Secular Sweden – the Foxhole where Atheists Thrive

In times of crisis, when in a life-threatening situation, or when exposed to the threat of serious illness, disability, or the mental and physical distress of losing, or nearly losing, someone close, research finds that people often turn to God (McCrae 1984; Bjork and Cohen 1993; Hood et.al 1996). Thus, as such situations de facto tend to trigger religious beliefs, there are no atheists in foxholes, as the saying goes. In this paper we challenge the state of the art by presenting the first large-scale, individual-level and long-term study of the impact of a major natural disaster on religiosity in a highly secular context – Sweden. In addition, we extend the focus of previous research into the political realm by investigating the long-term party-support for the political party of the Christian Democrats (KD) among natural disaster survivors. To do this, we rely on two data sources. First, we utilize unique and detailed individual-level administrative data of approximately 16,000 Swedish tourists who survived the 2004 Boxing Day-tsunami, that hit South-East Asia in 2004. We match this data against other individuals in the entire Swedish population registry on several pre-tsunami characteristics, allowing us to better isolate the effect of disaster exposure on measures of religiosity – collected via a survey. The survey is administrated to 10,000 individuals in our data (half of which were exposed to the tsunami), allowing for more nuanced long-term measures of religiosity and control variables, including exposure severity to the tsunami. In combination, these data allow us to study the long-term effect of natural disaster exposure on religious beliefs in a completely novel and unique way.  In summary, the tsunami survivors tend to believe less in God and they have a generally lower trust in the Church of Sweden, than their comparable but un-affected counterparts, almost 20 years after the disaster. They also attend religious service less, and are to a much lesser extent members of a religious organization. When it comes to the political party of the Christian Democrats (KD), there is, however no effect, neither regarding popularity or in terms of voter support. The party is equally uninteresting among both groups. These effects are all significant at the 95%-level and are driven by those who were severely exposed to the tsunami. Our study adds to the research field in several important ways. First, we are the first to present evidence of a negative effect on religiosity in the wake of having survived a life-threatening situation. Second, we present the first investigation into the long-term effects, 20 years after the disaster. Third, we are the first to extend the research field into the political realm. And fourth, our research design, methodological approach and unique individual-level data, allow us to aim for causal claims.