High-powered contracts, self-selection, and welfare in settings with externalities

We extend the experimental analysis of sorting and effort effects of high-powered contracts on welfare to situations with negative externalities. Participants solve brainteasers from Raven's matrices. The difference between right and wrong answers represents our measure of welfare per capita. We compare two contract schemes: fixed-wage and bonus contracts that reward subjects for the number of correct answers, regardless of the number of wrong answers. With fixed wages, selfish individuals have no effort incentive. With bonuses, they have incentives to answer as many questions as possible. The two contract schemes are further separated depending on whether participants self-select or are randomly assigned to a contract. The self-selection treatments correspond to cases where countries do not regulate contracts. The random assignment treatments mimic situations where countries either offer only bonuses or ban them. We find that bonuses generate lower welfare per capita than fixed wages as the higher effort incentives are outweighed by the detrimental effect of answering too many questions. However, due to productivity sorting, a general ban of bonuses does not increase welfare per capita compared to offering both contract schemes.