Institutional analysis helps explain how rules, roles, authority and information shape the choices people make together. Its strongest results locate an actual decision and show why a particular rule works or fails in a given context, rather than assuming every community should follow one governance model.
This guide develops the Institutional Analysis and Development framework through the seven categories of rules, common resources and original fictional governance cases. It distinguishes documented procedures from actual practices and explains why a correct calculation may still leave a difficult public decision.
THE GOLD STANDARD · INSTITUTIONAL ANALYSIS
Identify the rule, the choice and the outcome
Foundations · Seven rule types · Shared resources · Evaluation · Chapter index
Chapter index
Institutions, actors and rule categories · Chapters 1–12
- An institution is a set of rules in action
- Start from a decision that can be observed
- A written rule may not describe everyday practice
- Action situations reveal the mechanism
- Participants differ in knowledge and influence
- Position rules assign responsibilities
- Boundary rules decide who may participate
- Choice rules define available actions
- Aggregation rules shape collective decisions
- Information rules determine what is shared
- Payoff rules structure consequences
- Scope rules limit permissible outcomes
Shared resources, monitoring and governance structures · Chapters 13–24
- Use seven rule types as a diagnosis, not a slogan
- The environment can defeat a perfectly written rule
- Community conditions shape how rules function
- Common-pool resources are not inevitably ungoverned
- Worked case: an allocation shortfall
- Worked case: the maintenance contribution problem
- Monitoring makes agreements observable
- Sanctions should address a verified problem
- Dispute resolution provides a correction route
- Polycentric arrangements involve several centres of decision
- Adjacent decisions can constrain the focal decision
- Rules need revision when circumstances change
Outcomes, institutional learning and transfer · Chapters 25–36
- Judge an institution by outcomes, not paperwork
- A baseline is necessary before claiming improvement
- Survey denominators are part of the result
- Unintended consequences should be anticipated
- Rights and power should not disappear behind averages
- Informal practices can help or harm
- Rules can change how information is reported
- An institutional diagnosis should name its weakest link
- A policy can be understood without being trusted
- Institutional analysis must remain open to correction
- Change the environment and transfer the method
- Frequently asked questions and onward learning
1. An institution is a set of rules in action
Institutional analysis examines the rules, shared expectations, authority and incentives that shape repeated human decisions. An institution can be a formal court or service, but it can also include the practices through which people allocate a resource, agree on maintenance or resolve disputes. The analysis concerns what rules actually do, not the appearance of an organisational chart.
The Ostrom Workshop Institutional Analysis and Development framework provides a structured method for linking actors, situations and outcomes. Its value is in finding a testable explanation, not asserting that one ownership arrangement or management style works everywhere. It must be adapted to the actual community and physical setting.
2. Start from a decision that can be observed
Instead of asking whether a whole town has good governance, ask how a community room is allocated, how a water quota is set or how a pump repair receives approval. Identify the participants, time period, rules and outcome. A narrow question can be answered with evidence; a national stereotype or slogan about institutions cannot.
Separate making a policy from applying it, auditing it and challenging its result. The same organisation may contain several connected action situations, with different legal responsibilities. A slow service might reflect limited resources, unclear authorisation or an inaccessible information process rather than a lack of willingness to cooperate.
3. A written rule may not describe everyday practice
A policy might guarantee equal access while actual users encounter undocumented exceptions, outdated instructions or informal priorities. Researchers distinguish formal rules from rules in use. Some adaptations help people respond to unusual needs, while others cause inconsistent decisions or exclusion; the difference requires evidence.
Compare official documents, appropriately obtained interviews, observed processes and service records. Ask who knows the rule, whether it can be followed, and who decides on an exception. A mismatch is a reason to investigate, not automatic evidence of misconduct. The goal is to understand why participants made the choices they did.
4. Action situations reveal the mechanism
An action situation identifies participants, positions, available outcomes, information, control, choices and costs or benefits. Elinor Ostrom’s original action-situation work provides a foundation. A researcher can use the model to move from an abstract complaint to a concrete description of who can make which decision.
A fictional water group has several action situations: households request supplies, technicians monitor equipment, a committee sets schedules and an authorised official reviews complaints. Each decision affects the others, but they need not have the same participants. An analysis that mixes all four together may recommend a solution no one has authority to implement.
5. Participants differ in knowledge and influence
People affected by one rule do not always have the same resources or decision rights. Residents, farmers, engineers and regulators may all rely on a catchment, yet only some can alter the schedule or view particular records. A study that refers to one undifferentiated ‘community’ can hide the experiences that matter.
Map who participates, who is absent and how each person’s needs enter the process. A family with limited digital access faces a different practical problem from an officer who operates the online form. The Sociology and Information Ethics roots help distinguish the effects of social position and information design.
6. Position rules assign responsibilities
Position rules identify roles such as applicant, chair, monitor, technician or reviewer. Authority belongs to the position under the governing arrangement, not automatically to whoever understands the problem best. A maintenance worker may see an urgent defect but lack power to approve the repair budget.
Trace the route from observation to authorised action. Who can report, verify, decide and notify others? When the ordinary official is unavailable, does the system have a legitimate escalation route? A reliable institution connects technical knowledge with appropriate decision authority instead of relying on personal favours.
7. Boundary rules decide who may participate
Eligibility rules may determine whether a renter, visitor, resident or registered organisation can apply or vote. Such distinctions can have legitimate purposes, but can also leave affected people without representation. An analysis should identify the rule and its practical impact rather than pronounce every boundary automatically fair or unfair.
Ask what documentation is required, how exceptions work and which alternative routes are available. A rule that is easy to satisfy for one group may be inaccessible to another. The relevant legal standard, purpose and evidence all matter to the conclusion.
8. Choice rules define available actions
A participant may be permitted to apply, inspect, approve, allocate or appeal, but not every role can perform every action. Choice rules define this scope. A system may intentionally separate technical inspection from final approval, yet can create a bottleneck if there is no authorised response during a service emergency.
An analyst should map each choice available under ordinary and exceptional conditions. If an equipment failure requires an urgent response, who can act safely and lawfully? Rules and resources should be understood together, as emphasised by Systems Thinking.
9. Aggregation rules shape collective decisions
Unanimity, majority voting and decisions by authorised representatives combine individual preferences in different ways. Unanimity may protect participants from being overridden but can delay action; a majority can reach decisions more quickly but needs safeguards for people who would otherwise be excluded.
There is no universal decision rule that maximises legitimacy and efficiency for every situation. Describe who is eligible to participate, the outcome being decided and which legal or ethical constraints apply. A committee’s decision about maintenance is not necessarily governed by the same process as a national authority’s legal regulation.
10. Information rules determine what is shared
Information rules affect which records, reasons and updates reach participants. A publicly announced allocation may be hard to follow if its formula is hidden or the available data are too old. A decision can be technically correct but practically opaque to people who need to understand or challenge it.
Transparency has legitimate limits. Publishing aggregate withdrawals may support accountability without disclosing private household records. A trustworthy arrangement explains the procedure and the relevant outcome while protecting information whose disclosure is not authorised or necessary.
11. Payoff rules structure consequences
Payoff rules shape costs and benefits attached to choices. Contributions to maintenance may cost time or money, while access to a shared resource can create personal benefit. If one person expects everyone else to cover the cost, individual incentives may conflict with maintaining the common system.
Not every participant seeks only financial gain. Duty, trust, social relationships and lawful obligations may influence choices. Model those factors where evidence supports them, and be honest about the assumptions. A numerical incentive calculation is not a reliable prediction of every person’s motivation.
12. Scope rules limit permissible outcomes
Scope rules define which outcomes the situation may produce. A neighbourhood committee may decide opening hours for a public garden without authority to change statutory safety standards. An online service may assign bookings but not create new legal eligibility categories.
Recommendations fail when they assign responsibilities beyond the legitimate powers of the participants. A good institutional analysis specifies which level can decide, which approval route applies and what remains outside the situation’s scope.
13. Use seven rule types as a diagnosis, not a slogan
Position, boundary, choice, aggregation, information, payoff and scope rules form a practical set of questions in the IAD tradition. When an application is delayed, the analyst can ask whether the problem lies in unclear authority, an exclusionary boundary, missing information or a choice requiring too many approvals.
Apply the seven questions to a school library’s room-booking service. Who can request a room, who approves, which decisions are collective, which records are visible and what happens if the room becomes unusable? The categories are useful because each points to a different potential repair.
14. The environment can defeat a perfectly written rule
Physical constraints shape choices. A water system may have fewer available units during drought; a library may have a broken lift; a fishery may have changed breeding conditions. An excellent policy cannot by itself supply a resource that does not exist or make unsafe infrastructure usable.
Record the capacity, variability and operational state of the resource. Distinguish an institutional question from a technical or environmental limitation while examining how they interact. The Infrastructure Studies and Ecology roots help connect policy with material conditions.
15. Community conditions shape how rules function
Shared language, expectations, local knowledge and past cooperation can influence whether participants understand a rule and trust its application. But these attributes are not permanent national traits. Institutions must be investigated with evidence about the people and circumstances actually involved.
Interview and survey methods need adequate coverage. A small group of confident speakers at a meeting may not represent the families who depend on a service but cannot attend. A credible analysis records whose experience is missing and what additional observation could address that gap.
16. Common-pool resources are not inevitably ungoverned
A common-pool resource is often difficult to exclude users from and subtractable in use: one withdrawal leaves less available for others. Such a resource can still have allocation, restoration and monitoring rules. The label should not be confused with unrestricted open access.
The Ostrom Workshop explains why shared resources must be analysed in their physical and social setting. The success of a particular arrangement is a research question, not proof that public, private or community management will always outperform alternatives.
17. Worked case: an allocation shortfall
Imagine a fictional reservoir providing 1,000 units in a dry season. Households request 450 units, farms request 500, and an ecological reserve is proposed at 200. Requests total 1,150, exceeding the available amount by 150. The calculation identifies scarcity but does not identify the group that should bear the shortfall.
A legitimate decision may require evidence about legal entitlements, environmental minima, vulnerability, measurement reliability and the authority of the deciding group. The institutional model asks who can choose, how reasons are communicated and how a mistake can be reviewed. Arithmetic does not determine justice.
18. Worked case: the maintenance contribution problem
A fictional pump delivers benefits worth 300 units to each of four users if repaired. Total benefit is 1,200, while the maintenance cost is 800. Splitting cost equally yields 200 per person and an illustrative net benefit of 100 each; aggregate net benefit is 400. That does not establish that people will voluntarily cooperate.
A participant may expect others to pay, lack resources, distrust the estimate or be bound by another contribution rule. Institutional analysis identifies the actual incentives and obligations rather than inferring a person’s character from nonpayment. It also asks who verifies completion of the repair.
19. Monitoring makes agreements observable
A quota may be difficult to apply consistently when the relevant withdrawals cannot be measured or the records are inaccessible. Monitoring can provide evidence about what was used, who maintains the system and whether an agreed condition was met. But a logged number is only as dependable as its definition and measuring method.
A reading might be disputed because of equipment failure or incomplete sampling. An authorised review process should examine the evidence before assigning consequences. See Measurement and Research for how records and uncertainty contribute to confidence.
20. Sanctions should address a verified problem
Rules may attach consequences to violations, but increasing punishment is not a universal solution. A failure to comply can reflect deliberate action, misunderstanding, technical breakdown or an inaccessible requirement. Proper authority and proportionate response are necessary before conclusions about fault are made.
A system that punishes everyone who reports a problem may discourage useful information. Institutional evaluation should examine unintended effects on trust and safety, not merely count penalties administered. The Ethics root supports analysis of fairness and responsibility.
21. Dispute resolution provides a correction route
Users may disagree over eligibility, meter values, allocation or procedure. A known process for checking records and reviewing a decision can prevent ordinary problems from becoming sustained distrust. A right of appeal mentioned only in a dense policy document may be inaccessible in practice.
Document who receives the challenge, which evidence is accepted and how the outcome is explained. Not every dispute requires a courtroom, but formal legal rights and court processes are relevant where applicable. See Judicial Studies for the specialised judicial context.
22. Polycentric arrangements involve several centres of decision
A neighbourhood water group, municipal utility and environmental authority may each control different parts of a shared system. Multiple centres can bring specialised knowledge and local adaptation, but also create coordination gaps if responsibility is unclear. More committees are not inherently better governance.
Map which choices belong to which level and how rules interact. A local group may know that a pump is failing while another institution controls funding. A solution limited to changing local voting procedures will miss the authority needed to fix the physical problem.
23. Adjacent decisions can constrain the focal decision
The people distributing water may depend on other actors to purchase equipment, approve repairs or monitor environmental effects. These connected action situations may set limits on what the focal committee can accomplish. A well-organised meeting does not repair a pump when the required budget remains unavailable.
Draw separate situations for provision, allocation, maintenance and review. Trace information and decisions between them. The IAD framework helps show why focusing only on user preferences can hide the institutional bottleneck that actually determines the outcome.
24. Rules need revision when circumstances change
Population growth, technology, climate and equipment conditions can make an older rule unsuitable. Institutional stability is useful when it preserves reliable expectations; it is harmful when it prevents legitimate learning from new evidence. Revisions should follow authorised procedures with clear communication.
When a new policy appears to improve service, check whether the definition of the measured outcome also changed. A digital portal may count completed forms differently from a paper process. A time trend without comparable data can mislead about the value of reform.
25. Judge an institution by outcomes, not paperwork
Meeting counts, signed forms and inspections are activities rather than proof that a service meets its purpose. A programme can hold many meetings while a broken pipeline still prevents households receiving water. A completed report does not show that users understand or can challenge a decision.
Define outcomes that match the public task: reliable access, timely repairs, legitimate rule application and appropriate treatment of affected people. Keep separate metrics when they describe different dimensions. A single score that hides severe failure for one group can be misleading.
26. A baseline is necessary before claiming improvement
If complaints fall after a new rule, rainfall, staffing or equipment may also have changed. A before-and-after comparison records an association, but cannot by itself isolate the rule’s causal effect. A useful study identifies plausible alternatives and the observations that might distinguish them.
Repeated measurements, comparison groups or phased implementation can help where ethically and practically appropriate. None is automatically a perfect experiment. The Statistics and Scientific Method roots provide tools for evaluating how strongly a finding supports a causal claim.
27. Survey denominators are part of the result
Suppose a fictional committee surveys 80 households. Sixty reply, and 42 respondents say that a new allocation schedule is clear. The share among respondents is 42 ÷ 60 = 70%. The share of all households invited who explicitly gave that answer is 42 ÷ 80 = 52.5%. The views of the other 20 are unknown.
It would therefore be misleading to announce that 70% of all households approve the rule. Clarity is also different from approval: a user may understand a rule but believe it is unfair. An evidence-grounded survey defines exactly which experience and population the question measures.
28. Unintended consequences should be anticipated
A faster digital process may exclude people without devices. A strict maintenance penalty may discourage reporting of uncertain defects. A rule designed to reduce resource withdrawals can create pressure to conceal use when measurement and participation are poor. These are hypotheses to examine, not guaranteed outcomes.
Before adopting a reform, identify who benefits, who bears a new burden and which behaviour the rule might encourage. Test plausible failure modes and create review processes. Good institutional design includes correction after implementation rather than celebrating the rule’s publication as the final achievement.
29. Rights and power should not disappear behind averages
A community may show an improved average outcome while a small vulnerable group loses meaningful access. Institutional analysis therefore records who can participate, who has authority and which rights or obligations constrain choices. An aggregate gain is not a complete answer to a fairness question.
A numerical model may quantify the total benefit of maintenance but cannot by itself authorise taking resources from a group whose entitlement must be protected. The Ethics and Government roots help place analytical findings beside lawful and normative responsibilities.
30. Informal practices can help or harm
Participants may develop practical routines not found in formal regulations. Some can speed up cooperation and share local knowledge; others can give preferential access to insiders or conceal decisions from people outside the group. It is not accurate to describe all informal rules as corrupt or all traditional rules as wise.
Document actual actions and their consequences. Compare the informal process with the written one, then test which users it serves and where it creates uncertainty. A good reform may preserve useful flexibility while replacing discretionary practices that cannot be fairly reviewed.
31. Rules can change how information is reported
A performance target can create incentives to change classifications rather than improve the actual service. For example, an office rewarded for closing requests quickly may mark a problem complete before the user can use the service. The reported count improves while the underlying need remains unresolved.
Audit what qualifies as a completed task, which records are retained and how users can correct mistaken closures. This is a measurement problem linked to a payoff rule. It helps explain why organisational targets need independent outcome checks and cannot be trusted solely from the dashboard they are designed to improve.
32. An institutional diagnosis should name its weakest link
A failed outcome may arise from scarce resources, unclear authority, poor information, inadequate incentives or an actual breach. These explanations may interact. A good diagnosis identifies the earliest point at which the intended action becomes impossible or misleading rather than scattering recommendations across every part of the system.
Write the expected action, the necessary conditions, the observed deviation and the evidence supporting the most plausible cause. Then propose a bounded correction and test whether the outcome changes. The Systems Thinking root deepens this method of locating dependencies.
33. A policy can be understood without being trusted
People may understand an allocation formula but distrust how meter readings are collected or how exceptions are approved. Others may trust a committee yet misunderstand its criteria. Communication, fairness perceptions and actual implementation are distinct features, even when they influence one another.
A good institutional evaluation can ask separate questions about comprehension, transparency, access to review and service results. Treating a positive satisfaction score as proof of lawful compliance would be a category error, just as treating a complaint as proof of a specific rule violation would be.
34. Institutional analysis must remain open to correction
An analyst may prefer community management, public administration or market arrangements based on previous experiences. The framework does not remove such perspectives automatically; its discipline lies in making hypotheses testable and considering evidence that could challenge them. A theory that cannot explain its own limits becomes an ideology rather than a useful tool.
Report uncertainties about sampled participants, changing rules and unobserved decisions. If a reform produces unexpected effects, revisit the original assumptions instead of redefining the metrics to protect the preferred story. The Research root develops practices for evidence correction.
35. Change the environment and transfer the method
Move the fictional water-allocation example to a school library with limited study rooms. The scarce resource is now space and time rather than water. The relevant actors, schedules, accessibility needs and service failures differ. The seven rule questions still help identify positions, eligibility, choices, joint decisions, information, costs and permitted outcomes.
Now assume the booking system becomes digital. Which users may struggle to enter, how do staff correct records and who can authorise a change when equipment fails? A capable learner preserves the analytical structure while replacing assumptions that belonged specifically to a reservoir.
36. Frequently asked questions and onward learning
Is every institution a government office? No; stable rules and practices occur in many settings. Do common resources inevitably collapse? No. Does a published rule prove everyone follows it? No. Is a majority decision automatically legitimate? Its legal and ethical context matters. Can a numerical shortfall decide who loses access? Not without the relevant authority, priorities and safeguards.
Continue into Systems Thinking, Sociology, Government, Ethics, Statistics and Infrastructure Studies. The Ostrom Workshop and Ostrom’s action-situation paper are primary references. All worked data are fictional. Learning transfer is explored in The Gold Standard of Education: Learning Transfers into Life.
| Request | Units |
|---|---|
| Households | 450 |
| Farms | 500 |
| Ecological reserve proposal | 200 |
| Total demand | 1,150 |
| Available | 1,000 |
| Shortfall requiring a legitimate decision | 150 |
| Quantity | Units |
|---|---|
| Four individual benefits of 300 | 1,200 |
| Total maintenance cost | 800 |
| Simple aggregate net benefit | 400 |