Is the S&P 500 a Momentum Trade? I Measured It.

The claim shows up every time a handful of large stocks leads the market: the S&P 500 is really just a momentum trade. Josh Brown put a version of it on television recently, and it is repeated constantly on finance Twitter. The reasoning is intuitive. Cap weighting lets winners grow into bigger positions. The top of the index is full of companies that went up a lot. And in the first half of 2026, a dedicated momentum index returned roughly three and a half times what the S&P 500 did.

Here is what makes this claim different from most market arguments: it is testable. Momentum has a precise definition, a published construction rule, and a public monthly return series going back to 1927. If the S&P 500 were a momentum trade, it would load on that factor.

So I measured it. Over the same window, an actual momentum fund loads +0.42 on the momentum factor. The S&P 500 loads −0.0003. Across 425 rolling three-year windows since 1988, the S&P 500’s momentum loading has never once reached a fifth of a real momentum fund’s.

Four different things people mean by momentum

Most of the disagreement here dissolves once the word is pinned down, because the claim is true under one reading and false under the others.

Academic cross-sectional momentum ranks stocks against each other on trailing return and buys the winners while shorting or avoiding the losers. Kenneth French’s momentum factor, the standard yardstick, is built from “six value-weight portfolios formed on size and prior (2-12) returns,” where the portfolios are re-formed every month andMom = 1/2 (Small High + Big High) − 1/2 (Small Low + Big Low).¹ “Prior (2-12)” means the eleven months ending one month ago: the most recent month is deliberately skipped. The mechanism is a ranking rule plus a monthly reallocation. For the underlying literature, see our guide on growth versus momentum, which covers why the two get conflated.

A long-only momentum index does the same thing inside a universe. The S&P 500 Momentum Index scores every S&P 500 member on its volatility-adjusted twelve-month price change excluding the most recent month, keeps roughly the highest-scoring 100, weights them by market capitalization multiplied by their momentum score, and rebuilds the portfolio twice a year.² Note how close that signal is to French’s. This is a real momentum product, and its existence is itself evidence: S&P would have little reason to publish it if the parent index already did the job.

Time-series momentum, or trend following, asks whether an asset’s own recent return has been positive and sizes exposure accordingly. It is a different strategy on a different unit of analysis, covered in our guide on managed futures and trend following.

“A market full of recent winners.” This is the colloquial usage, and it describes an environment rather than a strategy. It is a legitimate thing to say. It is also the reading that no regression can settle, which matters later.

Cap weighting does not buy the winner

This is the mechanical heart of the argument, and it is where the critique goes wrong.

Take a two-stock index. Company A is worth $600 and Company B is worth $400, so the weights are 60% and 40%. Now A rises 50% and B does not move. A is worth $900, B is still worth $400, and the total is $1,300. A is now 69.2% of the index and B is 30.8%.

Count the shares that changed hands: zero. A fund that already held both companies in proportion did not need to buy more A. The A shares it already owned became more valuable. The weight moved because the price moved, and the portfolio did nothing at all.

Compare that to an equal-weight fund holding the same two companies. To get back to 50/50 it must sell $250 of A and buy $250 of B. It sells the winner. And a momentum rule would go the other way and buy more A precisely because A went up. Cap weighting is the only one of the three that trades nothing.

Our guide on why the S&P 500 is not really passive notes in passing that the index is self-rebalancing, with winners growing and losers shrinking without anyone forecasting them. That single property is what the momentum critique misreads as active winner-chasing. For how deliberate rebalancing rules differ, see how often to rebalance.

The tool below runs all three rules on the same five holdings. The column to watch is turnover, which is exactly 0.0% for cap weighting no matter what the returns are.

Try the “Momentum reverses” preset and you will see equal weighting buy the best performer rather than sell it. That is not a glitch. Equal weighting trims whatever sits above 1/N, which is usually the recent winner but not always, and in that preset the best performer is the smallest position. Cap weighting is the only rule whose turnover stays at zero regardless of which holding led.

One more distinction worth keeping straight. Market capitalization is not a momentum score. A company can hold a large index weight after years of flat or negative returns, and a small S&P 500 member can have outstanding twelve-month momentum while remaining a rounding error in the index. Market cap measures current value; momentum measures a path. The two correlate loosely and are not interchangeable.

So measure it

Mechanics establish that cap weighting does not chase winners. They do not establish that the resulting portfolio has no momentum exposure, because exposure is an outcome rather than an intention. That is an empirical question, so here is the empirical answer.

I joined French’s monthly momentum factor to S&P 500 total-return data and ran Carhart four-factor regressions, which regress a fund’s excess return on the market, size, value, and momentum factors. The momentum coefficient is the loading: how much of the fund’s behavior is explained by the winners-minus-losers factor, after accounting for its market exposure. This is the same test that Summitward’s portfolio page runs on user holdings.

Those windows differ, which matters, because a loading measured over 1988 to 2026 and one measured over 2015 to 2026 are not the same measurement. So I re-estimated the S&P 500 over each momentum fund’s exact window. Over SPMO’s span the S&P 500’s momentum loading is −0.0003 (t = −0.04). Over MTUM’s it is −0.002 (t = −0.25). Same months, same factor, same regression: one is +0.42 and the other is zero.

And the recent run does not change it. Estimated from January 2023 forward, the S&P 500’s momentum loading is +0.000 (t = 0.04). From January 2024 forward it is +0.004 (t = 0.20). Even during the stretch that prompted the claim, the exposure is not there.

The loading has never gotten close, in 38 years

A single full-sample number can hide a lot, so here is every trailing three-year window in the sample. Each point is the S&P 500’s momentum loading over the prior 36 months, so the first window ends in January 1991 and the last in May 2026. The dashed line marks what a real momentum fund measures.

Across all 425 windows the loading stays between −0.081 and +0.034. It is statistically indistinguishable from zero in 81% of them, and positive in only 33%. There is no three-year stretch in 38 years, including the dot-com run and the entire mega-cap era, in which the S&P 500 looked like a momentum fund by this measure.

Two honest caveats about that full-sample −0.017. First, it is statistically significant. With 460 months and an R-squared of 0.994, a very small loading can be estimated precisely, so significance here is a statement about measurement precision rather than about importance. Significant and negligible are both true, and reporting only one of them would be misleading. Second, it is negative, which is the opposite direction from the claim.

That negative sign shows up in the raw correlation too. Monthly correlation between the S&P 500’s excess return and the momentum factor from 1988 to 2026 is −0.28: +0.07 in the 1988-1999 stretch, −0.52 across the 2000s, −0.20 in the 2010s, and −0.28 since 2020. On the full value-weighted US market back to 1927 it is −0.34. When momentum has done well, the index has tended to do slightly worse.

There is a reason for all of this that is worth stating plainly rather than leaving for a reader to catch. The S&P 500 has no momentum tilt largely because it is approximately the market portfolio, and the market cannot tilt toward a subset of itself. Cap weighting is what makes it the market. That is not a dodge around the question; it is the answer to it. The critique’s specific assertion is that cap weighting manufactures a momentum tilt, and the data says it does not.

The last four weeks

While I was writing this, the market ran the experiment. Momentum had a spectacular first half of 2026 and then gave back half of it in four weeks.

The momentum trade fell about 13%. The S&P 500 fell 0.94%. The equal-weight S&P 500 rose and closed at a new high on July 28. If the index were the momentum trade, it would have unwound with it.

What this is not: a momentum crash in the sense Daniel and Moskowitz document, where the strategy breaks down in panic states following a market decline and while the market rebounds.³ That pattern does not fit here. Momentum peaked on June 22, after the S&P 500’s June 2 high, and the index sits only about 2% below its own peak while breadth improves. This looks like a rotation into the rest of the market, not a panic. The narrower point stands on its own: two series that diverge by twelve percentage points in four weeks are not one trade.

What the critique gets right

A one-sided answer would be worth less than the argument it replaces, and three parts of the critique survive.

Concentration is real. The ten largest positions in an S&P 500 fund were about 36.8% of assets on July 28, 2026, counting Alphabet’s two share classes separately.⁴ Owning the index means a large, undiversified bet on a small number of businesses that share customers, suppliers, and valuation assumptions. The measurement above does not argue otherwise, and our guide on concentration risk has the measurement framework.

Worth noticing, though: that figure was above 40% earlier in the year. Concentration fell by several percentage points without any index fund selling a share, for the same reason it rose without any fund buying one. The mechanism runs both directions, which is precisely the argument against calling it winner-chasing.

Index membership does buy after success. Companies become eligible for the S&P 500 only after they get large, liquid and profitable enough, which usually means after the stock has already done well. Funds then buy at inclusion. This is a genuine buy-after-success channel, and it is the strongest version of the critique. But it has weakened sharply: Greenwood and Sammon find the abnormal return around S&P 500 additions fell from an average of 3.4% in the 1980s and 7.6% in the 1990s to 0.8% over 2010-2020, “statistically indistinguishable from zero,” with deletions at −0.6%.⁵ Additions are also rare relative to the whole portfolio, so this is a small channel rather than a monthly reallocation. Our guide on index funds and mega-IPOs covers the inclusion machinery in detail.

Factor exposure is not fixed. Every number above is a measurement over a stated window, not a law. If index composition changed enough, the loading could move, and the honest position is that this needs re-measuring rather than settling once. That is an argument for checking, which is what the chart above is for.

And the reading my data does not test. If the claim is that investor behavior has become momentum-like, that flows chase performance, that breadth is narrow, or that the same crowded names dominate sentiment, then a factor regression on index returns does not refute it. Those are statements about how people are acting, and this guide measures how the index is built and how it behaves. The claim is wrong as a statement about the S&P 500’s construction and its measured factor exposure. It may be defensible as loose shorthand for a market environment, and I would not pretend a regression settles that version.

Equal weight is a different bet, not a neutral one

The usual prescription that follows the momentum critique is to switch to the equal-weight S&P 500. It does reduce single-company concentration, and the same regression shows what else it does.

RSP loads −0.104 on momentum (t = −7.0), roughly six times more negative than the cap-weighted index. It also picks up +0.207 on value and +0.106 on size. So the fix for an index supposedly too driven by momentum is a fund that is substantially more anti-momentum, and it comes bundled with value and small-cap tilts you may or may not want. Mechanically that follows from the design: the equal-weight index holds the same 500 companies and resets every constituent to the same weight each quarter,⁶ which means systematically trimming whatever went up.

That is a defensible active position. It is not a neutral one, and it carries higher turnover, higher costs, and more taxable distributions in a taxable account. Fewer eggs in the largest basket is not the same thing as better diversified, since equal weighting leaves you in the same country, the same currency, and the same economic cycle while adding exposure to smaller and less profitable companies.

What I recommend for DIY investors

Do not sell a low-cost cap-weighted index fund because someone called it a momentum trade. The label is wrong, and acting on it means changing your allocation for a reason that does not survive measurement, which is market timing with extra steps.

Do take the concentration question seriously, because it is the real one hiding inside the wrong label. Concretely:

• Measure the whole household. The number that matters is how much of your wealth and future income depends on the same group of companies. That means adding up employer stock, unvested equity, and career risk alongside whatever your funds hold.
• Diversify along axes that actually differ. International developed and emerging equities, bonds, and deliberate size, value, or profitability tilts change your risk. Adding a second US large-cap fund does not. See the case for global diversification.
• Use new contributions first. Directing new money to underweight assets is the cheapest rebalancing tool and avoids realizing gains.
• Do not buy a momentum fund because momentum just won. A momentum sleeve can be defensible for an investor who understands factor cyclicality and holds it through long stretches of underperformance. Buying it after a +36% half is performance chasing at the fund level, and the four weeks in the table above are what that risk looks like.
• Write the policy down before the next narrative. Decide your maximum acceptable exposure to US mega caps in advance, then rebalance to it on a rule rather than on commentary.

Who the S&P 500 works well for

Investors who want cheap, tax-efficient exposure to large US companies, who accept market-cap weighting, who understand it is not a global portfolio, and who hold their other exposures elsewhere. It is also a perfectly reasonable single equity choice inside a 401(k) with a poor menu.

Who needs something different

Anyone whose paycheck, equity compensation, and portfolio all point at the same handful of mega-cap technology companies, since their true exposure is far larger than a fund statement suggests. Investors treating the S&P 500 as a complete retirement portfolio rather than one asset class among several. And if the index’s current company and sector concentration is more than you can tolerate, the remedy is deliberate diversification into different risks rather than a different US large-cap weighting scheme. See human capital risk for tech workers.

How Summitward helps

You can run this regression on your own holdings rather than take mine on faith. Summitward’s portfolio analysis runs CAPM, Fama-French three-factor, Carhart four-factor, and Fama-French five-factor regressions against Ken French’s data, reporting each factor’s beta, t-statistic, p-value, and the model R-squared, with per-holding rows as well as a portfolio total. Momentum is one of the factors reported, so the question “do I actually have momentum exposure” has a number attached to it for your portfolio specifically.

The concentration tab computes your Herfindahl-Hirschman index, effective number of positions, and largest position, and flags holdings above a threshold you set. And the free Portfolio X-Ray needs no login: paste in tickers and weights and it will tell you which of your funds occupy the same slice of the market and what they add up to, which is usually the fastest way to find out that four tickers are one bet. Note that it groups funds by category rather than looking through to individual share-level holdings, so it will tell you that VOO and QQQ overlap heavily without computing your exact look-through weight in any one company.

Frequently asked questions

Is the S&P 500 a momentum trade?

No, on the two readings that can be tested. It is not constructed as one: it selects on size, liquidity, and profitability rather than trailing return, and it does not reallocate toward winners. And it does not behave as one: measured against the standard momentum factor over the same window in which an actual momentum fund loads +0.42, the S&P 500 loads −0.0003. If the phrase is meant loosely, as shorthand for a market led by a few recent winners, that is a description of an environment and this measurement does not speak to it.

Does an index fund buy more of a stock when it goes up?

Not for stocks it already holds. The weight rises because the shares already owned became more valuable, which requires no transaction. Index funds do trade, but for other reasons: additions and deletions, share issuance and buybacks, mergers, float changes, and investor cash flows. Incoming contributions are invested at current weights, which does mean new money buys more of what has risen, but that is a property of the contribution rather than a rebalancing rule.

If the loading is statistically significant, isn't there some momentum exposure?

The full-sample loading is −0.017 with a t-statistic of −4.8, so it is distinguishable from zero and it is negative. Significance here comes from a long sample and a very well-fitting regression, which let a tiny coefficient be pinned down precisely. A loading of −0.017 still contributes almost nothing to returns next to a real momentum fund’s +0.42. Both facts belong in the answer.

Should I switch to the equal-weight S&P 500?

Only if you want its factor tilts on purpose. RSP measures −0.104 on momentum, +0.207 on value, and +0.106 on size, so it is an active bet against momentum with value and size attached, plus higher turnover and more taxable distributions. It reduces single-company concentration, which is a real benefit. It does not make you globally diversified.

Has the S&P 500 ever had meaningful momentum exposure?

Not in this data. Across 425 rolling three-year windows from 1991 to 2026, the loading stayed between −0.081 and +0.034, never reaching a fifth of what a dedicated momentum fund measures. It was indistinguishable from zero in 81% of those windows.

Is QQQ a momentum fund?

No. QQQ measures −0.063 on momentum. Its large loadings are elsewhere: strongly negative on value, which is a growth bet, and above 1.0 on the market. The Nasdaq-100 selects on listing venue rather than on performance. See why I avoid QQQ.

Does high concentration mean a crash is coming?

Concentration raises how much your outcome depends on a small number of companies. It is not a timing signal, and treating it as one has cost investors more than the concentration itself. You do not need a market forecast to decide that a portfolio is more concentrated than you want.

Key takeaways

• Cap weighting lets winners grow; it does not buy them. A 60/40 index becomes 69/31 when the larger holding rises 50%, with zero shares traded. Equal weighting would sell the winner and a momentum rule would buy it. Turnover of exactly zero is the signature.
• Measured against the standard momentum factor, the S&P 500 loads zero. Over the identical window, SPMO loads +0.42 and the S&P 500 loads −0.0003. Across 425 rolling three-year windows since 1988 the loading never left the band −0.081 to +0.034.
• The correlation runs the other way. Monthly correlation between S&P 500 excess returns and the momentum factor is −0.28 since 1988, and negative in every subperiod after 1999.
• Four weeks in 2026 made the point. Momentum funds fell 13% to 15% from June 30 to July 28 while the S&P 500 fell 0.9% and equal weight hit a new high.
• Concentration is the real risk inside the wrong label. The top ten were about 36.8% of an S&P 500 fund in late July 2026. Address that with global diversification and a written policy, not by swapping one US large-cap weighting scheme for another.
• Equal weight is an active factor bet. It is six times more anti-momentum than the index it replaces, with value and size tilts and higher tax cost attached.

Related guides

• Growth vs. Momentum covers the momentum literature and the four momentum ETFs worth knowing.
• The S&P 500 Is Not Passive on committee governance and the discretion inside the index rules.
• Concentration Risk for HHI, effective positions, and the sell-versus-hold math.
• Fama-French Factors explains what a factor loading is and how to read a regression.
• The Tech Bro Portfolio on why several US large-cap funds are one bet.
• Index Funds and Mega-IPOs on float adjustment and how companies enter the index.
• Managed Futures and Trend Following for time-series momentum, the other kind.

Sources

1. Kenneth R. French, “Detail for Monthly Momentum Factor (Mom)”, Data Library, Tuck School of Business. Six value-weight portfolios formed on size and prior (2-12) returns, re-formed monthly, with NYSE median size breakpoints and 30th/70th NYSE percentile return breakpoints applied to NYSE, AMEX and NASDAQ stocks. Monthly factor data through May 2026 used for all regressions in this guide.
2. Invesco S&P 500 Momentum ETF, summary prospectus filed December 19, 2025, describing the S&P 500 Momentum Index as approximately 100 S&P 500 stocks with the highest momentum score, weighted by market capitalization multiplied by momentum score, and reporting 99 constituents as of October 31, 2025. The momentum score as a volatility-adjusted twelve-month price change excluding the most recent month, the 9%/3x weight cap, and the semi-annual March and September rebalance are described in this 2020 filing. S&P Dow Jones Indices’ own methodology pages block automated access, so index descriptions here are cited to SEC filings that reproduce them.
3. Kent Daniel and Tobias J. Moskowitz, “Momentum crashes”, Journal of Financial Economics 122(2), November 2016, 221-247. Momentum crashes “occur in panic states, following market declines and when market volatility is high, and are contemporaneous with market rebounds.”
4. State Street Global Advisors, SPDR S&P 500 ETF Trust holdings, as of July 28, 2026. The ten largest listed positions sum to 36.79%, counting Alphabet’s Class A and Class C shares as two of the ten; counting Alphabet once, the top ten distinct companies would be slightly higher. Cross-checked against stockanalysis.com at 36.71% as of July 27, 2026. These are ETF weights, which track the index closely but are not identical to it.
5. Robin Greenwood and Marco Sammon, “The Disappearing Index Effect”, NBER Working Paper 30748, December 2022. Sample 1980-2020. Abnormal returns around S&P 500 additions averaged 3.4% in the 1980s, 7.6% in the 1990s, 5.2% over 2000-2009, and 0.8% over 2010-2020, “statistically indistinguishable from zero”; deletions were −0.6% over 2010-2020.
6. S&P 500 Equal Weight Index description reproduced in a Citigroup pricing supplement filed April 2, 2026: the index holds the same constituents as the S&P 500 and “at each quarterly rebalancing, each constituent … is allocated the same weight as every other constituent.”
7. All factor loadings, rolling betas, correlations, and 2026 return figures in this guide are the author’s own calculations, reproducible with scripts/extract_momentum_betas.py in the Summitward repository. Regressions are ordinary least squares on monthly excess returns with classical standard errors; return figures come from daily closing prices. They are measurements over stated windows and are not forecasts.