I am pleased to announce that I will deliver a two-day online course with Timberlake on Local Projections for Time-Series and Panel Data Analysis: Methods, Applications, and Replication in Stata, on 27–28 August 2026. The course will take place online via Microsoft Teams and will focus on the practical use of local projections for applied macroeconomics, finance, and policy-oriented empirical research.
Local projections have become an essential tool in empirical macroeconomics. Since Jordà’s seminal contribution, they have offered a flexible way to estimate impulse responses horizon by horizon, without requiring the full specification and estimation of a multivariate dynamic system such as a VAR. This makes them particularly attractive for applied researchers interested in dynamic responses to shocks, nonlinearities, heterogeneous effects, and policy interventions.
The course is designed as a practical and applied introduction to local projections. We will begin with the foundations of time-series local projections, before moving to panel local projections, state-dependent and time-varying specifications, and LP-DiD approaches for dynamic treatment analysis in staggered intervention settings. Throughout the two days, the emphasis will be on the links between econometric design, identification, inference, interpretation, and replication in Stata.
Several applied examples will be discussed during the course. These include the estimation of impulse responses in time-series settings, panel local projections with fixed effects and clustered inference, the role of international reserves in responses to terms-of-trade shocks, and the dynamic response of interest rates to geopolitical risk shocks. We will also discuss common pitfalls in local projection designs, including weakly identified shocks, loss of precision at long horizons, interaction terms, regime dependence, and graphical communication of dynamic effects.
A central objective of the course is to make local projections directly usable for applied work. Participants will work with Stata workflows based on lpirf, locproj, and LP-DiD approaches. The course will combine methodological discussion, empirical examples, and hands-on replication exercises, with the aim of helping participants design, estimate, interpret, and report credible local projection analyses.
The course is particularly relevant for master’s and PhD students, applied researchers, research assistants, analysts, and policy researchers working with macroeconomic, financial, or panel data. Prior exposure to econometrics, regression analysis, and basic time-series or panel-data methods is recommended.
Registration details are available on the Timberlake course page. The sessions will be recorded and made available to registrants for 90 days after the course ends.
Course page and registration:
Local Projections for Time-Series and Panel Data Analysis: Methods, Applications, and Replication in Stata
27–28 August 2026, online via Teams, delivered by Prof. Jamel Saadaoui, Université Paris VIII Vincennes-Saint-Denis: https://timberlake.co/uk/local-projections-for-time-series-and-panel-data-analysis.html