Handing It On

Volume I — Transition: From Here to the Living Economy

Nine movements, one successor.

THE PLATE

A brass key resting on an old wooden table, a shaft of sunlight falling across it. — Plate I.11The Second Set of Keys.A handover is not a ceremony. It is the moment the building stops needing you, and the only way to find out whether it has arrived is to be there and not intervene.

THE LETTER

You began this volume with a Monday. Ten chapters ago, the question was what one person does inside a real organisation, with a real quarter and a real signature limit, to move it toward an economics that builds its stocks instead of drawing them down. You found the positive core and costed it. You built the first instrument. You read a balance sheet as a living system, ran a pilot that paid, brought the money with you, opened a second circle, got through the stall, shook hands with an institution, and grew the thing without losing what made it worth growing.

This chapter is the last one and it asks the only question that has been waiting underneath all the others: does any of it survive you?

Not as sentiment. As structure. Because the honest position, at this point in the volume, is that everything you have built is currently load-bearing on one person's continued presence and one person's continued preference — and a preference is not a property right. The next owner is not bound by your convictions. The next chief executive is not bound by your memos. A buyer can undo eleven chapters of work in a single afternoon and be entirely within their rights, because you never wrote the rights down.

So the work of this chapter is to write them down. Three things, in this order: the memory — what you know that nobody has recorded; the mandate — who decides after you and under what constraint; and the ownership — the document that makes the first two enforceable rather than hoped for.

We will look at the structures that have actually done this, as filed, not as described in a press release: Ernst Abbe's foundation statute of 1889, Robert Bosch's separation of money from votes, Patagonia's two trusts of September 2022, the British employee ownership trust, the American ESOP, and the perpetual purpose trust and the one line of Delaware code it depends on.

And we will do the arithmetic that most of the enthusiastic literature will not do for you, which is what a lock costs. It costs something real, it has starved enterprises that did not compute it first, and the number that tells you whether your enterprise is one of them takes about four minutes to work out.

— The Editors

DISCOVERY

What is already working

Start with the oldest one, because it is the one everybody else copied.

Carl Zeiss, Jena, 1889. Ernst Abbe — the physicist who gave optics the Abbe number and the sine condition — inherited the Zeiss works on the death of his partner and did something with them that had no precedent. He wrote a statute. The Carl-Zeiss-Stiftung took ownership of the firm, and the statute, substantially settled by 1896, bound it: profit shares for employees, a limit on the ratio between the highest salary and the lowest, an eight-hour day, paid holiday, a pension fund, and a provision that no individual could ever again own the enterprise. Abbe, who could have become one of the richest men in Germany, made himself an employee of his own foundation. The structure is one hundred and thirty-six years old and Zeiss is still there.

Robert Bosch, Stuttgart, 1937 and 1964. Bosch wrote a will and left instructions. The structure that came out of it in 1964 is the single most instructive ownership document in industrial history, because of what it separates.

  holder                                        capital %   votes %
  ------------------------------------------------------------------
  Robert Bosch Stiftung (charitable foundation)     94.00      0.01
  Robert Bosch Industrietreuhand KG (trust)          0.00     93.00
  Bosch family                                       6.00      7.00

Read that table twice. The foundation owns nearly all of the money and casts almost none of the votes. The industrial trust casts nearly all of the votes and owns none of the money. The foundation cannot sell the company because it cannot outvote anyone; the trust cannot enrich itself because it holds no capital. Each is the other's constraint, and neither can defect. Bosch is a group with revenue in the region of ninety billion euros, and no shareholder alive has a personal financial reason to break it up.

Patagonia, Ventura, September 2022. The Chouinard family transferred the company. Two percent of the shares — the entire voting class — went to the Patagonia Purpose Trust, a perpetual purpose trust whose stated purpose is to keep the company's values in force. Ninety-eight percent — the non-voting class — went to the Holdfast Collective, a 501(c)(4).

The detail that matters is the one least reported. The family paid about $17.5 million in gift tax on the transfer of the voting class, and because a 501(c)(4) is not a charity, the gift of the other ninety-eight percent earned no charitable deduction at all. They also did not receive the roughly three billion dollars a sale would have produced, on which they would have owed some seven hundred million in tax. The structure cost them money. It did not save them any, and anybody arguing this is a tax strategy has not read the filing.

John Lewis, London, 1929 and 1950. John Spedan Lewis settled the shares of his father's business into trust for the people working in it. Partners, not employees; a written constitution; a partnership bonus paid every year from

We will come back to John Lewis in the arithmetic, because it is also the

clearest case of what a lock costs.

Organically Grown Company, Oregon, 2018. A produce distributor converted to the first American perpetual purpose trust structure of its kind, with capital provided partly through RSF Social Finance. It is here because it is the proof that the structure is available to firms of ordinary size, not only to Stuttgart.

And the scale of the question. Project Equity's estimate is that around 2.9 million businesses in the United States are owned by people over fifty-five, employing on the order of thirty-two million people. Every one of those is a succession event waiting to happen, and the default outcome, on John Ward's long-standing figures, is that thirty percent of family businesses reach the second generation, twelve percent the third and three percent the fourth.

Those percentages are usually quoted as a warning. Read them the other way, as conditional probabilities, and they are a design brief: the second handover is harder than the first, and the third is hardest of all. Whatever you build has to survive not you, but your successor's successor.

THE ARITHMETIC

What works, what does not, and where the line sits

The Balenciaga cut, and it is the whole chapter: succession is not a governance decision. It is a financing decision, and the number is computable before you file anything.

Here is why. An enterprise grows out of three sources: retained profit, new debt, and new equity. Every steward-ownership structure in the Discovery movement above removes the third one permanently. That is not a side effect — it is the point of the lock. But it means the enterprise's maximum sustainable growth rate is now fixed by arithmetic, and the arithmetic is Robert Higgins's, from 1977:

        g*  =  ROE  x  b

  ROE  return on equity, on beginning-of-period equity
  b    retention ratio — the share of profit not distributed

That is it. A firm earning twelve percent on equity and retaining every penny of it can grow at twelve percent a year, forever, and not one point faster without borrowing more or selling something. Retain seventy percent instead and the ceiling falls to 8.4 percent.

    ROE    retention     ceiling
  --------------------------------
     8%         70%         5.6%
    12%         70%         8.4%
    12%        100%        12.0%
    16%         70%        11.2%
    20%        100%        20.0%

Now put the ceiling beside the growth your business actually has to hold to keep its position, and you have the only test that matters:

      ROE x b   >=   the growth rate the business must hold

At a twelve percent ceiling: mature retail needing three percent a year is comfortably financeable from inside. Branded consumer at seven percent is fine. Precision optics at eleven is tight and workable. Automotive electronics at fifteen percent is short by three points a year and semiconductor fabrication at twenty-two is short by ten. A shortfall of three points a year, compounded, is not a strategic disagreement. It is the slow loss of a market.

Which brings us to the honest negative, and there are two.

The first is John Lewis, and it is a real cost paid by real people. The Partnership stated an investment requirement of about a billion pounds. Its profit before tax in the year to January 2024 was fifty-six million, after a loss before tax of two hundred and thirty-four million the year before.

  £1.0bn required  /  £56m annual profit  =  17.9 years

Eighteen years of profit to fund a five-year plan, with no equity to issue because the shares are in trust and that is the entire arrangement. The partnership bonus, paid every year from 1953, was not paid in 2022, 2023 or 2024. In 2023 the chair publicly explored selling a minority stake; the structure did not bend, and the capital did not arrive either. That is what a lock costs when the ceiling is below the requirement, and no amount of conviction about ownership makes the eighteen years into five.

The second negative is sharper, and it will decide more transactions than the first. The British employee ownership trust carries a genuine tax relief: under TCGA 1992 s.236H, introduced by the Finance Act 2014, a seller of a controlling interest to an EOT pays no capital gains tax at all, against a main rate of twenty-four percent. That sounds decisive. Do the present value and it is not.

An EOT rarely pays cash on the day. It pays out of future profits, over years, unsecured, and typically at a discount to a trade sale because there are no synergies and no competitive tension. Take a ten million pound business, a twenty percent discount, and a seller discounting deferred money at eight percent:

  trade sale, £10.00m, CGT 24%                 ->  £7.60m, today
  EOT at £8.00m, paid over 1 year   PV £7.41m  ->  -£0.19m
  EOT at £8.00m, paid over 3 years  PV £6.87m  ->  -£0.73m
  EOT at £8.00m, paid over 6 years  PV £6.16m  ->  -£1.44m

And the break-even discount — the most useful line in this chapter for anyone about to negotiate one:

  paid over 1 year    survives a discount of  17.9%
  paid over 3 years   survives a discount of  11.5%
  paid over 5 years   survives a discount of   4.8%
  paid over 6 years   survives a discount of   1.4%
  paid over 8 years   survives a discount of  -5.8%

The tax relief is not what decides an employee ownership transaction. The deferral is. Past about five years the relief has been entirely consumed by the time value of the seller's own money, and the seller is worse off than they would have been taking a taxed trade sale on the day. This is not an argument against employee ownership. It is the instruction: shorten the payout or shrink the discount. Both are negotiable and both are usually left un-negotiated because everybody in the room is being polite about purpose.

Two more figures worth carrying. In the United States, the reason the Bosch structure was unavailable for half a century is IRC s.4943: a private foundation may hold at most twenty percent of a business. Bosch's foundation holds ninety-four — 4.7 times the American cap. The Philanthropic Enterprise Act of 2017 added s.4943(g), effective 2018, with a narrow exception on four conditions: one hundred percent ownership, not acquired by purchase, all net operating income distributed to the foundation within 120 days of quarter end, and independent operation. It was written for Newman's Own and it is available to anyone who fits it.

And the duration trap. Uniform Trust Code s.409 caps a non-charitable purpose trust at twenty-one years. A working generation is about twenty-five. The default American purpose trust expires before the first successor retires — 0.84 of a generation. Delaware's 12 Del. C. s.3556 removes the limit entirely, which is why the American conversions are Delaware trusts holding operating companies chartered elsewhere. The filing state is not paperwork. It is the instrument.

DREAM

What becomes ordinary

In the enterprise that has done this, the founder's departure is a calendar event rather than a crisis, and the reason is that the three transfers happened years apart and each was finished before the next began.

The memory went first. There is a document — it is about sixty pages and it took a hundred and twenty hours over several years — that holds the twenty standing decisions of the business and why each was taken, the ten relationships that cannot be re-created by an introduction, the five numbers that tell you whether the thing is well, and one page on what the business is for. It is read by every new director in their first fortnight. It is revised annually by whoever currently holds each decision, so it is never a relic. Nobody calls it a legacy document. They call it the operating memory, and they use it on Tuesdays.

The mandate went second. The board has a written constraint, not a written wish: a list of things it may not do without a specific consent from a specific holder, and that list is short enough to be remembered — sell the enterprise, change what it is for, change the constraint itself. Everything else, the board decides, and it decides freely, including things the founder would not have chosen. That freedom is deliberate. A successor who may only do what the founder would have done is not a successor; they are an impersonation, and they will leave.

The ownership went last, because it was the part that needed the other two to be true first. The economic majority sits with the people who do the work, through a trust that pays them and can be explained on one page. The voting control sits where it cannot be bought. And the enterprise raises the capital it needs without asking either of them for permission, because it raises it one level down — against a named asset, in a named subsidiary, with a named lender — exactly as Carl Zeiss did when it floated Meditec in 2000 and reformed its statute in 2004 so that its operating companies could reach a market at all.

Nobody in the building finds any of this remarkable. The founder comes in on Thursdays. She is asked fewer questions each year and she has stopped counting, which is the surest sign the transfer worked.

DESIGN

The structure that gets there

Three transfers, and the order is not negotiable.

Transfer one: the memory. Start now, finish before you announce anything.

Do not attempt completeness. Between nine and ten thousand non-routine judgements are made across a long tenure and roughly eighty-five percent of them are undocumented — call it eight and a half thousand decisions living in one head. You are not going to write those down and you should not try. Write four things:

The twenty standing decisions. The choices that are re-made every year and never revisited — this supplier, this margin floor, this refusal to enter that market. For each: what was decided, what the alternative was, and what would have to change for the answer to change. That last clause is the whole value. It is what turns a rule into a judgement the successor can exercise.
The ten relationships. The people for whom an introduction is not a transfer. Name them, name what is actually owed in each direction, and hand them over in person, one at a time, over a year.
The five numbers. The instruments you actually read to know whether the business is well. Not the pack. The five.
One page on what it is for. Written by you, in your own sentences, and short enough that it goes into the trust deed as a recital.

A hundred and twenty hours across an eighteen-year tenure is 6.7 hours a year. That is the entire price of the memory transfer, and it is the cheapest thing in this chapter.

Transfer two: the mandate. The constraint, not the wish.

A short reserved-matters list, held by a holder who cannot profit from breaching it. Three items is usually right: disposal of the enterprise, alteration of purpose, alteration of the constraint. Resist the fourth item, and the fifth. Every additional reserved matter converts a successor into an applicant, and the structures that have lasted a century reserve almost nothing.

Transfer three: the ownership. Separate the two rights.

This is Bosch's insight and it is available to firms of any size. Economic rights and control rights are different things and they should be held by different parties with different interests. Control goes to a perpetual purpose trust — Delaware, s.3556, so it does not expire — holding a single non-economic share with the reserved-matters veto and nothing else. Economics go to the people: an employee ownership trust in the United Kingdom, an ESOP in the United States, a cooperative where the law supports it, a foundation where s.4943(g) fits.

And the capital answer, which is the part most schemes omit. Compute ROE × b before you file. If it clears your required growth, lock it. If it does not, do not choose a different trust — the trust is not the problem. Do what Zeiss did: lock the parent, raise capital one level down, against a named asset, never against the purpose. A locked parent with a financeable subsidiary is a solved problem. A locked parent with no subsidiary and a shortfall of three points a year is John Lewis.

DESTINY

How it holds when you stop pushing

It holds on three things and it fails on four, and both lists are short enough to check annually.

It holds because the constraint is in a deed, not a culture. Cultures are inherited by whoever is hired next. Deeds are not.

It holds because the reserved matters are few. A structure that permits nothing is broken within a decade by people who need to run a business. A structure that reserves three things can be honoured indefinitely because honouring it costs almost nothing day to day.

It holds because the capital route was designed at the same time as the lock. The two decisions are one decision. Taken separately, the lock always comes first — it is the emotionally satisfying one — and the capital question arrives four years later as a crisis.

Here is where it fails. It fails when the founder stays on the trust board and the successor discovers that the mandate came with a shadow. It fails when the purpose in the deed is written so broadly — "to act in the long-term interests of the company" — that it constrains nothing and any buyer's lawyer can satisfy it. It fails when the trust has one trustee, because a single trustee is a single point of capture and every century-old structure has at least three with staggered terms. And it fails, most often and most quietly, when the enterprise's required growth rises above its ceiling and nobody recomputes the inequality, so the lock stops protecting the purpose and starts strangling the thing the purpose was about.

Recompute ROE × b every year. It is the annual health check of the ownership structure, it takes four minutes, and no standard governance calendar contains it.

DELIGHT

What it feels like

There is a specific and underrated pleasure in being asked fewer questions.

It arrives slowly. For the first months after a handover the calls keep coming and you answer them, and each one is quietly gratifying in a way you would not admit. Then there is a week when three things are decided without you that you would have decided the same way, and a week when something is decided without you that you would have decided differently — and it works. That second week is the one that changes how you feel about the whole enterprise, because it is the first evidence that you built something rather than performed something.

And there is the pleasure of the document itself. Writing the twenty standing decisions is strange and absorbing work: you sit down expecting to record what you know and you find out what you know, which are not the same activity. Half the entries surprise you. Two of them turn out to have been wrong for six years and you fix them on the spot, because writing down what would have to change for this answer to change is a test the answer can fail.

The best of it is the plainest. Somebody who was not there when you started can now explain, accurately and in their own words, what this place is for. They did not learn it from you. They read it. And it survives you, which was the point.

OPERATIONALIZE THIS

At the level of finance — the transfer instrument itself

This movement is not a description of an instrument. It is the instrument, in the order the documents get signed.

The structure: a stapled succession deed. Three components, executed together, so that no one of them can be completed while the others are pending.

Component one — the Purpose Trust. A non-charitable perpetual purpose trust, settled in Delaware under 12 Del. C. s.3556 (or Wyoming, New Hampshire or South Dakota; the UTC default of twenty-one years is fatal and must be escaped deliberately). It holds one non-economic share with a veto over exactly three reserved matters: disposal of the enterprise or any controlling interest, alteration of the stated purpose, and alteration of this list. It holds no other rights, receives no dividend and has no board seat. Trustees: three to five, with staggered terms, at least one with no prior relationship to the founder, and a stated succession method for trustees themselves — the omission that has voided more of these than any other.

Component two — the Economic Trust. The employee vehicle. In the United Kingdom, an EOT under TCGA 1992 s.236H acquiring more than fifty percent; note the Autumn Budget 2024 reforms, which require trustees not to be controlled by the former owners and extend the relief clawback window to four tax years after disposal. In the United States, an ESOP under ERISA, with IRC s.1042 rollover for a C-corporation seller of thirty percent or more, and the S-corporation structure where the plan is the sole shareholder — a 100 percent ESOP-owned S corporation has a tax-exempt shareholder, so roughly 29.6 percent of pretax income stays in the business every year and services the debt that bought it. Where a foundation is the right holder, IRC s.4943(g) permits 100 percent ownership on its four conditions.

Component three — the Operating Memory, as a condition precedent. This is the piece nobody staples in and it is why it belongs here. The memory document — twenty decisions, ten relationships, five numbers, one page of purpose — is a deliverable of the transaction, certified complete by the incoming chair, and release of the first tranche of consideration is conditional on it. An unpaid promise to write it later is not written later.

The balance-sheet treatment. For the company, deferred consideration to an EOT or a leveraged ESOP is a liability at amortised cost, and the interest unwind runs through finance costs — model the covenant headroom for the full deferral term before signing, not after. For the seller, deferred unsecured consideration is an asset of uncertain timing; discount it at your own rate, not at zero, which is the single most common error in these transactions. Where a share is issued to the purpose trust for nil consideration, it is a non-economic instrument and should be disclosed as a control right in the notes rather than recognised as equity of value.

The counterparty. Internal first: the trustees of the economic vehicle are your counterparty, and they owe duties to the beneficiaries, not to you. Appoint an independent trustee and let them negotiate against you properly — a transaction where the seller picks both sides is the transaction the tax authority reads first. Then the lender: for a leveraged EOT or ESOP, a bank that has done one before is worth a point of margin over one that has not.

The number that decides it. One inequality, on the front page:

        ROE  x  retention ratio    >=    required growth rate

Worked, for the example firm in the figures module: 0.14 × 0.85 = 11.9% against a required 9.0 percent — headroom of 2.9 points, so lock it. If the headroom is negative, the answer is not a different trust. It is Zeiss's answer: lock the parent, finance one level down.

The first ninety days.

Day	Action	Artifact
1–15	Compute `ROE × b` against required growth for five years	The one-page capital test
16–30	Draft the purpose statement and the three reserved matters	The recital
31–45	Choose the filing state; instruct trust counsel	Term sheet
46–60	Independent valuation; model the deferral and its PV	The break-even discount
61–75	Appoint independent trustees; they instruct their own adviser	Trustee consents
76–90	Begin the operating memory; twenty decisions listed, five drafted	The memory, opened

The memory is begun in the first ninety days and finished before completion. It is the only part of this that cannot be bought, drafted by counsel, or done in a hurry — and it is the part that makes the rest of it worth signing.

APPRECIATIVE QUESTIONS

Twelve, for a room

Discovery — what is already working

Think of a handover in this organisation that went unusually well — a role, a client, a piece of craft. What did the person leaving actually do, and could we describe it precisely enough to do it again?
What do we already do here that would continue perfectly well if every one of us left tomorrow? What makes those things different from the rest?
Which decisions in this business are genuinely well documented, and who wrote them down — what made that person bother?

Dream — what becomes possible

Imagine somebody joining in ten years who never met any of us. What would we want them to be able to read in their first fortnight, and what would it say?
If the three things this enterprise may never do were written into a deed tomorrow, what would they be — and what freedoms would we be deliberately leaving to whoever comes next?
If ownership sat with the people doing the work, what is the first conversation that would be easier, and who would be having it?

Design — what we build

What is the growth rate this business must hold to keep its position, and what is the highest rate we can finance from inside? Who here knows both numbers?
Where in this group would we raise capital one level down, against a named asset, without touching the centre?
Who are the ten relationships that cannot be transferred by an introduction, and what would a proper handover of one of them look like?

Destiny — how it holds

What would have to be true for this constraint to still be honoured when nobody remembers who wrote it?
Who checks the ownership structure each year, in the way we check the accounts — and if nobody does, whose calendar should it go into?
What is the first sign we would see that the structure had begun to strangle the thing it was built to protect, and who would be free to say so?

WORKS CITED

Abbe, E. (1896). Statut der Carl-Zeiss-Stiftung. Jena. (Statute as revised 1896; reformed 2004.)

Blasi, J. R., Freeman, R. B. and Kruse, D. L. (2013). The Citizen's Share: Reducing Inequality in the 21st Century. Yale University Press.

Chouinard, Y. (2022). "Earth is now our only shareholder." Patagonia, Inc., 14 September 2022.

Employee Ownership Association. The Employee Ownership Report and annual sector figures. Brough, UK.

Gelles, D. (2022). "Billionaire No More: Patagonia Founder Gives Away the Company." The New York Times, 14 September 2022.

Higgins, R. C. (1977). "How Much Growth Can a Firm Afford?" Financial Management, 6(3), 7–16.

HM Revenue & Customs. Taxation of Chargeable Gains Act 1992, ss. 236H–236U, as inserted by the Finance Act 2014 and amended following the Autumn Budget 2024.

Internal Revenue Code, 26 U.S.C. §§ 512(e)(3), 1042, 4943 and 4943(g) (Philanthropic Enterprise Act of 2017).

John Lewis Partnership plc. Annual Report and Accounts, years to January 2023 and January 2024. London.

Kelso, L. O. and Adler, M. J. (1958). The Capitalist Manifesto. Random House.

National Center for Employee Ownership. ESOPs by the Numbers. Oakland, CA. (Counts derived from Form 5500 filings.)

Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press.

Project Equity. The Case for Employee Ownership: Small Business Closure Crisis. Oakland, CA.

Robert Bosch GmbH. Annual Report, and Robert Bosch Stiftung, Ownership Structure. Stuttgart.

Uniform Law Commission. Uniform Trust Code, § 409 (Noncharitable Trust Without Ascertainable Beneficiary).

Veblen, T. (1904). The Theory of Business Enterprise. Charles Scribner's Sons.

Veblen, T. (1923). Absentee Ownership and Business Enterprise in Recent Times. B. W. Huebsch.

Ward, J. L. (1987). Keeping the Family Business Healthy. Jossey-Bass.

Delaware Code, title 12, § 3556 (Trust for a Noncharitable Purpose).

Note on figures. The Bosch capital-and-vote split, the Patagonia transfer arithmetic, the Higgins sustainable-growth ceiling, the John Lewis investment ratio, the EOT present-value and break-even-discount tables, the ESOP retention figure, the s.4943 multiple, the s.409 duration and the generational survival rates are all computed in lib/verify/I_11.py and reproducible with python3 lib/verify.py I.11. The EOT discount rate, deferral periods and trade discount are stated assumptions and are labelled as such in the module; replace them with your own before using the table on a real transaction.