The short answer
- TWR (time-weighted return) measures how well the investments performed, ignoring the timing and size of deposits and withdrawals. Use it to judge a manager or strategy and to compare against a benchmark.
- XIRR (money-weighted return) measures what the investor actually earned on the money they had invested, given when cash went in and out. Use it to judge the outcome of the owner’s decisions, including when to add money.
- CAGR is a smoothed annual growth rate between two values. It is only meaningful when there were no external cash flows in between, or when it is applied to a flow-adjusted series such as a TWR index.
A worked example
An investor starts with 100,000. In year one the investments gain 10%, so the portfolio is worth 110,000. The investor then deposits another 100,000, making 210,000. In year two the investments fall 10%, leaving 189,000.
| Measure | Calculation | Result |
|---|---|---|
| TWR | 1.10 × 0.90 − 1 | −1.0% over two years |
| XIRR | Solve 100,000(1+r)² + 100,000(1+r) = 189,000 | about −3.7% per year |
| Naive CAGR | (189,000 / 100,000)^(1/2) − 1 | +37.5% per year, which is wrong |
The investments themselves lost only 1% over the two years, which is what TWR reports. But the investor put the second 100,000 in just before the loss, so they ended with 189,000 after contributing 200,000, a real loss of 11,000. XIRR reflects that timing and is lower. The naive CAGR treats the deposit as profit and produces a number that looks excellent and is meaningless.
How TWR is built correctly
TWR chains sub-period returns, and a new sub-period begins at every external cash flow: a deposit or withdrawal. Internal movements such as buying a security with cash already in the account, receiving a dividend or a stock split are not external flows and must not break the chain. A common error is treating a corporate-action adjustment as a deposit, which distorts the series.
Industry standards such as the GIPS framework describe how firms should calculate and present returns. CapLens follows the same chain-linking principle, but we do not claim verification against any standard.
Common pitfalls
- Calculating TWR on month-end values when large flows occurred mid-month.
- Comparing a manager’s TWR with an investor’s XIRR and concluding that one is wrong.
- Annualising a short period, which magnifies noise.
- Ignoring currency: a portfolio in several currencies needs returns measured in a stated base currency (see multi-currency reporting).
What this looks like in CapLens
The CAGR and performance reports show TWR, XIRR and CAGR side by side, derived from nightly snapshots and transaction-derived cash flows, with a reconciliation so each figure can be traced. See the reporting tour.