<img src="https://ws.zoominfo.com/pixel/pIUYSip8PKsGpxhxzC1V" width="1" height="1" style="display: none;">

5 Questions to Learn if Fair Lending Regression Analysis is Right for You!

author
3 min read
Jul 14, 2016

Regression Analysis is a powerful tool to improve your Fair Lending compliance, and we've recently seen an increase in questions about what it is and who it can help. In this post, you'll learn the five questions you need to ask if Regression Analysis is the right tool for your institution.

To accurately understand your Fair Lending risk, you need to understand the story your data tells about your institution. Why? Because the goal of Fair Lending is to ensure that similarly situated individuals are treated equally all the way through the crediting process; disparities in applications, originations, denials and more across demographic groups may indicate a difference in treatment or impact. That's why the regulators will start with data analysis to identify any disparities when they evaluate your Fair Lending risk.

There are many different approaches for conducting data analysis. One popular option is Fair Lending Regression Analysis. Wondering if it's right for you? These 5 questions will help you discover the answer.

1. Do you want to know what your data says about you?

This is kind of a trick question, because every compliance officer at every institution should want to know the story their data tells about them.

That's because the best defense is a good offense. When the regulators analyze your data, they will begin to develop a story about your institution and your Fair Lending performance. It's important to know the story your data says, so that you can explain or defend any disparities and mitigate any risks proactively.

Regression is a powerful tool because it provides insight about the relationship between different variables. In Fair Lending, it can help determine the statistical relevance of any potential disparities. That means it can isolate certain variables and help you determine whether (and how much) they impacted the lending activity. 

2. Do you have lending disparities? 

Sometimes, a disparity might be easy to explain - maybe only a handful of people applied and one was denied, or it was a fixable error in the data. Sometimes, a disparity may be more difficult to understand. That's where regression can help.

Fair Lending regression analysis is perfect for analyzing loan data to determine if a prohibited basis factor, such as race, gender, ethnicity, and/or age, had an impact on the decision to make or price a loan. If your denial or price disparity rates are high, regression analysis can provide clarity and insight into your Fair Lending risk.

Regression models will also estimate the pricing and/or probability of denial to help identify outliers, or situations where what occurred was different than what was expected. This is sometimes called Matched Pairs, or Comparative File Analysis. With individual file reviews and comparisons, you can examine why disparities may have occurred for those Protected Class loans.

 [Exciting Update: Matched pairs analysis is a standard feature of Nfairlending. To learn more about Nfairlending, our updated compliance analysis software, click here.]

3. Are you a high-volume lender?

Regression analysis is more effective with larger data sets, so we recommend analyzing approximately 1,000 records or more. Regression for Fair Lending compliance can be used to analyze any loan data - from credit cards to auto loans - so it's not just about HMDA records.

For smaller data sets, sophisticated regression modeling is not necessary to fully understand your Fair Lending risk. Still, data analysis should be a part of your compliance management program.

4. Can your data support regression analysis?

This may seem obvious, but regression won't work when there is little or no variation in the independent or dependent variables. For example, if all or none of your applicants are members of a protected class, or all  or none of your applicants were denied, regression won't be able to determine whether those variables were involved in the loan decisioning or pricing. That said, this is usually not a problem in larger data sets.

Likewise, good data integrity is essential for any kind of analysis, including regression.

5. Are you ready to know the story your data tells?

If you have high disparities, sufficient loan records, and good data quality, Fair Lending regression analysis may be a great option for you. It is a deep dive analysis that will provide clarity and help you reduce Fair Lending risk unlike any other method.

If you decide to use Ncontracts for your Fair Lending Regression Analysis, our team of compliance experts will review the results with you and provide guidance for improving your Fair Lending compliance program. 

Ncontracts Viewpoint: Regression analysis is a powerful Fair Lending compliance tool, but it's not a fit for everyone. If you have questions about regression analysis and whether it's right for you, contact Ncontracts today.

Want to learn more? We conduct regression analysis on credit card, mortgage, indirect auto and other loan types for clients around the country. From gathering the data to reviewing the results, we'll be with you every step of the way to help, support, clarify and strengthen.

And for more guidance, check out our article on Lending Compliance Management

 

How to Build a Strong Fair Lending & Redlining Compliance Management System


Subscribe to the Nsight Blog