Evolving Post-World War II U.S. Inflation Dynamics

For postwar U.S. data, this paper uses Bayesian methods to account for the four sources of uncertainty in a random coefficients vector autoregression for inflation, unemployment, and an interest rate. We use the model to assemble evidence about the evolution of measures of the persistence of inflation, prospective long-horizon forecasts (means) of inflation and unemployment, statistics for testing an approximation to the natural-unemployment-rate hypothesis, and a version of the Taylor rule. We relate these measures to stories that interpret the conquest of U.S. inflation under Volcker and Greenspan as reflecting how the monetary policy authority came to learn an approximate version of the natural-unemployment-rate hypothesis. We study Taylor's warning that defects in that approximation may cause the monetary authority to forget the natural-rate hypothesis as the persistence of inflation attenuates.