A decade ago, selling a stake in a closed-end fund read as a distress signal: an LP exiting at a discount because it suddenly needed cash. Today it is a routine portfolio-management tool and the largest source of liquidity in private markets, where the time to a real exit has stretched well beyond a fund's nominal ten-year life. Full-year 2025 secondary volume is estimated at $226–240 billion depending on whose methodology you use — and both figures are records.
Two different markets need separating at the outset. One is the secondary market in private company shares, where you buy an interest in SpaceX or OpenAI from an employee or an early angel; that market has its own guide. The other is the secondary market in fund interests: what changes hands is not a portfolio company's share but an LP position in a private fund, with everything attached to it. This article is about the second market.
It splits into two transaction types, and the distinction matters. An LP-led deal is initiated by the investor: it sells its interest and exits. A GP-led deal is initiated by the manager: it moves assets from an old fund into a new one and puts existing investors to an election. In the first case the GP is a gatekeeper; in the second it sits on both sides of the table. That is the source of the conflict of interest around which the last few years of market practice have been built.
LP-led: how a fund interest is sold
What transfers in an LP-led deal is broader than it looks. The buyer takes on not only the funded interest but also the unfunded commitment — the obligation to answer future capital calls. Pricing is struck against NAV as of an agreed reference date (also called the cut-off date), and between that date and closing the price is adjusted for distributions received and for every capital call drawn in the interim.
You cannot simply sell. The LPA will almost always require the GP's written consent, often restrict transfers to quarterly dates, allow the GP to demand legal opinions on the buyer's status, and push the manager's costs onto the seller. Rights of first refusal held by the GP or the other LPs are checked separately. Two documents carry the deal: a purchase and sale agreement between seller and buyer, which lists the excluded obligations the buyer does not assume, and a tripartite transfer agreement to which the GP is a party.
The discount to NAV depends on strategy. Jefferies puts average 2025 pricing for LP portfolios at 87% of NAV: buyout at 92%, private credit at 91%, venture and growth at 78%, real estate at 70%. The spread reflects cash-flow predictability and the quality of the reported mark — where NAV is struck conservatively and corroborated by transactions, the discount is narrow.
GP-led and the continuation fund: why a manager buys from itself
The classic setup: a fund is running out of time, but one or two assets in the portfolio are clearly unfinished — they need another three to five years and capital for bolt-ons. Too early to sell to a strategic buyer, impossible to hold in the old fund. The answer is a continuation fund (or continuation vehicle, CV): a new structure managed by the same GP that buys the asset out of the old fund at an agreed price.
The capital comes from dedicated secondaries funds — Lexington, Ardian, StepStone, Coller — and from those LPs in the old fund who choose to stay. Each existing investor faces a mandatory election: cash out at the transaction price, or roll over into the new vehicle. The rollover itself comes in two flavours — on the CV's new terms, or through a status quo option under which the roller's management fee and carry do not increase and carried interest on its interest does not crystallise. ILPA insists that such an option be offered; in practice, in a 2024 sample only about 17% of deals offered a true status quo, and roughly 90% of LPs ultimately take the cash.
CV terms themselves have standardised by 2026. Management fees in the overwhelming majority sit at 50–100 basis points of invested capital, well below the classic 2%. Carry is nearly always tiered and tied to the specific asset's performance, frequently with dual IRR and MOIC hurdles. The base term runs around five years with two one-year extensions. More than 90% of deals carry unfunded commitments, and in roughly two-thirds of cases a rolling LP must fund its pro rata share.
The conflict of interest, and how the market clears it
In a GP-led deal the manager stands on both sides: seller, acting for the old fund, and buyer, acting for the CV. There are three pressure points. Price: mark it too low and the GP benefits new investors at the expense of old ones. Crystallisation: the deal converts the old fund's paper gain into real carried interest for the manager, giving the GP a direct cash incentive to get the transaction done. New terms: the CV carries its own fee layer, so a rolling LP risks paying twice for the same asset.
The industry's answer is procedural. The GP takes the deal to the LPAC for formal conflict clearance — ILPA's view is that this should happen even where the LPA contains an anticipatory waiver, and that the LPAC should have at least ten business days. Price is validated externally through a fairness opinion or an independent valuation, and better still through a competitive process with multiple bids. Accrued carry attributable to selling LPs should, on ILPA's recommendation, be rolled into the new fund rather than taken in cash.
The regulatory frame turned out shorter than expected. In August 2023 the SEC adopted the Private Fund Adviser Rules, one of which — the Adviser-Led Secondaries Rule (Rule 211(h)(2)-2) — expressly required a registered adviser to obtain a fairness opinion or valuation opinion for every GP-led transaction and to disclose its material business relationships with the provider of that opinion. On 5 June 2024 the Fifth Circuit, in National Association of Private Fund Managers v. SEC (No. 23-60471), vacated the entire package: the Advisers Act does not give the SEC that authority. The SEC's 2026 rulemaking agenda contains no re-proposal.
The result is paradoxical: the rule is gone, the practice remains. A fairness opinion today is a market standard rather than a legal requirement, and it rests on three supports. The Advisers Act antifraud provisions are still in force; the LPA and Delaware fiduciary duties bite directly; and ILPA has filled the regulatory vacuum — the 2023 guidance plus a refreshed draft out for public comment through 5 August 2026, which proposes an election period of no less than 30 business days, a competitive process backed by independent valuation, and no fee increases for rolling investors.
What happens when the procedure is treated as a formality was on display in ADIC v. EMG. In December 2025 the Abu Dhabi Investment Council went to the Delaware Court of Chancery to challenge the transfer of a 30% stake in Ascent Resources into a continuation fund: one week for the LPAC to review materials, a refusal to allow an in-camera session, and a mismatch between what the LPAC was shown and what prospective buyers saw. The parties agreed to pause the deal. The lesson reads plainly: the fight is not about whether a fairness opinion exists, but about the process that produced it.
The market in numbers
2025 was a record year on both counts. Jefferies counted $240 billion (up 48% on 2024), of which LP-led was $125 billion and GP-led $115 billion. Evercore reports $226 billion (up 41%): $120 billion LP-led, $106 billion GP-led. The gap is a question of counting methodology; the direction is identical.
The structural shift came in the first half of 2026: volume passed $120 billion and GP-led deals crossed half the market for the first time, at roughly 54%. Within GP-led, more than half the volume sits in single-asset CVs — the transfer of one, strongest asset. At the same time buyer capacity is tightening: secondaries dry powder fell about 10% over the half, and coverage of annual volume approached 1.0x. Negotiating leverage is moving toward buyers, and further narrowing of discounts should not be assumed.
What this means for a family office
If the office is already an LP and receives a GP-led proposal, the work starts with valuation. Run your own number on the asset independently — even roughly, on multiples — and compare it to the deal price rather than treating the fairness opinion as absolution: who wrote it, what relationships with the GP were disclosed, and whether there was a competitive process all matter. Then three questions: who pays the transaction costs (often the fund itself, which means the LPs); how the CV's new terms differ from the old ones and whether a status quo option is on the table; and what unfunded commitment a rollover brings with it. Separately, the election window: less than a few weeks alongside a thin data room is a signal, not a formality.
If the office comes in as a buyer, secondaries solve two problems. They shorten the J-curve: you acquire an already-formed portfolio of known assets, so the early-year trough is compressed and capital comes back sooner. And they give vintage diversification — access to fund years long closed to primary subscription. The price of that is information asymmetry: the GP and the seller know more about the asset than the buyer, and in a GP-led deal the seller also runs the business. Which is why the NAV discount is less a bargain than the price of risk. Access typically runs through specialist secondaries funds or feeder structures and turns on qualified investor status.
The tax side is deliberately sketched here rather than worked through, but two points belong in the file. A transfer of a fund interest is a transfer of a partnership interest: the buyer inherits carryover rather than market basis, and the allocation of profit across the transaction period is a documents question. For US funds, Section 1446(f) applies: a buyer purchasing from a non-US seller must withhold where the partnership has effectively connected income, and the withholding base includes assumed liabilities, so exemption certificates and three years of K-1s are requested well in advance. For tax-exempt investors there is a separate thread — UBTI, where the fund uses leverage.