Betting Against Beta: With Empirical Analysis and Further Development of the Strategy

Abstract

This thesis investigates the empirical performance and practical applicability of the Betting Against Beta (BAB) strategy, which challenges the conventional Capital Asset Pricing Model (CAPM) assumption that higher systematic risk, as quantified by beta, is associated with higher expected returns. While theoretical frameworks propose a positive linear relationship between beta and return, empirical findings indicate a flatter Security Market Line,suggesting that low-beta assets tend to outperform on a risk-adjusted basis. Building on the foundational work of Frazzini and Pedersen (2014), we replicate the BAB strategy utilizing ex ante beta estimates, constructing long positions in low-beta portfolios and short positions in high-beta portfolios across five equity markets: Denmark, Germany (DAX), the United States (NASDAQ 100), the United Kingdom (FTSE 100), and Japan (Nikkei 225).To enhance the original strategy, we implement a stochastic dominance (SD) filter based on Kolokolova and Xu (2024), which facilitates the exclusion of assets that are subject to stochastic domination. This methodology has the potential to improve the portfolio’s Sharpe ratio and alpha. We rigorously assess the robustness of both the traditional and SD-enhanced BAB strategies by incorporating real-world frictions, such as transaction costs, borrowing fees, and short-selling constraints. Additionally, we conduct extensive sensitivity analyses by varying return horizons, dominance periods, and shrinkage intensities to evaluate how different parameter choices affect performance.Our findings confirm that BAB strategies consistently generate significant positive alpha,and that stochastic dominance filtering enhances performance in most settings. Furthermore,we demonstrate that BAB returns are particularly robust during bear markets, thereby highlighting the strategy’s potential as a defensive tool. Nonetheless, we emphasize that the implementation of BAB requires careful consideration of cost structures and access to leverage. Overall, the results underscore both the theoretical soundness and the practical challenges associated with exploiting the low-beta anomaly through BAB-based strategies.