A Measure of Price Pressures

Abstract

The Federal Reserve devotes significant resources to forecasting key economic variables such as real gross domestic product growth, employment, and inflation. The outlook for these variables also matters a great deal to businesses and financial market participants. The authors present a factor-augmented Bayesian vector autoregressive forecasting model that significantly outperforms both a benchmark random walk model and a pure time-series model. They then use these factors in an ordered probit model to develop the probability distribution over a 12-month horizon. One distribution assesses the probability that inflation will exceed 2.5 percent over the next year; they term this probability a price pressure measure. This price pressure measure would provide policymakers and markets with a quantitative assessment of the probability that average inflation over the next 12 months will be higher than the Fed’s long-term inflation target of 2 percent.