27 Jan 2009

Pigouvian Tax Fail?

NW asks:
I understand very well the rationale for taxation to reduce pollution such as greenhouse gases. However, from an economics point of view, is that the only arsenal for controlling pollution? Are there instances when Pigouvian taxes just don't work, and that an outright ban is needed? If so, how do you justify using a ban instead of a Pigouvian tax?
These questions are easy to answer in theory ("it depends!") but hard to answer in practice. Here are a few things to consider:
  1. What is price elasticity before and after a tax is imposed? If it's low before and high after, then a tax is MOST effective. If it's low and then low, then a tax may not change behavior "enough."
  2. What is being taxed? What is being regulated? Are they the same things as the targeted behavior? If we want people to drive less but tax gas, then people may switch to fuel-efficient cars and drive the SAME amount.
  3. How do psychology and social dynamics matter? When the French government banned cigarettes in bars and restaurants, everyone ignored it.
  4. Where does the tax revenue go? If the revenue goes to the general fund or a special interest group, the tax may be too high, or never change -- even when behavior does. My favorite example of this was the phone excise tax to pay for 1898 Spanish American War many wars. That tax was finally rescinded in 2005 -- after consumers had paid the Feds $300 billion.
  5. Is there an offsetting benefit that a ban would destroy? People like smoking cigarettes (and drinking alcohol), so a ban would not be effective. Taxing alcohol is one way to reduce demand, without driving the trade underground (e.g., Prohibition). Those who still support a ban on drugs (I don't) are ignoring this important behavioral consideration.
  6. Some bans or regulations are considered superior to taxes because the current behavior is so harmful. Thus, we have the ban on CFCs, murder, organ sales, etc. As you can see from this list, a ban can be effective, useful and harmful (respectively), so consistent rhetoric does not necessarily lead to consistent results.
  7. Also note the biggest difference between a tax and a ban -- a tax will increase price by a known amount but reduce behavior by an unknown amount. A ban will change behavior by a known about, but at a cost that is unknown.
  8. As part of this difference, note that taxes are much easier to understand and trace, while bans and other regulatory actions tend to be more opaque -- especially if there are "formulas" and "adjustments" that can be manipulated by lobbyists and legislators for personal gain.
  9. All of these points assume that taxes and regulations are implemented and collected in accordance with the law. If they are not (e.g., black markets, corruption, etc.), then it's unclear which is better. It's easier to hide unpaid taxes but black markets can exist for banned items or behavior.
Everything boils down to the costs and benefits of actions -- tax, regulate, "educate", or do nothing.

The reason that I advocate higher prices for water (and carbon taxes) is that the case for a ban on outdoor watering, long showers (SUVs), etc. is far weaker than the case for raising the price to reduce water (or carbon) consumption. Of course, these price increases must be large enough for people to notice!

The reason that I advocate per-capita rebates of high taxes is because I do not want to "feed the beast" of government with big revenues.

Bottom Line: Use the simplest instrument possible. Make sure everyone understands it. Implement the instrument transparently. Observe the change in behavior. Adjust.
Readers wanting a longer treatment of this topic should read this 66 page policy paper [PDF] on environmental taxes that is part of the Mirrlees Review of UK taxation policy. Here's part of the executive summary:
The case for using taxes, charges and emissions trading schemes (rather than regulation) to help
achieve environmental goals is primarily a matter of cost-efficiency. Economic instruments may be able to achieve a given level of environmental protection at lower cost, by providing incentives for polluters to choose the most cost-effective abatement mechanisms and by encouraging the greatest abatement effort from those polluters for whom it is least expensive. Economic instruments also provide ongoing incentives for innovation in pollution control; they may also be less prone to influence by polluters themselves than regulations negotiated case-by-case with individual firms. However, they are not a panacea. They can encourage costly avoidance activities, such as illegal waste dumping, and in some cases they may have significant distributional consequences, placing heavy burdens on the poor. They are most useful when wide-ranging changes in behaviour are needed across a large number of polluters – the costs of regulation in such cases are large, and the efficiency benefits of economic instruments are likely to be greater. Little will be gained, however, making the tax structure too sophisticated when the environmental costs are low.