In the aftermath of the 2008 American housing crisis and the subsequent global economic recession, John Derbyshire wrote the following in the National Review, “The folk losing their homes? Are victims not of ‘predatory lenders,’ but of government-sponsored — in fact government-mandated — political correctness.” The ‘political correctness’ Derbyshire was referring to was the 1977 Community Reinvestment Act or the CRA. The CRA is a piece of federal housing legislation that was passed to encourage large, federally insured banks to increase the provision of financial services to the communities in which they were chartered to exist. Passed alongside a series of comparable housing and community development law, the CRA was part of a broader attempt by congress to attend to the growing concerns surrounding the prevalence of financial inequity. Banks based in metropolitan areas were systematically taking in deposits from surrounding lower-income communities but making their financial services (mortgages, small business loans, etc.) widely inaccessible to those same communities.
In order to redirect lending, the CRA relied on an incentive system. It granted four federal financial agencies and offices the power to review and grade how bank’s lending portfolios (where lending was going and in what volume) aligned with the goal of the CRA. These institutions would then consider that grade when approving future requests for mergers, acquisitions and a series of other federally regulated banking activity. But for the first decade after its passage, the CRA resulted in zero penalties and was widely considered just federal lip service on the concerns of financial inequity. Enter 1989 and the Financial Institutions Reform, Recovery and Enforcement Act (FIRREA). FIRREA remains to be one of the most substantial overhauls of federal financial law and regulation, carrying in it a vast array of implications for the financial sector. Of particular significance to the CRA, FIRREA resulted in the publicization of mortgage lending data from federally insured banks. Following FIRREA, community groups, watchdog agencies and journalists were able to use mortgage data to better hedge the CRA against targeted banks. Community members threatened boycotts and urged lawmakers to hold these banks accountable, a concept known as ‘regulation from below’. With increased pressure from the general public, 1989 marked the first ever penalty given out under the CRA.
Riding on the momentum of FIRREA, in 1995 the CRA was reformed to provide a more structured and rigorous grading process. Included in this process was the lending test, which measured the volume of lending going to CRA-eligible borrowers inside a banks chartered area. Borrowers could be “eligible” for the lending test in two ways; they could make an income that was less than 80% of the metropolitan area’s median income or live in a census tract that had an average income under that cutoff. Following 1995, the impact of the CRA grew more noticeable, with greater penalties being given out and the banking industry being increasingly cognisant of compliance.
How did we get from the story above to the comment made by John Derbyshire? The connection exists in the ongoing debate between academics on the topic, with two mutually-exclusive groups butting heads. The first group is represented by a series of research concluding that the incentive system created by the CRA coerced banks into abandoning safe and sound lending practices in order to meet quota-like lending marks. They back this with data correlating CRA eligibility (as determined by the lending test) with higher delinquency rates prior to 2008. The second group is a denunciation of that claim, arguing instead that the data shows no relationship between the CRA and delinquency, with even some evidence suggesting that CRA eligible lending performed better in the aggregate.
So which side is right? And why such sustained disagreement? Some disagreement arises out of a research challenge with the Act; isolating the impact of CRA eligibility from the exogenous effect of income on loan performance. Using a thorough dataset of a sample of approximately 60,000 loans made by a large lending institution in Southern California between 2002 and 2006, I used a regression discontinuity design to get around this challenge. The discontinuity method allowed me to compare two borrower groups that were just on either side of the 80% cutoff point in the lending test. Given such a small difference in income, any major difference in delinquency rates are likely the effect of the CRA. What I found was that there was no major difference and that the income group just above the 80% cutoff actually performed slightly worse.
Beyond this method, I performed statistical analyses that showed higher probability of default among both low-income borrowers outside of low-income census tracts and high income borrowers inside low-income census tracts–but the probability of default among the latter group is far more substantial. This builds a different narrative around the impact of the CRA. As opposed to low-income borrowers defaulting at higher rates, it seems that the more significant default rates were associated with high-income borrowers buying homes in developing neighborhoods, most likely in search of a profit. That the CRA encouraged lending to affluent housing speculators was certainly not the intention of the act but rather suggests a flaw in the policy design. While claims like John Derbyshire are outlandish given the results of the discontinuity, the broader statistical analyses I performed draw attention to the importance of policy design. Policy makers need to be more specific when defining who they intended to serve and consider how policy may be abused. Otherwise we will continue to see federal incentive systems be blatantly manipulated.
Leave a Reply