【主题】 Portfolio Choice with Subset Combination of Characteristics
【报告人】吴轲（副教授，人民大学）
【时间】12月23日（周一）15:30-17:00
【地点】 经济学院楼701室
【摘要】We propose a novel portfolio strategy based on complete subset combination (CSC) of a large number of firm characteristics. The CSC strategy is easy to implement under mean variance utility and can be adapted to investors with more general utility functions. Monte Carlo simulations suggest it offers favorable efficiency and robustness tradeoff in portfolio decision. Empirical application to US individual stocks using 92 characteristics shows that CSC strategy achieves desirable certainty equivalent return and Sharpe ratio. It also outperforms alternative combination or characteristics selection strategies with commonly used machine learning tools. The portfolio value of CSC remains significant both statistically and economically net of transaction costs and after correction for publication bias of characteristics based anomalies.
