Power is the capacity to shape what happens.
In one line: power works when a person, group or institution can alter choices, rules, resources, information, incentives or access strongly enough to change another person’s behaviour or the final outcome.
Evidence boundary: Power is studied across political science, sociology, economics, psychology and organisational research, and no single definition captures every form. This article uses an operational public model: power changes the feasible choices, payoffs, information or decision rights available to actors. It does not assume that all influence is abusive or that formal authority is the only kind of power.
Power is often imagined as command: one person tells another what to do. That is only the most visible form.
Power can also decide which choices exist, who enters the room, what information becomes visible, whose definition of the problem becomes official, which resources are available and what happens to someone who refuses.
What Is Power?
Power concerns the ability to affect outcomes despite uncertainty, competing interests or resistance.
Resources/authority/information → options and constraints → decisions → action → consequences → feedback → stronger, weaker or redistributed power.
1. Formal Authority Gives Decision Rights
Parents, teachers, managers, judges and governments can hold formal authority defined by roles and institutions.
Authority specifies what a person is permitted to decide and what others are expected to accept. Legitimate authority can reduce coordination cost because every disagreement does not need to be renegotiated from zero.
But authority should remain bounded by role, law, evidence and accountability. “I have authority” is not the same as “every use of my authority is justified.”
2. Resources Create Power Because Others Need Them
Money, land, technology, expertise, labour, infrastructure and access can all produce bargaining power.
If one actor controls something others strongly depend on, that actor can often influence terms. Dependence matters as much as possession.
Power therefore grows not only from having resources but from how difficult those resources are to replace.
3. Information Changes Power
People who know more about the situation can frame choices, predict consequences or detect risks that others cannot.
An expert may possess informational power because others cannot easily evaluate the claim independently. A platform may possess power because it controls data about users. A student may be disadvantaged if everyone discusses their performance while they do not understand the criteria being used.
Transparency can redistribute power by making relevant information available to more actors.
4. Agenda-Setting Power Decides What Gets Discussed
One of the strongest forms of power appears before the final decision.
Who decides which problem enters the meeting? Which options are considered realistic? Which evidence is admissible? Which issue never reaches the agenda?
If only two options are presented, the person choosing between them may have less power than the person who designed the option set.
5. Rules Shape Power Before Any Individual Acts
Institutions distribute decision rights, protections, obligations and access.
The World Bank’s governance framework describes policy arenas where actors with unequal power interact inside changing rules. Unequal power can produce exclusion, capture and clientelism, especially when some actors can influence rules while others struggle even to enter the decision process.
Power is therefore partly personal and partly structural.
6. Networks Multiply Power
Relationships can provide access to information, resources, legitimacy and coordinated action.
A person with modest individual resources may become powerful inside a strong coalition. A highly resourced individual may become weaker if isolated from the networks through which decisions actually move.
Power can therefore be distributed across relationships rather than stored entirely inside one actor.
7. Legitimacy Makes Power Cheaper to Exercise
People often comply voluntarily when they regard authority or procedure as legitimate.
This reduces the need for constant coercion. A trusted teacher does not need to threaten students every minute. A legitimate institution can coordinate behaviour through accepted rules.
Legitimacy can be damaged when rules are arbitrary, selectively applied or impossible to challenge.
8. Incentives Translate Power Into Behaviour
Power often operates by changing the payoffs around action.
A manager can allocate bonuses. A government can tax or subsidise. A school can attach privileges or consequences to behaviour. A platform can make one type of content more visible than another.
The World Bank notes that governance change can occur by shifting the incentives, preferences and contestability of powerful actors rather than by rewriting formal structures alone.
9. Power Asymmetry Changes Bargaining
When one side can walk away easily and the other cannot, negotiation is not symmetrical.
Workers, consumers, students, tenants, suppliers and citizens can enter relationships with very different fallback options. Those differences affect what people can safely refuse.
Understanding power therefore requires asking not only “Did both sides agree?” but “What alternatives did each side realistically have?”
10. Power Can Produce Capability
Power is not inherently harmful.
Parents use authority to protect young children. Governments coordinate public goods. Teachers set learning conditions. Leaders allocate resources. Institutions enforce rules that prevent stronger actors from imposing costs on weaker ones.
The important question is what the power is used to make possible, for whom and under what safeguards.
11. Unchecked Power Corrupts Information
When people fear the consequence of contradiction, feedback becomes less reliable.
Bad news is softened. Metrics are selected to please. Uncertainty is hidden. The powerful actor may then become less informed precisely because everyone has learned to protect them from uncomfortable evidence.
This is a systems problem: excessive power can reduce the information quality available to the person who holds it.
12. Contestability Creates a Correction Path
Contestability means decisions and positions can be challenged by other actors rather than becoming permanently insulated.
Appeals, elections, competition, audits, independent review, open evidence and alternative providers can all create forms of contestability.
The World Bank treats greater contestability as one route through which excluded actors can influence policy arenas and power asymmetries can be mitigated.
13. AI Is Creating New Power Concentrations
The World Bank’s 2026 World Development Report highlights how control over frontier models, advanced chips, data centres and capital can create new power asymmetries among countries, governments, corporations and individuals.
This makes AI a power question as well as a technology question: who controls the infrastructure, who owns the data, who sets the rules, and who can realistically refuse the system?
14. Education Should Teach Power Without Teaching Cynicism
Students should learn to notice authority, incentives, agenda-setting and unequal access without concluding that every relationship is merely manipulation.
A useful analysis asks: What power exists? What legitimate job does it serve? What limits it? Who can challenge misuse? Who receives the consequence?
That produces better judgement than either blind obedience or automatic suspicion.
The Whole Power Chain
Resources/authority/information → option set → incentives and constraints → decision rights → action → consequences → legitimacy and feedback → challenge or reinforcement → redistribution or concentration of power.
A Useful Metaphor: Power Changes the Shape of the Room
Some people choose where the doors are. Some control the microphone. Some can leave easily. Some need permission to enter. Some decide what counts as evidence.
Power is not only who speaks loudest inside the room. It is also who designed the room and who can redesign it.
Power at Three Zoom Levels
Micro: one interaction
Who can influence the choice, and what happens if someone says no?
Meso: one organisation
Who controls information, resources, promotion, rules, agenda and escalation?
Macro: institutions and society
Which groups can shape rules, enter decision arenas and distribute benefits or costs across others?
How Power Fails
- Authority inflation: a legitimate role is used to justify decisions outside its proper scope.
- Capture: rules or institutions are bent toward the interests of a narrow group.
- Information suppression: people hide evidence because contradiction is unsafe.
- Agenda monopoly: important issues cannot enter the decision process.
- Dependency exploitation: one side uses the other’s lack of alternatives to impose unreasonable terms.
- Legitimacy theatre: formal procedure exists but meaningful challenge is impossible.
- Power blindness: outcomes are described as freely chosen while structural constraints remain invisible.
How Power Is Repaired
Clarify decision rights. Make evidence and criteria visible. Align authority with responsibility. Create independent challenge routes. Reduce unnecessary dependency. Protect the ability to report bad news. Expand meaningful participation where affected people are excluded.
The repair target is not “no power.” Complex systems need power to act. The target is power that remains bounded, legible, answerable and correctable.
What Parents and Students Should Notice
- Who can make the decision?
- Who controls the information used to make it?
- What realistic alternatives do affected people have?
- Which rules limit the powerful actor?
- Can people question the decision safely?
- Who bears the downside if the decision is wrong?
- Does the power increase another person’s capability or merely increase dependence?
Power Over, Power To and Power With Are Different Mechanisms
Power over concerns the ability to constrain or direct another actor. Power to concerns capability: the ability to make something happen that otherwise could not happen. Power with emerges when people coordinate resources, information or legitimacy so that collective capability exceeds what each actor could achieve alone.
A government can use power over through law enforcement, power to through infrastructure and public-service capacity, and power with when citizens, institutions and firms coordinate around a shared objective.
These forms can coexist. A system becomes misleading when every use of power is described only as domination or, in the opposite direction, when capability-building language hides coercive control.
Latent Power and Exercised Power Are Not the Same
An actor can possess power without using it. The credible possibility of action may be enough to change behaviour.
A regulator rarely needs to prosecute every violation if firms believe enforcement is possible. A manager may not need to reject every request if employees know the role has final approval authority. A country may derive bargaining power from capabilities it hopes never to use.
Potential action → others update expectations → behaviour changes before the power is visibly exercised.
Credibility Determines Whether Threats and Promises Carry Power
A threat that cannot realistically be carried out has weak power. A promise from an actor unable or unwilling to deliver also has weak power.
Credibility depends on capability, incentives, reputation, institutional constraints and whether carrying out the stated action would still make sense when the moment arrives.
This creates an important distinction between declared power and credible power.
Power Often Comes From Control of Dependencies
Dependency is relational. Actor A becomes more powerful over Actor B when B strongly needs something A controls and has few acceptable alternatives.
The same resource can therefore create very different power in different contexts. A rare specialist has greater bargaining power when replacement takes months than when many equivalent specialists are available immediately.
Dependence rises when need is high, substitutes are weak, switching is costly and time pressure is severe.
Exit, Voice and Alternatives Change Power
An affected person has more practical power when they can leave, challenge, appeal, switch provider, organise collectively or reach an independent authority.
Formal consent can therefore overstate real freedom when exit is prohibitively costly. Conversely, a system with meaningful alternatives can constrain powerful actors even without constant direct supervision.
Healthy institutions usually preserve more than one correction route because any single route can itself be captured.
Veto Power Is Different From Agenda Power
An actor with veto power can block a proposal. An actor with agenda power can influence which proposals are considered in the first place.
These powers have different effects. A veto player may prevent change while lacking the ability to initiate alternatives. An agenda setter may shape outcomes strongly while never issuing a final command.
High-resolution analysis therefore maps initiation, agenda, amendment, veto, implementation and review separately.
Implementation Power Can Defeat Formal Decisions
A policy can be approved centrally yet altered substantially during implementation. Front-line discretion, budgets, technical systems, staffing, interpretation and local priorities determine what the rule becomes in practice.
This means the person who formally decides and the people who operationally shape the outcome can occupy different power layers.
Power maps should therefore follow the full chain from decision right to real-world receipt.
Power Can Shape Preferences, Not Only Choices
Power sometimes operates by changing the menu of available choices. At deeper levels, institutions, culture, advertising, professional training or information environments can influence what people come to regard as normal, desirable or even imaginable.
This does not make every preference artificial. It means preference formation itself occurs inside social and informational systems.
A careful analysis therefore separates what a person chose, which alternatives were available, and how the preference landscape was formed.
Control of Measurement Creates Power Over Reality’s Representation
The actor who chooses the metric can influence what becomes visible. If performance is defined only by one indicator, work not represented by that indicator can disappear from the decision layer.
This is why independent measurement, multiple evidence routes and receiver feedback matter. Otherwise the same actor may control the action, the metric and the interpretation of whether the action succeeded.
Power becomes harder to correct when it controls both the world-changing action and the evidence used to judge that action.
Countervailing Power Prevents One Centre From Becoming the Whole System
Countervailing power means another actor or institution has enough independence and capability to challenge, constrain or expose the first.
Examples include independent courts, auditors, professional bodies, competing firms, unions, free media, boards, ombuds functions and internal escalation channels. The exact institution depends on the domain.
The design principle is broader: important power should rarely depend on the goodwill of the same actor it is meant to constrain.
Separation of Powers Trades Speed for Error Containment
Dividing initiation, approval, execution and review among different actors can slow decisions. That friction is sometimes deliberate because it reduces the chance that one mistake, bias or private interest controls the entire chain.
Too much fragmentation can create paralysis. Too little can create unchecked concentration. The appropriate balance depends on consequence, reversibility and urgency.
Emergency Power Changes the Normal Control Ratio
Crises can justify temporary concentration of authority because delay itself may be dangerous. But emergency powers create a second problem: how and when does the system return to normal?
Strong emergency architecture therefore defines scope, trigger, review, evidence requirements and expiry rather than treating extraordinary authority as self-terminating.
Power Reproduces Itself Through Institutions and Inheritance
Resources, networks, status, ownership and access can persist across time. Power today can therefore alter who possesses power tomorrow.
A dominant platform can use current scale to collect more data and reinforce future scale. Wealth can purchase education, networks and investment opportunities. Institutional positions can shape succession rules.
This creates cumulative processes in which small initial advantages can widen unless countervailing mechanisms preserve contestability.
Power Can Be Concentrated Yet State Capacity Still Be Weak
Concentration of authority does not guarantee capacity. A leader may possess broad formal power while institutions lack competent staff, accurate data, logistics or implementation ability.
This distinction matters because ability to command and ability to produce the intended outcome are different variables.
Healthy systems need enough legitimate power to act and enough operational capability to make action real.
A High-Resolution Power Audit
- Outcome: What can this actor actually change?
- Source: Does power come from authority, resources, information, status, network position, expertise or control of alternatives?
- Type: Is this power over, power to or power with?
- State: Is the power latent or currently exercised?
- Credibility: Can threats, promises or commitments actually be carried out?
- Dependency: Who needs whom, and how replaceable is each side?
- Exit: Can affected actors leave or switch?
- Voice: Can they challenge, appeal or organise collectively?
- Agenda: Who decides which questions and options become visible?
- Veto: Who can block action?
- Implementation: Who can reshape the decision during execution?
- Measurement: Who defines success and controls the evidence?
- Preference formation: Who shapes what actors regard as normal or possible?
- Countervailing power: Which independent actor can constrain or expose misuse?
- Concentration: How many independent failure points exist before one actor controls the whole chain?
- Emergency: Are extraordinary powers bounded by trigger, scope and expiry?
- Inheritance: Does present power reproduce future power automatically?
- Capacity: Can the actor actually implement what it commands?
- Receiver: Who experiences the consequence and can their evidence travel back?
- Correction: What happens when the powerful actor is wrong?
Connect Power to the Wider eduKateSG Mechanism Estate
- How Rules Work — how power becomes stabilised, constrained or reproduced through formal expectations.
- How Markets Work — how alternatives, concentration and switching costs change bargaining power.
- How Leadership Works — how legitimate direction differs from mere control.
- How Information Works — why control of observation, provenance and visibility changes decision power.
- How Responsibility Works — why authority without answerability creates structural failure.
Continue Through eduKateSG
Evidence and Further Reading
The World Bank’s World Development Report 2017: Governance and the Law provides the core governance evidence used here: unequal power in policy arenas can create exclusion, capture and clientelism, while change can occur through shifts in incentives, preferences, beliefs and contestability. The World Development Report 2026 extends the power question into AI, examining how control over models, chips, data centres, capital and surveillance capabilities can reshape power across countries, governments, firms and individuals.
Frequently Asked Questions
Is all power bad?
No. Power makes coordinated action possible. Parents, teachers, institutions and governments all need legitimate forms of power. The important questions are scope, purpose, accountability, proportionality and correction.
Is power the same as authority?
No. Authority is formally recognised decision power attached to a role. Power is broader and can also come from resources, expertise, information, networks, status or control over alternatives.
How can weaker actors gain power?
Possible routes include better information, stronger alternatives, collective organisation, legal protection, access to decision arenas, independent review and reduced dependence on one powerful actor.
Final compression: Power works by changing who can choose, who can refuse, which options exist, which information matters and who carries the consequences—so healthy systems make power strong enough to act but bounded enough to remain answerable to reality and people.