Fiscal Studies (1998) vol. 19, no. 4, pp. 375–402 Equity and Ecotax Reform in the EU: Achieving a 10 per cent Reduction in CO2 Emissions Using Excise Duties

TERRY BARKER and JONATHAN KÖHLER*

Abstract

This paper considers the distributional effects of imposing additional excise duties on energy products according to carbon content. The assumed duties escalate from 1999 to 2010 and achieve levels reducing CO2 emissions by 10 per cent below baseline by 2010 for 11 EU member states. By 2010, real personal disposable incomes are 1.6 per cent above baseline and employment is 1.2 per cent above, assuming that the change is tax-revenue-neutral. The study concludes that the changes will be weakly regressive for nearly all the member states in the study if revenues are used to reduce employers’ taxes and strongly progressive if they are given back lump-sum to households.

JEL classification: C53, D12, H22, Q48.

I. INTRODUCTION This paper inquires into changes in the distribution of expenditure and income that would follow the introduction of revenue-neutral tax changes to reduce CO2 emissions across the European Union (EU). Consumer expenditures on environmentally-sensitive goods and services (ESGSs) are examined and the effect on these expenditures of imposing different excise duties on energy products calculated. The consequences for incomes and employment are also considered. The analysis concentrates on the distribution across different groups

*Department of Applied Economics, University of Cambridge. The authors would like to thank Paul Ekins and two referees for their useful advice. Fiscal Studies of households, classified by their level of consumption expenditure and by socio- economic group. The results come from the E3ME, an environment–energy– economy model for the EU, which generates results for 11 member states of the EU up to the year 2010. Reductions in greenhouse gas (GHG) emissions, in particular CO2, are once more an important issue in international policy. The 1998 Kyoto protocol has set a specific target for GHG emission reductions for the EU of 8 per cent below 1990 levels on average between 2008 and 2012. Given the need to achieve emission reductions, the use of market-based instruments is now widely accepted as an approach that brings the efficiency of markets to bear on the problem of sustainable development in an effective way. The concepts and implementation issues are discussed in European Environmental Agency (1996), EFILWC (1996) and Ekins (1997). The European Commission has proposed additional excise duties on energy products in place of the carbon/energy tax, although the rates proposed are too low to make much contribution to the target. Nevertheless, ecotax reform remains on the tax agenda, and even if new proposals are made for emissions trading, the simulation of tax policies is helpful in assessing the effects of auctioning permits, since in theory these should have much the same effects (Cramton and Kerr, 1997). However, there have been two main objections to the use of fiscal policy to reduce CO2 emissions, such as increased excise duties on energy products: the first is that such taxes will increase industrial costs and weaken the competitive position of energy-intensive industries; the second is that the taxes are inequitable, an objection that had its most powerful political expression in the defeat of the UK government’s proposal to increase VAT on domestic fuels in 1994. On the first objection, the literature has found no evidence of significant reductions in international competitiveness that would follow ecotax reforms (Barker and Köhler, 1998). However, there are still widely differing opinions about the distributional consequences of ecotax reform, partly because they depend on the way in which the revenues from proposed legislation are recycled in the economy. This paper addresses the issue by considering a reduction in distortionary labour taxes and comparing the results with recycling via lump-sum payments to all households. The literature is very sparse in this area, and much of it is focused on the effects of a carbon tax. For instance, it is frequently argued that the distribution effects of a carbon/energy tax will be largely regressive: the burden of the tax will fall disproportionately on lower-income households (see Poterba (1991), Dower and Zimmerman (1992), Smith (1992), Commission of the European Communities — CEC — (1992) and Johnson, McKay and Smith (1990) for discussions). The basis upon which it is usually assumed that carbon/energy taxes are regressive is intuitively obvious: lower-income households tend to spend a larger proportion of total household expenditures on fuel for domestic energy services (i.e. heating, hot water, cooking and lighting).

376 Equity and Ecotax Reform in the EU

A recent review of the literature and state of the art in research on equity effects of environmental policies in relation to climate change has been made by the OECD (1995). Three of the conclusions of this review are: 1. ‘With regard to income distribution effects, empirical studies suggest that a national carbon tax or trading programme would be at least mildly regressive ... in many OECD countries, although there is some evidence that such programmes might actually be progressive in developing countries.’ 2. ‘Perhaps the major conclusion of the empirical studies is that distributional effects need to be considered very carefully as policy proposals move from concept, to design, and then to implementation.’ 3. ‘The specific mitigation or compensation strategies also need to be spelled out more clearly. If taxes are collected, how exactly would the revenues be “recycled” or used?’ The three most detailed quantitative studies for national economies of the equity effects of a carbon tax are by Symons, Proops and Gay (1994) for the UK, Hamilton and Cameron (1994) for Canada and Cornwell and Creedy (1996 and 1997) for Australia. These are static, mostly partial equilibrium analyses, making use of input–output tables and consumer demand systems to calculate the effects of carbon taxes under different assumptions about recycling. The conclusions are that the magnitude of a carbon tax to reduce CO2 emissions — for example, by 20 per cent — is large, but transfer payments can be adjusted to compensate for the regressivity of the tax without decreasing total revenues (see, for example, Cornwell and Creedy (1996, p. 35)). However, there is no technological substitution in these models, and no revenue-recycling through reductions in employment taxes. The most comprehensive study of the expenditure distribution effects of carbon/energy taxation in Europe is that undertaken by the European Commission (Smith, 1992; CEC, 1992). The results are also discussed in OECD (1995). The study mainly relied on EU household expenditure surveys for six countries, usually for 1985 (Spain, Italy, the Netherlands, France, Germany (1983) and Ireland (1987)). The analysis was mainly static, assuming (a) no indirect effects, (b) that any taxes do not change demand patterns, (c) that all the tax is passed on to prices and (d) that there is no revenue-recycling. Under these assumptions, the ratio of regressivity — defined as carbon tax payments as a percentage of household total expenditure of the poorest quartile over the corresponding percentage for the richest quartile — was calculated for 11 countries. The conclusion was that carbon and energy taxes were weakly regressive for most countries, but more strongly regressive for the UK and Ireland. The analysis (CEC, 1992) goes further in considering indirect effects in some illustrative calculations for France and Germany, which confirm the static calculations (CEC, 1992, p. 134). Smith was able to go further in considering the

377 Fiscal Studies effects of recycling either as lump-sum payments or across-the-board reductions in tax rates, but only for the UK in 1988 (Smith, 1992, pp. 264–6) and again without considering indirect and dynamic effects. The main limitation of the EC study is that it concentrates on one side of environmental fiscal reform (EFR) — the effects of tax increases on expenditures; the other side of EFR — the use of revenues to reduce employers’ taxes and increase employment — is not treated specifically. EFR is intended to increase employment and reduce unemployment as well as improve environmental performance; it is likely that the extra employment will have favourable effects on the distribution of income, given that low-income households are likely to benefit more than other households from reductions in unemployment. In the large quantitative studies of the effects of the EC’s proposed carbon/energy tax (Data Resources Incorporated (DRI), 1991 and 1992) or of a substantial package of environmental policies (DRI, 1994), distributional questions have been incidental, rather than fundamental, components of the analysis. The reason is partly the lack of data and the poor quality of the data available, although consistent data covering the EU are becoming available. The paper is organised as follows. The next section introduces the E3ME model and describes its development to incorporate expenditure distributions. It also describes the data on ESGSs. The scenarios for ecotaxation and their relationship to the 1997 proposed Directive on additional excise duties on energy products are detailed in Section III. Section IV gives the macroeconomic context for the detailed results and Section V sets out the increases in prices for goods and services bought by the socio-economic groups in the study and the effects of these increases on real disposable incomes. Section VI concludes.

II. THE E3ME MODEL AND DATA The energy–environment–economy model for Europe, E3ME, is a sectoral, regionalised, econometric model of the EU (CEC, 1995 and 1998). It is a disaggregated time-series, cross-section econometric model, which is distinct from the alternative methodology of (computable) general equilibrium (CGE) models, although it has benefited from the techniques used in CGE models — for example, in calibration on recent data. The model has been constructed by a team of partner institutes across Europe, with Cambridge Econometrics acting as co-ordinator of the project.1 The model treats member states as distinct economic

1The E3ME project is supported by the CEC, DGXII for Science, Research and Development, under the EC Non-Nuclear Energy Programme (JOULE-THERMIE) 1994–98, contract no. JOS3-CT95-0011. Participating organisations (contractors and associate contractors only) are: Cambridge Econometrics; Chambre de Commerce et d’Industrie de Paris; ERECO (Brussels); Bureau Federal du Plan (Brussels); Statistics Norway; University of Oulu; INFRAS (Zürich).

378 Equity and Ecotax Reform in the EU entities interacting with one another; but at the same time it is one model, giving the benefits of common classifications, definitions and methodology, and with equation estimates and results capable of being aggregated to the European level. In this it differs from the suite of DRI econometric models used to analyse possible EU E3 policies (DRI, 1991, 1992 and 1994) or the HERMES model (see CEC (1993)) used to analyse the EC’s proposed carbon/energy tax. In some respects, E3ME is similar to some national models used for E3 policy analysis (Cappelen, 1991; Barker and Lewney, 1991). Previous model applications have been limited in two respects. First, they have relied either on an integrated EU macro-model (for example, QUEST used in CEC (1992)) so that they fail to capture important sectoral effects, or on a suite of national models (HERMES or DRI (1991)) so that they fail to capture proper integration at the sectoral level. The use of E3ME as an integrated sectoral model addresses these problems. Second, the work on CO2 abatement has been almost entirely divorced from that on SO2, an approach that is clearly inappropriate since many measures targeting one of these kinds of emissions will affect the other. Since policy needs to encompass a package of measures that will tackle both kinds of emissions, it requires a tool for analysis that models instruments and emissions consistently, including unintended side-effects. The current operational version of the model includes 11 member states (the EU-12 without Greece), 30 industries, 27 consumer categories, 17 fuel users and 11 fuels. It provides substantial detail on CO2 emissions. The model solves for output, prices, employment, fuel demand and supply each year to 2010 in considerable detail under various assumptions about fiscal and exchange rate policy. The key assumptions of the model are as follows: • employment can adjust freely in response to demand (i.e. there is no full- employment assumption); • the additional duties affect prices only; there is no independent response via R&D, innovation or other factors affecting competitiveness; • prices are determined according to estimated responses from time-series equations with long-run full passing-on of unit costs (i.e. prices are not required to clear markets, but are largely administered); • wage rates are determined in a wage-bargaining framework such that all taxes and duties are paid by the final consumer in the long run; • economic activities and prices in the rest of the world are unaffected by the introduction of the additional duties in Europe; the world oil price is unchanged between projections; and • exchange rates and interest rates are all unchanged between projections. The model has been extended by introducing two new classifications or sets for the analysis of household expenditures: 10 ESGSs and 14 socio-economic groups classifying households on the basis of consumption expenditure, by

379 Fiscal Studies occupation — manual labour or unemployment — and by main source of income. These classifications have been chosen after consideration of the data available on the distribution of expenditures (EUROSTAT, 1992 and 1993) and the detail available on consumers’ expenditures in EUROSTAT accounts. They concentrate on socio-economic groups that are generally considered to be the poorer sections of society and which may suffer considerable increases in the prices of vital goods (for example, heating for pensioners) if ecotaxes are introduced.

1. The Treatment of Nominal Incomes by Socio-Economic Group At present, E3ME distinguishes the following nominal income and expenditure flows to the personal sector by region: wages and salaries by industry, social security benefits, residual incomes, direct income tax payments and employees’ social security payments. The wages and salaries are derived from wage and employment equations for 30 industrial groups (see Barker and Gardiner (1996)). Residual income flows are also estimated from equations. Rates for social security benefits, for taxes and for social security payments are all calculated from historical data and held constant in the projection at the last observation values, unless further information about changes in rates is available. This assumption implies that these flows will rise at the same rate as regional wages and salaries, since the rates are in relation to wages and salaries. Information on the income characteristics of the socio-economic groups is required in order to attribute these flows to the groups. Data for total gross income, net income, social security income, pensions income, and income tax and National Insurance payments are taken from the Luxemburg Income Study. See Köhler (1998) for a detailed treatment of the data. The nominal incomes per head from the E3ME dataset for 1988 are then converted to expenditures using the average saving ratios by region. They are then compared with the total nominal expenditures for the socio-economic groups from the EUROSTAT surveys and ratios between the two calculated and stored. These ratios represent the differential saving ratios between the groups and the average for each region as well as differences in definitions, grossing-up and other discrepancies. It is assumed that these differentials and residuals remain a constant proportion of nominal expenditures over the projection period. Since there are comprehensive data for one round of surveys only, the question arises as to the constancy of the proportions of income from the different sources. They may be expected to change gradually over time (for example, as the proportion of pensioners in the population grows) and they may change over the economic cycle as unemployment rises and falls. Some evidence on the behaviour of the coefficients is given for the UK in Table 1. The table shows that, from 1983 onwards, there is no obvious pattern for the proportion of

380 TABLE 1 Wages and Salaries as Source of Household Income in the UK, and Unemployment Rates Per cent Year of FES 1996– 1995– 1994– 1993 1992 1991 1990 1989 1988 1987 1986 1985 1984 1983 97 96 95 Unemployment rate 7.5 8.2 9.3 10.3 9.7 8.1 5.8 6.3 8.1 10.0 11.1 10.9 10.7 10.5 Wages and salaries Lowest quintile 3.6 4.1 3.2 3.6 3.2 4.9 3.1 3.7 3.3 4.4 3.9 2.6 3.4 3.9 Second quintile 23.9 22.5 22.7 21.3 25.3 25.4 29.6 27.7 27.6 26.4 28.2 26.9 29.5 28.9 Sources: Family Expenditure Survey; Monthly Abstract of Statistics; Annual Abstract of Statistics. Fiscal Studies wages of the lowest quintile and perhaps a slight decline for the second quintile. Taking the last business cycle as running from approximately 1984 to 1995, there is also no obvious relationship between the unemployment rate and the wage source of income.

2. The Treatment of Real Expenditures and Price Indices of ESGSs by Socio- Economic Group The 1988 survey results, as aggregated, extended and processed by Köhler, Luhmann and Wadeskog (1997), provide data on all households’ expenditures on environmentally-sensitive goods and services. The main data source used was the EUROSTAT Family Budgets (EUROSTAT, 1992 and 1993), which contain detailed household consumption data based on surveys carried out in the 12 member states in 1988. The data cover 11 of the 12 EEC states, but detailed data for (West) Germany are not available in this source, so the German survey (Statistisches Bundesamt, 1994) was used. The EUROSTAT data include expenditures of households grouped by expenditure categories, by socio- economic groups and by the main income source of the household. The data across the expenditure categories provide information on the distribution of consumption expenditure, and the income-source groups enable the ‘at-risk’ groups — those whose income is mainly derived from social transfers and state pensions — to be analysed. These data provide the only source of expenditure data that have been calculated on a consistent basis for the EEC states. The data do not identify expenditure on vehicle fuels as an individual category. These data were assembled from national data sources. The data for Germany have been converted into the EUROSTAT categories. The EUROSTAT data are limited in that numbers of households in the various surveys are not given and the data are given for adult-equivalents (converted from the household membership using the OECD scale), so further analysis cannot be undertaken. This also means that no account can be taken of effects due to household composition, but Smith (1992, p. 252) suggests that this will not give rise to major biases. The EUROSTAT expenditure categories concentrate on low expenditure groups: the highest category is ‘1.6 × the average expenditure over all households’. In the UK 1988 Family Expenditure Survey (Central Statistical Office, 1990), this was found to cover 2,134 of the 7,265 households surveyed, so the wealthy parts of the populations are not described in much detail. However, since the objective of this analysis is to concentrate on the low expenditure/income groups which may be disadvantaged by new ecotaxes, the organisation of the data is well suited to this purpose. Another problem is that households in low expenditure groups may have a high income but be frugal, instead of having their expenditure constrained by a low income. This would mean that the low expenditure categories were not necessarily

382 Equity and Ecotax Reform in the EU

FIGURE 1 Distribution of Expenditure Proportions on Domestic Fuels, 1988

16 16

14 14

12 12

10 10

8 8

6 6

4 4 Percentage of total expenditure Percentage of total expenditure 2 2

0 0 <0.4 0.4Ð0.6 0.6Ð0.8 0.8Ð1.2 1.2Ð1.6 >1.6 <0.4 0.4Ð0.6 0.6Ð0.8 0.8Ð1.2 1.2Ð1.6 >1.6 Expenditure group Expenditure group (proportion of average expenditure) (proportion of average expenditure)

Belgium Ireland Luxemburg Spain Denmark Italy Netherlands West Germany France Portugal UK

Sources: EUROSTAT Family Budgets, 1992 and 1993; EUROSTAT calculations; Statistisches Bundesamt, 1994; Bundesministerium für Wirtschaft, 1996; UK Family Expenditure Survey, 1988; Italian household survey, 1988; Belgian household survey, 1988; Spanish household survey, 1988; EU project PL950582 Environmental Fiscal Reform. composed of households that were poor. The distribution of expenditures compared with incomes will also be skewed by differences in saving rates for different income groups. These issues are discussed further in Köhler, Luhmann and Wadeskog (1997). Köhler, Luhmann and Wadeskog describe the data in detail, and Figures 1 and 2 present expenditure proportions on domestic fuels2 and householder transport.3 The percentage of households’ budgets spent on domestic energy decreases as they spend more overall, i.e. the poor spend a larger percentage of their budgets on energy. For most countries, the budget portion spent on energy decreases from a range of 3.54 per cent in Spain to 9.95 per cent in Denmark for

2Given in the EUROSTAT data as gas, electricity and other fuels. 3Calculated as vehicle purchases + purchased transport from the EUROSTAT data with expenditures on petrol and diesel fuel added from national data.

383 Fiscal Studies

FIGURE 2 Distribution of Expenditure Proportions on Transport, 1988

20 20

18 18

16 16

14 14

12 12

10 10

8 8

6 6

Percentage of total expenditure 4 Percentage of total expenditure 4

2 2

0 0 <0.4 0.4Ð0.6 0.6Ð0.8 0.8Ð1.2 1.2Ð1.6 >1.6 <0.4 0.4Ð0.6 0.6Ð0.8 0.8Ð1.2 1.2Ð1.6 >1.6 Expenditure group Expenditure group (proportion of average expenditure) (proportion of average expenditure)

Belgium Ireland Luxemburg Spain Denmark Italy Netherlands West Germany France Portugal UK

Sources: Same as for Figure 1. the 0.4–0.6 × average expenditure group to a range of 2.09 per cent in Spain to 4.36 per cent in Denmark for the highest expenditure group. The pattern of transport expenditure is in sharp contrast to that of expenditures on domestic energy: for all countries, it increases with overall expenditure. In Italy, Portugal and the UK, there is a particularly sharp increase between the 1.2–1.6 expenditure group and the highest expenditure group (>1.6 × average expenditure). Most households spend more of their budgets on transport than on energy: for households in the 0.8–1.2 expenditure group, the spending on transport is between 6.26 per cent (Belgium) and 11.09 per cent (France) compared with a variation between 2.74 per cent (Spain) and 6.50 per cent (Denmark) for expenditure on energy. The high expenditure groups spend very heavily on transport: in Denmark, France, Italy, Luxemburg and Portugal, more than 15 per cent of household expenditure is on transport for the highest expenditure group. The E3ME database and projections of consumers’ expenditures per head in 27 categories, when converted to the same ESGS classification as above, provide a second set of estimates. These data are compared in 1988 and the ratios stored.

384 Equity and Ecotax Reform in the EU

The E3ME projections of consumers’ expenditures can then be converted to ESGS projections in constant 1988 prices. The corresponding price indices of ESGS by region, converted to 1988=1.0, can also be calculated from the price indices of consumers’ expenditures for the 27 categories. Two strong assumptions are then necessary to derive the real expenditures on ESGSs by socio-economic group and the corresponding price indices. First, it is assumed that each group responds to price and income changes in the same way as the average household in each region. Second, it is assumed that the prices of the ESGSs move the same way for all groups in a region, i.e. although there may be differentials in the prices paid, the differentials remain constant over the projection. The real expenditures on ESGSs, by socio-economic group, region and year, can then be calculated and summed to give totals by group, domestic heating by group or householder transport by group. The price indices for ESGSs, socio-economic group, region and year can then be calculated and weighted to give the group indices. This method preserves the characteristic spending patterns of the socio- economic groups as given by the 1988 survey results and shows how each group is affected by the changes in prices of ESGSs following an increase in environmental tax rates. The nominal expenditures by the groups can then be deflated by the price indices to give real expenditures by region and year.

III. ASSUMPTIONS FOR THE PROJECTIONS 1. Additional Excise Duties on Energy Products The European Commission (Press Release, Brussels, 12 March 1997, reported in the Financial Times) has proposed a Directive to widen the scope of the minimum rate system to include all energy products. This proposal combines two requirements made by the Council of Ministers: the first is to present new proposals for energy taxation following the failure of the carbon/energy tax proposal to be agreed; the second is to reduce distortions in the energy market caused by different rates of excise duties on fuels in different member states. The proposal is intended to give member states the flexibility to differentiate rates of taxation on the basis of environmental criteria, while complying with the minimum rates; it encourages member states to avoid increasing the tax burden and to use extra revenues to reduce taxes on labour. The new proposal extends the present framework of harmonising duties on oil products to cover coal, gas, electricity and other energy products. New minimum rates starting on the dates 1 January 1998, 1 January 2000 and 1 January 2002 are proposed for motor fuels and heating fuels. A wide range of exemptions is proposed: fuels used by airlines, products used on navigation within Community waters and special cases. Member states may choose to exempt renewable energy sources. Special refunds are proposed for firms whose energy costs are in

385 Fiscal Studies excess of 10 per cent and 20 per cent of production costs. The proposal is being considered by the Council of Ministers and governments but has yet to be adopted.

2. Treatment of Additional Energy Duties versus a Carbon Tax E3ME can include both a carbon/energy tax and excise duties. The main difference between them is the inclusion of a carbon tax in the basic prices of the energy industries. Excise duties on energy products are intended to be paid further down the chain of production, with the inputs into electricity and oil- refining untaxed, and electricity itself subject to tax; in this case, the basic prices of the energy industries are largely unaffected, except inasmuch as their labour costs and own use of electricity and fuel are taxed. The scenarios analysed in this paper consist of excise duties and not the carbon/energy tax. Additional excise duties on energy products have been treated in E3ME specifically as a new set of excise duties on purchases by industries and consumers of coal, oil products, gas and electricity, with the tax being assumed to be passed fully on to prices, the same treatment as for other excise duties.

3. Scenarios In the projections reported in this paper, a set of escalating additional excise duties for EU member states has been calculated for the period 1999–2010, graduated according to the carbon content of the energy products being taxed (with electricity taxed on the basis of the carbon content of its generating input fuels, but other exemptions and special treatment of the proposed Directive not included). These rates are assumed to escalate in constant absolute amounts each year and are calculated, in the main projection, to achieve an overall reduction of 10 per cent in CO2 emissions below base-case projections by 2010 for the total for all 11 member states included in the current version of the model (EU-11). (For a comparison, the 2010 rates of tax to achieve the 10 per cent reduction correspond to a rate of $16, 1999 prices, per barrel oil-equivalent (boe) compared with the $10 per boe, 1993 prices, for 2000 proposed by the Commission for a carbon/energy tax.) An indication of the magnitude of these additional taxes is also given by the ex-ante estimates of their revenues as a proportion of GDP in current prices. These estimates are shown in Table 2 for the years 2000, 2005 and 2010. For the EU-11, they rise from 0.4 per cent of GDP in the year 2000 to 2.2 per cent by 2010. These estimates can be compared with those of Smith (1994) for the proposed $10 per boe carbon/energy tax, which are also ex ante but ignore any exemptions or effects of economic growth and baseline changes in economic structure. Given these differences in treatment, the percentages are very similar, with the differences in rates between countries reflecting their CO2 emissions per unit of GDP and the exemptions of the additional excise duty proposal. The increases can also be compared with

386 Equity and Ecotax Reform in the EU

TABLE 2 Ex-Ante Revenues from Additional Excise Duties as a Percentage of GDP (current prices) Per cent 2000 2005 2010 Smitha $10/barrel tax Belgium 0.4 1.2 2.0 1.7 Denmark 0.5 1.4 1.9 1.0 France 0.3 0.8 1.4 1.0 Ireland 0.7 2.3 4.0 1.8 Italy 0.4 1.4 2.5 0.9 Luxemburg 0.8 2.4 3.6 2.6 Netherlands 0.4 1.5 2.4 1.6 Portugal 0.6 2.1 4.1 1.3 Spain 0.3 1.2 2.1 1.2 UK 0.6 2.2 4.0 1.3 West Germany 0.4 1.2 1.9 1.3

EU-11 0.4 1.3 2.2 aSmith (1994) does not specify a year. Sources: EU project PL950582 Environmental Fiscal Reform; Smith (1994, p. 282) quoted in OECD (1995). existing rates of tax on energy and labour. In 1993, the EU-15 levied taxes on production factors (labour, capital, energy and environment) of 42 per cent of GDP (European Environmental Agency, 1996, pp. 24–5); the taxes on energy were relatively small — some 3 per cent of GDP — while the taxes on labour were about 20 per cent of GDP. The energy tax share may have risen slightly since 1993, but the increase of some 2.2 percentage points represented by the additional excise duties required to reduce CO2 emissions by 10 per cent would be a significant switch in the sources of tax revenue, raising the energy share and reducing the labour share, if employers’ contributions were reduced to keep the switch tax-revenue-neutral. The revenues from these additional duties are assumed to be spent by each member state in reducing employers’ contributions to social security schemes so that the overall effect is to keep the total tax revenues from the duties and the contributions constant, i.e. the tax change is tax-revenue-neutral in this restricted sense; however, overall tax revenues and public sector borrowing may well change as other tax revenues and spending change in response to the changed economic conditions.

387 Fiscal Studies

This main projection requires a baseline projection for comparison and has been supplemented by three other projections which help to explain the results. E3ME has therefore been solved for: The baseline: A projection of business-as-usual growth in output, employment and CO2 emissions, with a calibration of the growth of energy demand to yield an overall increase in CO2 emissions of some 7 per cent above 1990 levels by 2010 for the 11 member states covered in the model (EU-11). Projection 1: Additional excise duties on energy products in proportion to their carbon content imposed at escalating rates from 1999 sufficient to reduce EU-11 CO2 emissions by 10 per cent below baseline by 2010. These duties are tax-revenue-neutral via reductions in employers’ contributions to social security. Projection 2: The additional duties of projection 1 but without any reductions in employers’ contributions. Projection 3: The additional duties of projection 1 but excluding additional duties on fuels for road transport; tax-revenue-neutral. Projection 4: The additional duties of projection 1 but excluding additional duties on fuels for heat and power (i.e. including additional duties on fuels for road transport only); tax-revenue-neutral. Projection 5: The additional duties of projection 1 but tax-revenue-neutral via recycling of revenues by an equal lump-sum payment to all households, weighted by the number of adult-equivalents and without any reductions in employers’ contributions. By comparing these projections with the baseline and with each other, an assessment can be made of the effects of the policies on income distributions, as well as on GDP, employment and CO2 emissions.

IV. EFFECTS ON GDP, EMPLOYMENT AND CO2 EMISSIONS Before presenting the detailed effects on income and expenditure by socio- economic group, it is helpful to give some macroeconomic results by member state. The top panel of Table 3 shows the effects of additional duties on energy products without recycling of the revenues, i.e. with the revenues being used to reduce public borrowing. The results clearly show small but mixed effects for GDP and employment and large effects (over 10 per cent overall) for CO2 emissions. The overall effects for GDP are positive but negligible; those for employment are negative. The reason for the mixed GDP results is that, in general, imports fall more than exports, because the EU is a substantial net importer of energy products bearing the higher taxes; gross output falls below the baseline in every member state, although the falls are most pronounced in the

388 Equity and Ecotax Reform in the EU energy industries. The outliers in GDP effects are the Netherlands (–2.0 per cent below base) and Belgium (+0.8 per cent). Exports from the Netherlands fall more than imports and the decline in GDP is exacerbated by a decline in investment in energy-intensive industries (oil and metals). Imports into Belgium fall more than exports and consumers’ expenditure is increased by wage rates being exogenous (an assumption imposed for Belgium and Luxemburg to reflect behaviour in these labour markets).

TABLE 3

Effects on GDP, Employment and CO2 Emissions, 2010 Additional Duties, without Recycling (Projection 2 – Baseline)

GDP Employment CO2 emissions Per cent Thousands Per cent Million tonnes Per cent of carbon Belgium 0.8 14.2 0.3 –6.3 –20.8 Denmark 0.0 1.3 0.0 –1.2 –7.6 France 0.7 –49.8 –0.2 –14.8 –14.3 Ireland –0.4 0.0 0.0 –1.3 –10.9 Italy 0.3 –75.4 –0.3 –14.3 –11.7 Luxemburg 0.3 0.4 0.2 –0.2 –11.3 Netherlands –2.0 –11.1 –0.2 –3.1 –6.6 Portugal –0.5 2.2 0.1 –0.8 –4.6 Spain –0.2 –60.6 –0.4 –8.7 –12.5 UK –0.5 –109.8 –0.4 –16.6 –10.0 West Germany 0.2 –199.8 –0.6 –17.3 –8.1

EU-11 0.1 –488.3 –0.3 –84.7 –10.6

389 Fiscal Studies

Additional Duties, with Recycling (Projection 1 – Baseline)

GDP Employment CO2 emissions Per cent Thousands Per cent Million tonnes Per cent of carbon Belgium 1.9 116.8 2.9 –6.1 –20.2 Denmark 1.3 14.8 0.6 –1.1 –7.3 France 1.4 202.1 0.7 –14.7 –14.2 Ireland 0.8 22.1 1.5 –1.3 –10.9 Italy 1.7 553.7 1.9 –13.1 –10.6 Luxemburg 1.0 2.9 1.2 –0.2 –10.7 Netherlands –0.8 69.7 1.2 –2.8 –5.8 Portugal 0.6 52.6 1.3 –1.0 –5.6 Spain 1.2 203.9 1.4 –8.0 –11.4 UK 1.8 426.7 1.4 –15.0 –9.0 West Germany 1.4 245.2 0.7 –16.8 –7.8

EU-11 1.4 1,910.4 1.2 –80.1 –10.0 Note: This and subsequent tables show results to various degrees of significance; this should not be taken to imply the accuracy of the projections. The bottom panel of Table 3 shows the effects on GDP, employment and CO2 emissions of the duties with revenues recycled by reducing employers’ contributions. For the EU-11, GDP is 1.4 per cent above base by 2010, employment 1.2 per cent above and CO2 emissions exactly 10 per cent below by design. The usual experience in most EU economies in the last 20 years has been ‘jobless’ growth, with employment growing much less than output. However, the projection has employment growing at a similar rate to output, through the effect of reductions in the costs of labour to employers. This causes firms to substitute labour for other inputs to production and also reduces total production costs, providing a stimulus to the economy. Furthermore, since labour taxes are often distortionary, a reduction in these taxes will reduce the distortion and provide a further stimulus. These factors explain the increase in GDP relative to the baseline. GDP increases in all regions except the Netherlands (which experiences some loss of international competitiveness), with the largest increase, of over 1.9 per cent (i.e. 0.2 per cent per annum 1999–2010), for Belgium. The ‘double dividend’ of increases in employment and reductions in emissions is evident for all the member states covered by the simulations. The increases in employment are very uneven across states, ranging from 0.6 per cent for Denmark and France to 2.9 per cent for Belgium. The Belgian result is partly due to the assumption that wage rates in the economy are rigid, so that reductions in non-wage costs have maximum effect on total labour costs. The reductions in CO2 are very similar to those in the top panel of the table, showing that they are the effect of the excise duties rather than of the reductions in employment taxes. Since the

390 Equity and Ecotax Reform in the EU economy is operating at a higher level of activity, the reduction in emissions is slightly lower for most countries. The estimates of positive double dividends agree with other empirical studies of carbon and energy tax in the EU (DRI, 1991 and 1994; Ekins, 1997); the theoretical literature of the topic is reviewed by Bohm (1997) who concludes that carbon taxes ‘cannot be regarded as distortionary in any conventional sense of the term’ (p. 122) and that if the tax revenues are used to reduce a tax on employment or the tax is an alternative to command-and-control policy, ‘there exists a double dividend’ (p. 121).

V. EFFECTS ON EXPENDITURE AND INCOME OF SOCIO- ECONOMIC GROUPS The effects of additional ecotaxes on vulnerable socio-economic groups come from: • the increases in prices of the goods and services they buy; • the increase in their income from the use of the revenues from the taxes; • the reductions in environmental damages they experience; and • the indirect effects from the interactions of the policies on the economy and the environment. The immediate impact of additional duties on energy products is to increase prices of domestic energy, as shown in the middle panel of Table 4. However, although the price of householder road fuels also increases, the effects on householder transport and overall average price levels (shown in the last and first panels of Table 4 respectively) are mixed, because the reductions in employers’ contributions are passed on by employers in the form of lower prices of all goods and services bought by consumers. These include prices of new vehicles, which form a large proportion of households’ expenditure on transport. For seven of the 11 countries (not Belgium, France, the Netherlands and Portugal), the domestic energy price increases are weakly progressive, with the higher expenditure groups facing the higher increase in prices because they have a higher share of electricity in their fuel use (and, in many countries, electricity is generated using coal). However, the increases in prices of householder transport are strongly regressive, i.e. the lower expenditure groups tend to face larger price increases. This result is, perhaps, counter-intuitive. It happens because the increases in prices of vehicle fuels are outweighed by reductions in prices of air transport and new vehicles (as a result of the reductions in employment costs). Spending on these is much more progressive than spending on vehicle fuels, which does not exhibit a strong trend across the different socio-economic groups (Köhler, Luhmann and Wadeskog, 1997).

391 Fiscal Studies

In order to assess the impacts on real incomes, the real expenditures on these goods and services and the nominal incomes of the groups have to be taken into account. The real expenditures vary strongly across expenditure groups, as shown in the earlier discussion of the data, with domestic energy spending being regressive and spending on householder transport being progressive. The effects of the increases in nominal incomes across expenditure groups will offset these increases in nominal expenditures, depending on the sources of income and the increases in occupational wage rates of the wage earners in the different groups. Assumptions affecting these income changes are made in the projections: namely, (a) that social transfers and pensions will be increased at the same rate as the increase in wage rates, (b) that the reductions in employers’ contributions are across-the-board and not targeted at the lower-paid and (c) that the proportions of income coming from wages and salaries do not alter with the level of unemployment. With these assumptions, the overall effects on the real personal disposable incomes (real PDIs) of the socio-economic groups can be estimated. The results are presented for the differences between projection 1 and the baseline in the first panel of Table 5. Note that since no estimates of the numbers of households or adult-equivalents are given for each group, it is not possible to provide any

392 TABLE 4 Changes in Projection 1 Prices for Socio-Economic Group Expenditure, 2010 (percentage difference from baseline) Change in Overall Average Prices Per cent All Expenditure group Social Pensions households with proportion of average expenditure: transfers <0.4 0.4–0.6 0.6–0.8 0.8–1.2 1.2–1.6 >1.6 Belgium –1.05 –0.23 –0.68 –0.87 –1.14 –1.27 –1.57 –0.64 –0.86 Denmark –0.04 0.41 0.10 0.10 –0.05 –0.10 –0.13 0.08 0.10 France –0.54 –0.52 –0.54 –0.54 –0.53 –0.57 –0.56 –0.52 –0.48 Ireland 1.61 2.67 2.38 2.09 1.70 1.29 0.94 2.57 2.00 Italy –1.34 –1.24 –1.20 –1.23 –1.28 –1.34 –1.50 –1.22 –1.16 Luxemburg –0.78 –0.65 –0.76 –0.77 –0.77 –0.79 –0.85 –0.53 –0.79 Netherlands –0.16 0.68 0.12 –0.05 –0.15 –0.26 –0.35 0.14 –0.10 Portugal –0.63 –0.01 –0.52 –0.64 –0.53 –0.70 –0.76 –0.64 –0.33 Spain –0.16 0.67 0.28 0.01 –0.17 –0.34 –0.56 0.14 0.19 UK 1.77 1.93 2.25 1.94 1.65 1.61 1.14 2.44 1.71 West Germany –1.13 0.39 0.25 –0.05 –0.62 –1.02 –1.20 –0.59 –0.74 Change in Domestic Energy Prices Per cent All Expenditure group Social Pensions households with proportion of average expenditure: transfers <0.4 0.4–0.6 0.6–0.8 0.8–1.2 1.2–1.6 >1.6 Belgium 25.56 25.83 26.11 25.80 25.06 25.30 24.72 25.82 26.25 Denmark 13.97 13.66 13.17 14.12 13.78 14.42 14.13 14.77 12.39 France 6.31 6.24 6.33 6.16 6.34 5.93 5.92 5.87 6.88 Ireland 33.71 33.15 33.31 33.49 33.81 33.91 34.13 33.57 32.83 Italy 16.30 13.56 14.52 15.17 16.25 17.48 18.94 18.08 16.71 Luxemburg 10.90 10.05 10.49 10.49 10.92 11.34 11.43 11.73 11.69 Netherlands 22.68 22.95 22.88 22.65 22.63 22.76 22.49 22.03 22.37 Portugal 11.15 11.15 11.15 11.15 11.15 11.15 11.15 11.15 11.15 Spain 32.45 31.71 31.68 31.94 32.45 32.80 33.37 32.35 33.41 UK 24.47 21.03 23.56 25.12 25.20 26.40 24.86 22.96 21.43 West Germany 17.51 16.64 16.64 16.64 16.65 16.69 16.77 16.68 16.85 Change in Householder Transport Prices Per cent All Expenditure group Social Pensions households with proportion of average expenditure: transfers <0.4 0.4–0.6 0.6–0.8 0.8–1.2 1.2–1.6 >1.6 Belgium 0.03 0.86 0.68 0.49 0.31 –0.27 –0.41 0.52 –0.06 Denmark 1.37 3.83 2.27 2.78 1.81 1.00 0.65 2.40 2.52 France 0.99 1.75 1.87 1.67 1.05 0.66 0.59 1.45 1.24 Ireland 2.07 5.24 4.25 4.08 2.80 1.62 1.00 3.64 1.87 Italy 0.35 3.14 3.14 3.01 2.61 1.38 –0.91 0.51 1.37 Luxemburg 0.21 1.89 0.87 0.53 0.25 0.05 0.02 0.95 –0.07 Netherlands 0.71 2.12 1.01 1.16 0.80 0.59 0.31 0.66 0.62 Portugal 1.38 3.31 3.60 3.40 3.96 2.46 0.37 0.91 4.45 Spain –0.37 0.49 0.36 0.04 0.02 –0.55 –0.77 –0.07 –0.07 UK 4.70 7.49 7.03 6.70 6.54 6.55 2.60 7.43 5.34 West Germany –0.83 1.11 0.38 –0.34 –0.85 –1.06 –1.18 0.04 –0.78 TABLE 5 Changes in Real Personal Disposable Income, 2010 (percentage difference from baseline) Change in Real PDIs: Additional Duties (Projection 1 – Baseline) Per cent All Expenditure group Social Pensions households with proportion of average expenditure: transfers <0.4 0.4–0.6 0.6–0.8 0.8–1.2 1.2–1.6 >1.6 Belgium 3.65 2.96 3.36 3.51 3.75 3.88 4.17 3.25 3.47 Denmark 0.82 0.42 0.71 0.70 0.84 0.88 0.91 0.71 0.68 France 0.81 0.80 0.81 0.81 0.81 0.84 0.83 0.79 0.76 Ireland 2.71 1.67 1.96 2.24 2.62 3.03 3.37 1.78 2.36 Italy 2.42 2.31 2.28 2.31 2.37 2.42 2.58 2.33 2.23 Luxemburg 1.82 1.72 1.82 1.82 1.81 1.83 1.90 1.58 1.83 Netherlands 1.54 0.73 1.29 1.45 1.54 1.65 1.74 1.26 1.49 Portugal 1.34 0.73 1.23 1.34 1.24 1.40 1.46 1.34 1.05 Spain 0.84 0.01 0.40 0.66 0.84 1.01 1.23 0.53 0.48 UK 3.33 3.18 2.86 3.17 3.45 3.49 3.95 2.69 3.42 West Germany 2.01 0.53 0.67 0.95 1.51 1.91 2.08 1.48 1.63 Change in Real PDIs: Additional Duties excluding Road Fuels (Projection 3 – Baseline) Per cent All Expenditure group Social Pensions households with proportion of average expenditure: transfers <0.4 0.4–0.6 0.6–0.8 0.8–1.2 1.2–1.6 >1.6 Belgium 3.28 2.54 2.96 3.14 3.38 3.51 3.78 2.89 3.07 Denmark 0.68 0.23 0.54 0.54 0.70 0.75 0.78 0.56 0.50 France 0.42 0.34 0.40 0.42 0.42 0.44 0.45 0.43 0.36 Ireland 2.48 1.35 1.69 2.01 2.40 2.78 3.13 1.55 2.13 Italy 1.87 1.72 1.77 1.80 1.85 1.88 1.99 1.79 1.77 Luxemburg 1.42 1.31 1.38 1.40 1.42 1.44 1.50 1.29 1.37 Netherlands 1.38 0.55 1.10 1.27 1.38 1.48 1.58 1.08 1.31 Portugal 1.24 0.40 0.98 1.12 1.19 1.28 1.40 1.28 1.30 Spain 0.05 –0.77 –0.37 –0.11 0.09 0.22 0.41 –0.22 –0.27 UK 2.85 2.64 2.35 2.69 2.99 3.02 3.43 2.15 2.94 West Germany 1.35 –0.19 –0.04 0.27 0.85 1.25 1.42 0.83 0.95 Change in Real PDIs: Road Fuel Duties (Projection 4 – Baseline) Per cent All Expenditure group Social Pensions households with proportion of average expenditure: transfers <0.4 0.4–0.6 0.6–0.8 0.8–1.2 1.2–1.6 >1.6 Belgium 0.39 0.43 0.42 0.40 0.39 0.39 0.41 0.38 0.42 Denmark 0.24 0.32 0.28 0.27 0.24 0.23 0.23 0.26 0.30 France 0.39 0.45 0.41 0.38 0.38 0.40 0.38 0.35 0.40 Ireland 0.23 0.32 0.27 0.23 0.22 0.24 0.24 0.23 0.23 Italy 0.56 0.60 0.52 0.51 0.53 0.54 0.60 0.55 0.47 Luxemburg 0.40 0.41 0.43 0.42 0.39 0.39 0.39 0.28 0.46 Netherlands 0.17 0.18 0.19 0.18 0.16 0.17 0.16 0.18 0.18 Portugal 0.11 0.33 0.26 0.24 0.06 0.12 0.07 0.07 –0.25 Spain 0.80 0.79 0.78 0.79 0.76 0.81 0.83 0.76 0.76 UK 0.50 0.55 0.52 0.49 0.48 0.49 0.54 0.56 0.50 West Germany 0.67 0.72 0.71 0.69 0.67 0.67 0.68 0.66 0.69 Fiscal Studies

EU-11 totals, except for all households in the EU-11 which are calculated to have an increase of real PDI of 1.6 per cent above base by 2010. The most striking result is that every expenditure group in every member state in the analysis benefits from the tax shift, with the extent of the benefit ranging from the lowest of 0.01 per cent for real PDI in households with under 0.4 of mean expenditures in Spain to the highest of 4.17 per cent for the highest expenditure group in Belgium. The results for vulnerable groups (‘social transfers’ and ‘pensions’) are similar to those for the bottom two expenditure groups, reflecting their relatively low expenditures and incomes. In all countries, vulnerable groups benefit, but generally less than the average of all households. The differences between countries tend to be much larger than the differences between expenditure groups in the same country.

TABLE 6

The Distributional Effects of Taxes to Cut CO2 Emissions: Ratios of Regressivity With recycling No recycling Overall tax Domestic Petrol tax only Lump-sum Excise duties energy tax only recycling only Belgium 1.008 1.008 1.000 0.969 1.011 Denmark 1.002 1.002 1.000 0.919 1.004 France 1.000 1.000 1.000 0.997 1.002 Ireland 1.014 1.014 1.000 0.990 1.014 Italy 1.003 1.002 1.001 1.000 1.003 Luxemburg 1.001 1.001 1.000 0.940 1.002 Netherlands 1.004 1.005 1.000 0.967 1.006 Portugal 1.002 1.004 0.998 0.991 1.003 Spain 1.008 1.008 1.000 0.999 1.009 UK 1.011 1.011 1.000 0.993 1.013 West Germany 1.014 1.015 1.000 0.973 1.013 Notes: Ratio of regressivity calculated as percentage change in personal disposable income of >1.6 expenditure group over the percentage change in personal disposable income of 0.4–0.6 expenditure group; a value greater than 1 indicates regressivity. Projections: Overall tax — Projection 1: duties on carbon to reduce CO2 emissions by 10 per cent by 2010; recycling by reductions in employers’ social security contributions to make duties revenue- neutral. Domestic energy tax — Projection 3: duties on domestic energy as in overall tax, no duties on vehicle fuels; recycling by reductions in employers’ social security contributions to make duties revenue-neutral. Petrol tax — Projection 4: duties on vehicle fuels as in overall tax, no duties on domestic energy; recycling by reductions in employers’ social security contributions to make duties revenue-neutral. Lump-sum recycling — Projection 5: the additional duties of projection 1, but with recycling of revenues via lump-sum payments to households instead of the reduction in employers’ contributions. Excise duties only — Projection 2: the additional duties of projection 1, but without any reduction in employers’ contributions.

398 Equity and Ecotax Reform in the EU

The middle and last panels of Table 5 show the two components of the overall change — the increases in duties excluding those on petrol and diesel, and the increases in duties on petrol and diesel alone; in each case, the revenues are recycled via reductions in employers’ contributions. The purpose of showing these results is to demonstrate that the overall result of weak regressivity is due to the increases in taxes on domestic energy; if road fuels are taxed on their own, the outcome is weakly progressive for most EU economies. The much smaller effects for petrol and diesel are due to the fact that these fuels have much smaller carbon contents in relation to their costs to the final purchaser than most fuels, so an additional tax related to their carbon content has a smaller effect in terms of increases in their percentage cost to final users. Table 6 brings together the results for regressivity by presenting ratios of indices for the top expenditure group to the second lowest (expenditures 0.4–0.6 of the mean); the lowest expenditure group appears to contain very few households for many countries in the EUROSTAT sample, so results for this group are regarded as too erratic and unrepresentative. The indices are the percentage changes given in Table 5 plus 100 to transform the negative changes. A regressive effect is shown by the index for regressivity in Table 6 being above 1.000. Results are presented for the three projections in Table 5 plus the projection with recycling of revenues via lump-sum payments to households and the projection without recycling, showing the effects of the additional duties. The results are dominated by the effects of domestic energy taxes, which are weakly regressive; there is no overall pattern if road fuel duties alone are imposed. Comparing the projection without recycling in the last column to the overall tax with recycling in the first column, it can be seen that the effect of recycling by means of reductions in employers’ social security contributions slightly reduces the ratios of regressivity. Overall, it can be seen that the package of measures that is examined here is regressive across expenditure groups, but only to a small degree. In comparison, recycling of revenues via lump-sum payments to households can be seen to have a progressive effect. However, such a scheme would be unlikely to be adopted and, if used instead of a reduction in labour taxes, would reduce the overall economic benefits of the tax reform.

VI. CONCLUSIONS This study is intended to contribute to the literature on the equity effects of carbon taxation in the EU. In comparison with the earlier studies (Smith, 1992; CEC, 1992), • it uses more recent household survey data for 1988; • it is a comprehensive analysis including indirect effects of intermediate demand, consumption, investment and trade, as well as employment, prices and wage rates;

399 Fiscal Studies

• it provides a forward-looking, fully dynamic, integrated analysis for 11 member states of the EU year by year to 2010; • it is based on the additional excise duties proposed for energy products in March 1997; and • it is tax-revenue-neutral, not through lump-sum repayments to consumers, but through reductions in employers’ contributions to social security schemes. In order to reach the results, a set of assumptions have been made, the most important of which are: that the real expenditure characteristics of the socio- economic groups of the study vary in the same way as those for all households, country by country, particularly the responses to changes in relative prices of energy products; that the proportions of income from different sources for each group remain constant; and that social security and other transfers are increased in line with any increase in wage rates as a result of the tax changes. The main findings of the study are: • In all countries, the tax changes lead to an increase in personal disposable income from the baseline case for all socio-economic groups; there is also an increase in employment in all countries. • The eventual effects on the distribution of incomes associated with carbon taxation in the EU are not nearly so regressive as the initial impact on expenditure suggests. • The most regressive impact is on West Germany, the UK and Ireland, although the impact is weak. • The regressive effect comes from the impact of taxation on domestic energy consumption. • If revenues are recycled through reductions in employer taxes, all member states and all socio-economic groups in the study experienced an increase in real disposable income; the UK, Italy, Belgium, Ireland and West Germany stand to gain the most, but the largest gain is only an increase in the growth rate of 0.3 per cent a year for Belgium over 1999–2010. • The scale of the revenues is such that the changes can easily have a progressive effect, but this is not likely to be via the market responses and is only likely to take place by deliberate policy — for example, (a) targeting reductions in employer taxes on employment of the lower-paid, (b) using some of the revenues to improve the energy efficiency of domestic fuel use by lower income groups or (c) using the revenues to raise incomes of vulnerable groups directly via social security payments. To summarise, the results are dominated by the overall effects of the fiscal change on income and employment. By 2010, real personal disposable incomes are 1.6 per cent above baseline and employment is 1.2 per cent above, assuming that the change is tax-revenue-neutral, with employers’ contributions to social

400 Equity and Ecotax Reform in the EU security schemes being reduced to compensate for the additional revenues from excise duties. Governments are assumed to index social security benefits and state pensions in line with the increase in wage rates following the change. However, without deliberate further policy changes to redirect the extra incomes towards the more vulnerable socio-economic groups, the change will be weakly regressive for nearly all the member states in the study. Although low- expenditure households might be expected to experience an increase in real income following this limited ecotax reform, high-expenditure households experience an even greater increase.

REFERENCES

Barker, T. and Gardiner, B. (1996), ‘Employment, wage formation and pricing in the European Union: empirical modelling of environmental tax reform’, in C. Carraro and D. Siniscalco (eds), Environmental Fiscal Reform and Unemployment, Dordrecht: Kluwer. — and Köhler, J. (eds) (1998), International Competitiveness and Environmental Policies, Cheltenham: Edward Elgar. — and Lewney, R. (1991), ‘A green scenario for the UK economy’, in T. Barker (ed.), Green Futures for Economic Growth, Cambridge: Cambridge Econometrics. Bohm, P. (1997), ‘Environmental taxation and the double dividend’, in T. O’Riordan (ed.), Ecotaxation, London: Earthscan. Bundesministerium für Wirtschaft (1996), Energiedaten ’96. Nationale und internationale Entwicklung, Bonn: Bundesministerium für Wirtschaft. Cappelen, Å. (1991), ‘MODAG: a medium term macroeconometric model of the Norwegian economy’, Oslo, Central Bureau of Statistics, Discussion Paper no. 67. Central Statistical Office (1990), Family Expenditure Survey 1988, London: HMSO. Commission of the European Communities (1992), ‘European economy: the climate challenge — economic aspects of the Community’s strategy for limiting CO2 emissions’, European Economy, no. 51. — (ed.) (1993), HERMES: Harmonised Econometric Research for Modelling Economic Systems, Amsterdam: North Holland. — (1995), E3ME: An Energy–Environment–Economy Model for Europe, EUR 16715 EN, Luxemburg: Office for Official Publications of the European Communities. — (1998), The Single Market Review, Subseries VI, Volume 4: Employment Trade and Labour Costs in Manufacturing, (by Cambridge Econometrics), Luxemburg: Office for Official Publications of the European Communities. Cornwell, A. and Creedy, J. (1996), ‘Carbon taxation, prices and inequality in Australia’, Fiscal Studies, vol. 17, no. 3, pp. 21–38. — and — (1997), Environmental Taxes and Economic Welfare, Cheltenham: Edward Elgar. Cramton, P. and Kerr, S. (1997), ‘How and why to auction CO2 emissions permits’, University of Maryland, Working Paper. Dower, R. C. and Zimmerman, M. B. (1992), The Right Climate for Carbon Taxes: Creating Economic Incentives to Protect the Atmosphere, Washington DC: World Resources Institute.

401 Fiscal Studies

Data Resources Incorporated (DRI) (1991), The Economic Impact of a Package of EC Measures to Control CO2 Emissions, Final Report prepared for the Commission of the European Communities, November. — (1992), Impact of a Package of EC Measures to Control CO2 Emissions on European Industry, Report prepared for the European Commission. — (1994), Potential Benefits of Integration of Environmental and Economic Policies, Report prepared for the European Commission, DG XI, Dordrecht: Graham and Trotman and Kluwer. EFILWC (1996), Environmental Taxes and Charges: National Experiences and Plans, Dublin: European Foundation for the Improvement of Living and Working Conditions; Luxemburg: Office for Official Publications of the European Communities. Ekins, P. (1997), ‘On the dividends from environmental taxation’, in T. O’Riordan (ed.), Ecotaxation, London: Earthscan. European Environmental Agency (1996), Environmental Taxes: Implementation and Environmental Effectiveness, Luxemburg: Office for Official Publications of the European Communities. EUROSTAT (1992), Family Budgets: Comparative Tables, ‘Denmark, Greece, France, Ireland, Luxemburg, Netherlands’, Luxemburg: EUROSTAT Publications. — (1993), Family Budgets: Comparative Tables, ‘Belgium, Germany, Spain, Italy, Portugal, UK’, Luxemburg: EUROSTAT Publications. Hamilton, K. and Cameron, C. (1994), ‘Simulating the distributional effects of a carbon tax’, Canadian Public Policy, vol. 20, pp. 385–99. Johnson, P., McKay, S. and Smith, S. (1990), The Distributional Consequences of Environmental Taxes, Commentary no. 23, London: Institute for Fiscal Studies. Köhler, J. (1998), ‘Imputed data for expenditure and income in the E3ME model’, University of Cambridge, Department of Applied Economics, Environmental Fiscal Reform Project, Working Paper no. 11. Köhler, J., Luhmann, H.-J. and Wadeskog, A. (1997), ‘Expenditure on environmentally sensitive goods and services: household spending across the EU’, University of Cambridge, Department of Applied Economics, Environmental Fiscal Reform Project, Working Paper no. 1. OECD (1995), Climate Change, Economic Instruments and Income Distribution, Paris: Organisation for Economic Co-operation and Development. Poterba, J. (1991), ‘Tax policy to combat global warming: on designing a carbon tax’, in R. Dornbusch and J. M. Poterba (eds), Global Warming: Economic Policy Responses, Cambridge, Mass.: MIT. Smith, S. (1992), ‘The distributional consequences of taxes on energy and the carbon content of fuels’, European Economy, Special Edition no. 1: The Economics of Limiting CO2 Emissions, pp. 241–68. — (1994), ‘Who pays for climate change policies? Distributional side-effects and policy responses’, in The Economics of Climate Change: Proceedings of an OECD/IEA Conference, Paris: Organisation for Economic Co-operation and Development. Statistisches Bundesamt (1994), Einkommens- und Verbrauchsstichprobe (Income and Consumption Survey) 1988, Fachserie 15, Heft 5: Aufwendungen privater Haushalte für den Privaten Verbrauch, Wiesbaden: Statistisches Bundesamt. Symons, E., Proops, J. and Gay, P. (1994), ‘Carbon taxes, consumer demand and carbon dioxide emissions: a simulation analysis for the UK’, Fiscal Studies, vol. 15, no. 2, pp. 19–43.

402