One fund ranked against nineteen vintage peers on IRR, TVPI and DPI, with both Excel quartile methods, and why a top-quartile claim can rest on a formula choice.
A quartile ranking compares a fund with the upper quartile, median and lower quartile of funds of the same vintage and strategy, on the same measure and the same valuation date. On the illustrative 2018 peer set below, a fund with a 17.9 per cent net IRR clears the upper quartile of 17.80 per cent on one quartile method and misses the 18.20 per cent threshold on the other, while its TVPI ranks second quartile and its DPI third.
Worked in full in Private Equity Investor Relations by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
Almost every fundraising deck claims a top-quartile track record, and almost every experienced limited partner recomputes the claim. The arithmetic is simple. What makes the result fragile is the set of choices behind it: which measure, which peer set, whether the fund counts itself, and which percentile formula. An investor relations team that has run the numbers every way can say exactly what is true; one that has run them once is exposed in the first due diligence call.
Nineteen illustrative buyout funds of vintage 2018, valued at the same date. Sorted by net IRR, in per cent:
−1.5, 2.8, 5.0, 6.9, 8.1, 9.6, 10.4, 11.5, 12.2, 13.0, 14.3, 15.1, 16.5, 17.4, 18.2, 19.6, 21.0, 24.8, 31.5.
Our fund: net IRR 17.9 per cent, TVPI 1.74x, DPI 0.56x.
QUARTILE.INC: position = (n − 1) × p, counted from zero, interpolated between neighbours.
QUARTILE.EXC: position = (n + 1) × p, counted from one, interpolated between neighbours.
In Excel, with the peer IRRs in B2:B20:
=QUARTILE.INC(B2:B20,3) returns 17.80
=QUARTILE.EXC(B2:B20,3) returns 18.20
With 19 peers, the inclusive method lands at position 13.5: halfway between the 14th and 15th values, 17.4 and 18.2, giving 17.80. The exclusive method lands at position 15.0, exactly on 18.2. The same nineteen numbers produce two upper quartiles 0.4 points apart, and our fund sits between them: 0.1 points above one and 0.3 points below the other.
| Measure | Our fund | Upper quartile | Median | Lower quartile | Quartile |
|---|---|---|---|---|---|
| Net IRR, % | 17.9 | 17.80 | 13.00 | 8.85 | 1st |
| TVPI, x | 1.74 | 1.83 | 1.58 | 1.37 | 2nd |
| DPI, x | 0.56 | 0.84 | 0.60 | 0.42 | 3rd |
The fund ranks first, second or third quartile depending on which measure is quoted. The IRR may owe something to early timing or a subscription line; the TVPI is ordinary; the DPI, the only one of the three made of cash, is below the median. An allocator reading only the first row would draw the wrong conclusion about the second two.
Benchmark providers publish quartiles on their own universe, which may or may not contain the fund being ranked. Adding the fund itself to the set moves the threshold. On the inclusive method it lifts the upper quartile from 17.80 to 17.98 per cent, because a value near the top is added; on the exclusive method it eases from 18.20 to 18.13 per cent, still above the fund. On IRR the four combinations give:
| Peer set | Method | Upper quartile | Our fund's quartile |
|---|---|---|---|
| 19 peers | Inclusive | 17.80 | 1st |
| 19 peers | Exclusive | 18.20 | 2nd |
| 20, including our fund | Inclusive | 17.98 | 2nd |
| 20, including our fund | Exclusive | 18.13 | 2nd |
One combination in four supports the top-quartile claim. That is not dishonest in itself, but it is the kind of fact a limited partner will find, and a claim that survives only one method is better stated as a rank: sixth of 20, or a percentile rank of 73.7 per cent: 14 of the other 19 funds are below it. Excel's =PERCENTRANK.INC(B2:B21,17.9) on the set including the fund returns 0.736, because it truncates to three digits.
Choosing the measure after seeing the ranks. Quoting the IRR because it is the only first-quartile figure is the most common version, and the one allocators are trained to spot. In this case the fund's honest description is: top quartile on IRR on one method, second quartile on TVPI, below median on DPI.
A good DDQ answer gives the measure, the benchmark provider, the vintage and strategy definition, the peer count, the valuation date, the method, and all three quartiles. It takes a table, not a sentence, and it costs the manager nothing when the result is strong.
Compute the thresholds with both quartile functions, with and without the fund, on IRR, TVPI and DPI, at the same valuation date as the peers. If the fund is top quartile on all of them, say so. If it is not, a rank and a percentile are more defensible than a label. The way one fund's return can be stated six true ways is shown in six ways to state the same fund's return, and the workbook behind it is the free workbook for this case.
A fund at or above the 75th percentile of its peer group: same vintage, same strategy, same measure, same valuation date. In the worked case the upper-quartile net IRR of 19 peers is 17.80 per cent on the inclusive method, so a 17.9 per cent fund qualifies, but only by 0.1 points and only on that method and peer set.
Use whichever the benchmark provider uses, and say which. They interpolate differently: on 19 peers the inclusive upper quartile is 17.80 per cent and the exclusive one 18.20 per cent. When a fund sits between the two, the ranking depends on the formula, and a rank such as sixth of 20 is the more defensible statement.
Often. IRR rewards early distributions and short holding periods, TVPI rewards total value. In the worked case the fund is first quartile on IRR, second on TVPI at 1.74x against an upper quartile of 1.83x, and third on DPI at 0.56x against a median of 0.60x. Allocators ask for all three.
This article is one calculation from Private Equity Investor Relations. 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 453 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.