Benchmarking operational risk stress testing models

Filippo Curti, Marco Migueis and Robert Stewart

Abstract

The Federal Reserve System's Comprehensive Capital Analysis and Review (CCAR) requires large bank holding companies to project operational losses under stressed macroeconomic scenarios, such as deep recessions characterized by high unemployment. However, the relationship between macroeconomic downturns and operational losses is influenced by innumerable factors, which are challenging to quantify and difficult to forecast. Therefore, the models designed to measure stressed operational losses are simplifications of complex processes, so multiple benchmarks should be used to guide primary model development. When modeling highly uncertain processes, benchmarks provide different conceptions allowing for alternative perspectives and reasoned judgments. This paper outlines several approaches to benchmarking operational loss projections under stressed scenarios using both accounting metrics and historical loss experience. These benchmarks provide a tool for bankers, regulators and practitioners to gauge the reasonableness and conservatism inherent in stressed operational loss projections, thus bringing a concrete structure and a panoramic view to the operational risk CCAR process.