Case Study: Fair Lending Analysis of Indirect Auto Loan Pricing

This report details a fair lending analysis conducted for “First Regional National Bank” examining potential gender-based disparities in indirect auto loan pricing. Using econometric modeling with variables like credit score and loan amount, the analysis compared female applicants (target group) to male applicants (control group). The results, presented in regression tables for individual dealers and a pooled dataset, found no statistically significant pricing differences between genders after controlling for relevant factors. This suggests loan decisions were based on established criteria, not gender. The study was conducted by Premier Insights, Inc.

The information in this case study is based on the provided report and is intended to illustrate Premier Insights, Inc.'s analytical capabilities. The names of the financial institution and market areas have been changed for confidentiality purposes.                                 

Introduction

Premier Insights, Inc. (Premier) conducted an in-depth analysis of indirect auto loan pricing for First Regional National Bank (the “Bank”) as part of the Bank’s fair lending management program. This case study demonstrates the application of advanced econometric modeling to ensure fair lending practices. The analysis covered loan data and focused on potential disparities in pricing between male and female applicants.

Background

The Bank, like many financial institutions, wanted to ensure its lending practices were equitable. Premier was engaged to perform a rigorous statistical analysis to determine whether there were any statistically significant disparities in loan pricing based on gender. The analysis was designed to identify any such disparities and, if found, provide an understanding of the underlying factors.

Methodology

• Data Scope: The analysis included indirect auto loans purchased from dealerships in the Greater Capital City and surrounding areas, where the Bank had established relationships. The loans were all purchased by the Bank with full recourse.
• Target and Control Groups: Female applicants were designated as the target group, and male applicants served as the control group. These designations were determined using a proxy based on the Census name database.
• Econometric Modeling: Multivariate regression analysis was the primary tool used. This method allows for the simultaneous evaluation of multiple factors to isolate the influence of each independently.
• Model Variables: The regression models included factors considered in loan pricing: credit score, loan-to-value (LTV) ratio, loan amount, age of collateral, and loan term. Loans were analyzed both individually by dealer and as a pooled dataset, with controls for each dealer.
• Addressing Missing Data: If data were missing, models were initially estimated without those variables. If a variable proved significant, a control was introduced to account for the missing data rather than removing the observation.
• Statistical Testing: The models were used to test for statistically significant differences in pricing based on gender. The analysis determined the direction of any differences, with a positive coefficient indicating differences favoring the male control group and a negative coefficient favoring the female target group.

Findings and Results

• Overall Analysis: The regression analyses revealed that gender was not a statistically significant factor in loan pricing when other relevant criteria were considered.
• Combined Dealer Analysis: The combined dataset of all dealer loans did not show a statistically significant difference in loan pricing between male and female applicants, once the control variables were added.
• Individual Dealer Analysis: Results from individual dealerships were also largely consistent, showing no significant gender-based disparities in pricing. Some dealers had too few observations to draw statistical conclusions.
• For example, at Bassett Auto, LLC, the male-female difference in the predicted rate was -0.257 with a p value of 0.413 when considering only female borrowers and -0.193 with a p value of 0.539 when adding control factors.
• At Matthews Super Mart Auto Sales, Inc., the male-female difference in the predicted rate was 0.302 with a p value of 0.093 when considering only female borrowers and 0.045 with a p value of 0.736 when adding control factors.
• Statistical Significance: The p-values associated with the male-female differences were not statistically significant in most cases, reinforcing the finding that pricing was independent of gender.

Conclusion

The findings from this analysis demonstrate that the loan decisions made by the Bank were based on appropriate criteria, such as credit score, LTV, loan amount, age of collateral and term, and not on the gender of the borrower. The study found no evidence of disparate treatment based on gender in the indirect auto loan pricing at the Bank.

Value of Premier Insights, Inc. Services

This case study exemplifies Premier Insights, Inc.’s ability to perform rigorous, data-driven analyses for financial institutions. By utilizing advanced econometric modeling techniques, Premier can assist banks in ensuring fair lending practices. This helps banks mitigate risk and comply with fair lending regulations. The ability to control for various factors simultaneously, provides a clear understanding of the key drivers of lending decisions, and an ability to demonstrate compliance.

About Premier Insights, Inc.

Premier Insights, Inc. is a research and consulting firm specializing in fair lending, Community Reinvestment Act (CRA), and marketing research for the banking industry. Our expertise in econometric modeling and data analysis, combined with our in-depth knowledge of banking regulations, provides invaluable support to financial institutions.

This case study highlights the type of rigorous analysis and valuable insights Premier Insights can provide for its clients.