August 11, 2009
Paying for design flaws
Updates on what this is going to cost you and me.
Let's start with Fannie Mae, the government-sponsored enterprise that was allowed to function as a quasi-private company from 1968 to 2008 and is currently under conservatorship of the Federal Housing Finance Agency. Prior to conservatorship, Fannie earned a profit two ways. First, it used borrowed funds to purchase mortgages that it held directly. Because investors perceived the GSE's debt to be implicitly backed by the federal government, Fannie's borrowing costs were very low, and it had an incentive to engage in arbitrage on a huge scale, borrowing cheap and buying as many mortgages as regulators would allow. As of June 2009 these assets came to $793 billion. Second, Fannie would bundle mortgages into securities on which it provided a guarantee of timely payment of principal and interest, in exchange for which it received a modest fee, and assumed a staggering off-balance-sheet liability. As of June 2009, Fannie's guarantees involved an additional notional liability of $2.4 trillion. Fannie's equity was never remotely sufficient to make such guarantees credible, and this game, too, is best interpreted as arbitraging an implicit government obligation to pick up the tab should things turn sour.
Total nonperforming loans in our guaranty book of business were $171.0 billion on June 30, 2009, compared with $144.9 billion on March 31, 2009, and $119.2 billion on December 31, 2008.
This doesn't mean that taxpayers are now out $171 billion, because the salvage value on these properties is not zero. On the other hand, more defaults could be ahead.
Fannie's GSE brother, Freddie Mac, had over $2 trillion in mortgages held outright or guaranteed, last time I checked. Freddie reported $15 billion in non-performing assets that it holds outright and an additional $62 billion in non-performing assets on which it has issued guarantees (Table 2 in their second-quarter 10Q). But on top of this is the imminent bankruptcy anticipated from mortgage originator Taylor, Bean & Whitaker, following last week's allegations of "irregular transactions that raised concerns of fraud". Freddie Mac filed this report (hat tip: CR):
TBW accounted for approximately 5.2% and 2.7% of our single-family mortgage purchase volume activity for full-year 2008 and the six months ended June 30, 2009, respectively. We are in the process of determining our total exposure to TBW in the event it cannot perform its contractual obligations to us. The amount of our losses in such event could be significant.
And the Federal Housing Administration is stepping in where Fannie and Freddie leave off. A recent report from the Office of Inspector General noted that the FHA's current single-family insured exposure totals over $560 billion and invites preventable fraud. TBW was the third biggest originator of FHA loans, with the overall delinquency rate on FHA-guaranteed loans now up to 7.42%.
Then there's the $1 trillion in mortgage exposure that Ginnie Mae is about to take on.
Posted by James Hamilton at August 11, 2009 08:27 PMdigg this | reddit
Please add Ginnie Mae to the list of government crimes. See
Posted by: E. Barandiaran at August 11, 2009 09:25 PM
"Among the FHA, Ginnie, Fannie and Freddie, nearly nine of every 10 new mortgages in America now carry a federal taxpayer guarantee."
That's in the article you link to under Ginnie Mae, and I thought worth highlighting.
Posted by: Tom at August 11, 2009 09:38 PM
Idle speculation: Where would mortgage rates be if there was no implicit government guarantee and/or securitization mechanism provided by Fannie/Freddie/etc ?
Posted by: Mike at August 11, 2009 10:01 PM
You mention the GSE's low cost of capital. This is due to the lack of credit risk associated with government backing. There was a huge crowding out effect whereby commercial lenders could not compete with the GSEs. Consequently, lenders began to seek out non-traditional revenue sources (subprime, Alt-A, etc). Once they were given a taste of the profitability in non-traditional lending it began to explode in scale. Make no mistake, the growth of the GSEs and the Federal Government's burning desire to grow home ownership beyond it's natural limit were significant distortions in the marketplace and major contributors to the housing bubble.
Posted by: T-Dub at August 12, 2009 04:37 AM
Trillions are hard for me to relate to; so many zeros. But I think a trillion works out to about $10,000 per household.
Posted by: Rich at August 12, 2009 08:08 AM
Hey Mike, that's the point exactly. Without the governments manipulation of the mortgage market, we may never have had a bubble of the same size. Higher mortgage rates would have made houses less affordable. I know that its un-American to root for less affordable housing, but at some point, we as a country have to realize that just because it made sense to increase home ownership from 35% to 60%, it may not make sense to keep going. Housing is a terrible investment. It ties up capital. It has a low real return. It requires large amounts of leverage. It has no productivity. So yeah, without FNM/FRE/FHA/GNMA, mortgage rates would have been, and would still be, higher. But I'm in favor of that!
Posted by: joe at August 12, 2009 08:26 AM
@T-Dub: That's a whole new variant of the traditional right-wing "poor people caused the meltdown" logic. The GSEs *forced* the lenders to start making terrible (but highly profitable) loans. Wow, forced to make higher profits!
By that logic, drug dealers could claim that, "due to my inability to compete in the job market, I was forced to start selling crack. It's not my fault!"
At this point, the GSEs and FHA are the primary reason anybody's been able to get ANY financing over much of the last 6-12 months, as the lenders refused to make any loans they couldn't sell to the government.
Posted by: Matt at August 12, 2009 09:11 AM
When I was in business school, back in the mid-90's, our econ professor railed against the GSEs. He had been under-assistant deputy secretary in Reagan's Treasury or something.
He predicted this thing was going to blow up. It was his case study of moral hazard.
Posted by: Joseph Somsel at August 12, 2009 09:24 AM
GSEs were (well, caused anyhow) significant distortions in the market? Yes, but let's be clear about what distortions they caused. Matt has a reasonable point. When private lenders had a larger share of the mortgage market, that did not prevent other lenders from learning of the joys of subprime lending. But mostly they didn't. GSEs have been around lots longer than subprime lending. We don't have good correlation, much less evidence of causation. "Make no mistake" begins to sound like salesmanship, hand-waving to distract from the weakness of the argument.
A notable change that correlates somewhat better with the emergency of a large subprime mortgage market is the underfunding, disparaging and institutional undermining of financial regulation. If a mortgage market dominated at the high-quality end by large, federally hocus-pocused institutions had continued under more stringent regulation, we may never have seen a Countrywide or an IndyMac.
No, just because the prior arguments trying to blame anti-discriminatory and low-cost housing efforts failed doesn't mean we need to fall for this one.
What distortions can reasonably be put down to GSEs? Artificially low borrowing costs and "too big to fail" thinking encouraged systemic risk. Systemic risk strongly implies that "too big to fail" will prove self-fulfilling. That is not isolated to the GSEs, by the way. It led to distortions in the almost-fully-private financial sector, too.
Posted by: kharris at August 12, 2009 09:55 AM
Mike, Joe and James are on target. But understating. The total FF losses put to taxpayers may be about $600 billion. Taking their current nonperforming just gets a fraction of the eventual loss from further writedowns and asset undervaluation. They own or guarantee over 5 trillion that will never aver be worth what they are carrying it at.
FHA and Ginnie are the worst ofenders for new accruals, doing nearly half of current new loans recently. You can still get that 3% down. If that is a problem the Nehemia foundation and one other will spot you the down payment as a charity. Billions are going to this good cause, or maybe its money well spent by the housing community to get Uncle Sugar to cover the other 97%?
Posted by: AWH at August 12, 2009 10:12 AM
"At this point, the GSEs and FHA are the primary reason anybody's been able to get ANY financing over much of the last 6-12 months, as the lenders refused to make any loans they couldn't sell to the government."
EXACTLY! I realize the long-term distortionary effects of the GSE, but the current issue is the implosion in the residential housing market. Talk to anyone in this market and they will tell you that agencies ARE the market. If these entites fail or dramatically slow origination, the result could be catastrophic. There would functionally be NO lending. What is left of the residential market and the vast majority on many peoples savings would be gone, overnight.
The current short term policies should be aimed at slowing the bursting of the bubble so homeowners can repair savings. Long term, the government guarantee must go.
Posted by: JD at August 12, 2009 11:01 AM
Let's see, we have had the most massive debt problem in our nation's history and what is the government solution. Well more debt of course. After all we have to stimulate more consumption to pull us out of this mess, don't we?
Just as an aside if you add all that has been committed for stimulus and the exposure of Fannie, Freddie, Ginnie, and FHA you are almost exactly at 2008's GDP. Wow, what a country!!
Professor, your analysis is great, but more frightening than most realize.
Posted by: DickF at August 12, 2009 01:54 PM
I don't see how all this could have happened:
"the risk to the government from a potential default on GSE [Fannie Mae] debt is effectively zero ... the expected cost to the government of providing an explicit government guarantee on $1 trillion in GSE debt is just $2 million."
-- Joseph Stiglitz, 2002 (via Falkenblog.)
If even Nobel winners can't see design flaws any better than this, who can?
Going forward, who will give us our better designs and regulations for future?
(Yes, we all know that Joe and everyone else can see design flaws looking backward. Joe himself will be the first to tell you, and everyone else will be telling you right after.)
Posted by: Jim Glass at August 12, 2009 02:18 PM
Jim Glass: You neglected to mention the assumptions on which Stiglitz made those stress tests, based on year 2000 data before the housing bubble.
"... assume that the risk-based capital standard is enforced effectively by Office of Federal Housing Enterprise Oversight (OFHEO), which regulates the GSEs" and "assume that Fannie Mae and Freddie Mac hold enough capital to withstand the stress test imposed under the capital standard."
It turns out that the managers of Fannie and Freddie violated both of those assumptions. Between 2000 and 2007 they increased their leverage from about 35 to over 60.
Stiglitz can't be faulted if the corporate managers and shareholders, in their greedy quest for higher profits, violated the capital requirements that he stipulated for the stress test.
Posted by: Joseph at August 12, 2009 08:24 PM
I agree with DickF. I think that the govt has demonstrated that they are going to continue to throw money at these failed institutions to try to clean up their mess. This will continue to have negative consequences for our currency, as their main remedy for all of these problems is money printing and stimulus, which is essentially debt. And unfortunately one of the only ways for most people to protect and potentially prosper from this is by investing in those assets that should rise from all the money printing, such as gold and other commodities. Here is a good discussion on a gold mining company, Premier Gold that should benefit from a rise in the gold price. It is unusually well-diversified for an exploration company and does not contain the same risks as many other junior miners that possess a single project.
Posted by: jturner at August 13, 2009 06:25 AM
hempton has some good posts on the GSE's accounting and how he has modelled their portfolios
Posted by: oops at August 13, 2009 01:41 PM
I must quarrel with the idea that housing is "not productive." Housing provides a flow of services (shelter) whose absence would make people very unhappy (at least outside of San Diego).
It may be reasonable to angue that the US has too much housing, and so that returns to housing are too low. But that is much different from being "non-productive."
Posted by: Richard Green at August 13, 2009 02:06 PM
JD, what would you propose as an alternative? Since we've been spending ourselves silly during expansions, do you recommend another pro-cyclical reaction to a severe recession?
Posted by: Mickey at August 13, 2009 05:27 PM
"Jim Glass: You neglected to mention the assumptions on which Stiglitz made those stress tests... that the risk-based capital standard is enforced effectively by Office of Federal Housing Enterprise Oversight (OFHEO)..." etc.,
*Assumptions* in stating the risk of defualt was zero, and the cost of guaranteeing $1 trillion was $2 million?
The GSE's equity-to-debt-and-guarantees ratio (1.5-to-98.5 at the time of our host's earlier analysis) was dynamite that they were built over, just as if one stored dynamite in one's basement. And it had been described as such by many people for years.
Of course, one can calculate that the risk that the dynamite one is standing upon will blow up is exactly zero if one *assumes* regulators will always do a top job, and managers will always do a top job, and no politician will ever start playing games with operations, and there will never be an unexpected earthquake or other event, and nobody will ever drop a careless match -- and basically assume that all the things that could cause the dynamite to go off will never happen.
But even so, building one's house over dynamite still generally would be considered a "design flaw", I think.
The point of risk analysis is calculating downside risk NOT assuming no bad thing will ever happen. Assuming the bad things will never happen makes the whole exercise pointless. As is demonstrated by his statements versus subsequent reality.
"Stiglitz can't be faulted if [regulators and management screwed up]"
And Stiglitz had strong reason to have such great faith in regulators and management as to calculate zero risk, right?
After all, he'd just witnessed the wonderful regulatory successes of the Clinton/Rubin/Reno bubble years - the Dot.com bubble, Enron, Worldcom, the stock option swindles, etc. ... and there were the banking management-regulatory successes of the S&L industry ... and he greatly admired the astute regulatory & policy advice administered out of Washington world-wide after the 1997 financial crisis ... and I think he had a few things to say about regulated special intersts getting in bed with politicians in a few cases.
So of course I can see how reasonable it was for him to say: assuming nothing ever goes wrong regulating- or management- or seriously otherway-wise, the risk that the dynamite will go off is zero.
And if he takes a look at the smoking crater today he'll probably think: I was right! Something went wrong to cause this.
Heck, the risk that Microsoft will go broke in two years isn't zero. Ask the people who used to work at Enron.
Which gets to my larger point, which was not just to take a shot at Joe, no matter how much he deserves 'em.
One of my favorite stats is that 50% of NFL football games are determined by luck, chance.
No fan ever believes such a thing, no big-name sports commentator ever says it. But it is so, and it is pretty obvious -- e.g. it's why sports betting books managing big money using supercomputers to set odds can't predict more than 75% winners (not against the odds, just W-L). The favorite wins 50% of the time, the other 50% is a coin flip, so the favorite wins 75%. And that's the highest rate anyone can predict.
Point is: with NFL football games you have 100% knowledge of all the data factors and rules in advance -- and still fully 50% of outcomes are chance.
The world economy is rather more complex than NFL football, with rather more unknowns in all directions and dimensions. If chance events determine 50% of outcomes of all football games simple as they are, is it possible they might have played a major role in the last three years and will in the future?
If so , it makes a big difference in regulatory design -- if you want to avoid future "design flaws".
If the world is determinate and you know everything about it, no chance in play, you can set up "forcing" determinative regulations -- and might even calculate from them that risk of future events like GSE defaults are zero.
But if the world is full of chance events, developing processes you can't fully understand until after the fact (if then), and many interactions too complex to calculate -- you'll want a whole different regulatory system, with a lot more flexibility and diversity, and you'll never say the risk of anything is "zero".
And if the world is of the second type, but you put in a regulatory scheme of the first type, you are setting up to create serious future problems via serious design flaws.
In which regard, just as no NFL football fan or professional pundit ever says, "fully 50% of game outcomes are chance, random", but instead is much more likely to say, "I have an explanation for everthing, and here it is..." where are the economic pundits and politicians saying: "You know, there really are significant forces at play that I personally, and we all, just don't understand and certainly can't predict, so in our regulatory reforms..."?
Is anyone saying such a thing? Stiglitz? Krugman? Anybody?
If so, I don't see many, and the "I know everything..." ones seem to have them outnumbered 99-1.
Posted by: Jim Glass at August 15, 2009 11:01 PM