Author
The research community at the Institute includes visiting scholars, consultants, economists, research analysts, and research assistants. These scholars bring a diversity of backgrounds, interests, and expertise to research that deepens our understanding of economic opportunity and inclusion as well as policies that work to improve both.
What caused the mortgage crisis that tipped the U.S. into the Great Recession? One intuitive story remains prevalent more than a decade later: A housing bubble was inflated by ill-advised lending to subprime borrowers.
On the heels of the financial crisis, Institute visiting scholar Stefania Albanesi was studying personal bankruptcy reform at the New York Fed when she sensed this subprime narrative “wasn’t squaring with what we were finding.” Borrowers with low credit scores had not taken out a higher share of mortgages or defaulted much more than they had historically. “What we did see during the foreclosure crisis was a lot of high-credit-score borrowers defaulting on their mortgages,” she said, especially those whose good credit had allowed them to purchase investment properties.
Albanesi says this false lesson—unduly pinning the Great Recession on subprime homebuyers—makes mortgages even harder to get for people with lower credit scores, who tend to be younger, lower-income, and non-White.
This research also led Albanesi to question the mechanics of credit scores themselves. The widely accepted scores assigned to almost every American—by for-profit companies using opaque formulas—seem to have missed the mark on predicting the risk of consumer default. She wondered: Could artificial intelligence do better? Albanesi plans to continue exploring this question during her visit at the Institute, extending research launched with former graduate student Domonkos Vamossy.
The short answer so far: A.I. seems far superior. “Based on our model, conventional credit scores misclassify 30 to 50 percent of consumers,” Albanesi said, especially those with low credit scores. “Being categorized as ‘deep subprime’ if you are in fact subprime or near-prime, for example—that’s a very big difference in the borrowing conditions that you face.”
The magazine of the opportunity & Inclusive Growth institute
Their “deep machine-learning” algorithm digests massive amounts of consumer data to derive credit scores that appear fairer and more accurate than those that presently govern our financial lives. One might hope that the credit, lending, and loan securitization industries would share these goals; Albanesi is more optimistic that policymakers could be motivated by the findings.
This work on credit scores and first-time homebuyers complements her other projects that illuminate the economic plight of young people. How does student debt postpone the decision to start a family and take out a mortgage? Why does today’s labor market push “boomerang” college kids back home with their parents?
“As economists, we should really care,” Albanesi said. “What happens in the five to 10 years after you graduate from college or high school can shape your entire life’s trajectory.”
This article is featured in the Spring 2023 issue of For All, the magazine of the Opportunity & Inclusive Growth Institute
More scholar spotlights from this issue
Jeff Horwich is the senior economics writer for the Minneapolis Fed. He has been an economic journalist with public radio, commissioned examiner for the Consumer Financial Protection Bureau, and director of policy and communications for the Minneapolis Public Housing Authority. He received his master’s degree in applied economics from the University of Minnesota.
