Footnote:
Nach Informationen von SSRN wurde die ursprüngliche Fassung des Dokuments August 1, 2023 erstellt
Description:
Being able to accurately estimate key macroeconomic indicators prone to lagged releases remain crucial for policymakers to enact appropriate policy responses, including during times of volatility. Recently, studies have implemented Machine Learning algorithms into Gross Domestic Product (GDP) nowcast, and its components, demonstrating promising out-of-sample predictive performance. This paper nowcasts private consumption (PC), which comprises 60.2% of Malaysia’s GDP in 2022, with a large set of 87 macroeconomic variables, using a framework that combines Least Absolute Shrinkage and Selection Operator (LASSO), and principal component analysis (PCA) layers. The framework performed better in nowcasting PC for the pre-COVID period (1Q 2005 to 4Q 2019) than for the with-COVID period (1Q 2005 to 1Q 2023), and with some observable improvements during the endemic transition (1Q 2022 onwards). The framework serves as a useful nowcasting tool, particularly during periods of mild fluctuations. Further research is warranted to identify models with promising predictive power during and after periods of high volatility, including over a longer term