Page 495 - untitled

466 Dissenting Statement

35
2010. It is also instructive to compare the Lehman analysts’ estimate that the 2006
vintage of subprime loans would suff er lifetime losses of 19 percent under “stressed”
conditions to other, later, more informed estimates. In early 2010, for example,
Moody’s made a similar estimate for the 2006 vintage and projected a 38 percent
loss rate aft er the 30 percent decline in housing prices had actually occurred. 36
Th e Lehman loss rate projection suggests that the analysts did not have an
accurate estimate of the number of NTMs actually outstanding in 2006. Indeed, I
have not found any studies in the period before the fi nancial crisis in which anyone—
scholar or fi nancial analyst—actually seemed to understand how many NTMs were
in the fi nancial system at the time. It was only aft er the fi nancial crisis, when my AEI
colleague, Edward Pinto, began gathering this information from various unrelated
and disparate sources that the total number of NTMs in the fi nancial markets
became clear. As a result, all loss projections before Pinto’s work were bound to be
faulty.
Much of the Commission majority’s report, which criticizes fi rms, regulators,
corporate executives, risk managers and ratings agency analysts for failure to perceive
the losses that lay ahead, is sheer hindsight. It appears that information about the
composition of the mortgage market was simply not known when the bubble began
to defl ate. Th e Commission never attempted a serious study of what was known
about the composition of the mortgage market in 2007, apparently satisfi ed simply
to blame market participants for failing to understand the risks that lay before them,
without trying to understand what information was actually available.
Th e mortgage market is studied constantly by thousands of analysts,
academics, regulators, traders and investors. How could all these people have
missed something as important as the actual number of NTMs outstanding? Most
market participants appear to have assumed in the bubble years that Fannie and
Freddie continued to adhere to the same conservative underwriting policies they
had previously pursued. Until Fannie and Freddie were required to meet HUD’s AH
goals, they rarely acquired subprime or other low quality mortgages. Indeed, the
very defi nition of a traditional prime mortgage was a loan that Fannie and Freddie
would buy. Lesser loans were rejected, and were ultimately insured by FHA or made
by a relatively small group of subprime originators and investors.
Although anyone who followed HUD’s AH regulations, and thought through
their implications, would have realized that Fannie and Freddie must have been
shift ing their buying activities to low quality loans, few people had incentives to
uncover the new buying pattern. Investors believed that there was no signifi cant
risk in MBS backed by Fannie and Freddie, since they were thought (correctly, as
it turns out) to be implicitly backed by the federal government. In addition, the
GSEs were exempted by law from having to fi le information with the Securities and
Exchange Commission (SEC)--they agreed to fi le voluntarily in 2002--leaving them
free from disclosure obligations and questions from analysts about the quality of
their mortgages.
When Fannie voluntarily began fi ling reports with the SEC in 2003, it disclosed
35 Fannie Mae, 2010 Second Quarter Credit Supplement, http://www.fanniemae.com/ir/pdf/sec/2010/
q2credit_summary.pdf.
36 “Moody’s Projects Losses of Almost Half of Original Balance from 2007 Subprime Mortgage
Securities,” http://seekingalpha.com/article/182556-moodys-projects-losses-of-almost-half-of-original-
balance-from-2007-subprime-mortgage-securities.