Replication Data for: Fulling Discontent: The Politics of Carbon Taxation in Canada

Research hypotheses: Trade-sensitive and carbon intensive sectors and workers are more likely to reject carbon taxation than those who are sheltered and cleaner. Individual level dataset: Data from the 2019 and 2021 Canadian Election Study (CES). Assess the relationship between respondents' employment profile and their views on carbon taxation and environmental policy more broadly. Respondents are geocoded at the provincial level and are asked to specify in writing their occupation. Manually coded respondents' 2-digit NAICS occupation classification and matched them with the trade and emission-intensity data. Employ modified specification to account for job-specific trade-sensitivity (1) and carbon-intensity (2). Data on sector-specific GHG emissions and exposure to trade are provided by Statistics Canada. Ecological data: To account for job and constituency-specific carbon emissions, I leverage Canada's GHG inventory compiled by Statistics Canada. The inventory breaks down yearly emissions by NAICS industrial sectors at the provincial level. Using Statistics Canada's employment data, I first divide our emission data by industrial and provincial employment profiles in order to account for each jobs's carbon intensity at the provincial level. I then match these data with constituency-specific employment profiles to account for overall emissions at the riding-level. Insofar as employment data at the constituency-level are drawn from the national quinquennial census of 2011, 2016 and 2021, I perform linear extrapolations in order to balance the dataset, thus covering the full 2015-2021 period. Another factor behind \"carbon layoffs'' is exposure to trade. I first leverage Statistics Canada's trade flows data. Trade flows are provided both at the industry-level on a province by province basis, whereas employment profile is broken down at the constituency-level. Following Yamazaki (2017), I build a constituency-specific trade-sensitivity index. Data on constituency-level sociodemographic characteristics --unemployment, minority share, BA share, car commuters' share, household income and population density-- all come from the Canadian Census. Linear extrapolations are performed in order to balance the dataset. I leverage Election Canada's constituency-level data to account for the vote share captured by anti-carbon tax parties (the Conservative Party and the People's) between 2015 and 2021. Regional gas prices stem from both Kalibrate and Statistics Canada. The yearly share of non-carbon energy sources (hydro, renewable, nuclear) in the provincial electricity mix is provided by Statistics Canada. Finally, constituency-level belief in the existence of anthropogenic climate change comes from the 2016 and 2018 Canadian Climate Opinion Maps, from the Yale program on Climate Change and Communication.