Climate target acceptability under increasing fuel prices and compensation

Subjects were told that Finland has a target of halving transportation emissions by the year 2030, and that pursuing this target may lead to higher per-liter prices for the fuel that their primary vehicle uses. They were presented with a lower price scenario where emissions stay at their current level, and a higher price scenario under which the emissions target is achieved. They were then given an estimate of how much more they would pay per month in fuel bills at their current consumption level under the higher price scenario, and asked if they would accept it as a cost of achieving the target. The main intervention randomized the per-liter prices presented in these two scenarios, leveraging the substantial uncertainty over price development at the time of the survey: in the months leading up to the start of the trial in late 2022, Finnish gasoline and diesel prices had swinged rapidly between ? and 2.5 euros per liter, inducing salient public debate as well as loosening of climate policy by the government with the stated aim of providing price relief.

The respondents who rejected the higher price were asked in a follow-up if receiving compensation for the higher fuel costs would make them accept the higher price. The amount of compensation was determined as a randomized fraction of their initial cost-estimate, though the respondents were only presented the resulting sum in euros per month. To obtain symmetric information from the respondents that initially accepted the higher price, they were asked if they would still accept the higher price if it was accompanied by increases in other taxes that may need to be raised to cover the costs of pursuing the target, amounting to a randomized monthly euro sum that was emphasized to not depend on their fuel consumption.

Finally, the respondents were asked a series of questions on their attitudes and beliefs on climate policy and future transportation choices, e.g. whether or not they want Finland to pursue abatement in the transportation sector in the first place, and what kind of a vehicle they expect to drive in the future. These questions did not include any randomized treatments, except for a question on the perceived necessity of higher fossil fuel prices as a tool to achieve the target. In this question, half of the respondents were randomly read a short introduction describing the consensus of experts on the cost-effectiveness of pricing emissions before eliciting their opinion on the matter.

2022-10-25

2022-12-03

Share of respondents accepting different price-scenarios before compensation or extra taxes
Share of respondents accepting different price-scenarios after compensation or extra taxes

Share of respondents saying that they think higher fossil fuel prices are necessary to achieve the emissions target

The main intervention randomized the price scenarios in three ways. The most important dimension of randomization was the difference in prices between the low- and high-price scenarios. This was randomly chosen to be either 0.5, 1 or 1.5 euros per liter. Secondarily, we randomized the status quo price between 2 and 2.5 euros per liter in order to test if anchoring respondents to a different starting price would influence responses. These were justified as approximate average prices over the entirety of 2022 and the summer of 2022, respectively. Finally, to test if framing the cost as foregone savings instead of future payments would matter, some respondents were asked if they would accept prices staying at their current level to achieve the target while the low-price scenario abandoning the target would yield a price decrease. All in all, with the three dimensions of randomization, there were eight treatment cells that a respondent could end up in (prices increasing from (1) 2 euros per liter by 0.5 euros per liter, (2) 1 euros per liter or (3) 1.5 euros per liter, or from 2.5 euros per liter by (4) 0.5 euros per liter or (5) 1 euros per liter, or prices decreasing from 2.5 euros per liter by (6) 0.5 euros per liter or (7) 1 euros per liter, or (8) from 2 euros per liter by 0.5 euros per liter). Each cell was presented with ex-ante equal probability, eventually yielding 8 roughly equally-sized groups.

In the follow-up question that introduced the accompanying transfers, the compensations were randomized with equal probability between 50% and 75% of the initial cost estimate determined by the respondents treatment cell and their monthly fuel consumption (which was calculated based on background questions about their primary vehicle's fuel efficiency and annual kilometers driven). The randomization in terms of cost-shares was aimed at producing variation efficiently by ensuring that those with initially high costs would get higher offers in absolute terms, and nobody would be overcompensated. However, in the opposite direction no natural upper bound was available, and it was less clear if those who initially accepted higher costs would accept more or less additional costs. Thus, the additional taxes, were simply randomized with equal probability between fixed sums of 50 euros or 200 euros per month.

The final intervention included in the survey concerned a question on whether or not the respondent thought that higher fossil fuel prices are necessary for Finland to achieve its transportation emissions target. A randomly chosen half of the respondents were first read the following introduction (translated here from the original Finnish text): "Increasing fossil fuel prices above the free market price through taxation or emissions trading is widely considered by economists to be the most cost-effective way to reduce carbon dioxide emissions for society." They were then asked (and the other half directly without the introduction): "Do you personally think that achieving the emission target requires high prices for fossil fuels?"

Randomization done in office by a computer (implemented by Kantar, the survey company we partnered with to conduct the interviews)

Individual

No

2 000 individuals

2 000 individuals

Total final sample size: 2 145
Main question: between 268 and 272 respondents per treatment cell
Follow-up: 1 214 respondents rejected the scenario in the main question, leading to
615 respondents offered 50% compensation and 599 offered 75% compensation and
475 respondents offered 50€ extra tax and 456 offered 200€ extra tax

Secondary information treatment: 1 081 receiving the introduction, 1 064 not