Why do autocrats allow legislatures, parties and elections? These nominally democratic in- stitutions are thought to be fundamental pillars of democracies, yet they have been common features of authoritarian regimes in the post-World War II period. I argue that autocrats use these institutions to deter the threat of being overthrown by regime elites. More specifically, they contribute to co-optation of potential rivals by sharing power and spoils, gather informa- tion about the opposition and credibly reinforce the perception of the ruling clique’s right to continued rule. Several recent studies have illustrated a relationship between institutions un- der authoritarianism and the longevity of authoritarian rule. However, although most scholars concede that regime elites are both the most pervasive threat to authoritarian rulers and the targets of co-optation strategies, studies of institutions under authoritarianism overwhelmingly focus on the diverse category of regime breakdowns. In order to account for this drawback of previous research, I turn my attention to coup attempts, a threat to incumbents that in most cases involves regime elites, and thus more precisely captures the core argument.In accordance with previous research addressing regime longevity, I find that autocrats ruling with nominally democratic institutions are less likely to be subjected to both coup attempts and successful coups. The findings provide evidence for the proposed theoretical relationship in which authoritarian leaders credibly share their power with potential rivals in a legislature, a regime support party and display strength through elections. However, I find no evidence that institutional differences between autocracies with some variant of nominally democratic institutions in place differ in their propensity to be subjected to a coup attempt.In sum, the results strengthen the claim that nominally democratic institutions under au- thoritarianism are effective tools for co-optation and deterrence of regime elites. It is clear that the institutional characteristics of autocracies matter for the propensity to be subjected to a coup attempt, and should therefore be further scrutinized, especially whether differences in in- stitutional design matters. Moreover, the approach of this thesis underlines that disaggregating regime breakdown and concentrating on more homogenous events such as coup attempts is a fruitful endeavor in order to more accurately test complex theories.