BPCE - Risk Report - Pillar III 2020

CREDIT RISKS

RISK MEASUREMENT AND INTERNAL RATINGS

INTERNAL RATINGS-BASED APPROACHES – RETAIL CUSTOMERS For retail customers, Groupe BPCE has established standardized internal ratings-based methods and centralized ratings applications used to assess the credit quality of its loan books for better risk supervision. For the Banque Populaire and Caisse d’Epargne networks, they are also used to determine capital requirements under the Advanced IRB method. The probability of default of retail customers is modeled by the Risk division, based in large part on the banking behavior of the counterparties. The models are segmented by type of customer, distinguishing between individual and professional customers (with or without balance sheets) and according to products owned. The counterparties in each segment are automatically classified using statistical models (usually logistic regression models) into similar and statistically separate risk categories. Probability of default is estimated for each of these categories, based on the observation of average default rates over the longest period possible so as to obtain a period representative of the possible variability of the observed default rates. These estimates are systematically adjusted by applying margins of conservatism to cover any uncertainties. For comparison purposes, risk reconciliation is carried out between internal ratings and agency ratings. Loss given default (LGD) is an economic loss measured by incorporating all inherent factors in a transaction as well as the costs incurred during the collection process. LGD estimation models for retail customers are applied specifically to each network. LGD values are first estimated by product, and based on whether or not any collateral has been provided. Other factors may also be considered secondarily, where they can be used to statistically distinguish between degrees of loss. The estimation method employed is based on the observation of marginal collection rates, depending on how long the customer has been in default. The advantage of this method is that it can be directly used to estimate LGD rates applied to performing loans and ELBE rates applied to loans in default. Estimates are based on internal collection histories for exposures at default over an extended period. Two margins of conservatism are then systematically added: the first to cover estimate uncertainties and the second to mitigate any economic slowdown effect. Groupe BPCE uses two models to estimate EAD. The first estimates a Credit Conversion Factor (CCF) for off-balance sheet exposures. This model is automatically applied when off-balance sheet exposures are deemed material ( i.e. exceeding the limits set for each type of product). The second estimates a flat increase in the balance sheet for non-material off-balance sheet exposures.

INTERNAL RATINGS-BASED APPROACHES – NON-RETAIL CUSTOMERS Groupe BPCE has comprehensive systems for measuring non-retail customer risks, using either the Foundation IRB or Advanced IRB approach depending on the network and the customer segment. These systems can also be used to assess the credit quality of its loan books for better risk supervision. The rating system consists in assigning a score to each counterparty. Given the Group’s cooperative structure, a network of officers is responsible for determining the customer’s rating for the Group based on the uniqueness of the score. The score assigned to a counterparty is usually suggested by a model, then adjusted and validated by risk function experts after they perform an individual analysis. This process is applied to the entire Non-Retail portfolio, except the new models reserved for Small Enterprises (SEs), which are automatically rated (as with the Retail portfolio). The counterparty rating models are mainly structured according to the type of counterparty (corporates, institutions, public sector entities, etc.) and size of the company (measured by its annual revenues). When volumes are sufficient (SMEs, ISEs, etc.), the models rely on statistical modeling (logistic regression methods) of customer defaults, combined with qualitative questionnaires. Otherwise, expert criteria are used, consisting of quantitative factors (financial ratios, solvency, etc.) derived from financial data, and qualitative factors assessing the customer’s economic and strategic components. With respect to country risk, the system is based on sovereign ratings and country ratings that limit the ratings that can be given to non-sovereign counterparties. The Non-Retail rating scale is built using past Standard & Poor’s ratings to ensure the direct comparability in terms of risks with the rating agencies. For the new SE models, specific scales were defined for each model used to perform regulatory calculations. These scales are connected with the Non-Retail rating scale for internal risk management. For statistical models, the calibration of probabilities of default on the scales defined for regulatory calculations is based on the same principles as those set out for retail customers (in particular the historic representation of default rates, as well as the estimation of uncertainty margins).

5

91

RISK REPORT PILLAR III 2020 | GROUPE BPCE