A news-based climate policy uncertainty index for China

Using a deep learning algorithm, the MacBERTmodel, this study constructs indices of China's climate policy uncertainty (CCPU) at the national provincial, and city levels for the first time. The CCPU indexes are based on the text mining of news from a set of major newspapers in China.Our results show that these indices can effectively depict the dynamic evolution of CPU and regional heterogeneity in China. The CCPU dataset canprovide a useful source of information for government, academics, and investors to understand the dynamics of the climate policies in China. These indices can also be used to investigate the empirical relationship between CPU and other socio-economic factors in China.