One measures the decisions about what to hold; the other measures what actually happened to your money, including when you added it. They can disagree enormously.
13 min read · Level 4
The standard teaching is that the time-weighted return is the fair one because it isolates skill from the accident of when money arrived, and that framing quietly concedes something it should not. Deciding when to commit money is not an accident and it is not outside the investor's control - it is one of the few decisions an investor genuinely makes. Treating contribution timing as noise to be stripped out is convenient for the industry, which is measured on the stripped figure, and it is convenient for the investor, because the stripped figure is usually the more flattering one. A more demanding reading is that the time-weighted return is the correct measure of a manager and the money-weighted return is the correct measure of a person, and that anyone who quotes only the first about their own portfolio has chosen a figure that excludes their own contribution to the outcome. The two cases here differ by 90,000 of real money and by nothing else, and only one of the two measures noticed.
As soon as money moves in or out mid-way, 'what return did I get?' stops having one answer, and the two answers it does have are both correct.
The time-weighted return chops the period at every cash flow, computes the return of each segment, and multiplies the segments together. Because each segment is measured on whatever was there at the time, the size of a contribution has no effect on the answer. It measures the performance of the holdings, independent of when money arrived.
The money-weighted return is an IRR: the single rate that makes every contribution, withdrawal and the closing value balance. It is affected by size and timing, heavily, because a large contribution just before a good period counts for more than a small one before a bad one. It measures what happened to your money.
The first answers 'how did these holdings perform'. The second answers 'how did I do'. Neither is a better version of the other.
Because the gap between them is where an uncomfortable and useful truth lives. A fund can report a perfectly honest time-weighted return that almost none of its investors experienced, because money arrived after the good period and left after the bad one. That is not a criticism of the fund, and it is not a criticism of the investors - it is the arithmetic of a measure that is deliberately blind to timing meeting a group of people whose timing was not blind.
And because the two are used for different jobs by convention, and comparing across the convention is where errors happen. Published fund performance is time-weighted, because it is the only fair way to hold a manager responsible for what they control. A private fund's own performance is money-weighted, because the manager chooses when to call and return capital. Putting a time-weighted fund figure beside a money-weighted private one and treating the comparison as meaningful is a common and expensive mistake.
Educational, not personalised financial advice. Nothing in the Investor Academy is a recommendation to buy, sell or hold anything — what is right for you depends on your own circumstances and the rules where you live.
The same holdings, the same returns, one difference in when money was added.
Case one. Start with 100,000. Year one returns +25%, so the balance is 125,000. Add 200,000, taking it to 325,000. Year two returns -20%, so it ends at 260,000.
Time-weighted: 1.25 x 0.80 - 1 = 0.00%
Money-weighted: the rate solving -100,000 - 200,000/(1+r) + 260,000/(1+r)^2 = 0, which is -10.2633%. In closed form, 1 + r = (-10 + the square root of 360) / 10.
Money in was 300,000 and money out was 260,000, so 40,000 was lost. The time-weighted figure of exactly zero is true and describes the holdings. The money-weighted figure of -10.26% is also true and describes the investor.
Case two. The identical returns, in the opposite order. Start with 100,000. Year one returns -20%, so the balance is 80,000. Add 200,000, taking it to 280,000. Year two returns +25%, so it ends at 350,000.
Time-weighted: 0.80 x 1.25 - 1 = 0.00%
Money-weighted: the rate solving -100,000 - 200,000/(1+r) + 350,000/(1+r)^2 = 0, which is +12.1320%. In closed form, 1 + r = (-4 + the square root of 72) / 4.
Money in was 300,000 and money out was 350,000, so 50,000 was made.
So: identical holdings, identical annual returns, identical time-weighted return of exactly 0.00% in both cases - and money-weighted returns 22.40 percentage points apart, from -10.26% to +12.13%. The entire difference is that in case one the large contribution went in before the bad year and in case two before the good one.
That is the whole lesson, and it has a blunt consequence. The manager's figure is the same in both cases, and it should be, because the manager did the same thing. The investor's outcome differs by 90,000 of real money. Nobody was misled, no figure was wrong, and the two people would each describe the same two years completely differently and both be telling the truth.
The first mistake is comparing your own money-weighted result with a fund's published time-weighted one and concluding something about the fund. Those are different measures of different things, and the gap between them is mostly a statement about your own contributions.
The second is the reverse: judging your own outcome by the time-weighted figure. It is deliberately blind to when you added money, and when you added money is a decision you made and are responsible for.
The third is using a money-weighted figure to evaluate a manager who does not control the cash flows. A manager of an open fund cannot decide when subscriptions arrive, so holding them to a measure driven by subscriptions is holding them to something they cannot affect.
The fourth is chaining segments carelessly. A time-weighted return requires the portfolio to be valued at every cash flow, not at convenient intervals, and a segment that spans a contribution is contaminated by it.
The fifth is concluding that a gap between the two means somebody behaved badly. It usually just means money moved, and any movement at all will produce some gap.
How did the holdings perform, or how did my money do. The first is time-weighted, the second is money-weighted, and answering the wrong one produces a figure that will be quoted for years.
Each contribution and withdrawal ends one segment and begins the next. A time-weighted return computed on quarter ends when money moved mid-quarter has one contaminated segment, and the contamination is exactly what the measure exists to exclude.
A large gap says your timing of contributions dominated your holdings. That is worth knowing and it is invisible in either figure alone - both cases here have a time-weighted return of exactly 0.00%.
Where the manager decides when capital is called and returned, money-weighted is the fair measure of them. Where the investor decides, time-weighted is the fair measure of the manager and money-weighted is the honest measure of the investor.
One line: 'time-weighted' or 'money-weighted'. Most of the confusion in this subject is not analytical but a missing label, and the label costs nothing.
Any time money went in or out during the period being measured, which is almost every real portfolio and every fund with contributions.
Neither is needed for a single lump sum held untouched from start to finish, where the CAGR is exact and both measures reduce to it. And a money-weighted return on a portfolio you did not control the cash flows of tells you about the cash flows, not about the portfolio.
A finding that recurs whenever fund flows are compared with fund performance is that the average money invested in a strategy experiences a lower return than the strategy's published figure. The mechanism needs no bad behaviour by anyone: assets under management are small during the early period and largest after a strong run, so a poor stretch is suffered by far more money than the good stretch was enjoyed by. The published time-weighted figure is entirely correct, and it weights the years when almost nobody was there exactly the same as the years when almost everybody was.
When a strong published figure was achieved on a small asset base and the assets grew afterwards, the return most investors received is lower than the headline. Ask how large the fund was during the period that produced most of the record.
In practice most disputes about a time-weighted return come down to when the portfolio was valued rather than to how the segments were multiplied. Valuing at month ends while a large contribution landed mid-month puts part of the new money into a segment that was supposed to measure only the old, which is exactly the contamination the measure exists to remove. The error is small in any one period and it accumulates, and it is invisible in the final figure because the arithmetic is impeccable either way.
Before trusting a time-weighted figure, ask on what dates the portfolio was valued. A measure whose whole purpose is to isolate cash flows is only as good as its willingness to value on the day they happened.
Create a free account to read the rest and apply it to your own investing.