2.3  Market based instruments

Two policy instruments are suggested in Figure 1. Price or quantity can be used to achieve the same result. With perfect information and certainty about the relevant MB and MD functions, a regulatory agency can auction off rights to Q* and obtains P* or set a price P* with the result Q* – the outcome is the same. One important difference is the revenue effect. With a tax, revenues go to government whereas with quotas, either government could auction them off and get revenue equal to the capitalised value of the tax or government could grandfather them and not enjoy any revenues.

2.3.1  Polluter pays

The polluter pays principle, endorsed by OECD in 1972, is a direct descendent of the Pigouvian tax scheme used to establish efficiency within the context of competitive equilibrium. Figure 1 shows how polluter pays would work. Industry’s profit maximising equilibrium is B. If a fictional environmental agency were to price use of the river according to the MD schedule then efficiency would result. The agency would only have to set a charge of P^* to achieve the efficient outcome Q*.

Setting political considerations aside, knowledge of marginal damages is one of the more significant barriers to achieving efficiency. Recognising this difficulty resulted in an alternative approach that seeks to minimise the cost of meeting a standard. Returning to Figure 1, a standard set by the agency is shown as Q_R units of pollution. If the agency set a charge of P_R per unit of pollution then the environmental outcome Q_R would result. Two points should be emphasised. First, it would be highly unlikely that the agency has the information needed to set the efficient target (Q*). Second, whether or not the target is met depends on the charge – if too high (P > P_R) then target will be exceeded, if too low (P < P_R) the outcome will not meet the target. The flexibility required to perfectly achieve the target may be difficult to achieve because proposed changes would be decided in the political market.

Economists favour the polluter pays approach, relative to regulation, because it tends to produce the desired result at lowest cost to society and provides an on-going incentive to implement less polluting technology. Rather than trying to achieve economic efficiency the aim is reduced to achieving the desired outcome at least cost to the economy. Figure 3 shows two firms 1 and 2 with an unregulated equilibrium of . Let us assume that the agency wants to reduce emissions to 2e^*. Requiring each firm to emit e^* will achieve the desired target. This is not a cost-minimising level because the marginal costs of reducing emissions are not equal MC₁ (e*) >MC₂ (e*). A charge of t* will provide the necessary opportunity cost to each firm. If firm i pollutes then it pays t*, the alternative is to treat wastes at a marginal cost of MC_i and avoid paying the charge. A charge will result in . The information needed to achieve this result is considerably less that the Pigouvian ideal. The agency need only specify the target 2e* and set the charge. There is no need to know each firm’s marginal abatement costs. The agency must of course monitor and measure discharges but there is no a priori reason why this must result in a net cost to the agency over and above the CAC alternative.

Figure 3 – Achieving a target at least cost

A charge levied on polluters for use of the environment provides an incentive to change behaviour in ways that can promote sustainable development. According to the OECD evidence of their effectiveness is given by the higher responsiveness of energy demand to changes in energy prices in the longer than in the shorter run (OECD, 2001). For example, fuel taxes differentiated according to environmental criteria on gasoline (leaded and unleaded) and diesel fuels (ordinary and ultra-low sulphur) have led to a reduction in the use of the most-polluting fuels.

The difficulty of measuring and monitoring externalities has limited the wider application of environmental taxes. Ideally, taxes based on emissions should reflect differences in environmental impacts. Most taxes are levied on the products associated with pollution rather than the environmental impact of pollution. Two diesel vehicles could consume the same amount of fuel – and therefore pay the same charge – but one generates less pollution. In some cases fuel characteristics (eg, sulphur, lead content) may be a reasonably good proxy for the externality.

The idea proposed in Figure 1 shows the tax being directly associated with the level of pollution from the industry. In a real-world situation, the production of externalities arise from particular technologies and behaviour. The environmental tax base can be defined in a number of ways. For example, the tax might be levied on an input or output. Regardless of whether it is defined on an input (eg, diesel) or an output (eg, sulphur) it is highly likely that the tax is imperfectly correlated with the level of emissions that we want to control. In some cases we can measure and monitor emissions and the best (efficient) solution is to tax emissions provided, of course, the causal relationship with damages is known.

Charges create an element of certainty in prices of the externality but the behavioural response depends on price elasticity of demand and income. For example, in New Zealand the short-run price elasticity of petrol is –0.11 and short-run income elasticity is 0.57 (Hughes, 1980). Although dated, these results tell us that demand is relatively price inelastic and petrol is a normal good. In general, a “problem” with multiple dimensions requires use of a combination of taxes. For example, externalities associated with traffic congestion might be more efficiently dealt with using a combination of congestion pricing and environmental taxes.

Figure 4 shows that transportation fuels and vehicle related charges provide the largest tax base within the OECD. Thus it would appear that polluter pays has a limited range of use, especially if one considers the possibility that governments may be targeting the tax base simply for revenue raising and not necessarily for improving environmental quality. Environmental taxes per capita are highest in Europe (eg, Denmark, Norway US$1,200-1,400) and relatively low in the USA and NZ (US$250).

Choice of the correct tax base is sometimes difficult. For example, consider choosing to levy a tax on diesel. The impact of the tax on environmental quality will depend on the interaction of many variables that are exogenous to the use of diesel, such as climate, neighbourhood and time of day and, the age of the vehicle, its maintenance and driving behaviour.

Conceptually, each variable could act to determine the environmental impact of diesel use. A tax on diesel will imperfectly discriminate among different vehicles and drivers. It is not necessarily true that there will be a trade-off between efficiency and equity. Suppose that a diesel tax leads to efficiency gains by bringing marginal costs and marginal benefits closer together. There is no reason why the tax should increase inequality – this is an empirical issue.

Most stylised versions of taxation show externalities being corrected independently of the effects on the distribution of welfare in the economy. The reason is that this version assumes government makes lump sum transfers to redistribute income and welfare. As Sandmo (2000) points out, the use of taxes is to correct for an externality – one target, one instrument. The difficulty associated with the distribution of welfare is that government does not have the information to individualise lump-sum payments.

Figure 4 – Revenues raised from environmentally related tax-bases

Source: OECD 2002

There are, at least, two aspects to the distributional issues associated with environmental policy. To illustrate, consider imposing a tax on diesel and assume that the waste associated with diesel combustion impacts both air and water quality. First, it is quite conceivable that the distributional impact of an increase in air quality will differ from the impact of an increase in water quality. Second, a tax on diesel will have distributional effects that depend on the income elasticity of diesel (impact on households) and the ability of producers (impact on business) to pass the tax on.

A double dividend has been attributed to environmental taxes. The idea behind a double dividend is that if we implement more environmental taxes then other taxes can be cut and this can lead to other non-environmental gains. The underlying assumption is that government revenue stays the same. Economic theory does not give unqualified support to the view that both dividends will be positive. Although this is likely to be the case for the environmental dividend but there is no general case for the tax revenue dividend to be positive. To say anything with confidence requires further assumptions about the initial state of the tax system and the demand interrelationships among goods and services.