Ecological Economics 199 (2022) 107507 Contents lists available at ScienceDirect Ecological Economics journal homepage: www.elsevier.com/locate/ecolecon Analysis Climate concern and policy acceptance before and after COVID-19 Stefan Drews a, *, Ivan Savin a, b, Jeroen C.J.M. van den Bergh a, c, d, Sergio Villamayor-Tomás a, e a Institute of Environmental Science and Technology, Universitat Autònoma de Barcelona, Spain b Graduate School of Economics and Management, Ural Federal University, Yekaterinburg, Russian Federation c ICREA, Barcelona, Spain d School of Business and Economics & Institute for Environmental Studies, VU University Amsterdam, the Netherlands e Institute of Geosciences, Department of Cartography, Federal University of Minas Gerais, Brazil A R T I C L E I N F O A B S T R A C T Keywords: It remains unclear how COVID-19 has affected public engagement with the climate crisis. According to the finite- Climate change pool-of-worry hypothesis, concern about climate change should have decreased after the pandemic, in turn Climate policy reducing climate-policy acceptance. Here we test these and several other conjectures by using survey data from Coronavirus 1172 Spanish participants who responded before and after the first wave of COVID-19, allowing for both Policy support Panel study aggregate and within-person analyses. We find that on average climate concern has decreased, while acceptance of most climate policies has increased. At the individual-level, adverse health experiences are unrelated to these changes. The same holds for negative economic experiences, with the exception that unemployment is associated with reduced acceptance of some policies. Complementary to the finite-pool-of-worry test, we examine three additional pandemic-related issues. As we find, (1) higher climate concern and policy acceptance are associated with a belief that climate change contributed to the COVID-19 outbreak; (2) higher policy acceptance is asso­ ciated with a positive opinion about how the government addressed the COVID-19 crisis; (3) citizens show favorable attitudes to a carbon tax with revenues used to compensate COVID-19-related expenditures. Overall, we conclude there is support for addressing the global climate crisis even during a global health crisis. 1. Introduction and show sufficient concern about it. A logical consequence of such concern is that citizens express greater acceptance of climate policies There has been much speculation about whether the COVID-19 aimed at emissions reduction (Bouman et al., 2020b). A main reason pandemic has affected support for mitigating climate change. Com­ why virtually no country is on track to achieve the Paris climate targets mentators have pointed at similarities, differences and other connec­ is that they lack sufficiently strong climate policies. This can be partly tions between these two global crises. While some argue that responses explained by inadequate public support or even resistance against such to COVID-19 can accelerate the implementation of ambitious climate policies (Carattini et al., 2018; Kyselá et al., 2019; Savin et al., 2020). policies and a transition to a low-carbon economy (Markard and Rose­ The COVID-19 crisis presents a unique natural experiment to test if a nbloom, 2020; Engström et al., 2020), others are more skeptical (Gosens major threat to health and the economy affects how people feel about and Jotzo, 2020; Mukanjari and Sterner, 2020). Focusing on the atti­ climate action and policy. There are reasons to expect both negative and tudes of citizens, various theoretical and opinion articles suggest that the positive effects. A detrimental effect of the pandemic may be based on COVID-19 pandemic presents opportunities as well as challenges for the psychological argument that individuals have a so-called “finite pool public acceptance of climate policy (Howarth et al., 2020; Botzen et al., of worry” (Weber, 2006). According to this idea, once a concern about 2020; Reese et al., 2020; Bouman et al., 2020a). one issue (e.g., the health or economic situation due to COVID-19) goes The present study empirically examines how citizens’ perceptions up, for another (e.g., climate change) it goes down, with potentially and attitudes regarding climate change and climate policies have been negative consequences for associated behaviors and policy preferences. affected by the COVID-19 pandemic and associated measures, such as Although the idea of a finite-pool-of-worry is frequently mentioned in confinement and travel restrictions. For societies to be able to combat research on climate change perceptions (e.g. Whitmarsh, 2011; Van Der climate change, citizens need to accept the seriousness of the problem Linden, 2017), to our knowledge it has actually not seen much empirical * Corresponding author at: ICTA, Edifici Z, UAB Campus, 08193 Bellaterra, Spain. E-mail address: stefan.drews@uab.cat (S. Drews). https://doi.org/10.1016/j.ecolecon.2022.107507 Received 8 November 2021; Received in revised form 14 March 2022; Accepted 27 May 2022 Available online 1 June 2022 0921-8009/© 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by- nc-nd/4.0/). S. Drews et al. Ecological Economics 199 (2022) 107507 testing, neither generally nor in the context of the recent COVID-19 2. Hypotheses pandemic. One recent study for the UK finds limited evidence for the finite-pool-of-worry hypothesis (Evensen et al., 2021). Drawing on This study involves three main research aims. The first is to test the survey data with the same respondents before and after COVID-19, the finite-pool-of-worry hypothesis (Weber, 2006), which suggests that the authors find little difference in the perceived severity of climate change pandemic reduced climate concern and policy acceptance over time and other beliefs, such as the anthropogenic nature of the climate crisis. among individuals who are concerned about, or affected by, different Related research examines how perceptions of climate change can aspects of the pandemic. As mentioned above, the finite-pool-of-worry weaken in the advent of particular events, such as the global economic hypothesis tells us that people can only pay attention to a limited crisis of 2008 (Capstick et al., 2015). number of concerns and new concerns can then crowd out existing ones The pandemic might, however, also create positive effects on climate (ibid). Thus, we predict that on the aggregate level both climate concern concern and policy acceptance. One pathway could be that climate (Hypothesis 1a) and policy acceptance (Hypothesis 1b) will be lower change is perceived as a contributing factor to the outbreak of COVID-19 after COVID-19. In addition, we expect that individuals with stronger or to the spread of infectious diseases more generally (Smith et al., 2014; negative economic- and/or health-related experiences and perceptions van Wijk et al., 2020). A second possibility is that people’s evaluations of regarding COVID-19 are more likely to show a decrease in their climate how the COVID-19 crisis has been managed affects how they judge other concern (Hypothesis 2a) and policy acceptance over time (Hypothesis public policies, such as those to address climate change. This is related 2b). to, but distinct from, previous research showing that if citizens consider An alternative proposition, complementary to the finite-pool-of- the government to be untrustworthy or inefficient, they may show lower worry hypothesis, is that concern and acceptance may have increased policy support (Hammar and Jagers, 2006; Savin et al., 2020; Kitt et al., over time for people who perceive that climate change has actually 2021). contributed to the outbreak of COVID-19. The impact of climate change Three main survey-based approaches are available to examine on the spread of infectious diseases was already acknowledged by the changes in climate attitudes and perceptions. The first one is to collect IPCC (Smith et al., 2014). Research also shows that perceiving and data at one specific point in time on climate perceptions and variables communicating health consequences of climate change can increase that might have caused changes, such as perceptions of recent economic public engagement (Akerlof et al., 2010; Myers et al., 2012; Amelung trends (Kachi et al., 2015; Bakaki and Bernauer, 2018). The second, et al., 2019). Moreover, a multinational survey conducted in March repeated cross-sectional design, uses data from different samples and 2020 found that a majority of the participants believed that climate different points in time to assess changes or differences at the aggregate change influences infectious diseases (van Wijk et al., 2020). This leads level (Scruggs and Benegal, 2012). The third and arguably best approach us to expect that individuals who perceive climate change to be a factor is to examine temporal changes for samples of the same individuals, contributing to the outbreak of COVID-19 show an increase in climate known as longitudinal or panel studies, which allows assessing within- concern (Hypothesis 3a) and policy acceptance (Hypothesis 3b). individual changes (Mildenberger and Leiserowitz, 2017; Jenkins- Another proposition is that there is not just a perceived link between Smith et al., 2020). perceptions of the two problems (COVID-19 and climate change) but The present COVID-19 situation presents an opportunity to also between their solutions. For example, research shows that increased contribute to this literature. This study examines how economic, health support for a plastic charge is linked to increased support for other and other aspects of COVID-19 have affected perceptions and policy environmental policies – which has been termed ‘policy spillover’ attitudes associated with climate change. While giving attention to (Thomas et al., 2019). In the present case, people who favorably assess aggregate-level results, the thrust of our study consists of analyzing the government’s response to tackle the COVID-19 crisis may have individual-level changes of climate perceptions and attitudes. To this increased their acceptance of climate policy (Hypothesis 4), regardless end we use the same sample of Spanish survey respondents (n = 1172) of their concern. from before and after the start of the COVID-19 pandemic. This goes The second research aim is to examine whether changes in climate beyond the approach adopted by the earlier mentioned UK study concern are related to changes in policy acceptance due to COVID-19. (Evensen et al., 2021) by uncovering several underlying factors of po­ An implication of the finite-pool-of-worry hypothesis would be that tential changes over time. Our focus on Spain is relevant given that it lower climate concern translates into lower policy acceptance (Hy­ belongs to the countries that have been most affected health-wise by the pothesis 5). This is based on prior cross-sectional research demon­ COVID-19 pandemic and economically by associated preventive mea­ strating that climate concern is associated with climate policy sures (Section 3.2.2 provides more details about the Spanish context). acceptance (Bouman et al., 2020b). While it is plausible to assume that We have three research aims. First, the core of this paper is to detect this relationship also holds in a dynamic setting, some authors have changes in climate concern and policy acceptance, and how these are stressed the need for empirical testing of this assumption (Capstick et al., associated with experiences and perceptions regarding health-, eco­ 2015; Mildenberger and Leiserowitz, 2017). nomic-, climate- and policy-related aspects of COVID-19. Second, we The third research aim puts the previous two aims in a broader assess whether changes in climate concern are associated with changes context. Past research shows that policy attitudes are affected by in policy acceptance. A distinctive feature of this study is that we ac­ numerous other factors than climate concern, including policy design count for within-person changes over time, which allows assessing how and framing (Drews and van den Bergh, 2016). Specifically, perceived an external event such as COVID-19 affect attitudes and beliefs. To economic and social co-benefits from climate action can increase policy complement this analysis, our third aim is to analyze how features of support (Bain et al., 2012). For example, a carbon tax tends to be viewed climate policy – notably social and health co-benefits – matter in more favorably by citizens when its revenues are recycled for specific attenuating the potential negative effects of COVID-19 on policy purposes (Beiser-McGrath and Bernauer, 2019), such as to alleviate acceptance. Specifically, we examine whether the acceptance of a car­ regressive distributional impacts of the policy, compared to a situation bon tax is strengthened when the tax not only addresses climate change where the revenues are unspecified or used for the general governmental but presumably also helps to tackle COVID-19. Addressing this third budget (Carattini et al., 2018; Maestre-Andrés et al., 2019, 2021). Based research aim means deriving insights about post-COVID-19 policy atti­ on this insight, we derive two expectations in the context of the tudes as we cannot assess changes over time, that is, we conduct a cross- pandemic. First, a carbon tax has been argued to support the recovery sectional analysis. from COVID-19 by using the resulting revenues to stimulate the econ­ omy and compensate for extra COVID-19 expenditures (Burke et al., 2020; Galbraith and van den Bergh, 2020; Mintz-Woo et al., 2020). Hence, we expect that acceptance of a carbon tax that explicitly uses 2 S. Drews et al. Ecological Economics 199 (2022) 107507 revenues for addressing the COVID-19 crisis will be higher than for an and geographical distribution, making the survey sample representative unspecified use of the revenues (Hypothesis 6a). Second, we are inter­ of the general population on these characteristics (Table A1 shows a ested in examining whether spending money to mitigate the conse­ comparison of key survey characteristics with census data). The quences of COVID-19 pandemic represents an opportunity to raise response rate was 59% with a sample size of 2004 participants over 18 acceptance of climate policy, particularly among people who have years old and the average completion time was 15 min. This first survey become less concerned about climate change (i.e., expectedly due to serves as the pre-COVID-19 baseline. The second survey was not origi­ increased concerns about COVID-19). Thus, we expect that the positive nally planned by the research team, but the unfolding COVID-19 crisis difference between acceptance of a carbon tax with COVID-19 and un­ several months after the first survey presented an opportunity to study specified revenues will be higher for people who have reduced their both individual-level as well as aggregate changes of perceptions and climate concern (Hypothesis 6b). attitudes. The data were obtained in June 2020 with a two-step sampling A summary of the research framework and hypotheses is provided in procedure. We first invited individuals who already participated in the Fig. 1. As stressed in Section 1, our focus is on understanding the 2019 survey, resulting in 1172 respondents (58% response rate). The individual-level changes of climate concern and policy acceptance sample was then supplemented with new respondents to achieve a (Hypothesis 2–5, middle panel in Fig. 1). This focus is accompanied by sample size of at least 2000, while assuring representativity on the same tests of propositions related to the aggregate-level changes (Hypothesis characteristics as for the first survey. Among the new participants the 1a/b, upper panel), and cross-sectional differences of post-Covid-19 response rate was 82%. policy acceptance (Hypothesis 6a/b, lower panel). Table A2 in the Appendix compares the overall samples in 2019 and 2020 with the subsample of respondents who participated in both sur­ 3. Method veys on a broad range of characteristics, including key socio- demographics, income and employment, experience with COVID-19 as 3.1. Data well as climate concern and acceptance of carbon taxation. In particular, we run a Kolmogorov-Smirnov test (with the null hypothesis that the For this study we draw on two online surveys from Spain, conducted two samples are drawn from the same distribution) and a Wilcoxon rank by using a panel of the survey company “Netquest”. Both surveys were sum test (equivalent to the Mann-Whitney test, with the null hypothesis approved by the University’s ethics committee. The first one was run in that one distribution is not stochastically greater than the other). As one July–August 2019. Sampling was done by using quotas on age, gender can see, the subsample of repeated survey participants is very similar to Fig. 1. Framework and hypotheses of this study. Note: The focus lies on individual-level changes of climate concern and policy acceptance (Hypothesis 2–5, middle panel). This is accompanied by tests of hypotheses related to aggregate-level changes (Hypothesis 1a/b, upper panel), and cross-sectional differences of post-COVID-19 policy acceptance (Hypothesis 6a/b, lower panel). 3 S. Drews et al. Ecological Economics 199 (2022) 107507 the overall samples of the two surveys and differs significantly only with 3.2.2. Explanatory and control variables respect to gender (slightly fewer women in the subsample participating The 2020 survey included various questions to measure people’s in both surveys) and age (higher average age of those who participated experiences and perceptions related to health, economic, climate and in both surveys). Since the differences are few and small, we can policy aspects of COVID-19. One variable captures participants’ threat conclude that the subsample of 1172 respondents is fairly comparable to perception related to COVID-19. Two variables measure respondents’ the representative full sample for Spain in 2019 and 2020. Given that past health problems due to COVID-19, either for themselves or their our main interest is in examining within-person changes, our analysis family and friends. Another variable measures people’s psychological henceforth focuses on this subsample of respondents. stress due to the pandemic and associated confinement measures. Two survey items cover economic aspects, namely respondents’ perceived 3.2. Variables income loss during the pandemic, as well as their current employment situation. The latter variable was divided into whether a respondent was 3.2.1. Climate change concern and policy acceptance unemployed (which may have been the case even before COVID-19) or The following survey items were present in both surveys and are used on a temporary workforce reduction program (Spanish “ERTE”) initi­ as the two main dependent variables to achieve the three research aims ated due to COVID-19. One variable covers the perceived effect of of this study. First, an item measured climate change concern by asking climate change on the outbreak of COVID-19. Finally, we measure how “How concerned are you about climate change?” (5-point Likert scale people evaluate the government’s efforts to respond to the pandemic. with response options: “Not at all”, “a little”, “quite a lot”, “much”, “very Table 1 shows all these variables and their question wordings, coding, much”). descriptive statistics and how they relate to the individual-level hy­ Second, policy acceptance is conceptualized and measured here as potheses 2–4 explained in Section 2. the acceptance of a carbon tax. Specifically, we measured acceptance for In addition to the tests of hypotheses 2a/b related to the finite-pool- various types of the policy in order to understand the role of the tax of-worry through the survey-based measures of experiences and per­ revenues. The questions on carbon taxation were introduced by this ceptions, we examine the influence of objective experience of COVID- sentence: “Under the Paris Agreement from 2015, each country, 19. To this end, we use official data reported by the Spanish govern­ including Spain, must implement policies to reduce their CO2 emissions, ment on the number of reported new COVID-19 cases per day in every which contribute to climate change. One major proposal to achieve autonomous community. This data is then cumulated for the period until emissions reduction is by implementing a carbon tax on fossil fuels the launch of the survey (23rd of June), normalized by the number of whose combustion is the main cause of CO2 emissions”. The first ques­ inhabitants in each of the autonomous communities and linked to the tion asked “How acceptable do you find a carbon tax?” (5-point Likert survey respondents own reported location (see Fig. 2 visualizing the scale: “completely unacceptable”, “somewhat unacceptable”, “neither spread of COVID-19, and Fig. A1 in Appendix for further information). unacceptable nor acceptable”, “somewhat acceptable”, “completely As noted before, our survey respondents were invited using quotas for acceptable”), without specifying any use of the tax revenues (labelled autonomous communities. The combined data can be considered as a “unspecified” henceforth). In both surveys, respondents were then made rough proxy of objective exposure to contagion for the survey aware of the fact that carbon taxes generate revenues. Three follow-up respondents. questions were “How acceptable do you find the carbon tax if its reve­ To test hypotheses 5 and 6b, one explanatory variable needs to be nues were…” (all measured on the same 5-point Likert scale): constructed, namely change in climate concern. This is done by calcu­ lating the absolute difference between the post- and pre-values of (a) “used to compensate low-income households” (“PoorHH”), climate concern. Finally, our control variables are age, gender, educa­ (b) “used to support the development of climate projects” tion, household income, trust in politicians, and political orientation. (“Climate”), Their descriptive statistics can be found in Table A1 in the Appendix. (c) “returned in equal amounts to all households as compensation” (“AllHH”). 4. Results In addition, the 2020 survey had one new item that offered an 4.1. Understanding changes in climate concern and policy acceptance explicit COVID-19 option: We start with the analysis of changes in climate concern and policy (d) “used to pay for the extra public expenditures resulting from the acceptance. First, we briefly consider aggregate indicators derived from COVID-19 crisis” (“COVID-19”). the data, and then look at the individual-level effects hypothesized above. Fig. 3 shows the pre- and post-COVID-19 averages of concern Both climate concern and the policy acceptance variables are used to about climate change as well as acceptances of the four types of carbon test aggregate-level hypotheses 1 and 6a. For testing individual-level taxes (Unspecified, PoorHH, Climate, AllHH). The figure and an hypotheses 2–4 and operationalize the change in climate concern and accompanying pairwise Mann-Whitney test that explicitly compares policy acceptance in our analysis, we employ the so-called “residualized responses from the same individuals for stochastic dominance demon­ change approach” (Castro-Schilo and Grimm, 2018). This involves using strate that climate concern has decreased (M = 3.48 vs. M = 3.2, p < the post-COVID-19 values of climate concern or policy acceptance as 0.001, see Table A3). The corresponding effect size Cohen’s d = 0.255 dependent variables, and the associated pre-COVID-19 values as inde­ falls into the category of small (0.2) to medium (0.5). However, it is pendent variables in regression analyses. worth noting that a similar percentage of people indicated that they are One additional question asked respondents to freely allocate revenue at least “somewhat concerned” about climate change: 79% in 2019 uses (“What percentage of the total carbon tax revenues (100%) would versus 77% in 2020. Fig. 4 shows that the average reduction of climate you prefer to allocate for each of the 4 proposed options? Please make concern is mainly driven by the lower share of people who are “very sure that the total amount is equal to 100%.”). Since this question had much” concerned about climate change, suggesting that high concern only three revenue use options in the 2019 survey, perfect comparability has decreased. This finding supports our Hypothesis 1a. In contrast with is not possible. However, we use this question to complement our the latter, policy acceptance has statistically significantly increased for analysis of within-person changes involving the variables described three of the four types of carbon taxes, with one tax (unspecified reve­ before. nues) having a small-to-moderate effect size (d = 0.259), while the others have d values below 0.2 (Table A3). Only a carbon tax with revenues going to all households is not viewed significantly differently 4 S. Drews et al. Ecological Economics 199 (2022) 107507 Table 1 demographic variables. Table 2 reports results from an ordered logis­ Variables from the post-COVID-19 survey used to examine associations with tic regression using the residualized change approach mentioned in changes in climate concern and policy acceptance. Section 3.2.1. It regresses the post-COVID-19 level of climate concern Label of Question Coding of Mean Related (column 2) and policy acceptance (columns 3–6) on pre-COVID-19 variable wording response (standard hypothesis levels as well as on the explanatory and control variables mentioned options deviation) in the Methods section. Note that three variables (peer health issues, or % household income, political orientation) together have many missing Perceived How serious of 1 = not at all, 2 3.38 H2a/b data (n = 426). To check the robustness of our results with the larger COVID-19 a threat do you = little, 3 = (1.00) sample size, we exclude these variables from the analysis and report the threat think COVID-19 considerable, 4 is to you and = much, 5 = results in Tables A5–A8 in the Appendix. Overall, the results are very your family? very much similar across the two regression models. Personal Have you had 2 = yes, 0.68 (0.28) H2a/b Our results show that neither objective exposure to COVID-19, nor health serious health serious; 1 = self-reported personal, peer health issues, or psychological stress due to issues issues due to yes, mild; 0 = COVID-19? no COVID-19, is associated with changes in climate concern or policy Peer health How many of 0–60 1.07 H2a/b acceptance (Table 2). While the sign of the coefficient of these variables issues your family (2.66) is sometimes in the expected negative direction, they are not statistically members and significant. However, the perception that COVID-19 is a threat for re­ friends have spondents or their families is – against expectations – associated with had serious health issues increases in climate concern and acceptance of a carbon tax with un­ due to COVID- specified use of revenues (p < 0.01). Furthermore, economic factors such 19 (excluding as income and job loss do not show consistent effects either. However, it yourself)? is noteworthy that unemployed people have reduced their acceptance of Psychological To which extent 1 = none, 2 = 2.53 H2a/b a carbon tax with climate revenues (p < 0.05). Overall, there is no or stress have the little, 3 = quite (1.12) COVID-19 some, 4 = very little support for Hypotheses 2a and 2b. pandemic and much, 5 = very With respect to the additional factors, we find a fairly consistent the much effect for people’s belief that climate change has contributed to the confinement outbreak of COVID-19, increasing both climate concern (p < 0.001) and caused any psychological acceptance of almost all types of carbon taxes (p < 0.05). Thus, there is stress on you? support for Hypotheses 3a and 3b. Moreover, people’s evaluation of the Household How did the 1 = very 2.54 H2a/b performance of the government in tackling the COVID-19 crisis is income loss COVID-19 crisis negatively, 2 = (0.75) positively associated with increases in acceptance of a carbon tax with affect the net negatively, 3 = unspecified revenues. The latter gives partial support for our Hypothesis monthly no effect, 4 = income positively, 5 = 4 (the statistical significance of these results is even more frequent in the (including very positively full sample, as shown in Table A5 in the Appendix). salary and/or Despite general increases of policy acceptance and the decrease of rents) of your climate concern, these aggregate findings do not rule out the possibility household? ERTE What is your 1 = Worker 7.33% H2a/b that individual-level reductions in climate concern have also reduced current job affected by a policy acceptance, which would be an implication of the finite-pool-of- situation? temporary worry hypothesis. To examine this, we conduct a regression analysis of suspension of change in climate-policy acceptance on change in climate concern, in employment two steps. The first involves change in climate concern (operationalized (Spanish “ERTE”), as the absolute difference between the post- and pre-values) as the only 0 otherwise predictor next to pre-COVID-19 acceptance level. The second step in­ Unemployed What is your 1= 12.71% H2a/b volves controlling for other COVID-19-related variables and socio- current job Unemployed, demographic characteristics. Results are shown in Table 3. We do not situation? 0 otherwise Perceived To what extent 1 = Totally 2.66 H3a/b find that decreases in climate concern are significantly associated with effect of do you agree agree, 2 = (1.26) decreases in acceptance of a carbon tax, neither in the simple model nor climate with the partially agree, the model with all variables. Hence, we reject Hypothesis 5. following 3 = neutral, 4 statement? = partially “Climate disagree, 5 = 4.2. The role of co-benefits from climate policy for addressing the COVID- change has totally disagree 19 crisis contributed to the COVID-19 Next, we examine to what extent policy acceptance is affected by outbreak” climate policy being presented as having a co-benefit in terms of Assessment of Overall, how 1 = very 2.70 H4 government would you rate negatively, 2 = (1.28) responding to the COVID-19 crisis. For this purpose, we first consider the negatively, 3 = acceptance of a carbon tax whose revenues are used to compensate the performance of neutral, 4 = public expenditures resulting from COVID-19 and compare this policy the government positively, 5 = with the four other revenue uses (Fig. A2 in Appendix). This analysis to deal with very positively COVID-19? draws only from the post-COVID-19 survey data. Results show that using revenues for COVID-19 has a slightly higher public acceptance than when the revenues remain unspecified (M = 3.39 vs. M = 3.31, p < 0.05 after COVID-19. All this suggests that the relationship between COVID- in pairwise Mann-Whitney test, Cohen’s d = 0.062, see Table A4), or 19, climate change concern and policy acceptance is more complex than when the revenues are transferred to all households (M = 3.23, p < 0.01, one would expect. We thus reject Hypothesis 1b. d = 0.129). The former results lend support to Hypothesis 6a. An To corroborate these and obtain additional insights, it is necessary to additional finding is that there is considerably more acceptance of a examine individual-level effects while controlling for socio- carbon tax that uses revenues to finance climate projects than of a 5 S. Drews et al. Ecological Economics 199 (2022) 107507 Fig. 2. Spread of COVID-19 cases per 10′ 000 inhabitants in Spain. Fig. 3. Changes of climate concern and acceptance of carbon taxes before and after first COVID-19 wave. Note: Results include averages and error bars depicting +/− 2 s.e. carbon tax that financially supports COVID-19 expenditures (M = 4.04, choice was between only three options (PoorHH, AllHH, Climate), more p < 0.01, d = 0.552). Moreover, expenditures related to COVID-19 have than half of the sample preferred a combination of the three revenue a similar acceptance as distributing them among poor households (M = uses, with a somewhat greater share of revenues allocated to climate 3.33). In other words, the carbon tax with COVID-19 revenue use per­ projects than to the other two uses (Maestre-Andrés et al., 2021). Post- forms relatively well compared to other uses of the carbon tax revenues. COVID-19, when given the additional option to allocate money to the In addition, respondents were asked how they would allocate tax COVID-19 revenue use, again over 50% of respondents prefer a roughly revenues to different uses if they had complete freedom to decide about equal distribution among all available uses. Taken together, this shows this. The results are summarized in Fig. 5. Before COVID-19, when the that the COVID-19 revenue use is received as a valid alternative to other 6 S. Drews et al. Ecological Economics 199 (2022) 107507 Fig. 4. Comparison of climate-change concern and acceptance of carbon tax under different use of revenues before (2019) and after COVID-19 (2020). 7 S. Drews et al. Ecological Economics 199 (2022) 107507 Table 2 The effect of COVID-19 on changes in climate concern and policy acceptance. Independent variables Post-COVID-19 climate concern Post-COVID-19 carbon tax acceptance, by revenue use Unspecified Poor HH Climate All HH Pre-COVID-19 concern or acceptance 3.69*** 2.01*** 2.04*** 2.13*** 1.81*** Objective COVID-19 exposure 1.00 1.00 1.00 1.00 1.00 Perceived COVID-19 threat 1.43*** 1.26** 1.09 1.02 0.92 Personal health issues 1.26 0.96 0.99 0.72 1.23 Peer health issues 0.99 0.98 1.01 1.03 0.97 Psychological stress 1.02 0.90 0.97 1.07 1.00 Income loss 1.11 1.03 0.91 1.07 1.12 Unemployed 1.06 0.75 1.29 0.56* 0.88 ERTE 0.58 0.64 0.96 0.77 1.24 Perceived effect of climate 1.32*** 1.21** 1.14* 1.19** 1.11 Assessment of government 1.09 1.23** 1.16 1.13 1.08 Control variables Age 1.00 1.00 1.01 1.01 0.99 Gender 0.90 0.88 1.17 1.11 1.10 Education 0.95 1.12 1.09 1.14* 0.91 Household income 0.96 1.12 0.97 0.90 0.96 Political orientation 0.91* 0.87*** 0.82*** 0.88** 0.97 Trust in politicians 1.07 1.38* 0.96 1.05 0.63*** Pseudo R2 0.49 0.40 0.29 0.27 0.18 Note: Ordered logistic regression. Results indicate odds ratios. *p < 0.05; **p < 0.01; ***p < 0.001; N = 746. Nagelkerke Pseudo R2 was used. revenue uses. 5. Discussion As explained above, reduced climate concern is expected to lead to reduced policy acceptance at the individual-level. However, such an 5.1. Individual-level findings effect might be attenuated by a policy that accounts explicitly for COVID-19 concerns, such as the carbon tax with COVID-19 revenue use. The main objective of this paper was to examine how several To test this hypothesis (6b), we examine whether changes in climate individual-level factors are linked to changes in climate concern and concern are associated with the difference between acceptance of a policy acceptance. The first main insight is that we do not find clear COVID-19 carbon tax and the tax with unspecified revenues. Specif­ evidence that reductions of climate concern are associated with various ically, we test a regression equation of the following form: health and economic effects of COVID-19. Negative experiences of the / Carbon tax supportCOVID Carbon tax supportunspecified = αChange in climate concern + βX + ε (1) respondent or its family members and friends had no effect whatsoever on changes in climate concern and policy acceptance either. Thus, we expect that the ratio of acceptances of a carbon tax on LHS That said, we also found that the perception of COVID-19 as a threat of Eq. (1) will be higher for those people who have reduced their climate to oneself or one’s family is associated with higher climate concern as concern, i.e. coefficient α is expected to be negative and significant. well as higher acceptance of a carbon tax with unspecified revenues. Results indicate that while the coefficient of change in climate concern is These findings contrast with the finite-pool-of-worry hypothesis. They negative, it is not statistically significant (Table 4). suggest and confirm what has been found in another (unpublished) As a last step, we seek to provide complementary insights about the study by Sisco et al. (2020), namely that threat perception about one role of the COVID-19 revenue use in comparison with the other carbon issue may spill over on another issue. The difference in effects of threat taxes. To this end, we run a set of regressions of the post-COVID-19 perception versus other COVID-19 health variables may be explained by levels of tax acceptance, allowing us to understand how COVID-19 ex­ the former variable being more forward-looking, and therefore con­ periences and perceptions are linked to this policy option. Results in nected with thoughts on policy solutions, whereas the latter examined Table 5 show that higher COVID-19 threat perception is associated with past experiences. acceptance of the carbon tax under the COVID-19 revenue use (p < Economic effects due to COVID-19 show little consistency. However, 0.001), but has no significant effect on the acceptance of any other we found that unemployed people showed less acceptance of a carbon carbon taxes. There are three notable findings regarding economic ef­ tax with climate revenues. This may be viewed at least as partial evi­ fects: Unemployment and income loss – somewhat surprisingly – are dence for economic insecurity leading to less support for some climate associated with lower acceptance of the COVID-19 tax (both p < 0.05). policies, which in turn may be considered as indirect evidence for the Unemployment is also linked to lower acceptance of a carbon tax with finite-pool-of-worry. Overall, though, our findings are consistent with climate revenues (p < 0.01). A positive assessment of the government’s research using a similar within-person research design (Mildenberger effort to address the crisis is associated with higher acceptance of carbon and Leiserowitz, 2017), suggesting that the so-called economy-envi­ taxes with unspecified and COVID-19 revenues (p < 0.01 and p < ronment tradeoff for attitudes may have been overemphasized by other 0.001), while the perceived effect of COVID-19 on climate is linked to research relying on cross-sectional or aggregate-level analyses (e.g., higher acceptance of all considered carbon tax revenue uses. Scruggs and Benegal, 2012). In other words, economic worries are not a major factor associated with reduced climate concern or policy accep­ tance. It is worth acknowledging, however, that here we focused on 8 S. Drews et al. Ecological Economics 199 (2022) 107507 personal economic situations and did not cover people’s perceptions of Berlin Rubin, 2022). However, causality of the relationship might well the current and future general situation of the economy. For example, a be reverse, that is, people become more worried about climate change study for Germany shows that a perceived deterioration of the general over time and attribute other kinds of problems to it. Further experi­ economic situation due to the COVID-19 crisis has a significantly mental research could try to substantiate these findings. negative effect on the acceptance of climate-oriented economic stimulus programs (Engler et al., 2021). This may be consistent with the fact that 5.2. Aggregate-level findings and synthesis we found lower acceptance of the tax with COVID-19 revenue use among people who lost income and jobs. Our finding that aggregate climate concern has decreased can be Another insight is that changes in climate concern are unrelated to viewed as evidence for the finite-pool-of-worry hypothesis, despite its changes in policy acceptance. This finding contrasts the strong evidence rather small effect size. It also appears to contrast with results of a study from cross-sectional studies (e.g. Bouman et al., 2020b) and therefore for the UK which showed that climate beliefs are stable before and after explains why some authors have called for distinguishing static and COVID-19 and that, if anything, there is a “finite-pool-of-attention” dynamic relationships of climate perceptions (Capstick et al., 2015; (Evensen et al., 2021). The latter study analyzed Twitter data, finding Mildenberger and Leiserowitz, 2017). It could be, for example, that that people have communicated less about climate change after COVID- policy acceptance is not responsive to small reductions in climate 19. This suggests that conceptualizations and operationalizations mat­ concern taking place at the upper end, that is, when declining from very ter, which also becomes evident when considering our second main high to still high levels of concern. It may also have to do with our finding, namely that aggregate policy acceptance of carbon taxation measurement of a more generalized climate concern, which according to with distinct revenue uses tends to increase over time. the literature tends to have not as strong effects as more personal worry It is evident that climate concern and policy acceptance are different about, or perceived seriousness of, climate change (Van Der Linden, concepts. Acceptance may increase for reasons unrelated to concern. 2017). One such reason is that policies may (be perceived to) have co-benefits. A few comments on the two additional variables of interest are in We showed that on average a carbon tax with a revenue explicitly order. A positive assessment of the government’s effort to address the framed to pay for COVID-19 expenditures is equally or slightly more crisis is associated with higher acceptance of carbon taxes with un­ accepted than other revenue uses except climate projects. In addition, specified revenues, but not for the other explicit revenue options results indicate that this revenue is more positively viewed by people (PoorHH, AllHH, Climate). This could explain the relatively high in­ who perceive COVID-19 as threat, but more negatively by people who crease of acceptance of this type of carbon tax (revenue use) at the have lost income. One interpretation of the latter is that respondents aggregate level. In other words, people may have observed that the may have understood this revenue use not so much as an economic but a government is doing a good job in tackling the health crisis and may health response to COVID-19. The wording of the survey question was then find it more acceptable to introduce a carbon tax even if the rev­ somewhat ambiguous in this respect. In any case, given that these results enues are not clearly earmarked. This may be evidence for what has are cross-sectional, we can only speculate that any co-benefits have been termed a positive “policy spillover” (Thomas et al., 2019). contributed to aggregate-level increases of policy acceptance. In line with our expectations was that the belief that “climate change It is difficult to answer why aggregate climate concern has somewhat has contributed to COVID-19” is significantly associated with higher decreased, although we do not find this decrease to be linked to the climate concern and most cases of policy acceptance. This finding is individual-level variables examined here. It is of course possible that consistent with other recent, cross-sectional evidence on people’s sub­ changes are the result of influences unrelated with COVID-19. For jective attribution of climate change to COVID-19 for countries, namely example, data for the first survey in 2019 was collected at the end of July New Zealand (Thaker and Cook, 2021) and the US (Wong-Parodi and just a few days after a heat wave in Spain (see Fig. A1 in the Appendix). Table 3 The effect of change in climate concern on change in policy acceptance. Independent variables Post-COVID-19 carbon tax acceptance, by revenue use Unspecified Poor HH Climate All HH M1 M2 M1 M2 M1 M2 M1 M2 Pre-COVID-19 acceptance 2.12*** 2.02*** 2.21*** 2.04*** 2.25*** 2.13*** 1.84*** 1.81*** Change in climate concern 1.06 1.04 0.97 1.08 0.98 1.00 1.01 0.99 Objective COVID-19 exposure 1.26 1.09 1.02 0.92 Perceived COVID-19 threat 0.95** 0.98 0.72 1.23 Personal health issues 0.98 1.01 1.03 0.97 Peer health issues 0.90 0.97 1.07 1.00 Psychological stress 1.03 0.91 1.07 1.12 Income loss 0.75 1.28 0.56 0.88 Unemployed 0.64 0.96 0.77* 1.24 ERTE 1.21 1.14 1.19 1.11 Perceived effect of climate 2.02** 2.04* 2.13** 1.81 Assessment of government 1.23** 1.16 1.13 1.08 Control variables Age 1.00 1.01 1.01 0.99 Gender 0.88 1.17 1.11 1.10 Education 1.12* 1.10 1.14* 0.91 Household income 1.12 0.97 0.90 0.96 Political orientation 0.87*** 0.82*** 0.88** 0.97 Trust in politicians 1.38* 0.96 1.05 0.63*** Pseudo R2 0.24 0.40 0.20 0.34 0.18 0.41 0.13 0.18 Note: Ordered logistic regression. Results indicate odds ratios. *p < 0.05; **p < 0.01; ***p < 0.001; N = 1172 in Model 1 (M1), N = 746 in Model 2 (M2). Nagelkerke’s Pseudo R2 was used. 9 S. Drews et al. Ecological Economics 199 (2022) 107507 Fig. 5. Preferred allocation of revenues from a carbon tax. Respondents were asked to express the share (in %) of total revenues to be allocated to each revenue use. The left panel shows results for 2019, the right panel for 2020 with an additional COVID-19 option. Research suggests that such experiences can (temporarily) increase have influenced especially the policy-related variables. Future research climate change concern (Howe et al., 2019). This could mean that our will have to answer these questions. baseline measure was rather high. The available data of respondents’ locations does not allow for proper controls in this respect. In contrast, the overall increase of climate policy acceptance may be the result of 5.3. Limitations increasing attention about climate change due to the UN climate talks in Spain in December 2019 or the Friday for Future movement. All this Some limitations of our study are as follows. The first is that it can points to a more general consideration for the interpretation of our only draw on single items to measure changes of the key constructs. That findings, namely the issue of timing. The post-COVID-19 data was is, single-item measures, such as for climate concern and policy accep­ collected after the end of the first wave. Results may differ with a larger tance, might be somewhat unreliable compared to composite indices number of people being affected health-wise and/or economically throughout subsequent waves, while the emergence of vaccinations may Table 5 Acceptance of a carbon tax with revenue use for COVID-19, compared with other types of revenue use. Table 4 Independent Carbon tax acceptance, by revenue use Analysis of the additional acceptance of a carbon tax due to a COVID-19 revenue variables COVID- Unspecified Poor Climate All HH use. 19 HH Additional carbon tax acceptance due to COVID-19 Climate concern 1.01 1.94*** 1.08 2.20*** 0.90 revenue Objective COVID-19 1.00 1.00 1.00 1.00* 1.00 (Intercept) 1.20*** 1.77*** exposure Perceived COVID-19 Change in climate concern 0.01 − 0.06 1.40*** 1.11 1.07 0.91 0.91 threat Pre-COVID-19 acceptance − 0.14*** Personal health Objective COVID-19 exposure 0.00 1.00 0.93 0.99 0.69 1.35 issues Perceived COVID-19 threat 0.03 Peer health issues 1.00 1.00 1.02 1.04 0.97 Personal health issues 0.10 Psychological stress 0.94 0.88 0.95 1.05 1.02 Peer health issues 0.00 Income loss 0.82* 0.88 0.92 0.95 1.15 Psychological stress − 0.01 Unemployed 0.66* 0.70 1.29 0.50** 0.80 Income loss − 0.03 ERTE 0.86 0.94 0.90 0.97 1.20 Unemployed − 0.03 Perceived effect of ERTE 0.13 1.13** 1.16** 1.15* 1.13* 1.14* climate Perceived effect of climate 0.00 Assessment of Assessment of government − 0.00 1.34*** 1.28** 1.14 1.12 1.06 government Control variables Control variables Age 0.00 Age 1.01 1.00 1.00 1.01 0.99 Gender 0.15* Gender 1.50** 0.93 1.21 1.19 1.25 Education − 0.05 Education 0.96 1.22*** 1.03 1.13* 0.86** Household income − 0.03 Household income 0.96 1.08 0.91 0.93 0.92 Political orientation 0.01 Political orientation 0.92* 0.89** 0.80*** 0.88** 0.96 Trust in politicians − 0.09 Trust in politicians 0.97 1.67*** 1.00 1.13 0.66*** Adj. R2 − 0.00 0.08 Pseudo R2 0.14 0.16 0.16 0.26 0.07 Note: Results indicate OLS regression coefficients. *p < 0.05; **p < 0.01; ***p Note: Ordered logistic regression. Coefficients indicate odds ratios. *p < 0.05; < 0.001; N = 1172 in Model 1 (M1), N = 746 in Model 2 (M2). **p < 0.01; ***p < 0.001; N = 746. Nagelkerke’s Pseudo R2 was used. 10 S. Drews et al. Ecological Economics 199 (2022) 107507 (Kyselá et al., 2019). For example, prior research on within-person we find that citizens’ positive assessment of the government’s action to changes of climate beliefs and attitudes has shown that survey address COVID-19 is associated with higher acceptance of some climate response errors, such as paying insufficient attention to questions, are policies. In the concrete case of a carbon tax, our study suggests that related to variability in responses (Jenkins-Smith et al., 2020). While we policy makers can consider allocating at least some of the tax revenues examined effects of survey completion time and did not find any sig­ for COVID-19 purposes, as citizens view this option favorably, next to nificant relation with the magnitude of change in climate concern or other conventional revenue uses. The results of our case study for Spain policy support (results are available upon request), the existence of other – a country that was hit relatively hard by the pandemic – offers an such effects cannot entirely be ruled out. Second, to test the finite-pool- optimistic conclusion for policy: even in times of a global health crisis, of-worry hypothesis more comprehensively, it would have been ideal to public support for addressing the global climate crisis remains fairly not only measure changes in climate concern, but also changes in non- stable. climate concerns. This was of course not possible, given that our base­ line survey was not designed with such a research question in mind. Third, although our focus here was on carbon taxes with a broad set of Declaration of Competing Interest social and climate revenue uses, carbon pricing is generally one of the more contested climate policies (Rhodes et al., 2017). Future research The authors declare that they have no known competing financial could investigate changes in other, less polarizing, types of climate interests or personal relationships that could have appeared to influence policies. the work reported in this paper. 6. Conclusions Acknowledgements This study examined how COVID-19 has affected concern about This work was funded by an ERC Advanced Grant from the European climate change and acceptance of climate policy in Spain. Using data for Research Council (ERC) under the European Union’s Horizon 2020 the same survey participants before and after COVID-19, we find that the Research and Innovation Programme [grant agreement n◦ 741087]. pandemic has somewhat reduced aggregate-level concern about climate Ivan Savin acknowledges financial support from the Russian Science change. In contrast, public acceptance of a carbon tax with different Foundation [RSF grant number 19-18-00262]. Sergio Villamayor- revenue uses has slightly increased over time. At the individual level, Tomas’ work was supported by a Marie Curie IF grant (nr. 660089), however, we do not find reductions in climate concern to be related to Ramon y Cajal Fellowship (RyC-2017-22782), financial support from health or economic effects of COVID-19. This means limited evidence for the Federal University of Minas Gerais´ Visiting Professor program a finite-pool-of-worry, at least when measured through experience- (Contrato n◦ 253/2020), and the financial support from the Spanish based proxies. It rather appears that the perception of one threat Ministry of Science, Innovation and Universities, through the “María de (climate change) spills over onto another threat (COVID-19). Similarly, Maeztu” programme for Units of Excellence (CEX2019-000940-MEU). Appendix A Table A1 Comparison of sample (N = 1172) with census data. Variables Description Mean (SD) or % Spanish population Gender Dummy: female 47.44% 50.99% Age 18 to 88 years old 49.30 (14.10) 43.59 Monthly household 1 (less than 1000€) to 5 (More than 2.64 (1.14) translating into a value between ± €2600 2385 euros per month income 4000€) Education 1 (Less than 5 years of school) to 8 5.05 (1.30); 89.92% of the sample have medium professional or 63.8% have a medium professional or (University) higher studies higher studies* Political orientation 1 (left-wing) to 10 (right-wing) 4.54 (2.44) 4.6 (2.0)* Notes: Sampling was done by using quotas on age, gender and geographical distribution, making the survey sample representative of the general population on these characteristics. The rest of variables are compared with census data from the Spanish National Institute of Statistics (www.ine.es) unless another source is indicated. * Source: Centro de Investigaciones Sociológicas (CIS), Barómetro de Julio 2020. Table A2 Results of the Kolmogorov-Smirnov (K-S) and Wilcoxon rank sum tests for differences in samples. 11 S. Drews et al. Ecological Economics 199 (2022) 107507 Mean of Mean p- Cohe Mean of Mean of p- Cohe repeated of all n’s d repeated p- n’s d p-value value all 2200 value 1172 2004 (effec 1172 value (effect Variable (K-S (Wilc responde (Wilco respon- respon t size) responde (K-S size) test) o-xon nts in -xon dents in -dents nts in test) test) 2020 test) 2019 in 2019 2020 Gender 1.47 1.51 0.280 0.047 0.073 1.47 1.51 0.288 0.049 0.071 (female=1) Age 48.45 45.15 <0.001 <0.001 0.223 49.30 45.50 <0.001 <0.001 0.258 Region 8.30 8.32 0.999 0.950 0.006 8.27 8.29 0.999 0.990 0.003 Education 5.09 5.02 0.408 0.089 0.060 5.05 5.02 0.977 0.411 0.029 Climate 3.48 3.51 0.992 0.376 0.032 3.20 3.20 0.999 0.845 0.004 concern Acceptance of carbon tax 2.97 3.00 0.999 0.647 0.018 3.31 3.34 0.983 0.648 0.019 (unspecified use) Personal health issues - - - - - 2.93 2.92 0.999 0.348 0.033 COVID-19 Employ- - - - - - 2.53 2.44 0.228 0.346 0.047 ment status Household 4.48 4.44 0.986 0.377 0.028 2.64 2.58 0.967 0.210 0.050 income Political 4.49 4.47 0.999 0.792 0.006 4.54 4.44 0.917 0.278 0.041 orientation Trust in 1.62 1.64 0.999 0.548 0.025 1.71 1.70 0.999 0.822 0.011 politicians Note: cases where the differences are found to be significantly different are marked with grey shading. Table A3 Comparing results for climate concern and policy acceptance before (2019) and after COVID-19 (2020). Type of sample 2019 2020 Pairwise Mann-Whitney test’s p value Cohen’s d (effect size) Climate concern 3.475 3.201 <0.0001 0.255 (3.409–3.542) (3.143–3.259) Carbon tax acceptance (Unspecified revenue use) 2.973 3.314 <0.0001 0.259 (2.894–3.051) (3.239–3.389) Carbon tax acceptance (PoorHH) 3.187 3.332 <0.0001 0.123 (3.121–3.253) (3.261–3.403) Carbon tax acceptance (Climate) 3.887 4.043 <0.0001 0.152 (3.826–3.946) (3.983–4.103) Carbon tax acceptance (AllHH) 3.230 3.227 0.971 0.003 (3.164–3.296) (3.152–3.302) Note: Error bars with ±2 s.e. are reported in parentheses. P-values below 0.05 indicate significant differences between data samples between data samples. Table A4 Comparing carbon tax acceptance under different revenue uses in 2020. Type of non-COVID revenue use Mean acceptance of revenue use Mean of COVID-19 revenue Mann-Whitney test’s p value Cohen’s d (effect size) Unspecfied revenue use 3.314 3.393 0.0427 0.062 (3.239–3.389) (3.318–3.469) PoorHH 3.332 3.393 0.054 0.049 (3.261–3.403) (3.318–3.469) Climate 4.043 3.393 <0.0001 0.552 (3.983–4.103) (3.318–3.469) AllHH 3.227 3.393 0.0001 0.129 (3.152–3.302) (3.318–3.469) Note: Error bars with ±2 s.e. are reported in parentheses. P-values below 0.05 indicate significant differences between data samples. 12 S. Drews et al. Ecological Economics 199 (2022) 107507 Table A5 The effect of COVID-19 on changes in climate concern and policy acceptance (variables with many missing data excluded). Independent variables Post-COVID-19 climate concern Post-COVID-19 carbon tax acceptance, by revenue use Unspecified Poor HH Climate All HH Pre-COVID-19 concern/acceptance 3.77*** 1.85*** 2.07 2.03*** 1.81*** Objective COVID-19 exposure 1.00 1.00* 1.00 1.00 1.00 Perceived COVID-19 threat 1.47*** 1.18** 1.14* 1.12 0.97 Personal health issues 1.26 1.03 0.97 0.76 0.94 Psychological stress 0.98 0.86* 0.96 0.99 1.09 Income loss 1.12 0.99 0.96 1.18* 1.12 Unemployed 1.06 0.88 1.29 0.63** 0.98 ERTE 0.73 0.77 1.14 0.68 1.27 Perceived effect of climate 1.32*** 1.24*** 1.11* 1.14** 1.06 Assessment of government 1.21*** 1.46*** 1.36*** 1.23*** 1.05 Control variables Age 1.01 1.00 1.00 1.00 1.00 Gender 1.04 0.84 1.12 0.91 1.13 Education 0.91* 1.02 1.01 1.05 0.90* Trust in politicians 1.00 1.25* 0.96 1.15 0.73** Pseudo R2 0.49 0.34 0.24 0.21 0.16 Note: Ordered logistic regression. Coefficients indicate odds ratios. *p < 0.05; **p < 0.01; ***p < 0.001; N = 1155. Table A6 The effect of change in climate concern on change in policy acceptance (variables with many missing data excluded). Independent variables Post-COVID-19 carbon tax acceptance, by revenue use Unspecified Poor HH Climate All HH Pre-COVID-19 acceptance 1.85*** 2.06*** 2.02*** 1.81*** Change in climate concern 1.03 0.99 0.97 1.01 Objective COVID-19 exposure 1.00* 1.00 1.00 1.00 Perceived COVID-19 threat 1.18** 1.14* 1.12 0.97 Personal health issues 1.02 0.97 0.76 0.93 Psychological stress 0.86** 0.96 0.99 1.09 Income loss 0.99 0.96 1.18* 1.12 Unemployed 0.88 1.29 0.63** 0.98 ERTE 0.77 1.14 0.68 1.26 Perceived effect of climate 1.24*** 1.11* 1.14** 1.06 Assessment of government 1.46*** 1.36*** 1.23*** 1.05 Control variables Age 1.00 1.00 1.00 1.00 Gender 0.84 1.12 0.91 1.13 Education 1.03 1.01 1.05 0.90* Trust in politicians 1.25* 0.96 1.15 0.73*** Pseudo R2 0.34 0.24 0.21 0.16 Note: Ordered logistic regression. Coefficients indicate odds ratios. *p < 0.05; **p < 0.01; ***p < 0.001; N = 1155. Table A7 OLS regression of the additional acceptance of a carbon tax due to a COVID-19 revenue use (variables with many missing data excluded). Additional carbon tax acceptance due to COVID-19 revenue (Intercept) 1.53*** Change in climate concern − 0.13 Pre-COVID-19 acceptance − 0.07*** Objective COVID-19 exposure 0.00 Perceived COVID-19 threat 0.05* Personal health issues 0.00 Psychological stress 0.02 Income loss 0.01 Unemployed − 0.07 ERTE 0.10 Perceived effect of climate − 0.02 Assessment of government − 0.04 Control variables Age 0.00 Gender 0.12* (continued on next page) 13 S. Drews et al. Ecological Economics 199 (2022) 107507 Table A7 (continued ) Additional carbon tax acceptance due to COVID-19 revenue Education − 0.03 Trust in politicians − 0.02 Adj. R2 0.07 Note: *p < 0.05; **p < 0.01; ***p < 0.001; N = 1155. Table A8 Acceptance of carbon tax with COVID-19 revenue use compared to other types of carbon taxation (variables with many missing data excluded). Independent variables Carbon tax acceptance, by revenue use COVID-19 Unspecified Poor HH Climate All HH Climate concern 1.03 1.96*** 1.12 2.20*** 0.93 Objective COVID-19 exposure 1.00 1.00** 1.00 1.00 1.00 Perceived COVID-19 threat 1.34*** 1.01 1.09 0.98 0.96 Personal health issues 0.88 1.03 0.97 0.70 1.01 Psychological stress 0.96 0.84** 0.98 0.98 1.13* Income loss 0.91 0.90 0.96 1.09 1.18* Unemployed 0.82 0.89 1.39 0.61** 0.93 ERTE 0.94 0.94 1.20 0.85 1.30 Perceived effect of climate 1.16*** 1.24*** 1.12* 1.08 1.07 Assessment of government 1.38*** 1.46*** 1.43*** 1.22*** 1.06 Control variables Age 1.00 1.00 1.00 1.00 0.99 Gender 1.30* 0.89 1.18 0.96 1.27* Education 0.94 1.10* 0.95 1.07 0.85*** Trust in politicians 1.08 1.50*** 1.02 1.25* 0.75** Pseudo R2 0.11 0.30 0.10 0.22 0.06 Note: Ordered logistic regression. Coefficients indicate odds ratios. *p < 0.05; **p < 0.01; ***p < 0.001; N = 1155. Nagelkerke Pseudo R2 was used. Fig. A1. Evolution of contagion and death cases in Spain (Confinement: 14 March - 21 June; survey: 23 June - 3 July). Source: Spanish National Epidemiology Center https://cnecovid.isciii.es/covid19/#documentaci%C3%B3n-y-datos (contagion cases), and Europa Press, www. epdata.es (death cases) 14 S. Drews et al. 15 Fig. A2. Comparison of carbon tax acceptance under different revenue uses after COVID-19. Ecological Economics 199 (2022) 107507 S. Drews et al. Ecological Economics 199 (2022) 107507 References Kitt, S., Axsen, J., Long, Z., Rhodes, E., 2021. The role of trust in citizen acceptance of climate policy: comparing perceptions of government competence, integrity and value similarity. Ecol. Econ. 183, 106958 https://doi.org/10.1016/j. Akerlof, K., DeBono, R., Berry, P., Leiserowitz, A., Roser-Renouf, C., Clarke, K.-L., ecolecon.2021.106958. Rogaeva, A., Nisbet, M.C., Weathers, M.R., Maibach, E.W., 2010. Public perceptions Kyselá, E., Ščasný, M., Zvěřinová, I., 2019. Attitudes toward climate change mitigation of climate change as a human health risk: surveys of the United States, Canada and policies: a review of measures and a construct of policy attitudes. Clim. Pol. 19, Malta. Int. J. Environ. Res. Public Health 7, 2559–2606. https://doi.org/10.3390/ 878–892. https://doi.org/10.1080/14693062.2019.1611534. ijerph7062559. Maestre-Andrés, S., Drews, S., van den Bergh, J., 2019. Perceived fairness and public Amelung, D., Fischer, H., Herrmann, A., Aall, C., Louis, V.R., Becher, H., Wilkinson, P., acceptability of carbon pricing: a review of the literature. Clim. Pol. 19, 1186–1204. Sauerborn, R., 2019. Human health as a motivator for climate change mitigation: https://doi.org/10.1080/14693062.2019.1639490. results from four European high-income countries. Glob. Environ. Chang. 57, Maestre-Andrés, S., Drews, S., Savin, I., van den Bergh, J., 2021. Carbon tax acceptability 101918 https://doi.org/10.1016/j.gloenvcha.2019.05.002. with information provision and mixed revenue uses. Nat. Commun. https://doi.org/ Bain, P.G., Hornsey, M.J., Bongiorno, R., Jeffries, C., 2012. Promoting pro-environmental 10.1038/s41467-021-27380-8. action in climate change deniers. Nat. Clim. Chang. 2, 600–603. https://doi.org/ Markard, J., Rosenbloom, D., 2020. A tale of two crises: COVID-19 and climate. 10.1038/nclimate1532. Sustainability: science. Pract. Policy 16, 53–60. https://doi.org/10.1080/ Bakaki, Z., Bernauer, T., 2018. Do economic conditions affect public support for 15487733.2020.1765679. environmental policy? J. Clean. Prod. 195, 66–78. https://doi.org/10.1016/j. Mildenberger, M., Leiserowitz, A., 2017. Public opinion on climate change: is there an jclepro.2018.05.162. economy–environment tradeoff? Environ. Politics 26, 801–824. https://doi.org/ Beiser-McGrath, L.F., Bernauer, T., 2019. Could revenue recycling make effective carbon 10.1080/09644016.2017.1322275. taxation politically feasible? Sci. Adv. 5, eaax3323. https://doi.org/10.1126/sciadv. Mintz-Woo, K., Dennig, F., Liu, H., Schinko, T., 2020. Carbon pricing and COVID-19. aax3323. Clim. Pol. 0, 1–9. https://doi.org/10.1080/14693062.2020.1831432. Botzen, W., Duijndam, S., van Beukering, P., 2020. Lessons for climate policy from Mukanjari, S., Sterner, T., 2020. Charting a “green path” for recovery from COVID-19. behavioral biases towards COVID-19 and climate change risks. World Dev. 105214 Environ. Resour. Econ. 76, 825–853. https://doi.org/10.1007/s10640-020-00479-0. https://doi.org/10.1016/j.worlddev.2020.105214. Myers, T.A., Nisbet, M.C., Maibach, E.W., Leiserowitz, A.A., 2012. A public health frame Bouman, T., Steg, L., Dietz, T., 2020a. Insights from early COVID-19 responses about arouses hopeful emotions about climate change. Clim. Chang. 113, 1105–1112. promoting sustainable action. Nat. Sustain. 1–7 https://doi.org/10.1038/s41893- https://doi.org/10.1007/s10584-012-0513-6. 020-00626-x. Reese, G., Hamann, K.R.S., Heidbreder, L.M., Loy, L.S., Menzel, C., Neubert, S., Bouman, T., Verschoor, M., Albers, C.J., Böhm, G., Fisher, S.D., Poortinga, W., Tröger, J., Wullenkord, M.C., 2020. SARS-Cov-2 and environmental protection: a Whitmarsh, L., Steg, L., 2020b. When worry about climate change leads to climate collective psychology agenda for environmental psychology research. J. Environ. action: how values, worry and personal responsibility relate to various climate Psychol. 70, 101444 https://doi.org/10.1016/j.jenvp.2020.101444. actions. Glob. Environ. Chang. 62, 102061 https://doi.org/10.1016/j. Rhodes, E., Axsen, J., Jaccard, M., 2017. Exploring citizen support for different types of gloenvcha.2020.102061. climate policy. Ecol. Econ. 137, 56–69. https://doi.org/10.1016/j. Burke, J., Fankhauser, S., Bowen, A., 2020. Pricing Carbon during the Economic ecolecon.2017.02.027. Recovery from the COVID-19 Pandemic. Policy Brief. Savin, I., Drews, S., Maestre-Andrés, S., van den Bergh, J., 2020. Public views on carbon Capstick, S., Whitmarsh, L., Poortinga, W., Pidgeon, N., Upham, P., 2015. International taxation and its fairness: a computational-linguistics analysis. Clim. Chang. 162, trends in public perceptions of climate change over the past quarter century: 2107–2138. https://doi.org/10.1007/s10584-020-02842-y. international trends in public perceptions of climate change. Wiley Interdiscip. Rev. Scruggs, L., Benegal, S., 2012. Declining public concern about climate change: can we Clim. Chang. 6, 35–61. https://doi.org/10.1002/wcc.321. blame the great recession? Glob. Environ. Chang. 22, 505–515. https://doi.org/ Carattini, S., Carvalho, M., Fankhauser, S., 2018. Overcoming public resistance to carbon 10.1016/j.gloenvcha.2012.01.002. taxes. Wiley Interdiscip. Rev. Clim. Chang., e531 https://doi.org/10.1002/wcc.531. Sisco, M., Constantino, S., Gao, Y., Tavoni, M., Cooperman, A., Bosetti, V., Weber, E., Castro-Schilo, L., Grimm, K.J., 2018. Using residualized change versus difference scores 2020. A finite Pool of worry or a finite Pool of attention? Evidence and for longitudinal research. J. Soc. Pers. Relat. 35, 32–58. https://doi.org/10.1177/ qualifications. Working Paper. 0265407517718387. Smith, K.R., Woodward, A., Campbell-Lendrum, D., Chadee, D.D., Honda, Y., Liu, Q., Drews, S., van den Bergh, J.C.J.M., 2016. What explains public support for climate Olwoch, J.M., Revich, B., Sauerborn, R., 2014. Human health: impacts, adaptation, policies? A review of empirical and experimental studies. Clim. Pol. 16, 855–876. and co-benefits. In: Field, C.B., Barros, V.R., Dokken, D.J., Mach, K.J., https://doi.org/10.1080/14693062.2015.1058240. Mastrandrea, M.D., Bilir, T.E., Chatterjee, M., Ebi, K.L., Estrada, Y.O., Genova, R.C., Engler, D., Groh, E.D., Gutsche, G., Ziegler, A., 2021. Acceptance of climate-oriented Girma, B., Kissel, E.S., Levy, A.N., MacCracken, S., Mastrandrea, P.R., White, L.L. policy measures under the COVID-19 crisis: an empirical analysis for Germany. Clim. (Eds.), Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global Pol. 0, 1–17. https://doi.org/10.1080/14693062.2020.1864269. and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Engström, G., Gars, J., Jaakkola, N., Lindahl, T., Spiro, D., van Benthem, A.A., 2020. Report of the Intergovernmental Panel on Climate Change. Cambridge University What policies address both the coronavirus crisis and the climate crisis? Environ. Press, Cambridge, United Kingdom and New York, NY, USA, pp. 709–754. Resour. Econ. 76, 789–810. https://doi.org/10.1007/s10640-020-00451-y. Thaker, J., Cook, C., 2021. Experience or attribution? Exploring the relationship between Evensen, D., Whitmarsh, L., Bartie, P., Devine-Wright, P., Dickie, J., Varley, A., Ryder, S., personal experience, political affiliation, and subjective attributions with mitigation Mayer, A., 2021. Effect of “finite pool of worry” and COVID-19 on UK climate change behavioural intentions and COVID-19 recovery policy support. J. Environ. Psychol. perceptions. PNAS 118. https://doi.org/10.1073/pnas.2018936118. 77, 101685. Galbraith, E., van den Bergh, J., 2020. Carbon tax to aid economic recovery. Nature 581, Thomas, G.O., Sautkina, E., Poortinga, W., Wolstenholme, E., Whitmarsh, L., 2019. The 262. English plastic bag charge changed behavior and increased support for other charges Gosens, J., Jotzo, F., 2020. China’s post-COVID-19 stimulus: no green new Deal in sight. to reduce plastic waste. Front. Psychol. 10, 266. https://doi.org/10.3389/ Environ. Innovat. Soc. Trans. 36, 250–254. https://doi.org/10.1016/j. fpsyg.2019.00266. eist.2020.07.004. Van Der Linden, S., 2017. Determinants and measurement of climate change risk Hammar, H., Jagers, S.C., 2006. Can trust in politicians explain individuals’ support for perception, worry, and concern [WWW document]. Oxford Res. Encycloped. Clim. climate policy? The case of CO2 tax. Clim. Pol. 5, 613–625. https://doi.org/ Sci. https://doi.org/10.1093/acrefore/9780190228620.013.318. 10.1080/14693062.2006.9685582. van Wijk, M., Naing, S., Franchy, S.D., Heslop, R.T., Lozano, I.N., Vila, J., Ballesté- Howarth, C., Bryant, P., Corner, A., Fankhauser, S., Gouldson, A., Whitmarsh, L., Delpierre, C., 2020. Perception and knowledge of the effect of climate change on Willis, R., 2020. Building a social mandate for climate action: lessons from COVID- infectious diseases within the general public: a multinational cross-sectional survey- 19. Environ. Resour. Econ. 76, 1107–1115. https://doi.org/10.1007/s10640-020- based study. PLoS One 15, e0241579. https://doi.org/10.1371/journal. 00446-9. pone.0241579. Howe, P.D., Marlon, J.R., Mildenberger, M., Shield, B.S., 2019. How will climate change Weber, E.U., 2006. Experience-based and description-based perceptions of Long-term shape climate opinion? Environ. Res. Lett. 14, 113001 https://doi.org/10.1088/ risk: why global warming does not scare us (yet). Clim. Chang. 77, 103–120. https:// 1748-9326/ab466a. doi.org/10.1007/s10584-006-9060-3. Jenkins-Smith, H.C., Ripberger, J.T., Silva, C.L., Carlson, D.E., Gupta, K., Carlson, N., Whitmarsh, L., 2011. Scepticism and uncertainty about climate change: dimensions, Ter-Mkrtchyan, A., Dunlap, R.E., 2020. Partisan asymmetry in temporal stability of determinants and change over time. Glob. Environ. Chang. 21, 690–700. https://doi. climate change beliefs. Nat. Clim. Chang. 10, 322–328. https://doi.org/10.1038/ org/10.1016/j.gloenvcha.2011.01.016. s41558-020-0719-y. Wong-Parodi, G., Berlin Rubin, N., 2022. Exploring how climate change subjective Kachi, A., Bernauer, T., Gampfer, R., 2015. Climate policy in hard times: are the attribution, personal experience with extremes, concern, and subjective knowledge pessimists right? Ecol. Econ. 114, 227–241. https://doi.org/10.1016/j. relate to pro-environmental attitudes and behavioral intentions in the United States. ecolecon.2015.03.002. J. Environ. Psychol. 79, 101728 https://doi.org/10.1016/j.jenvp.2021.101728. 16