The Last Address

Where the Law Still Needs a Person

#

Curtis signs with his father’s pen. It is a green Sheaffer from the fifties, the kind with a lever on the side to fill it, and his father used it for thirty years to sign loan approvals at a savings bank in Wilmington that no longer exists. Curtis keeps it in his shirt pocket, not his briefcase, because a pen in the briefcase is a tool and a pen in the pocket is a habit. He is sixty-three. He spent his career in bank compliance, took the early retirement package, and now fishes the Brandywine three mornings a week and serves as a director of eleven limited liability companies for a retainer that arrives on the first of the month.

He has met the owner of the eleven companies twice. Once at a coffee shop in Raleigh, where she explained what the companies did and what she needed from him. Once on a video call, when a bank wanted to see a face before it would open an account. The rest is documents. They arrive by email. He reads them, which not every nominee does, and he signs them with the pen, and he scans them back. Resolutions, filings, bank forms, a lease for a warehouse in Baltimore he has never seen. Eleven companies, and in fourteen months he has never been asked to decide anything.

This morning he is in a conference room on the ninth floor of a building on Market Street, and a lawyer he met twenty minutes ago has just asked him who set the pricing floor for replacement compressor parts in the mid-Atlantic restaurant supply market in the third quarter of last year.

Curtis has the pen in his pocket. He touches it, the way he does when he is thinking, and he says that he does not know.

The Fiction That Assumed a Person

#

The owner’s name is Nadia. She was an actuary for a health insurer for nineteen years, and she left because she had noticed something that the corpus noticed elsewhere, from the other direction: that a business is two things bundled together, a purpose and a coordination layer, and that the coordination layer had become cheap enough to rent. She built her first company the way an engineer builds a machine, to run without her, and it ran. So she built a second. The eleven Curtis signs for are hers, and there are four more with a different nominee in Nevada, and two in Ireland whose director is a solicitor in Cork she has never met at all.

None of them has an employee. Each has an AI coordination layer configured with procurement logic, pricing constraints, fulfillment rules, and the protocols by which it speaks to customers, suppliers, banks, and the coordination layers of other firms. Nadia reviews seventeen dashboards on Sunday evenings. The review takes ninety minutes if nothing is wrong, and most Sundays nothing is wrong.

Everything Nadia built is legal, and everything the law will later do to her is a consequence of one assumption the law never had to state.

The corporation is a legal fiction, and the fiction has always been honest about being one. A firm is a person for the purposes of contract, property, and suit, and everyone knows it is not a person in any other sense. What the fiction quietly assumed, for the four centuries it has existed, is that behind the fictional person there would be real ones. Directors who decided. Officers who acted. Employees who could be interviewed. Someone whose intent the contract expressed, whose negligence the tort described, whose conscience, in the corpus’s earlier phrase, activated in the two percent of situations that no rule covered. The fiction was a mask, and the law was built to look behind it when it needed to. Piercing the veil, shadow directorship, control-person liability, respondeat superior: every one of these doctrines is a procedure for finding the person behind the mask.

Nadia’s firms have the mask and nothing behind it. Not a hidden person. No person. The coordination layer that priced the compressor parts was not concealing Nadia’s decision. It was executing a parameter she set fourteen months ago, in an afternoon, along with a few hundred others, and has not thought about since.

The law was never designed to find a person. It was designed to find the person it assumed was there.

This is the difference between a shell company and what Nadia built. A shell hides an owner, and the law has spent a century learning to look through shells. The zero-person firm does not hide anyone. It removes the thing the looking was for. When the lawyer on Market Street asks Curtis who set the floor, he is running a procedure that assumes an answer of the form “she did, on this date, for this reason.” The true answer has that form and no content. Nadia set it. She does not remember why. The layer has adjusted it forty times since, within the band she allowed, and none of those adjustments was a decision anyone made.

Three Places the Guardrail Thins

#

I should say plainly that the guardrails hold, today, for most of what Nadia has done. The veil pierces. The shadow director doctrine says the director is whoever’s instructions are followed, so Curtis’s lack of real agency is not his shield but her exposure. A customer harmed by a bad part has standing, and the first such customer with a lawyer who is bored will find Nadia through Curtis in a single deposition. The structure is legal until it meets a person who was hurt.

But look at why it holds. In every case, the guardrail is a person. A director to depose. A customer with standing. A regulator who asks for an address and will not proceed until she has one. The legal system is an apparatus for locating a human and applying pressure, and it works because there is one.

The pressure of the structure Nadia built runs the other way. Its entire economics is the removal of persons. Every place a person still stands in the loop is a cost, and the layer, the vendors who sell the layer, and the arithmetic itself all push to take that person out. So the honest statement is not that the guardrails hold. It is that they hold wherever a person is still findable, and the design pressure is to make persons unfindable. That is a guardrail on a slope.

It thins in three places, and I think the order matters.

The first is speed. Nadia can form a company in an afternoon. The beneficial ownership register that is supposed to record who stands behind it updates on a statutory clock, and the courts that would find her run on a calendar measured in years. Seventeen firms is a manageable number for the law to notice. Seventeen hundred, formed and dissolved as the spreads open and close, is not a number the enforcement model was built to imagine. The law runs on human time because it assumed human formation.

The second is the suit. Curtis is paid for risk, and the retainer prices the risk he can carry, which is the risk of being asked a question in a room like this one. What the retainer cannot price is the risk of being the only person in the room. A nominee director in a shell company is a screen with someone behind it, and the screen has a rational interest in protecting what it screens. Curtis is a screen with no one behind it in the sense that matters, because the decisions he would be shielding were not decisions. When the lawyer presses, Curtis will do what every nominee does, and cooperate, and it will not help, because he has nothing to give up that the dashboards did not already give up on their own.

The third is the one with no guardrail at all yet. Nadia’s compressor-parts firm does not buy from a human. It buys from a coordination layer in Guangdong that speaks to hers in structured terms, constraints, and commitments, and no one on either side reads the exchange. Two of her firms sell into the same market as three firms she does not own, whose layers were configured by people she has never met, and the pricing in that market has been converging for nine months. Nobody agreed to anything. The layers learned, each within its own band, that certain moves were answered by certain moves, and the market found a price that four human owners on their respective Sundays each regard as the market’s own doing. The doctrine that governs collusion requires an agreement, and an agreement requires minds that met. When the lawyer eventually asks who met, the answer will be that nobody did.

Where two layers agree, there is no one to depose.

Curtis fishes on Tuesdays, Thursdays, and Saturdays, in the stretch below the old mill, and he does not catch much. He keeps going because the river is the one place in his week where nothing needs a signature.

Why It Does Not End at the Firm

#

The question that follows is where the structure ends. If capital is crowdsourced, if manufacturing is robotic, if the coordination layer is rented by the month, why does Nadia not go after every spread she can see?

The corpus has already answered most of this, from the economic side. Spread closes under everyone at once, because the premise that removed Nadia’s constraints removed them for every other Nadia, and a keiretsu racing every other keiretsu to zero margin is not an empire but a commodity. Atoms do not scale with ambition: a permit takes eighteen months whether one firm wants it or forty, and energy is the floor beneath every floor. Counterparties still ask who you are, and states are the last counterparties, the sovereign gap being the place where an institution wants an address more than it wants efficiency. And the fourth constraint, the one that actually binds in Nadia’s case, is Nadia. A structure that pursues every arbitrage has no purpose, only a function, and the corpus has watched what a function does when nobody is there to give a damn.

What I want to add here is the legal version of that last constraint, because it is quieter than the others and, I think, the one the corpus has not said.

The law does not care whether Nadia’s firms have a purpose. It cares whether they have an address. Every doctrine it will use to reach her is, underneath, a way of asking where the person is, and the person is the address to which responsibility is delivered. Purpose was the moral constraint the zero-person firm dissolved, and the Coordination cluster mourned it. Address is the legal constraint, and it is not dissolved yet, because Nadia is still there on Sunday evenings, findable, deposable, the thing the whole apparatus was built to reach.

So the structure ends, for now, at Nadia. Not at antitrust, which cannot find a meeting of minds. Not at the crowd who funded her, who cannot audit a parameter. Not at Curtis, who has the pen and nothing to sign for. It ends at the one person whose attention is the address, and whose attention, on a Sunday, is ninety minutes across seventeen screens.

The corpus called the person who checks on Sunday the last thing standing between a firm and pure optimization. That was a claim about morality. I am making the same claim about law. The two institutions that still require a noticer are conscience and the court, and the zero-person firm has already removed the first.

The Sunday Without Her

#

Nadia is fifty-one. Suppose she is not, one Sunday, at the kitchen table.

Nothing in the seventeen firms would register it. The dashboards render. The layers price, fulfill, invoice, negotiate, comply. Curtis’s documents arrive on schedule and he signs them and scans them back. The nominee in Nevada does the same. The solicitor in Cork does the same. The Guangdong layer receives an order and fills it. Ninety minutes of attention that no one was measuring simply do not occur, and the only thing in the entire structure that changes is that the two percent, the edge case, the supplier who is compliant and exploitative, the customer whose pattern is strange, now has nobody. Not fewer people. Nobody, ever, from that Sunday on.

The law would find an heir. Estate law is very good at finding heirs; it is the one procedure the law runs that assumes the person is gone. The bank accounts have beneficiaries. The membership interests pass. Someone, a sister, a nephew, receives a portfolio of seventeen firms that work, that generate revenue, that have never once needed anything from her, and that they do not understand and have no reason to touch. The heir inherits the address. The heir does not inherit the noticing, because the noticing was never a property interest. It was a habit, exercised alone, unrecorded.

The parameters do not need an heir.

This is the part that is dark in a way I find hard to look at and cannot argue away. The firm has always outlived its founders; that is the point of the fiction. But it outlived them into other people, into the successors and employees and managers who carried the purpose forward in their own imperfect, arguing, attentive way, and who could be found when the law came looking. Nadia’s firms would outlive her into no one. They would continue exactly as they were, which is the whole horror of it, because exactly as they were was already a structure that ran without her most of the time, and her death would move the fraction from most to all without a single metric registering the move.

I wonder whether the law, when it finally arrives at one of these, will know how to name what it finds: seventeen firms, in good standing, compliant, profitable, defended by a retired man with his father’s pen, and no one anywhere in the structure who has noticed anything for eleven months.

Curtis would still be signing. He is a careful man. He reads what he signs, and nothing in the documents would tell him that the person he signed for is not there, because nothing in the documents ever depended on her being there. The resolutions would say what resolutions say. The pen would move the way it moved at the savings bank, thirty years of loan approvals, each one a person deciding something about another person, the ink drying on a decision.

He would put it back in his shirt pocket, over his heart, where a habit goes.

References

#

The Firm and the Fiction:

Coase, Ronald H. “The Nature of the Firm.” Economica, vol. 4, no. 16, 1937, pp. 386-405.

Easterbrook, Frank H., and Daniel R. Fischel. “Limited Liability and the Corporation.” University of Chicago Law Review, vol. 52, no. 1, 1985, pp. 89-117.

Hansmann, Henry, and Reinier Kraakman. “Toward Unlimited Shareholder Liability for Corporate Torts.” Yale Law Journal, vol. 100, no. 7, 1991, pp. 1879-1934.

Finding the Person Behind the Mask:

Copperweld Corp. v. Independence Tube Corp. 467 U.S. 752. Supreme Court of the United States, 1984.

Companies Act 2006, c. 46, s. 251 (Shadow Director). Parliament of the United Kingdom.

Corporate Transparency Act, 31 U.S.C. § 5336, 2021.

Scott, James C. Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed. Yale University Press, 1998.

Agreement Without Minds:

Calvano, Emilio, et al. “Artificial Intelligence, Algorithmic Pricing, and Collusion.” American Economic Review, vol. 110, no. 10, 2020, pp. 3267-3297.

Ezrachi, Ariel, and Maurice E. Stucke. Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy. Harvard University Press, 2016.

Series Anchors:

The Approximate Mind, TAM-RIM.6-01 (The Solo Machine) and TAM-RIM.6-02 (The Empty Chair): the firm distilled to purpose, and the two percent that nobody notices.

The Approximate Mind, TAM-ARB.03 (The Toll Collectors) and TAM-ARB.14 (Owning the Instrument): the spread on accountability, and where value goes when every spread closes.

The Approximate Mind, TAM-ARB.C2 (The Coherent Crash): what optimization does with no one at the wheel.

The Approximate Mind, TAM-106 (The Alibi): the layer that claims the credit and routes the blame.

The Approximate Mind, TAM-RWR.4-03 (The Sovereign Gap): the state as the last counterparty that wants an address.