Perceptions of Corruption: An Empirical Study Controlling for Survey Bias

Abstract

Survey data on corruption are widely used to construct corruption indices, but are hardly questioned. How do individual experiences shape perceptions of corruption? Do more work-engaged respondents perceive corruption as a bigger obstacle to business operations than others? What role does answer bias play in corruption surveys? This article brings together several strands of literature to discuss these questions, and tests them empirically with survey data from Bangladesh and Sri Lanka. We find that individuals who are more work engaged report corruption as a bigger obstacle. So did respondents who were previously exposed to corruption. We control for possible answer bias by implementing a randomized response technique, and find that corruption tends to be under-reported. The effects of work engagement and prior exposure to corruption are more pronounced when the bias indicator is considered, and again become stronger once we control for answer bias affecting past experiences with corruption. JEL: C83, D03, D73, O12