In this study, we investigate whether we can identify a probit model with macroeconomic variables to forecast the monthly excess return signs of the U.S. oil and gas industry index by mining a big macroeconomic variable dataset designed by McCracken and Ng (2015). Three different information criteria and a Forward Sequential Variable Selection Algorithm are combined to select the most ”important” macroeconomic variables for prediction. A static probit model with 14 monthly macroeconomic predictors is found to predict one-month ahead excess returns of the U.S. oil and gas industry. We also show that the predictors in this model are related to the future performance of the S&P 500 index, the WTI crude oil price and the U.S. foreign exchange rate against the major currencies, which are the key risk factors of the oil and gas industry identified by previous studies. Active trading strategies based on the static probit model can generate higher Sharpe ratios than a buy-and-hold strategy. The forecasting ability of the identified model is found to be robust to different industry classification systems. The empirical results have important implications for investors and policy makers.
Liu, J., & Kemp, A. G. (2017). Forecasting the Sign of U.S. Oil and Gas Industry Stock Index Excess Returns Employing Macroeconomic Variables. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.2990880