Code and data for "Macroeconomic shocks, downside risk and regional heterogeneity in China's pilot carbon markets: Evidence from a quantile factor-augmented vector autoregression"

This record contains all data and MATLAB code needed to reproduce the results of the paper. The Excel workbook data/QFAVAR_data.xlsx holds: • the raw daily carbon market data (average transaction price, volume and turnover) of the seven Chinese pilot carbon markets: Beijing, Tianjin, Shanghai, Chongqing, Hubei, Guangdong and Shenzhen; • end-of-month stock market values and the constituents of the carbon-neutrality concept; • the macroeconomic series (producer price index, industrial value added, manufacturing PMI, consumer expectations index and electricity consumption), all downloaded from the iFinD database (Tonghuashun); • the constructed monthly series for January 2015 to July 2026, the model input and the robustness data sets. The MATLAB code (written and tested in R2022a) rebuilds the monthly data from the raw downloads and estimates the quantile factor-augmented vector autoregression (QFAVAR) of Korobilis and Schröder (2025) by Gibbs sampling. It runs the baseline model, 13 robustness checks, the out-of-sample forecast evaluation and the specification comparison, and produces every table and figure that is based on data. Each step can be switched on or off in MAIN.m; steps that are not run again take their results from results_paper/, which contains the output of the runs reported in the paper. See README.txt for instructions.