What Cost Does

Every system of governance rests on an implicit claim about cost, a claim so deeply embedded in the structure that it is rarely stated and almost never examined. The medieval king could govern because the cost of raising an army was high enough that only those with vast landholdings could do it, and the landholders' interests could be bargained into a stable arrangement: military protection in exchange for taxation, loyalty in exchange for justice, obedience in exchange for the promise that the king's power would be exercised within limits that custom defined and force occasionally enforced. The democratic state could govern because administering territory required revenue, revenue required taxation, and taxation could be conditioned on representation. The state's need for money gave citizens bargaining power. The liberal market could coordinate because the cost of information was high enough that firms and intermediaries were needed to discover prices, match buyers with sellers, enforce contracts, and bear the risk of transactions between strangers. Those intermediaries could be regulated, taxed, and held to account because they occupied positions that were visible, territorial, and expensive to maintain.

In each case, the costliness of the governing function was the mechanism by which governance was connected to legitimacy. Costly governance created chokepoints, and chokepoints created accountability surfaces: places where the flow of power became narrow enough that the governed could stand in the way and demand terms. A king who needed taxes could be compelled to grant a charter. A state that needed revenue could be compelled to permit a parliament. A firm that needed customers could be compelled to honor a contract. An intermediary who needed clients could be compelled to submit to regulation. The connection between cost and legitimacy was never automatic, never clean, never without violence and corruption and centuries of struggle in which the powerful resisted every constraint and the powerless died for the constraints they won. But the connection existed because costliness created dependency, dependency created the possibility of constraint, and constraint was the raw material from which constitutional order was slowly, painfully, built.

Cheaper coordination can remove a particular occasion for bargaining. Two agents may arrange a transaction without the people formerly employed to negotiate it. The transaction still uses equipment, contractual authority, and institutions that recognize its consequences. A supply chain can be rerouted before its participants have time to respond. What has disappeared is an opportunity to intervene in the operation, not every dependency through which its principal can be required to answer.

The danger is a separation between the ease of acting and the difficulty of contesting the act. A principal can obtain coordination at machine speed while those affected wait for a forum able to reach it. The institutional task is to bind the continuing authority and resources behind the operation before that delay becomes a license to impose consequences.

The question that follows is not economic but political, and it is the oldest question in constitutional theory, restated in a medium that its original formulators could not have imagined. What constraints can be imposed on coordination that does not need permission? What accountability is possible for processes that terminate before any human can inspect them? What does freedom mean when the interference is real, the effects are severe, and the immediate act came from an execution that has already ended, while the institution responsible for deploying it disputes what it must do next?

These are old questions in new form. Every constitutional tradition has grappled with the problem of invisible power: power that shapes lives without presenting a face that can be petitioned, a body that can be imprisoned, a decision that can be reversed. The feudal serf's labor was extracted by a lord who was at least visible. The factory worker's conditions were set by an owner who at least had a name and an address. What is new is that the invisibility is architectural rather than conspiratorial. No one is hiding. The process simply operates at a tempo and a scale that human perception cannot follow and human institutions were not designed to govern.

The bill of exchange made truth portable. Different filters selected which computational work to continue. Cheaper operations changed some of the places where coordination occurred. What minimum constitutional architecture makes coordination at machine speed compatible with human freedom?