How The World Works | Externalities

A factory can pay for fuel, machinery, wages, rent and transport.

Its accounts may balance perfectly.

Then smoke leaves the chimney.

The smoke does not stop at the property line to ask whether the neighbours were part of the transaction.

A driver enters a crowded road and adds a little more delay to everyone behind. A nightclub produces music for paying customers and noise for someone trying to sleep two buildings away. A researcher discovers something useful, publishes it, and other people build on knowledge they did not fully pay to create. One household maintains a beautiful garden and passers-by enjoy a street they never funded.

The common structure is simple and enormous:

The decision happens here. Part of the consequence lands somewhere else.

Economists call that family of effects externalities.

Quick Read

An externality exists when an action imposes a cost or creates a benefit for others that is not fully reflected in the decision-maker’s private costs, benefits or market price.

Negative externalities include cases such as pollution, unwanted noise and congestion. Positive externalities include knowledge spillovers and other benefits that spread beyond the person or organisation that paid to create them.

The key problem is a boundary mismatch. The person making the decision optimises inside one boundary, while the consequences travel across a larger one.

That can make activities with external costs happen more than they would if all costs were counted, while activities with external benefits may happen less than they would if all benefits were rewarded.

The difficult work begins after the definition: measuring the spillover, identifying who receives it, deciding which effects are genuinely external, and choosing a remedy that does not create a larger problem than the one it is trying to solve.

The One-Sentence Answer

Externalities work when an action changes the welfare, opportunity or environment of people outside the transaction, but the decision-maker does not fully face those outside costs or receive those outside benefits, so private incentives diverge from the wider social effect.

The Boundary Problem

Most decisions require a boundary.

A household considers its budget. A company considers its costs and revenue. A driver considers travel time. A developer considers the economics of a project. A researcher considers funding, effort and expected reward.

But reality does not promise to respect the boundary used in the spreadsheet.

Air moves. Water flows. Noise travels. Traffic queues. Ideas spread. Disease can spread. Knowledge crosses organisational borders. Land use changes neighbouring land. One person’s security practices can affect others in a network. One company’s technical standard can increase or reduce compatibility for an entire ecosystem.

An externality appears when the consequence boundary is wider than the decision boundary and the spillover is not fully brought back into the decision.

Private Cost and Social Cost

Suppose producing one additional unit of a good costs a firm $10 in labour, materials, energy and other expenses.

That is part of the private cost.

Now suppose producing that unit also creates $3 of harm outside the firm that the firm does not pay for.

The wider social cost is $13.

If the decision-maker sees only $10, behaviour is guided by an incomplete price signal.

This does not automatically tell us what policy should be used. But it tells us why the private decision can diverge from the wider outcome.

Private Benefit and Social Benefit

The same logic works in the other direction.

A company funds research because it expects some private return. Yet part of the knowledge may spill into universities, suppliers, competitors or later innovations that the original company cannot fully capture.

The private benefit is smaller than the total benefit generated across society.

If decision-makers receive only part of the reward, they may invest less than would be desirable from a wider social perspective.

This is why positive externalities matter just as much as negative ones. A world map concerned only with harm would miss half the mechanism.

Negative Externalities

A negative externality imposes an uncompensated cost on others.

Pollution can impose environmental and health-related costs beyond the polluter’s accounts.
Congestion means one additional road user can add delay to many other travellers.
Noise can reduce the comfort or productivity of people who did not choose the source.
An insecure digital service can create risks for people elsewhere in a network.
Poorly controlled waste can impose clean-up or ecological costs on communities downstream.

The defining feature is not that the action has a bad side effect. The defining feature is that part of the cost falls outside the decision and is not fully internalised.

Positive Externalities

A positive externality creates an uncompensated benefit for others.

Basic research can generate ideas that later innovators use.
Open technical knowledge can help other builders solve problems faster.
A well-maintained property can improve the experience or value of a surrounding area.
Education can create benefits beyond the individual learner through productivity, civic participation and knowledge transmission, although the size and nature of such spillovers require empirical care.
A compatible standard can create value for participants who did not design the original system.

Positive externalities create a mirror-image problem: socially valuable activity may be underprovided because the creator cannot collect the whole return.

An Externality Is Not Just Any Effect on Someone Else

This distinction prevents the concept from becoming uselessly broad.

If a new bakery opens and customers switch away from the bakery next door, the incumbent loses revenue. Someone else was affected, but economists would not automatically call every competitive loss a technical externality. Much of that effect travels through market prices and ordinary competition.

Externalities are especially concerned with effects that occur outside the relevant price or contractual mechanism.

That is why the phrase unpriced spillover is often useful, even though real cases may involve partial pricing rather than no pricing at all.

The Externality Chain

action → intended private outcome → spillover pathway → outside receiver → unpriced cost or benefit → distorted incentive → repeated behaviour → aggregate social outcome

Each link matters.

What exactly is the action? Through what physical, informational or social route does the effect travel? Who receives it? Is the effect additional at the margin? Is it already reflected in a price, contract, tax, liability rule or regulation? Does it accumulate? Does it cross borders? How certain is the attribution?

Without those questions, “externality” can become a label attached to any disliked consequence.

Local Optimisation Can Create Global Failure

Externalities expose a general systems problem.

Each actor can behave sensibly inside a local objective while the combined system performs badly.

A driver chooses the road because it is fastest for that driver. Thousands of drivers do the same. The road becomes congested. A company minimises disposal cost. Many companies make the same calculation. A shared environmental resource degrades. Each participant’s action may be understandable from the private viewpoint.

The problem is that the objective function is too small for the consequence field.

This is why externalities connect directly to How Optimisation Works. Optimising the wrong boundary can make a system more efficiently wrong.

Pollution: The Classical Example

Pollution remains the standard example because the boundary mismatch is easy to see.

A producer receives revenue from the product. If the producer does not bear the full cost of emissions, some production costs are effectively transferred to people or environments outside the transaction.

The IMF’s explanation of externalities uses precisely this private-versus-social-cost distinction: when pollution costs are not borne by the producer or consumer, market prices do not capture the full social cost.

The important analytical move is not moral outrage. It is accounting at the correct boundary.

Congestion: The Cost That Appears in Other People’s Clocks

Congestion is a beautiful example because the external cost is time.

One additional vehicle may experience a private travel cost, but it can also slightly slow many other vehicles. The driver does not personally receive an invoice from each delayed traveller.

At low traffic volumes, the additional delay may be negligible. Near capacity, one more vehicle can impose much larger marginal delay.

This teaches another lesson: externalities are often nonlinear.

The same action can create very different external costs depending on system state.

Knowledge: The Benefit That Refuses to Stay Put

Ideas behave differently from many physical goods.

Once knowledge exists, other people may learn from it, imitate parts of it, combine it with other ideas or use it to ask new questions. The original researcher or company may capture some value through patents, secrecy, expertise, reputation or first-mover advantage, but not necessarily all of it.

NBER research on basic research and innovation policy treats these knowledge spillovers as a central reason private incentives may differ from social returns.

This is one reason societies fund universities, public research and other knowledge institutions. The precise policy mix is contestable, but the mechanism is clear: socially useful knowledge can travel beyond the original payer.

Externalities Can Travel Across Time

The receiver does not have to exist at the same moment as the decision-maker.

Some actions create delayed costs or benefits. Infrastructure maintenance deferred today may increase failure risk years later. Environmental damage may accumulate slowly. Research conducted now may produce benefits decades later. Debt, land use and resource depletion can transfer consequences across generations.

Time creates a special difficulty because future receivers cannot bargain in the present in the ordinary way.

This is not automatically an externality in every intertemporal case, but the lens becomes powerful whenever the people making the decision do not fully bear later consequences.

Externalities Can Cross Borders

Air, rivers, oceans, cyber networks, supply chains and financial systems do not stop neatly at political boundaries.

A country may receive the private benefit of an activity while part of the cost travels elsewhere. Or one country may invest in knowledge, disease surveillance, environmental protection or standards whose benefits spread internationally.

Cross-border externalities are hard because the decision-maker and receiver sit under different authorities. A domestic rule cannot always internalise a global effect by itself.

This is where externalities meet treaties, international governance and collective action.

Externalities Can Hide Inside Supply Chains

A final product can look clean because the visible transaction is far from the production process that created its external costs.

The buyer sees a price. Upstream, land, energy, water, labour, waste and transport systems may carry consequences that are only partly reflected in that price.

This does not mean every supply-chain impact is an externality. Wages, energy and transport costs may already be priced. The analytical job is to identify which consequences fall outside existing contracts or prices.

Tracing externalities therefore often requires following the chain beyond the point of sale.

Externalities and Data

Digital systems create new spillover patterns.

One person may consent to upload information that reveals something about relatives, colleagues or contacts. One organisation’s weak security can create risks for partners. One platform’s design can change information quality across a wider public environment.

These cases are not always clean textbook externalities because contracts, law and consent structures vary. But the boundary question remains useful:

Who is affected who did not meaningfully participate in the decision?

Modern systems increasingly need to ask that question because information travels cheaply and consequences can propagate quickly.

Externalities and Power

Externalities are not distributed randomly.

Some actors have more ability to push costs outward and protect benefits inward. Others have less ability to refuse exposure, move away, negotiate compensation or influence regulation.

A purely aggregate calculation can therefore hide distribution.

Suppose an activity creates $100 of private benefit and $80 of external cost. It may still look positive in a crude total. But who receives the $100? Who bears the $80? Are the costs concentrated on a small community while benefits are widely distributed? Are the affected people able to participate in the decision?

Efficiency and fairness are distinct questions. Both matter.

Externalities and Markets

Markets coordinate through prices.

When prices omit relevant social costs or benefits, they send incomplete signals.

A product appears cheaper than its full social cost. Demand can therefore be higher than it would be under fuller accounting. A socially beneficial activity produces benefits its creator cannot capture, so private investment may be lower than the wider benefit would justify.

This is one reason externalities are described as a form of market failure.

But “market failure” does not mean “government intervention automatically succeeds.” Any remedy has its own information requirements, administrative costs, political incentives and unintended consequences.

Serious analysis compares imperfect alternatives rather than comparing an imperfect market with an imaginary perfect regulator.

Internalisation: Bringing the Consequence Back

To internalise an externality means changing the system so the decision-maker faces more of the relevant outside cost or receives more of the relevant outside benefit.

The concept is powerful because it is mechanism-neutral. Internalisation can happen through many routes:

taxes or charges;
subsidies;
liability;
regulation;
property rights;
bargaining;
tradable permits;
standards;
technical redesign;
information disclosure;
collective provision;
social norms or contracts.

The correct tool depends on measurability, transaction costs, distribution, enforcement, uncertainty and institutional capacity.

Pigouvian Taxes: Put a Price on the Outside Cost

Arthur Pigou’s classic approach was to tax activities with negative externalities so the private decision reflects more of the social cost.

In the idealised version, the tax equals the marginal external harm.

This aligns incentives without requiring the regulator to dictate every individual action. Actors remain free to reduce the activity, adopt cleaner technology or pay the charge when the activity is still worth doing.

The difficult word is equals.

Real external costs can be uncertain, location-dependent, nonlinear or politically contested. Measuring the correct marginal harm is often difficult. A theoretically elegant tool can therefore become a practical measurement problem.

Subsidies: Reward the Outside Benefit

Positive externalities create the opposite logic.

If an activity produces benefits that spread beyond the payer, subsidies or public funding can increase activity toward a level closer to the wider social benefit.

Research funding is a common example because basic knowledge can produce spillovers that private firms cannot fully capture.

But subsidies also need discipline. If policymakers cannot distinguish high-spillover activity from ordinary private benefit, money can be misallocated. Positive externality is not a magic phrase that makes every subsidy wise.

Coase: Sometimes the Boundary Can Be Bargained

Ronald Coase changed the discussion by focusing attention on property rights and transaction costs.

If rights are clear and affected parties can bargain at very low cost, they may sometimes negotiate a mutually beneficial arrangement around the externality.

That insight is often simplified into “markets can solve externalities.” The more useful lesson is conditional.

Bargaining becomes difficult when thousands or millions of people are affected, harms are hard to measure, parties cannot be identified, information is asymmetric, enforcement is costly or future generations are involved.

Coase’s contribution therefore connects externalities directly to transaction costs: the real-world cost of finding parties, negotiating, specifying rights, monitoring agreements and enforcing them.

Regulation: Set a Boundary Instead of a Price

Sometimes society does not want to price every unit of harm. It sets limits.

Emission standards, zoning rules, safety requirements, noise limits and technical standards can restrict activities that impose external costs.

Regulation can be appropriate when harm is severe, measurement is difficult, thresholds matter or society wants a minimum floor rather than a continuous price trade-off.

But regulation creates its own questions: Is the rule targeted? Can it be enforced? Does it adapt to better technology? Does it unintentionally block low-harm alternatives? Are small actors burdened disproportionately?

See How Regulation Works.

Tradable Permits: Make the Total Constraint Explicit

Another approach sets a total quantity and allows rights to the constrained activity to be traded.

The system separates two decisions:

How much total activity or pollution is acceptable?
Who can reduce it most cheaply?

In theory, trading can move reductions toward lower-cost opportunities while respecting the overall cap.

Again, design matters. Monitoring, allocation, enforcement, market power and the geographical distribution of harm can all affect outcomes.

Technology Can Remove the Spillover Path

Economics is not the only way to internalise an externality.

Sometimes engineering changes the causal pathway itself.

Filters reduce emissions. Better insulation reduces noise. Waste-treatment systems reduce downstream contamination. Cybersecurity architecture reduces propagation of digital risk. Public transport can change congestion dynamics. Interoperability can reduce ecosystem lock-in.

This is a powerful reminder: do not assume every social problem must be solved through prices or prohibitions. Sometimes the best intervention changes the physics or interface of the system.

Measurement Is the Hard Part

The phrase “make polluters pay” is simple. Calculating exactly what they should pay can be extraordinarily difficult.

Externalities often require answers to questions such as:

What is the marginal harm from one additional unit?
Does harm depend on location or time?
Who is exposed?
How long does the effect last?
What would have happened anyway?
Are effects reversible?
Are there thresholds beyond which harm rises sharply?
How should uncertain future effects be valued?
How do we avoid counting the same harm twice?
What evidence would change the estimate?

Externality policy is therefore inseparable from evidence, models, uncertainty and monitoring.

Average Cost Can Hide Marginal Cost

Many decisions happen at the margin.

The relevant question is not always “What is the average harm from traffic?” but “What additional harm does one more trip create at this time and place?”

On an empty road, the congestion externality of one extra vehicle may be tiny. Near capacity, the same vehicle can create larger delay.

This is why pricing and regulation based on broad averages can miss the actual causal structure.

Externalities Can Be Reciprocal

Coase also pushed economists to see that many social-cost problems are reciprocal.

A noisy activity harms a neighbour who values quiet. But banning the activity also imposes a cost on the producer and customers. The problem is not solved analytically by declaring one side nonexistent.

Society must decide which arrangement creates the better total outcome, subject to rights, fairness and institutional constraints.

This does not mean all harms are morally symmetrical. It means the policy comparison should count consequences on all sides.

Externality Shifting: Solving It Here, Moving It There

A system can appear to reduce an externality by exporting it.

A city becomes cleaner because dirty production moves elsewhere. A company reports lower direct emissions while upstream suppliers carry more. A digital service reduces visible moderation costs by shifting burden to users. A waste policy moves disposal across a border.

This is why boundary selection matters so much.

If measurement stops where the organisation stops, the externality may disappear from the report while remaining in the world.

Externalities and Path Dependence

External costs can accumulate into infrastructure.

If an activity is underpriced for decades because some costs remain external, investment patterns may grow around that activity. Housing, transport, factories, skills and supply chains adapt.

Later, even if society begins pricing the externality more accurately, transition can be difficult because people made real investments under the old rules.

This is the bridge to How The World Works | Path Dependence.

The past cost boundary helped build the present path.

Externalities and Defaults

A default can influence how often an activity occurs. If the activity carries an externality, the default indirectly changes spillovers.

Imagine a digital service that defaults to high data sharing. The immediate design question concerns choice architecture. The wider question asks whether one person’s default disclosure affects others. The mechanisms are distinct but connected.

See How The World Works | Defaults.

Externalities and Incentives

Externalities are fundamentally an incentive-alignment problem.

The decision-maker responds to one set of costs and benefits while society experiences another.

Internalisation attempts to bring those sets closer together.

This is why How Incentives Work is a natural companion: externality policy often changes prices, rewards, liabilities or constraints so private behaviour responds to wider consequences.

Externalities and Public Goods Are Related, Not Identical

A public good is typically defined by non-rivalry and non-excludability: one person’s use does not significantly reduce another’s, and it is difficult to exclude non-payers.

Externalities are spillover effects outside the decision or price.

The concepts often interact. Public goods can create positive spillovers, and free-rider problems can reduce private provision. But they are not synonyms.

Keeping the distinction lets us ask the correct question: Is the main problem an unpriced effect, non-excludability, non-rivalry, or some combination?

Externalities and Common-Pool Resources Are Also Different

A common-pool resource is difficult to exclude people from but rivalrous in use: one person’s extraction reduces what remains for others.

Overuse can involve externalities because each user may not bear the full cost imposed on others. But again, the concepts should remain distinct.

Good systems thinking does not flatten neighbouring ideas. It keeps boundaries clear so connections become meaningful.

The Distribution Test

After estimating total cost and benefit, ask who receives each part.

Does the decision-maker receive most of the benefit?
Are costs concentrated on people with less bargaining power?
Do benefits spread widely while one group pays?
Are future generations carrying part of the cost?
Do different neighbourhoods experience different exposure?
Can affected people exit, negotiate or be compensated?

An externality can be small in aggregate and severe for a particular receiver.

That is why averages do not finish the analysis.

The Uncertainty Test

Some externalities are well measured. Others are uncertain.

Uncertainty does not mean “ignore the effect.” Nor does it mean “assume the worst estimate is true.”

It means represent uncertainty honestly.

What is known?
What is estimated?
What range is plausible?
Which assumptions drive the result?
How reversible is the harm?
What new evidence should trigger revision?

Externality policy under uncertainty is therefore partly a risk problem.

The Externality Audit

Define the action. What decision or activity are we analysing?
Define the private boundary. Which costs and benefits does the decision-maker currently face?
Trace the spillover. How does the effect leave that boundary?
Identify the receiver. Who or what experiences the outside consequence?
Check pricing. Is the effect already reflected in a contract, price, tax, liability or rule?
Measure the margin. What does one additional unit of activity change?
Check time and place. Does the effect vary by location, congestion, season or system state?
Check distribution. Who gains and who pays?
Check uncertainty. Which parts are measured and which are modelled?
Check substitution. Would a remedy simply move the activity elsewhere?
Compare tools. Price, standard, property right, bargaining, technology, subsidy, information or direct provision?
Observe return. Did the intervention reduce the actual external harm or increase the actual external benefit?

When the Externality Lens Fails

The lens fails when every unpleasant consequence is casually labelled an externality.

Competitive losses, ordinary price changes, private disappointment and contractual costs are not automatically externalities. Nor is every public benefit evidence of a positive externality large enough to justify subsidy.

Ask whether the effect truly escapes the decision-maker’s price or contractual boundary.

Precision protects the concept from becoming moral vocabulary instead of causal vocabulary.

A Better Design Principle: Follow the Consequence to Its Receiver

One of the easiest ways to miss an externality is to stop observing at the point where the transaction completes.

The product was sold. The journey was made. The waste left the building. The data was uploaded. The research paper was published.

Operationally, the event looks finished.

Causally, it may only be beginning.

Follow the smoke. Follow the queue. Follow the discarded material. Follow the information. Follow the idea. Follow the downstream maintenance. Follow the person who receives the consequence without appearing on the invoice.

That is where the hidden accounting begins.

How Externalities Connect to the Rest of the World

Markets: prices coordinate decisions, but externalities mean prices can omit wider effects.
Incentives: private rewards and costs may not align with social rewards and costs.
Power: some actors can push costs outward more easily than others.
Regulation: rules can set boundaries where prices alone are insufficient.
Risk: future and uncertain external harms require probabilistic judgement.
Monitoring: effects must be measured at the receiver, often over time.
Path dependence: underpriced activities can build long-lived infrastructure around themselves.
Friction: transaction costs determine whether affected parties can bargain or switch.
Standards: shared technical rules can reduce harmful spillovers or create beneficial compatibility effects.
Innovation: knowledge spillovers can make private and social returns diverge.

Questions a Reader Can Now Ask

Who makes the decision?
Who receives the consequence?
Which costs appear on the decision-maker’s ledger?
Which costs do not?
Which benefits spread beyond the payer?
Is the spillover physical, informational, social or temporal?
Is it already priced or regulated?
Does it grow near a threshold or capacity limit?
Who bears the marginal effect?
Could bargaining realistically solve it?
Would a tax, subsidy, standard, liability rule or technology change the mechanism?
Could the proposed solution merely move the externality elsewhere?
How would we know the receiver actually improved?

Frequently Asked Questions

Are externalities always bad?

No. Negative externalities impose outside costs; positive externalities create outside benefits. Knowledge spillovers are a classic positive example.

Is every side effect an externality?

No. The concept is most useful when a cost or benefit affects people outside the transaction and is not fully reflected in the relevant price or incentive. Ordinary competitive effects through prices are not automatically technical externalities.

Does an externality always justify government intervention?

No. The existence of a market failure does not prove a particular remedy will improve the outcome. Policy has information, enforcement and administrative costs. The relevant comparison is among feasible alternatives.

What does “internalise the externality” mean?

It means changing the decision environment so the actor faces more of the external cost or receives more of the external benefit—for example through prices, rights, liability, regulation, subsidies, standards or technical design.

Why are externalities difficult to measure?

Because effects may be dispersed, delayed, uncertain, nonlinear or difficult to attribute. The relevant marginal impact can also vary by time, place and system state.

Research Basis and Further Reading

International Monetary Fund, “Externalities: Prices Do Not Capture All Costs”.
Ronald H. Coase, Nobel Prize materials on transaction costs and property rights and the institutional structure of the economy.
Ufuk Akcigit, Douglas Hanley & Nicolas Serrano-Velarde, “Back to Basics: Basic Research Spillovers, Innovation Policy and Growth”, NBER.
Wolfgang Keller, “Knowledge Spillovers, Trade, and FDI”, NBER.
International Monetary Fund, “What Are Global Public Goods?” for the relationship among spillovers, public goods and free-rider problems.

The Larger Idea

A receipt is a beautiful object because it draws a boundary around an exchange.

This much was bought. This much was paid. These were the parties.

The world is less tidy.

The sound leaves the room. The smoke leaves the chimney. The traffic delay enters other people’s clocks. The idea leaves the laboratory. The waste moves downstream. The code connects to other systems. The benefit spreads. The cost spreads.

And suddenly the transaction has more receivers than the receipt records.

That is the deep usefulness of externalities as a world lens.

They teach us to distrust boundaries that are administratively convenient but causally incomplete.

If a consequence escapes the decision, follow it until you find the person, place or future that receives it.

How The World Works | Externalities — When Consequences Escape the Decision