A new ETF that builds on a different way of measuring stock momentum should better capture the lightning in a bottle, writes Samuel Lee of Morningstar ETFInvestor.

Several readers have asked me for momentum fund options. Until now, I could only point them to AQR's offerings with much confidence.

But with the introduction of iShares MSCI USA Momentum Factor ETF (MTUM), there's now a credible ETF alternative. This fund is a cut above its ETF competitors.

First, it's far cheaper than any other momentum strategy out there, with a 0.15% expense ratio. Second, it's seeded with $100 million of the Arizona State Retirement System's money, which will likely encourage stability and secondary-market liquidity. And finally, I like its methodology.

Momentum is the tendency for performance to persist. Barring Treasury bonds, momentum has been found in nearly every market and asset class studied. It's the most convincing refutation of the efficient-markets hypothesis.

Most investors prefer to sell winners to lock in gains and hold onto losers with the hope of eventually breaking even, a bias called the disposition effect. The combination of anchoring and disposition effect means market prices underreact to new information, leading to price trends.

More powerfully, when we observe these price trends, we extrapolate them into the future and pile in, creating a self-fulfilling prophecy in which a trend is further extended by bandwagon investors. Eventually, the delusion ends, and that price trend suddenly reverses.

MTUM tracks the MSCI USA Momentum Index, which applies a twist to the traditional momentum strategy.

In most academic studies, momentum is measured by 12-month trailing price return, excluding the most recent month, for 11 months of price return. The most recent month is excluded because stocks exhibit short-term mean reversion: The best-performing stocks in the past month often underperform the subsequent month.

The MSCI version for each stock measures six- and 12-month price momentum, excluding the most recent month. However, unlike the academic studies, the six- and 12-month windows are just moved back a month. The six-month signal is calculated by using months two to seven, the 12-month signal uses months two to 13.

The returns are divided by the annualized standard deviation of the stock's trailing three-year returns. Then they're converted to z-scores, a standard method for normalizing data into a "bell curve" distribution. The z-scores of the six- and 12-month signals are averaged to produce a final momentum score. Then the score is multiplied by the stock's market cap to come up with a final weight.

The overall effect of looking at risk-adjusted return and averaging two signals is to ensure the consistency of returns. Research has indicated that stocks that smoothly rise or fall are more likely to exhibit momentum than stocks that suddenly rise or collapse.

The index is rebalanced twice a year, which keeps turnover down but weakens the fund's momentum tilt. The index has a conditional rule that checks each month to see whether the current month's volatility has steeply risen versus the past month's volatility. If the condition is met, the index rebalances, but only using the six-month signal.

This rule was implemented to protect against momentum crashes. During bear markets, momentum strategies switch to defensive stocks and miss out on the subsequent recoveries led by cyclical stocks.

The index is, naturally, a back-tested strategy. It seems to be an improvement over the traditional approach, represented by the AQR Momentum Index, which rebalances quarterly, using the traditional 12-month ex-recent month return signal.

Given the weight of the evidence, I think the odds favor MTUM earning some momentum premium in the future. Even if it's nil, the fund is so cheap it's unlikely to hurt you much.

Subscribe to Morningstar ETFInvestor here...

Related Articles:

The Holy Grail of Investing

Momentum Stock Entry Signals

Here's a Bone You Can Pick