Pollution Taxes and the Double Dividend
A common idea is that a Pigouvian pollution tax has a “double dividend”: it reduces pollution, and it raises revenue (see Tullock 1967). The revenue is raised efficiently, because it allows other, distorting, taxes to be reduced. That is undoubtedly true. What has been in the literature since the 90’s is how high the pollution tax should be set. Should it be set to the marginal cost of pollution damage, or higher, or lower?
One’s first thought, and what seems to me must be right, is to set it exactly equal to the marginal cost. Inefficiency arises when a price is set below its marginal cost, and if pollution is one cost of production and is not incorporated in the price, surely that good is overproduced.
What has been argued, though, is that if there is already a distortion in the economy because of a tax on labor, the pollution tax should be set less than the marginal cost of the externality (see Bovenberg & de Mooij, 1994 and Fullerton & Metcalf, 1998).
Let’s think what happens in a model with two goods, A, which is clean, and B, which is dirty. Each requires 1 unit of labor, but in addition, B destroys X units of labor per unit of production via an externality. The government has levied a sales tax of T per unit on goods A and B (an equal tax like this makes it equivalent to an income tax of T). People choose the amount of labor, L, that they supply. Surely, in this case, the pollution tax should equal X.
The model commonly used is different. It has no destructive effect of pollution on labor. Instead, X is a bad that enters the utility function directly. It assumes that utility is separable in A, B, L, and X, since otherwise all sorts of effects can get gotten from cross-elasticities. Here, it is said, the pollution tax should be set to below the marginal cost of the externality.
Perhaps the problem may lie in how “marginal cost of the externality” is defined. It should be the amount people would be willing to pay to reduce the bad activity slightly. That amount will decline if the value of cash is less– which it will be if prices rise because of a pollution tax.
I don’t understand the idea properly, though, even after reading Bovenberg & de Mooij, 1994 and Fullerton & Metcalf, 1998). It seems odd that given the choice of a distorting tax on a polluting industry and distorting tax on all industries, we should choose to tax all industries. The paradox probably arises because because a big distorting tax on one product distorts more than two smaller distorting taxes that raise the same revenue from two products.
References:
A.C. Pigou, The Economics of Welfare (4th ed. 1932).
A. Lans Bovenberg & Ruud A. de Mooij, Environmental Levies and Distortionary Taxation, 84 Am. Econ. Rev. 1085 (1994).
Don Fullerton and Gilbert E. Metcalf, Chicago-Kent L. Rev. 73: 221 (1998) Environmental Taxes and the Double-Dividend Hypothesis: Did You Really Expect Something for Nothing?
Gordon Tullock, Excess Benefit, 3 Water Resources Res. 643, 643 (1967).