The Coase Theorem, formalized by George Stigler from Coase's 1960 'Problem of Social Cost,' states that under conditions of zero transaction costs and well-defined property rights, private parties will bargain to efficient resource allocation regardless of how rights are initially assigned. A polluting factory and downstream fishermen will negotiate the optimal pollution level whether the factory has the right to pollute or the fishermen have the right to clean water—when bargaining is costless, the efficient outcome emerges. The theorem's radical implication is not that government intervention is unnecessary (Coase emphasized transaction costs are rarely zero) but that transaction costs, not property rights assignment, determine whether markets can solve coordination problems. When transaction costs are high, market bargaining fails and governance structures (regulation, hierarchical organization, institutional mechanisms) must substitute. AI reduces some transaction costs (execution, coordination) while raising others (evaluation, judgment quality)—a pattern the theorem helps diagnose.
The theorem emerged from Coase's challenge to the Pigouvian tradition in welfare economics, which held that externalities (costs or benefits falling on parties outside a transaction) required government intervention—taxes on pollution, subsidies for beneficial activities, regulation of nuisances. Coase demonstrated that this conclusion rested on an implicit assumption: that transaction costs made private bargaining impossible. Remove that assumption—imagine a world where the factory and fishermen can costlessly negotiate—and the externality resolves itself through exchange. The fishermen, if they value clean water highly enough, can pay the factory to reduce pollution. The factory, if it values production highly enough, can pay the fishermen to tolerate pollution. The efficient outcome (the pollution level where marginal cost equals marginal benefit) will emerge regardless of who initially held the right.
The theorem's significance lies not in its description of a zero-transaction-cost world—which Coase acknowledged does not exist—but in its revelation of what prevents efficient outcomes in the actual world: transaction costs are the binding constraint on market coordination. When they are low (well-defined rights, few parties, symmetric information, low bargaining costs), markets work. When they are high (ambiguous rights, many affected parties, informational asymmetries, costly negotiation), markets fail and institutions must substitute. The AI moment has created exactly this pattern in the knowledge economy: transaction costs for execution have approached zero (anyone can prompt an AI to produce code, analysis, content), while transaction costs for judgment—evaluating whether the output serves purpose, verifying quality beneath smooth surfaces, coordinating collective evaluation under velocity pressure—have intensified. The Coase Theorem predicts the governance response: execution migrates to market-like relationships (direct AI tool use), judgment remains governed hierarchically (internal deliberation within firms), and hybrid structures (the vector pod) emerge to manage the interface.
The theorem also illuminates the platform dependency problem. Knowledge workers and AI platforms have theoretically well-defined rights (the terms of service specify what each party can and cannot do), but the transaction costs of exit are asymmetric. A worker switching platforms bears high costs (lost workflow knowledge, relearning, adaptation period). The platform losing one worker bears negligible costs (revenue diversified across millions of users). This asymmetry means that even with clear property rights, bargaining will not produce efficient governance of the relationship—the platform can opportunistically exploit its position because the worker's exit threat is not credible. The Coasean analysis implies that regulation reducing switching costs (portability standards, interoperability requirements) or governance structures protecting workers from opportunistic platform behavior (collective bargaining, user councils, regulatory oversight) are necessary when transaction cost asymmetries prevent market mechanisms from functioning.
The theorem appears (though Coase never used the term) in 'The Problem of Social Cost,' published in the Journal of Law and Economics in 1960. Coase developed it partly in response to a seminar at the University of Chicago where he presented his ideas to a hostile faculty audience including Milton Friedman and George Stigler. By the end of the evening, he had persuaded them—Stigler later formalized the argument as the Coase Theorem. The theorem's simplicity made it extraordinarily influential: it reshaped environmental economics, tort law, property rights theory, and regulatory design. It also generated substantial confusion: readers who took the zero-transaction-cost assumption literally missed Coase's actual point, which was that transaction costs are almost never zero and that their magnitude determines when markets work and when institutions must substitute. Williamson's career can be read as the systematic elaboration of this second, more practically consequential implication.
Zero transaction costs produce efficiency. When bargaining is costless and rights are clear, private parties will negotiate to efficient resource allocation regardless of initial rights assignment.
The assumption reveals the constraint. The theorem's value is not describing a zero-transaction-cost world but revealing that transaction costs are what prevent efficient market outcomes in the actual world.
High costs require governance. When transaction costs make bargaining prohibitive, institutions (regulation, hierarchical organization, collective mechanisms) must substitute for market exchange.
AI has created asymmetric transaction costs. Execution costs near zero, judgment costs high and rising—producing the organizational reorganization Williamson's framework predicts.
Efficiency requires appropriate governance. The normative implication: institutions should be designed to economize on the transaction costs that are actually high, not the costs that theory assumes are low.