Abstract- Post-COVID economic challenges have made it essential to reassess the nowcasting models used for estimating India’s Gross Domestic Product (GDP). One of the key difficulties in this context is the availability of diverse High-Frequency Indicators (HFIs) that exhibit different lead–lag relationships with GDP. To address this issue, the present study employs a Mixed-Frequency Multi-Factor Vector Autoregression (VAR) model, which allows the joint incorporation of multiple types of HFIs. The model includes a wide range of indicators such as nominal variables, survey-based indicators, labour market indicators, and measures of real economic activity. Most existing nowcasting models for India are limited to producing aggregate GDP estimates and do not explicitly account for the individual contributions of each HFI to the GDP nowcast. To overcome this limitation, this study separately quantifies the impact of each HFI on revisions in GDP nowcasts as new economic information becomes available.Furthermore, the unprecedented economic disruptions caused by the COVID-19 pandemic generated extreme outlier values in several HFIs, leading to distortions in model parameters. To mitigate this challenge, the Oxford Stringency Index along with a novel data transformation technique has been incorporated. This approach reduces the model’s sensitivity to large economic shocks and enhances its ability to handle unexpected events without overreacting. Consequently, this transformation offers a significant contribution to the forecasting literature, particularly in contexts characterized by extreme observations.
Dipak Arjun (Sat,) studied this question.