In short, based on their empirical analyses, Przeworski et al. claim that the exogenous version holds true; the endogenous version, however, is false: Development makes democracies endure, but it does not make them more likely to emerge. Przeworski et al. find that no democracy failed above a per capita GDP of $6055, Argentina’s level in 1975. Referring back to Lipset (1959, 1960), they argue, somewhat vaguely, that this is so because wealth lowers the distributional conflicts within society “through various sociological mechanisms” (Przeworski et al. 2000, 101), but fail to be more explicit. Moreover, similar to Lipset (1994), Przeworski et al. emphasize the role of growth, arguing that growth performance is nearly determinant for democracies to survive, but not so for dictatorships (Przeworski et al. 2000, 109). However, as is true so often for their extensive empirical analyses, they do not offer a theoretical explanation of the mechanism as to why this is so. Yet, economic factors alone are not sufficient to account for the fates of democratic and authoritarian regimes. Democracies are found to be less stable when (1) they are more unequal to begin with, (2) inequality increases, (3) when labor receives a lower share of the value added in manufacturing. The same holds for dictatorships: they are more vulnerable to breakdown when inequality is high, and especially when they are poor. Nonetheless, these patterns must be taken with skepticism, since data on income inequality is scarce and differs in operationalization, especially for countries that have undergone transition. Most importantly, however, Przeworski et al. reject the endogenous hypothesis that economic development brings about democracy. According to them, democracies come into being almost randomly, with similar chances at all levels of development. How Przeworski et al. arrive at this conclusion, however, remains puzzling: Throughout the analyses contained in the book, the estimated coefficient for the level of economic development indicates a positive and statistically significant effecton transitions to democracy, albeit smaller than for the exogenous version. Their strict rejection of the endogenous theory remained influential on verbal grounds alone, and the inattentive or statistically less skilled reader was convinced by their argument.In a powerful attempt to rebut Przeworski et al., Boix and Stokes (2003) point out a number of shortcomings of the study. Having demonstrated the apparent misinterpretation of the estimated coefficient on endogenous democratization, they rebuke the validity of the underlying assumptions of Przeworski et al.’s model: If the exogenous hypothesis was true, as is claimed by Przeworski et al., countries are more likely to remaindemocratic at high levels of economic development. Thus, even if transitions to democracy occur randomly at all levels of development, after a long enough period of time, there are simply few cases left to undergo a transition, especially at higher levels of development.