A1102
Title: Nowcasting recession risk in the US and the Euro Area
Authors: Francesco Furno - Amazon Web Services (United States) [presenting]
Domenico Giannone - University of Washington (United States)
Abstract: Timely coincident recession risk indicators are presented for the United States (US) and the Euro Area (EA) at a monthly frequency. The indicators are constructed by estimating a parsimonious Bayesian logit based on two predictors, which summarize financial conditions and real economic activity. The composite indicator of systemic stress (CISS) is selected to measure financial conditions, as well as the US PMIs and the EA economic sentiment index (ESI) to summarize real economic activity. The predictors are available immediately after the month of reference concludes. Back-testing the indicator over the periods 1980-2021 for the US and 1985-2021 for the Euro Area reveals a 96\% and a 92\% in-sample accuracy and 95\% and 88\% pseudo-out-of-sample, respectively. The indicators are more accurate than popular indicators such as the Sahm-Rule, especially at determining when the economy leaves a recession, and complement spread-based indicators, which are good at forecasting instead of nowcasting recessions.