ESSAYS ON EMPIRICAL ASSET PRICING

Abstract

In essay 1: This paper measures the time-varying provision of liquidity by buy-side customers (e.g., mutual funds and pension funds), relative to bond dealers, in corporate bond markets using a structural vector autoregression (SVAR) model. As indicated by my simple theory model, shocks to the relative willingness of customers and bond dealers to provide liquidity affect, in opposite directions, the choice of bond dealers between market-making (principal) and matchmaking (riskless principal) transactions. Motivated by this model, my SVAR empirically disentangles these shocks to customers versus bond dealers. My SVAR-derived patterns of these structural shocks provide fundamental insights into the mechanics in corporate bond markets following recent events, such as exposing the increased role of buy-side customers for liquidity provision after the many regulatory changes following the 2008 financial crisis. Furthermore, my empirical approach generates “factors” that provide an improved time-series asset-pricing model for yield spreads of corporate bonds of different credit ratings. In essay 2: We consider an approach to derive the conditional expectation of return quantities under the real-world probability measure, exploiting the form of the projected stochastic discount factor. Our treatment is formulaic in that the real-world expectation can be synthesized from the prices of the risk-free bond, the asset, and options on the asset. The method is free of distributional assumptions, and we use it to study empirical questions related to (i) conditional probability of a disaster and return upside and (ii) spanning hypothesis in the Treasury market. We examine empirical consistency and show that our theoretical treatment is relevant. In essay 3: Based on data until the mid 2000s, oil price changes were shown to predict international equity index returns with a negative predictive slope. Extending the sample to 2015, we document that this relationship has been reversed over the last ten years and therefore has not been stable over time. We then posit that oil price changes are still useful for forecasting equity returns once complemented with relevant information about oil supply and global economic activity. Using a structural VAR approach, we decompose oil price changes into oil supply shocks, global demand shocks, and oil-specific demand shocks. The hypothesis that oil supply shocks and oil-specific demand shocks (global demand shocks) predict equity returns with a negative (positive) slope is supported by the empirical evidence over the 1986--2015 period. The results are statistically and economically significant and do not appear to be consistent with time-varying risk premia.