- Consistent with earlier findings, @AEIecon’s Aparna Mathur finds that a carbon tax is regressive.
- The carbon tax burden would comprise 3.5% of the income of the poorest decile of households.
- In the consumption approach, the carbon tax is substantially less regressive.
This paper analyzes the distributional implications of an illustrative $15 carbon tax imposed in 2010 on carbon in fossil fuels. We analyze its incidence across income classes and regions, both in isolation and when combined with measures that apply the carbon tax revenue to lowering other distortionary taxes in the economy. The analysis first uses an input-output table approach to estimate the effect of the carbon tax on consumer prices, assuming that the tax is passed through fully to retail prices. Then, using Consumer Expenditure Survey data on consumption patterns, we estimate the burdens across households, assuming no behavioral response to the new prices.
Consistent with earlier findings, we find that a carbon tax is regressive. Taking into account both direct and indirect energy costs, the carbon tax burden would comprise 3.5 percent of the income of the poorest decile of households and only 0.6 percent of the income of the highest decile. In the consumption approach, the carbon tax is substantially less regressive, with the ratio of average taxes paid by the bottom and top deciles equal to about 1.7.
In the tax swap simulations, we subtract the burden of other taxes that the carbon tax revenue could displace, such as the corporate and personal income taxes, and compute the net effect on households. We analyze revenue-neutral tax shifts under three assumptions about how those other taxes lower households' capital and labor income: all borne by labor, all borne by capital, and a 50/50 split. Although all of the tax swaps lower the overall burden of the carbon tax (as a share of household income) on the poorest two deciles, tax swaps also exacerbate the regressivity of the carbon tax on the high end. This means that the benefit to the highest income households of the reduction in other taxes is greater than their share of the burden of the carbon tax.
Results suggest that if policymakers direct about 11 percent of the tax revenue towards the poorest two deciles, for example through greater spending on social safety net programs than would otherwise occur, then on average those households would be no worse off after the carbon tax than they were before.
The degree of variation in the carbon tax incidence across regions (with no offsetting tax decreases) is modest; the maximum difference in the average rate across regions is 0.45 percentage points of income. Of the tax swaps, the labor tax swap results in the least variation in net burdens across regions, with a maximum difference across regions of 0.5 percentage points of income. In contrast, the capital tax swap produces a maximum difference across regions of 1.66 percentage points. This suggests that capital incomes are more unevenly distributed across regions than labor incomes.