An Analysis of Italian Bad Loans: Determinants, Forecasts and Transmission to the Real Economy

Abstract

Since 2009 conditions in Italian credit market have been experiencing a dramatic worsening, reflecting the two most severe recessions since the Great Depression. Probability of default for non financial firms has reached unexpected values, while deterioration in credit portfolios has spread. This paper studies bad loans for Italian non financial firms during the last twenty years. We propose different linear and non-linear methodologies focusing on short and long-term determinants, forecasting properties and dynamic responses. Our empirical results suggest that macroeconomic and financial, but also, specified lenders and borrowers variables affect bad loans. Linear and non-linear models augmented with financial variables and asset prices produce better out-of-sample forecasts. A dynamic response analysis shows that default rates move with a cyclical pattern, falling after a positive shock in macroeconomic and financial variables. Moreover, a positive shock in bank credit or in default rate does not produce a clear feedback effect from credit to the real economy. In a non-linear framework, this happens only when the default rate is above a critical value, suggesting a possible breakdown in the transmission of credit to the real economy when credit quality is weak.