HEFM

Overview

HEFM stands for Herramienta de Evaluación Fuera de Muestra — or Pseudo Out-of-Sample Evaluation Tool (POSET) in English.

This platform was developed in 2021 in response to a key question: How can we assess whether a given SVAR model is a reliable guide for macroeconomic analysis? Since then, various methodological approaches have been designed to quantify both the predictive accuracy and explanatory robustness of these models.

The site presents a concrete implementation of that effort, grounded in statistical model selection techniques, particularly time series cross-validation. The platform enables users to evaluate forecast errors and examine the consistency of structural responses across different estimation windows.

Purpose

HEFM serves as a tool to diagnose the performance of SVAR models along two critical dimensions: their ability to predict key macroeconomic variables out of sample, and their ability to explain the empirical regularities in the data through structural responses.

It does this by combining pseudo out-of-sample forecasting exercises with analysis of the dispersion in impulse-response functions (IRFs) across multiple estimation windows. This dual focus allows practitioners to identify models that are not only statistically robust, but also consistent with the theoretical causal mechanisms embedded in their structure.

For a deeper look at the underlying assumptions and implementation details, see the Resources section.

Documentation and Resources

Looking for the economic logic behind the tool? Visit the Resources section.