Quantitative Style Investing
52 Pages Posted: 19 May 2016
Date Written: May 18, 2016
Abstract
I introduce a systematic portfolio choice solution that significantly beats a benchmark market portfolio by an average of 34.2% per year after transaction costs. The corresponding annual Sharpe ratio is 1.97 per year compared to 0.42, over 4.7 times the size of the benchmark. A more conservative sample that excludes micro cap stocks yields an annual Sharpe ratio 2.18 times the benchmark. I construct my solution by applying multivariable cross-sectional regressions of six key stock characteristics, to aggregate forecasting signals from multiple sources. I apply simple filtering techniques to reduce estimation and sampling error, use only information known at time t, and predict expected returns. I validate the procedure by achieving commensurate results as prior studies when forming portfolios from decile sorts. However, by sorting stocks by expected returns into more extreme portfolios, i.e. 25 and 50 portfolios, I am able to further enhance performance gains over existing works.
Keywords: active management, stock-picking, return predictability
JEL Classification: G11, G12, G17, C13
Suggested Citation: Suggested Citation