Skip to content

Loss data included in EBA's recommended implementation of SMA


In the summer of 2019, the EBA formally provided their Basel III advice. Read on for a round-up of their recommendations.

EBA's recommended implementation of SMA

On 2 July 2019, the EBA published “Call for advice on Basel III implementation: key findings from the impact assessment and policy recommendations”, the summary results of their work on the Basel III implementation assessment and QIS impact study. The EBA will formally deliver its advice to the European Commission and publish a full report around the end of July 2019.

Headline recommendations

The key recommendations are that:

  • In line with ORX benchmarking, the introduction of the Standardised Measurement Approach (SMA) will bring significant capital increase, primarily focused on larger banks
  • When EU legislators adopt the BCBS SA, “the discretion to set ILM equal to 1 is not applied”, meaning that loss data will be used in the SMA capital calculation

In addition, the recommendations contain a significant number of qualitative items which relate to standards, risk management practice and Pillar 2.

This follows recent APRA guidance which stated that the Australian implementation of the SMA calculation will not use internal loss data. This confirms that there will not be jurisdictional consistency in implementation of Pillar 1 capital standards.

SMA details

The EBA stated their analyses showed that “a bank’s past operational losses are an effective indicator of the current-year operational losses”, and, “the volatility of the capital requirements is mainly driven by the BIC rather than by the ILM”. Additionally, they recommended allowing a bank-specific ILM for smaller intuitions which fall into Bucket 1.

The paper also hints at a longer implementation timeline than has been implied before, which they say would smooth "cliff effects" caused by sharp increases in capital and allow for more accurate loss data histories to be collected.

Qualitative recommendations

Perhaps surprisingly, two-thirds of the 36 recommendations provided are qualitative and largely relate to elements outside the scope of Pillar 1 capital. They are wide-ranging, covering definitions of risk types, framework requirements, loss data quality and Pillar 2. Of particular interest is the recommended use of the four elements of the AMA (with a change in the label of BEICF to KRI) for ICAAP purposes.

Table 1: Qualitative recommendations
Qualitative recommendations
Summary of the main recommendations

Definitional requirements

Set of recommendations aiming to update and harmonise definitions in the relevant regulatory products, including  clarifications on model risk, legal risk and ICT risk.

Governance and organisational requirements on loss data

  • Criteria to build the loss dataset

Criteria for ensuring the completeness and the quality of the loss data set.
  • Operational risk framework

Requirements on governance, reporting and control of operational risk.
  • Supervisory review of data quality

Requesting supervisors to perform periodical reviews of loss data set.


Disclosure standards on operational risk losses for Bucket 2 and 3 banks.

ICAAP and Pillar 2

On the use of internal data , scenarios analysis, external data and key risk indicators in ICAAP to ensure greater effectiveness in the management and control of operational risk and more granular measurement and better allocation of funds across the organisation.

Business indicator - FINREP mapping

A mapping (Level 2 text) should permit the association of the BI items to the FINREP items to enable European banks to calculate the BI accurately, consistently and with minimal effort.

EU-wide stress test

In other regulatory news, the EBA recently published an updated methodology for its stress tests. The requirement for operational risk is consistent, maintaining the same split focus on conduct and non-conduct forecasting. It mentions an expectation to use AMA models for capital projections, and AMA loss data reporting guidelines.

The final methodology will be published at the end of the year. The stress test exercise will be launched in January, with results published by the end of July.


Contact us  

To find out more about this topic, get in touch with us.

Find out more