Title: Development of Short-term Forecasting Models for Non-oil GDP

Number: 04/2024

Author: Nazrin Ramazanova, Ogtay Gahramanov

Language: Endlish

Date: 2024

Abstract: Short-term forecasting of non-oil GDP is essential for preparing its medium-term projections. This paper evaluates the performance of three key forecasting models—Vector Autoregression (VAR), Mixed Data Sampling (MIDAS), and Dynamic Factor Models (DFM)—in predicting quarterly growth of non-oil GDP. The empirical results show that most models outperform the AR(1) benchmark in quarterly forecasts. To leverage the strengths of each model, the study also examines combined forecast results. A forecast combination based on mean squared error (MSE) weights improves forecast accuracy compared to individual models. Overall, the findings highlight the advantages of a multi-model forecasting strategy for enhancing the accuracy of short-term GDP growth predictions in Azerbaijan’s dynamic economic environment. The results emphasize the practical benefits of using combination methods to strengthen the reliability of short-term forecasts.

Keywords: short-term forecasting, non-oil GDP, forecast combination

JEL Classification: C32, E52, O40

Development of Short-term Forecasting Models for Non-oil GDP

409.29 KB