Note: I didn’t read the details of this too thoroughly, although it’s probably a good idea to know this piece and Przeworski and Limongi’s “Modernization: Theories and Facts” if you’re going to answer a Q on democratisation”
_Main Argument:_ The authors challenge Przeworski and Limongi’s refutation of endogenous democratization. Development is both an endogenous and an exogenous cause of democracy. Development in poor and middle-income countries increases the probability of democracy for both endogenous and exogenous reasons: development increases the probability of the transition to democracy and the probability that an existing democracy will sustain itself. The endogenous effect of development is attenuated at high levels of income.
_Method:_ Testing a previous theory with empirical data.
== Notes ==
A critique of Przeworski and Limongi’s “Modernization: Theories and Facts” (1997)
– P&L reconsider the classic proposition that economic development favours democracy, identifying an ambiguity in this proposition: Why do we observe a higher proportion of democracies among rich than poor countries?
– Is it because development increases the likelihood that poor countries will undergo a transition to democracy? (“endogenous” theory)
– Or is it because development makes democracies, once established, less likely to fall to dictatorships? (“exogenous” theory)
– P&L offer evidence that the exogenous theory holds, the endogenous one fails
– They (Boix & Stokes) argue that modernization theory (at least its endogenous variant), has no empirical basis
P&L’s Conclusion: “once established, democracies are likely to die in poor countries and certain to survive in wealthy ones” (1997:167).
_Boix & Stokes’ Tests of Robustness:_
_Robustness Problem 1: Dwindling Numbers_ – they observe few transitions to democracy at high levels of income and infer that income does not cause such transitions, whereas this observation is in fact consistent with endogenous democratization
_Robustness Problem 2: Sample Selection_ – in 1950, the distribution of regimes was not random but highly correlated with per capita income; Second, growth patterns are not randomly distributed either
In short, democratization is a process endogenous to development, but this fact is less salient when we look only at a post-1950 sample
_Robustness Problem 3: Omitted Variables_ – the mobility of capital/ oil-rich countries