Forecasting US inflation using Markov dimension switching

This study considers Bayesian variable selection in the Phillips curve context by using the Bernoulli approach of Korobilis (2013a). The Bernoulli model, however, is unable to account for model change over time, which is important if the set of relevant predictors changes over time. To tackle this problem, this paper extends the Bernoulli model by introducing a novel modeling approach called Markov Dimension Switching (MDS). MDS allows the set of predictors to change over time. The MDS and Bernoulli model reveal that the unemployment rate, the Treasury bill rate and the number of newly built houses are the most important variables in the generalized Phillips curve. Furthermore, these three predictors exhibit a sizeable degree of time variation for which the Bernoulli approach is not able to account, stressing the importance and benefit of the MDS approach. In a forecasting exercise the MDS model compares favorably to the Bernoulli model for one quarter and one year ahead inflation. In addition, it turns out that the performance of MDS model forecasting is competitive in comparison with other models found to be useful in the inflation forecasting literature.