A news-based macro uncertainty index for Italy

Abstract

In this study, we introduce an innovative high-frequency uncertainty index derived from an extensive dataset of Italian news sources. Our index methodically sorts news items by their macroeconomic significance and underlying sentiment. We focus on a period of analysis, from July 2017 to July 2021, characterized by pronounced uncertainties, notably the emergence of the 5 Star-Lega government coalition in 2018, which introduced significant political unpredictability, and the onset of the COVID-19 pandemic. We uncover a significant correlation between our newly developed uncertainty index and measures of sovereign risk (such as the BTP-Bund spread) during the tenure of the 5 Star-Lega government. However, this correlation diminishes following the initiation of the European Central Bank’s quantitative easing program, underscoring the substantial influence of political uncertainty on sovereign risk before the ECB’s intervention. The index is applied to dissect the Italian macroeconomic business cycle using a mixed-frequency model that integrates both financial and real economy indicators. Additionally, the study presents a nonlinear model leveraging the index to assess its predictive power in forecasting the Italian business cycle in recent years. The findings provide substantial evidence of the index’s predictive capabilities, highlighting its effectiveness in foreseeing shifts in the business cycle.