Introduction
The numbers on a fund's performance page come from three different traditions. The multiples are fund accounting: distributions to paid-in capital (DPI), residual value to paid-in capital (RVPI), and total value to paid-in capital (TVPI) measure cash returned, value still held, and their sum against what limited partners (LPs) have put in. The internal rate of return (IRR) comes from corporate finance and adds time. The public market equivalent (PME), best known in the ratio form Steven Kaplan and Antoinette Schoar published in 2005, comes from academic research and asks whether the same cash flows would have earned more in a stock index. Built for different questions, they can disagree about the same fund, and the gaps between them are where timing, borrowing, and generous valuations show up.
Which question LPs ask first has changed with the cycle. After years of thin exits, Bain's 2026 Global Private Equity Report describes LPs asking for evidence of strong IRRs and steady DPI, not IRR alone. For a private capital advisory (PCA) banker the same numbers are also pricing inputs, because a secondary buyer pays for what a fund still holds, not for what it has returned.
The Multiples: DPI, RVPI, and TVPI
Paid-In Capital as the Common Denominator
All three multiples divide by paid-in capital, the cumulative amount LPs have actually contributed, not their commitments. Paid-in includes capital called for management fees and expenses as well as investments, so fees weigh on every multiple from the first call. Distributions are what LPs receive after the general partner's (GP's) carried interest, and residual value is their share of the fund's net asset value (NAV), also reported net of accrued carry.
- DPI (Distributions to Paid-In Capital)
The ratio of the cumulative distributions a private fund has paid its limited partners to the capital they have paid in. Also called the realization multiple, it counts only value actually returned, so a DPI of 1.0x means LPs have received back exactly what they contributed.
RVPI divides the remaining NAV by the same paid-in capital, and TVPI adds the two, giving realized plus unrealized value per dollar contributed:
The split matters more than the total. The table follows one hypothetical buyout fund through its life, with figures in millions of dollars:
| Fund age | Paid-in | Distributions | NAV | DPI | RVPI | TVPI |
|---|---|---|---|---|---|---|
| Year 3 | 60 | 5 | 61 | 0.08x | 1.02x | 1.10x |
| Year 6 | 90 | 45 | 108 | 0.50x | 1.20x | 1.70x |
| Year 9 | 100 | 120 | 60 | 1.20x | 0.60x | 1.80x |
| Year 12 | 100 | 175 | 0 | 1.75x | 0.00x | 1.75x |
Early on, almost all of TVPI is RVPI, the GP's valuation of companies it has not sold. By year nine most of the value is cash, and at liquidation DPI and TVPI are the same number. The DPI column also traces the J-curve described in the fund lifecycle article, staying near zero while capital is called.
That is why how fund NAV is set matters to every multiple containing it, and why LPs in a slow exit market lean on DPI, the only multiple that does not depend on a valuation; the exit routes that pressure pushes sponsors toward are compared in PE exit strategies: sale, IPO, or recap.
MOIC Versus TVPI
Multiple on invested capital (MOIC) is the deal team's version of the ratio: total value divided by the capital invested in a company or portfolio. The arithmetic matches TVPI; the perspective does not. MOIC is usually quoted gross, on money that went into companies, while TVPI is usually net, on everything LPs paid in, after fees, expenses, and carry.
Suppose LPs have paid in 100, of which 90 bought companies and 10 covered fees and expenses, and the portfolio is now worth 225. Gross MOIC is 2.5x. If the fund is fully caught up in its waterfall, accrued carry takes 20% of the 125 of profit, leaving LPs 200 on 100 paid in: a net TVPI of 2.0x. The deal-level version of these metrics is covered in LBO returns: IRR, MOIC, and cash-on-cash.
IRR: Adding Time, and Opening the Door to Timing Effects
Multiples ignore time: 1.8x in four years and in twelve look identical. IRR fixes that with the annual rate that sets the net present value of all LP cash flows to zero, treating the current NAV as a final distribution:
Funds usually report a since-inception IRR, from the first cash flow to the latest quarter, which makes it path-dependent: early cash flows stay in the calculation for good, and an unrealized fund's IRR rests partly on its NAV.
Same Multiple, Different IRR
In each of four simplified patterns, an LP pays in 100 at the start; the columns show what comes back and when.
| Pattern | Distributions | TVPI | IRR |
|---|---|---|---|
| A: early exit | 180 in year 4 | 1.80x | about 15.8% |
| B: late exit | 180 in year 8 | 1.80x | about 7.6% |
| C: quick flip, then late exit | 50 in year 1, 110 in year 8 | 1.60x | about 9.2% |
| D: late exit, lower multiple | 160 in year 8 | 1.60x | about 6.1% |
A and B show the time value of money: the same multiple earns half the IRR when the cash takes twice as long, a trade-off unpacked in this comparison of IRR, MOIC, and cash-on-cash. C and D show a ranking reversal. Both return 1.6x, but C's year-one 50 lifts its IRR by about three points, enough to rank it above B, a fund that made its LPs more money.
Early Distributions and the Reinvestment Assumption
Pattern C is why managers prize an early realization: the lift to since-inception IRR persists however slowly the rest of the portfolio exits, because that cash flow never leaves the calculation. IRR also treats early distributions as if they kept compounding at the fund's own rate, the reinvestment assumption, though an LP handed 50 in year one may have nowhere to earn that return. A high IRR on a modest multiple usually means a few early exits, not a portfolio that compounded.
Subscription Lines: Higher IRR, Lower Multiple
A subscription line lets a fund buy companies with a short-term loan secured on LPs' uncalled commitments and call the capital later. Shortening the time LPs' money is outstanding raises IRR; the interest and fees reduce the multiple. The Institutional Limited Partners Association (ILPA) illustrated the trade-off in its June 2017 guidance on subscription lines: in a simplified example, delaying the first capital call by two years raised IRR from 6.62% to 7.98% while TVPI fell from 1.45x to 1.35x. It also cited an analysis by Cobalt of 498 funds in which the median IRR uplift was 206 basis points by year three, shrinking to 35-45 basis points by the end of fund life.
The uplift is largest early, exactly when a manager markets its next fund on an interim IRR. ILPA therefore recommended reporting net IRR with and without the facility, and accruing the preferred return from the date the line is drawn. How the line changes an LP's cash planning is covered in capital calls, distributions, and the LP cash-flow problem.
Gross Versus Net, With and Without the Facility
Gross performance measures the portfolio: cash flows between the fund and its companies, before fees, fund expenses, and carried interest. Net performance measures the LP's experience, after all of them. The spread is the price of the fund, widened by fees early and by carry once the fund is well in the money, as the worked example in the fees, carry, and waterfall article shows.
ILPA's performance template, released in January 2025 for adoption from the first quarter of 2026, asks for both, each split by the impact of fund-level subscription facilities. That gives four views of one fund:
- Gross, without the facility: the investments, as if capital had been called on each deal date.
- Gross, with the facility: the same investments, including the line's timing and cost.
- Net, without the facility: the LP's return had capital been called directly.
- Net, with the facility: the LP's actual cash experience, the usual headline.
Mixing the boxes is the classic presentation problem. For US-registered advisers, the Securities and Exchange Commission (SEC) staff's February 2024 marketing FAQ treats a gross IRR without the facility, shown only beside a net IRR with it, as a violation of the marketing rule. How a placement agent reconciles gross and net figures is covered in the PPM, data room, and LP due diligence.
PME: Measuring a Fund Against the Public Market
IRR and TVPI say whether a fund made money, not whether it beat the alternative: a 12% IRR earned while listed equities compounded at 14% is a poor result. Setting the fund's IRR against the index's annual return is unreliable, because capital moves at irregular dates. The public market equivalent instead runs the fund's own cash flows through the index.
- Public Market Equivalent (PME)
A family of measures that compare a private fund with a public index by investing the fund's own contributions and distributions in the index on the same dates. The Kaplan-Schoar PME (KS-PME) is the ratio of distributions plus NAV to contributions, each compounded at the index return; a value above 1.0 means the fund beat the index net of fees.
In the Kaplan-Schoar form, each contribution C and distribution D is grown to the end date T using the index level I, with the remaining NAV counted as a final distribution:
Take an LP that pays in 100 at the start and receives 60 in year three and 90 in year six: a TVPI of 1.5x and an IRR of about 9%. If the index returned 10% a year, the contribution grows to about 177 by year six and the distributions, each compounded from its own date, to about 170. The KS-PME is about 0.96: the fund made money, yet its LPs ended with roughly 4% less than the index would have given them.
Applied to real data, the measure reframed the debate over buyout returns. Using Burgiss cash-flow data on nearly 1,400 US funds, Robert Harris, Tim Jenkinson, and Steven Kaplan found the average US buyout fund beat the S&P 500 by 20% to 27% over its life, more than 3% a year, while average venture funds outperformed in the 1990s and lagged in the 2000s.
Direct Alpha and the Choice of Index
KS-PME covers the whole life, so it does not say how long the outperformance took. Direct Alpha, proposed by Oleg Gredil, Barry Griffiths, and Rüdiger Stucke in 2014, annualizes it: the IRR of the fund's cash flows after each is compounded to the end date at the index return. In the example it is about minus 0.9% a year, a 9% fund against a 10% index. With real index returns, which vary year to year, the answer also depends on whether calls and distributions fell before or after rallies, which a simple IRR comparison cannot see.
The benchmark index is the other judgment call. In the same sample, the average buyout PME was 1.20 against the S&P 500 but 1.07 against the Russell 2000 Value, an index closer to the smaller, value-oriented companies buyout funds tend to own. A European pension would benchmark a European program against a European or global index, and a PME is only as meaningful as its index.
Quartile Rankings and Vintage-Year Comparison
Absolute returns also need a peer group. Funds raised in the same year invest and exit through the same markets, so LPs compare a fund with others of its strategy and vintage year, using data from providers such as Cambridge Associates, Burgiss, Preqin, and PitchBook. Its position in that set is its quartile ranking.
- Top-Quartile Fund
A private fund whose performance ranks above the 75th percentile of a benchmark peer group, usually funds of the same strategy and vintage year, on a stated metric such as net IRR, TVPI, or DPI as of a stated date. The label depends on the dataset, the metric, and the date chosen.
That last sentence is where the label gets stretched. A manager can quote the metric on which it ranks best, the dataset where its peers look weakest, and its strongest quarter, and a young fund's ranking rests mostly on untested RVPI. ILPA's 2017 paper added a subtler distortion: by delaying the first call, a subscription line can shift a fund's quartile and, where vintage is dated by first cash flow, its vintage-year classification. A fair comparison holds strategy, vintage, metric, dataset, and date constant, and shows DPI and PME beside IRR.
LPs rank funds to predict the next one, and the evidence for that is weaker than pitch books suggest. Harris, Jenkinson, Kaplan, and Rüdiger Stucke found strong buyout persistence on final performance, but little or none after 2000 when previous funds were judged on the performance available at fundraising, the information an LP actually has. Venture persistence held on either basis.
How Secondary Buyers and Advisors Read the Same Track Record
A primary LP deciding whether to re-up reads the whole record. A secondary buyer of the same fund interest reads little of that history: it buys the remaining NAV and any unfunded commitment, so its return depends only on how much the portfolio still distributes, when, and what it pays today. The advisor has to present one set of numbers that works for both.
| Reader | Main question | What it weights most |
|---|---|---|
| Primary LP re-upping | Should we back this manager again? | Net IRR, TVPI, DPI, and PME against vintage peers |
| Secondary buyer of an LP interest | What will the remaining assets pay, and when? | RVPI, NAV credibility, distribution timing, unfunded commitments |
| Continuation vehicle investor | What is this asset worth from today? | Entry price, company prospects, exit path |
| Advisor running the sale | Can every bidder compare the numbers? | Gross or net, facility impact, reference date, realized versus borrowed cash |
The differences show up most clearly in three situations.
Buyers Price RVPI and the Timing of Future Cash
Take an LP that committed 50 to the fund in the earlier table. In year six it has paid in 45, received 22.5, and holds a NAV of 54: DPI 0.50x, RVPI 1.20x, TVPI 1.70x. A buyer bidding 90% of NAV pays a purchase price of 48.6 and takes on the remaining 5 of unfunded commitment. If it expects 63 of distributions, its own multiple is 63 on 53.6, about 1.18x, and treating the calls as paid up front, its IRR turns on timing: roughly 8% a year if the cash arrives within two years, about 4% if it takes four. The fund's 1.70x and lifetime IRR matter only as evidence of how reliable the marks and the GP's exit pace have been, the judgment set out in how secondary buyers assess a fund interest.
The seller reads the trade as a conversion of RVPI into DPI at 90 cents on the dollar: its DPI jumps from 0.50x to about 1.58x, which is also its final TVPI, down from 1.70x the quarter before. Whether that is a good trade depends on what else the cash can do, which is why the conversation is framed as a discount to NAV rather than an IRR, as pricing LP interests explains.
Why a Continuation Vehicle's IRR Resets
A continuation vehicle (CV) splits one asset's history into two track records. For the old fund, the sale is a realization at the transfer price: selling LPs' cash counts as distributions, lifting DPI, and the asset's contribution to the old fund's IRR and TVPI is fixed at that price. The CV starts a new record at the same price, with a TVPI near 1.0x and an IRR clock that begins at closing. For rolling LPs, whether the rollover appears as a distribution and a new contribution depends on how the deal and its reporting are structured; their real experience runs across both vehicles.
The reset cuts both ways. A strong asset earns a fresh return on a tested price, but a manager can market two clean records, neither showing what continuing investors earned, so an LP deciding whether to roll should ask for look-through performance across both vehicles. The DPI lift is part of the appeal: in Dechert's 2026 Global Private Equity Outlook survey, 46% of respondents said they were using GP-leds or CVs to deliver distributions. How the transfer price meets carry and the GP's rollover is covered in CV economics.
Distributions Funded by NAV Loans
A NAV loan, secured on the portfolio, can raise DPI without selling anything. If the year-six fund in the table borrows 10 and distributes it, DPI rises from 0.50x to about 0.61x, while NAV, net of the loan, falls to 98 and RVPI to about 1.09x. TVPI is unchanged at 1.70x on the day and drifts lower as interest accrues. ILPA's 2024 guidance on NAV facilities notes that such distributions materially move IRR and DPI, are often recallable, and leave LPs rebuilding a synthetic DPI to see what the manager has realized. An advisor preparing an LP sale or a fundraise should show realized DPI separately from borrowed DPI and put the loan in the NAV bridge bidders see; terms and consents are covered in NAV lending in practice.
Every metric on the page converges at the end of a fund's life. When the last asset is sold, RVPI falls to zero, TVPI equals DPI, the IRR stops moving, and the PME rests on cash alone. Until then every figure except DPI blends cash with an estimate, and timing, borrowing, and marks push them in different directions. A primary LP decides how much of that blend it believes. A secondary buyer prices the part still unresolved, the gap between TVPI and DPI, and the advisor's work is making that gap legible on one consistent basis.
