The Value of Accurate Weather Forecasts : Social Sentiment Responses Reflected in Social Media in China

This paper combines real weather data with the city-daily weather forecasts broadcast by the government, as well as the sentiment index expressed by posts on the popular social media Weibo in China, and through an interactive regression design analyzes the differential sentiment responses to temperatures under different sizes and signs of daily temperature forecast errors. My main results have suggested that more accurate temperature forecasts lead to smaller shifts towards unhappiness caused by the cold temperatures. The same effect does not play out under hot weathers, unless in cities with higher income, greater long run temperature forecast accuracy, or during holidays. My study also suggests that additional negative sentiment shocks are likely related to cold or heat alarms issued according to the national forecasts, resulting in that positive forecast errors have greater marginal effects on sentiment than negative errors during the cold temperatures. Overall, these results meet the intuition that advanced forecast technology provides more accurate daily temperature forecasts, and adds to great social benefits in China in terms of improving people’s subjective well-beings as expressed by social media sentiments. In the current time under climate change, when extremal weather events are expected with greater frequency into the future, my work would help to provide an insight of the value of developing a modern weather forecasting system that can benefit billions of people in the long run