The Second Exit

Continuation vehicles have gone from an end-of-fund clean-up tool to one of the main exit channels in private markets. The commercial story is familiar by now. A sponsor moves one or more portfolio companies from an older fund into a new vehicle, gives existing limited partners a choice between cash and continued exposure, and brings in fresh capital to fund the purchase and the next leg of growth.

The tax story is less familiar. A continuation transaction can look economically like a sale, legally like a restructuring, and, for selected investors, like a tax-deferred rollover — and all three descriptions can be true at once. Partnership divisions, disguised sales, debt shifts, basis adjustments, carried interest, withholding, blockers, transfer taxes, and each investor’s own tax status can matter as much as the headline price.¹

That is the focus of this alert. The thesis is broader than tax: a continuation fund is a miniature capital market formed around one sponsor-controlled asset or portfolio. It discovers a price, supplies liquidity, transfers risk, extends duration, resets governance and carry, and reallocates tax attributes. Tax is not a footnote to that product design. It is one of the prices being negotiated.

Executive summary

• The market is now structural. Lazard counts $233 billion of global secondary volume in 2025 ($116 billion GP-led) and a record $124 billion in the first half of 2026, with about $61 billion of GP-led deals and a full-year forecast near $275 billion. Jefferies puts 2025 GP-led volume at $115 billion, of which continuation vehicles were 89 percent. Methodologies differ; direction does not.
• Single-asset vehicles are the flagship format. They ran 54 percent of GP-led deployment in the first half of 2026, and 87 percent of that volume priced above 90 percent of NAV. The market has also developed standards: roughly 27 percent of launched deals failed or paused, and software CVs now get a colder look.
• For rolling LPs, the intended federal result is deferral. Not forgiveness. The usual route is the partnership-division rules — a deemed §721 contribution of assets to the CV and a §731 distribution of CV interests — with the built-in gain preserved and tracked under §704(c).
• For cashing-out LPs, the transaction is meant to be a taxable realization. Character, timing, and withholding change with the asset: stock, partnership interests, credit paper, real estate, and blocker shares do not produce the same answer.
• Incoming investors care about basis. A high purchase price without a matching inside-basis adjustment is tax drag. Section 754 elections and §743(b) adjustments are economic terms, and should be negotiated like them.
• The sponsor has two carry questions. What happens to accrued carry in the legacy fund at the CV price, and how is the new CV promote struck? Section 1061, the profits-interest revenue procedures, and the decision to roll crystallized carry sit at both stages.
• Asset class changes the tax engine. Venture CVs can destroy §1202/QSBS benefits — a stake that grew after the 2025 OBBBA rewrite. Credit CVs bring OID, market discount, and ordinary income. Real-estate and infrastructure CVs bring FIRPTA, recapture, transfer taxes, and liability-allocation work.
• Returns need a two-stage lens, and now a third. The legacy fund records an exit at the CV price; the CV starts a fresh underwriting period at the same price; rolling LPs deserve a continuous look-through return from original purchase to final exit. Any single IRR hides the economics.
• The early performance record is mixed in a useful way. HEC Paris and Evercore find CV returns broadly comparable to buyout benchmarks, with a narrower spread for single-asset vehicles. Comparable is not superior — especially net of a second promote.

01 What Exactly Is a Continuation Fund?

A continuation fund — usually called a continuation vehicle, or CV — is a new investment vehicle formed to acquire one or more assets from an existing fund managed by the same sponsor. Existing LPs get a liquidity election: sell for cash, roll some or all of the position into the CV, or (in some deals) choose among rollover classes. New investors supply the cash that funds the purchase and, often, follow-on capital.

Stop thinking of it as one transaction. It is four transactions wearing one term sheet:

• The legacy fund is monetizing an asset at an agreed value.
• Selling LPs are converting an illiquid fund position into cash.
• Rolling LPs are trying to keep the same exposure without paying tax merely because the wrapper changed.
• New investors are making a fresh investment at the CV price, with newly negotiated governance, fees, and carry.

The sponsor sits across all four seats at once. That is why the market has built fairness opinions, LPAC consent, competitive price discovery, and structured election mechanics around these deals: the GP sets a price at which it is both seller and buyer, and everyone at the table knows it. The process is usually robust. It is still worth reading.

Swipe sideways for the other columns.

02 From Zombie Funds to a Core Exit Channel

The first generation of continuation funds lived in the least glamorous corner of private equity: old funds, difficult assets, no obvious buyer. The modern market is nearly the reverse. Sponsors use CVs for companies they do not want to sell yet.

Lazard estimates total secondary volume of $152 billion in 2024 and $233 billion in 2025, with GP-led transactions at $116 billion in 2025. Jefferies counts a similar $240 billion total and $115 billion of GP-led volume, of which continuation vehicles were 89 percent; average CV size in 2025 was about $900 million, and twenty-nine GP-led deals cleared $1 billion.

The first half of 2026 kept the pace. Lazard puts first-half volume at $124 billion — a new first-half record, up roughly 28 percent year over year — including about $61 billion of GP-led activity, with trailing-twelve-month volume of $260 billion (double the 2021 market) and a full-year forecast near $275 billion. Continuation funds were 84 percent of GP-led deployment; single-asset vehicles alone were 54 percent. Jefferies estimates that nearly 80 percent of the top 100 sponsors by AUM had completed a CV by the end of 2025, and that GP-led secondaries were roughly 14 percent of sponsor-backed exit volume.

The CV is no longer a substitute for an exit. It is an exit channel — and, at 14 percent of sponsor-backed exits, a load-bearing one.

The secondary market, per Lazard ($B)

Swipe sideways for the other columns.

GP-led volume ran roughly 84–89% continuation vehicles in 2025–26, depending on whose survey you read.

Sources: Lazard 2025 Secondary Market Report (Feb. 2026); Lazard Interim 2026 Secondary Market Report (Aug. 2026). GP-led CV share per Lazard and Jefferies.

WHY THIS MATTERS FOR TAX

As CVs become repeatable products rather than bespoke rescues, the tax workstream has to become as institutionalized as the financing and fairness workstreams. A sponsor running five CVs cannot treat basis, withholding, blocker leakage, rollover mechanics, and audit history as one-off closing surprises. The market has industrialized price discovery and capital formation; tax should not still be artisanal. The better framing is that every CV reallocates five things at once: liquidity, duration, control, incentive economics, and tax basis. If one of those reallocations is mispriced, someone is subsidizing someone else.

03 Why We Expect the Market to Keep Growing

Several forces point the same way. Private equity still carries a large stock of mature unrealized assets, while the recovery in M&A and IPOs has been uneven — especially for sponsor exits. LPs need DPI; sponsors want more time with their best companies; and secondary buyers have raised enough dedicated capital to make the CV a repeatable financing market rather than a bespoke rescue. Jefferies estimated $327 billion of dedicated secondary capital in 2025, and Lazard reports that 40 percent of surveyed secondary investors now manage evergreen or ‘40 Act vehicles. That matters because semi-liquid capital can write larger, more frequent checks and can turn secondaries into a standing source of liquidity. Credit and infrastructure are following buyout down the same path. In other words, the CV is growing for the same reason any capital market grows: more issuers, more buyers, more standardized terms, and more ways to finance the trade.

Forecasts should be treated as forecasts. Jefferies has sketched a path toward roughly $300 billion of annual secondary volume, while Lazard’s current 2026 forecast is about $275 billion. We would not bank a point estimate. We would bank the direction — and, more important, the change in market structure. Once repeat sponsors, evergreen capital, credit specialists, infrastructure specialists, and CV-to-CV exits all participate, the question is no longer whether continuation funds survive the exit drought. It is what share of private-market liquidity they permanently own.

04 The Tax Map: Four Taxpayers, Four Objectives

Swipe sideways for the other columns.

A well-designed transaction lets these objectives coexist. A poorly designed one makes one constituency pay for another constituency’s tax result — quietly, and usually a few years after everyone has framed the closing dinner photos.

05 Alternative Paths, and the Authorities Behind Them

Most market commentary treats “the CV structure” as if there were exactly one. There is a dominant pattern; there is no single mandatory form. The choice of path decides who recognizes gain, which partnership continues, whether holding periods tack, whether QSBS survives, and who inherits the old fund’s audit history. Counsel should write the intended characterization into the step plan before the election memo goes out to LPs.

The authorities below are the working set. They do not all point the same way. That is the point.

The dominant hybrid: division plus a taxable cash-out

The workhorse is a partnership division under Treas. Reg. §1.708-1(d), overlaid with a cash-out for selling partners intended to be a sale under §741 or a disguised sale under §707(a)(2)(B) and Treas. Reg. §1.707-3.

In the common assets-over form — the default if no form is specified, per Treas. Reg. §1.708-1(d)(3)(i) — the legacy fund is deemed to contribute the transferred assets to the CV for CV interests, generally under §721(a), then distribute those interests to partners. Rolling partners keep them. Selling partners are treated as disposing of them for cash supplied by the new investors, often through a seller partnership or a redemption so the buyer can take certificates from a U.S. partnership rather than from a global LP roster.

The elegance is real. So is the fragility. Each deemed step has to qualify on its own facts. Section 721(a) does not apply if §721(b) treats the CV as an investment company. Section 731 does not protect a distribution of money — or of marketable securities treated as money under §731(c) — in excess of outside basis. Section 752 treats a drop in a partner’s share of liabilities as money. Sections 704(c)(1)(B) and 737 can pull pre-contribution gain forward if property moves again inside seven years.

“Tax-free rollover” is a conclusion. It is not a closing-checklist item.

Assets-over versus assets-up

Treas. Reg. §1.708-1(d)(3) offers two respected forms. Assets-over is simpler commercially; title to the portfolio company never has to pass through the partners. Mixing-bowl exposure under §§704(c)(1)(B) and 737 can attach to the deemed contribution and to later movement of the property, and the successor rules (Treas. Reg. §§1.704-4(d)(2), 1.737-2) can treat the CV as stepping into the contributing partner’s shoes.

Assets-up (Treas. Reg. §1.708-1(d)(3)(ii)) requires the divided partnership — which must be a continuing partnership — to distribute assets to partners so that they own them under local law, after which the partners contribute the assets to the recipient partnership. The form is heavier: real-estate transfer taxes, change-of-control covenants, lender consents, and securities-law restrictions can make it impractical. It can still be the better tax form when a particular partner must receive particular §704(c) property, or when an assets-over contribution would create an investment-company problem that a targeted assets-up allocation would avoid.

If the parties do not specify a form, the regulations impose assets-over. Silence is a choice.

Which partnership continues: the 50 percent test

Under §708(b)(2)(B) and Treas. Reg. §1.708-1(d)(1), a resulting partnership continues the prior partnership if its members held more than 50 percent of the capital and profits of the prior partnership. Otherwise it is a new partnership.

The test is not a technicality. A continuation can inherit the old partnership’s EIN, §754 election, accounting methods, and — more awkwardly — its pre-closing history under the BBA centralized audit regime. An imputed underpayment for a pre-CV year can land on a vehicle that incoming investors think of as new; a vehicle the investment committee calls “new” can still answer for the old fund’s audit like an heir who accepted the estate without reading the inventory. Documents can push that risk back to legacy sellers, mandate a §6226 push-out, or back it with RWI. They cannot wish the statute away.

In many single-asset CVs the rolling group, GP included, is well under 50 percent of the old fund — Jefferies reports average LP rollover near 15 percent — so the CV is usually a new partnership. Full-fund restructurings and high-rollover deals can flip the answer. Model the test. Do not assume it.

Interest sale without a division

A second family of deals never bothers with a division. New investors buy partnership interests — in the legacy fund, in a special-purpose seller partnership, or in the CV after a recapitalization — and selling LPs take cash under §741, subject to §751 look-through. Rolling LPs simply keep what they have. Rev. Rul. 84-52 remains the starting point for recapitalizations of a continuing partnership: the partner is treated as exchanging its old interest for a new one under §721, with basis and holding period carrying over.

This path can be cleaner when the asset should not move; it can be worse when new investors want a clean vehicle, fresh governance, and a new waterfall. A purchase of old-fund interests also drags more of the old fund’s liabilities, elections, and audit file into the buyer’s life, unless the purchased interests are isolated through a new upper-tier partnership with a §754 election.

Taxable sale and optional reinvestment

The blunt instrument: the legacy fund sells the asset for cash, distributes proceeds to all LPs, and those who wish to continue subscribe into the CV. Gain is recognized under §1001; the CV takes cost basis; holding periods restart. QSBS original-issuance status dies at the fund level and is not reborn for rolling LPs. Related-party loss disallowance under §§267 and 707(b) can apply if the deal is a loss and common control exists — common sponsorship alone does not answer the control test; ownership has to be mapped.

Some investor bases prefer this path. Tax-exempts and non-U.S. investors holding through blockers, and taxable LPs with high basis or usable losses, may care more about a clean basis reset than deferral. A taxable sale is also the honest structure when nobody can get comfortable with §721(b), §752, or mixing-bowl risk.

Merger treatment

Where the CV is combined with the legacy fund rather than split from it, Treas. Reg. §1.708-1(c) supplies merger forms that parallel the division forms: assets-over as default, assets-up if the title work is done. The continuation test looks to which partnership’s partners end up owning more than 50 percent of the resulting partnership. Cash paid to exiting partners can be treated as sale proceeds for the redeemed interests under Treas. Reg. §1.708-1(c)(4) if the agreement so designates. Without that designation, the cash can be analyzed as boot and pulled into the disguised-sale rules.

Disguised-sale architecture

Whatever label the deck uses, Treas. Reg. §1.707-3 asks whether a contribution of property and a transfer of money are, in substance, a sale; transfers within two years are presumed to be one unless the facts clearly establish otherwise. Treas. Reg. §1.707-5 governs liabilities; §1.707-6 addresses purchases of interests. A well-built CV uses these rules deliberately on the selling side and stays outside them on the rolling side. That split is a drafting and funding problem: cash from new investors must reach sellers without tainting the rolling partners’ contributions, and debt reallocations must not manufacture deemed money under §752 for people who thought they were only changing wrappers.

Swipe sideways for the other columns.

A DRAFTING RULE

Pick the form. Name it in the contribution and purchase agreements. Designate how seller cash is characterized under Treas. Reg. §1.708-1(c)(4) or the disguised-sale regulations. The regulations will invent a form if you do not, and the invented one is rarely the one you wanted.

06 Selling LPs: A Real Exit Means Real Tax

The selling LP is not trying to defer anything; the selling LP wants cash. The work is making the taxable result follow the economics. In many U.S. structures, cash from new investors enters the CV and funds a redemption or a purchase of interests held through a seller partnership. The amount of gain is only the start. Character is often the larger number.

• Section 751 can convert part of partnership-interest gain into ordinary income to the extent attributable to unrealized receivables and inventory-type items. In tiered structures, look-through matters; credit assets and depreciation recapture are the usual sources.
• Corporate stock can carry §1248 consequences for certain foreign corporations. Depreciable assets bring §§1245 and 1250 recapture.
• Debt relief — a reduction in the seller’s share of liabilities — increases amount realized or creates a deemed cash distribution under §752.
• State and local sourcing may not follow the federal capital-gain answer, especially for operating partnerships, real estate, and carried interests. A California or New York taxable LP can owe a state bill the federal model never saw.

Selling at a loss is a different problem: related-party rules under §§267 and 707(b) can disallow or defer losses in sponsor-controlled structures.

07 Rolling LPs: Deferral Is Not a Basis Step-Up

For a rolling LP, “tax-free” usually means tax-deferred. If the old fund held an asset worth $100 with a tax basis of $30, moving it into a CV does not erase the $70 of built-in gain. It changes where that gain is tracked.

Section 704(c) is the mechanism that keeps new investors from inheriting the old investors’ pre-CV gain. The CV tracks the difference between tax basis and book value and, when the asset is later sold, allocates the historic built-in gain back to the partners who economically owned it before the rollover. In a single-asset CV that layer is conceptually one number. In a multi-asset CV each asset carries its own layer, holding period, and character.

Tax allocations and cash can then separate. A rolling LP can be allocated historic gain when an asset is sold even if the waterfall or reserves delay the related cash — phantom income, with real estimated-tax deadlines. Holding period in the CV interest, holding period in the underlying asset, tax capital, and outside basis all need to be mapped; those numbers feed §1061, loss utilization, and later distributions.

08 New Investors: Basis Is an Investment Term

A new investor pays today’s price and wants today’s tax basis, at least for its own economics. If the CV holds low-basis assets and no investor-specific adjustment is made, the new investor pays tax on appreciation that accrued before it arrived — a subsidy no one intended and no one priced.

Where the transaction is treated as a purchase of partnership interests, a §754 election permits a special inside-basis adjustment under §743(b) for the transferee. The adjustment does not rewrite the partnership’s common basis; it is personal to the buyer. If a new investor pays $100 for an interest economically attributable to an asset with $30 of common basis, a properly allocated §743(b) adjustment is the difference between a cost-basis investment and an investment still carrying someone else’s $70 of gain.

The negotiation then moves to administration: who pays for the valuation, the allocation, and the annual tax accounting? Sponsors increasingly treat those costs as a beneficiary-pays item when only the incoming class receives the step-up. Fair enough — but write it down.

09 The §721(b) Investment-Company Trap

Section 721(a) is broad. Section 721(b) is the exception that matters in multi-asset and financial-asset CVs: a contribution to a partnership is taxable if the partnership, were it incorporated, would be an investment company under the corporate rules. Diversification, and the presence of stocks, securities, and cash around the contribution date, are the statutory concern.

A single-asset CV holding one private operating company usually has a cleaner path. A vehicle holding a basket of marketable stocks or fund interests does not. The irony is commercial: the more diversified transaction can be the more fragile nonrecognition transaction — partnership tax’s little joke at modern portfolio theory’s expense. The fact to test is the exact asset mix immediately after the contribution, not the marketing description of the strategy.

10 Marketable Securities, Mixing Bowls, and Seven Years

Partnership tax remembers. The rules are built to stop partners from contributing appreciated property and then using distributions to finish what is economically a sale:

• Section 731(c) treats certain marketable securities as money for distribution purposes.
• Section 704(c)(1)(B) can trigger built-in gain if contributed property is distributed to another partner within seven years.
• Section 737 can trigger pre-contribution gain if a contributing partner later receives other property.

These rules do not make CVs unworkable. They make post-closing asset movement a formation issue. Follow-on acquisitions, in-kind distributions, spin-offs, and internal reorganizations should be modeled at closing — not discovered in year three by whoever inherited the file.

11 Debt Can Turn a Rollover into Cash

Section 752 is easy to explain and easy to miss in a board paper. A decrease in a partner’s share of partnership liabilities is a distribution of money; if that deemed money exceeds outside basis, §731 gain follows.

The trap appears when the legacy fund or an underlying portfolio partnership is levered and the CV changes who bears the debt. It also appears when NAV financing, a subscription facility, or acquisition debt arrives around the deal. The arithmetic fits on a term sheet: a partner with $10 million of outside basis whose share of liabilities falls from $20 million to $6 million has received a deemed $14 million distribution and recognizes $4 million of gain — at a closing everyone described as tax-free. Debt finances the transaction and reallocates tax basis; those two lives should be modeled together, in the tax workstream — not in a financing appendix nobody on the LPAC reads. The Northline case study below puts numbers on this.

12 Carried Interest: Two Bites, Two Analyses

The old carry

At the CV price, the legacy fund has hit a realization point for the transferred asset, and accrued carry is measured under the old waterfall. Market practice still favors a substantial rollover of that accrued carry, even where the economics have crystallized, because alignment sells better than a GP cash-out.

Section 1061 generally requires a three-year holding period for favorable long-term capital-gain treatment on gain allocated through an applicable partnership interest. In a properly structured CV, the relevant underlying asset holding period may carry through the restructuring — but the exact steps matter, and a taxable sale-and-repurchase can restart the clock. Profits interests raise a separate safe-harbor issue: Rev. Proc. 93-27 does not apply to certain dispositions within two years of receipt, and Rev. Proc. 2001-43 depends on continued satisfaction of its conditions. Rolling or restructuring a recently issued profits interest therefore deserves its own analysis even when the economics barely moved.

The new carry

The CV typically has a new waterfall, a new hurdle, and a new measurement period. Morgan Lewis’s 2026 review of 169 CVs found tiered carry in 79 percent of deals, with 60 percent using both IRR and MOIC thresholds. Evercore reports that super-carry — a promote above 20 percent — appeared in about 35 percent of first-half 2026 CV volume, up from 15 percent in full-year 2025. Management fees have compressed; 97 percent of the Morgan Lewis sample charged 1 percent or less. The market is trading fee for promote, which tells you where the parties expect the value creation to sit.

This is a genuine second incentive period. The sponsor can realize carry on value created in the legacy fund and earn a fresh promote on value created after the CV price. That can be fair. It is fair only if LPs can see, in the waterfall, that the old carry is charged on pre-CV value and the new carry only on value created after that price.

THE CARRY QUESTION LPS SHOULD ASK

If I roll, am I paying carry twice on the same appreciation — or is the old carry crystallized only on pre-CV value and the new carry charged only on new value? The answer lives in the distribution section, not the marketing deck. Well-run processes can point to the paragraph.

13 Blockers, Non-U.S. Investors, and Withholding

An existing tax-exempt or non-U.S. investor may have invested through a corporation to block UBTI or ECI. If the blocker holds an appreciated asset, that investor may want to transfer the stock, keep the corporation, and avoid an entity-level asset sale. A new investor may want the opposite: a fresh blocker (or none) and a basis step-up in the asset. Buying old blocker stock can mean inheriting low inside basis and a latent corporate tax. Parallel ownership — old blocker for one class, new blocker for another — is a common truce, and it works only if the sponsor can keep the classes economically aligned after leakage.

The blocker can also become a pricing issue. A predominantly taxable buyer may value an interest held through an existing blocker at less than an economically identical unblocked interest because the buyer inherits latent corporate tax and may lose the benefit of asset-level basis. A predominantly blocked seller may resist giving up value for a structure it needed for its own tax profile. There is no universal “blocker discount.” Any adjustment is normally part of the broader negotiation over price, structure, tax attributes, and bargaining power. The important point is that the blocker is not merely a tax-compliance question; in the wrong buyer-seller mix, it becomes a valuation question.

The seller may be a partnership; the ultimate investors may be global. The agreement therefore needs a withholding architecture:

• Section 1446(f) can require withholding on transfers of partnership interests by non-U.S. persons when the partnership has effectively connected gain.
• Section 1445 can apply to transfers involving U.S. real property interests and USRPI-heavy partnerships.
• A seller-partnership structure can let the immediate seller deliver a Form W-9; the seller partnership can still owe obligations to its own foreign partners.
• Treaties and sovereign exemptions do not automatically solve partnership-level ECI or FIRPTA. The chain has to be tested link by link.
• Transfer taxes and stamp duties sit outside the income-tax system and still close deals, especially where real property or local entities move.

Gross-up provisions, certificates, refund mechanics, cooperation covenants, and the allocation of a failed withholding position belong in the election materials — not in a side letter signed the night before closing by whoever was still awake.

14 The Venture Problem: §1202 After the OBBBA Rewrite

Qualified small business stock can exclude a large share of gain for eligible noncorporate taxpayers, and the stakes went up in 2025. For stock issued after July 4, 2025, the One Big Beautiful Bill Act raised the per-issuer exclusion cap from $10 million to $15 million (inflation-indexed from 2027), raised the issuer gross-asset ceiling from $50 million to $75 million, and replaced the five-year cliff with a tiered schedule: 50 percent exclusion at three years, 75 percent at four, 100 percent at five, with the non-excluded slice under the early tiers taxed at the 28 percent rate. Stock issued on or before July 4, 2025 keeps the old rules. Venture portfolios now run a dual-track QSBS book, share class by share class, issuance date by issuance date.

None of that generosity survives careless structuring. QSBS depends on original issuance, and the partnership rollover rules do not preserve it. Section 1202(h) protects certain transfers — gifts, transfers at death, certain distributions of stock from a partnership to a partner — but not a contribution of QSBS by one partnership to another. The §1045 regulations are blunt: stock contributed to a partnership is not QSBS in the transferee partnership’s hands, because the transferee did not acquire it at original issue. Practitioners have flagged the standard assets-over CV as a way to burn some or all of the §1202 benefit. Nor can a rollover upgrade old stock into the new regime: §1202(i) keeps pre-OBBBA stock on pre-OBBBA terms even through tax-free exchanges.

The trade can be expensive, and the arithmetic deserves a line in the model. If a rollover strips QSBS status, the eventual gain is regular long-term capital gain — 20 percent plus the 3.8 percent NIIT — rather than gain a completed five-year hold would have excluded entirely. A founder-LP with $14 million of exclusion-eligible gain on post-2025 stock therefore stands to lose roughly $3.3 million of federal benefit. (The special 28 percent §1202 rate applies only to the included slice of gain on stock that remains QSBS under the 50 and 75 percent tiers; it has no role once the stock has ceased to be QSBS.) The LP avoids tax today and gives up an exclusion that would have sheltered the eventual exit; deferral that costs a permanent exclusion is a loan you paid to make. For a venture CV, the §1202 analysis belongs at the top of the structuring memo. If the exclusion is large, a taxable sale may be the worse federal headline and the better economic result, or an interest-sale path that never contributes the stock to a new partnership may be the only path worth taking.

QSBS IS A DEAL-MODEL ITEM

Before moving venture assets into a partnership CV, quantify the exclusion at risk — under the right regime for each share’s issuance date, since post-July 4, 2025 stock carries the larger $15 million / $75 million parameters. Deferral is not automatically valuable if the price of deferral is permanent exclusion.

15 Credit, Real Estate, and Infrastructure

Private-credit CVs are growing quickly — Evercore counted GP-led deals at roughly $17 billion of $20.4 billion of credit-secondary volume in the first half of 2026 — and their tax profile is not a buyout profile with a different label:

• Accrued but unpaid interest can be ordinary income.
• OID keeps accruing on a yield-to-maturity basis whether or not cash shows up.
• Market discount can turn economic appreciation into ordinary income on disposition or principal payment.
• Loan modifications around the CV can be deemed exchanges under the debt-modification regulations.
• Distressed names raise COD, bad-debt, and worthless-security questions.
• Section 751 sits at the center when partnership interests hold debt with ordinary attributes.

Real estate and infrastructure add recapture, FIRPTA, transfer taxes, debt-driven outside basis, and, in some infrastructure deals, tax-credit and partnership-flip mechanics. A CV should not trigger recapture or reallocate incentives merely because the sponsor extended the hold.

16 Tax Indemnities, RWI, and the BBA

Continuation-fund paper has moved toward an M&A model: deeper diligence, broader representations, representation-and-warranty insurance appearing more often. Incoming investors are buying a seasoned asset with a tax history they do not control.

The negotiating question is who owns pre-closing tax risk. The usual toolkit is an excluded-obligation indemnity, RWI, a special escrow, a covenant to make a §6226 push-out election, and a rule that sends tax refunds back to the party that economically paid the tax. The question sharpens when the CV is the tax continuation of the legacy partnership — a tax identity can survive a transaction the investment committee describes as a new fund.

17 Single-Asset versus Multi-Asset

Swipe sideways for the other columns.

18 The LP Election Is an After-Tax Allocation of Capital

An LP choosing between cash and roll should not stop at whether it still likes the company. The CV price is a new investment decision. The useful question is: if I did not already own this asset, would I buy it today at this valuation, with this leverage, this duration, this governance, this carry schedule, and this tax basis? Familiarity is not an investment thesis, and deferral is not alpha.

For a taxable LP the roll has an embedded financing benefit: the government leaves the deferred tax inside the investment until the gain is recognized, so the investor can compound gross rather than net proceeds. That benefit can be meaningful when basis is low and the expected hold is long. But it competes with a newly validated entry price, concentration risk, another three-to-five-year duration, and a fresh carry schedule. Lazard’s 2025 return-target chart shows secondary buyers underwriting at approximately 2.2x / 20.6% net MOIC / IRR for single-asset CVs and 2.0x / 19.3% for multi-asset CVs. A rolling LP should compare its after-tax expected return with those new-money hurdles — not with the return the old fund already earned.

THE ROLL DECISION

Compare the after-tax value of selling today and reinvesting the net proceeds elsewhere with the after-tax expected value of rolling the gross pre-tax exposure into the CV. Then run the same comparison at a common risk assumption. Deferral is worth more when embedded tax is large, the expected hold is meaningful, and the alternative requires paying tax first. It is worth less when the CV price is full, leverage has less room to help, fees and carry reset hard, or the investor’s basis is already high. The election is an after-tax capital-allocation decision, not a referendum on the sponsor.

19 Case Study: Four Books on the Same Deal

Abstract rules hide the collisions, so here is a hypothetical — simplified on purpose, but close enough to a live single-asset buyout CV to show how the four tax stories interact. Figures are rounded; state tax is ignored in the base case and noted at the end. The deal is an assets-over division with a seller partnership for the cash electors and a §754 election in the CV.

The company and the funds

Harbor Partners Fund IV, a 2018-vintage buyout fund, acquired Northline Services in January 2019. Northline is a U.S. partnership for tax purposes; Fund IV holds the interest directly.

Swipe sideways for the other columns.

The $30 million of basis accretion since entry is retained earnings plus amortized transaction costs. Built-in gain at the CV price is $720 million ($1,000 − $280). Harbor has not delevered at the fund level; the $400 million of company debt is allocated to Fund IV partners under §752.

Fund IV’s partners split as follows at the CV price. The GP’s $80 million includes $20 million of remaining GP commitment and $60 million of carry value crystallized under Fund IV’s waterfall (20 percent over an 8 percent preferred) at the $1 billion equity price. For the case we treat the $60 million as earned on pre-CV value and rolled in kind.

Swipe sideways for the other columns.

New investors pay $550 million in cash for the selling LPs’ 55 percent and negotiate a 20 percent promote over a 1.25x MOIC / 12 percent IRR compound hurdle, with 25 percent super-carry above 2.0x. Rolling LPs and the GP take corresponding CV interests; the CV elects §754; the parties specify assets-over form. For purposes of the hypothetical, assume the rolling LPs and GP together held 45 percent of Fund IV’s capital and profits immediately before the division. They own 45 percent of the CV after closing, so the CV does not satisfy the more-than-50-percent continuation test and is treated as a new partnership. Incoming investors do not inherit Fund IV’s EIN or its old audit years. They still want indemnity and RWI for pre-closing tax of the company itself.

Northline: one company, two underwrites, three IRRs

``` BITE ONE — FUND IV BITE TWO — THE CV $250M to $1.0B equity · 4.0x · legacy carry $1.0B to $1.4B · 1.4x · new promote

$250M entry ──────────────▶ $1.0B CV price ──────────────▶ $1.4B 2030 exit Jan 2019 (both an exit and an entry) Jan 2030 §704(c) layer stamped here: $720M built-in gain waits for the rollers ```

The hypothetical Northline arc. The 2026 price is simultaneously Fund IV’s exit and the CV’s entry. That is the central measurement problem with CV track records: one sponsor can report a realization and a new investment around the same valuation event, while a rolling LP experiences one continuous exposure.

Closing day, party by party

Selling LP — Meridian Taxable LLC. Take a taxable U.S. seller owning one-fifth of the selling group: $110 million of equity value, $30.8 million of outside basis (already including its $44 million share of partnership liabilities), and $44 million of allocated debt. Amount realized is $110 million of cash plus $44 million of liability relief, or $154 million. Adjusted outside basis is $30.8 million — do not add the $44 million again; §752 liability basis is already inside outside basis. Gain is therefore $123.2 million. Assume 12 percent of the gain — about $14.8 million — is §751 ordinary income from depreciation recapture and other hot assets, with the remaining $108.4 million treated as long-term capital gain. If NIIT applies, federal tax is roughly $6.0 million on the ordinary slice plus $25.8 million on the capital slice: about $31.8 million total, leaving roughly $78.2 million of net cash from the $110 million equity check.

If the same LP is a California resident, add state tax. At a 13.3 percent top marginal rate, the incremental California tax on a $123.2 million gain can exceed $16 million before considering federal deductibility limitations and the exact sourcing facts. The after-tax gap between a California seller and a Texas seller is large enough to change the election for some taxable family offices. The election memo should say so — and should use the investor’s actual residence, basis, character mix, and state sourcing rather than a generic federal rate.

Rolling LP — Oakridge Family Holdings. Oakridge owns one-tenth of the rolling group: $37 million of equity value, $10.4 million of outside basis (including its $14.8 million share of company debt), and $14.8 million of allocated company debt. The fact that Oakridge’s allocated liabilities exceed its $10.4 million outside basis is not itself an inconsistency; partnership liabilities can support outside basis even where the partner has negative tax-basis capital. If the assets-over steps hold and the debt allocation is preserved, Oakridge recognizes nothing at closing. Its carryover outside basis in the CV interest remains $10.4 million, its share of liabilities stays $14.8 million, and the $26.6 million of historic built-in gain remains tracked under §704(c).

Now change one fact. Suppose the CV adds $100 million of new holdco financing and uses $40 million of it, economically, toward seller proceeds and costs. If Oakridge’s allocated debt falls by $12 million against $10.4 million of outside basis, §752 plus §731 produce $1.6 million of gain at closing. The rollover is no longer clean. This is the board-level §752 point, and it is why debt models belong in the tax workstream.

Oakridge’s holding period in Northline tacks. Section 1061 is not Oakridge’s problem (it holds no API), and QSBS is not in play — Northline is a partnership, not C-corporation stock acquired at original issue.

New investor — Westbrook Secondaries. Westbrook writes $200 million of the $550 million of new money for a 20 percent CV equity interest and is allocated 20 percent of Northline’s $400 million debt, or $80 million. If the transaction is documented as a purchase of CV partnership interests from the seller partnership, Westbrook’s outside basis is therefore $280 million: $200 million of cash cost plus $80 million under §752. Its share of the CV’s common inside basis is approximately $56 million, producing a preliminary upper-tier §743(b) adjustment of about $224 million. Because the CV’s principal asset is itself an interest in the Northline partnership, that adjustment initially attaches to the lower-tier partnership interest; whether and how the benefit reaches Northline’s underlying assets requires a separate tiered-partnership analysis, including the election status and applicable rules at both tiers. The point is simple even when the schedule is not: a buyer paying today’s equity price should model today’s outside basis, including debt, against the common inside basis it actually inherits — and should not assume a headline §754 election automatically creates asset-level depreciation basis where the structure is tiered.

Westbrook also negotiates three unglamorous points worth more than a 10-basis-point fee tweak: the purchase mechanics must actually support §743(b); the CV will maintain the §754 election and complete the initial §743/§755 study; and pre-closing company tax is an excluded obligation backed by RWI plus a $15 million special escrow. It also specifies who pays the annual compliance cost of maintaining a large transferee-specific basis schedule. Basis is valuable only if someone can still prove it four years later.

The GP — Harbor Partners. Harbor rolls $20 million of commitment and $60 million of crystallized Fund IV carry into 8 percent of the CV. First question: is the $60 million a taxable disposition of a profits interest, a book-up and continuation of the old interest, or a new grant? The parties treat the roll as a nonrecognition exchange of a seven-year-old profits interest for a CV profits interest tied to the same asset, with a $60 million capital account after the book-up. Rev. Proc. 93-27’s two-year disposition limit is behind them; §1061’s three-year period is behind them on the old value, provided the asset’s holding period tacks. A sloppy recap the IRS could treat as a sale-and-regrant would reset both clocks to zero.

Second, the new promote: 20 percent above the hurdle, 25 percent above 2.0x, applied only to post-CV appreciation. It does not reach back to the $720 million of pre-CV built-in gain. Rolling LPs should be able to verify that sentence in the distribution section — not infer it from the summary materials.

Four years later

In January 2030 the CV sells Northline at $1.80 billion of enterprise value. Net debt is still $400 million; equity value is $1.40 billion; post-CV appreciation is $400 million. Ignore transaction costs and the unused follow-on reserve. The waterfall has cleared the 1.25x hurdle. Harbor’s new carry on the $400 million of post-CV profit is $80 million; the remaining $320 million is shared 55 / 37 / 8 among new investors, rolling LPs, and the GP’s committed and rolled-carry capital. We keep the arithmetic flat so the tax points stay visible.

Swipe sideways for the other columns.

Three collisions now show up in numbers — and they are the reason a CV cannot be understood from the headline purchase price alone.

• Oakridge receives 2030 cash that includes its $37 million of rolled value plus its share of the $400 million of new profit — and is allocated the $26.6 million §704(c) layer that has been waiting patiently since 2026. Deferral worked; forgiveness did not. If the waterfall reserves cash for indemnities or escrow, Oakridge can have phantom income in the exit year.
• Westbrook’s return is a four-year compounding problem from a $1.0 billion equity entry, not from Fund IV’s $250 million entry. A 40 percent equity gain in four years is respectable and is a different investment from the 2019–2026 ride. Westbrook’s tax result tracks that second ride only if the interest-purchase mechanics, §754 election, §743(b) calculation, §755 allocation, and later reporting all work. Kill the adjustment — or fail to substantiate it — and the buyer can inherit tax history it did not economically earn.
• Harbor reports a strong Fund IV realization at 4.0x gross equity and, separately, a CV result that depends on the $1.0 billion starting point. Both numbers are true, and both belong in the track record — side by side, with the CV price visible between them.

Use the case as a template, not as advice. Change the rollover percentage and the 50 percent test flips. Change the debt allocation and the rolling LP has closing-day gain. Change the asset to C-corporation QSBS and the assets-over contribution may burn the exclusion. Change the seller to a non-U.S. partner and §1446(f) withholding appears. The four books have to close. If they do not, the missing money comes out of someone at the table.

20 Did the Second Bite Outperform?

This is the question the market should ask more often. The honest answer, in August 2026: the evidence is finally becoming useful, but it is still too young and too marks-heavy to call the second bite a proven source of alpha. The better question is not whether CVs are ‘good.’ It is whether a new investor — and separately a rolling taxable LP — is being paid enough for the price, concentration, duration, governance, carry reset, and tax profile at the CV entry point.

What the question even is

A CV creates two reportable investment periods around one company. The legacy fund measures return from original purchase to the CV price; the CV measures return from the CV price to the next exit. Different entry values, leverage, fees, carry, duration — different investments. A sponsor can produce an excellent legacy realization and a dull CV return without contradiction, and the inverse happens too: a CV bought on a conservative mark in a closed IPO window can later look clever. Asset quality and investment return are not the same thing. A wonderful company can be a poor purchase at the wrong price; ask anyone who bought Cisco in March 2000.

What the studies actually show

Oliver Gottschalg’s HEC Paris paper on 2018–2023 vintages assembled 386 continuation funds with performance data for 297. The preliminary analysis finds little return difference by geography, industry, or single- versus multi-asset format. Comparing 133 buyout funds from the 2020 vintage with 149 simulated multi-asset portfolios built from single-asset CVs, he finds comparable returns and a narrower spread of outcomes for the single-asset vehicles. A companion reading of the same research program, reported in the trade press, puts mean TVPI for the sample near 1.56x and mean net IRR near 18.8 percent.

Evercore’s ongoing work with HEC, updated through 2018–2024 vintages, has been summarized as a 23.6 percent average net IRR for CVs, with single-asset vehicles performing in line with buyout funds; an earlier cut put average single-asset TVPI at 1.499x against 1.513x for 2019-vintage buyouts. Those two sentences are often folded into a claim that CVs “outperform.” They do not say that. They say that, on the marks and realizations available so far, a portfolio of CVs has looked like a portfolio of buyout funds, with less dispersion in the single-asset sleeve. Treat the high interim IRRs with care: short duration plus unrealized marks produces large IRRs that may not survive a full harvest cycle, and DPI in the early cuts was still low — as one would expect for vehicles designed to hold.

Lincoln International’s company-level work is a different lens. In 93 single-asset CV transactions from 2024 through mid-2025, the transferred companies were older holdings (5.7 years on average in the observed subset) with stronger revenue and EBITDA growth, higher margins, and premium multiples versus broader private-market benchmarks. That supports the claim that sponsors send good, mature companies into CVs. It does not tell the LP what buying those companies at the CV price will earn.

How to read the evidence without fooling yourself

• Selection bias is the feature. Sponsors put their better remaining assets into single-asset CVs, so a sample of those assets should look better than the average leftover. The LP already owned many of them; the decision is price and terms, not whether Northline is a decent business.
• Survivorship and mark-to-model bias are still large. Most 2018–2023 CVs have not fully realized; TVPI embeds sponsor marks at the CV and, later, at the CV’s own NAV. The first heavy realization years will tell us more than another year of interim IRRs.
• Fee and carry resets change the LP’s share of the same company. A 20 percent promote on post-CV value — and, in a growing share of deals, super-carry above it — is a real haircut on the second ride. Morgan Lewis’s 79 percent tiered-carry finding and Evercore’s 2026 super-carry print belong in the return model, not a governance appendix.
• Average LP rollover near 15 percent is itself evidence. Most eligible legacy capital is therefore monetized rather than rolled. That may reflect liquidity needs rather than a view on CV alpha. Revealed preference is still not a footnote.
• CV-to-CV exits are no longer theoretical. Lazard flags continuation-to-continuation deals as an emerging structural tool; Onex’s process around Ryan is a public example. A third measurement period, a third carry clock, another layer of basis — and a measurement and disclosure question that compounds with each vehicle. The continuous look-through return becomes more valuable, not less, as the chain lengthens.

Swipe sideways for the other columns.

For rolling LPs there is a third metric: the continuous look-through return from original purchase through the CV exit. It is the most honest number because it covers the whole period the LP owned the company, and it should sit next to the legacy-fund IRR and the standalone CV IRR. Anything less can make one asset look like two unrelated success stories.

OUR PREFERRED REPORTING

Show three numbers: (1) legacy-fund return through the CV transaction; (2) standalone CV return from the CV price; and (3) continuous look-through return for rolling LPs. Add, for taxable LPs, an after-tax roll-versus-sell exhibit on actual basis and character. The last of those is still rare. It should not be.

So did the second bite outperform? On the public evidence, it has so far behaved like a buyout product built from above-average companies purchased at above-average prices, with less dispersion in the single-asset sleeve and a second incentive period to fund. That can be an excellent product. It is not yet evidence that retaining the company beats selling it. The next wave of realizations will test whether superior asset selection overcomes the arithmetic of a higher entry multiple, concentration, fresh duration, and another carry clock. Until then, tax is one of the few sources of return the parties can engineer with high confidence on day one — which is precisely why basis, rollover mechanics, withholding, and carry design belong in the investment model rather than an appendix.

21 What We Would Require in the Tax Workstream

• A step plan stating the intended federal characterization of every legal transfer, including the form of division or merger and the designation of seller cash.
• A partner-by-partner model of outside basis, §752 liabilities, and expected gain at closing, with a toggle for the 50 percent continuation test and a state-residence toggle for taxable partners.
• An asset-by-asset schedule of tax basis, fair value, holding period, and character, including §751 attributes.
• A §721(b) investment-company analysis on the asset mix immediately after the contribution.
• A §704(c) methodology and a plan for maintaining the layers for the life of the CV.
• A §754 / §743(b) model for incoming investors, including who bears the compliance cost.
• A blocker map showing old basis, latent entity-level tax, and class-by-class ownership.
• A withholding matrix under §§1445, 1446(a), and 1446(f), plus state and foreign regimes.
• A carried-interest memo covering crystallized carry, the roll, §1061, and profits-interest history under Rev. Proc. 93-27.
• A QSBS schedule for any venture or growth asset — dual-track for pre- and post-OBBBA issuances — prepared before an assets-over structure is chosen.
• A BBA audit-risk allocation, including push-out mechanics and pre-closing indemnity or RWI.
• An after-tax roll-versus-sell exhibit for taxable LPs, using actual basis and the actual waterfall.

22 What We Would Watch Next

• CV-to-CV exits, and the third layer of basis, carry, and history they create.
• Retail and evergreen capital, which will pressure tax reporting, liquidity mechanics, and blocker design.
• Private credit, which may force a more standardized approach to ordinary-income assets and tax basis.
• QSBS guidance or litigation, as venture CV activity — and the fatter post-OBBBA exclusion — makes the mismatch between §721 and §1202 harder to ignore.
• Realized CV returns, once the 2019–2021 vintage vehicles finish their original terms.
• Tax disclosure to LPs. Sophisticated investors will ask for an after-tax election analysis. They should.

23 Our Take

Continuation funds solve a real capital-markets problem. They create liquidity without forcing every investor to sell, let LPs with different horizons choose different paths, bring fresh capital to mature assets, and give sponsors a way to separate the decision to monetize from the decision to stop owning. That is economically useful — and it is why the product is becoming permanent.

The same flexibility makes them unusually tax-intensive. A CV is a negotiated separation of investors who want to leave from investors who want to stay, with new capital entering at the exact price where old capital exits. Partnership tax has to decide which pieces are sales, which are contributions, which basis follows whom, which holding periods survive, and which deferred gains travel forward. Capital markets determine the price. Tax determines how much of that price each constituency actually keeps.

The most useful tax question is not whether “the CV is taxable” — the transaction is too layered for one adjective. Better: taxable to whom? At what step? On what amount and character? With what basis afterward? And who bears the risk if the intended characterization is wrong?

Sponsors that answer those questions early will price deals more accurately, give LPs better choices, reduce closing friction, and avoid leakage that no later financial engineering can recover. The first generation of CV analysis asked whether a sponsor could sell an asset to another vehicle it manages and still run a fair process. The next generation should ask a harder question: did the transaction create superior after-tax, risk-adjusted value for the investors who chose each path? A flagship CV process should be able to answer that question before closing — and then report the answer after exit.

Selected research and primary authorities

Lazard, 2025 Secondary Market Report (Feb. 2026) and Interim 2026 Secondary Market Report (Aug. 2026): global and GP-led volume, single-asset and pricing mix, dry powder, and the ~$275B 2026 forecast.

Jefferies, 2025 Global Secondary Market Review (Feb. 2026): CV size, 89% CV share of GP-led volume, sponsor adoption, average LP rollover, CV share of sponsor-backed exits.

Evercore Private Capital Advisory, 2025 and first-half 2026 secondary market reporting: GP-led mix, single-asset volume, credit-secondary GP-led share, super-carry frequency.

Morgan Lewis, Annual Continuation Vehicles Report 2026: terms across 169 CVs — fees, tiered carry, hurdles, term, key person.

Oliver Gottschalg, HEC Paris, “Continuation Funds”: Performance and Determinants, 2018–2023 Vintages (2025), HEC Paris Research Paper SPE-2025-1584; Evercore/HEC updates through 2018–2024 vintages as reported in the trade press.

Lincoln International, Single-Asset CVs: How They Stack Up (Dec. 2025): operating and valuation characteristics of companies placed into SACVs.

CFA Institute Research and Policy Center, Continuation Funds: Ethics in Private Markets (2025); ILPA, Continuation Funds: Considerations for Limited Partners and General Partners.

Daniel A. Zygielbaum & Jennifer A. Fitzgerald, Continuation Funds: Tax Issues, Practical Tax Strategies (Jan. 2025); DLA Piper, Continuation Funds: Preserving Tax-Free Rollover Treatment Under IRC §721 (2026); Deloitte, M&A Tax Talk: Continuation Funds; RSM, Continuation Vehicle Transactions Can Bring Complexities for PE Funds.

One Big Beautiful Bill Act, Pub. L. 119-21 (July 4, 2025), amendments to §1202 (tiered holding periods; $15M cap; $75M gross-asset ceiling) for stock issued after July 4, 2025.

Statutes: 26 U.S.C. §§267, 704, 707, 708, 721, 731, 737, 741, 743, 751, 752, 754, 1001, 1045, 1061, 1202, 1245, 1248, 1250, 1445, 1446, 6226.

Regulations and rulings: Treas. Reg. §§1.704-3, 1.704-4, 1.707-3, 1.707-5, 1.707-6, 1.708-1(c), 1.708-1(d), 1.737-1, 1.743-1, 1.751-1, 1.752-1, 1.1045-1; Rev. Rul. 84-52; Rev. Proc. 93-27; Rev. Proc. 2001-43; IRS Publication 515.

Important notes

This alert addresses U.S. federal income tax. State, local, and non-U.S. consequences can change the result and must be analyzed by investor and jurisdiction. Tax treatment depends on the exact legal steps, governing documents, asset types, investor profiles, and elections. The Northline figures are a hypothetical used to show mechanics; they are not a prediction and not a description of any client matter.

Market-volume and performance estimates differ among advisers because methodologies and samples differ. Where leading sources disagree, this alert names the source rather than pretending the numbers reconcile.

C . A . S .

Download the alert (PDF, 578 KB)

This alert is for general information only and is not tax, legal, investment, or accounting advice. Consult your own advisers before acting. Carlos A. Schmidt, MBT, MBA, CPA · carlos@lapresallc.com · (917) 558-6393.