Will your Factor Deliver? An Examination of Factor Robustness and Implementation Costs. Beck et al (2016)

The authors are curious to know which factors in the literature are most robust; so they searched the SSRN database and found the illiquidity, low beta, value, momentum, size, and quality factors to be most studied and relevant.  So they performed robustness tests for each of those factors over the 1967-2014 (in the US) and 1987-2014 (in international stock markets) period.

For the low beta factor, they found no statistically significant difference in returns between low beta and high beta stocks in the US.  They did however find that the sharpe ratios of low beta portfolios were higher than the sharpe ratios of high beta portfolios.  Internationally, there is no significant difference in returns between low and high beta stocks; although as was found in the US, the sharpe ratios of low beta portfolios were higher than those of high beta portfolios.

For the value factor, they found a statistically significant difference in returns between value and growth stocks in the US.  Also, they find that the sharpe ratios of value portfolios were higher than the sharpe ratios of growth portfolios.  Internationally, there is a significant difference in returns between value and growth stocks as well as sharpe ratios (except in the UK).

For the size factor, they found a statistically significant difference in returns between small and large cap stocks in the US.  They did however find that the sharpe ratios of small portfolios were not significantly higher than the sharpe ratios of large cap portfolios.  Internationally, there is no significant difference in returns between small and large cap stocks, nor in the sharpe ratios.

For the momentum factor, they found statistically significant difference in returns between small cap past winner and loser stocks in the US, but not in large cap.  They found the same scenario in regard to the sharpe ratio, where small cap past winners outperformed past losers, but large cap past winners did not outperform past losers.   Internationally, they found the same scenario where small cap past winners outperform past losers, but not large cap, on both a nominal and risk-adjusted basis.  Japan, however, is not a good place for momentum investing.

For the illiquidity factor, they found a statistically significant difference in returns between illiquid and liquid stocks in the US.  They also find that the sharpe ratios of illiquid portfolios were higher than the sharpe ratios of liquid portfolios.  Internationally, there is no significant difference in returns between illiquid and liquid stocks; the same conclusion is reached on a risk-adjusted basis, where illiquid portfolios' sharpe ratios are not statistically higher than those of liquid portfolios ex-US.

For the quality factor, they found no statistically significant difference in returns between quality and junk stocks in the US.  The same conclusion is reached on a risk-adjusted basis, where quality portfolios' sharpe ratios are not statistically higher than those of junk portfolios in the US.   Internationally, there is no significant difference in returns between quality and junk stocks; although  the sharpe ratios of quality portfolios were higher than those of junk portfolios in Europe ex-UK and Global.

Next the authors look at some more risk measures for each of the factors in long-only portfolios.  For large caps, they find low beta portfolios seem to have the least volatility, and momentum portfolios have the highest; small caps are more volatile than large caps, with the value factor being the most volatile.  The value factor for both large and small caps has the largest draw-down and period of underperformance (with the exception of quality, which has a 10.8 year period of underperformance).

Next the authors look at some more risk measures for each of the factors in zero-cost portfolios. They find the low beta and quality portfolios perform the worst in regard to draw-down, volatility and underperformance; this is likely caused from the short side underperforming the market in bull market periods. 

Finally, the authors look at which of the factors still provide excess returns after taking into account transaction costs.  They find that only the value, small cap low beta, and small cap illiquidity factors still add value after taking into account transaction costs.  All other factors erode value when taking into account transaction costs.

Beck, N., Hsu, J., Kalesnik, V., & Kostka, H. (2016). Will Your Factor Deliver? An Examination of Factor Robustness and Implementation Costs. Financial Analysts Journal, 72(5), 58–82.

The multifactor investing framework has become very popular in the indexing community. Both academic and practitioner researchers have documented hundreds of equity factors. But which of these factors are likely to profit investors once implemented? We find that many of the documented factors lack robustness. Size and quality, two of the more prominent factors, show weak robustness, whereas value, momentum, illiquidity, and low beta are more robust. Further examining implementation characteristics, we find that liquidity-demanding factors, such as illiquidity and momentum, are associated with significantly higher trading costs than are other factors. Investors may be better off accessing these factors through active management rather than indexation.