Posted by on Feb 15, 2016 in Articles & Advice, Blog, Columns, Featured |

Image Credit: Christophe Vorlet

By Jason Zweig |  Feb. 12, 2016  1:58 pm ET

Many investors are searching for a refuge in the latest stock-market storm.

Ever since the financial crisis of 2008-2009, fund companies have been pushing “smart-beta” funds. These are bundles of companies that have tended to be less risky and more lucrative than the market as a whole, because of cheaper stock prices, higher dividends or other factors.

But, argues a new study, their low risk in the past may lead to higher risk in the future. A report by Research Affiliates, a firm in Newport Beach, Calif., whose strategies are used to manage about $155 billion, says some smart-beta techniques have outperformed mainly by becoming dangerously expensive.

The smart-beta menu is mushrooming: So-called fundamental, equal-weight, high-dividend, low-volatility, high-momentum, low-beta or high-quality strategies are all mechanical ways of assembling bundles of companies other than by their total stock-market value, as conventional index funds do, or by piecemeal selection, as traditional stock pickers do.

Each is also backed by research purporting to show it has beaten the market. But can so many disparate strategies all work?

Morningstar, the investment-research firm, estimates that $72.9 billion in new money poured into such funds last year. By the end of 2015, they held $565 billion in assets — triple their level five years earlier.

“Much of this money is just chasing recent past performance,” says Robert Arnott, chief executive of Research Affiliates. The flood of new buyers has inflated prices, jacking up returns and attracting still more new buyers. Looking at a dozen smart-beta approaches, Research Affiliates found that nearly all were cheaper than the rest of the market a decade ago but now trade at premium prices. (Mr. Arnott and his firm have been proponents of several of these strategies and earn fees associated with them.)

Consider low-volatility stocks, which fluctuate less than average. From the 1970s through the early 2000s, they were about half as expensive as the overall market. Today, reckons Research Affiliates, they are about 20% more costly than other stocks. Such companies trade at an average of 2.8 to 2.9 times book value; the market as a whole is at 2.4 times that measure of net worth.

Take away this recent price surge, and low-volatility stocks offer virtually no extra return, says Mr. Arnott. “All the value they’ve added has come from getting more expensive,” he warns. Pay too much for good qualities like high dividends or low volatility, and future returns will be lower.

Across many fields, studies like those that make the case for smart beta can be contaminated by luck, bias or confirmation of the researchers’ own beliefs.

Research on 100 studies in psychology found in 2015 that more than 60% couldn’t be replicated. Similar results have been found in medicine and economics. Campbell Harvey, a professor at Duke University and president of the American Finance Association, estimates that at least half of all “discoveries” in investment research, and financial products based on them, are false. Because researchers can test a virtually unlimited number of possibilities, the odds are high that winning strategies are the result of luck alone.

By analyzing the effect of rising valuations on returns, the new Research Affiliates study seeks to isolate another element that can lead past results to seem better than they were.

Brian Nosek, a psychology professor at the University of Virginia and executive director of the Center for Open Science, a nonprofit seeking to improve research practices, has spent much of the last decade analyzing why so many studies don’t stand up over time.

Because researchers have an incentive to come up with results that are “positive and clean and novel,” he says, they often test a plethora of ideas, throwing out those that don’t appear to work and pursuing those that confirm their own hunches.

If the researchers test enough possibilities, they may find positive results by chance alone — and may fool themselves into believing that luck didn’t determine the outcomes.

Some smart-beta strategies may also fail to stand up over time. After years of outperformance, “equal-weight” funds, which hold the same amount of each stock in a market index, stumbled in 2015 and are lagging again this year.

Investors can take a hint from science, where new platforms are springing up to test whether research is robust. Scientists are increasingly “preregistering” their hypotheses in advance and publishing the results regardless of whether the final data confirm their theories.

Likewise, investment firms should report which techniques they are testing and how, as well as which end up failing and why.

Because computers and access to data are proliferating, the odds that strategies are based on “statistical flukes without theoretical support” is rising, warns Marcos López de Prado, a senior managing director at Guggenheim Partners, an investment firm in New York that manages about $240 billion. “For now, our best shot is to educate the public,” he says, “because not everyone in the industry is going to come clean.”

Mr. Arnott of Research Affiliates says investors should be particularly wary of strategies that have recently become richly priced — including many specializing in high-dividend, low-volatility and “high-quality,” or robustly profitable, stocks.

Be on your guard if a fund’s holdings have a higher ratio of price to earnings or book value than they did in the past, or than the market does now. You don’t want smart beta to leave you feeling stupid.


Source: The Wall Street Journal