Forecasting inflation for India with the Phillips curve: evidence from internet search data

Abstract

This paper forecasts inflation based on the New Keynesian Phillips curve for India using internet search-based inflation expectations. We find that compared to inflation forecasts based on traditional time-series estimation method like AR(1) or household-survey based inflation expectations, our model that employs inflation expectations of agents based on real-time inflation-related internet search, has better inflation forecasting performance. From a policy standpoint, this underscores the importance of analysing big data for the purpose of forecasting inflation.