Intervention brings gains beyond the market itself

1. The main benefit is that the quantity produced moves closer to the quantity society would choose once third parties are counted.
2. The harm itself falls, so cleaner air, less waste and better health are gains that never appear in any firm's accounts.
3. A tax also raises revenue, which can fund the treatment or the clean-up the activity caused, or anything else in 3.5.1.
4. A subsidy can raise consumption of a good with external benefits permanently, because a household that installs a heat pump keeps it for years.

Every policy costs money to run and enforce

1. Someone has to collect the tax, check the claims or police the rule, and that administration cost is paid whether or not the policy works.
2. A subsidy is a direct call on the budget, so it either raises borrowing or takes the place of other spending.
3. That is the opportunity cost, covered in 1.2.3, and the case for the policy has to beat the best alternative use of the same money.

The burden falls unevenly across households

1. An indirect tax takes a larger share of a low income than of a high one, which makes it regressive in the sense set out in 3.5.7.
2. A subsidy can run the same way, because the household that can find the rest of the cost of a heat pump is rarely the poorest one.
3. So a policy can correct the externality and widen the differences in 3.3.3 at the same time, which is why who pays is part of the judgement.

Getting the level wrong creates new problems

1. Setting a tax above the external cost cuts the activity below what society wants, and setting it below leaves part of the harm in place.
2. Nobody knows the external cost exactly, so the level is an estimate and the risk of getting it wrong cannot be removed.
3. Unintended consequences follow too, as buyers switch to an untaxed substitute or a taxed producer moves the activity abroad.

Reaching a judgement on whether to intervene

1. It depends on the size of the harm, because a small external cost is not worth an expensive scheme to collect and police.
2. It depends on whether the harm can be priced, because a tax needs a number and regulation does not.
3. It depends on what the money would otherwise buy, since the opportunity cost is the real test of a subsidy.
4. It depends on who ends up paying, because a policy that corrects an externality by taxing the poorest hardest has bought its gain at a cost worth naming.
5. Overall: intervention is worth it where the harm is large, measurable and enforceable, as with tobacco and carbon emissions, because there the benefit clearly beats the administration and the opportunity cost, but for small or hard-to-value externalities the safer judgement is that the cure costs more than the problem.