Malaysia
Forecasting and policy analysis in central banks increasingly rely on models that go well beyond simple autoregressions. Policymakers need tools that can capture persistent trends, shifting regimes, parameter instability and the growing use of high-frequency and mixed-frequency indicators. This course focuses on intermediate and advanced econometric techniques directly relevant to macroeconomic modelling and forecasting in a central banking environment.
The programme builds on standard multivariate time-series foundations and moves towards state-space methods, Bayesian econometrics, regime-switching models and machine-learning approaches for forecasting. The emphasis is on how these frameworks can be used in practice to inform policy decisions under uncertainty, assess model robustness and communicate results credibly. Teaching will combine lectures with hands-on empirical exercises using real or realistic central-bank data, as well as case studies from SEACEN member central banks and other monetary authorities.
At the end of the course, participants should be able to: (1) specify and estimate dynamic econometric models suitable for macroeconomic and financial forecasting. (2) Analyse long-run relationships among macroeconomic variables and use them in forecasting frameworks. (3) Formulate and estimate state-space and unobserved-components models using the Kalman filter. (4) Work with time-varying parameter models and apply them. (5) Understand the core building blocks of Bayesian econometrics, including priors, posteriors and MCMC algorithms. (6) Use panel-data methods.
The course is intended for staff of central banks and monetary authorities with more than three years of professional experience, who are already familiar with introductory time-series econometrics. Participants should hold at least a Master’s degree in Economics, Finance or Statistics; those with a PhD are welcome. A solid grounding in statistics and econometrics, and prior exposure to forecasting work, will enable participants to benefit fully from the material. Only participants who complete 80% of the sessions will receive a certificate of completion.
We will be using several software packages in this course. Participants will need Eviews, Matlab, Stata, and R for most empirical illustrations.
All potential topics will be delivered through a combination of lectures and exercises/workshops:
Dynamic Multivariate Econometric Models: Estimation and Forecasting of Short-run versus Long-run relationships
© 2025 The SEACEN Centre
198201000672 (80416-M)
