Production-based asset pricing: A cross-industry study

This paper empirically tests the production-based asset pricing model (PCAPM) using industry-level data. Specifically, I examine the pricing relationship between physical investment returns and equity returns by testing the hypothesis (spanning assumption) that the payoff space of physical investment spans the payoff space of financial securities. Instead of following the traditional approach of testing a linear physical investment factor pricing model, I propose a new testing procedure based on entropic principles and no-arbitrage constraints. The proposed nonparametric method recovers and compares the state price densities for physical investment returns and equity returns. The new testing procedure has two main advantages over the traditional approach. First, the new method is consistent with no-arbitrage conditions while the traditional approach leads to a stochastic discount factor that may take negative values. Second, the new method alleviates the joint hypothesis test problem suffered by the traditional approach. I apply the proposed procedure to the following six industries: mining, construction, manufacturing, transportation, communication, and public utilities. The empirical results show that the state price density recovered from physical investment returns is able to correctly price the corresponding equity returns. This provides supporting evidence that the spanning assumption holds at the cross-industry level. Furthermore, this study highlights the fact that physical capital investment conveys important information for pricing financial securities. Hence, to explain both the time series and the cross-sectional variations in equity returns, it is useful to develop models that explicitly address the impact of key production characteristics on financial asset pricing.