In 1968 the ecologist Garrett Hardin argued in “The Tragedy of the Commons” (Science 162, 1243–1248) that a resource open to all is doomed: each user rationally takes as much as they can, the costs are shared by everyone, and the resource collapses. Hardin saw only two escapes – privatize the resource, or hand it to a coercive central authority (“Leviathan or private property”).
The political economist Elinor Ostrom spent four decades showing that this dichotomy is false. Studying real irrigation systems, fisheries, forests and pastures that had been managed sustainably for centuries, she documented a third path: self-governance by the users themselves. In 2009 she became the first woman awarded the Nobel Memorial Prize in Economic Sciences, “for her analysis of economic governance, especially the commons” (shared with Oliver Williamson).
Her central correction: Hardin’s tragedy describes open access (no rules, no boundary, no accountability), not a commons. A commons is a common-pool resource governed by an identifiable community under rules that the community makes and enforces. The interesting question is not whether commons can survive but which institutional arrangements let them survive – the subject of her 1990 book Governing the Commons (Cambridge University Press).
From the cases that endured, Ostrom distilled eight design principles – institutional features that long-lived, self-governed commons share. They are descriptive regularities, not a blueprint, but they hold up: a 2010 review by Michael Cox and colleagues, “A Review of Design Principles for Community-based Natural Resource Management” (Ecology and Society 15(4):38), re-examined 91 studies and found the principles broadly empirically supported, refining the original eight into eleven.
The refinement is worth reading rather than citing, because it says which principles carry the empirical weight. Cox and colleagues coded 91 studies covering 77 cases, and every principle came out with at least twice as many supportive as unsupportive cases; a Fisher’s exact test between the presence of a principle and reported success was significant at the 5% level for all of them except principle 8, nested enterprises, which reached only 10%. Their reformulation splits principles 1, 2 and 4 in two and leaves 3, 5, 6, 7 and 8 exactly as Ostrom wrote them – which is how eight becomes eleven.
Boundaries divide into user boundaries and resource boundaries (the first strongly supported, the second only moderately); congruence divides into congruence with local conditions and the proportionality of appropriation to provision; and monitoring divides into social monitoring, users watching one another, and environmental monitoring, users watching the resource. Cox notes that this last split is deliberately not the one used in the coding, where the two halves were the presence of monitors at all (moderately supported) and whether those monitors were themselves community members (very strongly supported) – the single strongest result in the review, and the one that most directly contradicts the instinct to hire an outside auditor.
The eighth principle points to Ostrom’s wider thesis, set out in her Nobel lecture “Beyond Markets and States: Polycentric Governance of Complex Economic Systems” (2009). Complex resource problems are best handled not by a single global optimizer but by many overlapping, semi-autonomous decision centres – polycentricity – that each operate at their own scale and interact through rules of mutual adjustment. No unit is sovereign over the whole; robustness comes from redundancy and local adaptation rather than central design.
This is a direct rejection of the Leviathan-or-market framing. It also anticipates a structural pattern that reappears in subDAOs and working groups: a treasury or protocol that grows past what one assembly can competently govern tends to fracture into nested units with delegated scope.
A DAO treasury is a textbook common-pool resource: a shared pool that any member can draw on, where one member’s spending diminishes what remains for the rest. That makes Ostrom’s principles a useful lens for reading DAO mechanism design – and for reading how DAOs fail, which is usually a failure to satisfy one of them:
Where the principles are hard to satisfy on-chain, the failure modes are predictable: absent monitoring incentives, a treasury is quietly captured; absent a real conflict-resolution arena, disputes escalate straight to forks and exits. The recurring lesson from institutional economics holds here – with positive transaction costs, the rules and who gets to make them determine the outcome.
Alex Tabarrok, an economist at George Mason University, uses commons in a wider sense in “Blockchains and the Opportunity of the Commons” (Marginal Revolution, June 2018). His commons is an unowned place, platform or protocol through which people meet, communicate and trade, such as TCP/IP, HTTP, GPS or the English language. He contrasts it with privately owned platforms such as Uber, Facebook and YouTube, whose owners maximize the value of the company rather than the value to the people who use them.
A blockchain, he argues, lets a platform's creator commit credibly never to take that value while still being paid, through the issue and appreciation of the platform's token, so tokens make it worth building a commons; his example is LBRY, a video platform on an unowned protocol and a firm he advises. His caveat is that decentralized institutions such as markets and language have almost all evolved rather than been designed, and a designed one has to be close to right from the start, because changing it later means a costly fork. Platform economics more generally is covered under network effects and platforms.
Ostrom's fourth and fifth principles assume that someone monitors the commons and imposes the sanctions, and evolutionary game theory asks why anyone would, since punishing a free rider costs the punisher. Karl Sigmund, a mathematician at the University of Vienna and one of the pioneers of evolutionary game theory, compared two answers with Hannelore De Silva, Arne Traulsen and Christoph Hauert in “Social learning promotes institutions for governing the commons” (Nature, 2010). Under peer punishment, members fine free riders themselves after the event. Under pool punishment, members pay into a sanctioning fund before the joint effort begins, the equivalent of hiring a police force.
Peer punishment is cheaper, since it costs nothing while everyone cooperates, but it invites second-order free riders: members who contribute to the common good and leave the punishing to others. Pool punishment exposes them, because a refusal to pay into the fund is visible even when nobody defects. In the authors' model, once second-order free riders are punished as well, pool punishers do better than peer punishers, and the sanctioning institution emerges from trial, error and the imitation of successful examples alone, with no altruism and no higher authority. Pool punishment gives up some efficiency in exchange for stability.
In a 2018 conversation with Edge, Sigmund names the weakness of the institutional answer: the sheriffs, referees and judges who enforce the rules can be bribed. Peer punishment risks feuds that escalate into a war of all against all; institutional punishment risks corruption. In DAO terms, a council or security committee paid from the treasury to police it is a form of pool punishment, with the same stability and the same exposure to bribery described under vote markets and governance bribery.
The principle a DAO is most likely to satisfy on paper and lose in practice is the third: most of those affected by the rules can participate in changing them. Token voting satisfies it by construction, and then the DAO discovers that an electorate cannot be summoned for routine business and starts handing recurring decisions to committees, stewards and auto-adjusting modules. Each of those grants is rational on its own terms, and collectively they move rule-making out of the arena where the affected users sit.
Ostrom's framework does not forbid this – principle 8 explicitly expects layered, nested decision centres in any large system – but it does supply the test. A nested tier is legitimate in her account when the users retain the ability to constitute and re-constitute it; a mandate that never expires and is never revisited is a tier the users no longer choose. That is the distinction standing authorizations and delegated mandates examines mechanism by mechanism, and the reason it turns out that DAOs bound delegated authority by magnitude, direction and frequency but almost never by duration. Read against principle 3, an expiry is not an extra safeguard; it is the thing that keeps the delegation a delegation.
The measurable version of the failure is not a lost vote but the absence of votes beside continuing treasury activity, which is exactly the diagnostic Ostrom's monitoring principles ask a commons to run on itself – and the review above found that monitoring works best when the monitors are the users. See also voter apathy, council dissolution, and hyperstructures, the limit case in which there is no rule-making tier left to capture because the protocol was built without one.
Caper’s answer leans hardest on Ostrom’s first and sixth principles – a hard boundary and a cheap way out of a dispute. Every member who has voted and still holds tokens has a credible, priced exit: rather than being outvoted and locked in, a member who has taken part can leave and withdraw their canonical share of the common treasury. Both conditions are asserted on the way in, so neither is rhetorical – the call aborts on an empty vote-token bucket and again on an empty governance-token bucket. That share is not negotiated or discretionary – it is the same weight the protocol uses for voting, fixed by formula.
The exit is not free, and it is worth being exact about where the cost sits: the treasury share itself is paid out untaxed, but the leg that converts the departing member’s governance tokens back to XRD pays the skim every other sale pays, at the rate the curve carries at that point, and the skim goes to the caper’s own treasury, its founder and the Caper Commons rather than to an operator. The treasury’s part lands after the exit share is paid, so the leaver is not paid a slice of their own skim.
What matters for Ostrom’s purposes is not that the exit is costless but that its price is fixed and public: a faction that captures a majority cannot strand a minority in a treasury it no longer trusts – the minority walks with its share. In Ostrom’s terms, a guaranteed exit is both a clear boundary and a low-cost conflict-resolution mechanism, and it changes the collective-action calculus before any vote is cast. See what is a caper and rage-quit and exit rights.