Here's why:

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.)

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.

Dean does say that Treasury should be given credit for releasing the data that informed the stress tests -- whether or not you believe in the test, we now have a much clearer idea of bank balance sheets. But he's worried, as I am, that if this scenario proves too rosy and the banks both pass the stress test and prove unable to survive the next few years, that the administration's credibility will be seriously eroded. It's bad to cry "wolf," but it may be even worse to cry "not-wolf."