Could a new paper published by researchers from the Federal Reserve Bank of St. Louis finally answer the question of whether affordable housing laws caused the subprime crisis?
There are good reasons why this idea has held on for so long. While political biases may be behind its refusal to die, honest observers can also see some logic in the camp of the those who blame Uncle Sam for the subprime crisis. After all, the Community Reinvestment Act did encourage lenders to give poor people more access to credit. And Congress definitely did set explicit targets for Fannie Mae and Freddie Mac to help increase mortgage lending to areas with high levels of minorities and the poor to support affordable housing.
It doesn’t take a big leap to think that the government’s nudges may have led to mortgages being made to people who couldn’t afford to repay them. It makes for a nice, tidy explanation. And heaven knows how we love nice and tidy. But, as Rubén Hernández-Murillo, Andra Ghent and Michael Owyang show, this is actually not what happened.
Their research asks whether there’s evidence that lenders changed their behavior to meet CRA, Fannie or Freddie-mandated goals by:
- making more subprime loans than they otherwise would have?
- lending to riskier borrowers than they otherwise would have?
- charging risky borrowers less than they otherwise would have?
The CRA, Fannie and Freddie have specific metrics for evaluating whether a particular loan — or a loan-turned-mortgage-backed-security — actually qualifies towards their affordable housing or credit access goals. These metrics are generally related to the racial and / or income composition of a certain census tract or individual borrower and serve as strict cut-offs in regards to program goals. For example, if a bank makes a loan to someone whose income is less than or equal to 80% of the median income for that “metropolitan statistical area” then that loan will count towards the bank’s mandate CRA target (that is, if it’s covered by the CRA; not all financial institutions are.) If their income is ever so slightly over this 80% bar, then a loan to them would not help the bank hit its CRA target.
The very presence of these kind of black-and-white thresholds allowed the researchers to use a regression discontinuity approach to see whether there were any statistically significant increases in the number of loans made to people who fell just inside of the threshold as opposed to those who fell just on the outside of it. If the banks were strongly influenced by their desire to meet affordable housing targets then they should have made more of the kind of loans that would qualify and less of the kind of loans that would not qualify. These relative changes should show up as anomalies in an otherwise continuous data pattern, but the researchers’ examinations of 722,157 securitized subprime mortgages made in California and Florida from 2004-2006 found no evidence of this.
The same held true for the other two research questions. Hernández-Murillo, Ghent and Owyang did not find that lenders gave lower (i.e. better) prices to borrowers falling inside the threshold vs. those falling outside. Nor did they find evidence that lenders lent more often to qualifying riskier borrowers (those who would go on to be seriously delinquent within the first two years after receiving the loan) than they did to riskier borrowers whose loans would not qualify.
So, to recap, the subprime story went something like this: foolish borrowers borrowed and foolish lenders lent. The evidence appears to show that lenders did not change their behavior in order to hit the affordable housing targets that the government imposed on them. While the government’s programs to encourage affordable housing may have other flaws they did not, at any rate, directly cause the subprime crisis. Myth busted.