A Family Creed Is Not Enough — You Also Need Family Law

Why do some family rules still hold after ten generations, while others cannot outlast even one? Most people will answer: the difference lies in how thoroughly the clauses are drawn. Not necessarily. Sheng Xuanhuai’s testament was drafted with far more precision than Fan Zhongyan’s thirteen rules, and it was taken apart in little more than a year; the Fan family’s thirteen, with no shortage of holes, governed for nine hundred years. The fate of a family’s rules is not decided by the fineness of the drafting.

The year 1064, Kaifeng. Fan Chunren, a palace censor, sent up a memorial to Emperor Yingzong of Song. What it reported was his own family’s business.

His father, Fan Zhongyan, had been dead for twelve years. In his last years the father had bought a thousand mu of charitable fields in Suzhou, roughly 160 acres, and drawn up thirteen rules with his own hand to provide for the whole clan. Now something had gone wrong. The memorial put it plainly: “Among the sons of the several branches there are now those who do not follow the rules. The prefecture and the county have no imperial statute to invoke, and the family itself is hard pressed to bring them to account. Over five to seven years, it has been drifting toward ruin.” Someone in the family was breaking the rules; the government had no legal text to cite; the family could not discipline him on its own; and within a few short years, the Charitable Estate (yizhuang) was plainly headed for ruin.

The thirteen rules said how to punish. They did not say what to do when the punishment would not stick.

Fan Chunren asked the emperor to close the gap: let the local courts accept cases against estate members who broke the rules. Yingzong approved. That same year, Fan Chunren had his father’s thirteen rules carved in stone and set at the foot of Tianping Mountain.

One set of family rules: the father wrote the clauses, and the son added the teeth. For the next eight hundred years and more, the rules kept working.

A creed is not a law

The texts a family hands down come in two kinds.

What Zeng Guofan left is the first kind. The “eight foundations” of his family teaching, among them: study founded on exact scholarship, the serving of parents founded on gladdening their hearts, the keeping of a household founded on never rising late, the holding of office founded on never taking money. Add the three characters he pressed on his children: filial devotion, diligence, forbearance. Add the eight-character formula handed down from his forebears, and nearly fifteen hundred family letters besides. The Zeng household recited the ancestral teachings at dawn every day; in two hundred years the family produced more than two hundred figures of note, and the teachings can claim no small part of the credit.

Yet search these texts from end to end and you will not find one clause on how property is to be divided, not one on how the man in charge is to be chosen or a rule-breaker punished. Zeng Guofan did not need them: he laid up no silver, and the family had little common property to govern.

What Fan Zhongyan left is the second kind. In the thirteen rules there is not one sentence of high principle; every clause transacts business. How rice is issued: branch by branch and head by head, one sheng a day for every mouth, no distinction between the sons of wives and the sons of concubines. How money is paid: thirty strings of cash for marrying out a daughter, twenty if she remarries; twenty for taking a wife, nothing for taking a second. How the accounts are squared: each branch holds a passbook for the rice it draws, the steward holds a master ledger, and at the end of every month the two books are set against each other. How the people are checked: should the steward dare divert funds, “any branch may detect it and compel him to make the sum good” — every branch had the right to expose him, and whatever was verified he repaid in full. Even the famine year had its place in the clauses: in good years, store up two years’ grain; when the stores fall short, funerals are paid before weddings.

The Zeng texts teach a person how to conduct himself. The Fan text stipulates how money moves, how accounts are squared, and how a violation is made good. The first kind is jiaxun, the family creed: it answers what we believe. The second is jiafa, the family law: it answers what happens when something goes wrong. A family needs both. Most families have only the first.

What one man writes, and what a family writes

April 1916, Shanghai. As he lay dying, Sheng Xuanhuai dictated his last will: the estate’s yearly income to be split into ten shares, five for good works, five divided among his five branches; the principal held in common, never to be broken up. He had designed the machinery of execution as well: the Zhishantang (Hall of the Highest Good) Council, with eight eminent men as councilors and another eight as advisers, and a board that met every week, major decisions requiring the consent of all.

For precision, this design stood far above the Fan family’s thirteen rules. As for its end: the five branches balked at every turn; within months the general manager and his deputy had both resigned and the council had stalled; after little more than a year the family compromised, and it was agreed that half the estate would be divided among the five branches by the drawing of lots.

The fault lay not in the clauses but in where the clauses had come from. From first to last, the rules were fixed by one man; from first to last, the sons of the five branches were people being informed, not people being consulted. A rule no one took part in making is a rule no one truly owns.

The sixteen articles of Kikkoman, the Japanese soy-sauce house, came by a different road. Eight families merged in 1917; the family constitution was not finalized until 1926, eight or nine years in between. The drafting was itself one long negotiation: who would hold the pen, whose wording would prevail, whether some clause might hide one branch’s private advantage; item by item, all of it had to be talked into agreement. The sixteen articles that came out of that talking, all eight houses accepted.

The contemporary example is Lee Kum Kee of Hong Kong. Around 2002, Lee Man Tat set up a family council of seven, he and his wife and their five children, with the chair rotating among them. The council met once a quarter, four days at a sitting, sixteen days in a year; arriving late cost a fine of two thousand Hong Kong dollars. Lee Kum Kee’s family constitution was argued out clause by clause at these meetings: no divorce and no extramarital affair, on pain of automatic departure from the board; fifth-generation members who want to join the company must first work outside it for three to five years, then apply and be assessed by exactly the standards set for outsiders; the chairman must be a member of the family, while the chief executive may be hired from without. Amending the constitution takes the approval of more than seventy-five percent of the family council.

The paper’s teeth

Rules argued out and written down are still not the end of it.

Fan Chunren’s way was to borrow the teeth of the state: petition the throne, and join a private family’s rules to the enforcing power of the government. The modern family’s way is to take the clauses of its constitution that matter most and land them in documents that carry legal force.

When the Ford family took its company public in 1956, it wrote the family’s bottom line into the corporate charter: Class B shares may be held only by family members, their descendants, and the family trusts, and once transferred to an outsider they are downgraded automatically into common stock; so long as the family’s holding of B shares does not fall below the number fixed in the charter, the family together commands forty percent of the votes. This is not a gentlemen’s agreement but a charter provision, and in seventy years no one has shaken it. Hermès’s H51 is the same: an agreement binding the family to act in concert, under which 50.2 percent of the shares, out of the family’s 62.8 percent, are locked for twenty years, in black and white, with liability for breach.

The profession keeps a standard layering for all this: vision and values are written into the constitution and held up by conviction; clauses that touch ownership and material interest must land in legally enforceable documents.

In PwC’s 2023 global survey of family businesses, seventy-four percent of the families said their members trusted one another; those that had built a mechanism for resolving conflict came to nineteen percent. The other eight in ten, when trouble comes, can only improvise.

After the writing

The most common way for a constitution to die is to be finished and locked in a drawer.

The Fan Charitable Estate’s nine hundred years were walked out by keeping the rules with one hand and amending them with the other. In 1083, after someone found a loophole in the tenancy of the land, the clan added a clause: no one may take a tenancy on the charitable fields, and doing it under a borrowed name carries the same penalty. When the estate was rebuilt in 1196, a scholarship for the imperial examinations was added: subsidies for members traveling to sit the examinations, half rates for those who sat repeatedly, and money returned by anyone who registered and then, without cause, did not go. Even “registering and not sitting” had been thought of. By the late Qing the rules had accumulated past one hundred, and Fan Zhongyan’s original thirteen still stood at the head of the book.

Of the Mars family’s Five Principles, the one at the core, Mutuality, began as a letter Forrest Mars wrote to the company in 1947; in the 1980s the third generation compiled it into text. By the fourth generation the text has not moved. What moves is the interpretation: the same principle, put to work on questions the generation before never met.

How to draft a family constitution

Gather up the experience of these houses and bring it down to the level of operations, and it comes to five essentials.

First, distinguish the creed from the law, and give each its own text. Values, family culture, and expectations of conduct belong to the creed layer: they are written in narrative language and carry no penalties. Employment, equity, dividends, exit, and the resolution of conflict belong to the law layer: every clause there must be executable and adjudicable. The two layers may live inside a single constitution, but the nature of each provision must be unmistakable. Writing matters of law in the language of the creed is a common cause of failure: the clauses hold expectations but no standards, and when trouble comes there is nothing to judge by.

Second, form a drafting group that spans the branches, and allow it a generous timetable. Five to eight people is the right size, covering every branch and every generation. The rhythm common in the field is one concentrated session every six to eight weeks, and twelve to eighteen months to a finished text. The drafting is itself the building of consensus, and its value is not less than the text’s; a document handed down unilaterally by the family head, however complete its clauses, can hardly escape the Sheng family’s ending. Where disagreements run deep, an outside facilitator can be engaged to chair the sessions and see that every side states its position in full.

Third, four kinds of procedural clause must go in. One, an employment policy for family members: the preconditions for entering the business and the standards by which candidates are assessed. Lee Kum Kee’s practice is three to five years of outside work first, with application and assessment held to exactly the standards used for non-family hires. Two, equity and exit: who is qualified to hold shares, how the shares are priced at exit, and whether the family holds a right of first refusal to buy them back. Three, conflict resolution and the right of interpretation: disputes go first through the family’s internal procedure, and only when that fails to outside legal channels; where the final right to interpret the clauses rests should be stated in writing. Four, an amendment procedure: set the thresholds for proposing and passing amendments; Lee Kum Kee set passage at more than seventy-five percent of the family council. A constitution without an amendment clause either petrifies or gets rewritten at will.

Fourth, land the key clauses in legal documents. In most jurisdictions the family constitution itself carries no enforceable power. Its clauses touching ownership and major interests must each come to rest in arrangements that do: equity and dividend provisions written into the shareholders’ agreement and the company’s articles, the walling-off of assets accomplished through trusts, agreements on marital property set down in prenuptial contracts, the employment policy folded into the company’s personnel rules. The constitution binds consensus; the legal documents supply force; neither layer can be spared. Ford’s Class B provision sits in the corporate charter, and Hermès’s H51 is a shareholders’ agreement. The hardest rules of both houses were never left at the level of the constitution.

Fifth, set a mechanism for rereading and revision. At least once a year, the constitution should be brought out at the family meeting and read against the year’s actual affairs, to check how it has been kept; every three to five years it should be reviewed whole, and the clauses that have gone stale updated by the amendment procedure. A constitution written once and never touched again is little different from no constitution at all.

A closing thought

Zeng Guofan said: “A family’s way lasts not on the strength of a moment’s office and rank, but on the strength of family rules that reach far.”

Of the “family rules” in that sentence, the Zeng family built half, and with the creed alone raised two hundred years of talent. The Fan family built the whole of it, creed and law standing side by side, and walked nine hundred years.

To judge whether a set of family rules will hold, do not look at how handsomely it is written. Look at three things: whether the whole family argued it into being together; whether it has been joined to the teeth of the law; and whether, in each year after the writing, someone still takes it out and reads it against the family’s affairs.

Fan Chunren’s memorial supplied the second.

---

Case sources: The Nine-Hundred-Year Trust: Fan Zhongyan and the Fan Charitable Estate; The Plow and the Book: The Family Wisdom of Zeng Guofan; Spend the Interest, Never the Principal: The Last Will of Sheng Xuanhuai; Eight Families, One Brew: Kikkoman’s 360-Year Experiment in Restraint; A Bottle of Sauce and a Book of Law: Five Generations of Lee Kum Kee; The Watch Still Runs: Five Generations of Fords and the Art of Repair; The Saddle Stitch: Six Generations of Hermès and the War for Its Soul; The Candy Empire That Refused to Go Public

Academic references

• Craig E. Aronoff & John L. Ward (2011) Family Business Governance: Maximizing Family and Business Potential, Palgrave Macmillan
• PwC (2023) Global Family Business Survey: Transform to Build Trust
• Ivan Lansberg (1999) Succeeding Generations: Realizing the Dream of Families in Business