Private equity fund economics explained: fees, carry, waterfalls and DPI
A private equity fund makes money for its managers in two ways: a fee for running the fund and a share of the profits. Understanding how those are calculated, and how performance is reported along the way, is basic fluency for anyone working in or around a fund. This guide explains the pieces and works through an invented example.
How does the GP get paid?
A fund is usually a limited partnership. Limited partners (LPs) commit capital, and the general partner (GP) manages it. The GP earns two things.
Management fee. An annual percentage, commonly quoted around 1.5% to 2% for buyout funds, though practice varies. During the investment period it is typically charged on committed capital, and afterwards often on invested capital at cost. Fees are paid out of LP money, so they reduce what is available to invest.
Carried interest ("carry"). A share of the fund's profit, conventionally 20%, paid only if the fund clears certain tests. It is the part that ties the GP's reward to results.
What are the hurdle and the catch-up?
The hurdle (or preferred return) is a minimum annual return LPs must receive before the GP shares in profit, often around 8% compounded. The catch-up then lets the GP receive most or all of the next distributions until it holds its full 20% of total profit. With a full catch-up, a fund that clears the hurdle comfortably ends up paying carry on all profit, not only the part above the hurdle. Without a catch-up, carry is paid only on profit above the hurdle. Terms differ by fund.
What is the difference between a European and an American waterfall?
A European (whole-fund) waterfall pays out in order: return of all contributed capital to LPs, then the hurdle, then the catch-up, then an 80/20 split. The GP is paid carry only after LPs have their whole investment back plus the hurdle.
An American (deal-by-deal) waterfall calculates carry on each investment as it is sold. It pays the GP earlier, and if early deals win and later ones lose, the GP may have received more than 20% of final profit. To deal with that, agreements usually include a clawback, where the GP returns the excess, and sometimes an escrow. Both models exist, and practitioners disagree about which is fairer to LPs.
Worked example: the Fenwick Ridge fund
The fund and its figures are invented. To keep the arithmetic readable, assume LPs contribute 100 million in total (fees included) at the start, the fund returns a single distribution of 220 million at the end of year 5, the hurdle is 8% compounded, carry is 20% and there is a 100% catch-up. Figures are in millions.
Step 1: capital and hurdle. LPs first receive 100 back. The hurdle amount is 100 x 1.08^5 = 146.93, so the preferred return is 46.93.
Step 2: catch-up. Total profit will be 220 - 100 = 120. For the GP to hold 20% of profit (pref plus catch-up), the catch-up is 46.93 x 0.20 / 0.80 = 11.73. The GP gets 11.73 and the LPs nothing in this tier.
Step 3: the split. What remains is 220 - 100 - 46.93 - 11.73 = 61.33. The LPs receive 80% (49.07) and the GP 20% (12.27).
Result.
| Item | LPs | GP |
|---|---|---|
| Capital returned | 100.00 | 0.00 |
| Preferred return | 46.93 | 0.00 |
| Catch-up | 0.00 | 11.73 |
| 80/20 split | 49.07 | 12.27 |
| Total | 196.00 | 24.00 |
The GP's carry is 24.00, which is exactly 20% of the 120 profit. The LPs receive 196.00, a net multiple of 1.96 times against a gross 2.20 times. As annual returns, gross is 2.21/5 - 1 = 17.1%, and net to LPs is 1.961/5 - 1 = 14.4%. The gap between those two figures reflects carry only, because fees are already inside the 100 million contributed.
Without a catch-up, carry would be 20% of (120 - 46.93) = 14.61, to show why GPs negotiate for it.
American comparison. Suppose the fund had just two deals, and the first sold for a profit of 80 while the second lost 20. Taking carry at the first sale, ignoring the hurdle, the GP would receive 20% x 80 = 16. A whole-fund test on the final net profit of 60 would give 12. The extra 4 is what a clawback is designed to return.
What does the J-curve look like?
Early on, performance reported to LPs typically dips below cost. Fees are charged from day one, while investments are held at cost or written up slowly, and few exits have happened. As companies grow and are sold, value rises and cash comes back. Plotting returns against time gives a J shape.
An invented fund (100 million called over three years, 220 million returned by year 8) might look like this at a few dates:
| Year | Paid-in | Distributed | Remaining value | DPI | TVPI | IRR |
|---|---|---|---|---|---|---|
| 1 | 70 | 0 | 66 | 0.00x | 0.94x | -10% |
| 3 | 100 | 0 | 112 | 0.00x | 1.12x | 5.5% |
| 5 | 100 | 60 | 85 | 0.60x | 1.45x | 9.8% |
| 7 | 100 | 170 | 30 | 1.70x | 2.00x | 14.5% |
| 8 | 100 | 220 | 0 | 2.20x | 2.20x | 15.7% |
What do DPI, TVPI and IRR each tell you?
- DPI (distributions to paid-in) is cash actually returned ÷ capital paid in. It is realised and cannot be argued with.
- RVPI (residual value to paid-in) is the remaining reported value ÷ paid-in. It rests on the GP's valuations.
- TVPI (total value to paid-in) is DPI plus RVPI. Early in a fund's life it relies heavily on valuations.
- IRR is the annual return that makes the net present value of the cash flows zero. It depends on timing: it rewards early returns of cash, and it can be flattered by short holding periods or subscription credit lines. Money multiples ignore timing. Read them together.
What mistakes do people make?
- Mixing up commitments and paid-in capital.
- Comparing IRRs of funds at different ages.
- Treating TVPI as cash.
- Applying the hurdle to profit only, rather than to capital contributed from the date it was paid in.
- Forgetting that the example above assumes one cash flow in and one out, which real funds do not have.
About this guide: it is original to Charta Learning, uses an invented fund and has not yet been reviewed by an experienced practitioner. Real partnership agreements differ on fee bases, offsets, hurdle compounding, catch-up percentage, timing of calls and distributions, and clawback mechanics.