I was sent a review copy of A Crisis of Beliefs, by Nicola Gennaioli and Andrei Shleifer (henceforth GS). They say that the financial crisis of 2008 illustrates a theory of expectations formation in which market participants both place too much weight on recent news and in some circumstances ignore tail risk.
We know from Tetlock, whose name does not appear in the index, that a good forecaster puts a lot of weight on baseline information–characteristics that are more universal and permanent. Inefficient forecasters instead tend to focus on information that is more recent and local. GS argue that financial market participants are inefficient forecasters.
So far, what I like about the book:
1. The writing is clear.
2. Years ago, I contrasted two classes of theories of the 2008 financial crisis. One I called “moral failure” and the other I called “cognitive failure.” The theory that GS builds falls within that latter class, which is the one on which I would place more weight.
3. GS take seriously data that comes from surveys of the expectations of market participants. They are not afraid to find fault with the rational expectations hypothesis.
What I don’t like:
GS use standard economic modeling methodology, as opposed to Bookstaber’s agent-based modeling. See my review of The End of Theory. In particular, I think that institutional details are important, and Bookstaber’s rich depiction of different classes of market participants is better than a standard mathematical model. Also, I don’t like the idea of collapsing divergent expectations into a single representative agent. Getting away from the representative-agent model is a point in favor of Frydman and Goldberg. Note that Bookstaber, Frydman, and Goldberg do not appear in the index, either.
I have found that high prices in coastal California led to mass immigration. About 2% of homeowners were selling and moving out of the cities annually at the peak – a tremendous amount of selling pressure. That immigration to less expensive markets was the chief source of buying pressure in cities like Phoenix.
I’m curious. Does their survey data include those who were selling and moving away, or does it only include surveys of buyers and owners?
I would add the thought that “cognitive failure” can provide the temptation and fuel for what ultimately becomes a “moral failure”.
The focus on knowledge with shallow roots in the present rather than deeper ones in the past, the unreasonable confidence that “of course it is going to work” the follows from that contracted understanding, and the desire not to miss out on something in the (speculative) future, while all aspects of cognitive failure, are a recipe for recklessness, first allowing, and then participating in, all manner of little frauds and fast dealing, and general lack of good faith which are the ground of moral failure.
Smart investments become dumb investments if too much money chases them. It wasn’t a bad idea to start companies to build the internet, but buying those companies at early 2000 valuations was a terrible idea. It’s not a bad idea to lend money for home mortgages, but lending huge sums for crappy mortgages was a bad idea. I’d say the problem is less moral or cognitive than systemic – as an investment class gains momentum, it simultaneously becomes too short-term profitable for any public company to abstain from and too overvalued to be a good investment in the medium term.
If I were to try to fix the problem, I’d try to add mechanisms that let bearish opinions speak. More short-selling, for one.
You can short housing. Rent.
Also, during the bubble there was a significant portion of owners moving away from expensive cities. Maybe 1/3 of home sales were from owners moving to cheaper cities, which also is a kind of shorting.
We do have some mechanisms for letting bearish opinions into the markets, but they don’t seem to be strong enough to prevent bubbles (or “momentum trading”, if you’re the sort of person who calls bad students “exceptional”).
On the theories about the Financial Crash, I’m surprised that you don’t place more weight on “AND”, and the intersection of both. The moral failures are clear up thru the 2006 housing boom and bust: 0% downpayments; special buying loans of interest only for a few years; failure to confirm income – plus dishonesty by buyers at claiming higher income than was true (asking and getting “Liar Loans”).
Lying speculators are more of a moral problem than a cognitive one. Yet the cognitive failure to simulate a big reduction in house prices was huge. The failure to read the actual terms of the MBS / CDO agreements, terms that autistic Dr. Michael Burry was actually willing to read, was both types of failure. The lawyers and rocket science analysts had jobs to analyze what the terms meant, and to analyze the risk. Their risk analysis was a big failure.
One of the confusing things is that the house buying bubble ended in 2006, but the Financial Crisis didn’t really get going until 2008. The near-total failure to update the risk models is almost all cognitive failure, but their models didn’t take into account speculator liars who were quite willing to walk away from their last 0% down, 106% loan value (to cover closing costs) non-owner occupied home loans. The walking away caused problems, but those problems started small.
Liar loans are a moral problem, like shoplifting. Business models that don’t take them into account are cognitively inferior, like a retail shop whose profit depends completely on no shoplifting (and almost no security).
There was a little scare in 2015. I think the lending models have added enough security to keep the problem manageable. They didn’t have enough in 2002-2008.
https://www.bloomberg.com/news/articles/2015-09-08/liar-loans-redux-they-re-back-and-sneaking-into-aaa-rated-bonds