Nowcasting India's Quarterly GDP Growth: A Factor-Augmented Time-Varying Coefficient Regression Model (FA-TVCRM)

Area of Science:

• Econometrics
• Macroeconomic Forecasting
• Time Series Analysis

Background:

• High-frequency economic data is crucial for timely decision-making.
• Traditional GDP indicators are released with a significant lag.
• Nowcasting, using high-frequency data to project current economic conditions, is essential.

Purpose of the Study:

• To develop and evaluate a novel Factor Augmented Time Varying Coefficient Regression (FA-TVCR) model for nowcasting Indian GDP growth.
• To assess the model's ability to handle structural breaks and large datasets.
• To compare the FA-TVCR model's performance against established nowcasting techniques.

Main Methods:

• Development of a Factor Augmented Time Varying Coefficient Regression (FA-TVCR) model.
• Estimation using 19 and 28 high-frequency economic indicators.
• Comparison with Dynamic Factor Model (DFM) and Autoregressive Integrated Moving Average (ARIMA) models.
• Evaluation using in-sample and out-of-sample Root Mean Square Error (RMSE) and Diebold-Mariano tests.

Main Results:

• The FA-TVCR model demonstrated superior performance over DFM and ARIMA models in terms of RMSE.
• FA-TVCR consistently outperformed DFM in predictive power.
• FA-TVCR showed comparable forecast accuracy to ARIMA under normal conditions but excelled during economic shocks like the COVID-19 pandemic.

Conclusions:

• The FA-TVCR model offers a robust and accurate approach to nowcasting economic indicators like GDP.
• This novel method effectively integrates information from numerous high-frequency indicators.
• The FA-TVCR model is particularly valuable for forecasting during periods of economic volatility and structural breaks.