Time-Weighted vs Money-Weighted Returns: Whose Math Wins

Time-weighted returns measure how a strategy performed with the effect of deposits and withdrawals stripped out. Money-weighted returns measure what a particular pool of capital actually did, timing included. Neither is the true percentage. They answer different questions — one about the manager's decisions, one about the owner's — and the common error is asking a single number to answer both.

The distinction was settled, more or less by accident, by American pension committees in the 1960s. A plan might hire two outside managers running similar mandates. At year-end one showed a higher percentage than the other. But the plan itself had been sending contributions in all year, unevenly, and those contributions landed in the two accounts on different dates and in different amounts. If one manager happened to receive a large deposit just before a strong quarter, that manager looked brilliant. If the other received it just before a weak one, that manager looked careless. The committee had no way to tell whether it was looking at skill or at the accident of its own calendar.

A young academic named Peter Dietz had identified the flaw in his doctoral work on pension fund measurement, and his objection was impossible to unsee once stated. The plan sponsor decided when contributions arrived and when benefits went out. The manager did not. Grading the manager on a rate that moved with the timing of those checks was like grading a plant manager on headquarters' decision to double the workforce during a slow month. Two entirely different jobs were being blended into one grade. The Bank Administration Institute later published a study of the problem and gave the alternative its lasting name: the time-weighted rate of return, a figure built so that the fund's own deposits and withdrawals could not move it. That recommendation, made for the convenience of pension committees, is why nearly every statement an exited founder receives today quotes a time-weighted figure — and why almost nobody asks which figure it is.

Why pension committees wanted a number they could not blame on themselves

The mechanism is easier than the vocabulary, and the whole subject turns on it.

Picture a year as a reel of film, and every deposit or withdrawal as a cut. Between the cuts, the camera records only the percentage change of whatever assets were on screen. A strong stretch counts as a strong stretch whether the account held a modest sum or a very large one. A weak stretch counts as a weak stretch on exactly the same terms. At the end of the year those stretch-by-stretch percentages are chained together — each one multiplied through the next, so that a gain compounds on the one before it — to produce a single rate for the whole period.

Because every stretch gets a vote equal to its length in time rather than to the dollars that happened to occupy it, a large wire in November cannot drown out what happened from January through October. The size of the flows never enters the calculation. Only their dates do, and only as boundaries between sub-periods. The result describes the portfolio as if a single unit of money had sat in it untouched from the first day to the last. It is the return of the strategy, not the return of any particular owner's dollars.

That is exactly what a committee comparing two managers wanted, and it is what the Global Investment Performance Standards — the voluntary industry rulebook most reputable managers of public securities adopt when presenting results to prospective clients — treat as the default. The standards exist so that when two firms each present a figure for a strategy, a prospective client is comparing like with like, rather than one firm's fortunate deposit timing against another's unfortunate timing.

Founders will recognise the move from the operating side. Judging a division head on operating margin rather than on total profit is the same idea. A division that received a large capital injection from headquarters in March will produce more absolute profit than one that did not, and that tells you nothing about which manager is better. Strip out what headquarters did; look at what the manager did with what they were given. Time-weighting is operating margin for portfolios.

What money-weighting records that time-weighting is built to ignore

The other method predates the first by centuries, because it is how anyone with a ledger naturally thinks. Take every deposit and every withdrawal, each with its date, and find the single annual rate at which those cash flows would have had to grow to arrive at the closing balance. That rate is the money-weighted return, also called the dollar-weighted return, and usually computed as an internal rate of return, or IRR — the constant rate that makes the timed cash flows explain the ending value.

Here the size of every flow matters enormously. A large sum that sat through a strong stretch pulls the rate up far harder than a small one would. A large sum that sat through a weak stretch pulls it down just as hard. Two families invested in the identical strategy over the identical twelve months will report different money-weighted returns if they moved capital on different days. Their time-weighted returns will match, or come very close, because the strategy did not change. Only the calendar of their money did.

The clearest illustration is two siblings sharing an adviser, a strategy, and a year. One transfers the bulk of an inheritance in January. The other waits, watches a decline, and transfers a similar sum afterwards. The time-weighted number on both statements describes the same strategy and will be effectively the same. The money-weighted numbers will diverge, possibly sharply. The adviser did not become more or less competent between the two wires. One sibling's capital simply occupied a different stretch of the year. Anyone who fires the adviser on the strength of the second sibling's figure is repeating the mistake the pension committees identified: grading the manager on a check the manager did not write.

Buried in that illustration is a quiet test worth remembering. When no money moves in or out, the two methods agree exactly. Divergence between them is never a calculation error and never evidence of dishonesty. Divergence is the timing of the cash, and the size of the gap is a direct measure of how much that timing mattered.

This is not a theoretical curiosity. Morningstar has for many years published research comparing the time-weighted returns funds report with the money-weighted returns their actual investors experienced,¹ and the comparison has consistently found a gap, with investor outcomes trailing the funds' own reported figures.¹ The magnitude moves from period to period and varies by category, but the direction has been stubborn. The cause is behavioural rather than mathematical: money tends to arrive after a fund has done well and leave after it has done badly, so the average dollar spends more of its life in the fund's weaker stretches than its stronger ones. The fund's time-weighted figure is honest. The investors' experience is honestly different. An earlier piece on whether time in the market beats timing approached this from the market's side; the return gap is the same phenomenon seen from the statement.

The mandate question hiding behind the worse number

The strongest objection arrives right about here. If the money-weighted figure is worse, is the manager not still at fault — for putting new money to work on an unlucky day, or for failing to hold it back until conditions improved?

That is a mandate question, not a methodology question, and confusing the two is a common source of friction in advisory relationships. If the agreed instruction was to invest incoming capital promptly, the timing of the deposit belonged to whoever sent the wire, and the manager executed. If the instruction was to stage the money in over some months, the staging was a decision the manager made, and it deserves to be examined as a decision — discussed, questioned, judged on its reasoning — rather than blended into a single percentage and labelled performance. Where a manager holds genuine discretion over deployment, a money-weighted figure is the fair measure, because the calendar was theirs. Where the mandate was to stay invested, smuggling the family's own timing into the score is a category error.

Keeping the two apart is how a family avoids replacing a competent manager over a calendar the family controlled, or retaining a weak one because a fortunate wire happened to land before a rally. It is one of the more useful things to settle at the outset of a relationship, alongside the questions worth asking a wealth manager about incentives and process: who owns the timing decision, in writing, because that answer determines which arithmetic should be used to grade the relationship.

A second objection follows: that the return gap proves nothing about reporting method, only about investor behaviour, and that a disciplined owner who never chases or flees would see the two figures converge. That is correct, and it is precisely the point. The distance between an owner's money-weighted return and the manager's time-weighted return is a measurement of the owner's own timing decisions. It is a report card on the client rather than on the manager. Which is exactly why most statements do not display it prominently, and why the families who would learn most from seeing it rarely ask.

A third objection is more temperamental, and common among people who have run a balance sheet: percentages are a distraction, and the only number that matters is the change in dollars on the page. But dollars without a rate say nothing about how large the capital base was or how long it sat there. A modest percentage on a very large pool produces an impressive-looking dollar figure while representing an inefficient use of the wealth; a mild wobble on that same pool produces an alarming dollar figure that means very little. A percentage without a method does not say whose decisions it is scoring. Both absences are incomplete. The useful habit is not to crown a favourite metric but to insist the method be labelled — and that both methods appear whenever the cash flows were large enough to matter.

Same business, three shareholders, three different answers

Anyone who has raised capital already understands this split without the terminology. A company's operating performance — the growth of its revenue and its cash generation, year after year — is one thing. What any given shareholder made from it is another, and that depends almost entirely on when the shareholder bought in and at what price. The seed investor, the growth-round investor and the person who bought shares on the day of the public listing all owned the same business. Their money-weighted returns from that identical business can run from spectacular to disappointing, because their capital arrived at different moments and paid different prices for the same underlying cash flows.

On the operating side, founders never confuse these. Nobody looks at a late-round investor's mediocre outcome and concludes the company was mediocre. Yet the same people, once they become allocators, routinely read a manager's time-weighted figure as though it described their own capital, or read their own money-weighted figure as though it described the manager. The vocabulary changes when the money changes sides, and the instinct that served them well as operators quietly goes missing.

When the manager owns the clock: the private-market inversion

There is one large corner of the investment world where money-weighted returns are the standard rather than the exception, and it is instructive precisely because it inverts the 1968 logic without breaking it. Private equity, venture and private real estate funds report their results as IRR.² The reason is structural: in those vehicles the manager, not the investor, controls the timing of the flows. Investors commit capital, but the fund draws it down through capital calls when the manager decides to buy something, and returns it through distributions when the manager decides to sell. Because the manager is choosing when money goes in and comes out, a measure that includes timing is the fair one. Time-weighting would be unfair in the other direction — it would strip out the most consequential decision the manager actually made.

The rule, in other words, was never "always time-weight." It was: measure the party who controlled the cash.

But the same sensitivity that makes IRR appropriate in private funds also makes it sensitive to engineering, and the mechanism appears in nearly every private-market allocation a family office reviews. A fund can arrange a borrowing facility — commonly called a subscription line — secured against investors' uncalled commitments. Rather than calling capital on the day of a purchase, the fund borrows from a bank to fund the acquisition and calls investors' capital months later to repay the loan. The underlying business performs exactly as it would have. The eventual profit is the same. But investors' money was formally at work for fewer months, and because IRR is an annualised rate, a given profit earned over a shorter span produces a higher percentage. The presentation standards have responded by pushing funds to show the figure both with and without the effect of such borrowing, which is worth asking for.

This is also why experienced allocators ask for a second figure alongside IRR: the multiple on invested capital, which is simply total money returned divided by total money put in, with no annualising and therefore no sensitivity to the clock. A high IRR paired with a modest multiple usually means the money was not outstanding for long, whether through quick exits or through borrowing. A high multiple paired with a modest IRR usually means patient capital held for a long time. Neither number alone tells the story. The pair does.

The year after a sale, when the two numbers part company

For someone contributing a steady amount each month to a large balance, the flows are small relative to the account and arrive evenly, so the two methods sit close together and the distinction is academic.

Proceeds from a business behave the opposite way. They arrive as a closing wire, then a holdback released later, then a substantial tax payment going out, perhaps an earnout landing a year after that. Interspersed are decisions to buy a building, fund a foundation, or seed a second operating company. Every one of those is an external cash flow — capital entering or leaving for reasons that have nothing whatever to do with whether the manager chose well that quarter — and every one is large relative to the account it passes through. That is exactly the condition under which time-weighted and money-weighted figures tell genuinely different stories about the same twelve months.

If a large tax payment goes out just before a strong stretch, the money-weighted return will read worse than the time-weighted one, because fewer family dollars occupied the good months. If a delayed earnout lands just before a weak stretch, the money-weighted return will read worse again, for the opposite reason. In neither case did the manager's decisions change. The family's calendar did. The year of the exit is already the most consequential allocation window most founders will ever pass through. It is also the window in which reading the wrong percentage is most likely, and most expensive in judgement.

The label matters more than the number

The portable principle is small and durable: a return percentage is an answer to a question, and the question has to be named before the number can be used. Time-weighted answers how did this strategy perform. Money-weighted answers what did this particular capital do. One is the right measure when the client controls the cash, which is the normal arrangement in public markets. The other is the right measure when the manager controls the cash, which is the normal arrangement in private funds.

So when the statement arrives, the first question is not whether the number is good. It is which number it is, and who controlled the timing during the period it covers. Where the flows were large — and after a sale they always are — both figures belong on the page, and the gap between them is not an error to be reconciled but information to be read. It reports, quite precisely, how much of the year's outcome came from the portfolio and how much came from the calendar.