Elinor Ostrom · Beyond Markets and States: Polycentric Governance of Complex Economic Systems
The most famous trap in the economics textbook is the Prisoner's Dilemma: two suspects are questioned separately, each does best by betraying the other, and both end up serving longer. For decades economists used it to explain why a group of people cannot hold on to a shared pasture, fishery or forest. On December 8, 2009, in a lecture hall at Stockholm University, a 76-year-old political scientist knocked the floor out from under that analogy in two sentences:
Public investigators purposely keep prisoners separated so they cannot communicate. The users of a common-pool resource are not so limited.— Elinor Ostrom, Nobel Prize Lecture, 2009
Elinor Ostrom was the first woman to receive the economics prize, and her doctorate was in political science — UCLA's economics department had discouraged her because she had no mathematics.
Her target was the biologist Garrett Hardin's 1968 essay "The Tragedy of the Commons." "Picture a pasture open to all": each herdsman gains the full value of one more animal while the cost of overgrazing is shared by everyone, so each adds another, and another, until the pasture is ruined. Two exits were on offer: let the state run it, or divide it into private property. Market or state, pick one.
Ostrom spent half a century looking at real pastures. Her team screened more than 500 case studies scattered across anthropology, history and engineering and found that Nepali irrigation systems built and run by farmers grew more rice and shared water more fairly than government-run ones, and that fishing villages, irrigation districts and forest communities had kept the same resource going for long stretches using rules no outsider could read. She reproduced it in the lab. Put eight strangers in a "common-pool" game and forbid them to talk, and they grab even more than theory predicts. Let them talk face to face — "cheap talk," which game theory says should change nothing — and overharvesting drops sharply. Let them design their own penalties, and they reach 90 percent of the optimum.
From the successful cases she distilled eight "design principles" — clear boundaries, rules that fit local conditions, sanctions that start small, and so on. But the conclusion she pressed hardest was the less comfortable one: there is no panacea. Her named example was Individual Transferable Quotas being recommended for every fishery on earth; handing resources to government, privatizing, or decentralizing can each fail on its own. She was not proposing to replace the state with the community. She was arguing that governance can have many centers, nested inside one another. That is what "polycentric" in the title means.
What happened next? The core claim held. In 2010, Cox and colleagues reviewed 91 studies that had tested the principles and found them well supported; most of the criticism was theoretical rather than counter-evidence from the field. In 2019, three scholars outside her circle, writing in the Journal of Economic Perspectives fifty years after Hardin, traced his error to two conflations: a resource with the way it is governed, and unmanaged open access with a commons. But the idea has been pulled in two directions she explicitly rejected. Some treat the eight principles as a recipe to copy — the very panacea she warned against. Others read her as a standard-bearer against government and privatization; in 2019 two scholars wrote a paper to correct that. She was never either. And the lecture is plain about the limits: "Many groups have struggled and failed."
Section four opens with a question used as a heading: "Are Rational Individuals Helplessly Trapped in Social Dilemmas?"
The audience was economists, and every one of them knew the Prisoner's Dilemma: two suspects questioned separately, each better off betraying the other. For half a century the model had been carried over to overfishing, overgrazing and pumped-dry aquifers — everyone knows restraint would help, nobody dares to restrain first. Ostrom did not attack the model's mathematics. She attacked a piece of stage scenery the model takes for granted:
Public investigators purposely keep prisoners separated so they cannot communicate. The users of a common-pool resource are not so limited.— Elinor Ostrom, Beyond Markets and States, 2009
The prisoners are trapped because someone deliberately stops them from talking. Fishers, herders and irrigators pass each other on the same road every week. Take down that wall, and whether the dilemma is still a dilemma becomes an empirical question to be settled in the field, not a logical necessity. She went on: modeling people as waiting for outsiders to change the rules was "a retrogressive step."
Elinor Ostrom was born in Los Angeles in 1933 to a family of simple means and was, in her words, a poor kid in a rich kid's school at Beverly Hills High. Looking for work after college, she found employers' first question was whether she could type and take shorthand. UCLA's economics department discouraged her from a PhD: she had no mathematics, because as a girl she had been advised against courses beyond algebra and geometry. Political science admitted her, in a class of 40 that included four women; after they started, they learned the faculty had held "a very heated meeting" criticizing the committee for admitting women at all and offering them assistantships.
In 1965 her husband, the political scientist Vincent Ostrom, was offered a full professorship at Indiana University, and she "tagged along as it was very hard for any department to hire a woman in those days." When the department later needed someone to teach Introduction to American Government on Tuesdays, Thursdays and Saturday mornings at 7:30 a.m., it hired her as a visiting assistant professor. As a graduate student she had been assigned to study the West Basin, a groundwater basin under the Pacific coast of Los Angeles County: heavy pumping, seawater intrusion, and a large group of public and private water producers trying to agree on what to do — "Without knowing I was studying a common-pool resource problem." Hardin's essay would not appear until three years after she defended her dissertation.
In October 2009 the economics prize went to her and Oliver Williamson, hers "for her analysis of economic governance, especially the commons." She was the first woman to win it, and a political scientist. When she took the podium in Stockholm that December she was 76, and her subject was her own fifty-year road: California water agencies, police departments in 80 U.S. metropolitan areas, irrigation canals in Nepal, forests in about a dozen countries. Two and a half years later, on June 12, 2012, she died in Bloomington.
She begins with the old worldview. Mid-twentieth-century social science ran on three simple parts: two optimal organizations (the market for private goods, the state for everything else), two kinds of goods (Samuelson's private and public), and one kind of person (fully informed, fully rational, counting only their own payoff). Reality that didn't fit was treated as a mess to be cleaned up. Her first example is not a pasture but a city. The prevailing view held that a metropolitan area with dozens of police departments and water agencies was "chaotic" and ought to be consolidated into one government. In the 1970s she and colleagues studied policing in 80 metropolitan areas: direct services like patrol were more efficient with many small producers, indirect services like crime labs with a few large ones — a mix of large and small worked best.
We demonstrated that complexity is not the same as chaos in regard to metropolitan governance.— Elinor Ostrom, 2009
Then she doubles the types of goods. Sort goods on two axes — how hard it is to exclude people, and whether one person's use subtracts from another's — and alongside private and public goods you get toll goods such as theaters and clubs, and the crucial fourth type: common-pool resources, hard to exclude like public goods, yet depletable like private ones. Forests, water systems, fisheries and the atmosphere all sit in that box.
Only now does she bring Hardin on stage. His pasture "open to all" was widely accepted because it matched the Prisoner's Dilemma prediction of no cooperation; people came to presume that every common-pool resource was owned by no one and that only government could impose new rules from outside. Then come the two sentences from "the moment": the prisoners are kept apart; resource users are not.
The evidence moves along three lines. The first is the literature. When the U.S. National Research Council convened scholars from many disciplines in the mid-1980s, it became clear that anthropologists, historians and engineers had long documented communities governing their own resources — but the studies were scattered across fields and countries and had never been read together. Her team spent more than two years on a coding manual and screened over 500 case studies; only a few recorded the actors, strategies, resource condition and rules together, and what made it into the database was 44 inshore fishing groups and 47 irrigation systems. Every self-organized success had boundary rules, allocation rules, active monitoring and local sanctions. And the "grim trigger" — punish one defection with permanent retaliation, a staple of repeated-game theory — they "did not find a single case" of.
She expected the next step to be statistics showing which specific rules went with success. On a 1988 sabbatical in Bielefeld, Germany, she hit a wall:
Finally, I had to give up the idea that specific rules might be associated with successful cases.— Elinor Ostrom, 2009
They had recorded 27 kinds of boundary rule and 112 kinds of choice rule. So she moved up a level of generality and proposed the eight "design principles" — not principles the fishers carried in their heads, but the regularities that separated long-lived institutions from failed ones.
The second line is the laboratory. With Roy Gardner and James Walker, she had eight subjects divide tokens between "Market 1," with a fixed return, and "Market 2," a common pool whose return depended on everyone's investment. Theory predicted 8 tokens each, 64 in total; if the group put in only 36, everyone would earn more. Subjects invested even more than the Nash prediction. Then one thing changed: after ten rounds they could talk face to face. Game theory says this does nothing, since no one enforces the promises — yet joint returns rose sharply. When they could pay to fine one another, they sanctioned far more than the predicted zero, and overuse of sanctions ate the net gains. The group that first discussed and chose its own sanctioning system got the best result of any of the experiments: 90 percent of optimal after fines.
The third line is the field. The Nepal irrigation database grew to 229 systems, and farmer-governed systems outperformed government ones on physical condition, water reaching the tail end, and agricultural productivity. In 163 forests, those legally designated as government protected areas showed no statistical difference in forest density from other forests; what did track forest condition was whether local users monitored.
She closes in two places. One is a limit on her own case:
On the other hand, we cannot be overly optimistic and presume that dilemmas will always be solved by those involved. Many groups have struggled and failed— Elinor Ostrom, 2009
The other is policy: rather than designing institutions "to force (or nudge) entirely self-interested individuals," the core goal of public policy should be "to facilitate the development of institutions that bring out the best in humans." And a methodological farewell: when the world is not well described by a simple model, improve the theory so as "to be able to understand complexity and not simply reject it."
Hardin's essay was not really about pastures. The 2019 Journal of Economic Perspectives retrospective points out that its subtitle was "The Population Problem Has No Technical Solution," that its main target was population growth, and that Hardin called the "freedom to breed" "intolerable" and argued for mutual coercion. The pasture was an allegory, and its first line — "Picture a pasture open to all" — already builds in the conclusion. It describes unmanaged open access, not a commons with members, boundaries and rules. Ostrom's unbundling of property rights in the lecture is aimed squarely at this: users may hold some of five rights — access, withdrawal, management, exclusion, alienation — and scholars who assumed that no right to sell meant no property rights simply could not see the local rules.
The evidence base is thinner than "far more frequently than expected" suggests. The founding meta-analysis rests on 44 fishing groups and 47 irrigation systems, all second-hand case studies written by others and filtered from more than 500 to find enough complete variables. When the printed lecture cites Tang's irrigation figures, they do not add up: "Of the 47 irrigation systems … 12 were managed by governmental agencies of which only 40 percent (n = 7) had high performance. Of the 25 farmer-managed, over 70 percent (n = 18)." Twelve plus 25 is not 47, and 7 of 12 is not 40 percent. The direction is unaffected, but the figures on the page have passed through several hands.
The forest finding rests on an unusual measurement. "Density" in those 163 forests was not computed from satellites or a census; the forester who had just supervised plot measurements rated it on a five-point scale, as a footnote in the lecture explains. And the finding that protected areas did not differ from other forests is narrow: the legal label of protection, by itself, is unrelated to forest condition; what matters is whether someone on the ground is watching. It does not mean protected areas are useless.
She walked back the status of the design principles herself. The lecture uses the version revised by Cox and colleagues in 2009, splitting boundaries into user and resource boundaries and monitoring into monitoring users and monitoring the resource. Footnote 5 reads:
The term “design principle” has confused many readers. Perhaps I should have used the term “best practices” to describe the rules and structure of robust institutions.— Elinor Ostrom, 2009, footnote 5
The lecture says Cox and coauthors analyzed "over 100 studies" and that "two-thirds" confirmed the principles. When the review was published in 2010, the figure was 91 studies, with an average evaluation of 3.73 — slightly below 4, "moderately supportive." Both numbers belong here: same direction, but the lecture's version is rounder.
Where the lecture sits in her thinking. The last two sections are really about what comes next: replacing the all-knowing rational actor with boundedly rational people who learn and internalize norms, putting trust at the center of collective-action theory, and building a larger social-ecological systems framework. Polycentricity was not a conclusion. It was a set of blueprints left for the next builders.
HELD UP · as of 2026-10
The core claim — the commons is not doomed, self-governance works under identifiable conditions, single blueprints fail — has survived later reviews and an independent retrospective and is now mainstream. But it has been pulled in two directions she explicitly rejected: the eight principles as a universal recipe, and Ostrom as a champion against the state.
December 2010. Cox, Arnold and Villamayor-Tomás published in Ecology and Society the review she had cited in advance: 91 studies, and the principles "well supported empirically"; every principle had at least twice as many supportive as unsupportive cases. The lowest-scoring group was nine "abstract" studies — the criticism was mostly theoretical. The most-criticized principle was the first, clear boundaries; Turner wrote that in agro-pastoral settings "Rules of access are often politically malleable and spatial boundaries fluid." A caveat: this is work Ostrom adopted directly in the lecture, so it is not an independent check.
2016. Araral had criticized the meta-analysis in a paper titled "Ostrom, Hardin and the commons" (this page could not read it and does not restate its argument). Cox and colleagues replied in Environmental Science & Policy, correcting what they saw as factual errors while conceding that the critique had hit two real problems: the difficulty of unpacking complex variable relationships, and the conflation of diverse outcomes — challenges, they said, for the whole commons field. The same year, Baggio and colleagues ran a configurational analysis of 69 irrigation, fishery and forestry cases and found that which principles matter depends on the resource: clear social boundaries for mobile resources like tuna, monitoring for static ones like forests or water in canals; only congruence with local conditions and proportionality between investment and extraction mattered regardless. They also wrote that the principles "have often been viewed as a panacea," and argued that this did not reflect Ostrom's intent.
2019. Frischmann, Marciano and Ramello, from law and economics, published "Tragedy of the Commons after 50 Years" in the Journal of Economic Perspectives — the retrospective on this page from outside Ostrom's group. Their verdict: in light of the Ostroms' work, Hardin made two conceptual mistakes, conflating a resource with its governance and conflating open access with a commons, and "Hardin's distorted perspective on the commons often persists to this day in economic discussions." The same year, Sarker and Blomquist, in the Journal of Institutional Economics, addressed the opposite misreading: her critical stance had led people to see her work as "anti-Hardin, anti-tragedy, or more specifically, anti-statist and anti-privatization," and it was neither. The lecture itself is the evidence: government recognition of local rights and nested layers are both among her principles.
Today's consensus runs roughly like this: Hardin's world of anonymous people who never talk does exhaust resources, and Ostrom herself produced that tragedy again and again in the lab; but it is a special case, not the normal one. What remains open is the harder half. Are the design principles causes of success, or traits of the successful? Can what works in a small community scale up to a global common pool like the atmosphere? The lecture ends with a to-do list, not a finished map.