A new bakery opens and the old bakery loses customers. A reader publishes a book and a stranger dislikes the world it describes. A factory adds smoke to the air. All three acts reach beyond the person who performs them, but that resemblance leaves almost everything important undecided. The lost customer, the offended preference and the damaged lung enter different relationships. If every unwelcome consequence were enough to establish an unpaid bill, the concept of an externality would lose its power to distinguish them. Its useful edge lies in the work that follows the discovery of an effect: identifying what has changed, whose decision omitted it and what an alternative arrangement could improve.
The competitor's loss begins with a price or a choice inside the market. A customer can gain from the same transaction that deprives an established seller of a return. Treating only the seller's loss as an outside cost would give the incumbent a claim against the ordinary operation of exchange. It would also leave the buyer's gain out of precisely the account that was supposed to become more complete. This does not make every competitive change harmless. Production, quality, working conditions and the use of resources can all change. It means that the movement of a receipt by itself does not tell us how the total opportunity has changed.
A fixed-quantity comparison makes the point without asking it to prove too much. Suppose ten identical units are produced at a cost of six each and each is worth fourteen to its buyer. At a price of ten, buyers gain forty and the seller gains forty. At a price of eight, buyers gain sixty and the seller gains twenty. The eighty of gains from trade is unchanged. Nothing has changed here except the price dividing it: quantity, quality, costs and the people counted all remain fixed. If production also imposes an unpriced harm of thirty, that harm reduces the total to fifty at either price. The redistribution through the price and the damage outside the transaction require separate entries.
Prices can nevertheless transmit an effect that matters for more than the division of a fixed gain. A borrower facing a binding financing constraint may have to sell assets after a fall in income. If many borrowers sell into the same market, the price can fall, requiring further sales to raise the money they need. An individual seller may take the price as given while contributing to the movement that tightens everyone else's position. In Korinek's model, this mechanism can move productive assets away from the people who use them more productively. The relevant imperfection is the financing constraint together with the effect of sales on prices. When the constraint is loose, the same mechanism does not operate. Calling an effect pecuniary identifies its route through prices; it does not settle its consequences under every institutional condition.
Even a clear outside cost need not disappear at the best attainable quantity. Suppose an activity produces private net benefit of twenty times its quantity minus the square of that quantity, and imposes damage of four for each unit. Quantity is divisible and nonnegative; the money-valued schedules include all effects being counted. Private choice reaches ten units. Counting the damage lowers the best quantity to eight, where the additional private benefit equals the additional damage. Total net benefit rises from sixty to sixty-four. Damage still amounts to thirty-two. Its persistence means that the remaining activity is worth more than the harm it causes under these assumptions. Eliminating the activity altogether would also eliminate the larger benefit it provides.
A charge of four per unit would make the private chooser select eight, while raising thirty-two in revenue. The improvement is four, not thirty-two, and the charge does not make the residual harm unreal. It changes the reason for undertaking another unit. Nor does the example establish that some positive amount of harm must always be accepted. If damage per unit were twenty or more in the same model, the best nonnegative quantity would be zero. The boundary is a comparison at the relevant margin. Neither the existence of damage nor the presence of a price supplies the answer in advance.
The remedy has an edge of its own. A rule that costs seven in additional real resources to obtain the four of improvement in this example would lower the counted total by three. A more effective rule might pass that test, and a right or a distributional purpose could furnish a reason not represented in the stipulated money sum. The narrow calculation establishes only what this particular means contributes to this particular account. It cannot turn an expensive procedure into evidence that no one was harmed. Coase's insistence on the cost of operating an arrangement is useful precisely because it makes a remedy answerable to the same attention to consequences that exposed the original omission.
Assigning a property right changes the starting position without automatically completing the account. Someone must know what the right covers, discover an infringement and make an agreement or a remedy effective. Ownership of the chimney does not by itself settle every interest in the air, just as ownership of a field does not specify every permitted use of it. Rights define actions people may take in relation to one another. Their allocation can make bargaining easier, place an obligation on the party able to avoid a harm, or leave the parties unable to reach an arrangement. The costs and consequences depend on the rights actually specified and the institutions that give them force.
The most difficult edge concerns what should count as a claim at all. A person can dislike another person's reading, worship or association. That preference is a fact about the person who holds it, but turning it into a charge on the other person's freedom requires a further judgment. One can give a sincere dislike a number without giving its holder a right to control its object. The same distinction matters when injury, dignity or an inherited obligation resists an easy monetary description. An economic account can clarify opportunities and consequences; a decision about whose conduct may be priced also defines the terms on which people live together. Treating every expressed aversion as a compensable loss would smuggle that decision into the arithmetic.
The concept remains useful when it makes these judgments visible. An inquiry can ask whether an effect is already reflected in a contract, whether it alters a real opportunity, whether a constraint makes an otherwise ordinary price movement costly, and whether the proposed rule respects the interests it is meant to reconcile. A disputed value is a reason to examine evidence and alternatives. It is not a reason to enter zero without argument. Equally, the language of spillover cannot make every proposed charge justified. The account must identify both the consequence being omitted and the change that would answer it.
An effect on someone else begins the inquiry; it does not decide the claim.
The externality has brought the stranger into the account, but it cannot by itself tell us everything the stranger is owed. A price can make one omitted consequence bear on a decision. A rule can protect an interest that a bargain will not reach. A shared institution can hold together what separate choices leave exposed. Each response needs a boundary, a purpose and an account of what it changes. The next volume returns to a buyer and a seller who stand inside the same transaction but know different things about it. The account can omit a consequence because it falls on a stranger; it can also fail because one party cannot see what the other already knows. The argument remains open.