The Federal Reserve this afternoon released “The Supervisory Capital Assessment Program: Design and Implementation” (the “Report”), which describes in general terms the nature of the “stress tests” applied to the 19 largest U.S. banking organizations (those with assets in excess of $100 billion at year-end 2008). The release followed reported meetings on Friday morning between Federal Reserve officials and senior executives of the 19 banks to review the preliminary findings of the stress tests. The final results are expected to be provided to the firms on or about May 4, but it remains to be seen which elements, if any, of the stress test results will be made public.

The concept of the stress tests was first announced on February 10, as part of the Treasury Department’s Capital Assistance Plan (“CAP”). (Treasury released further details about CAP on February 25.) CAP is designed as a supplement to last fall’s Capital Purchase Program; CAP will provide additional Treasury investments in the largest U.S. banking institutions that may need additional capital to continue lending, even in the face of future losses. While each of the 19 largest firms have already received capital investments from Treasury, no funds have yet been invested under CAP itself.

According to the Report, the stress test has been designed to determine whether any of the 19 largest banking firms – all presumably systemically significant – need an additional capital “buffer.” This buffer is the additional capital that any of the firms might need to cover future losses yet remain sufficiently capitalized. Put another way, the buffer reflects the difference between the amount of capital left in a firm after expected future losses occur and the amount of capital deemed necessary for the firm to remain in sound condition.

The Report describes generally the assumptions used in the stress tests, the information provided by the firms, and the nature of the analysis. The highlights of the tests include the following:

  • The test for each institution involved two assessments: the impact on capital based on a “baseline scenario” set of assumptions about future performance of the U.S. economy and the impact based on an “adverse scenario” set of more pessimistic assumptions.
  • The baseline scenario assumes a decline in real GDP in 2009 of 2% and an increase of 2.1% in 2010; civilian unemployment of 8.4% in 2009 and 8.8% in 2010; and a decrease in house prices of 14% in 2009 and 4% in 2010.
  • The "more adverse" scenario assumes a decline in real GDP in 2009 of 3.3% and an increase of 0.5% in 2010; civilian unemployment of 8.9% in 2009 and 10.3% in 2010; and a decrease in house prices of 22% in 2009 and 7% in 2010.
  • It is important to note that the "more advere" scenario was not designed as a “worst case” analysis. Rather, the assumptions are those that the Federal Reserve believes have a 10% probability of occurring. This view may be generous. One school of thought is that the current recession is the result of events that were thought to have less than a 10% probability.
  • The Report also contains one useful nugget of information about predicting the performance of residential mortgage loans. The Federal Reserve looked at various features of already troubled loans and found that predictive factors included the type of mortgage product, FICO scores, the range of loan-to- value ratios, and geography. Not predictive, however, were the level of documentation of a loan and the year of origination – suggesting that concerns about “no doc” or “low doc” loans and about loans originated in 2006 and 2007 may have been overstated.
  • The useful predictive factors are not entirely reflected in the administration’s new loan refinancing and modification programs. It may be helpful to adjust these programs to account more carefully for the risks of delinquency.