A Simple Equation

New research shows why preventing another mortgage crisis might come down to cultivating better basic math skills.
May 20, 2010
Print this page

The precise cause of the subprime mortgage crisis remains subject to debate, but one fact is clear: homebuyers are far more likely to default on certain types of mortgage contracts than on others. Adjustable-rate mortgages (ARMs), for example, have a much higher probability of entering delinquency and foreclosure than other types of subprime mortgage contracts.

One popular explanation is that even as the subprime mortgage market made credit available to consumers who had previously been shut out of the housing market, lenders took advantage of unsophisticated consumers by steering them towards the riskiest mortgage products. These vulnerable borrowers, who often lack even elementary financial skills, took on attractive but risky subprime mortgages that they couldn’t afford and on which, in many cases, they subsequently defaulted.

Professor Stephan Meier worked with Kristopher Gerardi of the Federal Reserve Bank of Atlanta and Lorenz Goette of the University of Lausanne in Switzerland to see if they could find evidence that the most financially naïve consumers were ushered toward the most treacherous mortgages.

The researchers surveyed subprime mortgage borrowers listed on registers of deeds in three northeastern U.S. states, asking participants to complete a three-part survey that assessed respondents’ numerical ability (that is, basic math skills), economic literacy and cognitive ability.

It’s important to view numerical ability as one facet of financial and economic literacy rather than as synonymous to it. “The ability to understand basic mathematical concepts may be important for economic literacy,” Meier explains, “but it does not reflect whether or a person understands basic concepts about the economy, like what compound interest is or what inflation does, which we consider economic literacy.” Nor is a college degree a guarantee of facility with numbers. “Yes, there’s a much lower probability that a person with a college degree can’t divide 300 by two — the easiest of the questions we ask on the numerical ability survey,” he says. “But even people with college degrees can have difficulty dealing with numbers.”

The researchers matched the survey scores with the deeds and repayment history of each participant to see which mortgages were current, delinquent or in foreclosure, and what types of mortgages each individual held. The comparison revealed an extreme correlation between numerical ability scores and delinquencies: borrowers with the lowest numerical ability were delinquent almost 24 percent of the time, compared to 12 percent for borrowers with the highest numerical ability. Borrowers with the lowest numerical ability had a 21 percent rate of foreclosure compared to about 7 percent for the highest ability group.

Significantly, the default rates for those in the lowest numerical ability group were similar across all types of mortgages, rather than being more heavily skewed toward the most risky types. That result implies that the relationship between high rates of default and low rates of numerical literacy is not due to borrowers taking on too much debt or being systematically steered to the riskiest mortgages.

These consumers may not have been more susceptible than were their savvier peers, but, Meier suggests, they were susceptible to their own lack of financial know-how. “People with low numerical ability are not just mismanaging their money when it comes to mortgages; they are mismanaging their finances in general,” he says. “We’ve all been confronted with a recession that required everyone to readjust their budgets, spend less and think more about savings. These are people who just don’t have the skills to do that, and they’ve become trapped in a cycle of financial vulnerability.”

The public policy challenges, then, are difficult at best. In Meier’s view, much-touted plain vanilla mortgages are probably not the answer for the most vulnerable borrowers, because even the more straightforward terms of these mortgages would not address the dearth of basic math skills and overall lack of financial literacy skills. He suggests that lenders consider applying a more rigorous screening process for prospective borrowers, including simple testing for basic math skills. Better yet, lenders could offer support to the most vulnerable borrowers, such as helping with budgeting, providing tools that make it easier to track budgets and providing courtesy reminder calls as payment due dates approach.

In the short run it may be tempting for lenders to squeeze as much money as they can out of customers who don’t know better. “But in the long run, those customers are more likely to default, and there go the profits,” Meier says.

Other research has shown that a surprisingly large proportion of Americans lack a basic level of financial and economic literacy. And, Meier points out that the financial decisions everyone faces have become much more complex than they once were. “We choose between defined benefit or defined contribution retirement plans. Everyone has a credit card now. And — at least during the housing boom — everyone could buy a house,” he says. “An investment in financial education would be a good one because the returns would only increase as financial decision making becomes even more complicated.”

Stephan Meier is assistant professor of management at Columbia Business School.

Stephan Meier

Stephan Meier is an Associate Professor at Columbia Business School. He holds a PhD in Economics from the University of Zurich, was previously a senior economist at the Center for Behavioral Economics and Decision-Making at the Federal Reserve Bank of Boston and taught courses on strategic interactions and economic policy at Harvard University and the University of Zurich. His research interest is in behavioral...

View full profile

Read the Research

Kristopher Gerardi, Lorenz Goette, Stephan Meier

"Numerical Ability Predicts Mortgage Default"


View abstract/citation
View Research