Stated versus Revealed Preferences on Carbon Tax Revenue Recycling

confidence: high · 4 sources

⚠
Working research — not a recommendation.

This page records what the evidence says a policy does, not what anyone should do about it. It has been synthesized from the cited sources but has not been independently reviewed.

What the evidence says

Stated support for a carbon tax with a uniform dividend was low in the French survey. Only 10 percent of respondents approved of the reform and 70 percent disapproved. Much of this follows from beliefs about who gains. While 70 percent of households were objectively expected to gain, only 14 percent thought they would, a 56 point gap. Only 19 percent thought the policy would benefit the poorest households, and a large majority perceived a progressive scheme as regressive. The beliefs are tied to approval. In the instrumental-variable estimates, believing one does not lose increases acceptance by more than 50 percentage points. Believing the tax is environmentally effective increases approval by more than 40 percentage points.

Those beliefs are hard to move with information about one's own outcome. Among respondents who first thought they would lose, only 12 percent switched to "win" after customized feedback that they would win. Awareness of the actual mechanism is also low. In a Swiss survey of 1,012 respondents, only a quarter knew how the revenue was redistributed.

On how people want revenue used, the stated-preference reviews agree on a ranking. Earmarking for environmental projects ranks first. Social cushioning for vulnerable groups ranks second. Revenue-neutral cuts in other taxes rank last. A review of the literature finds that most studies show no clear public preference for using revenue to reduce regressivity, and that many people prefer environmental projects. The same review finds little trust in governments to use the revenue well. A choice experiment finds that earmarking increases the probability that respondents choose the tax. Respondents also value environmental earmarking more than redistribution earmarking in marginal willingness to pay.

Stated preferences move with information. Information about carbon tax effectiveness reduced doubt about the tax and decreased demand for environmental earmarking. Information about distributional effects led respondents to recognize the favorable distributional properties of lump-sum transfers. The claims give no effect sizes for either shift.

Evidence on revealed preference for carbon tax revenue recycling is thin. The claims offer two analogues from other taxes. After a Swiss region introduced pricing of refuse by the bag, residents were willing to pay 70 percent more than they had stated before. Several congestion-charge cities saw many ex ante opponents change their minds after the charge reduced road use. The claims do not measure how people respond once a carbon tax with a given recycling design is in place.

Where the evidence splits

The reviews disagree on how much weight equity-based uses of revenue carry. One review finds no clear public preference for using revenue to reduce regressivity. Another finds social cushioning is the second most preferred option, and a choice experiment finds that distributional information moves preferences toward lump-sum transfers and social cushioning. The sources point to a design difference: the distributional-information experiment told respondents about distributional effects, and the sources reporting no clear preference do not say whether respondents had that information. The claims do not test this directly.

The latent-class claims are inconsistent with their own quote. The claims describe one segment as preferring environmental earmarking and another as preferring redistribution earmarking. The quoted text says both segments prefer earmarking for income redistribution over the general government budget. The claims do not settle this, and the segment descriptions are unsupported on this page.

The French survey shows that stated opposition is strong and hard to move. The refuse-bag and congestion-charge studies show stated opposition giving way after implementation. These are different taxes, designs and countries, and the sources on this page do not say whether the pattern carries over to carbon taxes.

Thresholds and dose-response

No threshold established in the current sources.

Open empirical questions

Findings

A carbon tax with a uniform dividend has low stated approval, and few people believe they would gain

confidence — direction: highconfidence — magnitude: highconsensus: thin. This is one survey.
mechanism
respondents expect to lose and expect the policy to be regressive. They doubt that the dividend offsets their costs.
effect
10 percent approve and 70 percent disapprove. 14 percent believe they would gain, against 70 percent of households objectively expected to gain. 19 percent believe the poorest would benefit.
conditions / moderators
results are for French respondents in a survey with a €110 per adult annual dividend. A large majority perceived the scheme as regressive although it was progressive.

Beliefs about personal loss and about environmental effectiveness increase approval

confidence — direction: highconfidence — magnitude: mediumconsensus: thin. This is one study.
mechanism
respondents who believe they do not lose, or that the tax reduces emissions, are more likely to accept the reform.
effect
believing one does not lose increases acceptance by more than 50 percentage points. Believing the tax is environmentally effective increases approval by more than 40 percentage points.
conditions / moderators
exact point estimates are not given. Identification uses a fuzzy regression discontinuity and instrumental variables.

Feedback that a respondent will gain decreases the belief that they will lose only for a small share

confidence — direction: highconfidence — magnitude: highconsensus: thin. This is one study.
mechanism
customized information tells respondents they are expected to win. Most do not update.
effect
12 percent of initial "lose" respondents switched to "win".
conditions / moderators
randomized information treatment; the claim covers respondents who initially thought they would lose.

Stated preferences rank environmental earmarking first and revenue-neutral tax cuts last

confidence — direction: highconfidence — magnitude: mediumconsensus: settled for the ranking in stated preferences, with exceptions noted.
mechanism
respondents prefer revenue to support further emissions reductions. Revenue-neutral cuts to other taxes are the least popular option.
effect
earmarking is the most preferred option, social cushioning is second, and cutting other taxes is least preferred. Earmarking increases the probability that respondents choose the tax. Environmental earmarking is valued more than redistribution earmarking in marginal willingness to pay.
conditions / moderators
one source notes exceptions, such as a Sweden example. A review finds that most studies show no clear preference for redistribution to reduce regressivity.

Information shifts stated revenue-use preferences

confidence — direction: highconfidence — magnitude: mediumconsensus: thin
mechanism
information about tax effectiveness reduces doubt about the tax and lowers demand for environmental earmarking. Information about distributional effects shows respondents the favorable distributional properties of lump-sum transfers.
effect
information about effectiveness decreases preference for environmental earmarking. Distributional information increases preference for lump-sum transfers and social cushioning. No magnitudes are given.
conditions / moderators
choice experiments, reported in one review. Effect sizes are not quantified.

People report little trust in governments to use carbon pricing revenue well

confidence — direction: highconfidence — magnitude: mediumconsensus: thin
mechanism
not established in the claims.
effect
direction is negative for trust. No magnitude is given.
conditions / moderators
not established. The claim does not say whether trust was an outcome or a causal variable.

Awareness of the actual recycling mechanism is low after years of implementation

confidence — direction: highconfidence — magnitude: highconsensus: thin
mechanism
not established in the claims.
effect
only a quarter of 1,012 respondents were aware of the mode of redistribution.
conditions / moderators
one Swiss survey (INFRAS, 2015). Tier A in this page's classification.

Implementation of other Pigovian charges increases acceptance relative to stated ex ante positions

confidence — direction: mediumconfidence — magnitude: mediumconsensus: thin. The refuse-bag result comes from one study with a control group.
mechanism
residents experience the charge and its effect, such as reduced road use or less congestion.
effect
willingness to pay for refuse bag pricing increases 70 percent after implementation relative to stated willingness. Many ex ante congestion-charge opponents changed their minds. No magnitude is given for congestion charges.
conditions / moderators
refuse pricing and congestion charges, not carbon taxes. The congestion-charge evidence is qualitative and magnitudes vary across cities.