Policy Research Working Paper 10914 Why Did Support for Climate Policies Decline in Europe and Central Asia? Alexandru Cojocaru Michael Lokshin Iván Torre Europe and Central Asia Region & Poverty and Equity Global Practice September 2024 Policy Research Working Paper 10914 Abstract This paper investigates trends in willingness to pay higher to pay to combat climate change declined over this period. taxes to combat climate change in countries of Eastern and The paper tests several hypotheses that could explain the Central Europe and Central Asia between 2016 and 2023. deterioration of public readiness to support climate change Using data from the Life in Transition Survey, it shows that policies. The most likely explanation is the growing politi- despite increasing attention from policy makers, scientists, cization of the climate change agenda in the region. and the media, the average shares of respondents willing This paper is a product of the Office of the Chief Economist, Europe and Central Asia Region and the Poverty and Equity Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http:// www.worldbank.org/prwp. The authors may be contacted at mlokshin@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Why Did Support for Climate Policies Decline in Europe and Central Asia? Alexandru Cojocaru, Michael Lokshin and Iván Torre1 JEL: O12, P28, Q54, Q58 Keywords: Climate change, Europe and Central Asia, public opinion, government policies. 1 Michael Lokshin is Lead Economist in the Chief Economist Office of the Europe and Central Asia Region of the World Bank. Alexandru Cojocaru is Senior Economist at the Poverty Global Practice of the World Bank. Ivan Torre is Senior Economist in the Chief Economist Office of the Europe and Central Asia Region of the World Bank. This paper’s findings, interpretations, and conclusions are entirely those of the authors and do not necessarily represent the views of the World Bank, its Executive Directors, or the countries they represent. The authors thank Miguel Purroy for excellent research assistance. 1. Introduction Since 1896, when Swedish scientist Svante Arrhenius discovered that doubling the amount of carbon dioxide in the atmosphere would increase Earth’s surface temperature by between 5 OC and 6OC, climate change has been the target of a burgeoning number of studies and approaches originating in the natural, human, and social sciences, and economics (Nunes and Ferreira 2022). By the 21st century, the overwhelming majority of scientists had come to believe in the evidence presented in these diverse works and were focused on finding solutions to mitigate climate change impacts. The increase in scientific confidence and media coverage does not appear to have translated into a parallel increase in public concern about climate change or support for climate change-related actions. Public opinion about climate change has fluctuated over the last 50 years. Knowledge and awareness of climate change grew in the 1980s and early 1990s. From the mid-1990s to the mid- 2000s, the variability of public concerns and opinions increased. The period between the mid-to late-2000s was characterized by declining public attention and rising skepticism in some countries, with growing polarization both within and across countries. The 2010s witnessed a stabilization of public concern about climate change (Capstick et al. 2015). The period from 2016 to 2024 has been described as one in which “opinion leader influence” played an important part, as social media increasingly shaped people’s perceptions of environmental issues (Baiardi 2022). This disconnect between popular attitudes and the growing body of scientific evidence stems from the complex combination of emotions, personal experiences, trust, values, worldviews, and political ideologies that influence individuals’ attitudes and beliefs about climate change (Slovic 2016, Libarkin et al. 2018). As the full ramifications of climate change remain uncertain and have yet to fully manifest themselves, individuals find it challenging to accurately weigh the potential costs against the prospective benefits of enacting climate policies (Crispino and Loberto 2024). The Science Comprehension Thesis postulates that because people do not know what scientists know and are not able to process information the way scientists can, they fail to take climate change as seriously as scientists believe they should (Weber and Stern 2011). The urgency of the moment calls for coordinated commitment and political will on the part of the international community to cut greenhouse gas emissions to counteract increasingly severe climate impacts (IPCC 2023). These efforts require broad support and engagement of global populations. 2 Public perceptions of climate change are the major determinants of countries’ climate and energy response. Understanding the drivers of concerns about climate change is essential to designing effective communication strategies, democratic policies, and socially robust technologies (Whitmarsh and Capstick 2018). This paper investigates recent trends in attitudes toward climate change in Eastern and Central Europe and Central Asia based on data from the 2016 and 2023 rounds of the Life in Transition Survey (LiTS). It shows that despite increasing attention from policy makers, scientists, and the media, the proportion of respondents willing to pay to combat climate change declined significantly between the two rounds. We propose and test several hypotheses that could explain the deterioration of public support for climate change policies in the region. This paper contributes to the literature on perceptions of climate change in several meaningful ways. It is among the first to quantitatively demonstrate the growing influence of right-wing populism on support for climate policies across countries in Europe and Central Asia. It systematically tests several hypotheses to explain the decline in climate policy support, including competing priorities, personal experiences with extreme weather, and political polarization. The paper contributes to a better understanding of factors influencing public opinion on climate change. The paper is also one of the first to draw on data covering the period of profound transformations of the political and economic landscape in Europe and Central Asia, including as a result of the COVID-19 pandemic, the sharp increase in the cost of living, and the Russian Federation’s invasion of Ukraine. The paper is organized as follows. Section 2 reviews the literature. Section 3 outlines our hypotheses. Section 4 describes the data and presents some descriptive results. Section 5 presents the main results of our multivariate analysis. Section 6 summarizes the paper’s main findings and proposes policy implications. 2. Literature review The decline in public concern about climate change in the late 2000s is reported by Ratter et al. (2012) in Germany, where the proportion of people who feel that climate change poses a threat almost halved (from 17 to 9 percent) between 2008 and 2011. Climate skepticism was reported to be on the rise in Great Britain (BBC 2010). A continuous decrease in concern about climate change 3 was also reported around the same time among citizens of the EU (Eurobarometer 2009), Canada, and Australia (Puglise and Lyons 2010). Franzen and Volg (2013) analyze trends in 33 OECD and non-OECD countries, where they find a moderate decline in environmental concerns over the two decades before 2010. More recent research, conducted from the late 2010s to today, indicates heterogeneous but mainly elevated climate change concerns. Marlon et al. (2022) find increases in the importance of climate change issues and perceived harm in every state in the US between 2008 and 2020. They show that policy support grew more in liberal states but remained unchanged elsewhere. According to Lachapelle and Borick (2022), between 2011 and 2021, public opinions in Canada and the US shifted toward acceptance of the anthropogenic nature of climate change. Both countries registered only moderate increases in support for policy interventions. However, the degree of politicization and polarization in the debate on nature and optimal policies toward climate change was much greater in the US. Combining data from Eurobarometer and Twitter, Crispino and Loberto (2024) show that in Italy in 2022, attention to climate change had increased but sentiments toward climate change deteriorated. Kenny (2024) reveals that prioritization of environmental issues in the UK rose throughout the 1980s, falling dramatically in the early 1990s and recovering by 2001 before falling in subsequent years. This decline intensified following the financial crisis of 2008. Support for environmental issues recovered during the last decade and increased between the Spring of 2018 and the Winter of 2019. 3. Data We use data from the 2016 and 2023 rounds of the Life in Transition Survey (LiTS) conducted by the European Bank for Reconstruction and Development (EBRD) and the World Bank (EBRD 2023). The survey covers the transition economies of Europe and Central Asia and several comparator countries in Western Europe, the Middle East, and North Africa. It included 34 countries in 2016 and 39 countries in 2023. The survey is based on a nationally representative sample of around 1,500 households per country in the 2016 round and around 1,000 households 4 per country in the 2023 round. For this analysis, we used a subset of 30 countries surveyed in both rounds of the LiTS. 2 We capture support for tax-financed investments in climate change actions by examining responses to the question “Would you be willing to give up part of your income or pay more taxes if you were sure that the extra money was spent to combat climate change?” We create a dummy with a value of 1 if the respondent responded yes and 0 if he or she responded no or don’t know. This dummy variable is the main dependent variable of our analysis. Our data also contain information on age, gender, education, marital status, religion, and type of residence, which we use as controls in our empirical analysis. We supplement the data from the LiTS with the Google Trends index, which is calculated by taking the total number of searches for a specific term at a given location and within a given time frame, dividing it by the total number of all searches conducted at that location and during the period, and then scaling this ratio with respect to the highest value in the series. This normalization process precluded us from comparing search interest indexes over time (i.e., Mavragani and Ochoa 2019). To overcome this problem, we present the comparisons of the changes in the index with the changes in the search index for the topics we consider relevant and comparable with the issue of climate change. We also use administrative data on several environmental risks at the primary sampling unit (PSU) level, 3 capturing means, standard deviations, ranges of temperature and precipitation, information on air pollution, and the severity of flooding. We use PSU-level statistics on the risk factors for the seven years before each survey. We derive a weighted average of each risk factor with weights linearly inversely related to the time from the survey date (the further away the year is from the survey date, the lower the weight). We also tested specifications using principal component analysis or quadratic weights to aggregate seven years of risk factor information. Qualitatively, 2 This subset comprises countries that belonged to the former Soviet Union (except Ukraine and Turkmenistan), countries of the former Yugoslavia, Germany, Greece, Mongolia, Türkiye, and countries in Central and Eastern Europe that transitioned from a planned to a market economy during the 1990s. We excluded from the sample six countries in the Middle East and North Africa that were only included in the last round (Algeria, Jordan, Lebanon, Morocco, Tunisia, and the West Bank and Gaza). 3 We have this administrative data for all countries in our sample except 9 countries in the Middle East, North Africa, and Western Europe. 5 these specifications produce results that are similar to the results of the linear weighted aggregation function. Temperature is represented as the mean temperature in degrees Celsius, and precipitation is measured as the mean rainfall in millimeters per hour. Both temperature and precipitation data come from NASA’s Global Land Data Assimilation System Project (Beaudoing and Rodell 2020). Air pollution is measured by the concentration of ground-level fine particulate matter (PM2.5), in micrograms per cubic meter, with data obtained from the NASA Socioeconomic Data and Applications Center (Hammer et al. 2022). The Flood Severity Index is based on data from the Global Active Archive of Large Flood Events by the Dartmouth Flood Observatory at the University of Colorado (Brakenridge 2024). 4. Documenting the change Figure 1 shows the changes in the willingness to pay more taxes to combat climate change between 2016 and 2023. The proportion of the population supporting higher taxes for that purpose declined in the 30 countries in our sample, from 36.6 to 30.7. This 5.9 percentage point (16.1 percent) decline resulted from an average drop of almost 12 percentage points for 20 countries in our sample (countries below the 45-degree line in Figure 1) and an average increase of about 6.3 percent in 10 countries (countries above the 45-degree line). The proportion of the population willing to pay part of their income or higher taxes to combat climate change declined among virtually all groups (Table 1) by more than 20 percent for men and only 12 percent for women. By age, the most significant decline was among 40 – 49 (23.9 percent) and 30 – 39 (17.9). Among respondents 70 and older, willingness to pay did not change between the two survey rounds. The differences between the two rounds exceeded 20 percent among people with upper-secondary education, post-secondary education, and bachelor’s degrees. Among respondents with primary or lower levels of education, willingness to pay to combat climate change rose, but the number of respondents in these groups is small. The decline in willingness to pay is most pronounced among married people (a 17.1 percent drop), single respondents (17.0 percent), Christians (20.6 percent), and people with other religious affiliations (21.5 percent). The decline is greater in urban than rural areas. 6 To identify the drivers of these changes, we applied an Oaxaca-Blinder decomposition (Oaxaca 1973). This method decomposes the difference in the proportions of the population between 2016 and 2023 into the component attributable to the group differences in the magnitudes of the determinants of the support for higher taxes, the component attributable to the group differences in the effect of these determinants, and the interaction term. We model the willingness to pay higher taxes to combat climate change in two survey rounds (denoted 1 and 2) as a function of individual characteristics. The linear model of such a relationship can be expressed as follows: 1 = 1′ 1 1 + and 2 2′ 2 2 = + , ( ≠ ) (1) where, 1 is the willingness to pay by individual i in round 1, 2 is the willingness to pay by individual j in round 2, 1 2 1 and are vectors of individual characteristics in two rounds, and 2 are vectors of estimated coefficients, and 1 and 2 are mean-zero error terms. Then, the difference in the willingness to pay between the two rounds can be evaluated as: ∆ ( ) = (1 ) − � 2 1 ′ 1 2 � = ( ) − ( )′ 2 (2) � correspondingly, we can rewrite Equation (2) as follows: � and Replacing ( ) and () with ��� � = � ∆ �� � 1 − ��′ 2 + 2 �� � �′ (1 − 2 ) + � 2 ��� 1−�� �� ′ 2 � (1 − 2 ). (3) The first term in Equation (3) represents the “endowment effect,” the part of the difference in the willingness to pay explained by the differences in average characteristics between the two rounds. The second component, the “coefficient effect,” is the part of the difference attributable to changes in the returns on characteristics. The last component, the “interaction effect,” is the part of the difference attributable to the simultaneous effect of differences in endowments and coefficients. Table 2 presents the results of Oaxaca-Blinder decomposition. The negative sign of the decomposition component indicates that the corresponding changes between the two rounds of the survey led to a reduction in the difference in willingness to pay. The positive sign indicates that the corresponding changes increased the difference in willingness to pay. The 5.3 percentage point difference in the share of the population willing to pay to combat climate change can be decomposed into the effect of changes in the endowments (–8.8 percent of the difference), changes in the returns or coefficients (118.6 percent of the difference), and the interaction term (–9.8 7 percent of the difference). 4 The contributions of these components are statistically different from zero, indicating that the change in coefficients is the primary factor explaining the difference in willingness to pay between the two survey rounds. The contribution of the change in endowments – the demographic shift between the two rounds – is much smaller and works in the opposite direction, reducing the difference. These results suggest that the change in willingness to pay may not have been the result of a compositional change (with groups intrinsically less willing to pay increasing their share in the population and groups intrinsically more willing to pay decreasing their share) but rather the result of a shift in the population’s intrinsic preferences. Using the Oaxaca-Blinder decomposition, we can also identify how much each variable (or group of variables) contributes to the components of the decomposition. 5 About 62 percent of the −8.8 percent contribution of endowments to the differences in shares is explained by changes in religious composition, 17.9 percent by the changes in marital status composition, and 14.5 percent by changes in household size between the two rounds. Age appears to be the most significant factor in the contribution of coefficients to changes in the proportion of people willing to pay for climate change (209.3 percent). To some degree, the age effect is counterbalanced by the negative effect of the intercept (−164.7 percent). Changes in the coefficients on household size and type of residence positively and significantly affect the differences in the population shares. 5. Explaining the change The results of the Oaxaca-Blinder decomposition confirm that the decline in the willingness to pay to combat climate change is not an artifact of inter-temporal changes in the population structure (population aging, urbanization, changes in educational attainment). Rather, it reflects changes in attitudes related to support for climate change action. In this section, we test four hypotheses related to the possible drivers of these attitudinal changes: (i) a secular decline in the willingness to pay, (ii) changes in the relative salience of competing priorities, (iii) changes in the level of political polarization in society, and (iv) changes caused by personal experience with climate risks. This list of hypotheses is not exhaustive and is driven, in part, by the data available in the LiTS. 4 The difference in population shares reported in Table 2 is smaller than the difference reported in Table 1 because some observations with missing values were dropped from the analysis. 5 The binary form of the dependent variables requires using the non-linear, normalized decomposition approach by Yun (2005). 8 Although it is not exhaustive, we believe it captures key factors that may explain the decline in support for climate change actions between 2016 and 2023. Secular decline in willingness to pay hypothesis. The decline in willingness to pay to combat climate change could be a part of the overall decline in willingness to pay higher taxes for any policy. In principle, this willingness could decline if individuals became more risk-averse and changed their spending and saving patterns after facing the series of adverse shocks that hit Europe during the period between the two survey rounds. However, the share of respondents willing to pay more taxes for improvements in education and health services rose in 2023 (Figure 2). We, therefore, reject this hypothesis. Competing priorities hypothesis. During the period between the two survey rounds, countries in Europe and Central Asia faced multiple challenges, including an influx of refugees from the Middle East and North Africa in the early 2010s and from Ukraine in 2022, the COVID-19 pandemic, political tension and security concerns, the war in Ukraine, and the cost-of-living crisis (Lokshin et al. 2024). These competing priorities may have diverted public attention and resources away from climate change, making it more challenging to build support for higher taxes to combat climate change. The “finite pool of worry” concept postulates that people have a limited capacity for worry or concern (Weber 2006). Issues perceived as more threatening or relevant can crowd out concern for climate change. The media play a crucial role in influencing public discourse and the choice of priorities. Climate change receives less attention when other crises and events dominate media coverage. Brulle et al. (2012) report that US climate change concerns declined significantly during the 2008 financial crisis. The immediate threat of the COVID-19 pandemic and measures to contain its spread and mitigate its economic and social impact shifted priorities away from the more gradual and longer-term threats posed by climate change (Helm 2020). At the same time, the pandemic increased awareness of the interrelationships of global risks, including climate change, and boosted support for policies to address these challenges (Bouman et al. 2021). Security concerns and the humanitarian crisis caused by the war in Ukraine diverted public interests away from climate change (Tooze 2022, Genschel et al. 2024). A report by the European Investment Bank (2022) argues that the decline in the share of respondents identifying climate change as the 9 most severe global challenge − from 29 percent in 2021 to 23 percent in 2022 − is related to the beginning of the war in Ukraine. There is no direct means of testing the competing priority hypothesis, so we draw on indirect evidence. We use responses to the following questions about general attitudes about respondents’ living situation: To what extent do you agree with the following statements: “The economic situation in our country is better today than around 4 years ago”; “I have done better in life than my parents”; “My household lives better nowadays than around 4 years ago”; “All things considered, I am satisfied with my life now”; “Children who are born now will have a better life than my generation”; “All things considered, I am satisfied with my job as a whole.” Responses to these questions take five values ranging from “strongly disagree” to “strongly agree.” They capture a general sense of worry about respondents’ livelihoods. We assume that this general sense of worry would crowd out concerns about climate change, as they are less immediately linked to the respondent’s livelihood. Table 3 displays the aggregate responses. Respondents’ assessment of their life situation improved between 2016 and 2023 across all six dimensions, with the differences in averages being positive and statistically significant. Figure 3 shows qualitatively similar results at the country level. The response averages are higher in 2023 than in 2016 (the markers are above the 45-degree line), indicating that, on average, respondents’ life situation across these six dimensions improved in most countries, and the general sense of worry may have declined. In theory, this result should be associated with an increase in concerns about climate change – but this is not what we observe. We substantiate these results with Google trend data. The competing priority hypothesis would require the relative number of searchers about climate change to decline between 2016 and 2023. But the opposite is observed: Between 2016 and 2023, the Google trend index of search topics increased by 31 percent for environmental and climate change topics and 47 percent for global warming; it rose by 36 percent for health, 31 percent for education, 16 percent for immigration policies and border issues, and by 2 percent for public safety. We find analogous changes in the index at the country level. 6 6 These results are available from the authors on request. 10 This, albeit indirect, evidence allows us to conclude that the decline in willingness to pay for programs to combat climate change cannot be explained by the decrease in interest or attention to climate change because more pertinent problems pushed this agenda away. Political polarization hypothesis. Changes in willingness to pay for climate change could be related to the politicization of the climate change agenda. Substantial research demonstrates that partisanship and political ideology are essential determinants of attitudes toward climate change (e.g., Dunlap et al. 2016). Value orientation and identity are often crucial to selecting and perceiving information about complex challenges like climate change, and they can be more important than the cognitive ability to understand such issues (Kahan et al. 2012). Recent elections have seen the rise of right-wing populism in many Western democracies in Europe. The influence of right-wing populist parties, which oppose climate change and hamper climate policies at the local and national levels, continues to grow (Nagu et al. 2024). For example, support for policies to slash carbon emissions to net zero by 2050 declined in 19 of 27 EU countries, with Finland, Estonia, and Czechia registering a drop of up to 15 percentage points (Mathiesen et al. 2024). The decline in willingness to support climate change actions through higher taxes may, then, be related to the increasing polarization of the climate change agenda in Europe and the growing influence of right-wing populist parties in the region. We follow the conceptual framework of Eatwell and Goodwin (2018) and Mudde (2007), who define four pillars (the “Four-Ds”) of national populism. Distrust refers to the growth of skepticism and lack of faith in the traditional elites and institutions. Destruction relates to the perceived erosion of national identity, traditional values, established ways of life, and the fear of losing what are considered core aspects of the nation’s character and heritage. Deprivation addresses economic anxieties and concerns about job insecurity, stagnant wages, and declining living standards. It is often related to globalization and technological change. De-alignment sees established political parties as increasingly similar and unresponsive to voters’ concerns; as a result, voters become more willing to support new radical political movements. This framework helps identify characteristics that differentiate right-wing populists from other groups of the population. Supporters of right-wing populism often portray migrants as economic competitors and cultural threats, associating them with concerns about national security and crime. Migrants are frequently used to symbolize unwanted globalization and loss of national sovereignty. 11 Minorities may be perceived as failing to integrate or having divided loyalties. They are also sometimes viewed as beneficiaries of “special treatments” such as affirmative action or diversity policies at the expense of the majority. Their cultural practices are often framed as incompatible with national values. Some movements emphasize traditional gender roles and family values; others claim to protect women’s rights, particularly in contrast to perceived threats from immigrant cultures. Women’s advancement in society may be seen positively or as a threat to traditional social structures by others (Dietze and Roth 2020). Populists are often skeptical about existing redistribution systems, which they view as corrupt, and feel that “natives” are being deprived while others (often immigrants or minorities) unfairly benefit from redistribution (Halikiopoulou and Vlandas 2022). Right-wing populist economic attitudes are more nuanced. Supporters distrust elites, whom they see as manipulating competition for personal gains, and oppose global competition that threatens local industries. Although strongly supporting private property rights, they favor domestic ownership and are skeptical of foreign control over land or critical industries. Their stance on market economies is mixed: They support free markets for domestic businesses but are wary of the impact of globalization on national interests. Based on this framework, we create two indices. The right-wing populism index is constructed as an equally weighted sum of respondents’ responses to questions related to attitudes toward migrants, women, family values, and redistribution policies. The economic conservatism index comprises respondents’ opinions towards private ownership, competition, and the market economy. 7 Responses to each question are normalized by the mean response in 2016. Higher values of the index correspond to more right-wing views. The right-wing populism index rose from 0.99 in 2016 to 1.06 in 2023, reflecting the growing influence of right-wing ideologies in the countries in our sample, which grew by about 7 percent. 7 The right-wing populism index is constructed as an equally-weighted average of the responses to the following questions: whether immigrants make a valuable contribution to the national economy or are a burden on the social protection system; the extent to which respondents agree or disagree that women are as competent as men in executive positions; whether it is better for everyone involved if men and women earn money and take care of the home and children; and preferences regarding government spending to reduce inequality. It ranges from 0 (the most left-wing ideology) to 1 (the most right-wing ideology). The economic conservatism index is constructed as a weighted average of responses to questions about whether private ownership of business and industry rather than government ownership should be increased, whether competition is viewed as beneficial or harmful, and whether respondents believe a market economy is preferable to a planned economy. 12 That increase could explain the decline in willingness to pay to combat climate change if the strength of the relationship between political ideology and willingness to pay did not significantly decline between the two years. 8 Table 4 shows the results of the FE regression estimation of the model, in which we regress willingness to pay for climate change on the right-wing populism and economic conservatism indexes and respondent’s characteristics separately for the 2016 and 2023 samples. Respondents with stronger right-wing views are significantly less likely to be willing to pay to combat climate change in both years. However, the equality of the coefficients of that variable in the 2016 and 2023 regression cannot be rejected (χ2 = 0.49, Prob = 0.482). At the same time, we find no significant effect of the economic conservatism index on our dependent variable. We observe an increase in the share of right-leaning respondents from 2016 to 2023 but no significant differences in the effects of such ideologies on individual willingness to pay. 9 So, the decline in willingness to pay could be attributed to the growing share of right-leaning respondents from 2016 to 2023. 10 Personal experience of weather and climate-change events hypothesis. As the impacts of climate change become more severe and widespread, personal experiences may play an essential role in shaping public attitudes toward climate change. Exposure to extreme weather events can make the issue more salient and reduce the psychological distance often associated with climate change (McDonald et al. 2015). Bergquist et al. (2021) and Marlon et al. (2021) demonstrate that experiences of extreme weather events are associated with increased support for climate policies in the US. Similar results have been found in the UK (Demski 2020), and in Germany, Norway, and Sweden (Ogunbode 2020). Siedlecki and Weier (2022) indicate that the effect of personal experiences on the support of climate change policies is moderated by perceived abilities to cope with and adapt to climate risks, which could depend on proximity to the resilience infrastructure. 8 If the influence of right-wing ideology increases, but the correlation between the right-wing ideology and support for climate change policies weakens, the net effect is ambiguous. 9 We find no significant differences between the rounds of the LiTS in the effects of the political polarization index on the willingness to pay for education, health, fight inequality, biological loss, and pollution. 10 Estimating the model shown in Table 4 on a sample that excludes countries of Central Asia (Armenia, Azerbaijan, Kazakhstan, the Kyrgyz Republic, Tajikistan, and Uzbekistan) produces results that are qualitatively similar to those of the main model; coefficients on the political polarization index are negative and significant for both years but not statistically different from each other. 13 If the intensity of climate change events, such as extreme temperatures, precipitation, and floods, declined from 2016 to 2023, we might expect a corresponding reduction in public demand for climate actions and willingness to support such policies. That support would also change if climatic events affected people differently in the periods covered by the two survey rounds. To test the personal experience hypothesis, we use administrative data on several environmental risks at the primary sampling unit (PSU) level. 11 The top panel of Table 5 shows changes in the means of these risk factors between 2016 and 2023. The levels of precipitation remained relatively constant over that period. The temperature, however, increased across all three dimensions. The level of air pollution and flood severity index were higher in 2016 than in 2023. The bottom panel of Table 5 shows the coefficients on the risk factors when we add them to the specification in Table 4. None of the factors was significantly different from zero in both years. Only the flood severity index significantly but negatively affects the willingness to pay in 2016. Our results fail to confirm the personal experience hypothesis, in terms of both differences in levels and changes in the effects of risk factors on willingness to pay for climate change actions. Extended hypothesis tests Does the Oaxaca-Blinder decomposition of the changes in willingness to pay higher taxes to combat climate change confirm our hypothesis results? Table 6 shows the decomposition in which we extended our baseline specification in Table 2 by adding indexes of right-wing populism and economic conservatism and the set of the PSU-level variables representing the environmental risks. The results are consistent with our findings. The right-wing populism index is the only factor that significantly affects the difference in willingness to pay between the third and fourth rounds of the LiTS (Specification 2). The effect of this variable comes from the changes in “endowment,” or, in other words, from the increased share of respondents who report embracing a right-wing populist ideology. Neither the economic conservatism index nor the environmental risk indicators have a meaningful effect on that difference (Specifications 3 and 4). 11 We have administrative data for all countries in our sample except nine countries in the Middle East, North Africa, and Western Europe. 14 6. Conclusions This paper uses data from the Life in Transition Survey to investigate trends in public support for climate change policies in Eastern and Central Europe and Central Asia between 2016 and 2023. Our analysis reveals that despite increasing scientific consensus and media attention, the share of respondents willing to pay higher taxes to combat climate change declined significantly between 2016 and 2023. This decline was heterogeneous across countries and demographic groups, with the most substantial decreases observed among men, people 30 to 49 years old, and people with higher education levels. Changes in the willingness to pay cannot be explained by changes in the structure of the population during this period, such as aging or urbanization. Our analysis shows that changes in attitudes, not demographic shifts, primarily caused the decline in support. We formulate and test several hypotheses to explain this trend. Our empirical evidence confirms that the decline is not part of a general decrease in willingness to pay for public policies. Personal experiences with extreme weather also cannot explain the change. The rise of right-wing populism emerges as the most plausible explanation for the declining support for climate policies. Our analysis reveals a negative correlation between right-wing populist attitudes and willingness to support climate change policies across both survey rounds. These findings have important implications for policy makers in the region. The growing influence of right-wing populism presents a significant challenge to building and maintaining public support for climate action. Our results suggest that policies to combat climate change may need to be framed and communicated in ways that resonate with a broader spectrum of political ideologies and appeal across political divides. This study highlights the critical linkages between political ideologies, public opinion, and support for climate policies. As the urgency of climate action grows, understanding and addressing these dynamics are critical for developing effective regional climate strategies. The challenge for policy practitioners is to carefully navigate the evolving political landscape, while building public support for impactful climate action. 15 Future research could examine strategies for maintaining or increasing climate policy support in contexts of rising populism. Exploring policy approaches that bridge the growing ideological divide could inform more effective and widely supported climate strategies. 16 References Baiardi, D. (2022). “What do you think about climate change?” Journal of Economic Surveys, 37: 1255-1313. BBC News. (2010). “Climate skepticism ‘on the rise,” BBC poll shows, February 7, 2010 (http://news.bbc.co.uk/2/hi/science/nature/8500443.stm). Beaudoing, H. and M. Rodell. (2020). GLDAS Noah Land Surface Model L4 monthly 0.25 x 0.25 degree V2.1, Goddard Earth Sciences Data and Information Services Center (GES DISC), Greenbelt, Maryland, USA. Bergquist, M., Nilsson, A., and N. Harring. (2021). “Experiencing a severe weather event increases concern about climate change.” Frontiers in Psychology, 12, 546. Bouman, T., Verschoor, M., Albers, C. J., Böhm, G., Fisher, S. D., Poortinga, W., and L. Steg (2021). “When worry about climate change leads to climate action: How values, worry and personal responsibility relate to various climate actions.” Global Environmental Change, 71, 102384. Brakenridge, G. (2024). Global Active Archive of Large Flood Events. Dartmouth Flood Observatory, University of Colorado, Boulder, USA. Brulle, R., Carmichael, J., and J. Jenkins. (2012). Shifting public opinion on climate change: An empirical assessment of factors influencing concern over climate change in the US, 2002–2010. Climatic Change, 114(2):169–188. Capstick, S., Whitmarsh, L., Poortinga, W., Pidgeon, N., and P. Upham. (2015). “International trends in public perceptions of climate change over the past quarter century.” WIREs Climate Change, 6:35-61. Crispino, M., and M. Loberto (2024). “Do people pay attention to climate change? Evidence from Italy.” Journal of Economic Behavior and Organization, 219: 434-449. Dietze, G. and J. Roth (2020). Right-Wing Populism and Gender: European Perspectives and Beyond. Transcript. Verlag, Bielefeld. Demski, C., Capstick, S., Pidgeon, N., and A. Spence. (2020). “Experience of extreme weather affects climate change mitigation and adaptation responses.” Climatic Change, 158(1): 149-164. Dunlap, R., McCright, A., and J. Yarosh. (2016). “The political divide on climate change: Partisan polarization widens in the U.S.” Environment: Science and Policy for Sustainable Development, 58(5): 4–23. Eatwell, R. and M. Goodwin. (2018). National Populism: The Revolt Against Liberal Democracy. Pelican Books. EBRD (European Bank for Reconstruction and Development). (2023). Life in Transition Survey. London. Eurobarometer. (2009). Europeans’ Attitudes Towards Climate Change (Special Eurobarometer 322, European Commission, November 2009). European Investment Bank. (2022). EIB Climate Survey 2022. Luxembourg: European Investment Bank. 17 Franzen, A., and D. Vogl. (2013). “Two decades of measuring environmental attitudes: A comparative analysis of 33 countries.” Global Environmental Change, 23(5): 1001-1008. Genschel, P., Limberg, J., and L. Seelkopf. (2024). “The climate crisis, policy distraction and support for fuel taxation.” European Journal of Political Research. https://doi.org/10.1111/1475-6765.12687 Hammer, M., van Donkelaar, A., Li, C., Lyapustin, A., Sayer, A., Hsu, N., Levy, R., Garay, M., Kalashnikova, O., Kahn, R., Brauer, M., Apte, J., Henze, D., Zhang, L., Zhang, Q., Ford, B., Pierce, J., and R. Martin. (2022). Global Annual PM2.5 Grids from MODIS, MISR and SeaWiFS Aerosol Optical Depth (AOD), 1998-2019, V4.GL.03. NASA Socioeconomic Data and Applications Center, Palisades, New York, USA. Halikiopoulou, D., and T. Vlandas. (2022). Understanding right-wing populism and what to do about it. Vienna: FES Regional Office for International Cooperation - Democracy of the Future, Viena. Helm, D. (2020). The environmental impacts of the coronavirus. Environmental and Resource Economics, 76, 21-38. IPCC (The Intergovernmental Panel on Climate Change) (2023). AR6 Synthesis Report: Climate Change 2023, IPCC Geneva, Switzerland. Kahan, D., Peters, E., Wittlin, M. et al. (2012). “The polarizing impact of science literacy and numeracy on perceived climate change risks.” Nature Climate Change 2: 732–735. Kenny, J. (2024). “The Changing Prioritization of Environmental Protection in Britain: 1982– 2019.” Government and Opposition, 59(1):168-186. Lachapelle, E., and C. Borick. (2022). “A decade of comparative Canadian and American public opinion on climate change.” North American Colloquium Climate Report Series. Libarkin, J., Gold, A., Harris, S., McNeal, K., and R. Bowles. (2018). “A new, valid measure of climate change understanding: Associations with risk perception.” Climatic Change, 150: 403–416. Marlon, J., Wang, X., Bergquist, P., Howe, P., Leiserowitz, A., Maibach, E., Mildenberger, M. and S. Rosenthal (2022). “Change in US state-level public opinion about climate change: 2008–2020,” Environmental Research Letters, 17(12). Mathiesen, K., Camut, N., Weise, Z., Cooper, C., and L. Guillot. (2024). “Bears, cars and angry farmers fuel green backlash,” Politico, February 22, 2024. Marlon, J. R., Wang, X., Mildenberger, M., Bergquist, P., Swain, S., Hayhoe, K., and A. Leiserowitz. (2021). “Hot, dry days increase perceived experience with global warming.” Global Environmental Change, 68, 102247. Mavragani, A. and G. Ochoa. (2019). “Google Trends in Infodemiology and Infoveillance: Methodology Framework.” JMIR Public Health and Surveillance 5(2):13439 McDonald, R., Chai, H., and B. Newell. (2015). “Personal experience and the ‘psychological distance’ of climate change: An integrative review.” Journal of Environmental Psychology, 44:109-118. Mudde C. (2007). Populist Radical Right Parties in Europe. Cambridge University Press. Nagu, N., Strzalkowski, P., Kuner, L., and B. Alvarez-Villar. (2024). “Right-wing populists challenge Europe’s climate efforts.” Clean Energy Wire. June 07, 2024. 18 Nunes, L., and D. Ferreira. (2022). “Perception of Climate Change Effects over Time and the Contribution of Different Areas of Knowledge to Its Understanding and Mitigation.” Climate 10(7). Oaxaca, R. (1973). “Male-Female Wage Differentials in Urban Labor Markets.” International Economic Review, 14: 693–709. Ogunbode, C. A., Demski, C., Capstick, S. B., and R. Sposato. (2020). “Attribution matters: Revisiting the link between extreme weather experience and climate change mitigation responses.” Global Environmental Change, 62, 102117. Puglise, A., and L. Lyons (2010). Australiens’ Views Shift on Climate Change. (http://www.gallup.com/poll/141782/Australians-Views-Shift-Climate-Change.aspx). Ratter, B., Philipp, K., and H. von Storch (2012). “Between hype and decline: recent trends in public perception of climate change.” Environmental Science & Policy, 18: 3-8. Siedlecki, M., and J. Weier. (2022). “The Role of Vulnerability and Perceived Adaptive Capacity in Shaping Climate Change Risk Perceptions: A Systematic Review.” Frontiers in Psychology, 13, 839198. Slovic, P. (2016). “Informing and educating the public about risk.” In P., Slovic, (Ed.), The perception of risk: 220-236, Routledge. Tooze, A. (2022). “How the world is paying for Putin’s war in Ukraine.” Foreign Policy, 241: 60-71. Weber, E. and P. Stern. (2011). “Public understanding of climate change in the United States.” American Psychology 66: 315-328. Weber, E. (2006). “Experience-based and description-based perceptions of long-term risk: Why global warming does not scare us (yet).” Climatic Change, 77(1–2):103–120. Whitmarsh, L., and S. Capstick (2018). “2 - Perceptions of climate change.” Editor(s): Clayton, S., and C. Manning, Psychology and Climate Change, Academic Press. Yun, M. (2005). “A simple solution to the identification problem in detailed wage decompositions.” Economic Inquiry, 43(4):766–72. 19 Figure 1: Willingness to pay more taxes to combat climate change, 2016 and 2023. 20 Figure 2: Willingness to pay more taxes to combat climate change, improve education, reduce inequality, and improve healthcare, 2016 and 2023. 21 Figure 3: Changes in attitudes about life between 2016 and 2023. 22 Table 1: Proportion of respondents willing to pay more taxes to finance policies to combat climate change. LiTS 2016 and 2023. 2016 2023 Difference Percent change Total 0.366 0.307 0.059*** 16.1 Gender Male 0.389 0.309 0.081*** 20.7 Female 0.348 0.307 0.042*** 12.0 Age group Under 20 0.440 0.397 0.043 9.8 20-29 0.411 0.373 0.038 *** 9.3 30-39 0.406 0.334 0.073 *** 17.9 40-49 0.407 0.310 0.097*** 23.9 50-59 0.376 0.310 0.067 *** 17.7 60-69 0.307 0.273 0.034*** 11.1 70 and older 0.243 0.236 0.007 2.7 Education No degree / no education 0.311 0.352 −0.041 −13.0 Primary 0.304 0.361 −0.057 *** −18.6 Lower secondary 0.293 0.251 0.042*** 14.3 Upper secondary 0.377 0.296 0.081 *** 21.6 Post-secondary non-tertiary 0.372 0.295 0.077 *** 20.7 Tertiary (not a university diploma) 0.396 0.346 0.050 *** 12.7 Bachelor’s degree or more 0.467 0.362 0.104*** 22.4 Master’s degree or PhD 0.381 0.326 0.054*** 14.3 Marital status Single (never married) 0.369 0.306 0.063*** 17.0 Married 0.405 0.336 0.069 *** 17.1 Widowed 0.265 0.244 0.021** 7.8 Divorced 0.252 0.235 0.017 6.8 Separated 0.322 0.400 −0.077** −24.0 Religion Atheist/agnostic/none 0.276 0.234 0.042*** 15.3 Buddhist 0.565 0.505 0.060 ** 10.7 Jewish 0.408 0.556 −0.147 −36.1 Christian 0.318 0.252 0.065*** 20.6 Muslim 0.491 0.440 0.051 *** 10.4 Other 0.299 0.235 0.064 ** 21.5 Type of residence Rural 0.379 0.327 0.052*** 13.8 Urban 0.357 0.294 0.063 *** 17.7 Note: *** indicates that the difference between 2016 and 2023 proportions is statistically different from zero at the 1 percent level, ** statistically significant at the 5 percent level, * statistically significant at the 10 percent level. 23 Table 2: Oaxaca-Blinder decomposition of changes in proportion of respondents willing to pay higher taxes to combat climate change. Coefficient Standard Error Percent of component Overall Sample 2016 0.368*** 0.028 Sample 2023 0.316 *** 0.022 Difference 0.053 *** 0.021 100.0 Endowments − 0.005 ** 0.006 −8.8 Coefficients 0.062 *** 0.021 118.6 Interaction −0.005 ** 0.005 −9.8 Endowments 100.0 Age −0.000 0.000 9.4 Education 0.000 0.003 − 5.8 Marital status −0.001 0.001 17.9 Religion −0.003 0.004 62.9 Household Size −0.001 0.001 14.5 Monthly Income −0.000 0.000 2.7 Urban 0.000 0.000 − 1.5 Coefficients 100.0 Age 0.131 *** 0.063 209.3 Education 0.004 0.030 5.9 Marital status 0.001 0.013 1.3 Religion −0.003 0.022 −5.1 Household Size 0.020 0.020 31.7 Monthly Income − 0.001 0.001 − 1.2 Urban 0.014 0.011 23.0 Intercept − 0.103 0.076 −164.7 Interactions 100.0 Age 0.000 0.001 −1.3 Education − 0.007 * 0.004 131.0 Marital status 0.000 0.001 −6.7 Religion 0.001 0.002 −28.7 Household Size −0.000 0.001 6.6 Monthly Income 0.000 0.000 −2.6 Urban − 0.000 0.001 1.8 Note: *** statistically significant at the 1 percent level, statistically significant at the 5 percent level, * statistically ** significant at the 10 percent level. Standard errors are clustered at a country level. 24 Table 3: Average agreement with life situation statements. 2016 and 2023. LiTS Round 2016 2023 Statement Mean Std. Error Mean Std. Error The political situation in [COUNTRY] is better today than four years ago 2.571 (1.161) 2.633*** (1.323) I have done better in life than my parents 3.223 (1.138) 3.335 *** (1.191) My household lives better nowadays than four years ago 2.842 (1.136) 2.991*** (1.260) All things considered, I am satisfied with my life now 3.254 (1.119) 3.484 *** (1.142) Children who are born now will have a better life than my generation 3.203 (1.177) 3.292*** (1.260) All things considered, I am satisfied with my job 3.329 (1.122) 3.883 *** (0.973) Note: *** statistically significant at the 1 percent level. 25 Table 4: Willingness to pay to combat climate change and right−wing populism and economic conservatism indexes. 2016 and 2023. Right-wing populism index Economic conservatism index 2016 2023 2016 2023 Coeff. Std. Err Coeff. Std. Err Coeff. Std. Err Coeff. Std. Err Right-wing populism index -0.211*** 0.018 -0.196*** 0.015 0.003 0.003 -0.003 0.004 Female -0.016* 0.009 -0.008 0.007 -0.006 0.008 0.002 0.007 Age groups 20-29 years old Reference category 30-39 years old 0.008 0.012 -0.022 0.014 0.000 0.015 -0.026* 0.015 40-49 years old 0.015 0.013 -0.024 0.016 0.017 0.014 -0.033** 0.016 50-59 years old 0.005 0.013 -0.016 0.015 0.004 0.015 -0.024 0.016 60-69 years old -0.010 0.017 -0.024 0.016 -0.019 0.016 -0.034** 0.016 70+ years old -0.015 0.018 -0.013 0.016 -0.030 0.019 -0.031** 0.016 Education Primary Reference category Lower secondary 0.001 0.016 -0.031 0.018 * -0.003 0.015 -0.010 0.018 (Upper) secondary 0.037** 0.017 -0.014 0.016 0.048*** 0.014 -0.000 0.013 Post-secondary non-tertiary 0.061 *** 0.021 0.020 0.023 0.063 *** 0.019 0.028 0.023 Tertiary (not a university diploma) 0.090*** 0.020 0.032 0.026 0.094*** 0.018 0.036* 0.021 Bachelor’s degree or more 0.080*** 0.025 0.026 0.020 0.078*** 0.026 0.034** 0.017 Master’s degree or PhD 0.091*** 0.023 0.078*** 0.022 0.096*** 0.019 0.089*** 0.021 Marital status Single (never married) Reference category Married -0.002 0.014 0.004 0.013 -0.008 0.015 0.005 0.013 Widowed -0.035 0.016 ** -0.031 0.016 * -0.033 0.016 ** -0.029* 0.017 Divorced -0.038** 0.019 -0.009 0.013 -0.034* 0.019 -0.013 0.014 Separated -0.020 0.024 0.079 0.031 ** -0.030 0.024 0.090** 0.036 Religion Atheistic/agnostic/none Reference category Buddhist 0.131 *** 0.021 0.119 *** 0.020 0.109*** 0.022 0.092*** 0.017 Jewish -0.087 0.062 0.339** 0.135 -0.073 0.071 0.401*** 0.083 Christian -0.021 0.020 -0.006 0.023 -0.017 0.018 -0.001 0.024 Muslim 0.072 0.039 * 0.025 0.036 0.068 0.045 0.031 0.038 Other 0.006 0.040 -0.048 0.046 0.017 0.040 -0.014 0.034 Household size 0.002 0.005 0.006 0.005 0.005 0.005 0.005 0.005 Urban -0.014 0.011 -0.012 0.012 -0.003 0.012 -0.014 0.012 Income ladder 0.025*** 0.004 0.013*** 0.004 0.028*** 0.004 0.015*** 0.004 Intercept 0.330*** 0.054 0.616*** 0.039 0.122** 0.053 0.422*** 0.038 Number of observations 20,855 21,659 22,499 23,347 Note: *** statistically significant at the 1 percent level, ** statistically significant at the 5 percent level, * statistically significant at the 10 percent level. Standard errors are clustered at a country level. 26 Table 5: Environmental indicators at the PSU level. 2016 and 2023. Direct comparisons 2016 Change 2022 Precipitation Mean 0.004 = 0.004 Standard deviation 0.002 = 0.002 Range 0.007 = 0.006 Temperature Mean 0.296*** < 0.398 Standard deviation 0.378*** < 0.436 Range 1.098*** < 1.256 Air pollution (PM2.5 per m3) 0.411*** > 0.189 Flood severity index 0.002*** > 0.001 Regression results Coefficient Std. Error Coefficient Std. Error Precipitation Mean 0.039 1.363 0.975 1.559 Standard deviation −1.251 3.438 1.372 3.793 Range −0.287 0.958 0.364 1.171 Temperature Mean 0.021 0.029 0.008 0.019 Standard deviation −0.003 0.018 0.011 0.013 Range 0.003 0.005 −0.001 0.006 Air pollution (PM2.5 per m ) 3 −0.021 0.023 0.015 0.052 Flood severity index −5.570** 2.601 −0.980 6.369 Note: In the lower panel other controls are the same as in Table 4. *** statistically significant at the 1 percent level, ** statistically significant at the 5 percent level, statistically significant at the 10 percent level. Standard errors are * clustered at the country level. 27 Table 6: Oaxaca-Blinder decomposition of changes in proportions of respondents willing to pay higher taxes to combat climate change for different hypotheses. Baseline Specification 2 Specification 3 Specification 4 Coeff. Std. Err Coeff. Std. Err Coeff. Std. Err Coeff. Std. Err Overall Sample 2016 0.368*** 0.028 0.381*** 0.028 0.376*** 0.027 0.351*** 0.028 Sample 2023 0.316 *** 0.022 0.315 *** 0.022 0.322 *** 0.022 0.314*** 0.022 Difference 0.053 0.021 ** 0.066 *** 0.022 0.054 0.021 ** 0.037 0.024 Endowments −0.005 0.006 0.014 0.008 −0.002 0.007 0.036 0.031 Coefficients 0.062*** 0.021 0.062*** 0.020 0.067*** 0.021 0.054 0.035 Interaction −0.005 0.005 −0.010 0.006 −0.010 0.006 −0.052 0.039 Endowments Age −0.000 0.000 −0.000 0.000 −0.000 0.000 −0.000 0.001 Education 0.000 0.003 0.003 0.003 0.002 0.003 0.002 0.003 Marital status −0.001 0.001 −0.001 0.001 −0.001 0.001 −0.001 0.001 Religion −0.003 0.004 −0.002 0.005 −0.002 0.004 −0.006 0.005 Household size −0.001 0.001 −0.000 0.002 −0.000 0.001 −0.002 0.002 Monthly income −0.000 0.000 −0.000 0.001 −0.000 0.001 0.000 0.001 Urban 0.000 0.000 −0.000 0.000 0.000 0.000 −0.001 0.001 Right−wing populism index 0.015*** 0.004 Economic conservatism index −0.000 0.000 Environmental risk indicators 0.044 0.033 Coefficients Age 0.131** 0.063 0.110* 0.064 0.136** 0.060 0.134* 0.071 Education 0.004 0.030 0.013 0.032 0.002 0.031 −0.013 0.029 Marital status 0.001 0.013 0.006 0.014 −0.001 0.013 −0.011 0.015 Religion −0.003 0.022 0.007 0.024 −0.002 0.023 −0.006 0.024 Household size 0.020 0.020 0.003 0.020 0.020 0.021 0.018 0.021 Monthly income −0.001 0.001 0.008 0.006 0.010 0.007 −0.001 0.001 Urban 0.014 0.011 0.004 0.009 0.017 0.012 0.023* 0.012 Right−wing populism index 0.005 0.021 Economic conservatism 0.010 0.007 Environmental risk indicators 0.027 0.023 Intercept −0.103 0.076 −0.095 0.082 −0.125 0.071 * −0.117 0.088 Interaction Age 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.001 Education −0.007* 0.004 −0.007* 0.004 −0.007* 0.004 −0.006 0.004 Marital status 0.000 0.001 0.001 0.001 0.000 0.001 −0.001 0.001 Religion 0.001 0.002 0.001 0.002 0.001 0.002 0.004 0.003 Household size −0.000 0.001 −0.000 0.000 −0.000 0.001 −0.001 0.001 Monthly income 0.000 0.000 −0.004 0.003 −0.005 0.004 −0.000 0.001 Urban −0.000 0.001 0.000 0.000 −0.000 0.001 0.002 0.002 Right-wing populism index −0.000 0.001 Economic conservatism 0.001 0.001 Environmental risk indicators −0.050 0.039 Note: *** statistically significant at the 1 percent level, ** statistically significant at the 5 percent level, * statistically significant at the 10 percent level. Standard errors are clustered at a country level. 28