🇸🇬Family Office: SFO/MFO Structure, Jurisdictions, Governance and Succession

Concept

A family office is the structure that manages the capital and affairs of one family (single family office, SFO) or several (multi family office, MFO): investment, legal and tax support, succession, the education of heirs and sometimes everyday life. It is the headquarters of Flag 4, tying holdings, foundations and trusts into a single system.

The idea grew out of the practice of industrial dynasties: the Rockefeller office, traditionally regarded as the first modern family office (1882), ran the family's affairs as a single whole. Today the format is booming — according to Deloitte, there are around 8,030 single family offices worldwide against roughly 6,130 in 2019, and their number may rise to about 10,720 by 2030; assets under management from $3.1 to $5.4 trillion. Families increasingly want to manage capital systematically and gather scattered advisers under one roof.

The capital and budget levels below are practitioner benchmarks, not statutory thresholds: where the line falls inside the range depends on the asset mix and on which functions the family keeps in-house.

Section map: which question to start from, and where the answer is.

What a family office does

The set of functions varies, but the core is stable: consolidated accounting and reporting across all assets, an investment strategy and allocation across asset classes, tax and legal support, succession and the preparation of heirs, and philanthropy. Larger offices add real estate management, art collections and sometimes concierge-level service. The broader the mandate, the more expensive the infrastructure — and the more important it is to set out in advance what the office handles itself and what it delegates to outside contractors.

SFO or MFO

An SFO serves one family and is usually justified from roughly $100–250 million in capital: it is expensive to run, but it gives full control and confidentiality. The line is set not by the headline sum but by the asset mix and the cost ratio: a minimally viable team costs much the same whatever the size of the portfolio, so on $100 million a $1 million budget already means about 100 bps a year. An MFO works for several families and spreads the cost of the team and infrastructure — a sensible entry point while a dedicated office is not yet economically justified. The choice between them comes down to a balance of cost, control and privacy; many families start with an MFO and move to an SFO as their capital grows.

An SFO's budget usually starts at $1–2 million a year: team salaries, audit, legal, IT and data storage. Even a small office's team means a CIO or investment director, a financial controller and a lawyer; the rest is outsourced to banks, asset managers and auditors. That is exactly why an MFO is attractive at the outset: the same expertise without the full fixed overhead. As assets grow and the structure becomes more complex, the family moves to its own office, where control and confidentiality outweigh the saving.

Jurisdictions

The jurisdiction is chosen for its combination of tax regime, infrastructure — banks, asset managers, lawyers — and reputation. For Russian and international families the tone is set by a trio: Singapore, the UAE and Switzerland. The first two have built regimes squarely for the family office; the third wins on the maturity of its market. The choice of hub is usually tied to the principals' citizenship and residence, banking access and where the family actually lives — which affects both tax residency and the structure's resilience to sanctions.

Singapore

Singapore offers incentives under sections 13O and 13U of the Income Tax Act: a fund's income from designated investments is exempt from tax. The terms of both regimes are set by the MAS framework of 5 July 2023.

The regimes differ on the asset threshold and the size of the team; their other requirements coincide. Capital is pooled through a VCC structure; for more, see the overview of wealth planning in Singapore. The primary source on the incentives is the MAS website.

UAE

In Dubai there is the DIFC: the Family Wealth Centre has operated since 2023, and the Family Arrangements Regulations 2023 (enacted 31 January 2023) replaced the former single family office regime. The entry threshold is around $50 million in total net assets (including real estate and operating businesses), and the office manages without a separate DFSA licence.

ADGM does run its own SFO regime: a single family office is a controlled activity (Rule 8 of the Commercial Licensing Regulations (Controlled Activities) Rules 2025) licensed by the ADGM Registration Authority, and a pure SFO needs no FSRA financial-services permission.

The ADGM wealth gate is family net assets of at least USD 10 million (Rule 13 of the Commercial Licensing Regulations (Conditions of Licence and Branch Registration) Rules 2026, published and in force 24 April 2026; the 2024 edition set USD 30 million measured on investable/liquid assets). SPVs (restricted scope companies) and foundations sit under that regime as ownership layers.

A family foundation in the UAE, where the conditions are met, is treated as tax-transparent and pays no corporate tax (Ministerial Decision No. 261 of 2024).

Switzerland

Switzerland's strength is its mature private banking infrastructure, political stability and reputation; there is no special family office regime here. Wealthy families combine the office with lump-sum taxation (forfait) for the principals themselves and a residence permit. Competition to these hubs comes from Hong Kong, which since 2023 has introduced its own tax incentive for single family offices and is fighting for the same capital.

Hong Kong

Hong Kong's answer is the FIHV (family-owned investment holding vehicle): a family investment vehicle's income from qualifying transactions in Schedule 16C assets is charged profits tax at 0%. The threshold is HK$240 million of assets under the management of an ESF Office (eligible single family office), which is expected to show real substance: at least two full-time employees and HK$2 million of operating expenditure in Hong Kong a year. There is no application to a regulator: the concession works by self-assessment in the profits tax return, and a single-family office generally manages without an SFC licence.

The Inland Revenue (Amendment) (Preferential Tax Regimes for Funds, Family-owned Investment Holding Vehicles and Carried Interest) Bill 2026 broadens the list of qualifying assets — digital assets, private credit, real estate outside Hong Kong, carbon credits, insurance-linked securities, interests in non-corporate private entities and precious metals. It was gazetted on 12 June 2026 and introduced into the Legislative Council on 24 June 2026; as at early September 2026 it has not been enacted and remains under consideration.

Once enacted, the amendments apply to years of assessment beginning on or after 1 April 2025 — retrospectively from 2025/26. Conditions and risks are set out in the Hong Kong FIHV article.

The hubs compared

Four hubs in two cuts; "—" means the regime simply has no such parameter. First, what the regime demands at the door.

Threshold and presence decide whether a regime is open at all. The second cut is what it gives and how it is obtained.

The tax effect is delivered by Singapore and Hong Kong, while DIFC and Switzerland win on infrastructure and access rather than on rate.

Where the Office Sits When the Family Is Scattered

The jurisdiction sections above compare hubs as if the family lived in one place. The harder, commoner case is a family spread across three countries — and then the office's location is read against the members' residences, not against a league table. The working matrix, in the order that decides cases:

• Members' tax residences first. Every beneficiary's and controller's residence country taxes, attributes and reports against the office's structures: the residence map is the gravitational centre of the whole stack — see tax residency basics. An office jurisdiction is chosen so that no member's residence turns it into a trap.
• CFC and management perimeters. A member's residence with controlled-foreign-company rules can attribute the office's holding-layer profits to that member; and an office genuinely managed from a member's high-tax country can be dragged into residence there — place of effective management is a fact, not an address. Both are covered in economic substance and the holding structures material.
• Banking and licensing perimeter. The office is regulated where its activity is performed and its clients sit — not where its licence was convenient; and the banks read the members' passports and residences, not the office's flag (see the map of banks by jurisdiction).
• Reporting follows the people. CRS and FATCA report the accounts to where the members live; a scattered family means scattered reporting — assumed, not managed around (CRS overview).

The answer pattern that survives: locate the office where its staff and decisions genuinely are, in a hub whose regime fits the family's actual assets (13O/13U, FIHV, DIFC above), and clear every member's residence against the structure before fixing the jurisdiction. Where members live in treaty-weak or CFC-heavy countries, the office location compensates — not the other way round.

The licensing perimeter: when an office stops being a family office in law

Every regime above assumes the office is a family office in the regulator's sense. That status is a perimeter, not a label: it is held on conditions, and ordinary commercial decisions — taking a friend's money alongside the family's, admitting a long-serving executive as an investor, letting the family tree run one generation too far — can put the office outside it. Crossing the line does not fine the family; it converts the office into a regulated investment adviser, fund manager or AIFM, with authorisation, capital, conduct and reporting duties attached, and it normally happens without anyone deciding to become regulated. What counts as the trigger differs by jurisdiction, and — the part that catches families running an office in two places — the tests are not versions of one test.

United States: clients, ownership, holding out. The Dodd-Frank Wall Street Reform and Consumer Protection Act (Pub. L. 111-203, s.409(a), 21 July 2010) inserted s.202(a)(11)(G) into the Investment Advisers Act of 1940 (15 U.S.C. s.80b-2(a)(11)(G)), excluding "any family office, as defined by rule, regulation, or order of the Commission". The definition is 17 CFR 275.202(a)(11)(G)-1, and its three conditions are cumulative. The office must have no clients other than family clients, with a one-year transition where an interest passes involuntarily on a death or a key employee's departure (para. (b)(1)); it must be wholly owned by family clients and exclusively controlled, directly or indirectly, by family members or family entities (para. (b)(2)); and it must not hold itself out to the public as an investment adviser (para. (b)(3)). "Family member" reaches all lineal descendants of a common ancestor and their spouses or spousal equivalents, "provided that the common ancestor is no more than 10 generations removed from the youngest generation of family members" (para. (d)(6)) — a hard numeric ceiling, and the one condition a long-lived dynasty fails by doing nothing at all. A narrow grandfather in para. (c) preserves offices that were already advising certain officers, directors and employees who are accredited investors, and certain family-owned companies, before 1 January 2010.

European Union and United Kingdom: the test is external capital. AIFMD (Directive 2011/61/EU) does not define family offices, and they do not appear in the Article 2(3) list of excluded entities; the exclusion lives in recital 7, which says that "[i]nvestment undertakings, such as family office vehicles which invest the private wealth of investors without raising external capital, should not be considered to be AIFs in accordance with this Directive". The FCA reads it the same way and gives the reason: at PERG 16.2, Q2.50 it explains that the AIF definition "does not cover an arrangement in which the persons raising and providing capital are the same", and looks for a family relationship between the investors, no capital raised outside that relationship, and money and relationships that predate the vehicle. Article 3(1) of the Directive adds a separate way out where a fund's only investors are the manager or its parent undertakings or subsidiaries. Note what this test does not ask: neither the number of generations nor who controls the office matters — only whether outside capital came in. Failing it does not create a need for a "family office licence", because there is none; it makes the vehicle an AIF that someone must be authorised to manage — see the fund regulatory perimeter and, for what the vehicle is treated as, entity classification.

Hong Kong: three limbs and an intra-group carve-out. The SFC's published position on family offices is that a licence is needed only where three things are true together — the services are a regulated activity, they are carried on as a business, and that business is carried on in Hong Kong. Type 9 (asset management) carries an intra-group carve-out a single family office can use where it serves only its related entities: its wholly owned subsidiaries, the holding company that holds all its issued shares, or that holding company's other wholly owned subsidiaries. Separately, the SFC treats an arrangement that is not run as a business — one "not receiving any income, other than reimbursement of operating expenses from the family" — as generally outside licensing, which makes the office's own fee arrangement a licensing fact and not merely an accounting one. Multi-family offices get neither concession: they serve more than one family and are run as commercial ventures, so the three limbs are met in the ordinary way. And the FIHV concession described above is a tax regime carrying no licensing relief — the two questions are answered separately, on different facts.

Singapore: from a case-by-case exemption to a notified class exemption. Singapore offices historically relied on exemptions MAS granted on the facts of each office. MAS's revised framework for single family offices takes effect on 15 June 2026 and replaces that with a class exemption: no capital markets services licence, but notification — a Notice of Commencement of Business within 14 days of starting operations in Singapore, an account with a MAS-licensed bank, and an annual return. Offices already operating before 15 June 2026 have a transition period and must notify by 15 June 2027; afterwards individual exemptions are granted only in exceptional cases. The planning point is that the Singapore relief is now conditional and evidenced: a missed notification is a compliance failure with a date on it, not a formality.

The Gulf: the registrar licenses the office, not the financial regulator. In the DIFC the Family Arrangements Regulations 2023 route the office to the DIFC Registrar rather than the DFSA, and the office needs no DFSA financial-services licence; the thresholds and the registration mechanics are in the DIFC family office. In ADGM a single family office is a controlled activity licensed by the Registration Authority (Rule 8 of the Commercial Licensing Regulations (Controlled Activities) Rules 2025), again without FSRA financial-services permission. In both, the licensing question and the tax question are put to different bodies and answered on different facts.

Read the table as a set of trip-wires rather than a ranking. Each regime is broken by a different fact, so an office that is safe in one place can be regulated in another on the same set of events: an outside investor breaks the American test and not the Hong Kong one; a fee threatens the Hong Kong description and not the American exclusion; a fifteenth generation breaks the American rule and is irrelevant everywhere else. The practical discipline is to keep a written list of the conditions each of the family's offices actually relies on, and to test every new client, investor, fee and family admission against that list before it happens rather than at the next audit.

The office's own tax character: cost centre or trade or business

Everything above looks at the office from the family's side. A tax authority looks at it from the other side and asks a different question: is this entity carrying on a business, or is it the family spending money on its own investments? The answer decides whether the office's payroll, premises, technology and professional fees are deductible at all — which, on a budget that starts at US$1–2 million a year, is not a rounding error. It is also the one question where the structure of the office, rather than the choice of hub, does the work.

Same activity, two answers. In Higgins v. Commissioner (US Supreme Court, 3 February 1941, 312 U.S. 212) the taxpayer ran extensive holdings of securities and real estate from rented offices with paid staff and deducted the salaries and expenses under s.23(a) of the Revenue Act of 1932 as the expenses of carrying on a trade or business. The Court refused. He "merely kept records and collected interest and dividends from his securities, through managerial attention for his investments", and "[n]o matter how large the estate or how continuous or extended the work required may be, such facts are not sufficient, as a matter of law" to make that a trade or business: scale does not convert investing into trading. In Lender Management, LLC v. Commissioner (US Tax Court, 13 December 2017, T.C. Memo. 2017-246, docket nos. 25617-15 and 25618-15, Kerrigan J., years 2010–2012) a family office doing recognisably the same work came out the other way. The court found that Lender Management "did not simply make investments on behalf of the Lender family group. It provided investment advisory services and managed investments for each of its clients individually"; that it advised clients it did not control; and that it was paid a profits interest in each of the investment LLCs in exchange for the services it provided, rather than a return on its own capital. On those facts it "did substantially more than keeping records and collecting interest and dividends" and was carrying on a trade or business within IRC s.162. What separates the two decisions is not size, family relationship or effort — it is whether the office has clients it does not control and is paid for a service rather than on its own money.

Why the answer matters more than it used to. After Higgins, Congress created a deduction for the expenses of producing income — now IRC s.212 — and for decades an individual who could not reach s.162 could still reach that. The route is closed. IRC s.67(g), inserted by the Tax Cuts and Jobs Act (Pub. L. 115-97, s.11045, 22 December 2017), disallowed miscellaneous itemized deductions for tax years beginning after 31 December 2017 and before 1 January 2026; Pub. L. 119-21, s.70110 (4 July 2025) removed the end date and redesignated the provision as s.67(h), so the disallowance now runs on indefinitely. For an individual the s.212 fallback no longer exists, and the whole question is whether the office's activity is a trade or business under s.162.

The holding does not travel. Lender Management is a United States decision on its own facts. It says nothing about an office that simply manages the family's own money and recharges its costs: that office sits on the Higgins side of the line whatever it is called, and calling it a "management company" does not move it. Nor does the case decide the position outside the United States, where the usual shape is a service company charging the family's entities a fee with a mark-up — and the live questions become whether the mark-up is at arm's length (transfer pricing) and whether the paying entities can deduct what they pay. Where the family includes US persons, the reporting and control consequences that follow are set out in US tax controls for family offices.

Governance

The working centre of governance is the investment policy statement (IPS): the document in which the family fixes the capital's objectives, the target allocation across asset classes, restrictions, liquidity requirements and authority — who decides, and within what limits. Until there is an IPS, the "strategy" is in practice the manager's memory and the latest thread of correspondence. The document lives in the investment committee — the regular forum where family and office check the portfolio against the strategy; the CIO sets the agenda, delivers the allocation and answers for the choice of managers. The basic division is simple: the family decides the "what and why" — objectives, risk appetite, major transactions; the office executes the "how" — instrument selection, rebalancing, reporting. The minimum configuration of roles:

• investment committee — approves the IPS and deviations from it;
• CIO — answers for the allocation, the managers and the reporting;
• principal or family council — objectives, risk appetite, the office's mandate.

The values layer — who sits on the committee, how family branches vote, what happens when generations change — is fixed one level up, in the family constitution; the IPS is its investment annex.

Succession

The defining challenge of the decade is the "great wealth transfer". Cerulli estimates that by 2048 around $124 trillion will change hands in the United States alone: roughly $105 trillion to heirs and a further $18 trillion to charity. That is why the focus falls on preparing the next generation and formalising governance: without a succession plan both the capital and the office itself lose their footing as generations change.

The instruments are chosen to fit the family: trusts and a Private Trust Company for control and asset protection, private foundations and a family constitution for governance rules, life insurance for liquidity to cover taxes and buy-outs. The family office ties them into a single mechanism and makes sure the documents do not drift away from the actual decisions. How those instruments relate to one another and to the matrimonial regime is set out in the family perimeter map.

Family office as a service

The classic SFO, with its own staff and a budget upwards of a million dollars a year, is the top end of the spectrum. The family office function is needed by a family long before a dedicated office is economically justified, and it is met by a service model — the outsourced or virtual family office. It works on a hub-and-spoke principle: a single coordinator holds the whole picture, and specialists plug in as the task requires. In practice this function comes down to three roles.

A single point of entry

A single point of entry takes financial and legal questions. The family gets one address to which any request is sent: a bank's compliance query on a personal account, a real estate transaction, a tax notice from another jurisdiction, the choice of an investment platform. From there the question is routed to the relevant specialist, but responsibility for the outcome and the deadlines stays at one point. This spares the family the familiar cost of coordinating a dozen advisers who do not talk to one another.

An administrator of legal entities

A family's capital is usually packaged into companies: a family holding, an SPV for real estate or an aircraft, operating businesses, funds and trusts. Each entity has its own calendar of obligations: annual reporting and audit, renewal of registered agents and directors, updating KYC files at banks, UBO registers, substance requirements. The family office keeps this calendar across all the structures at once — the family sees a consolidated picture instead of a stream of letters from registrars in five jurisdictions.

A personal adviser to the principal and the family

The foundation of this role is accumulated context: the adviser knows the family members' residences and passports, their marriage contracts, their plans for heirs, the history of past decisions. So a new question — a relocation, the sale of a business, a gift — is resolved with the whole picture in view. The longer the family–adviser relationship lasts, the more valuable it becomes: knowledge of the family outlives changes of banks, asset managers and jurisdictions.

Changing a provider is not moving the family office

Families say "we are changing our family office" when they mean any of several different things: replacing the multi-family office that coordinates everything, firing one discretionary manager, moving a custody account to another bank, or ending a trustee's appointment. These are not the same act, and treating them as one is the commonest way a transition goes wrong. A family office is not a single supplier; it is a stack of separate legal relationships — advisory, discretionary management, fiduciary office, custody, administration — each created by its own contract or instrument, each terminating on its own terms, and each moving (or not moving) different things.

The people around a family's capital are usually collapsed in conversation into one word — "the provider" — but law keeps them apart. An adviser recommends and has no power to act. A discretionary manager trades the portfolio under a mandate that grants defined authority. A fiduciary — a trustee, protector, private trust company or foundation council — holds legal title or a duty of loyalty and cannot simply be "switched" like a vendor. A custodian holds the assets and settles trades, and under custody law is structurally separate from whoever instructs it. Deciding what a transition changes starts with deciding which of these four you are actually replacing.

Who does what: adviser, fiduciary, manager, custodian

The single most useful document to draw before any transition is a map of who holds what authority over which asset, under which instrument. In an SFO these functions are largely internal (the office's own CIO, controller and lawyer) with custody and execution outsourced; in an MFO the coordinating layer is the external firm while managers and custodians sit below it; a purely external ("virtual") office buys every function from third parties. Whatever the shape, the functions stay legally distinct.

This table separates the functions that clients most often merge. The point of comparison is not who is "best" but what each relationship actually carries — because that determines what ends, and what moves, when you replace it.

The lesson of the table is that replacing the coordinating office is the lightest change — it moves information and administration — while replacing a custodian is the heaviest, because it moves the assets. Replacing a fiduciary is a different category again: it is a change of legal office effected by deed, governed by the trust or foundation instrument, and it does not follow the mechanics of firing a supplier. This is why the family charter and the investment policy statement matter at a transition: they record which authorities exist and who may change them, so a handover does not depend on one person's memory.

Ending an engagement: mandates, authority, data and records

Terminating a relationship cleanly means acting on four fronts at once, because ending the contract does not automatically end the powers it created.

Mandates and notice. The engagement or management agreement sets the notice period, whether either side may terminate for convenience or only for cause, and what fees survive termination. Read it before anything else: the notice clock, not a market convention, sets the earliest clean exit. There is no universal statutory transfer period — where an agreement is silent, the practical floor is whatever the custodian's transfer process and the settlement of open trades require.

Authority revocation. A discretionary mandate is carried by a limited power of attorney or a discretionary-authority form lodged with the custodian. Terminating the IMA and revoking that authority at the custodian are two separate steps; until the revocation reaches the custodian, the outgoing manager can still trade the account. The revocation should be dated and confirmed by the custodian, and any standing instructions (recurring transfers, sweep rules, dividend handling) reviewed at the same moment, because they outlive the manager.

Data ownership and export. Who owns the consolidated reporting, the performance history, the KYC files and the entity records is a contractual question, and it is answered badly by silence. A well-drafted engagement gives the family a right to its own data, an export in a usable format, and a defined handover on exit; many standard forms do not. Consolidated performance data assembled by the office is often the hardest thing to recover, because it lives in the provider's system, not in any custody statement.

Records. Regulated participants must keep their own books regardless of the client's move, and those retention duties are a reason a provider keeps records, not evidence that the client will receive them. Under the US Investment Advisers Act an adviser must preserve its books and records for at least five years from the end of the fiscal year of the last entry, the first two in an appropriate office, and — critically for a transition — an adviser that ceases business must arrange to preserve the records for the rest of that period and notify the SEC in writing where they are kept (17 CFR 275.204-2). In the UK an FCA firm must keep orderly records of its business, and MiFID-business records for at least five years (SYSC 9.1). A Singapore capital-markets licensee is bound by the record-keeping and conduct duties of the Securities and Futures (Licensing and Conduct of Business) Regulations. The client's takeaway is narrow but important: the outgoing firm's regulatory retention is not a substitute for the family obtaining its own copies before the relationship ends. What the family should be holding on its own side, and why a tax authority later reads the gap as an evidential problem rather than an administrative one, is set out in family office records and tax evidence.

What stays with the custodian

The reason "switching provider" is not "moving assets" is the structure of custody. Under the SEC custody rule a client's funds and securities are held by a qualified custodian — a bank, a registered broker-dealer, a futures commission merchant or a qualifying foreign institution — in a separate account in the client's name, or in an account holding only clients' assets; and the custodian, not the adviser, sends the account statements (17 CFR 275.206(4)-2). Because the assets sit under the custody agreement and not under the advisory or management mandate, changing the adviser or the manager leaves the assets where they are. Only a decision to change the custodian moves them, and that is a separate, heavier transaction — see securities custody and, for how the account's law is fixed, booking centres.

Unfinished trades. A control date rarely lands on an empty book. Trades executed but not yet settled belong to the outgoing mandate and must settle where they were placed; in the US the standard cycle has been T+1 since 28 May 2024, so a trade done on the cutover day settles the next business day and cannot simply be "handed over" mid-flight. Corporate actions in progress, unsettled subscriptions and redemptions in funds, margin and securities-lending positions, and pending dividends all have to be identified and left to complete or unwound deliberately, not discovered after the account has been transferred.

Standing instructions and the account itself. If the custodian is being kept and only the manager is changing, the account number, its booking jurisdiction and its standing settlement instructions usually persist — which is the easy case. If the custodian is being changed, the assets move by an account transfer on the custody rails (in the US, broker-to-broker transfers run through the ACATS system; other markets have their own), positions move in kind where possible, and only genuinely non-transferable holdings are sold. In-kind transfer preserves the tax basis and avoids forced disposals; a cash transfer forces sales and can trigger tax — so which route is used is a decision, not a default.

The transition, end to end

A clean transition is organised around a single control (cutover) date and a period of parallel reconciliation either side of it, so that at every moment it is clear who may act, where the assets are and which records are authoritative. The order below is the logic, not a calendar; the elapsed time is whatever the notice periods, the custody transfer and the open trades require.

1. Map the stack. List every mandate, authority, custody account, fiduciary office, entity and data set, and mark for each what it holds and what its termination terms are. This is the responsibilities matrix above, made concrete for the family.
2. Read the exits. Extract each agreement's notice period, termination trigger, surviving-fee clause and data-handover obligation. The longest binding notice sets the earliest clean cutover.
3. Onboard the successor first. The incoming provider's KYC, account opening and mandate should be substantially complete before the outgoing one is switched off, so authority never lapses into a gap. Successor onboarding is the step most often underestimated and most often the real bottleneck.
4. Set the cutover date and run in parallel. For a defined window both sets of records are reconciled against the custodian's statements — the independent source of truth — so positions, cash and cost basis agree before authority changes hands.
5. On the date: revoke and re-grant, in sequence. Revoke the outgoing authority at the custodian and lodge the successor's, confirm the revocation is effective, and re-point standing instructions. Assets that are not moving stay put; assets that are moving begin their account transfer.
6. Let open trades settle. Unsettled trades complete where they were placed; corporate actions and fund subscriptions/redemptions in flight are tracked to completion.
7. Close the money and the calendar. Pro-rate and settle fees to the cutover date, recover deposits or retainers, and hand over the tax and reporting calendar — CRS/FATCA classifications and filing dates, entity annual returns and audits, UBO and substance obligations, capital-gains and withholding reporting — so no deadline falls in the gap between providers. A missed filing during a handover is a classic avoidable failure.
8. Take delivery of data and records. Obtain the exported data, the performance history and the entity files, and confirm the outgoing firm's own regulatory retention separately.

Liability allocation. Because authority and duty change hands on the cutover, so does responsibility. The outgoing manager answers for trades and advice up to the date it held the mandate; the successor answers from the moment its authority takes effect; the custodian remains answerable for safekeeping throughout, as its duty never depended on who was instructing it. Errors discovered after transfer are allocated by when they occurred, which is why the parallel reconciliation and the dated revocation matter as evidence, not just as housekeeping. Where a fiduciary is retiring, the retiring trustee will usually seek an indemnity and the discharge terms are set by the trust instrument — a matter closer to business succession and succession planning than to swapping a supplier.

When the outgoing provider will not cooperate

Two failures recur, and both are managed by not depending on the outgoing provider's goodwill.

The provider will not release data. A firm may withhold consolidated reports or delay the export — sometimes over disputed fees, sometimes through inertia. The structural protection is that the assets and the custody statements do not belong to it: the custodian holds the assets and issues the statements directly, so the family can rebuild positions, cash and cost basis from the custodian even if the office's own reporting is withheld. What is genuinely hard to reconstruct is the office's assembled performance history and internal analysis — which is exactly why the data-ownership and export right should be a term of the engagement from the start, not a favour asked at the exit. A fee dispute is resolved on its own track and should not be allowed to hold the assets hostage, since the custodian, not the office, controls them.

A key person is unavailable. Much of a family office's value is the accumulated context held by one relationship manager or lawyer — residences, marriage contracts, the history of past decisions — and that knowledge can walk out with a departure or a death. The defence is institutional memory that does not live in one head: the family charter and the investment policy statement recording objectives, authorities and decision rules; a maintained register of entities, mandates, custody accounts and the reporting/tax calendar; and copies of the underlying documents held by the family, not only by the provider. A transition tests whether that memory was ever externalised. Where the missing person was a fiduciary, continuity is a matter of the trust or foundation instrument — the mechanism for replacing a trustee or protector who cannot or will not act — not of a service contract.

Purpose-directed capital: who sets the purpose, who owns it, who grants, who controls, and how it ends

Most of what a family office manages is the family's own capital, held for the family's benefit. A separate and frequently misunderstood layer is purpose-directed capital — money committed to an aim outside the family: a charitable mission, a defined non-charitable purpose, an endowment or scholarship, a collection held for the public. The moment capital is dedicated to a purpose, the questions of ownership, benefit and control come apart, and the family that funded it keeps influence without keeping the economic interest. This map routes each of those questions to the guide that answers it; the deep charitable, tax and cross-border treatment lives in family philanthropy: foundations, DAFs and cross-border giving, and the vehicles themselves in the trust and foundation studies.

The distinctions the map keeps apart are the ones clients most often merge. Setting the purpose (donor intent) is not owning the capital, which passes to the vehicle on a completed gift; owning is not deciding a grant, which a fiduciary or a committee does within the purpose; deciding is not controlling execution, which runs after the money leaves through monitoring and, if it is diverted, recovery; and none of these is changing the purpose, which the law deliberately makes hard once the dedication is complete.

Two boundaries are worth stating outright. First, changing the vehicle's provider is not changing the purpose or the ownership: replacing a trustee, an administrator or a discretionary manager is the transition described above in this article, and it moves authority or administration, not the dedication of the capital. Second, control is not a right of return — the family that sets a purpose and funds a vehicle keeps the influence its reserved powers give it (a protector's veto, an adviser's recommendation, a seat on the council), but not the economic interest it gave away; recovering the capital takes a reserved reversion or a legal claim, not merely displeasure with how the purpose is being pursued. The charitable, tax and cross-border detail behind every row is in family philanthropy.