With the recent blue wave of Democratic success in retaking the House of Representatives, we are likely to see a return of many policy proposals typically associated with the left side of the political aisle. One of those proposals is surely to include a carbon tax of some kind. Proponents argue that the carbon tax — especially one that is made revenue-neutral by a reduction of existing taxes commensurate to the revenue raised by the new carbon tax — is a win-win. Taxing a “bad,” they argue, reduces harm to the environment and collects revenue for other necessary government spending. The taxing of such “bads” is called a Pigouvian tax, and economic analyses of their effects typically forget one of the foundational principles of the science.

A great economist once wrote that economics looks not only at “the immediate but at the longer effects of any act or policy; it consists of tracing the consequences of that policy not merely for one group but for all groups.” The “longer effects” of a particular policy are what economists refer to as the “unseen,” both because they occur in the future and because they are obscured by the miasma of interactions with other aspects of the economy. The standard economic analysis of a carbon tax fails to take account of all that is unseen, and in so doing vastly overstates its efficacy.

The standard analysis of the distortionary effect of a tax is that the narrower the tax base the greater the distortion. This makes basic logical sense — if the government wants to raise $1 billion in revenue, it could raise the marginal tax rate on personal income in each bracket. Or, as one among many alternatives, it could raise the marginal tax rate on only those taxpayers who earn over $250,000 per year. Notwithstanding any ethical arguments, the latter option is more distortionary in a purely economic sense; because with the smaller base, in order to raise $1 billion, the increase in marginal rates would necessarily be greater and any negative effects on work hours, rates of saving, etc., will be more pronounced. Similarly, Pigouvian taxes are targeted at a smaller tax base than other, more general taxes on labor or capital, and are correspondingly more distortionary than those generic alternatives.

But there are causes of distortions beyond how broad or narrow the base of taxation. The distortionary effect of preexisting taxes — such as the income tax and the payroll tax — can be amplified by the imposition of a carbon tax. Intuitively, a carbon tax makes business inputs and some capital goods more expensive, and so the distortions of a wage or income tax are amplified by a new carbon tax. More specifically, the economic cost of raising a dollar of revenue through a carbon tax is higher than raising a dollar through ordinary income taxes, because carbon taxes are most intensely targeted at the narrow tax base of individual commodities or emissions from specific industries. There is a corresponding distortion of intermediate inputs that raises the price of commodities. This price increase reduces the real return to both workers and capital — a given nominal wage or distribution of earnings and profits has less purchasing power than before.

But when there are preexisting taxes on wages and profits, a carbon tax acts as a tax on both labor and capital, imposed at a higher level of the structure of production. Because the carbon tax is more narrowly targeted than these general taxes, the dollar for dollar economic distortion it causes is correspondingly larger. The reason for this is intuitive but also relatively abstract. Notwithstanding purely environmental benefits, carbon taxes are more distortionary than general income or payroll taxes because they serve as a specific tax on carbon-intensive intermediate inputs. A tax on intermediate inputs is more distortionary than a tax on output because the former distorts the “input mix” of production processes. A tax on final goods or services primarily affects prices for consumers in final product markets, but a tax that falls largely on intermediate inputs is narrower and distorts production processes that in turn drive the composition of the vastly more numerous and valuable final product markets, amplifying the distortions already presented by the existence of general income and payroll taxes.

This result is best illustrated with a hypothetical example. Assume the economy is emitting 10 billion tons of carbon dioxide and there is no carbon tax. Next, assume it is possible to accurately calculate the social cost of carbon, and that cost is $10 per ton. Therefore, if no carbon tax is imposed, carbon emissions will cause $100 billion annually in increased social cost. With no other taxes, the optimal carbon tax is $10 per ton of emissions. For the sake of making the math easier, assume this optimal tax incentivizes the economy to cut its carbon footprint in half, to $50 billion. However, the carbon tax imposes real costs on the economy as well, as consumers and producers alike strive to use less carbon-intensive techniques. We will say this harm to the economy is $10 billion. Therefore, there is a net benefit of the carbon tax of $40 billion.

Now we will run the example again, this time with an income tax. Assume an income tax causes a deadweight loss distortion of $100 billion. Basic economics and tax theory tell us that as you increase the tax rate, deadweight loss increases more than proportionally. Here, we are not implementing an income tax from scratch but on top of and in conjunction with the carbon tax of $10 per ton, so there is an implicit tax rate that is at least 10% higher than before. Therefore, the actual deadweight loss generated by the income tax itself, generated by its interaction with the carbon tax, will increase by more than 10% — say, to $120 billion. This $120 billion deadweight loss is in addition to the $10 billion in distortions generated by the carbon tax — the indirect effect of the carbon tax, generated by the tax interaction effect, is $20 billion in deadweight loss.

With the numbers given, and under traditional, static analysis, instituting a $10 per ton carbon tax still makes economic sense if the receipts from the tax are returned to citizens via lump-sum payments. Imposing the tax reduces social costs by $50 billion, while the economy endures damages of $10 billion (deadweight loss directly from the carbon tax) plus the $20 billion from the tax interaction effect, for a total distortion of $30 billion. However, the $50 billion gain for the economy ($50 billion reduction in social cost) minus this $30 billion distortion still supposedly leaves the economy with an apparent net gain of $20 billion from the imposition of the carbon tax.