TVPI, DPI, and RVPI: Fund-Level Return Metrics for LPs

Key Takeaways

Why DPI Matters More in 2026

For most of the 2010s, LPs evaluating private-equity and real-estate funds led with IRR and treated TVPI as the secondary multiple. DPI — the share of paid-in capital that had been distributed back as actual cash — was the third metric on the page, mentioned mostly when a fund was approaching the end of its life. That ordering has flipped. In 2026, DPI is the metric LPs ask about first when re-up decisions hit the desk.

The numbers explain why. Per McKinsey's 2026 Global Private Markets Report, distributions ran at approximately 6% of AUM in the 12 months ending June 2025, against a 16% average from 2015–2019 and a 14% 10-year average through 2024. Five-year rolling DPI/AUM hit its lowest recorded level in 2025. Bain's 2026 LP survey of ~300 LPs now ties DPI with MOIC as the second-most-important allocation metric — IRR still leads, but its primacy has narrowed. The cash LPs aren't receiving still exists in fund NAV; it just hasn't been realized and isn't available for new commitments.

This article is the practitioner reference for the fund-level multiple cluster: the formulas, the TVPI = DPI + RVPI identity that organizes everything, a worked $80M CRE fund walked through its full 10-year lifecycle, net- vs-gross mechanics, 2026 benchmarks, and the decision framework for when each multiple matters most.

The TVPI = DPI + RVPI Identity

The intuition matters. DPI is the past: capital that has come back to the LP's account, available for redeployment, recyclable into the next vintage. RVPI is the present-and-future: the NAV the GP currently reports on hold-stage assets, subject to realization through future sales. TVPI is the combined claim — how the fund is doing at any point in time, totaling realized and unrealized value.

The identity also clarifies why DPI has become the focal metric. A fund with TVPI of 1.8x looks healthy until you decompose it: TVPI of 1.8x with DPI of 1.4x means the LP has received 1.4x of paid-in capital and has another 0.4x in residual NAV waiting to be realized. TVPI of 1.8x with DPI of 0.3x means the LP has received only 0.3x and has 1.5x in unrealized NAV that may or may not convert at the marked value. Same headline; very different risk profiles.

The Formulas

• DPI = Cumulative Distributions ÷ Paid-In Capital. Realized cash returned to the LP divided by capital actually drawn (not capital committed; the distinction matters). At fund maturity, DPI equals the fund's final realized multiple. Mid-life, DPI describes how much of the fund is in the rearview.

• RVPI = Residual NAV ÷ Paid-In Capital. Net asset value of remaining portfolio positions divided by paid-in capital. RVPI is GP-reported (or third-party-appraised); it represents unrealized value subject to mark-to-market revisions and ultimate realization at sale.

• TVPI = (Cumulative Distributions + Residual NAV) ÷ Paid-In Capital. The overall scorecard. Mathematically identical to DPI + RVPI.

All three use paid-in capital (capital drawn) in the denominator rather than commitment (capital pledged but not yet called). This is the standard institutional reporting convention — ILPA, Cambridge Associates, Preqin, and Pitchbook all use paid-in. The distinction is consequential because a fund that has drawn 60% of commitments looks different on a paid-in basis than a commitment basis, particularly early in the j-curve.

The J-Curve

Private-fund multiples typically follow a j-curve shape over the fund's life: TVPI dips below 1.0x early, recovers through the middle of the fund's life, and rises through realization. The dip happens because management fees and transaction costs are paid before meaningful value creation has occurred — a year-one LP statement might show paid-in capital of $20M, distributions of $0, and NAV of $18M (reflecting the impact of fees and immature investments). TVPI = 0.9x, DPI = 0.0x, RVPI = 0.9x. The LP is technically underwater.

The j-curve recovers as portfolio assets mature, NOI grows (for CRE), and a few early realizations bring DPI off zero. By year four or five of a 10-year fund, TVPI typically crosses 1.0x and continues climbing. DPI starts low and rises through the harvest period; RVPI peaks mid-life as portfolio NAV builds, then declines as positions realize and convert to DPI. At maturity, RVPI approaches zero and TVPI converges to DPI.

Two CRE-specific notes on the j-curve. The dip is shallower than buyout: CRE assets generate current cash flow from rents starting in year one or two, which keeps the j-curve from dipping as far as buyout, where portfolio companies often produce no distributable cash for the first few years. RVPI moves with cap rates: a CRE fund's residual NAV depends on the cap rates at which appraisers (or the GP) mark the portfolio. A 50-bps cap rate move can shift RVPI by 10–15% without any operational change, which is a meaningful source of TVPI volatility unique to CRE funds.

Worked Example: $80M CRE Fund Lifecycle

A $80M closed-end CRE value-add fund. 8-year investment period followed by a 2-year harvest. Year-by-year, the three multiples and the underlying IRR look approximately like this:

Reading the table: in year 1, the fund is in the j-curve dip — $20M called, $0 distributed, $18M NAV (reflecting 10% drag from fees and acquisition costs on still-immature assets). By year 4, paid-in is fully drawn at $80M, $15M has been distributed, and NAV has built to $95M as the value-add business plans take hold — TVPI 1.38x, but DPI is only 0.19x. Between years 6 and 7, the harvest period inflects: distributions cross paid-in capital (DPI = 1.0x), NAV is being realized into cash. At maturity in year 10, the LP has received $145M in cash (DPI 1.81x), $5M of residual NAV remains for a long-tail wind-down, and TVPI has converged to 1.88x. IRR settles around 14%.

The pattern matters. A re-up decision in year 4 (TVPI 1.38x, DPI 0.19x) is a different conversation than one in year 7 (TVPI 1.75x, DPI 1.0x). The TVPI numbers are similar; the cash-realization profile is dramatically different.

Net vs. Gross

Every multiple in this cluster has a gross version and a net version. Gross means before fund management fees and carried interest — the GP's view of what the deals actually produced. Net means after fees and carry — what the LP actually receives. Public benchmarks (Cambridge Associates, Preqin, Pitchbook) report net by default. Marketing materials sometimes use gross to flatter results.

For buyout funds, the gross-to-net delta typically runs 10–20% per Waveup's 2026 benchmarking — a 3.5x gross TVPI delivers approximately 2.5–2.8x net depending on fund structure. For CRE the delta is wider on value-add and opportunistic funds (heavier promote structures) and tighter on core funds (lower fees and lower performance loadings). The spread is determined by management fee percentage, preferred return rate, catch-up provisions, and the carry percentage above pref.

The same logic applies to DPI and RVPI. Net DPI is what the LP has in their pocket. Net RVPI is the residual value the LP will receive after future fees and carry. The TVPI = DPI + RVPI identity holds on a net basis the same as on a gross basis; the inputs are just scaled by the fee impact. When evaluating a fund, always ask which is reported — and prefer net for benchmarking, gross for evaluating GP deal-selection skill.

When Each Metric Matters Most

All three multiples describe the same fund; they emphasize different dimensions of performance.

• DPI matters most for: re-up decisions, GP fundraising track records, secondary sale pricing, and late-stage fund evaluations. A GP raising a successor fund without a meaningful DPI on the predecessor has to defend the absence; an LP committing to a follow-on commits real cash and wants to see real cash returned. DPI is the metric that compounds — it recycles into the next vintage. RVPI doesn't.

• RVPI matters most for: mid-life fund monitoring, manager-skill assessment on the unrealized portion, and cap-rate stress testing for CRE funds. A fund showing 1.5x RVPI in year 5 with 0.5x DPI is sitting on most of its return in unrealized NAV — the LP needs to assess whether the GP's marks are defensible, whether comp data supports the cap rates, and what happens to RVPI under a 50–100 bps cap rate stress. RVPI is also the metric most subject to manipulation: GP marks on illiquid assets are opinions until they're transactions.

• TVPI matters most for: overall scorecard at fund maturity, cross-fund comparison, and top-quartile benchmarking. When a fund is being compared to its vintage peers, TVPI is the headline. When underwriting a fund commitment, the realistic TVPI projection is the answer to "what should I expect to end up with."

Combining the three: TVPI tells you the score, DPI tells you how much is in the bank, RVPI tells you what's still in play. An LP scorecard that quotes only TVPI is incomplete. The full picture requires all three.

TVPI vs. MOIC Reconciliation

TVPI and MOIC are mathematically siblings. MOIC at the deal level: total distributions divided by total equity contributed. TVPI at the fund level: (total distributions + residual NAV) divided by paid-in capital. Same construction; different scope. The two converge as a fund nears the end of its life (when RVPI → 0 and the fund's TVPI equals the LP-weighted average of deal-level MOICs net of fund-level fees).

The practical distinction: MOIC is what an individual deal produces; TVPI is what an LP holds across the fund's portfolio of deals minus fees. A fund with five deal-level MOICs of 2.0x might produce a TVPI of 1.7x after fees and carry. The MOIC tells you whether the deals worked; the TVPI tells you whether the fund structure delivered them to the LP intact. The deeper treatment is in the sibling article on equity multiple and MOIC.

The 2026 Macro Context

Four data points anchor why the fund-level cluster matters more in 2026 than any time in the last decade:

• DPI compression. McKinsey's 2026 Global Private Markets Report documents distributions running at ~6% of AUM in H1 2025, against a 16% average from 2015–2019 and a 14% 10-year average through 2024. Five-year rolling DPI/AUM at the lowest recorded level in 2025. LPs are receiving roughly half the realized cash they received in the late 2010s — while paying the same management fees on the same NAV.

• Hold-period extension. Per Bain's 2026 Global Private Equity Report, median holds at exit now sit near seven years, up from five to six in 2010–2021. Roughly 40% of buyout-backed companies globally are held longer than five years (vs 29% in 2019). CRE has experienced a parallel extension as the refi-environment forces sponsors to hold past planned exit windows. Both extensions push DPI right on the timeline.

• LP preference shift. Bain's 2026 LP survey of ~300 LPs places DPI tied with MOIC as the second-most-important allocation metric. IRR still leads but its primacy has narrowed materially since 2019. The same survey shows roughly two-thirds of LPs would prefer to wait for an improved DPI multiple over accepting near-term liquidity at a discount.

• Secondary market growth. Per Preqin and J.P. Morgan, secondary transaction volume hit $226B in 2025 (+41% YoY), with GP-led continuation vehicles growing 62% YoY. Secondaries function as a DPI alternative — LPs sell residual NAV (RVPI) at a discount to convert it to cash (DPI). The growth of this market is now a structural feature of LP liquidity rather than a workaround.

A subsidiary note on NAV facilities: per Proskauer / Moonfare data, the NAV-loan market is projected to grow from ~$100B currently to ~$600B by 2030. GP use of NAV loans to fund LP distributions fell ~90% in H2 2023 after ILPA pushback; current primary use (~45% of facilities) is follow-on investments. The footnote LPs should know: NAV-loan-funded distributions inflate reported DPI without underlying value creation. Ask whether DPI is operationally generated or NAV-loan-funded.

2026 Benchmarks

Per Cambridge Associates' Q4 2025 Private Equity Index and Real Estate Index, mature-vintage benchmarks look approximately like this:

The DPI compression in the 2018–2020 vintages is visible directly: median net DPI of 0.4–0.6x at mid-life is substantially below the 0.8–1.0x that comparable vintages were showing in 2018–2019. The TVPI compression is smaller (1.4–1.6x vs ~1.8x for prior vintages) because RVPI has held up — but the unrealized portion of those TVPI numbers is now a much larger share, raising the question of how much of the headline ultimately converts to cash.

Five Mistakes LPs Make

• Treating TVPI and DPI as interchangeable. A 1.8x TVPI with 1.4x DPI and a 1.8x TVPI with 0.3x DPI are two very different funds. The first has delivered most of its return; the second is still sitting on it. Always decompose.

• Trusting GP-reported RVPI without scrutiny. RVPI is the GP's mark on hold-stage assets. Independent appraisal, transaction-based comp data, and stress tests on cap-rate assumptions are the LP defense against optimistic marks. The most common source of disappointment in PE/CRE fund commitments is RVPI that didn't convert at the marked level.

• Comparing gross-quoted to net-quoted multiples. If one fund reports 3.5x and another reports 2.8x, the headlines aren't comparable until you confirm both are net or both are gross. The gross-to-net delta of 10–20% can flip the comparison.

• Ignoring vintage when benchmarking. A 2018-vintage fund showing 1.5x TVPI at year 6 looks different than a 2014-vintage fund showing 1.5x TVPI at year 8. Vintage matters because market conditions, deployment pace, and realization windows differ across cycles. Cambridge Associates and Preqin publish vintage-specific benchmarks; use them.

• Not asking whether DPI was operationally generated or NAV-loan-funded. NAV-loan distributions inflate DPI without underlying realizations. They show up as paydowns in subsequent quarters. Ask the GP directly.

Do It in Apers

Related Articles

• Equity Multiple and MOIC: When the Multiple Matters More Than IRR — the deal-level cousin to TVPI.
• IRR Calculator and Formula for Real Estate — the time-weighted complement.
• IRR Formula and the Reinvestment-Assumption Question — methodology pillar on IRR.
• Cash-on-Cash Return: Levered vs. Unlevered — the current-yield metric.
• IRR Sensitivity Analysis and Stress Testing — how to test fund-level IRR under stress.
• Preferred Return Mechanics — how preferred return shapes net DPI and net TVPI.
• Promote / Carried Interest, Step by Step — the gross-to-net mechanics.

FAQ