2018 REGISTRATION DOCUMENT AND ANNUAL FINANCIAL REPORT

2018 Registration document and annual fi nancial report - BNP PARIBAS 347

5RISKS AND CAPITAL ADEQUACY PILLAR 3

Credit risk

➤ TABLE 29: CREDIT RISK-WEIGHTED ASSETS MOVEMENTS BY KEY DRIVER (EU CR8)

In millions of euros

RWAs Credit risk Capital Requirements Credit risk

Total of which IRB

approach Total of which IRB

approach

1 January 2018 504,298 243,398 40,344 19,472

Asset size 25,489 17,590 2,039 1,407

Asset quality (30,056) (26,411) (2,404) (2,113)

Model update 3,484 3,604 279 288

Methodology and policy 2,278 2,278 182 182

Acquisitons and disposals (2,475) (25) (198) (2)

Currency 977 1,985 78 159

Others (144) (95) (12) (8)

31 DECEMBER 2018 503,851 242,323 40,308 19,386

Credit risk-weighted assets decreased by EUR 0.5 billion in 2018, primarily as a result of the following:

■ an increase of EUR 25 billion in line with business activity, including an increase of EUR 10 billion in Domestic Markets, EUR 9 billion in International Financial Services and EUR 6 billion in CIB;

■ an improvement in asset quality resulting in a decrease of -EUR 30 billion, with improvement in risk parameters and the implementation of three securitisation programmes;

■ an increase of EUR 3 billion relating to the update of models;

■ an increase of EUR 2 billion linked to methodological changes, mainly related to the macro-prudential measure of mortgages in Belgium;

■ a net fall of EUR 2 billion related to scope changes, in particular following the disposal of First Hawaiian Bank (FHB) and the acquisition of the core banking activities of Raiffeisen Bank in Poland.

CREDIT RISK: I NTERNAL R ATINGS B ASED A PPROACH (IRBA)

The internal rating system developed by the Group covers the entire Bank. The IRBA framework, validated in December 2007, covers the portfolio described in Approaches used to calculate capital requirements in the section entitled Exposure to credit risk.

The Group has developed specific internal models adapted for the most common categories of exposure and clients in its loan portfolio. BNP Paribas bases these developments on internal data gathered over long periods. Each of these models is developed and maintained by a specialist team, in conjunction with relevant RISK and business line experts. Moreover, verifi cation is performed to ensure compliance with the fl oors set by the regulation on these models. The Bank does not use models developed by external suppliers.

Counterparty rating (or the Probability of Default) and the Loss Given Default is determined either using purely statistical models for portfolios with the highest degree of granularity (loans to individuals or to very small enterprises) or a combination of models and expert judgement based on indicative values.

Loss Given Default is defi ned as the loss that the Bank would suffer in the event of the counterparty s default in times of economic downturn, as required by regulations. For each transaction, it is measured using the recovery rate for a senior unsecured exposure to the counterparty, adjusted for any risk mitigation techniques (collaterals or guarantees ). Amounts recoverable against these mitigants are estimated each year using conservative assumptions as well as haircuts calibrated to refl ect economic downturn conditions.

The Bank models its own conversion factors on fi nancing commitments by using internal default data. Conversion factors are used to measure the off-balance sheet exposure at risk in the event of a default. This parameter is assigned automatically depending on the transaction type for all portfolios and therefore, is not determined by the Credit Committees.

The breakdown of the main models used by the Group, their characteristics and main exposures covered are presented below.