The Index Is More Active Than It Looks

Indexing is often described as the decision not to decide. I think that misses the most important part.

An index fund is passive only for the person holding it. Underneath, it is a decision to rely on prices produced by competing investors, to accept an index provider’s methodology, to own changing corporate leadership through market-cap weights, and to inherit both the market’s successes and its errors — rather than make every security-level judgment yourself. The investor makes one simple decision; the market keeps making millions underneath it. This essay is about those millions of decisions: where they come from, why they are hard to beat, and where owning their aggregate result still leaves you exposed.

Why this matters to you

If you invest for yourself, this is about what an index fund actually owns — why broad exposure is powerful, why it can still fall by half, and why the benchmark is not a financial plan. Allocation, horizon, taxes, and your own behavior still decide most of your outcome.

If you advise others, it is about a distinction the index makes unavoidable: security selection is not portfolio planning. Cheap benchmark exposure is the easy part; matching risk to a client, and keeping them invested through a drawdown, is where the value moves.

If you allocate at scale, it is about where broad beta is enough and where it is not — how benchmark construction shapes exposure, and why a liability, constraint, or benchmark mismatch has to justify the cost, model risk, and governance burden that customization adds.

By the end, the case is built from the ground up: why investors seek equity exposure, why broad ownership matters, where prices and index weights come from, why the result is so hard to beat, where indexing works best, and what it cannot solve.

The argument in four points

1. Broad ownership captures productive enterprise without requiring you to identify every future winner — which matters because most of the market’s long-run wealth comes from a handful of companies no one can reliably name in advance.
2. Index funds depend on active price discovery and designed benchmark rules. They do not exist outside active markets: they inherit prices that competing investors produce and weights that a methodology assigns.
3. Low costs and the rarity of persistent manager skill make indexing a powerful default — the average active dollar earns the market return before costs and less after, and top funds have rarely repeated at better than chance.
4. The default still inherits what the market carries: valuation, concentration, drawdowns, methodology choices, and the fact that a benchmark is not, by itself, the right portfolio for a particular investor.

Passive for whom?

“Passive investing” is a phrase doing five different jobs, and the argument only makes sense once they are separated.

The fundholder is passive: they buy a fund, add to it, and make no security-level decisions. The index provider is not passive — it maintains a methodology, applies eligibility rules, and in some indexes exercises committee judgment about membership. The fund manager implementing the index is not passive either: tracking an index through inflows, redemptions, corporate actions, and membership changes is an operational discipline. Beneath all three, active market participants — analysts, portfolio managers, quantitative funds, arbitrageurs, speculators — compete to price every security, and companies themselves compete in the real economy for customers, talent, and capital.

The investor may be passive. The system underneath is not. Everything that follows unpacks that — beginning not with the fund, but with the reason anyone wants to own equities at all.

Why own equities?

An equity is a residual claim on a productive enterprise: after a company pays its suppliers, workers, lenders, and tax authorities, whatever remains belongs to its shareholders. Own a broad slice of the corporate sector and you own the residual of the economy’s productive activity — the part left over after every prior claim is settled. That residual is volatile, which is exactly why it has been paid to bear.

The reward, when it comes, arrives through identifiable channels: corporate earnings and their reinvestment, dividends, share buybacksShare buyback (repurchase)A company buying back its own shares, which lifts per-share figures by shrinking the share count.Full definition →, and the market’s willingness to pay a higher or lower multiple for the same dollar of earnings. Behind those sit productivity growth, new products, and expansion. It is tempting to compress all of it into “a growing economy lifts stocks” — and that is where the argument for equity ownership most often goes wrong.

Economic growth and shareholder return are related but not identical, and the wedge between them is not small. Several forces drive them apart:

• investors can overpay at the start, so even strong growth delivers a poor return;
• the gains can accrue to workers, consumers, or governments rather than shareholders;
• new share issuance dilutes existing owners;
• much of an economy’s fastest growth happens inside private companies that list late or never; and
• currency and benchmark composition reshape what a shareholder actually earns.

Jay Ritter (2005) put the paradox in one number: across sixteen countries from 1900 to 2002, the correlation between real per-capita GDP growth and real equity returns was negative, −0.37. Fast-growing economies did not reward shareholders more; if anything, slightly less.

So the case for equity ownership is sturdier than “buy stocks because the economy grows”: a diversified claim on productive enterprise has, historically, been compensated for the risk it carries. If productive ownership can compound capital over time, the next question is whether history shows shareholders capturing that compounding — and how far to trust that history.

The long-run U.S. record is the most cited piece of evidence, and Jeremy Siegel assembled its most famous version, tracking the real, inflation-adjusted total return of a dollar since 1802: U.S. equities compounding near 6.8% a year, long-term government bonds 3.3%, the dollar itself losing about 1.4% a year to inflation since 1802. Over two centuries that turns one dollar in stocks into something on the order of two million in purchasing power, while the same dollar held as cash withers to about four cents. But a compounded endpoint hides how the money was actually made. Exhibit 1 plots a reconstructed 1871–2023 core of that record from Shiller’s historical monthly series and the author’s calculations — and the climb is anything but smooth.

That last caveat is not a footnote. Edward McQuarrie’s 2024 Financial Analysts Journal study, rebuilding U.S. stock and bond returns from digitized archives back to 1792, does not overturn “stocks for the long run” so much as puncture its precision. His own long-run real equity return (about 6%) is close to Siegel’s, but the pre-1871 data were thin, survivorship-prone, and appear to have understated nineteenth-century bond returns — and once broader, failure-inclusive data are used, stocks and bonds delivered roughly equal returns across long stretches of 1793–1941. The premium is regime-dependent rather than a stable constant: real over the full 227 years, but nearly gone if the best dozen are removed. Nor is the United States the neutral case — the Dimson–Marsh–Staunton data (the Triumph of the Optimists tradition, extended in the 2025 UBS Global Investment Returns Yearbook) show equities beat bonds in every one of thirty-five markets since 1900, and also that the U.S. was among the strongest, a survivor whose century should not be naively extrapolated.

Read the record for what it is: strong, internationally corroborated evidence that bearing equity risk has been rewarded over the long run — the equity case. It is not proof that any particular index is optimal, that equities are safe at every horizon, or that the next century resembles the last. The record supports owning productive enterprise broadly; it does not, by itself, prove that market-cap indexing is the way to do it. That case has to be built, and the rest of this essay builds it.

Growth, cycles, and creative destruction

Compounding happened through the crashes and the churn — and a broad index quietly re-composes itself around both, with no forecast required.

The drawdowns in Exhibit 1 are not noise around the trend line; they are how the two centuries were actually lived. That jagged equity path ran through 1873, 1929 and the Depression, the 1970s inflation, 2000, and 2008 — expansions and contractions, wars and policy shocks, valuations stretching and snapping back. And the order of those years matters: a long-run average is a destination, while sequence-of-returns risk is the journey, and drawing down through a crash is a different experience from accumulating into one.

The distinction is easy to get backwards. Indexing does not work because markets have bull and bear phases; long-run compounding occurred through them, as the growth of earnings, reinvestment, and successful innovation outweighed repeated failures and contractions in the record observed. A bear market is not the engine of the return — it is the systematic risk being borne, the thing diversification cannot remove and the reason the reward existed at all. So terminal wealth can be enormous and the path can still include losses deep enough to end a plan, or a nerve; horizon, withdrawals, and behavior decide which an investor lives.

And the composition of the winners kept changing underneath it. Look at what the market owned across those two centuries and the churn is astonishing. In 1900, railroads were roughly 63% of U.S. stock-market value; today they are under 1%, and about four-fifths of 1900’s market capitalization sat in industries now small or gone. The sectors that dominate the index now — technology, healthcare, much of modern energy — were essentially absent then. Joseph Schumpeter named the process in 1942: the “perennial gale of creative destruction,” the incessant industrial mutation that revolutionizes the economic structure from within, destroying the old arrangement as it builds the new one. Only five of the hundred largest U.S. companies of 1917 were still in the top hundred decades later; roughly half the top hundred of 1970 had been replaced by 2000.

For a broad, cap-weighted index this churn is not a problem to be managed but a property that comes for free. New companies list and, if they succeed, rise in weight; declining companies shrink and eventually drop out; the index’s economic composition drifts toward whatever the economy is actually rewarding, with no analyst forecasting the transition. That is the deep reason a broad index can participate in innovations that were unimaginable when an investor first bought it.

That churn has an internal twin, one that runs through the shareholder’s own holdings. When a company is acquired its owners receive stock in the buyer; when one spins off a division they receive the spin-off — the name changes while the economic claim survives inside the successor. Jeremy Siegel and Jeremy Schwartz (2006) measured this on the original S&P 500: they held the 500 companies in the index as of March 1957 and kept only what those firms became — every merger share, every spin-off — adding none of the roughly thousand names the index took on afterward. That “totally passive” portfolio of the originals compounded at 11.40% a year through 2003 against 10.85% for the continuously updated index, and beat it in nine of ten sectors with a higher risk-adjusted return. Not every old firm made it — Bethlehem Steel went bankrupt, and U.S. Steel, General Motors, and Alcoa lagged badly — but corporate disappearance is rarely shareholder extinction: Mobil’s owners became ExxonMobil’s, and a line of food companies folded into Philip Morris. It is one more layer of activity working beneath a decision to sit still — operated by no index holder, and noticed by almost none.

But innovation’s gift to investors is far less reliable than its gift to society. Transformative technology routinely enriches consumers and workers while punishing the shareholders who financed it: new industries over-invest and over-compete — nineteenth-century railroads were built far past what traffic could support, and after the 1893 panic companies owning a quarter to a third of U.S. rail mileage went bankrupt. The cleanest modern case is aviation, which Warren Buffett in 2007 called a bottomless pit of capital with no durable advantage “ever since the days of the Wright Brothers.” (Tellingly, the surviving railroads still beat the market even as the industry shrank — so the lesson is over-building, not that progress is bad for owners.)

Innovation, then, explains how the pie grows and how the index quietly re-composes itself around the growth. It does not tell an investor which company will capture that growth, keep its margins, avoid dilution, survive competition, and not already have the outcome priced in. No one solves that forecasting problem reliably — which is precisely where diversification stops being a preference and becomes a necessity.

The rare-winner problem

The strongest quantitative argument for broad ownership is also the least intuitive. Long-run individual-stock outcomes are extremely positively skewed: most stocks do poorly over their lifetimes, and a thin sliver does so well that it carries the whole market. The asymmetry is structural — a stock can lose at most everything, while a compounder’s upside has no ceiling, so a few enormous winners can outweigh armies of failures.

Hendrik Bessembinder (2018) measured this against the full record of the Center for Research in Security Prices (CRSP), the academic registry of U.S. common stocks, from 1926 through 2016:

• Only 42.6% of stocks beat one-month Treasury bills over their full lifetimes; the single most common lifetime outcome was a total loss.
• The best-performing ≈4% (1,092 of roughly 25,300 companies) account for the entire net wealth creation of the U.S. stock market above Treasury bills.
• Just 90 companies — about a third of one percent of all that ever listed — account for over half of that net wealth creation.

These aggregates are reproduced from the paper rather than recomputed: the underlying data are CRSP-licensed, and this site does not publish charts it cannot regenerate from source. The portfolio implication is sharp. If a handful of companies you cannot reliably name in advance will produce most of the market’s wealth, then any concentrated portfolio’s dominant risk is missing them entirely — and holding a single randomly chosen stock underperformed the market in Bessembinder’s simulations 96% of the time. Broad ownership converts that miss-risk into participation: from the point a winner enters the index, the portfolio holds it, and cap weighting grows the position on its own. Business success and failure feed the weights automatically, with no one voting on which companies deserve to matter.

The claim should not be over-extended. A broad index does not capture every winner from the start of its life: companies IPO late, spend years outside eligibility rules, list abroad, get acquired before inclusion, or do their fastest compounding as small caps outside a large-cap index. A total-market fund narrows those gaps; nothing eliminates them. The rare-winner argument is a case for breadth, not a promise of omniscience — and breadth is where portfolio theory begins.

Diversification and portfolio construction

Diversification is the portfolio response to that uncertainty, and its logic predates index funds by decades. Modern Portfolio TheoryModern Portfolio Theory (MPT)Markowitz's framework: judge every holding by what it does to the whole portfolio.Full definition → (MPT), from Markowitz (1952), reframed the problem: the unit of decision is not the security but the portfolio, and a portfolio’s risk depends not only on how volatile each holding is but on how the holdings move together — their correlationCorrelationHow closely two assets move together, from −1 to +1. The engine of diversification.Full definition →, or in the raw units the mathematics uses, their covarianceCovarianceCorrelation with the units left in — the raw input portfolio risk math actually uses.Full definition →.

The practical consequence is diversificationDiversificationSpreading capital across holdings so no single company's fate decides the outcome.Full definition →. Combine holdings that do not move in lockstep and the portfolio’s swings are smaller than the average of its parts, because company-specific misfortunes partly offset. Push that across enough holdings and idiosyncratic riskIdiosyncratic riskCompany-specific risk — the part diversification can actually eliminate.Full definition → — the failed product, the fraud, the lost contract — shrinks toward zero.

What remains is systematic riskSystematic riskMarket-wide risk that diversification cannot remove — the risk that gets paid.Full definition → — market risk, in plain terms: recessions, rate shocks, crises, the movements that hit nearly everything at once. No amount of diversification removes it. A diversified portfolio in a bear marketBear marketA decline of 20% or more from a peak.Full definition → goes down. Exhibit 2 puts numbers on that boundary.

The closed form behind the curves, for readers who want it, is

σ_p = σ · √( 1/N + (1 − 1/N) · ρ )

where σ_p is the portfolio’s volatility, σ is each asset’s volatility, N is the number of holdings, and ρ is the correlation each pair of holdings shares. In words: volatility falls as holdings are added, but never below a floor set by the correlation the holdings share. At ρ = 1 the expression equals σ for every N — the flat line — and lower correlation lowers the floor the curve approaches.

Diversification tells you to own many things. Markowitz’s framework goes further: it defines the set of portfolios that cannot be bettered — more expected return only by taking more volatility, less volatility only by giving up return — the efficient frontierEfficient frontierThe set of portfolios offering the most expected return for each level of risk.Full definition →. Exhibit 3 sketches it under stated assumptions, and, just as important, marks what the theory refuses to prove.

Two honest caveats belong here, because MPT is routinely made to promise more than it does. First, everything in the framework runs on estimates of expected returns, volatilities, and correlations, drawn from a history that does not sit still — and an optimizer that takes the estimates literally produces portfolios that look precise and are not. Second, MPT supports diversification and covariance-aware construction; it does not prove that the S&P 500, or any particular index, is the optimal portfolio, or that mean and variance capture everything an investor cares about.

Asset-pricing theory pushes the logic to its limit. If company-specific risk can be shed for free, bearing it should earn nothing; the only risk left to be paid is the kind diversification cannot remove. The Capital Asset Pricing ModelCapital Asset Pricing Model (CAPM)The model tying an asset's expected return to its beta against the market portfolio.Full definition → (CAPM), from Sharpe (1964), turns that into an equilibrium in which an asset’s expected return above the risk-free rate depends only on its betaBetaHow much an asset tends to move per 1% move in the market.Full definition → — its exposure to the market portfolioMarket portfolioTheory's portfolio of all investable assets, weighted by value. No real fund holds it.Full definition →. Three qualifications matter. That market portfolio is not the S&P 500: it contains all assets — stocks, bonds, real estate, private businesses, even human capital — so a public-equity index is only a proxy, which is Roll’s (1977) point that a test against a proxy is a test of the proxy. And decades of factor evidence (Fama–French, and its own fragility to publication and replication — McLean and Pontiff, 2016; Hou, Xue, and Zhang, 2020) show average returns varying in ways one market beta does not explain. The model’s vocabulary has survived better than its predictions. Portfolio theory says to own a broad market portfolio; it does not say where the prices to build it come from.

Prices are produced, not given

An index fund inherits its prices; it never makes them.

Nobody hands an index fund a list of good businesses. What it receives is prices, and prices are produced — by fundamental analysts revising earnings models, portfolio managers sizing positions, quantitative funds trading signals, executives buying back stock or issuing it, lenders repricing credit, customers and suppliers shifting the fundamentals, arbitrageurs closing gaps, and speculators taking the other side. Price discoveryPrice discoveryThe market process that turns competing judgments and trades into a price.Full definition → is the name for what all of that adds up to: a market price is the negotiated output of disagreement, continuously revised — not a declaration of objective truth.

This is where the efficient-market hypothesisEfficient Market Hypothesis (EMH)Prices reflect available information because investors compete to act on it; hard to beat doesn't mean always right.Full definition → (Fama, 1970) matters — and where it does not. Efficiency is not an assumption that investors are omniscient or agree; it is an outcome of their competition: they seek information, form differing expectations, and trade; prices adjust; and visible mispricing attracts further research and capital. What indexing needs is not that prices be right — they are demonstrably not always right, as Shiller (1981) showed from volatility alone and this site’s Shiller CAPE note maps across 155 years — but that they are hard to beat: they embody an enormous quantity of competing, capital-backed analysis, and outdoing it persistently, after costs, is rare. Grossman and Stiglitz (1980) sharpen this to a paradox: prices could never be perfectly informative, because then no one would be paid to produce the research that makes them so. Markets are partly self-correcting, not perfectly — cost, disagreement, risk, and the limits of arbitrage keep every error from vanishing at once. Active research is not the enemy of indexing; it is the input.

It is also a resource indexing consumes without paying for, which raises the obvious question: does price quality erode as the passive share climbs? The evidence is genuinely mixed — Coles, Heath, and Ringgenberg (2022) find index investing reduces information production while leaving price informativeness unchanged, while Sammon (2024) finds rising passive ownership has measurably slowed how fast prices incorporate information ahead of earnings. The Grossman–Stiglitz logic says the damage should be partly self-limiting, because thinner competition makes mispricing larger and analysis better paid; how far the passive share can rise before that strains is an open question this essay does not settle. What is settled is the direction of dependence: index investors do not produce these judgments — they inherit the prices produced by participants who do.

From Bogle to benchmark rules

For most of the twentieth century, all of the above was academic and institutional; the first working index portfolios were run for pensions, not people. Around 1971, Wells Fargo’s Management Sciences group — with academic consultants including Eugene Fama and Myron Scholes — built an equal-weighted account tracking every NYSE stock for the Samsonite pension, and American National Bank of Chicago and Batterymarch were building institutional index accounts in the same years. Indexing already existed; what it lacked was a door ordinary investors could walk through.

John Bogle built that door. He founded Vanguard in 1975 and, on August 31, 1976, launched the First Index Investment Trust — the first index mutual fund aimed at individuals. Its underwriting was, in his own word, “an abject failure”: against a target near $150 million it raised a little over $11 million. Bogle did not invent indexing — he freely credited the academics, casting Paul Samuelson as “the Professor” and himself as “the Student” — but he made it buyable, wrapped in Vanguard’s client-owned, at-cost structure, in which the funds own the manager and scale flows back to fundholders as lower fees. He turned a theory about the market portfolio into something a household could own. The product was simple; its construction still required rules deciding which securities entered, and in what proportion.

“The index” is not a fact of nature; it is a product with a methodology, and the methodology is where the discretion lives. The chain from analysis to weight runs in two streams, and Exhibit 4 draws both.

Providers set eligibility rules — domicile, listing, size, liquidity, float minimums, sometimes profitability — apply float adjustmentFloat-adjusted market capitalizationMarket value counting only the shares the public can actually trade.Full definition → so weights track what is buyable, and run membership upkeep: scheduled reconstitutionIndex reconstitutionHow index membership is re-formed under published rules — scheduled in some families, as-needed in others.Full definition → in some families, as-needed changes in others (the S&P U.S. indices have no fixed reconstitution date), with a committee retaining explicit discretion in the S&P family. The S&P 500 is, precisely, a methodology-governed, float-adjusted, committee-overseen large-cap index — a designed object approximating the U.S. large-cap market, not a photograph of it. Two consequences follow. Index changes are not verdicts on business quality — companies leave through acquisition, spin-off, or reclassification as much as decline, and additions are eligibility events, not endorsements (though once priced ones: S&P 500 addition was worth roughly a 3% abnormal return (a price jump beyond what the market alone explains) in the classic studies of Shleifer and of Harris and Gurel (1986), a premium Greenwood and Sammon (2025) show has since collapsed toward zero). And indexes are not interchangeable: market capitalization, valuation, fundamentals, and eligibility are related but distinct, so “it’s an index fund” is the beginning of due diligence, not the end.

The index as an adaptive portfolio

Return now to the innovation thread. Market-cap weightingMarket-capitalization weightingWeighting holdings by company value, so prices — not a manager — set the portfolio.Full definition → — in practice float-adjusted — has a property no other scheme shares: the portfolio’s weights update themselves. When a holding’s price rises, its weight rises with it. No trade occurs, no rebalance is triggered; the market has simply revalued the same shares. Exhibit 5 walks the mechanism through three fictional companies.

The consequences follow directly. A cap-weighted fund never has to sell a position merely because it appreciated, which keeps turnover and its costs low; winners are allowed to run, so a company that compounds for a decade grows from a small weight into a dominant one without the fundholder identifying it; and decliners shrink themselves out of relevance on the same mechanism. This is how the index adapts to creative destruction without a macro forecast — it changes after investors and businesses reveal which companies are capturing the future, not before. It is also why a broad index contains companies displaying every characteristic the factor literature has named — value and growth, winners and losers — without targeting any of them: containing is not targeting, and broad ownership is not a factorFactor investingDeliberately tilting a portfolio toward traits linked to differences in average returns.Full definition → bet.

The ledger has another side, and adaptation names it. Because weight follows price, cap weighting assigns its largest weights after prices have risen; it is not valuation-neutral, and it can ride optimism and mispricing upward into real concentration. Winner-retention is emphatically not the academic momentum factorMomentum factorBuying recent winners and selling recent losers — a deliberate sort, not a side effect.Full definition → of Jegadeesh and Titman (1993), which deliberately ranks stocks on past returns and rebalances on a schedule; cap weighting sorts nothing and rebalances nothing, it simply declines to interfere. But “declines to interfere” cuts both ways: the mechanism that captures genuine winners also increases exposure to whatever the market currently prizes, exactly when it is most richly priced. No current market shows that double edge more clearly than artificial intelligence — worth a brief look as a case study, not a forecast. What is instructive about the AI build-out is not whether the valuations are right (this essay takes no view) but that it makes the geography of a modern value chain visible, and with it both the strength and the limit of indexing.

A single AI accelerator is a relay across borders. U.S. firms design the chips — NVIDIA’s data-center revenue hit $193.7 billion in the year ended January 2026 — but Taiwan’s TSMC fabricates and packages them (advanced nodes were 74% of its 2025 wafer revenue), South Korea’s SK Hynix and Samsung and U.S.-based Micron supply the memory stacked beside them, and none of it is possible without the Netherlands’ ASML, the sole maker of EUV lithography, or Japan’s equipment suppliers. No one country owns the chain, and each link is a single point of dependence. The capital is concentrated too: the largest U.S. hyperscalers — the giant cloud buyers, Amazon, Microsoft, Alphabet, and Meta — spent on the order of $400 billion of capital expenditure in 2025, and Taiwan’s exports jumped 34.9% that year to a record $640.7 billion.

Two disciplines matter for an indexing essay. First, the boom is real but narrow: within the same Taiwanese trade data that show semiconductors surging, base metals, chemicals, and plastics fell — the AI wave did not lift all boats, and “the market” gains are really a handful of names. As of August 4, 2026, the ten largest S&P 500 constituents were 38.01% of the index, NVIDIA alone about 7.7%. Second, and this is the whole point: economic gains are not the same as investor returns, and are not evenly shared. A cap-weighted benchmark quietly concentrates its holder into that same handful of AI-linked names — not because of any view that “AI is good for everything,” but because price times shares put them there. The benchmark an investor chooses decides which slice of a global transformation they actually own, and that concentration is itself the risk: a repricing of a few names now moves the whole index. An index adapts as the value chain evolves; it does not protect its holder from the possibility that the adaptation has run ahead of the earnings.

Why active management is difficult

After watching an index delegate both selection and reweighting, the natural question is why not improve on it by choosing actively. The first answer is arithmetic. Sharpe’s 1991 note “The Arithmetic of Active Management” is an identity, not a forecast: active and passive investors together own the market, the passive segment holds it in market proportions and so earns the market return before costs, and therefore the aggregate of all active positions must also earn the market return before costs — after which the higher fees, trading costs, and taxes of active management leave the average active dollar behind. Exhibit 6 draws the ordering.

The identity binds the average, not the tails, and it leans on its definitions — “passive” meaning holding the market in market proportions. Pedersen (2018) sharpens the objection: real index portfolios must trade, so the equality is exact only in a market that never changes. What the arithmetic settles in the strict case, evidence must establish in the real one — and the survivorship-corrected SPIVA scorecards (S&P Indices Versus Active) do. In the Year-End 2025 U.S. scorecard, 78.78% of large-cap funds trailed the S&P 500 over one year, 85.59% over ten, and 92.89% over twenty. Those are fund counts; the asset-weighted companion, letting the average dollar speak, shows active doing better than the average fund — it edged the index in 2025 — while still trailing over the decade by about 1.2 points a year. And the winners rarely repeat: of the large-cap funds in their category’s top quartile in 2021, about a fifth held it in 2022 and none across 2023–2025 — at or below the roughly 0.4% a coin-flip predicts for four straight top-quartile years.

Read the evidence for what it is: an empirical hurdle, not an impossibility theorem. The arithmetic does not say every active manager loses; skill can exist, and some managers clear the hurdle for long stretches. What the base rates describe is the bet an investor accepts when paying active fees — and how much harder “identify the persistent winner beforehand” is than “observe that winners existed.” Observing skill afterward is easy. Buying it beforehand is the hard problem.

Where indexing works — and where it strains

Indexing is only as good as the market underneath it.

Everything above is really a set of conditions. Broad, low-cost ownership represents a market well when:

• the market is broad and diverse in its listed companies;
• meaningful free float is available to buy;
• competitive price discovery, reliable reporting, and strong investor protections make prices worth inheriting;
• liquidity is deep, and trading, custody, and implementation frictions are low; and
• a rules-based benchmark reasonably represents the economy it tracks.

Indexing does not need markets to be perfectly efficient — the argument has conceded they are not — only competitive, investable, and diversified enough that owning all of it cheaply is a good way to own the opportunity set. The developed large-cap markets clear that bar comfortably. Not every market does.

“Emerging markets” is not one thing, and the honest analysis is about boundary conditions rather than a verdict. The two dominant index providers make the developed-versus-emerging line an explicit, rules-based judgment about market accessibility, not just size: MSCI grades every market on economic development, size and liquidity, and a market-accessibility score built from eighteen measures, while FTSE Russell runs a parallel twenty-two-criterion “quality of markets” matrix. Tellingly, both relax their requirements for lower tiers — FTSE demands fair minority-shareholder treatment and the absence of foreign-ownership limits only for its top classifications, formally tolerating their absence below. The line is a defensible convention, not a fact of nature: FTSE classifies South Korea as developed while MSCI still calls it emerging.

Those relaxed conditions are exactly the ones that degrade a cap-weighted index. Foreign-ownership limits (Saudi Arabia and the UAE at 49%, China A-shares capped at 30%) wall off part of the market; non-convertible currencies and settlement frictions raise the cost of holding it; large strategic and state owners thin the float and blunt minority-shareholder rights. The result is a benchmark that maps poorly to its economy and concentrates hard: as of December 30, 2025, the MSCI Emerging Markets index held 76.8% of its weight in just four countries — and Russia, which had been as much as ~6.5% of it around 2009, was reclassified to a standalone zero in 2022 — a clean illustration of governance and political tail risk arriving all at once.

The tempting inference — that inefficient markets must therefore be easy to beat actively — does not follow. Greater apparent inefficiency is necessary but not sufficient for realizable alpha — the outperformance an investor can actually capture after costs: research is costlier, information access is uneven, trading costs and capacity limits bite, governance risk can overwhelm valuation work, and capital controls can prevent implementing a good idea at all. SPIVA’s emerging-market results swing sharply by window — a majority of funds beat the benchmark in the first half of 2025, and a majority trailed in other periods — which is the signature of noise, not a dependable edge. An emerging-market index remains useful exposure; it simply owns a narrower, more concentrated, less liquid, and more politically shaped slice of activity than a developed-market index does, and that is a boundary condition worth pricing rather than a reason to abandon the tool.

What indexing cannot solve

None of the above makes an index fund safe, and this section is part of the argument, not a disclaimer: a tool is only understood when its failure modes are. The ledger, in one view:

Indexing can reduce	Indexing still inherits
Company-specific risk	Market-wide risk and drawdowns
Dependence on any one manager or forecast	Valuation and bubble risk
Fees, turnover, and research burden	Concentration in whatever has risen
The risk of missing the rare big winners	Sequence, currency, and country risk
Security-selection error	Index-methodology and benchmark choices

The bull-and-bear mechanism deserves stating precisely. An index participates in the expansion and contraction of the market it owns; it does not distinguish a justified bull market from a speculative one, and it does not exit merely because valuation looks high. It holds the market’s judgment all the way up and all the way down — not an implementation error, but what owning the market means. Whether today’s prices are stretched is an empirical question the 155-year CAPE record is this site’s map of; a bear market does not invalidate indexing, it reveals the systematic risk that diversification across companies cannot remove.

The rest of the ledger the table cannot carry. Narrow, thematic, leveraged, and inverse products carry the word index while abandoning most of what this essay argues for. Indexing concentrates voting power — by 2019 the three largest index managers cast roughly a quarter of the votes at S&P 500 companies, and Bebchuk and Hirst (2019) argue their incentive to steward is weak. And the largest failure modes are the investor’s own: dollar-weighted returns have persistently trailed the funds investors held (Morningstar’s “Mind the Gap”), home bias concentrates a portfolio in one economy, a retiree drawing down through a crash faces sequence risk a long accumulator does not, and the historical record itself is a survivor’s — the U.S. was one of the most successful markets in the global sample (Dimson, Marsh, and Staunton, 2002). Above all, a benchmark is not a financial plan: an excellent index fund can be the wrong portfolio for a particular liability, horizon, or risk tolerance, because matching the market is a virtue only for an investor whose objective is, in fact, the market.

Those failure modes are also why the investors best equipped to buy the whole market often, deliberately, do not — not because they think indexing fails, but because their objectives differ. A pension owes defined benefits on a schedule; an insurer holds regulated capital against liabilities; an endowment runs a spending policy; a family office carries legacy positions and tax constraints. Objectives like these arrive with duration targets, liquidity needs, currency exposures, and drawdown limits a market benchmark knows nothing about. So they customize:

• liability-driven investing that matches assets to obligations;
• currency hedging;
• factor tilts inside explicit risk budgets;
• completion portfolios around positions that cannot be sold;
• private markets; and
• tax-aware implementation.

The customization answers their constraints, not a defect in indexing — and the public-equity sleeve inside these structures is, often enough, itself indexed. But it carries its own bill: higher fees, model and manager-selection risk, governance burden, illiquidity, and more room for committee error, since false precision is a risk category of its own. Customization earns its place only when a defined objective the index cannot meet outweighs that bill.

My conclusion

Trace the whole inquiry and the case assembles itself. Equity ownership has historically rewarded bearing the risk of productive enterprise — through innovation, reinvestment, failure, and cycles, not instead of them. Because a few unnameable winners carry the result, diversification and broad ownership are the rational response to not knowing which they are. Competing investors produce the prices; index methodology turns those prices into weights; Bogle made the structure cheap and buyable; and cap weighting lets the portfolio evolve as corporate leadership changes. The method still inherits valuation, concentration, market, behavioral, and benchmark risk; its strength varies with the quality of the market underneath it; and institutions rightly customize when their objectives differ from generic wealth accumulation. What persuaded me is not any single theory but the convergence — portfolio theory, the pricing machinery, the cost arithmetic, the skew of long-run outcomes, and the scorecards all point the same way independently.

So here is where I land. I do not think indexing is powerful because markets are always right. I think it is powerful because future winners are genuinely hard to identify, competing investors make prices hard to beat, and costs are one of the few investment variables an investor controls with certainty. For an individual, that makes a broad index a strong default — not a complete plan. For an advisor, it shifts the work from predicting securities toward building a portfolio the client can actually hold. For an institution, it sets the benchmark that customization has to justify leaving. Indexing is not the absence of active judgment; it is the decision to own the aggregate outcome of millions of judgments, cheaply, while accepting that the market’s mistakes, concentrations, and drawdowns come with it. That is why I think it is perhaps the simplest sophisticated strategy in investing — and why I would still never mistake it for a financial plan.

It also fits how the work on this site is organized. In the framework, economics describes the environment; fundamentals evaluate businesses; quantitative analysis tests the evidence; technical analysis reads the market’s revealed judgment — the same judgment cap weighting turns into portfolio weights. Indexing, seen that way, is a decision to own the pooled output of everyone else’s competing analysis rather than to make every judgment personally. That does not retire active research — Grossman and Stiglitz’s point stands, and the process this site follows is active research.

Method and caveats

This essay is a synthesis of existing theory and published research — it contains no original empirical test, which is why it is labelled a research essay rather than a research note. Exhibit 1 is a reconstructed real-wealth path derived from Robert Shiller’s Yale series (S&P composite price, dividends, CPI, and the long Treasury yield, 1871–2023); the ~10-year Treasury series is an author-constructed approximation. The frozen derived series and chart generator ship with the note, but the original workbook-to-CSV transformation pipeline does not, so the note does not claim full source-to-output reproducibility. Exhibits 2–6 are the author’s constructions from stated closed forms, stated assumptions, or fictional-by-design examples, with their generator code and computed inputs shipped here. The Bessembinder figures are quoted from the paper’s published aggregates rather than recomputed, because the underlying CRSP data are licensed. Current-market figures carry their individual as-of dates, the most recent being August 4, 2026.

Sources · 35 references

Theory

• Markowitz, Harry. “Portfolio Selection.” The Journal of Finance, vol. 7, no. 1, 1952, pp. 77–91. https://doi.org/10.1111/j.1540-6261.1952.tb01525.x
• Markowitz, Harry. “Foundations of Portfolio Theory.” Nobel Lecture, December 1990. https://www.nobelprize.org/prizes/economic-sciences/1990/markowitz/lecture/
• Sharpe, William F. “Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk.” The Journal of Finance, vol. 19, no. 3, 1964, pp. 425–442. https://doi.org/10.1111/j.1540-6261.1964.tb02865.x
• Sharpe, William F. “The Arithmetic of Active Management.” Financial Analysts Journal, vol. 47, no. 1, 1991, pp. 7–9. https://doi.org/10.2469/faj.v47.n1.7; author-hosted copy: https://web.stanford.edu/~wfsharpe/art/active/active.htm
• Pedersen, Lasse Heje. “Sharpening the Arithmetic of Active Management.” Financial Analysts Journal, vol. 74, no. 1, 2018, pp. 21–36. https://doi.org/10.2469/faj.v74.n1.4
• Roll, Richard. “A Critique of the Asset Pricing Theory’s Tests, Part I.” Journal of Financial Economics, vol. 4, no. 2, 1977, pp. 129–176. https://doi.org/10.1016/0304-405X(77)90009-5
• Fama, Eugene F. “Efficient Capital Markets: A Review of Theory and Empirical Work.” The Journal of Finance, vol. 25, no. 2, 1970, pp. 383–417. https://doi.org/10.1111/j.1540-6261.1970.tb00518.x
• Grossman, Sanford J., and Joseph E. Stiglitz. “On the Impossibility of Informationally Efficient Markets.” The American Economic Review, vol. 70, no. 3, 1980, pp. 393–408. https://www.aeaweb.org/aer/top20/70.3.393-408.pdf
• Schumpeter, Joseph A. Capitalism, Socialism and Democracy. Harper & Brothers, 1942 (ch. “The Process of Creative Destruction”).

Historical returns and equity ownership

• Siegel, Jeremy J. “The Equity Premium: Stock and Bond Returns Since 1802.” Financial Analysts Journal, vol. 48, no. 1, 1992, pp. 28–46. https://doi.org/10.2469/faj.v48.n1.28
• Siegel, Jeremy J., and Jeremy D. Schwartz. “Long-Term Returns on the Original S&P 500 Companies.” Financial Analysts Journal, vol. 62, no. 1, 2006, pp. 18–31 (the “Total Descendants” study; see also Siegel, The Future for Investors, Crown Business, 2005). https://doi.org/10.2469/faj.v62.n1.4055
• Siegel, Jeremy J. Stocks for the Long Run. 6th ed., McGraw-Hill, 2022.
• McCaffrey, Paul. “Stocks for the Long Run? New Evidence, Old Debates.” CFA Institute Research Foundation, 2025 (reproducing Siegel’s 6th-edition figures and summarizing the McQuarrie critique). https://rpc.cfainstitute.org/research/foundation/2025/stocks-for-long-run-new-evidence-old-debates
• McQuarrie, Edward F. “Stocks for the Long Run? Sometimes Yes, Sometimes No.” Financial Analysts Journal, vol. 80, no. 1, 2024, pp. 12–28. https://doi.org/10.1080/0015198X.2023.2268556
• Ritter, Jay R. “Economic Growth and Equity Returns.” Pacific-Basin Finance Journal, vol. 13, no. 5, 2005, pp. 489–503. https://doi.org/10.1016/j.pacfin.2005.07.001
• Buffett, Warren E. Berkshire Hathaway Chairman’s Letter, 2007. https://www.berkshirehathaway.com/letters/2007ltr.pdf
• Dimson, Elroy, Paul Marsh, and Mike Staunton. Triumph of the Optimists: 101 Years of Global Investment Returns. Princeton University Press, 2002; and UBS Global Investment Returns Yearbook 2025 (summary edition).

Empirical evidence

• Bessembinder, Hendrik. “Do Stocks Outperform Treasury Bills?” Journal of Financial Economics, vol. 129, no. 3, 2018, pp. 440–457. https://doi.org/10.1016/j.jfineco.2018.06.004
• Fama, Eugene F., and Kenneth R. French. “Common Risk Factors in the Returns on Stocks and Bonds.” Journal of Financial Economics, vol. 33, no. 1, 1993, pp. 3–56. https://doi.org/10.1016/0304-405X(93)90023-5
• Jegadeesh, Narasimhan, and Sheridan Titman. “Returns to Buying Winners and Selling Losers.” The Journal of Finance, vol. 48, no. 1, 1993, pp. 65–91. https://doi.org/10.1111/j.1540-6261.1993.tb04702.x
• McLean, R. David, and Jeffrey Pontiff. “Does Academic Research Destroy Stock Return Predictability?” The Journal of Finance, vol. 71, no. 1, 2016, pp. 5–32. https://doi.org/10.1111/jofi.12365
• Hou, Kewei, Chen Xue, and Lu Zhang. “Replicating Anomalies.” The Review of Financial Studies, vol. 33, no. 5, 2020, pp. 2019–2133. https://doi.org/10.1093/rfs/hhy131
• Shiller, Robert J. “Do Stock Prices Move Too Much to Be Justified by Subsequent Changes in Dividends?” The American Economic Review, vol. 71, no. 3, 1981, pp. 421–436.
• Coles, Jeffrey L., Davidson Heath, and Matthew C. Ringgenberg. “On Index Investing.” Journal of Financial Economics, vol. 145, no. 3, 2022, pp. 665–683. https://doi.org/10.1016/j.jfineco.2022.05.007
• Sammon, Marco. “Passive Ownership and Price Informativeness.” Management Science, 2024. https://doi.org/10.1287/mnsc.2023.00836
• Shleifer, Andrei. “Do Demand Curves for Stocks Slope Down?” The Journal of Finance, vol. 41, no. 3, 1986, pp. 579–590; and Harris, Lawrence, and Eitan Gurel, same issue, pp. 815–829.
• Greenwood, Robin, and Marco Sammon. “The Disappearing Index Effect.” The Journal of Finance, vol. 80, no. 2, 2025, pp. 657–698. https://doi.org/10.1111/jofi.13410
• Bebchuk, Lucian, and Scott Hirst. “Index Funds and the Future of Corporate Governance.” Columbia Law Review, vol. 119, no. 8, 2019, pp. 2029–2146. https://columbialawreview.org/content/index-funds-and-the-future-of-corporate-governance-theory-evidence-and-policy/
• S&P Dow Jones Indices. SPIVA U.S. Scorecard, Year-End 2025 (Report 1a and Report 4) and U.S. Persistence Scorecard, Year-End 2025 (Report 2). https://www.spglobal.com/spdji/en/spiva/article/spiva-us/
• Morningstar Research. Mind the Gap — annual studies of investor dollar-weighted versus fund returns. Cited qualitatively.

Indexing history, methodology, and market structure

• Bogle, John C. “The Professor, the Student, and the Index Fund,” 2011. https://johncbogle.com/; and The Vanguard Group corporate archive, “Indexing since 1976.” https://corporate.vanguard.com/
• S&P Dow Jones Indices. S&P U.S. Indices Methodology, July 2026 edition. https://www.spglobal.com/spdji/en/documents/methodologies/methodology-sp-us-indices.pdf
• MSCI. Market Classification Framework and Global Market Accessibility Review, 2025. https://www.msci.com/index-methodology
• FTSE Russell. Equity Country Classification Process, 2026. https://www.lseg.com/en/ftse-russell
• NVIDIA, TSMC, SK Hynix, ASML, and the major U.S. hyperscalers, fiscal-2025/2026 results; Taiwan Ministry of Finance, “Annual External Trade Report 2025”; State Street SPDR S&P 500 ETF holdings (index-concentration figure, as of August 4, 2026). Company and official filings, cited with as-of dates.

---

Ken Capital is an independent investment-research portfolio. This essay is personal research and not investment advice.