In the spotlight
The short- and long-term effect of a nuclear disaster on voting outcomes
What are the political effects of a nuclear disaster? In this paper (Mehic 2023a), I use data from the 1986 Chernobyl accident to evaluate the voting outcomes of the Swedish Green Party (MP), which was founded in 1982, and elected to the national parliament in 1988, two years after the disaster.
By Adrian Mehic
There are several reasons for why this question is important. First, with climate change, extreme weather, sudden temperature fluctuations, and other adverse effects on local environments are likely to become more salient. A second reason is the importance of nuclear power; one of the chief arguments against its use is the risk of accidents. Finally, the Chernobyl accident, being one of the major events of post-World War II Europe, had profound political consequences in most Western countries. Importantly, in several countries, green parties were elected to national parliaments, and in France, Finland, and Germany, these parties were junior members of government coalitions already by the mid-1990s. As a direct consequence of the accident, the Italian government decided to begin to shut down the country’s nuclear power plants in 1988. The Italian decision to abandon nuclear power is very similar to the one made by the German government after the 2011 Fukushima disaster.
In Swedish politics, the future of nuclear power has been a major issue for several decades. Together with the fact that the election of new parties to the Swedish parliament is exceptionally rare – the MP’s election to parliament in 1988 was the first time since 1932 that a new party entered parliament, this makes the setting in the paper interesting also for Swedish political history.
Due to rainfall and adverse wind conditions, Sweden was the second-most affected country by Chernobyl fallout, after the Soviet Union. This can be seen as a natural experiment, providing exogenous variation in the level of fallout between municipalities. Immediately after the accident, authorities conducted large-scale aerial measurements of fallout. It is therefore possible to measure the level of fallout in each of the 284 municipalities. In five municipalities, the average level of local fallout exceeded 37 kilobecquerels per square meter, an internationally utilized threshold to define an area as ‘contaminated’. These five municipalities are of particular interest in the analysis.
In the paper, I study how the local variations in fallout contributed to support for the MP using data from parliamentary elections. The at the time recently established MP had a high-profiled anti-nuclear agenda. Thus, we would expect to see higher vote shares of the MP in fallout-affected areas. I can thus estimate the “Chernobyl effect” for each election year and for each municipality; that is, the share of the MP vote that can be explained by local variations in fallout. The results show that there was a significant increase in MP voting in contaminated municipalities. One standard deviation higher fallout, increased MP voting by around 0.2 standard deviations. While there is a positive Chernobyl effect in all elections from 1988 to 2018, the coefficient was only statistically significant until the 1994 elections. Looking specifically at the five most-affected municipalities, I find that the MP vote share was around one percentage point higher than it would have been if they had not been affected.
To examine mechanisms, I combine the data on voting outcomes and on local fallout levels with data from an annual survey randomly distributed to households. This allows me to examine attitudes towards nuclear power. To measure pre-Chernobyl attitudes to nuclear power, I use municipality-level voting outcomes from the 1980 referendum on the future of nuclear power in Sweden.
These results reveal several clues about the mechanisms behind the findings on the MP voting. First, the decline in support for nuclear power after Chernobyl was more pronounced in high-fallout areas, which is in line with the increase in MP support. The decrease in support for nuclear power can, thus, be explained by voters in fallout-affected areas reporting to be significantly more worried about adverse consequences of nuclear power. Notably, however, respondents in high-fallout areas were particularly concerned about the adverse consequences for health, and air and water quality. They also worried about problems for future generations, which may relate to problems with long-term final storage of nuclear waste, rather than to the risk of a new accident. Another finding is that voters who frequently read the local newspaper were more worried than respondents in general. This is likely due to the heavy focus placed by local media in affected areas. As an example of this, the words “cesium” and “Chernobyl” are still (2019) over-represented in local print media in fallout-affected municipalities.
Overall, the findings in the paper suggest that the Chernobyl accident had profound effects on voting outcomes, and that the effect on the MP vote remained for almost a decade after the accident.
More from the author
This research was undertaken as part of the author’s PhD studies at the department of Economics at Lund University (Mehic 2023b). The thesis includes three additional chapters. One examines the role of the 2015 refugee crisis in explaining the increase of the vote share of the anti-immigration Sweden Democrat party at the local level. This chapter shows that while there is an overall positive effect of immigration on the anti-immigration vote, the results are heterogenous depending on the pre-influx characteristics of the municipalities. The second thesis chapters is on the heterogenous consequences of hybrid education at the university level. It shows that women from high-socioeconomic (SES) families benefit from this mode of instruction. The final chapter examines the role of SES and beauty in explaining grade outcomes in university. High SES and beauty affect own and peer grades positively. As potential mechanisms, the study finds evidence of both a direct spillover mechanism, as well as an indirect, marriage-market channel, through which exposure to physically attractive peers increases the well-being of students not currently in a romantic relationship.
Mehic, Adrian (2023a). “The Electoral Consequences of Environmental Accidents: Evidence from Chernobyl.” Unpublished manuscript.
Mehic, Adrian (2023b). Essays in Political Economy and Economic Sociology. Lund: Lund Economic Studies 236.