A tax works best when demand responds

1. How far the quantity falls depends on the size of the tax and on the price elasticity of demand, covered in 2.2.6, and the elasticity is the part the government cannot choose.
2. Where demand is inelastic, as with petrol and tobacco, the tax raises a lot of revenue and changes the activity very little.
3. A tax can still work in that case if it is banded, because the producer can escape it by changing the product rather than the buyer changing the purchase.

A subsidy works only if take-up follows

1. Lowering the price does not complete the job, because the household still has to decide to act and to find the remainder of the cost.
2. Take-up is weakest where the amount left to pay is large, which is why heat pump grants reach far fewer homes than the number eligible.
3. Some of the money also goes to people who would have bought the good anyway, so part of the spending buys no change in behaviour at all.

Regulation depends on enforcement, not the rule

1. A ban or a legal charge can change behaviour within weeks, which is faster than any tax or campaign manages.
2. That speed only appears where the rule is policed, since an unenforced limit changes nothing and costs the government its credibility.
3. Regulation also applies to everyone equally, so it does not let the better-off carry on by simply paying the charge.

Information provision is cheap but works slowly

1. Information changes what people want rather than what they can afford, so its effect lasts once the campaign has stopped.
2. It is also the slowest tool, because a change in what counts as normal behaviour takes years and cannot be dated to a single measure.
3. Its real value is as part of a package, since UK smoking policy stacked duty, the ban in enclosed public places, plain packaging, NHS campaigns and free stop-smoking services together.

Reaching a judgement on how well policies work

1. It depends on how responsive demand is, because an inelastic good needs regulation or reformulation rather than a price signal.
2. It depends on whether the harm can be measured, because a tax can only be set against an external cost that somebody has put a number on.
3. It depends on enforcement, because the strictest rule on paper achieves nothing if breaking it carries no consequence.
4. It depends on how long is allowed, because a campaign judged after a year and a ban judged after a month are not being tested on the same terms.
5. Overall: the policies with the strongest UK evidence behind them are the ones that changed the product or made the alternative easy, such as the sugar levy and the bag charge, while taxes on inelastic goods mainly raise revenue and information alone shifts behaviour only over many years.