Article Abstract:
Holding politicians accountable through reelection has long been a focus of empirical research, yet results are mixed in terms of whether electoral accountability works in practice. I offer a new theory of voter behavior to explain why electoral accountability may break down. Where voters perceive a greater likelihood of malfeasance in a second term, information about good first-term performance becomes irrelevant to predicting second-term performance. Instead, voters turn to horizontal accountability institutions for assurance that reelected incumbents will perform well. I test the argument in Peru using a conjoint experiment and regression discontinuity design. I demonstrate not only a mayoral incumbency disadvantage, but also that voters prefer challengers even over high-performing incumbents. I then show that voter preferences are affected by beliefs about increasing corruption and the low likelihood of good performance being repeated and that voter trust in horizontal accountability institutions attenuates the anti-incumbency bias.
