The one, three and five-year returns an investment committee asks for, and why the opening mark decides them.
A horizon IRR treats the NAV at the start of the window as a capital call, takes the actual calls and distributions inside the window, and treats the NAV at the end as a final distribution. On an illustrative nine-year-old fund with 320 of NAV five years ago, 410 distributed since and 85 left, the five-year horizon IRR is 13.98 per cent against a since-inception IRR of 11.07 per cent. Mark the opening NAV 10 per cent lower and the same five years show 17.81 per cent.
Worked in full in Private Markets Performance by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
A since-inception IRR answers one question: what did the money earn from the first call to today? It cannot tell an investment committee what the fund did over the last year, the last three or the last five, which is how every other asset class in the portfolio reports. A horizon IRR (also called a point-to-point or period IRR) answers that by opening the window at a NAV rather than at the first call. It is the only way to put a private fund in a table beside a listed index's one, three and five-year returns, and it is also the most manipulable number a private fund can report, because the opening NAV is a valuation.
The cash flows are those of a fictional nine-year-old buyout fund, annual and at year end. The year-end NAVs are illustrative assumptions added for this calculation.
| Year | Contributions | Distributions | Year-end NAV |
|---|---|---|---|
| 0 to 3 | 240 | 0 | n/a |
| 4 | 30 | 0 | 320 |
| 5 | 0 | 40 | 330 |
| 6 | 0 | 90 | 290 |
| 7 | 0 | 120 | 215 |
| 8 | 0 | 100 | 140 |
| 9 | 0 | 60 | 85 |
| Total | 270 | 410 |
The contributions in years 0 to 3 were 60, 75, 60 and 45. Since inception, the fund's IRR with the year 9 NAV as a terminal flow is 11.07 per cent and its TVPI 1.83x.
Rule: horizon IRR = IRR of ( −NAV at the start of the window; distributions − contributions in each period of the window; + NAV at the end, added to the last period ).
Five years (end of year 4 to end of year 9): −320, 40, 90, 120, 100, 60 + 85 = 145. IRR = 13.98%.
Three years (end of year 6 to end of year 9): −290, 120, 100, 145. IRR = 12.01%.
One year (end of year 8 to end of year 9): −140, 145. IRR = (60 + 85) / 140 − 1 = 3.57%.
In Excel, on a row of net flows with the NAVs in place: =IRR(F10:K10) where F10 holds =-NAV_year4 and K10 holds =dist_year9-contrib_year9+NAV_year9. With dated quarterly flows use =XIRR(values, dates) and put the opening NAV on the window's start date.
The year 4 call of 30 is not in the five-year window: it is already inside the opening NAV of 320. Counting it twice, once in the NAV and once as a flow, is a common spreadsheet error and lowers the rate. The opening flow is the NAV after every call and distribution on that date.
| Window | Opening NAV | Distributions in window | Value out incl. closing NAV | Window multiple | Horizon IRR |
|---|---|---|---|---|---|
| 1 year | 140 | 60 | 145 | 1.04x | 3.57% |
| 3 years | 290 | 280 | 365 | 1.26x | 12.01% |
| 5 years | 320 | 410 | 495 | 1.55x | 13.98% |
| Since inception | n/a | 410 | 495 | 1.83x | 11.07% |
The five-year figure is 2.92 points above the since-inception rate, though it covers the same distributions. The difference is the starting point: the window opens after the fund's J-curve, when 270 of calls had become a NAV of 320, a TVPI of only 1.19x. The fees and early write-downs of years 0 to 4 sit outside the window. The one-year figure is low because the last year carried little but a small exit and a mark-down; it says almost nothing about the manager.
| Year 4 NAV | Five-year horizon IRR | Change, basis points |
|---|---|---|
| 288 (10 per cent lower) | 17.81% | 383 |
| 320 (as reported) | 13.98% | 0 |
| 352 (10 per cent higher) | 10.69% | -330 |
Not one cash flow changed, yet the five-year return spans 10.69 to 17.81 per cent. A conservative mark at the start of a window flatters every horizon return that begins there; an aggressive one depresses it. Since-inception IRR is immune, because it starts from cash. That asymmetry is why a horizon IRR should never be presented without the since-inception figure beside it.
A horizon IRR is not a time-weighted return. Chaining the five annual returns, (NAV + distributions − contributions) / opening NAV each year, gives 12.20 per cent a year here, not 13.98. The IRR weights the years by the capital in the fund, and most of it was there in the early, stronger years of the window.
Annualising the change in NAV. Between year 4 and year 9 the NAV fell from 320 to 85, an annualised -23.29 per cent, while the investor received 410 in cash and earned 13.98 per cent. A NAV series on its own describes a fund returning capital, not one losing money; distributions must be in the calculation. The other mistake is reading a five-year horizon IRR on a fund in harvest as evidence about the manager's next fund: the window has been chosen to start after the costly years.
To compute a horizon IRR, open with the NAV as a negative flow, take the window's cash flows, and close with the NAV. Here that gives 13.98 per cent over five years, 12.01 over three and 3.57 over one, against 11.07 since inception, with an opening mark that moves the five-year figure by up to 383 basis points. The since-inception rate, the multiples and the public market equivalents on these cash flows are computed in the free workbook for this case. On why fund rates are not additive in the first place, see why fund IRRs cannot be averaged.
A since-inception IRR starts at the first capital call, so it includes the fees and write-downs of the early years. A horizon IRR starts at a NAV. In the worked case the five-year horizon IRR is 13.98 per cent while the since-inception IRR on the same fund is 11.07 per cent, because the window opens after the J-curve.
Because it is the only investment the window recognises. Marking the opening NAV 10 per cent lower, 288 instead of 320, lifts the five-year horizon IRR from 13.98 to 17.81 per cent without changing any cash flow; marking it 10 per cent higher cuts it to 10.69 per cent.
No. The time-weighted return chains each year's return regardless of how much capital was invested; the horizon IRR weights each year by the capital in the fund. On the same five years the chained time-weighted return is 12.20 per cent a year and the horizon IRR 13.98.
This article is one calculation from Private Markets Performance. The book takes the same case from first principles to the decision, chapter by chapter, and every figure it prints is a live formula in the free companion workbooks.
Get the book on Amazon →Free companion files
Also on Amazon UK · Amazon Germany · Amazon France · Amazon Canada
Reading guide: private equity and private markets → · All 681 articles →
If this book helped, or didn’t, a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.