Examining the effect of cost information and framing on support for methane regulations in Europe

Methane abatement policies will play an important role in mitigating climate change given the high global-warming potential of methane compared to carbon dioxide. Yet evidence on public attitudes and support for methane regulations is lacking. In partnership with the Clean Air Task Force, we develop an original nationally representative survey of four European countries (N = 5629) to show variation in public opinions about methane emissions and policy to reduce them. Using a framing experiment, we test variation in these preferences as a function of policy impacts on cost, global climate change, local pollution, or energy security. We find largely null effects across the board: attitudes are remarkably durable to varying treatments, suggesting that support for methane regulations is not sensitive to cost information and policy framing. The results from this survey provide a much-needed baseline for public attitudes about methane abatement and will inform existing debates on what information is and is not effective in generating support for ambitious methane policy.


Introduction
Curbing methane emissions is crucial to addressing the climate crisis. Methane is a short-lived but highly potent greenhouse gas, which is estimated to have contributed to 0.5 • C of global warming relative to the pre-industrial period (IPCC AR6 Working Group I Report). International agreements, such as the Global Methane Pledge and nationally determined contributions in the Paris Agreement, are among the most significant commitments to reduce methane emissions (IEA 2023). As with other global climate agreements, success will depend on domestic public support for methane regulations (Bernauer and Gampfer 2015, Tingley and Tomz 2022).
The politics around regulating methane are likely to be distinct from the more well-studied case of carbon dioxide (CO 2 ), due to methane's dual character as an externality and an input, the public's everyday experience with methane gas, and the proximity of problems caused by methane pollution. But scholars have paid relatively little attention to the politics around regulating methane. Considerable research addresses public opinion on climate policies generally, including domestic climate action (e.g. Howe et al 2015, Bernauer and Beiser-McGrath 2016, Stokes and Warshaw 2017, Bergquist et al 2020b and international climate cooperation (e.g. Bechtel and Scheve 2013, Bechtel et al 2019, Genovese 2020. The vast majority of these studies assesses public support for efforts to reduce CO 2 emissions, whereas very little public opinion research has focused on methane. In this study, we report results from the first crossnational survey of public attitudes toward methane regulations. We examine baseline levels of knowledge and beliefs about methane as a component of natural gas, as well as support for regulating methane emissions from natural gas imports, in four European countries. Next, we draw on the literature assessing public support for CO 2 -reduction policies, and on similarities and differences between CO 2 and methane, to assess the role of cost considerations and benefit frames on support for aggressive regulations on gas imported into the European Union (EU). We find high support for aggressive regulations, but we find no distinguishable sensitivity to cost information or benefit framing.

Methane as a distinctive case in climate politics
Methane is distinct from other greenhouse gases in at least four ways that may imply differences in public attitudes and reactions to cost information and policy framing. First, methane is economically useful as an energy and heating fuel (as well as a nuisance byproduct of oil extraction, landfills, and manure management). Methane that is lost to the atmosphere is therefore perceived as wasteful and could lend itself to cross-partisan coalitions supporting regulations that would reduce waste (Rabe 2021).
Methane's value as an input contrasts with other greenhouse gases like CO 2 and nitrous oxide, which are pure externalities. The emission of these other pollutants, which have little economic value as inputs, imposes costs on society that are not accounted for in the costs of production or (therefore) in the prices of products. At a simplistic level, this means that regulating CO 2 and other non-methane emissions adds costs in the short term. Regulation that mitigates methane pollution, by contrast, may reduce costs in the short term by reducing waste of a valuable input.
Second, consumers use methane directly as a component of natural gas, for example as a heating, cooking, and transportation fuel. Personal, direct, and daily experience could make natural gas-and by extension the regulation of methane-quite salient to consumers. In Germany and the US, for example, bans on gas appliances have become political flash points (Chazan 2023, Joselow 2023. To be sure, the ubiquity of other fossil fuels to modern life and economic production can also make their regulation salient to consumers. The crucial distinction is that the regulation of other fossil fuels targets them as externalities associated with production, rather than targeting the inputs themselves 3 . How will this distinction shape public attitudes about methane as compared with CO 2 and other greenhouse gases? This question motivates the present study and (we hope) future work on this topic.
Third, the localized nature of costs associated with methane pollution also distinguishes it from other greenhouse gases. Specifically, unlike CO 2 emissions that primarily affect global climate change, methane emissions spur localized health effects. The Climate and Clean Air Coalition estimates that human-caused methane emissions and their copollutants are responsible for over 1 million deaths per year globally, due to their contribution to the formation of ground-level ozone (Michanowicz et al 2021. As a result, the health effects of methane emissions create local policy problems akin to those caused by purely local pollutants like soot and particulates (Sprinz andVaahtoranta 1994, Ansolabehere andKonisky 2016). From this perspective, methane resembles non-CO 2 gases like chlorofluorocarbons (CFCs) and hydrochlorofluorocarbons (HCFCs). CFCs and HCFCs cause ozone depletion that leads to cancer and crop loss and can be easier to justify regulating on a short-term cost-benefit basis (Bernauer 2013). The success of global treaties like the Montreal Protocol to phase out ozone-depleting emissions may provide a framework for similar agreements around methane abatement (Butler et al 2016, Ocko et al 2021). In contrast scholars have theorized (with mixed empirical support) that the perceived temporal and spatial distance of problems caused by CO 2 has been an obstacle to building support for aggressive policy to mitigate climate change (Spence et al 2012, Brügger et al 2015, 2016, Schuldt et al 2018. Consistent with these theories, the temporallyand spatially proximate nature of health problems caused by methane emissions might boost support for methane regulations. Finally, reducing methane emissions from petroleum production has implications for security policy, given global reliance on belligerent states for imported oil and gas (van de Graaf and Colgan 2017). Following the 2022 invasion of Ukraine, concerns about financially aiding and abetting the Russian regime coincided with rising demand for 'responsibly sourced gas'-natural gas that is produced while minimizing methane leakages (Project Canary 2021, Halff et al 2023). This resulted in broad support for import bans on Russian oil and gas in Europe, which-in addition to the growing policy debate over the European Green Deal and the EU Carbon Border Adjustment Mechanism-likely made the issue of methane regulations highly salient to the public when our survey was fielded in Summer 2022. The timing of our study offers a valuable opportunity to analyze attitudes toward methane policies during a confluence of energy crisis and energy transition. This type of context is likely to emerge elsewhere as governments face difficult choices over decarbonization. To be sure, the increased salience of gas may also impact the study's replicability (e.g. outside of Europe or, counterfactually, in previous time periods), a topic to which we return in the conclusion.
These distinctions between methane and other greenhouse gases may influence baseline attitudes in addition to changing the way that people respond to information about the costs and benefits of regulations. In the US, for example, there is wider support for methane taxes on industry relative to support for carbon taxes. The US Inflation Reduction Act of 2022 passed with significant methane fees (ranging from $ 900 to $ 1500 per ton) on oil and gas facilities (McCormick 2022), despite state-level resistance to including methane emissions in severance taxes on production (Rabe et al 2020).

Describing public knowledge and beliefs about methane
We extend upon this observation in the US, by examining baseline knowledge, beliefs, and policy support relevant to methane in the EU. In early August 2022, we fielded a nationally representative survey with 5629 respondents in France, Germany, Italy, and Poland. The survey was administered online using the YouGov platform. The full survey instrument is included in the supplementary materials.
We selected cases with political relevance to EU energy regulation. Germany, France, and Italy are the three most populous countries in the EU, and the largest consumers of gas (BP 2023). Poland is the fifth-most populous country and the sixth largest gas consumer. We include Poland additionally because of its special political relevance as a historic fossil fuel producer and the source of the largest energy-related methane emissions in the EU (IEA 2023). While none of the four countries is a major global oil and gas producer, Germany and Italy are the top two oil refiners in the EU (BP 2023). Both France and Poland have significant but largely untapped shale reserves, the development of which (to replace imports from Russia) emerged as a subject of debate in 2022-23 (Harper 2023). All of these features mean that these four countries are pivotal to the debate over energysector methane regulations in the EU.
Figures 1 and 2 visualize public knowledge and beliefs about natural gas and methane. Together they suggest that public knowledge about methane is not particularly strong, but that concern about its association with climate change is quite high. Figure 1 summarizes the results when we asked respondents which two phrases they most associate with either 'natural gas' or 'methane' (randomly assigned). The figure shows that word choice matters: with the exception of Italy, respondents were more likely to draw an association with home heating and cooking if the words 'natural gas' were used instead of 'methane.' Notably, in Italian, the word metano ('methane') is often used interchangeably with gas naturale ('natural gas'). This echoes similar findings in the United States (Lacroix et al 2021) and is important vis-a-vis our claim that methane may be salient with the public due to its importance in daily life. Figure 1 shows that the veracity of this claim is sensitive to word choice. Conversely, respondents were more likely to associate methane, rather than natural gas, with both air pollution and greenhouse gases (again, with the exception of Italy). Supplementary figure S1 reinforces these findings, by showing the limited extent to which the public recognizes that methane is the primary component of natural gas. The links between 'natural gas' and 'methane' appear tenuous and conditional in the public mind. Thus, the salience of local health problems and global climate changes associated with methane may also be sensitive to word choice. Figure 2 visualizes public understanding of the relationship between methane and climate change. Substantial majorities in all four countries understand that methane in the atmosphere has increased in recent years. Moreover, as suggested by the strong associations between methane and the phrases 'greenhouse gas' and 'fossil fuel' (figure 1), the public recognizes that methane poses a problem for the climate. The bottom panel of figure 2 shows that large majorities of respondents view methane as a major or minor problem for the climate.
Next we examine whether this perception of methane as a climate problem translates into public support for methane regulations. Figure 3 shows public support for several alternatives for addressing energy-sector methane, each of which is the subject of policy debate in the EU (e.g. EC COM/2020/663; COM/2021/805. See European Commission (2022) for an overview of proposals). The figure shows strong support for each type of methane policy. Such strong support suggests that preferences for reducing methane pollution are not sensitive to the typical objections to EU-level climate regulations as 'technocratic imposition' over national-level regulations (Wurzel et al 2016, Skjae rseth 2018. Support for a fee on imported oil and gas from countries with high rates of methane emissions is noticeably lower than support for the other policies. This is consistent with similar comparisons of public support for climate and transportation fee-based policy alternatives in China, Germany, and the United States

Effect of cost information and benefit framing on support for methane regulations
Given high baseline support for regulations, how does the public respond to information about the costs and benefits of such regulations? Scholars have focused considerable attention on how information and framing shape public support for aggressive climate action in general. This literature has shown that the public is sensitive to information about the economic costs and benefits of policies to reduce greenhouse gas emissions or promote a transition to cleaner energy sources (  et al 2022). The public seems less responsive to communication that emphasizes the non-economic benefits of climate policies, such as public health improvements and climate risk reduction (Bernauer and Beiser-McGrath 2016). We examine whether these findings hold in the case of methane regulations. Additionally, we ask whether the public is sensitive to an energy security frame. Security might be especially pertinent in the case of methane due to the geopolitics of the gas industry, particularly concerns about Russian dominance of the European gas market and Russia's aggression toward Ukraine and other eastern European states (Tollefson 2022).
We use a survey experiment to examine the effects of cost information and benefit framing on public support for methane regulations. Our analysis follows a pre-analysis plan that was preregistered with the Evidence in Governance and Politics Initiative (Bergquist and Mahdavi 2022). The experiment begins with a statement introducing the debate around methane regulations in the EU. We focus on the application of the regulations to imported gas, as a way of distinguishing the stringency of proposed regulations. This distinction is drawn from debates (precedent and contemporary to our study) around extending the European Commission's proposed methane regulations such that they would apply to imported gas (CATF 2021).
Respondents were then block-randomized (by country and political ideology) into one of six experimental conditions or to a control condition. The six experimental conditions were created by interacting a dichotomous cost information variable with a three-way benefit frame variable. The cost information conditions correspond to information that the regulations would either raise or lower household heating and energy costs. The benefit frame conditions correspond to an emphasis on one of three benefits associated with the regulations: mitigating climate change, addressing problems associated with localized methane pollution, or improving energy security in the EU. Thus, each respondent outside the control group was exposed to either of two informational blurbs about the cost of the proposed methane regulations, followed by one of three frames emphasizing particular benefits of the regulations. The control group only read the introductory statement. The full text of our experimental treatments is included in the supplementary materials.
We visualize our experimental results in figure 4 by showing 90% and 95% confidence intervals around the point estimates from three regressions of our outcome on an indicator variable for each condition. We control for political ideology, gender, age, income, education, and country fixed effects. Tabular results, including p-values adjusted for multiple hypothesis tests, are available in supplementary tables S6-S8.
Given our randomization design, we first examine the effect of each cost x benefit treatment combination. We compare the effect of each of our six treatment conditions with a control condition in which respondents were not shown any information about the costs or benefits of methane regulations. To examine heterogeneity of framing effects between cost-information conditions, we use the Wald test to assess whether the coefficients for each benefit frame are different in the cost-increasing and cost-reducing Figure 2. Knowledge, perceptions, and concern about methane: the figure shows respondents' understanding of methane as a gas in the atmosphere (top panel), and perception that methane is or is not a problem for the climate (bottom panel). Question wording is included in the supplementary materials. condition. We do not find evidence that individuals in either of the cost-information conditions are differentially responsive to any of the benefit frames (left panel, figure 4).
We next examine sensitivity to cost information, by marginalizing over the benefit-frame conditions. Once we marginalize across the three benefit-frame conditions, we do not find evidence that support for methane regulations is sensitive to information about cost impacts (right panel, figure 4). Compared to a control condition without cost information, respondents are slightly less likely to support regulations if told that energy costs would be higher, but the effect is not statistically distinguishable from zero either before or after applying the Benjamini-Hochberg adjustment (average marginal effect = −0.020, p BH = 0.28). Moreover, this two percentage point (pp) decrease in support is substantively small, as compared with the difference in support between respondents with left-leaning and non-left-leaning ideology (9 pp, shown in table S6). These effects are also small compared with prior studies of costsensitivity in support for CO 2 -reduction policies. For instance, in a recent study of US residents, support for comprehensive climate, economic, and social policy declines by 12-21 pp in response to information that the policy would impose costs on households (Bergquist et al 2020b, 4-5) 4 . The effect of the low-cost condition is nearly zero and not statistically significant (AME = −0.007, p BH = 0.61).
We also examine the effects of the benefits frames by marginalizing across the two cost-information 4 Note that, in contrast with the Bergquist et al (2020b) and other studies focused on CO2-reduction policies, we did not provide respondents with a specific numeric cost of methane regulations as we do not have a baseline reference for increased or decreased household costs. Here, support for a methane fee has been marginalized across responses to questions in which we randomly varied information about the consumer cost and use of revenues from such a fee. Responses did not vary substantially across these conditions. Figure 4. Effect of cost information and benefit framing on support for stronger methane regulations: effects are estimated using OLS regression, with HC2 standard errors, of policy support on cost information or benefit framing (right panel), or on the interaction of cost information and benefit framing (left panel). The dependent variable is a dichotomous indicator, where 1 indicates support for methane regulations and 0 indicates opposition. Thin lines show 95% confidence intervals, and thick lines show 90% confidence intervals. The models include controls for political ideology, gender, age, income, education, and country. Tabular results are available in supplementary tables S6-S8.
conditions. We find little evidence that benefit framing impacts support for methane regulations (right panel, figure 4). Emphasizing neither the local health benefits (AME = 0.003, p BH = 0.85) nor global climate benefits (AME = −0.014, p BH = 0.51) of reducing methane emissions has a noticeable effect on policy support. There is a slightly negative effect of emphasizing the energy security benefits of reducing methane emissions, but once adjusted with the Benjamini-Hochberg correction for multiple comparisons, the result is not statistically significant (AME = −0.030, p BH = 0.11). Moreover, the direction of this effect is inconsistent with our alternative hypothesis that the security treatment would increase support for the policy.
We pre-registered a number of conditional effects as exploratory hypotheses to gauge heterogeneity in the treatment effects. For each of the three treatments discussed above, we assess whether effects are conditional on country, ideology, age, education, and income. In addition to the ability to explore heterogeneous treatment effects, this assessment allows us to examine whether our overall effects are systematic null effects, or whether they mask effects that vary across relevant subgroups (Holbein et al 2021).
Only five of 165 tests are statistically significant, roughly what we would expect by chance alone if the null hypothesis were true (supplementary tables S11-S13). We therefore infer that there is no evidence for subgroup heterogeneity, suggestive of systematically null effects across five salient respondent characteristics. In addition to these pre-registered analyses, we explore whether effects vary across respondents with higher and lower levels of knowledge about methane (supplementary table S10). Here too, we find little evidence that our null effects mask treatment-effect heterogeneity.

Discussion
In this study, we present results from the first crossnational study of public attitudes toward regulating methane in the energy sector. Public knowledge about methane is not particularly strong, and, with the exception of Italians, European publics appear not to equate 'natural gas' with 'methane.' Still, we find remarkably high levels of concern about methane as a climate problem, and support for proposed EU regulations on methane pollution from the oil and gas sector. This finding is somewhat unsurprising, in light of methane's value as an input to the energy system, consumer familiarity with it as a fuel, and the proximity of health problems associated with methane pollution. We also show that public support for methane regulations is quite resistant to change in response to cost information or to framing that emphasizes climate, health, or security-related benefits of the policy. The resilience of public support for methane regulations in the face of information that the regulations would increase costs stands in stark contrast to public support for CO 2 regulations, which is highly costsensitive. We do not find evidence that the lack of detectable effects masks heterogeneous effects within groups of respondents or between different combinations of cost information and benefit frames. Our results suggest that EU policymakers have substantial latitude to enact stringent regulations on methane in the energy sector. On the other hand, differing associations with 'natural gas' and 'methane' suggest that the salience of methane and support for policy to regulate it may be sensitive to wording.
We hope this study inspires future research assessing the contours of support and opposition for methane regulations around the world. First, future work should assess public attitudes about methane regulations in other sectors such as agriculture and coal production. Consumers may be particularly supportive of regulations on gas production because of methane's direct value as an energy input and the salience of energy security concerns in light of the 2022 Russian invasion of Ukraine. In other sectors, methane is more similar to CO 2 and nitrous oxides in its character as a byproduct of economically productive activity. Future work should investigate whether public support for methane regulations in these sectors is as high or as unresponsive to information and framing. Second, the high level of baseline support for methane regulations raises the concern that a ceiling effect prevents our design from detecting increases in support in response to benefit frames. The study thus paves the way for scholars to develop survey instruments gauging support for methane regulation in a way that maximizes variance across responses. Relatedly, future work could investigate how support for methane policy varies according to policy design details that we did not explore in this paper. Our study shows quite high levels of baseline support for methane regulations in general, but we would not interpret our study to suggest that 80%-90% of the residents of these countries would support any methane regulation that the European Parliament might consider. Instead, in the spirit of a large body of existing research on climate policy (e.g. Stokes and Warshaw 2017, Wicki et al 2019, Bergquist et al 2020b, Bechtel et al 2021, future work should investigate support for specific policy instruments and combinations of policy attributes. This effort is crucial to understanding public support for domestic and international methane-reduction initiatives, which are a critical component of climate action given methane's high global-warming potency. Third, we hope that future work will expand the geographic scope of our analysis to examine the contours of public support in other EU countries and beyond Europe. Such research will provide important tests of how the topical salience of gas-related policy-whether because of interstate conflict, decarbonization efforts, unexpected market shocks, or a combination of these factors-plays a role in the formation of public attitudes about methane regulations.

Data availability statement
The data and code needed to replicate the findings from this study are available in the Harvard Dataverse: https://doi.org/10.7910/DVN/OMUYTO. anonymous reviewers for helpful comments on earlier versions of this research.

Ethical statement
This research was conducted in accordance with the principles embodied in the Belmont Report and was approved by Georgetown University's Institutional Review Board, study number 00005492. All participants gave their written informed consent to participate in the study.