Archived

This content is available here for research, reference, and/or recordkeeping.

Date Available

7-24-2026

Year of Publication

2026

Document Type

Doctoral Dissertation

Degree Name

Doctor of Philosophy (PhD)

College

Business and Economics

Department/School/Program

Finance and Quantitative Methods

Faculty

David Sovich

Abstract

This dissertation studies how technology, market structure, and contract design shape outcomes in consumer credit markets, with a particular focus on auto lending and leasing. In the first chapter, I investigate the performance of automated underwriting systems relative to human underwriters in the auto loan market. While prior studies document the advantages of automated underwriting, I show that its performance deteriorates under conditions of heightened data uncertainty. Exploiting the COVID-19 pandemic as an exogenous shock and using a combination of difference-in-differences and regression discontinuity designs, I find that automated underwriting performs significantly worse than human underwriting during this period, particularly for higher-risk borrowers whose income and employment were more likely to be disrupted. These findings highlight the limitations of algorithmic decision-making in environments that fall outside the scope of historical training data.In the second chapter, I examine how tax subsidies are passed through to consumers in the auto lease market using a novel dataset and a tax policy change in the state of Georgia. I find that auto dealers capture a substantial share of the subsidy, while consumers use about half of their tax savings to upgrade to more expensive vehicles. In contrast to prior work in consumer credit markets, I find no evidence that borrower characteristics, including credit scores and past experience, explain heterogeneity in passthrough. Instead, the results suggest that market structure in the auto lease market is the primary determinant of subsidy incidence.In the third chapter, I study the causal effect of interest rates on ex-post default in the indirect auto loan market. Using lender-specific discontinuities in loan pricing as a source of quasi-exogenous variation, I find that a 100-basis-point increase in interest rates raises the auto loan default rate by 41 basis points. This effect is concentrated among liquidity-constrained borrowers. I find no evidence that the relationship is driven by lower default costs or stronger strategic default motives. The results indicate that borrowers’ ability to pay, rather than their incentive to default, is the main mechanism linking higher interest rates to default.

Digital Object Identifier (DOI)

https://doi.org/10.13023/etd.2026.359

Archival?

Archival

Download

Share

COinS