Tuesday, March 03, 2009

The Unpredictability of This Crisis

As this crisis wears on, fewer and fewer people remember being surprised by it. For example, in a Blogginheads episode, Megan McCardle and Dean Baker both noted they anticipated the collapse even though Baker called for a 10% decline in housing back in 2002, which is about 20% below December 2008 level. But last April, as these problems were becoming apparent, I was at a meeting by the National Bureau of Economic Research, where all the top financial economists got together. Markus Brunnemeier gave a talk on the housing bubble and noted that about $200B in value had been destroyed via the housing price decline, and this represented only a couple percent in the stock market. The implication was the market had already overreacted. The consensus was this was correct, and I must admit I was no different.

I think what is most surprising about this is the accelerator mechanism that propelled a housing bubble into so many other sectors. It destroyed the market value of all sorts of assets, and seems caught in a positive feedback loop. A related puzzle is that the crisis seemed to start in the US, but the US equity market has declined less than most other countries, and our currency has strengthened. One of my favorite bank analysts, Tom Brown, called for a bank stock bottom last summer. I don't feel bad not understanding or anticipating this, because I know many thoughtful people missed it too, and that very few crises are understood in real time. Sure, with hindsight, I'm seeing connections, how a total lack in trust has created declines in market values which cause declines in banks that have to mark more and more of their assets to market, but realistically going back to last year I don't see how I could have anticipated this other than being a permabear, or knowing about the prevalence of NINJA loans. Indeed, maybe that is the missing variable, that one could not have predicted this problem without knowing how crazy mortgage underwriting became.


Anonymous said...

From NN [Identity having problems in my browser]:

Do you think speculative bubbles become easier to identify in the future when you have a few rules of thumb? You might be interested in the book 'Patterns of Speculation' by a econophysicist that analyzes speculative bubbles throughout history from a quantitative perspective.

Anonymous said...

i'm worse. i anticipated it and got pinched anyway. was in cash '05,'06 and '07, then convinced bernanke would print maybe $300-$500b and dilute everyone.

he starts cutting, they bail bear in the spring, market drops 15% he starts flooding, mkt recovers some. i put 10% in and spend the summer itching b/c couldnt find anything 'cheap'. in september i can't take the dilution anymore and put another 20% in. and then the bottom fell.

my only 'consolations' are they used 10t so far and buffet got taken too.

Anonymous said...

Do we need to pinpoint a trigger, a root a cause?

IMO the trigger doesn't matter because if the market is vulnerable to positive feedback loops then the slightest perturbation will eventually lead to crisis.

I speculate little about what the world will look like in 100 years but I would be shocked if we do not suffer from periodic economic declines.

Obviously there are good and bad policy responses to disruptions but eliminating major fluctuations seems impossible to me because the more "safe" you make the system, the more people then panic when the system appears to start breaking down.

J said...

I (like others) estimated losses caused by bad mortgages around $ 300 bn, most of it sold to foreign investors. The market could absorb it easily, and its over-reaction seemed (and was, I maintain) was excessive.

At that point I should have realized that a general panic and run on the bank was developing and I should have sold everything and buy Krugerrands.

Once the panicked run on the banks had developed, what should I have done? I did nothing.

Now the situation is that money is abundant again but people is paralysed by fear. Will confidence return as a flash flood (like in the crisis ten years ago) or be rebuilt gradually along the years (as in 1929)?