Three Essays on Corruption and Collusion

<p>This dissertation studies corruption and collusion with data derived from a laboratory experiment and household data. In Chapter 1 I study experimental procurement auctions with bribery and a public reserve to test for the tacitly collusive equilibrium described by Compte et al. (2005). Three sellers compete for 40 periods to sell a single item to a computerized buyer who accepts bribes and determines ties in bids and bribes randomly. In the closing periods, only 13.5% of auctions display the collusive equilibrium, but 58.7% of selling prices are noncompetitive. In comparison with simulated predictions for auctions that are corrupt but competitive, the mean selling price is 6.2% higher, efficiency is 35.2% lower, and the mean subject profit is 464% higher. Confusion leads to imperfect collusion, though some subjects learn to bid higher by observing bids. Men are more likely to bid the reserve. In Chapter 2 I present a method to detect corruption using only household data. I apply stochastic frontier (SF) analysis to measure the degree to which corrupt Chinese households underreport their income in comparison with other households, assuming the resultant differential is illegal income. Corrupt households on average underreport their income by 10%. I compare my results and method to those of Zhong (2018), who uses the same data but another method. Our results are similar, though only SF analysis 1) provides evidence of statistical significance, and 2) addresses endogeneity. My method provides an easy way to quantify the relative corruption between groups, regions, and countries. In Chapter 3 I apply the method of Chapter 2 to an Indonesian dataset. I find that the true incomes of public-sector households are, on average, about 50% higher than their reported income. I then divide the sample to support the findings of Martinez-Bravo et al. (2017), who exploit the fact that district mayors of the Suharto regime could finish their terms during the democratic transition, leading to exogenous variation in corruption exposure. When I restrict my sample to shorter-exposed districts, my measurement falls to 37.1%; when I restrict my sample to longer-exposed districts, my measurement rises to 56.2%. </p>

Name thesis:degree_name

Doctor of Philosophy in Economics (PhD)

Level thesis:degree_level

Dissertation

Year dc:date.available

2018

Author dc:creator

• Redwine, Arlo

Advisor dc:contributor.advisor

• McGee, Peter J.

Contributors dc:contributor

• Gaduh, Arya B.
• Hao, Li

• Collusion
• Corruption
• Economics
• Behavioral Economics
• International Economics
• Regional Economics

Repository record dc:identifier

https://scholarworks.uark.edu/etd/2971

OAI identifier oai:identifier

oai:scholarworks.uark.edu:etd-4527