On the GSEs (again)

This is an excerpt from my commentary at the Berkeley-UCLA conference on the mortgage meltdown:

Fannie and Freddie’s management teams did unseemly things with respect to accounting, it is very hard to argue that they behaved worse with respect to risk management than investment banks or regulated commercial banks. According to the firms’ monthly volume summaries, their delinquency rates on single-family mortgages remain below 1.6 percent as of November 2008; according to the Mortgage Bankers Association, the overall delinquency rate for that time for the single-family market was 3.93 percent for prime loans, and more than 18 percent for subprime loans. Again, this does not necessarily reflect virtuous management, but rather the fact that the GSEs’ regulator, then the Office of Federal Housing Enterprise Oversight, while often accused of being a weak regulator, actually prevented the firms from engaging in the worst sorts of underwriting behavior. Also quite remarkable is the fact that the delinquency rate for Fannie and Freddie multi-family loans remains at around a basis point.
Among the consequences of this regulation is that the firms lost market share (Figure 1). The pink line in the graph is Fannie and Freddie’s share of mortgage debt outstanding. Note that while it declined sharply from 2002 to 2006, the private label market gained the market share that the GSEs lost. Under the circumstances, it is hard to make the case that the GSEs were the fundamental cause of the mortgage crisis, although many critics would like to think so.
I should disclose that I worked at Freddie Mac for around 15 months. One of the things about the place that was quite striking to me is how seriously its staff took mortgage underwriting. The credit models for the prime book (the business Freddie should have stuck to) were sophisticated, and the arguments about how to do underwriting were at once passionate and scientific. The people responsible for modeling credit risk were, by any standard, well qualified to do so. The chief risk officer of the company at the time discouraged senior management from expanding beyond the prime mortgage business. The GSEs arguably performed their job better than FHA, which has always had limited resources for developing underwriting models.
Senior management of the GSEs was under tremendous pressure to expand their business lines beyond prime mortgages because of the above documented loss of market share. This led both companies, and particularly Freddie Mac, to expand investment into Alt-A mortgages, and it was these mortgages that caused Freddie Mac so much trouble. Had OFHEO been a stronger regulator, or had Freddie Mac been statutorily prohibited from making Alt-A mortgages, the company would still be solvent.
This phenomenon had nothing to do with Freddie Mac’s portfolio per se; even if the GSEs had been in the guarantee business alone, they still would have been under pressure to increase their business. Securitizers in the pure private market put fee generation ahead of due diligence when determining whether to fund mortgages. Keys et al. (2008)showed that loans that were more easily securitized received less lender scrutiny than those that were more likely to be held in portfolio.
If we are going to have GSEs, their re-emergence should rest on four pillars. First, as Jaffee and Quigley suggested in an earlier paper, their cost of funds should reflect the risk they take. This could be accomplished through a tax on new debt issuance. Second, GSEs should be stringently regulated so their products do not depart from high standards of underwriting. Third, to assure one and two happen, GSEs should be forbidden from lobbying. Finally, minimum capital requirements need to be higher, although this (along with the tax on new debt) will raise mortgage costs.
This does not mean the end of GSEs as we know it, but rather a roll-back to where they were in the middles 1990s. Recent events make it clear that the economy cannot rely on the purely private sector to fund 30 year fixed rate mortgages.

The Surprise
If one looks at commentary from the earlier part of this decade on the GSEs, one finds that most of the concerns about them involved market (interest rate) risk, rather than credit risk. Because nominal house prices rose nationally every year between the end of World War II and last year, it was hard to imagine that mortgages would induce a credit crisis. Certainly no empirical model could have predicted the events of the past few years.
This has powerful implications for how we think about capital going forward. Among other things, it suggests that the model-based capital standards proposed in Basel II are not sustainable. It also suggests that there is no substitute for rigorous and admittedly somewhat arbitrary minimum capital ratios for all institutions that lend. This will inevitably mean that the economy will lose out on some positive net present value opportunities. But it also means we will be far less likely to find ourselves in the current situation