How Single Family Offices Run Direct and Co-Investment Programs: A Guide for GPs

The Direct-Investing Story Is Real, but It Is Not a Straight Line

Ask a placement agent about single family offices and the conversation tends to start with the same claim: families have stopped writing blind-pool commitments and now want to "go direct." The claim is half right. Family offices do hold a large share of their private equity exposure through direct deals and co-investments, and they have done so for years. But the published survey data shows a cycle, not a one-way migration, and the most recent numbers show direct allocations being trimmed rather than expanded.

That nuance matters to anyone raising capital. A GP that pitches a family office as a pension plan with a smaller ticket will miss how the office decides; one that assumes every family wants control deals will miss the offices that want co-investment exposure precisely because they cannot staff a direct team. This analysis covers the survey data, team build-out, sourcing, structuring, the research on co-investment quality, and tax, and closes with what GPs and placement teams should take from it. Survey figures describe each survey's own respondent sample, not the whole family office universe.

---

Part I: What the Surveys Actually Show

The UBS series: directs peaked in 2021

The longest consistent public time series comes from the UBS Global Family Office Report, which splits private equity into "direct investments" and "funds / funds of funds." The 2025 edition, based on 317 UBS family office clients surveyed between January and April 2025, reported the global average strategic allocation to direct private equity at 9% in 2019, 10% in 2020, 13% in 2021, 9% in 2022, and 11% in both 2023 and 2024. Fund and fund-of-funds private equity moved from 7% in 2019 to 10% in 2024.

Three things stand out in that series:

• Directs were larger than funds in most years. In 2024, the UBS 2025 report put total private equity at 21% of the average portfolio, split 11% direct and 10% funds / funds of funds.
• The peak was 2021. Direct allocations hit 13% that year, then fell back to 9% in 2022.
• The planned direction was down. Family offices that intended to change their allocation planned to take direct private equity to 8% in 2025 while holding funds at 10%. UBS attributed the trimming to subdued capital markets and acquisition activity alongside high financing costs. One US family office executive quoted in the report put it plainly: if you do not have an exit, you do not have any cash to recycle.

The UBS Global Family Office Report 2026, based on 307 clients surveyed online between 22 January and 30 March 2026, confirms that the trim happened. Its 2025 strategic allocation shows 8% of the average portfolio in direct private equity and 9% in private equity funds / funds of funds. That puts the fund sleeve slightly ahead of the direct sleeve in the UBS global average, which the series had last shown in 2022.

Regional differences are large. The UBS 2025 report noted that US family offices cut direct private equity from 21% of their portfolios in 2023 to 14% in 2024. In the 2026 report, US offices show 10% in direct private equity and 10% in funds for 2025.

Other bank surveys: participation remains broad

Allocation percentages tell you how much capital sits in directs. Participation rates tell you how many offices are active at all, which is the more relevant number for a placement team building a target list.

• Citi Wealth 2025 Global Family Office Report (346 respondents from 45 countries, surveyed June and July 2025): 70% of respondents said they were engaged with direct investments, and four in ten of those reported increased or significantly increased activity over the past year. Citi also reported that private equity saw the most positive activity.
• BNY Wealth 2025 Investment Insights: Single Family Offices (282 respondents, fielded January 6 to February 3, 2025, offices managing at least $250 million): 64% of family offices expected to make six or more direct investments in the coming year. The same report recorded a 52% increase in respondents citing "alignment of interests" as a consideration compared with the previous year.

Reading the data correctly

Together, the surveys describe a population where direct and co-investment activity is common and sticky, but where the capital devoted to it flexes with the exit cycle. The UBS 2025 report also found that 37% of respondents expected a significant or moderate increase in direct private equity over the next five years (34% for funds / funds of funds), which UBS said could reverse the trimming some offices planned for 2025.

For a GP, the practical reading is that an office's appetite for co-investment in any given year can depend heavily on its own liquidity position, not just on its stated strategy. Recent realizations can say as much as stated appetite.

---

Part II: Building a Direct and Co-Investment Program

What a direct program actually requires

A fund commitment needs a manager-selection process and a monitoring function. A direct program needs, at minimum, sourcing, underwriting, legal negotiation, portfolio monitoring, valuation, and an exit plan for every position. Co-investment sits in between: the sponsor does the sourcing and leads the diligence, but the family office still needs to make a fast, independent decision and absorb the legal and administrative work of each vehicle.

The staffing data suggests many offices are thinly resourced for that load.

Headcount is lean. The UBS Global Family Office Report 2025 found that family offices employ 12 people on average, with wide variation; 6% of the US offices surveyed employed more than 50 staff. The 2026 edition adds that non-investment professionals (operations, accounting, legal support, lifestyle services) make up 40% of total staff numbers.

Cost is mostly people. UBS's 2025 report put the global average pure operating cost at 41.1 basis points of assets under management in 2024, with staff costs accounting for two-thirds (66%) of those pure expenses. For offices with more than $1 billion in AuM the average was 35.1 bps. The J.P. Morgan 2026 Global Family Office Report, based on 333 family offices across 30 countries, reported that average annual operating costs for offices with more than $1 billion in assets now exceed $6.6 million.

Understaffing is the named constraint. The BNY Wealth 2025 report found that understaffing is "emerging as a significant bottleneck for direct investing programs," with 44% of US offices citing it as a key challenge, an 83% year-over-year rise. For non-US offices, BNY identified a different challenge, limited scalability: in certain geographies, the set of potential local investees is comparatively small, prompting some skepticism about building the capability.

Three operating models

Described generally, and not attributed to any particular family, direct programs tend to follow one of three patterns:

1. Co-investment-led. The office keeps a fund portfolio and uses relationships with its GPs to access co-investments. The key skills are fast screening and legal turnaround. BNY's 2025 report describes this route explicitly, noting that working with a GP who sources opportunities, supports diligence, and structures transactions "offers an appealing mechanism to start building such in-house knowledge."
2. Sector-specialist direct. The office concentrates on industries where the family has operating history, often adjacent to the business that created the wealth. The UBS 2026 report notes that 77% of surveyed families still had an active operating business, which is a source of both deal flow and operating judgement.
3. Institutional direct platform. Some very large offices run something closer to a private equity firm, with dedicated deal teams, portfolio operations staff, and board representation. This is typically the most resource-intensive model to run.

GPs should not confuse the frequency of an office's direct activity with the depth of the team behind it.

The regulatory perimeter: why a family office cannot simply open its deals to friends

US single family offices typically rely on the SEC's family office rule, Rule 202(a)(11)(G)-1 under the Investment Advisers Act of 1940, to stay outside the definition of "investment adviser." The rule has three conditions: the office has no clients other than family clients; it is wholly owned by family clients and exclusively controlled by family members or family entities; and it does not hold itself out to the public as an investment adviser.

The definition of "family client" includes family members, certain former family members, key employees, and qualifying trusts, estates, charitable organizations, and wholly owned entities. It does not include unrelated families. The consequence for club deals: a single family office can invest alongside other families, but if it starts managing a vehicle for them, advising them and taking fees or carry, it risks stepping outside the exclusion. Syndication to outside investors is therefore often structured so that each family invests directly, or through a vehicle run by a registered or exempt-reporting adviser.

---

Part III: Sourcing — Sidecars, Club Deals, and Direct Origination

GP co-invest sidecars

The most common entry point is the co-investment that a fund sponsor offers to its limited partners and selected third parties alongside a fund deal. The sponsor typically has more equity to deploy than it wants its main fund to hold, either because of concentration limits or because it wants to win a larger deal without over-weighting one position.

The ILPA Principles 3.0 (2019), a widely referenced LP-side governance framework, set expectations for how GPs should run this process. Among them:

• GPs should disclose to all LPs, in advance, a framework for how co-investment opportunities, interests, and expenses will be allocated among the fund and participating co-investors, including whether any prioritization applies.
• Where co-investment rights have been granted in side letters, GPs should disclose the existence of those arrangements to all LPs.
• All suitable opportunities should first be allocated to the fund if they fit its strategy and the fund has remaining commitments.
• When presenting a co-investment, GPs should give prospective co-investors the strategic reasoning for including a co-investment tranche rather than allocating the whole amount to the fund.
• Broken-deal expenses should in most cases be shared pro rata across the fund, co-investment vehicles, and parallel vehicles.

These principles matter because they define what a sophisticated family office will ask for. An office that has been an LP in institutional funds will expect a written allocation policy, a clear explanation of why the deal is being syndicated, and an expense-sharing approach that matches ILPA's.

The SEC has also looked at this area. A June 23, 2020 Risk Alert from the then Office of Compliance Inspections and Examinations, "Observations from Examinations of Investment Advisers Managing Private Funds," reported that staff had observed advisers that disclosed a process for allocating co-investment opportunities but failed to follow it, and advisers that had agreements to give certain investors co-investment opportunities without adequate disclosure to other investors. In 2015, the SEC charged Kohlberg Kravis Roberts & Co. with misallocating more than $17 million in broken-deal expenses to its flagship funds; according to the SEC, co-investors, including KKR executives, benefited from the firm's deal sourcing but were not allocated any share of those expenses during a six-year period ending in 2011. KKR agreed to pay nearly $30 million to settle, including a $10 million penalty, without admitting or denying the findings.

For a family office co-investor, expense allocation is not a back-office detail: the co-investment documents should say where broken-deal costs and fees go.

Club deals: families investing alongside each other and alongside institutions

Club deals involve several long-term investors acquiring or recapitalizing a company together. Family offices can bring patient capital and operating or brand relationships.

The Merlin Entertainments consortium is a well-documented example. On 28 June 2019, a consortium of KIRKBI Invest A/S (a wholly owned subsidiary of KIRKBI A/S, the ultimate owner of the LEGO brand), funds managed as part of Blackstone's long-dated "Core" private equity strategy, and Canada Pension Plan Investment Board announced a recommended offer for Merlin Entertainments plc at 455 pence per share in cash. KIRKBI was already a 29.58% shareholder. According to the announcement, the offer was made through a newly incorporated company, with KIRKBI and the Blackstone/CPPIB group each owning 50% on completion. The release noted that KIRKBI had held a strategic shareholding in Merlin since selling LEGOLAND Parks to the company in 2005, and that KIRKBI and Blackstone private equity funds had jointly controlled Merlin in the eight years before its 2013 listing.

The structure shows what GPs should expect when a family office is a club partner: a strategic interest that predates the deal, equal ownership rather than a passive minority, and an emphasis on a long holding horizon.

Direct origination and long-held control positions

At the other end of the spectrum, some offices hold large positions they have built over many years. In September 2021, Four Seasons Hotels and Resorts announced that an affiliate of Cascade Investment had agreed to acquire half of the 47.5% stake held by an affiliate of Kingdom Holding Company for $2.21 billion in cash, increasing Cascade's ownership from 47.5% to 71.25% at a $10 billion enterprise value. The announcement noted that Cascade first invested in Four Seasons in 1997, when the company was public, and that founder Isadore Sharp would retain his 5% stake.

That transaction was not a GP co-investment but a purchase of shares from another existing shareholder of a private company. It shows the deal flow that offices with established positions can access without a sponsor, and why such offices may not be the right target for a standard sidecar pitch. A GP should also be clear about which route it is offering: experienced offices will not mistake a syndicated sidecar for a "direct" deal once they see the governance package.

---

Part IV: Structuring — SPVs, Minority Protections, and Governance Rights

The co-investment vehicle

Co-investments are usually made through a special purpose vehicle (SPV) controlled by the sponsor, through which co-investors hold an indirect interest in the portfolio company alongside the main fund. Alternatively, large co-investors may hold shares in the acquisition vehicle directly. The choice determines what the family office actually owns and what rights attach.

Issues a family office will typically examine in an SPV:

• Economics. Whether the sponsor charges a management fee or carried interest on the co-investment, and on what basis. ILPA Principles 3.0 acknowledge that GPs may charge fees on co-investments unless the LPA prohibits it, and ask that GPs disclose how such fees affect allocation between the fund and co-investors.
• Expense allocation. How broken-deal, transaction, and ongoing vehicle costs are shared, and whether fee income collected from the portfolio company is offset.
• Exit alignment. Whether the co-investment vehicle is required to exit at the same time and on the same terms as the fund. Sidecars are commonly built this way, which protects co-investors from being left behind but removes their ability to hold longer.
• Information rights. What reporting the co-investor receives, how often, and whether it includes company-level financials or only vehicle-level statements.
• Follow-on capital. Whether the co-investor has the right, or the obligation, to fund add-on acquisitions or rescue financings. ILPA asks GPs to disclose how follow-on investments related to co-investments will be allocated.

Minority protections in direct and club deals

When a family office invests directly as a minority holder, rather than through a sponsor's SPV, the shareholders' agreement carries most of the protection. The standard toolkit, described generally:

• Board seat or observer right, depending on the size of the stake and the other holders' tolerance.
• Reserved matters (consent rights) over items such as new equity issuance, major acquisitions or disposals, changes to the business plan, related-party transactions, debt above agreed thresholds, and changes to the constitutional documents.
• Pre-emption rights to protect against dilution in future rounds.
• Tag-along rights so the minority can sell on the same terms if the majority sells.
• Drag-along provisions, which the majority will want and the minority will want to limit with price floors or timing conditions.
• Exit mechanics, such as rights to require an IPO or sale process after a defined period, which matter more to family offices that may want to hold longer than the lead investor.

Comparing the three routes

The table below summarizes the main differences. Where a cell is qualitative, it is because published sources do not support a single representative number.

---

Part V: The Adverse-Selection Debate

The central question for any co-investor is whether the deals offered for syndication are as good as the deals the sponsor keeps. A sponsor has more information than its co-investors, and it decides which deals need outside equity. If it systematically offers the deals it is least confident in, or the largest deals that strain its own fund, co-investors will underperform. Published research does not settle the question, but it does frame it.

Fang, Ivashina and Lerner (2015). In "The disintermediation of financial markets: Direct investing in private equity," published in the Journal of Financial Economics (vol. 116), the authors studied two decades of direct private equity investments by seven large institutions. They found that direct investments outperformed public market indices, but that co-investments underperformed the funds they were made alongside, which the authors attributed to apparent adverse selection in the transactions available to co-investors. Solo transactions, where the institution invested on its own, outperformed fund benchmarks, and the authors pointed to the investors' ability to resolve information problems as an important driver.

Braun, Jenkinson and Schemmerl (2020). In "Adverse selection and the performance of private equity co-investments," also in the Journal of Financial Economics (vol. 136), the authors used a large sample of buyout and venture co-investments and compared them with the other deals in the same funds. In contrast to the earlier study, they found no evidence of adverse selection: gross return distributions for co-investments and other fund deals were similar. They also found that co-investments generally have lower costs to investors, and their simulations suggested that reasonably sized portfolios of co-investments significantly outperform fund returns on a net basis.

Lerner, Mao, Schoar and Zhang (2022). In "Investing outside the box: Evidence from alternative vehicles in private equity," in the Journal of Financial Economics (vol. 143), the authors used custodial data to study alternative vehicles, a category that includes co-investment and other non-main-fund vehicles. They reported that such vehicles reached about 40% of private equity commitments by 2017, that their average performance matched the broader private equity market but trailed the sponsoring partnerships' main funds, and that LPs with stronger past performance tended to invest in better-performing alternative vehicles, which the authors attributed largely to preferential access for top LPs.

What this means for a family office co-investor

The research points to three practical conclusions:

1. Selection is a variable, not a constant. Whether co-investments underperform appears to depend on the sample, the period, and which investors are in it. The Lerner et al. finding that better LPs get better vehicles suggests that access quality is itself unevenly distributed.
2. Fee savings are real but not sufficient. Lower fees improve net returns only if gross returns are not worse. A family office that accepts every sidecar it is offered is relying on the sponsor's selection.
3. Diversification within a co-investment program matters. The Braun et al. simulations concern reasonably sized portfolios of co-investments, not single deals. An office that makes two or three co-investments a year does not get the same statistical benefit.

What this means for GPs

A GP that can show, with data, that its syndicated deals have performed in line with its retained deals answers the question before it is asked. The SEC's 2020 Risk Alert findings make the same point from the regulatory side: examiners looked at whether disclosed allocation processes were actually followed, so a written, followed, and disclosed allocation policy can also serve as a marketing asset.

---

Part VI: Tax Considerations at a General Level

Tax shapes which vehicle a GP should offer. What follows is general description, not tax advice; outcomes depend on facts and jurisdiction.

Pass-through versus corporate holding. US sponsor SPVs are commonly partnerships for tax purposes, so income and gains flow through to the family's investing entities on a Schedule K-1. Families with non-US members, or tax-exempt entities such as private foundations, may need different treatment for the same deal, which is why parallel or blocker vehicles are offered alongside some SPVs.

Deductibility of the family office's own costs. For individuals, investment expenses under section 212 of the Internal Revenue Code are miscellaneous itemized deductions. The 2017 tax law suspended those deductions for 2018 through 2025, and Public Law 119-21 (enacted July 4, 2025) removed the end date, so the suspension now applies to taxable years beginning after December 31, 2017 without a sunset. This has raised interest in structuring the family office as a business whose expenses fall under section 162. In Lender Management, LLC v. Commissioner, T.C. Memo. 2017-246, the Tax Court held that a family office that managed investment LLCs for members of one family, and received a profits interest for doing so, was engaged in a trade or business. The analysis was fact-specific, and family relationships attract heightened scrutiny.

Profits interests and carried interest. Where a family office takes a profits interest in a vehicle it manages, section 1061 of the Code generally requires a holding period of more than three years, rather than one year, for gain on an applicable partnership interest to be treated as long-term capital gain.

Qualified small business stock. Direct investors in early-stage C corporations look at section 1202. For stock acquired after July 4, 2025, the statute now provides a 50% exclusion after three years, 75% after four years, and 100% after five years; the per-issuer cap is the greater of $15 million or ten times basis; and the issuer's aggregate gross assets limit is $75 million, with both thresholds indexed for inflation after 2026. Whether stock qualifies depends on the issuer meeting the statute's corporate, gross asset, and active business tests.

Private foundations co-investing with the family. Families that invest through or alongside a private foundation face the excess business holdings rules in section 4943. Broadly, a foundation's permitted holdings in a business enterprise are 20% of voting stock, reduced by the holdings of disqualified persons such as the family; this can rise to 35% where an unrelated party effectively controls the business, and holdings of no more than 2% of both voting stock and value are generally disregarded. A co-investment that makes sense for the family's taxable entities can create problems if the foundation participates too.

Cross-border families. Non-US offices investing in US operating businesses through partnerships may face effectively connected income and withholding issues, a common reason for blocker corporations.

The practical takeaway for GPs: a family office's tax profile can be more varied than a pension fund's, and a co-investment offering that only comes in one vehicle may exclude part of the family's capital.

---

Part VII: The Intelligence Implication — What GPs and Placement Teams Need to Know

Family offices can be attractive co-investors: patient capital, flexible decision-making, and, in many cases, an operating perspective from a family business. They are also highly varied, and the survey data shows their appetite moves with their own liquidity.

1. Qualify the office's program before pitching the deal. Is the office co-investment-led, sector-specialist, or a full direct platform? An office with a small investment team needs a different package and timeline from one that runs its own deals.

2. Ask about liquidity, not just strategy. The UBS series shows direct private equity allocations rising to 13% in 2021, falling back, and then being trimmed again for 2025 as exits slowed. Offices that recycle realizations into new directs will be selective when realizations are scarce.

3. Lead with your allocation policy. ILPA Principles 3.0 and the SEC's 2020 Risk Alert have made co-investment allocation a governance topic. Show the written policy, explain why each deal is being syndicated, and be ready to show how syndicated deals have performed relative to retained deals.

4. Be explicit about economics and expenses. State whether there is a fee or carry on the co-investment, how broken-deal and transaction costs are shared, and how any portfolio company fees are treated. The KKR broken-deal case remains a reference point for co-investors.

5. Respect the family office rule. A single family office relying on the exclusion generally cannot act as an adviser to unrelated families. A request to "bring its network" that would require it to organize and advise other investors can run into that limit.

6. Offer vehicle flexibility. Families may need to invest through taxable entities, trusts, foundations, or non-US structures. A co-investment that can accommodate parallel or blocker vehicles reaches more of the family's capital.

7. Match timelines to the team. BNY's 2025 survey lists time and understaffing among the top challenges in direct investing. A lean office may need earlier sight of the pipeline to meet a short co-investment window.

8. Map the people, not just the entity. In a dozen-person office, the decision on a co-investment may rest with a CIO, a principal, and an outside adviser. Knowing who holds those roles, and how long they have held them, is often the difference between a meeting and a declined email.

Wealthmetrica's family office directory is built to support that mapping, with firm-level and contact-level data drawn from regulatory filings and public sources; the data methodology page explains how records are compiled and refreshed. For GPs that also distribute through advisers, the RIA directory covers the registered firms that frequently sit alongside family offices as outsourced CIOs and co-investors.

---

Key sources

• UBS, Global Family Office Report 2025 (317 respondents; survey January to April 2025)
• UBS, Global Family Office Report 2026 (307 respondents; survey January to March 2026)
• J.P. Morgan Private Bank, 2026 Global Family Office Report (333 family offices, 30 countries)
• Citi Wealth, 2025 Global Family Office Report, press release of September 16, 2025
• BNY Wealth, 2025 Investment Insights: Single Family Offices Report (282 respondents)
• ILPA, Principles 3.0 (2019)
• SEC Office of Compliance Inspections and Examinations, Risk Alert: Observations from Examinations of Investment Advisers Managing Private Funds (June 23, 2020)
• SEC Press Release 2015-131, KKR broken-deal expenses (June 29, 2015)
• 17 CFR 275.202(a)(11)(G)-1, Family offices (Investment Advisers Act of 1940)
• Fang, Ivashina and Lerner, Journal of Financial Economics vol. 116 (2015)
• Braun, Jenkinson and Schemmerl, Journal of Financial Economics vol. 136 (2020)
• Lerner, Mao, Schoar and Zhang, Journal of Financial Economics vol. 143 (2022)
• KIRKBI, Blackstone and CPPIB, announcement of recommended offer for Merlin Entertainments plc (June 28, 2019)
• Four Seasons Hotels and Resorts, announcement of Cascade Investment affiliate stake acquisition (September 8, 2021)
• Internal Revenue Code sections 67, 1061, 1202 and 4943; Lender Management, LLC v. Commissioner, T.C. Memo. 2017-246

Wealthmetrica provides intelligence on investment advisers, family offices and the people who run them. This analysis draws on public regulatory filings, statutes and published research; it is not legal, tax, or investment advice.