An integrated tax-subsidy policy for carbon emission reduction

Abstract

A carbon tax is often cited by economists as an effective instrument to mitigate greenhouse gas emissions, but there is little political interest in the United States. In light of this political unpopularity, we develop and examine a net-revenue constrained carbon tax and subsidy program. The optimal revenue constrained tax and subsidy schedule based on our utility maximization model taxes energy sources with high emissions to energy price ratio, and subsidizes sources with low emissions to energy price ratios. This approach may be more palatable than a traditional carbon tax because it can change the relative price of low and high emissions energy sources while providing a mechanism to limit net tax increases and energy price increases. We find that a constrained tax/subsidy program provides welfare gains relative to a no-tax scenario. Welfare gains are estimated to be 1% and 36% of the welfare gains from a Pigouvian tax for the motor fuels industry and electric power industry, respectively. In contrast, subsidies for low-emitting energy sources funded from general tax funds rather than from high-emission energy tax revenues lead to welfare decreases substantially below our proposed tax/subsidy policy approach.

Introduction

A variety of public policy approaches for reducing greenhouse gas (GHG) emissions are being considered, developed, and adopted internationally. Carbon cap and trade programs have developed in the European Union (Convery, 2009), and the eastern United States (Aulisi et al., 2005). Similar programs are in development under the Western Climate Initiative, with the participation of several Western States and Canadian Provinces. A Low Carbon Fuel Standard, which focuses on carbon fuel intensity rather than total emissions from fuel production and consumption, was formally adopted by California in 2009. Carbon taxes have been adopted in various countries such as Sweden (Hammar and Jagers, 2007), and jurisdictions such as the Canadian Province of British Columbia (British Columbia Ministry of Small Business and Revenue, 2008). Renewable energy portfolio programs, renewable fuel standards, and renewable fuel subsidies have been adopted widely, with greenhouse gas mitigation among the stated justifications for adoption.

A carbon tax is often cited by economists as an effective instrument to address externalities associated with GHG emissions from energy production and use (Tol, 2005). Price incentives are usually found to provide higher welfare benefits than quantity instruments for dealing with global climate change problems (Fischer and Newell, 2008, Parry and Pizer, 2007, Newell and Pizer, 2003, Hoel and Karp, 2002, Pizer, 2002). However, despite the efforts of advocates for carbon taxes, considerable inertia currently exists in the U.S. at the national level for the adoption of a carbon cap and trade program, and many policymakers see the implementation of a carbon tax as politically infeasible in the United States, especially in the short run. There are several oft-stated political concerns about carbon taxes, but they include a reticence for adding “yet another” clearly identifiable tax on taxpayers, and the concern that they will lead to increases in energy prices (Metcalf, 2009).²

In contrast, subsidies for renewable energy, and in particular biofuel production, have been broadly applied by the federal and state governments in the United States.³ Historically, these subsidies for biofuels are funded from general tax funds, mostly from income and labor taxes, and not from fuel tax revenues.⁴ However, the result of such a subsidy in biofuels may lead to lower blended fuel prices than if no subsidy were provided, which may result in an increase in blended fuel consumption and conceivably even higher greenhouse gas emissions. Consequently, the cost of reducing GHG emissions with this instrument is large.

In this paper we examine a policy approach that could be interpreted as a compromise between the political concern over “more taxes” and higher consumer energy prices, but can reduce the costliness of traditional subsidies as a means to decrease GHG emissions. We develop a model of GHG-based subsidies for low carbon energy sources that are funded solely by carbon taxes on high carbon energy sources. This approach can alleviate the standard concerns mentioned above regarding both carbon taxes and traditional subsidies. The proposed instrument has three main political advantages over a traditional Pigouvian tax. First, it could be designed to be revenue neutral within the energy industry so as to alleviate concerns over additional taxes. Second, it would reduce the upward pressure on overall energy expenditures to varying degrees because it lowers some energy prices. Finally, it will tend to be a more cost-effective way of reducing GHG emission than a subsidy from general funds alone.

To formally examine the characteristics of this type of policy approach, we develop a model to maximize social welfare subject to an exogenously determined net tax revenue constraint. We derive the optimal market incentive instrument that internalizes the effects of greenhouse gas emissions given a net revenue constraint. This revenue constraint can be implemented such that there is no net increase in total energy industry tax revenues, where all revenues from carbon taxes are used to fund carbon-based energy subsidies. We investigate the cost and benefits of such a policy through simulations.

A standard Pigouvian tax schedule on GHGs is increasing in emissions, and is characterized by being separable such that the tax on one good associated with an externality is not dependent on the emissions of other goods in the economy (Sandmo, 1975, Kopczuk, 2003). In our setting, taxes on externalities are not separable because the magnitude and even the sign of a tax (a subsidy being a negative tax) on one form of energy depends, in part, on the relative emissions intensities of the other forms of energy. Three important factors determine the tax schedule if a revenue neutrality constraint is imposed within the energy sector: the net tax revenue target, the share of output to total industry output from all energy sources, and the relative marginal damages from pollution per unit of the relative price of the good.

There are existing policies with characteristics similar to our proposed policy approach. For example, Gainesville, Florida has implemented a program to impose a surcharge on consumption of electricity from the grid. The tax revenues from the consumption surcharge are then used to fund the purchase of electricity generated by privately owned solar panels at three times the standard rate thereby subsidizing solar energy sources (New York Times, 2009).

Florida's program is an example of a category of policy instruments called feebates.⁵ Feebates are also applied in the context of automotive markets and based on relative fuel economy of vehicles (Greene et al., 2005). Taxes are imposed on low mileage cars and tax rebate on high mileage cars. For a continuum of emission intensities across goods, a monotonic feebate schedule revolves around a pivot, at which point fees turn into rebates as emissions decline.⁶

To our knowledge none of the feebate structures actually implemented or examined in the related literature have been formally derived from a foundational optimization model.⁷ Such an approach provides insights into subtleties of revenue-constrained instruments. For example, based on our welfare maximization model, we arrive at a conditionally optimal feebate structure that depends on a set of relative market characteristics. We can derive a monotonically increasing tax/subsidy system based on emissions similar to the feebate schedules often described in the literature if output prices are identical across sectors. However, if output prices or production scale differ across sectors, the tax/subsidy schedule is not monotonic in emissions intensity.

The revenue-recycling literature also has some commonalities with our approach. For example, Parry (1998) and related research examines the trade-offs of taxing an environmental externality to reduce tax on labor. There are two key structural differences between the revenue-recycling literature and our model. First, this literature uses labor as the source of revenue when subsidies are provided for an environmentally beneficial good, and as the destination of revenues (as reductions in existing labor taxes) when a tax is imposed on an environmentally harmful good. In contrast, we construct an instrument in which the revenue sources and expenditures are all within the energy industry, and labor plays a more indirect role. Second, the revenue recycling literature relies on an exogenously determined tax revenue level that is satisfied by the sum of pollution and labor taxes and relies critically on the assumption that an optimal Pigouvian tax can be implemented in this environment. In our model, pollution tax revenues collected in the industry may be larger than the net tax revenue constraint imposed in the energy industry. This implies that the Pigouvian tax rate may not always be fully implementable. Feasible implementation of a pure Pigouvian tax is necessary for independent targeting of emissions with taxes (Kopczuk, 2003) and this condition is not met in our model. The consequence in our case is that the taxes are simultaneously dependent on all emissions rates within the sectors of polluting industries.

The rest of the article is organized as follows. Section 2 presents the model. Section 3 provides simulation methods and results for the electric power and the motor fuel industries. Section 4 concludes the article.