A1166
Title: Forecasting the tail behavior of banks risk sentiment
Authors: Paulo Rodrigues - Universidade Nova de Lisboa (Portugal) [presenting]
Joao Nicolau - ISEG and CEMAPRE (Portugal)
Adriana Cornea-Madeira - ISEG and CEMAPRE (Portugal)
Abstract: The aim is to provide novel approaches for modelling, forecasting and identifying the drivers of the tail behavior of time series. Using a new triple adaptive lasso approach, the relevance of different macrofinancial proxies are analyzed for forecasting the right tail extreme values of Bank Risk Sentiment Tracker (BRST) series, a synthetic indicator of market sentiment for individual listed banks. Specifically, considering a sample of 21 banks from the Euro area, the UK, and the US, the statistical significance of global and regional market volatility and, skewness measures as well as monetary and macro conditions are evaluated via inflation expectations and the yield curve spread. An extensive and novel forecasting exercise is performed in order to evaluate how well these variables perform in anticipating the risk dynamics traced by the BRST. The focus is on two forecast horizons: One-day ahead, $h=1$, and five-days ahead $h=5$.
