How to Assess Strategy ETFs

The first question to be asked of any strategy ETF: How does it make money? If you're buying a strategy in the hopes of earning market-beating returns, you better have a very good reason to justify your belief. The ETF providers sure don't. If they did, they'd quit the ETF industry, open up hedge funds, leverage their ideas to the hilt, and mint money. In other words, their comparative advantage is in selling products, not developing cutting-edge investing ideas. So what if the ETF came out with shiny brochures and eye-popping back-tested returns? The very existence of a supposedly market-beating strategy in an ETF should make you wary: The strategy is often 1) repackaged risk; 2) not very well supported with theory or data; or 3) overcrowded. You want a good, intuitive story as to why a strategy will continue to create excess profits, despite everyone knowing about it.

Where's the Theory?

Theory, which makes disprovable predictions, separates science from voodoo. Simple cause-and-effect stories don't cut it. Ideally, a theory should have been vetted by multiple independent and credible researchers. A theory should make predictions that later research confirms. You might find that level of rigor is too harsh for vetting an ETF. But if no hard thinking were required to pick a good strategy, we'd all be fabulously wealthy.

A good theory has several features. It's intuitive. It has mounds of supportive data. And it's persuasive, cohering with other features of the world that you're familiar with. Two of the most well-known properties of financial markets have ample theory and data backing them:

Momentum, or the tendency for prices to trend, has been found in virtually every market studied. It likely arises from cycles of investor underreaction and overreaction, and the tendency to herd. This story is well-supported by experimental data. Momentum persists because shorting it is very dangerous to arbitrageurs, who risk getting wiped out (value-manager GMO suffered massive withdrawals in the late 1990s betting against the tech bubble). In fact, rational investors may hop onto trends, strengthening them.

Value, or the tendency for stocks cheap by fundamental accounting metrics to outperform stocks expensive by such metrics, owes much to investors overreacting to bad news. However, value stocks also tend to exhibit negative momentum, meaning it may be profitable for arbitrageurs to short-sell value stocks in the short run. As with momentum, value can be dangerous to arbitrageurs. Traditional value indexes and dividend strategies all load up on the value factor, so good ETFs abound.

Risk Story?

In many cases, strategy ETFs repackage well-known "risk" factors (called such by convention, though no added risk may be involved) associated with higher returns, such as size, value, and momentum. One of the most popular alternate-weighting ETFs in the US, Guggenheim S&P 500 Equal Weight (RSP), equal-weights S&P 500 stocks and rebalances quarterly. It has handily beaten the S&P 500 since its 2003 inception. However, once you credit back its systematic mid-cap overweighting and tiny value tilt, it's actually generated zero outperformance over that period. You could've paid less for a mid-cap ETF and gotten the same performance. Historical simulations of equal-weighting covering multidecade spans and international stocks have shown zero value added after controlling for size and value exposure.

Where's the Data?

Investors are often wholly dependent on the inadequate materials put out by providers, a state of affairs possible because many are content to draw erroneous conclusions from just a few years of returns or low-quality research. Properly analyzing a strategy requires properly analyzing data. And a few years of returns isn't going to cut it. Not even a decade, most likely. The markets have a huge component of randomness. Coming to a well-reasoned, sound conclusion on a strategy often requires testing it on many decades of history, and not just in the U.S., in order to rule out chance.

For example, dividend strategies have been tested in many markets, to good effect. London Business School professors Elroy Dimson, Paul Marsh, and Mike Staunton have found above-average risk-adjusted returns for high-yield stocks with 110 years of U.K. stock market data, 84 years of U.S. data, and at least 19 more markets with at least 25 years of data. Dividend stocks have beaten the benchmarks in almost every market studied, and their returns aren't just from a few anomalous periods, but persist throughout the decades. To be fair, dividend strategies are often just repackaged value tilts, though they may be much purer than what you could get with a conventional value index.

Of course, even if a strategy passes this test, it's no slam dunk.

Overcrowding

The problem with exploiting a rigorously tested strategy is that many others have the same idea. Strategies can and often do become overcrowded. There's no good public data on when a strategy gets overcrowded, so you'll have to make reasoned guesses on whether a strategy is or isn't in vogue. One metric to assess a strategy's capacity is the liquidity of its underlying holdings. Another is whether every investor under the sun, especially institutions, is enthusiastic about it. Think long-only commodity futures indexes circa 2007 or managed futures now. However, if you've done your homework, you'll have a lot more conviction in the strategy and the fortitude to stick to it when everyone else is rushing for the exits.

What Passes This Test?

Aside from value and momentum, there are few other strategies that offer reasonable expectation of excess returns. Of the recent crop of strategy ETFs, low-volatility ones are the most promising.

Over long periods, low-volatility strategies have outperformed high-volatility strategies on risk-adjusted bases. As with the value and momentum effects, it's been found everywhere. Granted, low volatility overlaps a bit with value. It probably works due to benchmark-hugging by active managers everywhere, and lottery-seeking tendencies among investors. Low-volatility sectors, such as utilities, are trading at historically high valuations. Institutions are now keenly looking at managing their risks, whether through defensive equities or risk-weighted (risk parity) strategies. However, over the long run, the factors that have led to low volatility's outperformance (benchmark-hugging managers and lottery-seeking behavior) look set to continue.

A good sniff-test is whether you'd be OK sticking with a strategy for the next 20 years. If not, chances are the data haven't persuaded you.

An Ode to Skepticism

Analyzing the market and strategies is necessarily an exercise in playing the odds. No strategy will work all the time, and no way of picking strategies will be foolproof. A skeptical, scientific mindset will tilt the odds in your favor. When that rare good idea comes along, you'll have the conviction to stick with it during the inevitable (sometimes decade-long) bad spells. You'll need every advantage to get, because you're now a portfolio manager.

Samuel Lee is an ETF analyst with Morningstar