Introduction

Kaplow and Shavell (2002) offer a systematic critique of the view associated with Weitzman (1974) that the choice between corrective taxes and quantity targets is highly contingent in a world of uncertainty. They emphasize that this conventional understanding, heavily represented in texts and surveys on environmental economics as well as in more general textbooks, rests on a poorly appreciated and dubious set of assumptions.

First, in the Weitzman story, taxes are constrained to be linear, even though marginal harm is taken to be nonlinear (typically, rising in the quantity of emissions), so the Weitzman tax instrument violates the basic Pigouvian prescription that the marginal tax rate equal marginal harm. Second, taxes are taken to be fixed for all time, even though any error in setting the (constrained-to-be-linear) tax rate would become immediately apparent to the government. Relaxing either assumption restores the superiority of taxes over quantity regulation. Moreover, neither assumption was well motivated by Weitzman or in the subsequent literature. Nonlinear taxes are simple to state in many settings, and tax rates (really prices) of all sorts – whether set by government or the market – are routinely adjusted in response to changed circumstances.

Second, Kaplow and Shavell (2002) extend their analysis to permit schemes. On one hand, this instrument has the familiar and important advantage over command-and-control quantity regulation that a given quantity target is achieved at minimum cost. On the other hand, like familiar schemes, the quantity is set by fiat, once and for all.