Poor air quality has well-documented adverse impacts on health and the environment, but also causes important psychological, economic and social effects. A high level of air pollution can impair cognitive functioning and trigger negative emotions such as anxiety and anger. Recent findings suggests that these psychological impacts have knock-on effects on people’s decision making. In a new IZA paper, Luna Bellani, Stefano Ceolotto, Benjamin Elsner and Nico Pestel study the effect of air pollution on a high-stakes decision made by millions of people at a time, namely voting in parliamentary elections. In elections, people decide on the same issue on the same day but in different locations, which means that they are exposed to different levels of air pollution when making their decision.
Incumbent parties lose support on days with dirty air
The study uses county-level data on 64 federal and state parliament elections in Germany between 2000 and 2018, combined with daily measures of air pollution and weather conditions. The measure of local air pollution is the daily average concentration of particulate matter (PM10), one of the most frequently used indicators for suspended particles in the air. The outcome of interest is the vote share of the parties forming the incumbent government coalition on the day of the election. Voting for incumbent parties can be seen as an expression of support for the status quo. Relative to voting for opposition parties, voting for the current government also represents the less risky option.
The paper’s main finding is a negative effect of air pollution on the vote share of the incumbent parties and a corresponding increase in the vote share of established opposition parties. An increase in the ambient concentration of PM10 by ten micrograms per cubic meter – an increase that would commonly occur in many German cities – reduces the vote share of the incumbent parties by two percentage points and increases the vote share of established opposition parties by 2.8 percentage points. These are large effects, given that the typical drop in support for the incumbent government is around five percentage points.
Effect of air pollution driven by voters’ unhappiness with current government
The authors document similar effects in two large-scale representative surveys from Germany. One is a well-known monthly opinion poll carried out on behalf of German public television (Politbarometer). The results show that on days with higher pollution in a respondent’s region, respondents report a lower intention to vote for the incumbent federal government and a greater intention to vote for the opposition. At the same time, the results indicate a lower approval of the current government’s policies, while approval of the opposition is unaffected.
A second piece of evidence comes from the German Socio-Economic Panel Study (SOEP). Again, on interview days with higher air pollution, respondents show weaker identification with the current federal government and stronger identification with the opposition.
Emotions as a plausible mechanism affecting decision making
Using the SOEP survey data, the authors provide evidence that emotions are an important explanation why voters respond to poor air quality. On days with elevated levels of air pollution, respondents are more likely to be worried, feel angry and sad, and are less likely to feel happy. In contrast, there is no evidence that air pollution affects people’s perceptions of the current state of the economy or their own economic situation.
The authors view their findings as evidence of a behavioral bias. At the levels measured in Germany, the concentration of PM10 is not directly noticeable. People may notice a higher concentration through symptoms such as irritation of the airways or coughing. But since the same symptoms could also be caused by factors other than particulate matter, it is unlikely that people deliberately choose to change their voting behavior because they are exposed to high air pollution. A more likely explanation is that air pollution has an unconscious effect on voting, for example by affecting a person’s emotions or health, which in turn affects how they process information and make decisions.