On page 144, we write, “The financial crisis was not primarily due to Fannie and Freddie.” That sentence is followed by a footnote (77) that refers to an article by Charles Calomiris and Peter Wallison that, in fact, argues that the crisis was due to Fannie and Freddie. (The article title, which we give in the footnote, is “Blame Fannie Mae and Congress for the Credit Mess.”) This is obviously a mistake on our part.
This is what happened. In the final proofs that we were able to review, that sentence used to read: “We do not subscribe to the theory that the financial crisis is primarily due to Fannie and Freddie,” with the same footnote at the end. In that context, it’s clear that we are citing the Calomiris and Wallison article as an example of that theory.
In the final edits, we decided that this “we do not subscribe to the theory” language was unnecessary scaffolding and an unnecessary use of the first person, so we deleted it. The problem is that without the word “theory,” now it superficially looks like we are citing Calomiris and Wallison to support our position. Since the word “theory” vanished from the text, we should have put it in the footnote, as in, “For an example of this theory, see . . .”
The support for our assertion that “the financial crisis was not primarily due to Fannie and Freddie” is obviously not in the footnote; it’s in the next two pages of the main text.
Anyone with the faintest familiarity with our writings would realize that we could not possibly be citing Calomiris or Wallison as support, as either of them would certainly acknowledge. And even without an explicit signal in the footnote, I would expect most thinking readers, seeing the title of the Calomiris-Wallison article, to realize instantly that we are citing Calomiris and Wallison as intellectual opponents, not support. Nevertheless, Gene Epstein referred to this mistake as “outright dishonesty” in an article for Barron’s (behind a paywall). I guess everyone is entitled to his opinion.
Update: A reader kindly sent me the full text of Epstein’s article. Basically he says that we are wrong, because Fannie and Freddie held the majority of subprime loans. Here’s the claim:
“Figures provided by [Edward] Pinto, now a consultant to the mortgage-finance industry, are staggering. By mid-2008, nearly half, or 27 million, of the 55 million first mortgages in the U.S. were high-risk, classified according to his criteria as subprime or ‘Alt-A.’ Of those 27 million, 19.2 million were accounted for by government, whether by virtue of being held or guaranteed by Fannie or Freddie (12 million), held by the Federal Housing Authority (5 million), or held by private institutions (2.2 million) under the requirements of the Community Reinvestment Act and HUD.
“That leaves 7.8 million risky mortgages issued solely by Wall Street.”
Looks pretty compelling, right? Well, only until you realize that Pinto’s definition of subprime is one he made up himself. (See the words “by his criteria” in the quotation above.) In his December 2008 Congressional testimony,* he said there were 25 million high-risk loans. In that analysis, he said there were 17 million subprime loans, of which Fannie/Freddie held 5.7 million, or 34 percent. But drop down to PDF page 56 and you see that this 5.7 million is made up of:
- 0.8 million are “subprime private label mortgage backed securities”
- 4.9 million are “‘prime’ loans < 660 FICO”
In other words, Fannie and Freddie held zero subprime whole loans by the conventional definition (loans denominated as subprime by the issuer, which is the definition used by LoanPerformance). They held 4.9 million loans that Pinto, in his wisdom, decided to call subprime, and subprime-backed MBS that are equivalent to 0.8 million loans.
Now, Pinto knows more about mortgages than I do. Maybe loans with a FICO below 660 should have been counted as subprime all along, and maybe they should have gone into the LoanPerformance subprime database. But they weren’t. The fact remains that Fannie and Freddie had standards, and there were some loans they would not buy. Maybe those standards were too low; I would probably agree with Pinto on that one. But that’s a different issue. Pinto’s “data” don’t contradict the fact — pointed out by many people more knowledgeable than I — that Fannie and Freddie simply could not legally buy or guarantee the worst of the subprime loans.
(As for the 0.8 million loans in subprime MBS that Fannie/Freddie bought — that was a real problem, and one we point out on page 146 of the book.)
* Epstein’s numbers are slightly different from those in the December 2008 testimony, so I imagine Pinto updated his numbers. But they are close enough that the underlying principles are almost certainly the same.
Filed under: Corrections