↵ select ↓ ↑ navigate esc close

Paper Feed: Prize Lecture by Elinor Ostrom

Paper Feed ·

Listen: https://serve.buildtall.systems/83af3bf1d6d01784578106a8f2cb7a9d2039422febf0b8bcfa4a62dfb32e4859.mp3

Source: https://serve.buildtall.systems/5d11eb59b78e94bef479e66216078231dc803aa9bedd907314a96aba0e53c5f1.pdf


This episode covers the Prize Lecture by Elinor Ostrom, delivered on December 8, 2009, in which Ostrom looks back over roughly half a century of research on how people govern shared resources such as fisheries, forests, irrigation systems and groundwater, and the public goods that communities and governments provide. The lecture is Ostrom's own account of that work and of what it has overturned.

First, the record: OpenAlex has no record of this document. No work in the index matches its title and authors, so this episode makes no claim about how often it has been cited. The document is a lecture rather than a single study. Ostrom, then of the Workshop in Political Theory and Policy Analysis at Indiana University and of the Center for the Study of Institutional Diversity at Arizona State University, traces the path of her research and that of colleagues, and sums up what it found about how people manage common-pool resources, which are resources many people can use but that are hard to keep others from using, and public goods at several scales. Its central argument is that the older division of the world into markets and states cannot account for the many arrangements people actually build, and that polycentric systems, with many overlapping centers of decision-making, deserve attention. The evidence it draws on is broad but uneven. It reports a 1970s study of police services across 80 metropolitan areas, a screening of more than 500 case studies that produced comparable sets of 44 inshore fisher groups and 47 irrigation systems, laboratory experiments with small groups of university students, field experiments with more than 200 forest users in Colombia, and later Nepal irrigation and forest studies covering samples of 152, 163 and 178 forests or user groups. The strongest parts are the large cross-site comparisons and the repeated pairing of laboratory and field tests. The thinner parts are the case-study sets, which were assembled from work by scholars with different methods and vocabularies, and the lecture itself admits that the many variables acting at once in the field make single effects hard to isolate. The text does not state any funding source, and it names no journal or other venue beyond the lecture and its date.

Ostrom traces the path of her work over roughly half a century, beginning with her graduate studies in the late 1950s. She calls the early efforts to understand the water industry in California formative. She worked alongside Vincent Ostrom and Charles Tiebout as they formulated the concept of polycentric systems, meaning many overlapping centers of decision-making, for governing metropolitan areas. Her own first close study was of a large group of private and public water producers facing an overdrafted groundwater basin on the coast, pumped faster than it could refill, while saltwater intrusion threatened the basin's long-term use. Groundwater of this kind is a common-pool resource, one that many users draw on and that is hard to keep others from using, so who governs it and how shapes whether it lasts.

In the 1970s she joined colleagues in studying polycentric police industries serving U.S. metropolitan areas. The lecture says this work found that "the dominant theory underlying massive reform proposals was incorrect." Its central finding was that

Metropolitan areas served by a combination of large and small producers could achieve economies of scale in the production of some police services and avoid diseconomies of scale in the production of others.

Economies of scale means that producing more lowers the cost of each unit, and diseconomies of scale means that producing more raises it. Ostrom presents this as evidence that a mix of large and small providers can be efficient, which the reform theory had not allowed for.

These early empirical studies led over time to the Institutional Analysis and Development (IAD) framework, a common approach consistent with game theory. The lecture says that this framework let the researchers carry out a variety of empirical studies, including a meta-analysis of existing work.

Ostrom's lecture first reviews a body of field and laboratory work on common-pool resources, which are resources that many people can use but that are hard to keep others from using, such as forests and irrigation systems. Carefully designed experiments in the lab, which tested precise combinations of structural variables, found that isolated, anonymous individuals overharvest from such resources.

The result that stands out is what happened when participants were allowed to talk. The lecture says that simply allowing communication, or “cheap talk,” enables participants to reduce overharvesting and increase joint payoffs, contrary to game-theoretical predictions. Large studies of irrigation systems in Nepal and of forests around the world, the lecture argues, challenge the presumption that governments always do a better job than users in organizing and protecting important resources.

Many scholars are now building new theory. The lecture describes one core effort as developing a more general theory of individual choice that recognizes the central role of trust in coping with social dilemmas. A clear set of findings has emerged at the microsituational level, meaning the small settings in which individuals make their own decisions, about the structural factors that make increased cooperation more or less likely. Broader field settings are far more complex, so the lecture calls for more configural approaches, which study how factors combine, to explain what helps or hinders self-organized efforts within multilevel, polycentric systems, where many overlapping centers of decision-making operate at once.

For policy, the lecture draws a practical conclusion. Institutional rules need to fit a specific social-ecological setting, and the lecture is direct about the alternative: “One-size-fits-all” policies are not effective.

The lecture then turns to what it calls the earlier world view of simple systems. In the mid-twentieth century, the dominant scholarly effort was to fit the world into simple models and to criticize any institutional arrangement that did not fit. Ostrom names Herbert Simon and Vincent Ostrom among the scholars around the world who challenged the basic assumptions of that period.

The first assumption concerns two optimal organizational forms. The market was seen as the best institution for producing and exchanging private goods. For nonprivate goods, the older view held that one needed “the” government to impose rules and taxes, forcing self-interested individuals to contribute necessary resources and refrain from self-seeking activities. Without a hierarchical government to induce compliance, self-seeking citizens and officials would, in this view, fail to supply efficient levels of public goods such as peace and security, at multiple scales. Public goods are goods that are shared by many people and are difficult to withhold from any of them.

The same logic led scholars to recommend a single governmental unit for metropolitan areas. Such a unit, it was argued, would reduce the “chaotic” structure of metropolitan governance, increase efficiency, limit conflict among governmental units, and best serve a homogeneous view of the public. The lecture's point is that this split between market and government explained exchange of strictly private goods, but it has not adequately accounted for the internal dynamics within private firms. Nor does it adequately deal with the wide diversity of institutional arrangements found in practice.

The lecture sets out the two basic ideas it will work against. The first is Paul Samuelson's classic division of goods. In his definitional essay of 1954, Samuelson split goods into two types. Pure private goods are excludable, meaning an individual can be kept from consuming them unless they pay, and rivalrous, meaning that whatever one person consumes, no one else can. Public goods are the reverse on both counts: they are nonexcludable, which the lecture describes as impossible to keep those who have not paid from consuming them, and nonrivalrous, so that what one individual consumes does not limit consumption by others. The lecture notes that this division fitted a wider picture of the institutional world, in which private property was exchanged in markets and government-owned property was run through public hierarchies. In that picture, people were seen mainly as consumers or voters.

The second idea is a model of the individual. The lecture says that mainstream economics and game theory generally accepted that individuals are fully rational. Such a person is presumed to know every strategy open to them in a situation, to know which outcome each strategy leads to given how others are likely to behave, and to hold a ranking of those outcomes by their own preferences, measured as utility. The rational strategy in every situation is then to maximize expected utility. The lecture observes that utility was first conceived as a way of combining many external values on one internal scale, but in practice it came to be equated with a single externalized measure, such as expected profits. Ostrom's judgment is that this model has produced useful and empirically validated predictions about exchange transactions for goods with specific attributes in a competitive market, but not across a range of social dilemmas. She says she will return to the theory of individual behavior in section 7A.

The lecture then turns to the early efforts to build a fuller understanding of complex human systems. It describes how the mid-twentieth-century view of simple systems has slowly been changed by extensive empirical research and by a framework consistent with game-theoretical models, one that can be applied to a broad array of questions. The first of these efforts is the study of polycentric public industries. The lecture describes empirical work on how citizens, local public entrepreneurs and public officials engage in diverse ways of providing, producing and managing public service industries and common-property regimes at multiple scales. In its words, this work has generated substantial knowledge that is not explained by the simple models that preceded it.

In 1961, Vincent Ostrom, Charles Tiebout and Robert Warren introduced the concept of polycentricity to address a question about metropolitan areas. Many public and private agencies were providing and producing public services there, and other scholars had charged that the result was chaotic. The three authors wanted to test whether it was instead a productive arrangement. They describe a polycentric system as one with "many centers of decision making that are formally independent of each other." Whether those centers actually act independently, however, is a question to be settled case by case:

Whether they actually function independently, or instead constitute an interdependent system of relations, is an empirical question in particular cases.

The authors explain when such a set of jurisdictions could count as a system. To the extent that they take each other into account, enter into contracts and cooperative arrangements, or turn to central mechanisms to resolve conflicts, the jurisdictions "may function in a coherent manner with consistent and predictable patterns of interacting behavior."

To study this, the authors drew on the idea of a public service industry, borrowed from industrial economics, which treats the agencies that supply a service as a field whose performance can be compared. In the 1960s, several studies of water industry performance in different parts of California found substantial evidence that multiple public and private agencies had found productive ways of organizing water resources at several scales. These findings ran against the view that many governmental units without a clear hierarchy must produce chaos. The studies pointed to three mechanisms that raise productivity in these polycentric metropolitan areas. First, small and medium-sized cities were more effective than large cities at monitoring the performance of their citizens and the relevant costs. Second, citizens who were dissatisfied with a service could "vote with their feet" by moving to jurisdictions closer to their preferred mix and cost of public services. Third, local incorporated communities could contract with larger producers and change those contracts if they were not satisfied, while neighborhoods inside a large city had no comparable voice.

In the 1970s, the same line of work was extended to policing and public safety. Across 80 metropolitan areas, the researchers found that although many police departments served these areas, duplication, meaning more than one department serving the same set of citizens, rarely occurred. They also did not find the widely held belief that having more departments in one metropolitan area made policing less efficient. In their later summary of this work, the authors state that the

most efficient producers supply more output for given inputs in high multiplicity metropolitan areas than do the efficient producers in metropolitan areas with fewer producers.

In other words, the best-performing police forces were not the ones in areas with the fewest providers.

The lecture then turns to what the research showed about metropolitan governance. It reports that technical efficiency was also higher in metropolitan areas where a small number of producers supplied indirect services such as radio communication and criminal laboratory analyses, and it says these findings allowed the researchers to "reject the theory underlying the proposals of the metropolitan reform approach." In the lecture's words, "complexity is not the same as chaos in regard to metropolitan governance." The lecture says this lesson carried forward into later empirical studies of polycentric governance, meaning systems with many overlapping centers of decision-making, of resource and infrastructure systems across the world. The second half of the section concerns how goods are classified. The lecture says that studying how people cope with diverse public problems led the researchers to reject Samuelson's two-fold classification of goods. Buchanan, in 1965, had already added a third type, which he called "club goods": goods that groups could provide to themselves on a small scale, nonrivalrous in use, while excluding nonmembers from their benefits. The lecture then describes further modifications the authors proposed, which they argue affect the incentives individuals face. It lists four. First, the term "rivalry of consumption" is replaced with "subtractability of use." Second, subtractability of use and excludability are treated as varying from low to high, rather than as either present or absent. Third, a fourth type is added explicitly: "common-pool resources," which share subtractability with private goods and difficulty of exclusion with public goods. The lecture names forests, water systems, fisheries, and the global atmosphere as examples of immense importance for human survival. Fourth, the club good is renamed a "toll" good, since many goods with these characteristics are provided by small-scale public as well as private associations. A figure in the lecture gives an overview of these four broad types, and the lecture notes that each contains many subtypes. A river and a forest are both common-pool resources, yet they differ substantially in how mobile their resource units are, how easy they are to measure, how long they take to regenerate, and in their spatial extent, number of users, and other attributes.

Ostrom begins with what years of fieldwork showed her.

When one engages in substantial fieldwork, one confronts an immense diversity of situations in which humans interact.

She contrasts two scenes. Riding as an observer in a patrol car in the central district of a large American city at midnight on a Saturday, one sees one pattern of human interaction. In a suburb on a weekday afternoon, as school lets out, one sees another. In both, the observer watches an official produce a public good, local safety, but the people involved differ in age, sobriety, their reasons for being there, and what they are trying to accomplish. Those differences shape how the police officer behaves.

The second scene involves a different kind of problem. Private water companies, city utilities, private oil companies and local citizens meet in varied settings to assess who is to blame for overdrafting a groundwater basin, which has caused massive saltwater intrusion, and to decide what to do next. Ostrom writes that these people "all face the same problem", the overdraft of a common-pool resource, yet their behavior changes with the setting. They meet monthly in a private water association, face each other in a courtroom, and go to the legislature and eventually to the citizens to sponsor a Special Replenishment District. Groundwater basins, lakes, irrigation systems, fisheries and forests are the standard examples of common-pool resources.

The text also sorts goods by two properties: how subtractable their use is, and how difficult it is to exclude potential beneficiaries. Along these lines it names four types. Common-pool resources are one. Private goods, such as food, clothing and automobiles, are another. Public goods, such as peace and security, national defense, knowledge, fire protection and weather forecasts, form a third. Toll goods, such as theaters, private clubs and daycare centers, make up the fourth.

The central claim of this stretch follows from these cases.

These and many other situations observed in irrigation systems and forests in multiple countries do not closely resemble the standard models of a market or a hierarchy.

In other words, the settings Ostrom studied do not fit the two textbook models that dominate much of social science.

To make sense of this variety, Ostrom describes the work of colleagues associated with the Workshop in Political Theory and Policy Analysis. They developed the IAD framework, a nested set of building blocks that social scientists can use to understand human interactions and outcomes across diverse settings. The framework draws on earlier ideas. It builds on Commons's concept of transactions, Popper's logic of the situation, Allport's collective structures, Goffman's frames, and Schank and Abelson's scripts. It also draws inspiration from the work of Koestler and Simon, who both challenged the assumption

The author's approach rests on the view that human behavior and outcomes are "entirely based on a small set of irreducible building blocks." The text then sorts out the vocabulary of the field. Frameworks, theories and models are often used interchangeably, but the author nests them, moving from the most general set of assumptions down to the most precise.

The framework sits at the top of that nest. It is meant to hold the most general set of variables an institutional analyst might want to use, across settings as different as markets, private firms, families, community organizations, legislatures and government agencies. The author says it "provides a metatheoretical language to enable scholars to discuss any particular theory or to compare theories." In other words, it is a shared vocabulary above any single theory, so that scholars starting from different places can still talk to one another and compare their work.

Below the framework comes the specific theory. An analyst uses a theory to decide which parts of the framework matter for explaining a given set of outcomes, and how those parts relate. The text gives micro-level theories that fit the framework: game theory, microeconomic theory, transaction cost theory, and public goods and common-pool resource theories. At the most precise level are models. In the author's words,

Models make precise assumptions about a limited number of variables in a theory.

Scholars use models to work out the formal consequences of particular assumptions about what motivates the actors and how the situation they face is structured.

The framework is built as a cluster of linked variables, and any one of them can be broken down further depending on the question being asked. Its core is the action situation, the setting in which participants interact and make choices. That situation is shaped by external variables, and the text names three broad groups. The first is biophysical conditions, the physical features of the resource or environment, which in some analyses can be simplified to one of the four types of goods the author defines earlier in the lecture. The second is the attributes of a community, such as its history of prior interactions, how alike or varied its members are in key respects, and the knowledge and social capital of those who take part or are affected. The third is rules-in-use. These are the shared understandings of who "must, must not, or may take which actions affecting others subject to sanctions." The rules can change over time, either as people carry lessons from one setting into another or when they deliberately change the rules through collective or constitutional choice.

The external variables together shape the action situation and produce patterns of interaction and outcomes. Participants in the situation evaluate those outcomes, and so may scholars studying it. The results then feed back on the external variables and on the action situation itself, so the system keeps adjusting over time.

The lecture now turns to how the working parts of an action situation relate to formal game theory. The lecture states that

The internal working parts of an action situation are overtly consistent with the variables that a theorist uses to analyze a formal game.

In practice this has let colleagues pair the framework with formal game theory models, which strip a situation down to a simplified but interesting set of variables and then derive testable conclusions from them. Examples cited include work by Acheson and Gardner, by Gardner and colleagues, and by Weissing and Ostrom, along with agent-based models, which are computer simulations in which many individual actors follow rules and interact. Ostrom is candid about the limits of this approach:

It is not feasible to develop a formal game (or even an ABM) to analyze the more complex empirical settings with many variables of relevance affecting outcomes and of importance for institutional analysis.

The alternative she offers is that

It is possible, however, to use a common set of structural elements to develop structured coding forms for data collection and analysis.

The same shared variables can also guide experiments across many situations of interest to political economists, to examine why particular behavior and outcomes occur in some situations and not in others. To specify a game and predict its outcomes, the theorist must posit seven things: the characteristics of the actors, including the model of human choice the theorist adopts; the positions they hold, such as first mover or row player; the set of actions available at specific points in a decision tree; the amount of information available at each decision point; the outcomes that actors jointly affect; the functions that map actors and actions at those decision points into intermediate or final outcomes; and the benefits and costs attached to linking the actions chosen with the outcomes obtained. A footnote credits the game theorist Reinhard Selten with many hours of productive discussion in the early 1980s, when the framework was first being developed.

The lecture describes the working parts of what it calls an action situation, the setting in which people interact and their choices produce outcomes. These internal parts are the pieces the framework examines one by one. Ostrom credits a common framework used across a wide diversity of studies with a greater cumulation of understanding: findings from many studies can be built up together rather than remaining separate. The framework, called the IAD framework, which stands for Institutional Analysis and Development, places the situation under study inside a wider set of external variables. In the lecture's words,

The IAD framework overtly embeds a particular situation of interest in a broader setting of external variables, some of which can be self-consciously revised over time.

In other words, the surrounding conditions are not treated as fixed.

The next section asks, in its heading,

Are Rational Individuals Helplessly Trapped in Social Dilemmas?

A social dilemma is a situation in which individually sensible choices leave a group worse off, such as the overharvesting of a shared resource like a fishery or grazing land, or the underprovision of local public goods. Ostrom argues that the classic assumptions about rational individuals "hide the potentially productive efforts of individuals and groups to organize and solve social dilemmas." Classic models have often treated people caught in a Prisoner's Dilemma, a standard game in which each player is tempted to act against the common interest, as always trapped, with no capacity to change the structure they face. The lecture calls this "a retrogressive step in the theories used to analyze the human condition."

Ostrom's position is that whether people can transform the external variables shaping their situation differs widely from case to case. In the lecture's words,

It is an empirical condition that varies from situation to situation rather than a logical universality.

To make the contrast concrete, she notes that

Public investigators purposely keep prisoners separated so they cannot communicate. The users of a common-pool resource are not so limited.

When analysts present the people they model as trapped inside perverse situations, the lecture says, they tend to assume that outside scholars and public officials can analyze those situations from beyond them.

The conventional approach, as the lecture describes it, works in a fixed order. Outside observers study the situation, work out why it produces outcomes nobody wants, and propose what changes in the rules people actually follow would let them do better. Then external officials are expected to impose an optimal set of rules on the people involved. The assumption is that the momentum for change has to come from outside the situation, rather than from the self-reflection and creativity of those inside it to restructure their own patterns of interaction.

To show what this assumes, the lecture quotes the economist Sugden. Most modern economic theory, Sugden writes, describes a world presided over by a government, and sees that world through the government's eyes. In his words,

Private individuals, in contrast, are credited with little or no ability to solve collective problems among themselves.

The lecture uses this to argue that the picture leaves out what ordinary people can do for themselves, and so gives a distorted view of important economic and political issues.

The lecture then turns to Garrett Hardin's 1968 portrayal of a common-pool resource, meaning a shared resource such as a pasture open to all, where it is hard to keep people out and one person's use reduces what is left for others. In Hardin's account, users are trapped in an inexorable tragedy of overuse and destruction. The lecture says this view was widely accepted because it matched a familiar prediction from game theory: in a Prisoner's Dilemma, a standard game in which each player does better by acting selfishly even though both would gain from cooperating, or in other social dilemma games, people will not cooperate. The portrayal captured the attention of scholars and policymakers across the world. Many presumed that all common-pool resources were owned by no one. It followed that government officials had to impose new policies from outside to prevent destruction by users who, in the lecture's phrase, could "not do anything other than destroy the resources on which their own future" depended.

The lecture then makes a distinction between the dramatic incidents that drew public attention and a quieter body of work. Studies by anthropologists, economic historians, engineers, historians, philosophers, and political scientists had examined local governance of small to medium-scale common-pool resources over long periods of time. Many theorists and public officials did not notice them. The lecture explains why the findings did not add up to a body of knowledge:

Cumulation of the knowledge contained in these studies did not occur, due to the fact that the studies were written by scholars in diverse disciplines focusing on different types of resources located in many countries.

The lecture describes what changed this. The National Research Council, abbreviated NRC, set up a committee in the mid-1980s to assess diverse institutional arrangements for effective conservation and utilization of jointly managed resources. The committee brought scholars from multiple disciplines together, and the account goes on to describe how that work was carried out.

The work that followed the panel's call for case studies went further than counting examples. The National Research Council report opened the way to a range of studies using many methods, and the Workshop ran a longer research program that coded and analyzed common-pool resource case studies written by other scholars. In the lecture's words,

Finding multiple cases where resource users were successful in organizing themselves challenged the presumption that it was impossible for resource users to solve their own problems of overuse.

The next aim was to learn what those successful cases had in common, not just that they existed.

The Workshop's framework for complex systems, the IAD framework, became the basis for this stage. The researchers had already used it to study complex urban systems and to build a common language for linking the parts of such systems, so they used it to organize the work. It served as the foundation for a coding manual that recorded the same set of variables for each case. Ostrom describes this as an immense effort, with more than two years spent on the final manual. The central difficulty was that case study authors from different disciplines used different vocabularies. In her words,

A key problem was the minimal overlap of variables identified by case study authors from diverse disciplines.

The team read and screened over 500 case studies to find a small set that recorded information about the actors, their strategies, the condition of the resource, and the rules-in-use. The lecture notes that other meta-analysts have faced the same screening burden. One review screened over 100 articles to analyze 31 forest management cases, and another screened nearly 1,200 studies to reach a set of 268 cases of tropical forest cover change.

The screening produced two sets of comparable data. The first covered 44 subgroups of fishers harvesting from inshore fisheries. None of these fishery groups were government-managed. Eleven of them, or 25 percent, were not organized in any way. The other 33 had a range of informal rules that defined who could fish in a particular location and how harvesting was restricted.

The second set covered 47 irrigation systems, managed either by farmers or by a government. Of the 12 government-managed systems, only 40 percent, seven of them, had high performance. Of the 25 farmer-managed systems, more than 70 percent, 18 of them, had high performance. The lecture adds that rule conformance was a key variable affecting the adequacy of water over time. The breakdown given here accounts for 37 of the 47 systems, and this part of the text does not say how the remaining ten were classified.

The lecture adds a qualification to that picture. Alongside the significant levels of cooperation it found, it also found some support for earlier theoretical predictions of no cooperation in particular settings. Where people do not know one another, cannot communicate effectively, and so cannot build agreements, norms, and sanctions, the aggregate predictions of models built on rational individuals in a noncooperative game receive substantial support. As the lecture puts it,

These are sparse environments and full rationality appears to be a reasonable assumption in them.

On the other hand, the lecture reports that the capacity to overcome dilemmas and create effective governance "occurred far more frequently than expected." Whether it happened depended on the structure of the resource itself, and on whether the rules users developed were linked effectively to that structure. In all the self-organized systems the lecture studied, users had created boundary rules that determined who could use the resource, choice rules for allocating the flow of resource units, and active forms of monitoring and local sanctioning of rule breakers.

Yet one strategy was absent. The lecture states that it did not find a single case where harvesters used what it calls the "grim trigger" strategy, a form of punishment that many theoretical arguments had proposed as a way for individuals to solve repeated dilemmas. In a grim trigger strategy, a player cooperates until others defect, and then refuses to cooperate ever again.

The lecture then turns to property rights. Resource economists, it notes, have used the term "common property resource" for fisheries and water resources. The lecture argues that combining "property" with "resource" introduced considerable confusion between the nature of a good and the absence or presence of a property regime. A common-pool resource, it says, can be owned and managed as government property, private property, community property, or owned by no one.

A further reason for overlooking the property systems local users had developed, the lecture suggests, was that many scholars presumed that unless users possessed alienation rights, meaning the right to sell their property, they did not have any property rights at all.

The lecture credits a conceptual move to Schlager and Ostrom, who drew on the earlier work of John R. Commons to treat property rights as bundles of rights rather than a single right. A meta-analysis of existing field cases helped identify five rights that users of a common-pool resource might hold cumulatively. The first three are access, the right to enter a specified property; withdrawal, the right to harvest specific products from a resource; and management, which covers transforming the resource and regulating it, with the list continuing beyond that point. A footnote explains that access rights have puzzled some scholars. Its example is buying a permit to enter a public park, which gives the permit holder the right to enter and enjoy hiking and other nonharvesting activities for a defined period of time.

Ostrom's lecture moves from property rights to the rules that shape the settings where people use shared resources. She first notes that thinking of property rights as bundles has become common.

Conceiving of property-rights bundles is now widely accepted by scholars who have studied diverse property-rights systems around the world.

In this bundle, exclusion is the right to decide "who will have access, withdrawal, or management rights," and alienation is the right to "lease or sell any of the other four rights."

The lecture then turns to people who hold specific property rights to a resource. They also face more basic rules that shape the action situations they are in. An action situation is the setting in which participants choose actions, and those choices produce outcomes. The lecture says that "actors who have specific property rights to a resource also face more fundamental rules that affect the structure of the action situations they are in."

The authors' meta-analysis of case studies found, in their words, "an incredible array of specific rules used in different settings." Examples include who could withdraw how many units of a resource, and where and when; what information every user had to provide; and which costs and benefits came with which actions. To make this variety comparable, the authors turned to the Institutional Analysis and Development framework, usually called the IAD framework, a common tool for studying institutions. They had already identified seven working parts of an action situation, so they reasoned that seven broad types of rules might act as external influences on those parts.

The first type is boundary rules, which specify how actors are chosen to enter or leave positions. Position rules set out the positions and how many actors hold each one. Choice rules assign which actions are open to an actor in a given position. Information rules specify the channels of communication among actors and what information "must, may, or must not be shared." Scope rules specify "the outcomes that could be affected." Aggregation rules, such as majority or unanimity rules, specify how the decisions of actors at a decision point are combined into intermediate or final outcomes. The last type is payoff rules, which specify how benefits and costs are distributed to actors in positions. The lecture attributes this typology to earlier work.

A useful way to think about institutional rules, Ostrom explains, is to ask which part of an action situation a rule affects. In her framing, rules work as outside variables that act directly on the elements of that situation, and a figure in the lecture shows each rule type connecting to the particular element it changes. She then turns to the size of the task. The team's analysis identified seven broad types of rules rather than one or two, and she acknowledges that this has been "upsetting to scholars who wanted to rely on simple models of interactions among humans." The count of types was only the start, because the team also found many variants within each type. The case study authors, she reports, described 27 boundary rules in use in at least one common-pool resource setting. Some of these specified diverse forms of residence, organizational membership, or personal attributes that are either ascribed or acquired. The choice rules were more numerous still, with 112 different versions. Most were built from two parts: an allocation formula that sets where, when, or how resource units may be harvested, and a specific basis for putting that formula into effect, such as the amount of land a person holds, historical patterns of use, or assignment by lottery.

Ostrom then addresses the question that had driven the research: which rules are associated with resource institutions that survive over the long run. Her plan was to code the cases of successful and failed systems and then run careful statistical analysis to identify which specific rules went with success. She had not yet fully absorbed how many rules and how much diversity the team had recorded. During a sabbatical in 1988 at a research group led by Reinhard Selten at the Center for Interdisciplinary Research at Bielefeld University, she searched for rules that would work across ecological, social, and economic environments. The rules linked to success or failure varied extensively from site to site. Her conclusion is candid:

Finally, I had to give up the idea that specific rules might be associated with successful cases.

Moving up a level of generality, Ostrom looks at the broader institutional patterns that separated long-lasting systems from failed ones. She uses the term "design principle" for these patterns. She is careful to say that the fishers, irrigators, pastoralists and others did not consciously hold these principles in mind when they built systems that survived. Her aim was to identify a set of core underlying lessons that set the long-sustained regimes apart from the cases of failure.

She notes that the principles are described at length in her earlier work, so she offers a brief updated list, as developed by Cox, Arnold, and Villamayor-Tomás in 2009. The first two principles concern boundaries. Users need

Clear and locally understood boundaries between legitimate users and nonusers.

The resource itself also needs clear boundaries that separate it from the larger social-ecological system it sits within, meaning the linked web of people and environment around it.

The next principles concern fit and fairness. Rules for taking from the resource, which the author calls appropriation, and rules for maintaining it, called provision, should be congruent with local social and environmental conditions. Taking and maintaining should also line up: the distribution of costs should be proportional to the distribution of benefits. Decision-making is next. Most people affected by a resource regime should be authorized to take part in making and changing its rules.

Monitoring comes next, in two forms. Users, or people accountable to them, should watch how much is being taken and contributed. Those same people should also watch the condition of the resource itself. When rules are broken, the penalties should be measured. In the list's words,

Sanctions for rule violations start very low but become stronger if a user repeatedly violates a rule.

Disputes also need a route to resolution. The list calls for "Rapid, low-cost, local arenas" for settling conflicts among users or with officials.

Two final principles concern the relationship with higher authority. The government should recognize that local users have the right to make their own rules. And where a resource is closely tied to a larger system, the governance should be arranged in layers: "governance activities are organized in multiple nested layers."

The author argues that these principles appear to capture the core factors that affect the chance that an institution built by resource users will survive over the long term. She points to Cox and colleagues, who analyzed more than 100 studies by scholars who tested how well the principles explained success or failure across many different common-pool resources, which are resources many people share and that are hard to fence off.

In a footnote, she acknowledges that the term "design principle" has confused many readers. She suggests that she might have used the phrase "best practices" instead, to describe the rules and structure of robust institutions.

Ostrom then turns to how well the design principles have held up against other research. She writes that studies "confirm that robust resource systems are characterized by most of the design principles and that failures are not." Some studies found the principles inadequate, she notes, but their authors read them very rigidly and felt that successful systems showed more flexibility than the principles allow.

She also concedes a weakness in the original wording. In three instances the principles were too general and did not distinguish between ecological and social conditions. For that reason she has adopted the improvements to principles 1, 2, and 4 that Cox and coauthors suggested.

The lecture then reaches its fifth point, conducting experiments to study common-pool resource problems. A common-pool resource is a shared stock, such as a fishery, a forest, or an irrigation system, from which many people draw and from which it is hard to keep anyone out. Ostrom notes that a large number of field cases in which users overcame social dilemmas and kept a resource in use over the long term "successfully challenged the presumption that this was impossible." The difficulty is that many variables act at once in the field, so the effect of any single one is hard to isolate.

Two tools were used to get at this. The first was game-theoretical modelling, a way of working out the outcomes a set of variables would be predicted to produce. Ostrom points to earlier modelling work by Weissing and Ostrom and by Ostrom and Gardner. The second was to test precise combinations of variables in a controlled setting, which she says was just as important.

The experimental work took place in university laboratories. Roy Gardner and James Walker joined in an extended effort to build and test well-specified game-theoretical models consistent with the Institutional Analysis and Development framework, the analytical structure behind Ostrom's research. The first experiments used a static baseline, "as simple as could be specified without losing crucial aspects of the appropriation problems facing harvesters in the field."

Each of eight subjects received an endowment of tokens and could split it between two markets. Market 1 paid a fixed return. Market 2 was the common-pool resource, and its return depended on what all the subjects put into it together. Subjects saw only aggregated information, so no one could see what any individual had chosen. The shared return followed a quadratic production function, a curve that rises with total investment and then falls off, based on Gordon's classic model of a fishery from 1954.

The baseline was a commons dilemma. Game theory predicted substantial overuse of the resource, while a much better result for everyone was available if the subjects could bring their joint investment down.

Ostrom then describes the baseline experiments, in which subjects played the common-pool resource game over repeated rounds. Game theory predicted that subjects would invest according to the Nash equilibrium, the outcome in which no one can do better by changing only their own choice. In this game that meant 8 tokens each, 64 in total. Subjects could earn considerably more if they reduced their allocation to a total of 36 tokens. Instead, subjects in baseline experiments with multiple decision rounds substantially overinvested, putting in even more tokens than predicted, so the joint outcome was worse than the Nash equilibrium. A footnote adds useful context: in simple repeated public goods experiments, subjects tended to start above the predicted level and slowly approached the equilibrium from a higher level, while in common-pool resource games they started well below it and approached it from below.

Next Ostrom turns to face-to-face communication, building on earlier public goods research. After ten rounds without talking, subjects were allowed to meet as a group before returning to their terminals to make private decisions. The predicted outcome was the same as in the baseline, because a promise to cooperate is what economists call cheap talk. Nothing binds the promise, and no external "third party" ensured that it was kept. Subjects used the discussion to work out strategies and to agree, where possible, on how much each person should invest. After each round they saw the group's aggregate investment but not what each individual had done, so they could tell whether the total exceeded what had been agreed. In many rounds people kept their promises, and in others some defections occurred. Subjects also used the aggregate figures to scold unknown fellow subjects when the total was higher than agreed. Ostrom concludes that the opportunity for repeated face-to-face communication was "extremely successful in increasing joint returns." She notes that this matches a large number of studies showing that face-to-face communication helps people solve many kinds of social dilemmas.

The last part of this section concerns sanctioning. In many field settings, resource users have devised formal or informal ways of punishing one another when rules are broken. Ostrom points out that this behavior is not consistent with the theory of norm-free, complete rationality. So the researchers asked whether subjects in a controlled setting would use their own money to punish other subjects. After ten rounds of the baseline game, subjects were told they could pay a fee in order to impose a fine on another subject. The text sets up this comparison and the sanctioning results that follow it.

Subjects who were allowed to sanction one another did raise their gross benefits, but Ostrom reports that they substantially reduced their net returns, due to "the overuse of costly sanctions." Sanctions were mostly aimed at those who had defected, though a few appeared to target low contributors as a form of revenge by players who had already fined themselves. In a further design, participants could communicate and then decide whether to adopt a sanctioning system of their own. Those who chose to adopt one reached "the highest returns achieved in any of the common-pool resource laboratory experiments," about 90 percent of optimal once the fines for the few defections were subtracted.

The lecture draws a clear lesson from these laboratory results. Noncooperative game theory's predictions are roughly supported only when participants do not know the reputations of the others in a common-pool resource dilemma and cannot communicate with them. When subjects talk face to face, they frequently agree on joint strategies and keep to those agreements, which substantially raises their net returns. Communication that lets people design their own sanctioning system brings them close to optimal returns.

The work then turns to field experiments carried out by colleagues in Colombia, asking whether experienced villagers who depend on a resource make decisions about the "time spent in a forest" in a way that is mathematically consistent with the laboratory designs. Cardenas ran field experiments in rural schoolhouses with more than 200 users of local forests, reframing the decisions as "harvesting trees." With and without face-to-face communication, the outcomes were broadly consistent with those found among university students.

In another design, Cardenas, Stranlund, and Willis ran ten rounds of baseline experiments with resource users from five villages, then allowed them to communicate face to face for the following rounds. In five other villages, participants were told after the baseline rounds that a new regulation would require them to spend no more than the optimal time in the forest each round. Inspections occurred with a probability of 1/16 per round, which Ostrom describes as "a low but realistic probability for monitoring rule conformance in rural areas in developing countries." Anyone found over the limit had a penalty subtracted.

Ostrom also notes that related work extends beyond this setting. Henrich and colleagues conducted field experiments in several countries to test whether a much broader set of participants would also use punishment in public goods experiments, and earlier field experiments on social dilemmas were run in fifteen small communities. Similar findings come from public goods experiments, where punishers typically target low contributors.

The lecture then turns to a variant of the experiments in which a penalty applied to an individual's payoff, but the penalty was not revealed to the others in the group. Subjects in that setting increased their withdrawal levels compared with outcomes when face-to-face communication was allowed and no rule was imposed. Ostrom notes that other scholars have also found that externally imposed regulation, which would in theory produce higher joint returns, "crowded out" voluntary behavior to cooperate. She then describes public goods experiments by Fehr and Leibbrandt with fishers who harvest from an "open-access" inland lake in northeastern Brazil. The researchers found that a high share, 87 percent, of the fishers contributed in the first period, and that contributions leveled off in the periods that followed. They also examined the mesh size of each fisher's nets and found that those who contributed more in the experiment used nets with bigger mesh. Larger mesh lets young fish escape, grow larger and reproduce at a higher level than fish caught while still small, so cooperation in the field game matched cooperation seen in a real common-pool resource dilemma. Ostrom quotes the authors' conclusion that the

fact that our laboratory measure for other-regarding preferences predicts field behavior increases our confidence about the behavioral relevance of other-regarding preferences gained from laboratory experiments.

Summing up the experimental work on common-pool resources, which are shared resources such as fisheries and irrigation water from which it is hard to exclude users, she writes that many predictions of the conventional theory of collective action do not hold. More cooperation occurs than predicted, "cheap talk" increases cooperation, and subjects invest in sanctioning free-riders. The experiments also establish that motivational heterogeneity exists in harvesting or contribution decisions as well as in decisions on sanctioning. The lecture then moves from the laboratory to the field. Having conducted meta-analyses of case studies and experiments, Ostrom says her team needed field studies that could use the institutional analysis and development, or IAD, framework to ask consistent questions about key variables across sites. The first such effort grew out of a 1988 visit to Nepal, which turned up many written studies of farmer-built and maintained irrigation systems, as well as some government-constructed and managed ones. With Ganesh Shivakoti and Paul Benjamin, she revised the common-pool resource coding manual for the Nepal Irrigation and Institutions project. When the team coded the existing cases, they again found numerous "missing variables" that the original authors had not discussed, so colleagues made several trips to Nepal to revisit the described systems and fill in the missing data.

Ostrom describes how the irrigation work was checked and extended. Her team went back into the field to verify the data in the original study, and while there it added new cases to the data set. Lam then analyzed this larger collection and developed three performance measures that could be applied to every system: the physical condition of the irrigation works, the quantity of water available to farmers at the tail end of a system at different seasons of the year, and the agricultural productivity of each system. Controlling for differences in the environment, Lam found that systems governed by the farmers themselves performed significantly better on all three measures.

Ostrom explains the likely reasons. On farmer-governed systems, farmers communicate with one another at annual meetings and informally on a regular basis, develop their own agreements, appoint monitors, and sanction those who do not follow their rules. As a result, the lecture says, farmer-managed systems are likely to grow more rice, distribute water more equitably, and keep their systems in better repair than government systems. The result is not uniform. In her words, farmer systems "do vary in performance," but few perform as poorly as government systems once other relevant variables are held constant.

The pattern held as the work grew. Colleagues visited and coded still more irrigation systems in Nepal, and an expanded database of 229 systems again confirmed the earlier finding that farmer-managed systems perform better. Ostrom writes that

our findings are not unique to Nepal.

She points to other scholars who have documented effective farmer-designed and farmer-operated systems in Japan, India, and Sri Lanka.

The lecture then turns to forests. In 1992, Marilyn Hoskins, who headed the Forest, Trees and People Program at the Food and Agriculture Organization of the United Nations, asked colleagues at the Workshop to draw on the irrigation experience and develop methods for assessing the impact of different forest governance arrangements across several countries. After two years of intensive development and review by ecologists and social scientists around the world, the team produced ten research protocols. These were designed to gather reliable information about forest users, about how forests are governed, and about the ecological conditions of the sampled forests.

Those protocols became the basis of a long-term collaborative network, the International Forestry Resources and Institutions research program, known as IFRI. It has centers in Bolivia, Colombia, Guatemala, India, Kenya, Mexico, Nepal, Tanzania, Thailand, Uganda, and the United States, with new centers being established in Ethiopia and China. Ostrom calls IFRI unique among efforts to study forests: "it is the only interdisciplinary long-term monitoring and research program studying forests in multiple countries owned by governments, private organizations, and communities." The discussion of forests as a common-pool resource, a resource many people share and can be hard to protect from overuse, is only beginning here.

Forests matter for more than timber. They store carbon, hold much of the world's biodiversity, and support rural livelihoods in developing countries. A common policy answer for protecting them is government-owned protected areas, which the text calls a "favorite" policy recommendation. The authors ask whether that approach actually works, and they review a series of studies that test it.

One study, by Hayes in 2006, used data from the International Forest Resources and Institutions program, known as IFRI. A forester or ecologist who had supervised the measurement of trees, shrubs and ground cover in a random sample of forest plots rated each forest's density on a five-point scale. The sample included 163 forests. Seventy-six were government-owned and legally designated as protected, and 87 were public, private or communally owned lands used for a range of purposes. The paper reports:

No statistical difference existed between the forest density in officially designated protected areas versus other forested areas.

Government ownership, in other words, did not show up as denser forest.

The rating method is explained in a note. Comparing raw tree measurements across ecological zones would mislead, because average tree diameter is strongly shaped by rainfall, soils, elevation and other factors that vary widely between zones. So the forester is asked for a judgment instead, on a scale running from "very sparse to very abundant."

A study by Gibson, Williams and Ostrom in 2005 looked at 178 forest user groups and their monitoring behavior, meaning how closely they watched over the forest. It found a strong correlation between the level of monitoring and the forester's assessment of forest density. That link held even after the authors controlled for, or statistically held constant, whether users were formally organized, whether they depended heavily on the forest, and how much social capital the group had.

The text then turns to Chhatre and Agrawal, who in 2008 examined changes in the condition of 152 forests under diverse governance arrangements. They looked at forest size, collective action to improve the forest, the size of the user group, and how much local people depended on the forest. The paper quotes their finding that

forests with a higher probability of regeneration are likely to be small to medium in size with low levels of subsistence dependence, low commercial value, high levels of local enforcement, and strong collective action for improving the quality of the forest.

A second analysis by Chhatre and Agrawal, in 2009, focused on the tradeoffs and synergies between carbon storage in forests and their contributions to livelihoods. They found that larger forests were more effective at improving both carbon and livelihood outcomes, particularly where local communities also had high levels of rule-making autonomy.

Recent work by Coleman in 2009, and by Coleman and Steed in 2009, points the same way. Those studies identify local users' investment in monitoring as a major variable affecting forest conditions. They also find that when local users are given harvesting rights, they are more likely to monitor illegal uses themselves. The text notes that other focused studies stress the same relationship between local monitoring and better forest conditions.

The lecture points out that the legal designation of a forest as a protected area "is not by itself related to forest density." A forest can be protected on paper and still be thin and degraded. Detailed field studies of monitoring and enforcement, as they actually happen on the ground, show how hard it is to achieve high levels of forest regrowth without the active involvement of local forest users.

The research described here looked at forests under government, private, and communal property regimes. Those forests sometimes met social goals such as biodiversity protection, carbon storage, or improved livelihoods, and at other times they did not. The type of ownership alone does not settle the outcome. The lecture also warns about decentralization, the handing of authority down to local levels. Its own wording is that "when governments adopt top-down decentralization policies leaving local officials and users in the dark, stable forests may become subject to deforestation." In other words, a policy that shifts control without informing the people closest to the forest can undermine forests that were doing well.

From this, the lecture draws its central conclusion about forest governance. In its words, "it is not the general type of forest governance that is crucial in explaining forest conditions; rather, it is how a particular governance arrangement fits the local ecology, how specific rules are developed and adapted over time, and whether users consider the system to be legitimate and equitable." Three things matter: fit with the local ecology, rules that change over time, and whether the people involved see the system as fair. The lecture points readers to a fuller overview of that research program for the details.

The lecture then turns to where theory now stands. After half a century of its own empirical work and that of many other scholars, it asks directly: "where are we now?" and

What have we learned?

Its answer is that the older picture of "rational, but helpless, individuals who are trapped in social dilemmas" is not supported by a large number of studies using diverse methods. A social dilemma is a situation where what is good for each person separately leads to a poor result for the group.

The lecture is careful not to swing too far the other way. It says that "we cannot be overly optimistic and presume that dilemmas will always be solved by those involved." Many groups have struggled and failed. It also argues that simple policy fixes can fail too. Handing a resource to a government, privatizing it, or, more recently, decentralizing it may not work on its own.

The task that follows, in the lecture's terms, is to build theory that can explain and predict "when those involved in a common-pool resource dilemma will be able to self-organize." A common-pool resource is a shared resource, such as a fishery, a forest, or a groundwater basin, from which it is hard to keep people out and easy to take too much. The lecture stresses that this requires "a better theoretical understanding of human behavior as well as of the impact of the diverse contexts that humans face."

The lecture then turns to developing a more general theory of the individual. It recalls its earlier discussion of three levels of generality at which social phenomena are studied. Frameworks such as the Institutional Analysis and Development framework, usually called the IAD, are, in the lecture's words, "meta-theoretical devices that help provide a general language" for organizing the many studies of common-pool resources. They do not themselves explain outcomes. They give researchers a shared vocabulary so that studies in different places can be compared.

The lecture separates two words that are often used as if they meant the same thing: theory and model. A theory, as the lecture describes it, is an effort to build understanding by making core assumptions about specific working parts of frequently encountered phenomena and predicting general outcomes. A model is narrower. Models are "very specific working examples of a theory," and they are "frequently confused with being theories themselves."

The lecture's main example is rational choice. It notes that Alchian pointed out long ago that what is called "rational choice theory" is "not a broad theory of human behavior" but rather "a useful model to predict behavior in a particular situation," namely "a highly competitive market for private goods." Predictions from that model, the lecture says, are empirically supported in open markets for private goods and in other competitive environments, with experimental work by Holt, by Smith and Walker, and by Satz and Ferejohn cited in support. So the model is worth keeping for competitive settings where outcomes are excludable, meaning people can be kept from them, and divisible, meaning they can be split into portions.

The lecture then says that no single theory of human behavior has yet been formulated and tested successfully across a variety of settings. Scholars are instead positing and testing assumptions that are likely to sit at the core of future developments. The lecture lists three. The first is that boundedly rational individuals, meaning people whose information and reasoning are limited, can learn fuller and more reliable information in repeated situations when reliable feedback is present. The second is the use of heuristics, or rules of thumb, in everyday decisions. The third is that people have preferences about benefits for themselves, and also norms and preferences about benefits for others.

On information, the lecture is blunt. The assumption that individuals have complete information about all their available actions, the strategies others will likely adopt, and the probabilities of each outcome, "must be rejected in any but the very simplest of repeated settings." When people interact over time, it is reasonable to assume they learn more accurate information about their options and about what others are likely to do. Still, the lecture warns that some highly complex common-pool resource settings, where many users draw on a shared resource, "approach mathematical chaos," a phrase it attributes to Wilson and colleagues in 1994. In those settings, users cannot gain complete information about every likely combination of future events.

Heuristics are rules of thumb that people learn over time because they work relatively well in a particular setting. The lecture uses fishers as its example. Through heuristics, they end up "fishing for knowledge," a phrase it credits to Wilson in 1990, and learn to recognize diverse clues about environmental processes that they must take into account. When people interact repeatedly, heuristics can approach "best-response" strategies, the choices that do best given what others are doing, and come close to local optima, meaning the best results reachable from where they stand. The lecture adds a caution: in eras of rapid change or sudden shocks, heuristics may not enable people to achieve high payoffs.

Finally, the lecture turns to norms, which it describes as "internal valuations that are negative or positive related to specific actions such as lying or being brave in particular situations." In other words, a norm is a personal sense of whether an action is good or bad, held inside the person rather than imposed from outside.

Internal norms, the authors suggest, may show up as the weight a person gives to actions and outcomes in a particular setting. Among these are norms about valuing outcomes that others achieve. Fehr and Schmidt propose that people dislike unequal outcomes and so hold an internal norm of "inequity aversion." Axelrod is cited for the idea of meta norms, held by people who care whether others follow the norms a group has developed, and who thereby make those norms more likely to be followed. Leibbrandt, Gneezy, and List are cited for the finding that people who regularly work in teams are more likely to adopt norms and to trust each other than people working alone. Frohlich and Oppenheimer posit that many people adopt norms of fairness and justice. The authors also note that people do not all share the same norms or the same view of a situation, and that they can differ widely on whether a way of sharing costs is fair. But simply assuming that people hold norms is not enough, the authors argue, to predict behavior in a social dilemma, especially in very large groups with no arrangements for communication. Even with strong preferences to follow norms, the authors quote de Oliveira, Croson, and Eckel:

observed behavior may vary by context because the perception of the 'right thing' would change.

Context shapes how people learn about their situation and about the others they deal with. Individual differences matter, but the context of interactions also shapes behavior over time. The authors draw a parallel from biology, where an organism's appearance and behavior are shaped by the environment in which it develops. Some plants grow large, thin leaves in low light to catch more of it, and narrow, thicker leaves in high light to conserve water. Certain insects develop wings only when they live in crowded conditions, where they are likely to run short of food. Pfennig and Ledón-Rettig are quoted on this point:

Such environmentally contingent development is so commonplace that it can be regarded as a universal property of living things.

The authors draw the lesson for social science directly: "Social scientists also need to recognize that individual behavior is strongly affected by the context in which interactions take place rather than being simply a result of individual differences." The argument then turns to trust. Arrow, long ago, pointed to trust among participants as the most efficient mechanism for improving transactional outcomes. Yet collective-action theory, the authors observe, has paid more attention to payoff functions, the formulas that set what each person stands to gain, than to how individuals build trust. The sentence on trust runs on past where this text ends.

Trust matters here too. A social dilemma is a situation in which each person is tempted to hold back, even though everyone would do better if all cooperated. The lecture notes that empirical studies nonetheless confirm the important role of trust in overcoming such dilemmas. It then explains how its updated assumptions about learning and norm-adopting individuals can account for how people come to trust one another. On this view, trust grows when feedback mechanisms reinforce positive or negative learning, which leads to more cooperation and higher benefits for the group.

The argument goes beyond the idea that people simply adopt norms. In the lecture's words,

It is not only that individuals adopt norms but also that the structure of the situation generates sufficient information about the likely behavior of others to be trustworthy reciprocators who will bear their share of the costs of overcoming a dilemma.

In other words, the setup of a situation can tell people whom they can count on to pay their share. From this, the lecture concludes that "in some contexts, one can move beyond the presumption that rational individuals are helpless in overcoming social dilemma situations."

A diagram in the lecture makes the same point, and its caption states that

Microsituational and broader contexts of social dilemmas affect levels of trust and cooperation.

The lecture then turns to what it calls the microsituational level of analysis. It is blunt that saying context matters is not enough: "Asserting that context makes a difference in building or destroying trust and reciprocity is not a sufficient theoretical answer to how and why individuals sometimes solve and sometimes fail to solve dilemmas." To answer that question, it separates two kinds of context. The first is a microcontext, meaning the specific features of the action situation in which people interact directly. The second is the broader context of the social-ecological system, meaning the natural resources and wider community in which groups make their decisions.

The lecture points to one advantage of laboratory and field experiments here. The researcher designs the microsetting in which the experiment takes place, so the features of the situation can be varied on purpose. It reports that a growing body of such results, summarized in a review it draws on, establishes that particular attributes of a microsituation affect how much cooperation participants achieve. The lecture says this holds for both public goods dilemmas, where a benefit is shared by all, and common-pool resource dilemmas, where a shared resource such as a fishery or grazing land can be depleted by overuse.

Ostrom then lists six features of a situation that make people more likely to cooperate in a microsetting. These are the kinds of variables researchers have been manipulating in laboratory experiments.

The first is communication. When face-to-face talk is possible, participants read facial expressions, physical actions, and the way others phrase their words, and they use these cues to judge whether others can be trusted. The second is reputation. Knowing the past history of people who may be strangers before the interaction begins makes cooperation more likely. The third is a high marginal per capita return, or MPCR, which is the payoff a person gets back from each unit they put into a shared pot. When that return is high, each person can see that their own contribution makes a bigger difference, and that others are more likely to notice this too.

The fourth is the ability to enter and exit. If people can leave a situation at low cost, they are not stuck being what the lecture calls a sucker, someone who keeps cooperating while others take advantage. Others can also see that cooperators may leave and join other situations if their cooperation is not returned. The fifth is a longer time horizon. Participants who expect to interact over a long period can anticipate that more can be earned through cooperation than over a short one. The sixth is sanctioning that the participants themselves agree on. The lecture notes that outside sanctions or imposed systems may reduce cooperation. But when participants agree to a sanctioning system, "they frequently do not need to use sanctions at a high volume and net benefits can be improved substantially."

Ostrom notes that other microsituational variables are being tested in experiments around the world. She sums up the central core of these findings in one sentence: "when individuals face a social dilemma in a microsetting, they are more likely to cooperate when situational variables increase the likelihood of gaining trust that others will reciprocate." A social dilemma is a situation in which each person's sensible choice, taken alone, leaves the group worse off.

The lecture then turns from the laboratory to the field. People managing common-pool resources in the field, meaning shared resources that many people use and that are hard to fence off, are also shaped by a broader set of contextual variables. These concern the attributes of the social-ecological system, or SES, in which they interact. Ostrom reports that a group of scholars in Europe and the United States is working to extend a framework that links the Institutional Analysis and Development framework, the IAD, with its interactions and outcomes at the micro level, to this wider set of variables observed in the field.

The lecture's footnote adds detail on who is doing this work. Scholars at several European institutions, including the Stockholm Environment Institute, the International Institute for Applied Systems Analysis, Delft University of Technology, the University of Zurich, and the Potsdam Institute for Climate Impact Research, have held meetings to plan using a common framework to study a variety of resource systems. That framework was first developed by Ostrom in 2007. Scholars at the Workshop in Bloomington and at the Center for the Study of Institutional Diversity at Arizona State University are also expected to take part. The footnote identifies a core problem these scholars have raised: "the lack of cumulation across studies on diverse natural resource systems as well as humanly engineered resources."

Ostrom places the action situation inside a larger structure. She describes it as "an Action Situation generating Interactions and Outcomes that are affected by and affect a Resource System, Resource Units, Governance System, and Users," and those users in turn affect and are affected by social, economic, and political settings and related ecosystems. In her terms, a resource system is the stock being drawn on, such as a fishery or a groundwater basin, and resource units are what gets taken from it, such as fish or litres of water.

The lecture includes an overview of the highest tier of variables that appear in all field settings. Ostrom notes that this tier can be unpacked several times when a researcher studies a specific question, but that "there is not enough time or space to undertake a thorough unpacking" in a lecture of this kind.

She then compares two kinds of research. "Experimental researchers have reached a higher level of agreement about the impact of microsituational variables" on the incentives, levels of trust, and behavior of individuals in dilemma situations, she writes, "than exists among field researchers." Microsituational variables are the details of the immediate decision setting, as opposed to the wider social and ecological context. Field researchers, by contrast, have less consensus on how the broader variables shape behavior.

The broader variables also rarely act alone. "Few SES variables have a fully independent impact on the action situations that participants face and their likely behavior," Ostrom writes. Which variables matter most depends on what one wants to predict. Some matter for interactions such as monitoring, conflict, lobbying, and self-organization. Others matter for longer-term outcomes such as overharvesting, the regeneration of biodiversity, and the resilience of an ecological system to disturbances caused by humans and by nature.

For the specific question of whether users of a shared resource will organize themselves, Ostrom draws on many field studies.

A set of ten variables have been identified across many field studies as impacting the likelihood of users self-organizing in order to overcome a common-pool resource dilemma.

A common-pool resource dilemma arises when many people draw from a shared resource that is hard to fence off, so each person has an incentive to take more than their share. The ten variables include three features of the resource system: its size, its productivity, and its predictability. They also include how mobile the resource units are, and whether the users have collective-choice rules they can adopt to change their own operating rules. Four attributes of the users complete the list: how many users there are, whether there is leadership or entrepreneurship among them, how much they know about the system, and how important the system is to them.

Ostrom closes this stretch by naming a central challenge. Linking the broader contextual variables with the microcontextual ones is, in her words, "one of the major tasks facing scientists who work".

Ostrom turns to what she calls complexity and reform, where her account of how people manage shared resources meets public policy. She notes that the economic and social sciences have moved well past the older picture of two optimal organizational forms, two types of goods, and one model of the individual. Extensive empirical research, she says, documents the range of settings in which people solve common-pool resource problems on their own, when those solutions last over long periods, and how larger institutional arrangements either strengthen or weaken the ability of people at smaller scales to solve problems efficiently and sustainably. Common-pool resources are shared assets such as fisheries, irrigation systems, forests and lakes, where many users draw on one stock and none can easily keep the others out. She is candid that no single well-developed theory yet explains all the diverse outcomes, whether they arise in controlled laboratory experiments or in the field. Still, she says considerable agreement exists, and there is no single normative theory of justice that can be applied unambiguously to all such settings. Two elements stand out as central to solving social dilemmas: building trust among users, and writing institutional rules that match the ecological systems being used. She highlights what she calls a surprising but repeated finding, in the lecture's words, "the surprising but repeated finding that users of resources that are in relatively good condition ... do invest in various ways of monitoring one another," which she links to the core problem of building trust. The lecture then turns to policy. "Unfortunately," she writes, policy analysts, public officials, and scholars who still apply simple mathematical models to field settings have not absorbed these lessons. Her main target is the habit of recommending one fix for an entire category of resource.

All too often a single policy prescription - such as Individual Transferable Quotas (ITQs) - is recommended for all resources of a particular type, such as all fisheries.

An ITQ is a fishing quota system that gives individual fishers shares of an allowable catch, which they can trade. The lecture grants that several ITQ systems work well, but says that adapting the broad idea to a particular place "involves multiple years of hard work by the fishers involved as well as the government officials." A footnote points to a study of twenty-five mining camps in the American Southwest that links group size, acceptance of norms of cooperation, and support for property rights. The lecture's central lesson for policy follows.

The most important lesson for public policy analysis derived from the intellectual journey I have outlined here is that humans have a more complex motivational structure and more capability to solve social dilemmas than posited in earlier rational-choice theory.

She contrasts this with the approach that has dominated for much of the past half century, which has aimed at

Designing institutions to force (or nudge) entirely self-interested individuals to achieve better outcomes.

Her alternative is stated as follows: "Extensive empirical research leads me to argue that instead, a core goal of public policy should be to facilitate the

The lecture turns from theory to a research agenda. It asks researchers to study how polycentric institutions, meaning systems with many overlapping centers of decision-making, shape the wider work of building institutions that bring out the best in people. The question it poses is how such institutions "help or hinder the innovativeness, learning, adapting, trustworthiness, levels of cooperation of participants, and the achievement of more effective, equitable, and sustainable outcomes at multiple scales." On modeling, the lecture takes a measured position. Simple mathematical models remain useful, it says, where they capture enough of the core underlying structure and incentives to predict outcomes usefully, and researchers should keep using them in those settings. When the world being explained is not well described by a simple model, the author argues that we "must continue to improve our frameworks and theories so as to be able to understand complexity and not simply reject it."

The main takeaway is that people facing social dilemmas are not necessarily helpless. Across the cases, experiments and field studies the lecture reviews, users often organize themselves, monitor one another and sanction rule breakers, and farmer-managed irrigation systems and locally monitored forests often did better than government-run or unmonitored ones. From this the lecture concludes that a core goal of public policy should be to help people build their own institutions, rather than assuming that outside officials must impose rules. The limits are stated plainly. Ostrom warns that we cannot be overly optimistic and presume that dilemmas will always be solved, that many groups have struggled and failed, and that handing a resource to government, privatizing it, or decentralizing it may not work on its own. She also concedes that the design principles she describes were too general in some places and have been revised. Open questions remain. Researchers still need a theory that explains and predicts when those involved in a common-pool resource dilemma will be able to self-organize, and field researchers agree less than experimental researchers about how broader social and ecological variables shape outcomes. The lecture closes by calling for research on how polycentric institutions help or hinder innovation, learning, trust and cooperation, and it argues that frameworks and theories must keep improving so that complexity is understood rather than rejected.

  1. Prize Lecture by Elinor Ostrom