Market-Cap Weighted Indexes Are Chasing the Top: Passive Investing Has Active Trading Hidden Inside
Market-cap weighting isn't passive — it weights by price and buys high, sells low. Index reconstitution alone costs about 15 basis points a year in trading costs; fundamental-weighted indexes have beaten the cap-weighted value index by more than 2% a year for 20 years.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Market-cap weighting means weighting by price
Rob Arnott uses a thought experiment to expose the logic of cap weighting: suppose someone told you about a strategy — you buy a company only after its market cap crosses a certain threshold, on average buying when it has risen 75% relative to the market and trades at twice the market's valuation; the sell discipline is just as simple, you sell when its market cap falls below the threshold, on average selling at half the market's valuation and 7000 basis points below the market. Would you buy it? The answer is "Hard pass." But that is the active side of indexing: cap weighting is weighting by price, the more expensive the bigger the weight. The index is called passive, yet its 5% turnover portion "looks like an ultra-high-speed growth fund manager on drugs."
— Rob ArnottIndex reconstitution is legal front-running
S&P index funds already hold about 25% of the total market cap of all stocks in the index. Because index funds are obsessed with zero tracking error and must buy at the closing price on the day of inclusion, hedge funds buy ahead and resell to the index funds — what Arnott calls "legal front-running," a pattern he says he wrote about back in 1986. The cost is quantifiable: if you could trade at the moment S&P announces its decision rather than the moment it takes effect, you would gain 15 basis points a year. At 3% to 5% annual turnover, that works out to 300 to 500 basis points of cost per stock per trade. His metaphor: a herd of elephants squeezing through a revolving door.
— Rob ArnottKicked out and let back in, you lose both ways
The flip-flop problem of inclusion and deletion has specific statistics: about 28% of stocks that get added are deleted within ten years; of the stocks that get deleted, nearly half are added back within ten years. On average, a stock is added only after outperforming by 75 percentage points; if it falls out of favor and gets kicked out, it is deleted when it is down 7000 basis points. Arnott stresses the asymmetry: up 75 then down 70 doesn't put you back at zero, it leaves you down 50 — and you weren't in for that 75 gain, while you were fully present for the 70 decline. The deleted-then-re-added case is more extreme: in the year before deletion it underperformed by about 3500 basis points, then it outperformed by 180 percentage points and roughly tripled relative to the market before being added back.
— Rob ArnottPick stocks by economic footprint, not market cap
The fundamental-weighting approach makes the index "deliberately mirror the economy rather than deliberately mirror the market": instead of selecting and weighting by market cap, it looks at how big the business itself is — sales, profits, net assets (today with an intangible adjustment added), and shareholder distributions of dividends plus buybacks, averaging the four measures. Take NVIDIA: its share of total economic profits is a decent chunk, but not the 7% to 8% market-cap weight — it's in the 2% range; by sales it's also in the 2% range; by dividends or net assets it's small enough to ignore. Averaged out it's about 1% to 1.5% of the economy, so it qualifies for inclusion and gets included at that proportion. The result is to downweight a favored growth stock to its current economic size and to upweight an unloved cheap value stock to its economic size, giving the portfolio a pronounced value tilt.
— Rob ArnottRAFI has beaten the value index over 20 live years
The fair benchmark for RAFI is the cap-weighted value index, not the total-market index. Schwab, Invesco and PIMCO together hold more than $100 billion in RAFI assets; the strategy is neither new nor small, having launched about 20 years ago. The comparison: RAFI has beaten the cap-weighted value index by 2% to 2.5% a year on average (compounded). Arnott points to a counterintuitive fact: growth indexes have outperformed massively by getting ever more expensive relative to fundamentals, while the underlying fundamentals of value indexes and growth portfolios (sales, profits, book value, dividends) have grown roughly in step since the turn of the century. In other words, relative performance has diverged for 25 years, about 10,000 basis points cumulatively, but the fundamentals have moved side by side.
— Rob ArnottEqual weighting works, but the stock pool is the weak spot
Asked why not simply equal-weight, Arnott says equal weighting is a perfectly legitimate approach: it carries a strong small-cap tilt (small companies get the same weight as NVIDIA and ExxonMobil), a strong value bias (cheap and expensive get the same weight), and rebalancing alpha — trim what has risen, top up what has fallen. Equal weighting's one Achilles' heel is "equal-weight what?": if you equal-weight the S&P, your equal-weighted pool contains the companies that have already risen enough to qualify for inclusion while excluding the companies that have fallen enough to be cheap, so the resulting portfolio actually tilts toward high-multiple stocks. So over the long run equal weighting performs about the same as fundamental indexing, but with much higher volatility.
— Rob ArnottA value tilt has to survive a 38% shortfall
Fundamental indexing launched live in 2005, value peaked two years later in 2007, and then it underperformed brutally until the summer of 2020, followed by repeated bottoming and bouncing — outperform, crash, outperform again, crash again. By the end of 2025, the Russell value index had underperformed the Russell 1000 by 3800 basis points peak to trough, meaning you are 38% poorer than a plain Russell or S&P index investor. Arnott says this is a terrible headwind for any strategy with a value tilt. But RAFI has rebalancing alpha: if a stock soars and the fundamentals don't support it, trim; if a stock plunges and the fundamentals are intact, add. Relative to the value index, RAFI has outperformed by slightly more than 2% annualized live, and compounding over 20 years means being more than 50% wealthier than the cap-weighted value index, with tracking error of about 2.5% and roughly three winning years out of every four.
— Rob ArnottIn their own words · checked verbatim
On average, I'm buying them when they're up 75% relative to the market in the last year and trading it twice the market multiple.
Rob Arnott2:05
The 5% looks like a hyper growth manager on crystal meth.
Rob Arnott3:07
It's legal front running.
Rob Arnott4:11
It's a herd of elephants trying to go through a single revolving door.
Rob Arnott6:13
Now, if you gain 75 and lose 70, you aren't back where you started. You're down 50.
Rob Arnott7:15
The growth indexes have outperformed hugely. but they've outperformed by dint of becoming more and more expensive relative to fundamentals.
Rob Arnott11:24
You were 38% poorer than a simple Russell or S & P index investor.
Rob Arnott15:40
Figures
| Share of total index market cap held by S&P index funds | about 25% | 4:11 |
| Index reconstitution trading cost (extra if trading at the announcement price) | 15 basis points a year | 6:13 |
| Per-stock per-trade cost at that annual turnover | 300 to 500 basis points | 6:13 |
| Share of added stocks deleted within ten years | about 28% | 6:13 |
| Share of deleted stocks added back within ten years | nearly half | 6:13 |
| Average outperformance of added stocks before inclusion | 75 percentage points | 7:15 |
| Average loss relative to the market of deleted stocks at deletion | 7000 basis points | 7:15 |
| RAFI assets (Schwab, Invesco, PIMCO combined) | more than $100 billion | 10:20 |
| RAFI annualized excess return vs. cap-weighted value index | 2% to 2.5% | 10:20 |
| Russell value peak-to-trough underperformance vs. Russell 1000 | 3800 basis points | 15:40 |
Glossary
- cap weighted index
- An index that assigns weights by company market cap, so the bigger the market cap the bigger the weight.
- fundamental index / RAFI
- An index that weights by economic fundamentals such as sales, profits, net assets and dividends plus buybacks rather than by market cap.
- rebalancing alpha
- The return from periodically resetting to target weights: trimming what has risen and adding to what has fallen.
- tracking error
- The volatility of the deviation between a portfolio's return and its benchmark's return.
- flip-flops
- The back-and-forth of stocks being added to an index, then deleted, and sometimes added back again.
How to listen
Investors and asset managers who care about the hidden costs of index funds and who are allocating between value and growth — especially anyone trying to understand the flaws of market-cap weighting.
The roughly one-minute show intro and background setup at the start can be skipped; go straight to the thought experiment beginning at 1:02.