The impact of Company Leverage on the CAPM and Parametric Portfolio Construction
This thesis is divided in two chapters, in which I analyze the impact of company leverage on stock returns and optimal portfolios. In the first chapter, I derive a CAPM for levered equity from the unlevered one-factor CAPM (or asset CAPM), correcting for the presence of debt at both the individual company and market level. I show that the levered representation of the one-factor asset CAPM contains at least three factors conditioning its beta to the market excess return: leverage, debt maturity, and asset volatility of the rm relative to the market. The conditioning factors are correlated to the Fama-French SML and HML factors and give a theoretical explanation of why certain factor models work. The second chapter uses the conditioning factors of the first paper as the characteristics to run the parametric portfolio approach by Brandt, Santa-Clara and Valkanov and extend it to a continuous time setting. I 1) extract characteristics from a portfolio of credit-risky single stocks, 2) solve the dynamic programming problem for portfolio weights, and 3) show that accounting for credit risk in a multi-period framework achieves better Sharpe ratios than naive strategies such as EW and VW as well as results that are comparable to those of portfolios built on size, book-to-market, momentum and gross profitability.
Supervisor: Raman Uppal, EDHEC Business School
External reviewer: Michael Brandt, Duke University
Other committee member: Abraham Lioui, EDHEC Business School