This paper uses over 130 financial variables, available at a weekly frequency, to produce short-term forecasts of the expansion and recession phases of the Italian business cycle, with a focus on the most recent period affected by the Covid-19 pandemic. Financial data incorporate the expectations on growth prospects and are thus useful leading indicators of real economic activity.
An empirical application to 2020 and the first part of 2021, marked by the Covid-19 pandemic, shows that the proposed approach produces better forecasts than those obtained with models estimated using single financial variables. The forecast accuracy is comparable to that of models which exclusively employ real variables, both quantitative (industrial production) and qualitative (PMI indicators). The proposed approach could be integrated into a central bank's macroeconomic modelling toolbox, combining the information content of financial and real variables.
Published in 2023 in: Journal of Forecasting, v. 42, 7, pp. 1569-1593.
