Monday, December 20, 2010

Russ Roberts Interviews Joe Nocera

Russ Roberts interviews Joe Nocera about his book "All the Devils are Here". I thought this book was OK, but it spent a lot of time discussing the background of major players, who all seemed to either grow up 'on the wrong side of the tracks' or went to Groton. Am I the only person who had middle class parents, average in most ways? Perhaps that's my exceptionalism--extreme conventionality!

One important take-away seem to be, send all the econ bloggers a free copy of your book, because I think it was uniformly rated well in praise surrounding extended quotes (see Tabarrok here, Kling here).I can tell you that there is little upside to slamming a book on your blog, but you might receive a nice note from the publisher, or some nice links from the popular author if you have nice things to say. Book reviews are like stock analyst recommendations pre Sarbanes-Oxley.

Here's a snippet Nocera and Russel's take-away of they big mover in the 2008 crisis:
My interpretation of that has to do with modeling. I think one of the things that happened over the course of the last 20 years is the core idea that you could model away risk. One of the things that was striking in [Greenspan's] apology was when he said: Several people have won Nobel Prizes for the work that has turned out to be flawed.

In other words, 'models' were the main problem. We shouldn't believe so much in models. This is not insightful, because it leads to infinite regress. If you believe models are imprecise, and so add some uncertainty around them, this too is a model. Sure, you can keep compounding your uncertainty, but eventually this leads to avoid lending altogether, which is clearly suboptimal. But then Nocera mentions this pertaining to MBIA:

One of the things we said to them, we asked them about, was their models. They said: Actually, our models were pretty accurate. If we plugged in a 20% decline in housing prices, they showed the world blowing up. But we just assumed that that could never possibly happen! Famous paper by the Orszag brothers and Joseph Stiglitz, who was a Nobel Laureate, showing how remote it is that Fannie or Freddie would ever go bankrupt. Absolutely true.

So, the problem was not essentially models, hubris, or securitization, but rather this singular assumption: that housing prices, nationally, would not fall significantly year-over-year. It is much more parsimonious explanation; it explains more with less.

4 comments:

Anonymous said...

People do learn from experience, unfortunately the learning is too specific. The lesson we've collectively learned from the housing bubble is "housing bubbles can occur" - not "bubbles of any kind can occur".

Robert Johnson said...

Of course this means the 'rational actor' hypothesis is bunk. Lot's of people have said so, but this crystallizes down to one particular irrational belief.

Why would people believe that housing prices couldn't drop precipitously? Are prices really hard to determine objectively? No stable long term relationship exists between wages and housing prices, for example?

And if THAT'S true, then how DO you make rational lending and investing decisions? I'm really asking.

Eric Falkenstein said...

Robert: People believed it because prices had not, historically, ever fallen nationally in nominal terms since data had been collected. That's not a good reason, but I think it is the main reason. Now, lots of AAA stuff is based on this kind of reasoning, and historically,AAA stuff has a very low default rate, so it's not insane. With hindsight, clearly the underlying risk of the obligors changed materially ,with lots of no-money-down, no doc loans, but that wasn't obvious if you weren't in the business. Moody's actually stated that it wasn't their job to look at that information, a statement they regretted.

A lot of lending crises were based on bad assumptions about collateral--the internet, the commercial real-estate circa 1990, oil lending in the 1980's. I know old regulators who say this is perennial, so perhaps there's a large cost to assuming otherwise in the boom

Patrick R. Sullivan said...

I love the bit at the end where Nocera laments that housing is entirely dependent on the federal govt., and that needs to change. While he just wrote a polemic against the Republicans on the FCIC for essentially pointing out the same thing.