29 Oct “The Mother of All …”
Part II walks us backwards through key inflection points in monetary history from 2008 to 1574, towards the end of the Lyons (France) trade fairs. Some cities in France rose and fell on the strength of selling special products (in Beaucaire, at the Foire de la Madeleine, it was goods from around the Mediterranean-see Braudel, 1988) but the fairs of Lyons attained staying power through the supply of trade credit. Towards the end, Henry IV began tapping merchant credit to fund France’s war debts, triggering a series of bankruptcies, and the credit fairs moved to Genoa. Amato and Fantacci explain how the establishment of the Bank of England in 1694 shifted the credit risk for war finance from the king to the economy, thereby linking money, debt and sovereign interest tightly together, and making finance be “for finance’s sake.”
(The authors leave out some other crucial elements in the finance story–how cargo finance went on balance sheet and became the modern corporation. And I will leave war finance out of this article, not because it’s wholly irrelevant but because it’s not my main point.)
Securitization has a special status in “finance for finance’s sake.” It promises quantifiable returns on risky repackaged debt, and a rational investor will always prefer securitization to trade or enterprise finance. So will good borrowers. Securitization lowers the cost of capital by reducing uncertainty to a somewhat-quantifiable measure for which the investor is paid fairly given what is known at origination. The borrower keeps the change, a kind of reward for creating value. The safer the credit, the better for the borrower. In traditional middle-market lending, it’s the reverse. The safer the credit, the better for the lender.
Alas, many professionals have learned about securitization from the media or politicians (or banks) and don’t know the crucial difference between securitization and on-balance sheet corporate debt. On-balance sheet corporate debt disclosures are annualized. They hide defaults and losses by rolling over the problem ad infinitum, until the problem becomes too big to hide. EG, a bank loan portfolio has an average maturity of 5 years and annualized loan loss reserves of 4%, then these reserves are not covering the 60% of likely future losses. I.e., the average life is 2.5 years x 4% – 4% = 6%; and 6%/10% = 60%. That is assuming the 4% projection is honest…but who’s checking?
By contrast, securitization is an amortizing model. Every period, interest and principal cash flows are counted. This makes it possible to compare ongoing performance to the initial promise. If the risk turns out to be higher, the loss will hit investors based on the rules. Unexpected losses beyond a certain measure can’t be rolled over…unless there’s a loophole in the deal, or unless the securities can be re-securitized.
In short, the difference is a bit like two strategies for estimating the number of jelly beans in a jar. Corporate finance works by guessing. Securitization works by counting.
In the litigations I was involved with right after the Crisis, the case theories tended to focus on violations of lending protocols: loans, not securities. As more information became available, it became clearer that violations of lending protocols were motivated by violations higher up the power hierarchy, i.e., in the rating process. Over time, as lawyers and judges became more knowledgeable about securitization market practice, the case theories focused more on market standards violations.
It’s much easier to blame securitization and call for a return to fundamentals that everyone can understand, than it is to learn about technical subtleties that influence the structure of incentives. The problem is, finance isn’t for “everyone.” If we want to get a better grip on regulating financial behavior today, we need to understand not only how money is produced but also how finance people think. They run the finance show.
Many people accepted at face value the Fed’s explanation for imposing QE, as an application of lessons-learned from the Crash of 1929. But those of us who understand securitization also knew the Fed can’t roll over its underwater RMBS and RMBS CDOs. The SPVs that issue these bonds are not little banks that can grow their way out of losses; they are self-liquidating pass-throughs. The Fed booked a huge loss as a result of the bailout, then turned around and did with key rates what the banks did with LIBOR: lowered the Fed Funds rate to unrealistically low levels to manufacture positive spread. That’s not stimulus. It’s called filling a hole.
Because securitization was designed specifically not to roll over debt, it caps the banks’ ability to make money from money in the long term. Potentially it might have been used in creative ways to circulate money in the real economy and not just inflate consumer indebtedness. But that’s hard. In 2007, it was not an attractive option. There was easier money to be made by convincing the licensed credit rating agencies to violate their own risk-measurement standards. It wasn’t hard to do because very few people knew enough to spot the flaws in the rating system for structured finance, except the ones mining the flaws. With the right monetary incentives, the opacity of standards made the logic for the raters to break their own rules a no-brainer.
How did the agencies break their own rules? Quite simply, by allowing amortizing ABS and RMBS to be rolled over into revolving structured finance vehicles so as to delay the recognition of problems. These revolving vehicles originally were designed to help the banks short-circuit Basel limitations on leverage without losing origination income. They went by names like CDO (collateralized debt obligation), ABCP (asset-backed commercial paper) and SIV (structured investment vehicle), a CDO and ABCP hybrid.
CDOs never worked as advertised. In 2002-3, they were used to refinance fallen angel and junk corporate bonds. Many insurers who bought them suffered big losses, but it was in no one’s interest to disclose the scale of the problem. I only learned about it then in private conversations. ABS, RMBS repackaging in CDOs was raised as a possibility but consistently rejected by the raters until the early 2000s. Then, RMBS backed by subprime began hitting the skids. Because the default risk of CDOs is benchmarked against the corporate default benchmark (which does not reflect total PD or EL) and because RMBS ratings are infrequently updated, putting failing RMBS with still-high ratings into CDOs delayed the recognition of the performance problem until the risk could change hands. ABCP had the same flaw. SIVs combined the worst features of both.
The historical record is dismal, but we don’t need history to prove the mathematics: Start with $1. If the expected loss is 20%, that gives you $0.8. Repackage, and you have $.8 x .80. Repackage again and again, x .80 x .80. At the limit, the value to the last buyer is $0. Amato and Fantacci are right to call it the “the end of finance.”
When Nixon terminated dollar-gold convertibility and declared, “I am now a Keynesian in economics,” he set in motion a raft of incentives for banks that were anything but Keynesian. Most of all, he freed banks to produce, measure, and price debt capital at will. In any other industry this would be called cooking the books. But the allure was irresistible, and it seemed to enhance American soft power. Few in my generation who benefited from good educations could avoid the seduction of easy money, including me.
Then, securitization came along and actually imposed a limit on bank freedom by putting risk measurement in third-party hands. The most effective measurement tool was a numerical scale of impairment that Moody’s used on new originations, from the very beginning of the market until around the time of its IPO. (Read more here.) Had the same measurement rules been applied to both the primary and secondary structured market, excessive subprime origination would have been cut short. History would have taken a very different path.
Since the Crisis, we’ve learned to distrust many things we used to trust, including those institutions designated by the U.S. government to measure credit risk. But if we want to make debt sustainable, we need to recover enough trust somehow to begin believing once more that the rules are in place for a payment promise to be honored.
Securitization by the rules is safer and arguably fairer than single obligor lending, because it establishes a reasonably fair game for both sellers and buyers of credit risk so long as it isn’t allowed to devolve into a game of rent-seeking where capital always has the advantage. Securitization didn’t give birth to rent-seeking, and default didn’t just begin when the market started calling borrowers sellers or lenders investors.
So, please don’t throw out the baby with the bathwater. What Nixon did isn’t the mother of all securitizations. Absent the capacity of the market to regulate itself, and arguably that of regulators to enforce a level playing field between Main Street and Wall Street, securitization is an effective financial market leverage balancing mechanism. But it needs something much simpler, more transparent and powerful to work properly. An arm’s-length public yardstick, and competitive vigilance.