Why Identical Credit Scores Mask Delinquency Risk

A new study from MIT reveals that a person’s repayment patterns on loans vary strongly by geography and in particular by the hometown where a person grew up. 

Parents also play a role. When it comes to credit scores, you don’t outgrow your parent’s influence. Parents and their credit scores strongly influence an adult child’s credit scores.

In the research from MIT, economist Nathaniel Hendren and other co-authors studied access to credit in the United States. According to the research, stark credit score gaps exist based on race, parental income, geography, and education, and these gaps exist even after the authors controlled for income. 

Credit score gaps already exist at the beginning of adulthood, and they get only modestly smaller as people age. Where do changes in credit scores begin for young adults? The divergence in credit scores among young adults usually begins with small delinquencies on phone or utility bills.

The study found that credit scores underestimate race and class differences in future delinquencies. Black and Hispanic consumers are more likely to fall delinquent than white consumers with the same credit score.

An analysis of consumers ages 35 to 42 with identical 650 credit scores revealed that 61% of Black individuals have a 90-day or more delinquency compared with 47% of white individuals, 52% of Hispanic individuals, and 39% of Asian individuals.

The study defines delinquency as being 90 or more days past due on any line of credit. This delinquency definition aligns with the target estimand of the credit score.

Where consumers live influences behaviors in credit and labor markets, including debt repayment and credit access. A child that grows up in a place where adults repay debts are more likely to repay their own debts as adults regardless of their income.

When it comes to credit scoring, place stumps race. The study found places with higher credit scores for white individuals were places where individuals of other races had higher credit scores as well.

Friendships also makes a difference. In a place where a child with low family income has a friends of higher income, something interesting happens. The low-income child will be less likely to be delinquent as an adult.

Subprime lenders already take on borrowers with weaker credit scores and payment histories. These subprime customers also have a greater risk of falling delinquent on their accounts.

Whether a borrower pays a credit account on time impacts a lender’s default and delinquency rates. These rates are key factors in how a subprime lender will price credit. Lenders also need to consider how to set their financial reserves to offset any losses.

A subprime customer’s payment history also affects a subprime lender’s profitability because late or missed payments result in a decline in interest income and increases in collection costs.

Even factors that are not financial in nature — such as hometown and parent credit score and credit behavior — can predict the likelihood of repayment. These non-financial predictors can help lenders assess risk and look beyond the traditional credit score when assessing present and future customers. 

Nonfinancial factors also can help to improve and refine underwriting models used by lenders.

The models would be able to tell the difference between borrowers who may appear risky but are actually able to repay versus those borrowers who are more likely to default. A more refined underwriting model would improve a subprime lender’s portfolio performance. 

By identifying characteristics that improve credit scores, such as parent credit scores and credit behavior and the repayment norms of hometowns, subprime lenders can gain a competitive edge by reducing the number of losses from defaults, expanding their lending into underserved segments, and offering more competitive rates.

According to the study, credit scores may understate or overstate the credit risks for different ethnic and racial groups. That’s why a traditional credit score could identify a consumer as having the wrong credit risk, and subprime lenders using that scoring model could overcharge or undercharge for that risk.

Both overcharging or undercharging for credit risk impacts a subprime lender’s profitability and exposure. Recognizing biases in credit scoring may result in better and fairer pricing models that account for risk but don’t exclude good borrowers.

A recent research project by professors at the University of Illinois Urbana-Champaign pointed to widespread bias in both credit scoring and in mortgage lending. 

A person’s parents and the hometown they grew up in has a big influence on their repayment patterns as an adult. The first delinquency for young people may be on small bills such as phones and utilities.

Bias in credit scoring may result in lenders overcharging or undercharging for a credit risk affecting a lender’s profitability and exposure.