Venue / Hotel
We investigate different measures and components of inflation and their roles in predicting market excess returns. We dissect consumer price index (CPI) and personal consumption expenditures price index (PCEPI) by energy, food, and core, and examine corresponding inflation rates. We apply neural network/deep learning and investigate the predictive performance. We find that inflation rate based on the CPI’s food component, along with stock variance and scaled market price, has 5% monthly out-of-sample R2 in predicting S&P 500 excess returns. For a single-factor model, the inflation rate on CPI’s energy component alone has more predictive power than the core inflation rates.