ronald coase won the nobel for writing two papers. the first explained why firms exist. the second — the one everyone's heard of and nobody gets right — was about a cow eating a farmer's corn. the paper is usually taught as "regulation is unnecessary, markets sort externalities out." this is almost exactly backwards. what coase actually said is something stranger and more useful: in a world of zero transaction costs, who you assign a right to doesn't matter. in the real world, it's almost all that matters.
here's the setup. a rancher's cows wander onto a farmer's land and eat his corn. the usual response is: the rancher is creating an externality, he should pay damages, maybe the government should fine him. coase pointed out that this framing embeds a choice nobody argued for. why is the rancher wronging the farmer and not the other way around? why is the farmer's right to corn-free pasture more fundamental than the rancher's right to let cows roam? the economics doesn't care. both uses of the land have value; the question is which use is higher. if you give the rancher the right to roam, the farmer pays him to fence. if you give the farmer the right to corn, the rancher pays him damages. either way, the land ends up doing its highest-value thing. this is the coase theorem in its useful form.
the part people skip is "in a world of zero transaction costs." our world is not that world. in the real one, bargaining costs time, lawyers, lobbyists, and trust. when coase said the initial assignment doesn't matter, he meant it as a limiting case to illustrate what does matter: the assignment matters exactly because transaction costs are nonzero. who has the baseline right determines who has to organize the bargain, who has to sue, who has to hire the lawyer. in practice, the initial assignment is the whole game. coase was not arguing that regulation is unnecessary. he was arguing that regulation should pay more attention to the cost of enforcing whatever it decides.
the deeper reading of coase, which is the one i find most useful, is that almost every "externality" people complain about is a property right nobody has written down yet. pollution is an externality because no one owns the atmosphere. noise is an externality because no one owns ambient sound. congestion is an externality because no one owns the marginal seat of road space. data privacy is an externality because no one owns the data a surveillance camera captures as you walk past. in every one of these cases, the problem is not that markets have failed — it's that there is no property to have a market in.
you can test this by running the thought experiment. if someone owned the atmosphere, polluters would pay them, and pollution would be priced at its social cost. that's carbon markets, broadly. if someone owned road space, drivers would pay them, and peak-hour traffic would thin. that's congestion pricing. if someone owned ambient noise, noisy bars would pay residents, and the resulting volume would be the one residents were willing to accept. that's what noise bylaws are, clumsily. each of these is a "market failure" that evaporates the moment you remember to define what's being bought and sold.
the flip side is that bad property-right assignments cause a lot of what we call market failure. patents that are too strong block research that would produce more value than the patent protects. patents that are too weak underfund research. zoning that gives incumbent homeowners veto rights over new construction produces housing shortages. the "market" didn't fail; the right was misallocated. this is a mechanism-design problem, not a market-vs-regulation one, and it's the framing i wish more of the policy world used.
the short heuristic i use, when someone says "we have an externality here." first question: what is the thing nobody owns? second: is there a transaction cost that would eat the gains from bargaining even if someone did own it? third: given those answers, what's the cleanest assignment of rights that produces the outcome we want? if you can't answer those, you're about to legislate a symptom.
coase is harder than people give him credit for. "markets solve externalities" is the bad pop version. "rights matter more when transactions are expensive" is the real thing. if you take only one framing from economics into your political reasoning, take that one.