David Brooks, homogenous viewpoints and systemic risk
April 6, 2009
In Friday’s New York Times Op-Ed, sociologist-columnist David Brooks has grouped existing accounts of the credit crisis into two interesting broad narratives. The greed narrative and the stupidity narrative. Brooks’ piece highlights an important danger that I have also found in my research — homogeneity.
But first — Brooks’ two takes on the crisis. The first, which broadly maps on to the dominant view in orthodox economics, argues that the source of the problems was agency conflicts. This, I should note, is by now the dominant view of the Internet bubble. In my view, it also misses the actual complexity of valuing a uncertain technology such as e-commerce.
The second, “stupidity” narrative, is a lot more interesting. Brooks writes:
The second … revolves around ignorance and uncertainty. The primary problem is not the greed of a giant oligarchy. It’s that overconfident bankers didn’t know what they were doing. They thought they had these sophisticated tools to reduce risk. But when big events — like the rise of China — fundamentally altered the world economy, their tools were worse than useless.
But why were bankers overconfident? Brooks places the emphasis on technology and the homogenization of perspectives:
To me, the most interesting factor is the way instant communications lead to unconscious conformity. You’d think that with thousands of ideas flowing at light speed around the world, you’d get a diversity of viewpoints and expectations that would balance one another out. Instead, global communications seem to have led people in the financial subculture to adopt homogenous viewpoints. They made the same one-way bets at the same time.
What to make of it? I am not very persuaded by the “information availability” argument. I see people ignore information every day, particularly when it does not support their preexisting beliefs.
But I think that Brooks is on to something with his emphasis on homogeneity. In a recent article, David Stark and I concur with this concern. We also offer an alternative view on how homogeneity arises in the financial market.
As we see it, homogeneity the direct outcome of the growth in financial models brought about by the quantitative revolution. Models, as Donald MacKenzie made clear, are not simply used for forecasting but also for stress-testing the trades. This is done by translating market prices into the estimates that rival traders (“the market”) are making. Interestingly, the translation is also done with the use of models, but models in reverse.
But this reflexivity, we claim, can lead to even greater systemic risks. How? The rise of models offers traders a way to align their bets. If, by any chance, every arbitrageur turns out to be wrong on a particular trade, the positions of their rivals will provide them with a false sense of confidence. Such greater confidence will then lead them to double up their exposure, causing potentially catastrophic losses.
Why is this interesting? First, it explains the concentration of arbitrage losses around a few cases of merger cancellation. Second, it explains how traders that do not see each other experience similar losses: they were both looking at the same spread plots. Third, it provides an explanation for overexposure that does not require any form of psychologically extreme disposition, like “greed.” Fourth, it points to an interesting paradox of the quantitative revolution: the more precise, careful and reflexive traders become –thanks to financial models– the greater the potential for systemic risk.