The End of Comfortable Keynesianism
Wednesday, October 12, 2011
I was delighted to learn yesterday that Tom Sargent and Chris Sims were awarded this year’s Nobel Prize in Economics. I had the good fortune to study macroeconomics briefly with Sargent and to be a faculty colleague of Sims’s for a number of years.
Tom Sargent’s graduate macro course was hard. I don’t mean hard in the “I hate equations and graphs make me break out in hives” kind of hard. I mean hard in the “I’m normally really good at this stuff and it all used to make sense, but what’s going on?” kind of hard. In the world of undergraduate economics, from whence my classmates and I had all emerged, there were Keynesian truisms that one could master. The state of the economy was set by the intersection of aggregate demand and aggregate supply curves. An adept policy maker could, by pulling the right levers and twisting the right knobs, shift these curves and thereby set the economy on the proper course. Implicitly, to such a policy mastermind, the economy was populated with individuals who acted in reliable, predictable ways.
This assumption that people behaved in reliable, predictable ways was often equivalent to assuming that people in the economy were stupid and could be repeatedly fooled.
That behavioral assumption was the point of attack for Sargent and other proponents of the “rational expectations” school. This assumption that people behaved in reliable, predictable ways was often equivalent to assuming that people in the economy were stupid and could be repeatedly fooled. If you wanted to spur the economy, just apply a burst of stimulus spending or pump up the money supply. When the economic agents in the economy—say, gullible store owners—saw customers coming through the door, flush with the new cash, they would conclude that happy days were here again and ask their suppliers to ramp up production; the economy would then spring to life. Those store owners wouldn’t stop to ask whether the stimulus would be paid for by higher future taxes, or whether the newly printed money would cause inflation, thereby undercutting its value. They would just suffer the rude surprises later on.
In rational expectations models, the people are smarter; they know what’s going on. If you offer them goodies today, paid for by taxes tomorrow, they look at both sides of the ledger, not just one. To the dismay of graduate students, this makes the math much harder. It also undercuts some of the old verities. At a dinner with Sims, when I was just coming out of graduate school, I made some mention of aggregate demand. He asserted that there was no such thing. This was deeply unsettling, even after my exposure to teachers like Sargent and John Taylor. More importantly, the rational expectations approach implies that the challenges are much greater for the economic policy maker, who now needs to worry about savvy economic counterparties who understand the game.
Why should we set aside our old, nice, simple economic models in favor of ornate new ones? Only because the old ones were not working very well. It was no coincidence that the rational expectations approach to macroeconomics emerged amid the stagflation of the 1970s. Under the old, comfortable Keynesian reasoning, there was not supposed to be stagflation; the policy maker was supposed to be able to choose between high inflation and low unemployment, or low inflation and high unemployment. The unpalatable combination of high inflation and high unemployment wasn’t supposed to be on the menu.
Why should we set aside our old, nice, simple economic models in favor of ornate new ones?
In the wake of the global financial crisis, these Keynesian qualms were forgotten or set aside. Policy makers feverishly yanked levers and spun knobs. Financial journalists weighed in on the proper calibration of stimulus to adjust aggregate demand. And the only plausible explanation for why Keynesian stimulus did not revive the economy was that we did not do enough of it. [See John Cogan and John Taylor for an alternative view].
In a snarky column this week, the Financial Times applauds Sargent and Sims for ultimately seeing the error of their earlier ways:
Prof Sargent now recognises that people get confused from time to time, while Prof Sims models the fact that we all can only take in so much information while thinking about the future.
Just because economists get it wrong doesn’t mean they should stop trying.
This presumes that the most striking hubris of recent years lay with those who credited the public’s intelligence, not with those who presumed the public would react like sheep. I never interpreted the teachings of Sargent or Sims to say that economic actors were impervious to error. Rather, the lesson I took was that counting on those actors to err systematically and repeatedly could lead us seriously astray in our policies.
Philip I. Levy is a resident scholar at the American Enterprise Institute.
FURTHER READING: Levy also writes “Free Trade’s Rude Awakening,” “Obama’s Great Buffett Confusion,” “How Do Jobs Numbers Work?” and “Will Exports Save the U.S. Economy?”
Image by Rob Green | Bergman Group