Combining official and Google Trends data to forecast the Italian youth unemployment rate

Abstract

The increased availability of online information in recent years has aroused interest in the possibility of deriving indications for many kinds of phenomena. In the more specific economic and statistical context, numerous studies suggest the use of online search data to improve the forecasting and nowcasting of official economic indicators with a view to increasing the promptness of their circulation. The purpose of this work is to investigate if the use of big data can improve the forecasting of the youth unemployment rate – estimated in Italy on a monthly basis by the Italian National Institute of Statistics – by means of time series models. The time series used are those of the Google Trends query share for the keyword offerte di lavoro (job offers) and the official labour force survey data for the Italian youth unemployment rate since 2004. Two different models are estimated: an ARIMA model using only the official youth unemployment rate series and a VAR model combining the former series with the Google Trends query share. The results show that the use of Google Trends information leads to an average decrease in the forecast error.