The boundary of the firm
In 1937 Ronald Coase asked why firms exist. If markets allocate resources so efficiently, why does every task not pass to the best available counterparty under a fresh contract? Why gather work inside an organization at all?
The answer was cost. A transaction requires someone to find a counterparty, establish what is wanted, agree the terms, supervise the result, and settle whatever the agreement failed to anticipate. When those costs exceed the cost of directing the same work inside an organization, the work moves inside. When internal direction becomes more expensive than an external transaction, it moves out. The boundary of the firm is therefore not fixed by nature. It is drawn by the relative cost of coordinating within and transacting without.
That argument has survived because it describes more than contracting. It describes the hidden price of shared understanding. Work becomes difficult at the point where what one person knows must be transferred into a form another person can use. A customer explains a need to one office. The office condenses it for another. The second office turns the condensation into an instruction, the instruction joins a queue, and a specialist eventually acts on a description several steps removed from the fact that began the process. Each interface performs two operations at once: lossy compression and delay. Some context is discarded so that the rest can travel, and what travels waits to be received.
An organization chart can be read as a map of this tax. The boxes hold specialized context. The lines mark the places where context must be compressed, transmitted, queued, and reconstructed. Large firms do not pay the tax because they are badly managed. They pay it because knowledge has historically been expensive to move without distortion, and the firm is the best institution yet devised for bearing that expense.
The price is now falling.
The collapsing interface
Systems can increasingly hold a body of operating context intact and make it available at the point of action. The effect is not merely that an old sequence runs faster. Parts of the sequence cease to be necessary. When the original account of a problem and the context required to judge it can meet in the same place, the chain of summaries between them shortens. The queue thins with the chain. Judgment remains, but much of the machinery that once carried judgment from one specialized head to another does not.
Coase’s boundary moves when this happens. A function that required a department because its context was costly to assemble may require only a competent steward once that context can be preserved. Work once held together by layers of internal coordination can be performed by smaller organizations and exchanged across their boundaries. The firm contracts toward the work for which direction, responsibility, and judgment still belong under one roof.
This does not imply the disappearance of the firm. It implies a change in its efficient size. The modern corporation was shaped partly by the cost of moving knowledge through human interfaces. Reduce that cost and the shape changes. The movement is already visible wherever a small body can act with a completeness of context that previously belonged only to a large one.
There is a simple way to identify the businesses most exposed to this movement. Their product is the tax itself.
The business of the tax
The consulting archetype is the clearest case. An organization cannot move context cleanly among its own parts, so it hires another organization to collect the fragments, interview the holders, reconcile their accounts, and return a compressed version in a form senior judgment can absorb. At its best, this supplies rare judgment and independent authority. Much of the time, however, the economic unit is the transfer itself: meetings, workshops, committees, interviews, and presentations whose billable substance is the movement of context between heads.
The model is not defective. It answers a condition in which internal coordination is expensive. But a business that earns its return from the cost of transferring context cannot be indifferent to a collapse in that cost. Genuine judgment remains scarce. Responsibility remains valuable. The broad pyramid of labor devoted to gathering, translating, and carrying context does not enjoy the same protection. When the interface becomes cheap, a business whose product is the interface stands on the wrong side of the change.
Another institutional form stands on the right side. It does not repeatedly carry context from one private copy to another. It crystallizes the context once.
The common record
Markets have always contained facts that every participant needs and no participant can keep alone. The state of a fleet, the quality of a cargo, the ownership of a right, and the standing of a counterparty all have the same awkward properties. They are shared, consequential, and liable to decay. The wasteful arrangement is for every participant to maintain a private version, each assembled at full cost, each incomplete in a different way, each becoming stale in silence.
The register is the market’s answer. Context is gathered once, resolved once, kept true by a steward, and made available to everyone who relies on it. Its economy is not simply one of scale. The register prevents the market from paying repeatedly for inferior copies of the same truth. Its essential service is the maintained common record, and its authority comes from accepting a burden no private copy can accept: to adjudicate among conflicting accounts and continue to stand behind the result as the world changes.
This is crystallized context. It is not a summary passed down a chain and reconstructed at every stop. It is a durable reference to which the chain can return. The cost of maintaining the truth is concentrated in the steward; the benefit radiates across the market. Where the context is shared and its decay is costly, the register is not merely another provider of information. It is the institution that removes thousands of parallel coordination burdens by holding one of them properly.
The fall in coordination cost therefore does not weaken the register. It strengthens the case for it. The same force that permits firms to become smaller produces more firms, and more firms produce more boundaries across which work must pass. What was once an instruction between departments becomes a transaction between organizations. Each new transaction raises a prior question before price, performance, or settlement can be considered: who is on the other side, really, and who stands behind them?
The counterparty question multiplies with the boundaries. Demand for the common record grows in proportion to the force shrinking the organizations that consult it.
A modern recurrence
The pattern appeared with unusual clarity in the credit markets of the early 2000s. Every credit desk kept private marks for instruments that did not trade on an exchange and had no dependable consensus close. The marks were necessary, costly to produce, and mutually inconsistent. Each desk held a decaying private account of the same market, and in unsettled conditions the differences between those accounts became more than an inconvenience. A market cannot measure its risk cleanly when every participant carries a different private truth about the price of that risk.
A consortium formed around contributed credit marks. Participating dealers sent their marks to a single steward, which blended them and returned the composite to the participants. The result was better than any one desk’s private estimate because it contained information no one desk possessed. The firm that became part of the reference-data establishment began not with proprietary facts but with a disciplined exchange: many partial accounts went in, one more authoritative account came back.
The decisive clause was that contribution was the prerequisite for receipt. A participant received the composite only by supplying its own marks. This solved the cold beginning that defeats many common records. The first contributors did not donate information to a service that others could consume without them. Contribution and consumption were made the same act. The arrangement excluded the pure free rider and gave each participant a reason to strengthen the record it wanted to consult.
From there the mechanism became a ratchet. Each additional contributor widened the information in the composite. A better composite increased the disadvantage of relying on a private mark. That disadvantage recruited the next contributor, whose information improved the composite again. Once the common record crossed a threshold of participation, staying outside no longer preserved independence. It preserved an inferior view of the market.
This is a general law of registry formation, not a peculiarity of credit. A shared record has little value before the shared facts arrive, while the first holder of those facts has little reason to contribute to an empty record. Receipt conditioned on contribution joins the two sides of the problem. The record ignites because access to the common context requires adding to it.
The government of the governed
The same arrangement carries a warning. A consortium of the governed is a magnificent ignition mechanism and a dangerous permanent government.
Founding contributors know the problem, hold the needed information, and can deliver immediate adoption. They can also come to treat the common record as an instrument of their collective interest. The risk is structural rather than moral. The participants who supply the record are not identical to the market that must rely on it, and the interests of an early group do not remain identical to the interests of every later user. A reference governed indefinitely by the parties it describes or advantages may preserve participation while losing neutrality.
The steward must therefore become more than the agent of its founders. It must serve the market, including participants who arrived later, parties outside the founding circle, and users whose interests differ from those of the contributors. Governance that helps a record begin can prevent it from becoming canonical if it hardens into privilege. A common record earns authority only when the market believes that no member can bend the answer.
The two lessons belong together. Contribution as a condition of receipt can solve the registry’s cold start. Neutral stewardship is what allows the resulting institution to outgrow the consortium that started it. One explains ignition; the other explains legitimacy.
The credit market supplied a modern demonstration of both the power and the danger. The larger question now belongs to every market still paying thousands of times for partial, decaying accounts of the same consequential truth. Institutional identity is such a market. Its participants carry private maps of firms, funds, affiliates, and people; every map differs, and every map begins to decay as soon as it is drawn. The need for a common record rises as organizational boundaries multiply, yet the facts from which that record must be made remain scattered among the parties who need it.
The historical law is plain. The application is not. What would this look like for the register of who is who?