Photo by merci mockup on Unsplash
The short version
- MTUM has compounded at 16.3% over ten years, but its risk is negatively skewed: the danger sits in rare, sharp reversals rather than in day-to-day volatility.
- The academic record on momentum crashes (Daniel & Moskowitz 2016, Barroso & Santa-Clara 2015) describes exactly the failure mode a long-only momentum ETF inherits — a violent snap-back after a panic, when yesterday's laggards lead.
- Bottom line: MTUM is a defensible satellite tilt for an investor who understands its tail shape, not a set-and-forget substitute for a broad-market core.
Momentum is one of the most durable anomalies in the empirical asset-pricing literature — documented across decades, markets, and asset classes. It is also one of the few factors that periodically detonates. The returns look smooth for years and then surrender a large share of accumulated gains in a matter of weeks. The iShares MSCI USA Momentum Factor ETF (MTUM) packages the U.S. large- and mid-cap version of this factor into a single, cheap, liquid vehicle. The question worth asking is not whether momentum works — the ten-year record says it has — but what shape its risk actually takes, and whether that shape fits the role you would ask it to play.
What MTUM actually holds
MTUM tracks the MSCI USA Momentum Factor Index. The index does not simply buy stocks that went up. It ranks large- and mid-cap U.S. names by risk-adjusted price momentum — trailing 6-month and 12-month excess returns, each scaled by realized volatility — and weights the highest scorers. That volatility scaling matters: it is a deliberate attempt to tilt toward steadier winners rather than the most explosive movers, because the most explosive movers are precisely the names that reverse hardest.
Reconstitution happens semi-annually, in May and November, with a rules-based provision for an ad-hoc rebalance when realized market volatility spikes. That schedule is the single most important structural fact about the fund, and I will return to it. A portfolio that only refreshes its holdings twice a year is, by construction, slow to recognize when its winners have stopped winning.
The data on the table
| Metric | MTUM |
|---|---|
| Name | iShares MSCI USA Momentum Factor ETF |
| Expense ratio | 0.15% |
| AUM | $29.0B |
| Inception | 2013-04-16 |
| Distribution yield | 0.5% |
| 5Y CAGR | 14.3% |
| 10Y CAGR | 16.3% |
| 5Y annualized volatility | 21.6% |
| Max drawdown (trailing 5Y) | -32.3% |
Return and risk figures are computed from adjusted daily closes via yfinance, pulled 2026-07-16. Expense ratio, AUM, inception, and yield are from the iShares fund page (iShares MSCI USA Momentum Factor ETF fact sheet). At 0.15%, MTUM is inexpensive for a single-factor product, and its $29.0B in assets removes any real concern about bid-ask spread or closure risk — two frictions that quietly erode returns in smaller factor funds.
Why the volatility number is the wrong number to watch
A 21.6% annualized volatility looks unremarkable — a shade above a broad U.S. equity index in the same window. If volatility were the whole story, MTUM would read as a slightly racier version of the market. It is not, and the reason is the part the standard deviation hides.
The defining risk of momentum is not dispersion; it is negative skew. Daniel & Moskowitz, in their 2016 paper Momentum Crashes, show that momentum returns are conditionally left-skewed: the strategy delivers many small-to-moderate positive months and a handful of catastrophic ones. The crashes cluster in a specific setting — after a sharp market decline, in the early, violent phase of a rebound. In that moment the stocks that had fallen hardest (which a momentum portfolio has either sold or never held) rocket off the bottom, while the defensive names momentum rotated into lag badly. A long-only fund cannot short anything, but it inherits the same wound from the opposite side: it is concentrated in exactly the names about to be left behind.
Momentum does not crash in the depths of a panic. It crashes on the way out, when the market snaps back and yesterday's losers lead the recovery.
This is why the -32.3% trailing five-year drawdown deserves more weight than the tidy volatility figure. Standard deviation treats an up-move and a down-move symmetrically; it cannot see that momentum's losses arrive concentrated, correlated, and at the worst possible time — precisely when an investor is most tempted to capitulate. Barroso & Santa-Clara (2015) made the constructive version of this point: because momentum's own volatility is forecastable, scaling exposure by recent volatility historically tamed the tail. MTUM's index applies a related idea at the stock level through its risk-adjusted ranking, but volatility scaling reduces crash severity — it does not abolish it.
The semi-annual rebalance is a feature and a lag
Here is the non-obvious second-order effect. MTUM's twice-a-year reconstitution is usually sold as a virtue: it holds down turnover, keeps the tax-cost ratio low, and prevents the fund from chasing noise. All true. But the same slowness means that between rebalances the portfolio can drift meaningfully from a live momentum signal. In a regime change — the fast reversals that produce crashes — the fund keeps holding last cycle's leaders for weeks or months after those names have rolled over.
The ad-hoc volatility-triggered rebalance exists to blunt exactly this, and it is a thoughtful piece of index design. Yet it fires on realized volatility, which is a lagging read of stress. With the VIX sitting at 16.5 as of 2026-07-14 (FRED, asof 2026-07-14), the market is in a calm regime and the trigger is dormant — which is the environment in which momentum quietly accumulates the concentrated bets that a future reversal would punish. The lesson is not that the rules are flawed; it is that no semi-annual, rules-based process can be fully in front of a crash that unfolds in days. An investor holding MTUM is implicitly accepting that lag as the price of low turnover.
Where MTUM plausibly fits — and where it does not
The ten-year CAGR of 16.3% is real and was earned across a genuinely varied decade: a long bull market, a pandemic shock, an aggressive hiking cycle, and the disinflation that followed. That is a stronger regime sample than most single-factor funds can claim, and it argues against dismissing momentum as a backtest artifact. Initially I expected the live record to lag the pre-inception simulations most factor products lean on; the realized 10-year number holds up better than that skepticism assumed.
What the record does not do is make MTUM a core holding. A factor tilt with left-skewed returns belongs in the satellite sleeve — sized so that a -32% drawdown arriving at an inconvenient moment does not force a behavioral error. Readers weighing momentum against other factor implementations may find the adjacent comparisons useful: the trade-off between rules-based and AI-driven momentum in AMOM Explained, the quality-versus-momentum question in QRFT vs AMOM, and the broader framing of factor tilts against a plain index core in MOAT vs QUAL. Each reinforces the same discipline: decide the role first, then choose the vehicle.
Scenarios where MTUM fits
- Reader in their 30s, broad-market core already in place, comfortable with a tilt: a modest MTUM satellite (single-digit percent of equity) adds a documented factor premium without threatening the portfolio's center of gravity. The crash risk is survivable at that size.
- Reader who checks accounts often and reacts to drawdowns: MTUM is a poor fit. The negatively skewed loss profile is designed to test exactly the behavior that leads to selling the bottom of a reversal.
- Reader seeking the single fund to hold forever: also a poor fit. A single-factor product is a supplement to a diversified core, not a replacement for one.
Editor's read
Momentum is one of the better-evidenced factors in the literature, and MTUM implements it cheaply, liquidly, and with a sensible volatility-aware ranking. If I held it, it would be a small satellite tilt whose sizing already assumes a -32% drawdown will eventually show up at the worst time — because the whole risk of momentum is that its losses are concentrated, not diffuse. Bought at that size and with that expectation, the ten-year record is attractive. Bought as a core, or by an investor who will sell into the first violent reversal, it is a mistake the data can see coming.
The editor does not hold MTUM at the time of writing.
Frequently asked questions
What is a "momentum crash"? It is a rare, severe stretch of momentum underperformance that occurs in the early phase of a market rebound, when the stocks that fell hardest lead the recovery and momentum portfolios — positioned in the prior cycle's winners — get left behind. The academic reference is Daniel & Moskowitz (2016), Momentum Crashes.
Does MTUM's -32.3% drawdown mean it is riskier than the S&P 500? Not by volatility, which is only modestly higher. The distinction is shape: momentum's losses are negatively skewed and arrive concentrated in reversals, so the tail risk is worse than the standard deviation suggests even when the two funds' volatility looks similar. This analysis does not include a same-window comparison against a broad index, which would sharpen the point.
How often does MTUM change its holdings? Semi-annually, in May and November, with a rules-based provision for an extra rebalance when market volatility spikes. Low turnover keeps costs and taxes down but leaves the fund slow to shed winners that have rolled over between scheduled dates.
Is 0.15% a reasonable fee for this exposure? For a single-factor ETF, yes. It is well below typical active factor pricing, and MTUM's $29.0B in assets means implementation frictions like bid-ask spread and closure risk are negligible. Fee is not the deciding variable here; risk shape is.
Should momentum be a core holding? The role it fits best is a satellite tilt sized so that a large, badly timed drawdown does not force a sale. A single-factor fund with left-skewed returns is a supplement to a diversified core, not a substitute for one.
Key takeaways
- MTUM's 16.3% ten-year CAGR was earned across multiple regimes, which is real evidence the factor is not merely a backtest artifact.
- The risk that matters is negative skew, not volatility: momentum's losses are concentrated in sharp market reversals, and the -32.3% drawdown reflects that tail better than the 21.6% volatility does.
- Semi-annual reconstitution keeps costs low but leaves the fund structurally lagged during the fast regime changes that produce crashes.
- At 0.15% and $29.0B, cost and liquidity are not concerns — sizing and holder behavior are.
- MTUM fits as a modestly sized satellite tilt for an investor who has priced in the crash; it does not fit as a core or a single-fund solution.
Methodology. Return, volatility, and drawdown figures are computed from adjusted daily closing prices via yfinance, pulled 2026-07-16; the trailing windows are five and ten years to that date. Expense ratio, AUM, inception, and distribution yield are from the iShares fund page. Macro figures (10-year Treasury 4.58%, effective fed funds 3.63%, VIX 16.5, CPI YoY 3.7%) are from FRED, asof dates 2026-06-01 to 2026-07-14. Factor claims reference Daniel & Moskowitz (2016) and Barroso & Santa-Clara (2015).
What this analysis can and can't tell you. A single-fund trailing-window read cannot substitute for a full comparison against a broad-market benchmark, and it does not model a fresh momentum crash as a forward stress test — the ten-year sample contains stress episodes but no guarantee the next reversal resembles them. Single-regime risk and the live-versus-backtest gap both remain.
This article is for educational purposes and does not constitute personalized financial advice. See the full Disclaimer.