Managing credit risk with info‐gap uncertainty

Purpose - The paper aims to provide a quantitative methodology for dealing with (true) Knightian uncertainty in the management of credit risk based on information‐gap decision theory. Design/methodology/approach - Credit risk management assigns clients to credit risk categories with estimated probabilities of default for each category. Since probabilities of default are subject to uncertainty the estimated expected loss given default on a loan‐book can be subject to significant uncertainty. Information‐gap decision theory is applied to construct optimal loan‐book portfolios that are robust against uncertainty. Findings - By choosing optimal interest‐rate ratios among the credit risk categories one can simultaneously satisfy regulatory requirements on expected losses and an institution's aspirations on expected profits. Research limitations/implications - In the analysis presented here only defaults over specific time frames have been considered. However, performance requirements expressed in terms of defaults and profits over multiple time frames that allow for transitions of clients between credit risk categories over time could also be incorporated into an information‐gap analysis. Practical implications - An additional management analysis tool for applying information‐gap modeling to credit risk has been provided. Originality/value - This paper provides a new methodology for analyzing credit risk based on information‐gap decision theory.