The fact that a carbon tax is an environmentally and economically efficient instrument for reducing emissions is often highlighted, but the equity story is also of importance: who bears the burden of the tax? This paper addresses the question of the distributional burden of a carbon tax.

The authors show that it is not only the income measure that matters for the incidence of the tax, but also the underlying distribution of income across a society. They look at the Swedish carbon tax on transport fuel and its impacts on different income groups over the period 1999–2012. The overall trend is towards an increase as time goes on in the burden borne by lower income groups – i.e. in regressivity – which is highly correlated with the increase in income inequality that has occurred in Sweden in recent years.

The authors find evidence to support their hypothesis that for expenditure on necessities, rising income inequality increases the regressivity of a tax on consumption. To mitigate climate change, a carbon tax should be applied to goods that are typically characterised as necessities: transport fuel, food, heating and electricity. Carbon taxation is thus likely to be regressive in high-income countries, and increasingly so, the more unequal the distribution of income.

Given the importance of carbon taxation for reducing emissions globally, policymakers therefore need to design carbon tax policy that includes revenue-recycling mechanisms, reductions of distortionary taxes, or other means to offset the regressive effect.

Key points for decision-makers

  • This is the first empirical study of carbon and fuel tax incidence that looks at a longer time period than just one specific year.
  • The authors examine if there is empirical evidence to support the common assertion that carbon taxes are regressive, and what are the most important determinants of carbon tax incidence – i.e. the factors in where the burden will fall.
  • They study the Swedish carbon tax on transport fuel, using data from 1999–2012 on carbon tax expenditure from a large annual household expenditure survey: the tax mainly affects the transport sector and so the analysis focuses only on the carbon tax part of households’ expenditure on gasoline/petrol and diesel.
  • Carbon tax burden is measured as the percentage of a household’s income that is spent on the tax. The authors use measures of annual income, measured as disposable income in any given year; and lifetime income, where total expenditure in a year is used as a proxy.
  • The differences in size of the carbon tax budget share across income groups determines the distributional effect. If the budget share decreases with a move up the income distribution, the tax will be regressive, and the incidence will be progressive if the budget share increases with income.
  • Measured against annual income the tax had regressive impacts; progressive impacts were recorded when using a measure of lifetime income.
  • The authors find that the most important variable explaining variations in regressivity over time in Sweden is changes to income inequality, measured by the Gini coefficient – where 0 equates to complete equality and 100 to complete inequality. They show that at a Gini below 22 the Swedish carbon tax is progressive, and that above a Gini of around 30 the tax is regressive. Income inequality has grown in Sweden since the tax’s implementation in 1991 (to 26.9 in 2012), leading to a more regressive outcome from the tax over time.
  • In analysing previous studies of gasoline tax incidence across OECD countries the authors find a similar strong correlation between regressivity and income inequality: for those countries, below a Gini of 24 a gasoline tax will be progressive and above 29 it will be regressive, using both annual and lifetime income.
Keep in touch with the Grantham Research Institute at LSE
Sign up to our newsletters and get the latest analysis, research, commentary and details of upcoming events.