The U-shaped cost of digital innovation: Behavioural adaptation in 1,960 Chinese innovation-board firms

Replication package for the manuscript 'The U-shaped cost of digital innovation: Behavioural adaptation in 1,960 Chinese innovation-board firms'. Contents: Analysis_Code/: 14 Python scripts implementing the full estimation pipeline (baseline panel regressions, U-shape turning points and Fieller confidence intervals, Lind-Mehlum tests, instrumental-variable endogeneity checks, heterogeneity and robustness analyses, TFP estimation via OP/ACF, extended data tables and figures). Results/: 12 JSON result files containing every number reported in the manuscript tables and extended data (baseline, robustness, heterogeneity, endogeneity, Lind-Mehlum tests, TFP, and the authoritative manuscript-numbers file). README.md: package structure and reproduction instructions. All coefficients and turning points were cross-validated in four software stacks to 1e-8; clustered standard errors agree within 0.02%. Primary financial data are publicly available through the East Money datacenter-web interface; raw scraped records are not redistributed per the provider's terms of service.