Dissecting strategies with diversified factor exposures reveals some questionable components
Those new to factor investing may choose to play it safe by going for a multi-factor product. While the main factors can outperform in the long term, their performance is cyclical; getting exposure to multiple factors can create a smoother ride.
But the performance of a multi-factor product is only as good as its underlying parts. And, as a new note from FactorResearch explains, those parts will likely differ from one issuer to another.
“[T]he factor selection, definitions, and portfolio construction are discretionary choices of the ETF issuers, which have a meaningful impact on performance,” wrote Nicolas Rabener, managing director of FactorResearch.
Performing an analysis of multi-factor smart-beta ETFs in the US from 2014 to 2018, FactorResearch found exposure to common equity factors that are in line with academic and industry standards — as well as some exposures that do not align well with academic research on factor investing.
“The high exposure to the size, value, and low-volatility factors is to be expected … but the exposure to growth is unusual, as there is no theoretical foundation that supports positive excess returns from this factor,” Rabener said.
He noted that the assets under management for growth-linked single-factor ETF products eclipsed those for value last year, suggesting that investors are overlooking academic research that supports the case for value.
He added that the data also showed a lack of exposure to momentum and quality, the other factors supported by research. The underexposure to quality may be explained by the fact that quality can be defined in various ways, making measurements of exposure to the factor difficult. But since momentum follows highly homogeneous definitions across index providers, Rabener said, the lack of exposure is harder to account for.
To give an idea of how multi-factor ETFs perform compared to the broader stock market, he presented an analysis of an equally weighted index of US-listed multi-factor ETFs against the broad stock market from 2013 until 2018. The index performed comparable to the stock market until 2016, but underperformed after that.
“The lack of outperformance can likely be explained by management fees and factor exposure,” he said. Citing the ongoing price war between ETF issuers, he noted that fees for plain-vanilla equity ETFs have sunk to near-zero levels, with the first zero-fee ETFs just on the horizon. While factor-based ETFs are much cheaper than active mutual funds, they still charge between nine and 70 basis points in fees per year, negatively impacting performance compared to an almost zero-fee index.
An analysis of factor exposures showed high tilts toward size and moderate exposure to growth, value, and low-volatility factors. Growth stocks reportedly have done well since 2016, but those exposed to the low-volatility factor were essentially flat, while small and cheap stocks outperformed.
To confirm whether exposures to more classically supported factors would lead to better performance, he showed an analysis of a simple multi-factor portfolio based on value, size, momentum, quality, and low volatility. Aside from equalizing allocations with monthly rebalancing, the simulation included transaction costs of 10 basis points on stock level when creating the factors, a 0.5% annual management fee, and exposure only to stocks with market caps that exceeded US$1 billion.
“Although this portfolio did not generate higher returns than the market, it showed less underperformance that can likely be explained by a more diversified factor exposure,” he said.
