The administration's argument for the stress tests' stringency is made on page seven of the results. There's even a helpful graph:


In short, the stress tests are assuming a loan loss rate -- that is to say, the percentage of loans a particular firm has written-off as non-recoverable -- higher than that of the Great Depression. An adverse scenario, in other words, that makes Franklin Delano Roosevelt look like a wimp. That seems comfortingly stringent. On the other hand, it's not clear that the two periods are directly comparable. Administration critics will tell you that Depression-era loans were less heavily leveraged. Mortgages had 50 percent downpayments. There was no housing bubble. This downturn won't be worse than the Great Depression in its total effect on society, but the loss rate is a misleading indicator. And in pushing that indicator, the administration is being a bit slippery. (And it's worth noting that some metrics in this recession have outperformed similar measures from the Great Depression.) A more direct comparison between current conditions and the "adverse scenario conditions is quite a bit scarier.

It's sort of hard to adjudicate this debate. If the next two years prove extremely grim and the capital buffers set by the stress tests prove comically inadequate, then we'll be able to say pretty confidently that the stress tests were not stringent enough. Conversely, if the next two years see slow improvement and the stress tests describe an adverse scenario that never quite occurs, they'll look pretty good in retrospect. At the moment, we can say with some confidence that the adverse scenario seems a whole lot likelier than the baseline scenario.

But like most economic issues, objective reality only has a bit part in the drama. There were two competing objectives with the stress tests. To describe what is likeliest to happen and to affect what is likeliest to happen. Manageable results that reassure the market and save the banks might unlock the funds of scared investors, who will in turn invest in and strengthen the banks, who will in turn accelerate their lending, which will in turn help businesses, which will in turn blunt unemployment, which will in turn lower foreclosures, which will in turn lower the loss rate, which will in turn make the relative optimism of the stress tests look good. But none of that will happen if the tests are so optimistic that they mis-describe reality, don't save the banks, and reduce the administration's credibility for future policy interventions. That line between describing reality and changing it, however, is a hard one to walk.

You know who I wouldn't want to be? Timothy Geithner.