Daily Growth at Risk: financial or real drivers? The answer is not always the same
We propose a daily growth-at-risk (GaR) approach based on high-frequency financial and real indicators for monitoring downside risks in the US economy. We show that the relative importance of these indicators in terms of their forecasting power is time varying. Indeed, the optimal forecasting weight...
| Authors: | , , |
|---|---|
| Format: | article |
| Status: | Versión aceptada para publicación |
| Publication Date: | 2024 |
| Country: | España |
| Institution: | Universidad de Barcelona |
| Repository: | Dipòsit Digital de la UB |
| OAI Identifier: | oai:diposit.ub.edu:2445/213280 |
| Online Access: | https://hdl.handle.net/2445/213280 |
| Access Level: | Open access |
| Keyword: | Risc (Economia) Variables aleatòries Aprenentatge automàtic Valor (Economia) Risk Random variables Machine learning Value (Economics) |
| Summary: | We propose a daily growth-at-risk (GaR) approach based on high-frequency financial and real indicators for monitoring downside risks in the US economy. We show that the relative importance of these indicators in terms of their forecasting power is time varying. Indeed, the optimal forecasting weights of our variables differed clearly between the Global Financial Crisis and the recent Covid-19 crisis, reflecting the dissimilar nature of these two events. We introduce LASSO, elastic net, and adaptive sparse group LASSO into the family of mixed data sampling models used to estimate GaR and show how they outperform previous candidates explored in the literature. Moreover, equity market volatility, credit spreads, and the Aruoba–Diebold–Scotti business conditions index are found to be relevant indicators for nowcasting economic activity, especially during episodes of crisis. Overall, our results show that daily information about both real and financial variables is key for producing accurate point and tail-risk nowcasts of economic activity. |
|---|