Research
How political elites and parties compete over climate change, and how it became politicised.
- PublishedExecutive Climate Change Attention: Toward an Indicator of Comparative Climate Change Attention
- PublishedL'opposition aux politiques climatiques comme stratégie électorale. La conflictualisation de la compétition partisane sur les enjeux écologiques en France (1981-2024)
- Book chapterPartis non écologistes ↗
- Working paperSelective Accommodation: How Conservative Parties Respond to Radical Right Contestation of Climate Action in Europe
- Working paperDisplacing Climate Competition: Electoral Yield and Far-Right Opposition to Climate Policy
How what governments and parties propose on climate shapes citizens' support for policies and their vote.
- PublishedMore than Symbols: The Effect of Symbolic Policies on Climate Policy Support
- PublishedBetween Goals and Means: The Electoral Consequences of Parties' Climate Proposals
- Working paperWho Listens to the Party? How Political Parties Shape Voters' Issue Priorities
- Working paperNo Party to Rule Them All: Green Delegitimisation and the Erosion of Issue Ownership on Climate Change
Open comparative datasets on what parties post, where, and about which issues.
Climate politics: Elites, party competition and politicisation
- Executive Climate Change Attention: Toward an Indicator of Comparative Climate Change Attention2026Research question
How much attention do national leaders give to climate change when they address the world, and how has it changed since the 1980s?
Every year, each country's executive addresses the United Nations General Assembly. We trained a language model to spot sentences about climate change in these speeches, giving a comparable measure of climate attention for 195 countries since 1985.
Key findingClimate change has gone from a marginal topic to a standard part of leaders' speeches at the UN: before 2007, only one speech in five mentioned it; since 2015, 85% do. But attention is very unequal: it is highest in the countries most exposed to climate change, such as small island states, and lowest in Arab states.ResultsEach country's value is the share of sentences in its leader's annual speech at the UN General Assembly that a language model classifies as being about climate change. Small island states are hard to see on the map: see the ranking. Averages are simple means across countries.
What explains the differences?TimingAttention jumps in 2006–2007, falls back during the Great Recession, and rises again after 2018, tracking media coverage of climate change.
Vulnerability↑More vulnerable countries talk more about climate change: small island states lead, Arab states trail. The effect holds mainly for middle-income countries.
Cost of acting↓Leaders of economies whose growth relies more on emissions, and of poorer countries, tend to talk less about it.
Politics↑Liberal democracies talk more about climate, and left-leaning leaders more than right-leaning ones. In rich democracies, stronger green parties go with more attention.
Associations from panel models with country and year fixed effects: they describe who talks more, not causal effects.
ContributionComparative research lacks a measure of how much governments prioritize climate change that covers most countries over a long period. This paper introduces the Executive Climate Change Attention indicator, built from UN General Assembly speeches with a fine-tuned transformer model, and shows its value by examining the drivers of climate attention and its link with national climate policy outcomes. The indicator and the model are openly available.
Abstract
This article introduces the Executive Climate Change Attention indicator. This unique comparative measure assesses the importance given to climate change by executives in a total of 195 countries since 1985 through their annual individual government speeches at the United Nations General Assembly (UNGA). To compute this indicator, we trained a transformer model to classify sentences related to climate change in the government speeches at the UNGA. We illustrate the usefulness of the indicator by examining the drivers of climate change attention at the UNGA and the link between climate change attention and national climate policy outcomes.↑ Back to overviewCite
Grossman, E., & Jan, M. (2026). Executive Climate Change Attention: Toward an Indicator of Comparative Climate Change Attention. Global Environmental Politics, 26(2), 137–150. https://doi.org/10.1162/glep.a.1
@article{grossman2026executive, title = {{Executive Climate Change Attention: Toward an Indicator of Comparative Climate Change Attention}}, author = {Grossman, Emiliano and Jan, Malo}, journal = {Global Environmental Politics}, year = {2026}, volume = {26}, number = {2}, pages = {137--150}, doi = {10.1162/glep.a.1} } - L'opposition aux politiques climatiques comme stratégie électorale. La conflictualisation de la compétition partisane sur les enjeux écologiques en France (1981-2024)2025Research question
Why have French right-wing parties started to oppose ecology, after decades of near consensus?
I measure how often each party supports or opposes ecology in its election manifestos since 1981 and in its social media posts since 2010, then code what the opposition targets and model when it rises.
Key findingUntil the late 2010s, French parties broadly agreed on ecology, and even the far right rarely opposed it. Since 2021, the right has turned it into a line of conflict: on social media, the RN now posts more against ecology than for it, and Reconquête's 2024 manifesto opposed ecology more than it supported it.ResultsSupport and opposition are the share of each manifesto's sentences (manual coding) or of each year's social media posts (supervised classifiers, with every opposed post checked by hand) taking that position on ecology. Targets: manual coding of all LR, RN and Reconquête posts opposing ecology since 2010.
What drives the opposition?Salience↑The RN opposes ecology more when other parties talk it up: climate becomes a wedge issue against mainstream parties.
Costlier policies↑New, more constraining climate policies, from low-emission zones to the EU Green Deal and the 2035 car ban, gave the right more targets.
Social movements↑The Yellow Vests and the 2024 farmers' protests signalled a potential electorate: in the month after the Yellow Vests began, LR's share of anti-ecology content was 17 points higher.
Competition on the right↑Each right-wing party opposes ecology more when the others do, and LR, the RN and Reconquête oppose it in very similar terms.
Costs, not scienceThe opposition is framed around purchasing power, for drivers, farmers, rural residents and homeowners. Questioning climate science is almost absent.
Drivers come from weekly regressions on LR and RN social media content and from manual coding of every post opposing ecology.
ContributionResearch on issue competition mostly shows mainstream parties converging on issues raised by niche parties such as the Greens. This article identifies the conditions under which parties instead turn a consensual issue into a line of conflict: high salience combined with costlier policies. It also shows the radical right broadening its issue agenda beyond immigration, and the value of social media data and supervised classifiers for tracking party discourse.
Abstract
This article analyzes the recent rise of conflict in party competition over ecological issues, which had long been the subject of broad political consensus. It argues that opposition to environmental issues stems from the combination of the growing salience of climate issues and the transformation of public policies, whose economic costs are increasingly perceived as high. While incentives for opposition were initially weak, these changes offered parties the opportunity to mobilize groups considered to be the “losers” of the ecological transition. This hypothesis is tested through an analysis of the salience of ecological issues and the positions taken by French political parties in their manifestos and on social media.Cet article analyse la récente conflictualisation de la compétition partisane sur l’écologie, jusque-là surtout marquée par un traitement consensuel de ces enjeux. Il soutient que l’opposition à l’écologie résulte de la combinaison d’une saillance accrue des enjeux climatiques et de la transformation des politiques publiques, dont les coûts économiques sont perçus comme plus élevés. Alors que les incitations à l’opposition étaient initialement faibles, ces changements ont offert aux partis l’opportunité de mobiliser les groupes considérés perdants de la transition écologique. Cette hypothèse est testée à partir de l’analyse de la saillance et des positions des partis français sur l’écologie dans leurs programmes et sur les réseaux sociaux.↑ Back to overviewCite
Jan, M. (2025). L'opposition aux politiques climatiques comme stratégie électorale. La conflictualisation de la compétition partisane sur les enjeux écologiques en France (1981-2024). Revue française de science politique, 75(3), 459–490. https://doi.org/10.3917/rfsp.753.0459
@article{jan2025lopposition, title = {{L'opposition aux politiques climatiques comme stratégie électorale. La conflictualisation de la compétition partisane sur les enjeux écologiques en France (1981-2024)}}, author = {Jan, Malo}, journal = {Revue française de science politique}, year = {2025}, volume = {75}, number = {3}, pages = {459--490}, doi = {10.3917/rfsp.753.0459} } - Selective Accommodation: How Conservative Parties Respond to Radical Right Contestation of Climate Action in Europe2026↑ Back to overview
Abstract
Do conservative parties adjust their issue strategies on climate change in response to radical right contestation, and if so, under which conditions? Radical right parties increasingly contest climate action across Europe, yet unlike on immigration or other cultural core issues, they do so on an already salient consensus where they lack strong issue ownership and where conservative parties hold prior commitments making outright convergence costly. I argue that conservative parties respond through selective accommodation, increasing criticism of specific climate policies while maintaining broader support for climate goals, but that this adjustment is conditional on radical right electoral success and conservative opposition status. Using supervised text classification to measure pro- and anti-climate salience in party social media communication across 14 European countries, combined with panel regressions and case studies of Germany, the Netherlands, and the United Kingdom, I find that radical right anti-climate politicization is associated with increased conservative anti-climate salience only under these specific conditions, while pro-climate salience remains stable. These findings suggest that radical right influence extends beyond cultural core issues, albeit in conditional ways. - Displacing Climate Competition: Electoral Yield and Far-Right Opposition to Climate Policy2026↑ Back to overview
Abstract
Far-right parties have increasingly made opposition to climate action central to their appeals, although climate scepticism long remained peripheral to the party family. This paper asks why opposition to climate action became salient, and why it emerged only after support for climate action had become established. I argue that the electoral yield of climate opposition depends on the level at which competition is organised: opposing broadly supported climate goals offers far-right parties little, but as climate commitments are translated into identifiable policies, opposing particular means becomes more attractive. Using the PartySOME data on parties' social media communication and a case study of Denmark, I examine whether far-right opposition follows the broader expansion of climate support, and whether it shifts towards concrete policy means, addresses groups less central to far-right communication, and foregrounds economic concerns alongside cultural and populist appeals.
Climate politics: How elites shape public opinion
- More than Symbols: The Effect of Symbolic Policies on Climate Policy Support2026Research question
Can symbolic policies, measures with a strong message but little material impact, increase public support for costly climate policies?
Costly climate policies often meet public resistance. We argue that pairing them with a symbolic policy, such as ministers taking the train instead of the plane, makes them more popular by changing what the costly policy signals rather than what it costs.
Key findingAnnounced alone, a lower highway speed limit was backed by 45% of respondents; paired with ministers taking the train instead of the plane, by 69%. For a carbon tax rise, support went from 20% to 32% when paired with a yearly citizens' climate proposal debated in Parliament.ResultsEach bar is the share of respondents who gave one of the two positive answers on a 4-point scale, among people randomly shown that announcement (3,300 respondents in France, 660 per announcement, survey-weighted). Whiskers are 95% confidence intervals; the dashed line marks the costly policy announced alone. The paper reports the same comparisons on the full 4-point scale, with controls.
ContributionResearch on public support for climate policy has mostly looked at material compensation for those who bear the costs. This paper identifies a non-material route: symbolic policies change the message a costly policy conveys, raising perceived fairness, elite exemplarity, and government credibility. It introduces the concept of symbolic policies and tests it with survey experiments and qualitative interviews in France.
Abstract
As climate change effects become increasingly salient, the need for stringent climate policies becomes more pressing. The implementation of such policies is often met with resistance from the public due to their perceived costs and distributional implications. Scholars have mostly focused on material compensations to increase public support among policy losers. This article goes beyond the existing literature by showing how what we term symbolic policies can enhance support for costlier policies. We define symbolic policies as policies sending meaningful messages to the public but having low material impacts. We argue that without changing the material costs that climate policies impose, symbolic policies increase public support by altering the message that costly policies convey. We demonstrate our argument using survey experiments and qualitative interviews conducted in France, showing that symbolic policies can significantly increase support for costly climate policies and increase perceptions of fairness, elite behavior, and government credibility.↑ Back to overviewCite
Tallent, T., Jan, M., & Sattelmayer, L. (2026). More than Symbols: The Effect of Symbolic Policies on Climate Policy Support. American Political Science Review, 120(3), 1106–1126. https://doi.org/10.1017/S0003055425101408
@article{tallent2026more, title = {{More than Symbols: The Effect of Symbolic Policies on Climate Policy Support}}, author = {Tallent, Théodore and Jan, Malo and Sattelmayer, Luis}, journal = {American Political Science Review}, year = {2026}, volume = {120}, number = {3}, pages = {1106--1126}, doi = {10.1017/S0003055425101408} } - Between Goals and Means: The Electoral Consequences of Parties' Climate Proposals2026Research question
Do candidates win or lose votes by supporting or opposing climate action, and does it matter whether they talk about general goals or concrete policies?
A representative sample of French voters repeatedly chose between two hypothetical presidential candidates whose programmes varied at random, including their stance on climate change as a goal and on a concrete policy: banning new petrol and diesel cars from 2035.
Key findingRejecting climate action is costly: a candidate calling to stop fighting climate change was chosen 33% of the time, against 58% for one pledging to continue. But on concrete policy, the winning position is mixed: investing in electric cars without banning petrol and diesel cars (57%) beats adding the 2035 ban (48%) and going back to petrol and diesel (46%).ResultsVoters chose between two candidates at a time (1,811 respondents, 9,055 choices, ELIPSS panel, July 2024). Each bar is the share of times a candidate taking that position was chosen, averaging over all the other randomly varied proposals (taxes, immigration, democracy). Above 50%, the position helps; below, it hurts. Whiskers are 95% confidence intervals.
Who rewards what?Denial↓Rejecting climate action is penalised in every partisan group, from Green supporters (19%) to radical-right supporters (43%), who still prefer a candidate pledging to continue (56%).
Top priorityMaking climate the top priority adds little: Green supporters reward it slightly more than continuing (67% vs 63%), radical-right supporters less (50% vs 56%).
Mixed position↑"No ban, but invest in electric cars" scores highest in every partisan group; among Green supporters, it is level with the ban.
The ban↓Radical-right, conservative and non-partisan voters penalise the 2035 ban, and no group, not even Green supporters, rewards it over investment alone.
Party groups are based on the party respondents feel closest to: they describe core supporters, not each party's potential electorate.
ContributionStudies of climate politics usually place parties on a single scale from opposing to supporting climate action. This paper distinguishes the direction of a position from its specificity, goals versus concrete policy means, and shows that they carry different electoral consequences. This helps explain why parties can endorse climate action in principle while contesting specific instruments, and why pro-climate parties face limited electoral incentives to advance ambitious, specific proposals.
Abstract
How do political parties’ climate proposals impact their electoral fortunes? While research has explored public support for climate policies and the electoral consequences of implementing them, evidence remains limited on how parties’ climate proposals influence their support. This article conceptualises party positions on climate action along two dimensions: the direction of the proposal (supporting or opposing climate action) and its specificity (emphasising goals or concrete policy means). Drawing on a conjoint experiment in France, it analyses how climate proposals on these dimensions affect voter support for hypothetical candidates. Results show that supporting climate goals is electorally rewarding, while climate denial is strongly penalised across partisan lines. However, backing ambitious, concrete climate policies yields no electoral gains and may even provoke voter pushback, whereas avoiding such measures can be advantageous. Overall, the findings suggest pro-climate parties face limited electoral incentives to advance ambitious, specific climate proposals.↑ Back to overviewCite
Jan, M. (2026). Between Goals and Means: The Electoral Consequences of Parties' Climate Proposals. West European Politics. https://doi.org/10.1080/01402382.2026.2674875
@article{jan2026between, title = {{Between Goals and Means: The Electoral Consequences of Parties' Climate Proposals}}, author = {Jan, Malo}, journal = {West European Politics}, year = {2026}, doi = {10.1080/01402382.2026.2674875} } - Who Listens to the Party? How Political Parties Shape Voters' Issue Priorities2026↑ Back to overview
- No Party to Rule Them All: Green Delegitimisation and the Erosion of Issue Ownership on Climate Change2025↑ Back to overview
Data & party communication: Parties on social media
- Issue Competition on Social Media: A New Comparative Dataset on Parties' Issue Agendas2026
Building on PartySOME, we classified 4.6 million posts by 219 parties in 18 European countries into the 21 policy issues of the Comparative Agendas Project, giving a monthly measure of what each party talks about on Facebook, Twitter and Instagram.
219parties18European countries3platforms4.6Mposts classified21policy issues, monthlyResultsIssue shares are the share of a party's posts classified into each Comparative Agendas Project issue (posts without policy content count in the total). Family lines average parties' yearly shares; years with fewer than 20 posts are left out. Party families from ParlGov.
What the data showIssue ownershipParties talk most about the issues they own: radical-right parties about immigration and crime, Greens about the environment, the left about the economy, civil rights and labour.
Shocks↑Agendas react to events: health takes over after the COVID-19 outbreak, and defence, energy and foreign affairs rise after Russia's invasion of Ukraine.
Immigration↑When mainstream parties, left or right, pay more attention to immigration, radical-right parties rise in the polls the following month.
The immigration result comes from monthly models of radical-right polling with country and month fixed effects.
ContributionStudies of issue competition have mostly relied on manifestos, which only capture election periods. PartySOME II offers conceptually standardized, cross-platform and cross-national measures of parties' issue attention, month by month, opening up the study of how parties react to events, to each other and to public opinion between elections.
↑ Back to overviewAbstract
Understanding how parties compete by shaping agendas and promoting issue priorities is central to the study of party competition. Existing research relies heavily on manifestos, limiting analyses to election periods and offline contexts, while newer datasets lack cross-national breadth, temporal depth, and conceptual harmonization. We introduce a comparative dataset of party issue agendas derived from 4,600,000 social media posts for 219 parties across 18 European countries on Twitter, Facebook, and Instagram. Using supervised text classification aligned with the Comparative Agendas Project, we generate monthly measures of salience over more than a decade, providing conceptually standardized, cross-platform, and cross-nationally comparable indicators of issue attention. We validate our measures through comparisons over time, across party families, major political shocks, and existing databases, and illustrate the dataset's leverage by showing how mainstream parties' immigration attention is associated with radical-right support gains. This dataset substantially expands the empirical infrastructure for studying dynamic issue competition. - PartySOME: A Comprehensive Dataset on Political Parties' SOcial MEdia Activity2025
With Luis Sattelmayer, we collected the social media activity of 498 parties in 37 countries on Facebook, Instagram, Twitter and YouTube, from the creation of each account to the end of 2023: how much they post, on which platform, and when.
498parties37countries4platforms6.7Mposts and videos2005–2023from account creationResultsPartySOME v1.0: 498 parties in 37 countries, from account creation to the end of 2023. Instagram accounts are dated by their first post. Some existing accounts, especially on Twitter after the platform closed its data access, could not be collected, so account counts are a lower bound and posting in 2023 is partly incomplete.
What the data showBigger parties↑The larger a party's vote share, the more it posts, on every platform: resources matter.
Radical right↑Radical-right parties post significantly more than all other party families on Twitter and Instagram.
Populists↑Populist parties post more on Facebook, Twitter and YouTube, but not on Instagram.
Elections↑Activity rises as elections approach on every platform, and accelerates sharply in the final months.
From regression models of monthly posts by platform, with party vote share, government participation, party family, populism and time to the next election.
ContributionMost research on parties' social media use covers one country, one platform or one election campaign. PartySOME is the most comprehensive comparative dataset of parties' social media activity, openly available, making it possible to study platform adoption, shifts between platforms, campaign dynamics, and how online activity relates to party success.
Abstract
What motivates political parties to use social media, and how does this usage correlate with outcomes such as party success? As social media platforms have become increasingly important for political communication, these questions have gained prominence in the literature on both party politics and social media politics. In this paper, we contribute to these discussions by introducing a new dataset on the social media activity of 498 political parties across 37 countries on four platforms, Facebook, Instagram, Twitter, and YouTube, from account creation to the end of 2023. The dataset provides rich aggregated data on posting volume and timing, enabling the study of long- term patterns in platform adoption, communicative patterns, and parties’ online activity. The paper presents the data, provides descriptive and inferential findings on how political parties engage with social media, and outlines several research avenues enabled by this dataset.↑ Back to overviewCite
Jan, M., & Sattelmayer, L. (2025). PartySOME: A Comprehensive Dataset on Political Parties' SOcial MEdia Activity. Party Politics. https://doi.org/10.1177/13540688251388763
@article{jan2025partysome, title = {{PartySOME: A Comprehensive Dataset on Political Parties' SOcial MEdia Activity}}, author = {Jan, Malo and Sattelmayer, Luis}, journal = {Party Politics}, year = {2025}, doi = {10.1177/13540688251388763} } - Not in the Picture: Representation and Gender Gaps in Parties' Visual Communication2026↑ Back to overview
Abstract
Research on women's political representation measures who holds office. Yet few citizens have an accurate picture of a party's parliamentary composition, and what they encounter far more directly are the images parties circulate. Party imagery is also less constrained than institutional inclusion since portraying women requires no nomination and takes no seat from anyone else. We study how parties represent women in their own visual communication, examining how often women appear (presence), in what thematic contexts (placement), and how visually foregrounded they are when they do (salience). Using computer vision on 273,000 images posted by 175 parties on Instagram across 18 countries from account creation to the end of 2023, we detect and gender-classify every face, identify party leaders using visual embeddings, and match images to the captions' policy domains. We find that women account for only 21% of all detected faces. Visual presence of women is correlated with parties' institutional gender composition but consistently remains lower. It varies across party families, and is higher under women leaders even when the leader's own face is excluded, though it reaches only a third at its highest. Where women appear, their placement is clearly gendered by policy domain, and their faces are larger and more central than men's. Parties show women prominently and centrally, but rarely. This has implications for how representation is measured and studied. In the visual communication that parties send to the public, we find a visual gender gap in which women are largely absent.