The Efficiency of a Carbon Tax: Broadly Accepted and Broadly Wrong
Friday, January 31, 2014
The standard assumption about the superior efficiency of a carbon tax relative to bans and energy consumption standards is deeply problematic for both scientific and political reasons.
Making an argument broadly accepted among economists, Sita Slavov, my colleague at the American Enterprise Institute, recently wrote that a carbon tax “set to reflect the spillover costs of carbon emissions” would be a policy more efficient than such interventions as light-bulb bans and energy consumption standards for appliances, sometimes called “command-and-control” policies. The tax would allow consumers to reduce the emission of greenhouse gases (GHG) attendant upon their energy consumption patterns in ways that minimize the costs of doing so. Bans and performance standards, on the other hand, allow for far less flexibility, and thus would achieve a given GHG reduction at a cost higher than necessary.
This standard argument is deeply problematic for reasons both scientific and political. Science first: Let us defer the issues raised by the poor predictive record of the Intergovernmental Panel on Climate Change climate models, the fierce debate in the climate and atmospheric journals about the direction and magnitude of feedback effects, and the problems inherent in the temperature record. Instead, assume that the models’ predictions about temperatures and weather anomalies in 2050 and 2100 — even the most apocalyptic — are correct. The following table shows the amount of warming averted by 2100 were GHG emissions to be reduced by, respectively, the United States and the industrialized world (the OECD90: North America, Western Europe, Australia, New Zealand, and Japan) under the highest temperature sensitivity assumption used by IPCC (a doubling of GHG concentrations would cause an increase of 4.5 degrees Celsius). The model used here is the MAGICC/SCENGEN climate simulator developed at the National Center for Atmospheric Research, and used by both IPCC and the Environmental Protection Agency.
In short: A carbon tax (or, for that matter, any other climate policy), whether imposed by the United States unilaterally or by the industrialized world, would have virtually no effect on temperatures over the course of this century. (For a discussion of the actual political/wealth redistribution impetus underlying “carbon” policies, see this.) Note that the standard deviation of the annual global average surface temperature (once the trend is removed) is about a tenth of a degree; accordingly, in most cases the reduction in temperatures would be difficult to measure. It simply cannot be the case that a policy — whether a tax or an alternative command-and-control approach — rationalized on the basis of a purported adverse environmental effect (“externality”) but engendering no practical effect on that externality can be “efficient.” A carbon tax or other such intervention — regardless of one’s view of the underlying climatology — would be all cost and effectively no benefit, although the wealth redistribution effects can be significant, as just noted.
What incentives do policymakers have to choose an ‘efficient’ carbon tax?
Let us turn now to the political (or public choice) problem inherent in the standard assumption about the superior efficiency of a carbon tax relative to bans and consumption standards. Slavov assumes, as previously noted, a carbon tax “set to reflect the spillover costs of carbon emissions,” an approach reflecting a conceptual experiment utterly irrelevant in terms of the institutional realities shaping environmental policies generally and “carbon” policies in particular. In other words, Slavov — and much of the economics profession — assumes the adoption of an efficient carbon tax and then argues that it is superior to an inefficient command-and-control approach. But what incentives do policymakers have to choose that “efficient” carbon tax? Slavov here makes the same error inherent in the broad conventional economic view of environmental policy: Market incentives lead away from efficiency in resource allocation — too much pollution will be produced — but government somehow has incentives to correct for that distortion by choosing the efficient carbon tax.
Support for a carbon tax must be based upon that assumption; without it, the supposed superior efficiency of the carbon tax disappears. And indeed the assumption is not tenable: A carbon tax would be hidden in the costs of goods and services. This would result in a standard fiscal illusion problem: For any given voter (or for the median voter), the economic cost of the tax would be difficult to measure. Political competition thus would be less effective in terms of constraining the magnitude of the tax, and so imposition of the tax would be accompanied by less political accountability for its impacts. This effect would be exacerbated by the problem of political time horizons on the part of politicians shorter (or a political discount rate higher) than those of the (median) voters. That adverse outcome would be strengthened by (1) the weakening of the political parties (which are long-lived political institutions) caused by campaign finance restrictions (that is, incumbent protections), (2) gerrymandered districts in the House of Representatives, and (3) political competition among interest groups for the revenues yielded by the carbon tax.
A carbon tax, whether imposed by the United States unilaterally or by the industrialized world, would have virtually no effect on temperatures over the course of this century.
The shorter political time horizon (the next election) is likely to yield too little attention to the adverse long-run effects of policy choices — in this context, a carbon tax that is too high. The gerrymandering process (“wasting” the votes of the political opposition) is likely to result in less two-party competition in districts than otherwise would be the case, and thus a reduction in political accountability (although the effect in a House of Representatives or a Congress acting collectively is less clear). Both of those conditions are likely to yield a carbon tax higher than that “set to reflect the spillover costs of carbon emissions,” and the same is true for interest-group competition for the revenues.
A tax higher than that correcting for (“internalizing”) the assumed carbon externality is inefficient, and there is no obvious ex ante reason to believe that it would be better (less inefficient) than the inflexible command-and-control policies criticized by Slavov and others. Indeed, the opposite may be true: Command-and-control policies emerge from a complex process of political competition, and because they are visible — people notice when they cannot purchase their preferred light bulbs or when appliances cost more — there are good reasons to hypothesize that they are likely to reflect, however crudely, the marginal costs and benefits of the policies. Notice that the Obama administration has been unable to induce even a Democratic Congress to enact “carbon” policies — they are being promulgated in a regulatory process not obviously legal — and that the famous light-bulb ban has not been enforced because of legislative riders written by Congressman Michael Burgess and passed as part of the annual budget process.
Slavov concludes by arguing that a carbon tax might “address climate change while … raising revenue that can be used to lower other taxes.” Since it is implausible that the “incidence” (the distribution of the economic burden) of a carbon tax would be the same as that of those “other taxes,” Slavov here slips away from the efficiency rationale for a carbon tax because we cannot judge the relative efficiencies of the carbon and other taxes unless we know what the latter are. An assumption that a carbon tax would be the more efficient is deeply problematic, both because of the fiscal illusion problem noted above and because there is no particular reason to believe that the demand for fuels is superior to, say, income taxes as a proxy for the benefits of general-fund public services. And even in a standard tax “neutrality” analytic framework, precisely what is the basis for the implicit assumption that a carbon tax — which, again, would have virtually no effect on the supposed externality — would yield fewer allocational distortions than the taxes that it would replace?
Notice that the Obama administration has been unable to induce even a Democratic Congress to enact ‘carbon’ policies.
Instead, substitution of one tax for another is a classic redistribution process, one that is likely to require an expansion of government so as to compensate the losers, even if only partially. Accordingly, even in Slavov’s conceptual framework, not all of the revenue would be used to “lower other taxes.” Instead, net government revenues — or, more broadly, government allocation of resources — would have to increase, and unless one believes that government resource use is more productive than that of the private sector, “efficiency” writ large is not prominent among the likely outcomes.
John Maynard Keynes once noted that “Practical men, who believe themselves to be quite exempt from any intellectual influence, are usually the slaves of some defunct economist.” To whatever degree that is true, economists build upon the intellectual foundations that they find, and it sometimes is the case that too little reworking is undertaken. That is decidedly the case with the conventional analysis of externalities, and it is my hope that the ongoing debate at AEI over carbon taxes might engender a reconsideration of assumptions both standard and too facile.
Benjamin Zycher is a resident scholar at the American Enterprise Institute.
FURTHER READING: Zycher also writes “The Climate Change Climate Keeps Changing, but the Carbon Tax Is Eternal,” “The Unbearable Lightness of the Climate Change Industrial Complex," and “The 'Science' of Global Warming.” Michael M. Rosen reveals “Greens’ Irrational Fear Flies Again.” Steven F. Hayward investigates “‘Environmental Justice,’ EPA Style.”
Image by Dianna Ingram / Bergman Group