Local projections for applied macroeconomics

This repo contains lecture materials for a short course (~5 hours) on the practical implementation of local projections for applied macroeconomics.

The slides emphasize hands-on coding tools and intuitive concepts rather than math. I prioritize simplicity of implementation in Stata over cutting-edge econometric performance (but I give references to cutting-edge papers). My running example is based on the empirical application in Bilal & Känzig (2026).

This is not a first course on local projections. I assume basic familiarity with concepts such as impulse responses and some exposure to basic local projections and structural VARs. Good additional resources include Diego Känzig's lecture slides, the chapters in the 2016 Handbook of Macroeconomics by Valerie Ramey and Stock & Watson, and the 2025 JEL article by Jordà & Taylor.

The materials have not yet been thoroughly vetted. If you spot any errors or confusing parts, I would be grateful if you could create a GitHub Issue or email me.

Contents

• Lecture 1 (Stata do-file)
  • Control variables, identification by timing restrictions
  • Standard errors, pointwise confidence intervals
  • Variance decompositions
  • Proxies/instruments
  • Other identification schemes
• Lecture 2
  • Comparison between local projections and VARs
• Lecture 3 (Stata do-file)
  • Simultaneous confidence bands
  • Shrinkage estimation
  • Bias correction
  • Bootstrap
• Reading list

Omissions

The course currently does not cover the following important topics, though that may change in the future:

• Basic introduction to dynamic causal effects and local projections
• Introduction to structural VARs
• Panel data
• Nonlinear specifications
• Weak instruments

Acknowledgements

This material is based upon work supported by the NSF under Grant #2238049 and by the Alfred P. Sloan Foundation.