Does electoral accountability matter? Evidence from a term-limit reform

This paper provides the first empirical evidence isolating the effect of electoral accountability by exploiting within-term variation in reelection incentives while holding constant both selection and competence. We exploit an unexpected and retroactive reform in Chile that introduced term limits and removed reelection eligibility mid-mandate for one third of mayors. Using high-frequency municipal procurement data from 2016 to 2024, we rely on the construction of indicators of misuse of public resources based on discretionary and opaque procurement. We find that the sudden loss of reelection incentives increases the share of discretionary procurement by 9-17\%, concentrated in clientelist and low-transparency categories such as private-use items and consultancy services, with no corresponding rise in core service areas such as health or education. In a secondary analysis, we examine the effects of term-limit-induced turnover in the subsequent mandate. While discretionary procurement shares revert to pre-reform levels after incumbent exit, procurement outcomes do not improve beyond baseline. Overall, the results provide direct evidence of moral hazard in politics and reveal an asymmetry in the short-run accountability--turnover trade-off: the costs of removing electoral incentives are immediate and sizable, while the short-term benefits of forced turnover are limited.