Recent empirical studies have shown incumbency disadvantage in many young and/or poor democracies, such as Brazil, India, and most post-communist Eastern European countries. This stands in marked contrast with the evidence from most mature democracies, most notably the U.S. Taken as a whole, this body of evidence suggests that incumbency effects transcend the type of incumbent (party or candidate), level of government (local or national), or type of office (legislative or executive), but that broadly the sign of the effect differs between developing and developed democracies. This puzzle has received little attention. The extant theoretical literature on electoral accountability mainly characterizes incumbency advantage. This paper addressed the gap in this literature by building a simple agency model with adverse selection in which incumbent replacement is linked to a well-known stylized difference between developed and developing democracies -- ability of institutions to curtail rent-seeking by politicians. A key premise is that institutional weakness in developing democracies may allow incumbents to generate increasing rents during their tenure. In this setup, I show that incumbent retention is negatively related to the quality of politicians and institutional weakness. The intuition is that voters try to minimize future rent-extraction by not allowing the politicians to exploit developing rent-extraction networks. Also, incumbency disadvantage is easier to sustain as the average quality of politicians decreases. Moreover, when institutions are weak and all politicians are rent-seeking, retention rate is lowest as the pool of politicians improves. This is because the voter is confident that future rent-extraction will be lower in case of incumbent replacement. Thus, institutional weakness can lead to a perverse equilibrium in which even positive political selection is ineffective at inducing democratic accountability. It also highlights the importance of institutional reform. I extend the model to look at the effects of various institutional reforms, party discipline, and ideological polarization.
