CMStatistics 2022: Start Registration
View Submission - CFE
A1317
Title: Nowcasting inflation using web searches Authors:  Marco Colagrossi - European Commission, Joint Research Centre (Italy) [presenting]
Luca Barbaglia - European Commission Joint Research Centre (Italy)
Sergio Consoli - Joint Research Centre (JRC) (Italy)
Abstract: Researchers are increasingly exploiting unconventional data sources to study phenomena for which timely and high-frequency data are not readily available. We exploit web search data obtained through a non-public Google API to improve fore- and now-casting performances of consumer price indexes (CPIs) across EU Member States. Starting from Google Search categories, we retrieve over seven thousand topics (which are aggregations of several web search queries that could be assigned to the same semantic domain) broadly related to the economy. Finally, using machine learning models, we show how by including the relevant set of searches, it is possible to improve fore- and now- casting performances of several CPIs across Europe.