Enhancing model robustness and compliance
The organisation required a comprehensive diagnostic review of its ECL framework. This included re-engineering key model components and implementing an automated solution to improve accuracy, strengthen governance and enhance operational efficiency across the ECL computation process.
Re-engineered models and enabled automation
Grant Thornton Bharat conducted a detailed assessment of the existing ECL framework, including a comprehensive review of probability of default (PD), loss given default (LGD) and exposure at default (EAD) methodologies to understand model architecture, underlying assumptions and computational logic.
We evaluated the conceptual soundness and implementation of macroeconomic overlays and forward-looking adjustments to ensure alignment with Ind AS 109 requirements. Our team also critically assessed model design, calibration techniques and validation parameters to strengthen stability, predictive capability and suitability for credit risk estimation.
In addition, we reviewed segmentation methodologies to enhance risk differentiation and appropriateness of categorisation across portfolios. Based on these insights, we redesigned key modelling components, including PD and credit conversion factor (CCF) approaches, aligning them with industry practices to improve overall model robustness and reliability.
Improved model reliability and operational efficiency
The engagement resulted in a comprehensive diagnostic report outlining key gaps and actionable recommendations aligned with regulatory expectations and leading practices. Enhanced segmentation approaches improved risk differentiation, while recalibrated models strengthened predictive capability and reliability.
The implementation of an automated ECL computation framework significantly reduced manual effort, improved accuracy and accelerated reporting timelines, enabling a more efficient and controlled credit risk management process.
