Getting back to academically validated factors
BY Yaelle Gang | March 12, 2019
With factor investing reaching high levels of popularity, a new paper by Scientific Beta is cautioning that a proliferation of new factors that aren’t academically validated could lead to unintended exposures and misunderstandings about risk exposures.
“Investors can choose to rely on standard factors that have survived the scrutiny of countless empirical studies and have been independently replicated and validated,” noted the paper. “Alternatively, they can choose to forego this free due diligence and take on the risk of selecting a provider-specific factor definition, which is somewhat similar to taking on the risk of selecting an active manager.”
Commonly used factors include size, momentum, value, profitability and investment, the paper noted.
Yet many providers are using other proprietary factors, which can come along with a host of issues, noted the paper. For example, commercial factors are based on complex composite definitions, providing flexibility that can be used to seek out factors with the highest performance in a dataset. This can lead to the rise of spurious factors, which can work well in smart datasets but are useless in reality.
“I think the key issue is really that investors should be concerned about the risk that analysis can come up with the wrong factors,” says Felix Goltz, head of applied research at the EDHEC-Risk Institute and research director at Scientific Beta, noting there are limitations to what you can conclude from analyzing data.
“You have to be very careful when you conclude something from a small dataset in terms of really how meaningful that’s going to be. So there’s a big risk of ending up with the wrong factors when your analysis is limited to one investment analysis and one dataset.”
Factor investing is experiencing a lot of product innovation, says Goltz. “We just looked closely and were actually surprised when taking a closer look that there was a clear mismatch between what the industry often refers to as factor investing and what actually the evidence says about factors.”
The novelty of these products may attract investors, he adds, but they should be focusing on reliability.
If providers say their factors are grounded in academic research, they should be using and validating the factor definitions across different independent studies and using a risk-based explanation to support the factors’ existence, the paper said.
“I think the important recommendation for investors is really to ask questions about the robustness behind the research that led to these factors and really to ask questions about external validation of different factors,” says Goltz.
There will always be debate, he adds, but at least investors can seek out what’s robust and has been replicated independently. “It could be the case that the innovative factor, the enhanced factor, is truly an enhancement, but maybe you should just wait until that has been confirmed by others and that has been also shown to persist in other datasets because otherwise you just have a big risk that you’re relying on false results.”