AUTHORITARIAN REGIMES, DECISIONAL CONSTRAINTS AND THE INITIATION OF DISPUTES

The concept of decisional constraints invokes images of elections, public opinion polls, public demonstrations, and powerful legislatures. Democratically elected leaders are cautious about going to war because they do not want to lose the next election, suffer from plummeting public support in the polls, have to respond to widespread public demonstrations, and/or answer hostile questions in congressional hearings due to costly and unsuccessful wars (Gaubatz 1999; Reiter and Stam 2002; Russet 1993). Increasingly, the institutional literature on the democratic peace emphasizes the efficiency of democratic institutions in mobilizing resources for war, the transparency of the political process, which makes democratic threats more credible; the ability of democratic armies to seize the initiative on the battlefield; and the vigorous debate within democracies, which makes it more likely that they will avoid foolish or risky policies (Bueno de Mesquita et al. 1999; Fearon 1994; Lake 1992; Reiter and Stam 2002; Schultz 2001).

None of the three types of authoritarian regimes experiences the type of decisional constraints emphasized by the first wave of literature on political institutions and the democratic peace. Autocratic leaders usually need not need worry about reelection. They can repress demonstrations, ignore public opinion, and abolish uncooperative legislatures. It was the lack of these types of decisional constraints that led us to deemphasize such constraints as a potential source of peace among specific types of authoritarian regimes. Among the three types we have examined, however, single-party regimes are the ones that make the most consistent claims to represent the majority of the population and they incorporate large numbers of people into the regime through recruitment into the party (Geddes 1999). Can this rhetorical commitment to the will of the majority and inclusion of large numbers of people into the governing party compensate for the lack of competitive politics and autonomous representative institutions? The need to respond to the concerns of members of the party led us, in “Dictatorial Peace?” to classify single-party regimes as somewhat constrained in contrast to personalist regimes, which we considered the least constrained (PBS, 18). The lack of meaningful elections, political space for protest, a free press, and independent legislatures, however, led us to suggest that decisional constraints would have only a medium–low chance of generating a peace among single-party regimes.

Single-party regimes, however, may be closer to democracies than to personalist regimes on some of the dimensions emphasized by the more recent institutional literature on the democratic peace. Most importantly, single-party regimes have demonstrated substantial war-fighting capabilities. Only one pure single-party regime has lost a war since 1945, and that was at the hands of another single-party regime, when the Soviet Union crushed Hungary in 1956. Pure single-party regimes have won five wars and fought to eight draws, with many of the draws involving two single-party regimes (Reiter and Stam 2002, 56–57). The war-fighting effectiveness of single-party regimes could reduce conflict through effective deterrence. It could also encourage these regimes to initiate disputes against states with weak and vulnerable regimes, a temptation that we also suggested might afflict democracies. As long as chief executives must gain the approval of the other leaders of the party and important decisions are vetted in a politburo or similar institution, it is possible that single-party regimes may exhibit some of the same caution and restraint as democracies.

Personalist dictatorships share none of these characteristics with single-party regimes or democracies. In “Dictatorial Peace?” we focused on a different kind of institutional constraint, one imposed by the poor war-fighting effectiveness of the armed forces in personalist regimes (PBS, 17–19). As Reiter and Stam note, personalist regimes have not won a single war in the post-WWII era, with the exception of the Ogaden War, in which two hybrid regimes with some personalist characteristics defeated a pure personalist regime. In contrast, democracies have won most of their wars. One would expect, as we suggested in “Dictatorial Peace?” that the leaders of personalist regimes would think twice before initiating any disputes, given their vulnerability. Yet we did not find strong evidence in “Dictatorial Peace?” that personalist regimes avoided participation in militarized disputes (PBS, 24). Pairs of personalist dictatorships were no less likely to engage in disputes than were mixed dyads. The different types of constraints experienced by democracies and personalist regimes did not produce an avoidance of disputes among democracy–personalist dyads. Indeed, as Reiter and Stam have shown, personalist regimes have picked numerous fights with democracies.

This finding leaves us with an interesting puzzle. Why do personalist regimes start fights they are likely to lose? Reiter and Stam suggest that personalist dictatorships pick on democracies because they see them as unwilling to pay high costs in war. Dictators expect democracies to back down in the face of a threat to use force or give up the fight once the costs become too high. “Dictators hope to win by simply outlasting their cost-averse democratic opponents” (Reiter and Stam 2003, 333–37). Thus, democratic constraints encourage personalist regimes to target democracies. But do personalist regimes also start fights with regimes that do not experience the same kind of decisional constraints that democracies do? If not, the unique political constraints found in democracies may be the key factor in accounting for the behavior of personalist regimes. If personalist regimes initiate disputes against regimes other than democracies, however, this may reflect the general importance of the lack of decisional constraints in such dictatorships. The lack of decisional constraints in personalist regimes allows dictators to engage in aggressive policies for several reasons. First, opponents of war have no institutionalized mechanisms for influencing policy. Second, dictators can lose in war and remain in power because there are no institutionalized mechanisms for punishing leaders for policy failure (Bueno de Mesquita and Siverson 1995). Finally, dictators tend to be surrounded by sycophants who tell the leader what he wants to hear, which suggests that personalist regimes may be among the least likely to calculate accurately the consequences of a belligerent foreign policy.

Examining the conflict behavior of dyads containing personalist dictators and single-party regimes will help us to deepen our understanding of institutional constraints and international conflict and address the principal themes of Reiter and Stam's contribution. If single-party regimes start more militarized interstate disputes against personalist dictatorships, this could indicate that we have found the relationship we had expected to find among democracy–personalist dyads, with capable single-party regimes selecting vulnerable personalist regimes as their targets. If personalist dictators start more conflicts against single-party regimes, it may mean that the limited institutional constraints found in single-party regimes make them more peaceful in general than other types of authoritarian regimes. In contrast, the lack of even limited constraints may lead personalist regimes to pick fights against regimes that are likely to defeat them in war.

ANALYZING SINGLE PARTY–PERSONALIST DICTATORSHIP DYADS

To address these issues, we added two dummy variables to the directed dyad data set constructed by Reiter and Stam. The first, Single Party/Personalist, is coded 1 if state A, the potential initiator, is a single-party regime and state B, the potential target, is a personalist dictatorship. The second, Personalist/Single Party, is coded 1 if state A is a personalist regime and state B is a single-party regime. As indicated by the logistic regression presented in Table 1, single-party regimes are significantly less likely to target personalist regimes than the norm. Personalist regimes were more likely to challenge single-party regimes than the norm but this result is of only limited statistical significance (p < .10, one tailed test). As with the analyses in “Dictatorial Peace?” and “Identifying the Culprit,” these results are included in models that control for contiguity, major power status, relative power, alliance, economic development, and temporal dependence.

The finding for single party–personalist dyads is especially interesting compared to the findings for democracies. Democracies are significantly less likely to initiate disputes against a particular authoritarian regime only in the case of military regimes, and that finding, as noted by Reiter and Stam (Table 1) (2003, 333–37), is not statistically significant when rare events logit is used to estimate the model. The finding that single-party regimes are unlikely to initiate disputes against personalist regimes is significant at the .01 level. Single-party regimes have started only 16 disputes with personalist regimes, while personalist dictators have started 43 disputes with single-party regimes. This finding suggests that single-party regimes may be less likely to initiate disputes in general than are other types of authoritarian regimes, a possibility that should be explored in future research. The positive, weakly significant finding for personalist initiators deepens the puzzle surrounding such regimes. Perhaps personalist regimes target both democracies and single-party regimes because they perceive them to be institutionally constrained and reluctant to go to war. Or perhaps the lack of institutional constraints in personalist regimes allows them to pick fights without making careful calculations about the likely outcomes. Further research will be needed to untangle this puzzle.