How different is the judicial prosecution of corrupt elites across regime types? It is widely believed that democracies are better than authoritarian regimes at holding corrupt politicians accountable, but this claim has not been systematically formalized nor tested. My dissertation addresses this question in three interventions, each of which makes a distinctive contribution to the discipline. First, using a formal model of judicial prosecution of corrupt political elites, I show that executives in democracies are more likely than those in authoritarian regimes to prosecute other politicians. This, however, can be attributed to the fact that authoritarian leaders can rely on intimidation tactics to control and curb opposition, and not necessarily because democracies increase oversight of those in power. Contrary to what existing research has suggested, my model shows that politicized prosecution, or the use of prosecution as a way to control or curb opposition, is more commonly found in places with democratic characteristics than in authoritarian regimes. My second intervention is empirical: I test the implications of my theory using original data of judicial investigations conducted against Mexican governors in the last 27 years. The data encompasses a period that maximizes subnational variation of regime type, while the outcome of interest, judicial investigations, is recorded at the individual level. This design enables me to analyze occurrences of judicial processes as well as instances of impunity across institutional settings. The empirical analysis shows that prosecutions in states with democratic institutions are more frequent, and that they are consistent with a politicized use of the judicial system. To my knowledge, this is the first dataset that observes prosecution at the individual level across regime types. Finally, I study the link between prosecution and accountability as understood by the citizens. I study two scandalous prosecutions in Mexico: the 1989 arrest of La Quina, the leader of the oil union and the michoacanazo, a large sting operation conducted in 2009 by president Felipe Calderón. Both arrests came about in very different institutional contexts, one during the golden authoritarian years and the other a decade after the inaugural elections of 2000. My analysis, however, shows that public accounts of these cases are similarly ambivalent: both arrests were considered authoritarian moves aimed at controlling political opponents, yet simultaneously legally adequate and desirable. I conclude by connecting these three interventions to call into question two widely held assumptions in the literature: that judicial prosecution of political elites are brought by different mechanisms in democracies than in authoritarian regimes, and that judicially prosecuting corruption is always an instance of accountability.