The following excerpt from “Guaranteed to Fail” is reprinted with permission from Princeton University Press and from the authors, Viral Acharya, Matthew Richardson, Stijn van Nieuwerburgh and Lawrence J. White.
Selection 1: The Race to the Bottom
While there is little doubt that the housing goals played an important role in shifting Fannie Mae and Freddie Mac’s profile to riskier mortgage loans, it remains an interesting question whether Fannie Mae and Freddie Mac deliberately chose to increase the riskiness of the loans that they bought 2004 onward or whether they were forced to do so by the U.S. Congress, which wanted to promote home ownership.
While the public/private nature of the GSEs leads to a moral hazard problem even in normal times, the question is whether moral hazard was exacerbated by the astronomical growth of the subprime market segment.
As pointed out earlier, the GSEs saw consecutive increases in their low- and moderate-income, special affordability, and underserved areas goals in each of 1996, 1997, 2001, 2005, 2006, 2007, and 2008. However, the largest increases took place in 1996 and in 2001, outside of the rapid growth of the 2003 period and onward.
Moreover, the target increases in 2005, 2006, and 2007 were more modest, yet that is when most of the increase in riskiness took place. Finally, Fannie and Freddie missed one or more of their mission targets on several occasions, without severe sanctions by the regulator, suggesting that adherence was largely voluntary.
Former FHFA director James Lockhart testified that both Fannie and Freddie “had serious deficiencies in systems, risk management, and internal controls.” Furthermore, “there was no mission related reason why the Enterprises needed portfolios that totaled $1.5 trillion.” He chalked it up to “the Enterprises’ drive for market share and short-term profitability.” In fact, in testimony to the Financial Crisis Inquiry Commission on April 9, 2010, former Fannie Mae CEO Daniel Mudd admitted as much:
“In 2003, Fannie Mae’s estimated market share of new single-family mortgage-related securities was 45%. By 2006, it had fallen to 23.7%. It became clear that the movement towards nontraditional products was not a fad, but a growing and permanent change in the mortgage marketplace, which the GSEs (as companies specialized in and limited to, the mortgage market) could not ignore.”
Similar language can be found in Fannie Mae’s own strategic plan document, “Fannie Mae Strategic Plan, 2007-2011, Deepen Segments – Develop Breadth,” in which the company outlined its 2007 onwards strategy:
“Our business model – investing in and guaranteeing home mortgages – is a good one, so good that others want to ‘take us out’… Under our new strategy, we will take and manage more credit risk, moving deeper into the credit pool to serve a large and growing part of the mortgage market.”
The data tell the story. From 1992 to 2002, Fannie Mae and Freddie Mac were clearly major participants in high risk mortgage lending. Nevertheless, the period 2003-2007 represented a significant shift.
For comparison purposes, we restrict ourselves to the size of mortgages at or below the conforming limit level. For example, from 2001 to 2003, for mortgage loans with LTVs greater than 80% and/or FICO scores less than 660, Fannie Mae and Freddie Mac represented respectively 86%, 80% and 74% of this high risk activity.
From 2004-2005, this changed as both the dollar volume and share of high risk lending of conforming size loans moved towards the private sector, with $168 billion (and a 26% share) in 2003 to $283 billion (and a 52% share) in 2004 and $330 billion (and 58% share) in 2005.
Consistent with the race to the bottom, Fannie and Freddie responded by increasing their high-risk mortgage participation by recovering a majority share of 51% in 2006 and an almost complete share of the market in 2007 at 87%. Equally important, as a percentage of its own business, Fannie and Freddie’s risky mortgage share increased from 25% in 2003 to 36% in 2007.
Even more telling, if the above analysis is restricted to the very highest risk mortgage loans, i.e., those with LTVs [loan to value ratios] >90% and FICO<620, [the data] shows an almost identical “race-to-the-bottom” pattern in Fannie and Freddie’s share during the 2003- 2007 period, culminating in a doubling of these particularly risky mortgages from $10.4 billion in 2006 to $20.3 billion in 2007.
The SEC 10-K credit-risk filings of Fannie Mae are also revealing of the deterioration in mortgage loans that were purchased by the GSEs during the 2004-2007 period, either for their own portfolios or to be sold off to others. For example, 17% of the 2006 and 25% of the 2007 mortgages that Fannie bought had a loan-to-value ratio in excess of 80%.
The fraction of loans with CLTVs greater than 95% went from 5% in 2004 to 15% in 2007. The borrowers also had lower credit scores: 17.4% of 2006 loans and 18% of 2007 loans had FICO scores below 660. A relatively large share was ARMs (16.6% in 2006 and 9% in 2007) or interest-only loans (15.2% in both years). The Alt-A fraction of purchases was 21.8% in 2006 and 16.7% in 2007, up from 12% in 2004.
Finally, non-full documentation loans went from 18% in 2004 to 31% in 2007. If anything, Freddie Mac’s credit-risk profile was worse than Fannie’s. In 2004, 11% of the loans that Freddie bought had CLTVs above 100%, which increased to 37% by 2007.
Interest-only loans grew from 2% to 20%, and low-FICO-score loans from 4% to 7%. As a final indication of its all-in approach to mortgage lending in 2007, note that mortgage loans with both FICO<620 and LTV>90% reached $20.3 billion, essentially double that of any other year.
Clearly, the quality of GSE loans deteriorated substantially from 2003 to 2007. It seems that the GSEs were able to stretch the concept of a prime, conforming loan much beyond what its regulator had intended.
© 2011 Princeton University Press.
