Why Most Funds Fail

The kindest thing that can be said about the active fund industry is that its failures are measurable. A fund exists to beat something: an index, a cash rate, a stated objective. The comparison can be made after fees, and it can be made honestly, which means counting every fund that existed at the start of the period rather than only the ones still standing at the end. Done that way, the measurement is not flattering. Done any other way, it is not a measurement.

Consider what an honest count requires. Take every fund in a category on the first day of a long period. Some will close, some will merge into siblings, some will change their mandate and quietly move to another shelf. If the scorekeeper keeps only the funds alive on the last day, the record is a list of survivors, and survivors are by definition the ones that did well enough to survive. The longer the period, the larger the cut, and the more generous the surviving average becomes. Any comparison worth reading follows the dead to the end of the period and charges their results to the category that held them.

When the count is made that way over long horizons, the record is the one the arithmetic below predicts: active management, taken as a whole, trails the index it was hired to beat. That outcome does not say that active managers are unintelligent, or lazy, or poorly resourced. The people in this industry are among the best-compensated analytical talent anywhere, working with data and tools few institutions can match. It says something colder: that the activity itself, taken as a whole, does not survive contact with its own benchmark once costs and closures are honestly counted. A league table of survivors always looks respectable. The honest table is the one that keeps the empty rows.

Private funds fail in a different shape. A public fund that disappoints usually dwindles; a private fund that disappoints tends to close, return what is left, and stop reporting. That last step matters more than it sounds, because many measures of private-fund performance are built from what funds choose to report. A fund that stops reporting takes its worst months with it, and a fund that starts reporting after a strong launch brings only its best ones. Both habits push the reported average up, and neither is visible from the average itself. An illustration makes the size of the problem concrete. Suppose a hundred funds each have the same true expected return, and in a given year the worst-performing tenth of them close and stop reporting. The average of the ninety that remain is higher than the true average by roughly the gap between the closed tenth and everyone else, every year, compounding. The figures here are illustrative; the mechanism is not. In a population where closure is common, mortality is not a footnote to performance. It is the central fact about it.

There is a popular version of this argument that we should examine rather than repeat, because it is repeated constantly: the claim that only a small minority of funds beat the stock market. As an outcome over a strong decade for equities, something like that is easy to believe, because a diversified fund that holds less than all of its capital in equities will usually trail an all-equity index when equities are rising quickly. But as a judgment it is the wrong test, and saying so matters to us, because a firm that complains about a rigged scoreboard only when it loses is not trustworthy when it wins. Many private funds never promised to beat an all-equity index. They promised steadier returns, smaller drawdowns, results that do not depend on the stock market rising. Judging them on the equity index alone is like judging a goalkeeper on goals scored. The honest verdict is narrower and still severe: measured against benchmarks matching their own stated job, the arithmetic still says most funds will fall short over long horizons. The popular claim is wrong about the test and broadly right about the outcome.

Why? Not one reason, and anyone who offers one reason is selling something. The arithmetic comes first. Active investors, taken together, hold the market; before costs, the average actively managed dollar must therefore earn the market's return, and after costs it must lag by exactly the amount of the costs. That is not a finding that could be overturned by better data. It is an identity, and it holds in every market where active and passive money coexist. Every fee, every spread, every unit of market impact is paid out of a gross result that starts, in aggregate, at zero relative to the benchmark. A fund therefore does not need a little skill. It needs enough skill to pay its entire cost structure and then keep going, every year, while competitors with similar data and similar models compete the same opportunity down. Costs compound against gross skill the way interest compounds against a borrower: quietly, and on the whole balance.

An illustration shows how quietly the cost side does its damage. Take an illustrative fund whose total costs, fees and trading together, come to two percent a year, and suppose its gross results exactly match its benchmark. After ten years the investor holds about eighteen percent less than an investor who paid nothing, and after twenty years about a third less, purely from compounding the cost. To merely break even with a costless alternative, the fund needs gross skill of two percent a year, every year, against competitors who are trying to take the same two percent away. Put that way, the surprising fact is not that most funds fall short. It is that the industry is so often described as if falling short were the exception. We set our own cost structure with this arithmetic in front of us, because every point of cost is a point of skill that must be found again each year before the investor sees anything.

Then the opportunity itself moves. A real edge, once found, attracts capital, and capital is the edge's natural predator: the more money trades a pattern, the faster the pattern is priced away and the more the trading itself moves prices. This is crowding, and it explains a rhythm every allocator recognises: a strategy's best years are usually its smallest, and the flood of money that arrives to celebrate them is the beginning of the decline. Capacity is the same fact stated from the other side. Add style drift, the slow migration away from whatever the fund was actually good at, usually toward whatever is currently raising money. Add the selection built into the industry's marketing: the backtests that get shown are the survivors of a private graveyard of backtests that did not, and a simulation chosen because it is beautiful carries the beauty of the selection, not of the market. Add key-person risk, the fund that is one person's judgment wearing an institution's clothes, and operational failure, the unglamorous category (fraud, error, broken controls) that ends funds no market view could have saved. And add the investors' own timing, a subject taken up in the companion piece in this Library: capital arrives after the good years and leaves after the bad ones, so even a genuinely skilled fund is often largest exactly when its opportunity is smallest.

What, then, do the survivors have in common? Reasoned from the causes above, the answer points away from brilliance and toward posture. The funds that persist tend to have costs low enough that the arithmetic is not a sentence; a defined edge they can state in plain language and defend when it underperforms, as every real edge periodically does; size discipline, meaning a willingness to refuse money that would dilute the opportunity; validation habits that treat a beautiful backtest as a suspect rather than a trophy; and investors whose horizon matches the strategy's, so that capital is a stabiliser rather than a weather event. None of this is glamorous. All of it is rare, because each item on the list requires declining something attractive: fee income, asset growth, a compelling simulation, impatient capital.

There is one more failure that the numbers cannot show, and it is the one we think about most: the fund that measures itself against a benchmark it chose after the fact. A strategy that underperforms equities can be reframed as a diversifier; one that underperforms bonds can be reframed as a growth engine. Every reframing is individually plausible, and together they make a fund unfalsifiable. Our answer is to state, before a single position is taken, what the strategy is for, what it should be compared with, and what result over what horizon would tell us it is not working. A standard written in advance can be failed. That is the only kind worth having.

We publish this essay as a new firm, and we want to be exact about what it is and is not. It is not a claim about us. Our fund has not begun trading. We have no track record, and nothing in these pages should be read as implying that we belong to the surviving minority; that question will be answered by audited results over years, or it will not be answered at all. What this essay is, is the bar, stated publicly, with the evidence attached. We built our working standards against that bar: validation before capital, explanations a person can state, diversification maintained as a discipline rather than reported as a count, size treated as a risk, and terms designed for investors who think in years. If we fail, we intend to fail against a standard we published in advance, in full view of the data that predicts failure is the likely outcome. That seems to us the minimum honesty the scorecard deserves.