Replication package for: When Is a Move in Long-Run Expectations Significant? Anchoring and Short-Horizon Components in Mexico

Data and code that reproduce every table, figure and number in the paper "When Is a Move in Long-Run Expectations Significant? Anchoring and Short-Horizon Components in Mexico". A bivariate state-space model, estimated by maximum likelihood with the Kalman filter on monthly data for Mexico (January 2002 to August 2026), separates a latent anchoring gap, measured by long-run survey expectations of Banco de México's expert survey, from a persistent short-horizon component of twelve-month expectations, and attaches to the anchoring gap a quasi-real-time band that includes hyperparameter uncertainty. The package contains the raw data (Banco de México SIE, INEGI, FRED), the Python scripts that build the analysis data set, the R pipeline (run_all.R) that produces all estimates, robustness checks, forecasts, detection exercises, tables and figures, and all outputs. See README.md for software requirements, data sources and step-by-step instructions. The VIX series (Cboe) is not redistributed; README.md explains how to download it.