MEFMI conducted a virtual regional workshop on Nowcasting Gross Domestic Product (GDP) Using Mixed-Frequency Models from 22 to 25 June 2026, aimed at strengthening participants’ capacity to generate more timely estimates of economic activity.
Timely assessments of economic activity are critical for evidence-based policymaking. However, conventional macroeconomic forecasting often relies on quarterly GDP data that may be published with significant time lags. Mixed-frequency models help address this challenge by combining high-frequency indicators with lower-frequency national accounts data to produce more timely estimates of current economic conditions.
The workshop equipped participants with technical skills to construct, estimate and interpret mixed-frequency nowcasting models. Areas covered included traditional bridge equations, Mixed Data Sampling (MIDAS) regressions, unrestricted MIDAS models, forecast evaluation techniques, and the application of Mixed-Frequency Vector Autoregression (VAR) and Bayesian Vector Autoregression (BVAR) models using EViews.
Through hands-on practical sessions, participants applied these techniques to real-world datasets, strengthening their ability to generate timely estimates of GDP and other key macroeconomic variables.
A total of 81 officials from nine MEFMI member countries participated in the workshop, including participants from the Confederation of Zimbabwe Industries through the MEFMI Business Development Unit.
The skills developed through the programme are expected to support stronger short-term forecasting, evidence-based policy analysis and more responsive economic decision-making within participating institutions. MEFMI continues to strengthen macroeconomic management capacity across the region through practical training in modern analytical tools.
by Frank Chansa