Inflation in the 1970s in the U.S.: Misspecification, Learning and Sunspots

In the late 1960s and into the 1970s, the United States experienced a burst of inflation the origins of which seemed hard to uncover. This paper advances the idea that the Fed simply got the model wrong. We assume that the true model of the economy is a variant of the standard New Keynesian model, but the Fed estimates its Phillips curve with a reduced-form equation, consistent with common practice. We show that a central bank that learns its model by recursive least squares would have arrived at the erroneous conclusion that its output-inflation trade-off was better than is truly the case. This statistical inference error induces a policy error that is serious enough to induce the sunspot equilibrium that Clarida, Gali and Gertler (1999) argue was true in the 1970s. The Volcker disinflation is then seen as a bold stroke to rule out sunspot equilibria and restore stability of inflation expectations. An implication of this is that the observed higher volatility of the economy in the 1970s, in comparison with the period after the Volcker disinflation, is really a manifestation of having erroneously assumed away sunspots which shows up as non-constant variances of fundamental shocks during the earlier period.