Most news outlets seem anxious to join the Treasury's PR campaign in pronouncing the banks essentially healthy based on the stress test results. There is of course enormous uncertainty around the course of the economy over the next few years, and the results of these stress tests may well prove to be an accurate assessment of the banks' health, but there are some reasons for believing that the stress tests are likely to prove too lenient. 1) Fraud in mortgage issuance -- we know that many of the loans issued in this period involved fraud, more often on the lenders' side than the borrowers. In these cases, for example where the mortgage application grossly overstates the buyers income or the appraisal hugely overstates the market price of the house, default rates will be far higher than would be expected even in bad economic times. Also, recovery rates will be far lower if the original appraisal price was inflated. 2) Unemployment -- in their negative scenario, the stress tests assumed a year-round average unemployment rate of 8.9 percent for the 2009 and 10.3 percent for 2010. The economy is on track to have a much higher unemployment rate, as it is likely to hit 9.0 percent in April. My best guess for a year-round average would be 9.4 percent for 2009 and probably around 10.5 percent for 2010. (These numbers assume no second stimulus, but of course Congress will not sit back and just let the unemployment rate go through the roof.) 3) House prices -- the negative scenario assumes that house prices, as measured by the Case-Shiller 10-City index fall 22.0 percent in 2009. Prices in this index have been falling at a 24 percent annual rate in recent months. Given the massive inventory of unsold homes, It is reasonable to expect that this rate of price decline could continue at least through 2009. What difference would harsher assumptions make? The projected loss rate on first mortgages increases by 45 percent between the baseline scenario and the negative scenarios in the stress tests. The baseline scenario assumes an 8.4 percent unemployment rate for 2009 and 8.8 percent for 2010 (some serious stimulus here), compared to the 8.9 and 10.3 rates in the negative scenario. The rate of house price decline in the baseline scenario was 14 percent in 2009 and 4 percent in 2010, compared to 22 percent and 7 percent in the negative scenario. So, if my somewhat more negative numbers prove accurate let's assume that it increases losses by about 20 percent. That comes to an additional $120 billion in losses. That would mean that instead of having to raise $75 billion, these banks would have to raise $195 billion. That's a qualitatively different picture. So, are the stress tests worthless? They did provide a much clearer picture of the position of individual banks than we had previously. It is worth noting that this is a 180 degree shift from the original course pursued by Treasury Secretary Henry Paulson last fall. Paulson tried to conceal the situation of individual banks, putting a cloud over all of them. Treasury also should be credited for disclosing many of the specifics of the stress tests so it is possible to do a quick (or more in depth) analysis of its assumptions and explore the implications of alternative assumptions. Still, it is hard not to conclude that these stress tests and certainly the PR campaign around them, were intended to paint as positive a picture as possible of the banks' financial condition. If this picture proves to be wrong, then it means that we will have unnecessarily delayed the clean-up of the financial system. It will also be bad political news for the administration (Geithner and Summers will presumably be joining the ranks of the unemployed). Of course, the big second stimulus package that Congress will pass this summer, will save both the banks and the administration.
--Dean Baker
