Market Segmentation and Non-Parametric Asset Pricing
I propose to examine market segmentation between equity and bond markets. Although under the no-arbitrage principle equity and bond markets should prove integrated, realworld frictions may induce some degree of market segmentation. To assess the extent of this segmentation, I will use non-parametric estimators of the stochastic discount factor (SDF). I propose to make four contributions in this work. First, the non-parametric methods I will use surmount several econometric limitations of previous such investigations. Second, I propose a novel machine learning-based SDF estimator. Third, I intend to examine time variation in the extent of segmentation between equity and bond markets, which previous work has not empirically tested. Fourth, I will use dual-asset-class SDF estimates to examine cross-asset-class trading signals, which have immediate practical applications.