Mastering Empirical Macroeconometrics

A Practitioner’s Guide from Panel Data to Structural Models

A comprehensive hands-on curriculum for learning empirical macroeconometrics — panel econometrics, local projections, VARs, difference-in-differences, Bayesian methods, DSGE estimation, and causal machine learning.
Author
Affiliation

Bijoy Ratan Ghosh

University of Virginia

Published

January 2026

Welcome

This is a comprehensive, hands-on curriculum for learning empirical macroeconometrics—from foundational panel data methods through Bayesian VARs and structural DSGE estimation.

Who This Is For

This curriculum is designed for:

  • PhD students in economics who want deep understanding of the methods they use
  • Applied researchers who want to go beyond “running the code” to understanding why methods work
  • Policy economists at central banks and research institutions

Philosophy

Each module follows a consistent structure:

  1. Theory Block: Mathematical foundations, key theorems, identification assumptions
  2. Monte Carlo Lab: Simulate known DGPs → estimate → verify properties (coverage, power, bias)
  3. Applied Example: Apply methods to real macroeconomic data
  4. Diagnostics Checklist: What to check, what can go wrong, red flags

The goal is not just to use these methods, but to understand them deeply enough to defend your choices to a skeptical audience.

Curriculum Overview

Phase Topic Modules Language
1 Foundations Panel Econometrics, Identification R
2 Dynamic Methods Local Projections, VAR/SVAR R
3 Treatment Effects Staggered DiD, Synthetic Control R
4 Bayesian Econometrics Foundations, BVAR, Panel Methods Python
5 Structural Macro DSGE Foundations, Estimation MATLAB/Python
6 Causal ML Regularization, Causal Forests Python

Motivation

Modern empirical macroeconomics requires fluency in a wide range of methods. A single research project might require:

  • Panel regressions to establish baseline correlations across countries
  • Local projections to trace dynamic responses to shocks
  • Staggered difference-in-differences to handle heterogeneous treatment timing
  • Vector autoregressions to characterize transmission mechanisms
  • Bayesian methods for shrinkage and structural identification
  • Machine learning to uncover heterogeneous effects

Rather than treating these as black boxes, this curriculum develops the intuition needed to choose, implement, and defend each method.

Getting Started

Start with Module 1: Panel Data Econometrics and work through sequentially. Each module builds on previous concepts.

TipReproducibility

Most examples use publicly available macroeconomic data (Penn World Table, IMF IFS, World Bank WDI). Code is fully reproducible—clone the repository and render locally.

Technical Requirements

  • R (≥4.0) with packages: fixest, lmtest, sandwich, ggplot2, dplyr
  • Python (≥3.9) with packages: numpy, pandas, statsmodels, pymc, arviz
  • MATLAB (optional) for DSGE modules, or use Dynare with Octave
  • Quarto (≥1.3) for rendering