Comonotonicity and the Cross-Section of Expected Stock Returns

Dependence in financial market plays a key role in various areas, such as risk analytics and portfolio management, etc. So far, the asset-pricing literature has focused on correlation as the dependence measure, which is the linear dependency between stock prices. In this paper, we provide new evidence of the importance of the nonlinear characteristics in stock returns. 

We use the model-free measure for implied dependence (MFID) which was introduced in Dhaene et al. (2012). The theory of comonotonicity is used to construct a synthetic comonotonic stock market index, which corresponds with the theoretical, extreme situation of positive dependence. The MFID measures the dependence by comparing the option prices of the comonotonic index with the observed index option prices.

We exploit the asset-pricing implications of MFID through a portfolio-sorting strategy. We find stocks with high exposure to innovation in MFID deliver low expected returns relative to stocks with a low exposure cross-sectionally. These abnormal returns cannot be explained by linear correlation, implying existence of nonlinear dependence premium on the top of correlation premium, or other standard risk factors. Moreover, the dependence premium is robust to several empirical setups.