New York – A recent paper, “Using Prediction Markets to Track Information Flows: Evidence from Google“, uses the accuracy of forecasts made in an internal prediction market to come to some interesting conclusions about market biases: in this market, where employees could place bets on the outcomes of internal corporate events at Google, the researchers find that traders were generally over-optimistic about outcomes, especially on days when Google’s stock appreciated, and under-priced extreme or “long-tail” outcomes. They also find that newly-hired employees were most likely to be over-optimistic, while grizzled veterans were generally better calibrated. Finally, they present evidence that physical proximity of market participants tended to cause greater correlation between their trades, and informal associations through social networks and interest groups are also a good predictor of correlated trading activity.
For a company like Google, internal information sharing is an extremely desirable thing. For a financial company, it would seem to us that too much information sharing and proximity may lead traders to bandwagon, analysts to lose their skepticism, and risk managers to adopt an over-optimistic groupthink mentality that blinds the company to long-tail risks. In fact, the hiring practices of many investment banks may look like the perfect recipe to generate all of the biases found by the Google researchers: hire people straight out of college in the middle of a bull market, give them options- or stock-based compensation, cluster them all in a giant trading floor in New York or suburban Connecticut, plug them into a constant stream of information that makes it hard to ignore short-term market moves, and have them work such long hours that their social lives tend to revolve around their colleagues. In Palo Alto or Mountain View, such a system may be optimal for generating motivation and innovation, but on Wall Street it seems reasonable to conclude that it would lead to over-optimistic firms that take on too much risk for a short-term return, while systematically under-pricing long-tail risk.
Surely, this isn’t a complete explanation for the excesses of a bull market, but it will be interesting to see if some financial institutions start to hire only veteran traders and analysts and place them in private offices as far as possible from the tumult of New York (Nebraska has been a good choice for some). To paraphrase Seinfeld, if every practice you have is wrong, then the opposite would have to be right.