Real-Time Nowcasting of CPI and Food Inflation with Google Trends-Evidence from the US, UK, and India (2012–2025

This study develops and evaluates real-time nowcasting models for headline and food inflation using Google Trends data across three major economies: the United States, United Kingdom, and India over the period 2012–2025. Employing mixed-frequency econometric techniques including MIDAS regression and regularized bridge models, we demonstrate that incorporating weekly Google search intensity for inflation-related keywords significantly improves nowcasting accuracy compared to traditional time-series baselines. Our strict real-time evaluation framework accounts for data release lags and information availability constraints, ensuring practical relevance for policymakers. Diebold–Mariano tests with Holm–Bonferroni correction confirm that Google Trends-augmented models systematically outperform seasonal-naive and ARIMA benchmarks, with mean absolute errors reduced by 15-30% across countries. The methodology provides timely inflation signals particularly valuable during periods of economic uncertainty.