Economics / Market structure
When the Price Becomes the Argument
A rising price can be evidence that a business is becoming more valuable. It can also become a reason for more money to follow it.
By [Alias]
November 7, 2026
Those two processes can coexist for a long time. Earnings improve, investors respond, prices rise. Then some of the response acquires a life of its own: the allocation that performed well attracts another contribution; the benchmark becomes harder to depart from; yesterday’s result becomes part of tomorrow’s demand.
Where does recognition end and reinforcement begin?
Persistent valuation-insensitive inflows can help push a concentrated market beyond plausible cash-flow and discount-rate assumptions when counterparty capacity is limited, while the gains they support can reinforce common exposures and magnify losses if those conditions reverse.
The hypothesis this piece pursues. Not a forecast.
That is the hypothesis worth pursuing. Not that passive investing inevitably produces a bubble, or that a quiet market is preparing a punishment. Rather, a market can become dependent on a pattern of demand while interpreting the resulting prices as independent confirmation of its judgment.
It is an economic question with a philosophical nuisance attached: how much confidence should we place in a signal that our own behavior is helping to produce?
Start with the money that actually moves
The total assets of passive funds are not the same thing as new demand. Their value can rise because existing holdings appreciate, without another subscription arriving. A very large stock of patient ownership can generate very little trading.
Index funds grew inside the fund industry, not to half the equity market
Index funds’ 52% share is a share of long-term fund assets across all asset types, not equity-market ownership. The four long-term-fund observations are sparse points, not an interpolated series. The second panel covers domestic-equity mutual funds only, excluding ETFs and CITs.
Source: Investment Company Institute, Figure 2.5 and Table 42. Units: share (%).The relevant starting point is incremental net flows and the transactions they require. Even then, a subscription to an index fund may be financed by withdrawing from an active equity fund. What was acquired matters, but so does what was relinquished. Net inflows to a fund are not automatically net new money into equities.
Index inflows broadly offset active outflows in aggregate
2016–25 index inflows of $2.93 trillion offset active outflows in aggregate; the switching is not traced, and the figures include reinvested dividends. Index mutual funds +$95bn plus index ETFs +$2,837bn; active mutual funds −$3,400bn plus active ETFs +$582bn. The four-way sum is +$114bn. AUM growth is not flow.
Source: ICI monthly flow tables. Units: USD billions, cumulative from January 2016.Nor is concentration one thing. As of October 31, 2024, the largest twenty companies represented 46% of S&P 500 weight, according to S&P Dow Jones Indices. That is a historical measure of asset-value concentration, not a 2026 reading, a passive-flow estimate or a map of who would transact under stress. [1]
Tracker concentration rose; sparse history challenges a simple passive-origin story
Top-10 whole-fund concentration reached 40.8% in December 2025, then 38.0% in June 2026. The 1965 provider figure of 38.2% is a separate sparse point. The tracker is a proxy, not the official index. Provider company-level weights and fund issuer weights differ, so no connecting line or splice is drawn between constructions.
Source: fund N-PORT holdings; index provider publication, 2026.Ownership concentration concerns who holds the assets. Common exposure concerns which forces affect those holdings. Correlated response concerns what the owners do when those forces change. One cannot read the last directly from the first.
Registered index funds remain minority owners of US stocks
Domestic index equity funds held 19% of US market capitalization in 2025. Active domestic-equity funds held 11%; other investors 69%. Published integer-rounded shares sum to 99% in 2025 and are not rescaled. Ownership does not measure trading or price setting.
Source: Investment Company Institute. Units: share of US market capitalization (%).Still, these things can interact. A market diversified across many names may be less diversified across discount rates, financing conditions or the continued willingness to maintain a familiar allocation.
Demand meets somebody else’s balance sheet
Suppose valuation-insensitive net subscriptions persist. The fund implementing them needs exposure according to its mandate, not a fresh judgment that each constituent has become inexpensive.
Someone must accommodate that demand. Every transaction has another side, but the other side does not arrive with unlimited risk capacity at yesterday’s price. An informed investor may reduce a position as it becomes expensive. Taking a much larger position against the prevailing price can require financing, tolerance for losses and patience that not every balance sheet has.
If that capacity is limited, prices may have to move substantially to induce an adjustment. Transactions at the new price then revalue holdings that never traded.
The market’s capitalization is not a vault containing the cash that would be available if everyone tried to leave.
Xavier Gabaix and Ralph Koijen’s inelastic-markets work develops the general case for flows having substantial valuation effects when institutions adjust their portfolios only modestly in response to prices. Passive-specific research goes further. [2]
Valentin Haddad, Paul Huebner and Erik Loualiche estimate an institutional demand system using US portfolio holdings from 2001–2020. Active investors respond more aggressively when surrounded by less price-responsive investors, but their estimated response does not fully offset the loss of elasticity associated with more passive ownership.
This is the first opening for overshoot: repeated demand meets a supply of counterparty risk-bearing that is willing, but not infinitely willing.
Persistence matters. A temporary transaction need can be absorbed by someone prepared to wait for it to pass. A sustained change in demand asks someone to carry a different portfolio for longer. What looks like an obvious correction opportunity can become an expensive commitment if the pressure continues before anything reverses.
Why the biggest names might move more
There is a tempting shortcut here, and it is wrong. A capitalization-weighted fund does not mechanically have to acquire more shares after a stock appreciates. Its existing holding already appreciates with it. Likewise, proportional inflows do not mechanically make the largest stocks a larger fraction of the index.
A disproportionate effect needs another channel: different stocks must encounter different resistance to the demand.
Hao Jiang, Dimitri Vayanos and Lu Zheng offer one such mechanism. In their model, active investors facing strong demand for a large firm’s shares must bear more firm-specific risk to resist the price pressure. As the price rises, the absolute risk associated with that exposure can grow, making further resistance more costly.
Their empirical tests, using quarterly S&P 500 index-fund data over 1996–2020, find larger passive flows associated with greater outperformance by the largest constituents, consistent with the mechanism. The model and evidence give the concentration argument an actual economic channel. They do not tell us how much of a particular valuation is unjustified. That requires the next, harder step.
The largest CRSP stocks show widening value-weight leadership
A proxy, not S&P 500 equal weight. Both portfolios use the largest NYSE-breakpoint size decile in the CRSP NYSE/AMEX/NASDAQ universe. The five-year annualized value-weight minus equal-weight return gap reached 5.1 percentage points in May 2026. Relative performance does not establish passive causation.
Source: CRSP. Units: percentage points per year, rolling five-year annualized.When support becomes dependence
Fundamentals are not a secret immutable price that a sufficiently stern analyst can retrieve. They involve uncertain future payouts and the rates at which those payouts should be discounted. Better earnings prospects, lower required returns or less risk can justify a much higher price.
Earnings rose with matched weights, but a relative-P/E premium remains
42.4% weight against 35.2% of earnings is a matched sample with 96.1% coverage, not the whole index. A relative P/E of 1.36 is not proof of mispricing. In 2025 the positive-net-income and ex-Berkshire variants are 1.40 and 1.44. Unmatched issuers are removed from both sides; losses shrink the net-income denominator and lower the ratio.
Source: SEC financial-statement frames pulled November 7, 2026; year-end weights.Even a flow-driven increase can partly reflect a lasting change in who bears risk and what return they require. Price pressure is not automatically mispricing.
The proxy rate cushion is thinner than in 2020–21
An excess CAPE yield of 0.56% is a proxy, not a DCF verdict. It was about 3.8% in December 2020 and 3.1% in November 2021, against −1.1% in December 1999. It does not rule out explanations based on expected growth, equity risk premia or how real yields are measured.
Source: Robert Shiller, online data. Units: yield (%).Nor must the underlying business stand still. Cheaper equity financing can support investment that eventually improves cash flows. Some expectations help bring about the conditions that justify them. The distinction is between a financial process that enables productive earning power and one that increasingly needs additional demand to hold its price.
The bubble hypothesis begins when prices outrun a defensible range of cash-flow and discount-rate assumptions, yet persistent demand keeps the discrepancy from correcting. The support increasingly depends on additional investors accepting the price, rather than on plausible payouts supporting it.
CAPE is high on its valid historical sample, which is not a bubble test
CAPE for October 2026 is 40.71, the 98.9th percentile of 1,750 valid months. The valid sample runs January 1881 to October 2026; empty 1871–80 rows are excluded. High valuation neither dates a crash nor proves passive causation or a current bubble.
Source: Robert Shiller, online data. Units: CAPE ratio.That is more than the observation that valuations are high. It is a claim about what must continue for them to remain high.
Now suppose those gains validate the allocation that helped produce them. A saver sees a successful habit. An active manager sees the cost of departing from the benchmark. An institution sees a process that has repeatedly delivered acceptable results. None needs to decide that fundamentals no longer matter.
The feedback can operate through perfectly ordinary decisions: maintain the contribution, limit the deviation, avoid explaining another period of relative underperformance. If those decisions sustain valuation-insensitive demand, the gains become both an outcome and an argument for repeating the behavior.
The index itself need not chase its winners for this loop to form. It can arise in decisions about which investment vehicles receive new savings, which benchmark defines acceptable performance, and how much deviation an active portfolio is permitted. This is a hypothesis about the allocation around passive strategies, not a claim about the index’s internal mechanics.
ETF-share trading is not a count of underlying-stock trades
In-kind ETF creation and redemption does not mechanically force underlying purchases. In 2025 domestic-equity ETF gross issuance plus redemptions totaled $8.3 trillion, 5.6% of $148.1 trillion in company-stock trading value. ETF shares were 28% of secondary trading volume, a different denominator. Authorized-participant inventory, netting and hedging break any one-to-one mapping.
Source: Investment Company Institute. Units: share (%).This is the interpretive extension: evidence of success can become partly self-generated. A real earnings improvement may start the process; subsequent demand may carry the price further than that improvement warrants. The two are not mutually exclusive, which makes the distinction difficult and the question worth asking.
A bubble need not be a market full of people who know nothing. It can be a market in which enough people find it costly to insist on what they think they know.
The conditions can change together
The upward process does not guarantee an abrupt reversal. Subscriptions can slow without turning into withdrawals. Earnings can catch up. Patient holders can remain patient.
But if expectations weaken, flows change or financing tightens, concentrated exposures can leave a market unusually sensitive to the adjustment. Common holdings alone do not cause synchronized sales; similar obligations can turn common exposure into correlated action.
Implied and trailing volatility can diverge sharply
VIX looks forward 30 calendar days; realized volatility looks back 21 trading days. The gap is not a forecast error, and VIX includes a variance risk premium. The same-date comparison is descriptive; neither the jumps nor the earlier calm are attributed to passive flows.
Source: FRED series SP500 and VIXCLS; Cboe is the VIX provider. Units: annualized volatility (%).Market liquidity is the ability to transact without a large price impact. Funding liquidity is the ability to finance positions and meet obligations. Markus Brunnermeier and Lasse Pedersen show how the two can reinforce each other when intermediary capital is constrained. [3]
A loss reduces the equity supporting a leveraged position. Collateral demands or risk limits can require exposure to be reduced. Counterparties protecting their own balance sheets may offer less capacity just when it is most needed. Lower transaction prices then affect other portfolios and collateral values.
The important change is the sensitivity. An investor well inside a financing limit can tolerate a loss without acting. Near the limit, an additional modest move can activate a requirement to transact. The trigger can be small because earlier developments have already brought the position close to a boundary.
There is nothing inevitable about the whole chain. Its force comes from the links being able to reinforce one another, not from each link being universally present.
Not every surprise is the same surprise
A financial system can accumulate fragility endogenously: through its own positions, incentives and financing arrangements. A catalyst is whatever activates an adjustment. A severe tail outcome is the scale of the resulting loss. Those are three different things.
Calling every crash a black swan can conceal the very thing this argument asks us to examine: the vulnerabilities assembled before the news arrived. Conversely, even a market with recognizable weaknesses can encounter an external event that genuinely exceeds what its participants imagined.
Large realized-volatility spikes predate passive funds
This is CRSP value-weighted total-market realized volatility including dividends, not the S&P 500 or VIX. Spikes in 1929 and 1987 challenge any claim that severe volatility is a new passive-era phenomenon. Descriptive history, not a causal test. The daily series is shown here as the maximum within each month.
Source: CRSP. Units: annualized 21-trading-day volatility (%).Resonance offers a useful scientific provocation here. In stochastic resonance, an appropriate amount of noise can strengthen a nonlinear system’s response to a weak external signal. The question it lends us is what the system was capable of amplifying. Financial constraints already give us a mechanism without requiring a physics label. [4]
What an amplification episode actually establishes
During the UK gilt-market stress of September–October 2022, leveraged liability-driven investment strategies faced additional collateral demands as long-dated bond prices fell. Some could not obtain fresh collateral quickly enough and had to reduce positions in an illiquid market, adding to pressure on prices. The Bank of England stepped in with temporary purchases.
A large gilt shock met a temporary purchase backstop
The UK 2022 gilt stress is an amplification case, not passive-equity bubble evidence; the initial shock was large. Fitted 30-year spot yields are not individual gilt quotes. Purchases totaled £19.26bn; the broader LDI, pension and insurance sector’s more than £36bn of net sales is a different measure and window.
Source: Bank of England yield-curve archive and operation results; APF report, January 2023.Gabor Pinter’s transaction-level study estimated more than £36 billion of net gilt sales by the broader liability-driven-investment, pension and insurance sector between September 23 and October 14, 2022. [5]
This is evidence of financing and market-liquidity amplification, not evidence of a passive-equity bubble. The episode illustrates how obligations can change market behavior; it does not establish the source of an equity valuation overshoot or prove that small disturbances always cascade.
The strongest objection is a functioning market
The alternative account deserves its full strength. Perhaps the largest companies have become more valuable because their earnings prospects improved. Perhaps lower discount rates or lower perceived risk explain valuations. Perhaps the supposed excess is a disagreement about the future, not a price awaiting correction.
Informed active investors can supply shares when demand raises prices, redirect capital and improve price discovery. The greater the apparent discrepancy, the greater the potential incentive to investigate it. Firms can also issue equity. Counterparty capacity can expand rather than remain fixed while passive assets grow.
Robin Greenwood and Marco Sammon supply an important challenge to a simple more-indexing-means-more-distortion story. Studying historical S&P 500 additions and deletions, they find that associated abnormal returns became much smaller in 2010–2020 despite the growth of indexed assets. [6]
Slow, stable passive ownership can also damp trading. A long-term unleveraged holder has no obligation to respond to a decline as a leveraged trader would. Steady contributions can supply demand during weakness. Lower management and turnover costs are genuine benefits, not evidence that investors have surrendered their judgment.
The hypothesis remains consequential precisely because neither account wins by definition. If prices mainly reflect improved earning power and durable risk-sharing, concentration may be the record of success. If a material part of the rise depends on continued valuation-insensitive demand meeting limited resistance, concentration is the shape of the exposure.
The danger is not agreement itself. It is agreement becoming hard to distinguish from the price support generated by repeatedly acting on it.
When a rising price is offered as proof that the allocation was right, the question is what would still justify that price if the expected stream of new demand weakened.
Readings this piece keeps in view
The quantities charted above, at their latest reading. They are descriptive, and none of them dates anything.
Index funds’ share of US market cap
19
%, 2025
Active domestic-equity funds 11%; other investors 69%.
Largest-decile VW minus EW, five-year
5.1
pp/yr, May 2026
The widest value-weight leadership in the CRSP sample.
Excess CAPE yield
0.56
%, Oct 2026
Against 3.8% in December 2020 and −1.1% in December 1999.
Shiller CAPE
40.7
Oct 2026
98.9th percentile of 1,750 valid months since 1881.
Sources and as-of notes
1
S&P Dow Jones Indices, S&P 500 constituent weights as of October 31, 2024.
Historical measure, not a 2026 reading
2
Gabaix and Koijen, “In Search of the Origins of Financial Fluctuations: The Inelastic Markets Hypothesis”; Haddad, Huebner and Loualiche on institutional demand elasticity, 2001–2020 holdings.
3
Brunnermeier and Pedersen, “Market Liquidity and Funding Liquidity.”
4
Krawiecki and Hołyst, stochastic resonance in an interacting-agent financial model, 2003.
5
Pinter, transaction-level study of LDI, pension and insurance gilt sales, September 23 to October 14, 2022; Bank of England operation results and APF report, January 31, 2023.
6
Greenwood and Sammon on S&P 500 additions and deletions, 2010–2020.
Data
ICI Fact Book 2026; CRSP 202609 vintage; Shiller online data, October 2026 with estimated CPI; FRED SP500 and VIXCLS; SEC financial-statement frames pulled November 7, 2026.
Source list abbreviated for this design mock