Past venture capital performance can contain useful information about future results. Research has found persistence among VC managers, including when the analysis uses information available during fundraising. That makes a track record relevant to diligence. It does not make a previous top-quartile result a promise about the next fund.
The harder question is how much of the record an investor could have known at the time—and whether the people, strategy and investment opportunity remain comparable.
What the research measures
Two studies approach persistence from different directions:
| Research | Unit of analysis | Main finding relevant to investors |
|---|---|---|
| Harris, Jenkinson, Kaplan and Stucke, November 2020 working paper | Successive US buyout and VC funds | VC persistence remains when prior performance is measured at fundraising; buyout evidence is substantially weaker on that basis |
| Nanda, Samila and Sorenson, August 2017 working paper | Individual VC investments | Early successful exits predict subsequent exit success; the authors find patterns consistent with improved access to deals |
The Harris and co-authors paper uses institutional-investor cash flows net of fees and carry. It distinguishes performance known at fundraising from later outcomes.
The evidence weakens after 2000. In that subsample, the comparison across all four prior public-market-equivalent (PME) quartiles is not statistically significant; the top-versus-bottom comparison reaches the 10% level. Quartile-based selection therefore needs particular caution.
The Nanda, Samila and Sorenson paper studies investment-level outcomes. An IPO or acquisition is an exit indicator; it is not itself a measure of an LP's net return. Those two studies therefore provide complementary evidence rather than interchangeable statistics.
Fundraising creates an information problem
Imagine a manager raising Fund II while Fund I still holds most of its portfolio. Several years later, Fund I achieves an exceptional exit and earns a top-quartile ranking.
An analyst looking backwards might classify Fund II as the successor to a top-quartile fund. The LP deciding whether to invest in Fund II did not yet know about that exit. The relevant evidence was Fund I's record on the decision date.
That is look-ahead bias: using later information to make an earlier decision appear easier than it was. In manager selection, even an entirely accurate final performance table can mislead if the wrong date is attached to the investment decision.
Request dated historical reports where available. A current fundraising deck may show what happened eventually without showing what was known when investors committed.
A ranking needs a comparison group
“Top quartile” means the top quarter of a defined group, using a particular measure and reporting date. It is incomplete without those details.
A VC fund may rank differently on net IRR, total value to paid-in capital and distributions to paid-in capital. An older fund with realised exits is also difficult to compare with a younger fund whose return consists largely of estimated holdings.
The chance baseline matters when interpreting transition tables. In an evenly divided ranking, 25% of funds fall in the top quartile and 50% are above the median. Those are different tests. A manager exceeding one baseline does not automatically exceed the other by the same amount.
Check the data population too. A study of successive funds necessarily has a different scope from one that includes managers who never raise another vehicle. Sample selection affects how readily the result can be applied to the next emerging manager in an LP's pipeline.
Persistence can have more than one explanation
Investment judgement is one possible source of repeat success. Other candidates include founder relationships, access to allocations, sector expertise, reputation and the ability to recruit or support portfolio teams.
In the Nanda study, early success was associated with later investment success, and the pattern was consistent with better access to opportunities. The finding does not isolate an immutable personal talent for picking companies. It also does not establish that every investor can buy exposure to the advantage at any price.
For an LP, the practical question is whether the mechanism behind an attractive record remains present. A departing partner, much larger fund or move into a new stage may alter it.
Turn the record into diligence questions
Start with attribution. Who sourced the investments, made the decisions and remained involved through the outcome? Establish which parts of the record belong to the team now raising money.
Next, examine repeatability. If an early-stage manager raises a fund five times larger, can it preserve its entry prices, ownership targets and investment pace? If the strategy changes, assess the new opportunity on its own merits instead of assuming the earlier ranking carries across.
Finally, connect reported returns to investor experience. Review DPI, TVPI and remaining value, the treatment of fees, and the benchmark used. A high multiple dominated by one unsold holding warrants different questions from a similar multiple supported by several cash distributions.
Past performance is evidence to investigate. The investment decision rests on how that record was produced, what remains repeatable and the terms on which the next investor can participate.
FAQ
Does past VC performance guarantee a good next fund?
No. Treat the record as evidence about a process, then assess whether the team, strategy, access and terms remain comparable. A ranking does not show which elements produced the result or whether a new LP can obtain the same exposure.
What does the research imply for first-time VC funds?
The Harris study's first-time funds performed respectably relative to successor funds in its sample; the comparison does not establish significant first-time outperformance. It challenges automatic exclusion, rather than removing the need for diligence on a new manager.
Why use performance known at fundraising?
That is the information an LP could actually assess before committing. Using a successful exit that occurred years later makes the decision appear easier in hindsight. Ask for dated reports and distinguish subsequent outcomes from the evidence available at the time.
Is an investment-level exit record the same as LP performance?
No. Counting acquisitions or IPOs does not establish the proceeds earned on the securities held. Net fund performance also incorporates contributions, remaining value, costs, carry and timing. Use research findings at the level they measure rather than converting an exit-success statistic into an LP return.
