Identifying Information Asymmetries: New Methods and Evidence from a Randomized Field Experiment

This paper estimates the prevalence of asymmetric information in a consumer credit market using a field experiment of our design. A major South African lender issued 60,000 direct mail offers where the interest rate was randomized along two different dimensions — an initial “offer rate†on the direct mail solicitation, and a weakly lesser “contract rate†the applicant received after responding to the solicitation and agreeing to the initial offer rate. These two dimensions of random variation in interest rates, combined with the large sample (including 6,200 accepted offers) and complete knowledge of the Lender’s information set, will enable us to identify the prevalence and impacts of specific types of private information. Specifically, our setup distinguishes adverse selection from moral hazard/repayment burden effects on repayment and profitability, and thereby generates unique empirical evidence on the sources and importance (or lack thereof) of asymmetric information.