Department of the Environment

About us | Contact us | Publications

About us header images - leftAbout us header images - centreAbout us header images - right

Publications archive

Disclaimer

Key departmental publications, e.g. annual reports, budget papers and program guidelines are available in our online archive.

Much of the material listed on these archived web pages has been superseded, or served a particular purpose at a particular time. It may contain references to activities or policies that have no current application. Many archived documents may link to web pages that have moved or no longer exist, or may refer to other documents that are no longer available.

Taxation and the Environment

Environmental Economics Seminar Series
Department of the Environment, Sport and Territories
March 1996
ISBN 0 642 24878 8

Background paper

Mick Common
Centre for Resource and Environmental Studies
Australian National University

Introduction

The idea of taxing environmentally damaging activities, such as pollution generation, in order to reduce their scale has a long history in economics. So does investigation of the question of how government might best raise the revenue it needs to finance its expenditures. Until recently these were seen as quite separate issues, but now they are increasingly seen as linked, especially in relation to the environmental impacts of fossil fuel use.

This paper is intended to raise some of the issues that it is hoped that the seminar will address. The next section reviews the main features of the theoretical literature on taxation as an instrument for pollution control. Section 3 looks at some of the issues that arise when environmental damage is considered as a potential tax base. Such consideration has been largely focussed on carbon dioxide emissions from fossil fuel combustion, and section 4 reviews some features of the recent Australian debate on carbon taxation. Section 5 offers some concluding comments.

Taxation as an instrument of environmental policy

Environmental economics books (for example Baumol and Oates 1988, Common 1988, Tietenberg 1992) typically introduce pollution as

an example of market failure, due to external effects where private property rights are incomplete. The market failure can, it is argued, be corrected by taxing emissions from all sources at a rate per unit equal to the marginal external cost arising at the allocatively efficient level of total emissions. An alternative approach, deriving from the Coase theorem, is to create the missing private property rights. This would avoid the problems involved in determining the proper rate for (Pigouvian) taxation. Since these problems are formidable, it is unfortunate that the property rights creation approach cannot work where the pollution problem is public in character - which is the typical situation. An alternative property rights approach, which is not vitiated by the publicness of the damage, involves identifying the efficient total of emissions, and then issuing tradeable emissions permits in amount equal to the efficient total.

When economists talk about the 'optimal' level of pollution, what they mean is the level at which the marginal costs and benefits of emissions abatement are equal. If an Environmental Protection Agency, EPA, is to identify this allocatively efficient level of pollution as its target for emissions control, it needs to know the marginal costs and benefits of abatement. It can then determine either the tax rate to set or the quantity of permits to create. Generally this information is not available to it. While economists have devoted a lot of time and effort to devising means for estimating the marginal costs and, especially, benefits of pollution abatement, most recognise, with regret, that control to optimal levels is not feasible. The policy relevant questions are about methods of control to achieve what economists call 'arbitrary standards'. This terminology merely means that the policy target does not derive from a precise balancing of the marginal costs and benefits of abatement. It does not, necessarily, mean that it is arbitrary in the sense of having been adopted impulsively or capriciously. Arbitrary standards, in this sense, may be the result of a great deal of scientific research and political deliberation.

In recent years, consequently, attention has largely focussed on the question of the 'best' instrument for pollution control to arbitrary standards. Much of this analysis also applies to control to optimal standards. The literature arising is extensive (see Cropper and Oates 1992). What follows is merely an overview of some of the core propositions which inform economists' views (see, for example, Common 1988 for a fuller introductory treatment).

The major classes of instrument to be considered here are:

l. Command and control, also referred to as regulation, or direct control, of emissions source behaviour. This can take two main forms:

2. Price incentive modification. Three principal options exist here:

a. emissions taxation

b. tradeable emissions permits

c. payment for emissions abatement

Economists display a strong presumption in favour of price incentive instruments, especially a and b. This derives from analysis of theoretical models of the behaviour of firms discharging wastes, as follows. Suppose that there are a number of firms in an area discharging wastes of some kind, that an EPA is established with all necessary powers, and that, after due deliberation, its objective is an arbitrary x per cent cut in total emissions. Now assume:

(i) that all of the firms are technically efficient cost minimisers;

(ii) that all of the firms are price-takers;

(iii) that the environmental protection agency has complete information on the costs of abatement in each firm;

(iv) that the input and output prices facing all of the firms are determined in competitive markets; and

(v) that the EPA can costlessly monitor emissions, and costlessly enforce compliance.

Given these assumptions, the least cost theorem says that total abatement costs are minimised where the EPA taxes emissions from all firms at a uniform rate per unit. Actually, given these assumptions, this theorem does not, as it is widely understood to, establish a presumption in favour of emissions taxation over command and control of the form that specifies allowable emissions levels. The EPA could achieve the same target at the same minimum cost by: paying the firms at a uniform rate per unit abatement; setting the emissions level for each firm; creating emissions permits tradeable as between firms with the total amount of emissions permitted equal to the arbitrary standard for total emissions.

The real problem with the command and control option is in regard to assumption (iii). If regulation is to involve the same abatement cost total as uniform taxation, then the levels of abatement enforced by the EPA will have to differ across firms such that more is done by those firms where abatement costs are lower. For the EPA to be able to issue the individually tailored regulations involved, it would have to know the marginal abatement cost function for each firm. This is generally infeasible, and the operational version of command and control is seen as involving each firm cutting back emissions by the same x% so as to achieve an overall reduction. Because firms' abatement cost functions differ, uniform percentage cutbacks would not be efficient.

The least cost property of uniform emissions taxation derives from the fact that it loads total abatement across firms such that those for which it costs less do more. Each firm would abate up to the level where its marginal abatement cost was equal to the tax rate, so as to minimise its costs inclusive of tax. As a result, all firms would be operating with the same marginal abatement cost. Given this, there would be no reallocation of the total abatement cost as between firms that could reduce the total cost of abatement. The EPA does not need to know each firm's abatement cost function to identify the differential abatement targets for each firm that minimise total cost. The reaction of the firms to the uniform tax identifies the least cost solution.

However, while it is dropping assumption (iii) that gives uniform taxation the cost advantage over command and control, doing this also generates a problem for uniform taxation. This is

that without this assumption, uniform taxation is not guaranteed to achieve the target level of total abatement, so it is not dependable. It will be dependable if the EPA sets the right uniform tax rate. But, in order to do this it needs to know the abatement cost function for every firm. In the absence of such information, it can set a tax rate which will achieve some overall abatement level at the lowest cost that is possible. But, the achieved level will not, except by chance, be the x% reduction in total emissions desired. The EPA could proceed by trial and error, setting a tax rate, observing the outcome, and adjusting the rate up or down as necessary. This would involve additional adjustment costs.

Tradeable permits have the attraction that where assumption (iii) is not applicable, they are both least cost and dependable. They are dependable because the total quantity of permits created by the EPA is equal to the emissions total that corresponds to the desired x% cutback. They are least cost because the market in permits, allocates them to firms where it is cheaper to buy a permit than to abate. Firms where abatement is relatively cheap will abate rather than buy permits. The loading of total abatement across firms so as to minimise total cost is automatically generated by the cost minimising behaviour of the firms. This outcome is crucially dependent on assumptions (ii) and (v). Where there are few firms, one may be able to exercise power in the permit market and distort its operation. Monitoring firms' compliance, ensuring that their actual emissions correspond to their permit holding, may involve the EPA in substantial costs. Note that such potential problems aside, the dependability and least cost properties of permits hold irrespective of how the permits are initially allocated by the EPA.

If the number of polluting firms is assumed constant, EPA payment for abatement at a uniform rate involves the same abatement costs as uniform taxation. The difference between the two instruments is in the direction of flow for transfer payments. Taxation generates EPA/government revenue, subsidisation is a charge on revenue. The distinction between abatement costs and transfer payments is important. Abatement costs are real social opportunity costs. Transfer payments are re-allocations of command over resources. Affected firms, of course, see tax payments as costs, and generally oppose taxation as a control instrument. Economists, on the other hand, generally oppose subsidisation of pollution abatement.

If assumption (v) is dropped, then a proper analysis of the costs associated with alternative instruments should account for monitoring and enforcement costs. In some pollution problems, such as those arising from emissions from fossil fuel combustion, monitoring and enforcement costs will be lower if some input is used as the point of control, rather than emissions as such. It has been shown (Common 1977) that if the relationships between the input level and emissions are known to the EPA, the least cost property of uniform emissions taxation carries over to input taxation. This is a reasonable assumption in the case of, for example, carbon dioxide releases in fossil fuel combustion.

The economic analysis of the choice of instrument for pollution control to arbitrary standards typically uses only the criteria of cost minimisation and dependability. Clearly there are a number of other relevant criteria for the evaluation of alternative instruments (see, for example, Bohm and Russell 1985). One might be interested in equity, for example. The least cost property of taxation/tradeable permits does not imply that these instruments are generally equitable.

Many non-economists often seem attracted to a class of instrument not listed above, which economists refer to as 'moral suasion'. This class is a mixed bag, but generally it involves attempts to change the information and/or preferences of economic agents, without also adopting command and control measures or changing price incentive structures. Examples are: subsidisation of research and development for alternative technologies; finance of basic research; finance of campaigns to raise public awareness; product labelling requirements; voluntary EPA-emissions source agreements on emissions targets. Economists are generally dismissive of moral suasion approaches, though there appears to have been relatively little systematic research on their cost-effectiveness in relation to other instruments. Pollution control practice has largely ignored the recommendations of economists in favour of price incentive systems, and has relied dominantly on command and control, especially in the form of the regulation of process and/or equipment. A recent major survey of environmental economics asked the question: what has it contributed to the design and implementation of policy? The answer given was:

We have seen some actual programs of transferable emissions permits in the United States and some use of effluent charges in Europe. And with the enactment of the 1990

amendments to the Clean Air Act, the US has introduced a major program of tradeable allowances to control sulfur emissions - moving this country squarely into the use of incentive-based approaches to regulation in at least one area of environmental policy. But, at the same time, effluent charge and marketable permit programs are few in number and often bear only a modest resemblance to the pure programs of economic incentives supported by economists (Cropper and Oates 1992 p729).

Interest in price incentives for pollution control appears to be growing. The authors of this survey state that:

it is our sense that we are at a point in the evolution of environmental policy at which the economics profession is in a very favourable position to influence the course of policy. As we move into the 1990s, the general political and policy setting is one that is genuinely receptive to market approaches to solving our social problems. Not only in the United States but in other countries as well, the prevailing atmosphere is a conservative one with a strong disposition toward the use of market incentives, wherever possible, for the attainment of our social objectives (Cropper and Oates l992 p730).

James (1993) surveys the situation in Australia in relation to the use of economic instruments for environmental objectives.

The environment as tax base

The literature on taxation as an instrument for environmental policy generally ignores the use of the revenues arising, at least in the formal analysis (see Baumol and Oates l988). A recent major survey (Cropper and Oates l992), for example, notes that emissions taxes are 'sources of public revenue' which 'in these days of large budget deficits...promise a new revenue source to hard-pressed legislators', but does not explore the matter beyond a few illustrative figures. It is also the case that until recently the public finance literature paid rather little attention to damage to the environment as a potential tax base (see for example Atkinson and Stiglitz 1980, a standard graduate text). There is not, that is, a well established literature integrating environmental and taxation concerns on which economists can draw for policy advice.

The last few years have seen the beginnings of such a literature, given increasing awareness of the pervasiveness of environmental externalities, and the revenue potential arising thereby. Interest has centred on the environmental damages associated with fossil fuel use, and latterly on carbon dioxide emissions especially. This interest has not been confined to academic circles, nor has advocacy of environmental damage as a tax base been confined to radical commentators. In the last five years, for example, The Economist has carried leaders as well as features arguing for energy/carbon as a tax base.

The basic appeal of environmental damage as a tax base is frequently referred to as a double dividend. The idea here is that a revenue neutral tax base shift to environmental damage can offer both environmental improvements and reduced welfare costs of taxation. Working within the public finance/optimal taxation framework, Goulder (1994, copies available on request) distinguishes three double dividend propositions:

Weak Form: By using revenues from the environmental tax to finance reductions in marginal rates of an existing distortionary tax, one achieves cost savings relative to the case where the tax revenues are returned to taxpayers in lump-sum fashion.

Intermediate Form: It is possible to find a distortionary tax such that the revenue neutral substitution of the environmental tax for this tax involves a zero or negative gross cost.

Strong Form: The revenue-neutral substitution of the environmental tax for typical or representative distortionary taxes involves a zero or negative gross cost (emphasis in the original).

Here 'gross cost' means the welfare cost of the tax initiative when no account is taken of any welfare benefits arising as the result of environmental impacts due to the tax base switch. Hence, where the gross cost is zero, the welfare change is positive so long as there is some welfare benefit from environmental impact. Goulder puts these propositions this way in order to highlight what he sees as the policy appeal of double dividend propositions - if true. they indicate that there is a welfare case for environmental taxation, without requiring that the size of the environmental welfare benefits be established at all precisely. All that is necessary is to know that these benefits have a positive sign. Given the difficulties of measuring these benefits, this is seen as a useful way of formulating the problem facing decision makers.

Goulder notes that the weak form of the double dividend proposition is, within the optimal taxation framework of analysis, effectively true by definition. It is also, from a policy perspective, largely uninteresting, as lump-sum grants are not a relevant benchmark in that context.

Goulder considers the status of the other two forms on the basis of a review of qualitative theoretical models and numerical simulation models, concentrating on the strong form. The main features of his conclusions are:

The existing theoretical analyses of this issue cast doubt on the strong double dividend claim. At the same time, the theoretical case against the strong form is not airtight...

Although the evidence is mixed, numerical results tend to militate against the strong double dividend claim. There is a need for more systematic numerical investigations of revenue-neutral environmentally motivated tax policies.

The difficulty of establishing the strong double dividend does not contradict the common numerical finding that an environmental tax can promote higher levels of national income when revenues are earmarked for capital formation. National income is distinct from welfare, and the double dividend notions (at least as defined here) concern welfare costs.

This last point is important for the assessment of the results obtained from applied general equilibrium models, such as ORANI, WEDGE, and MEGABARE for examples in relation to double dividend propositions. Such model results are strictly irrelevant to the propositions as usually formulated. This is not, of course, to say that they are irrelevant to actual policy deliberation.

There is another important point that needs to be made here. Lack of support for the strong form of the double dividend proposition does not imply that substituting environmental for other taxes generally involves a welfare cost. To quote Goulder again:

It is important to recognize that the difficulty of substantiating the strong double dividend claim is not an indication that green tax swaps are ill-advised. Even if gross costs are positive, the environmental benefits could outweigh these cost. possibly by a substantial margin.

This applies a fortiori to results from applied general equilibrium models.

The usefulness of the welfare economics/optimal taxation approach to policy analysis could be argued to be rather limited, given its reliance on strong assumptions about human behaviour and technology, and the sensitivity of many of its results to parameterisation of those assumptions (Atkinson and Stiglitz 1980). There could also be objections to its ethical basis in utilitarianism as consumer sovereignty. Barker (in Carraro and Siniscalco 1993) takes a more eclectic approach, to carbon taxation in particular, concluding that it is 'worthy of extended and detailed study' as:

At first sight it has all the hallmarks of a good tax: it tackles an accepted economic problem, helping to bring the private costs of emitting CO2 into line with social costs of global warming; its revenues can be expected to be elastic with respect to income, because energy demand tends to rise with income and it is not easy to substitute away from fossil fuels to alternative energy sources in energy supply; if the tax is implemented in a similar manner to taxes on hydrocarbons, it should be simple and cheap to administer; and on the basis of some of the studies published, its side effects on equity and growth are likely to be small or negligible as long as fiscal neutrality is observed and special provisions made for low income groups. Indeed, the carbon tax could become one of the main pillars of fiscal systems throughout the world in the 21st century, if necessary replacing income tax, VAT or import duties (or all three forms of tax) and improving social welfare.

The fossil fuels, combustion of which is the dominant source for gross CO 2 emissions, are non-renewable resources. Economists' interest in the taxation of non-renewable resources is longstanding. However, the motivation in this literature is different from that in regard to the environmental taxes, including carbon taxes, discussed above. It is the rent on resource extraction which has been seen as the tax base in this literature, rather than the quantity extracted. The attraction of resource rents as tax base is that they can, in principle, bear a rate of 100 per cent without distorting the intertemporally efficient depletion programmes which would emerge if there were competitive markets. Carbon taxation would be implemented with fossil fuel extraction /use as tax base, and would affect the depletion programmes emerging in competitive markets. Ingham et al (in Carraro and Siniscalco 1993) use simple theoretical models to address some of the issues arising, and conclude that carbon taxation should be rising over time, and added to other taxes. Goulder (1994) considers that 'the extent to which the burden of environmentally motivated taxes is imposed on exhaustible resources (and on resource rents) remains a very interesting empirical issue'.

The carbon tax debate in Australia

This seminar is concerned with environmental taxation generally rather than the specific issue of carbon taxation. However, it is that context in which, especially in Australia, the general issues have in fact been raised. Some brief comments on the recent debate in Australia, in relation to the issues canvassed above, are therefore in order.

The first point to be made is that Australia has, by signing the 1992 Climate Convention, undertaken to reduce its (non-ozone depleting) greenhouse gas emissions to 1990 levels by 2000. It is generally agreed that in the absence of further policy action this commitment will not be met. This policy target was not adopted because it was shown to be where marginal costs and benefits were equal. It is, like most environmental policy targets, in the nature of an arbitrary standard. Given that there are several greenhouse gases. and that the commitment is understood to refer to net rather than gross emissions, it is a complex standard which does not immediately translate into specific targets for gross, or net, emissions of individual gases (see the greenhouse gas inventory produced under the terms of the climate convention, AGPS 1994).

It is, however, generally understood that it implies a need for some reduction in net carbon dioxide emissions, and this is widely taken to mean some reduction in gross emissions. If this is accepted, then the operative policy issue is the choice of policy instrument. Much of the adverse commentary on carbon taxation does not address this issue. It argues against carbon taxation, that is, without proposing alternative means by which Australia might meet its international obligations (see, for example, Zeitsch and Chisholm 1994, commenting on Common and Hamilton 1994). Much of the commentary expressed as opposition to carbon taxation appears actually to be opposition to Australia's commitments arising from signing the Climate Change Convention, rather than a contribution to the debate on instrument choice.

If those commitments are accepted, then reasoned opposition to carbon taxation must address the question 'if not carbon tax, what?' The economics literature briefly sketched in section 2 above is the starting point for an examination of this question. I am not aware that the least cost theorem about emissions taxation, or its translation into an input tax theorem, has been disproved. It is, of course, possible that there are special features of the carbon dioxide case that violate the assumptions of the theorems in such a way as to overturn the least cost presumption for taxation. Two such possibilities - sink enhancement and technical inefficiency in fossil fuel use - are briefly discussed below. As noted above, taxation is not dependable, but given the non-specific arbitrary standard involved this is hardly a decisive objection. Based on the abatement cost literature, there is a presumption that uniform taxation of carbon dioxide emissions, implemented via fossil fuel taxation, will move things in the right direction, with whatever abatement occurs being achieved at the lowest total cost of abatement.

Carbon taxation would, unlike the alternative instruments, be a regular and ongoing source of government revenue. However, if Australia's international commitment is taken as given, double dividend arguments of the type reviewed in section 3 above are not relevant to the policy debate. The operative question is then not whether the gross welfare costs associated with some switching of the tax base are zero or negative, but whether those costs are lower than those involved in meeting the commitment in any other way. If the abatement costs are minimised with taxation, and the alternatives do not give rise to any revenue, it would seem intuitive that taxation involves lower gross welfare costs than the alternatives. The gross welfare costs associated with carbon taxation would be expected to vary with the use to which the revenue is put, and the question then arises as to which other taxes should be cut, or whether the deficit should be reduced, or which expenditures should be increased.

Applied general equilibrium models have been used to consider some alternative revenue uses in terms of impacts on national income, employment, etc (see, for Australian examples, Common and Hamilton 1994, Zeitsch and Chisholm l994, McDougall and Dixon 1994, McKibbin and Pearce 1994). Extracting a clear message from these exercises is very difficult since not only do the models differ, but so do the scenarios modelled in terms of Australian carbon tax rate/gross emissions reduction target. It might be noted here that with reference to the operative policy issue for Australia, how best to meet its international obligations, it is not clear that it is appropriate to model unilateral action by Australia (as in the examples cited above). All of the industrial nations signed the Climate Change Convention.

The commitment arising from that convention relates to net emissions. These can be reduced by sink enhancement as well as by reducing gross emissions. The least cost theorem addresses the question of the choice of instrument for reducing gross emissions, and it is this question that is the focus of most of the debate. However, one contribution has explicitly addressed the question of alternative instruments to meet Australia's treaty obligations, and has claimed that carbon dioxide sink enhancement would be a lower cost option than carbon taxation. According to the recent environmental policy discussion paper issued by the Liberal/National Coalition (McLachlan 1995), that commitment can be met 'at a fraction of the cost of a carbon tax' by 'the cessation of broad acre scrub clearing and a national re-afforestation programme'. This is clearly a very important claim.

The actual cost of such a sink enhancement programme is not given in the discussion paper. A telephone inquiry to the office of the paper's author, the Shadow Minister for Environment and Heritage, yielded a figure of $200 million per annum, said to be based on data supplied by Greening Australia. Using planting rates given in the discussion paper, this implies an average cost of $800 per hectare. My understanding is that Greening Australia's replanting cost estimates are for a range from $400 to $2000 per hectare, excluding land acquisition costs. It is not clear that the 'fraction of the cost' claim is based on either a proper assessment of the opportunity costs, or of the impact on the measures, such as national income, against which carbon tax has been costed by modelling exercises. Such a sink enhancement programme would not raise any revenue, and would, presumably, increase government expenditure.

There are a number of ways in which, in principle, carbon taxation and sink enhancement programmes could be combined. Liability for the former on the basis of fossil fuel use could be reduced where sink enhancement could be demonstrated. Revenues from the tax could be used to finance public tree planting programmes, subsidise private tree planting, or finance compensation for broad acre clearing foregone. Such uses of the revenue would be at the expense of tax base switching or deficit reduction. The relative merits of such alternatives do not appear to have been systematically examined either in theoretical or empirical work. Similar remarks apply to the use of carbon tax revenue to subsidise the development and take up of alternative energy sources.

Economic analysis of instrument choice assumes that inputs to production and consumption are used in a technically efficient way. It is assumed, that is, that producers and consumers do not behave wastefully and do not use more of any input than is necessary given existing technologies. Applied general equilibrium models make the same assumption. Situations where, for example, less of some fossil fuel could be burned for the same output and the same level of other inputs, are excluded by assumption. In the absence of such assumptions, there would be available 'free lunches' by virtue of improved technical efficiency. It is also assumed that producers and consumers select the combination of inputs that minimises their costs for a given level of output /utility, given the relative prices of the inputs.

It is inconsistent to cite estimates of the costs associated with carbon taxation from standard economic models as reasons for not adopting it, and to claim that there are 'no regrets' options for meeting the international obligation. If it can be met at no cost, then it can be met at no cost by taxation. Taxation would, of course, involve transfer payments, but while these are costs from the perspective of the firm/individual, they are not abatement costs, real resource/output costs, for Australia as a whole. Presumably the preference of Australian industry for a moral suasion/no regrets approach to meeting the arbitrary standard reflects such considerations, together with the assessment that there is currently technical inefficiency and/or a lack of full cost minimisation in energy use in Australia.

Such inefficiency has been alleged by some, non-economist, commentators for a number of years. It has been pointed out that Australia has higher carbon dioxide emissions per capita and per unit national income than most industrial economies. Industry sources have denied that this reflects technical inefficiency, citing special Australian circumstances, such as low population density or industrial structure. Again, one has to note that it is inconsistent to deny technical inefficiency and to claim that moral suasion can reduce emissions below what they would otherwise have been at no cost.

Models, numbers and policy

Bringing together the problems of environmental protection and public finance is a relatively new departure in policy analysis. Clearcut theoretical results giving un-ambiguous and general policy prescriptions are not available. Empirical modelling is a necessary input to the analysis and debate of particular issues. It is not, however, sufficient, and if not handled carefully can confuse rather than clarify the essential issues. In the current carbon tax debate in Australia there are really two separate issues. Should Australia have signed the Climate Change Convention? Given that it did, how should it meet its arising commitments? Estimates of the size of the gross cost of carbon taxation are, unless the relevance of the least cost theorem is explicitly rejected and /or alternative instruments considered, relevant to the first question, not the second. Standard economic analysis has it that taxation is the least cost way of reducing emissions. The point is not that it is illegitimate to debate the first question, but that it creates confusion to present numbers from models relevant to the first question as if they bore upon the second. Zeitsch and Chisholm (1994), for example, conclude from modelling results that 'one can only be sceptical of the merits of carbon taxation'. This conclusion is not based on modelling the impact of alternative instruments to meet Australia's treaty commitments. The only instrument considered is carbon taxation. Their scepticism is about the merits of the commitment.

Such scepticism is based on the following line of argument. Acting to cut carbon dioxide emissions will lower Australian living standards. There will be no benefits to set against this cost, given unilateral action by Australia. It has already been noted here that, unless other industrial nations do not honour their commitments, Australia will not be acting unilaterally. However, it remains true that the size of any benefits in terms of avoided climate change damage to Australia is impossible to quantify. This does not mean that it is zero. The arbitrary standard to which Australia has committed itself arose, like most such, neither impulsively nor capriciously, but as the outcome of much scientific and political deliberation. Presumably one would not want to set it aside lightly. Presumably one would want to be convinced that Australians generally would gain perceptibly from abandoning the commitment, or would suffer perceptibly from honouring it.

Estimates of the material standard of living cost involved in honouring the commitment vary according to the model used. Looking across a range of models and simulations, one is looking at an upper limit which is a once-off reduction in national income of 1 per cent, assuming unilateral action. To make some allowance for population growth assume this means 1.5 per cent per capita. Given growth at 2.5 per cent per capita, this means that after a decade with the tax the average Australian would be 26.1 per cent better off as measured by per capita national income, instead of 28 per cent better off without the tax. Judgements as to whether or not this represents an un-acceptable sacrifice can legitimately differ. Given the uncertainties and complexities attending the enhanced greenhouse effect problem, there is no good reason to accord the judgement of economists privileged status here.

It is useful, for perspective, to compare an estimated cost of one per cent of national income with the accuracy with which national income can be measured. National income accounting conventions are such that it can be measured in three different ways, each, in principle, producing the same number. In practice the three measures differ, as reported in published national income data. Table 2.2 of ABS 1993, for example, gives as 'statistical discrepancy' the difference between GDP measured on the income basis and GDP measured on the expenditure basis for the 22 years from 1970-71 to 1991-92. The average discrepancy is 0.7 per cent of the income based measure. For seven of the years it is 1.0 per cent or greater; in three years it is 1.5 per cent or greater. Estimates of the national income cost to Australia of using carbon taxation to meet its treaty commitment in terms of gross carbon dioxide emissions are of the same order of magnitude as the statistical discrepancy in measuring national income.

The costs associated with staying with the commitment would, of course, not be the same for all Australians. Questions of equity in current burden sharing arise here, as they do generally with costs incurred for environmental protection. There are, again both particularly and generally, questions of equity as between generations. Equity in relation to the environment is the subject of the next seminar in this series, and no doubt environmental/carbon taxation will be re-visited then. The seminar after that, the final one, is to deal with competitiveness and the environment. Again, it seems likely that this will involve some re-visiting of aspects of environmental/carbon taxation.

References

ABS. 1993. Australian Economic Indicators. January/February 1993. Canberra. ABS.

AGPS. 1994. Climate change: Australia's national report under the United Nations framework convention on climate change. Canberra, AGPS.

Atkinson, A. B. and Stiglitz, J. E. 1980. Lectures on public economics. Singapore: McGraw Hill.

Baumol, W. J. and Oates, W. E. 1988. The theory of environmental policy (2nd edition). Cambridge: Cambridge University Press.

Bohm, R. and Russell, C. S. 1985. Comparative analysis of alternative policy instruments. In Kneese, A. V. and Sweeney, J. L. (eds) Handbook of natural resource and energy economics Volume L Amsterdam, Elesevier.

Carraro, C. and Siniscalco, D. 1993. The European carbon tax: an economic assessment. Kluwer, Dordrecht.

Common, M.S. 1977. A note on the use of taxes to control pollution, Scandinavian Journal of Economics, 79:346-349.

Common, M. S. 1988. Environmental and resource economics: an introduction. Harwell: Longman.

Common, M. S. and Hamilton, C. 1994. Taxation of greenhouse gas emissions and employment. Appendix C in House of Representatives Standing Committee on Environment, Recreation and the Arts, Working with the environment: opportunities for job growth. AGPS, Canberra.

Cropper, M. L. and Oates, W. E. 1992. Environmental economics: a survey, Journal of Economic Literature, XXX:675-740.

Goulder, L. H. 1994. Environmental taxation and the 'Double Dividend': a reader's guide. Paper to the 50th Congress of the International Institute for Public Finance, 'Public Finance Environment and Natural Resources'. Cambridge, Mass.

James D. 1993. Australia's experience in using economic instruments for meeting environmental objectives. Canberra: Department of Environment Sport and Territories.

McDougall, R. and Dixon, P. 1994. Analysing the economy -wide effects of an energy tax: results for Australia from the ORANIE model. DEST Workshop: The role of economic analysis in assessing greenhouse impacts and responses, Canberra October 1994.

McKibbin, W. and Pearce D. 1994.G-Cubed and environmental policy analysis. DEST Workshop: The role of economic analysis in assessing greenhouse impacts and responses, Canberra October 1994.

Department of the Environment, Sport and Territories Logo