Trust vs foundation: structural choice in 2026

Two legal traditions, one strategic decision

The choice between a trust and a foundation is, at its core, a choice between two competing legal traditions that have never fully reconciled. The trust is a product of English equity jurisdiction, traceable to the Statute of Uses 1535 and refined through centuries of Chancery court decisions. The foundation — the Stiftung, fondation, fundación — is a civil-law institution with legal personality, no beneficial owners in the trust-law sense, and a governance structure that looks more like a special-purpose company than a fiduciary relationship. In 2026, with CRS data exchange now covering 120 jurisdictions, BEPS Pillar Two beginning to bite at the entity level, and a generation of wealth owners confronting cross-border family circumstances that earlier structures never anticipated, the structural decision has become more consequential than at any point in the past two decades.

The practical stakes are significant. According to the Global Family Office Report 2024 published by UBS and Campden Research, approximately 38% of single-family offices reviewed their primary holding structure in the preceding 24 months, with jurisdictional review the most commonly cited driver. The shift is not merely cosmetic. Choosing the wrong vehicle — or placing it in the wrong jurisdiction — can produce unintended deemed-distribution events, loss of treaty access for underlying assets, personal liability for trustees or council members, and exposure to fraudulent conveyance challenges that strip away decades of asset protection planning.

Legal-system roots and the recognition problem

A trust does not exist as a legal person. It is a relationship: a settlor transfers assets to a trustee, who holds legal title and manages those assets for the benefit of identified or identifiable beneficiaries. Beneficial ownership and legal ownership are deliberately split. This split is the trust's defining feature and simultaneously its greatest source of international friction. Civil-law systems, which cover most of continental Europe, Latin America, the Middle East, and East Asia, have no equivalent concept of bifurcated ownership. When a French notaire or a German Notar encounters a trust deed, they are dealing with a concept their legal system was not designed to accommodate.

The Hague Convention on the Law Applicable to Trusts and on their Recognition, concluded in 1985 and in force from 1992, was intended to resolve this impasse. It requires signatory states to recognize trusts validly created under the law of a chosen jurisdiction, even if that recognition conflicts with the domestic legal tradition. The Convention currently has 14 contracting states, including the United Kingdom, Australia, Canada, Italy, Luxembourg, Malta, the Netherlands, and Switzerland. Notably absent are Germany, France, Spain, Austria, and virtually all of Latin America. The practical consequence is significant: a trust established under Jersey law and governed by the Trusts (Jersey) Law 1984 will be recognized in Luxembourg and Switzerland, but a Swiss court order enforced in Germany may require a complete recharacterization of the beneficial interest before German courts will act on it.

The Hague Convention on Trusts solved a theoretical problem elegantly while leaving the most commercially important civil-law jurisdictions outside its scope. For families with German or French situs assets, the convention provides comfort that stops at the border.

Foundations, by contrast, possess legal personality. A Liechtenstein Stiftung under the Personen- und Gesellschaftsrecht (PGR) of 1926, a Panama Private Interest Foundation under Law 25 of 1995, or a Dutch Stichting under the Burgerlijk Wetboek is a legal entity. It can own property, enter contracts, sue and be sued, and open bank accounts in its own name. Civil-law jurisdictions understand it intuitively. A Swiss court, a German tax authority, or a Dubai notary will process a foundation's constitutional documents without the conceptual friction that a trust deed generates. This recognition advantage is not marginal — in practice, it reduces legal costs, shortens banking due diligence timelines, and makes cross-border asset transfers considerably more straightforward.

Jurisdiction selection: the primary filter

For trusts, the leading jurisdictions in 2026 remain Jersey, the Cayman Islands, New Zealand (for its perpetuity advantages), and the Cook Islands (for asset protection). British Virgin Islands trusts remain common but face incremental reputational pressure from FATF grey-listing episodes affecting the broader Caribbean. For foundations, Liechtenstein continues to set the technical standard: the PGR foundation law has been continuously modernized, the jurisdiction is FATCA- and CRS-compliant, it participates in the European Economic Area, and its supervisory authority — the Financial Market Authority (FMA) — applies a proportionate but credible regulatory framework. Austria's Privatstiftung under the Privatstiftungsgesetz 1993 is well-suited for European Union-domiciled families but carries specific restrictions on founder control that limit its flexibility compared to Liechtenstein. The Netherlands Stichting, while lacking dedicated private wealth legislation, is used by sophisticated structures because of the Netherlands' extensive double-tax treaty network — 96 treaties in force as of 2025.

Tax neutrality: where the structures diverge materially

Tax neutrality — the principle that the structure itself should not generate a tax cost beyond what would arise if assets were held directly — is achievable in both trust and foundation form, but the pathways differ and the risk profiles are not equivalent.

A discretionary trust in a zero-tax jurisdiction such as the Cayman Islands or the British Virgin Islands is nominally transparent from a tax perspective but creates classification complexity in the home jurisdiction of the settlor and beneficiaries. In the United States, a foreign non-grantor trust is subject to the throwback rules under IRC Sections 665 through 668, which impose an interest charge on accumulated income distributed to US beneficiaries — a punitive mechanism designed specifically to neutralize the tax-deferral advantage of offshore accumulation. A US person who is treated as the grantor of a foreign trust under IRC Section 679 must report the trust annually on Form 3520 and Form 3520-A, with penalties for failure reaching USD 10,000 per violation or 35% of the gross reportable amount, whichever is greater. For non-US families with US-connected beneficiaries, these rules fundamentally alter the trust's utility.

The foundation's legal personality creates a different tax classification problem. In Germany, a foreign private foundation may be treated as a Zwischengesellschaft — an interposed entity — under the Außensteuergesetz (AStG), with its income attributed directly to the German-resident founder or beneficiaries if the passive income threshold is exceeded and the jurisdiction does not meet the substantial activity test. The 2021 revision to the AStG, implementing the EU Anti-Tax Avoidance Directive (ATAD) II, tightened these rules materially. Similar controlled foreign corporation (CFC) regimes apply in France under Article 209 B of the Code Général des Impôts, in the United Kingdom under Part 9A of TIOPA 2010, and across most OECD members.

The more fundamental challenge facing both structures in 2026 is BEPS Pillar Two. The OECD's Global Anti-Base Erosion (GloBE) rules apply to multinational enterprise groups with consolidated revenues exceeding EUR 750 million. Most single-family offices fall below this threshold. However, where family enterprises — operating companies, real estate platforms, or investment management entities — are owned through the structure, the Pillar Two qualified domestic minimum top-up tax (QDMTT) and income inclusion rule (IIR) may apply. The Subject-to-Tax Rule (STTR), targeted at source-country withholding on certain payments, applies at a 9% rate to interest and royalty payments flowing to low-tax jurisdictions. Neither trusts nor foundations in traditional offshore centers are immune from this if they hold operating assets or receive intragroup payments.

BEPS Pillar Two does not specifically target wealth-holding structures, but families with operating assets generating consolidated revenues above EUR 750 million must model the GloBE impact before 2026 year-end. The penalties for miscalculation are not theoretical.

One area where trusts maintain a specific advantage is treaty access for capital gains and dividend flows. A Cayman Islands trust that holds shares through a Luxembourg holding company can access the Luxembourg-source country treaty network because the Luxembourg entity is the treaty counterparty, not the trust itself. A foundation in the same structural position achieves the same outcome, but certain treaty definitions of 'resident' and 'beneficial owner' — as clarified by the OECD Model Commentary's 2017 revision — require the foundation to demonstrate substantive management and control in the treaty jurisdiction. The 2023 update to the OECD Commentary on Article 1 further tightened beneficial ownership analysis for conduit arrangements, making substance planning more demanding for both vehicles.

Governance flexibility and the protector mechanism

Governance is where the trust and foundation most visibly diverge — and where family offices often underestimate the long-term implications of their structural choice.

A trust's governance is determined primarily by its deed and the applicable trust law. The settlor can confer extensive reserved powers on themselves — the power to appoint and remove trustees, the power to direct investments, the power to add or exclude beneficiaries — but doing so creates risk. Under the rule in Re Vandervell's Trusts (No 2) [1974] Ch 269 and more recently analyzed in Webb v Webb [2020] EWHC 1489 (Ch), excessive settlor control can collapse the trust back into a bare trust or result in the trust assets being treated as remaining in the settlor's estate for tax purposes. Jersey's Trusts (Jersey) Law 1984, as amended, explicitly permits reserved powers under Article 9A, providing statutory clarity that the English common law does not offer with equal precision. The Cook Islands International Trusts Act 1984 (as amended in 2004) goes further, permitting the settlor to retain investment powers, the power to revoke, and the power to change governing law without the trust being considered a sham.

The protector — a third party appointed to oversee the trustee, typically holding the power to remove and replace trustees and to consent to certain trustee actions — has become a near-universal feature of discretionary family trusts established in the past 15 years. It serves as a check on trustee discretion without re-vesting control in the settlor. The protector concept has no direct analogue in civil-law foundation law, though the foundation council (Stiftungsrat in Liechtenstein, conseil de fondation in French-speaking jurisdictions) can be structured with internal checks: mandatory quorum requirements, reserved matters requiring supermajority council votes, and the designation of a supervisory board distinct from the management board.

The foundation's governance advantage is legibility. A foundation charter and bylaws are documents that any corporate lawyer, banker, or regulator can read and process. The charter specifies the foundation's purpose, the beneficiary designation methodology, the council composition, and the amendment procedure. In Liechtenstein, the charter is filed with the public register, while the bylaws — which typically contain the beneficiary details — remain private. This combination of public constitutional clarity and private beneficiary protection has made the Liechtenstein Stiftung the reference standard for privacy-conscious, compliance-compliant wealth structuring in continental Europe.

Family governance integration

Increasingly, family offices use the structural choice to reinforce broader family governance frameworks. A foundation's constitutional documents naturally accommodate a family council, a protector equivalent in the form of a 'beneficiary assembly,' and formal amendment procedures that require multi-generational consent. The Liechtenstein foundation law permits the designation of a Begünstigter (beneficiary) who has enforceable rights against the foundation, or a Ermessensbegünstigter (discretionary beneficiary) who does not — a distinction that mirrors the trust law categories of fixed and discretionary beneficiaries but is expressed in terms a civil-law notary will recognize.

Trusts, in common-law jurisdictions with mature trust cultures, can achieve equivalent family governance depth through a combination of the trust deed, a letter of wishes, and a standalone family governance charter. The letter of wishes is not legally binding but is taken seriously by professional trustees — particularly in Jersey and Guernsey, where the Royal Court has repeatedly affirmed that trustees should give weight to the settlor's expressed wishes without being bound by them. The non-binding nature of the letter of wishes is both a feature and a limitation: it preserves trustee discretion while allowing the settlor to communicate family values and succession priorities informally.

Asset protection: comparable strength, different mechanisms

Asset protection is frequently cited as a primary motivation for establishing either a trust or a foundation. The honest assessment is that both structures offer meaningful but not absolute protection, and the differences between them are more about mechanism than outcome.

The core asset protection principle for trusts is alienation: once assets are transferred to a trustee and the transfer is not voidable, the settlor's creditors generally cannot reach them. The critical variable is the fraudulent conveyance or fraudulent transfer analysis. In Jersey, Section 11 of the Trusts (Jersey) Law 1984 provides that a trust is not invalid merely because a settlor retains certain powers, but a transfer made with the intent to defraud creditors is voidable to the extent necessary to satisfy the creditor's claim. The Jersey law imposes a two-year limitation period from the date of transfer for creditor claims, with a 'discoverability' extension to five years in limited circumstances. The Cook Islands sets the limitation period at two years from the date of transfer or one year from the date the creditor knew or should have known of the transfer — one of the shortest windows in any asset protection jurisdiction globally.

For foundations, asset protection operates through the legal personality barrier. Once assets are transferred to the foundation, they belong to the foundation as a legal entity — not to the founder, not to the beneficiaries, and not to the foundation council members. A creditor of the founder cannot pierce the foundation merely by demonstrating a debt unless the fraudulent intent at the time of transfer can be established. Liechtenstein's asset protection framework for foundations was strengthened by the 2009 revision to the PGR, which introduced an explicit provision that a foundation's assets are not subject to claims of the founder's creditors where the founder has validly relinquished control. The practical limitation is identical to trusts: transfers made in contemplation of insolvency or with fraudulent intent will be unwound under Liechtenstein's avoidance provisions, and the relevant EU Insolvency Regulation (Recast) — Regulation (EU) 2015/848 — creates significant cross-border coordination obligations that affect both vehicles where EU creditors are involved.

Asset protection is not a destination; it is a trajectory. The strength of either structure depends on how long ago the transfer was made, what control was retained, and whether the jurisdiction's fraudulent conveyance window has closed. A trust or foundation established in the shadow of litigation provides almost no protection at all.

One area where foundations have historically claimed an advantage is banking and investment account management. A foundation's legal personality means it can establish accounts in its own name without requiring a complex trustee-beneficiary disclosure chain. Under the OECD's CRS — now implemented through the Common Reporting Standard and Multilateral Competent Authority Agreement framework covering 120 jurisdictions — both trusts and foundations are reportable entities. However, the practical reporting chain differs. A trust is reported through the trustee's home jurisdiction, with the settlor, protector, and all discretionary beneficiaries flagged as 'controlling persons.' A foundation is reported through its home jurisdiction, with the founder, the foundation council members, and the beneficiaries similarly flagged. Neither structure provides opacity under CRS, but the foundation's single-jurisdiction anchor tends to produce cleaner, more internally consistent reporting than a trust with trustees in one jurisdiction and assets in several others.

The parallel structure: bridging legal traditions

For families whose complexity genuinely spans both legal traditions — typically ultra-high-net-worth families with assets in common-law and civil-law jurisdictions, beneficiaries holding multiple nationalities, and business interests requiring treaty network flexibility — neither a pure trust nor a pure foundation resolves the full structural problem. The answer increasingly adopted by advisors in Zurich, Geneva, Luxembourg, and the Channel Islands is a parallel structure: a trust that holds the shares or beneficial interest of a foundation, or a foundation that owns a corporate trustee.

The trust-owning-foundation model places a discretionary trust (often Jersey-governed) at the apex of the structure, with the trust owning 100% of the shares of a Liechtenstein or Dutch foundation established as an intermediate holding entity. The foundation's legal personality provides the banking and contracting interface in continental Europe; the trust's discretionary structure provides the beneficiary flexibility and protector oversight that civil-law foundations handle less elegantly. The additional compliance cost of this approach — a second set of governance documents, two separate regulatory frameworks, consolidated CRS reporting that must reconcile both layers — is real. Annual administration costs for such structures at reputable institutional providers typically run between EUR 80,000 and EUR 200,000 depending on asset complexity and activity levels.

The foundation-owning-trustee model inverts this: a Liechtenstein foundation owns a Jersey or Cayman corporate trustee entity, which in turn administers a series of trusts for different family branches. This approach is used where the family's primary legal connections are continental European but specific asset lines — private equity fund interests, US real estate, common-law jurisdiction business assets — require trust-form holding for practical or regulatory reasons. The foundation's council then exercises the foundation's shareholder rights over the corporate trustee, achieving a degree of family oversight that exceeds what a standard protector arrangement provides.

Practical decision framework for 2026

The structural decision should be driven by a systematic analysis of five factors, applied in sequence rather than simultaneously. The first is the legal-system exposure of the primary asset base: where are the assets located, and which legal system governs property rights over them? If more than 40% of asset value is concentrated in civil-law jurisdictions without Hague Convention membership, a foundation has a strong prima facie advantage on recognition grounds alone.

The second factor is the citizenship and residency profile of the family. US persons in the structure — whether as settlors, beneficiaries, or trustees — trigger the IRC's foreign trust reporting regime and effectively require a US tax attorney to certify the structure annually. The third factor is the governance philosophy of the family: families that want a close functional analogue to a corporate board structure, with clear constitutional documents and defined council terms, will find foundations more natural. Families that prize flexibility, the ability to adjust beneficiary classes across generations without formal charter amendment, and the protector mechanism as a family-controlled check will generally prefer trusts.

The fourth factor is the timeline of the asset protection need. Establishing either structure in response to a known, crystallizing risk is almost certainly futile. The structure must be established, and assets transferred, sufficiently in advance that the applicable fraudulent conveyance limitation period has expired before any creditor claim materializes. The fifth factor — and one that will grow in importance through 2026 and beyond — is the BEPS Pillar Two substance profile. Where the structure holds operating assets generating revenues that could eventually cross the EUR 750 million GloBE threshold, or where intragroup payments expose the structure to the Subject-to-Tax Rule's 9% threshold, a substance analysis for both the trust jurisdiction and any foundation jurisdiction is not optional. Regulatory minimum staffing requirements, local office space, and documented management activities will increasingly determine whether the structure's tax position survives OECD-aligned review.

The trust-versus-foundation decision in 2026 is not a question of which vehicle is superior in the abstract. It is a question of which vehicle is superior given a specific family's legal exposures, governance preferences, tax profile, and asset protection timeline. The answer will differ for a German industrialist family with assets concentrated in Bavaria, for a Hong Kong-based family with US beneficiaries and Southeast Asian operating businesses, and for a Brazilian entrepreneur seeking protection of liquid assets while retaining flexibility for future investment. What none of these families can afford is the assumption that the structural choice made for the previous generation remains optimal for the next one.