The Inheritance

The trade press studies the newest firms, because novelty is what it is paid to report. The House studies the oldest, for a reason that is methodological rather than romantic. Every governance framework in circulation today is a hypothesis about what will hold. These houses are the completed experiment. They are the only long-run evidence available of an institutional design that has actually been run — through the whole of the modern financial era, in real weather, with real money.

The sample is short and it should be read as short. Monte dei Paschi opened at Siena in 1472 as a charitable mount of lending and is still trading. Berenberg has operated at Hamburg since 1590, founded by two brothers who arrived as Protestant refugees from Antwerp. Hoare’s has traded in London since 1672 and at the same Fleet Street address since 1690. Metzler began at Frankfurt in 1674 as a cloth business. Rahn+Bodmer opened at Zurich in 1750 as a silk-trading house and found that its side business in bills of exchange was the better trade.

Two features of that list matter more than the dates. The first is that almost none of them began as a bank. They were goldsmiths, cloth merchants, silk traders, and charitable lenders who were trusted with money because they were already trusted with goods, and who became financial institutions by accretion. Credit followed reputation; it did not precede it. The second is the brevity of the list itself. Thousands of houses were founded alongside these, on comparable capital, in the same streets, and are gone without a trace. Survivorship here is not a bias to be corrected for. It is the measurement.

Note also what the list does not contain. There are no houses in it that grew fastest in their generation, none that were the most profitable in their city, and none that were admired for their innovations. The institutions that held those distinctions were, almost without exception, absorbed or wound up within a century of holding them. Longevity and pre-eminence are not the same prize, and the houses in this Standard are cited only for the first.

The method that follows is therefore comparative rather than admiring. Where a practice appears in one house it is noted; where it appears in all of them, in different centuries and different legal systems, it is treated as a finding.

The Source of the Plates

The houses examined in this Standard are the houses hung in The Long Gallery, the permanent exhibition of the House — fifty-seven plates in eleven rooms, in three wings, recording the objects, the systems and the people through which capital has passed since a table in the Athenian agora. The two rooms that concern this document are the last two.

Room X, The Sitters holds fourteen faces: Cosimo de’ Medici and Jacob Fugger, who assembled the machinery; William Paterson and John Law, who proposed two structures on one vision, of which one survived; Nathan Mayer Rothschild, George Peabody and Jay Cooke; Walter Bagehot, who wrote the rule; Hetty Green, who was liquid when nobody else was; J. P. Morgan, who was the lender of last resort before a central bank was; Siegmund Warburg, Paul Volcker, Muriel Siebert and Ray Dalio.

Room XI, The Houses holds fourteen institutions, hung in the order they were founded, from 1472 to 1995. Twelve are still trading; two are not. That ratio is the subject of this Standard, and the room states its own governing sentence plainly: it hangs there because the mechanism by which a house outlives its founders — or fails to — is what the room is about.

Room XI of The Long Gallery: fourteen houses, in the order they were founded, and what each is cited for. Almost every entry is cited for something the house did to its own ownership, liability or succession — a decision taken about itself rather than about a market, and settled at the beginning. That is the finding rather than an artifact of selection: the market did to all of them, more or less, the same things. Two of the seventeen are cited for failures, and both of those are examined in Section XII.

Note what the register does not contain. There are no houses here that grew fastest in their generation, none that were the most profitable in their city, and none admired for innovation. The institutions holding those distinctions were, almost without exception, absorbed or wound up within a century of holding them.

A house cannot be assessed by its balance sheet alone, because a balance sheet describes a moment and these institutions are assessed over centuries. Nor can it be assessed by its reputation, which is a lagging indicator maintained by people with an interest in it. The House reads an institution on the three planes named in the Abstract, and the reading takes the form of nine questions.

Every one of the nine can be answered on the day a firm is incorporated, and almost none of them can be answered honestly for the first time in year five. That asymmetry is why the old houses are useful to a firm that designs new ones: the questions they answered at their founding are the questions a principal is answering now, usually without noticing that a permanent answer is being given.

The eleven lessons that follow are grouped in the order of the planes. Lessons I to IV concern structure; V to VIII concern conduct; IX to XI concern time. Each closes with two obligations, because the House holds that a standard recommended to a client and not imposed on the recommender is worth precisely nothing.

The Evidence

Bordier & Cie has been run by the family that founded it at Geneva in 1844, into a seventh generation, and where its neighbours converted to corporate form it remained a partnership in which the partners are liable without limit — a position the firm treats as its principal statement about itself rather than as a legacy inconvenience. Brown Brothers Harriman, which dates itself from 1818, reached the same structure independently and holds it still: an unlimited-liability partnership among some forty partners, with deposits uninsured by design. Rahn+Bodmer at Zurich is one of the last Swiss houses of which the same is true. Berenberg has admitted fewer than forty personally liable partners in four centuries. And when Hoare’s reorganized in 1929 it became not a limited company but an unlimited one.

The Lesson

Risk is governed by the person who cannot walk away from it. No committee, no policy manual and no framework yet written produces the caution produced by an owner whose own house is pledged against the firm’s obligations. These houses did not copy one another — they are separated by oceans and centuries — and they converge on the same legal form anyway.

The modern corporate form exists to sever that link, and there are legitimate reasons for it. But the severance has a cost rarely priced at incorporation: once no one is personally exposed, the discipline must be manufactured elsewhere, in writing, with independence, and at expense — and maintained by people who are not exposed either. Note also what Bordier demonstrates about positioning: a structure becomes a market position the moment competitors abandon it, and what was ordinary in 1844 is now the whole differentiation without a word of marketing.

The Evidence

In December 1912, before a subcommittee of the United States House of Representatives investigating the concentration of money and credit, counsel put it to the most powerful private banker in America that commercial credit rests upon money or property. The answer was that “the first thing is character.” Pressed, he placed it ahead of money and everything else, and observed that money could not buy it.

The same proposition had been written into a shop front in Edo two hundred and forty years earlier. Mitsui Takatoshi opened the Echigoya store in 1673 on a principle that scandalized the trade: fixed prices, posted openly, cash only, no haggling, and no credit extended on reputation alone. Competitors sold to whoever they favoured, on terms settled by negotiation and rank.

The Lesson

Credit, as Study No. 05 established, is a permission rather than a service, and permissions are granted to counterparties rather than to collateral. Collateral is what a lender falls back on when the first judgment was wrong; it is the remedy, not the test. A house that has inverted the two has not made itself safer — it has moved the point of failure to where recovery is slowest and most public.

The Mitsui half of the lesson is the harder one, because it runs in the other direction. A fixed, posted price is a promise that the house will treat a stranger exactly as it treats a regular, and it is the single hardest promise for any institution to keep at scale. Character is examined in both directions or it is not a standard; it is a preference exercised by the stronger party.

The Evidence

Richard Hoare was apprenticed to a goldsmith in 1665, the year the plague emptied London, and took his freedom of the Goldsmiths’ Company in 1672 — the date his bank counts as its founding. He kept coin and plate safe for a fee, and the safekeeping became the business. The house he founded has never taken an outside partner; every partner today is his direct descendant, twelve generations on.

That is a refusal repeated, not a fact of history. In the decades either side of the First World War the London private banks were sold, one after another, for good reasons: Goslings into the merger that formed Barclays in 1896; Coutts to the National Provincial in 1920; Child & Co, founded 1664, in 1924; Drummonds, trading since 1717, to the Royal Bank of Scotland the same year. Hoare’s is the only one of them independent today, and the decision was taken by a generation that could not have known it would be vindicated.

The Lesson

The industry treats a declined transaction as a cost and a closed one as a result, which is why so many institutions cannot describe the business they will not write. But the houses in this Standard are defined more precisely by their refusals than by their transactions, and the refusals are of a particular kind: they are refusals of the most profitable thing available at the time.

Study No. 05 observed the same in a living structure: the private house could lawfully take the other side of its clients’ flow, and declines, because the whole relationship rests on the client’s belief that the house is not his counterparty.

The Evidence

The most consequential structural invention in the history of private banking was not an instrument. It was an arrangement of entities. From a single counting house at Frankfurt, the Rothschild family established separate houses at London, Paris, Vienna and Naples — each capitalized in its own right, each subject to its own sovereign, its own currency and its own law, and each able to trade with the others as a counterparty rather than as a branch. Nathan, the London brother, is the one the Gallery hangs, and the caption is the point: the network’s commercial advantage was information, but its structural advantage was insulation.

The Lesson

Separation of capital is a risk control, not a matter of prestige or tax. A single entity holding every activity of a house transmits any failure in any activity to all the others instantaneously and without a decision being taken. Separate entities require the failure to travel — through a guarantee, an intercompany loan, a shared licence or a common counterparty — and each of those channels is a thing that can be examined, capped and refused in advance.

Study No. 05 reached the same conclusion from the opposite direction, observing that every large operator in this industry is organized as a group: an institutional entity holding the liquidity and credit relationships, a distribution entity, and separate licensed entities isolating regulatory risk by market. That is containment, and it is the one lesson here that becomes more expensive the longer it is deferred, because capital is easy to separate before it is deployed and very difficult afterwards.

The Evidence

Berenberg has been owned by one family, in continuous descent, since two brothers left Antwerp as Protestant refugees in 1585 with nothing portable but their trade. The Gallery’s note on the plate is the lesson in a sentence: portraiture in a house like this was never decoration but a record of who stood behind the signature on the letter of credit — the closest thing the eighteenth century had to a certificate of good standing.

The habit is general. Hoare’s has kept its ledgers since the seventeenth century and holds the accounts of Pepys, Dryden and Austen in its own archive at Fleet Street. Berenberg numbers its financial years — the four hundred and twenty-eighth, the four hundred and thirtieth — which is a small typographical habit and a large institutional statement.

The Lesson

An institution whose people change every generation cannot keep its judgment in its people. It must keep it in documents: what was decided, by whom, on what evidence and against what objection. This is how a firm in 1930 could be governed by a lesson learned in 1830 without anyone alive remembering the loss that taught it.

The modern equivalent is not a museum. It is the practice Standard No. 01 took from the horological register: every object numbered, nothing destroyed, an extract retrievable a century later. It is a discipline of contemporaneous record — minutes that state reasons rather than outcomes, risk decisions logged at the time with the dissent attached, counterparty files recording why an arrangement was accepted and not merely that it was. A firm that cannot show why it decided cannot defend what it decided, to a regulator, to an acquirer, or to its own successors.

The Evidence

Gutzwiller has stayed deliberately small across five generations at Basel, holding to unlimited liability and to a client list it does not attempt to grow quickly. It is the least known of the houses in Room XI, and that is not an accident of fortune; it is the condition it selected. The oldest private houses in Europe are, without exception, the quietest institutions in finance: they do not name clients, do not publish case studies, and several have gone centuries without a public statement about any individual account.

The Lesson

This is usually read as secrecy. It is more precisely a rule about ownership: what the client tells the house belongs to the client, and the house has no licence to spend it. Discretion exercised for the client’s benefit is a service and coexists perfectly with complete transparency toward regulators, auditors and the client himself. Silence maintained for the house’s benefit — about its own economics, its own conflicts, its own counterparties — is opacity wearing discretion’s clothes, and the old houses are poor examples of it: they disclose their form, their liability, their partners and their accounts.

Gutzwiller adds the commercial half. Growth is a choice rather than a duty, and a house that decides its maximum size in advance is making a governance decision — usually the decision that keeps it alive, because a house that cannot decline work cannot keep a standard.

The Evidence

In 1750 two Zurich merchants set up a silk-trading business in a building on the Limmatquai known by its sign, the twisted sword. One of them had learned securities and bills of exchange on his travels and carried them on as a sideline; the sideline outgrew the trade. The house appears in the governmental calendar as a bank in 1805 and abandoned goods altogether in the 1850s. The Rahn name enters in 1882, the Bodmer name in 1917, the Bidermann name in 1951. The trade changed, the family names changed, the currency changed; the address did not. Hoare’s has traded at 37 Fleet Street since 1690 on the same principle.

The Lesson

Permanence is partly physical, and the reason is not sentimental. An address that cannot be moved in a weekend is a form of collateral: it says the house can be found, that its principals are resident where its clients are, and that it cannot be dissolved and reconstituted under another name when a difficulty arrives. Counterparties price this even when they do not say so.

The contemporary failure mode is the reverse and common enough to have its own economics. A firm selects a jurisdiction because incorporation is quick and the capital minimum low, discovers that no bank and no payment provider will serve a company of that description, and learns in month nine that it has bought a certificate rather than a seat. Study No. 05 states the correct sequence: establish which counterparties will serve the firm, then select the licence they will accept.

The Evidence

Julius Bär was founded at Zurich in 1890, more than a century after Rahn+Bodmer and three centuries after Wegelin, on the street where it still trades. Of the Swiss and German houses hung in Room XI it is the only one that has neither converted from that tradition nor been absorbed into a larger bank, and it is young enough that the choice is still in front of it: convert, merge, or stay as it is. Metzler has answered the same question one way since 1674; Hoare’s has answered it twelve times.

The Lesson

Ownership structure is a strategic instrument and is used as one by the houses that last. The listed form buys capital and liquidity and sells, in exchange, the horizon — because a listed institution is measured every quarter by people who may hold it for a week. The closed form sacrifices capital and liquidity and buys the ability to make a decision whose payoff is thirty years away.

Neither is superior in the abstract. What is not available is to take the capital and keep the horizon. The error the House sees most often is not a founder who sells too much equity; it is a founder who sells equity without noticing that he has simultaneously sold the tempo of every future decision, and who then attempts to run a generational institution to a quarterly clock.

The Evidence

On 23 July 1805 two Genevans, both under thirty, signed a partnership deed and opened a house in the old town. It was called De Candolle, Mallet & Cie; the name Pictet did not appear until 1841 and was not the firm’s style until 1926. What has carried it is not the family but the succession rule: a partnership share can be neither inherited nor bought, a partner is admitted only by unanimous vote of the others, and fathers and sons may not serve at the same time. Tenures are long and deliberately overlap, so the house is never held by a single generation at once. In two hundred and twenty years the firm has had fewer than fifty partners; Berenberg has admitted fewer than forty in four hundred and thirty.

The Lesson

Two disciplines are visible in both, and they are usually confused. The first is that partners are admitted rarely: scarcity of admission is what makes partnership a control rather than a title, and a house that promotes into ownership on tenure has converted its most important selection decision into a human-resources process.

The second is that ownership and management are separable and should be separated deliberately rather than by crisis. Metzler appointed personally liable partners from outside the family as early as the seventeenth century. The houses that survived their third generation opened management to talent while keeping ownership closed; those that failed at the third generation — the overwhelming majority, in every jurisdiction and every century — did the opposite, or did neither, and discovered the question at a funeral.

The Evidence

Mirabaud was founded at Geneva in 1819, in the settled years after Napoleon, and passed through seven generations of the same families before converting from a partnership to a corporate form in the present century, as Lombard Odier and Pictet also did. The Gallery treats that conversion as the significant modern event in all three houses, and notes that it is rarely discussed publicly, because it changed what a partner personally stands behind.

Set against seven generations the arithmetic of a modern listed retail broker: in a record year of operating income, marketing consumed something approaching half the entire cost base and net profit fell by a quarter. Nothing was mismanaged. The structure was operating exactly as designed, at the tempo it was designed for.

The Lesson

The unit of measurement is not a cultural preference. It determines the answer to every material question an institution faces. A house measuring in quarters will accept leverage it can service this year and a client it can monetize this year; a house measuring in generations declines both, and looks for several years like the weaker business.

Which is correct depends entirely on what is being built — and that is the decision, made once, that Study No. 05 argued is very nearly irreversible. The error is not choosing the short horizon. It is choosing it by default, through individually reasonable decisions about platforms, jurisdictions and counterparties, and discovering afterwards that the tempo of the institution was set by its vendors. When a house changes its legal form it changes the answer to the only question a client is really asking, which is who loses first — and that answer belongs in the disclosure rather than in the footnotes.

The Evidence

Geneva in 1796 was a poor place to open a bank: the old banking families had lent their fortunes to Louis XVI and lost them, the export trade had collapsed with the Revolution, and the city would be annexed by France within two years. Henri Hentsch opened a house in silks and commissions on 11 January; his second cousin Jean-Gédéon Lombard joined two years later and the firm was renamed. They fell out, the two lineages traded apart for two centuries, and merged again only in 2002. Across those years the name on the door changed repeatedly, as at every Geneva partnership, because the name was simply the list of the partners then liable. The institution was never the name. It was the deed underneath it.

The Lesson

A house that identifies itself with its current principals must be rebuilt each generation. A house that identifies itself with its constitution need not, and the choice is made at formation and almost never afterwards. This is the lesson that governs the other ten, and the one least available to a founder in the first years, when survival is the whole of the agenda.

Personal liability, refusal, discretion, the seat, the archive, the closed register of owners, the long horizon: not one of these pays for itself inside a founder’s tenure. Each is a transfer from the present generation to a later one, and each is therefore rational only for someone who has decided that the institution matters more than his own return from it.

The motto more than one European banking family borrowed from Virgil states it in four words — sic vos non vobis, thus you build, but not for yourselves.

Room XI read against the eleven lessons: what each house held, and how each ended. Three readings carry the argument of the grid.

• The surviving private houses are almost identical. They did not copy one another — they are separated by centuries, languages and legal systems, and in several cases by open commercial rivalry — and they converge anyway, on practices none of them described as a framework and none of them published. Convergence arrived at independently is the evidence for treating these as conditions of survival rather than preferences of taste, and it is why the House states them as a standard: what unrelated institutions discovered separately over four centuries can be adopted deliberately in a week.
• Lesson IV is the exception, and it is the modern one. Only Rothschild held it in full; the others were single houses in single cities and did not need it. Wegelin, destroyed by a foreign jurisdiction, is precisely the case that separation of capital and of perimeter exists to contain — which is why it is the lesson most often missing from a contemporary design.
• The two houses that failed score as well as the survivors. Barings holds nine of eleven for two centuries; Wegelin holds nine of eleven for nearly three. Neither was saved, because what would have saved them is not an inheritance and does not appear in this table at all. Section XII is about the two columns that are missing.

Room XI hangs fourteen houses; twelve are still trading. The two that are not failed in opposite ways, and between them exhaust how a house of this kind is lost.

Barings — The Desk Nobody Was Watching

Francis Baring was the third son of a German wool merchant near Exeter, deaf from youth, apprenticed in London at seventeen. By 1806 the firm bore his name alone. He financed the war against Napoleon, helped arrange the American purchase of Louisiana in 1803, and died in 1810 with the house at the height of its reputation.

A hundred and eighty-five years later it was insolvent. The losses were approximately £827 million, more than twice the firm’s available trading capital, accumulated by a twenty-eight-year-old trader in Singapore who both dealt and controlled the settlement of his own dealing, and who concealed the position in an error account. On 5 March 1995 it was sold for one pound. Two hundred and thirty-three years ended in a week, and the cause was the absence of one control: the independent risk and settlement function drawn as a dashed bar at the foot of Exhibit 5 in Study No. 05.

Wegelin — The Policy the Whole Partnership Stood Behind

Wegelin was founded at St Gallen in 1741 by a linen merchant and became the oldest bank in Switzerland. In January 2013 it pleaded guilty in a Manhattan court to helping American clients conceal at least $1.2 billion from the United States tax authorities, paid $57.8 million in restitution and fines, and announced in the same statement that it would cease to operate as a bank. Two hundred and seventy-two years ended in one plea, over conduct the firm said it had believed lawful under Swiss law at the time.

The Two Shapes of an Ending

Barings failed from one desk nobody was watching — a control that was absent. Wegelin failed from a policy the entire partnership stood behind, right up to the day a foreign court disagreed — a control that was present, functioning, and pointed at the wrong perimeter. The first is a governance failure and the second is a jurisdictional one, and no amount of the first kind of discipline prevents the second.

This section sits after the eleven lessons because they are otherwise capable of being read as a claim that heritage protects. Heritage is accumulated evidence that a structure has worked, and it becomes worthless the moment the structure is altered — which at Barings occurred not by decision but by growth. The question the House puts to every mandate it manages is therefore narrow: which activity here is conducted under a name earned by a different activity, and in whose jurisdiction is it actually conducted?

Each lesson, the instrument by which the House binds itself, and the requirement placed on the client institution. Read the centre column alone and it is a description of a firm; read the right column alone and it is a specification for an institution. The register exists because the two are the same document written twice, and because a client is entitled to ask the House to produce its own evidence for any line it is being asked to accept.

The register is deliberately symmetrical. Every obligation the House places on a client institution is one the House has first placed on itself, and can be asked to evidence. Where the two columns differ, they differ only in form: X-CHASE is not a regulated institution and does not hold client capital, so several of the client’s obligations have no exact counterpart in the House and are met by the nearest available discipline.

Note also which lessons are absent from the left column. Lesson IV, the separation of capital, binds the House only weakly, because a firm that holds no client money and carries no balance-sheet risk has little to separate. It binds the client institution completely. Asymmetries of that kind are stated rather than smoothed over, on the same principle that governs the rest of this Standard.

A standard that is not attached to a moment of decision is an essay. The eleven lessons of this document are attached to one moment, and it is early: the point at which a principal has decided to establish an institution and has not yet signed anything.

The reason is the finding of Exhibit 3. Every practice in this Standard is a decision about the inside of the firm — liability, ownership, separation of capital, succession, the seat, the record — and each of them is settled, in law and in practice, by documents executed in the first weeks: the constitution, the shareholders’ agreement, the group structure, the licence application, the first vendor contracts. After those, a lesson is not adopted. It is retrofitted, at a multiple of the cost, usually in response to an event.

The Sequence the House Holds To

Design — the eleven lessons are put to the principal as questions, and the answers are written into an architecture the principal owns: the group structure, the liability and ownership arrangement, the licensing perimeter, the banking sequence, the organizational design, and the capital deployment schedule that follow from them.

Launch — procurement begins against that architecture rather than before it. Counterparties are selected on the nine questions of Exhibit 2, and every arrangement is examined for the channel by which its failure would reach the rest of the group.

Management — the eleven are re-asked on a stated cycle, because the answers drift. The founder who was personally exposed at incorporation is frequently not exposed by year three, and the moment at which that changed is almost never recorded.

The House therefore holds that the structural work belongs at the beginning and should be conducted as a discrete, paid, fixed-scope engagement whose deliverable is an architecture rather than an introduction — the drawing board of Standard No. 01, arrived at here from four centuries of banking rather than from the bench. The Blueprint is not offered without charge, and the reason is contained in this Standard rather than in a commercial policy: a design produced at no cost is a design the principal has not been required to own, and an architecture nobody owns is the architecture the vendors will write.

None of this makes the engagement a governance product. The eleven lessons are not sold, and they are not the deliverable; they are the standard against which the deliverable is measured, in the same way that the nine questions of Exhibit 2 are not a service but the terms on which a counterparty is assessed. What a principal receives is an architecture. What this Standard describes is what the House will refuse to put its name to while building one.

The old houses are not models to be copied. Their markets are gone, their instruments are obsolete, their legal forms are in several cases unavailable, and their advantages were in part the advantages of a world with fewer competitors and slower news. Nothing in this Standard proposes that an institution founded today should be organized as a Hamburg cloth merchant was organized in 1590.

What they offer is narrower and more durable. They are the surviving record of which institutional decisions absorb error and which transmit it. Every one of the eleven lessons is a decision about the inside of the firm, taken before there was any evidence that it would matter, and maintained at a cost through periods when it plainly did not. That is the whole of the method: the protections were installed while they were cheap.

Nor were their circumstances easier. Each of these houses operated without limited liability, without deposit insurance, without a lender of last resort for most of its life, and frequently without a central bank or an enforceable law of contract across the borders it traded over. They carried more risk than a modern institution carries and they survived it, which is worth stating because the contemporary argument for abandoning these disciplines is almost always that conditions today are unusually demanding. They are not. They are unusually protected.

X-CHASE takes from them a firm’s worth of obligations and one governing proposition: that the structure of an institution is chosen once, by someone, at the beginning — and that if it is not chosen deliberately by the owner of the capital it will be chosen incidentally by his vendors, his platform, his jurisdiction, and his first liquidity agreement. The old houses chose. That is the entire difference between them and the thousands of firms founded alongside them whose names are not recorded anywhere.

Standard No. 01 took its method from houses that make objects; this one takes its conditions from houses that were institutions. The two arrive at the same place from opposite directions, which the House regards as the strongest evidence available that the place is correct.

Nothing in that proposition requires a long history to act upon, which is the only reason a firm of a few years is entitled to publish this Standard at all. The eleven lessons are not the reward of age. They are the conditions of it. Every one of them was available to the houses in this gallery on the day they opened, at a cost they could afford, and the reason so few of their contemporaries survive is not that the conditions were secret. It is that they were expensive at the moment they were cheapest.