A Tax Program Versus a Cap and Trade Program with Early Credit Incentives

In order to discuss the differences between a tax and a cap and trade program to control undesirable emissions (pollutants), some background information is needed. If a policy is to be economically efficient, knowledge about both the marginal cost of abatement of the pollutant and the marginal benefit of reducing the pollutant must be known. Equating the two measures derives an economically efficient solution. Rarely is there an explicit representation made of the marginal benefit of reducing the pollutant, however, because of the difficulty of quantification. Most often, a target is specified and the question asked is, “What is the best way to meet the target?” This is referred to as being a cost-effective solution to meet the target. It may not yield the most efficient level of reduction by the best path; rather it identifies the best path to arrive at a predetermined target.

To be efficient, a tax must be imposed on the pollutant directly, not on the output from which it comes or some other measure used as a proxy for the pollutant. Consider a tax applied uniformly across all polluters. Since the tax affects the price of the given good or service, the quantity of the pollutant will be adjusted; however, the amount by which the quantity will be affected is unknown. In this case, it is not possible to know whether a given tax can result in a specified target.¹⁸⁶

With a tradable permit system (often referred to as a cap and trade program), the policy mechanism acts on determining a given quantity of adjustment and the value of the tradable permits (which affects the price of a given good) needed to achieve the target quantity adjustment. Here the value of the tax, and the ultimate effect on prices are subject to uncertainty.

If the policy maker is certain about the shape of the marginal abatement cost curve, the tax and the cap and trade system will yield equivalent cost-effective solutions. That is, if the policymaker knows how much it will cost to achieve each incremental reduction in a pollutant, the system can be designed so that a tax and a cap and trade system will yield identical results. The primary distinction between the two programs is not associated with economic efficiency (or economic cost-effectiveness) but rather with the distribution effects of the two programs. It is entirely possible that the two will yield equivalent costeffective solutions to reducing pollution, but the distribution of the impacts can vary considerably, depending on how the permits are allocated. If permits are auctioned, it is equivalent to a tax. If permits are allocated at no cost, the same cost-effective solution is achieved, but there is a redistribution of income among the permit holders and between permit holders and affected non-permit holders. The holders of permits possess a commodity that takes on a value, which can be submitted to the appropriate authority for each unit of pollution emitted or sold to a polluter who has a higher cost of abatement.

The coverage of the policy assumed in the above discussion is the entire economy. If certain segments of the economy are excluded, reducing overall coverage, those units that have the lowest cost abatement profiles may be excluded, making the solution more costly than would otherwise be possible. Another cost is the administration of the program. Monitoring and reporting costs can be significant and, moreover, difficult to quantify. Because a tax program is the easiest to implement, it has a lower administrative cost burden.

Comparing Computable General Equilibrium and Macroeconomic Models

To assess the macroeconomic impacts of energy and environmental policies, two types of frameworks are generally employed. While both frameworks model the macroeconomy in a consistent fashion and should provide similar results in the long run, their orientation and objectives are different. Each framework provides relevant insights. In order to fully comprehend the nature of macroeconomic impacts presented in this study, which employs the macroeconomic framework, it is useful to understand the main features of each framework.

There is generally no role for macroeconomic stabilization policies in these CGE models, because the economy is assumed to be at full employment. Because demands and supplies of goods and services are expressed in terms of relative prices (these are essentially barter models), there is no “medium of exchange” role for money, nominal interest rates, and consequently the central bank. The real interest rate is determined by productivity and thrift in the economy. In most of these models, the government sector is limited in scope and rarely presented in detail. Generally the assumption is made that the government produces a good (public good) that is desired by households and thus enters their utility functions and taxes are the prices they pay for them. The government must balance its budget over the forecast horizon, because forward-looking rational agents (both taxpayers and bondholders) would have it no other way and would adjust their saving and spending behaviors otherwise.

Essentially these models are meant for short- to medium-term forecasting, and policy simulations can be run effectively. The adjustment costs of the economy to near-term events and policies and the emergence of disequilibria like labor unemployment and non-optimal use of capital resources are explicitly accounted for. These models often have detailed financial sectors and permit assessment of fiscal and monetary policy measures in reaction to government economic policies. Given that they are based statistically on an observed structure of the economy in the past, all forecasts and policy simulations with these models assume that this structure will prevail in the future. For short-term projections this assumption is plausible, but it becomes increasingly tenuous as the projection period lengthens.

The Macroeconomic Activity Module (MAM) of NEMS, developed by Global Insight, Inc. (formerly DRI-WEFA), is essentially such a macroeconomic model. This model can address transition effects of energy policies and has a more detailed government sector and a well-defined set of fiscal policy levers to address alternative policies related to the collection and redistribution of revenues from a tax. It is regularly used for short- to medium-term forecasting and policy simulations at EIA. Here, we make a distinction between forecasting and policy simulations. By forecasting we mean the most likely outcome in the future (short to medium term) given past behavior, existing policies, and the likely course of exogenous variables. Policy simulation (or simulation) means the likely outcome in the short to medium term as a result of a deliberate change in policy, all else being equal. While the macroeconomic model is useful in providing valuable insights about the likely short- to medium-term direction of changes in the future as a result of energy-related policies and legislation, given the shortcomings noted, the longer-term results should be viewed with caution.

Much of economic decisionmaking is forward looking and is based on expectations of future prices, policies, and the economic environment. Given their empirical orientation, in most macroeconomic models consumer and business expectations of prices, inflation, interest rates, etc., are formulated and statistically estimated as depending upon past experiences (adaptive expectations). In the present context, this largely retrospective approach is not wholly satisfactory to the rational expectations school who would argue that the announcement of the energy legislation would significantly influence expectations of inflation or growth prior to any realized change in prices or spending . Thus, the actual disruptions from a policy, announced well in advance, would be smaller than predicted by the macroeconomic models, because consumers and businesses would already have adjusted their behavior before the policy took effect. Because CGE models are essentially static models where prices always clear markets, the modeling of expectations of prices is less problematic, and they would tend to show smaller impacts. Whether expectations are adaptive or perfect is subject to empirical evaluation, and there is no consensus on this.

Both macroeconomic models and CGEs have their relative strengths and weaknesses. For studying short- to medium-term impacts of energy, environmental, and other policy changes, macroeconomic models provide insights about the disruption and transition costs involved over time as the economy fully adjusts to them. CGE models provide a useful perspective about the resulting long-run state of the economy.

Notes