Two Related Ideas, Often Confused
Before weighing the evidence, it helps to be precise about what is actually being tested, because "trend-following" and "momentum" get used almost interchangeably in casual conversation even though they measure different things. Trend-following asks whether an asset’s price is above or below its own recent trend — often approximated with a simple tool like a moving average, which smooths out the noise in daily price swings so a longer-term direction becomes visible. Momentum instead ranks a group of assets against each other and favors the ones that have outperformed over some recent window, typically three to twelve months.
| Feature | Trend-following | Momentum |
|---|---|---|
| Signal direction | Absolute — asset vs. its own price history | Relative — asset vs. other assets |
| Typical lookback | Often 10-month or 12-month moving average / absolute return | Usually 3–12 month ranked returns |
| Common construction | Long/short or long/cash, often across futures on stocks, bonds, commodities, currencies | Long the winners, sometimes short the losers, within a stock universe |
| Main appeal | Drawdown control; can go defensive when prices fall | Captures persistence in relative performance ("winners keep winning") |
| Main limitation | Whipsaws in choppy, non-trending markets | Vulnerable to sharp reversals ("momentum crashes") after steep declines |
The distinction matters because the two strategies are usually tested — and defended — using different kinds of evidence, and conflating them can make the case for either one sound stronger or weaker than it actually is.
Why the Evidence Is Taken Seriously
The credibility of both ideas rests less on any single result than on how many times, in how many different markets, researchers have found something similar. AQR’s A Century of Evidence on Trend-Following Investing is the anchor study here: using data on stocks, bonds, commodities, and currencies back to 1880, the authors constructed a diversified futures portfolio using momentum signals across 67 markets, sized to a constant volatility target, and found that the strategy delivered positive returns net of estimated transaction costs and fees across more than a century. Perhaps the most cited finding is behavioral rather than purely statistical: during the ten worst drawdowns for a traditional 60/40 stock-bond portfolio, the trend-following strategy posted positive returns in eight of those ten episodes.
Momentum has its own, separate research lineage. Jegadeesh and Titman’s 1993 paper on buying winners and selling losers put the effect on the map by showing that stocks with strong 3-to-12-month returns tended to keep outperforming over the following months, while laggards kept lagging. Rouwenhorst extended the test to a dozen European markets and found similar results, which is the kind of cross-market replication researchers rely on to rule out the possibility that a pattern is just a statistical fluke specific to one country or period. Perhaps most tellingly, Eugene Fama and Kenneth French — architects of the efficient market hypothesis, an academic framework that argues prices already reflect all available information — stress-tested momentum alongside several other candidate anomalies in their paper Dissecting Anomalies and, somewhat reluctantly, concluded that momentum survived where several other anomalies did not.
Two simpler, more accessible tests point in a similar direction. Meb Faber’s 2007 paper tested a rule as basic as holding an asset only when its price sits above its 10-month moving average, applied across stocks, bonds, real estate, and commodities since the early 1900s; it produced returns comparable to buy-and-hold but with notably shallower drawdowns. Wes Gray blended a 12-month absolute momentum rule with a moving-average rule across U.S. stocks, bonds, foreign stocks, REITs, and commodities, and framed the appeal in behavioral terms: investors’ appetite for risk shifts with recent experience, and a trend rule can act as a kind of release valve that helps people stay invested through a crash rather than abandoning a plan at the worst possible moment.
What the Studies Actually Establish — and Where They Stop
This is where a myth-busting lens earns its keep: acknowledging that the underlying pattern is well-documented does not mean every headline claim about it is warranted.
| What the research supports | What it does not establish |
|---|---|
| The pattern persists across asset classes and roughly 100+ years of data | That the same edge will repeat in the same form going forward |
| Trend-following posted gains in most of the worst 60/40 drawdowns tested | That it eliminates large losses or removes risk entirely |
| Effects replicate across regions (e.g., international momentum tests) | That a standard retail brokerage account can capture results identical to a futures-based, volatility-targeted portfolio |
| Momentum survives rigorous academic stress-testing better than several rival anomalies | That momentum should replace, rather than supplement, a core indexed holding |
Maximum drawdown — the largest peak-to-trough decline over a period — is the metric most of this research leans on to argue for trend-following’s behavioral value, but it is a useful, incomplete measure: it says nothing about how long a recovery took, how often false signals ("whipsaws") triggered unnecessary trades, or what happened in the many periods that weren’t among the worst ten.
The Retail Translation Problem
The strongest results in this literature come from long/short, futures-based, multi-asset portfolios rebalanced monthly and scaled to a target volatility level — a construction that is genuinely different from a person buying a handful of trending stocks in a brokerage app. AQR’s own century-long test explicitly incorporated estimated transaction costs and a hedge-fund-style fee structure to make the numbers realistic, which is a tacit admission that trading frequency and cost structure change outcomes materially. That matters for a retail investor for a simple reason: trend and momentum rules require regular rebalancing, and rebalancing has a cost — in commissions, in taxes on realized gains, and, where they exist, in transaction taxes designed specifically to discourage frequent trading. Research on financial transaction taxes notes that even modest levies scale linearly with how often an investor trades, which disincentivizes exactly the kind of rebalancing these strategies depend on. None of this proves the retail version fails; it simply means the gap between an institutional backtest and a personal account is not cosmetic.
A Satellite Question, Not a Referendum on Indexing
It’s worth remembering that a market-cap-weighted index fund already behaves like a crude momentum strategy in one sense — it lets winners grow to a larger weight and lets losers shrink, simply by tracking market value. That overlap is a reason to see trend-following and momentum less as a rival to indexing and more as one possible layer of diversification by strategy, alongside diversification by asset class and geography — something an investor might add if the complexity, turnover, and discipline required actually suit them, not something the evidence demands everyone adopt.
The Honest Bottom Line
The historical record for trend-following and momentum is genuinely more robust than most market folklore — replicated across centuries, asset classes, and continents, and grudgingly conceded even by efficient-market researchers. But robust evidence for a pattern’s existence is a different claim from evidence that it belongs in a particular person’s portfolio. The strongest results come from implementations — futures access, volatility targeting, long/short construction — that most retail investors don’t replicate exactly, and the friction of taxes and trading costs is not a footnote but part of the real return. Treating these strategies as an optional, well-documented tool rather than a verdict on indexing is the more defensible reading of the evidence.
