Disentangling the Low-Risk Effect - The Role of Volatility and Correlation in the Low-Risk Anomaly

Channels of the Low-Risk Effect:  The low beta effect can be separated into distinct low volatility and low correlation components. I implement low volatility (VOL) and low correlation (COR) factors using standard construction methodology and find in developed markets across four global regions that the low correlation premium can be explained by exposure to momentum stocks, while low volatility stocks resemble profitable firms with conservative investments. Additionally, I show that low volatility firms are associated with low dispersion of earnings forecasts, suggesting that the underperformance of low profitability and high investment firms results, at least partially, from diverging earnings expectations and arbitrage asymmetry.

Return, Risk, and Exposure of Low-Risk Portfolios: Low Volatility (LV), Global Minimum Variance (GMV), and Most Diversified (MD) portfolios provide a simple, cost-effective way to access low-risk premiums in global equity markets. The LV portfolio primarily benefits from the increased individual performance of low-volatility stocks, while the MD portfolio profits from enhanced diversification of low-correlation stocks, and the GMV portfolio balances both aspects. When carefully implemented, all three achieve comparable performance in large stocks, while the MD portfolio underperforms in small stocks, where low volatility is more prevalent than low correlation. I further compare their out-of-sample performance to ex-ante, net-of-cost optimal factor portfolios in US, European, and Asia-Pacific stock markets, finding that low-risk portfolios attain comparable out-of-sample performance, net of costs, as Fama and French five-factor plus momentum portfolios. Thus, low-risk portfolios present a viable alternative to long/short factor portfolios, offering simplicity, cost-efficiency, ease of implementation, and eliminating the need for leverage or shorting, while maintaining similar performance.