Cointegration, information transmission, and the lead‐lag effect between industry portfolios and the stock market

Abstract

This paper shows that lagged information transmission between industry portfolio and market prices entails cointegration. We analyze monthly industry portfolios in the US market for the period 1963-2015. We find cointegration between six industry portfolio and market prices. We show that the equilibrium error, the long-term common factor between industry portfolio and market cumulative returns, has strong predictive power for excess industry portfolio returns. In line with gradual information diffusion across connected industries, the equilibrium error proxies for changes in the investment opportunity set that lead to industry return predictability by informed investors. Forecasting models including the equilibrium error have superior forecasting performance relative to models without it, illustrating the importance of cointegration between the industry portfolio and market prices. Overall, our findings have important implications for investment and risk-management decisions, since the out-of-sample explanatory power of the equilibrium error is economically meaningful for making optimal portfolio allocations.