A News-based Policy Index for Italy: Expectations and Fiscal Policy
Fantozzi DanielaMuscarnera Alessio
CEIS Research Paper
In this paper we investigate the effect of a "news-based" policy shock on consumption and investments. To this aim, we construct a new measure of policy announcements, the Policy News Index (PNI), analyzing textual data from the most important Italian business newspaper (Il Sole 24 Ore). To disentangle news and noise from fiscal policy communication, we provide measures of newspaper coverage of policy announcements purged from uncertainty. Using a BVAR, we estimate the response of households and firms to a news on government spending. Results indicate that the "news" or "foresight" shock has delayed effects on government spending, consumption and investments. Agents receive mixed signals from newspapers about changes in government spending and they do not react before changes are fully in place due to a "lack of trust". Additionally, we show in a counterfactual exercise that the "confidence channel" plays a crucial role in in the transmission of the shock. Hence, our results provides evidence that outcomes are expectations-driven.
Keywords: News-based index, Textual data, Text mining, Fiscal foresight, Agents’ expectations, BVAR.
JEL codes: C32,C81,D80,E62
Date: Thursday 11 March 2021
Revision Date: Thursday 11 March 2021