Arnold Kling  

The Duelling Guarantee

PRINT
Oh, No!... More Fantasy Testimony...

The myth is that mortgage securitization reflects the genius of Wall Street. The reality is that it reflects the stupidity of the way that regulatory capital requirements are calculated.

The market for low-risk loans is dominated by Freddie and Fannie because the GSE's are required to hold less capital than banks for taking the same risk. The market for high-risk loans is dominated by private securitization because capital requirements encourage banks to buy securitized loans that they could never originate on their own. Freddie and Fannie were relatively minor players in the high-risk market, but what little they did undermined their safety and soundness, against the warnings of mid-level staff.

The economic fundamentals of the mortgage market are best understood using Robert Van Order's theory of the duelling guarantee.

I first introduced Bob Van Order when I wrote my chapter describing my early experiences at Freddie Mac. Van Order, along with Chet Foster, came up with the approach for calculating mortgage credit risk using an option pricing framework.

Van Order also had a theory to explain how mortgages ended up held by different institutions. He pointed out that depository institutions (banks as well as savings and loans) had their liabilities guaranteed by the government. So the fact that Freddie and Fannie enjoyed an implicit guarantee did not give the GSE's any inherent advantage.

Is it more economical for a depository institution to hold its own loans or to hold securities created by Fannie and Freddie? In the absence of regulatory incentives, the pure economics of the choice are as follows:

1. Fannie and Freddie have greater geographic diversification, which lowers their risk and their cost.

2. Fannie and Freddie have specialized expertise (including people like Van Order) to fine-tune credit risk modeling more accurately.

3. On the other hand, depository institutions who hold their own loans face fewer agency costs. Their loans are being originated by their own employees, not somebody else's. They can give their employees training and incentives to screen loan files carefully. When you buy loans from someone else, their employees are trained to "produce" loans. For the mortgage broker, every loan that gets passed on to a buyer represents production, whether the loan is underwritten properly, improperly, or fraudulently.

4. The loans held by depository institutions include fewer transaction costs. You do not pay Wall Street to package the mortgage into a security and to trade it. If you need to get a pension fund to proviide money to fund the loan, you can issue debt to the pension fund (of course, the pension fund won't enjoy the same guarantee as a depositor who stays under the $100,000 FDIC limit that was in place until the emergency legislation that was passed just recently).

Overall,I believe that factor (3) is most important. That is, Freddie and Fannie would not exist if all they had going for them was the same guarantee of liabilities that is enjoyed by depository institutions.

In practice, the main difference advantage enjoyed Freddie and Fannie concerns capital requirements. Their capital requirements are calculated using a methodology that differs from that of depository institutions.

Banks and savings and loans have to hold a minimum ratio of capital to risk-weighted assets. Even a low-risk mortgage loan, where the borrower puts down between 20 and 40 percent of the purchase price, has a 0.35 weight, where 1.0 would be the riskiest.

Freddie and Fannie do not use a risk-weighted capital ratio. For credit risk, their capital requirement is determined by a stress test. How much capital would they need to survive a specific downturn in house prices? This varies by the type of mortgage.

For low-risk mortgages (down payments of 20 percent or more), the capital required by the stress test is less than the capital required for banks in their risk-based capital formula. Because of these different capital requirements, Freddie and Fannie should dominate the market for low-risk loans, as indeed they do.

For high-risk mortgages, as I pointed out, the FDIC gives the advantage to private securitizers. That is, for a high-risk loan that is laundered through the mortgage securitization process, the FDIC requires less capital for a bank than even a low-risk loan originated and held the old-fashioned way.

I believe that the stress test makes high-risk loans prohibitively costly for Freddie Mac and Fannie Mae. The question is not why Freddie Mac and Fannie Mae were less active than private securitizers in the subprime market. The question is why Freddie and Fannie were involved at all. The answer, to repeat, is that senior management ignored safety and soundness considerations, in part to satisfy political pressure, in part to satisfy market share goals, and in part because they did not believe what their staffs were telling them about safety and soundness risks. On the latter point, history has shown that the staffs were correct.

In my opinion (and this may reflect my experience at Freddie Mac), the most appropriate way to calculate capital requirements is the stress test. If the stress test methodology were used for depository institutions, then in my judgment they would have no disadvantage in holding low-risk loans, and they would drive Freddie and Fannie out of the market. If the stress test were applied to high-risk loans and to private securities (rather than the moronic rating-agency ratings), then the market for high-risk loans (a) would be limited to loans originated and held by depository institutions and (b) would be much, much smaller.

In retrospect, if capital requirements had been rational, we would never have had the nontransparency of mortgage securities. Moreover, we would have had fewer mortgages with low down payments, which would have meant a smaller rise in house prices and a much less dramatic crash.

Why did the irrational pattern of capital requirements survive? Perhaps the folks at the FDIC were just plain stupid. I doubt that is the case. My guess is that there were plenty of career civil servants who were not happy with what the capital requirements were doing. But at higher levels, and with Congress, what mattered was the increase in home ownership and the campaign contributions from Wall Street, for whom the transaction costs of securitization represented profits.

That is the reality. It is not widely understood, and most people would rather believe a different narrative. The Left wants to blame deregulation. The Right wants to blame lending to minorities. But this blog is giving you the true narrative. I have not seen anyone refute it.


Comments and Sharing





COMMENTS (16 to date)
Gamut writes:

Sure, makes sense, but you do essentially end up making the point of the Right when you say "with Congress, what mattered was the increase in home ownership". With the will of Congress to provide mechanisms for housing the poor, policy at the FDIC was probably largely a mediator, rather than a cause. On the other hand, the notion of "deregulation" doesn't fit anywhere within this massively complex network of causes and effects. In this way, your insights thus far (and I have been reading them daily), really very much fall into the "Right" camp, but are much more nuanced than the mere ascribing of effect to a particular intent; You do a spectacular job of explaining the connection between the two. Which apparently the Right can produce, while the Left leaves to the imagination.

Gary Rogers writes:

Not only have I not found anyone to refute this, I have not found anyone else able to provide the details necessary to understand the problem. I guess I fall into the geek category when I think that our most important task right now is to understand the problem rather than pointing fingers. We need to understand so we can fix things without making them worse. I personally thank you for sharing your knowlege and experience and hope that the rest of the economic community follows your lead.

The alternative is to handle our current crisis the way we handled the great depression.

rhhardin writes:

Stress test, I take it, means taking into account that losses are not independent.

That ought to happen at every level of risk assessment. There's a profit opportunity!

Its failure shows up as a fallacy of composition whenever there's a financial panic.

That fail-safe systems fail by failing to fail safe is once again observed.

Tushar writes:

Another brilliant exposition by Arnold.
Thanks so much. This crisis has really shown who deserves to be shunned from respectability in the Economics profession (macroeconomists) and who really needs to be listened to (you).

The Snob writes:

Fantastic essay. However, it raises a question. If I recall correctly, this "duelling guarantee" market structure existed for at least a couple of decades before housing began to diverge from historical valuations in the mid- to late-90s. Was this divergence an emergent behavior that was going to blow up eventually, or was there an additional accelerant that caused it to start going into overspeed mode?

Arnold Kling writes:

Snob,
Fannie and Freddie did not grow instantly to 50 percent of the market, but the move took place quite quickly. See slide 3 of James Hamilton's presentation:
http://www.econbrowser.com/archives/2008/10/Hamilton_Rady.ppt

I think that private securitization took off in part because of credit scoring, which made it appear as if some of the agency costs of third-party lending had gone down (Don't believe our underwriter? Check the credit scores!). Also, you needed the positive feedback mechanism of lower down payments leading to higher prices leading to lower down payments to work its way through the system. It started gradually, then accelerated in the last few years.

Mogden,
I think other countries did experience property bubbles, and I wish I knew more about them. Were they as dramatic? If so, were there similar institutional changes taking place?

As to the financial crisis in other countries, I think they have less robust banking systems to begin with (highly concentrated banks, less experienced regulators). Their exposure may have been to U.S. shocks more than to problems in their domestic loan markets. but I don't know.

Les writes:

Thank you Arnold for by far the most complete and the most informed diagnosis I have seen anywhere.

I do suspect that some politicians were motivated by more than concern for more home-ownership, because Fannie and Freddie provided favors and money to them, and that top Fannie and Freddie management enjoyed large political favors in return.

So I don't see so much irrationality as I see mutual back-scratching and lack of an ethical compass.

TC writes:

Arnold Kling,

Thanks for your work on uncovering this regulatory arbitrage issue with FDIC.

What if we reformed our banking/financial sector in the following way:

We give banks two options. Option A is to be an FDIC insured bank. Let's call them LRBs (Low Risk Banks) Option B is to not be an FDIC insured bank. Let's call them HRBs (High Risk Banks).

LRBs would be required to invest all of their customer deposits into short term federal treasury bonds or cash. How's that for regulatory simplicity?

HRBs would be required to notify their depositors, on a semi-annual basis, that they are placing their money in a High Risk Bank and that their deposits are not insured by the FDIC. HRBs would be allowed to invest their customer deposits in any legal investment, including junk bonds, sub prime mortgages, no document loans.

Your thoughts?

TC writes:

I should have included in my proposed "reform" of the banking/financial sector some discussion of the rating agencies.

In my formulation High Risk Banks would have the option of purchasing AAA bonds or AA bonds or junk bonds. Still, they would be dependent, to some extent on the rating agencies.

Why did the rating agencies perform so badly in this crisis? How can we reform them so that they do better next time?

The Snob writes:

Arnold,

Thanks for the cite, that clarifies my question about the GSEs.

The next question I'm trying to get my head around is how the credit default swaps play into this. Those seemingly exploded in number relatively late in the game, and have been presented as growing out of all proportion to reality. The implication being that they act to multiply the effect of defaults well beyond what they would otherwise be.

As for bubbles in other countries, I hear it's quite bad in Spain.

RobbL writes:

Arnold,
All right, maybe we should have done something different with Fannie and Freddie, but unless I am missing something, we still would have had a problem. Most of the really toxic stuff that we are dealing with was not generated by them. It was, as you put it, the "geniuses" on Wall Street. Even if F and F were perfectly liquid and solvent we would still be hurtling off the cliff. Despite the crazy right, F and F are a small part of the problem.

Arnold Kling writes:

RobbL,
If Freddie and Fannie had been better managed, what would the world have looked like? One possibility is that the private securitization market would have been even more dominant. On the other hand, without backing from Freddie and Fannie, private securitization might have gone less far.

But regardless of what the private securitizers would have done, if Freddie and Fannie had kept their noses clean, today they would be sound and could be part of the solution instead of part of the problem.

Even if we blame the private securitizers more than Freddie and Fannie, we are still left with the fact that only the insanity of the capital regulations allowed private securitization to function. With a rational system of risk weights based on loss probabilities, the securitizers would never have been able to place their paper in bank portfolios, and I doubt that others would have stepped in to take it.

Jayson Virissimo writes:

Arnold, when is the book coming?

Richard Boltuck writes:

Arnold . . . because of your experience at Freddie, you know a lot about the history and incentives that led to the expansion of securitized high- and lower-risk mortgages. Could you also discuss the distribution of these holdings within the commercial banking system? My impression is that they are concentrated among selected large banks, now among the money-center banks, and much less so among smaller community banks, but this may be incorrect. (WaPo ran a story over the weekend suggesting that community banks have been strengthened by the diversion of deposits in the wake of public doubts about New York banks). Many thanks for this excellent account.

parviziyi writes:

In Spain, house prices have gone down sharply in recent times. They had been driven up by low interest rates, in part. Spain over the past decade underwent an economic boom especially a real estate boom, which, if Spain had its own central bank, would've pushed up the local interest rates in Spain for sure. But Spain's base rate is set by the ECB, and the ECB rate is primarily dictated by what's happening to the economies of Germany and France, not Spain. Now, the thing is, the Spanish banks today are in good shape despite the loan losses they are experiencing in the wake of the sharply falling house prices. They made plently of reserve provisions during the boom.

Each EU sovereign jurisdiction largely regulates its own banks, even though some basic EU-wide regulations are also in place. The Spanish bank regulatory regime is something I think the US can learn from, especially since Spain has the developed world's highest rate of homeownership (in 2002 85% of Spanish households owned their house, though the figure slipped a little since then because of the affordability problem connected to the rising house prices). One of Spanish banking's better features, in my opinion, is that all mortgage loans are variable rate loans. It means the banks don't have systemic interest-rate risk. This risk is instead borne by the customers and it is counter-cyclical for the economy (or would be if interest rates were dictated by Spanish cycles instead of German ones). When a bank is faced with interest-rate risk, it incentivizes the bank to sell its loans to the securities market.

inthewoods writes:

Excellent post - I would only comment on the end saying the left blames deregulation - I would just add that your post does not address the deregulation of the investment banking industry. The SEC allowed the IBs to exceed 12-1 leverage ratios. This, to me, is a key part of the crisis. A 30-1 leverage ratio means that a 10% down move will wipe out your capital.

Comments for this entry have been closed
Return to top