Friday, January 24, 2020 at 11:40am to 1:10pm
Ives Hall, 115
B07 Tower Rd, Ithaca, NY 14853, USA
Ming Li - University of Pennsylvania
Information and Corruption: Evidence from China's Land Auctions
Abstract: This paper examines how sellers' private information affects auction outcomes differently in two-stage auction and English auction, and how this difference affects auctioneer's incentive in choosing auction formats and gives rise to corruption in the context of China's land market. Using a theoretical model to endogenize the incentive of China's local governments, this paper finds that (1) two-stage auctions are more prone to corruption than are English auctions when information is asymmetric and (2) land with lower value faces a harder constraint for corruption. Consequently, local governments tend to use two-stage auctions on low-value land to maximize personal benefits and use English auctions on high-value land to maximize public benefits. Using a detailed data set that covers all land transactions in China between 2007 and 2017, I structurally estimate a common value auction model where bidders are asymmetric in information and private costs. Results show that land sold by two-stage auctions on average have a value lower than English auction by CNY 343=m2, explaining 43% of the observed price gap (selection effect); and the remaining 57% can be explained by the different bidding equilibrium of the two auction formats (corruption effect). Moreover, connected bidders have a significant information advantage over unconnected ones, allowing them to bid higher and win more often. The counterfactual results suggest that using English auction only and increasing public information disclosure, can both significantly reduce corruption, increase land revenues, and increase social welfare.