(portions excerpted from my September 18 Congressional Testimony)
The plan currently under consideration in Congress is bound to ultimately fail because it does nothing to resolve the fundamental conditions of valuation transparency and corporate and household leverage that are the origins of the crisis. To the naïve, the most basic consideration of value in a structured finance arrangement lies in the performance of the loans in the collateral pool. To the learned, that performance never truly mattered because those loans were never “sold” to the pool in any fashion typical of a normal “passive” transaction. Without resolving this basic idea of a true sale, we will never have market-based valuation in the structured finance sector, and the Paulson plan incentivizes continued growth along that paradigm.
To form a view of the objective “truly passive” sale, it is useful to use the automobile analogy. Suppose I sell you my car. I give you the car and receive cash in return. You drive away. If the engine falls out two weeks later you are out of luck. Classical caveat emptor applies. That is a true sale. From an accounting perspective, we can clearly take the car off the balance sheet and replace it with the cash at the time of the sale.
Now suppose I give you a warranty at the time of sale: for six months, if anything happens I will give you your money back. I should probably keep some amount of cash on hand to satisfy the obligation should you run into trouble. From an accounting perspective, it would make sense that I should not close the sale until the six months have passed. That is, I should hold reserves against the possibility that the car will break down. The entire car is still “on-balance sheet” (as a contingent liability) but so is the cash so we have no effect until the warranty expires.
Now, adding complexity, suppose that I give you a warranty that states I will fix the car – not give you your money back – if anything goes wrong. Perhaps I don’t have to reserve against the entire price of the car now, but only the cost of repairs. That could make sense, but the maximum cost of repairs is still the total value of the car. Even if I could estimate the probable cost of repairs, I may still want to remain aware of the total possible liability involved in honoring the warranty.
Back in the early development of securitization, and indeed still today with most Real Estate Mortgage Investment Conduits (REMICS), the underlying situation is that of the first example of the true sale above. A pool of loans is purchased from the seller and financed through RMBS. The pool is passive in the sense that the seller is legally restricted from swapping loans into and out of the pool.
But as different securitization structures and paradigms developed, sellers and investors saw the capacity for greater arbitrage through manipulating pools, rendering the passive view of many securitizations revealed in even FASB’s view toward its FIN46 post-Enron reforms obsolete. In 1989, early credit card securitizations showed the path forward. Because credit card accounts are, on average, outstanding less than a year it is necessary to structure a “revolving” period so that the pool of loans can be funded by bonds of longer maturity than the loans, themselves. During the revolving period, old credit card loans that have been paid off are replaced with new loans so that the pool balance remains relatively constant. Since the early 1990s, there were concerns with “cherry picking” the new accounts to be added so as to increase the credit quality of the pool. While such practice has never been officially confirmed or denied, it has appeared as if regulators allow such practices to encourage stable funding for the industry.
But the idea of selling the loans while not really passing along the full risk of the loans’ performance was too attractive for the rest of the world to pass up. Hence, Enron embarked on a similar endeavor, “selling” something for accounting purposes while retaining the economic risk. Of course, the risk was truly borne by Enron until the firm failed, resulting the spectacular disentangling of myriad funding conduits and instruments implemented in the process and sold throughout the world.
While FASB’s FIN46 revised accounting rules to preclude another Enron, it specifically exempted consumer credit securitizations from its rulemaking. Having had their future demonstrated so clearly by Enron, however, non-bank financial firms like New Century, investment banks like Bear Stearns, and even some banks and thrifts pursed the same strategy. Those firms sold pools of mortgages that bore little resemblance to the early REMIC structures. In the private-label RMBS, loans that didn’t perform well were repurchased in a ready and fluid fashion, more akin to an inappropriately off-balance sheet covered bond than any sort of passive true sale securitization envisioned by FASB.
Many sellers have voluntarily provided additional support to preserve the performance and bond ratings of their structured transactions. (Moody’s Investor’s Service, “The Costs and Benefits of Supporting “Troubled” Asset-Backed Securities: Has the Balance Shifted?” January 1997) Still, it would be egregious to maintain that securitization transfers no risk at all. After all, in the event of catastrophic asset quality problems, the seller may choose NOT to support a troubled deal, notwithstanding any legal responsibility to do so. In such a case, the asset-backed bond investors and any third-party credit enhancers, such as a bond insurer, would absorb the residual losses. By contrast, a portfolio lender would have to absorb all losses.
Trouble begins in this paradigm, however, when loan performance sours beyond the ability of the seller to support pool performance out of regular operating earnings. Then, the seller has to either increase earnings or stem losses. Since the seller’s earnings primarily rise through MAKING NEW LOANS TO GENERATE UNDERWRITNG FEES, the seller accelerates underwriting. Since better-qualified borrowers will most likely obtain cheaper loans from financially sound lenders, the seller targets down-market consumers – subprime borrowers – for the new business.
Of course, less creditworthy borrowers mean more losses. So the deterioration in loan performance that prompted the decline is met with more deterioration in loan performance. As the firm tumults down the death spiral, they attempt to modify loans using repayment and forbearance plans, while aggressively reaging loans to classify as much of the portfolio as “performing” as possible. Some lenders, upon realizing that they were unable to generate enough repayment and forbearance plans to feed the reaging process, resort to “amnesty” programs, wherein they merely wrote off the past due balance and called the loan current once again – sometimes without the delinquent borrower even knowing their loan had been awarded this “amnesty”!
Loan swapping under “representations and warranties,” therefore, together with loan modifications carried out through ongoing servicing allow the seller to readily absorb the loan risk that was purportedly sold. Hence, Nomura notes that, “…without audits or third-party oversight, an ABS servicer in financial distress may manipulate amortization triggers, divert deal cash flows, or otherwise misappropriate assets. (Nomura, “ABS Credit Migrations 2004,” December 7, 2004, p. 41)
The financial prospects for a seller that is unable to muster the resources to voluntarily support a securitization are grim. Such a seller would likely no longer receive any excess spread from the securitization trusts and might have difficulty raising external cash due to uncertainty over the asset quality of its serviced portfolio. Such a seller would surely not be able to issue again in markets any time soon. Hence, the seller can be reasonably expected to fail outright in the near term (Moody’s Investors Service, “Bullet Proof Structures Revisited: Bankruptcies and a Market Hangover Test Securitizations’ Mettle, August 30, 2002, p. 3)
Bondholders often have a legal right to replace the primary servicer with a backup servicer, since, “…the performance of securitized assets can be impaired by actions taken by a servicer in financial distress, but they usually need to do so before bankruptcy. The bankruptcy court may not allow replacement of the servicer since servicer rights may be viewed as a property right of the debtor’s estate. Investors thus may have no choice but to continue with the original servicer even if the quality of its servicing is poor. Even if the servicer is willing to give up servicing rights, those can often be difficult to transfer because they are tainted by the servicer’s malfeasance, often of too little value for a follow-up servicer to maintain on any reasonably profitable basis. (Moody’s Investors Service, “Bullet Proof Structures Revisited: Bankruptcies and a Market Hangover Test Securitizations’ Mettle, August 30, 2002, p. 4; Nomura, “ABS Credit Migrations 2004,” December 7, 2004, p. 41)
Servicing rights also create difficulties for a bankruptcy trustee – including the Federal Deposit Insurance Corporation – who seeks to liquidate the assets of a failed seller/servicer. The loan servicing rights are often the final asset remaining in the firm and the substantial potential for servicer malfeasance as the seller /servicer approaches bankruptcy can deteriorate their value significantly. Thus, firms like IndyMac are difficult to liquidate because no other servicer is willing to service the portfolio without substantial remuneration to insulate them from the losses and legal ramifications of unwinding the potential fraud and malfeasance left over from the previous distressed servicer. That is why the FDIC was left servicing the NextBank credit card portfolio, and that is most likely the case at IndyMac as well (and some of that is probably behind Bank of America’s purchase of Countrywide at a very favorable price).
In summary, “true sale” as it has been practiced does not make sense. To get at any meaningful FASB reforms, we need to go back to the principle of true sale. I sold you the car. Caveat emptor. Retaining servicing rights are like maintaining the car after the sale. Loan swap agreements are simple warranties. In order to improve accounting standards, therefore, we have to put a limit on the amount of money that can be spent on maintenance, i.e., loan servicing, after the sale. While FASB has maintained that acting under representations and warranties is not optional, they are interpreting that optionality only in the strict legal sense. There is always the “real” option of simply refusing to support the pool representations and warranties. While that may result in some lawsuits, in reality those can be beaten back for several years giving the firm a chance to restore performance and eventually meet their contractual obligations. Hence, we also need to financially value the option in the warranty.
The problem is a tragic collision of economics, finance, and accounting, where economic risk is placed where it is difficult to value financially and even the most complex accounting rules do not apply. While that does not augur for prohibiting any of the above arrangements in the long term, it does provide a rationale for constraining financial product developments so that they do not grow systemically large before finance and accounting can properly characterize their risks and returns. Hence, there may be reason to curb the over-reliance on financial innovations – certainly within the realm of financial institutions that receive Federal and State safety net protection – and require public reporting of such exposures and values to better align incentives for innovation with the need for financial stability.