16 January 2013

Pigouvian taxes do NOT produce deadweight losses

For my final exam question, I asked students to compare the pros and cons of cap and trade and Pigouvian taxes for reducing the negative externality from pollution. Many students gave useful answers, but many also made two big mistakes.

The first was to claim that an advantage was their propensity to reduce consumption (an excellent example of rewriting the question as the answer :).

The other was simultaneously frustrating and enlightening to me. Many students claimed that Pigouvian taxes created a deadweight loss, i.e., the tax would reduce surplus-generating activity.* That statement is true, in general, of fiscal taxes designed to generate revenue (e.g., income, property or expenditure taxes) that the government would spend elsewhere,** but it's NOT true of Pigouvian taxes that are designed to reduce behavior that generates harm that is not reflected in the price of the good being taxed, e.g., taxes on cigarettes to pay for additional health costs or taxes on fuels to reduce and/or ameliorate the costs of pollution.***

The trouble that my students encountered -- and many teachers of economics fail to clarify -- is that fiscal taxes distort prices to generate revenue while Pigouvian taxes correct prices to affect behavior. (We explore the tension between these two goals in this paper on groundwater taxes.)

Bottom Line: We use the same word ("tax") to refer to two different policy instruments. Fiscal taxes generate revenue with some reduction in efficiency; Pigouvian taxes generate revenue as they improve efficiency. (That's why they are called win-win, but don't tell that to the people creating the pollution!)

* Miscalibrated taxes of all types create deadweight losses from being set at the wrong level, but those losses are not present in theory.

** Deadweight losses will be lowest when behavior changes by the smallest amount, i.e., it's best to tax the most inelastic behavior.

*** For more on why politicians prefer command and control over Pigouvian taxes, read this paper [pdf] by Buchanan and Tullock. Buchanan just died; he and Tullock are responsible for much of Public Choice theory, i.e., the idea that politicians and bureaucrats may serve themselves, not the public interest. Here's my review of their brilliant book on constitutions and laws.


  1. Maybe they are way ahead of their level and are talking about the tax interaction effect!

  2. Lonesome Cowboy20 January, 2013 19:31

    David: What about general sales taxes or vats? If you tax all retail products x percent (as part of raising money to finance publicly provided goods and services), does that create the usual distortion and welfare loss suggested by the typical partial equilibrium analysis?

  3. @LC -- a VAT *does* create deadweight losses, but there's an argument (valid) that it's better to tax expense than income, given the "benefits"of being productive to society. That may be hard to hold up when one man's expense is another man's income. The VAT is also favored b/c it's harder to dodge, but a tax on property (real estate) is even harder to dodge.

  4. It is preferable to not have an externality altogether, right? The reason is that tax must create a deadweight loss. Just because the benefit is greater than the cost, does not mean that the little triangle disappears. This is what I don't get.

    1. Yes, no externality is better. If it's "removed" via technology, etc., then there's no tax. If you're trying to reduce it (via tax), then you're REDUCING the loss/waste from "too much" activity by increasing the price to the efficient level (in theory), so that DWL will be zero when the tax is set correctly...


  5. I always describe dead weight losses as a result of producing too much and too little, even before introducing externalities or taxes, to avoid them attaching DWL to a particular market failure or policy instrument. I also tell them that it is hard for governments to set the Pigouvian tax (constant or stepwise) equal to the true marginal external cost (not constant and not certain) so that setting it too low doesn't eliminate the DWL from too much production and setting it too high creates a new DWL from loo little production. It is a good concept for principles students who don't get to take the deeper dive, but I don't spend a lot of time in my environmental class on Pigouvian taxes as many tax instruments are aimed at charging producers of multiple products for environmental services and resources they would otherwise not volunteer to pay for.


Spam will be deleted. Comments on older posts must be approved.
If you're having problems posting, email your comment to me