Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Political Science

Committee Chair

Dwayne Woods

Committee Member 1

Eric N. Waltenburg

Committee Member 2

James A. McCann

Committee Member 3

Rosalee A. Clawson


Corruption is pervasive in many countries. However, the most effective punishment to deter corruption remains unknown. Some have suggested that the death penalty is required to reduce corruption and ensure clean government. However, research shows that the likelihood of punishment is the strongest factor, not the severity of the punishment. It requires large social costs, such as expenditure on policemen, court personnel, and specialized monitoring equipment, to improve the probability of punishment. Can elevation of the severity of punishment be a good substitute? The empirical results from this dissertation suggest that the probability of punishing corruption, i.e., how likely a corrupt public official is to be detected, investigated, apprehended, convicted, and ultimately sentenced, matters a great deal. This dissertation finds that if the probability of punishing corruption is low, even the most severe punishment cannot effectively deter corruption. Instead, a high probability, even of mild punishment, appears to work effectively. This dissertation examines corruption punishment in China, a country that relies heavily on the most severe form of punishment, as well as the United States, a country that implements milder forms of punishment to deter corruption. I conducted the comparison with individual-level corruption case data in both countries. It is inherently difficult to obtain data on the probability of corruption punishment. Thus, I created a novel inverted measurement for the probability of corruption punishment-- the duration of corruption. This measurement instrument enables comparisons between probabilities of corruption punishment. Using this novel measurement, I find that the probability of punishing corruption is substantially smaller in China than in the United States. Within each country, I also find evidence of institutional designs for compensating for the low probability of punishment with raised severity. The probability of punishment differs between subcategories of public officials within a country; however, the severity of punishment varies. I also find that implementation consistency plays a role. The graphed relationship between the total corruption amount and the severity of punishment is that of a concave parabola in China and linear in the United States. Although large amounts of missing data for individual cases may cause bias, our results show that, in contrast to relying on severity for deterrence, improving the probability of punishment consistently relates to the effectiveness of anti-corruption activity. Additionally, the case details show that public servants think rationally and respond to these probability differences.