Nerds on Wall Street: Math, Machines, and Wired Markets

2.2. Hits and Misses: Rational and Irrational Technology Exuberance

Peter Bernstein's book Capital Ideas (Free Press, 1993) tells the story of Bill Sharpe, who wandered Wall Street looking for enough computer time to run a simple capital asset pricing model (CAPM) portfolio optimization, while being regarded as something of a crackpot for doing so. Now these computations are routine, though not without problems arising from sensitivity to errors and noise. (These problems, in turn, are being addressed by still more computation, using resampling and other computer-intensive methods.) Yesterday's crackpot, of course, has become today's visionary Nobel laureate, as proven by the sheer rise in calculating power. It is instructive to look back at technological ideas in finance that have succeeded, some spectacularly so, and those that have been marginalized. This approach avoids the onerous task of actually predicting what will work in the future.

Some crackpots are just crackpots. The past 30 years have also seen their share of wacky ideas that remain wacky. In the 1980s, expert systems had produced some impressive results in diagnosing diseases, designing networks, and running factories. Business magazines called each one "the breakthrough of the century." Of course, Wall Street wanted in, thinking that thinking machines would do all of the thinking! A cover of Wall Street Computer Review magazine illustrated an extreme example of this expectation. It had Socrates standing on the ...