Is "low volatility effect" that anomalous?

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
School of Business | Master's thesis
Degree programme
Abstract PURPOSE OF STUDY The empirical relation between expected stock returns and volatility is currently a matter of debate. The purpose of this study is to investigate whether the low volatility effect is present in the alternative data set that excludes penny stocks. I mainly focus on total volatility or the diagonal of the covariance matrix. Idiosyncratic volatility is studied to a lesser extent of testing the efficiency of the Fama French three factor portfolios. DATA AND METHODOLOGY My data set includes the annually reconstituted top 1000 stocks by market capitalization tracked by CRSP with at least 24 months of return history from 1983 to 2013. I form equally weighted and market capitalization weighted portfolios by ranking stocks based on 48-month total volatility and 12-month idiosyncratic volatility with respect to the Fama French three factor model into decile portfolios, from bottom (decile 1) to top (decile 10) on a monthly basis. The post-formation decile portfolio returns are controlled for the Fama French three-factor exposure and are measured by various risk and return metrics. RESULT I find that the equally weighted top decile portfolio sorted by total volatility statistically outperforms the equally weighted bottom decile portfolio by 1.01% (t-statistic 2.94) in monthly average return. Size and value effect cannot account for the average returns of the decile portfolios. Irrespective of whether volatility adequately captures risk, I find that the top decile is fundamentally riskier than the bottom decile by various measures. Component analysis shows that the top decile is dominated by stocks of firms in computer programing and semiconductor device related industries while the bottom decile is dominated by stocks of firms in electric service related industries. Keeping all else equal, changing from the naïve equal weighting to the market capitalization weighting distorts the positive relation between average returns and volatility. The market capitalization weighted portfolios underperform both the equally weighted counterpart portfolios and the market factor portfolio in monthly average returns. The study shows that the low volatility effect is not present in the research data set and that equal weighting scheme exposes the outperformance of the top decile portfolio. The peculiarity of stocks of firms in the dominating industries in the top decile potentially explains why volatility increases monotonically with average return but not with alpha. In addition, that idiosyncratic volatility predicts returns shows that the Fama French factor model does not adequately capture the systematic factor exposure of the decile portfolios.
Total volatility, idiosyncratic volatility, equally weighted, market capitalization weighted
Other note