Characteristics-based Portfolio Choice with Leverage Constraints

We show that the introduction of a leverage constraint improves the practical implementation of characteristics-based portfolios. The addition of the constraint leads to significantly lower transaction costs, to a reduction of negative portfolio weights, and to a decrease in volatility and misspecification risk. Furthermore, it allows investors to implement any desired level of leverage. In this study, we include 12 characteristics, thereby extending the classical size, book-to-market and momentum paradigm. We report several key indicators such as the proportion of negative weights, Sharpe ratio, volatility, transaction costs, the transaction cost-adjusted certainty equivalent returns, and the Herfindahl–Hirschman index. Analyzing the sensitivity of these key indicators to the choice of multiple combinations of the 12 characteristics, to risk aversion, and to estimation sample size, we show that constrained policies are much less sensitive to these parameters than their unconstrained counterparts. Finally, for quadratic utility, we derive a semi-closed analytical form for the portfolio weights. Overall, we provide a comprehensive extension of characteristics-based portfolio choice and contribute to a better understanding and implementation of the allocation process.