How Regressive is a Price on Carbon?

01/01/2010
Featured in print Digest

A price on carbon could yield substantial government revenues, and careful recycling of these revenues could offset the regressive nature of a national GHG [greenhouse-gas] emissions policy.

Under either a cap-and-trade program that limits carbon emissions or a carbon tax that imposes an outright tax on these emissions, the poor may be among the hardest hit. Because they spend a greater share of their income on energy than higher-income families, households in the lowest fifth of the income distribution could shoulder a relative burden that is 1.4 to 4 times higher than that of households in the top fifth of the income distribution, according to a study by Corbett Grainger and Charles Kolstad. In Who Pays a Price on Carbon? (NBER Working Paper No. 15239), they show that the burden on the poorest households doubles when a price on carbon is targeted narrowly on energy consumption (and not other energy uses) rather than broadly across all industries. Our results suggest that the burden as a percent of annual income is much higher among lower income groups than higher income groups, the authors write.

Previous research already has suggested that a carbon tax would probably be regressive. This study furthers the analysis by making three key points. First, by linking the amount of carbon emissions from each industry to consumer expenditures by income group, the authors show that consumption differences explain the regressivity of a carbon tax. Assuming a levy of $15 per ton of carbon dioxide, which is in the range of current proposals in Congress, the authors calculate that the one-fifth of households at the bottom of the income distribution would spend an extra $325 a year. That's less than a third of what the one-fifth of households at the top of the income distribution would pay annually. However, households in the low-income group earn only one-tenth as much as those in the high-income group on average, so their burden relative to income would be almost four times higher.

Some economists argue that annual income, which changes over time, may be less accurate as a measure of household well-being than income measured over a lifetime. On the basis of lifetime income, the burden on the low-income households would be 1.4 times higher than it would be on their higher-income counterparts, this study finds.

The second key point is that calculations by household understate how regressive a price on carbon would really be. That's because households in the highest income quintile are much larger, averaging 3.1 persons, than those in the lowest quintile, which average only 1.8 persons. Accounting for those differences (and for economies of scale in household consumption), the authors calculate that the real impact of a carbon tax on a person in the lowest income quintile would be nearly five times more burdensome than for someone in the top income quintile. Using lifetime income in this calculation, the burden would be 2.2 times greater.

The third key point is that the regressivity of a tax on carbon depends on how broadly it's applied. If it's levied on all greenhouse-gas emissions, then the burden on the lowest-income fifth of households would be 3.25 times as high as the burden for the highest-income fifth (1.4 times as high based on lifetime income). Per capita, the burden would be about five times higher (over twice as high based on lifetime income). If the tax only applied to consumption of energy goods, then the burden on low-income households would climb to nearly four times that of their higher-income counterparts annually (1.6 times, using lifetime income). Per capita, it would soar to six times annually (2.6 times, based on lifetime income). The authors conclude "that the regressivity of the policy is driven largely by direct energy consumption.

They offer several caveats about their study, which assumes that all costs are passed on to the consumer and don't affect workers' wages or investors' returns. Consumption is held fixed, and no attempt is made to simulate how consumers' buying habits or companies' production practices would change if the price of carbon went up. Nor does this study consider whether the benefits of a carbon tax would affect households at different points in the income distribution disproportionately through an effect on climate change.

The authors calculate that a $15 tax per ton of carbon dioxide would raise as much as $79 billion a year. Congress could use some of those revenues to mitigate the regressive effects of the tax. For example, an income tax break of $119, $112, $105, and $76 to individuals in the first four income quintiles, respectively, would balance the burden to about 1 percent of net annual income for each group, and still leave nearly $50 billion in government revenues, the authors calculate. Or, the revenue could alleviate the burden of other regressive taxes, such as the payroll tax. A price on carbon could yield substantial government revenues, and careful recycling of these revenues could offset the regressive nature of a national GHG [greenhouse-gas] emissions policy.

-- Laurent Belsie