LAN SHI’s Research

June 2009

PUBLISHED AND ACCEPTED PAPERS IN REFEREED JOURNALS

“The Limit of Oversight in Policing: Evidence from the 2001 Cincinnati Riot” (July 2008, accepted for publication in Journal of Public Economics)

Oversight in policing involves investigating officers for complaints against them and punishing them if found guilty. Officers commit errors in policing and, since reducing the error rate is costly, they cut down policing to avoid complaints. This paper tests the hypothesis that oversight reduces policing by exploiting a quasi-experiment: In April 2001, a riot erupted in Cincinnati after a white office shot dead an unarmed African-American adolescent; the sharply increased media attention, a Justice Department investigation, together with a "racial profiling" lawsuit, exogenously raised the expected penalty of an officer's errors. Compared with the period from January 1999 to March 2001, arrests during the remaining months of 2001 fell substantially. The decline was more significant for offenses where the error rate was higher. Communities with a greater percentage of African-Americans experienced greater arrest reductions. Felony crime surged during the same period.

SUBMITTED PAPERS AND WORKING PAPERS

“Respondable Risk and Incentives for CEOs: the Role of Information-collection and Decision-making”, March 2009, revised and resubmitted

I propose a model where the agent can respond to risk: he can exert effort to collect information about the underlying state in order to make correct decisions. Such effort is more valuable in a riskier environment, and incentives can increase with "respondable" risk. Testing the model using data on CEOs, I find that incentives for CEOs increase with industry-wide risk, a measure of respondable risk. As the CEO is less able to collect information or has less discretion in acting on his information, the positive relation between incentives and risk decreases.

“Incentive Effect of Piece Rate Contracts: Evidence from Two Small Field Experiments”, May 2009

We conducted two field experiments in a tree-thinning setting. In one experiment, we switched the pay of a randomly chosen half (the treatment group) from hourly wages to piece rate pay. Workers in the control group were paid hourly wages throughout. In the second experiment, workers were switched from hourly to piece rate pay all at once. The difference-in-difference and before-after estimates suggest that the productivity increase was on the order of 20-23 percent. Although the sample size is small, the estimates are statistically significant and robust. While the quality did not drop, the study highlights the measurement costs in setting up the right level of piece rates.

“Relational Contract, Reputation Capital, and Forms of Explicit Contracts: Evidence from Information Technology Outsourcing” (joint with Anjana Susarla), June 2009

We augment existing studies of spot procurement contracts by introducing relational contracting. We first show that at intermediate interest rates, the form of procurement contract affects the parties' reneging temptation on a given relational contract, and hence affects the best relational contract that is sustainable. We also show that relational contracts make contracts less complete and complexity plays a lesser role in the choice between fixed-price versus cost-plus contracts with relational contracting. Using a novel data set on information technology outsourcing contracts, we find that lower reneging temptation, measured by narrowly varying alternative vendors' modification costs, favors relational fixed-price contracts. Second, we find that a vendor with high reputation capital in fair bargaining (cost-cutting) is more likely to be awarded a fixed-price (cost-plus) contract. Lastly, we find that complexity matters less in the choice between fixed-price versus cost-plus contracts with relational contracting.

"The Disciplining Effect of Concern for Referrals for Better Informed Agents: Evidence from Real Estate Transactions" (joint with Christina Tapia), February 2009, under review

Using the future residence of home sellers, we compare a seller who will relocate to another state and thus will likely not provide referrals with a seller who remains in the state and thus might bring referrals. We find that moving-out-of-state sellers' residences take more days to sell than staying-in-state sellers yet without any price benefits. Moreover, among moving-out-of-state sellers, an uninformed moving-out-of-state seller's residence stays on the market for fewer days and is sold at a lower price than an informed moving-out-of-state seller. We also find that a senior seller's house sells faster and for less. We interpret these findings together as supporting that i) a concern for referrals provides discipline to both shirking and manipulation of information by agents and ii) it is important that the client be informed in protecting her own interests in one-shot transactions.

WORK IN PROGRESS (TENTATIVE TITLES)

Competition, Opaque Information, Effect of Reputation (2008, fall)

“Tipping Afterward (Bonus Pay) or Beforehand (Efficiency Wage) in the Face of Uncertainty: Evidence from a Field Experiment” (2008, summer)

“Forgiving in Relational Contracts” (2008, summer)

“Does Monetary Incentives Crowd Out Intrinsic Motivation? Evidence from a Survey”