The Data Dividend: Why ECL Makes Data The New Capital For Banks

The regulatory environment is witnessing a major shift, moving from prudential norms to Expected Credit Loss (ECL)-based provisions. ECL is an objective and customised method to calculate credit provisions. It focuses on principle-based measurement based on an entity’s experience with asset quality and recovery. The ECL approach drives institutions to build internal assessment capabilities that can address asset quality concerns, including the impact of macroeconomic factors. It aims to ensure that adequate provisions are created to protect the balance sheet from credit shocks.

ECL relies on three foundational elements: data management, ECL modelling methodologies and disclosures. Data can significantly shape how banks are perceived by regulators and rating agencies. It also determines their financial strength and investor confidence in capital and debt. Data availability and quality reflect an institution’s true governance standards, beyond what is stated or claimed.

If the available data is unreliable or incomplete, banks may need to set aside additional provisions and buffers, which can weigh on their profitability and reduce their appeal to investors. In case there are insufficient datasets to make a rigorous analysis, the choice of modelling techniques may also not be commensurately sophisticated. This affects banks’ ability to refine customer segments based on risk homogeneity. Auditors may issue a qualified opinion on the financial statements if the estimates and underlying assumptions cannot be clearly explained. Disclosures on ECL in the annual report reflect governance standards and the strength of internal controls. In this context, data plays a key role as a hygiene factor and drives competitive advantage.

Data availability is a key component in deciding which approach can be used for ECL measurement. A lack of sufficient data may result in judgment-based triggers for the Significant Increase in Credit Risk (SICR) criteria and adopting a simplified estimation of Probability of Default (PD), Loss Given Default (LGD), Exposure at Default (EAD) and modelling macroeconomic impact. These subjective models are likely to face challenges and additional overlays by auditors and regulators to account for idiosyncratic and macroeconomic factors. With adequate historic data, statistical modelling techniques can be adopted to identify patterns, forecast default rates, and predict recoveries across economic cycles with reasonable accuracy.

Data granularity helps in the segmentation of the portfolio based on the homogeneous risks of each product/segment. In case granular segmentation is not possible, estimates tend to be either inflated to accommodate the varying risk levels within the portfolio or set aggressively, overlooking the full range of risk factors.

Data completeness improves the predictive capability of the models. When default data is adequate, models are less susceptible to volatility. This provides stability to provision estimates and a greater degree of certainty to capital planning. However, the selection of datasets must be consistent and justified. Choosing data to highlight favourable periods of credit trends or excluding certain periods/segments without an appropriate rationale may raise concerns about bias. Maintaining consistent data and modelling standards is therefore a minimum expectation to ensure the reliability of modelled outputs.

Data integrity and accuracy ensure that modelled outputs are relevant and credible for boards, auditors, rating agencies, regulators, and investors who rely on them for decision-making. Data traceability also supports the validation of modelled outputs and demonstrates strong governance standards. This makes the audit process more efficient, reduces the number of queries from regulators, and lowers the risk of overlays on provisions.

High-quality data enables the creation of Management Information System (MIS) dashboards for timely analysis of portfolio quality and prompting corrective action when required. It also helps auditors and regulators build confidence in the ECL approach, output, and governance. Furthermore, it enhances credibility and reduces the chances of rigorous scrutiny/challenges and potential financial penalties.

High-quality, granular data allows for more accurate estimation of expected credit losses and better assessment of portfolio risks, while supporting early detection of vulnerabilities. This enables institutions to undertake timely corrective actions and regulate credit exposure at the portfolio level. It also allows them to create appropriate buffers for potential losses, considering both portfolio quality and macroeconomic conditions. By aligning provisions with expected losses, banks free up excessive capital that can be used to support growth.

Data plays a vital role in ECL disclosures as well, providing investors with a clear view of how credit is managed within the institution. When modelled outputs are accurate and disclosures are consistent, they solidify investor confidence and signal strong governance.

Data is the fuel and is critical in making the ECL framework the future of credit-related provisions.

The article is authored by Aruna Pannala, partner at Deloitte India.

The views expressed in this article are solely those of the author and do not necessarily reflect the opinion of NDTV Profit or its affiliates. Readers are advised to conduct their own research or consult a qualified professional before making any investment or business decisions. NDTV Profit does not guarantee the accuracy, completeness, or reliability of the information presented in this article.