Terrorism, Government Surveillance and Individual Well-Being
Saturday, Jan. 7, 2017 8:00 AM – 10:00 AM
Swissotel Chicago, St Gallen 3
- Chair: Solomon W. Polachek, Binghamton University
The Long-Term Costs of Government Surveillance: Insights from Stasi Spying in East Germany
AbstractBased on official records from the former East German Ministry for State
Security, we quantify the long-term costs of state surveillance on trust, social ties and
economic performance. Using county-level variation in the number of informers in the
1980s, we exploit discontinuities at state borders to show that higher levels of government
surveillance led to lower levels of trust and a reduction in social interactions in postreunification Germany. Based on a second, complementary identification strategy that
accounts for county fixed effects, we further estimate the economic costs of spying. We find
that more intense surveillance caused lower self-employment rates, higher unemployment
rates and larger out-migration throughout the 1990s and 2000s. We provide evidence
that our effects are directly driven by government surveillance and not due to alternative
mechanisms. Overall, our results suggest that the social and economic costs of East
German state surveillance are large and persistent.
Individual Well-Being and the Allocation of Time Before and After the Boston Marathon Terrorist Bombing
AbstractWe quantify the effect of the 2013 Boston marathon bombing on Americans’ emotional feelings and time uses, exploiting data from the American Time Use Survey and Well-Being Module that was collected continuously in the days around the terrorist attack. We find that the bombing led to a significant drop of about 1.5 points in well-being (on a scale of one to six) for residents of the States close to Boston. This effect is much larger in size than the estimates in the literature of the reduction in subjective well-being due to a one percentage drop in GDP or to unemployment. The happiness of American women also fell significantly, by almost a point while we find no impact for American men overall. This is possibly due to stronger aversion to risk for women than men, as documented in some of the behavioral literature. According to our estimates, feelings of stress also increased significantly due to the bombing, while labor supply and other time uses were not affected, except for leisure hours that dropped in the aftermath of the bombing –in line with the prediction that outdoor activities, potentially exposed to terrorism, will drop. In contrast, we find no well-being or other time use effect of the deadly shootings at the Sandy Hook school a few months earlier, confirming that terrorism has a stronger impact on individuals than other violent deaths. In the long-run though the effects of the bombing vanish: terrorism has a sharp, but not permanent, impact on subjective well-being in countries where such attacks are rare.
Covering the Campaign: Automated Extraction of Election Events in 2014 South Africa
AbstractWhat constitutes coverage of "election events" by media during political campaigns? Advances in computation allow for automated coding of events in International Relations, but have thus far largely been confined to conflict events (e.g., GDELT). However, classifying events in the electoral context is poses new difficulties because what constitutes an "election event'' depends on theories of how, and to what extent, media events persuade voters during campaigns. Prior approaches in democratic politics differ in their conceptualization of the role of media in campaigns and how it affects political behavior and electoral outcomes, establishing no consistent framework for defining election coverage. We apply machine learning methods to classify election-related events in a corpus of more than 200,000 news stories and social media posts related to South Africa's 2014 election. We use a theoretically informed classification of election coverage to demonstrate how variations in definitions produce variation in "election-related'' or "election salient'' codings on a host of campaign activities (including violence, protest, and riots) by the media and social media, resulting in radically distinct representations of the electoral landscape. We discuss the challenges and opportunities this research and method pose for replication to study (electoral and non-electoral) events in other settings.
University of Rochester
- F5 - International Relations, National Security, and International Political Economy