Externalities and Social Costs
On this page
Prerequisites: Surplus, Taxes, and Subsidies
When a street-side workshop takes on an additional order, the buyer receives the goods and the workshop earns revenue, but the neighbor endures an extra night of noise. If one looks only at the accounts between the buyer and the seller, this impact has no price and does not automatically enter their decision-making. The omitted third-party cost is the starting point for understanding externalities.
An externality refers to a situation where one person's actions affect others, and these effects are not fully reflected in the prices or compensation faced by the decision-makers. Not all price changes constitute this type of market failure: a competitor lowering prices, causing you to earn less, is typically transmitted through market prices; noise damaging a neighbor's rest, however, may not be accounted for in this transaction.
Adding the Neighbor to the Original Marginal Comparison
Let daily output be approximated as continuous, measured in batches; the buyer's marginal value for the next batch is 100−10Q, and the workshop's marginal private cost is 20+10Q. Ignoring market power for now, the private quantity is determined where the two are equal: Q=4, at a price of 60.
Now, add an external noise damage of 20 per batch. This figure is a pedagogical setting for calculation ease; real damages require measurement and valuation, and are not automatically derived from decibel readings. The marginal social cost becomes 40+10Q.
| Output Level | Buyer's Value for Next Batch | Private Cost | Third-Party Damage | Social Cost |
|---|---|---|---|---|
| Q=2 | 80 | 40 | 20 | 60 |
| Q=3 | 70 | 50 | 20 | 70 |
| Q=4 | 60 | 60 | 20 | 80 |
At Q=3, the marginal value equals the social cost; increasing output further means the added value is less than the added resource and noise costs. Therefore, under this monetary valuation metric, the socially optimal quantity is 3, while the private market quantity of 4 is too high.
This does not mean society wishes for the workshop to cease production entirely. In the first three batches, there is still value exceeding social costs; allowing some production and some noise may be preferable to a complete ban. If the noise involves unacceptable health thresholds or rights constraints, those rules should be written directly into the feasible set, rather than pretending all impacts can be infinitely compensated with money.
Why Charging 20 Per Batch Changes Choices
Charging the workshop 20 per batch changes its marginal cost from 20+10Q to 40+10Q. When the buying and selling plans are compatible, the buyer pays 70, and the workshop keeps 50 after tax, resulting in a transaction volume of 3. The workshop still chooses based on its own returns, but the omitted cost has now entered the marginal account.
The government collects 20×3=60. This revenue is a transfer, not an indication that the noise damage has automatically disappeared. The remaining three batches still cause the set external damage of 60. The policy improvement comes from eliminating the output between the 3rd and 4th batches where costs previously exceeded values; in this example, the net benefit is the triangular area 1×20/2=10.
In the previous lesson without externalities, a tax reducing transactions would lose previously achievable gains; in this lesson, the eliminated transactions happen to have omitted third-party costs. These two conclusions do not conflict; the key is whether the no-tax baseline already included all relevant costs.
Such a corrective tax needs to approximate the marginal external damage at the target output level; simply averaging all historical damages does not necessarily yield the correct result. If the number of affected parties increases, and the damage per batch rises from 20 to 40, the optimal quantity would drop to 2; continuing to charge 20 would leave too much output. Policy rules require evidence-based updates.
Quotas, Fees, and Technical Requirements Solve Different Problems
In the simplified example with only one workshop and full information, directly limiting the total quantity to 3 can also achieve the same output. However, when there are multiple workshops with different noise-reduction costs, how quotas are allocated will affect resource costs.
Assume that for a given output level, one workshop incurs a cost of 5 per batch to install soundproofing facilities, while another incurs 30, both capable of eliminating the corresponding batch noise; the damage remains 20. If fees can be reliably charged based on actual uneliminated noise, the first workshop is willing to spend 5 to avoid paying 20, while the second would rather pay 20 than spend 30. Mandating that both workshops use the same soundproofing equipment might force the second workshop to pay more than the cost of the damage avoided.
This comparison fixes the output level and only illustrates the choice of abatement method; a complete solution must also consider output adjustments. Fees also have prerequisites: emissions must be measurable, responsible parties identifiable, and avoidance and collection costs controllable. When actual noise cannot be measured, equipment standards or time-of-day restrictions may be more enforceable; one cannot declare one tool superior in all environments based on a simple model.
Quotas provide clear numerical constraints, but how they are enforced, whether permits are tradable, and who receives initial rights still affect costs and distribution. Taxes provide price signals but do not guarantee that total emissions remain strictly at a certain value when demand suddenly increases. One should first decide what to control, and then select the rule.
Can the Neighbor and Workshop Negotiate Themselves?
If there is only one workshop and one neighbor, and both understand the damages and benefits, with clear rights and low negotiation and enforcement costs, they might change production or install soundproofing through compensation. This shows that negotiation can sometimes internalize external effects.
However, one cannot conclude from this that "government intervention is never needed." With a hundred neighbors, how do they collectively bid, who represents future residents, is there an incentive to understate damages, and can compensation be enforced? These factors may invalidate the conditions for low-cost negotiation. Initial rights also affect wealth and negotiation outcomes; even if an ideal model achieves efficiency under different rights allocations, it does not mean distribution is unimportant.
Moving the factory far away does not necessarily eliminate the externality; it may simply shift the burden to a different set of recipients. Evaluation must maintain consistent system boundaries, tracing the full path of affected third parties along with the behavior.
Positive Externalities Are the Other Side of the Same Ledger
If a company trains employees, and some skills spill over to benefit other firms through mobility, the trainer may not receive all social benefits, leading to potentially low private investment. However, not all training should be subsidized at the same rate: one must first confirm that additional benefits indeed spill over and would not otherwise be internalized by wages or contracts, and estimate how much additional training the subsidy generates rather than just replacing the company's original expenditure.
For both positive and negative externalities, the core steps are the same: identify the third party, clarify the impact path, estimate the marginal impact, check if existing prices or contracts already include it, and then compare specific rules. One cannot skip quantity, cost, and controls simply because something "is good for society."
Checking Conclusions Under New Conditions
After the workshop installs soundproofing, the noise damage per batch drops to zero, but the government continues to collect the original noise tax of 20 per batch. What happens?
Expand Reasoning
If other conditions remain unchanged, the previously omitted marginal damage has disappeared. Continuing to levy the same corrective tax would suppress output to an excessively low level. One must confirm the tax base, actual damages, and other policy objectives, rather than perpetuating the tax simply because "it was effective before." The cost of installing equipment should also be entered into the new production costs and not omitted from the new comparison.
Some add the 60 in tax revenue to the 10 in efficiency gains, claiming society gains 70 as a result. What is the problem?
Expand Reasoning
The 60 in tax revenue is transferred from private entities to the government; it is not itself new resources; the 10 is the net improvement from eliminating excessive production under the current metric. How the government uses the tax revenue may generate additional benefits or costs, but this should be evaluated as a separate explicit expenditure effect, rather than automatically adding all tax revenue to net benefits.
Sources and Further Reading
These references support concepts and statistical definitions; the numerical cases and diagrams are original synthetic teaching examples.