Composite GDP nowcasting using macroeconomic variables and electricity data

Area of Science:

• Economics
• Econometrics
• Macroeconomic Forecasting

Background:

• Accurate gross domestic product (GDP) forecasting, or nowcasting, is vital for effective macroeconomic regulation.
• Traditional models often rely on a limited set of macroeconomic indicators.
• Real-time data, such as electricity usage, offers a promising avenue for enhancing GDP nowcasting.

Purpose of the Study:

• To develop and evaluate a composite GDP nowcasting model.
• To integrate predictions from a dynamic factor model (using mixed-frequency indicators) and a regression model (using electricity data).
• To assess the model's performance in reducing forecast errors for quarterly GDP growth rates.

Main Methods:

• A composite model combining a dynamic factor model and a regression model.
• The regression model incorporates electricity consumption and changes in electricity capacity as predictors.
• Nowcasting is performed for year-on-year quarterly GDP growth rates using data from Fujian Province, China.

Main Results:

• The proposed composite nowcasting model demonstrated reduced forecast errors.
• The model effectively leverages correlations between GDP, macroeconomic indicators, and electricity data.
• Incorporating electricity capacity changes provided additional predictive power for future electricity demand.

Conclusions:

• The composite GDP nowcasting model offers improved accuracy over individual models.
• Electricity data, including capacity changes, is a valuable predictor for GDP nowcasting.
• This approach enhances macroeconomic forecasting by integrating diverse data sources.