Watch for fair lending risk when using alternative data

Clayton J. Mitchell, Kate Gutierrez-Wilson
7/14/2021
Watch for fair lending risk when using alternative data for lending

Using alternative data to inform lending decisions is a popular trend for banks and fintech companies exploring new ways to grow their business. Alternative data can be as simple as a customer’s history with a lender. Other examples include cash flow variability, social media metrics, and information related to rental data, cellphone payment history, and other publicly available records.

On the surface, many of these nontraditional credit metrics might seem like a benign way to reach new customers or to expand offerings, but alternative data also introduces the potential for fair lending risk. Many banks neglect to confirm that their alternative data models treat all customers fairly and ethically according to fair lending laws and regulations. And all it takes is one misstep to draw increased scrutiny from regulators and auditors.

You can benefit from using alternative data – but risks exist

You can benefit from using alternative data — but risks exist

Alternative data can provide valuable insights into audiences you might not be reaching, or it can give you a competitive edge in attracting new business. For example, you could use nontraditional data models to provide loans to creditworthy people who otherwise might not have had access due to damaged credit scores or thin credit files.

However, alternative data can also lead to unintentional exclusion. Let’s say your bank uses an algorithm that collects data for similarly situated audiences, and the results identify a group with strong credit history that could be targeted as potential customers. But the report might also unfairly omit people based on where they live or work. In this example, alternative data could create or worsen a potential fair lending risk.

When your company is considering a new product or service that uses alternative data, relevant stakeholders who can analyze the inputs and potential fair lending implications of alternative data should be included so that risks related to steering, redlining, and pricing aren’t inadvertently introduced. Further, evaluation of performance and impacts of alternative data models should be completed on an ongoing basis to identify potential unintended outcomes.

Collaboration is critical when it comes to fair lending and alternative data

Collaboration is critical when it comes to fair lending and alternative data

Fair lending issues often arise when your model development and sales teams are disconnected from the risk and compliance teams during the initial planning process. Without the proper background and preparation, fair lending professionals can have difficulty assessing the compliance of new products and services that involve alternative data.

Financial services companies also must monitor fintech and third-party relationships to make sure they aren't introducing uncontrolled fair lending risk. Including your risk and compliance team from the beginning will help confirm that the alternative data being considered is ultimately being used accurately, reliably, and fairly.

7 steps to help your company effectively use alternative data

7 steps to help your company effectively use alternative data

Your financial services company can use this seven-step cycle to help analyze and assess potential fair lending risks:

  1. Seek feedback from regulators so they won’t be surprised when you use nontraditional metrics.
  2. Document the business case for why you’re going to use alternative data.
  3. Evaluate and document the potential impact of model inputs related to fair lending risk.
  4. Establish key risk indicators and key performance indicators to predict potential discrepancies in new data models.
  5. Monitor outcomes and results for fair lending risk.
  6. Continually evaluate and adjust your data models.
  7. Repeat the process to account for changes in alternative data.

If you’re diligent about identifying and managing potential fair lending risk, your company can take advantage of nontraditional data models to attract new customers and help your business grow.

Crowe can help you understand fair lending risk that comes with alternative data

Crowe can help you understand fair lending risk that comes with alternative data

The fair lending landscape changes fast, and it can be hard to know where to start. Crowe fair lending risk specialists can help you determine the impact and potential scope of using alternative data. We combine data analytics with deep banking industry experience and regulatory knowledge to help you accurately assess fair lending risks and then develop the policies and procedures to address them.

Let’s connect

Have questions about how alternative data can affect your fair lending efforts? Get in touch – we’d be happy to talk.
Clayton J. Mitchell
Clayton J. Mitchell
Managing Principal, Fintech
Katie Gutierrez
Kate Gutierrez-Wilson
Financial Services Consulting