The IFRS 9 accounting standard requires the prediction of credit deterioration in financial instruments, i.e., significant increases in credit risk (SICR). However, the definition of such a SICR-event is inherently ambiguous, given its current reliance on evaluating the change in the estimated probability of default (PD) against some arbitrary threshold. We examine the shortcomings of this PD-comparison approach and propose an alternative framework for generating SICR-definitions based on three parameters: delinquency, stickiness, and the outcome period. Having varied these framework parameters, we obtain 27 unique SICR-definitions and fit logistic regression models accordingly using rich South African mortgage and macroeconomic data. For each definition and corresponding model, the resulting SICR-rates are analysed at the portfolio-level on their stability over time and their responsiveness to economic downturns. At the account-level, we compare both the accuracy and dynamicity of the SICR-predictions, and discover several interesting trends and trade-offs. These results can help any bank with appropriately setting the three framework parameters in defining SICR-events for prediction purposes. We demonstrate this process by comparing the best-performing SICR-model to the PD-comparison approach, and show the latter's inferiority as an early-warning system. Our work can therefore guide the formulation, modelling, and testing of any SICR-definition, thereby promoting the timeous recognition of credit losses; the main imperative of IFRS 9.
This R-codebase can be run sequentially using the file numbering itself as a structure. Delinquency measures are algorithmically defined in DelinqM.R as data-driven functions, which may be valuable to the practitioner outside of the study's current scope. These delinquency measures were formulated and empirically tested in Botha22, as part of a loss optimisation exercise of recovery decision times, as implemented in the corresponding R-codebase. A simulation study from Botha2021 also demonstrated these delinquency measures at length, with its corresponding R-codebase. Similarly, the TruEnd-procedure from Botha2024 and its corresponding R-codebase is implemented in the TruEnd.R script, which includes a small variety of functions related to running the TruEnd-procedure practically.
This R-codebase assumes that monthly loan performance data is available. Naturally, the data itself can't be made publically available given its sensitive nature, as well as various data privacy laws, particularly the Protection of Personal Information (POPI) Act of 2013 in South Africa. However, the structure and type of data that is required for reproducing this study, is sufficiently described in the commentary within the scripts. This should enable the practitioner to extract and prepare data accordingly. Moreover, this codebase assumes South African macroeconomic data is available, as sourced and collated by internal staff of the bank in question.
All code and scripts are hereby released under an MIT license. Similarly, all graphs produced by relevant scripts as well as those published here, are hereby released under a Creative Commons Attribution (CC-BY 4.0) licence.