postcommunist world resulted in protracted confrontation between relatively balanced powers. The regimes that emerged from these modes of transitions are not the most successful democracies but rather are unconsolidated, unstable, partial democracies.

The granting of autonomous status to minority populations has gained support among academics and practitioners alike as a way to solve, manage, and even preempt ethnic conflict. In spite of the enthusiasm for ethnofederalism, however, the provision of autonomy to minorities may actually increase rather than decrease the likelihood of conflict. Under certain political conditions, autonomy promotes the separate identity of the minority and increases its motivation and capacity to seek separation from the central state. This article presents a rudimentary theoretical framework identifying which qualities of autonomy solutions increase the likelihood of conflict. It discusses how autonomy relates to other factors conducive to conflict by studying minorities in the South Caucasus and examines the case of Georgia. In Georgia, there were five ethnic minority populations, two of whom—the Abkhaz and the South Ossetians—enjoyed autonomous status and were the only minorities to engage in armed conflict with the Georgian government. This article shows how autonomy, by empowering ethnic elites with control of statelike institutions and by enhancing factors such as leadership, economic viability, and external support, played a crucial role in the escalation of the conflicts in Abkhazia and South Ossetia. Conversely, the absence of autonomy mitigated separatist and secessionist sentiments among two of Georgia's other minority groups—Javakheti's Armenian and Kvemo Kartli's Azeri populations.