Leveraging Fintechs and Big Data in a Fair Lending Focused 2022

Over the past decade, financial technology companies (fintechs) have been on the rise, disrupting all segments of the financial industry. Their innovative technology, swift adaptation to market trends, and ability to create personalized and efficient customer experiences enabled them to reshape the way we look at, use, and provide financial services.

Banks Can Leverage Fintechs’ Advanced Technologies

At the onset, fintech companies and traditional financial institutions viewed each other solely through the lens of competition, fighting to maintain and gain valuable shares of the market. Over time, this competition turned to collaboration and foes turned to friends as banks and fintechs began to form partnerships, each bringing their competitive edge over the other to a mutual table. These relationships are created in variety of forms, ranging from a referral-based partnership to overall acquisitions. Commonly, banks and fintechs collaborate in such a way that both parties maintain equal control and responsibility over the full customer experience.

In forming these partnerships, banks can leverage fintechs to enhance their business across all segments. Fintechs create models using big data and artificial intelligence (AI) that can streamline many aspects of the business to minimize a bank’s labor costs, reduce human error, and increase the speed of accepting applications and underwriting loans. They also have the innate ability to develop consumer-centric apps, which in the age of smart phone technology is a key facet of consumer acquisition and retention. Rather than expending vast amounts of resources on developing their own models, banks are able to use those already developed by fintechs and implement them into their own customer base. This enables banks to create a more unique experience by tailoring their products and services to each individual customer. Additionally, this technology can provide a modernized customer service experience by providing answers, assistance, and even financial education through the click of a button. Fintechs also use machine learning, a branch of AI that uses big data to build models that update themselves without human interaction. Therefore, as customers’ needs change, the algorithm can automatically adapt and even predict future needs, providing consumers with a constant stream of personalized products and services.

Big data not only helps with customer experience, but it can help financial institutions improve in accuracy and compliance. This can be particularly beneficial to smaller banks by alleviating their struggle to find enough resources and bandwidth to stay current with regulatory changes and update their systems accordingly. The ability to quickly search through hundreds of documents and thousands of data points can help detect irregular patterns, areas of concern, and fraud. Fintechs reap their fair share of benefits as well; banks can provide fintechs with access to funds, consumer confidence, and a deeper knowledge of the regulatory compliance landscape.

Increasing Access to Credit for the Unbanked Community 

These partnerships are not only mutually beneficial, they are valuable to consumers as well. In addition to improved customer experiences, these partnerships can bridge a gap in lending by providing credit access to the previously underserved. In 2019, the Federal Reserve reported that 22% of American adults are unbanked or underbanked. These adults were more likely to have low income, less education, or be in a racial or ethnic minority group. Often, those that fall into these categories are unable to participate in the mainstream financial lending market, including the ability to access safe, affordable credit. Many unbanked Americans are continuously denied or fail to even apply for loans because they are credit invisible or have a limited credit history. With the aid of fintech and big data, banks have the opportunity to open the market to those currently left out. Increasing technology enables alternative data to be used to predict a borrower’s ability to repay. Algorithms can now rely on social media, utility payments, cash flow, and spending records to show creditworthiness. In fact, big data can provide algorithms with thousands of data points enabling them to perform a predictive analysis of a borrower’s repayment ability, leading to an increase in approval ratings regardless if a consumer has a bank account or traditional credit score. Therefore, these bank-fintech partnerships are expanding access to credit throughout the country by enabling lenders to expand into consumer markets that historically have been underserved.

However, these breakthroughs have not been without their limitations. Compliance challenges are mounting and are anticipated to become greater in both scope and complexity during the Biden administration, particularly regarding fair lending and fair servicing.

Existing and Emerging Regulatory Concerns 

At the center of the fair lending framework is the Equal Credit Opportunity Act (ECOA) and the Fair Housing Act (FHA). These statutes prohibit creditors from discriminating against consumers on the basis of race, national origin, gender, or marital status in any part of a credit transaction. While these laws are not new, financial institutions will encounter increased compliance challenges from the same technological advances that provide expanded lending options. As mentioned above, AI systems use thousands of data points to make decisions when extending credit. A problem ensues, however, when it cannot be clear which factors are being used to reach these decisions. One common compliance hurdle can be illustrated by ECOA’s adverse action requirement. ECOA mandates that when an adverse action, such as a denial of an application, is given to a consumer, a creditor must provide a written statement listing the reasons for such denial. Because of the copious amounts of data being analyzed for each decision, algorithms are unable to recite the deciding factors and provide an explicit, understandable explanation. This inhibits creditors from being able to provide applicants with the information required by the statute. This is often referred to as the “black box” problem, a situation in which users, creators, or even the algorithms themselves cannot decipher the complex innerworkings of their own decision making.

This “black box” problem leads to broader discrimination concerns in fair lending as well. While AI has the potential to remove human biases from decision-making, it is not free of its own. Although companies can ensure that their algorithms are not considering prohibited characteristics, the algorithms can still produce discriminatory results. Even if inputs into the model are not proxies for discrimination, nor the data as a whole, with such a complex algorithmic formula it can be unclear whether the data points interact with each other in way that results in them functioning as proxies for protected groups. To avoid this, algorithms must ensure that each data point has a concrete nexus to creditworthiness.

Looking ahead to 2022, financial institutions will need to keep a close eye on agency trends. Fair lending is one of the administration’s key priorities, and Biden’s appointees have started to emphasize this priority accordingly. In October, the DOJ, CFPB, and HUD announced their “Combatting Redlining Initiative,” describing it as one of the most aggressive and coordinated efforts to combat discrimination in lending. Regulators explained that they are not only focused on traditional redlining, but seemingly neutral algorithms that may reinforce biases that have long existed. The DOJ stated that it will be targeting lenders of all types and sizes, with CFPB Director Rohit Chopra echoing this in his remarks. In discussing AI in lending, he emphasized that the Bureau will look closely at both depository and non-depository intuitions.

For the Future

With all members of the financial ecosystem becoming perspicuously aware of rising challenges associated with advanced technology, they should not work in silos to combat discrimination in lending. Lawmakers issued general statements encouraging the use of technological advances to reach underserved market segments while concurrently marking their intent to enforce misuse of these techniques. There is not only a lack of clarity and guidance for financial institutions to lean on, but existing gaps in the regulatory framework that create uncertainty about whether advanced lending methods are consistent with fair lending laws. Financial regulation currently lacks the ability to incorporate algorithmic models into this framework, leaving little room for innovation and larger spaces for noncompliance. It is important for financial institutions and fintechs to not only work together but to collaborate with regulators to find solutions that encourage innovation and preserve compliance. By leveraging big data and fintech technology, banks can be at the heart of efforts to dismantle discriminatory lending and maintain compliance while simultaneously increasing access to safe, affordable credit.

Republished with Permission. As seen on Global Banking and Finance Review. This article was first published on December 28, 2021.