The debate over whether the asset class carries a premium concerns roughly two percentage points. The gap between an upper-quartile and a lower-quartile manager is closer to thirteen. No allocator owns the average, which makes dispersion, not access, the variable that decides the result.
The question of whether investors are compensated for illiquidity has absorbed more attention than it merits, because the answer does not change what an allocator should do next.
Consider the strongest version of each position. On one side, Cambridge Associates data show the US Private Equity Index ahead of the Russell 3000 Index public market equivalent by 219 basis points over ten-year horizons.¹ On the other, AQR's research on expected returns for illiquid assets puts the point estimate for a distinct illiquidity premium near zero and attributes much of the apparent premium to leverage, size, and sector composition.²
The distance between those two positions is roughly two percentage points of annual return. The distance between an upper-quartile and a lower-quartile private equity manager is 12.9 percentage points.³
That comparison is the argument. Whether the asset class carries a premium is a question about an average. No allocator owns the average. An allocator owns a small number of specific managers, and the spread between a good and a poor outcome within that selection is roughly six times the size of the premium under debate.
What does the asset class premium actually decide?
In public equities, the allocation decision dominates the selection decision. That is the empirical basis for indexing, and the numbers are not close. A study sourced through eVestment found the spread between upper- and lower-quartile public equity funds at approximately 1.5 percentage points.³ S&P and Burgiss data put the interquartile range for US small-cap growth funds at 1.2 percentage points annually over a twenty-year period.⁴ KKR's 2026 analysis explains the mechanism plainly: active public equity managers tend to hold similar portfolios with different company weightings, which produces dispersion of approximately 300 basis points.⁵
Choosing to own public equities is therefore worth far more than choosing who manages it. An investor who gets the allocation right and the manager wrong still captures most of the return.
Private markets invert that relationship. The same S&P and Burgiss data place the interquartile range for multi-stage venture capital at 18.2 percentage points.⁴ Private Equity International reports the spread between the strongest and weakest buyout managers in the 2022 vintage exceeding 13 percentage points, the widest since 2014, with upper-quartile managers delivering approximately 22 percent internal rate of return over the past decade against a median near 13 percent.⁶
The allocation decision in private markets is worth low single digits. The selection decision is worth low double digits. An investor who gets the allocation right and the manager wrong has captured very little, and may have captured less than a public index would have delivered at a fraction of the cost.
This gap is also widening rather than closing. McKinsey has characterized the current recovery as K-shaped, with stronger managers benefiting disproportionately while weaker ones struggle with exits and distributions.⁷ When the market-level return compresses, the selection decision does not become less important. It becomes a larger share of whatever return remains.
Why is private markets dispersion so much wider?
The honest answer is that a meaningful portion of it is mechanical rather than a measure of skill, and any argument that skips this step is incomplete.
A typical private equity fund holds roughly a dozen to twenty portfolio companies, all of them levered, concentrated in businesses far smaller than those in a public index. A public equity manager holds on the order of two hundred positions in a universe of larger, mostly unlevered companies. Concentration and leverage widen the distribution of outcomes on their own, before any question of manager capability arises.
Verdad's analysis pressed that point to its conclusion, finding that observed private equity dispersion is approximately what randomly constructed portfolios of roughly twenty levered micro-cap companies would produce, and that private equity managers trailed a simulated factor-matched benchmark by two to four percentage points annually, a gap consistent with the fee load.⁸
The implication is specific and it is uncomfortable for the standard industry argument. Wide dispersion proves that the selection decision is consequential. It does not prove that selection is rewarded. The most common claim in private markets marketing — dispersion is wide, therefore choose carefully — treats those two statements as the same one. They are not. Width is a necessary condition for selection to matter, and it is not a sufficient one.
Is the dispersion selectable?
This is the question that actually determines whether a private markets program is worth running, and the evidence is mixed in a useful way.
Harris, Jenkinson, Kaplan and Stucke drew the distinction that matters. Measured on final realized performance, persistence appears in the data. Measured on the information an investor would have possessed at the time of the next fundraise — the only information a commitment decision can actually use — they find little or no evidence of persistence for post-2000 buyout funds, and they find that skipping a vintage does not improve the odds.⁹ An investor gains little by knowing a current fund's relative standing when deciding whether to commit to the next one.
One asymmetry in that literature is more actionable than the headline. Where persistence does appear reliably, it appears at the lower end of the distribution. Weak managers repeat more dependably than strong ones.
That reframes the objective. In a thirteen-point distribution, moving a program from the lower quartile to the median captures more return, more reliably, than reaching for the upper quartile. Avoidance is the selectable edge. Prediction is the harder claim, and most programs are structured as though the reverse were true.
What remains selectable is the mechanism rather than the label. Return attribution on the prior fund, separating multiple expansion from operating change. Continuity of the individuals who produced the record at the deal level. Whether the strategy being raised is the strategy that generated the track record, which a materially larger fund is not. And what the manager declined, since rejection discipline is where competitive pressure shows up before it appears in results.
How do perpetual structures change the dispersion an allocator owns?
The shift toward perpetual vehicles is usually discussed as a liquidity decision. Measured against dispersion, it is something else entirely.
A 2026 market overview published by Hamilton Lane, one of the largest private markets managers, reports dispersion between upper- and lower-quartile evergreen funds of approximately 600 basis points over a trailing one-year period and approximately 300 basis points over three years, materially narrower than the equivalent closed-end range. The same analysis notes that roughly one third of closed-end buyout funds carried a loss over the trailing year, and that evergreen equity funds trailed public benchmarks over the period.¹⁰ MSCI and Morningstar have each observed that meaningful dispersion nonetheless persists within the evergreen universe.¹¹
Set those figures beside the closed-end numbers. A perpetual structure converts a selection decision worth twelve to thirteen percentage points into one worth three to six. At the three-year measure, evergreen dispersion is approximately the width of active public equity manager dispersion.
That is the trade, and it is a return decision rather than an operational one. Perpetual vehicles hold more positions, deploy continuously across entry points, and blend primaries with secondaries and co-investments. Each of those choices narrows the distribution at both ends, which is why the structure produces fewer losses and fewer exceptional results.
Whether the trade is favorable depends entirely on one thing. An allocator without a defensible selection process is better served by the narrower distribution, because that allocator was carrying dispersion risk without being compensated for it. An allocator with a selection process is giving up the reason for the allocation, and paying private markets pricing for public-equity-width outcomes.
Neither choice is wrong. What does not survive examination is selecting the structure on its liquidity terms and discovering the dispersion consequence afterward.
What should an allocator underwrite?
If dispersion rather than access determines the outcome, the diligence agenda changes. Four questions apply to any private markets position.
- What is the dispersion width of the specific opportunity set being accessed, and where in that distribution must the program land for the allocation to clear its public alternative net of all fees?
- What does the selection process produce that the market does not already price, and what evidence supports it beyond the track records of the managers already selected?
- How many managers can the program underwrite to that standard, and does the current position count exceed it? Each position beyond that number moves the portfolio toward the median.
- Where a structure narrows dispersion, what fee is being paid relative to what the narrowed distribution can deliver?
A program that cannot answer the second question should answer the fourth one differently. There is no dishonor in owning the median deliberately. The failure is owning it while describing the program as manager selection.
So where does the return actually come from?
BIP Capital has published on the premium question before, arguing that illiquidity is the mechanism generating the private markets return advantage and citing the Cambridge Associates comparison referenced above. That position and this one are compatible, and the reconciliation is worth stating directly.
Both accept that private markets have delivered returns above public benchmarks. They differ on mechanism: whether the lock-up produced the advantage, or whether leverage, size, sector composition, and manager selection produced it while the lock-up accompanied it. For an allocator, the disagreement resolves, because both readings arrive in the same place. Under the first, the premium is roughly two hundred basis points and available to anyone willing to accept the lock-up. Under the second, it is smaller. In either case it is a fraction of the twelve to thirteen points separating a strong manager from a weak one in the same vintage, pursuing the same strategy.
The asset class premium determines whether private markets belong in a portfolio at all. Manager dispersion determines what a particular allocation is worth. Those are different questions, and only the second one has an answer specific to the firm asking it.
For informational purposes only. Not investment advice, and not an offer to sell or a solicitation of an offer to buy any security.