A1033
Title: Portfolio choice with unsystematic risk
Authors: Paolo Zaffaroni - Imperial College London (United Kingdom) [presenting]
Abstract: A normative theory is developed for constructing mean-variance (MVE) portfolios, including the global-minimum-variance (GMV) portfolio, when unsystematic risk is compensated in a no-arbitrage setting. Two inefficient portfolios are identified: a portfolio that depends only on systematic risk factors and a portfolio that depends only on unsystematic risk, which, when combined, give mean-variance efficient portfolios. It is shown that, as the number of assets increases, the unsystematic risk portfolio" dominates the systematic risk portfolio" both in terms of the magnitude of its weights and in terms of its Sharpe ratio. A penalized estimator is then derived under a no-arbitrage constraint that allows for the econometric identification of the unsystematic risk portfolio, leading to a shrinkage-type estimator and demonstrating that it is equivalent to the robust portfolio" desired by investors averse to model misspecification. Finally, it is demonstrated that the theoretical insights lead to an economically and statistically significant improvement in out-of-sample performance.
