Sector betting in the gross profitability anomaly : a performance analysis and sector betting effects in the U.S equity market

Abstract

This thesis presents an analysis of quality investing; more specifically how gross profitability scaled by book assets can generate abnormal returns in the U.S market, as previously illustrated by Novy-Marx (2012). Our contribution is to explain these abnormal returns through a sector betting effect, which occurs while constructing portfolios on the gross profitability metric. In the first part of the analysis, we replicate Novy-Marx (2012). We utilize return- and accounting data from 1963 to 2016 and firstly investigate whether gross profitability predicts the cross section of returns. Our Fama-Macbeth regression show that gross profitability scaled by assets has roughly the same prediction power as book-to-market in terms of t-values. Furthermore, we construct a univariate portfolio sort using the same accounting metric. Our results yield a significant monthly alpha of 0.48% in the Fama French three-factor model. These results are very similar to what Novy-Marx (2012) finds. Next, we analyse the effect of sector betting within this portfolio sort. We complete three individual strategies to achieve this. Firstly, we demean gross profitability scaled by assets by the yearly sector average. This nearly eliminates under- or overweighting the respective sectors. Demeaning also leads to a reduction in portfolio performance in all our different performance measures. Secondly, we continue to complete the same portfolio sort within the sectors. Our results show that gross profitability to assets, does not generate abnormal returns in seven out of nine sectors. Thirdly, we complete a strategy where we are long low profitability stocks in high profitability sectors, and short high profitability stocks in low profitability sectors. This strategy yields a significant alpha of 0.48. Our results provide strong evidence that sector components of stock returns account for much of the individual gross profitability anomaly. Lastly, we combine gross profitability and book-to-market in two different strategies and test for performance. Our results suggests that controlling for profitability in a book-to-market sorted portfolio increases performance. Additionally, a combination portfolio of straight profitability and straight book-to-market displays less sector betting and reduced volatility