An Anatomy of Characteristics in Dynamic Trading
Finance & Accounting
|Speaker:||Hong Zhang, Tsinghua University|
|Date: ||Wednesday 17 February 2021|
|Time: ||13:00 - 14:00|
|Location: ||Online via Zoom (link available from J.Lashkor@exeter.ac.uk)|
We propose testing the joint and marginal power of characteristics in predicting returns via their contribution to a dynamic trading algorithm (e.g., Kyle 1985). Applying optimal trading portfolios to a sample of 147 characteristics over 1980-2019 confirms that characteristics can jointly deliver significant out-of-sample returns (23% annually). However, most characteristics (88%) fail to supply independent information to dynamic trading — removing these subsumed characteristics can further enhance the optimal portfolio returns to 33.7%. Our analysis further reveals a leading role played by Fama-French-Carhart factors as informative characteristics.