Pricing Carbon in the U.S.: A Model-Based Analysis of Power Sector Only Approaches

Abstract

In June 2010, as the prospects in the U.S. Senate for an economy-wide cap-and-trade bill dimmed, some proponents of climate policy began to push for a more limited-scope approach. One proposed way to limit the scope of the bill was to apply the cap-and-trade program only to the carbon dioxide (CO2) emissions from electricity generation. This paper uses an intertemporal computable general equilibrium (CGE) model of the world economy called G-Cubed to compare a power-sector-only climate policy with economy-wide measures that either place the same price on carbon or achieve the same cumulative emissions reduction as the program limited to the power sector. We first model a power-sector-only scenario (the Core Scenario) that broadly represents the emissions reduction ambition of a proposal offered by Senator Bingaman in July 2010. We calculate a linearly declining series of emissions caps for U.S. electricity generation from 2012 to 2030 that fall to 17 percent below 2005 levels in 2020 and 42 percent below 2005 levels in 2030. We calculate the Hotelling price path that achieves cumulative emissions equal to the sum of the caps, and we assume that all tax revenues are distributed lump sum to U.S. households. We then model a second scenario (the Same Price Scenario) in which the carbon price from the first scenario is applied to all fossil CO2 emissions in the US economy, not just CO2 from the power sector. Comparing this with the Core Scenario shows the incremental emissions reductions and incremental costs of expanding the policy from the power sector to the entire economy. The third scenario (the Same Emissions Scenario) calculates the Hotelling CO2 price path applied to all fossil CO2 that achieves the same cumulative reductions as the Core Scenario. Comparing it with the Core Scenario shows the consequences, for both carbon prices and economic costs, of using a narrow rather than a broad-based policy. To isolate the effects of U.S. policy, we assume the U.S. alone adopts these climate policies, with no comparable efforts abroad. We find both predictable and surprising results. As might be expected, the Core Scenario results in a carbon price in the power sector that is almost twice the economy-wide price that achieves the same cumulative emissions. In particular, the power-sector-only approach requires a price on CO2 that begins at $23 in 2012 and rises to $46 in 2030, whereas the economy-wide price begins at $13 in 2012 and rises to $25 in 2030. We find that a price on carbon only in the power-sector does not produce offsetting increases in emissions in other sectors. Rather, we find that carbon emissions outside the power sector fall slightly relative to baseline. This is because of the economic linkages between sectors and the consequences of higher electricity prices on economy-wide economic activity. More novel results involve the effect of the policies on trade, consumption, and measures of welfare. We find that all three policy scenarios reduce investment in the capital-intensive energy sector, which lowers imports of durable goods and strengthens the U.S. terms of trade. Thus we find important trade consequences for the U.S. of climate policy -- even the power-sector-only scenario, which one might think would have relatively low effects on terms of trade given that the U.S. electricity sector uses mostly non-traded fuels. In all scenarios, the relative price of consumption goods falls and U.S. GDP shifts toward consumption and away from investment. Together, the fall in the price of consumption goods and the lump sum rebate of carbon tax revenues are sufficient to raise real consumption relative to baseline. We attribute our findings to the structure of the G-Cubed Model, which has unique capabilities to model trade, currency, interest rates, and other financial market outcomes. We find that all of the policy scenarios produce an overall increase in consumption in the U.S. relative to baseline because the relatively lower price of traded goods more than offsets the declines in consumption due to higher embodied energy prices. Accordingly, we find that welfare measures based on consumption (such as equivalent variation) are higher in the policy scenarios than in the baseline by an amount that depends on how much the policy strengthens the U.S. dollar. In other words, in our model climate policy makes American consumers better off, not even counting its environmental benefits. In these simulations, the welfare improvement from climate policy is greatest in the Same Price Scenario, which is the most environmentally stringent and produces the strongest increase in the U.S. terms of trade. However, these results are sensitive to the trade elasticities in the model. If goods produced domestically are better substitutes for goods from abroad than the standard model assumes, then the policies would be less welfare-enhancing than these results suggest.