Far-right political support and clean technology adoption in Germany

Abstract The diffusion of clean technologies is critical for climate goals, yet political polarization influences adoption. Here we examine how ideology affects residential solar photovoltaic (PV) and electric vehicle (EV) uptake in Germany, drawing on responses from a three-wave individual-level survey, municipality-level observations of PV and EV registrations and election results. Supporters of the far-right Alternative für Deutschland (AfD) show significantly lower willingness to adopt EVs and, to a lesser degree, PV systems compared with supporters of the pro-environmental Bündnis 90/Die Grünen (GRÜNE) party. However, polarization is not static. The relative adoption gap between municipalities with strong AfD support and those with strong GRÜNE support declined from 76% to 31% for solar PV between 2000 and 2023, and from 84% to 60% for EVs between 2015 and 2019. Here, using an instrumental-variable strategy based on wolf attacks as a predictor of AfD vote share, we provide complementary evidence on the relationship between far-right voting and the adoption of clean technology. Counterfactual simulations indicate that, by 2023, solar PV diffusion could have been up to 1.3% higher and EV diffusion up to 2.9% higher in the absence of rising electoral support for the far right. Similar content being viewed by others Main The transition to a sustainable energy future necessitates a substantial reduction in fossil fuel use. Households contribute to over 60% of global greenhouse gas emissions through their direct and indirect consumption behaviour1, making the rapid scaling of clean technologies—such as solar photovoltaics (PVs) and electric vehicles (EVs)—a critical component of this effort2,3. Whereas early adoption was supported by subsidies4,5, durable longer-term impacts will probably hinge on changes in behaviours and preferences6,7. Non-monetary factors—such as ease of installation, aesthetics, charging infrastructure, peer effects and cultural acceptance—are critical for sustaining diffusion8,9,10,11,12,13. A key factor is the rise of political partisanship, which in some contexts has turned the adoption of clean technology into a ‘cultural war’, undermining public support14. A growing body of literature has found associations between political ideology and energy- and climate-related behaviours15,16, including the adoption of clean technologies17,18,19,20,21. From a social identity perspective, such behaviours can serve both to maintain consistency with political beliefs22 and to signal social and political affiliation16,23. However, despite a rapidly expanding empirical record, a fundamental limitation remains unresolved: the lack of credible identification strategies that isolate the effect of political ideology from correlated socio-economic and contextual factors. Political ideology is strongly intertwined with factors such as income24,25, education26, social status27, place28, urban–rural structure29 and housing characteristics30, many of which are difficult to observe comprehensively at fine spatial and temporal scales. As a result, existing estimates of ideological effects on clean technology adoption are probably biased by unobserved confounding. This identification challenge is compounded by several additional factors. Previous research often relies on self-reported adoption or stated preferences, which are prone to social desirability and recall biases and offer limited geographic and temporal resolution7. At the same time, studies based on observed adoption behaviour rarely incorporate complementary measures of individual-level intentions and beliefs, making it difficult to assess how stated preferences may align with observed diffusion patterns. Comparisons across different clean technologies are also scarce, despite substantial differences in their visibility, cost structures and socio-political salience. Finally, empirical evidence remains heavily concentrated in the USA16,17,23,31, raising questions about external validity across political systems and institutional contexts7. Taken together, this existing work leaves unresolved whether, and under what conditions, political ideology influences the diffusion of clean technologies. In this Article we address these research gaps through a three-step empirical approach. First, we analyse individual-level survey data (three waves; n = 2,356–3,092) to examine how political preferences are associated with clean technology adoption behaviour and intentions. Second, we characterize the geographic diffusion of solar PVs and EVs across German municipalities, quantifying partisan differences in observed adoption using a panel dataset with official registration records and voting behaviour at the municipality level (n = 10,733; 2000–2023 for PV and 2014–2023 for EV). Third, we examine whether changes in political support—particularly for the far-right Alternative für Deutschland (English: Alternative for Germany; hereafter referred to as AfD)—are associated with subsequent changes in clean technology adoption. To address endogeneity concerns, we use two complementary empirical strategies. First, we estimate panel regressions that relate within-municipality changes in political support to subsequent clean technology adoption, accounting for time-invariant municipal differences, national time trends and local policy or institutional variation through municipality and county-year fixed effects, as well as a rich set of time-varying controls. Second, we complement these models with an instrumental-variable strategy that uses exposure to wolf attacks on livestock as a predictor of AfD vote share. Previous research has linked wolf attacks to increases in electoral support for the far-right AfD party32, and we confirm a strong first-stage relationship in our setting. Across both approaches, we examine whether increases in far-right political support are associated with lower subsequent adoption of clean technologies. We focus on Germany, Europe’s largest economy, recognized both for its competitive automotive industry and its early adoption of residential solar PV at scale (Supplementary Note 1). Germany’s political system consists of a range of parties that span the political spectrum, with the strongest divide in terms of energy and climate-related policies between the pro-environmental Bündnis 90/Die Grünen (English: Alliance 90/The Greens; hereafter referred to as GRÜNE) party and the far-right AfD33 (Supplementary Note 2). Whereas the GRÜNE party advocates for aggressive decarbonization, the AfD opposes clean technology adoption34,35,36, denies human-driven climate change and claims that EVs are more harmful than combustion engines37,38. Having gained political traction in recent years, the AfD secured 10.3% of the vote in the 2021 federal election39 and 20.8% in 202540. In this Article we use vote shares for AfD and GRÜNE as empirical proxies for political ideology. Across these analyses, we find consistent evidence that far-right ideology is negatively associated with clean technology adoption, but that the strength of this association differs across technologies. Survey responses show substantially lower adoption intentions among AfD supporters, particularly for EVs, whereas municipality-level data reveal persistent partisan differences in spatial diffusion. Panel models further indicate that increases in AfD support are associated with subsequent declines in clean technology adoption, a pattern that is consistent across multiple specifications. We conclude with stylized counterfactual simulations that indicate clean technology diffusion in Germany would have been measurably higher had far-right support remained at its 2013 baseline. Results Adoption behaviour and intentions across party preferences We conducted a three-wave survey to assess how clean technology adoption and adoption intentions (hereafter referred to as adoption outcomes) vary with political party preferences among German residents. Supporters of the pro-environmental GRÜNE party were more likely to report adoption outcomes of both solar PV and EVs than supporters of other parties. Adoption outcomes generally increased between waves, especially between waves 1 and 2, which were separated by 17 months, compared with three months between waves 2 and 3. However, the partisan ordering of responses remained stable across waves, indicating that the pooled patterns are not driven by a single survey period. The magnitude of these differences varies notably by technology (Fig. 1a,b). Averaged across waves, GRÜNE supporters were 48% more likely than AfD supporters to report adoption outcomes for PV, compared with a substantially larger gap of 227% for EVs. AfD supporters exhibited the lowest adoption outcomes for EVs across all parties, whereas for PV they lagged behind GRÜNE supporters but were not the lowest across the political spectrum. Self-reported ideology on a left–right scale yields similar patterns: right-leaning individuals are less likely to report adoption outcomes, with a weaker association for PV (coefficient of primary interest (β) = −0.04, P = 0.044; Fig. 1c) than for EVs (β = −0.07, P < 0.001; Fig. 1d). To further quantify these relationships, we estimate logistic regression models of self-reported adoption outcomes controlling for socio-demographic, household and state-level characteristics. The results confirm a stronger alignment of EV adoption outcomes with political preferences relative to PV (Fig. 1e,f). AfD supporters are 18% less likely to report PV adoption outcomes, although this difference is not statistically significant (P = 0.072), whereas GRÜNE supporters are 35% more likely (P < 0.001) than supporters of other parties. For EVs, the partisan divide is substantially larger: AfD supporters are 67% less likely to report adoption outcomes (P < 0.001), with the effect increasing among politically engaged individuals. Across both technologies, adoption outcomes are positively correlated: individuals who have adopted PV are more likely to consider purchasing an EV, and vice versa. Adoption outcomes are further associated with higher income, younger age, higher education and home ownership. These findings are robust across alternative specifications and variable definitions (Supplementary Note 3.1). Differences in adoption outcomes are mirrored by systematic differences in underlying beliefs about clean technologies. When presented with statements on climate benefits, raw material availability, lifespan and safety risks, AfD supporters express markedly more negative views—particularly for EVs—than GRÜNE supporters and the overall sample (Supplementary Note 3.2). For example, 26% of AfD supporters consider replacing a combustion engine with an EV to be climate-friendly, compared with 61% of GRÜNE supporters; similarly, only 9% of AfD supporters perceive raw materials for EV production as abundant, compared with 31% among GRÜNE supporters. By contrast, perceptions of residential PV are more favourable across all groups, and partisan differences are substantially smaller. Additional longitudinal analyses of respondents observed across all three survey waves yield similar descriptive patterns but no statistically significant within-person party-affiliation effects, consistent with limited within-respondent variation in party preference over time. Partisan gap in the diffusion of clean technologies The partisan differences in individual-level survey responses are mirrored in spatial differences in municipality-level administrative records of clean technology diffusion. We find that diffusion of solar PV and EVs is highly heterogeneous across regions (Fig. 2a,b). Residential solar PV penetration is highest in southern states, whereas EV penetration follows a similar spatial pattern, with registration rates 84% higher in southern than in eastern states. Political differences are also pronounced (Fig. 2c,d). In 2023, solar PV penetration was 44% higher in GRÜNE-leaning than in AfD-leaning municipalities. However, solar PV penetration was highest in municipalities that were not strongly aligned with either party, consistent with the concentration of GRÜNE support in urban areas where rooftop solar adoption is constrained by higher renter shares and multi-family housing. EV penetration showed a clearer partisan gap with over 4% of registered vehicles being electric in GRÜNE-leaning municipalities in 2023, compared with 2% in AfD-leaning municipalities. To examine how the partisan deployment gap evolved over time, we group municipalities according to their average AfD–GRÜNE vote-share difference (Fig. 2e–h). Solar PV and EV penetration increased substantially across all groups over the study period, but growth was consistently weaker in municipalities with stronger AfD relative to GRÜNE support (Fig. 2e,f). Consequently, the absolute deployment gap between municipalities in the top and bottom quintiles of the AfD–GRÜNE vote-share difference widened in every observed year (Fig. 2g,h). In relative terms, however, this gap narrowed over time. For solar PV, penetration was initially about 76% lower in AfD-leaning municipalities, but this relative gap declined to roughly 31% by 2023. For EVs, the relative gap declined from around 84% in 2015 to 60% in 2019, but remained largely stable thereafter. Comparing the first nine observed years of diffusion for each technology indicates that the smaller relative partisan gap for solar PV may partly reflect its more mature stage of diffusion (Supplementary Fig. 9). These patterns partly reflect socio-economic and built-environment differences. Population density is negatively associated with solar PV penetration but positively associated with EV penetration, whereas both technologies are more prevalent in higher-income municipalities. Yet EV penetration remains comparatively low in AfD-leaning municipalities, even at higher income levels, indicating that income alone does not explain the partisan gap (Supplementary Note 4). Changes in far-right support and clean technology adoption We next estimate how changing political ideology is associated with municipality-level clean technology adoption. Our dependent variables measure penetration rates at the municipality-year level: for solar PV, the number of installed PV systems relative to the number of single- and two-family homes; for EVs, their share among registered private passenger cars. The main explanatory variables are lagged vote shares for AfD and GRÜNE, which capture opposing ideological positions on climate and energy policy. A central identification challenge is that political ideology is correlated with potentially unobserved socio-economic and infrastructural characteristics that can also shape clean technology adoption. We therefore estimate two complementary panel-data models: an ordinary least squares (OLS) and an instrumental variable (IV) specification (Table 1). The OLS specification relates within-municipality changes in political support to clean technology penetration, conditional on municipality fixed effects, county-year fixed effects and a comprehensive set of time-varying covariates. The IV specification uses local wolf attacks on livestock as an instrument for AfD support, as motivated by previous research linking such attacks to increases in far-right vote shares32. Wolf attacks predict AfD support, with first-stage diagnostics indicating strong instruments under heteroscedastic errors (Montiel Olea–Pflueger (MOP) effective F = 20.25–21.94; weak-instrument robust Anderson–Rubin P < 0.001) (Methods and Supplementary Note 6). Across models, higher AfD support is associated with lower penetration of both solar PV and EVs. The OLS estimates indicate that a 1-percentage-point increase in AfD vote share is associated with reductions of 0.02 percentage points in solar PV penetration and 0.03 percentage points in EV penetration (P < 0.001 for both). The IV estimates are larger, indicating reductions of 0.76 percentage points for solar PV and 0.18 percentage points for EVs among municipalities whose voting behaviour was affected by wolf attacks (P < 0.001 for both). The difference in magnitude between the OLS and IV estimates is not a contradiction: the OLS models estimate broader municipality-level associations across the full sample, whereas the IV identifies a local average treatment effect (LATE) from variation in AfD support induced by the locally salient shock of wolf attacks, tying the estimate to both a specific mechanism and a subset of municipalities (Methods). Results for GRÜNE vote shares point in the opposite direction, consistent with the interpretation that ideological polarization shapes adoption. The findings are robust to alternative ideology measures, alternative constructions of the adoption outcomes and a wide range of specification checks including a panel-matching approach (Supplementary Note 5). Whereas the negative association between AfD support and penetration rates remains statistically significant across all regional subsamples in the OLS panel specifications, the IV estimates are more sensitive to sample restrictions and are not robust to splitting the sample into Eastern and Western Germany41 (Supplementary Note 6). Nevertheless, the IV passes a wide range of common IV robustness checks42,43. Separating EVs into BEVs and PHEVs yields consistent coefficients and indicates that political ideology is more strongly associated with on BEV adoption compared with PHEV adoption (Supplementary Notes 5.4 and 6.7). Specifications interacting year indicators with party vote shares further indicate that the association between political ideology and clean technology adoption may have strengthened over time on average, consistent with a widening absolute partisan gap in technology diffusion (Supplementary Note 7). To quantify the potential implications of rising far-right support, we conduct counterfactual simulations (Fig. 3). For the OLS models, we estimate how clean technology diffusion may have evolved if the observed increase in AfD vote shares since 2013 had not occurred. The IV-based counterfactual addresses a narrower question: how diffusion may have differed in the absence of the wolf-attack-induced increase in AfD support. Because the IV identifies a LATE, these estimates are local to municipalities whose voting behaviour was shifted by wolf attacks and should not be interpreted as Germany-wide counterfactuals. For solar PV, the OLS model indicates that diffusion would be 0.06–0.16 percentage points higher (0.5–1.3% more installations) in the absence of increased AfD support, whereas the IV estimates imply 0.35–0.53 percentage points higher adoption (4.1–6.2% more installations) among municipalities affected by wolf attacks. For EVs, the corresponding estimates are 0.08–0.11 percentage points (2.1–2.9% more EVs) for the OLS model and 0.08–0.20 percentage points (2.3–5.8% more EVs) for the IV specification. Although the two approaches address different counterfactual questions and should not be compared directly, both indicate that rising support for the AfD is associated with economically meaningful reductions in clean technology adoption (Methods and Supplementary Note 8). Discussion Political polarization and clean technology diffusion Our findings show that political ideology is closely tied to clean technology adoption, but unevenly across technologies. Survey responses reveal much larger partisan divides for EVs than for solar PV: GRÜNE supporters were 48% more likely than AfD supporters to report PV adoption or adoption intentions, but 227% more likely for EVs. Regression models show a similar pattern relative to supporters of other parties: AfD supporters were 67% less likely to express EV adoption outcomes, whereas the PV association was smaller and not statistically significant. The spatial analysis of municipality-level registration records and voting-behaviour data shows that these individual-level differences are mirrored in observed diffusion patterns. Clean technology adoption varies substantially—municipalities with strong support for the AfD have lower penetration of both PV and EVs than municipalities with strong support for the GRÜNE party. The relative gap in penetration rates narrowed over time, even as absolute differences increased due to higher overall adoption rates. This distinction has political and practical relevance: as technologies continue to scale, even declining relative gaps can translate into larger absolute differences in the deployment of clean technology. Municipality-level panel estimates indicate that rising far-right support is associated with subsequent declines in clean technology adoption, even after accounting for spatially correlated socio-economic and infrastructural factors. The IV strategy strengthens this interpretation by using variation in AfD support associated with wolf attacks on livestock, but its estimates should be interpreted as a LATE for municipalities whose voting behaviour responded to these locally salient shocks rather than as population-wide average effects. The larger IV estimates may therefore reflect both the specific mechanism through which AfD support increased and the characteristics of affected municipalities, which are disproportionately rural and agricultural. Wolf attacks may mobilize voters around local land-use conflicts and make broader far-right positions, including opposition to climate policy and clean technologies, more politically salient. The difference may also reflect treatment-effect heterogeneity across municipalities. Taken together, the consistency of the negative association across OLS panel models, IV specifications and a wide range of robustness checks strengthens the conclusion that increasing far-right support is systematically linked to lower clean technology adoption. Our counterfactual simulations should be interpreted as illustrative rather than as precise quantifications, but they do indicate that the estimated associations are large enough to matter for the diffusion of both solar PV systems and EVs. Limitations Our study has several limitations that provide avenues for future research. First, our municipality-level analysis uses election vote shares as proxies for regional ideological trends. These measures capture spatial and temporal variation in political support, but they do not directly observe individual-level beliefs, climate attitudes or alignment with far-right platforms. Although our survey evidence links the aggregate patterns to individual-level partisan differences in technology perceptions and adoption intentions, the administrative records analysis should be interpreted as capturing regional political environments rather than individual voting behaviour. Second, each empirical approach has distinct inferential limitations. The linear panel models estimate within-municipality changes in clean technology adoption following changes in political support, while accounting for time-invariant municipal differences, regional time trends and local policy or institutional variation through county-year fixed effects. However, they remain observational and cannot fully rule out unobserved time-varying confounders. The IV analysis provides complementary evidence, but its estimates are more sensitive to sample restrictions and are not robust to separately estimating the model for Eastern and Western Germany41 (Supplementary Note 6.4). In addition, auxiliary analyses of the first stage provide supportive but not definitive evidence: a doubly robust difference-in-differences estimator yields a positive and statistically significant association between wolf attacks and AfD support, and a GRÜNE placebo outcome yields a near-zero and non-significant estimate, but the pre-trend Wald test rejects the parallel trends assumption (Supplementary Note 6.6). Although the IV specification passes a range of standard diagnostic and robustness checks42, we therefore interpret the IV results as complementary evidence alongside the panel-regression estimates, rather than as definitive causal estimates. Across approaches, however, the evidence consistently points in the same direction: increases in far-right voting are associated with subsequent decreases in clean technology adoption. Third, although we document variation in the partisan gap across technologies and over time, our analysis does not establish conclusively whether the relationship between changes in far-right voting and subsequent technology adoption changed over the study period. Models with time-varying coefficients are consistent with a widening association, but these estimates are exploratory and should be interpreted with caution (Supplementary Note 7). Relatedly, descriptive comparisons across more similar diffusion stages indicate that relative partisan gaps may decline as technologies mature and become more widely adopted (Supplementary Note 4), although these comparisons cannot distinguish maturity effects from technology-specific differences in costs, infrastructure needs, policy support or consumer meanings. Finally, our analysis is specific to consumer-level technologies in Germany. Utility-scale infrastructure, such as grid-scale solar, wind turbines and energy storage systems, is shaped by different planning, financing and regulatory processes and may follow different political dynamics. Moreover, the study period predates substantial far-right governing power, limiting our ability to assess how clean technology adoption may respond if far-right parties directly shaped policy implementation, infrastructure planning or local public discourse. Future research could further examine these dynamics across technologies, stages of technological maturity and institutional contexts. Methods Data sources This study utilizes two primary datasets: a survey dataset that includes self-reported political affiliations, socio-economic characteristics, technology adoption status and future intentions of respondents, and an observational dataset that includes clean technology adoption, socio-economic characteristics and relevant controls at the municipality level. To construct the survey dataset, we conducted a three-wave survey of German residents in December 2022, May 2024 and August 2024. The survey explored participants’ concerns about energy availability and costs, their support for various government policies and their willingness to adopt clean technologies, including EVs and solar PV (Supplementary Note 9). Socio-demographic and household information were also collected. The survey waves included 2,356 participants in December 2022, 3,092 in May 2024 and 2,500 in August 2024. Of these, 816 participants completed all three waves. Additional participants were recruited in later waves to ensure a representative sample for each wave. The internet-based survey, conducted by YouGov, targeted individuals aged 18 years and older. It was designed in English, translated into German and administered in German to participants. Party affiliations in the survey broadly align with the 2021 federal election results. The election versus survey shares, respectively, were: far-right (AfD), 10.3% vs 9.1%; pro-environmental (GRÜNE), 14.8% vs 18.4%; Social Democrats (SPD), 25.7% vs 25.4%; Conservatives (CDU or CSU), 24.1% vs 19.7%; Liberals (FDP), 11.5% vs 11.4%; and Socialists (Linke), 4.9% vs 6.7%. The study protocol was reviewed by the Ethics Committee of the University of Freiburg, which confirmed that formal ethical approval was not required because all data were collected anonymously and no personally identifying information was recorded. Participants provided informed consent under YouGov’s standard procedures, and no individual can be identified from the published results. To construct the observational dataset, we combined the following municipality-level data that were obtained from publicly accessible sources and public record requests. (1) Vehicle registrations were sourced from the German Federal Motor Transport Authority (Kraftfahrt-Bundesamt (KBA)) for the years 2014 to 202344. The scientific-use files contain the number of vehicles as of 1 January and we thus interpreted the 2024 file as the vehicle fleet of 2023. (2) Solar deployment data were extracted from the official registry of German solar PV systems (Marktstammdatenregister), which covers the period from 2000 to 202345. (3) Records of wolf attacks, voting data up to 2021 and shapefiles for municipalities were obtained from the repository of Clemm von Hohenberg and Hager32. The dataset comprises all reported wolf attacks in Germany (1998–2021) aggregated by election period at the municipality level for all German states except for Mecklenburg-Vorpommern and Rheinland-Pfalz; the state of Schleswig-Holstein was only observed until 2019. The AfD and GRÜNE vote shares for 2022–2023 were linearly interpolated between the 2021 and 2025 federal election results, obtained from the registry of the Federal Returning Officer (Bundeswahlleiterin); the 2025 result serves as an interpolation anchor only and falls outside the estimation sample. (4) Panel data encompassing income, unemployment rates, demographic composition and housing statistics were extracted from a database (Regionaldatenbank) managed by federal and state governments46,47,48,49,50. Whereas data availability varies across variables, it typically covers the period spanning from 2007 to 2022. (5) Data on the deployment of public EV charging stations between 2000 and 2023 were obtained from Germany’s Federal Network Agency (Bundesnetzagentur)51. (6) The shapefiles of postal code areas for the year 2022 were obtained from an open data portal of the German government52. Self-reported political affiliation and adoption willingness We examine the association between self-reported political affiliation and technology adoption status as well as future adoption intentions using logistic regression. The dependent variable is adoption willingness (AWi,w), a binary indicator equal to 1 if respondent i in survey wave w answered the question ‘Do you plan to purchase solar panels/an electric vehicle in the next year?’ with either ‘Have purchased’ or ‘Have not purchased but intend to purchase’, and 0 otherwise. The model is specified as follows: Our primary explanatory variables capture party affiliation and political orientation. In the baseline specification, AfD and GRÜNE are binary indicators denoting self-reported support for the respective party and the β coefficients capture their influence. We control for a rich set of individual-level characteristics, including gender (46% male), household income (50% with a monthly income of ≥€3,000), education (52.1% Abitur or higher), age (mean = 49.9), household size (mean = 2.25), car ownership (80% of households) and home ownership status (18%). These control variables are represented by the vector Xi,w, while γ denotes the associated coefficient vector. To account for persistent socio-economic differences across regions—particularly between Eastern and Western Germany—we include federal state fixed effects denoted by δstate[i]. Survey wave fixed effects (ϕw) capture systematic shifts in preferences over time and εi,w denotes an idiosyncratic error term. Our results are robust to alternative model specifications, including linear probability models and different combinations of control variables (Supplementary Note 3.1). In addition, we estimate the association between continuous political ideology and adoption willingness. Political ideology is defined in line with the Manifesto Project’s right–left (RILE) framework33, where left-leaning positions emphasize social equality, government regulation, redistribution and progressive social policies, whereas right-leaning positions emphasize market orientation, limited government intervention, lower taxation, national defence and conservative social values53. Survey respondents self-assessed their political orientation (PO) on a scale from 1 (strongly left-leaning) to 11 (strongly right-leaning) (mean = 5.5). We estimate the following specification: This complementary specification enables us to assess whether adoption intentions vary systematically along the ideological spectrum, independent of party affiliation. The variables are defined analogously to those in equation (1). Effect of far-right support on clean technology adoption To estimate the relationship between political preferences and the diffusion of clean technologies, we use panel-regression techniques and an IV strategy using municipality-level data for Germany. For solar PV, we measure the number of installed systems as a share of one- and two-family homes, and for EVs, we measure the share of EVs (battery electric and plug-in hybrid vehicles) among registered private passenger cars. Linear panel regression We begin with an OLS panel-regression framework to estimate the association between political preferences and clean technology diffusion across municipalities and over time: where m indexes municipalities and t years. The dependent variable TechSharem,t denotes the penetration rate of the given technology in municipality m and year t. The coefficients of primary interest, βAfD and βGRÜNE capture the associations between lagged support for the far-right AfD and pro-environmental GRÜNE parties, respectively, and subsequent technology diffusion. Political preferences are measured as the vote shares of the AfD and GRÜNE across federal and state elections. Vote shares are lagged by one year to mitigate concerns about simultaneity. Years without elections are linearly interpolated, and alternative constructions yield similar results (Supplementary Note 5). Results from specifications that allow the associations between political preferences and clean technology adoption to vary over time are reported in Supplementary Note 7. The vector Xm,t includes time-varying municipality-level controls, while γ denotes the associated coefficient vector. These include income per capita, the number of new buildings per capita, the share of single-family homes, population density, the share of population aged 65 years or older and the unemployment rate. For EV adoption, we also control for the availability of charging infrastructure, measured as public EV charging stations per capita. All control variables are normalized to the [0,1] interval. To account for differential pre-existing political environments, we include interactions between year fixed effects and baseline AfD support in 2013, λtAfDm,2013. This flexible specification enables municipalities with different initial levels of AfD support to follow distinct adoption trajectories over time, thereby controlling for underlying trends in technology diffusion that may be correlated with political preferences but would have occurred independently of the subsequent rise of the AfD. Municipality fixed effects δm absorb time-invariant characteristics, such as geography, long-standing infrastructure and persistent differences in housing stock. County-year fixed effects ϕcounty[m],t capture common shocks at the county (German: Landkreis) level, including local policy changes, technological progress and macroeconomic trends. The idiosyncratic error term is denoted by εm,t. We conduct a range of robustness checks, including alternative model specifications and a panel-matching approach (Supplementary Note 5). Instrumental-variable strategy To address potential endogeneity in political preferences, we complement the OLS panel models with an IV approach that uses variation in AfD vote shares associated with wolf attacks on livestock. Prior research shows that such attacks are associated with increased support for the AfD32. We estimate a 2SLS (two-stage least squares) model, restricting the sample to German states that experienced wolf attacks during the study period. In the first stage, we estimate: where WolfAttacksm,t−1 is an indicator equal to one if a wolf attack occurred in municipality m within the previous four years. All control variables and baseline interaction terms are defined analogously to the OLS panel specification. Fixed effects are included at the municipality (δm) and year (ϕt) levels. The first requirement for the IV strategy is relevance: wolf attacks must predict variation in AfD vote shares that is conditional on the controls and fixed effects. In this setting, the relevance of wolf attacks as an instrument is supported by their political salience in rural Germany. The resurgence of wolves in Germany, driven primarily by migration from Eastern European countries54, has resulted in a rise in livestock attacks55, sparking concerns about both safety and the livelihoods of rural communities. The AfD has amplified these concerns in political communication, and previous research has documented a positive association between wolf attacks and far-right vote shares. In our data, the coefficient on ‘WolfAttacks’ in the first-stage regression is positive and statistically significant (P < 0.001), with an effect size that is comparable to that reported by Clemm von Hohenberg and Hager32. We assess instrument strength using standard weak-instrument diagnostics. The MOP effective F-statistic, which accounts for heteroscedasticity and clustered dependence, is 20.25 for solar PV and 21.94 for EVs, exceeding the conventional benchmark of 10 (ref. 56). Weak-instrument robust Anderson–Rubin tests reject the null of no effect for both outcomes (P < 0.001) (Supplementary Note 6.5). These diagnostics indicate that weak identification is unlikely to drive the results. As an additional check of the first-stage relationship, we estimate a Callaway–Sant’Anna doubly robust difference-in-differences model57, treating first exposure to wolf attacks as the treatment (Supplementary Note 6.6). Wolf attacks are associated with a 1.3 percentage point increase in AfD vote share (P < 0.001). A placebo test using the GRÜNE vote share as the outcome yields near-zero and statistically insignificant estimates in both panels (EV: +0.1 percentage points, P = 0.60; PV: −0.1 percentage points, P = 0.59), indicating that the instrument captures shifts in far-right support rather than broad changes in political preferences. The second requirement is the exclusion restriction: conditional on controls and fixed effects, wolf attacks should affect solar PV and EV penetration only through their association with AfD support. This assumption is not directly testable, but several features of the setting make a direct effect on clean technology adoption unlikely. Municipality fixed effects absorb time-invariant differences in rurality, housing structure and settlement patterns, whereas year fixed effects account for national shocks such as energy prices and macroeconomic conditions. Time-varying controls further account for differences in population density, income and age structure. Direct economic effects are also likely to be limited: even the highest number of wolf attacks observed within a four-year period (67 attacks) is small relative to the size of the average rural municipality (2,288 inhabitants), and compensation payments reduce the financial burden on affected livestock owners54. We further examine the plausibility of the exclusion restriction using an out-of-sample test of the instrument’s first stage. Specifically, we assess whether wolf attacks also predict GRÜNE vote shares also before 2013, before the founding of the AfD and before EV adoption became empirically meaningful. Although this exercise is imperfect—wolf attacks were less frequent and political responses to wildlife conflicts may have differed in this earlier period or other parties may have also politicized wolf attacks—it provides a useful benchmark. Unlike in the main analysis time horizon after 2013, where wolf attacks are a strong and negative predictor of GRÜNE vote shares, the estimated relationship is weak and statistically insignificant in the pre-AfD period (Supplementary Note 6.3). This pattern is consistent with the interpretation that wolf attacks became politically salient through subsequent politicization, rather than reflecting a persistent underlying association between wildlife conflicts and environmental voting preferences. In the second stage, we estimate: where \(\widehat{\mathrm{AfD}}_{m,t-1}\) denotes the predicted AfD vote share from the first stage. The IV estimates should be interpreted as a LATE: they capture the effect of changes in AfD support on technology adoption for municipalities whose voting behaviour responded to wolf attacks. Because these municipalities may differ from the broader population, the IV estimates should not be interpreted as population-wide average effects but as complementary evidence to the broader associations estimated in the OLS panel models. Counterfactual simulations To illustrate the potential magnitude of the relationship between far-right political support and clean technology adoption, we construct a set of model-based counterfactual scenarios that quantify the potential number of foregone technology adoptions (TotalForegone). For the OLS panel specification, the counterfactual is based on the observed increase in AfD vote shares relative to their 2013 baseline (ΔAfDm,2023): We use the estimated coefficient on AfD vote share \({\beta }^{\mathrm{AfD}}_{\mathrm{OLS}}\) to calculate the difference between observed adoption and the adoption implied under the baseline-vote-share scenario. Exposure of municipality m in 2023 (Exposurem,2023), the last observed year, is measured using the number of one- and two-family homes for solar PV and the total vehicle fleet for EVs. The sum is taken over the set \(\mathcal{M}\), which includes all observed municipalities. The 95% confidence intervals were obtained using B = 10,000 bootstrap replications. These estimates should be interpreted as the adoption difference implied by the estimated association between AfD vote share and technology adoption under the maintained assumptions of the OLS panel model. They do not constitute causal estimates of the number of technologies that would have been adopted in the absence of the AfD’s electoral gains. For the IV specification, the interpretation of the counterfactual differs because the estimated coefficient, \({\beta}^{\mathrm{AfD}}_{\mathrm{IV}}\), identifies a LATE for the subset of municipalities whose voting behaviour is affected by the instrument. Applying the IV estimate to the aggregate national increase in AfD vote shares would therefore require extrapolating beyond the population for which the IV estimate is identified. To remain consistent with the scope of the IV design, we construct a counterfactual based only on the variation in AfD vote share induced by the instrument. Let πFS denote the estimated first-stage relationship between wolf attacks and AfD vote share. The corresponding counterfactual difference is given by: This calculation should be interpreted as an estimate of the adoption difference associated with the wolf-attack-induced component of AfD support within the estimation sample. It does not provide an estimate of the nationwide effect of changes in AfD support, nor should it be interpreted as the exact number of technologies that would have been adopted under an alternative political trajectory. Consistent with the local interpretation of the IV estimates, the calculation is restricted to municipalities that experienced at least one wolf attack during the study period, denoted as the set \(\mathcal{A}\). This avoids extrapolating the estimated LATE to municipalities outside the support of the instrument. To characterize statistical uncertainty, we perform a Monte Carlo simulation with 10,000 draws from the asymptotic sampling distributions of \({\beta }^{\mathrm{AfD}}_{\mathrm{IV}}\) and πFS, propagating uncertainty from both estimation stages to obtain 95% confidence intervals. Data availability German vehicle fleet and EV adoption data were acquired through the scientific-use files of Kraftfahrt-Bundesamt (KBA), the German Federal Motor Transport Authority, and are not publicly available (https://www.kba.de/DE/Statistik/Forschungsdatenzentrum/forschungsdatenzentrum_node.html). Solar panel installation data were sourced from Marktstammdatenregister, the official registry of German solar PV systems (https://www.marktstammdatenregister.de/MaStR). Wolf attacks, voting data and municipality shapefiles were acquired from the publicly available repository of Clemm von Hohenberg and Hager32 (https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/OOHLYX&version=1.0). EV charging station data were sourced from the Ladesäulenregister by the Bundesnetzagentur, Germany’s Federal Network Agency (https://www.bundesnetzagentur.de/DE/Fachthemen/ElektrizitaetundGas/E-Mobilitaet/Ladesaeulenkarte/start.html). The postal code areas were obtained from the open data portal of the German government (https://www.govdata.de/suche/daten/deutschland-postleitzahlen). Panel data on income, unemployment, age and buildings were sourced from the Regionaldatenbank database (https://www.regionalstatistik.de/genesis/online). We cannot share the raw EV adoption data due to the conditions of the data use agreement. Those interested in obtaining these data should contact the KBA directly. All other datasets needed to replicate the results, including the survey data, are available through the Open Science Framework at https://osf.io/hk85r/?view_only=4a13ba446f21434c8fc7d83e35f91cd3. Code availability Analyses were conducted in R (v4.4.3; fixest v0.14.0, WeightIt v1.7.0, MatchIt v4.7.2, ivDiag v1.0.6, did v2.3.0 and ggplot2 v4.0.2), with data preparation in Python (v3.12; pandas v2.3.3, polars v1.36.0 and geopandas v1.1.1). The code to replicate the results and figures is available through the Open Science Framework at https://osf.io/hk85r/?view_only=4a13ba446f21434c8fc7d83e35f91cd3. References Ivanova, D. et al. Environmental impact assessment of household consumption. J. Ind. Ecol. 20, 526–536 (2016). Ajanovic, A. & Haas, R. Dissemination of electric vehicles in urban areas: major factors for success. Energy 115, 1451–1458 (2016). Breyer, C., Koskinen, O. & Blechinger, P. Profitable climate change mitigation: the case of greenhouse gas emission reduction benefits enabled by solar photovoltaic systems. Renew. Sustain. Energy Rev. 49, 610–628 (2015). Sierzchula, W., Bakker, S., Maat, K. & van Wee, B. The influence of financial incentives and other socio-economic factors on electric vehicle adoption. Energy Policy 68, 183–194 (2014). Lévay, P. Z., Drossinos, Y. & Thiel, C. The effect of fiscal incentives on market penetration of electric vehicles: a pairwise comparison of total cost of ownership. Energy Policy 105, 524–533 (2017). Stern, P. C., Dietz, T., Nielsen, K. S., Peng, W. & Vandenbergh, M. P. Feasible climate mitigation. Nat. Clim. Change 13, 6–8 https://doi.org/10.1038/s41558-022-01563-7 (2022). Nielsen, K. S. et al. Realizing the full potential of behavioural science for climate change mitigation. Nat. Clim. Change 14, 322–330 (2024). Walzberg, J., Carpenter, A. & Heath, G. A. Role of the social factors in success of solar photovoltaic reuse and recycle programmes. Nat. Energy 6, 913–924 (2021). Archsmith, J., Muehlegger, E. & Rapson, D. S. Future paths of electric vehicle adoption in the United States: predictable determinants, obstacles, and opportunities. Environ. Energy Policy Econ. 3, 71–110 (2022). Gillingham, K. T. & Bollinger, B. Social learning and solar photovoltaic adoption. Manag. Sci. 67, 7091–7112 (2021). Bollinger, B. & Gillingham, K. Peer effects in the diffusion of solar photovoltaic panels. Mark. Sci. 31, 900–912 (2012). Hoehnke, H. et al. Structural barriers and policy pathways for a just clean energy transition. Nat. Rev. Clean Technol. 2, 327–347 (2026). Wussow, M. et al. Exploring the potential of non-residential solar to tackle energy injustice. Nat. Energy 9, 654–663 (2024). Middleton, N. New campaign to stop ‘EV culture wars and misinformation’. Fleet World https://fleetworld.co.uk/new-campaign-to-stop-ev-culture-wars-and-misinformation/ (2023). Stolle, D., Hooghe, M. & Micheletti, M. Politics in the supermarket: political consumerism as a form of political participation. Int. Polit. Sci. Rev. 26, 245–269 (2005). Gromet, D. M., Kunreuther, H. & Larrick, R. P. Political ideology affects energy-efficiency attitudes and choices. Proc. Natl Acad. Sci. USA 110, 9314–9319 (2013). Mildenberger, M., Howe, P. D. & Miljanich, C. Households with solar installations are ideologically diverse and more politically active than their neighbours. Nat. Energy 4, 1033–1039 (2019). Burgess, M. G. et al. Supply, demand and polarization challenges facing US climate policies. Nat. Clim. Change 14, 134–142 (2024). Steg, L. Car use: lust and must. Instrumental, symbolic and affective motives for car use. Transp. Res. Part A Policy Pract. 39, 147–162 (2005). Brückmann, G., Willibald, F. & Blanco, V. Battery electric vehicle adoption in regions without strong policies. Transp. Res. Part D Transp. Environ. 90, 102615 (2021). Sintov, N. D., Abou-Ghalioum, V. & White, L. V. The partisan politics of low-carbon transport: why democrats are more likely to adopt electric vehicles than Republicans in the United States. Energy Res. Soc. Sci. 68, 101576 (2020). Steg, L., Bolderdijk, J. W., Keizer, K. & Perlaviciute, G. An integrated framework for encouraging pro-environmental behaviour: the role of values, situational factors and goals. J. Environ. Psychol. 38, 104–115 (2014). Brick, C., Sherman, D. K. & Kim, H. S. “Green to be seen” and “brown to keep down”: visibility moderates the effect of identity on pro-environmental behavior. J. Environ. Psychol. 51, 226–238 (2017). Helgason, A. F. & Rehm, P. Long-term income trajectories and the evolution of political attitudes. Eur. J. Polit. Res. 62, 264–284 (2023). Guillaud, E. Preferences for redistribution: an empirical analysis over 33 countries. J. Econ. Inequal. 11, 57–78 (2013). Stubager, R. Education effects on authoritarian–libertarian values: a question of socialization. Br. J. Sociol. 59, 327–350 (2008). Schoon, I., Cheng, H., Gale, C. R., Batty, G. D. & Deary, I. J. Social status, cognitive ability, and educational attainment as predictors of liberal social attitudes and political trust. Intelligence 38, 144–150 (2010). Cantoni, E. & Pons, V. Does context outweigh individual characteristics in driving voting behavior? Evidence from relocations within the United States. Am. Econ. Rev. 112, 1226–1272 (2022). Gimpel, J. G., Lovin, N., Moy, B. & Reeves, A. The urban–rural gulf in American political behavior. Polit. Behav. 42, 1343–1368 (2020). Hall, A. B. & Yoder, J. Does homeownership influence political behavior? Evidence from administrative data. J. Polit. 84, 351–366 (2022). Dokshin, F. A. & Gherghina, M. Party affiliation predicts homeowners’ decisions to install solar PV, but partisan gap wanes with improved economics of solar. Proc. Natl Acad. Sci USA 121, e2303519121 https://doi.org/10.1073/pnas.2303519121 (2024). Clemm von Hohenberg, B. & Hager, A. Wolf attacks predict far-right voting. Proc. Natl Acad. Sci. USA 119, e2202224119 (2022). Lehmann, P. et al. Manifesto Project Dataset (Version 2024a) https://doi.org/10.25522/MANIFESTO.MPDS.2024A (Manifesto Project, 2024). AfD-Fraktion im Deutschen Bundestag. Kostenfalle Wärmepumpe. AfD-Fraktion im Deutschen Bundestag https://afdbundestag.de/kostenfalle-waermepumpe/ (accessed 23 July 2026). AfD Landesverband NRW. Habecks Heiz-Irrsinn rächt sich: 80 Prozent der Wärmepumpen müssen wieder raus! AfD NRW https://afd.nrw/aktuelles/2023/07/habecks-heiz-irrsinn-raecht-sich-80-prozent-der-waermepumpen-muessen-wieder-raus/ (accessed 23 July 2026). Deutscher Bundestag, Parlamentsnachrichten. AfD: Wind und Solar bekämpfen, Landschaften schützen. Deutscher Bundestag https://www.bundestag.de/presse/hib/kurzmeldungen-984366 (accessed 23 July 2026). AfD-Fraktion Sachsen. Ladenhüter E-Auto: Irrweg Elektromobilität stoppen. AfD-Fraktion Sachsen https://afd-fraktion-sachsen.de/ladenhueter-e-auto-irrweg-elektromobilitaet-stoppen/ (accessed 23 July 2026). Deutsche Presse-Agentur. Elektroauto: Sächsische AfD hält Elektromobilität für einen Irrweg. Heiseonline https://www.heise.de/news/Elektroauto-Saechsische-AfD-haelt-Elektromobilitaet-fuer-einen-Irrweg-9220074.html (accessed 23 July 2026). Bundestagswahl 2021 Ergebnisse. Die Bundeswahlleiterin https://www.bundeswahlleiterin.de/bundestagswahlen/2021/ergebnisse.html (accessed 23 July 2026). Bundestagswahl 2025 Ergebnisse. Die Bundeswahlleiterin https://www.bundeswahlleiterin.de/bundestagswahlen/2025/ergebnisse.html (accessed 23 July 2026). Sonntag, N. The East in wolf’s clothing. Wolf attacks correlate with but do not cause far-right voting. Elect. Stud. 96, 102957 (2025). Lal, A., Lockhart, M., Xu, Y. & Zu, Z. How much should we trust instrumental variable estimates in political science? Practical advice based on 67 replicated studies. Polit. Anal. 32, 521–540 (2024). Montiel Olea, J. L. & Pflueger, C. A robust test for weak instruments. J. Bus. Econ. Stat. 31, 358–369 (2013). Kraftfahrt-Bundesamt. Bestand im Zentralen Fahrzeugregister 2025 (Off-Site, Variante S). Kraftfahrt-Bundesamt https://doi.org/10.25525/KBA-FDZOFF.ZFZRBS.2025.1 (2025). Marktstammdatenregister. Bundesnetzagentur für Elektrizität, Gas, Telekommunikation, Post und Eisenbahnen. Datendownload | Marktstammdatenregister (Version: October 1, 2024) https://www.marktstammdatenregister.de/MaStR/Datendownload (accessed 23 July 2026). Bevölkerung nach Geschlecht und Altersgruppen (17) - Stichtag 31.12. - regionale Tiefe: Gemeinden. Regionaldatenbank Deutschland https://www.regionalstatistik.de/genesis//online?operation=table&code=12411-02-03-5&bypass=true&levelindex=1&levelid=1788332176056#abreadcrumb (accessed 23 July 2026). Arbeitslose nach ausgewählten Personengruppen – Jahresdurchschnitt – regionale Tiefe: Gemeinden. Regionaldatenbank Deutschland https://www.regionalstatistik.de/genesis//online?operation=table&code=13211-01-03-5&bypass=true&levelindex=1&levelid=1788332470538#abreadcrumb (accessed 23 July 2026). Fertigstellungen neuer Wohngebäude und Wohnungen in Wohngebäuden nach Zahl der Wohnungen – Jahressumme – regionale Tiefe: Gemeinden. Regionaldatenbank Deutschland https://www.regionalstatistik.de/genesis//online?operation=table&code=31121-01-02-5&bypass=true&levelindex=1&levelid=1788332511019#abreadcrumb (accessed 23 July 2026). Bestand an Wohngebäuden und Wohnungen in Wohn- und Nichtwohngebäuden – Stichtag 31.12. – regionale Tiefe: Gemeinden. Regionaldatenbank Deutschland https://www.regionalstatistik.de/genesis//online?operation=table&code=31231-02-01-5&bypass=true&levelindex=1&levelid=1788332331736#abreadcrumb (accessed 23 July 2026). Lohn- und Einkommensteuerpflichtige, Gesamtbetrag der Einkünfte, Lohn- und Einkommensteuer – Jahressumme – regionale Tiefe: Gemeinden. Regionaldatenbank Deutschland https://www.regionalstatistik.de/genesis//online?operation=table&code=73111-01-01-5&bypass=true&levelindex=1&levelid=1788332415733#abreadcrumb (accessed 23 July 2026). Ladesäuleninfrastruktur. Statistische Ämter des Bundes und der Länder Gemeinsames Statistikportal https://www.statistikportal.de/de/nachhaltigkeit/ergebnisse/ziel-9-industrie-innovation-und-infrastruktur/ladesaeuleninfrastruktur (accessed 23 July 2026). Deutschland: Postleitzahlen. GOVDATA https://www.govdata.de/suche/daten/deutschland-postleitzahlen (accessed 23 July 2026). Laver, M. J. & Budge, I. Measuring policy distances and modelling coalition formation. In Party Policy and Government Coalitions (eds Laver, M. J. & Budge, I.) 15–40 https://doi.org/10.1007/978-1-349-22368-8_2 (Palgrave Macmillan, 1992). Reinhardt, I. & Kluth, G. Wolfsverursachte Schäden, Präventions- und Ausgleichszahlungen in Deutschland 2021 (Dokumentations- und Beratungsstelle des Bundes zum Thema Wolf, 2022). Khorozyan, I. & Heurich, M. Large-scale sheep losses to wolves (Canis lupus) in Germany are related to the expansion of the wolf population but not to increasing wolf numbers. Front. Ecol. Evol. 10, 778917 (2022). Staiger, D. & Stock, J. H. Instrumental Variables Regression with Weak Instruments. Technical Working Paper No. 151 https://doi.org/10.3386/t0151 (National Bureau of Economic Research, 1994). Callaway, B. & Sant’Anna, P. H. C. Difference-in-Differences with multiple time periods. J. Econom. 225, 200–230 (2021). Acknowledgements We thank H. Boudet and J. Flora for their assistance with the survey, and L. Rüde and M. Beihofer for assisting with the initial data acquisition. We also thank B. Clemm von Hohenberg for his valuable feedback about the research. Finally, we thank our colleagues at the Climate Action Research Lab for insightful discussions and feedback on research presentations. Funding This work was funded in part by the European Union under the HORIZON program “Climate-Resilient Development Pathways in Metropolitan Regions of Europe (CARMINE)” (award number 101137851) (M.W. and D.N.). The administration of the survey was supported by the Eva Mayr-Stihl Foundation (C.Z., R.S.-S., S.P. and C.K.). Open access funding provided by Albert-Ludwigs-Universität Freiburg im Breisgau. Author information Authors and Affiliations Contributions S.R. and M.W. were lead authors and contributed equally. S.R., M.W. and C.Z. conceptualized and designed the research; S.R., M.W., C.Z. and B.K. performed the research and analysed the data; S.R., M.W. and C.Z. wrote the initial paper draft; all authors reviewed and edited the paper. D.N., R.S.-S, S.P. and C.K. provided funding acquisition support. Corresponding author Ethics declarations Competing interests The authors declare no competing interests. Peer review Peer review information Nature Sustainability thanks Holly Caggiano, Florian Egli and Malte Toetzke for their contribution to the peer review of this work. Additional information Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Supplementary information Supplementary Information (download PDF ) Supplementary Figs. 1–11, Tables 1–28 and Notes 1–9. Rights and permissions Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. About this article Cite this article Reining, S., Wussow, M., Zanocco, C. et al. Far-right political support and clean technology adoption in Germany. Nat Sustain (2026). https://doi.org/10.1038/s41893-026-01935-3 Received: Accepted: Published: Version of record: DOI: https://doi.org/10.1038/s41893-026-01935-3