Transboundary vegetation fire smoke and expressed sentiment: Evidence from Twitter

Abstract

This paper examines the impact of transboundary vegetation fire smoke on the real-time sentiment of Twitter users in Southeast Asia, including countries such as Brunei, Indonesia, Malaysia, Philippines, Singapore, Thailand, and Vietnam. We leverage the exogenous variation in wind directions for identification. We find that an increase in upwind fires by one standard deviation reduces the sentiment score by 0.5 percent of a standard deviation (after netting out the impact of unobserved local socioeconomic factors). During peak fire seasons, our estimate translates into sentiment damages comparable to the average Sunday-to-Monday sentiment drop. The adverse sentiment impact exhibits significant variation across countries and intensifies with factors such as the number of upwind fires, income levels, proximity to fires, and limited adaptability on weekdays. We show that cross-boundary air pollution is the primary channel, with smoke from neighboring countries exerting a greater impact on sentiment than domestically produced smoke. These findings underscore the psychosocial costs and geopolitical tensions associated with cross-border air pollution spillovers.