Hussman on the Relaxing Stress Test

Full article

Last week, financial stocks enjoyed a powerful advance and short squeeze on the announcement of the results of the “stress test” of major banks. It is important to begin by noting that this was not a regulatory procedure with teeth. It was initially a response to Congressional demands to introduce greater objectivity into the use of public capital for these bailouts, and gradually morphed into nothing more than a “confidence building” exercise. And keeping with the emphasis on keeping the numbers happy, as opposed to providing full and fair disclosure, the Wall Street Journal reported on Saturday, “The Federal Reserve significantly scaled back the size of the capital hole facing some of the nation's biggest banks shortly before concluding its stress tests, following two weeks of intense bargaining. In addition, according to bank and government officials, the Fed used a different measurement of bank-capital levels than analysts and investors had been expecting, resulting in much smaller capital deficits.”

To some extent, it is not possible to get full and fair disclosure using the method that regulators used in the first place, since it relied on banks' self-estimates of their potential losses in a further economic downturn. These of course being the same banks that made the bad loans, and have already proved themselves vastly incapable of loss estimation and risk management. Moreover, the Fed only asked for loss estimates for 2009 and 2010, not beyond – “Each participating firm was instructed to project potential losses on its loan, investment, and trading securities portfolios, including off-balance sheet commitments and contingent liabilities and exposures over the two-year horizon beginning with year-end 2008 financial statement data.” This period specifically excludes the window where we can expect the majority of “second wave” mortgage losses to be taken, as it does not capture any losses that will emerge as a result of mortgage resets from mid-2010, through 2011, and into 2012.

The “stress test” procedure also conveniently excludes any potential mark-to-market losses during 2009 and 2010, as banks “were instructed to estimate forward-looking, undiscounted credit losses, that is, losses due to failure to pay obligations (‘cash flow losses') rather than discounts related to mark-to-market values.”
Now, just think of this for a minute. Even if you assume that the “risk-weighted assets” of the banks are about two-thirds of their total assets (as the stress-test does), we're still looking at $7.8 trillion in total assets at risk in these banks, and despite being on the edge of insolvency only weeks ago, we are asked to believe that they will need less than 1% of this amount – $74.6 billion – of additional capital even in a worst case scenario. How do the stress tests arrive at this conclusion? 1) They underestimate potential losses by minimizing the horizon over which the losses would have to occur, excluding potential mark-to-market losses and restricting the loan losses to “cash flow” losses only; 2) They define capital well beyond tangible sources, to include about double what is available as Tier-1 common; 3) They include $362.9 billion in “resources other than capital” – essentially pre-provision net revenue expected to be earned by the banks over the coming two years to absorb potential losses; 4) They report the capital buffer that would be required after massive dilution in the common stock of these banks has already occurred.

As an example, Citigroup comes in with $119 billion in capital ($22 billion as Tier-1 common). Total assets are over $2.1 trillion, but the stress test assumes “risk weighted” assets of less than half that. Citi projects losses in 2009 and 2010 of $104.7 billion in a scenario where the unemployment rate reaches 10.3%. Citi assumes that it will earn $49 billion during that period which would partially absorb those losses, and that it will obtain $87 billion in Tier-1 from other capital sources, presumably including $33 billion of preferred that it would be willing to convert to common. Of course, Citi's entire market cap is only $22 billion, so the “$5.5 billion” that Citi is reported to need under the stress test is what it would require after a 5-to-1 dilution in its common stock (87+22/22). Essentially, we've got a company with a common equity buffer of just over 1% of total assets, that just 8 weeks ago was on the verge of receivership, and investors are urged to believe that there are enough voodoo dolls in the vault to make the company solvent even in a further weakened economy.