ABSTRACT The literature on corruption, grounded in the principal‐agent framework, has long promoted transparency as a tool to reduce misconduct by improving oversight. Yet standard models assume that transparency either deters wrongdoing or, at worst, has no effect—they do not account for the possibility that it could increase corruption. This paper adopts an agent‐centered perspective, arguing that in settings with weak accountability, transparency influences outcomes only through its effects on bureaucrats, and may backfire by altering their beliefs about the risks of engaging in corruption. I test this argument in Vietnam, a single‐party regime with weak accountability mechanisms. Using data from 2011 to 2016 on citizen‐reported bribery across districts, I show that when province‐level transparency reports signal higher levels of corruption than local officials may have previously perceived, wrongdoing increases in the following year. The findings are consistent with a mechanism in which new information lowers the perceived risks associated with corrupt behavior. I rule out alternative explanations, including statistical reversion and citizen‐driven reporting effects, and show that the effects are strongest in areas where bureaucrats are more likely to be exposed to the information. These findings contribute to a growing literature that reconsiders the theoretical assumptions behind transparency reforms and identifies the conditions under which they may backfire.
Tuan‐Ngoc Phan (Sat,) studied this question.