Air Quality and Corporate Bond Spreads in China: Identification and Forecasting

This repository contains the replication package for the paper “Air Quality and Corporate Bond Spreads in China: Identification and Forecasting.” The dataset includes processed monthly panels of Chinese corporate bond yields, air quality indices (AQI), precipitation, and industrial electricity consumption for the period 2014–2024. All variables have been cleaned and merged to facilitate replication of the econometric and machine learning analyses presented in the paper. The accompanying Python scripts reproduce the main empirical results, including OLS, 2SLS, and difference-in-differences regressions, as well as the large–small model collaborative forecasting framework integrating econometric identification with machine learning prediction. Due to data licensing restrictions, raw bond yield data from iFind and ChinaBond cannot be shared; however, all processed panels and code necessary for replication are provided.