From modelling to forecasting bank profitability: Evidence from euro area banks

Abstract

Building on the established bank profitability literature, this paper develops models encompassing macroeconomic and banking variables and assesses their forecasting performance. Understanding factors underpinning bank profitability is important for the design and implementation of stress tests, which have become one of the integral elements of bank risk management frameworks and capital planning. Results from the data for euro area banks suggest that particular attention may need to be paid to quantifying and managing interest rate risk. In addition, with the identified relevance of capital position and credit risk, additional efforts may be needed to monitor the ensuing risks. Results also suggest that models relying solely on macroeconomic variables do not perform significantly worse in terms of forecasting accuracy compared to richer specifications and can beat the autoregressive models over the medium term. Bearing in mind the costs associated with development, validation, and maintenance of more complex models underpinned by richer specifications, the results point to the potential use of presented or similar models by regulators, supervisors, and banks in their respective roles in the context of supervisory review and evaluation processes.