Consumer Lending Discrimination in FinTech
Research Lunch Series — October 16, 2019
On Wednesday, October 16 the Center for Equity, Gender, and Leadership (EGAL) hosted their first Research Lunch Series event of the Fall 2019 semester, a presentation on “Consumer Lending Discrimination in the FinTech Era” by Adair Morse, Associate Professor of Finance at Berkeley Haas and a Fellow at the Berkeley Center for Law and Business.
Professor Morse’s research sought to estimate discrimination and understand the effect of FinTech (financial technology) algorithmic models used for housing loans. The research found that, while algorithms removed the bias experienced in face to face interactions, the proxy variables used by the algorithms could still result in discrimination. One example given was that a housing loan algorithm might look at an individual’s high school to determine the risk of loan default. The high school one attended can correlate strongly with wealth — which, under the law, is a valid measure to look at when reviewing default risk. However, high schools can also correlate with race or ethnicity, even after removing wealth factors. For this reason, it was determined that using high school data would punish or impact some minority households. Through this finding, the research shows algorithms do not completely eliminate bias and can even contribute to continued discrimination.
The talk generated a lot of interest and heated questions from students and faculty. Many were interested in Professor Morse’s research model and what could be concluded from her research. While the research is still ongoing, Professor Morse admitted that despite these findings, algorithmic models still reduce the amount of discrimination marginalized communities may face in person, and therefore it is better for these communities to seek out loans through this method.