Forecasting Inflation with Online Prices

Abstract

Online prices are becoming an increasingly popular micro data source in the economic literature. In this paper we introduce goods’ prices collected online through web scraping retailer’s websites as a new input to forecast the CPI inflation. We perform monthly and quarterly out-of-sample inflation forecasts in multiple countries, including the US and the UK, and find that models using online prices outperform models with offline data, from traditional benchmarks to pooled forecasts. The online anticipation presents an attractive feature to both policymakers and practitioners, in particular given the non-negligible delay in the official CPI release. Finally, we discuss reasons why the anticipation in online prices can take place.