When Fair Isn’t Fair:
Understanding Choice Reversals Involving Social Preferences
with J. Andreoni, B. Barton, D. Bernheim and J. Naecker
Journal of Political Economy, 2020, 128 (5)
In settings with uncertainty, tension exists between ex ante and ex post notions of fairness. Subjects in an experiment most commonly select the ex ante fair alternative ex ante and switch to the ex post fair alternative ex post. One potential explanation embraces consequentialism and construes reversals as time inconsistent. Another abandons consequentialism in favor of deontological (rule-based) ethics and thereby avoids the implication that revisions imply inconsistency. We test these explanations by examining contingent planning and the demand for commitment. Our findings suggest that the most common attitude toward fairness involves a time-consistent preference for applying a naive deontological heuristic.
Consumption Response to Credit Expansions:
Evidence from Experimental Assignment of 45,307 Credit Lines
American Economic Review, 2022, 112 (1), Lead Article
I design a large-scale field experiment that constructs a randomized credit limit extension isolating selection, anticipation, wealth, and interest rate effects and study the impulse responses on spending, contract choice, and balance sheets. Participants borrow to spend 11 cents on the dollar in the quarter of the limit increase, with a cumulative difference of 28 cents by the third year. The effects extend to those far from the limit, those who had the new limits as available credit, and those with a meaningful buffer of liquid assets. Participants near their limits borrow and spend when limits are relaxed but put off spending and save out of constraints under the counterfactual when limits are tight. The findings provide strong support for a buffer-stock interpretation that emphasizes the importance of precautionary saving.
Forbearance vs. Interest Rates:
Experimental Tests of Liquidity and Strategic Default Triggers
I use the random assignment of debt relief policies in a large-scale field experiment to test default models emphasizing liquidity and strategic behavior. In contrast to liquidity being the sole trigger, borrowers respond differently to a dollar reduction in current payments when delivered through forbearance or interest rate reduction: forbearance reduces payments twice as much, whereas delinquencies are more responsive to a rate reduction. Compatible with strategic behavior, borrowers default in response to changes in future payments orthogonal to solvency and liquidity. Compatible with the endogeneity of triggers, whether forbearance or interest rates are more effective, and defaults are strategic is tightly linked to borrower balance sheets. I characterize a single strategic default trigger whose location is influenced by distress, precaution, and assets. The findings have implications for targeting loan modifications and modeling the pass-through of interest rates.
Financing and Investment
with O. Kim
How does financing affect small-and-medium enterprise (SME) investment? Selection bias and correlated confounds between financing and investment opportunities make this question difficult to answer. We report findings from a business credit line experiment conducted by a large European bank that exogenously increased debt capacity for some SMEs but not others with otherwise similar characteristics. Our research design addresses the challenges of estimating investment to financing sensitivity with a random assignment of automatic credit line increases and a strong first stage for debt capacity. Even though businesses had substantial financial slack before the intervention, treated businesses drew down 35 cents on the dollar of the additional debt capacity after 12 months and 55 cents after 36 months of the experiment. Examining detailed financing and spending patterns shows that businesses used high-cost revolving debt to manage day-to-day operating costs and inexpensive term loans to finance investment. Businesses used a mix of revolving and term loans in the short run but committed the entirety of their expanded debt capacity to finance investment in the long run. We highlight businesses’ desire to preserve financial flexibility as an important source of financing frictions.
Purpose, Intrinsic Incentives, and Output:
a Field Experiment with Investment Advisors
with A. Thakor and R. Thakor
Can intrinsically motivating employees with an organizational purpose (that transcends profit maximization) raise employee productivity and firm profitability? We study this question through a large-scale natural field experiment with a randomized control group within a large multinational financial services firm. Our research design overcomes the usual endogeneity and reverse causality concerns that make it difficult to establish a causal relationship between organizational purpose (and other intrinsic motivators) on performance. The experiment invites 808 investment advisors via a lottery to a workshop that explains, reminds, and connects the employees to the firm’s organizational purpose. We then compare the clearly defined and individually attributable outcomes of the invited group to the 14,191 advisors in the control group, which are indistinguishable from the invited based on ten outcome variables going back four years. We track impulse responses using administrative data on the firm’s bottom line (i.e., assets under management, total profits), employee financial performance (i.e., customer acquisition, bonus eligible profits), productivity (i.e., the likelihood of meeting expectations, progression in the organizational level ladder), client satisfaction metrics, and surveys. Finally, we scrutinize several additional commonly invoked mechanisms through which the effect manifests, such as social preferences, motivation, clarity, and identity, and quantify how much the employees value working for a company with purpose (i.e., in terms of a salary equivalent).
Work in Progress
Optimal Pricing using Causal Machine Learning
with Janis Skrastins and David Sraer
We introduce a novel approach to pricing consumer loans. Our approach combines structural modeling, experimental identification, and causal machine learning. We use the approach to quantify how much a bank can increase its profits by offering personalized pricing and better understand the extent of adverse selection and how it varies with borrowers’ characteristics. First, we study the pricing problem from the lender’s perspective. The optimal price depends on several key elasticities: how demand (takeup) and costs (missed payments) change as the interest rate changes. Second, we use a randomized pricing experiment alongside instrumental variables techniques to identify these elasticities and then calculate the optimal price and the profits under this price. Third, we devise a novel algorithm that partitions the data to maximize the lender’s profits. In in-sample honest tests, the algorithm substantially improves profits compared to risk-based and homogeneous pricing. We test the profitability of this procedure out of sample by underwriting new loans priced using this method and alternative pricing methods (uniform, risk-based, and homogeneous).
Capital Budgeting and Reallocation in Banks
with Janis Skrastins