The odds against exceptionalism

Every concentrated portfolio is a claim of exception. Own six businesses instead of six hundred and you are asserting, whether you say it aloud or not, that you can name in advance the handful of companies that will not regress, not stall, not be commoditized, not be disrupted — that will keep compounding while the great average of enterprise grinds along beneath them.

It is worth knowing precisely how arrogant that claim is. There is a ledger.

The ledger says: most stocks are not merely mediocre, they are worse than cash. Across nearly a century of American markets, the majority of listed companies delivered lifetime returns below Treasury bills; on the order of four percent of stocks account for essentially all of the market’s net wealth creation above the risk-free rate. The ledger says: growth does not persist the way stories do. Of companies growing revenue at twenty percent or more, only a small fraction are still doing it five years later, and almost none at ten. The ledger says: paying for exceptionalism has its own gravity — buy a company priced at extreme multiples of sales and the historical distribution of what happens next is not kind, however the story reads. And the ledger says one more thing, quietly, about the survivors: even the true exceptions punished their owners along the way. The great compounders of the last half-century routinely fell by half; the best of them, at some point, by ninety percent. The reward for being right was the repeated opportunity to feel wrong.

This is the outside view, and most investors flinch from it, because the inside view is so much better company. The inside view is the thesis — the product, the founder, the flywheel, the TAM. It is specific, vivid, and yours. The base rate is none of those things. It is the fate of a thousand companies you never met, aggregated into a number that does not care about your conviction. Kahneman spent a career showing what happens when the vivid particular meets the dull distribution: the particular wins the argument and loses the outcome.

So the rationalist’s conclusion writes itself: buy the average, own everything, submit to the distribution. Indexing is base-rate worship formalized — and it works, for exactly the reason the ledger describes.

We decline that conclusion, and it matters to say why — because the same ledger cuts the other way. If four percent of companies created all the wealth, then everything depends on exceptions. The average that the index worships is not produced by averageness; it is produced by a tiny number of extraordinary businesses dragging the whole distribution upward while the majority decays. The index owns the exceptions only by accident, diluted a thousandfold with the decaying majority. The entire game — the only game that has ever created equity wealth — is the exceptions. Base rates do not say the exceptional is a myth. They say it is rare, and they say most claims to have found it are false. Those are different statements. Averages are the biography of the herd; nothing about a herd’s biography forbids the existence of the animal that leaves it.

What the odds actually demand is not humility as posture but humility as procedure. The claim “this company is an exception” is not forbidden — it is expensive. It has to be paid for, and the currency is specificity.

So at this firm the base rate is not a veto. It is a question that every thesis must answer before it earns a position: name your reference class, state its fate, and say precisely what licenses this company’s exemption. Not “the team is exceptional” — the ledger is full of exceptional teams that reverted. Not “the market is huge” — TAM is the inside view’s favorite costume. A license is structural: the distribution-and-data moat that makes the twentieth year of dominance cheaper to defend than the first; the reinvestment engine that has already survived a capital cycle that killed its peers; the switching costs measured in a customer’s own risk, not in our optimism. If the license cannot be stated in a sentence that would survive a skeptic’s cross-examination, the thesis is not an exception. It is a base rate wearing a costume.

And because we know which way self-regard leans, the cross-examination is not optional and not ours: every thesis at this firm faces a written prosecution before a verdict, and every verdict must answer it. The odds against exceptionalism are not an enemy of conviction. They are the whetstone. A conviction that has stated its reference class, heard the prosecution, and can still say this one, for this reason, at this size — that conviction has earned the right to be concentrated.

There is a temperament hiding under all this arithmetic, and we may as well own it. We are not neutral about exceptionalism. We think the rare, deliberate, compounding enterprise is the most interesting object in markets, and that a life spent indexing everything is a decision to never name what you believe. The odds are steep; that is what makes the claim worth making carefully. We do not resent the mountain’s height. The height is the point.

The base rate tells you what happens to those who climb casually.

Climb anyway. Rope in.