Don’t believe everything you read in finance journals or buy what investment quants are selling you in their newfangled funds.
That is because the anomalies that help professors get tenure and novelties that attract investors, at the end of the day, fail to match the time-tested investment verities whose worth has been long and widely established.
So say Research Affiliates analysts Jason Hsu and Vitali Kalesnik, who pine for the days they were taught the asset pricing four-factor model in graduate school now that “quant shops…use an 81-factor model to build equity portfolios,” the duo write in their current newsletter.
Ironically, while the number of factors — investment anomalies that bring extra return premiums to those who can successfully capture them — have proliferated in academia and investing, Hsu and Kalesnik have rather expected genuine factors to shrink.
An example of that is the small-cap premium — one of the four (the others are market risk, value and momentum) in the classic four-factor model.
A mistake in how researchers treated data lent the appearance of validity to a small-cap premium, but subsequent research has found the anomaly does not exist internationally nor has it been observed in the U.S. since the early 1980s.
So where do quants get all these new, exotic factors that ostensibly add return and diversification benefits to your portfolios?
They stem from “the dangerous combination of cheap computing power and overzealous young finance Ph.D. students,” write Hsu and Kalesnik, who add:
“If one runs 10,000 back-tests, one is bound to discover a few incredible factors which generate huge Sharpe ratios.”
But this is mere data mining, and these new factors have tended not to withstand closer scrutiny. One very recent analysis looking at 600 — count 'em, 600 — factors from both academic and practitioner sources found that 49% had zero or negative premiums “out of sample.”
While in theory, then, an investor has the same chance as a coin toss to profit from these factors, in the real world, they write “net of transaction and management fees, tragically, you will likely do worse than a monkey throwing darts.”
More rigorous studies of investment anomalies have turned up an old familiar list:
Value, low volatility and momentum anomalies are “very significant” factors; market and illiquidity are “significant” factors; and other factors both new (such as “idiosyncratic volatility”) and classic (such as “quality stocks” as measured by return on earnings or profitability) are “insignificant.”
On reason ROE and profitability are not significant factors is that they do not work outside the U.S. As such, they “are more artifacts of U.S. data than real sources of return premia,” Hsu and Kalesnik write.
Besides looking for factors that apply universally, the Research Affiliates duo look for factors that can withstand tweaks in their definitions. For example, the value anomaly works whether it is measured by book-to-price or earnings-to-price, and momentum works whether one uses a 12-month look-back or a 6-month look-back.
Only after an investor has thus determined which factors are “for real” does the question of how to best capture those return premiums come to play — through low-cost smart beta or high-fee active management.
The Research Affiliates analysts say that the momentum and illiquidity premiums are best captured through skilled active managers because those strategies require sophisticated and costly trading that can’t be easily replicated in an index format.
“On the other hand,” they write, “value and low-beta strategies require no more than 10% and 20% annual turnover, respectively.”
In their conclusion, the smart-beta-oriented analysts throw down the investment gauntlet before today’s highfalutin quants and their financial exotica:
“We will gladly bet a simple blend of market, value, low-beta and momentum exposures against anyone’s optimized 81-factor portfolio,” they write.
Related on ThinkAdvisor: