Distributions from private equity funds have been stuck near 14 to 15 percent of net asset value (NAV) since the exit market seized up in late 2022, leaving limited partners (LPs) holding large unrealized positions against marks that many privately question.[1] Relatedly, regulators have recently flagged aggressive marking practices as a concern, and lenders extending NAV-based credit facilities have pushed for stronger independent valuation rights as a condition of financing.[2] The common thread comes down to a simple question: how much should anyone trust a private equity (PE) fund’s reported NAV?
Academic research has studied this question in various ways over at least the last decade. On average, interim NAVs tend to be conservative rather than inflated.[3] However, managers have shown a documented tendency to manage the timing of markdowns around fundraising, delaying bad news until a new fund has closed.[4] These findings can coexist, as an average tendency across the industry does not preclude individual managers manipulating markdown timing for their own benefit, nor does it tell an LP much about any single fund in front of them. Indeed, research has found that this kind of behavior is concentrated almost entirely among underperforming and lower-reputation managers, while top-performing managers tend towards more conservative valuation reporting.[5] A recent working paper asks perhaps a more useful question regarding a PE fund: rather than just how accurate is its NAV, does the pattern of how a manager arrives at a NAV predict what happens next?
Beyond the point-in-time mark
Ege Ercan, Steven Kaplan, and Ilya Strebulaev take up that question directly in a National Bureau of Economic Research (NBER) working paper released in 2025.[6] Using a novel investment-level dataset from StepStone Group, the study tracks 8,331 U.S. buyout investments across 894 funds, and 6,822 U.S. venture capital (VC) investments across 563 funds, each followed from initial investment through exit.[7]
In the dataset, managers report each investment’s capital invested, distributions, and estimated NAVs on a quarterly basis. The paper defines two behavioral markers embedded in a manager’s reporting history. “Staleness” is a quarter in which an investment’s valuation is not updated at all. “Markdown frequency” is the share of quarters in which the valuation is revised downward. Both are pervasive, but to varying degrees depending on asset class. For each investment, the paper tracks the share of that investment’s quarterly reports that were stale or marked down, from its start through to a given point, then averages that share across all investments observed at that point. On average over an investment’s first five years, buyout staleness runs from about 37 percent in year 1 down to roughly 29 percent by year 3, before drifting back up slightly. VC staleness is far more persistent, remaining between 71 and 79 percent throughout an investment’s life. In other words, the typical VC investor sees a portfolio company’s valuation move in only about one quarter out of four (Figure 1).[8]
Figure 1: VC valuations remain unchanged (“stale”) more than twice as often as buyout valuations throughout the first five years of an investment’s life
What the pattern predicts
The paper’s central finding is that the reporting history of NAV, not just the latest reported snapshots, helps predict how an investment ultimately performs, and how long it takes to get there. Holding interim realized and unrealized returns constant, investments with a longer history of staleness or more frequent markdowns go on to deliver worse returns. The paper measures future performance as the smooth quarterly rate that would carry an investment from its current interim value to its actual eventual exit value, spread over the quarters actually remaining until exit. This measure is anchored to each investment’s real outcome rather than its most recent paper mark alone. For buyout investments, a portfolio company that has been completely stale through its second year is expected to earn a quarterly return roughly 13 percentage points lower than an otherwise identical investment with no staleness at all (Figure 2).[9]
A portfolio company marked down in every quarter through its second year is expected to earn a quarterly return about 14 percentage points lower than one that was never marked down.[10] The same stale investment is also expected to exit around 3.4 quarters later than average. In the sample, a typical buyout investment still unexited at the end of its second year has, on average, about 13 quarters left before exit, so an additional 3.4 quarters amounts to roughly 25 percent longer than that typical remaining holding period (Figure 2).[11]
Figure 2: Staleness and markdowns predict both weaker returns and slower exits for buyout investments[12]
Staleness and markdowns aside, interim realized and unrealized returns each carry a weakly positive relationship with future performance. Unrealized value becomes a meaningfully stronger predictor later in an investment’s life, while realized value (cash already returned) shows a modest positive signal throughout.[13] In other words, persistent absence of NAV updates or a run of markdowns can be perceived as warning signs to an LP, regardless of how healthy the current multiple looks.
Reading between the marks
For LPs, the practical implication is that a manager’s reporting behavior is itself a data point worth tracking, distinct from the headline NAV reported. Two funds reporting the same current multiple are not, based on this evidence, in equivalent positions: the fund with a longer run of stale or marked-down quarters behind it carries a meaningfully worse forward-looking profile, all else equal.[14] This argues for building staleness and markdown frequency into ongoing portfolio monitoring, rather than considering each quarter’s NAV update in isolation. This is the kind of longitudinal, cross-manager tracking that depends on normalized, comparable reporting, a capability that underlies our own work in private market simulation and benchmarking.
The same signal is arguably more important in the secondaries and continuation vehicle market, where reported NAVs already face heightened scrutiny amid slower distributions and rising deal volume.[15] A buyer evaluating a GP-led continuation vehicle, or an LP weighing a stake sale, is betting on whether today’s reported valuation reflects the true state of the underlying company, functionally the same bet this paper studies. A reporting history marked by staleness or frequent markdowns is a reason for a secondaries buyer to price more conservatively or for a seller to expect a wider discount, a topic we considered in more depth in a recent post.
Conclusion
The authors are careful not to claim they have identified why this predictability exists. Strategic delays of NAV updates by managers, differing degrees of valuation conservatism, unobserved differences in investment quality, and simple return momentum could all generate the same statistical pattern.[16] Still, at a moment when LPs are questioning whether reported returns match cash actually returned, the paper’s core message is a useful reminder: a single interim NAV is only part of the story. The historical valuation pattern (e.g., how often it moved, and in which direction) carries real information about where it is likely to end up.
Key takeaways
- Interim NAVs are conservative on average, but aggressive timing does occur, and it is concentrated almost entirely among underperforming and lower-reputation managers.
- The history of a manager’s reporting, not just the latest NAV, is informative: greater staleness or more frequent markdowns predict measurably worse future returns and slower exits.
- For LPs, this argues for tracking staleness and markdown frequency signals, separate from multiples themselves, especially when evaluating GP-led continuation vehicles or pricing a secondaries stake.
[1] Praxis Rock, “What LPs Want in 2026: DPI and Transparency,” March 2026, https://praxisrock.com/insights/what-lps-want-2026-dpi-transparency, accessed July 28, 2026.
[2] Sage Advisory, “Private Credit Markets Under Pressure: Key Risks and Investor Strategies for 2026,” January 2026, https://www.sageadvisory.com/article/private-credit-markets-under-pressure-what-investors-should-heed-going-into-2026; Ropes & Gray LLP, “NAV Facilities in 2026: Structuring, Governance and Market Practice Considerations for Sponsors,” April 2026, https://www.ropesgray.com/en/insights/viewpoints/102mrf1/nav-facilities-in-2026-structuring-governance-and-market-practice-consideration, both accessed July 28, 2026.
[3] Tim Jenkinson, Miguel Sousa, and Rudiger Stucke, “How Fair are the Valuations of Private Equity Funds?,” Oxford University working paper, February 27, 2013, https://ssrn.com/abstract=2229547.
[4] Indraneel Chakraborty and Michael Ewens, “Managing Performance Signals Through Delay: Evidence from Venture Capital,” Management Science 64, no. 6 (2018): 2875–2900; Brad M. Barber and Ayako Yasuda, “Interim Fund Performance and Fundraising in Private Equity,” Journal of Financial Economics 124, no. 1 (2017): 172–194, https://www.sciencedirect.com/science/article/pii/S0304405X17300016.
[5] Gregory W. Brown, Oleg R. Gredil, and Steven N. Kaplan, “Do Private Equity Funds Manipulate Reported Returns?,” Journal of Financial Economics 132, no. 2 (2019): 267–297, https://doi.org/10.1016/j.jfineco.2018.10.011; Barber and Yasuda, “Interim Fund Performance and Fundraising in Private Equity.”
[6] Ege Y. Ercan, Steven N. Kaplan, and Ilya A. Strebulaev, “Interim Valuations, Predictability, and Outcomes in Private Equity,” NBER Working Paper No. 33637, April 2025 (revised May 2026), https://www.nber.org/papers/w33637. A version of this paper was first circulated as a working paper in May 2024; the current draft substantially expands the underlying dataset and adds exit timing as a second predicted outcome.
[7] Ibid.
[8] Ibid.
[9] Ibid.
[10] Ibid.
[11] Ibid.
[12] Each bar shows the difference in predicted outcome between two otherwise identical investments: one with a staleness (or markdown) frequency of 100 percent up to that point, and one with a frequency of 0 percent, holding interim realized and unrealized returns fixed. Panel (a) shows the difference in expected quarterly return, in percentage points; panel (b) shows the difference in expected number of quarters until exit. Data drawn from Ercan, Kaplan, and Strebulaev, “Interim Valuations, Predictability, and Outcomes,” Tables 4 and 5 (buyout). All estimates shown are statistically significant at conventional levels (p < 0.10 or better) except the Year 1 markdown effect on exit timing, which is not statistically distinguishable from zero.
[13] Ibid.
[14] Ibid.
[15] Maureen Farrell, “Investors Warn of ‘Rot in Private Equity’ as Funds Strike Circular Deals,” New York Times, December 24, 2025, https://www.nytimes.com/2025/12/24/business/private-equity-continuation-funds.html, last accessed August 3, 2026.
[16] Ercan, Kaplan, and Strebulaev, “Interim Valuations, Predictability, and Outcomes.”
