ALLOCATOR SERIES FOR FUNDS OF FUNDS · No. 1 of 5
Underwriting a Venture Fund Before DPI Exists
We recently published a five articles series aimed at educating family offices on venture capital. We are now publishing a new five articles series aimed at fund of funds as they interact with the venture capital asset class. This is the first article of the series.
DPI is rightly the metric of this cycle — and it is silent exactly where the best entry points sit. Here is what actually predicts it.
Pascal Bouvier, MiddleGame Ventures · August 2026
Every allocator conversation in this cycle arrives, usually within the first ten minutes, at the same three letters: DPI. Rightly so. Industry data put 2023 distributions at a fourteen-year low as a share of fund value, and the recovery since has been modest. After a decade in which paper marks did most of the talking, limited partners want cash back, and the managers who return it have earned the pulpit. We wrote as much in the final piece of our family-office series: for a mature manager, net DPI by vintage is the single most honest number in the deck.
But the metric of the cycle has a blind spot, and sophisticated allocators know it. DPI is silent precisely where the most attractive entry points sit. A fund raised in the past five years has, by definition, almost nothing to distribute yet — venture funds are roughly ten-year vehicles, and meaningful liquidity rarely arrives before year seven. Screen on DPI alone and you have not de-risked your program; you have confined it to yesterday’s vintages and yesterday’s managers. The discipline worth building is different: knowing which observable behaviors, in a fund too young to show DPI, actually predict it.
Early DPI predicts less than you think
There is a second, less comfortable reason not to over-weight early distributions: they are a weak predictor even when they exist. One analysis of 71 funds from the 2005–2008 vintages found that the correlation between a fund’s year-five DPI and its final performance was just 0.22 — while year-five TVPI, the blended measure every LP has learned to distrust, correlated at 0.4. Early cash can even be manufactured. Selling a winner in year four produces a handsome interim DPI and, often, an amputated final multiple; a manager who trims its best position to decorate a fundraising deck is optimizing your dashboard, not your outcome. The honest reading: early DPI is a signal worth having, weakly. It should be augmented with other kpis.
Persistence belongs to judgment — so underwrite the judgment
What does predict? The most robust finding in the academic literature remains Kaplan and Schoar’s: in venture, performance persists across a firm’s successive funds, with correlations approaching 0.7 — far higher than in buyout. Persistence is a property of a venture team’s repeatable judgment, not of its brand or its fund number. Which reframes the underwriting question for a young fund into something answerable: is there evidence that the judgment is real, and is the fund constructed so that good judgment converts into returns? That evidence hides in six places, and none of them requires a distribution.
- Entry price. What the manager actually paid, deal by deal, against stage medians. Discipline at entry is the one return driver fully within a GP’s control, and it is visible from the first cheque onward.
- Ownership per euro. Two managers can deploy identical capital into comparable companies and own materially different stakes. Ownership is the raw material of a fund-returner; how efficiently it is bought is measurable from day one.
- The provenance of markups. Not the TVPI headline — its composition. What share of the unrealized book is priced by new external leads in arm’s-length rounds, at what step-ups, versus insider extensions and internal marks? An up-round led by a demanding outside investor is evidence. A SAFE extension priced by the existing syndicate is not.
- The loss ledger. Losses are intrinsic to early-stage investing; slow loss recognition is not. A manager who writes down failures promptly is demonstrating calibration — and protecting you from discovering the truth all at once in year eight.
- Reserve behavior. Where follow-on capital actually went. Doubling into positions with external validation is conviction; bridging the middle of the book out of sentiment is its opposite. The follow-on ledger is the most candid document a young fund possesses.
- Sourcing repeatability. Where the last twenty investments came from, and whether that origin is a system — a thesis, a geography, a network that regenerates — or a sequence of fortunate collisions.
Alignment: the tell that costs nothing to check
One more signal sits outside the portfolio entirely. The median GP commitment in venture hovers around 3% of fund size; what matters is less the percentage than its weight relative to the partners’ own balance sheets, and whether the economics of the firm force the GP to win only when you do. Fees at market, transaction fees offset, a whole-fund waterfall with a clawback — none of this requires a track record to verify, and a young manager who is disciplined here is telling you how it will behave when it matters.
The five questions
If you want the framework in operational form, these five questions do most of the work:
- Walk me through your entry valuations, deal by deal, against stage medians.
- Who priced your last five markups — new external leads, or you and your syndicate?
- Show me your write-downs and when you took them.
- Where has your follow-on capital gone, and what evidence triggered each cheque?
- Where did your last twenty deals come from — and show me two you passed on, and why.
Where this leaves the allocator
Europe’s largest fund-of-funds investor, the European Investment Fund, has been explicit that emerging managers can demonstrate performance well before liquidity arrives — through portfolio progression, entry discipline and the quality of subsequent financing rounds. That is the right standard, and it is the one we hold ourselves to. A GP who cannot yet show DPI can still show you everything that produces eventual DPI. Ask for that. The good ones will have it ready — and will enjoy the conversation more than the ones who lead with a blended IRR.
This is No. 1 in a five-part series for allocators. Next: portfolio construction for early-stage fintech — ownership, reserves, and follow-on logic.
Sources
PitchBook/NVCA data on 2023 distributions as a share of fund value (fourteen-year low), widely reported across LP coverage.
The VC Factory, “VC Funds DPI: How Long Until Venture Capital Delivers Outlier Returns?” — analysis of 71 funds, 2005–2008 vintages (thevcfactory.com/vc-funds-dpi).
Kaplan, S. and Schoar, A., “Private Equity Performance: Returns, Persistence, and Capital Flows,” Journal of Finance (2005).
DiligenceVault, “The Emerging Manager Fundraising Playbook 2026” — GP commitment norms (diligencevault.com).
EU.VC interview with David Dana, Head of VC Investments, European Investment Fund, on how emerging managers can show performance (eu.vc).
MiddleGame Ventures, “How to Choose a Venture Manager” (middlegamevc.com/updates) — companion piece in our Foundation Series.