Empirical Studies of Information, Market Power, and Policy Design

Stillerman, David

doi:https://doi.org/10.21985/n2-w6f3-pm56

Work

Empirical Studies of Information, Market Power, and Policy Design

Public

Download PDF

Download All Files (.zip)

This thesis consists of three chapters: two are empirical studies of policy design in small-business lending markets, while the final is a large-scale retrospective of retail mergers. Chapter 1 examines moral hazard in loan guarantee programs. To address credit constraints in small-business lending markets, policymakers frequently rely on loan guarantees, which provide lenders with insurance against default and aim to expand credit. Guarantees affect loan prices through two channels: (1) they alter the effective marginal cost of lending, and (2) they create a moral hazard problem, dampening incentives for lenders to collect information about borrower quality. The combination of these two effects implies that guarantee programs disproportionately benefit high-risk borrowers and may even harm low-risk borrowers. In this paper, I quantify these channels in the setting of the SBA 7(a) Loan Program and find that more generous loan guarantees lead to a decline in average loan prices, but that benefits are concentrated among high-risk borrowers. In fact, low-risk borrowers receive lower surplus as guarantees increase, and moving from the 90% guarantee observed in the data to a rate of 50% would raise their surplus by 2.5%. The heterogeneity in impact suggests that alternative policies that moderate the effect of bank moral hazard could increase aggregate borrower surplus. I propose a hybrid policy design with a 50% guarantee and a subsidy set such that government spending is fixed. When compared to the baseline of 90% with no subsidy, the alternative policy mitigates the redistribution from low to high-risk borrowers and leads to a 1.6% increase in borrower surplus, on average, and 0.1 percentage point (1.6%) decline in the program’s default rate. Chapter 2, which is joint work with Paul Kim, studies the incentive design of the Paycheck Protection Program, a large loan forgiveness program implemented during the Coronavirus pandemic. The program was executed through private lenders with the goal of assisting small businesses in keeping their employees on payroll during the COVID-19 pandemic. We develop a model of PPP lending to capture the government’s tradeoff between inducing bank participation and targeting funds for use on payroll. Using the model, we establish that both increasing subsidies and relaxing forgiveness standards are effective in expanding credit access to borrowers seeking smaller loans. However, their efficacy in targeting (i.e., providing funds to businesses who will use them on payroll) depends on the correlation between loan amounts and borrowers’ return to payroll. We test the implications of the model using policy variation from the PPP Flexibility Act, legislation that relaxed forgiveness standards. Consistent with the predictions of the model, the average loan amount falls by between 6 and 7% in the period following the policy change. Furthermore, marginal borrowers are more likely than inframarginal borrowers to use funds for payroll, so making forgiveness more accessible increases the average share of funds used for those purposes. Chapter 3, which is joint work with Vivek Bhattacharya and Gaston Illanes, is a large-scale retrospective of US retail mergers. In particular, we document the price and quantity effects of all US retail mergers from 2006–2017 associated with deals larger than $340 million. Prices increase by 0.49% on average for merging parties, with an interquartile range of almost 5%. Non-merging parties exhibit slightly smaller price changes on average. Total quantities decline on average by 3–5%, but there is even larger variation across mergers. We investigate the role of synergies and market power by analyzing the timing of price changes, the relationship with locations of production facilities, and measures of market structure. We collect data on merger enforcement (remedy proposals in our case), and through the lens of a simple model, we estimate that agency preferences are such that they aim to challenge mergers where prices are expected to increase by more than 3.7–5.6% overall, or about 8.1–8.8% for merging parties.

Creator