Key takeaways
- Over the 20 years to December 2024, 91.99% of active US large-cap funds trailed the S&P 500, and 94.11% of all active US domestic equity funds trailed the S&P Composite 1500 (SPIVA U.S. Year-End 2024).
- The odds are not uniform. Over the 20 years to June 2025, 94.03% of emerging-market funds underperformed, but only 70.45% of international small-cap funds did.
- Only 37.29% of the 2,373 US domestic equity funds that existed in mid-2005 were still alive 20 years later. SPIVA counts the dead ones. Most comparisons you see quoted do not.
- Asset-weighting narrows the gap but does not close it: over 20 years, US large-cap active funds returned 8.77% a year equal-weighted and 9.48% asset-weighted, against 10.73% for the S&P 500.
- Morningstar, measuring against real index funds that charge real fees, found 21% of active funds survived and beat their passive peers over the 10 years to December 2025 — and just 10% among US large-cap.
Two investors, one active vs passive decision, and somebody keeping score
Picture two people with the same money and the same two decades in front of them. One pays a manager to pick stocks, and follows the fund's letters with genuine interest. The other buys the whole market and goes back to their life. Which of them is holding more at the end?
It isn't a matter of taste, because somebody has been keeping score. The SPIVA Scorecard has been published twice a year since 2002 by S&P Dow Jones Indices — the company that owns the S&P 500 — and it does exactly one job. It lines active funds up against the index they claim to beat.
Its US scorecard for year-end 2024 says that over the preceding 20 years, 91.99% of active large-cap US equity funds lost to the S&P 500. Across all active US domestic equity funds, measured against the broader S&P Composite 1500, the figure was 94.11%.
So the stock-picker probably lost. That much you'd have heard third-hand. What you probably haven't heard is the part worth your time: the number swings hard depending on which market you ask about, SPIVA counts in a way almost nobody else does, and S&P's separate Persistence Scorecard is more damaging to active management than the headline ever was.
The odds swing from 70% to 100% depending on where you look
Start with the claim underneath all of this. "Active management doesn't work" is really a statement about efficient markets: public information is already in the price, so hunting for mispriced stocks doesn't pay for the hunt. If that were equally true everywhere, the failure rates would look the same everywhere. Do they?
Not remotely. Take SPIVA's US scorecard for the 20 years to June 2025. Against the S&P 500, 91.03% of large-cap funds underperformed. Against the S&P/IFCI Composite, 94.03% of emerging-market funds did — worse, in the segment most often called inefficient and therefore fertile ground for stock pickers. Then look at international small-cap funds, measured against the S&P Developed Ex-U.S. SmallCap: 70.45%. Same house, same method, same window. A 24-point spread.
Europe stretches it further. SPIVA's Europe scorecard for year-end 2024 puts 10-year underperformance at 67.09% for sterling-denominated UK small-cap equity funds — and 100% for Denmark equity funds, where not one active fund beat the S&P Denmark BMI over the decade. Euro-denominated global equity funds: 97.37%.
The honest version of the efficient-markets claim is narrower than the popular one. In large, liquid, heavily researched markets the index is brutally hard to beat. In some smaller corners the failure rate drops to two in three — still losing odds, but not the same bet as nine in ten. Whether those corners are proof that skill pays in less efficient markets is a claim with twenty years of scorecard evidence behind it, most of it unkind.
Most of the funds you could have bought in 2005 no longer exist
Here's what makes SPIVA's numbers look harsher than the ones a fund company will show you. Funds die. They get liquidated or merged into other funds and, as SPIVA puts it, this "usually occurs due to continued poor performance by the fund." Counting the dead funds back in is the whole point, and the size of what survivorship bias adds to a long record has been measured at 0.94 percentage points a year over a 20-year sample.
Now suppose you compare only the funds still standing today against the index. You have quietly deleted the losers from the sample — the ones that were on the shelf, that somebody did buy, and that were closed precisely because they were bad.
SPIVA refuses to do that. It fixes the fund universe at the start of the period and uses that as the denominator, because "for someone making an investment decision at the beginning of the period, these funds are part of the opportunity set." A fund that existed in 2005, performed badly and was shut in 2013 counts as a failure. You could have bought it in 2005. You'd have lived through the closure.
The scale isn't a rounding error. Of the 2,373 US domestic equity funds alive at the start of SPIVA's 20-year window ending June 2025, only 37.29% were still there at the end. Among large-cap funds, 34.70% of the 758 starters survived. Correcting for that is what makes SPIVA's numbers worse than a fund company's — and more honest.
Morningstar puts a face on it: of the active US large-growth funds that existed two decades ago, nearly 66% closed, and just 1% both survived and outperformed their average indexed peer.
Sit with that before the next section. A fund you could have picked at the start of that window was more likely to stop existing than to still be there at the end, and the closures weren't random. They clustered on the funds that had been losing.
Nobody can actually buy an index — the best objection, tested
Critics of SPIVA have one fair point, and it deserves a serious hearing. Fund returns in SPIVA are "net of fees, excluding loads". The index they are measured against is a costless abstraction that charges nothing and trades nothing. You cannot buy the S&P 500. You can only buy a tracker, which has an expense ratio, a bid-ask spread and some tracking error. The comparison is rigged, the argument runs, by a benchmark that doesn't exist.
So what happens when you take the rigging out?
Morningstar's Active/Passive Barometer was built to find out. Instead of an index, it measures each active fund against a composite of the actual passive funds in its category — so the hurdle "reflects the actual, net-of-fees performance of investable passive funds." Its year-end 2025 edition covers 9,248 funds holding roughly USD 26 trillion, about 67% of the US fund market. Cost cuts the same universe a different way: sorting funds by expense ratio quintile produced a strictly monotone ranking of survive-and-outperform rates in Morningstar's US equity data.
Removing the objection barely moves the answer. Over the 10 years to December 2025, 21% of active strategies survived and beat their passive counterparts. The category detail is where it bites.
| Category | Survived and outperformed |
|---|---|
| All active strategies | 21% |
| US large-cap | 10% |
| US mid-cap | 22% |
| US small-cap | 25% |
Different shop, different method, a genuinely investable benchmark — same conclusion. The index fund's fee is real, and far too small to rescue the average active manager. If you're choosing a US large-cap fund, one in ten is the base rate you're playing against, and that's before you've tried to work out which one.
Weight by money rather than by fund, and active management looks better
A second objection, and it's a good one. Counting funds equally treats a $50 million fund and a $50 billion fund as one vote each, when hardly anybody's money is in the first. Weight by assets and you measure what investors actually lived through.
SPIVA publishes both, and the objection is partly right. Over the 20 years to June 2025, US large-cap active funds returned 8.77% annualised equal-weighted, but 9.48% asset-weighted. Big funds did better. Morningstar sees the same thing: over the past 10 years, the average dollar in active funds beat the average active fund in 17 of 20 categories. In aggregate, people are picking the better and cheaper ones. Sorting on fund ratings does not close the same gap, because a star rating ranks funds inside their category rather than against the index: Vanguard found 39% of 5-star funds beat their style benchmark over the next 36 months, against 46% of 1-star funds. That asset-weighted premium is also the practical answer on fund size and performance: the largest funds are the cheapest ones, and the fee discount has outweighed the scale penalty.
But the S&P 500 returned 10.73% a year over that same 20 years. Asset-weighting closes about a third of the gap — from 1.96 points a year to 1.25 — and then it stops.
Take that last figure for a walk, because a point and a quarter a year sounds like nothing at all. Say you and your neighbour start on the same morning, you in the average asset-weighted active fund and your neighbour in the index, and neither of you touches a thing for 20 years. Use $100 as an illustration, since the arithmetic scales to whatever you actually hold. At the end, for every $100 your neighbour is sitting on, you have about $80. You are short by a fifth, and no bill ever arrived.
If skill exists, can you spot it in advance?
Underperformance rates tell you the odds of a random fund losing. They don't tell you whether skill exists. Perhaps a handful of managers are genuinely good and the whole job is finding them. That's the argument keeping active management alive, and S&P built a separate document to test it. The other prerequisite is that the fund is genuinely trying: the closet indexing evidence shows a meaningful minority of active funds hold too little active share to beat anything after fees.
The U.S. Persistence Scorecard asks whether winners keep winning. Its year-end 2024 edition reports that among top-quartile domestic equity funds as of December 2020, not a single fund in any reported category was still top-quartile four years later. The year-end 2025 tables extend that run, and our evidence guide to fund performance persistence reads the full quartile transition matrices behind them.
Lower the bar to the top half and it barely improves. By pure chance, a top-half fund would stay top-half for five straight years about 6.25% of the time — a coin flip, four times over. The actual figure for all large-cap funds was 2.4%; for mid-cap funds 0.8%. Worse than random. Of the funds that beat their benchmark in 2022, a cross-category average of just 8.3% kept beating it in each of the next two years.
The same document shows where the dead funds come from: 25% of fourth-quartile US domestic equity funds were merged or liquidated within five years, against 7% of top-quartile funds. Bad performance and disappearance are the same event, separated by a few years.
Notice what this does and does not prove. Not that no manager has skill — but that past relative performance is close to useless as a way of identifying that skill in advance. That happens to be the only tool a retail investor has. It leaves open a separate question, which is what selling an underperforming fund and buying last year's winner actually costs after tax. The other number a factsheet offers, fund manager tenure, measures how much data you have on a manager rather than how good the manager is.
Where the index fund evidence against active management is weakest
The evidence that cuts the other way belongs here too, and there is some.
Fees are the clearest lever anyone has found. Over the 10 years to December 2025, Morningstar reports, 31% of cheapest-quintile active funds beat their average passive peer, against 17% in the priciest. Cost roughly doubles the strike rate. It still doesn't get you past 50%.
Bonds hold up better than shares: Morningstar's fixed-income cohort posted a 42% 10-year success rate, the highest of any group it tracks, and active intermediate-core bond managers hit 55% in 2025.
And persistence isn't uniformly below-random. S&P's Europe Persistence Scorecard for year-end 2023 found the share of funds staying top-half over four consecutive years exceeded chance in 4 of 10 reported categories — though in five of six equity categories, no top-quartile fund from 2019 stayed top-quartile for four years.
What the SPIVA scorecard evidence cannot tell you
If you want to check any of this yourself, here is exactly what it's built from — and where it stops.
SPIVA measures a specific universe: retail mutual funds, in defined style categories, against S&P's own indices. It does not cover hedge funds, separately managed accounts, or most institutional mandates. The 20-year windows ending in 2024 and 2025 contain one long, unusually concentrated US bull market in which a handful of mega-caps drove index returns — a regime unkind to a diversified stock picker almost by construction, and under no obligation to repeat.
Style drift muddies the comparison too. Only 47.70% of the domestic funds that survived SPIVA's 20-year window still had the same style classification at the end, so some are judged against a benchmark they long since stopped resembling. And an underperformance rate describes a population, not any individual fund you might be looking at. Two decades is evidence about odds, not a law.
Short-run results swing hard, which is worth seeing before you treat "90%" as a constant. In the first half of 2025, only 22.33% of US small-cap funds underperformed the S&P SmallCap 600. Over 20 years, 87.81% did. The long horizon is where the pattern lives.
What would change the conclusion
Three things, concretely. First, a sustained stretch in which the index is no longer driven by a narrow group of mega-caps: the concentration punishing diversified stock pickers would ease, and the large-cap numbers would move first. Second, evidence of persistence that beats chance across categories and a full cycle, not one region in one window — the Persistence Scorecards publish twice a year, so the test is public. Third, if the cheapest-quintile success rate, 31% over the last decade, climbed through 50% as fee compression continues, the arithmetic driving all of this would have changed.
None of that has happened yet.
So where does it leave the two people we started with? With odds, not a rule. Over 20 years in US large-cap, the base rate of picking a fund that beat the index and survived was around one in ten, and past performance was close to worthless as a way of improving those odds. Whether that bet is worth taking depends on what you think you know that the record doesn't. The judgement is yours; the base rate is not in dispute.
The number to hold onto is the annualised gap, not the win rate — the same arithmetic that governs rebalancing costs, and the same reason to check what a widely-cited dataset measures before you act on it. LedgerTouch reports cost drag alongside returns, where a 1.25-point annual gap becomes visible.