Lehman Watch

[EDIT - Great article by Whitney Tilson on Einhorn and his battle with Lehman. I have been following Einhorn for a while and I am glad to see others come out in support of him. Of course, as fundamental analysis goes, Einhorn is far more sophisticated than I and you should give him a read, but I hope you enjoy this article anyway.]

Bloomberg is reporting that "wall Street banks, brokerages and hedge funds may report $460 billion in credit losses from the collapse of the subprime mortgage market, or almost four times the amount already disclosed, according to Goldman Sachs." That is a staggering number. In light of this ominous statement, I thought a closer look at Lehman is warranted. Recall that on the day of Bear's troubles, Lehman's stock fell to $30, around 50% on concerns it would be the next shoe to drop. It has since returned to the $45 range after reassurances, acquiring lines of credit, and a better than expected earnings report. The problem, however, is that one needs to look beyond the myopic backward looking headline numbers on the income statement and focus on the future earnings potential by evaluating the balance sheet and the repeatability of the earnings. To do this, there are two important accounting rules to keep in mind.

1. FASB Statement 157. Issued last October to clarify how firms book assets and liabilities, the "Statement emphasizes that fair value is a market-based measurement, not an entity-specific measurement" and "establishes a fair value hierarchy" that is intended to describe the method by which assets and liabilities are booked. Level 1 of the hierarchy describes "financial instruments that trade in an active market," whereas level 3 describes assets valued "using significant unobservable inputs." Level 1 assets are high quality instruments such as treasuries or agencies that are still liquid in this environment and level 3 assets are the various structured products for whom liquidity has dried up (sub prime CDO's come to mind).
   
   The nice thing about quoted prices is that they are very really fair value. Thus the value of level 1 assets, being marked-to-market, can generally be trusted. However, when it comes to level 3 assets, what raises immediate red flags are the "significant unobservable inputs" that are used, in essence, to mark-to-model. Having taken various financial engineering and computational finance classes here at Princeton, I can attest first hand to how sensitive prices can be to inputs. To illustrate, consider the Black-Scholes option pricing framework. Depending on what volatility is plugged into the formula, one can produce different prices; higher vol means higher price, and lower vol means lower price. In the case of Black-Scholes, the price may not change significantly, but when pricing more complex instruments using more advanced mathematical tools, prices become very sensitive to inputs. For example, the standard CDO pricing method is to use a Gaussian copula to capture correlation. I have played around with this a little and can tell you that changing the correlation affects the price quite radically, much more so than changing vols in the Black-Scholes model.
   
   In short, the booked value of level 3 assets is dubious. There does not necessarily need to be an intention to mislead. In the case of Black-Scholes, there is a firm anchor in historical vols, implied vols, or some form of stochastic vol estimate. But with a CDO, how does one determine the appropriate correlation? We have seen correlations go haywire in this environment. Further, the very nature of the instruments, that they require significant assumptions (Gaussian copula, Gumbel copula, or something else entirely?) to make the models tractable, can result in imperfect prices. Even if one wishes to fairly mark-to-model, the model itself may be completely divorced from reality.
   
   I have expressed my opinion earlier that "level 3" asset is just a euphemism for "coming write down" and I stand by it. Some more so than others, of course.
   
   
2. FASB Statement 159. This is a very interesting rule. It is described here:
   

   How does that work? If the market decides a company is a bigger credit risk and starts demanding fatter risk premiums to buy its debt, the value of its existing debt falls. Under a rule being phased in throughout corporate America known as Financial Accounting Statement No. 159, that same logic applies to a company’s own debt. Companies that mark their liabilities to a market price, as Wall Street usually does, thus record as revenue a drop in the value of their own debt obligations.

   In essence, they make money because they owe less.

   Accounting experts said the exercise is perfectly legitimate, particularly if firms that mark liabilities to market do the same with their assets. At the same time, it highlights one of the ironies of so-called fair value accounting. “If you have a liability that declines in value because your credit worsens, you have a gain,” said Stephen Ryan, associate professor of accounting at New York University’s Stern School of Business.

Lehman's Balance Sheet

So how does this apply to Lehman? Well, here are the problems I see. Actually, all of this was noted by others, I am just collecting and distilling here.

1. Leverage. Lehman is highly levered, and that has only increased recently: "It’s quite simple: The more leverage Lehman has, the less room assets have to fall to wipe out its equity." They are currently at 32x. Not to belabor the point, but leverage has caused serious problems for others.
2. Definition of equity. Apparently Lehman counts long term subordinated borrowings as equity. Excluding this, leverage is in the 40x range.
   
3. FASB 159. Lehman "earned" $600 million this quarter from their debt trading lower. I would hardly consider those kosher earnings. Net income was only $489 million.
   
4. FASB 157. Bennet Sedacca notes that Lehman has four times as many level 3 assets as capital. Given the dubious nature of those valuations, that is a cause for concern.
5. Anecdotal evidence. I interned last summer at Lehman in fixed income research/trading. I was surprised at the time how much of their business was built around mortgages and structured products. More surprising was that every presentation I attended seemed to imply that mortgages and structured products would continue to remain high growth areas, even as the climate was changing. Of course, I wouldn't put too much weight in any impression I gained from presentations and prospectives, but I would use it as a guide to focus more detailed research. Previously their growth has been centered around structured products. We should look under the hood more carefully to see if there are any latent concerns, or if they have not successfully pursued other growth opportunities and moved away from mortgage backed debt.
   

What concerns me the most are points 1 and 4. That kind of leverage doesn't require much of a revaluation on the level 3 assets before solvency becomes a problem. Based on the Fed's credible signal, a revaluation is certainly on the cards.