Hayek, the True Sale Doctrine, and the Origins of the Financial Crisis

Hayek.jpgHere is my theory du jour about the origins of the financial crisis, suggested by one of my students*: blame it all on the true sale doctrine or rather on its evisceration. Stick with me to the end, and I have some overly broad generalizations about expertise, property rights, and Hayek.

The “true sale doctrine” is not a staple of the law school curriculum. At best it makes a brief cameo in secured transactions and bankruptcy courses. Notwithstanding this academic obscurity, however, its failure may have had a big role in the current melt-down of the banking sector and with it the world economy. Here is the gist of the issue:

Securitization is the process by which financial assets (essentially promises to pay money in the future) are transferred from their original holder to a special purpose vehicle such as an LLC or business trust, which then issues securities entitling the holder to some fractional right to the income from the transferred assets. Hence, for example, a bank might transfer mortgage loans to an SPV, the SPV would then issue securities to investors, and the cash from the sale of these securities would flow back to the bank. The investors in the securities have two ultimately inconsistent goals.


First, they want to be certain that the underlying financial assets will actually perform. The risk is that the loans held in the SPV’s portfolio will default and there will be no money to pay the holders of securities. One way of getting around this problem is for the originator — that is the bank that initially sold the mortgage loans to the SPV — to provide “credit enhancements” in the form of letters of credit, put options, or the like. All of these devices ultimately consist of a promise that the bank will pay the SPV the face value of the underlying mortgage loan if the actual home buyer defaults.

The second thing that the owners of the securities want is a guarantee of “bankruptcy remoteness.” They want assurance that if the originator goes bankrupt the transfer of the financial assets to the SPV will be treated as a sale, and the transferred assets will not be sucked into the bankruptcy proceedings.

The true sale doctrine is supposed to mediate these competing goals. Under ordinary commercial law rules (UCC 9-109 to be precise) a transaction that walks, talks, and acts like a loan will be treated as such, regardless of what you call it. In a securitization transaction where all of the risk of default on the underlying financial assets is pushed back onto the originator the law empowers courts to re-characterize the sale of the assets to the SPV as a secured loan. The idea is that in reality there has been no sale. The originator has simply made a promise to the investors in the securities that their money will be repaid and the underlying financial assets are in effect collateral on the originator’s promise.

In short the true sale doctrine is supposed to make sure that treatment of a transaction hinges on the real allocation of risk, not on the label that the parties choose to attach to the transaction. In practice, the application of the true sale doctrine is a mess, but at least it makes all of the parties to the transaction sweat through an opinion letter. They know that if the originator retains too much risk, much of the value of securitization will be lost.

Unless of course, the originator is a bank. Banks get special rules, rules that don’t include the true sale doctrine.

Banks are not subject to the bankruptcy code. Rather, in the event of insolvency, they are subject to the Federal Deposit Insurance Act. Hence, the determination of whether the finacial assets securitized by banks are “bankrtupcy remote” is governed not by the UCC and its case law, but by the FDIC regulations promulgated in the CFR. Among those regulations is 12 CFR 360.6(b), which states:

The FDIC shall not, by the exercise of its authority to disaffirm or repudiate contracts under 12 U.S.C 1821(e), reclaim, recover, or re-characterize as property of the institution or the receivership any financial assets transferred by an insured depository institution in connection with a securitization or participation, provided that such transfer meets all conditions for sale accounting treatment under generally accepted account principles, other than the “legal isolation” condition as it applies to institutions for which the FDIC may be appointed as conservator or receiver, which is addressed by this section. (emphasis added)

The italicized language essentially repeals the true sale doctrine for banks. So long as your accountants call a transaction a sale, it will be treated as such regardless of the underlying distribution of risk. Why does this matter? Because potentially it allows the banks to engage in a massive amount of … er … lying, creativity, self-deception, what have you … on their balance sheets. When the subprime mortgages were sold the banks got a lot of cash that the could use to buy stuff that was then carried on their balance sheets as an asset, propping up their capital reserves to comply with regulations. The problem, of course, is that the underlying credit enhancements required by the securitization were a huge liability. Under the true sale doctrine, on the other hand, these transfers would have been treated not as sales but as loans to the bank, and the bank would have carried the obligation as a liability. The result was a huge incentive for banks to securitize subprime loans that they then in effect became the insurers of.

Now the folks at the FDIC who drafted this rule would no doubt respond with two arguments. First, they would note that the 12 CFR 360.6(b) doesn’t totally abdicate inquiry into the loan-sale distinction, it simply punts it to the accountants rather than the lawyers. The problem here, as I understand it, is that the accountants in effect said that if you have control over an asset you own it, even if the underlying risk of nonpayment resides elsewhere. They did this because the accounting rule also required “legal isolation.” In effect, the accountants punted that question back to the lawyers. However, when the FDIC decided to make the accountants the arbiters of the underlying property rights they punted only to the part of the accounting rule that did not include the legal isolation test.

The second response by the good folks at the FDIC, I am assuming, would be that the risk that was retained by the banks through their credit enhancement agreements with the SPV was itself accounted for on the balance sheets. In effect, they said “Don’t worry about the fact that the allocation of that risk no longer effects property rights, our expert accountants can look at the contracts and tell you what they are worth in the abstract.”

As it has happened, of course, the accountants valuation of banks’ liabilities have turned out to be catastrophically wrong, with the result that the dip in subprime mortgage performance, rather than giving institutions a bad quarter has sent them spiralling into insolvency. On the flip side, the fact that lawyers rather than accountants were the door keepers for true sale status at non-bank institutions may be one of the reasons why securitization by these companies has not pushed them into insolvency like the banks. In effect, the true sale doctrine created a counter-balancing incentive against the push that the institutions become total insurers of the assets they were purportedly selling.

Here is the theoretical pay off for those who have made it this far in the post. There is a deep Hayekian point here about the best way of dealing with the allocation of risk. The FDIC rules ultimately put their faith in expert opinion, in this case the accountants. We didn’t need to worry about the risk the banks were taking on because the accountants, through an intellectually heroic feat of calculation could value that risk. The true sale doctrine, in contrast, ultimately put its faith in a decentralized process driven by property rights. The true sale doctrine meant that the allocation of non-payment risk to the bank would have a real impact on the value of the property rights represented by the securities issued by the SPV. Push too much risk on to the originator, and you will be treated as a lender rather than a purchaser. As anyone who has looked at the doctrine will tell you, of course, the loan-sale distinction is slippery, ad hoc, and often comes down to a guess about what the judge had for breakfast. But at its heart, it was an attempt to match property rights to risk and let the market sort it out. In contrast, the FDIC put its faith in experts.

*The student in question is Mark Brandenburg. Potential employers should keep their eye out for his resume.

You may also like...

7 Responses

  1. Jeff Lipshaw says:

    Nate, I didn’t have to see your name on the RSS feed to know that this HAD to be your post.

    This is your key move:

    “One way of getting around this problem is for the originator — that is the bank that initially sold the mortgage loans to the SPV — to provide “credit enhancements” in the form of letters of credit, put options, or the like. All of these devices ultimately consist of a promise that the bank will pay the SPV the face value of the underlying mortgage loan if the actual home buyer defaults.”

    I don’t think your second sentence follows. Not all credit enhancements are created equal, at least in terms of remoteness, or true sale. I can see an argument that the SPV’s right to put the instrument back to the bank undercuts a true sale, but a letter of credit or a credit default swap is the promise of a third party to step up and pay. That’s not a promise of the bank to pay, and hence I don’t see that it undercuts the “sale.”

  2. Nate Oman says:

    Jeff: I completely agree. I am over simplifying here to ease the exposition, but obviously you can have credit enhancements that don’t amount to guarantees. On the other hand, credit enhancements that DO amount to guarantees by the originator such as put options that let SPV’s require that banks repurchase non-performing loans at face value WERE very common. The same would be true if a bank issued a letter of credit on its own originated assets.

    The point of the post is that the absence of a real true sale test for “bankruptcy”* remoteness in banking transactions eliminated an important incentive for originators to adopt credit enhancements that pushed risk on to third parties.

    Certainly it seems to be the case that one of the big problems the banks face is that many of them, despite following an originate-to-distribute model find themselves in the position of in effect insuring the loans they had supposedly unloaded. One — not the only or even the most important but, nevertheless, one — reason for this may have been that doing so didn’t undermine their ability to promise investors “bankruptcy” remoteness.

    *”Bankrtupcy remoteness” is the wrong term here, since banks aren’t subject to the bankruptcy code. Perhaps “insolvency proceedings remoteness”?

  3. Nate Oman says:

    Another de facto guarantee was to retain the equity tranche but book it at some inflated value based on an over-optimistic assessment of the risk of default on the underlying assets. When the assets fail to perform, you take a write down which is the equivalent of a payment.

  4. Bruce Boyden says:

    When was sec. 360.6(b) added?

  5. Nate Oman says:

    September, 2000

  6. Andrea Mertz says:

    Do you have any urls of a true sale opinion letter? I am currently doing research on these letters and need to read an example. Thanks.

  7. Martin Perrone says:

    How many true sales must a loan go through before it can properly be placed in a REMIC trust? (Properly securitized)