Case Study: Analysis of Bank Loan Pricing Disparities

This study analyzes a Federal Deposit Insurance Corporation (FDIC) investigation into potential gender-based loan pricing disparities for First Unity Bank. The study refutes the FDIC's findings of discriminatory pricing by demonstrating that a more comprehensive model, including factors like branch location and credit score, reveals no statistically significant difference in interest rates between male and female borrowers. The FDIC's model is criticized for its smaller, non-random sample size and exclusion of key variables.

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) was retained to conduct an independent analysis of loan pricing practices at First Unity Bank (the “Bank”), specifically regarding unsecured consumer loans originated during a two-year period. This analysis was prompted by an inquiry from the FDIC which suggested a potential gender-based disparity in loan pricing. The goal of our analysis was to determine whether such a disparity existed and if so, what factors contributed to it.

Background

The Bank had two branch locations, one in the town of Oakhaven and a newer branch in the town of Crestwood. The Bank's loan portfolio included unsecured consumer loans. The FDIC had identified a potential issue regarding the pricing of these loans, suggesting that female borrowers might have been charged higher interest rates than male or joint borrowers.

Premier Insights' Approach

• Data Collection and Preparation: Premier obtained data for unsecured consumer loans originated by the Bank in a period of two years. Loan excluded were employee loans, special promotion loans, an agricultural loan, and a secured loan from the analysis. This approach mirrors the methodology typically used in such analyses.
• Statistical Modeling: A regression model was constructed to analyze loan pricing, incorporating factors identified by the FDIC, including:
  • The existence of a deposit relationship
  • An indicator for auto-draft payments
  • Credit score
  • Branch location
• Model Execution: Three separate models were executed: one each for each year’s loans, and one for the combination of both years’ loans. Loan pricing was measured by the contract note rate.

Key Findings

No Gender-Based Disparity: Our analysis found no statistically significant disparity in loan pricing between male/joint borrowers and female borrowers when controlling for relevant factors. The coefficient for "female" in all models was not statistically significant.

• Impact of Branch Location: Branch location was a highly significant factor in loan pricing. The Crestwood branch, which was newer, tended to have lower interest rates compared to the Oakhaven branch. The Crestwood branch had an average interest rate of 10.75%, while the Oakhaven branch had an average interest rate of 13.46%.
• Significance of Other Factors: Other explanatory factors, such as credit score and deposit relationship, were also significant in the models.
• Limitations of FDIC Analysis: Premier identified key weaknesses in the FDIC’s analysis:
  • Incomplete Sample: The FDIC sample included only half of the originated loans.
  • Omission of Key Factors: The FDIC model did not account for branch location and omitted credit score. The FDIC model included many irrelevant factors, such as bankruptcies and medical collections.
  • Potential for Bias: The FDIC approach potentially attributed pricing differences to gender when they were actually the result of unaccounted factors.

Explanation of Branch-Based Pricing Differences

The Crestwood branch, opened in 2007, had newer loan officers who followed the Bank's cost of funds index more closely when setting pricing. The customer base at the Crestwood branch was also newer to the Bank. In contrast, loan officers at the Oakhaven branch had been with the Bank for a longer time and worked primarily with repeat customers.

Conclusion

Premier’s analysis demonstrated that when appropriate and relevant factors are considered, there is no statistically significant gender-based disparity in loan pricing at the Bank. The key factor in pricing differences was the branch location, with the Crestwood branch generally offering lower interest rates. The FDIC analysis was the result of incomplete data, the omission of key factors, and the inclusion of irrelevant variables. These weaknesses led to potentially spurious conclusions. This case study highlights the ability of Premier Insights, Inc. to conduct thorough analyses that provide accurate and reliable results.