Deciphering the Neo-Fisherian Effect

We propose a novel empirical strategy to detect the neo-Fisherian effect, which typically refers to the short-run increase in inflation associated with a long-lasting increase in the nominal interest rate. Our procedure is agnostic about the nature and persistence of the shock that gives rise to such an effect. It only requires this shock to explain the largest fraction of the conditional variance of the nominal interest rate at a long but finite horizon, while being orthogonal to a standard monetary policy shock. Using simulated data, we show that our methodology is remarkably reliable both asymptotically and in short samples. When we apply it to U.S. data, we find strong evidence supporting the existence of a neo-Fisherian effect, and show that the underlying shock accounts for the bulk of the variability of the nominal interest rate and inflation at essentially any forecasting horizon, while explaining a non-negligible fraction of output fluctuations at business-cycle frequencies. Bayesian analysis of competing New Keynesian models indicates that the data favor the interpretation of the identified shock as a liquidity preference shock rather than an inflation target shock as is usually suggested in the literature.