Betting Against Betting Against Beta: The Implications of Non-standard Methods, Evidence from South Korea

We evaluate the Frazzini-Pedersen (2014) Betting-Against-Beta anomaly in the Korean stock market following Novy-Marx and Velikov’s (2022) critique. Our empirical findings are as follows. First, the rank-weighted portfolio used in the BAB strategy is similar to the equal-weighted portfolio in the sense that it consists of higher small and microcap stocks and lower large stocks compared with the value-weighted portfolio in the Korean stock market. Second, the beta-neutral portfolio constructed with the rank-weighted betas is similar to the one with the equal-weighted betas in the deleveraging procedure, and the beta-neutral strategy with the rank-weighted portfolio overstates the profitability of the BAB strategy. Third, the FF-3 risk-adjusted return of the BAB strategy becomes insignificant after netting out the transaction costs in the Korean stock market. Fourth, the betas estimated by the method Frazzini and Pedersen suggest are heavily biased in a predictable way and the bias generates significant errors in beta hedging in the Korean market. Fifth, we cannot find any evidence supporting the beta compression hypothesis and/or the borrowing constraint hypothesis in the Korean stock market if betas are estimated in the standard methods observed in the literature.