COMPARATIVE ANALYSIS OF TWO-STEP AND QUASI MAXIMUM LIKELIHOOD ESTIMATION IN THE DYNAMIC FACTOR MODEL FOR NOWCASTING GDP GROWTH IN INDONESIA

Souisa, Gilbert Alvaro
Leiwakabessy, Reyner M.
Damayanti, Salma
Terim, Mohammad Zanuar F
Pelu, Shelma M

BAREKENG: Jurnal Ilmu Matematika dan Terapan
Vol 19 No 1 (2025): BAREKENG: Journal of Mathematics and Its Application

Economic activity data is needed quickly to make policy decisions, but this data suffers from publication delays. Gross Domestic Product (GDP) data is released within five weeks after the end of the quarter. An effort that can be made to provide such data is through nowcasting, which is forecasting in the current period using variables that have a higher frequency. This study aims at nowcasting GDP growth. The nowcasting method used is the Dynamic Factor Model (DFM) with Two Step (TS) and Quasi Maximum Likelihood (QML) estimation. The nowcasting results show that the DFM-TS model is better than the DFM-QML because it has a larger adjusted R-squared value and has the smallest RMSE value of 1.71035 compared to the DFM-QML value, which has an RMSE value of 1.71598.