Recent Research: Fintech increases financial inclusion and reduces discrimination, yet regulatory challenges lurk

A review of recent reports finds the rise of financial technology (fintech) has the potential to improve the financial health and literacy of the traditionally underbanked and decrease discriminatory practices as more people gain access to services and are included in financial markets. However, regulators face new challenges as a result of fintech.

Fintech can provide access to less costly, low-dollar financial services, increasing capital access to a greater number of people, according to the Atlanta Federal Reserve. Tools such as alerts when purchases may trigger overdraft fees and when payment due dates are nearing, budget building capabilities, and services such as online education programs that encourage better management of financial affairs and promote improved credit ratings, all have the potential to increase financial inclusion. Other fintech tools enable wealth management firms to not only decrease user fees and minimum investments by leveraging robo-advisors, but also provides comprehensive portfolio management and learning tools; each aspect decreasing the barriers to entering the market and increasing financial inclusion.

Recent research indicating that fintech may also be decreasing discrimination when compared to traditional providers of financial services. A recent study found that fintech algorithms used in assessing credit risk for mortgage loans discriminated against African American and Latinx borrowers 40 percent less often than traditional face-to-face lenders. This reduction in discriminatory practices has the potential to save African American and Latinx borrowers an average of up to $767 million annually in unnecessary interest payments.

However, another recent study found that credit risk assessment methods utilizing big data and machine learning (BDML) may not eradicate discrimination in fintech lending. Even though BDML algorithms assess a wider variety of factors than traditional FICO scores and are proficient in eliminating discriminatory decisions based on human prejudices and biases, these broader inputs are not themselves immune to the effects of systemic bias. That is to say, the long history of biased practices still results in statistically discriminatory outputs from BDML models. Nevertheless, the authors find that these models are more likely to produce better outcomes for groups traditionally subjected to discriminatory lending practices. The researchers also suggest that as BDML algorithms grow more sophisticated, increasing the use of alternative data sources will continue to reduce disparities in racial and ethnic financial inclusion.

While fintech has potential for increasing financial inclusion, myriad regulatory challenges lie ahead. Issues of data ownership and privacy, inadequate community reinvestment by companies with extremely geographically limited and concentrated physical presences, and the criminal use of fintech are all areas where regulation may be necessary to ensure and strengthen gains in equity and inclusion made possible by fintech.