I have two reasons to bet on Raphael's view:
(1) At the time the dumbest mortgage business was being done, FHA was out of the picture. While FHA's market share is typically in the neighborhood of 12-15 percent, during the period 2003-2007, its market share ranged from 3.77 to 9.66 percent. FHA did not lower its underwriting standards to that of the shadow banking sector (a sector that was not subject to the Community Reinvestment Act, by the way) in order to keep market share--the government insurance program was far more disciplined than the private sector.
FHA's market share increased dramatically in 2009 and 2010, in large part because the private sector abandoned the low downpayment market. In 2010 in particular, FHA gained market share despite raising its prices and tightening its underwriting. FHA was also ramping up its market share after house prices collapsed. While house prices have not been robustly rising since late 2008, they have not been falling precipitously either. One could argue that the private sector has been backward looking, while the public sector has been more forward looking.
(2) The second reason I have is more speculative, and is something that I am currently in the middle of researching, but I want to put it out there as a hypothesis (and a hunch). I suspect that there is such a thing as "burn-out" in default--if a household goes through a difficult time without defaulting, it becomes decreasingly likely to default. Part of the reason for this is amortization, but that is a small reason. More important, people who refuse to default even when their measured characteristics suggest that they should have revealed that they are "different," and in a manner that is unobservable.
Now again, in the interest of full disclosure, I should note that I did not forecast the size of GSE losses, so maybe I shouldn't be taken that seriously. But I think my first argument will stand up, and as I do more research, I will know more about the second.