Private Equity IRR: Reconcile Gross, Net, DPI and TVPI
Audit private equity returns with a complete gross-to-net cash-flow example, fee calculations, NAV sensitivity, and reconciled DPI, RVPI and TVPI.
Net private equity IRR must be calculated from the investor's net cash flows. You cannot obtain it by subtracting a management-fee percentage and carry percentage from gross IRR. Payment timing and the contractual fee calculation both affect the answer.
The IRR & XIRR calculator accepts annual or dated amounts. It calculates the return on the schedule you supply; it does not calculate a private equity waterfall or determine which expenses belong in a gross or net reporting measure. The examples here show how to build and reconcile those inputs before using the tool.
Establish what gross and net mean in the report#
A gross return commonly describes investment performance before specified fund fees, expenses and carried interest. A net LP return uses the investor's contributions, distributions and remaining net value. The reporting level matters: a portfolio-level measure and an LP-level measure may use different transactions and adjustments.
ILPA's Performance Template provides definitions and calculation guidance to make those reporting choices explicit. Use the report's methodology to reconcile its figures; a label alone is insufficient.
Fees and carry are governed by the investment documents. As Investor.gov explains, the offering documents and agreements should disclose the fees and expenses investors incur. A shorthand such as "2 and 20" does not specify the fee base, payment dates, preferred return, catch-up, waterfall or clawback.
A reproducible gross-to-net example#
This simplified, hypothetical five-year model has the following terms:
- $100,000 is invested at time zero and sold for $200,000 at year five.
- A $2,000 management fee is paid separately at the end of each of years one through five: $10,000 total.
- Carry equals 20% of the $100,000 investment profit, or $20,000, paid on exit. For this example only, the carry base excludes management fees and there is no hurdle or catch-up.
- There are no other expenses, interim distributions, taxes, borrowing or remaining assets.
These are the complete assumptions of an illustration, not standard terms for every fund.
| Year | Investment cash flow before fees | Management fee | Carry | Net LP cash flow |
|---|---|---|---|---|
| 0 | -$100,000 | $0 | $0 | -$100,000 |
| 1 | $0 | -$2,000 | $0 | -$2,000 |
| 2 | $0 | -$2,000 | $0 | -$2,000 |
| 3 | $0 | -$2,000 | $0 | -$2,000 |
| 4 | $0 | -$2,000 | $0 | -$2,000 |
| 5 | $200,000 | -$2,000 | -$20,000 | $178,000 |
Enter these annual schedules separately in the calculator, retaining the zero periods:
Gross investment schedule:
-100000, 0, 0, 0, 0, 200000
IRR = 14.86983550%
Net LP schedule:
-100000, -2000, -2000, -2000, -2000, 178000
IRR = 10.87707231%
The gross result also has a direct check: (200000/100000)^(1/5)-1. For net IRR, substituting the unrounded rate in this equation gives approximately zero:
-100000 - 2000/(1+r) - 2000/(1+r)^2 - 2000/(1+r)^3
- 2000/(1+r)^4 + 178000/(1+r)^5 = 0
The difference is 3.99276319 percentage points in this example. That difference is an output of the stated schedule. It is not a fixed deduction you can apply to another fund. Changing the sale date, fee dates, fee base or carry terms changes the result.
Reconcile DPI and TVPI without losing the separate cash legs#
For the example above, the LP contributes $110,000 in total: $100,000 for the investment and $10,000 in fees. The LP receives $180,000 after carry. The year-five IRR row nets the $180,000 distribution against the $2,000 fee paid on that same date, producing $178,000.
For the contribution and distribution ratios, retain those separate amounts:
Paid-in capital = 100000 + 5*2000 = 110000
Distributions after carry = 200000 - 20000 = 180000
Remaining NAV = 0
DPI = distributions / paid-in capital = 180000/110000 = 1.63636x
RVPI = remaining NAV / paid-in capital = 0.00000x
TVPI = DPI + RVPI = 1.63636x
Net undiscounted profit = 180000 - 110000 = 70000
Using $178,000 as the entire distribution and omitting the final fee contribution would preserve the net IRR cash flow but produce different, incorrectly netted ratios for this example. A useful reconciliation keeps both the transaction detail and the net amount by date.
The ILPA glossary defines DPI, RVPI and TVPI. For an actual report, also check its treatment of recallable distributions, recycled capital and in-kind distributions. Our example has none of those features.
Since-inception IRR can include money that has not been distributed#
A since-inception calculation can include the remaining net asset value as a terminal positive amount at the reporting date. That allows measurement before liquidation, but the NAV is a valuation rather than cash already received.
Consider a separate hypothetical fund with these annual amounts:
| Year | LP cash contribution or distribution | Remaining NAV included at reporting date |
|---|---|---|
| 0 | -$100,000 | — |
| 1 | -$50,000 | — |
| 2 | $20,000 | — |
| 3 | $40,000 | — |
| 4 | $0 | $140,000 |
The schedule [-100000,-50000,20000,40000,140000] produces 9.23629420% IRR. Paid-in capital is $150,000 and distributions are $60,000, so DPI is 0.40x, RVPI is 0.93333x, and TVPI is 1.33333x.
If the year-four NAV is instead $100,000, with every actual payment unchanged, IRR falls to 2.06085195%. DPI remains 0.40x and TVPI becomes 1.06667x. This is a sensitivity calculation, not a prediction: it isolates how the valuation assumption changes the reported return.
Use actual transaction dates and XIRR for a real irregular schedule. Keep the valuation date with the NAV, and avoid adding it twice if the reported terminal amount already includes a distribution.
Interpret the J-curve and benchmarks with the underlying data#
Early fees, expenses and valuation changes can produce weak early performance before a fund realizes investments. Later distributions or higher valuations may change that pattern. There is no guaranteed year when a negative early return becomes positive, and a J-curve description is not a reason to ignore losses or defer scrutiny.
For example, a one-year measurement with a $100,000 contribution and a $98,000 terminal net valuation has a -2% IRR even though no asset has been sold. That simple result says nothing about the eventual exit value. Review the valuation assumptions and cash use as well as the percentage.
A meaningful benchmark comparison needs the same strategy, vintage, reporting date and gross/net basis. State the data provider and peer group. A universal table of "good" percentages without those details can create a false comparison. A target in marketing material also measures a different thing from a realized or since-inception result.
Before comparing two reports, reconcile:
- Contributions, distributions and the valuation date to the reported IRR inputs.
- Fees, carry, expense allocations and any financing effects to the stated methodology.
- DPI, RVPI and TVPI to the same paid-in capital definition.
- Remaining NAV and its sensitivity to the amount of cash already distributed.
A high percentage is not a substitute for those checks. The IRR vs NPV guide explains why timing, scale and multiple roots also matter when comparing investment schedules.
No. Build the net schedule using the actual fee, carry, contribution and distribution amounts on their payment dates, then calculate its IRR. In the stated five-year example, 14.87% gross becomes 10.88% net; that gap is specific to its assumptions.
Not necessarily. The result can include the remaining NAV as a terminal value. In the example here, 9.24% IRR coexists with 0.40x DPI, because most of the reported value remains undistributed.
No. Enter the net cash flows after applying the fund's contractual terms. The calculator analyzes the supplied annual or dated schedule, reports NPV and detected IRRs, and can show MIRR for annual rows with stated rates.
References
Written by
Hassaan Rasheed
Web Developer & Content Researcher
Hassaan builds calculators and writes source-linked guides across the site's subject areas. Calculator methods and reference data are documented in each guide so readers can verify the underlying sources.
About the author and methods

