Our approach to EMF-12 differed in three important ways from the methodologies used by other participants. First, our results are based on a highly disaggregated intertemporal general equilibrium model of the United States. Using a disaggregated model allowed us to examine the effects of carbon taxes on narrow segments of the economy, such as particular industries or types of household. Second, all parameters in our model were obtained by econometric estimation using a data set spanning 39 years. Thus, the response of industries and consumers to changes in prices will be consistent with the historical record. Third, we model productivity growth at the industry level and allow it to be an endogenous function of relative prices. Allowing productivity growth to differ across industries permits the model to reflect a conspicuous feature of historical data. Together, these features allow us to shed light on several questions not addressed by most of the other participants. Some of our principle findings are as follows.
In the United States, the effects of a carbon tax will be very similar to the effects of a tax placed solely on coal. Of all fossil fuels, coal is the least expensive per unit of energy and produces the most carbon dioxide when burned. Thus, a tax levied on carbon emissions will raise the cost of coal-based energy far more in percentage terms than the price of energy derived from oil or natural gas. In response to this price change, the demand for coal will fall substantially. The demands for oil and natural gas will also decline, but by much smaller percentages. Almost all coal consumed in the U.S. is used to generate electric power. As the price of coal rises, electric utilities will convert some generating capacity to other fuels. However, substitution possibilities are fairly limited, particularly in the short run, so the tax will raise the price of electricity significantly. Consumers and firms will substitute other inputs for electricity, leading to a fall in electricity demand. Higher energy prices will lead to slower productivity growth, reduced capital formation, and a reallocation of labor to lower wage industries, all of which will cause gross national product to be lower than it would have been in the absence of the tax.
The tax rate needed to achieve a fixed absolute emissions target, such as maintaining emissions at 1990 levels, will depend on how fast emissions grow in the absence of the tax. Baseline emission growth, in turn, will depend on the rate of productivity growth, the rate of capital accumulation, the rate of growth of the labor force, any energy saving biases in technical change, and the path of world oil prices. More rapid economic growth will generally lead to higher baseline emissions and will thus require higher tax rates if emissions are to be held at a fixed absolute level. Moreover, deeper absolute cuts in emissions will require sharply increasing tax rates.
A carbon tax large enough to have much effect on emissions will raise tens to hundreds of billions of dollars annually. How this revenue is used will affect the overall economic burden of the tax. By using the revenue to lower highly distortionary taxes elsewhere in the economy, the tax could actually increase output and economic welfare. In the remainder of this chapter we present an overview of our model and discuss each of its findings in detail.