A UK carbon/energy tax

A UK carbon/energy tax The macroeconomic effects Terry Barker, Susan Baylis and Peter Madsen

A carbon~energy tax has been proposed by the European Commission to start as US$3 per barrel oil equivalent in 1993, rising to US$10 in 2000. The effects of this tax on the UK economy can be assessed by using a large-scale energyenvironment-economy model in which energy demand is treated by means o f aggregate equations for energy users and fuel share equations for electricity and each of the main fossil fuels. The electricity supply industry is represented in the model by simulating the characteristics of each existing generating station and of a range of new station types, and by allowing the model to choose the fuels used from relative fuel prices and prices o f capital inputs. This approach contrasts with the multimodel approach used in most studies of the effects of the tax. The tax is introduced as a tax on imports and domestic output of fossil fuels in proportion to their carbon and energy content, with the tax being passed on through the economic system depending on the use of the fuels. Thus the electricity industry is faced with higher prices for coal, gas and fuel oil and chooses to burn the different fuels according to their relative prices. Four scenarios are developed, which vary according to the way the tax revenues are recycled through reductions in V A T or in income tax rates and whether the tax is introduced throughout the European C o m m u n i t y or throughout the OECD. I f the tax is European, it is assumed that the energy intensive industries will be exempt. The outcome is that the tax is sufficient to stabilize C02 emissions over the period 19902005, at 12% below base levels by the end of the Terry Barker is at the Department of Applied Economics, University of Cambridge, Sidgwick Avenue, Cambridge CB3 9DE, UK; Susan Baylis and Peter Madsen are at Cambridge Econometrics Limited, Covent Garden, Cambridge CB1 2HQ, UK.

296

period. There is an overall saving o f energy compared to base and a further switch to gas in electricity generation. The effects on the macroeconomy are rather small and as the revenues are recycled, rather than saved by the government, G D P growth is likely to increase under the tax by some 0.2% above base. This confirms other studies in which revenues are recycled, but contradicts many results from studies in which revenue recycling is ignored or incomplete. The effects on the energy intensive industries are expected to be small in all the scenarios, but 14% of tax revenues are lost if the industries are exempt. Keywords: Carbon tax; Energy tax; UK energy policy This paper explores the macroeconomic effects of the carbon/energy tax proposed by the Commission of the European Communities (CEC). ~ It sets out an interpretation of how such a tax may be implemented in the UK and what its effects will be on CO2 emissions, economic growth, industrial structure and energy consumption under different assumptions about revenue recycling and international competitiveness. It assesses the net costs or benefits to the economy over a 15 year period of cutting back CO2 emissions. The study follows a number of others on the effects of variants of the tax on the economies of the European Community (EC) and the UK (Barker and Lewney, Capros et al, DRI, Karadeloglou, and Standaert). 2 In comparison with these earlier studies, it reports the results of the tax using a single model of the UK economy in which the detailed treatment of energy demand and supply are part of the model, rather than being the result of running two or more models in sequence or simultaneously. In contrast with most of these other studies, it examines the effects of recycling revenues through reductions in VAT, which offset any general infla0301-4215/93/030296-13~) 1993 Butterworth-HeinemannLtd

A U K carbon~energy tax

tionary effects which might otherwise lead to a reduction in GDP. In addition, the approach allows for the direct modelling of the proposed exemptions of energy intensive industries from liability for the tax. The paper begins by describing the modelling approach undertaken and compares it with other approaches recently used to analyse the economic impact of carbon/energy taxes. It then reports on the treatment of the proposed carbon/energy tax and other assumptions made for the modelling exercise. An assessment of the effect of the tax in reducing CO2 emissions is made, leading on to a discussion of the macroeconomic implications for the UK, as well as the impact on selected industries and the energy sector. Some conclusions from this study are given at the end.

Modelling approaches The Cambridge Multisectoral Dynamic Model of the British economy (MDM) 3 was originally designed to focus on the analysis and forecasting of changes in economic structure. Individual industries are the basis for many of the economic relationships in the model. This allows the simulation of policies at the level of each industry, and is what distinguishes a multisectoral model from a macroeconomic model, the latter often treating the whole economy as a productive unit. Industries are the behavioural agents in the model for industrial investment, employment, prices, wages and intermediate demand. Recent developments in the treatment of energy use in the model, with the inclusion of CO2 and other emissions from the burning of fossil fuels, enable it to be used to analyse the implications for energy use, the environment and the economy of carbon/energy taxes. The modelling reported here examines the specific version of the carbon/energy tax first proposed by the CEC in 1991 and differs from that published previously (Barker and Lewney, and Barker, Baylis and Lewney),4 which assumed a carbon tax as an escalating ad valorem tax on energy consumption, allocated across the different energy carriers according to their carbon content.

An energy submodel The development of an energy submodel has allowed a detailed analysis of the demand for energy and the substitution between fuels following the imposition of a carbon/energy tax. The projection of fuel use, distinguished by user and type of fuel, is then available to calculate emissions of CO2, allow-

ENERGY POLICY March 1993

ing for different qualities and uses of fuels. The solution for the economic variables yields changes in economic activity and general price levels to the energy submodel. The submodel first calculates total energy demand by sector and then the share of this energy demand met by each fuel type. These results can be expressed as changes in the input-output coefficients for the electricity and other industries, so providing a feedback to the main model.

A submodel of the electricity supply industry A simple model of the plant profile of the UK electricity supply industry (ESI) has been constructed for projecting future fuel inputs to the electricity industry. The main inputs from MDM to the ESI model are the projected growth in electricity output and the relative fuel prices given the assumed level of carbon/energy tax. The ESI model then produces fuel input requirements for the electricity industry ie the amount of each type of fuel used in million therms: This is determined by the existing and expected characteristics of the generating stations, including their expected lifetime, their efficiency and the type of fuel they burn. It is a simulation rather than an optimizing model, in which assumptions about, for example, nuclear plant availability can be made, and in which a set of relative fuel prices is only one factor in determining fuel use.

Differences of approach Macroeconomic models vary greatly in the amount of sectoral detail they include; most ignore it because they are mainly interested in aggregate output. There is usually no scope for predicting major substitutions in resource inputs and hence sectoral outputs. The HERMES set of models, used in the studies of Standaert and Karadeloglou,~ have a detailed representation of the supply side of the economy emphasizing the substitution possibilities between factors. Aggregate energy is treated as a separate factor input into the production branches which therefore allows analysis of specific sectoral implications from an energy tax. 6 However, for most countries, there is no breakdown of aggregate energy into fuel types, and it is measured in constant prices. 7 The model therefore does not enable different fuels to be taxed at different rates and the possibilities for fuel substitution are ignored. In order to represent the energy sector properly, HERMES has had to be linked to MIDAS, an

297

A UK carbon~energy tax

energy model. 'Operational multimodel systems' is the term used to describe the systems of models at present being used to analyse the effects of carbon/ energy taxes and related policies at the European level. These include the HERMES-MIDAS system of models and the DRI system of macroeconomic, industrial and energy models used in the study commissioned by the CEC. 8 MIDAS is an energy model which takes certain key economic measures as given and models energy demand and supply in great detail and computes a set of energy prices and energy balances. 9 The different energy technologies in MIDAS are calibrated econometrically. Using MIDAS on its own to analyse the effects of carbon/energy taxes reduces the feedback between the energy sector and the rest of the economy. The linkage of MIDAS with the macrosectoral model HERMES described above has improved this aspect. However, even if the consistency problem in linkage can be solved by successive iterative solutions of the component models, as reported in the DRI study, there remains a more basic problem with the multimodel approach that attempts to combine macroeconomic models with detailed industry or energy models. This problem is that the system cannot adequately tackle the simulation of policies intended to take effect at the level of individual economic units. Normally these systems are first solved at the macroeconomic level; the results for the macroeconomic variables are then disaggregated by the industry model. However, if the policy is directed at the level of industrial variables, it is very difficult (without substantial intervention by the model operator) to ensure that the implicit results for macroeconomic variables from the industry model are consistent with the explicit results from the macro model. It is, for example, very difficult to use a macro-industry two-model system to simulate the effect of exempting selective energy intensive industries from the carbon/energy tax.

Assumptions The proposed carbon~energy tax The CEC proposal is that member states of the EC introduce a carbon/energy tax of US$3 per barrel oil equivalent in 1993, rising by US$1 a year to US$10 per barrel in 2000. It is envisaged that the tax will be indexed in line with inflation from 1993 ie that it will be a real tax in 1993 prices; it is assumed that after the year 2000 the tax rate will stay at US$10 per barrel in 1993 prices. The tax rates are intended to be allocated across fuel consumption, 50% in

298

accordance with calorific value and 50% in accordance with carbon content. In order to model the effects of the CEC proposed rates of tax, they have to be converted into specific duties on oil, coal, gas and nuclear electricity, as shown in Table 1.1° The energy tax component is estimated by noting that there are approximately 60 therms per barrel of oil and by using the thermal equivalent to calculate the duties implied for coal, gas and nuclear/large-scale hydroelectricity. The carbon component is derived from the carbon contents of the different fuels, starting with the estimated emissions of CO2 in terms of weight of carbon in oil. The combined rates are then converted into sterling at the fixed exchange rate of US$1.75 to the pound sterling and further converted into specific duties in £1 per therm of each fuel. The duties are assumed to be levied on the net available use of energy, starting with the estimated net use for 1991 taken from the Department of Energy Digest of UK Energy Statistics, 1992 edition, 11 and incrementing this net use through the projection period by the forecast increase in demand for the fuels in the UK market (ie UK output plus UK imports minus UK exports).

A European versus an OECD tax Two sets of scenarios have been considered in the analysis: the first set assesses the tax as proposed and assumes that it is not taken up by OECD members outside the EC; the second assumes that the tax applies across the OECD area. In the case of a EC only tax, the announced proposals provide for exemptions or exclusions for energy intensive industries, such as iron and steel, non-ferrous metals, chemicals, cement, glass and other building materials and paper and board. The exclusions are intended to protect the price competitiveness of these industries in international trade. In the modelling, it has been assumed that these industries would initially incur the tax in purchasing fuels, but are rebated by the full amount of levy they have paid in their fuel bills.12 In addition the CEC has suggested that the excluded industries would be expected to embark on a programme of introducing energy saving measures in exchange for being relieved of the tax. One option which may be considered is that companies, in order to qualify for the rebate, must appoint energyefficiency managers and satisfy an energy-efficiency inspectorate that best practice methods are being adopted. This type of scheme is being followed in the Netherlands and is partly responsible for the success of Japanese industry in reducing fuel consumption after the large energy price increases of the 1970s. In order to model this effect, the energy

ENERGY POLICY March 1993

AUK

carbon~energy tax

Table 1. Assumed real EC carbon and energy tax rates (1993 prices), a Carbon tax rates US$ per barrel oil equivalent Oil 1993 1994 2000 2005

1.5 2.0 5.0 5.0

Gas

Nuclear electricity

1.14 1.52 3.80 3.80

0.0 0.0 0.0 0.0

Coal

Gas

Nuclear electricity

1.5 2.0 5.0 5.0

1.5 2.0 5.0 5.0

1.5 2.0 5.0 5.0

Coal 1.92 2.56 6.40 6.40

Energy tax rates US$ per barrel oil equivalent Oil 1993 1994 2000 2005

1.5 2.0 5.0 5.0

Total carbon and energy tax rates US$ per barrel oil equivalent

1993 1994 2000 2005

Oil

Coal

Gas

Nuclear electricity

3.0 4.0 10.0 10.0

3.42 4.56 11.40 11.40

2.64 3.52 8.81) 8.80

1.5 2.0 5.0 5.0

Notes: a Rates in 1995-99 are linear interpolations. Source: b Commission of the E u r o p e a n C o m m u n i t y , and Cambridge Econometrics.

intensive industries are assumed to adjust their fuel consumption as if they had to pay the new tax. In the OECD-wide tax scenarios, it is assumed that some agreement is reached across the O E C D along the lines of the E C tax so that there is no need for special treatment of the energy intensive industries and that all industries pay the tax. The competitiveness of O E C D industries with respect to the rest of the world is allowed to deteriorate in the projection.

and reduced interest rates. In all the scenarios these are held constant at the base forecast levels. Finally, in the scenarios reported below no allowance is made for any stimulation to innovation as a result of the changes. It would be possible to use some of the tax revenues to stimulate research on energy saving and alternative non-fossil energy sources and this would tend to increase underlying growth. Such effects have not been considered here.

Recycling the carbon~energy tax revenues Other assumptions One of the more important assumptions in any carbon tax analysis concerns the likely effect of the tax on the world oil market. If E u r o p e reduces its demand for oil significantly, then unless supply is also reduced (for example by O P E C acting as a swing producer) the world oil price will weaken. This will stimulate demand in other parts of the world and reduce the impact of the CEC measure on overall CO2 emissions. In this paper the world oil price is assumed fixed with, implicitly, O P E C reducing supplies to match any reduction in E u r o p e a n or O E C D demands. Other assumptions concern the levels of interest rates and the exchange rates. The base forecast includes the sterling devaluation of September 1992

ENERGY POLICY March 1993

A critical feature in all scenarios is the treatment of the revenues from the taxes. The CEC has proposed that the tax should be revenue neutral ie the revenues are used to reduce the rate of income tax or V A T , for example. 13 Alternatively, if the revenues are saved by the government in the form of reductions in public sector deficits, then this will depress the economy, certainly in the short term. If the revenues are all spent by the government, for example on non-fossil energy investment, then this would imply a large investment programme which could lead to macroeconomic imbalance and rapid inflation. This implies that any decision to introduce a carbon/energy tax would be taken at the highest political level with the close involvement of finance ministries.

299

A UK carbon~energy tax Since the carbon/energy tax is an indirect tax with its immediate and direct effects on the price level, in macroeconomic terms it is a close substitute for other indirect taxes. If there is a gradual switch from other indirect taxes to a carbon/energy tax, so as to minimize the effect on the general level of prices, then the inflationary and indeed all the macroeconomic effects of the carbon tax are reduced to a minimum. 14 Analyses which do not assume recycling of the revenues (the 'depression scenario' in this paper) typically either ignore the macroeconomic effects (making the results of limited relevance) or are dominated by the macroeconomic properties of the model used and the exact means assumed for the spending of revenues. If the indirect tax option is not chosen as part of the political implementation of the tax, then the macroeconomic results will be less predictable and the policy package should therefore be considered as part of the usual budgetary process of the government. Two sets of scenarios (in addition to the 'depression scenario') are explored below in the treatment of the carbon/energy tax revenues: in one set, all revenues are recycled by means of reductions in the standard rate of VAT, and in the other, the reductions are in the standard rate of income tax. The assumption is that taxes are set such that in each year the net effect is tax revenue neutral ie the revenues from the carbon/energy tax are used entirely to reduce other taxes, as proposed by the CEC. It should be noted that this assumption does not necessarily imply that the results will be neutral with regard to the PSBR, the balance of payments, inflation or GDP growth. The change in the mix of taxes will alter the macroeconomic balance as the economy responds to the new tax structure, leading to second-round effects on tax revenues and government expenditure, even though all the new carbon/ energy tax revenues are recycled as described. The CEC tax and the international competitiveness of UK industries Much has been made of the risks to international trade of introducing a unilateral carbon/energy tax. The CEC proposals make allowance for the industries most exposed to this risk by excluding them in some way from paying the tax, at least until some more general OECD-wide tax is adopted. In the analysis below the exclusion is implemented by providing the selected industries with a rebate of the carbon/energy tax components of their coal, oil, gas and electricity costs. The exclusion reduces the efficiency of the tax as explained below, but it does mean that the European and UK industries most

300

vulnerable to competition are protected in their markets. The competitiveness question also requires special modelling in the scenarios. The MDM model used in the analysis distinguishes 13 sources of imports and the same 13 destinations for exports. These world areas can be aggregated to establish an EC or an OECD bloc. In the two main scenarios reported below, in which the EC is assumed to introduce a unilateral tax, the proportions of UK imports and exports of each commodity to and from the rest of the EC have been calculated. These proportions have then been used to measure the extent to which import prices by commodity would rise after the tax is imposed throughout the Community, and then the extent by which UK export volumes by commodity would increase in price due to increased domestic energy costs. The argument is that EC prices will rise in the same proportion as UK prices following the introduction of the tax, and that these price increases will be reflected in their exports to other countries, and therefore also in imports by the UK from the rest of the EC. The export and import prices of fuels and of goods and services supplied by the rest of the world are assumed to remain unchanged. In the scenarios in which an OECD-wide carbon/energy tax is introduced, the proportions are calculated for UK imports from the exports to the rest of the OECD. In summary, the relative competitiveness of the UK with respect to its EC trading partners remains unaffected, while the effect of the change in relative trade prices between the EC and the rest of the world is to decrease the price competitiveness of EC countries. These effects are all relatively small, especially when the reductions in VAT prevent a stimulus to domestic prices and hence wage rates. The effectiveness of the tax A carbon tax is intended to bear most heavily on goods and services which release most CO2 in their production, yet some of the most energy intensive industries are to be excluded. Clearly this will limit the impact of the tax on energy consumption and CO2 emissions, but there are two additional problems. The first is that the industries which do not pay the tax will improve their competitive position in relation to those industries which do pay. There will therefore be some switching of demand towards the products of energy intensive industries - precisely the reaction that such a tax should avoid - although it is assumed that these industries adjust their fuel consumption as if they had to pay the new tax. The other problem is that companies which find them-

ENERGY POLICY March 1993

A U K carbon~energy tax Table 2. U K C O z emissions (million t o n n e s c a r b o n ) .

1970 1980 1990 1995 2000 2005

Base

E C tax with VAT offset

E C tax with income tax offset

O E C D tax with V A T offset

O E C D tax with income tax offset

181.0 164.5 160.0 152.0 157.9 166.7

148.3 147.2 146.6

148.3 147.2 146.7

148.2 147.0 146.4

148.2 147.0 146.5

% g r o w t h pa

1970-80 1980-90 1990-95 1995-2000 2000-05

-0.95 -0.28 -1.02 0.76 1.09

-1.51 -0.15 -0.08

-1.51 -0.15 -0.07

-1.52 -0.16 -0.08

-1.52 -0.16 -0.07

%di~rence~ombase

1995 2000 2005

-2.4 -6.8 -12.1

-2.4 -6.8 -12.0

-2.5 -6.9 -12.2

-2.5 -6.9 -12.1

-7.4 -8.1 -8.5

-7.4 -8.1 -8.4

%di~renee~oml~0

1995 2000 2005

-5.0 -1.3 4.2

-7.3 -8.0 -8.4

-7.3 -8.0 -8.3

Source: For historical data see UK Department of the Environment, Digest of Environmental Protection and Water Statistics, various issues; for projections see text.

selves paying the tax will try to be reclassified as exempt or eligible for rebates if at all possible, thereby limiting the impact. CO2 e m i s s i o n s

The CEC proposals should be put in a global context. The scientific assessment report prepared for the Intergovernmental Panel on Climate Change (IPCC) concludes that 'in order to stabilise concentrations [of CO2 in the atmosphere] at present day levels, an immediate reduction in global anthro-

pogenic emissions by 60-80 percent would be required', t5 This is a formidable challenge for the UK given, first, that a continuation of recent trends suggests that emissions will rise, and second that most of the present high concentration is due to the past activities of the developed countries, which are accordingly expected by the developing countries to do most to reduce their present emissions.16 Table 2 and Figure 1 illustrate several key points about UK CO2 emissions and the impact of the policies being proposed to reduce them. First, the UK has seen a long-term decline in emissions

Million tonnes carbon 170

/

165

160

155 /

EC tax with

',L- . ~ ' -

150

145

Figure 1.

i

t

~

,

~

1980

1985

1990

1995

2000

CO 2

ENERGY POLICY March 1993

2005

emissions in the UK. 301

A UK carbon~energy tax

primarily because of the decline in the coal industry; this shows in Table 2 in the decline of 12% in total or 0.6% pa 1970-90. Second, emissions are expected to grow in the 1990s partly because transportation demand for fuels is now a large proportion of total demand and is expected to continue to grow faster than GDP unless taxes are increased; 17 CO2 emissions are expected to rise by 4.2% between 1990 and 2005 in the base forecast. Thus any emissions policy aiming for stabilization has to be strong enough to change the upward trend in emissions. Third, the UK contributes about 160 Mt of carbon each year to the atmosphere in the form of CO2.18 This amounts to 2.5% of global emissions at present and the growth in these emissions is much less than in the rest of the world, so UK (or European or even OECD) policies alone will not stabilize emissions in the long term. As long as the countries introducing emission targets maintain a reasonable macroeconomic balance, the effects on world growth are not likely to be sufficient to make any serious impact on other countries' emissions. The rates proposed by the CEC appear high enough to stabilize emissions for the UK irrespective of which revenue recycling option is chosen. Figure 1 shows that the annual rate of CO2 emissions is expected to decline to 146.6 Mt carbon by 2005 under the VAT offset scenario, more than 8% down on the 1990 level and around 12% below base levels in 2005. Table 2 shows that the reductions in CO2 emissions in the four scenarios are remarkably similar. The similarity is partly due to the assumption that the exempt energy intensive industries adjust their fuel consumption as if they had to pay the new EC tax in order to get compensation. But even so these energy intensive industries fare better under the EC tax and therefore use more energy.

The CEC tax and the UK macroeconomy The revenues and subsidies implied by the tax In the scenarios, it is assumed that the tax is imposed as specific duties on coal, nuclear electricity, petroleum products and gas levied at the narrowest channels in the flows of production and distribution. In practice, this implies that the duties are paid by British Coal, the oil and nuclear companies and British Gas, with any importers of the fuels also paying duties. These duties will then be passed on to the purchasers of the energy products in the same way that duties on hydrocarbons are at present on to petrol and derv prices. Exports of all fossil fuels are assumed to be excluded from the tax (although of course they may be taxed in the importing country).

302

The electricity industry is assumed to pay the tax on its fuel consumption and the regulator is assumed to allow the industry to pass on the tax in higher prices to customers. Other regulated industries such as telecommunications are treated in the same way, except they are also allowed to pass on the extra increases from the new tax embodied in electricity prices. The total revenues less the rebates to energy intensive industries are recycled in two separate scenarios by reducing the standard rates of VAT or income tax by exactly the amount of the net receipts of the carbon/energy tax, but as pointed out above this does not imply overall budget neutrality due to second-round effects. Table 3 presents the tax rates, revenues and rebates in the base and in four scenarios (the main scenarios for the EC only tax and the corresponding scenarios for the OECD-wide tax) for 1995, 2000 and 2005. In general, under the OECD-wide tax net carbon/energy tax revenues are higher and the macroeconomic consequences larger than under the EC tax, because the energy intensive industries are included in the tax. By 2005 for the EC tax, VAT rates can be reduced from 17.5% to 15.7% or the income tax basic rate from 25% to around 23%. The rebates for the excluded industries become substantial by the end of the period at just below £2 billion, or around 14% of gross revenues.

Effects on inflation and growth Table 4 shows the differences from base in the four scenarios. The general feature is that when the carbon/energy tax revenues are recycled via VAT reductions then the economy grows faster and the rate of inflation is unchanged or even reduced, and when the recycling is via reductions in income tax, overall growth remains positive, but slightly smaller, and inflation is higher. These macroeconomic effects largely follow through to the changes in the components of GDP, the PSBR, the balance of payments and unemployment. However, the PSBR ratios are lower in both scenarios compared with the base, but for different reasons. In the case of the VAT scenario, the extra growth in the economy and lower wage rates and prices for government expenditure are sufficient to more than compensate for the loss in VAT revenue. In the case of the income tax scenario, the net effect of higher inflation on tax revenues and government expenditure is to reduce the PSBR. In a further scenario, not reported in detail here, VAT was reduced by the exact amount required to keep inflation at base rates (an inflation neutral scenario) with very similar results to those of

ENERGY POLICY March 1993

A U K carbon~energy tax

Table 3. The carbon/energy tax scenarios: rates and revenues in 1995, 2000 and 2005.

Tax rates (%)a 1995 VAT Income tax 2000 VAT Income tax 2005 VAT Income tax

Base

EC tax and VAT offset

EC tax and income tax offset

17.5 25.0

16.3 25.1/

17.5 23.9

16.1 25.0

17.5 23.7

17.5 25.0

15.4 25.11

17.5 23.0

15.0 25.0

17.5 22.7

17.5 25.0

15.7 25.(I

17.5 23.2

15.4 25.0

17.5 22.9

4.75 0.00 4.75 -5.00 -0.32

4.78 0.00 4.78 0.66 -3.15

11.48 0.00 11.48 - 11.43 0.13

11.68 0.00 11.68 2.07 -7.38

13.68 0.00 13.68 - 12.52 1.31

13.87 0.00 13.87 3.411 -7.99

Tax revenues (£ billion differences from base) 1995 Carbon/energy tax 4.75 4.78 Rebate 0.81 0.82 Net revenue 3.94 3.96 VAT -4.12 0.57 Income tax -0.17 -2.52 2000 Carbon/energy tax 11.51 11.68 Rebate 1.81 1.84 Net revenue 9.70 9.84 VAT -9.57 1.79 Income tax 0.29 -6.03 2005 Carbon/energy tax 13.70 13.88 Rebate 1.91 1.93 Net revenue 11.79 11.95 VAT - 10.73 2.91 Income tax 1.16 -6.86

OECD tax and VAT offset

OECD tax and income tax offset

Note: a Results are given to one decimal place to illustrate the differences between scenarios. Tax rates would in practice be rounded. Source: see text.

the VAT revenue neutral scenario. The extra growth is consumer led as revenue recycling directly affects consumers. Therefore, the larger net revenues under the OECD-wide tax lead to higher growth compared with the EC tax. The stronger growth also implies a more sizable deterioration in the balance of payments even though competitiveness is harmed less. The higher inflation from income tax offsets also worsens the balance of payments compared with the VAT reduction cases. If the government chose to keep all the carbon/ energy tax revenues for reducing the PSBR, as explored in the depression scenario, GDP would be 0.37% lower than base in the EC tax case by 2005 (Table 5), compared with 0.17% above base if revenues were recycled via VAT offsets (Table 4). However, retaining this large amount (around 10% of VAT revenues) would only improve the PSBR by 0.4 percentage points, as lower growth reduces tax income while higher unemployment increases income transfers.

ENERGY POLICY March 1993

Effects on industrial structure and the energy sector The impact on industrial structure from a carbon/ energy tax is illustrated in Figures 2 and 3. Figure 2 shows the effect on two energy intensive industries, iron and steel, and chemicals, under the EC tax where they are exempt, and under the OECD tax where they face competition from nonOECD producers. Both scenarios assume revenue recycling through VAT reductions. Not surprisingly the industries fare worse when they have to pay the tax, because they lose their privileged position in relation to other industries in the UK, as well as losing some export markets to non-OECD producers. However, the reductions in output are comparatively small, only 1.5% below base for iron and steel at maximum. Figure 3 shows the effects of the EC tax with VAT offset on gross output of four selected industries: mechanical engineering shows no significant effects;

303

A U K carbon~energy tax Table 4. Macroeconomic results. EC tax with VAT offset

EC tax with income tax offset

OECD tax with VAT offset

OECD tax with income tax offset

GDP 1995 2000 2005 Consumers' expenditure 1995 2000 2005 Consumer prices 1995 2000 2005

% difference from base 0.09 0.03 0.19 0.02 0.17 0.09

Balance of payments/GDP 1995 2000 2005 PSBR/GDP 1995 2000 2005

Percentage point difference from base -0.07 -0.13 -0.22 -0.16 -0.38 -0.47 -0.23 -0.56 -0.55

-0.30 -0.78 - 1.02

-0.15 -0.24 -0.24

-0.17 -0.31 -0.31

-0.23 -0.34 -0.28

Unemployment 1995 2000 2005

difference from -23 -50 -59

base (thousand) -1 -34 -2 -78 -5 -93

-9 -26 -36

0.09 0.13 0.22

0.04 0.06 0.18

0.20 0.28 0.35

0.14 0.29 0.51

0.29 0.43 0.50

0.22 0.49 0.75

-0.38 -0.24 0.05

1.02 2.21 2.29

-0.60 -0.55 -0.12

1.08 2.35 2.45

-0.19 -0.25 -0.19

Source: see text.

Table 5. The depression scenario: differences from base.

1995 2000 2005

GDP (%)

Consumer prices (%)

BOP/ GDP (% point)

PSBR/ GDP (% point)

CO2 emissions (%)

-0.19 -0.43 -0.37

1.11 2.31 2.29

-0.01 -0.11 -0.23

-0.47 -0.74 -0.64

-2.6 -7.1 -12.2

Source: see text.

% difference from base 0.5

Chemicals (EC) /

. . . . "--,.. . . .

...........

",.,

............. ':::=:: ........

"~::K:-::=::::::::=:~

.,,,, -0.5

-........... :.........

",,

•

.............

Iron & Steel (EC)

.::

--~ ..........

~------::-:__

--~-~migZiOECD)

-1

-1.5

-2

Iron & Steel (OECD) I

1990

I

I

I

~

l

1995

I

I

I

l

I

2OO0

I

t

I

I

I

2005

Figure 2. G r o s s o u t p u t of iron and steel and chemicals: E C tax versus O E C D tax.

304

ENERGY POLICY March 1993

A UK carbon~energy tax

% difference from base 1 Construction .. ............................................

0,5

Business services

..............

............, :~::~:-,.--_:::::::---------_" •

................................... -. . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

Mechanical engineering

"\ ...... . -. -0.5

-- ...................

. ..............

Petroleum products

-1.5

I

I

I

I

990

I

I

I

I

I

I

1995

I

I

I

I

l

2000

I

2005

Figure 3. Gross output of selected industries: EC with VAT offset. Million therms 120,0o0 Primary Electricity

100,OOnI 80,0C

Gas 80,0~

Oil

40,0(

Motor spirit/ Defy

20,0(

Coal 1990

Base case

2005

EC Tax 2005

OECD Tax 2005

Figure 4. Demand for primary fuels in the UK. construction benefits from the faster growth which raises investment; business services benefits from the general shift from energy intensive industries to the rest of the economy; and petroleum products suffers from loss of domestic markets through the imposition of the tax. The effects on the energy sector of the tax are summarized in Table 6 and Figure 4. Table 6 shows

the decline in total primary energy use below base levels. The reduction in energy use of about 5.6% in 2005 is appreciably less than that in CO2 emissions (about 12% in 2005) because of fuel switching towards less carbon intensive fuels such as gas. Figure 4 shows the composition of the energy demand in terms of the primary fuels. The base case shows that total energy demand is expected to be

Table 6. Total primary energy use.

Base

1995 2000 2005

1995 2000 2005

0.5 6.3 13.0

EC tax with VAT offset

EC tax with income tax offset

OECD tax with VAT offset

OECD tax with income tax offset

%di~renceffombase -1.2 -1.2 -3.3 -3.3 -5.6 -5.6

-1.3 -3.4 -5.7

-1.3 -3.4 -5.7

%di~reneeffom1990 -0.7 -0.7 2.8 2.8 6.6 6.6

-0.8 2.7 6.5

-0.8 2.7 6.5

S o u r c e : see text.

ENERGY POLICY March 1993

305

A UK carbon~energy tax

above 1990 levels in 2005, although coal demand is well down and gas demand is up. The effects of the EC tax and the OECD tax are very similar; there is an overall reduction in demand compared to base and the substitution of gas for coal in electricity generation is further accelerated.

GDP costs and benefits of slowing climate change Some surveys of the literature have largely taken it as axiomatic that measures to abate CO2 emissions will reduce economic growth (Boero, Clarke and Winters, Nordhaus). 19 A feature of these surveys is that the GDP effects are nearly all negative and that the reductions in CO2 are all assumed to be achieved by carbon/energy taxes, often at very high rates. The crucial macroeconomic assumption when it comes to estimating both short-term and long-term effects on GDP is that concerning the treatment of tax revenues; yet these surveys do not record systematically the assumptions adopted, often because on this has not been made clear in the original studies. The usual assumption has been that the revenues are accumulated by government or that they are recycled as lump sum transfers. In those cases where recycling assumptions are explored, the effects on GDP growth are dramatic. 2° Where revenues have been explicitly recycled by means of reductions in indirect tax rates, it is notable that there are many instances of positive GDP effects. 21 The cost estimates that are presented are generally not based on empirical evidence but are rather the outcome of assumptions built into particular modelling frameworks. Unless the carbon/energy revenues are precisely offset by reductions in VAT or other indirect tax revenues, the results will be affected by the responses of the model to increases in inflation and to a change in prices of carbon intensive relative to non-carbon intensive goods and services. The purpose of a carbon tax is essentially to change relative prices, not the general price level; moreover, it is to change relative prices from a position which can be defined as suboptimal from the point of view of the social costs of CO2 emissions. If the economy has unemployed resources, then the spending of the revenues from a carbon/energy tax can increase GDP and move the economy towards full employment. This is the approach adopted in this paper. If the economy is assumed to be in static equilibrium at full employment, then the effect of a change in relative prices is more problematic. If it is assumed that the pre-existing set of

306

relative prices is optimal, then by definition any change will move away from the optimum and therefore be costly. However, the existing relative prices are not necessarily optimal; for example, in many countries there are taxes on labour or on other goods and services increasing the prices above any optimum. In addition, it is much more important in a long-term analysis to model the effects of the tax on dynamic processes of economic growth. This is very difficult, and many studies have adopted static models of the allocation of resources (general equilibrium models), rather than developing endogenous growth models. In the EC tax, VAT offset scenario of this paper, 20.1 Mt carbon are abated by 2005 at a benefit of 0.17% GDP, ie £950 million at 1985 prices. Converting this to US dollars at 1989 prices using Nordhaus's assumption of US$1.9 to £1, the GDP benefit is US$44 per tonne of carbon abated, an average benefit in 2005 for a 12% reduction in CO2 emissions. Although there are many differences in the models and methods surveyed by Nordhaus, this result compares with a US$4 per tonne average GDP cost estimate from his regression line relating costs to abatement. 22 The more general social benefits are, of course, expected to be much higher, because in addition to the increase in GDP there are three other net benefits. First, there are benefits from reductions in other external costs associated with CO2 emissions, such as less acid rain, improved urban air quality, less atmospheric noise and dirt. Second, the recycling of revenues permits a reduction in the most distortionary of other indirect taxes as well as offsetting the worst effects on the distribution of income. And third, there is the long-term global benefit of slowing global warming.

Conclusions The European Commission's proposal for a combined carbon and energy tax harmonized across EC member states is an important step in recognizing the problem of global warming and devising policies to meet CO2 targets. An e c o n o m y - e n e r g y environment model has been developed in order to provide a framework for analysing the implications of such policies for the UK economy. The tax has been introduced into the model at the rates proposed and scenarios considered for the use of the revenues and the coverage of the tax to exclude and include energy intensive industries. The main conclusions to emerge from the study are as follows.

ENERGY POLICY March 1993

A U K carbon/energy tax

First, the tax is expected to achieve stabilization of CO2 emissions by 2000; indeed, with the 1990-92 recession and the increased use of gas in the 1990s by the U K electricity industry, it is likely that even without a tax, emissions will be below 1990 levels in the year 2000. Thus the official U K government target of stabilization by 2000 is likely to be achieved by the CEC's proposed new tax. Second, the use of the revenues from the tax is critical in the macroeconomic effects of the tax, but not in its effects on CO2 emissions. If there are unemployed resources, judicious recycling of revenues can lead to increases in both G D P and employment. If V A T is reduced then the inflationary consequences of the new tax (it is after all designed to raise energy prices) will probably be neutralized and, if the model and assumptions are correct, inflation will actually be reduced. If income tax is reduced, then inflation is likely to increase, although the extent of the acceleration depends on the attitude of wage negotiations to the increases in disposable income from the reductions in income tax. If past behaviour is a guide, then the income tax cut will only be partly reflected in lower wage claims. This disposition of the U K economy towards a higher inflationary response has also been found in other studies of EC-wide carbon and energy taxes. 23 Third, following from the importance of the use of revenues is the neutrality aspect of implementation. The scenarios considered here are tax revenue neutral ie in each year the net receipts have been spent in reducing another tax. In this, the study has followed the statements made by the Commission and the criteria adopted in the other EC-wide studies discussed above. However, it is clear that revenue neutrality may not be an appropriate criterion, at least for the U K economy, because it does not properly simulate the long-term budgetary process. The macroeconomic balance of different taxes and expenditures will be decided from year to year on the basis of current concerns, and revenue neutrality is of little importance. If studies are focused on the macroeconomic costs of the tax, it seems preferable to assume a change which is inflation neutral and to leave questions of short-term macroeconomic demand management aside. Fourth, the proposed exclusion of energy intensive industries such as chemicals and steel from the tax coverage, although understandable in terms of fears about E u r o p e a n competitiveness, will weaken the tax base and tend to discredit the tax by introducing more opportunities for tax avoidance. If the extra tax costs were to be fully rebated to these industries, the tax take is estimated to be reduced by

ENERGY POLICY March 1993

around 14% in 2005 and the tax is also less effective in reducing emissions. This consideration leads to two further important issues for the tax: a critical aspect of any implementation will be the treatment of the excluded industries; and the loss in effectiveness confirms the emphasis given by the Commission in seeking the cooperation of other members of the O E C D in policies to reduce emissions. Finally, the results show that issues of international competitiveness for an EC-wide tax, although of central concern to some export sensitive companies, are relatively insignificant in macroeconomic terms. One reason for this is that, even for the more energy intensive products, other costs (including transport) are less affected by the tax. Another reason is that, as products become increasingly sophisticated, price competitiveness becomes less important in holding and increasing market share. The authors are grateful to the ESRC and Cambridge Econometrics for financial support for the research reported in this paper. The ESRC is supporting the use of the Cambridge model in the project 'Policy Options for Sustainable Energy Use in the UK Economy' in the Department of Applied Economics under its Global Climate Change Initiative. Cambridge Econometrics is responsible for maintaining the Cambridge model.

1CEC, A Community Strategy to Limit Carbon Dioxide Emissions and to Improve Energy Efficiency, SEC(91) 1744, 1991. ~T.S. Barker and R. Lewney, 'A green scenario for the UK economy', in T.S. Barker, ed, Green Futures for Economic Growth: Britain in 2010, Cambridge Econometrics, 1991, pp 11-38; P. Capros, P. Karadeloglou, L. Mantzos and G. Mentzas, Impact of Energy Tax and Carbon Taxes on C02 Emissions, Report for the CEC-DGXII, Athens, July 1991; DRI, The Economic Impact of a Package of EC Measures to Control C02 Emissions, Final Report prepared for the CEC, November 1991; and DRI, Impact of a Package of EC Measures to Control C02 Emissions on European Industry, Final Report prepared for the CEC, January 1992; P. Karadeloglou, "Energy tax versus carbon tax: a quantitative macroeconomic analysis with the HERMES/ MIDAS models', in F. Laroui and J.W. Velthuijsen, eds, An Energy Tax in Europe, SEO Report No 281, Amsterdam, 1992, pp 127-152; and S. Standaert, 'The macro-sectoral effects of an EC-wide energy tax', in F. Laroui and J.W. Velthuijsen, eds, An Energy Tax in Europe, SEO Report No 281, Amsterdam, 1992, ~p 27-63. kT.S. Barker and A.W. Peterson, eds, The Cambridge Muhisectoral Dynamic Model, Cambridge University Press, 1987. 40p cit, Ref2, Barker and Lewney; T.S. Barker, S. Baylis and R. Lewney, 'Can Britain meet the Toronto COz target'?', Paper presented to Energy, the Environment and Economic Policy Conference, Cambridge, 1-2 July 1991. 50p cit, Ref 2, Standaert and Karadeloglou. 6For a more detailed discussion see E. Donni, P. Valette and P. Zagam6, HERMES, Harmonised Econometric Research for the Modelling of Economic Systems, North-Holland, Amsterdam, forthcoming. 7Only HERMES-BELGIUM has an interfuel substitution module necessary for explicitly introducing a carbon/energy tax. See Bureau du Plan, L'introduction d'une Taxe sur les Emissions de C02 en Belgique: Consequences Macroeconometriques et Politiques d'Accompagnement Possibles, Brussels, (90)FB-RDR/4554/ al/5954, 1990.

307

A UK carbon~energy tax 80p cit, Ref 2, Karadeloglou and DRI. 9Models only for the major EC countries currently exist. 1°The CEC has proposed specific energy tax rates in ECUs per kilowatt hour for nuclear and all electricity. However, in this study the energy tax rates have been calculated on the basis of the thermal inputs to generation, and thus the final rates per kilowatt hour vary according to the efficiency of the station used, including nuclear. In addition, a lower rate per kilowatt hour has been proposed for large-scale hydroelectricity, based on 100% efficiency. l~Department of Energy, Digest of UK Energy Statistics, HMSO, London, 1992. 12In the proposition accepted by the CEC on 27 May 1992 rebates ranging from 25% to 90% depending on the energy intensity of the industry were suggested. In this study a full rebate was assumed. 130p cit, Ref 1. 14R.C. Dower and M.B. Zimmerman, The Right Climate for

Carbon Taxes: Creating Economic Incentives to Protect the Atmosphere, World Resources Institute, August 1992. They argue that carbon taxes would benefit the US economy as a whole if adopted as part of a shift in the tax burden from things we want to encourage to things we want to discourage. 15IPCC, Scientific Assessment of Climate Change, Report Prepared for IPCC by Working Group 1, June 1990, World Meteorological Organisation and United Nations Environment Program-

308

me, p5. 16Dr Subroto, 'Outlining the Organization's response to the Rio asgDenda', OPEC Bulletin, September 1992, pp 4-10. epartment of Transport, National Road Traffic Forecasts (Great Britain), London, HMSO, 1989. 18All measures of CO2 emissions in this paper are given in terms of carbon content, following the practice in UK DoE environmental statistics. 19G. Boero, R. Clarke and L.A. Winters, The Macroeconomic Consequences of Controlling Greenhouse Gases: A Survey, UK Department of the Environment, Environmental Economics Research Series, HMSO, 1991; W.D. Nordhaus, 'Economic approaches to greenhouse warming', in R. Dornbusch and J.M. Poterba, eds, Global Warming: Economic Policy Responses, MIT Press, 1991. 2°Op cit, Ref 19, p 93, Boero et al; Ref 14, pp 14-16. 21See op cit, Ref 14, as well as CEC Directorate-General for Economic and Financial Affairs, The Climate Challenge Econo-

mic: Aspects of the Community's Strategy for Limiting C02 Emissions, European Economy, No 51, May 1992, and CEC Directorate-General for Economic and Financial Affairs, The Economics of Limiting C02 Emissions, European Economy, S2PwecialEdition No 1, 1992. .D. Nordhaus, 'The cost of slowing climate change: a survey', in The Energy Journal, Vol 12, No 1, 1990, p 63. 230p cit, Ref 2, Standaert and Karadeloglou.

ENERGY POLICY March 1993