Replication package for: Repricing Ahead of the Trade Data: AI Task Substitution, Services-Export Exposure, and the Market–Fundamentals Lag in Emerging Markets

Replication package for the manuscript "Repricing Ahead of the Trade Data: AI Task Substitution, Services-Export Exposure, and the Market–Fundamentals Lag in Emerging Markets" (submitted to Emerging Markets Finance and Trade). The package contains all raw data extracts, configuration files, derived datasets, and Python code reproducing every table, figure and in-text estimate in the paper, including the online appendix. All data are from free, public sources: equity prices from public market-data endpoints (US and NSE listings), quarterly balance-of-payments series from the IMF SDMX 3.0 API (BPM6 standard components), World Bank WDI indicators, occupational exposure scores from Eloundou et al. (2024) and ILO Working Papers 140/166, and vendor revenue and headcount hand-collected from company filings with per-firm-year source records. Usage: unzip and run `python3 code/run_all.py` (offline; rebuilds all results from the included raw data in roughly 30-45 minutes). `run_all.py --fetch` optionally re-pulls the raw data first. The README maps each program to the outputs it produces, states software versions, and describes the integrity checks that enforce agreement between manuscript, appendix and derived data. Analyses are deterministic (fixed seeds); results are bit-reproducible.