There is little doubt, among economists at least, that market mechanisms are required to achieve significant reductions of greenhouse gas emissions in a cost-effective way. Exhortations will not suffice, and regulatory measures typically do not recognize the different circumstances of different households and firms. By providing financial incentives to reduce greenhouse gases, market mechanisms let individual decision-makers determine the optimum manner and extent of their response.

There is less agreement, however, about the type of market mechanism to apply. Should greenhouse gas emission quotas be instituted, like the cap-and-trade system successfully used to reduce sulfur dioxide and other air contaminant emissions from industrial processes? Or should a carbon tax be imposed on goods and services that generate greenhouse gases?

The economic logic of a carbon tax is that it internalizes the external cost of greenhouse gas emissions – the incremental damage costs that the release of greenhouse gases imposes on all others now and in the future. In so doing, the tax will raise the direct cost people and businesses incur when they consume goods and services that generate greenhouse gases, and provide industry with an incentive to develop substitute products and processes that generate less greenhouse gases. That in turn will promote a more appropriate benefit-cost test in the decisions people and businesses make. They will have to determine whether the benefit they derive from any greenhouse gas–generating behaviour exceeds the costs they face – costs that include the internalized cost of the greenhouse gases emitted. The higher the tax, the more likely they will choose alternatives that generate less greenhouse gases.

The political argument in support of carbon taxes is that, like sin taxes, they are a charge on behaviour we want to discourage. Most importantly, proponents of carbon taxes argue, the revenues a carbon tax raises can be used to reduce taxes on income and sales – charges on effort and economic activity that we do not want to discourage. Indeed, a major selling feature of the carbon tax recently instituted in British Columbia is that it will be revenue neutral – in effect, substituting for more odious, inefficient forms of taxation.

The economic logic and supporting political argument for a carbon tax are compelling. However, they are not entirely valid. The economic logic underlying a carbon tax presumes there is a well-defined external cost to internalize. The problem is that the incremental damage costs associated with incremental greenhouse gas emissions in any given jurisdiction are very unclear.

Whatever consensus there may be on the meteorological implications of increasing releases and concentrations of greenhouse gases in the atmosphere, there is no consensus on the damage caused by an incremental tonne of emissions. The incremental damage of a tonne of emissions generated in any given jurisdiction at any point in time will depend first and foremost on the amount of emissions being generated elsewhere – in other words, what any tonne of emissions in any given jurisdiction at any point in time is incremental to. In particular it will depend on where we are in relation to key thresholds.

Also, the estimated economic consequences of the emissions will depend on what measures are assumed to be taken to adapt to climate change and what technological innovations in adaptive and offset measures take place. Further, the present value of the damage or adaptation costs, as the recent Nordhaus-Stern debate shows, depends critically on the discount rate we apply to future effects –
especially very distant future ones.

If there is any emerging consensus among economists, it is that the fundamental concern is not so much the expected damage from each incremental tonne of emissions, which is highly uncertain and may be relatively small in present value terms at any significant discount rate. The major cost of greenhouse gas emissions above what scientists are telling us is a tolerable maximum threshold has to do with the risk of future catastrophic, irreversible consequences that ever-increasing greenhouse gas emissions may cause. Such a catastrophe may be a low-probability “tail event,” but nonetheless it is a risk that people in wealthier societies do not want to take and poorer societies do not want imposed on them by the cumulative historical and current actions of others.

This suggests that the market failure our greenhouse gas market mechanism must address is not the need to internalize a well-defined external cost. Rather it is the need to live within a cap on the total amount of greenhouse gases we emit – a cap that in any given jurisdiction is consistent with the global cap and global allocation of emission rights we expect will ultimately be agreed on.

Carbon taxes could be designed to achieve a target cap on emissions. However, the level of tax required is highly uncertain, dependent not only on the price elasticity of demand on all emission-generating goods and services but also on all other factors governing the price of those goods and services, like the cost of crude oil. The tax needed at a $50 per barrel price of crude is markedly different from that required at $150 per barrel. The likelihood of significantly undershooting or overshooting the target is considerable.

The more direct and certain way to achieve an overall emission target is to impose a cap-and-trade system. If a carbon tax is chosen, it is only as a practical “second best” alternative where a cap-and-trade system is difficult or expensive to implement and enforce.1

And what about the political argument in favour of a carbon tax – that it constitutes a more efficient and desirable form of taxation than income or sales taxes? That argument presumes the greenhouse gas–generating behaviour giving rise to the payment of carbon taxes is costless. However, if we take seriously the rationale for imposing the tax in the first place – as a way, even a “second best” way, to achieve a specified emission target – and if we recognize that even under the most optimistic elasticity and price assumptions the taxes currently introduced in British Columbia and elsewhere will not achieve that target, the funds that are raised are not free. They cannot be dedicated to reducing other taxes unless the emission target is abandoned. In other words, the carbon tax revenues are needed for additional measures, such as international offset purchases and spending on green infrastructure, that will be required to supplement the effect of the tax alone.

To say that a carbon tax can be revenue neutral, with the funds used entirely to reduce income and other taxation, would be like saying the same for road tolls. Perhaps road toll revenues could be used entirely to reduce income and other taxation, but only if the tolls were sufficiently high to eliminate the need for investments in additional road and bridge capacity. Otherwise, at least some of the road toll revenues are needed to pay for the investments needed to accommodate the increase in traffic that occurs despite the tolls. One could say that the road tolls offset the increase in income and property taxes that would otherwise be needed to make the required road and bridge investments, but that is quite disingenuous. It means other taxes may not go up as a result of the introduction of road tolls, but it certainly does not mean other taxes can go down.

So, by all means we should support carbon taxes if the alternative is no market mechanism at all. But it is important to recognize that cap-and-trade is a preferable market mechanism and carbon taxes should be limited to those sectors where cap-and-trade is impractical to implement and enforce. And let us not pretend that carbon taxes are revenue neutral, at least if our emission targets are to be met.

Note

1 One advantage of a carbon tax is that it avoids the risk of having costs of reducing greenhouse gases rise above a reasonable level. The tax provides certainty to firms about the maximum cost that needs to be incurred. However, cap-and-trade systems can avoid the risk of excessive cost if designed with an elastic supply of quotas at what are deemed to be unacceptably high quota costs.