The Dynamic Effects of Taxes and Subsidies on Market Structure andEconomic Growth

It is becoming increasingly clear that the evaluation of policy measures that affect the growth performance of the economy requires adequate treatment of their effects on technological change. Environmental policy and regulations fit in this category. In this paper, I study the effects of taxes and subsidies on the growth path of the economy in a model where market structure is endogenous and jointly determined with the rate of technological change. An important reason for considering market structure is the argument proposed by Carlton and Loury that a Pigouvian production tax on polluting firms is not sufficient to achieve social efficiency when the number of firms is endogenous. I investigate the implications of this argument for an economy that exhibits endogenous growth. The analysis is positive. I ask questions like, What are the effects on growth and market structure of a production tax? What are the effects of the specific procedure with which the tax is introduced? For example, What are the effects of a tax that is announced today but implemented at a later date? Does this implementation lag matter, and if so, how? Similarly, What are the effects of a tax that is announced today but that will be introduced at a later date with some probability? In other words, What is the effect of uncertainty about the implementation of the tax? Finally, What is the effect of a tax that is announced and implemented today and then removed at a later date? Two sets of results emerge. First, one cannot unambiguously conclude that a production tax -- an instrument very similar to the carbon tax -- reduces environmental damages. If one focuses on steady state effects only, the tax leads to lower aggregate production, lower production per firm, more firms, and unchanged growth. If the damage function depends on both the level of activity per firm and the number of firms, the tax does not necessarily reduce damages. Second, the transitional dynamics of the model imply that the procedure of introduction of the tax matters. Moreover, the model exhibits hysterisis and temporary policies have permanent effects. A production tax that is introduced and then removed "traps" the economy in a steady state with lower growth, higher aggregate production, higher production per firm, and a larger number of firms. This policy is unambiguously bad: growth falls and both arguments in the environmental damage function increase. A tax that seems likely to be implemented at some future date -- so that agents assign a positive probability to the event -- but that then is not implemented has similar, although milder, effects. The analysis suggests that a government that intends to introduce the tax must pay attention to the timing and transparency of its actions. More generally, the analysis is suggests that much care must be taken in the specification of the dynamics of the model in policy simulations.