| The electoral competition literature links the fate of elected office-holders to the well-being of their power base, the medial voter. In chapter two, a model of coups d'etat links the dictator's fortune to the well-being of her power base, the kingmakers. Chapter three tests whether the power base of leaders depends on the mechanism by which leaders are removed from office.;The model in chapter two offers an explanation for the incidence of coups d'etat. There are kingmakers who own equal shares in an export firm and a dictator who collects a fixed rent from the citizens. The profits of the kingmakers in any period depend positively on the price of exports and on the effort exerted by the dictator to maintain a public good that facilitates the sale of exports outside the country. Profits for the dictator decrease with the effort level exerted to maintain the public good. If a coup is staged, then a kingmaker is chosen at random to become the new dictator. Each period the price of exports is realized. In equilibrium the dictator is ousted whenever the price is below some trigger price. The intuition is simple. The expected gain to a kingmaker of staging a coup derive from the chance of becoming a dictator next period. These gains do not depend on the price this period. The costs of staging a coup are the lost profits from exports this period that obtain because the dictator exerts zero effort. Thus, as the price of exports varies, the costs of staging a coup vary relative to the fixed benefits. In particular, when the price is low enough, the cost is lower than the benefit and so a coup is optimal.;Chapter three develops duration models to test whether incumbents, unconstitutionally removed from office, are accountable to a different power base than those expelled by constitutional means. The results suggest that economic elites of LDCs are the ones who determine the timing of unconstitutional changes since it is they who have investment capabilities. When investment falls the unconstitutional transition probability of being out of office increases. However, there seems to be no evidence that constitutional transfers are effected by consumption per capita. |
