Data sourced from Dakota Private Markets, the private fund performance platform powered by Dakota. Learn More | Request Access
Every private fund performance database has a population problem before it has a data problem. A database can only show you the funds it holds, and the funds it holds are not a random sample of the funds that existed.
The funds that stopped reporting, were folded into a successor vehicle, or quietly wound down are often the ones missing. That gap is survivorship bias, and it sits underneath every benchmark built from the data.
In this article, we'll cover how survivorship bias gets into a private fund database, why it is harder to see than in public markets, and the questions that tell you whether a database has dealt with it.
6 Bias in Private Fund Performance Databases
1. Private Funds Leave a Dataset Without Announcing It
A public fund that closes leaves a paper trail. A private fund that stops submitting numbers usually leaves nothing. There is no delisting event, no ticker that goes dark, and no obligation to report once the fund is past its useful life to the manager.
Most private fund datasets are assembled from what managers, LPs, and public disclosures make available. When a fund underperforms and the manager stops sharing figures, the record often stops updating or never gets added in the first place.
Dakota Private Markets solution: Dakota Private Markets gathers fund performance from GP reporting, LP disclosures, and public filings, so a fund's place in the database does not hinge on its manager choosing to keep submitting.
2. Reporting Is a Choice, and the Choice Is Not Neutral
Voluntary reporting adds a second filter. Managers are more likely to submit a fund when the numbers are worth showing, and less likely when they are not.
That produces two distortions that compound. Strong funds are over-represented because their managers want them seen. Weak funds are under-represented because their managers have no reason to volunteer them. A database built mainly on submissions will lean toward the first group without anyone having decided it should.
Dakota Private Markets solution: Every record in Dakota Private Markets is reviewed by Dakota's research team before publication, rather than auto-populated or taken at face value from a GP's marketing deck.
3. The Bias Lands on the Numbers You Rely On Most
The effect is not subtle where it matters. When weaker funds drop out, the median and quartile breakpoints move up, so the benchmark describes the funds that did well enough to stay in the sample rather than the full peer group.
Consider a hypothetical vintage with 100 funds. If the 20 weakest stop reporting, the remaining 80 produce a higher median than the original 100 would have. The fund you are evaluating is now measured against a bar that never described the real peer group.
The same problem hides in a manager's track record. A GP's "Fund I through Fund IV" history looks clean when a weaker earlier vehicle is missing, and an allocator reading it has no way to know what was left out.
Dakota Private Markets solution: Request access to Dakota Private Markets to see who is in the sample before you trust the median, across 18,000+ named funds and 159,000+ performance records, each reviewed by Dakota's research team.
4. Smaller and Emerging Managers Are Hit Hardest
Larger, established managers tend to have the reporting infrastructure and the incentive to keep data flowing. Smaller and first-time managers are more likely to drop out of datasets, which means the benchmarks for emerging manager comparisons can be the least reliable of all. Evergreen and semi-liquid vehicles carry their own version of the problem, because they never close and so never produce a clean final return.
Dakota Private Markets solution: Dakota Private Markets gives every fund and sponsor its own profile with full history and updates records daily, so smaller managers are not left to depend on their own reporting capacity to stay visible.
5. Five Questions That Expose Survivorship Bias
You cannot remove survivorship bias by reading a single number, but you can test for it. Ask any database provider:
Does the database keep funds that stopped reporting or were liquidated? If dead and dormant funds are dropped, the history is survivors only.
How does a fund enter the database? Self-submission, public disclosure, and independent research each carry a different selection risk.
Can you see a manager's full fund series? If Fund II is missing from a GP with Funds I and III, ask why.
Is every record dated? A stale record is a clue that a fund may have stopped reporting.
Is the peer group size shown? A median drawn from a handful of funds in a narrow cohort is the most exposed to missing data.
Dakota Private Markets solution: Dakota Private Markets is built on named, research-verified funds, so these are questions you can put to the data yourself instead of taking a provider's word for it.
6. Treat Every Benchmark as a Ceiling, Not a Midpoint
The practical answer is not to distrust all benchmarks. It is to read them with the bias in mind: when a database cannot show you who dropped out, treat its median as the upper end of a reasonable range for the peer group.
Dakota Private Markets solution: Custom benchmarking in Dakota Private Markets lets you build the peer group yourself by vintage, strategy, geography, and fund size, so you are not stuck reading a published median that describes a broader group than the fund in front of you.
About Dakota Private Markets
Dakota Private Markets, powered by Dakota, is a private markets intelligence and performance platform for institutional investors and investment professionals. It covers 18,000+ named funds and 159,000+ performance records across seven asset classes, with Net IRR, TVPI, DPI, and RVPI standardized for comparison, and every record reviewed by Dakota's research team.
Request access today to build your own peer group from named, research-verified funds in Dakota Private Markets and trace a manager's full track record from the first vehicle to the current raise.
