These codes come without technical support of any kind. The code is free to use, provided that the paper is cited properly.

Codes based on M. Marcellino & M. Pfarrhofer (2025): "Nonparametric Mixed Frequency Monitoring Macro-at-Risk" Economics Letters 255 112498, using state space MF-VARs estimated with precision sampling (including MF-BART). See also mf-bavart for related work.

Estimation is integrated with the alfred R-package to access real-time data from alfred.stlouisfed.org for the United States.

• setup_mfvar is the main estimation file which downloads and transforms data (requires user input)
• est_mfvar estimates the model and produces outputs (is sourced automatically from setup_mfvar)
• utils contains several helper functions (is sourced automatically from est_mfvar)

• Data settings are described in the file
• run_mean refers to the conditional mean estimation and approximation, choose from:
  • bart-hs estimates a BART model and uses the shrinkage-based linear approximation
  • bart-proj estimates a BART model and uses the projection-based approximation
  • lin estimates a linear BVAR model with horseshoe prior
• outlier refers to whether a homoskedastic (when FALSE) or outlier (when TRUE) specification is estimated

Model outputs (MCMC draws for the data) are saved in the directory results and associated charts for variables defined in output_var are collected in the directory plots.