Blog Post

Measuring the Early Effects of a Carbon Tax: Comparing Three Policies across Income Groups

Aug 7, 2014 | Hal Gordon

The Initial Incidence of a Carbon Tax across Income GroupsEconomists argue that the most efficient way to reduce greenhouse gas emissions is to raise the price of energy by introducing a price on carbon emissions, either through a tax or a cap-and-trade regime. A tax (or a cap-and-trade system that auctions off permits) has a secondary effect beyond providing incentives for reducing emissions: it raises a lot of new government revenue. Just what is done with this revenue weighs heavily on the distributional effects of a carbon price; in fact, it has greater relevance to the distributional outcome than the actual price on carbon emissions. In a new RFF discussion paper, with RFF’s Rob Williams and Richard Morgenstern, and Jared Carbone of the University of Calgary, we analyzed three different options for allocating the revenue from a $30 per-ton tax on carbon (which translates to $192 billion in the first year), and the effects on the distribution of the cost burden across income groups in the United States. These effects change over time. We focus here on the initial effects at the outset of the policy because these near-term consequences are the most salient in the contemporary political dialogue.

Most carbon tax proposals plan to return the money collected by the government to the people. For instance, California’s cap-and-trade scheme returns much of the revenue to utility customers with a biannual “climate credit.” The three policies we examine would recycle all carbon tax revenue—that is, the level of governments spending would not change. In one policy, an equal share of the revenue is returned to each resident in the form of a lump-sum rebate. In another, the revenue is used to pay for a cut in the marginal labor tax. And in a third, it pays for a cut in marginal capital taxes. The tax cuts are implemented as an equal percentage reduction in the tax rate at all income levels, although different approaches would have different effects. It has been well established that total economic efficiency is expected to be greatest when the revenue from a carbon tax is used to reduce the capital tax and next best when it is used to reduce the labor tax because, in each case, the policy substitutes for a preexisting distortionary policy. A lump-sum rebate does not share this characteristic. However, economic efficiency comes with particular distributional outcomes that should also be evaluated.

Although the capital tax recycling policy is most efficient, it is also the most regressive. The labor tax policy is somewhat less efficient, but spreads the welfare effects relatively evenly across income quintiles. The lump-sum rebate policy is the least efficient and has the most diverse distributional outcomes, but results in a strongly progressive outcome. In fact, roughly speaking, the lowest three income quintiles (60 percent of households) experience welfare gains under this policy, according to our analysis.

To measure economic welfare impacts by quintile, we used various government data sources to calculate the share of expenditures for 23 consumption categories and the shares of capital, labor, and transfer income for households in each quintile. We used a dynamic overlapping generation model of the US economy to predict the changes in quantity and price of these goods and sources of income in the first year after a $30 per ton carbon tax is implemented

Previous studies have found that the distributional impact of carbon taxes is regressive when analyzed by household income, mainly because poor households spend a larger percentage of their income on energy-related expenses. We find, however, that this regressive effect can be tempered or even reversed by changing the way revenue is allocated, but this also affects overall economic efficiency.

Welfare Change per Household by Quintile (Percent of Income)

The effects across income quintiles of each approach to recycling are illustrated in the figure, along with the effect on the mean household. (Note the income of the mean household is much greater than the median.) As illustrated in the figure, a policymaker would pick the lump-sum rebate if she cared most about what would be economically advantageous in year one for the majority of households, or if she wanted to reduce inequality. The policymaker would pick the labor tax swap if she wanted to balance efficiency with reducing the natural regressively of the carbon tax or if she wanted to have the most even effect across income quintiles. Finally, the policymaker would pick the capital income tax swap if she cared most about the total welfare (economic growth) of the economy.

Our team is continuing to examine the incidence of a carbon tax under different policy designs. An analysis of welfare changes by state and an analysis of distributional effects over time will be released in the coming months.