We use the 2003 Consumer Expenditure Survey and emissions estimates from an input-output model to estimate the incidence of a price on carbon induced by a cap-and-trade program or carbon tax in the US context. We present results on how much difference income deciles pay for a carbon tax as well as which industries see the largest increase in costs due to a carbon tax. We illustrate the main determinant of the regressivity: consumption patterns for energy-intensive goods. We find that a policy targeting CO2 from energy consumption is more regressive than a price on all emissions. Furthermore, on a per-capita basis a carbon price is much more regressive than calculations at the household level. We discuss policy options to offset the adverse distributional effects of a carbon emissions policy.