We specify a multiple-equation model with equilibrium-correction terms, which connect inflation to the wage share and the functional income distribution, while not excluding a priori variables that are typically found in existing empirical U.S. Phillips curve models. We estimate the model equations using automatic variable selection with low Type-1 error probabilities on a sample with quarterly data that starts in the 1960s. Conditional on a relatively small number of location shift indicators, the price and wage equations have relatively constant parameters. The model’s explanatory power is shown by dynamic simulations. Applied to the COVID-19 period, the model shows that wage growth was important initially but that other factors later also became important, in particular the broad increase in international prices. Out-of-sample simulation shows how well the model forecasts inflation since early 2023.
Written by Gunnar Bårdsen and Ragnar Nymoen. Published in Oxford Bulletin of Economics and Statistics, 2025