David R. Henderson  

Stiglitz and Orszags on Fannie Mae

Paul Collier on Democracy... Getting Your Storks in a Row...
The paper concludes that the probability of default by the GSEs is extremely small. Given this, the expected monetary costs of exposure to GSE insolvency are relatively small -- even given very large levels of outstanding GSE debt and even assuming that the government would bear the cost of all GSE debt in the case of insolvency. For example, if the probability of the stress test conditions occurring is less than one in 500,000, and if the GSEs hold sufficient capital to withstand the stress test, the implication is that the expected cost to the government of providing an explicit government guarantee on $1 trillion in GSE debt is less than $2 million. To be sure, it is difficult to analyze extremely low-probability events, such as the one embodied in the stress test. Even if the analysis is off by an order of magnitude, however, the expected cost to the government is still very modest.

In his recent Podcast interview with Russ Roberts, Charles Calomiris reveals something interesting that has been almost loss down the memory hole. Almost. Calomiris points out that Joe Stiglitz, Jonathan Orszag, and Peter Orszag were hired by Fannie Mae to write a paper in 2002 defending the claim that the odds of Fannie Mae ever getting into financial trouble were extremely low. The quote above is the abstract of their piece. Fannie Mae later pulled it off its website. Stiglitz still lists it on his CV (page 47 at the bottom) but I have been unable to find the whole paper on the web. I can find a link, but the link doesn't work. Can anyone provide the whole paper?

The whole title is, "Implications of the New Fannie Mae and Freddie Mac Risk-based Capital Standard," by Joseph E. Stiglitz, Jonathan M. Orszag and Peter R. Orszag. It's listed in Fannie Mae Papers, Volume 1, Issue 2, March 2002.

I know it's out there because commenter Greg quotes from the paper.

Peter Orszag, of course, is President Obama's Director of the Office of Management and Budget.

H/T to Jeff Hummel and Kishore Jethanandani.

Comments and Sharing

CATEGORIES: Finance , Regulation

COMMENTS (22 to date)
Pierre Lemieux writes:

I just put the original Stiglitz et al. article at http://www.pierrelemieux.org/stiglitzrisk.pdf.

Neal W. writes:

They know their paper was bunk. They just took the payday. Pathetic.

david writes:

A variety of sources.

Pierre Lemieux's upload also appears to be correct.

Chip writes:

You can also find it at the Wayback Machine

ed writes:

BTW, the podcast is really great.

David R. Henderson writes:

Thanks Pierre, David, Chip, and ed,
Isn't decentralized information great?

Frank Howland writes:

The podcast was really great, full of information and provocative ideas. Calomiris has an impressive grasp of financial history and a really intersting take on the current financial crisis.

However, I wish that Charles Calomiris and Russ Roberts had noted that R. Glenn Hubbard, a respected economist who served in the Bush administration as head of the Council of Economic Advisors, also wrote a puff piece for Fannie Mae in the same series ("Evaluating Liquidity Risk Management at Fannie Mae," Vol II, Issue 5, Nov., 2003). Hubbard has not deleted the file, which is on a Columbia web site. The paper does not appear on his C.V., though to be fair, it doesn't show up on earlier versions of his C.V. either.



The Foreword (by a Fannie Mae official) says:

"Professor Hubbard concludes that 'Fannie Mae’s overall risk profile is lower than that of other financial institutions,' and that 'a ‘liquidity crisis’ for Fannie Mae is an extremely remote possibility.'"

Furthermore, Hubbard writes: "Fannie Mae’s assets are more transparent than the loan portfolios of commercial banks and its asset portfolio value and earnings are less volatile."

Robert writes:

What are the latest estimates on how much Fannie and Freddie have cost (omitting their most recent activity when the government told them to buy any document with a signature on it) ??

Marc writes:

To be fair, if you read the study -- not just selective quotes -- Stiglitz and the Orszags are pretty careful about the limitations of their analysis. For example, they note that they are assuming "that the risk-based capital standard is enforced effectively by Office of Federal Housing Enterprise Oversight (OFHEO), which regulates the GSEs" and that "Fannie Mae and Freddie Mac hold enough capital to withstand the stress test imposed under the capital standard." There is lots of evidence that OFHEO was asleep at the switch and that Fannie and Freddie were playing games with the capital they were holding. The methodology seems right given the assumptions -- so I am not sure this is an important study to use for any purpose.

Greg ransom writes:

Now Stiglitz is taking Soros money.

What we nd it a data bank of economists on the take -- Summers took huge Wall Street payouts, so did Krugman from Enron and the NY Times.

How many other economists have one eye on were they might have a big money payout, or a powerful and prestigous political job

Stephen writes:

The only justification that Glaeser presents is summarized in his last sentence: "It could have been much worse." This is how low the standard for public policy has declined, not just for Harvard economists but for most economists, even for some of your colleagues at GMU. So to answer your question I think it is the kind of herd behavior that these same economists have denounced as the source of all bubbles. Most people are an easy prey to the bully behavior of few self-appointed prophets. They may talk a lot when they are having lunch together, but to outsiders they want to show how clever they are with great ideas (most are attempts to reinvent the wheel) and at the same time that they are part of a powerful tribe.

Vacslav writes:

Given their method, their conclusion about the probability of insolvency could even be correct. It is the "monetary cost" conclusion that strikes me as totally absurd: one trillion times 1/500000 is 2 million : correct and absurd at the same time.

Why? Because given the size of exposure the government is not risk-neutral. 1 trillion is a significant fraction of government obligations and that's why one can't say that the "expected monetary cost" of exposure to GSE insolvency is equal to the average cost.

LA-C writes:

Here you go:

a writes:

Marc: "There is lots of evidence that OFHEO was asleep at the switch and that Fannie and Freddie were playing games with the capital they were holding."

--Pray tell, Marc, could it be, perhaps, that Fannie and Freddie were playing with the capital in such a way BECAUSE they had Nobel-Prize winning economists writing hagiographic studies about how awesome their portfolio is?

Marc writes:

At Marc the First:

They may have qualified the analysis in the paper, but the analysis was trotted on at least 2 occasions by Fannie and Freddie at congressional hearings.

Page 24

Page 129

I would bet it is one of the reasons both agencies were able to kill all reform efforts in the years preceding the meltdown.

Stan writes:

So they were obviously wrong. So were a lot of economists in 2002. I'm pretty sure that was before much of the subprime craziness became mainstream. I know you didn't argue anything here, but it is somewhat implied. As for Obama hiring Orszag, that is troubling.

q writes:

there is absolutely no contradiction here.

the GSEs' 2002 portfolio was not the problem. that's a strong portfolio. it's held up in this crisis just fine.

the problem was that the 2005, 2006, and 2007 mortgages have gone bad.

between these dates there were changes in the markets and changes in GSE business practices.

why are you assuming that their 2002 analysis would hold in 2006, 2007, 2008, or 2009?

Peter Lentz writes:


"the GSEs' 2002 portfolio was not the problem. that's a strong portfolio."

Where did you pull that out of?

You imply that substantial subprime investment entered the portfolios after the Stiglitz-Orszag study. Here's the inconvenient (for Stiglitz) truth:

"Interestingly, subprime market growth in the 1990s occurred largely without the participation of Fannie Mae and Freddie Mac. The GSEs started showing interest in this market toward the end of the decade and now [2002] purchase A-minus mortgages as a regular part of their business. National Mortgage News, a trade publication, estimates their combined market share in 2001 grew by 74 percent, representing about 11.5 percent of all subprime loan originations in that year. Some market analysts estimate that GSEs will soon be purchasing as much as one-half of all subprime originations." Fishbein NHI Shelterforce Online, #124, Sept-Oct 2002

And, at that time, the subprime market was not insignificant, according to the St. Louis Fed Review Jan/Feb 2006:

"The growth of subprime lending in the past decade has been quite dramatic. Using data reported by the magazine Inside B&C Lending, Table 3 reports that total subprime or B&C originations (loans) have grown from $65 billion in 1995 to $332 billion in 2003."

To argue that the failure of the Stiglitz-Orszag report to appropriately evaluate the inherent risk within the GSE protfolios is disingenuous, at best. Does Stiglitz, the vocally strong proponent of regulatory efficacy, believe that the bureaucracy would be more astute than he in recognizing trouble?

q writes:

i dont have FNM data in front of me, but the cumulative (ie since inception) loss rates up to Q2 2009 are:

2000: 1.07%
2001: 0.77%
2002: 0.64%

i don't have pre-2000 data but there is no reason to think it is worse than 2000.

assuming a margin of, say, 60bp per year between inception and now, these were very profitable years for FRE.

i think these institutions would have been more or less fine if they hadn't got into subprime lending or bought third party securities.

in fact, it may turn out that they cost the taxpayer very little (80%-90% of their "losses" now are actually loss reserves and their net interest margin is very high at the moment).

q writes:

i meant to begin:

i dont have FNM data in front of me, but the cumulative (ie since inception) loss rates up to Q2 2009 at FRE are:

Peter Lentz writes:


I'm must be dense. I don't see that the cumulative loss rates you cite support your point. The subprimes, whenever they were written, became toxic when real estate values tanked. You can see that as the subprime share of FNM's book increased, so did the loss rates. By 2002 the GSEs' policy had heartily embraced subprimes.

Logically, the earlier subprime mortgages were likely to go underwater to a lesser degree because there was less "false" value in the subject properties in earlier stages of the bubble. They would be expected to have a better loss rate. That's why Stiglitz's failure to perceive the nature of the subprime threat and its potential to grow with the bubble is so important.

The GSEs had gotten into subprime lending in a big way at the time that Prof. Stiglitz examined them. He wrote them a clean bill of health--downplayed the risk in the extreme. What was the purpose of the Stiglitz-Orszag study? The charge was to evaluate the risk attending the GSE portfolios. Those portfolios were in the process of uploading huge volumes of subprime mortgages at an accelerating pace. The portfolios were infected and Stiglitz missed it.

As far as I know, there is no basis for saying that subprime mortgages written before 2003 were better risks than those written after? It's convenient to posit that underwriting standards happened to collapse right after the Stiglitz study. Even if this unlikely coincidence is so, the most he could argue is that he did not see the degree of risk (despite the increasing volume?), but it cannot be denied that he failed to recognize the malevalent nature of subprime mortgages growing in the portfolios. And he would have to admit that he failed to recognize that the risk of loss would necessarily grow exponentially as the bubble caused more subprimes to go under water. Remember, his pronouncement was virtually zero risk.

I see a big contradiction (your word) between Striglitz's failure to perceive the growing malignancy within the GSE portfolios and the faith that he has that a regulatory regime can be expected to prevent future debacles the nature of which we cannot perceive now any better than Striglitz could then.

Peter Lentz

Jim writes:

There is good evidence that underwriting standards did indeed decline precipitously after the Stiglitz study, with the worst years being 2005 through 2007.

Put another way, two things happened historically:
(1) Lenders did more subprime lending
(2) The nature and character of subprime lending changed

Point #1 was evident even in 2002. But it is point #2 that caused the problem. A subprime loan from 2002 was a different animal than a subprime loan from 2006. The first was far less likely to be a no or limited documentation loan, for example. The first was also likely to have a lower loan-to-value.

There is plenty of information out there about the decline in standards post 2002.

One can be deceived by the similarity of the label "subprime" into thinking that a 2002 subprime loan was the same as a 2006 subprime loan. They were not. Lending processes and standards across the industry had become lax to undreamt-of degrees between 2003 and 2006. This is part of the reason why default rates from 2002 loans, even in subprime, are lower. Though it is of course also true that there was less home price inflation baked into the 2002 mortgages than there was in 2006.

All in all, it does not seem obvious to me that one could have allowed for the change in loan standards in 2002. Seeing that subprime loans were growing in 2002 was not enough. More of the 2002 subprime loans would have been much less damaging to a portfolio than more of the 2005 subprime loans ultimately proved to be.

John Hempton, of the blog Bronte Capital, has a great series of posts on the GSEs that is worth reading. Go to http://brontecapital.blogspot.com/ and search for "Fannie Mae." The first in what is now a 9 part series on modelling Fannie and Freddie is here:



Comments for this entry have been closed
Return to top