An ETF That Buys High and Sells Higher

Depending on your perspective, Russell's new momentum exchange-traded funds are either useful or utterly insane. Every month, like clockwork, the ETFs buy the hottest, best-performing stocks and dump laggards. It sounds like a surefire recipe for buying high and selling low. Yet a naive momentum strategy has generated market-beating returns over long time periods in almost every stock market studied.

The ETFs come in two flavors: Russell 1000 High Momentum ETF HMTM and Russell 2000 High Momentum ETF SHMO. Just as their names imply, they screen for high momentum stocks in either the large-cap Russell 1000 or the small-cap Russell 2000. Every month, the ETFs rank stocks by their cumulative returns over the past 250 trading days, excluding the past 20 trading days. The best-performing stocks are chosen until the target portfolio has a total capitalization of 35% of their respective universes.

Momentum, or the tendency for performance to persist over the medium term, is an anomaly in the sense that traditional theories don't do a good job explaining it nor why it still persists decades after its discovery. After all, if everyone knows about a market-beating strategy, they'll arbitrage it away. Thankfully, we're not flying blind here. A great deal of brainpower and ink has been spent uncovering its mysteries.

Why Does It Work?
Narasimhan Jegadeesh and Sheridan Titman are credited by academics for discovering momentum, though practitioners had been exploiting it for decades by the time the duo's study came out in 1993. The pair found that a simple long-short strategy that every month bought the top 10% of best 12-month-performing stocks and short-sold the worst 10% of 12-month performers earned excess returns of about 12% a year. Subsequent research has uncovered momentum in virtually every market studied, including commodities, currencies, stocks, and bonds, and over wide-ranging periods. Thanks to this rich body of research, we have a good idea of why momentum exists and how it behaves.

The most convincing explanation lies with behavioral biases, rather than rational, risk-based theories. In light of surprising or extreme news, investors may "anchor" new price estimates to old prices, preventing prices from fully reflecting new information. Investors are also loath to realize losses, preferring to keep dogs until they break even, and are too quick to sell winners. Both biases prevent prices from instantly reflecting new information; instead, prices slowly adjust to fair value, creating sustained price movements. Once a trend is established, performance-chasers hop on. The trend eventually collapses after the market realizes it has overshot.

Momentum's Momentum
Momentum probably hasn't disappeared because the severity of the strategy's losses means leveraged momentum strategies eventually get wiped out, discouraging institutions from arbitraging it away. Even if it doesn't wipe you out, momentum hurts you at the worst possible time. The Russell-Axioma U.S. Large Cap High Momentum Index, HMTM's benchmark, lost a staggering 48% from June 2008 to March 2009. The strategy can also be a self-fulfilling prophecy, accentuated by investors who try to exploit it. Since momentum's discovery, it's actually become more powerful and volatile, at least in the U.S. stock market. We've seen evidence of momentum strengthening across asset classes in recent years. It's unlikely that momentum will disappear anytime soon, owing to its risks, the universality of psychological biases, and structural limits to arbitrage. But it may become more volatile and perhaps less profitable as investors pile in.

Two-Faced Unreason
Given the strategy's nasty blowups, why bother? AQR founder Cliff Asness demonstrated that momentum pairs well with value, muting some risk. According to AQR, from 1980 to 2009, a simple large-cap momentum strategy (very similar in construction to HMTM's) had a correlation of about negative 0.50 to the Russell 1000 Value Index. It's rare to find two positive-expected-return long-only stock strategies with negative correlations, let alone ones that have market-beating returns themselves. He argues that value and momentum are opposing manifestations of behavioral biases. In "Value and Momentum Everywhere", he, Tobias Moskowitz and Lasse Pedersen find that value and momentum share this negative correlation structure across asset classes and geography. The study lends strong support to Asness's original thesis. We agree that when paired with value stock exposure, momentum becomes much less dangerous.

Value and Momentum in Practice
The simplest way to implement a value and momentum strategy is to split your equity allocation in half and rebalance regularly. Russell 1000 High Momentum ETF pairs well with Russell 1000 value ETFs like iShares Russell 1000 Value IWD and Vanguard Russell 1000 Value Index VONV. Russell 2000 High Momentum naturally fits with Russell 2000 Value ETFs. Historically, however, Russell 2000 funds have lost lots of money due to synchronized buying and selling of relatively illiquid small-cap stocks during index reconstitution periods. Some may prefer less popular small-cap value ETFs like WisdomTree SmallCap Dividend DES and PowerShares FTSE RAFI US 1500 Small-Mid PRFZ as substitutes to the Russell value funds. Braver investors can overweight small-cap momentum and value ETFs. Small-cap stocks tend to drive much of the excess returns in value and momentum strategies.

Not the Holy Grail
Unfortunately, value and momentum aren't slam dunks. The ETFs available don't fully exploit the value and momentum effects. The most value- and momentum-laden stocks tend to be small and relatively illiquid. You probably won't reap anything close to 12% annual excess returns as Jegadeesh and Titman found. Frictional costs, taxes, and capacity constraints all conspire against investors seeking to exploit the phenomenon. At the very least, investors should keep the momentum funds in tax-sheltered accounts, lest they run the real risk of paying taxes on short-term capital gains distributions. We wouldn't be surprised to see a value and momentum strategy only modestly outperform the market net of fees.