Capitalism's Confidence Games

Capitalism's Confidence Games

4:18 PM Monday August 23, 2010
by Chris Meyer & Julia Kirby  |

Our work on the evolution of global capitalism rests on a core idea: systems with feedback and selection evolve. That's a foundational idea in evolutionary biology — it's the reason that Darwin's finches had different beaks — but we note that it is also true in political economy. In fact, socially constructed systems (like capitalism) evolve all the faster because feedback can reshape notions more readily than noses........

Yet it's also true that human cognition can prevent important feedback from being processed. In general, humans are adept at pattern recognition, but our strength at converting common observation into reasonable prediction has weaknesses. For one thing, as Chris was taught in business school in the 1970s, people in general underestimate the likelihood of the unlikely. Rarities that happen one time in a hundred (a corporate bankruptcy, say, or a flood, in a five-year period) are estimated to happen once in a thousand cases. As a result of this tendency — surprise! — we're constantly being surprised.

In a New York Times column called "Often Wrong, But Never in Doubt", University of Chicago behavioral economist Richard Thaler focuses on one unfortunate result of the persistent belief that the world is more predictable than it is: he notes that it leads to a dangerous overconfidence among executives. Citing a survey of CFOs, he remarks: "It may be neither troubling nor surprising that CFOs can't accurately predict the stock market's path ... What is troubling ... is that as a group, many of these executives apparently don't realize that they lack forecasting ability."

Thaler then turns to papers of UCLA's Richard Roll, who has investigated the role of CEO hubris in inflating acquisition premiums. It is well established that in most cases when one company has acquired another, the returns on that investment do not wind up justifying the purchase price. Why would it be the case that the acquirer's management would tend to expect substantially higher performance than they can ultimately realize? Roll's observation is that acquirers "have typically done very well in the recent past, leading their CEOs to the mistaken belief that their success can be replicated in takeover targets once they are in charge of them."

Worst of all, Thaler concludes that such errors in interpreting feedback may be endemic to the system:

We shouldn't expect that the competition to become a top manager will weed out overconfidence. In fact, the competition may tend to select overconfident people. One route to the corner office is to combined overconfidence with luck, which can be hard to distinguish from skill.

In some ways, this reminds us of a scam described in Nasim Taleb's Fooled by Randomness: Send a spam note to a million people with a stock tip. The stock will either go up or down. If it goes up, send another tip. Repeat five times, and, assuming every stock has a 50% chance of going up, you will have 3,125 people who think you're a genius. Ask them all for money. We know, as a group, they will underestimate the possibility that you could have made those predictions at random. You can take that to the bank.

Of course, CEOs don't begin with intent to defraud their boards (for the most part). The parallel is in the mechanism that ends up misallocating funds. People, as Taleb argued, want to believe the world is less random than it is, but we set up some of our systems so that random results are misinterpreted as information.

So while we're quite certain that feedback mechanisms are the place to identify forces for change in our economy, sometimes the results are perverse. Thaler's concern is an important one. Human systems — capitalism among them — can only evolve into better versions of themselves to the extent that we humans are able to learn from our mistakes
.