Cover Image
close this bookEnvironment, Energy, and Economy: Strategies for Sustainability (UNU, 1997, 381 pages)
close this folderPart 3 - Energy-economy interactions in stabilizing CO2 emissions
View the document7. Modelling economically efficient abatement of greenhouse gases
View the document8. Macroeconomic costs and other side-effects of reducing CO2 emissions
View the document9. The effects of CO2 reduction policies on energy markets
Open this folder and view contentsComments on part 3
View the documentAppendix to part 3 examining the macroeconomic effects of curbing CO2 emissions with the Project LINK world econometric model

Appendix to part 3 examining the macroeconomic effects of curbing CO2 emissions with the Project LINK world econometric model

Hung-yi Li, Peter Pauly, and Kenneth G. Ruffing

1. Introduction

Amid the considerable uncertainty surrounding the issue of global warming, two scientific facts are undisputed. The first is that carbon dioxide gas has been accumulating in the earth's atmosphere over the past 100 years. The second is that the gas traps heat from the sun's energy when it is absorbed by the earth and then re-radiated.

The first area of uncertainty is how much the earth's climate would heat up in response to the further accumulation of carbon dioxide in the earth's atmosphere. Most studies based on computerized climate models have attempted to examine the impact of a doubling of atmospheric carbon dioxide from 1990 levels by the year 2100 (a plausible rate of increase in the absence of significant policy change). A United Nations scientific advisory committee concluded in 1990 and again in 1992 that the average global temperature by 2100 would increase within the range of 3-8°F, with a central value of 4.5° (IPCC, 1990; Houghton et al., 1992). A study by researchers at New York University and Lawrence Livermore National Laboratory of two ancient climates, one much colder and one much warmer than the current global climate, concluded that a doubling of atmospheric carbon dioxide was associated with an increase of 4°F in the average temperature of the atmosphere (Martin Hoffert and Curt Covery, cited by Stevens, 1993). Despite numerous critiques of such studies faulting the reliability of the data and the models used, most climate researchers believe that there is a greater than 50 per cent chance that the climate will warm up by at least 3.5° over the next century.

The second area of uncertainty concerns the effects of the anticipated warming in the absence of measures to prevent it. These effects bear mainly on the impact on agriculture and forestry of changes in weather patterns and on the impact of a potential rise of about 2 feet in the sealevel from melting polar ice. Studies of potential impacts on the United States based on the physical consequences associated with a doubling of CO2 equivalent concentrations of greenhouse gases have been done by the Environmental Protection Agency of the United States and used as the basis for alternative calculations by others. Recent estimates by Nordhaus (1993), Cline (1992), and Fankhauser (1992) range from 1 to 1.3 per cent of annual national income. Morgenstern (1991) provides arguments that support a figure closer to the "high" end of the range as the more plausible.

Schelling (1992) has pointed out that all the estimates of potential losses have relied on models of gradual change. He raises the possibility that "some atmospheric or oceanic circulatory systems may switch to alternative equilibria, producing regional changes that are both sudden and extreme" He admits that "insurance against catastrophes is... an argument for doing something expensive about greenhouse emissions. But to pay a couple percent of GNP as insurance premium, one would hope to know more about the risk to be averted" (1992: 8), and he calls for climate research to focus on extreme possibilities rather than to continue to refine median projections.

Representative of the "no regrets" approach of policy advocates for modest action to curb carbon dioxide and other greenhouse gas emissions is the position of a panel of the US National Research Council. This body concluded in 1991 that "despite the great uncertainties, greenhouse warming is a potential threat sufficient to justify action now" (cited by Stevens, 1993). The emerging consensus on the seriousness of the threat of global warming and on the wisdom of adopting early measures resulted in the United Nations Framework Convention on Climate Change by which developed countries agreed to limit their CO2 emissions in the year 2000 to their 1990 levels. The Convention entered into force on 21 March 1994.

The potential for emissions of carbon dioxide to change the global climate is prompting policy makers to examine the costs and benefits of a carbon tax. Research focuses on several aspects. Among the most important are: (a) the relation between the size of the tax and the quantity of carbon emissions that are abated and (b) the relation between the size of the tax and its effect on economic activity.

Understanding these aspects of a carbon tax is advanced considerably using computer simulation models (Weyant, 1993). Most often, the quantity of emissions abated by a tax is analysed using energy demand models, and the effect on economic activity is analysed using economic models. This schism is associated with the strengths and weaknesses of existing models (Beaver and Huntington, 1992). Energy demand models focus on the determinants of energy demand but many treat the level of economic activity as an exogenous variable. Conversely, economic models simulate economic activity endogenously but exogenize many of the important determinants of energy demand.

The use of energy demand models and economic models to analyse the quantity of emissions abated and the effect on economic activity forces the modellers to adopt several assumptions. Two of the most commonly used assumptions are that the pattern of international trade is constant and that energy prices are exogenous. These assumptions are required because most of the models cannot simulate endogenously how a carbon tax affects international trade or energy prices. One cause for the inability to simulate the change in trade is the geographic level of coverage and the level of aggregation. Many of the models used to analyse the effects of a carbon tax include one nation; therefore, these models cannot simulate the effects of trade. Conversely, models that include the entire world aggregate economic activity by countries into a few regions (Hoeller et al., 1990; Beaver and Huntington, 1992). Such aggregation prevents these models from simulating accurately the international pattern of trade. Finally, many models do not simulate energy supply, demand, or OPEC behaviour explicitly; therefore they cannot simulate energy prices endogenously.

The inability to simulate the pattern of international trade or the determinants of world energy markets limits the models' ability to evaluate a carbon tax. A carbon tax has significant short- and long-run effects on the economic system. In the short run, its price and budgetary effects are important for an economy's cyclical position; in that it is no different from any other fiscal policy. In the medium to long term, it alters the competitive position of nations, which changes their structure of imports and exports and the composition of value added across sectors and the product spectrum. A carbon tax also changes the supply of and demand for energy and the price for energy, which affects the incidence of the carbon tax. These changes in trade and world energy markets have an important effect on the quantity of emissions abated by a carbon tax and the effect of a tax on economic activity. Models that cannot simulate the changes in international trade and world energy markets caused by a carbon tax tend to underestimate the quantity of emissions abated by the tax and misrepresent the effect of the tax on economic activity. For example, models that cannot simulate the determination of world energy prices endogenously cannot simulate the reduction in world energy prices due to the reduction in demand generated by a carbon tax, which offsets some of the incentive to lower carbon emissions. Similarly, models that cannot simulate the pattern of international trade endogenously cannot simulate the gains and losses to nations that do and do not adopt a carbon tax.

Part of this paper describes how changes in the pattern of international trade and changes in energy markets affect the quantity of carbon abated by a carbon tax and the effect of a carbon tax on economic activity. These effects are simulated endogenously using the world econometric model of Project LINK (LINK) with the trace gas accounting system (TGAS) (Kaufmann et al., 1991). In this system, the combination of a short- to medium-term globally consistent and complete international macroeconomic model with modules that endogenously simulate the world oil market (OIL-LINK) and international energy demands, inter-fuel substitution, and emission levels (TGAS) provides a modelling framework that is particularly suited to address problems of short- to medium-term adjustments to a carbon tax.

2. Carbon taxes, energy prices, and international trade

Most of the scenarios used to analyse the effectiveness of carbon taxes, and indeed the simulations presented in the next section, assume that the tax is adopted worldwide and that energy prices are exogenous. If these assumptions are relaxed, the forecast for economic activity and carbon emissions may change significantly. If carbon taxes are not adopted worldwide, the carbon tax may cause significant changes in the pattern of internal trade. If energy prices are determined endogenously by economic conditions, a carbon tax may cause significant changes in the pre-tax price of energy. Both of these changes have an important effect on the quantity of carbon emissions that are abated by a carbon tax and the effect of a carbon tax on economic activity.

The failure to simulate the effect of a carbon tax on the international pattern of trade causes models to omit an important mechanism that will determine, in part, the quantity of emissions abated by a carbon tax and the effect of a carbon tax on economic activity. A carbon tax affects the international pattern of trade regardless of the level of international cooperation. If all nations impose the same carbon tax, the fractional increase in the price of energy to end-users will vary among nations because of existing differences in energy taxes or subsidies (Kaufmann, 1991). Nations that have the highest level of existing taxes experience the smallest percentage increase in prices. Under these conditions, the competitive balance shifts slightly in favour of those nations that experience the smallest percentage increase in prices. But the effects on trade that result when the carbon tax is imposed universally are small compared with the effects that occur if only some nations adopt a carbon tax. Without global participation, energy prices rise significantly in nations that adopt a carbon tax (participating nations) relative to nations that do not impose a carbon tax (non-participating nations). This differential rate of change causes the cost of production to decline in non-participating nations relative to that in participating nations. Ceteris paribus, the change in costs tips the competitive balance among nations in favour of nations that do not adopt the carbon tax.

The shift in competitiveness generated by a carbon tax changes the pattern of international trade via two mechanisms. The relative decline in the cost of production allows non-participating nations to capture market share from participating nations. The differences in end-user energy prices generated by the carbon tax also induce some multinational firms to relocate production in non-participating nations. Both of these changes increase the level of exports and economic activity in non-participating nations relative to nations that adopt a carbon tax.

The failure to simulate world energy markets endogenously also causes existing models to omit an important mechanism that determines, in part, the quantity of emissions abated by a carbon tax and the effect of a carbon tax on economic activity. A carbon tax reduces the demand and/or price for energy. The effect on demand and/or price depends on the degree to which energy producers, especially members of OPEC, recognize and react to the change in demand. If producers do not recognize and/or anticipate the reduction in demand correctly and adhere to their original schedule for adding capacity, reduced levels of demand create excess capacity. This overhang exerts downward pressure on the wellhead price of oil, which also depresses the price for other forms of energy. Conversely, OPEC may recognize and/or anticipate the reduction correctly and slow its schedule for adding capacity. Under these conditions, lower levels of demand do not create over-capacity and prices may proceed along the path that was envisioned for a world without a carbon tax (Kaufmann et al., 1994).

The changes in the pattern of international trade and world energy markets that are generated by a carbon tax have an important effect on the quantity of carbon abated by a carbon tax and the effect of the carbon tax on economic activity. The effect on the quantity of emissions abated is measured by "leakages" from the tax. Leakages are defined as an increase in emissions by non-participating nations beyond levels that they would have emitted had no carbon tax been imposed. By definition, the quantity of carbon abated by a tax shrinks as the quantity of leakages increases. Leakages are minimized implicitly by models that treat the pattern of international trade and energy prices as exogenous variables. Endogenizing these variables allows a model to measure the size of leakages that are generated by several mechanisms. One important cause of leakages is the change in the pattern of international trade. The increase in economic activity by non-participating nations that is caused by expanding market share and location decisions by multinational firms increases carbon emissions by non-participating nations relative to a world in which there is no carbon tax. Another important source of leakages is the potential for changes in world energy markets. If energy producers do not recognize the reduction in demand and lower capacity accordingly, the reduction in price diminishes the incentive to reduce energy use per unit output, which increases carbon emissions relative to a world in which there is no carbon tax.

The changes in trade and world energy markets have offsetting effects on the impact of a carbon tax on economic activity. The increase in economic activity by non-participating nations exacerbates the negative effect of a carbon tax on economic activity in participating nations because the increase in economic activity is accomplished, in part, by moving production from participating nations to non-participating nations. On the other hand, the potential for lower energy prices may reduce the negative effect of the carbon tax on economic activity. If a carbon tax reduces energy prices, the reduction slows the transfer of money from energy-consuming nations to energy-producing nations. Retarding this flow increases income in energy-consuming nations, which increases demand and production (Marquez and Pauly, 1984). Even though trade effects offset the energy price effects, there is no a priori reason to believe that they cancel each other so that the net effect is small. As a result, models that do not simulate patterns of international trade and the formation of energy prices endogenously either underestimate or overestimate the negative effect of a carbon tax on economic activity by participating nations. In addition to overestimating the quantity of emissions abated by a tax and misrepresenting the effect of a carbon tax on economic activity, the failure to represent the effect of a carbon tax on the pattern of international trade and world energy markets causes the models to ignore the potential for strategic behaviours regarding a carbon tax. The potential for leakages from the tax reduces the effectiveness of the tax, but this loss of effectiveness creates winners and losers. Nations with the smallest rise in end-user prices or that do not adopt a carbon tax may "win" because they may reap the benefits of lower emissions while enjoying increased levels of economic activity and emissions. Conversely, nations with the largest rise in end-user prices may "lose" by reducing emissions less than they anticipated and suffering economic losses greater than they anticipated. The effect of a carbon tax on the world energy markets also creates winners and losers. Energy-consuming nations will "win" by shifting a portion of abatement costs to energy-producing nations, which will "lose" by paying for abatements accomplished by consuming nations via lower energy prices and/or lower levels of exports.

3. A policy simulation

The modelling framework

The LINK model has been described elsewhere, including in a paper prepared for the UNU Programme on Global Change and Modelling in 1991 (Hickman and Ruffing, 1991). It is a multi-country model that links separate national econometric models by means of models of merchandise trade, exchange rates, and commodity prices. We use the LINK-TGAS modelling system to study how changes in the pattern of international trade and world energy markets affect the quantity of emissions abated and the economic impact of a carbon tax. A detailed description of the basic structure of the system can be found in Kaufmann et al. (1991). The LINK-TGAS system is suited for this task because, in addition to standard domestic macroeconomic feedbacks and the modelling of energy use and emissions, it simulates endogenously the international pattern of trade, disaggregated by commodities, and world energy prices. The pattern of international trade is simulated on the basis of bilateral trade matrices by commodity groupings; trade shares can respond to technological changes, to changes in structures and preferences, as well as to changes in competitiveness. The LINK-TGAS system uses a combination of assumptions about monetary and fiscal policy, OPEC behaviour, and OPEC capacity to generate a forecast for economic activity and the pattern of imports and exports for the 79 nations and regions in the Project LINK modelling system. The modelling system also generates a forecast for carbon emissions by the 14 nations for which TGAS models are available (Canada, France, Germany, Italy, Japan, the United Kingdom, the United States, Brazil, China, Hungary, India, Korea, Mexico, and Poland). With the exception of the states of the former Soviet Union, this subset represents the most important emitters of CO2. Finally, the modelling system also generates a forecast for energy prices. The effect of a carbon tax on the world energy markets is simulated endogenously and is based on the level of supply, demand, OPEC capacity, and OPEC behaviour (Kaufmann, 1995). Together, these variables constitute a baseline that can be compared with different scenarios to determine the effectiveness of a carbon tax.

The baseline scenario

One of the difficulties in constructing a baseline scenario is predicting CO2 emissions for the next decade in the absence of additional policy measures. There are several reasons for this. First, there is the question of the overall rate of GDP growth that might be expected over the eight years to the year 2000. The LINK reference scenario was 2.7 per cent per year (2.4 per cent for the decade as a whole). This is close to the baseline scenario described together with high and low variants in a recent study by the United Nations (UN, 1992) making use of the Global Input-Output Model (GIOM). Secondly, over periods of a decade or more, the sectoral composition of output can be expected to change both within and among countries. Thirdly, the energy intensity of production in each sector can also be expected to change as economic agents continue to adjust with a long response lag to past changes in energy pricing policies and other measures.

As a consequence of the interplay of these factors, the United Nations study concluded that carbon from fossil fuels emitted by the developed market economies would increase only slightly under the baseline scenario, cumulatively by only 8 per cent between 1990 and 2000 (0.8 per cent per year) but by 36 per cent under the high-growth scenario (3.1 per cent per year). In the high-growth scenario, emissions of carbon dioxide per unit of GDP are assumed to fall by 16 per cent in the developed market economies between 1990 and 2000. This implies an overall income elasticity of 1.0. Without the change assumed in the energy intensity per unit of output, the income elasticity would be nearly 1.5. There are other studies that suggest low rates of growth of CO2 emissions in the United States. Several widely quoted studies cited in Morgenstern (1991) suggest a rate of growth of only 0.9 per cent per year in the United States from 1987 to 2000 with current policy commitments. In the LINK baseline, overall CO2 emissions by the G7 countries are projected to increase by 3.3 per cent per year from 1993 to 2000 (cumulatively by 30 per cent). Since the aggregate GNP of the same countries is projected to increase by 2.7 per cent per year over the same period, the implied income elasticity is about 1.2.

A macroeconomic policy simulation using the world econometric model of Project LINK

In the LINK policy simulation prepared for this paper, a carbon emission tax was imposed on each G7 country. The tax was set at US$30 per metric ton of carbon equivalent (tC) in 1993 and increased to US$100/tC in 2000. There are considerable variations in the percentage reductions in CO2 emissions per country, ranging from a cumulative decrease of 3 per cent in the case of Japan to 11 per cent in the case of the United States. For the G7 countries as a group the reduction from the 2000 baseline figure is 8 per cent. This is not quite a third of the reduction that would be necessary to freeze emissions in 2000 to their 1990 levels. Yet the reduction in GDP would be about 2 per cent. Recently, Weyant (1993) reported that studies undertaken for the Energy Modeling Forum 12 suggest that the long-term costs of stabilizing global carbon emissions could be in the neighbourhood of 4 per cent of world GDP by the year 2100.

Several studies reviewed by Morgenstern (1991) estimate that a 10-20 per cent reduction from base levels would require a carbon tax of US$10-75/tC. Some more recent studies suggest levels of carbon taxes toward the higher end of this range. The Edmonds-Reilly model, for example, calculated that a 14 per cent reduction in CO2 emissions from baseline levels in OECD countries in 2005 would require a carbon tax of US$55-56/tC in that year (Barns et al., 1992). The Global 2100 model developed by Manne and Richels calculated that an 11 per cent reduction in the baseline CO2 emissions of OECD countries in 2000 would require carbon taxes in the range of US$60-70/tC in the terminal year (Manne, 1992). The LINK calculations appear to be on the high side as far as the level of tax required to induce significant CO2 emissions reductions is concerned. However, it should be borne in mind that most other studies employ a general equilibrium framework to calculate the fully adjusted response to a hypothetical tax, as if there were no adjustment lag. The smaller response suggested by a dynamic econometric model suggests that the task of reducing CO2 emissions may be greater than commonly thought.

As may be seen from table 3A.1, the macroeconomic effects are fairly strong. A carbon tax that rises to US$100/tC by 2000 would leave GDP in G7 countries 2 per cent lower than in the baseline. Growth elsewhere would also be reduced but by less because in this scenario no direct taxes were imposed on these countries. The fact that inflation increases by an average of only about 0.35 of a percentage point over a period to 2000, while unemployment increases by about 0.6 of a percentage point in 2000, suggests that there would be scope for offsetting reductions in other taxes to mitigate the negative impact of a carbon tax on output and employment.

Table 3A.1 The LINK model scenario (impact expressed as percentage deviation from baseline)


GDP in world

GDP in G7 countries

World trade volume

Inflation in developed market economies

Unemployment in developed market economies

CO2 emissions in G7 countries

1993

-019

-0.21

-0.24

0.69

0.09

- 0.7

1994

-0.67

-0.87

-0.55

0.28

0.54

-3.7

2000

-1.39

-2.04

-0.76

0.47

0.59

-8.1

Note: A carbon tax of US$30 per metric ton of carbon equivalent is imposed in 1993 rising to US$100 in 2000.

The role of oil prices

The scenario described in the previous section compares the effects of a unilateral carbon tax on economic activity, carbon emissions, and energy prices. These results are generated via an exogenous tax assumption under exogenous energy prices, and we can identify the ways in which changes in the pattern of international trade and energy prices that are generated by a carbon tax affect the quantity of carbon abated by the tax and the effect of the tax on economic activity. The comparisons are also used to identify the gains and losses that are associated with adopting a carbon tax and the cost of emissions reductions borne by emitters as opposed to energy producers.

The following scenario uses all of the assumptions about monetary and fiscal policy from the reference scenario, but now is based upon a uniform US$40 carbon tax with an endogenous oil price response. This change in the size of the shock is of little consequence for the multiplier properties of the system. This scenario calculates the first-purchase price of energy endogenously using the assumptions about OPEC behaviour and capacity used in the reference scenario. The price of crude oil in this scenario is different from the price calculated in the reference scenario because the carbon tax changes the demand for oil. Demand is changed because the end-user price of energy increases in nations that impose a carbon tax and the increase in end-user prices changes the pattern of international trade, the level of economic activity, and the quantity of energy consumed per unit of economic activity.

In this scenario, all variables are solved endogenously. Because it uses the same assumptions about OPEC behaviour and additions to OPEC capacity, this scenario represents the effect of a carbon tax on world energy markets if OPEC ignores the effect of the carbon tax on oil demand. Under these conditions, oil prices fall owing to lower levels of capacity utilization. This reduction in energy prices allows the scenario to identify how changes in world energy markets affect the quantity of emissions abated by the tax and how changes in world energy markets affect the impact of the carbon tax on economic activity. The effect of a carbon tax on world energy markets and the impact of these effects on the quantity of carbon abated and the level of economic activity can be gleaned from this scenario in three steps. In the first step, the results from this scenario are used to determine the effect of a carbon tax on energy prices by comparing the price forecasts generated by this scenario with those of the reference scenario.

In the second step, the results of this scenario are used to determine the quantity of carbon abated by a tax and how changes in energy prices affect this quantity. The total amount of carbon abated by the tax is determined by the difference in total carbon emissions between the reference scenario and this scenario. The leakage from the tax through world energy markets is determined by subtracting the total amount of emission reductions forecast by this scenario from the total amount of emission reductions forecast by the reference scenario. A positive value indicates a leakage from the tax. A similar process is used to determine the effect of a carbon tax on economic activity. The total effect of a carbon tax on economic activity is determined by comparing the level of economic activity forecast by the reference scenario with the level forecast by this scenario. The effect of changes in world energy markets on the economic impact of a carbon tax is determined by subtracting the level of economic activity forecast by the reference scenario from the level of economic activity forecast by this scenario. A positive value indicates that changes transmitted through world energy markets reduce the negative effects of a carbon tax on economic activity.

The quantity of emissions abated by the tax depends on changes that occur in both participating and non-participating nations. In participating nations, the carbon tax reduces the quantity of carbon emitted by reducing emissions per unit output. This reduction may be reinforced or offset by changes in the level of economic activity by participating nations and by changes in the first-purchase price of energy. In non-participating nations, the carbon tax changes the quantity of emissions abated by changing the level of economic activity and by reducing the first-purchase price of energy. The sum of these effects determines the quantity of emissions abated in toto.

Endogenizing the oil price has several effects. The relative increase in G7 activity associated with a carbon tax per se would increase the oil price, but the tax provides incentives to substitute away from oil. On balance, this depresses the oil price. By the year 2000, the price path lies about US$3 per barrel below the exogenous price path of the original scenario. This exerts a slightly positive impact on the participating countries in terms of lower imported inflation, even though it depresses activity in non-participating energy-producing countries (mostly developing countries). In addition, the implied change in the relative price of energy induces an increase in the use of carbon energy sources and thus reduces the efficiency of the tax.

The quantity of emissions abated by participating countries is given by the level of emissions forecast by the endogenous scenario relative to the level of emissions forecast by the reference scenario. The size of these reductions varies from 1.5 to 4.7 per cent in 2000. The largest reductions occur in the United States and the smallest reductions occur in Japan and Germany. This differential rate of reduction is caused by differences in the fuel mix, the mix of economic activities, the size of the elasticities estimated from the historical record for individual nations, and differences in the percentage increase in end-user prices generated by the carbon tax. This last effect seems to be the most significant determinant of the reduction in emissions. For example, the United States has the largest percentage increase in end-user prices in 2000 whereas Japan and Germany have the smallest percentage increase in end-user prices.

The direction and size of the change in emissions depend on changes in economic activity caused by the tax, changes in world energy markets that reduce the first-purchase price of energy, and changes in economic activity caused by changes in world energy markets. The results of the three scenarios can be used to identify the change in emissions associated with each mechanism. Changes in the world energy markets that reduce the first-purchase price of energy are the second mechanism that changes the quantity of carbon abated by the tax. The effect of reductions in the first-purchase price of energy on the quantity of carbon abated is evaluated by calculating the percentage change in the emissions/GDP ratio forecast by the reference scenario relative to the ratio forecast by the endogenous scenario. An increase in the carbon emissions/real GDP ratio forecast by the latter relative to carbon emissions/real GDP ratio forecast in the former indicates that the reduction in first-purchase prices reduces the incentive to reduce energy use, which generates a leakage from the carbon tax in participating nations.

The third mechanism for leakages in participating nations is a change in economic activity caused by changes in world energy markets. The reduction in the first-purchase price of coal, oil, and natural gas changes economic activity (including imports and exports) and these changes affect emissions. The change in emissions caused by this effect is evaluated by calculating the percentage change in economic activity forecast by the endogenous scenario relative to the change forecast by the reference scenario.

Leakages in non-participating nations depend on the same mechanisms that cause leakages in participating nations: changes in economic activity caused directly by the carbon tax, changes in the first-purchase price of energy that cause changes in the quantity of carbon emitted per unit of economic activity, and changes in economic activity caused by changes in world energy markets.

The effect of a carbon tax on the international pattern of trade and world energy markets creates the potential for an array of strategic behaviours. Changes in trade make non-participating nations better off and their gains are reinforced by the negative effects of the tax on participating nations. Changes in world energy markets redistribute the costs of the tax such that a carbon tax reduces the economic well-being of energy-producing nations and, in some cases, energy-consuming nations are able to reduce the negative impacts of the tax by shifting a significant portion of the costs to energy-producing nations.

A carbon tax presents energy producers with a range of strategies that vary between two extremes. At one extreme, producers can ignore the effect of the tax on demand and add capacity at the rate that would have prevailed in a world without a tax. Under these conditions capacity utilization drops and the resultant overhang depresses the price of oil and other forms of energy. At the other extreme, producers can anticipate perfectly the effects of the tax on demand and retard additions to capacity so that utilization rates and real prices evolve along the same path that would have prevailed in a world without a tax. Under these conditions, the demand for oil and other forms of energy drops. Regardless of the strategy chosen. the economic well-being of energy producers is reduced. These issues are explored in detail in Kaufmann et al. (1993).

4. Conclusions

This study has been concerned with unilateral (non-global) carbon tax policies, and with an evaluation of the effects of the response on the world oil market on the efficiency of such a policy, i.e. the extent of oil-related carbon leakages and the short-run adjustment costs of such a policy. The results are based on simulations with the LINK-TGAS system, which combines a global macroeconometric model with modules for energy use and emissions analysis and energy price determination. The results are, of course, preliminary, but they con firm and complement studies of a longer-term nature (McKibbin and Wilcoxen, 1992; Piggott et al., 1993).

Among the findings of the study are the following:

· A unilateral (G7) carbon tax induces reductions in emissions in these countries; the effect of such a policy on non-OECD emitters is diverse, ranging from a slight reduction to increases up to 15 per cent.

· The results are critically dependent upon the responses of international energy prices, in particular the world oil price. A carbon tax in the G7 countries induces a medium-term reduction in the real price of crude oil by approximately US$2 per barrel.

· The oil price reduction generates a positive stimulus in G7 countries that significantly tends to reduce the short- to medium-term GNP loss to G7 countries.

· A unilateral G7 policy generates a trade balance improvement in these countries, at the expense of non-participating countries; oil-exporters among the G7 (Canada and the United Kingdom) experience deteriorations in the trade balance.

· The results are sensitive to assumptions about the strategic behaviour of energy producers, in particular of OPEC.

It is, of course, possible to reduce or even eliminate the negative activity effects of a carbon tax through the recycling of revenues and parallel stabilizing policies, such as a mild monetary expansion. In general, such policies increase the short-run effectiveness of a tax, but at a price of increased inflationary pressure. The results obtained in single-country analyses (Shackleton et al., 1992) have to be confirmed in future multi-country simulations. The present results already point towards a substantial potential for improvement in the efficiency of international carbon taxation through international cooperation.

References

Barns, D. W., J. A. Edmonds, and J. M. Reilly. 1992. Use of the Edmonds Reilly Model to Model Energy-Related Greenhouse Gas Emission. Paris: OECD, Economics Department Working Papers No. 113.

Beaver, R. and H. Huntington. 1992. "A comparison of aggregate energy demand models for global warming policy analyses." Energy Policy 20(6): 569-574.

Cline, W. R. 1992. Global Warming: The Economic Stakes. Washington, D.C.: Institute for International Economics.

Fankhauser, S. 1992. "Global warming damage costs: Some monetary estimates." London: Centre for Social and Economic Research on the Global Environment, mimeo, August.

Hickman, B. G. and K. G. Ruffing. 1991. "Project Link: Past, present and future." UNU Conference on Global Change and Modelling, November, Tokyo.

Hoeller. P.A., A. Dean. and J. Nicolaisen. 1990. A Survey A Survey of the Costs of Reducing Greenhouse Gas Emissions. Paris: OECD, Working Paper.

Houghton, J. T.. B. A. Callander, and S. K. Varney (eds.). 1992. Climate Change 1992: The Supplementary Report to the IPCC Scientific Assessment. Cambridge: Cambridge University Press.

IPCC (International Panel on Climate Change). 1990. Policymakers, Summaries of the Scientific Assessment of Climate Change, Report prepared for IPCC by Working Croup 1. Geneva, June.

Kaufmann, R. K. 1991. "Limits on the economic effectiveness of a carbon tax." Energy Journal 12(4): 139-144.

Kaufmann, R. K. 1995. "A model of the world oil market for Project LINK: Integrating economics, geology, and politics." Economic Modelling 12(2): 165-178.

Kaufmann, R. K., W. Gruen, and R. Montesi. 1994. "Drilling rates and expected oil prices: The own price elasticity of US oil supply." Energy Resources 16: 39-58.

Kaufmann. R. K., H. Y. Li, P. Pauly, and L. H. Thompson. 1991. "Global Macroeconomic Effects of Carbon Taxes: A Feasibility Study." Prepared for the US Environmental Protection Agency.

Kaufmann, R. K., H. Y. Li, P. Pauly, K. Seto, and J. Sweitzer. 1993. "A preliminary analysis of international aspects of a carbon tax: The role of the world oil market." Boston University and University of Toronto, mimeo, December.

McKibbin, W. J. and P. J. Wilcoxen. 1992. G-CUBED: A Dynamic Multi-Sector General Equilibrium Growth Model of the Global Economy: Quantifying the Costs of Curbing CO2 Emissions. Washington, D.C.: Brookings Institution, Discussion Paper in International Economics No. 98, September.

Manne, A. S. 1992. Global 2100: Alternative Scenarios for Reducing Carbon Emissions. Paris: OECD, Economies Department Working Papers No. 111.

Marquez, J. and P. Pauly. 1984. "OPEC's pricing policy and the international transmission of oil price effects." Energy Economics 6(4): 267-275.

Morgenstern, R. D. 1991. "Towards a comprehensive approach to global climate change mitigation." American Economic Review 81(2): 140-145.

Nordhaus, W. D. 1993. "Reflections on the economies of climate change." Journal of Economic Perspectives 7(4): 11-26.

Piggott, J., J. Whalley, and R. Wigle. 1993. "How large are the incentives to join subglobal carbon-reduction initiatives." Journal of Policy Modeling 15: 473-490.

Schelling, T. C. 1992. "Some economics of global warming." American Economic Review 82(1): 1-14.

Shackleton, R., M. Shelby, A. Cristofaro, R. Brinner, J. Yanchar, L. Goulder, D. Jorgenson, P. Wilcoxen, and P. Pauly. 1992. "The efficiency value of carbon tax revenues." Washington, D.C.: US Environmental Protection Agency, Working Paper, mimeo, May.

Stevens, W. K. 1993. The New York Times, Tuesday, 14 September.

UN. 1992. "Structural change in the world economy: Implications for energy use and air emissions. Report of the Secretary-General." A/47/388, mimeo, 7 October.

Weyant, J. 1993. "Cost of reducing global carbon emissions." Journal of Economic Perspectives 7(4): 27-46.