The Geithner Treasury team finally unveiled the details of its Stress Test methodology today (see Supervisory Capital Assessment Program). We will have to wait a few more weeks for the results. Based on early returns, the results are all but a foregone conclusion – most banks will likely pass with flying colors.
So in the absence of meaningful outcomes from the Stress Test, we might be better served by deconstructing the process itself. In this way we can better determine what it might mean for a bank to pass or fail, better assess the results to sensitivity in the data, and ultimately, determine whether the test itself even made any sense.
So what can be made of the process?
It has recently been argued that the “more adverse” scenario should actually have been the baseline (see Stress Test Scenarios or Stress Testing the Stress Test). I agree wholeheartedly. It has been clear for awhile now that the Treasury’s scenarios did not adequately reflect economic reality. There is even some evidence that the economy has already deteriorated to a point where it is currently worse off than the “more adverse” scenario.
That notwithstanding, there were two other details about the methodology (among others) that raise some concern.
First, from p. 4 of the official Stress Test document:
The BHCs [bank holding companies] were asked to estimate their potential losses on loans, securities, and trading positions, as well as pre‐provision net revenue (PPNR) and the resources available from the allowance for loan and lease losses (ALLL) under two alternative macroeconomic scenarios.
My comment: So if I understand this correctly, the Treasury is relying on banks to provide them with an assessment of their current troubles. This is like asking an alcoholic if he thinks he has a problem.
Second, on p. 3 of the document, the Treasury points out:
The economic value of loans in the accrual book is reduced through the loan loss reserving process when repayment becomes doubtful, but is not reduced for fluctuations in market prices, which may be driven by market liquidity considerations…As a result…the results of this exercise are not comparable with those that would evaluate such assets on a mark‐to‐market basis.
My comment: So here we are again in a mark-to-fantasy kind of world, where banks (which possess asymmetric information about the true quality of their underlying asset portfolio vis-a-vis the Treasury) might be able to favorably adjust the value of their assets. In my opinion, the best assessment of the underlying value of the assets on the books is what the market is willing to pay for those assets. The Treasury’s approach, in my opinion, leaves too much room for banks to claim, “It’s a liquidity problem.”
All this begs the question: While most banks will pass, does the Stress Test fail?