Real-Time Monitoring of the Indian Economy – An Alternative Approach
With more frequent and up to date dataflow, the daily/weekly indicators allow model-based monitoring of economic activity to be more real time, thus enhancing their relevance for policy making. Anomaly in the general trend of the indicators have driven existing models to go off the mark with exorbitantly high error component post the pandemic. In this paper, we extend our existing nowcasting model for quarterly GDP in India which is based on monthly indicators to include additional data in the form of daily/weekly indicators. We develop a factor augmented MIDAS model by integrating two composite indices – at monthly and weekly frequency, respectively, constructed using the dynamic factor model for the purpose of nowcasting quarterly GDP. This framework allows to leverage detailed and comprehensive coverage of economic activity through monthly indicators while also updating nowcasts more frequently based on latest information content embedded in the weekly indicators. The results show that the FA-MIDAS produces a lower out of sample forecast error compared with univariate ARIMA and standalone monthly indictor based DFM nowcasting models during the post-pandemic period.