Almost everything written about family offices is about starting one. What to put in it, whom to hire, which jurisdiction, which reporting system. Very little is written about the other end, which is strange, because a family office is a business with a defined purpose and businesses whose purpose has been served are supposed to close.
Instead they persist. The structure was built for a family of four with an operating company and a concentrated position. Fifteen years later the company is sold, the four have become eleven across five countries, the position is a diversified portfolio managed by two banks, and the office continues to exist because closing it was never anyone’s job. What follows is how to recognise that state and how to end it without damaging what the office was built to protect.
The office nobody decided to keep
A family office is rarely designed. It accretes. A trusted person from the operating business takes over the family’s administration. An entity is incorporated to employ that person. A second entity is added for a property. A reporting tool is bought and an annual audit begins. At no point does anyone approve a mandate, because no moment ever called for one.
That origin explains why closing is so hard. There is no founding document that says what the office is for, so there is no benchmark against which to judge whether it is still doing it. Discussions about winding down therefore become discussions about people, which families avoid, rather than discussions about purpose, which they could actually resolve.
The first useful exercise is retrospective. Write down, in one paragraph, what the office does that the family could not buy from the market or do itself. If that paragraph is hard to write, the question of whether to keep the office has already been answered, and what remains is the sequencing.
When the cost line stops matching the asset base
The cost of a family office is almost entirely fixed. Salaries, premises, entity maintenance, audit, software, professional retainers. The asset base that supports those costs is not fixed at all. It falls when a generation divides it, when a large asset is sold and distributed, when a market corrects, or when a family branch takes its share and goes.
The result is a ratio that deteriorates silently. Nobody notices the year the office stopped being proportionate, because the costs did not rise. The denominator fell. Families that review the arrangement annually catch this early. Families that review it when someone complains catch it several years late, by which point the office has consumed a visible part of the return.
There is a second, less obvious cost, which is the opportunity cost of complexity. Six entities require six sets of accounts, six filing calendars, six banking relationships and six sets of beneficial ownership declarations, because the trade register records the ultimate beneficial owners of each registered company separately. Each one is small. Together they consume the attention of the family’s most senior adviser, who is then not available for the decisions that actually matter.
The honest test is comparative rather than absolute. Price what the family would pay to obtain the same services from external providers, at the scale it now operates. If the internal arrangement is meaningfully more expensive and not meaningfully better, the structure is being kept for reasons that are not financial, and those reasons should be named out loud.
When the family is no longer in one place
Dispersion changes the arithmetic more than shrinkage does. A family office serving members resident in one country solves one tax problem repeatedly. A family office serving members in five countries solves five different problems, each requiring local advice that the office itself cannot give, and each interacting with the others.
Residence changes also break the assumption that the office is neutral. A structure administered from one jurisdiction, for beneficiaries resident in several, raises questions about where decisions are actually taken and where the entities are effectively managed. The Dutch General Tax Act does not answer that question by reference to the place of incorporation: under its article 4, where a body is established is judged according to the circumstances. What was a simple administrative convenience becomes a substance question, and substance questions are answered by tax authorities in the country that asks first.
Dispersion also erodes the thing the office was really for, which is shared decision making. Once the family members hold their shares separately, invest separately, and speak to their own advisers, the office stops being a common instrument. It becomes a service provider to a group of clients who happen to be related and who increasingly want different things from it.
The signal to watch is not geography but agreement. When the annual meeting stops producing decisions and starts producing minutes, and when the branches begin negotiating with each other through the office rather than talking directly, the office has become the medium of the disagreement rather than the solution to it.
When the office has begun to serve itself
Every organisation eventually generates work for its own staff. In a family office this happens quietly, because the family cannot easily judge whether the work is necessary. Reports are produced monthly because they were always produced monthly. An entity is maintained because unwinding it would take a year. A mandate is monitored that has not changed in five years.
A reliable diagnostic is the ratio of reporting to decisions. If the office produces a substantial reporting pack each quarter and the family takes two decisions a year, the pack is not decision support. It is an activity that justifies the structure that produces it.
A second diagnostic is what happens when a key person is absent for a month. If nothing suffers, the function is documentation rather than judgment, and documentation can be bought. If everything suffers, the family has a concentration risk that a wind down would actually solve rather than create.
None of this is a criticism of the people involved, who are usually loyal and competent and did not design the arrangement either. It is a reason to have the conversation with them rather than about them.
Winding down the office is not liquidating the wealth
The most common objection to closing a family office is the fear that closing it means selling everything. It does not. The office is the administrative and decision making layer. The wealth sits in holdings, portfolios, property and operating interests, and most of that can continue untouched while the layer above it is dismantled.
What actually ends is a set of employment relationships, a set of premises and systems, and usually a number of intermediate entities that existed only to organise the work. What continues is ownership. Confusing the two produces paralysis, because the family believes it is being asked to undo thirty years of building when it is being asked to stop paying for a coordination function it no longer uses.
The correct framing is substitution. Every function the office performs is either discontinued, absorbed by a family member, or purchased externally. Consolidated reporting can be bought. Bookkeeping and tax compliance can be bought. Custody and investment management already sit outside. Governance, meaning who decides what, does not disappear and must be rehoused deliberately rather than allowed to lapse.
What survives the office
Three things should be preserved in almost every case. The first is the top holding structure, if there is one and if it still works. A holding that owns the participations, receives the income and holds the history is usually the cheapest entity in the group to maintain and the most expensive to recreate. Dissolving it because it feels like part of the office is a common and costly error.
The second is the records. Book 2 of the Burgerlijk Wetboek requires books, records and data carriers to be kept for seven years, and that obligation does not evaporate because the entity is closing. A second and separate duty sits in the General Tax Act, whose article 52 obliges the persons subject to it to retain those data carriers for seven years unless tax legislation provides otherwise, so the two run alongside each other and satisfying one does not discharge the other. In practice the family needs far more than the legal minimum: acquisition costs, historic valuations, share transfer documentation, deeds, tax rulings and correspondence with authorities. A record that cannot be found when a disposal is taxed years later is worth precisely nothing.
The third is the relationships. Bankers, notaries, auditors and advisers who know the structure carry knowledge that is not written down anywhere. Where those relationships were held personally by a departing employee, they need to be transferred formally, with introductions, before the departure rather than after it.
The order in which things come apart
Sequencing is where wind downs go wrong. The instinct is to start with the visible cost, which is people, and that is exactly the wrong first step, because the people are the only ones who know how the structure works. Terminate first and the family spends the next two years paying advisers to reconstruct what an employee could have explained in an afternoon.
A workable order begins with a decision and a document. The family agrees, in writing, what the end state looks like: which entities survive, who decides what afterwards, and by when. Then the inventory: every entity, account, contract, licence, lease, mandate and filing obligation, with an owner assigned to each. Then the substitutions are put in place and tested while the office still exists.
Only then do the entities come apart, and they come apart from the bottom. Dormant subsidiaries first, then operating layers, then the intermediate holdings, then the office entity itself. Employment matters are handled in parallel and early in the discussion, with retention arrangements for the people needed until the end, because an adviser who learns of the wind down from a colleague will be gone before the inventory is finished.
The Dutch mechanics of dissolution and liquidation
For Dutch entities the framework sits in Book 2 of the Burgerlijk Wetboek. Dissolution is resolved upon, after which the entity continues to exist for the purposes of liquidation and ceases to exist at the moment the liquidation ends, under article 2:19. Where the entity has no assets remaining at the moment of dissolution, article 2:19 states that it ceases to exist at that moment, which is the fast route open only to an entity that has nothing left when the resolution is taken.
That route now carries a disclosure obligation of its own. Where an entity is dissolved by resolution and ceases to exist at the same moment, article 2:19b requires the board to file with the registers, within fourteen days of the dissolution, a balance sheet and a statement of income and expenditure for the financial year in which the entity was dissolved and for the preceding year if no annual accounts have yet been published for it, a description of the cause of the absence of assets and, where applicable, of how any assets were realised and the proceeds distributed and of why creditors remained wholly or partly unpaid, together with any annual accounts of earlier years still subject to an unfulfilled publication obligation. The board must then inform the creditors of those filings in writing without delay. The Chamber of Commerce adds that failing to file, or filing incompletely, is punishable as an economic offence and can support an application to disqualify the directors. Those provisions entered into force on 15 November 2023 and are due to lapse on 15 November 2027, a date already postponed once by royal decree while a permanent regime is prepared.
Where there are assets, a liquidator realises them, settles the liabilities and prepares an account and a distribution plan, which are deposited and open to objection by creditors and entitled persons for a period fixed by article 2:23b of the Civil Code before the distribution is made. The plan is required only where two or more persons are entitled to the surplus. The account and the plan are deposited at the office of the registers in which the entity is registered and at the entity’s own office, the liquidator announces in a newspaper where and until when they may be inspected, the documents lie open for inspection for two months, and within two months of the deposit and its announcement any creditor or entitled person may petition the court to object. The Chamber of Commerce sets out the same steps for practitioners and notes that the distribution plan may be implemented after two months if no objection is filed. Liquidation ends at the moment there are no assets left that are known to the liquidator.
Deregistration is not the same as leaving the tax system
The tax side runs on its own timetable, and it does not end when the register entry does. The Chamber of Commerce passes the deregistration to the Tax Administration, but it also warns in the same breath that returns must continue to be filed for as long as they appear in the business portal, and that the Tax Administration will write separately to say when the final turnover tax return is due. The Tax Administration makes the point from its own side: while the deregistration is being processed, a request to file a corporate income tax return still has to be answered, including for a year in which the entity carried on no activity at all.
One item genuinely has to be dealt with before the registry step rather than after it. A registration in the one stop shop for turnover tax has to be ended before deregistration at the Chamber of Commerce and before the electronic identification credentials lapse, failing which the last mandatory returns may no longer be capable of being submitted, with a risk of exclusion from the scheme for two years.
The fiscal unity deserves a model of its own before the resolution is passed rather than a note afterwards. Where an asset was transferred from one member of the unity to another at a time when its market value exceeded its book value, it is restated at market value immediately before the member leaves, and article 15ai of the corporate income tax act allows three escapes and no others: a transfer made in the ordinary course of business appropriate to the nature and size of the parties, a transfer of a business or of an independent part of one against an issue of shares by the transferee where at least three calendar years have since passed, and the passage of at least six calendar years. It is also worth confirming which unity is being unwound, since a fiscal unity for corporate income tax is not the same thing as a fiscal unity for turnover tax.
The last piece is the least interesting and the most likely to be neglected. Someone must be responsible for the archive after the entities are gone, for keeping it accessible, and for knowing what is in it. Book 2 already names that person: the books, records and data carriers of a dissolved legal person must be kept for seven years after it has ceased to exist by a custodian designated in or under the articles or by the general meeting, and that custodian must report his name and address to the registers within eight days of the custody beginning. The registers themselves retain the entity’s data for ten years after it ceases to exist. A wind down that is executed cleanly and leaves no custodian of the history has solved a cost problem and created an information problem that the next generation will inherit. The office should close. The memory should not.
This article is informational and does not constitute tax, legal or investment advice. Each engagement is subject to scope and applicable regulation.