The Sanction

The neighbor who sees an extra animal entering the pasture has a second decision to make. He can report the breach, confront its owner, or let the matter pass. The grass lost to the animal is shared among the users; the time, discomfort, and possible retaliation involved in objecting fall first on him. Enforcement therefore presents a smaller version of the problem it is meant to solve. Everyone may benefit from a rule that holds, while each would prefer another person to bear the expense of making it hold. The covenant must organize this work along with the grazing. A limit written beside the gate needs someone able to discover a violation, a consequence that can actually follow, and a procedure through which the accused can answer the charge.

Monitoring makes conduct visible, and a graduated sanction gives the community a range of responses to what it sees. Ostrom found both among the recurring features of durable resource institutions: users, or monitors accountable to them, observe conduct; penalties begin at a low level and increase with repeated breaches.¹ The distinction between an occasional lapse and persistent appropriation matters to the future of the agreement. A warning can clarify a misunderstood turn; a repeated refusal may require a fine or loss of access. Meanwhile, the other users learn whether restraint is being reciprocated. A rule applied selectively to the weak supplies quite different information from one applied consistently. The community is establishing expectations about its monitors as well as about its herdsmen, for authority to punish creates an opportunity to misuse punishment.

The future can itself supply an incentive to respect the present agreement. Consider two users who receive three each in a season of mutual restraint, one each when both overuse, and respectively five and zero when one overuses while the other restrains. Each values next season’s payoff at a fraction δ of this season’s. Suppose conduct is perfectly observed and both follow the same rule: cooperate until a breach, then overuse forever. Continued cooperation is worth 3/(1 − δ); a unilateral breach yields 5 + δ/(1 − δ). Cooperation is sustainable against this deviation when δ is at least one-half. At δ = three-quarters, twelve exceeds eight. At δ = zero, five exceeds three and the temptation prevails. The promised loss of a continuing cooperative return changes the calculation, provided the users value that return enough and expect the response to occur.²

This permanent retaliation is a severe instrument. One mistaken accusation destroys all subsequent cooperative gains in the example. Its deterrent power rests on perfect observation, an indefinitely continuing relationship, and the specified willingness to carry out the threat. With a commonly known final season and purely selfish, fully rational players, the same one-period incentives unravel cooperation backward from the end. Real communities have to investigate conduct whose causes are uncertain and decide whether a relationship can be restored. Ostrom’s account of measured responses describes participants reacting mildly to small departures while trying to identify and correct their source.³ Graduated enforcement serves that practical problem. Its particular schedule still has to deter the relevant breaches; the calculation for permanent retaliation establishes no automatic threshold for warnings, temporary exclusions, or forgiveness. Those arrangements have different future payoffs and require their own incentive comparison.

People also sometimes pay to punish when a future private return cannot explain the expense. In Fehr and Gächter’s published experiment, participants allocated tokens between themselves and a group project, then in one condition could spend their own tokens to reduce another participant’s payoff. Groups of four either remained together or were randomly reassembled. The design prevented individual reputation building across periods, and each treatment had a known ten-period horizon. Contributions were higher with punishment opportunities in both arrangements; the fixed groups approached high cooperation while the no-punishment condition declined.⁴ These results establish a capacity under specified conditions, leaving open how it will operate elsewhere. They also distinguish a willingness to bear punishment costs from the narrow future-return incentive of the preceding example. A community may be able to draw on several motives for enforcement, whose strength and direction it must learn from conduct.

The cost of using the sanction must enter the institution’s accounts. In Ostrom, Walker, and Gardner’s common-resource experiments, sanctioning without communication increased resource yields relative to the baseline comparison, yet returns fell below that baseline once punishment fees and fines were counted. Groups that could communicate and chose a sanctioning system achieved much higher net returns with few breaches. The choice was endogenous: the outcome of the groups adopting the arrangement does not establish what forced adoption would have produced for every group.⁵ Nor is every fine in a real commons a resource destroyed. A payment into a common treasury is a transfer; the time spent investigating, collecting, disputing, or retaliating uses resources. In the experiment both fees and fines reduced participant earnings. A field comparison must identify what is transferred, what is consumed, and what future damage is prevented before declaring enforcement worthwhile.

A successful sanction helps preserve a workable agreement, including the possibility of returning to it. Its severity alone says little about that achievement. The users need to know the rule, the evidence, the consequence, and how an error can be corrected; the person enforcing it needs support sufficient to make enforcement more than an occasional act of courage. These arrangements join the trust of the preceding plate to observable conduct. Their reach becomes harder to secure as those affected cease to share a gate, a meeting place, or an authority they recognize. The final plate, The Planet, carries the question outward: how a discipline learned among neighbors can operate when the resource and the consequences extend across the world.