Air quality and tourism orders: a quasi-experimental study

Using a unique dataset collected from a Chinese online travel agency, we empirically examine the causal effect of air quality on tourism orders by employing a two-step regression-discontinuity design (RDD). The RDD serves as a local randomisation approach that allows us to eliminate the endogeneity concern of air quality. We control for meteorological conditions, city and month fixed effects, and additional social factors specific to the tourist destination. Our empirical results show a range of 49.3% to 51.9% drop in travel orders on the day at the cutoff point where air quality breaks through light to moderate pollution levels (AQI = 150). Furthermore, our empirical results suggest that tourists who select different orders may respond differently to air quality. Our research provides critical empirical evidence, elucidating the additional cost incurred in the tourism sector due to poor air quality.