Pages Menu
Categories Menu

Posted by on Feb 25, 2009 in Economy | 5 comments

How Realistic Is The Bank Stress Test?

Some initial details were released about the parameters that the government is using to run its stress tests on all the major banks. There is a possibility that the results of this exercise will lead to the largest governmental restructuring of the U.S. financial system since the 1930s, but the whole thing could also just be hot air. If I had to venture a guess, I’d assume that they are desperately hoping it’ll be a formal exercise, but there is a good chance that they will be enormously surprised. Of course the whole test is highly dependent on what their projections assume, considering that the entire reason we got in this mess was because banks were laughably optimistic about what would constitute a “realistic downturn.” Calculated Risk has a post up that details the projections they are using and I’m going to steal the chart but he has some good graphs that you should click through to his site to check out.


I’m sure you’re all dying to read my armchair quarterback analysis so here it is. I’m actually pretty amazed, it seems pretty realistic based on how they defined it. The “adverse” level of -3.3% GDP for this year is probably more likely than only -2.0% but I don’t see how it could get much worse than that with the government stimulus. I think it’s way too early to even guess what next year’s GDP will look like so I won’t try. My initial impression was that the 8.9% unemployment rate for this year and 10.3% for next seems far from an “adverse” scenario and more like what I think we’d be lucky to see. But then I read that it said it was the “annual average” rate, and my opinion changed entirely. Even if jobs continued to be shed at the levels we’ve seen the last few months the entire year, then the yearly average will still be about 9.0%. I’m pretty pessimistic but even I think that is unlikely unless the financial system collapses completely (and I think that possibility shouldn’t be a part of a stress test). If you click through to the Calculated Risk link, he shows that the housing price projects will take housing back to historical levels for the mainstream scenario and a normal trough for the adverse one. Those levels have been what I’ve anticipated since ’06 so I can’t complain.

Of course I’m blogging so I can’t get away without being critical about something. I have to say I’m surprised by the reasonableness of their macroeconomic figures, but that I question the validity of the stress test dynamics. Specifically, the risk models are mathematically flawed, assume historical correlations hold and don’t do a good job of modeling feedback loops. So while I agree with the parameters I’m sure I would fiercely disagree with how they relate the actual default rates and asset values to the macroeconomic parameters. There inaccurate modeling is precisely what nearly caused the system to fall apart suddenly last fall. Not only are we dealing trillions of highly opaque securities that have never even been subjected to a normal recession, but we are in a period of deflation and massive international interconnectedness. So on one hand I have to give props to the government for recognizing the severity of the situation, while simultaneously assuming that the results will be completely wrong. We’ll see.

I would say that regardless of the outcome of the stress test, the optimistic sign is that the government finally is starting to look towards the long term with its eyes open and recognizes the futility of trying to prop up asset prices. I am cynical that they will make the right moves over the coming months, but I am cautiously optimistic that they will start approaching the problem from a strategic standpoint for the first time, and that is absolutely necessary no matter what the future holds.

Update: Notice that the “adverse” scenario envisions another 25% drop from current prices which means that we’ve only seen about half the decline so far. But Robert Shiller — who is the foremost expert on housing prices and bubbles — says that there is a good chance for a big overshoot of 10-20%. Shiller is not normally seen as a doom-and-gloomer either. The stresses on the banking system are only about 1/3 of the way done.