Friday, February 7, 2014
Victor Fleischer (San Diego) presented The Inferiority of Pigouvian Taxes at UCLA yesterday as part of its Tax Policy and Public Finance Colloquium hosted by Jason Oh, Kirk Stark, and Alexander Wu:
Pigouvian (or "corrective") taxes have become the favored policy instrument to address activities that cause negative externalities. There is considerable academic support for Pigouvian taxes on a wide range of products and activities, including carbon, gasoline, fat, high fructose corn syrup, financial transactions, executive pay, excessive zoning, and SUVs. Economists of all political stripes are therefore mystified by our politicians’ collective inability to see the merits of using Pigouvian taxes more frequently to address serious social harms.
Like the nearsighted Mr. Magoo, the politicians have stumbled to the right place. The problem with Pigouvian taxes is that the theoretical advantages of tax as an instrument of social engineering do not survive the journey from academic theory into actionable policy. The problem is that we typically cannot tax and thereby reduce the harmful externality directly, but rather must find a proxy activity to tax that we believe, if reduced, will also reduce harm. In the case of a carbon tax, there is a proxy (carbon production) for the harm (carbon emission) that fits well. But when there is great variation in the amount of harm caused by different individuals or firms that engage in the activity, a uniform excise tax harm will not achieve the desired marginal effect. This heterogeneity or “targeting” problem has not been adequately addressed in the literature.
The targeting problem results from how our political institutions work rather than from the economics of tax instruments as such. In a world with costless information and perfect political institutions working in the public interest, Pigouvian taxes would not be uniform. They could be varied with a level of precision that allows the benefits of reducing externalities to more than offset any deadweight losses associated with poor targeting. Except in the case where the variation among externality producers is related to income, however, the taxing authorities will not be well positioned to design or implement a tax, and other policy instruments (regulation, government spending, behavioral nudges, education, private insurance) are likely to achieve better results at lower cost.