by Mario Rizzo
Recently, some economists have proposed a new application of the Pigovian tax idea. This is to correct the newly-discovered problem of internalities. An internality is a side-effect on the future selves of a given individual. So, for example, if an individual eats too much now his future self, perhaps a decade or two hence, will experience costs (damages) – not certainly, but with some probability. Internalities are generated by lack of willpower, myopia or other decisionmaking deficiencies on the part of present agents.
Now let us consider some of the problems in estimating the correct internality tax.
First, we need to know the probability of harm. This is inextricably related to the course of medical technology in the future. How easily (cheaply) certain health problems will be cured or alleviated decades from now is obviously very hard to predict.
Second, we need to know how far the individual’s present behavior is from the optimum. The answer is not obvious. People do care about their future selves to one degree or another. They do have self-regulatory or self-control capacities. In the limit where people care about their future selves as much as the current selves, they will already be at the optimum, with no need of a Pigovian tax all at. In fact, a tax here will cause people to overshoot the optimum.
In the non-limiting and probable case that people care but imperfectly about their future selves, simply taxing to the extent of the observed internality will lead to excessive restriction of their (eating) activity. The Pigovian needs to know the optimum amount of the activity before he can impose the appropriate tax — a tall order. (The tax must equal the difference between the social cost at the optimum and the appropriately adjusted full costs borne by the current self who cares about his future selves.)
Third, it is unclear what caring sufficiently about one’s future selves would be. In the interpersonal case, the Golden Rule may be appropriate: Care about others to the same degree as you care about yourself.
In the intertemporal context, however, we are quite accustomed to discounting future benefits and costs. So what is the “correct” discount rate? Without getting into details here, there is no unique normative rate.
In view of these problems of carrying through the Pigovian program in a serious way, the consequence of attempting to correct the “internality problem” through taxation will be the capture of the policy by those who have another agenda. The most obvious one is the old paternalist agenda: Do this because I think this is best for you, not because you, at some deep level of preference, think this is best. And, depending on the specific issue, other partial-interest groups will seek to control the policy.
It seems to me that Pigovian taxation is one of those ideas that looks far better in textbooks than it does in the real world. This is the problem with teaching it to undergraduates. In a textbook all of the necessary information is right there in front of you. Students then come away from their textbook learning with a fantasy of what is feasible in terms of economic policy. Unfortunately, it is a fantasy that also afflicts even those who really should know better. What is their excuse?
A little learning…
Further reading: John Nye: “The Pigou Problem.”
Mario Rizzo and Glen Whitman: “Knowledge Problem of New Paternalism.”
UPDATE: Greg Mankiw posts a statement by a fellow Pigou Club member on the carbon tax. Take a look at Nye’s criticisms of Pigovian taxes and then see how well they are answered in the post on Mankiw’s blog.