“NATURE DOESN’T RUN VERY GOOD EXPERIMENTS.”

“NATURE DOESN’T RUN VERY GOOD EXPERIMENTS.” Back in the sixties when I was studying economics, there was enormous optimism about what computers would enable economists to learn by modeling the economy. Professor John Meyer brought a dose of realism to his students by saying that, “Nature doesn’t run very good experiments.” He explained that when an economic variable increased, usually a lot of other variables increased at the same time so that it was difficult to disentangle which variable or variables were the driving forces. Computers can do remarkable things in economics, including often successful trading of securities, but the problem that Meyer identified so early has been a tough one. Economists have made some progress. One reason that Steven Leavitt (of Freakonomics fame) is able to discover so many interesting empirical findings is that he is gifted at finding variables that move independently of other variables so that he is able to disentangle cause and effect.