What the evidence says
The sources on this page measure public preference over how carbon tax revenue is used. They do not measure emissions, prices, or household finances. All six claims come from two sources: a review by Carattini (2018) and a paper by Muth (2023).
Focus groups in five European countries (Denmark, France, Germany, Ireland and the United Kingdom) ranked three revenue uses. Earmarking revenue for further emissions reductions was most preferred. Social cushioning for vulnerable groups, such as low-income households and people in remote areas, came second. Revenue-neutral tax reductions came last. The review also reports that when there is a clear trade-off between environmental earmarking and socially progressive redistribution, people tend to prefer earmarking. These findings are associational. The focus-group evidence is qualitative and gives no magnitudes.
A Swiss discrete choice experiment (Carattini et al., 2017) tested what changes that ordering. Giving respondents information about the environmental effectiveness of different carbon tax designs decreases their preference for environmental earmarking. With that information, and information on distributional effects, the most favored uses were lump-sum transfers and social cushioning. The review names the mechanism: the information reduced voters' doubt about the effectiveness of carbon taxes, and that in turn reduced the demand for earmarking. Earmarking, on this account, was demanded mainly to make the tax reduce emissions.
Muth (2023) cites experimental research on a different design, an equal tax rebate to all families. In the US, that scheme increased carbon tax acceptability. A similar scheme in Sweden decreased public support. Muth's text does not state effect sizes for either experiment.
No claim here compares a carbon dividend with earmarking in a single head-to-head design across countries, and none gives numeric effect sizes. The evidence is strongest on direction and thin on magnitude. The evidence base is also narrow: European focus groups, one Swiss experiment, and two experiments cited secondhand.
Where the evidence splits
Earmarking is preferred over redistribution in the focus groups and in the trade-off literature the review cites. In the Swiss experiment, lump-sum transfers and social cushioning were the most favored uses once respondents had information. The sources explain this split. Respondents in the Swiss experiment received information about environmental effectiveness and distributional effects, and the review says that information reduced doubt about effectiveness and so reduced demand for earmarking. The sources do not say whether the focus-group participants had the same information.
Rebates increased acceptability in the US and decreased support in Sweden. The sources on this page do not explain the split. Muth's text gives no design, population, or period difference between the two experiments.