For his dissertation, Felix Wallis applied natural language processing to lawmakers’ arguments in legislative debates to explain the stark divergence between the British and US responses to climate change. His research has important implications for the relationship between electoral politics and climate action.
By Felix Wallis (BSc Politics and International Relations)
Climate change is one of the most urgent threats facing modern society. If unchecked, it will increase the likelihood and severity of floods, droughts, extreme heat, and poverty. When the United Nations passed the 1997 Kyoto Protocol – the first legally binding international climate treaty – the UK and the US were in similar positions. Both had analogous socio-economic policies, comparable political values, and enormous carbon footprints. However, over the last 27 years, their environmental policies diverged, with Britain committing to cutting 100% of its emissions by 2050 and the US failing to pass any federal climate statutes until 2022. So why have these countries displayed such contrasting responses to climate change, and what can this tell us about the future of international climate action?
Political Polarisation
I theorised that differences in the level of polarisation between British and American politicians can explain their different climate policies. Based on previous research, my argument relied on the idea that when politicians’ preferences are separated into tight groups along party lines, they become less willing to negotiate and more likely to undermine policy proposals. This idea suggested that extreme differences in the environmental stances taken by Congressional Democrats and Republicans undermined the US’s ability to pass climate change legislation. Meanwhile, Parliamentary consensus about the importance of addressing climate change facilitated Britain’s world-leading environmental policy.
My analysis supported this theory by revealing that Labour and Conservative politicians in the UK delivered speeches on climate change that were more similar to each other than the equivalent speeches produced by US Democrats and Republicans. Figure 1 illustrates this trend by showing a consistently higher level of similarity among climate change speeches from Hansard, a complete transcript of everything said in Parliament, compared to the US’s Congressional Record. Nevertheless, the trend is noisy and becomes less distinct over time, suggesting that my polarisation theory is missing other important dynamics about UK and US climate policy.
Figure 1: Labour and Conservative politicians deliver speeches on climate change that are consistently more similar to each other than the equivalent speeches produced by Democratic and Republican politicians.
Electoral Risk
My second theory suggested that the divergence in UK and US climate policy reflected the different electoral risks faced by their politicians. This argument relied on the premise that legislators prioritise maximising their vote share over pursuing their preferred policy agenda when the latter might jeopardise their chance of re-election. Therefore, according to this principle, the American public’s limited support for climate action and US corporations’ anti-environmental campaign financing discouraged its politicians from supporting federal climate policy. Conversely, British voters’ homogeneous environmental views and UK corporations’ support for emissions trading allowed its politicians to make ambitious climate proposals without worrying about losing an election.
Using machine learning, my second analysis supported the electoral risk theory by showing that American legislators reduced their discussion of pro-environmental topics more than British legislators during the 60 days preceding a general election (see Figure 2 for exemplar speeches). In addition, the model supported my polarisation theory by revealing a larger gap between how much left- and right-wing politicians discussed pro-environmental topics in the US than in the UK. However, alongside these confirmatory findings, my results suggested that anti-environmental topics – mentions of ‘Climategate’, ‘global warming alarmism’, etc. – also receive less attention in the lead-up to an election. This unexpected finding raises questions about the mechanism through which electoral risk affects climate policy, an area I encourage researchers to investigate in the future.
Examples of UK and US speeches contained within a pro-environmental topic identified by the machine learning model
UK Speeches
I agree with the hon. Lady’s point that volcanos have an effect on our climate. She has not mentioned—perhaps she was about to—that the emission of large amounts of dust from volcanos into the upper atmosphere also has a shielding effect on the earth. However, if we remove the effects of volcanos by mathematical means, we must agree that global warming is still taking place.
As has already been made clear, it is not in the IPCC reports, which err on the side of conservatism and are entirely consensual. However, everybody knows that there are step-change events that could completely transform the situation—for example, if the Amazon rainforest burned or if the Arctic tundra melted and released the enormous quantities of methane trapped in it; methane is a powerful greenhouse gas that would ratchet up the greenhouse effect enormously.
US Speeches
Thank you. Madam Speaker. My colleagues. here is a quote from a Newsweek article: “There are ominous signs that the Earth’s weather patterns have begun to change dramatically. and that these changes may portend a drastic decline in food production. with serious political implications for about every nation on Earth. The drop in food output could begin quite soon. perhaps only 10 years from now.” My…
Mr. Speaker. on Friday. the Intergovernmental Panel on Climate Change published their latest report. confirming that climate change is happening and that it is the result of human activity. The report was produced by 259 scientists from 39 different countries over the last 6 years. and it is the most comprehensive and authoritative assessment of the Earth’s changing climate. The report shows that c…
Implications for climate policy
My analysis found strong evidence of polarisation harming American climate policy and weaker evidence of electoral risk’s malignant effects. These findings have two broader implications for countries’ domestic climate policies and the development of international environmental agreements.
First, by showing that polarisation is linked to poor climate action, my findings suggest that countries’ environmental policies depend on two factors: having clear scientific evidence about the effects of climate change and having politicians who are willing to work together to address the issue. As such, even though the evidence for climate change is unequivocal, we should not expect countries to work together to reduce their emissions unless politicians from competing parties agree on its importance.
Second, my results suggest that domestic politics may pose a larger challenge to climate action than international negotiations. A particular concern is the potential for domestic politicians to stop worldwide emissions reduction simply because they want to obstruct their opposition. This likelihood makes it especially tricky to coordinate countries’ climate policies, as all their diplomats and members of parliament or congress must agree on any proposed collective action. As such, it may be preferable for countries to focus on individually doing as much as they can to prevent climate change, rather than trying to coordinate their actions with international partners.
Ultimately, my dissertation concluded that policymakers must prioritise bipartisanship to limit global emissions. With diminishing opportunities to prevent the irreversible impacts of climate change, this bipartisanship will become increasingly important for maintaining the health of human societies and Earth systems.
The views expressed in this post are those of the author.