Re-examining the Inequity of Risk Based Loan Pricing

Most credit unions today used risk-based pricing when granting loans.  That means the rate paid by each member is tiered from lower to higher based on their FICO or other scoring matrix.

This change from the initial co-op practice that all members be charged the same rate for the same loan was considered a major innovation by credit unions.  Consultants showed how to implement the program in the 1990’s. Compliant scoring models were empirically validated by experts.  One long serving CEO of a top 10 credit union said that implementing risk-based pricing for members was one of his most important contributions.

The practice is defended with two logics.   It expands credit union lending opportunities by qualifying more borrowers at the lower end of the risk scale.   Credit unions also claim they save these members money by charging a lower rate than other lenders would, if the credit union had not made the loan.

Today the vast majority of credit union leaders routinely accept these propositions.   Members deserve the rate their economic or past borrowing behavior merits.  Members with perfect credit should get better terms than members who have struggled with their finances.

The Reason Why Some Coops Do Not Follow Risk Based Pricing

The most public critic of this approach to consumer lending is Jim Blaine the retired CEO of State Employees (SECU) of North Carolina.

His reasoning follows:

The issue with risk-based loan pricing based on credit scores is that it imposes a real dollar penalty unjustly on folks in “the lower” credit tiers (tiers are usually A,B,C,D,E).

An easy example would be that say “D paper” with a statistically sound, model-projected 10% default rate might pay 12% for a car loan, while “A paper” folks with a projected 1% default rate pay only 6%. The real world penalty for the “D paper” folks is a 6% higher rate – not a minor cost!

The injustice arises because modeling cannot predict which 10% of the “D paper” folks will actually default. But, a substantial, unjustified interest rate cost (+ 6% in the example) is imposed on the entire D tier! Or in other words, 90% of the folks in the D tier (who the model has empirically proved will not default!) are paying an undeserved up charge – because, again, the model has validated that 90% of folks in the tier will actually pay!

 These folks have been unjustly charged because the model has profiled them into a class.

“Redlining” is an example of a financial “tiering” practice for real estate lending that has been discredited (and made illegal). “Redlining” imposed an unjust penalty on generally African Americans who lived in a particular neighborhood – many of these folks were unable to obtain loans at any cost – effectively a 100% profiling penalty!

Two Reasons To Review This Policy Now

Recent political dialogues have raised awareness of systemic inequalities that can accrue in society in critical areas of life: health care, employment options, education and housing.  These lead to structural inequity passed from generation to generation as accepted wisdom:  that’s just the way things are.

Financial services incorporate these histories when they underwrite members who may need financial assistance for one or more of these activities.

America is in an era of examining many past practices to understand the origins of present inequities.  Historical interpretations and beliefs are being re-evaluated.  New interpretations and additional data can lead to better, more equitable policy today.

Is risk-based pricing locking in systemic bias or is it fairly equating a member’s risk of default and pricing?   Each credit union will make its own decision.  It is far easier in moments of uncertainty to automate judgments via an impersonal model than to take time to understand an individual’s situation.

But shouldn’t every member-owner be evaluated on their character, capacity and circumstance? And if credit worthy, pay the same as other credit worthy members?

Should credit unions be more self-critical of lending practices that perpetuate disadvantages for those who have the least or know the least –and are frequent targets for predatory lenders?

The second reason to revisit this practice is that credit union have many years of portfolio performance from several economic cycles. Just as the impact of overdraft fees on members is now being reassessed, might it also be worthwhile for a credit union to review its risk-based loan pricing results?   Which members pay the highest rates? What is the loss rate of various credit tiers?  What might be the outcome if members in each class of loan had paid the same rate?  Should our policy be modified?

Finally, which members are most likely to need and use the credit union for their borrowing needs?    Is that group one for which the credit union is most able to make a real difference in their lives?

I don’t know the answers.   But if a credit union or analyst has done such an evaluation, I would be glad to share their analysis and any actions they may have taken.

 

Leave a Reply

Your email address will not be published.