TWR vs IRR: Which Return Should You Show a Client?
Show a time-weighted return when you are judging the manager, and a money-weighted return, usually an IRR, when you are describing what the investor actually earned. The dividing line is control. If the person being measured chose when the money moved, use IRR. If the client chose, use TWR, because measuring a manager on decisions they did not make is neither fair nor useful. Anything being compared to a benchmark or a peer group has to be time-weighted, because an index has no cash flows to weight.
This is not a technical dispute. Both numbers are correct and they are answers to different questions, which is exactly why presenting the wrong one starts arguments in client meetings that have nothing to do with performance. Below is the distinction, an example where the two methods disagree by 128 basis points on the same portfolio, a tool that picks for you, and what the standards actually require.
The difference in one table
| Time-weighted (TWR) | Money-weighted (IRR) | |
|---|---|---|
| Answers | How did the portfolio perform? | How did the investor’s money perform? |
| Cash flows | Removed from the result by construction | Change the result directly |
| Method | Split at each flow, geometrically link the sub-period returns | Solve for the single rate that reconciles the flows to the ending value |
| Needs | A valuation at every external cash flow | Dated cash flows, no interim valuations required |
| Standard for | Composites, benchmark comparison, manager evaluation, mutual fund and SMA reporting | Private equity, real estate, drawdown funds, and any question about a single investor’s experience |
| Also called | TWRR, geometric return | Dollar-weighted return, MWR, MWRR |
The same portfolio, two answers
A separately managed account over Q1 2026. It opens at $1,000,000, returns 8.00% through Feb 15, then 1.56% through Mar 31. The manager’s decisions are identical in both columns below. The only difference is what the client did on Feb 15.| Method | Client adds $200,000 | Client takes out $200,000 |
|---|---|---|
| Time-weighted return | 9.6875% | 9.6875% |
| Money-weighted return | 9.1090% | 10.3882% |
Which should you show?
Return method picker
Four questions about the number you are producing. Nothing is stored or sent anywhere.
1. Who controls the timing of money moving into and out of the portfolio?
2. Will this number sit next to a benchmark, an index, a peer group, or another manager’s record?
3. What kind of vehicle is it?
4. What is the reader being asked to do with it?
Time-weighted return
The client controls the timing, so the manager should not be measured on it. Answer the four questions above for your own situation.
The decision table
The same logic, if you would rather read it than click it.| Situation | Show | Why |
|---|---|---|
| Anything printed next to a benchmark or peer group | Time-weighted | An index has no cash flows, so a money-weighted portfolio figure is not a like-for-like comparison |
| Discretionary SMA, client funds and withdraws at will | Time-weighted | The manager did not choose the timing |
| Composite or firm-level track record | Time-weighted | Required under GIPS unless the exception below is met in full |
| Private equity, real estate, credit, or any drawdown fund | Money-weighted, since inception | The GP controls capital calls and distributions, so timing is part of the performance |
| A client asking what their own money earned | Money-weighted, with time-weighted alongside | Only the money-weighted figure reconciles to the change in their balance |
| An account funded mid-period with no starting balance | Money-weighted | Time-weighted and Modified Dietz both need an invested base to divide by |
Situation by situation
Discretionary separate accounts
Time-weighted, without exception. The client wires money in when they sell a business and pulls it out when they buy a house, and neither event says anything about the manager. This is the case the convention was invented for.Private funds and anything with capital calls
Since-inception IRR, reported next to the multiples. A fund that calls $5,000,000 over two years, distributes $4,000,000, and carries $4,200,000 of remaining NAV produces a 15.38% IRR and a 1.64x TVPI. Both get reported because neither is sufficient alone: IRR understands time but can be engineered by it, and TVPI cannot be engineered by time but does not understand it. The clearest illustration is a subscription line of credit. Delay both of those capital calls by twelve months, leaving every distribution and the ending NAV exactly as they were, and the IRR goes to 22.64% while TVPI stays at 1.64x. Same deals, same dollars, 726 basis points. If you are comparing two funds on IRR alone, you may be comparing their credit facilities.Client-facing statements
Most advisers who do this well show both, labeled. Time-weighted answers “is this strategy working,” money-weighted answers “what happened to my money.” A client who sees 9.69% in one place and 9.11% in another with no explanation will assume one of them is a mistake, and they will assume it is the higher one. One sentence usually settles it: “The 9.69% is how the portfolio performed. The 9.11% is what your money earned, which is lower this quarter because you added to the account in February, before a slower stretch.”Marketing materials and RFPs
Time-weighted, because the reader will compare it to something else. If your strategy genuinely warrants a money-weighted presentation, the comparison figure has to be a public market equivalent, meaning the index return restated using your own cash flow timing. Publishing an IRR next to a raw index return invites a challenge you will lose.Accounts that break the math
An account opening at zero and funded mid-period has no starting balance to divide by, so time-weighted return and Modified Dietz both fail on it while an IRR solves normally. An account whose invested balance crosses through zero during the period breaks the IRR instead, because more than one rate can solve the equation. Neither case should be reported without a footnote.What the standards require
GIPS
Under the 2020 GIPS standards a firm must calculate time-weighted returns. Money-weighted returns may be presented instead only when the firm has control over the external cash flows into the portfolio and the portfolios in the composite, or the pooled fund, have at least one of these characteristics: closed-end, fixed life, fixed commitment, or illiquid investments as a significant part of the strategy. Read the test as two gates in series. Control over the cash flows is necessary but never sufficient on its own. The 2020 edition widened this: the earlier standards reserved money-weighted returns for private equity and closed-end real estate, and that restriction is gone, so more alternative strategies now qualify. Presentation differs too. Time-weighted composites report annual returns for each period. A money-weighted presentation reports a single figure, since inception, annualized through the most recent annual period end.The SEC Marketing Rule
Rule 206(4)-1 does not tell a US registered investment adviser which method to use. It governs how whichever number you chose is presented. Any advertisement showing gross performance must also show net performance with at least equal prominence, in a format designed to facilitate comparison, calculated using the same methodology over the same period. For anything other than a private fund, performance must be shown for one-, five-, and ten-year periods, or since inception if the track record is shorter, each with equal prominence, and current through a period ending no earlier than the most recent calendar year end.
This is a general explanation of calculation methods, not compliance advice. If you are claiming GIPS compliance or preparing performance for regulatory review, confirm the current requirements with your verifier or compliance counsel. The SEC staff has updated its Marketing Rule FAQs more than once, most recently in 2025 and 2026, so check the current version before relying on a summary of it.
Four mistakes that cost real credibility
An IRR next to an index return
The most damaging one, because it looks reasonable. It will flatter you or damage you at random depending on when your clients happened to move money, and a sophisticated prospect will notice.Switching methods between periods
Reporting the quarter time-weighted and the year money-weighted, or changing method when the numbers stop looking good, destroys the comparability of your own track record. Pick a method per report type, document it, and leave it alone.Annualizing a short period
Excel’sXIRR always returns an annualized rate. On the quarterly example above it reports roughly 43.0%, which is a projection, not a result. Do not annualize anything shorter than a year for client reporting, and confirm the basis before deciding your software disagrees with your spreadsheet.
Two numbers with no sentence between them
Showing both methods is good practice. Showing both without saying why they differ converts a teaching moment into a suspicion. Label them, and explain the gap in the same breath.The practical takeaway
Ask one question first: did the person being measured choose when the money moved? If yes, IRR. If no, TWR. Then ask whether the number is going next to a benchmark, because if it is, the answer is TWR regardless. Everything else is detail.
Advantage Portfolio Hub calculates Modified Dietz returns from your posted transactions, across every entity, account, and holding, chained monthly for a since-inception figure, alongside a since-inception money-weighted return and the paid-in multiples. Both numbers, from one set of posted transactions, so the choice this page describes is a reporting decision rather than a data-collection project.
Start a free trial
First entity loaded in 20 minutes. No card required. 14-day trial.
Educational content. Advantage Portfolio Hub is a software provider, not an investment adviser, accountant, or law firm. Nothing here is investment, tax, accounting, or compliance advice.