How Do Right-Wing Populist Parties Influence Climate and Renewable Energy Policies? Evidence from OECD Countries

Announcing the repeal of his predecessor’s Clean Power Plan at a rally in September 2017, President Trump declared, “Did you see what I did to that? Boom, gone!”¹ This move exemplified what many climate policy advocates feared: a populist politician and declared climate skeptic reversing policies brought in under a previous administration.

But how widespread is such action? Does the rise of authoritarian nationalist populists (sometimes labeled “right-wing” populists) and their entry into government always have a negative effect on climate and low-carbon energy policies? If there are differences in how far populists getting into power affects such policies, what factors can explain the variation? These questions matter not just because of the direct effects on domestic outcomes but also because of their influence on the policies and emissions of other countries via the erosion of international cooperation (Sælen et al. 2020).

Interest in the links between populism and climate change has emerged in the last few years (for a recent review, see Forchtner 2019). Within this literature are a number of recent studies looking specifically at how right-wing populist parties (RWPPs) might actually affect climate and sustainable energy policies and outcomes once elected to legislatures and governments (Böhmelt 2021; Ćetković and Hagemann 2020; Huber et al. 2021; Jahn 2021). These studies show some influence of RWPPs, but with variation across countries and policy areas.

This article adds to the literature by taking a quantitative approach to measuring policy effects and widening the focus beyond European countries, on which much of the literature so far has focused. We assess the impact of RWPPs on climate and renewable energy policy in thirty-one OECD countries over the period from 2007 to 2018, combining data on the quality of policies with established data sets on right-wing populism and on parliaments and governments. This scope means we cover a group of postindustrial countries with a shared social and political context for the emergence of authoritarian populism, while at the same time going beyond the European focus of existing studies, allowing us to assess the role of electoral systems and European Union (EU) membership. We capture both the direct effects of RWPPs as part of governing cabinets and leadership and indirect effects through their representation in legislatures on other parties in government. Our key findings, which are robust to a set of other political, economic and environmental controls, are, first, that RWPPs do have a significant negative impact on climate policy, but not on renewable energy policy; second, that the impact of RWPPs on climate policy is mitigated by proportional representation (PR) and by membership of the EU; and third, that climate policy effects of RWPPs in majoritarian countries work via both executive and legislative channels.

Our results are broadly in line with, and provide independent verification of, findings in the wider literature. The muted effects of right-wing populism on climate policy in countries with PR and coalition government is consonant with Ćetković and Hagemann (2020), and the stronger effects on climate policy than on renewables policy is similar to Huber et al. (2021). Quantitative studies of the effects of populism have so far looked at outcomes rather than policies, and our research helps clarify intermediating mechanisms. Our results suggest that climate policy change can explain at least some of the links found between right-wing populism and greenhouse gas (GHG) emissions (Jahn 2021) and between populist leadership and per capita carbon dioxide emissions (Böhmelt 2021).

Our study has some limits. Unlike some recent studies (e.g., Huber et al. 2021), we do not include consideration of left-wing populist parties. We also do not attempt to differentiate between different types of RWPPs (Zulianello 2020), mainly because of the nature of the data set on party characteristics on which we draw.

In section 2, we review the existing literature on right-wing populism and climate and renewable energy policy. Section 3 describes our data and methodological approach. In section 4, we present the findings of the analysis. Section 5 concludes with a discussion of the wider implications of the analysis.

There is a long-standing literature examining the determinants of climate policy, including local air pollution, high-carbon interests, knowledge of climate change, levels of education, and the left–right positions of governments (Dolšak 2009; Fankhauser et al. 2015; Karapin 2016; Steves and Teytelboym 2013). There is also an established literature on the determinants of renewable energy policy suggesting positive roles for energy security concerns, EU membership, and renewable resources but a negative role for the strength of high-carbon interests (Cheon and Urpelainen 2013; Jenner et al. 2012; Marques et al. 2010). Schaffer and Bernauer (2014) find that PR electoral rules and federalism are positively associated with more ambitious renewable energy policy.

By comparison, the study of the relationships between populism, on one hand, and attitudes to climate and renewable energy policy, on the other, is relatively recent (Forchtner 2019). There is, of course, a large literature on populism, which is often taken to involve a basic cleavage in society between the “people” and a corrupt “elite,” and a belief that politics should be an expression of the will of the people (Mudde 2004, 2007). However, populism is also usually understood as a “thin” ideology that always comes combined with values from other political ideologies, including on questions of distributive conflicts, social values, and identity (e.g., Canovan 2001). These elements do not necessarily combine in ways that are consistent with those other ideologies; for example, what is labeled “right-wing populism” or “radical right populism” typically includes positions on state intervention in the economy that would be seen as left wing on conventional measures, but often combined with nativism and social authoritarianism.

Fraune and Knodt (2018) and Lockwood (2018) draw attention to a general tendency for RWPPs and individuals supporting them to express greater climate skepticism and, in some cases, hostility to policies supporting renewable energy, while also favoring the use of fossil fuels. Lockwood (2018) explores two potential reasons for these positions: that supporters of RWPPs tend to be those “left behind” by globalization and technical change and resentful at paying for climate policy through forms of environmental taxation and that right-wing populists have an ideological hostility to climate policy as an essentially cosmopolitan agenda. Much subsequent research has focused on the connection between support for RWPPs and climate skepticism at the level of the individual citizen (e.g., Huber 2020; Jylhä et al. 2020; Huber et al. 2021; Kulin et al. 2021; Lachapelle and Kiss 2019) and on populist party policy platforms on, and wider rhetoric toward, climate change (Forchtner et al. 2018; Hess and Renner 2019; Huber et al. 2020; Żuk and Szulecki 2020).

However, there has been less focus on the effects of populism on climate and energy policies and outcomes. Ćetković and Hagemann (2020) examine six Western European countries over the period from 2008 to 2018, using a case study approach. Among these countries, they find only limited effects of the rise of RWPPs, for a number of reasons. One is that such parties entered governments in only relatively few cases. A second is that even where RWPPs were in cabinets, in only one case (Norway) did they directly control the energy and climate ministry, reflecting the fact that climate change is often not the main concern of RWPPs (Lockwood 2018). Third, when RWPPs had electoral success as measured by seats in legislatures, this tended to push larger parties to form coalition governments with other parties that had progressive energy and climate platforms, leading to an improvement in policies. Finally, they find that the potential effects of RWPPs were conditional on the absence of a strong international climate regime and overwhelmed by the influence of major events like the Paris Agreement.

Huber et al. (2021) also adopt a case study approach to assess the role of populists in power in six European countries, specifically the actions of RWPPs in government in Austria and Poland. They find clearer opposition to climate commitments than to renewable energy policy, in line with Lockwood’s (2018) suggestion that RWPPs in some countries may be more ambivalent about the latter. One reason for RWPPs embracing renewables may be because a nationalist ideology heightens concerns about energy security, and in countries without domestic fossil fuel reserves, renewable energy provides a route to security. Some RWPPs also support some forms of renewable energy while rejecting others, such as France’s Rassemblement national, which endorses solar PV but is opposed to wind, which may be related to populist right ideas about the national landscape (Forchtner and Kølvraa 2015).

Finally, Jahn (2021) (looking at the EU28) and Böhmelt (2021) (looking at a wider sample of sixty-six countries) adopt a quantitative approach to assessing the effects of populism not on policies but on GHG emissions as outcomes, finding that these are significantly higher where RWPPs are in government.

In this article, we explore the influence of RWPPs on climate and renewable energy policies using a quantitative approach rather than the case study one used in the literature so far. Like much of the literature, we are mainly concerned with the effects of populist parties on climate and renewable energy policies, as opposed to leaders, as in Böhmelt (2021). This is principally because populist parties may have an effect on policies even if they do not lead governments, either through representation in parliament, representing an electoral threat to other parties, or as partners in coalition governments.

For climate policy, our sample includes thirty-one Organisation for Economic Co-operation and Development (OECD) countries for the period 2007–2018, while that for renewable energy policy includes twenty-five OECD countries for 2010–2018. Because our data allow us to expand the frame beyond Western Europe and compare countries with different electoral systems, our main focus is on assessing whether the influence of RWPPs on policy shows systematic variation across these dimensions. At the same time, unlike Böhmelt (2021), we focus on a set of OECD countries that share a common context for the rise of right-wing populism, that is, industrial decline and the emergence of new political cleavages (Iversen and Soskice 2019; Kriesi et al. 2006; Oesch 2015).

Drawing on this discussion of the literature, we propose the following basic hypotheses:

The existence of climate and renewable energy targets at the supranational level for EU member states implies that we might see the influence of RWPPs muted in EU countries compared with non-EU countries. While RWPPs in theory have some influence on these targets through co-decision mechanisms involving the European Parliament and the Council of Ministers, in practice, these institutions are still dominated by representatives from mainstream, nonpopulist parties, and this is reflected in the broad pro-climate action orientation of the EU. This observation leads to our next hypothesis:

We also expect differences across countries with PR and majoritarian electoral systems. In countries with PR systems, the representation of right-wing populist voters is likely to happen directly through the formation of RWPPs; such parties have a greater chance of entering legislatures, and so the incentive to form and vote for such parties is greater. This in turn suggests that where RWPPs enter government, they will do so as (typically junior) coalition partners. However, as Lockwood (2018) and Ćetković and Hagemann (2020) suggested, we expect that climate policy and renewable energy portfolios will not be a priority for RWPPs entering cabinets, and so again the policy influence relationship will be more muted.

By contrast, in countries with plurality and majoritarian electoral systems, which tend to lead to a few (often two) large parties and majority governments, we expect right-wing populists to enter government via an internal capture of the existing center-right party, in what Snow and Moffitt (2012) call “mainstream populism.” While such cases may be rare, when they do occur, we expect them to have a greater effect on all policy areas, including climate and renewables policy, since populists effectively capture the whole of government. The recent Trump presidency and Republican administration provide an easily recognizable example.² This yields a further hypothesis:

Finally, we consider how different channels of influence might operate. A general expectation might be that RWPPs have more influence when they have representation in the executive rather than just in the legislature. However, Ćetković and Hagemann’s (2020) findings suggest that both routes of influence are possibly weaker in countries with PR electoral systems that have a tendency toward coalition governments.³ As earlier, we expect that in majoritarian countries, it is rarer for populists to gain representation at the executive level, but when they do so, they have control over the whole of government and so can be expected to have more of an effect on policy. So, our final hypothesis is as follows:

This study adopts a quantitative approach to assessing the influence of RWPPs on climate and renewables policies, using multiple regression analysis on two panel data sets.

To assess the strength of climate policy, we draw on the Climate Change Performance Index (CCPI)⁴ published annually by Germanwatch, the New Climate Institute, and the Climate Action Network, also used by Ćetković and Hagemann (2020). For this study, we have data from 2007 to 2018 on thirty-one OECD countries. The CCPI is constructed from scores across four categories: GHG emissions, energy use, renewable energy, and climate policy. Here we focus on the climate policy score (CPS).

The CPS is based on an annual rating of domestic and international climate policy commitments and performance by “climate and energy policy experts from non-governmental organizations, universities and think tanks within the countries that are evaluated.”⁵ This data source has the limitation that it is based on a set of subjective assessments of policy, albeit from a number of experts. Comparability over time is also affected by the fact that the expert pool providing the data has been extended and altered over time. However, for the countries in our sample, there are multiple experts, which should improve the accuracy of the overall assessments. Moreover, if personal biases in expert assessments are not time varying at the country level, they will be captured by our use of fixed effects in our estimation.⁶

We do not draw on the CCPI for renewable energy because its indicator is largely a measure of growth in renewable energy capacity rather than of policy. Instead, we use a renewable energy policy score (RES) from the World Bank Regulatory Indicators for Sustainable Energy (RISE) database, which provides data from 2010 onward.⁷ The RISE database includes twenty-five countries out of the thirty-one in our sample covered by the CCPI, those excluded being the smaller states of Estonia, Iceland, Latvia, Lithuania, Luxembourg, and Slovenia. The RES is based on expert assessments of seven dimensions of renewable energy policy and regulation.⁸ So, the remarks made earlier about data based on expert evaluations for the CPS indicator also apply here. For comparability of results, we have rescaled both CPS and RES to lie between 0 and 10, with a higher score indicating a more pro-climate or pro–renewable energy policy, respectively.

Data on party representation among leaders, cabinets, and legislatures (the lower houses of representatives or parliaments in bicameral systems) is taken from the Parliaments and Governments Database.⁹ To identify which political parties can be counted as right-wing and populist (i.e., as RWPPs), we rely on two sources. For Europe, we use the PopuList project,¹⁰ established by leading experts on populism and the radical right. As PopuList does not cover non-European countries, for these, we have adopted our own coding scheme based on published studies of the nature of parties and governments in these countries, as detailed in the  appendix.¹¹

There is a timing issue involved in combining the political data with the policy scores and other data. New cabinets and legislatures are formed by elections that fall somewhere within a particular year, and reflecting this, the Parliaments and Governments Database is organized by cabinet, not year. The rest of our data, including from the CPS and RES, are annual. Specifically, the CPS for a particular year is generated by assessments of experts collected in the latter part of that year (September to November), while the RES indicator is produced in December of each year. Our approach to matching data is to allocate the start and end years of governments and legislatures in the following way: if an election falls in the first half of a calendar year (e.g., May 2014), then we count the government and legislature as starting from that year (i.e., 2014), but if it falls in the second half of the calendar year (e.g., September 2014), then we count these as starting from the next calendar year (i.e., 2015). This is because the CPS and RES scores will tend to reflect the effects of political circumstances over the first part of the year, in part because it often takes a little time for new governments to change policies.

For the direct route of potential influence, we construct two variables, rwp_pm and rwp_cabinet, using the Parliaments and Governments Database. The first is 1 if the leader (i.e., prime minister or president) of government comes from a RWPP, and 0 otherwise. The second is defined as the number of RWPPs that hold cabinet posts divided by the total number of parties that hold cabinet posts.¹² In the analysis, we use the average of these two variables, which we call rwp_exec, as a measure of RWPP control of the executive. This variable takes on values between 0 and 1. The indirect route of potential influence is measured by rwp_leg, the share of seats in the legislature taken by RWPPs. In the initial analysis, we use the simple average of rwp_pm, rwp_cabinet, and rwp_leg which we call rwp, as an overall measure of the political influence of RWPPs.¹³

In addition, we adopt a number of controls, which we classify as political, economic, and environmental. It is first important to control for the conventional left–right position of the government (Fankhauser et al. 2015; Lutz 2019). For consistency, we again use data from the Parliaments and Government Database, which scores parties on a 0 to 10 scale, from left to right. We define parties with a score of 4 or less as left wing and parties with between 4 and 6 as center, which ensures that roughly equal numbers of parties are in each of the three categories. Then we create variables left_pm, left_cabinet, left_leg, centre_pm, centre_cabinet, and centre_leg in exactly the same way as for RWPPs; for example, left_pm is 1 if the head of government is from a left-wing party. We again define left and centre to be the averages of the three variables; these capture the potential influence of left-wing and center parties, respectively. All political effects, by definition, will thus be measured relative to the baseline of non-populist right-wing parties.

Second, we include a dummy for EU membership and a measure of how the majoritarian the electoral rules of the country are. The measure of majoritarianism is constructed from the World Bank’s widely used Database of Political Institutions.¹⁴ Specifically, following Muttakin et al. (2021), we define our measure maj = 1 − Proportional representation + Plurality + Housesys, where Proportional representation is an indicator variable that takes on the value 1 if some candidates are elected based on a percentage of votes received by their party, and 0 otherwise; Plurality is an indicator variable that takes on the value 1 if legislators are elected through a majoritarian rule, and 0 otherwise; and Housesys takes a value 1 if most seats in the legislature are filled via plurality rule, and 0 if most are filled via proportional rule. Overall, maj can take on integer values between 0 and 3.

For economic controls, we include gross domestic product (GDP) per capita (gdp pc) and the unemployment rate (u). These variables are completely standard: the role of the unemployment rate is to pick up any effects of the economic cycle.

Finally, we include a number of environmental controls, selected on the basis of the existing literature on determinants of climate and renewable energy policy cited in section 2. We have a relatively short time dimension to the panel because our CPS and RES variables are not available before 2007, so we focus on controls that do not require us to drop observations.

For the analysis of climate policy, we use an index of local air pollution per capita (the sum of NO_x and SO_x emissions per capita, denoted lap pc), which captures possible demand for environmental improvements;¹⁵ carbon dioxide emissions per unit of GDP, denoted CO2 gdp, as an indicator of the strength of high-carbon interests, along with fuel exports as a percentage of total merchandise exports, denoted fuel exports; and finally, the proportion of twenty-five- to thirty-four-year-olds with tertiary education, denoted tertiary ed, because awareness of and support for climate policy have been shown to be associated with higher levels of education.

For the analysis of renewable energy policy, we include: u, gdp pc, tertiary ed and a measure of the share of fossil fuels in electricity generation, fossil elec, as an indicator of the strength of high-carbon interests in the electricity sector. Sources for all these data are given in the  appendix.

Summary statistics for all variables and details on their units of measurement are reported in Table 1. Out of 372 observations for the CPS sample, each representing a country and a year, RWPPs were in cabinets in 65. In twenty-seven cases, national political leaders were from RWPPs.

Figures 1 and 2 show the evolution of mean CPS and RES scores for all countries in our data set over the period, respectively. The CPS mean shows no particular trend, but that of the RES shows clear upward movement throughout the period. A standard test for stationarity shows that for both CPS and RES, we can reject the null hypothesis of a unit root in favor of stationarity.¹⁶ However, there is evidence of considerable persistence in both variables, a point to which we return later.

Figure 3 shows the basic relationship between CPS and rwp and between RES and rwp, respectively. Figure 3a suggests a possible negative relationship between CPS and rwp. We show that this relationship becomes much clearer when we control for country fixed effects and other political, economic, and environmental factors. From Figure 3b, the picture for RES is less clear.

As a next step, we look graphically at how the strength of the relationship between CPS and RES, on one hand, and rwp, on the other, is mediated by how majoritarian the electoral system is and whether the country is an EU member.

To do this, using a simple regression, we first calculate the marginal effect of rwp on CPS for different values of maj and eu. In the regression, we lag rwp by one year because, as explained earlier, there are lags in the formation of policy. We also use this lag specification in the more detailed regressions we discuss later. The plots in Figures 4 and 5 show marginal effects plus their confidence intervals. Figure 4a shows a clear relationship between maj and the marginal effect of lagged rwp on CPS; for values of maj greater than 1, the impact of maj is significantly negative. This is suggestive evidence for H4, that RWPPs have more impact on climate change policy in majoritarian electoral systems. A somewhat weaker positive effect of EU membership on the relationship between lagged rwp and CPS is shown in Figure 4b.

Figure 5 repeats this exercise for RES. Figure 5a shows that there is a relationship between the effect of lagged rwp on RES and the degree of majoritarianism of the electoral system. The effect is nonmonotonic, but it does suggest that in pure majoritarian systems (i.e., maj = 3), rwp has a negative impact on RES. Figure 5b shows that there is possibly a weak negative relationship between the effect of lagged rwp on RES and EU membership.

As already noted, the unit root tests indicate considerable persistence in both CPS and RES, so any regression that omits a lagged dependent variable will be misspecified. Instead, a generalized method of moments (GMM) estimator should be used. However, it is also known that in data with a small cross section, as we have here, GMM estimators can be severely biased and imprecise (Bruno 2005). We therefore use an estimator to deal with this problem that has been developed by Bruno and can be implemented in Stata (xtlsdvc). At the first stage, this estimator implements the Arellano–Bond estimator, which first-differences the regression to eliminate country fixed effects, and then instruments the lagged change in the dependent variable with further lags. At a second stage, xtlsdvc corrects the bias in the Arellano–Bond estimator.¹⁷

The main regression results for the CPS indicator are shown in Table 2. As already noted, all political variables are lagged by one year to allow for the policy-making process. Specification 1 (reported in column 1) is the basic regression with just the three political variables. Column 1 shows, as we might expect, that both center and left parties have a positive effect on the CPS score relative to the right-wing baseline. To interpret the coefficients, note first that the mean CPS score is approximately 5, so the effect of center (left-of-center) parties is to increase the CPS score by about 22 percent (16 percent) relative to the mean. The coefficient on rwp is negative and significant, consistent with H1, that RWPPs have a negative effect on climate policy. It indicates that a RWPP implies an approximate 24 percent reduction in the CPS index relative to the mean, which is comparable in magnitude, though opposite in sign, to the effect of center and left parties.

Specification 2 allows for the effect of right-wing populism on CPS to vary by electoral system. To do this, we create an interaction term rwp_maj, which is equal to rwp multiplied by maj when maj is greater than 1, and zero otherwise. The reason for this specification of the interaction term is that, as can be seen in Figure 4a, the marginal effect of maj on the relationship with rwp_exec is significant only when maj is greater than 1. The interaction term rwp_maj is significantly negative, consistent with H4. The effect is also large; in strongly majoritarian systems when both the head of government and all the cabinet posts are held by RWPPs, the CPS score falls by 58 percent of the mean.

On the other hand, for countries that score 0 or 1 on maj, the effect of rwp is insignificant. This latter finding is consistent with our hypothesis that in countries with PR, where RWPPs enter government, they will do so typically as junior coalition partners with limited numbers of cabinet seats. In such circumstances, we can expect them to prioritize portfolios other than climate and renewable energy policy, given the greater salience to date of issues like immigration for such parties. Supporting evidence for this interpretation comes from data on cabinet portfolios for European countries going back to 1993 from the Party Systems and Governance Observatory.¹⁸ This shows that of the forty-three cabinets containing RWPPs for which data are available, RWPP representatives held the environment portfolio in only nine cases, five of which were from Poland, where the RWPP Prawo i Sprawiedliwość (PiS) has been the largest single party.

Specification 3 allows for the effect of right-wing populism on CPS to vary by EU membership, reflecting the potential constraints of EU policy. The coefficient on the interaction of rwp and EU membership, variable rwp_eu, is significantly positive, consistent with H3. Inspection of the coefficients on rwp and rwp_eu suggests that the two effects more or less offset each other, and in fact, in both regressions 3 and 7, we cannot reject the hypothesis that the sum of the two coefficients is zero at a 5 percent significance level. The implication is that while RWPPs outside the EU have a strong negative effect on climate policy, RWPPs within the EU have no significant effect. Finally, in both regressions 2 and 3, center and left-wing governments have a clear positive impact on CPS, as might be expected.

We then introduce controls, which are generally insignificant in specifications 5, 6, and 7; this is probably because they do not add much to the explanatory power of country fixed effects. To check this, specification 4 runs a static version of specification 5 with all economic and environmental controls added, but with country fixed effects and the lagged dependent variable omitted. Several controls then become significant, and the signs are mostly in line with the existing literature. For example, unemployment reduces the quality of climate policy, as other policy priorities become more important during recessions; local air pollution—measuring a citizen demand for environmental improvement—has a positive impact; and finally, CO₂ per unit of GDP, measuring producer resistance to decarbonization, has a negative effect. However, per capita GDP and tertiary education have no significant effect, and fuel exports are significant, but the sign is positive rather than the expected negative effect. Finally, the main results from columns 1–3 concerning the effects of political variables are robust to the introduction of controls in columns 5–7; in fact, the coefficients hardly change in size or significance level.

The main regression results for the RES indicator are shown in Table 3. The regression specifications 1–9 are the same as for CPS. A first observation is that RES is highly persistent, with a coefficient on the lagged dependent variable of over 0.9. Looking across specifications 1–3, we see first that all political variables, including rwp, are insignificant, except left, which has a positive sign. As the mean value of RES is approximately 6.6, the coefficient of approximately 0.47 on left implies that left-wing parties increase the RES score by approximately 7 percent on average, which is about half the effect of left on the mean value of the CPS score. Overall, these results are thus consistent with H2—in particular, we find no effect of right-wing populism on renewables policy.

As in Table 2, specification 4 is a static regression without fixed effects or lagged dependent variable; surprisingly, our control variables, which are standard, are not significant here, so variables like GDP per capita do not appear to explain cross-country variation in RES. This finding may reflect the fact that renewable energy policy has converged across countries more in recent years, compared with data used in earlier studies. In specifications 5–7, which repeat specifications 1–3 with the addition of controls, only left remains highly significant.

So far, we have considered the aggregate influence of RWPPs on policy; we now examine if the executive or legislative channel is more important, addressing H5. Although one might assume the executive channel may be more important, it is not immediately obvious which one dominates. For example, the executive channel may be weak if RWPPs are given cabinet portfolios unrelated to the environment, and the legislative channel may be important if the seat share reflects high levels of populist support that push mainstream parties toward adopting populist policies.

Tables 4 and 5 show the results for the CPS and RES indexes, respectively. In each table, we look at the effects of the two different variables rwp_exec and rwp_leg separately, both in levels and via the interactions with maj and eu. We do not enter both variables in a “two-horse race,” as they are highly correlated; the correlation coefficient between rwp_exec and rwp_leg is 0.78. The resulting co-linearity leads to political variables being insignificant in such regressions. Also, in each regression, we include left and centre but omit the controls, since the signs and significance of the political variables of interest do not change if we add the controls.

For CPS, the overall effects of RWPPs on policy appear to be stronger through the legislative than the executive route; that is, the coefficient on rwp_exec in specification 1 is only significant at 10 percent, whereas the coefficient on rwp_leg in column 4 is significant at 5 percent and is three times as large. If we look at the interactions between these two channels and maj, we see that the effect of RWPPs in both legislatures and executives is stronger in majoritarian countries, but the relative effect via the executive is certainly no bigger. So, overall, we do not find any evidence in favor of H5 for climate policy. Finally, for renewables, we would expect from Table 3 that RWPPs do not affect RES either though the executive or the legislative channel, and the results in Table 5 confirm this.

In this article, we have investigated whether there is a systematic tendency for RWPPs to have a negative impact on climate and renewable energy policies across OECD countries. Specifically, we have assessed what happens to these variables when representatives of RWPPs are elected to legislatures, enter government, or become leaders.

Our analysis shows that there is a robustly significant negative relationship between the strength of RWPP representation in both in the legislature and executive and climate policy. However, we also find that this relationship is mitigated by EU membership and that the negative effect of RWPPs on climate policy is far more pronounced in countries with strongly majoritarian electoral systems, relative to that in countries with more proportionately representative systems.

These findings are consistent with our hypothesis that in countries with PR, where RWPPs enter government, they will do so as typically junior coalition partners with limited numbers of cabinet seats and a tendency not to prioritize portfolios relevant to climate policy. By contrast, in countries with majoritarian electoral systems, when RWPPs get into government, our finding is that they have a much greater influence on climate policy. However, these situations are relatively unusual; in our sample, only a few countries have strongly majoritarian systems, and the episodes in which right-wing populists were in power in these countries are few and far between.

In contrast to the finding of a robust relationship between RWPP representation and climate policy, there is no strong and significant overall relationship with renewable energy policy. Consistent with other studies, this would seem to suggest that RWPPs are more ambivalent about renewable energy than they are about climate policy or that they have limited influence over renewables policy, especially in countries with more PR-based electoral systems.

Our analysis has various implications. First, when right-wing populists come into power, they can be expected to be disruptive of climate policy. This finding applies not only at the domestic level but also at the international level, since our measure of climate policy comprises international effort as well as national policy (an obvious example is Donald Trump withdrawing the United States from the Paris Agreement in 2017).

Second, it is in countries with majoritarian politics outside Europe that climate policy can be the most vulnerable to the influence of right-wing “mainstream” populists coming to power. However, unlike the case in PR systems, where permanently populist parties tend to form, mainstream populism has itself been a more unstable phenomenon, so the challenge to climate policy from such cases may be episodic.

The third implication relates mainly to European countries with PR and coalition governments. Many of these countries have seen a long-term rise in RWPP representation, but in most cases, such parties are still in a minority position. If the fortunes of RWPPs continue to rise, it is possible that the mitigating effect of PR electoral systems on the relationship with climate policy will weaken or disappear. Larger, stronger RWPPs in governments can affect more policy areas; this is clear from countries like Poland. At the same time, it is far from clear that RWPPs in Europe have reached their high-water mark.

A fourth implication arises from the often-made observation that populism is a reactive ideology (Canovan 1999) that focuses on perceived crises (Taggart 2000). Thus far, the main focus of European RWPPs has been immigration (and in the United Kingdom, Brexit), but as mainstream and left-wing parties focus increasingly on climate change, this focus may shift, leading RWPPs to put greater effort into trying to control environmental portfolios and shaping climate policy where they get into coalition governments.

Finally, a fifth implication follows from the absence of a strong and significant effect of right-wing populism on renewable energy policy. This offers the prospect that despite hostility to a broad climate agenda, right-wing populists in power may still support some technologies that reduce emissions, helping in turn to lower the costs of mitigation.

Sources for data not given in the text are as follows:

• gdp_pc, u_rate from https://data.oecd.org/, last accessed February 23, 2022;

• lap pc, CO2 gdp from https://stats.oecd.org/Index.aspx?DataSetCode=AIR_EMISSIONS, last accessed February 23, 2022;

• fuel exports from https://data.worldbank.org/indicator, last accessed February 23, 2022;

• fossil share elec from https://www.oecd-ilibrary.org/energy/data/iea-electricity-information-statistics_elect-data-en, last accessed February 23, 2022;

• tertiary ed share from https://data.oecd.org/eduatt, last accessed February 23, 2022.

Most of the countries outside Europe in this study have majoritarian or plurality electoral systems that tend to work against the formation of separate significant RWPPs. Instead, these countries tend to see periods of “mainstream populism” (Snow and Moffitt 2012), during which populist factions and leaders within traditionally center-right parties gain control. The paradigmatic case of this is in the United States, with the rise of the Tea Party movement within the Republican Party over the 2010s and the emergence of Trump as a leader in 2017. For the purposes of the analysis here, we have adopted a coding scheme as in Table A.1, which is based on the following accounts of the nature of political parties in the relevant countries and periods:

• United States (Oliver and Rahn 2016)

• Australia (Snow and Moffitt 2012)

• New Zealand (Donovan 2020; Vowles and Curtin 2020)

• Canada (Snow and Moffitt 2012)

• Japan (Lind 2018; Warren 2019)