His APSR paper misreported the main regression, where he claimed to have included fixed effects, but did not. Running the regression as reported completely kills his findings. Moreover, other scholars have uncovered serious issues in Kung's dataset: https://www.tandfonline.com/doi/pdf/10.1080/13603116.2014.882560
His QJE paper, again, completely manipulated the politician promotion data. Once these obvious data errors are fixed, the corresponding findings completely go away: https://michaelwiebe.com/blog/2021/02/replications
And for most of his other economic history papers, he never shares his datasets, keeps asking for exemptions from the journal data policies despite the fact that his datasets are digitized from historical sources and are not even proprietary. When people email him asking for datasets for replication purposes, he keeps coming up with excuses not to share. Very suspicious.
People like these are the reason why no one trusts empirical works anymore.