Fig. 5: Leave-One-Out distribution of the synthetic control for pilot zones.

(1) Here we iteratively re-estimate the baseline model to construct a synthetic pilot province omitting in each iteration one of the provinces that received a positive weight in Table 2; where panels a–e represent Guangdong, Guizhou, Zhejiang, Jiangxi, and Xinjiang, respectively. This sensitivity check allows us to evaluate to what extent our results are driven by any particular control province. (2) The figure displays that the iterative synthetic estimates (gray lines) for each pilot province are consistent with the original estimates and do not show large fluctuations. (3) The results indicate that the estimates of previous analysis are fairly robust to the exclusion of any particular province from our sample of comparison provinces.