Macro Musings with David Beckworth Andrew Martinez on the Art of Forecasting
Feb 16, 2026
Andrew Martinez, former U.S. Treasury economist and assistant professor at American University, blends forecasting, time-series econometrics, and policy analysis. He explores the state of macro forecasting, AI’s promise and pitfalls for turning points, why policymakers value simple interpretable models, the SEP’s role in monetary surprises, and the construction of an NGDP expectations gap.
AI Snips
Chapters
Books
Transcript
Episode notes
From Diplomatic Aims To Forecasting Career
- Andrew Martinez describes his path from international affairs interest to forecasting through GW, IMF, and a PhD at Oxford under David Hendry.
- He recounts joining Treasury in 2019 and doing macro forecasting for policymakers during crises like the pandemic.
Simplicity Beats Black Boxes In Policy Work
- Simple, interpretable models often beat black-box ones in policy settings because decisionmakers need a clear story and causal mechanism.
- Forecasters therefore balance advanced methods in the background with transparent models for communication.
Modern Forecasting Combines Big Data And Robustness
- Modern forecasting mixes big-data factor models, machine learning selection, and time-varying parameter methods to handle structural breaks.
- Forecasters constantly manage the bias–variance tradeoff to avoid overfitting while capturing instability.
