CGA Guest Speaker: Juan Palacios, PhD
Economic behavior is driven by mostly unobserved factors, such as personality traits or affective states, which determine preference orderings and individual behavior. While economists and psychologists have developed in the past decades numerous experiments and survey instruments to measure those individual attributes, there is still a scarcity of scalable datasets available for researchers and policy makers to connect subjective traits with real world behaviors. This study uses social media platforms as an unsolicited poll to infer individuals' emotional state, and link it to individual behavior. We construct a unique dataset including the universe of posts [N=18.55 Million] of a cohort of over 500,000 individuals located in China from January 1st 2019 to June 30th 2020, using micro data from Sina-Weibo, the largest social network in the country. We use state$-of-the-art Natural Language Processing (NLP) tools to compute the emotional state expressed in each post by the subjects in our sample. Using this data, we classify individuals based on their average degree of fear expression in their social media posts during the peak of a COVID-19 virus outbreak in the country, to reveal their underlying latent emotional state, and then link it to their visitation rate to out-of-home venues using a difference-in-differences design. The results display the predictive power of fear on behaviors. The effects are quantitatively significant and long-lasting: individuals with an average levels of expressed fear above their city median display a 3.9% higher reduction in all out-of-home activities than individuals with levels below median fear within the same city for months in the aftermath of COVID-19 outbreaks. High risk consumption places like restaurants and indoor entertainment showed the largest drops in visits-14.9% additional reduction of high fear people, which remained stable for months.
Juan Palacios is currently a post-doctoral researcher at the Massachusetts Institute of Technology, in the Center for Real Estate. His research focuses on environmental economics, real estate and health economics. He is currently focusing on the impacts of environmental hazards, such as air pollution or airborne diseases, on the performance, well-being and behavior of urbanities. His work has been presented in international conferences such as Urban Economic Association, American Real Estate and Urban Economic Association, or the American Economic Association. Juan has been involved in a number of projects funded by institutions like Google, The U.S. General Services Administration, and the European Commission.
To view the presentation, click this Zoom meeting link with Password: 694239. Or join by telephone: +1 312 626 6799.