There Are No Permanent Winners in the Market. And That’s Why You Keep Picking Them.
 |
In 2000, the five largest companies in the S&P 500 were General Electric, ExxonMobil, Pfizer, Citigroup and Cisco.
Read that list again slowly. At the time it did not look like a list of companies. It looked like a description of how the economy worked, and it would have been unremarkable to build a portfolio around all five on the grounds that they were going to be there regardless.
What followed is a matter of record. GE was removed from the Dow Jones Industrial Average in 2018 after a long decline and was eventually broken into three companies. Citigroup came close to failing in 2008. Cisco spent two decades below its dot-com peak. Pfizer and Exxon survived and spent long stretches behind the index.
None of them did anything foolish enough to explain the size of the fall. They were simply at the top of a structure that does not keep anyone at the top.
The Throne Always Changes Hands
The rotation is not a story about individual failures. It is the arithmetic of the index itself.
Innosight's work on corporate longevity puts a number on it. The average tenure of a company in the S&P 500 was about 33 years in 1964. By 2016 it had narrowed to roughly 24 years. Their forecast has it falling to about 12 years by 2027 — and at the current rate of churn, around half of today's index will have been replaced within ten years.
Half. Not through crisis, not through scandal, mostly just through the ordinary process of one kind of business mattering less than it did and another kind mattering more.
Which reframes what the top of the list actually is. It is not a ranking of durable strength. It is a snapshot of which industry the economy is currently paying the most for — and the thing being photographed keeps moving.
The Index Is a Rule, Not a List
This is the part that gets skipped, and it is the one that does the work.
An index is not a portfolio someone assembled once. It is a set of rules for which companies belong, applied continuously. Companies that meet the criteria come in. Companies that stop meeting them go out. Nobody has to notice, approve, or act.
So when a leader falls out of relevance over a decade, an index holder is not left holding it all the way down in full weight. The position shrinks as the company shrinks, and the rising replacement grows into the space. The rotation happens inside the holding.
That is the whole mechanical argument, and it has nothing to do with predicting anything. Buying the index is not a bet that today's largest companies will keep winning. It is a decision to own whatever the rule selects, including the names that do not exist yet.
Buying a single company is the opposite decision, and it is worth stating plainly rather than pejoratively: it is a bet that this particular name will be among the roughly half that remain — and that you will still want to hold it when the reason you bought it has stopped being the reason it matters.
The Math Says Index. The Heart Says Something Else.
The data on this is not ambiguous, and it is not new.
S&P's own SPIVA scorecard, measured to the end of 2024, found that 89.5% of actively managed US large-cap funds underperformed the S&P 500 over the preceding fifteen years. These are full-time professionals with research budgets, direct access to management, and every tool that exists. Roughly one in ten beat the index over that span.
The retail record is not better.
And yet almost nobody organises their financial life around that finding. People know the number and buy individual companies anyway — including people who can quote the number.
That gap is usually described as a failure of discipline. It is more accurately described as a failure of the number to be about the right thing.
Nobody Refreshes an Index Fund at 2 A.M.
An index fund is a rule. Rules are not something a person can feel loyal to.
Nobody sits at a dinner table describing how their total-market fund moved 0.3% today. Nobody reads a fund's methodology document at midnight. Nobody argues with a family member about whether a broad index is fairly valued, or feels betrayed by it, or defends it to a skeptical colleague.
A company is different. A company has a founder, a product you can hold, a fight it is having with a rival, and a story about what the world will look like if it wins. Owning a piece of it is participation in something, in a way that owning a rule is not.
Behavioural research keeps arriving at the same finding: people are not primarily moved by expected returns. They are moved by narratives. Not because they are careless, but because a narrative is how humans decide anything at all under uncertainty — which career, which city, which person.
The mistake is not that people want a story. The mistake is putting the retirement inside it.
The Question That Replaces the Percentage
Here is where this piece used to give a number, and where it should not have.
There is no correct split between the part of a portfolio that follows arithmetic and the part that follows conviction. The right proportion depends on income stability, dependants, time horizon, what else you own, whether you have a mortgage, and how you actually behaved the last time something fell 30%. None of that is knowable from here, and a number printed on a page has the specific danger of sounding like it was calculated for the reader rather than made up for the average of everyone.
What travels instead is a question, asked before the purchase rather than after:
If this position went to zero, what would have to change about my life?
If the honest answer is nothing much — I would be annoyed for a year, the position is sized appropriately for a story, and the story is allowed.
If the honest answer touches retirement, a child's education, a house, or the ability to stop working at some identifiable point, then the position is not a story. It is load-bearing, and it has been placed in the part of the portfolio that was never designed to bear load.
The reason this works better than a percentage is that it scales itself. The same question gives a different answer for someone at 28 with decades of income ahead and for someone at 62 counting the years, without either of them needing to be told which category they fall into.
And it has to be answered in advance. Afterwards it stops being a question about sizing and becomes a question about whether to sell at a loss, which is a different and much harder question — the one where the cost of stepping out usually gets paid.
What This Looks Like From Korea
The rotation is easier to see from here, and harder to sit out.
Korea's own index has turned over its leadership more than once inside living memory, and because the market is smaller the effect is not subtle. A handful of names carry an outsized share of the whole thing, which means the index itself is doing less diversifying than the word "index" suggests — the problem of what happens when two companies are half your index.
There is a second layer specific to investors here. Buying foreign shares from Korea usually means buying the famous few — the names that make it across an ocean into local coverage. That is a concentrated position wearing the clothes of international diversification, and it carries a currency underneath it, so the number in the account answers to something besides the company.
None of that is an argument against it. It is an argument for noticing that "I own American technology" and "I own four American technology companies" are different sentences with different failure modes, and that only one of them survives the throne changing hands.
What This Page Used to Say
This post has been here since April, and most of it stands. One section did not, and it has been removed.
The original prescribed an allocation: a specific percentage range for broad index funds, a specific percentage range for individual companies, and — the part that should not have been written — six named securities as examples, three funds and three companies, printed as though they were the natural contents of each bucket.
That was wrong in two separate ways.
The percentages were invented. They were not derived from anything about the reader, and there is no version of this blog that knows enough about anyone to produce them. Publishing them in a piece otherwise built on public data gave them a borrowed authority they had not earned.
And naming the securities converted an idea into a shopping list. The whole argument of the post is that the specific names rotate. Printing six of them as the illustration undercut the thing being illustrated — and any reader who acted on it would have been buying the current occupants of the throne on the authority of an article explaining that the throne does not stay occupied.
The structure survives. The numbers and the tickers do not, and nothing replaces them, because the honest replacement is a question the reader answers about their own life.
Where This Stops Being Useful
It says nothing about whether any company is a good holding. Rotation is a property of the index over decades. It is not evidence about any individual name in either direction.
Owning the index is not the same as being safe. A broad fund still falls in a bad year, and it can stay down for a long time. What it removes is the risk of holding the wrong name to zero — not the risk of markets.
And "half will be replaced" is a forecast. Innosight's twelve-year figure is a projection from an observed trend, not a measurement of the future. The direction has been consistent for sixty years. The exact number is an estimate.
Where You Are Standing
Look at what you would call permanent. Whichever companies feel like fixed features of the world right now — that impression is a fact about when you started paying attention, not about the companies, which is the same mechanism behind the year you privately call normal.
Ask the zero question about each individual position you hold. Not as a mood check. As a sentence with an answer, written down.
Check whether your index is as broad as its name. If a small number of companies dominate it, some of the rotation protection you think you bought is not there.
And separate the two jobs in your own accounting, whatever the proportions turn out to be. The part that has to work is not the part that is interesting, and confusing them is how a portfolio ends up concentrated by accident rather than by decision.
A Closing Note
Every list of the largest companies in the world has been read, at the time, as a description of how things are. Every one of them turned out to be a description of how things were.
That is not a warning. It is the mechanism working correctly — an economy that stopped replacing its leaders would be a much worse problem than one that does it constantly.
The market has no permanent winners. It does have a permanent game, and the game is compounding through the changeovers rather than trying to guess who is next.
The only reliable way to lose it is to put the part that has to work into a single name, and then be right about everything except how long the world stays the shape you found it in.
Data as of 2026-08-10. Sources: S&P Dow Jones Indices SPIVA U.S. Scorecard, year-end 2024; Innosight Corporate Longevity Forecast; S&P 500 historical constituent rankings. This post is observation, not investment advice.
Related Posts:
When Two Companies Are Half Your Index
The Cost of Stepping Out — Why Markets Punish the Absent
Everyone Has a Year They Call Normal
A Bond Doesn't Fall the Way a Stock Falls
Real vs Nominal — Why Your Money Can Grow and Shrink at the Same Time
Visuals on this post are AI-generated. The author works with AI as a research and drafting assistant; topics, judgments, and final edits are the author's own. This post is observation, not investment advice. See full Disclaimer for details.
