Maxi Capitalist

A Family Office Is a Decision-Making Machine

2025-11-18T00:00:00Z

The common description of a family office is a portfolio with a staff attached. That framing is comfortable and wrong. It leads families to hire for markets and discover, years later, that their real problem was never returns. It was decisions.

A family office is a machine for making decisions about capital under uncertainty, and for keeping those decisions coherent as the people making them multiply and diverge. Judged that way, most of what matters has nothing to do with asset selection.

The problem it actually solves

A founder decides alone, quickly, with everything in their head. That works precisely as long as there is one founder. Add a spouse, three children, an in-law and a foundation, and the informal system that served one person becomes a source of friction for eight. The office exists to replace intuition that cannot scale with structure that can.

The founder's greatest asset is judgement. It is also the one thing that cannot be inherited.

The office is how judgement is converted into something transmissible: mandates, thresholds, review rhythms, and a written account of why things are done the way they are.

Three systems, not one

Underneath the label sit three distinct machines, and families get into trouble when they build one and assume they have all three.

Capital. Allocation, liquidity, reserves, reporting. The visible part, and the least likely to sink the enterprise on its own.
Governance. Who decides, within what limits, and how conflict is resolved. The load-bearing part, and the most often left implicit.
Continuity. Education, succession, and the slow work of preparing the next set of decision-makers before they are needed.

A family that indexes on the first and neglects the other two buys itself excellent quarterly reports and an unmanaged crisis on a ten-year fuse.

Why the private-bank analogy misleads

A private bank sells products and is paid on assets. A family office, done properly, is paid to say no: to unnecessary complexity, to correlated risk dressed as diversification, to the deal that is exciting rather than good. Its value is measured in mistakes avoided, which never appear on a statement. Confusing the two imports the wrong incentives into the one structure meant to be free of them.

What good looks like

A well-built office can survive a poor decade of markets because its purpose was never to win a decade. It was to make sure that no single decision, and no single death, could unmake the family's capital. It makes boldness survivable by making failure non-fatal.

That is a modest-sounding standard. It is also the one most fortunes fail.

Optionality Is the Ultimate Asset

2025-12-09T00:00:00Z

Prediction gets the attention. Optionality wins the decades. The difference is temperamental as much as technical: the forecaster tries to be right about the future, while the allocator with optionality arranges never to depend on being right.

The asset you cannot see on a statement

Optionality is the preserved ability to act when you choose to, rather than when you are forced to. It does not appear as a line item. It shows up only in its absence, at the exact moment a forced seller meets a patient buyer and discovers which one they are.

The purpose of reserves is not return. It is the right to refuse a bad price.

You buy optionality with things that look like drag in good years: liquidity that earns little, fixed costs kept deliberately low, leverage left unused, governance that does not force annual action. Each is a small, visible cost paid for a large, invisible freedom.

Convexity, quietly

The reason this compounds is asymmetry. An owner who cannot be forced to sell keeps the full upside of patience while capping the downside of panic. Over enough cycles, the advantage is not that they pick better. It is that their mistakes are recoverable and their opportunities are affordable when everyone else's are not.

In good years	In bad years
Looks conservative, slightly behind	Buys what others must sell
Pays for unused flexibility	Flexibility becomes the only thing that matters

The discipline

Optionality decays. Held too long without use, it becomes mere timidity, and reserves that are never deployed are just a slower way of losing. The skill is not hoarding freedom but spending it rarely and decisively, then rebuilding it. Preserve the option; do not fall in love with holding it.

The owners who last are not the boldest or the most cautious. They are the ones who kept the freedom to be either, and used it on their own schedule.

The Architecture of Capital

2026-01-14T00:00:00Z

Ask how a fortune was made and you will hear about an asset: a company, a property, a position taken early and held. Ask how a fortune was kept and the answer is always duller, and always structural. Assets create wealth. Architecture decides whether it survives contact with tax, transition and time.

Selection is overrated; structure is underrated

Selection is where attention goes because it is where the story is. But the same assets, held inside different structures, produce radically different outcomes across a generation. One arrangement passes cleanly to heirs; another triggers a forced sale to pay a tax bill. One survives a divorce or a creditor; another does not. The assets were identical. The architecture was not.

A structure is a decision about the future made while you still have the freedom to make it.

What the architecture is made of

Entities. Holdcos, trusts, foundations and the boundaries they draw around risk and ownership.
Jurisdiction. The law each entity answers to, and the political weather it stands in.
Governance. Who controls what, and how control passes.
Liquidity. Whether the structure can meet its obligations without dismantling itself.

None of these is an investment decision. All of them determine what your investment decisions are ultimately worth.

The test of good architecture

Good structure is legible, boring and acceptable: understandable to the family, unremarkable to a bank, and defensible to a regulator. Cleverness is a liability here. The arrangement that only its architect understands is the one that fails when the architect is gone, which is precisely when it is needed.

Build for the moment you will not be in the room. That is the only real test, and structure is the only thing that passes it.

Where Should a Family Office Live?

2026-08-06T00:00:00Z

Ask where a family office should be based and you will get a list of countries. That is the first mistake. The question is framed as a single destination, and framed that way it quietly assumes the family should pin its structure, its people and its future to one flag. The families who get this right rarely do that. They treat location as a set of decisions to be made separately, and they keep the freedom to change their minds.

The backdrop is real movement, not theory. Henley & Partners projected a record 142,000 millionaires relocating in 2025, with the United Kingdom forecast to lose 16,500 of them, the largest outflow it has ever recorded, and the UAE gaining a record 9,800. Capital is voting with its feet. The interesting work is understanding what it is voting on.

What actually drives the decision

Strip away the brochures and the choice comes down to a stable set of criteria. They are knowable, and they can be weighted against a family's own priorities.

Tax exposure. Not headline income tax, but the full picture: income, capital gains, dividends, and above all inheritance and exit taxes, which do the real damage across a generation.
Rule of law and enforceability. Whether the courts are independent, whether contracts and structures hold, and whether a trust or foundation created there is recognised where it matters.
Political and monetary stability. The probability that the rules, the currency and the government still look familiar in twenty years.
Personal safety and quality of life. The part families underweight until they live somewhere, then never stop weighing.
Financial ecosystem and talent. Depth of private banks, advisers, lawyers and hireable investment professionals. The Global Financial Centres Index is the standard objective measure.
Banking acceptance and reputation. Whether a structure from that jurisdiction opens accounts and clears correspondent banking without friction, and how it reads to a regulator or counterparty.
Substance and operating cost. What it genuinely costs to run a compliant office there, and the real staff and premises required to hold the tax treatment.
Succession treatment. Forced heirship, inheritance tax, and whether the family's chosen governance survives a death intact.
Mobility and family access. Residence and visa pathways for the principals and the next generation.
Regime durability. The most underrated criterion by far: how likely the deal is to still exist when you need it.

That last point deserves its own weight, because the recent record is brutal on anyone who treated a tax regime as permanent. The United Kingdom abolished its 226-year-old non-domiciled regime on 6 April 2025, replacing it with a four-year window and, for many, full exposure to inheritance tax. Portugal closed its Non-Habitual Resident regime to new applicants. Italy's flat tax on foreign income, introduced at 100,000 euros in 2017, was doubled to 200,000 in 2024 and raised again to 300,000 for 2026. Several Swiss cantons, including Zurich and Basel, abolished lump-sum taxation by referendum years ago. The lesson is not that these places are bad choices. It is that any single choice is a depreciating asset.

The objective landscape

Measured against those criteria, a short list of hubs does most of the work in serious conversations. None of them wins on every axis, which is the point.

Hub	Personal tax	Legal system	FO viable from	Ecosystem (GFCI 37)	Watch
Dubai (DIFC)	No income or capital gains tax	English common law, own courts	~USD 30-50m	12th, leading in region	Office entity itself taxed at 9%; 0% needs substance
Abu Dhabi (ADGM)	No income or capital gains tax	English common law, applied directly	~USD 30-50m	Regional top tier	Smaller ecosystem than Dubai; strong on foundations
Singapore	No capital gains tax	Common law, highly rated	S$20-50m in the fund	4th	13O/13U thresholds tightened in 2025; incentive sunsets 2029
Switzerland	Lump-sum (forfait) available	Civil law, very stable	~USD 150m	Zurich, Geneva top 20	Not all cantons offer forfait; higher cost base
Hong Kong	No capital gains tax	Common law	Comparable to Singapore	3rd	Political and China-exposure risk to weigh
Italy (Milan)	300k flat tax on foreign income	Civil law, EU	Residence play, not FO hub	Milan mid-tier	Flat tax raised twice in two years

A few honest nuances the table cannot hold. The Gulf leads on tax, personal safety and cost, and the UAE has removed real friction: a single family office in the DIFC no longer needs to register as a designated non-financial business, and the centre now runs a dedicated Family Wealth Centre. But the family office company there is a taxable service entity at 9 percent, not a tax-free wrapper, and on the standard rule-of-law and political-rights indices the region scores lower than Switzerland or Singapore. That is a trade a family should make with open eyes, not by accident.

Switzerland and Singapore sit at the opposite corner: deep ecosystems, strong courts, decades of predictability, at a higher cost and, in Singapore's case, after a visible tightening of the family-office rules following its 2023 money-laundering scandal. Hong Kong offers a first-class common-law financial centre with a discount that reflects exactly the risk you are being paid to take. Europe's flat-tax jurisdictions are residence plays for the principals, not homes for the office itself.

The unbundled alternative

Here is the move most advisers will not lead with, because it sells fewer relocations: do not relocate the family office as a single unit at all. Unbundle it.

A family office is really three layers that happen to be discussed as one. There is the ownership layer, where the assets legally sit, which rewards a stable, boring, well-recognised structure jurisdiction such as a Liechtenstein or Jersey foundation or a Singapore trust. There is the management layer, where decisions are actually made and the team has substance, which rewards ecosystem and talent. And there is the residence layer, where the principals physically live, which rewards tax and lifestyle. Nothing requires these three to share a country. Forcing them to is how families end up over-optimising one axis and quietly failing the others.

Then treat jurisdiction the way you treat a portfolio: acquire optionality and hold it in reserve. A second residence right in a different bloc, obtained before you need it. A standby structure, drafted and dormant, that can be activated in weeks rather than quarters. The families who moved fastest and cheapest out of the UK in 2025 were not the ones who reacted in April. They were the ones who had arranged the option years earlier and simply exercised it. Optionality, again, is the asset. Relocation is just one way to spend it.

None of this is a loophole, and it is not free. Unbundling raises the two hardest questions in cross-border planning: substance, because a tax treatment you cannot defend with real people and real activity is a liability waiting to be assessed; and management and control, because the office is the entity most exposed to being dragged into the tax net of wherever its decisions are genuinely taken. Add controlled-foreign-company rules, common reporting standard transparency, and the plain cost of running more than one seat, and the unbundled model is only for families with enough scale and enough governance to carry it. But for those families, it answers the real question, which was never "which country," but "how do we never again be trapped by one."

The point

Jurisdiction is not a destination you arrive at. It is a set of exposures you manage. The winners over the next decade will not be the families who picked the best country in 2025, because 2025's best country is already being re-priced by its own parliament. They will be the ones who arranged, quietly and in advance, never to depend on the answer.

The Virtual Family Office

2026-08-07T00:00:00Z

The family office used to be a building full of people. Increasingly it is a small coordinating core wired to a network of specialists and a single software platform. This is the virtual family office, and it is the fastest-growing shape the industry has, because it lets families with tens of millions rather than hundreds of millions run something that looks and behaves like a real office. The temptation is to see it as a discount. That is the mistake this piece is about. A virtual family office is not a cheaper family office. It is the same set of responsibilities distributed differently, and the distribution moves the hardest risk to a place most families do not watch.

What a virtual family office actually is

Strip away the marketing and a virtual family office is three things joined together. A coordinating core, usually one to three people or, at the smallest scale, the principal and a trusted chief financial officer. A network of external specialists, brought in as needed: an investment adviser, a tax firm, legal counsel, a custodian, an accountant, sometimes a concierge. And a data spine, a single reporting and collaboration platform that is meant to be the one place where the truth about the family's position lives.

The industry calls this hub-and-spoke, and the phrase is more than a diagram. It describes where the value sits. In a traditional office the value is in the team. In a virtual one the value is in the orchestration and the data: whether the core can direct a dozen outside providers coherently, and whether the platform actually consolidates what those providers produce into one honest picture. Get those two things right and a family running eighty million dollars can have institutional-quality oversight for a fraction of the cost. Get them wrong and you have a family paying a dozen invoices for a picture that is still fragmented.

The right way to think about it, and the way the better providers now frame it, is that a virtual family office is an operating model, not a product. You do not buy one. You design one: you define decision rights, service scope and data architecture first, and you choose vendors second. Families who buy the software first and think about governance later end up automating their confusion.

The legal structure underneath

Virtual does not mean informal, and it does not mean unstructured. Almost every serious virtual family office still sits inside a legal wrapper, typically a company or management entity owned by the family's trusts or holding structure. What is virtual is the staffing and the premises, not the entity.

The wrapper matters for a reason that catches families by surprise: regulation. In the United States, a family office that manages only one family's money is generally excluded from registering as an investment adviser under the family office exclusion, provided it is wholly owned and controlled by the family and advises no outsiders. Structure it carelessly, let it advise a cousin's separate wealth or take an outside client, and the whole office can tip into being a regulated investment adviser, with the cost and disclosure that implies. The virtual model does not remove this question. If anything it sharpens it, because a lean office is more tempted to share its lean infrastructure with a friendly family to defray cost, which is exactly the move that breaks the exclusion.

So the structural work is real even when the office is light. You still need an entity, you still need it owned and controlled correctly, and you still need the service relationships with your outside providers to be documented as arm's-length engagements rather than an informal set of favours. Where the office charges the family's fund a management fee, as is common in Asia and the Gulf, that fee has to be genuinely arm's-length and supported by transfer-pricing logic, or the tax authority will unwind it. The lightest office in the world still has to be a real one on paper.

How the parties actually work together

This is where virtual offices succeed or quietly fail, and it has nothing to do with technology. The question is control. In a traditional office, control is exercised through employment: the CIO works for the family, and the family can hire, direct and fire. In a virtual office, control has to be exercised through design, because none of the specialists work for the family exclusively. They have other clients, their own incentives and their own view of where their responsibility ends.

That means three things have to be written down that an in-house team would carry in its head. First, decision rights: who can commit the family to an investment, who signs, who is merely consulted, and what size of decision escalates to the principal or a committee. Second, accountability lines: which provider owns which number, so that when the consolidated report is wrong there is one throat to hold rather than a circle of advisers each blaming the next. Third, coordination: someone in the core has to own the seams between providers, because the failures in a virtual office almost never happen inside a provider's work. They happen in the handoffs, where the tax firm assumes the custodian reported the cost basis and neither of them did.

The uncomfortable truth is that a virtual family office asks more of the family's governance, not less. The office is cheaper because you removed the staff. The staff were also the people who quietly held the whole thing together. In a virtual model that connective work does not disappear. It moves to the coordinating core and to the documents, and if neither is strong enough, the family has bought fragmentation and called it efficiency.

Compliance: you can outsource the work, not the responsibility

The most dangerous sentence in a virtual office is "our provider handles that." Providers handle tasks. They do not absorb the family's obligations. Anti-money-laundering and know-your-customer checks, sanctions screening, automatic exchange of information under the Common Reporting Standard, the newer crypto-asset reporting rules: these attach to the structure and its beneficiaries, and a fragmented set of vendors makes them harder to satisfy, not easier, because no single party sees the whole picture.

A virtual office therefore needs someone, in the core or on retainer, whose job is to own compliance across the network rather than assume each provider covers their own corner. The failure mode is specific and common: every provider is individually compliant, and the family as a whole is not, because an account in one place and a structure in another were never reconciled against the same reporting obligation. Regulators do not accept "distributed responsibility" as a defence. The lighter the office, the more deliberate this oversight has to be.

The risk the model is really about: IT and cyber

Here is the part most families underweight, and it is the reason a virtual family office should be understood as a risk decision as much as a cost one.

A family office is already an unusually attractive target: concentrated wealth, and a trove of the most sensitive data imaginable, from financial statements to passports to health records, held by an organisation that is typically smaller and less defended than the banks it deals with. Now consider what the virtual model does to that target. It replaces one defended perimeter with a web of external providers and a set of connected platforms. The attack surface is no longer the office. It is the entire vendor network, plus every application quietly authorised to touch the data.

This is not theoretical. Research consistently finds that almost every organisation works with at least one vendor that has been breached in the last two years, and the family-office studies are blunt about it: the vendor network is now the attack surface. The most instructive recent breaches did not involve the family office being hacked at all. An application connected to its systems was, an integration nobody remembered granting, and the data walked out through a door the family did not know was open. In a virtual office, where the whole point is to connect many providers through a shared platform, that class of risk is not a side effect. It is the design.

The threats are ordinary and effective: phishing and business email compromise, social engineering now sharpened by deepfakes, ransomware, and above all third-party compromise. Family businesses report these at high rates, and yet fewer than half describe their cyber posture as robust, and a large share rely on basic controls while lacking the advanced ones that actually matter here: vendor governance, identity and access management, and a rehearsed incident-response plan. The stakes are not abstract. A large majority of firms say a successful breach would trigger direct loss of assets or withdrawal of trust, and for a family the currency is reputation, which takes decades to build and an afternoon to lose.

So the discipline a serious virtual office demands is specific. Multi-factor authentication and strict access controls everywhere, on the principle that no provider and no application gets more reach than its task requires. A real inventory of every integration and every third party that touches the data, reviewed regularly, because you cannot govern what you have not listed. Vendor due diligence that examines each provider's own security posture before onboarding, not after a breach. Data-centric encryption, so that exfiltrated data is useless. And an incident-response plan that names, in advance, who does what across IT, legal, finance and communications, rehearsed rather than filed. A virtual office that has not done this work has not saved money. It has borrowed it from a future it has not insured against.

Virtual versus a traditional single-family office

The honest comparison is a set of trade-offs, not a verdict.

On cost, the virtual model wins decisively. A fully staffed single-family office commonly runs from one to several million dollars a year, driven by salaries; a virtual office can run from tens of thousands to a few hundred thousand. That difference is what has widened the range of families for whom an office of any kind makes sense, from the traditional threshold of hundreds of millions down to the low tens.

On control and alignment, the traditional office wins. An in-house team serves one family, sits on the same side of the table, and can be directed and held accountable directly. A network of providers, however good, serves many masters and must be governed through contracts and coordination rather than loyalty. Alignment in a virtual office is manufactured, not given.

On talent and continuity, it is genuinely mixed. A virtual office can rent expertise a mid-sized family could never afford to employ full-time, and can swap a weak provider without the trauma of firing a trusted employee. But it also concentrates knowledge in the coordinating core and the platform, which is its own key-person and key-system risk, and it lacks the deep institutional memory that a long-tenured in-house team accumulates.

On confidentiality, the traditional office wins on paper and the virtual office wins on modern reality. Fewer people inside means fewer internal leaks; but more external connections means more exposure, and in an era where the breach comes through the vendor, the virtual model's wider surface is the greater practical risk. The right conclusion is not that one is private and the other is not. It is that the virtual office trades a small, controllable internal exposure for a large, distributed external one, and must spend on security what it saved on salary.

Where the virtual model fits, and where it fights the rules

Jurisdiction is where the virtual idea meets its sharpest constraint, and it is widely misunderstood. The most attractive family-office regimes are increasingly built on substance: they give you favourable tax treatment in exchange for genuine local presence. Singapore's Sections 13O and 13U require local investment professionals, minimum local spending and real capital deployment; the Qualifying Free Zone Person regime in the DIFC and ADGM in the UAE demands substance too. These regimes are designed to attract offices that are physically there. A pure virtual office, whose entire premise is minimal local staff and no real premises, sits in direct tension with them.

That tension resolves in one of two ways, and the distinction matters. Either the family runs a genuinely light office in a jurisdiction that does not demand substance for the treatment it wants, accepting more modest tax benefits in return for the lean model. Or, more commonly among the sophisticated, the family stops conflating two different questions. Where the office coordinates from and where the capital is structured are separate decisions, a theme explored in Where should a family office live. A family can keep a lean coordinating core wherever it is convenient while its fund and its tax structure sit in a substance-based regime that is staffed to meet the local test, with the two connected by proper service agreements. In that reading the virtual model is not an alternative to Singapore or the UAE. It is an operating layer that can sit on top of them, provided the family is honest about the fact that the favourable regime still has to be genuinely inhabited by someone.

The families who get this wrong try to claim a substance-based benefit while running a shell, and discover that regulators and tax authorities have grown very good at telling the difference. The families who get it right treat the virtual office as what it is: a way to distribute the work, layered onto a structure that still has to be real where it counts.

The point

A virtual family office is a structure decision wearing the costume of a cost decision. What you are really choosing is to hold the same responsibilities with fewer of your own hands, which means the parts that used to be carried invisibly by employees, the governance, the coordination, the compliance and above all the security, now have to be designed and paid for explicitly. Do that work and the virtual model is one of the most powerful developments in private wealth in a generation, putting real institutional capability within reach of families who could never have staffed it. Skip it, and you have not built a lighter family office. You have unbundled a heavy one and left the load-bearing pieces out.

The Best-Known Family Offices

2026-08-08T00:00:00Z

Everyone in this world has heard of a few family offices, and almost nobody can tell you how they are built. That is a shame, because the famous ones are the best textbook the field has. They are famous mostly for their scale, but scale is the least interesting thing about them. What they actually show, once you look past the numbers, is that there is no such thing as the family office structure. There is a small set of very different structures, and the best-known names are the clearest example of each.

A useful fact to hold before the tour: among the largest offices, roughly two thirds are single-family offices, around a fifth are family-controlled holding companies with a dedicated investment arm, and the rest are hybrids. Those are not three sizes of the same thing. They are three different answers to the same question, which is how a family keeps control of its capital while letting professionals run it. Every office below is one of those answers, made visible.

The concentrated stake with a diversification arm: Walton Enterprises

The largest family office in the world is the least glamorous. Walton Enterprises, based in Bentonville, Arkansas, runs the wealth of Sam Walton's descendants, estimated at north of two hundred billion dollars, which makes it larger than most sovereign wealth funds. Its structure is the template for every founding family that keeps its operating company.

At the centre sits the anchor: the family's roughly forty-four percent stake in Walmart, held with the voting control that keeps the company in family hands. Around that anchor, a separate vehicle, the Walton family's investment arm, diversifies the rest into public equities, low-cost funds, Treasuries and direct positions, so the family is not solely exposed to a single retailer. Layered on top are real estate, community and urban development, trusts, tax, legal, philanthropy and administration. The office is the operating system for a fortune whose gravitational centre is still the company that created it.

The lesson is that when the fortune is a control stake in a great business, the family office is built around protecting that stake, not replacing it. Everything else is a satellite.

The founder who diversified away from the source: Cascade Investment

If Walton is the office built to hold the company, Cascade Investment is the office built to escape it. Cascade, which manages Bill Gates's personal wealth, is probably the best-known family office of all, and its whole story is diversification. Run for decades by Michael Larson through an entity that also oversees the Gates Foundation's endowment, it methodically sold Microsoft stock and rebuilt the fortune in durable, cash-generating businesses: railways, waste management, hospitality, energy and, most famously, farmland, enough to make Gates one of the largest private farmland owners in the United States.

The structure here is a professionally staffed single-family office, reported to run to around a hundred people, deliberately low-profile, optimised to turn concentrated technology equity into a diversified portfolio of boring, resilient assets. It is the textbook case of a founder refusing to let his wealth remain a bet on the company that made him.

Cascade also teaches a quieter lesson about fame. Public reporting about its chief investment officer some years ago put a spotlight on the office and its people, a reminder that when a family's name is attached to a small, powerful organisation, the reputational and key-person risk is real. Most families conclude from examples like this that the right amount of publicity is none.

The principal-led venture vehicle: Bezos Expeditions

Jeff Bezos's office, Bezos Expeditions, is a different animal again. Where Cascade preserves and Walton protects, Bezos Expeditions ventures. It is a lean, principal-led single-family office, run since its founding by a small team, whose anchor remains Amazon stock but whose activity is direct investing at the frontier: aerospace through Blue Origin, media through The Washington Post, and a stream of bets on artificial intelligence, robotics, biotech and space.

This is the model many technology founders have adopted, Sergey Brin's Bayshore Global and Eric Schmidt's offices among them: the family office as a personal venture arm, moving quickly on conviction, treated less as a vehicle for preservation than as a way to keep building. Its structure is minimal on purpose. The value is in the principal's judgement and access, not in a large institution, which is exactly why decision-making stays close to the founder rather than delegated to a committee.

The publicly traded family holding: Exor

Cross the Atlantic and the shape changes completely. The Agnelli family, who founded Fiat in Turin in 1899, run their capital through Exor, a holding company listed on the Amsterdam exchange with a net asset value of roughly thirty-plus billion euros. Exor is, in effect, a publicly traded family office. It owns controlling or significant stakes in Ferrari, Stellantis, CNH, Philips, The Economist and Juventus, and it invests in venture through a dedicated arm.

The instructive part is the control mechanism. The family sits above the listed company through a private holding company, Giovanni Agnelli B.V., and uses loyalty and special voting shares so that a minority of the economics, around fifty-five percent, translates into a supermajority of the votes, roughly eighty-four percent. The family therefore commands the empire while sharing its economics with public shareholders, and it does so in full view. John Elkann, a great-grandson of the founder, runs it. This is how a European industrial dynasty stays in charge of a diversified group across generations without owning all of it, and without hiding.

The foundation at the very top: Investor AB and the Wallenbergs

Sweden's Wallenberg family solves the same problem, control across generations, with a different and even more durable device. They have controlled the listed investment company Investor AB since 1916, holding stakes in Atlas Copco, AstraZeneca, SEB, Saab and Ericsson. But the family does not sit directly on top. A set of family foundations does.

Because foundations cannot be sold, inherited away or seized, placing them at the apex of the structure makes control effectively perpetual: it outlives every individual heir, and no single family member can cash out the empire. The family holds a majority of the votes through a minority of the capital, exactly as the Agnellis do, but the top of the Wallenberg structure is an institution rather than a person. It is the purest expression of the oldest idea in this field: separate ownership from control, and make the control unsellable. Compare it with Exor and you have the two great routes to permanence, the family holding with loyalty shares, and the foundation at the top.

The hedge fund that became a family office: Soros and the converts

Some of the most famous names became family offices by subtraction. George Soros's operation, once the most watched hedge fund in the world, returned outside investors' money and now runs only the family's capital as Soros Fund Management. Steven Cohen's did the same after his firm's troubles, becoming a family office before later reopening to outside money.

The structure barely changes when this happens: the trading floor, the analysts and the systems stay. What changes is who the money belongs to and, crucially, the regulation. An investment firm that advises only one family can generally step outside the rules that govern one managing outsiders' money. This is the family office as the endgame of a great investor who no longer needs, or wants, other people's capital and the obligations that come with it.

The original, which became an industry: Rockefeller and Bessemer

Finally, the templates. The modern family office traces back to the office John D. Rockefeller built to manage his fortune, the ancestor of what became Rockefeller & Co and later Rockefeller Capital Management. Bessemer Trust grew the same way, out of the Phipps family's Carnegie Steel wealth. Both began as a single family's private office and, over time, opened their doors to serve other families, turning into multi-family offices and wealth managers in their own right.

This is the most common lifecycle in the field: a single-family office professionalises to the point where its capability exceeds one family's needs, and it becomes a business. The famous multi-family offices are, almost always, someone's single-family office that grew up.

What the famous names have in common

Step back and the through-line is clear. Structure follows the origin of the fortune and the intent behind it. A control stake in a great company produces a Walton. A founder wanting out of his own stock produces a Cascade. A builder who cannot stop building produces a Bezos Expeditions. A dynasty determined to stay in command across centuries produces an Exor or a Wallenberg. A great investor who is done with outside money produces a Soros. And any of them, professionalised far enough, produces a Rockefeller.

Two things recur in every case. The first is that the hard problem is never investment selection; it is control across generations, solved with holding companies, loyalty shares, foundations and trusts. The second is that fame is a cost. The offices that court attention, or have it forced on them, carry reputational and key-person risk that the quiet ones avoid, which is why the overwhelming majority of serious family offices you will never hear of at all. The most instructive thing about the famous few is that they wrestle with exactly the same structural questions as every anonymous office a thousandth their size. The scale changes. The architecture does not.

What a Trust Actually Is

2026-08-09T00:00:00Z

A trust is the most misunderstood instrument in private wealth, and also one of the most powerful. People imagine a bank account with rules, or a vault, or a tax trick. It is none of those. A trust is a relationship: an arrangement in which one person hands assets to another to hold and manage for the benefit of a third. That simple split, between the person who owns something on paper and the people it is really for, is what makes a trust do everything it does.

This series walks through the trust structures a family office actually uses, one idea at a time. We start with the foundations, because every later decision, dynasty planning, asset protection, choosing a jurisdiction, rests on getting these right.

The four roles

Every trust has three essential roles and often a fourth.

Settlor
creates & funds it Trustee
holds legal title Beneficiaries
hold the benefit

The settlor gives assets to the trustee to hold for the beneficiaries. A protector may oversee the trustee.

The settlor creates the trust and transfers assets into it. The trustee takes legal ownership and must manage those assets under a strict duty to act in the beneficiaries' interest, following the trust deed. The beneficiaries are the people the trust exists for. And the protector, increasingly common in family structures, is a person or committee who oversees the trustee and can hold reserved powers, such as replacing the trustee or vetoing certain decisions.

The magic is in the split between the trustee and the beneficiaries. The trustee owns the assets in the eyes of the law but cannot use them for himself. The beneficiaries enjoy the assets but do not legally own them. That separation of legal ownership from benefit is the entire engine of a trust.

How it works in practice

The rules live in the trust deed, the document that sets out who benefits, on what terms, and how much discretion the trustee has. That last point matters more than any other operational detail.

In a fixed trust, the beneficiaries have defined entitlements, half to one child, half to another. In a discretionary trust, the trustee decides who receives what and when, guided but not bound by the settlor's wishes. Almost every serious family trust is discretionary, because discretion is what gives the structure its flexibility and much of its protection: if no beneficiary has a fixed right to the assets, there is less for a creditor, an ex-spouse or a tax authority to attach.

Because the trustee holds real power, families guide them with a letter of wishes, a private, non-binding note explaining how the settlor would like discretion exercised. It steers without creating legal entitlements. We return to it in Part 5.

The distinction that changes everything: revocable versus irrevocable

If you remember one thing from this piece, make it this. Whether a trust is revocable or irrevocable determines almost everything it can and cannot do.

A revocable trust can be changed or unwound by the settlor at any time. Because the settlor keeps that control, the law treats the assets as still belonging to them. That makes a revocable trust useful for organising an estate and avoiding probate, but it offers no protection from creditors and no estate-tax benefit. What you can take back, the world can still reach.

An irrevocable trust cannot be freely undone. The settlor genuinely gives up control. In exchange for that surrender, the assets can leave the settlor's estate, escape future estate tax, and sit beyond the reach of later creditors. Everything a family office wants a trust to do, protect wealth, move it out of a taxable estate, hold it across generations, requires an irrevocable trust.

	Revocable trust	Irrevocable trust
Settlor keeps control	Yes	No
Assets leave the estate	No	Yes
Creditor protection	None	Strong
Estate-tax benefit	None	Yes
Typical use	Probate avoidance, organisation	Wealth preservation, protection, succession

The trade-off is stark and unavoidable: control or protection, not both. You cannot keep the right to take the money back and also claim it is out of your reach. Every family that uses trusts seriously eventually accepts this bargain, giving up control to gain everything else. The rest of this series lives on the irrevocable side of that line.

One more thing before we go on

The trust is a creature of common law, the legal tradition of the United States, the United Kingdom and the offshore centres built on it. Much of the world, continental Europe, Latin America, the Middle East, runs on civil law, where the trust does not neatly exist. Those families reach for a different tool that does the same job through a different mechanism: the foundation.

That is where we go next. In Part 2, trust versus foundation: two answers to the same problem, and how to know which one a family should be using.

Trust or Foundation?

2026-08-10T00:00:00Z

In Part 1 we said the trust is a common-law invention. That leaves a practical problem for a large share of the world's wealthy families. If you grew up under the civil law of France, Germany, Switzerland, Brazil or the Gulf, the trust is a foreign concept your own legal system barely recognises. For you, the equivalent tool is the foundation. Understanding both, and the line between them, is the difference between forcing the wrong structure onto a family and giving them one their own law and advisers actually understand.

Two mechanisms, one job

A trust, as we saw, is a relationship. There is no new entity; a trustee simply holds assets for beneficiaries. It has no legal personality of its own.

A foundation is the opposite in form and identical in purpose. It is a separate legal entity, like a company, but with a crucial difference: it has no shareholders and no owners. It owns its assets in its own name, and it exists to serve a purpose or a class of beneficiaries set out in its charter, governed by a council rather than a board answering to owners.

So the trust splits ownership from benefit through a person, the trustee. The foundation dissolves ownership entirely into an entity that answers to no one but its own charter. Both achieve the same end: assets are held, governed and passed on, without any individual owning them outright.

	Trust	Foundation
Legal nature	A relationship, no separate entity	A separate legal entity
Owner of the assets	Trustee holds legal title	The foundation itself
Run by	Trustee (with optional protector)	A council
Legal tradition	Common law	Civil law
Familiar to	US, UK, offshore families	European, Latin American, Gulf families
Typical uses	Succession, protection, dynasty planning	Holding, succession, philanthropy

When a family chooses which

The honest answer is that it usually comes down to background, not a grand legal analysis.

Common-law families, and their banks and advisers, are fluent in trusts. They think in settlors and trustees, and their courts have centuries of case law to draw on. For them the trust is the natural choice.

Civil-law families often find the trust alien and, worse, find that their home courts and tax authorities treat it with suspicion or refuse to recognise it. A foundation, which looks and behaves like the company structures they already know, sits far more comfortably. It also carries a psychological advantage: a founder can watch a foundation own the assets, with a council and rules, rather than having to trust an individual trustee to hold them.

There is a functional split too. Foundations are especially favoured for holding a family's companies and for philanthropy, where an enduring, self-governing entity with a stated purpose is exactly what is wanted. Trusts are the sharper instrument for discretionary succession and for asset protection, which we come to in Part 4.

The jurisdictions that matter

Foundations are not all the same, and where you establish one shapes what it can do.

Liechtenstein is the classic home of the private foundation, the Stiftung, with a century of law behind it and deep expertise in serving international families. Panama built a large industry around its Private Interest Foundation, valued for flexibility and cost. Increasingly, the United Arab Emirates has become a serious contender: the foundation regimes in the ADGM in Abu Dhabi and the DIFC in Dubai combine a common-law environment with a civil-law-friendly vehicle, which is why they have drawn so many family offices relocating to the Gulf. Jersey and Guernsey, unusually, offer both trusts and foundations, letting a family pick the instrument rather than the island.

For trusts, the map is different, and we cover it in Parts 3, 4 and 5: the US dynasty states, and the offshore centres led by the Cook Islands.

The takeaway

Do not start from the instrument. Start from the family. A common-law family with common-law advisers and assets should almost always use a trust. A civil-law family, especially one holding operating companies or building a philanthropic legacy, is frequently better served by a foundation their own world understands. The mechanisms differ; the goal, holding and governing capital without anyone owning it outright, is the same.

With the choice of vehicle settled, the rest of this series follows the trust, because that is where the two most demanding jobs live. Next, in Part 3, the structure built to defeat time itself: the dynasty trust.

The Dynasty Trust

2026-08-11T00:00:00Z

Most fortunes do not survive three generations, and tax is one reason why. In many countries, wealth is taxed each time it passes from one generation to the next. Pass a fortune down three times and the tax authority can take a large share three times over, quite apart from what heirs spend or lose. The dynasty trust is the structure built to stop that clock: to hold wealth for grandchildren, great-grandchildren and beyond, without being taxed at each handover, and without any single heir being able to break it up.

It is the most ambitious thing a family does with a trust, because it is a bet against time.

The idea

A dynasty trust is an irrevocable trust designed to last for many generations, sometimes forever. The founder funds it once, using their lifetime exemption from transfer and generation-skipping tax where available. From then on, the assets belong to the trust, not to any individual. Descendants become beneficiaries who can receive income and distributions at the trustee's discretion, but they never personally own the capital, so it is never in any heir's taxable estate and never exposed to any heir's divorce or creditors.

Without a dynasty trust, wealth is taxed at each transfer:

Gen 1 → tax → Gen 2 → tax → Gen 3

With one, the trust holds the capital and it is taxed once, at the start:

The dynasty trust holds for Gen 1 · 2 · 3 · ...

Keeping capital in trust replaces repeated transfer tax with a single event at funding.

The result is compounding left undisturbed. A pool that is never carved up by tax or divided among heirs at each death can grow across a century in a way that a fortune passed hand to hand almost never does.

The rule it had to defeat

For centuries, the common law refused to let anyone control property forever. The rule against perpetuities forced trusts to end within roughly a lifetime plus twenty-one years, on the principle that the dead should not rule the living indefinitely. That rule made true dynasty trusts impossible.

Then a competition began. To attract trust business, a number of US states simply repealed or drastically extended the rule. South Dakota led and remains the standard-bearer; Nevada, Delaware and Alaska followed, among others. In these states a trust can now last for centuries or in perpetuity. That single legislative change is why an American dynasty trust is so often domiciled in South Dakota even when the family has never set foot there. Offshore centres, including the Cook Islands and Nevis, now permit perpetual trusts as well.

Jurisdiction	Trust duration
South Dakota	Perpetual
Nevada	Up to 365 years
Delaware	Perpetual (personal property)
Alaska	Perpetual
Cook Islands / Nevis	Perpetual permitted

The lesson is that a dynasty trust is only as durable as the law of the place it lives. Choosing the jurisdiction is choosing how long the structure can last, which is why families domicile these trusts where perpetuity is allowed, regardless of where they themselves reside.

What it buys, and what it costs

The benefits are real: one taxable event instead of many, a century of undisturbed compounding, and protection at every generation from the creditors, divorces and misjudgements of individual heirs. For a family that thinks in generations, nothing else does quite this.

The costs are equally real, and worth naming. The founder gives up control permanently. The terms written today govern descendants not yet born, in a world no one can predict, which is why the trustee's discretion and a well-drafted deed matter enormously. And a structure meant to last forever can calcify: rules that made sense in one era can trap a family in another. Good dynasty trusts are drafted with flexibility, mechanisms to adapt, decant into a new trust, or replace trustees, precisely because forever is a long time.

A dynasty trust protects a fortune from time and tax. Our next structure protects it from something more immediate and more hostile: other people who want to take it. In Part 4, the asset-protection trust, and why the strongest ones are not American at all.

The Asset-Protection Trust

2026-08-12T00:00:00Z

Most trusts are built to pass wealth on. This one is built to defend it. An asset-protection trust exists for a single purpose: to place assets beyond the reach of future creditors, lawsuits, and claims. For a family whose wealth is exposed, an entrepreneur, a surgeon, anyone whose name invites litigation, it can be the difference between a bad event and a ruinous one. And here, unusually, the best structures in the world are not American. They sit on a scattering of small islands, and understanding why is the point of this piece.

The problem it solves

Most jurisdictions will not let you protect assets from your own creditors by giving them to a trust you still benefit from. If you are a beneficiary of the trust, a self-settled trust, your creditors can usually reach it. That is the default almost everywhere, and it is fatal to the whole idea.

A handful of jurisdictions deliberately broke that rule. They passed laws that do protect a self-settled trust, letting the person who created the trust also benefit from it while keeping its assets out of creditors' hands. Those places are where asset-protection trusts live.

Domestic, and why it is not enough

Several US states, Nevada, South Dakota, Alaska, Delaware, offer domestic asset-protection trusts, or DAPTs. They are cheaper and simpler than going offshore, and for modest exposure they have a place.

But they carry a structural weakness that no drafting can cure. A US court in another state can assert its reach, and the US constitution's requirement that states honour each other's judgments means a determined creditor has a path in. Bankruptcy adds another. A DAPT raises the cost of coming after you; it does not put you beyond reach. The various "bridge" and hybrid structures that promise to convert into an offshore trust when trouble comes share the same flaw: they remain within US court reach right up to the worst possible moment, when moving assets looks most like fraud.

Offshore, and why the Cook Islands lead

The strongest asset-protection trusts are offshore, and one jurisdiction stands above the rest. The Cook Islands, a self-governing nation in the South Pacific, essentially invented modern asset protection with its International Trusts Act of 1984. Four features do the work:

It does not recognise foreign judgments. A US court order means nothing there; a creditor must bring a fresh case in the Cook Islands, under Cook Islands law.
The statute of limitations is short, one to two years, so by the time most creditors think to look offshore, the window has closed.
The standard of proof is criminal, beyond a reasonable doubt, an almost impossible bar for a creditor to clear.
Four decades of case law have tested the regime under real pressure, and it has held, most famously when a US court held settlors in contempt but still could not compel the Cook Islands trustee to hand over the assets.

Nevis is the close second, cheaper and faster, with the same criminal standard of proof and two more deterrents the Cook Islands lack: a creditor must post a bond, often 100,000 dollars or more, before even filing suit, and the jurisdiction has abolished the court order that freezes assets during litigation. Its weakness is a thinner track record. Belize competes on speed and cost; the Cayman Islands and the Bahamas, despite their fame, are built for estate planning and funds, not for creditor defence.

	Cook Islands	Nevis	US domestic (DAPT)
Recognises foreign judgments	No	No	Yes, effectively
Standard of proof for creditors	Beyond reasonable doubt	Beyond reasonable doubt	Ordinary civil standard
Limitation period	1 to 2 years	1 to 2 years	Varies, longer
Extra deterrent	Longest case-law record	Creditor must post a bond	None
Strength	Strongest	Very strong	Limited

The limits that catch people out

An asset-protection trust is powerful, and it is not magic. Three hard rules decide whether it works.

First, timing is everything. Protection has to be in place before a claim arises, while you are solvent and untroubled. Move assets into a trust after a lawsuit or a debt has appeared and any court, onshore or off, can call it a fraudulent transfer and unwind it. These trusts protect against the future, never the present.

Second, you must genuinely give up control. The whole structure rests on the settlor not being able to pull the assets back. Keep too much control and a court can treat the trust as a sham. A US judge can even jail a settlor for contempt for refusing to repatriate assets, which is exactly why the trustee must be foreign and independent: the court can punish you, but it cannot force the trustee.

Third, this is not tax avoidance and not secrecy. A properly run offshore trust is fully reported to your home tax authority and disclosed under the automatic exchange of information. It changes who can take your assets, not whether you declare them. Anyone selling it as a way to hide money is selling a crime.

Handled right, an asset-protection trust is the strongest financial shield a family can hold. Handled wrong, set up too late, controlled too tightly, or sold as secrecy, it is worse than nothing. Which brings us to the last and most practical question in this series. In Part 5: how to actually choose a jurisdiction and a trustee, the mistakes that ruin trusts, and when a family should not create one at all.

Choosing and Running a Trust

2026-08-13T00:00:00Z

The first four parts of this series covered what trusts are and what they do: the mechanics, the choice between trust and foundation, the dynasty trust built to outlast generations, and the asset-protection trust built to defend against creditors. This final part is the one families actually get wrong. A trust is only as good as two decisions and one discipline: the jurisdiction you choose, the trustee you appoint, and the way you run it afterwards.

Choosing the jurisdiction

There is no best jurisdiction, only a best fit for the job. Five things decide it:

Purpose. Dynasty planning points to a perpetuities-friendly home like South Dakota; serious asset protection points offshore to the Cook Islands or Nevis; ordinary succession may need neither.
Legal strength and track record. A protective statute is worth little without decades of case law showing courts will actually apply it under pressure. Pedigree matters.
Tax neutrality. A good trust jurisdiction imposes no tax of its own, so the only tax that applies is the one in the family's home country. The trust changes protection and succession, not what you owe.
Stability. You are making a decision meant to last generations. The political and legal stability of the jurisdiction is not a detail.
Reputation and banking. A jurisdiction that banks and regulators view as a red flag will create friction on every account and every deal. The quiet, well-regulated centres are worth more than the aggressive ones.

Choosing the trustee

This decision is underrated and decisive. The trustee holds legal title to everything; the wrong one can be a disaster, and the right one is the structure's backbone.

An individual trustee, a trusted relative or adviser, is cheap and personal, and usually a mistake at any scale: mortality, conflicts and inexperience make them fragile. A corporate trustee, a professional trust company, brings permanence, expertise and independence, at the cost of fees and some distance from the family. And for larger family offices, there is a third option that has become the favourite: the private trust company, a company the family creates specifically to act as trustee of its own trusts. It keeps trusteeship close to the family and its governance while still meeting professional fiduciary standards. It is more to set up and run, and for a substantial multigenerational structure it is often worth it.

The letter of wishes

Because a discretionary trustee holds real power, the family guides them with a letter of wishes: a private, non-binding note explaining how the settlor would like discretion used, how to weigh education, business ventures, or hardship among beneficiaries. It steers without creating legal entitlements that would weaken the trust's protection. It should be revisited as the family changes. It is the quiet instrument that keeps a trust aligned with a family's intent long after the settlor is gone.

The mistakes that ruin trusts

Most failed trusts fail for the same handful of reasons:

Keeping too much control. The most common and most fatal error. A settlor who still effectively runs the assets invites a court to treat the trust as a sham and ignore it. The protection you wanted requires the control you must give up.
Setting it up too late. Protection established after trouble appears is a fraudulent transfer. The time to build the structure is when you do not yet need it.
The wrong jurisdiction or trustee. A protective statute undone by a weak trustee, or a dynasty trust in a state that still limits duration, is a structure that fails at the one moment it is tested.
Ignoring reporting. A trust is not secrecy. It must be declared to the family's tax authorities and disclosed under the Common Reporting Standard and, for US persons, FATCA. Getting the structure right and the reporting wrong turns a legitimate plan into a liability.

When not to use a trust

The most useful advice in this series is that sometimes the answer is no. A trust adds cost, complexity and irreversibility. It is the wrong tool when the estate is simple enough that a will and clear ownership will do; when the family will not accept the loss of control an effective trust demands; when a civil-law family would be better served by a foundation their own system understands; or when the real motive is secrecy or dodging tax, in which case a trust is not a solution but a trap. A structure you do not need, or will not run properly, is worse than none.

The through-line

Read together, these five parts make one argument. A trust is a machine for separating ownership from benefit, and everything it can do, move wealth out of a taxable estate, hold it across generations, shield it from creditors, flows from that single split and from the price it demands: control, surrendered in exchange for protection and continuity. Get the mechanics, the vehicle, the jurisdiction and the trustee right, run it with discipline and full disclosure, and a trust is the most durable structure private wealth has. Get any of them wrong, and it is an expensive way to feel protected while being exposed.

That is the whole of it. The architecture is not complicated. The discipline is.