Response to discussion on a Feasibility Study of an Integrated Reporting System under Article 430c CRR
Go back
For small reporting credit institutions, a principle of proportionality should apply.
In addition, the scope of the data collections should be limited to the data required by the regulation without extending it to management data and without anticipating future regulations.
Furthermore, we would also like to stress that if reporting institutions could “report once”, harmonised and standardised data, i.e., produce a single report from which the relevant authorities could extract the data they require, it would be of great benefit to reporting institutions and authorities alike.
We believe that the two areas among the four core areas considered for the creation of an integrated reporting that are key for the success of the project are a unique and common data dictionary and the related governance aspects. They should be considered first, before moving on to the other topics, i.e. the Central Data Collection Point and the level of granularity of data, that, while relevant, can be discussed at a later stage.
Besides, the EBA feasibility study is launched shortly after the ECB published the Cost-Benefit Assessment (CBA) questionnaire on the same matter related to an integrated reporting system. We would have hoped closer interconnection between the ECB's Integrated Reporting Framework (IReF) and the EBA's feasibility study. We question the differences in terms of scope - the ECB project is limited to statistical data whereas the EBA project addresses a broader scope covering prudential, statistical and resolution reporting – and in terms of progress of the two projects – the implementation horizon of the ECB’s IReF is 2024 – 2027, whereas the EBA study remains at a macro level analysis. The alignment of the implementation of both projects is essential. Different implementation timetables and scopes would imply that reporting institutions will incur costs without the full benefit of an integrated reporting. They would have to overhaul their IT infrastructure after the implementation of ECB’s IReF, once the integrated reporting system for supervision, statistics and resolution is defined by the EBA.
In addition, could be considered other requirements such as SRB valuation data set (for the common part aligned with Anacredit) and Loan Tape standard (for the common part with Anacredit).
Besides, one of the desirable objectives of an integrated reporting system is to reduce the number and frequency of ad hoc requests. While national data may be needed to fulfil national needs or ad hoc data may arise as the result of specific situations, we encourage authorities to have a clear commitment to issue fewer ad hoc requests and to reuse granular data already reported, so to provide a relief for the reporting institutions regarding their reporting burden.
Besides, it should be noted that banks consider the DPM as a reference to fill in templates and perform controls. The DPM could not be considered and used as a dictionary as it is a technical tool and does not contain any definition. It could only be appended to the data dictionary.
Accordingly, while we agree that the alignment of the semantic level (data definitions), the syntactic level (Metadata Model and Logical Data Model) and the infrastructure level (Database) is important, we believe that the starting point of the data dictionary is the alignment of the concepts and definitions used in of all the regulatory frameworks, i.e. prudential, resolution and statistics reporting and by all authorities at UE and national levels.
Besides, when a new data dictionary is to be implemented, it should be flexible enough to be implemented by all banks, easy to update, taking into consideration the costs of IT systems and internal processing updates when it comes to deciding on the frequency of data developments.
Furthermore, if reporting institutions could “report once”, harmonised and standardised data, i.e., produce a single report from which the relevant authorities could extract the data they require, it would be of great benefit to reporting institutions and authorities alike.
Processing data will facilitate reconciliation and controls, reducing data transformation
There is a strong need of consistent and homogeneous definitions that are clearly understandable to ensure the coherence of the reporting framework and to contribute to reduce the reporting requirement overlaps.
However, not all balance sheet items are relevant to be reported or can be explained at a granular level. Reporting on a granular basis is more appropriate for risk items, such as credit risk items. Other balance sheet items should remain at aggregated level within rules defined in accordance with the authorities.
The responsibility of the correctness of the data reported remains to the credit institutions as the data are derived from their systems.
The execution of the transformation rules should be under the responsibility of the authorities, as reporting institutions could not be responsible of transformation of data that are performed by authorities.
While under the push approach, the reporting institutions will oversee the transformation of the data they sent, transformation rules should be clearly defined jointly between authorities and reporting institutions.
Definition of protocols between authorities to enhance sharing data among themselves to contribute to the reuse of data and avoid duplication of same data.
• A common data dictionary with consistent definitions applicable to a large scope of data requested, i.e. statistical, prudential, resolution should be a prerequisite.
• A joint committee as a forum of authorities should be completed with the participation of the reporting institutions, in particular for the definition of the transformation rules from granular data to templates where the participation of reporting institutions would provide technical expertise. It is important as the responsibility of defining the transformation rules from granular data to templates must be exercised jointly by the authorities and the reporting institutions.
• Authorities must enhance data sharing among themselves.
• National and ad hoc requests should be limited to specific reporting to treat specific situations or activities. To avoid additional costs for reporting institutions to collect more granular data with persistent ad hoc requests and no burden reporting relief, we believe that a clear commitment of the authorities is needed to issue fewer ad-hoc requests and to reuse more the granular data they already receive.
1) Please explain which institutions you think should be considered by the Feasibility Study.
As a first step, we believe that the scope of the Feasibility Study should be limited to credit institutions that are required to report to CRR and national frameworks.For small reporting credit institutions, a principle of proportionality should apply.
2) Please explain which data collections you think should be considered by the Feasibility Study.
Data collections that should be considered are mainly those listed in the stocktake by the Feasibility Study, i.e. supervisory, resolution and statistics requirements at EU and national levels. Could also be included in the scope of the data collections other requirements such as SRB valuation data set (for the common part aligned with Anacredit) and Loan Tape standard (for the common part with Anacredit).In addition, the scope of the data collections should be limited to the data required by the regulation without extending it to management data and without anticipating future regulations.
Furthermore, we would also like to stress that if reporting institutions could “report once”, harmonised and standardised data, i.e., produce a single report from which the relevant authorities could extract the data they require, it would be of great benefit to reporting institutions and authorities alike.
3) Do you consider that the issues identified, the options proposed and the assessment approach taken throughout the discussion paper are relevant and complete? If not, please explain.
The EBA Discussion Paper on a Feasibility Study of an Integrated Reporting System under Article 430c CRR presents options for key areas in the development of an integrated reporting system across prudential, resolution, and statistical reporting in EU. It remains at a macro level of analysis and is neutral as to the technical results of an integrated reporting framework.We believe that the two areas among the four core areas considered for the creation of an integrated reporting that are key for the success of the project are a unique and common data dictionary and the related governance aspects. They should be considered first, before moving on to the other topics, i.e. the Central Data Collection Point and the level of granularity of data, that, while relevant, can be discussed at a later stage.
Besides, the EBA feasibility study is launched shortly after the ECB published the Cost-Benefit Assessment (CBA) questionnaire on the same matter related to an integrated reporting system. We would have hoped closer interconnection between the ECB's Integrated Reporting Framework (IReF) and the EBA's feasibility study. We question the differences in terms of scope - the ECB project is limited to statistical data whereas the EBA project addresses a broader scope covering prudential, statistical and resolution reporting – and in terms of progress of the two projects – the implementation horizon of the ECB’s IReF is 2024 – 2027, whereas the EBA study remains at a macro level analysis. The alignment of the implementation of both projects is essential. Different implementation timetables and scopes would imply that reporting institutions will incur costs without the full benefit of an integrated reporting. They would have to overhaul their IT infrastructure after the implementation of ECB’s IReF, once the integrated reporting system for supervision, statistics and resolution is defined by the EBA.
5) Do you confirm the findings presented in the stocktake? If you have additional information, please provide more specific details about the amount of data collected.
We share the preliminary conclusions related to the stocktake of the current reporting framework to be included in an integrated reporting system, i.e., prudential, statistical, resolution data.In addition, could be considered other requirements such as SRB valuation data set (for the common part aligned with Anacredit) and Loan Tape standard (for the common part with Anacredit).
Besides, one of the desirable objectives of an integrated reporting system is to reduce the number and frequency of ad hoc requests. While national data may be needed to fulfil national needs or ad hoc data may arise as the result of specific situations, we encourage authorities to have a clear commitment to issue fewer ad hoc requests and to reuse granular data already reported, so to provide a relief for the reporting institutions regarding their reporting burden.
8) Do you use one or more data dictionaries in your compliance and reporting processes?
Banks use multiple dictionaries, usually a combination of external reporting dictionaries from authorities and internal dictionaries.Besides, it should be noted that banks consider the DPM as a reference to fill in templates and perform controls. The DPM could not be considered and used as a dictionary as it is a technical tool and does not contain any definition. It could only be appended to the data dictionary.
9) What are the characteristics you think a data dictionary should have? Do you agree with the one referred to in this document? Do you think any characteristic is missing or should not be included?
Building up a common glossary should start from the existing regulatory frameworks and the assessment of the degree of common definitions and concepts that already exist. The common dictionary should also be elaborated in connection with the initiative of the Integrated Reporting Framework of the ECB. As considered in the study, the priority must be given to the principle commonly referred to “define once”, i.e., a common language and dictionary with consistent definitions applicable to a large scope of data requested including statistical, prudential and resolution data as it is not in the interest of either the reporting institutions or the authorities to pile up data dictionaries developed by different authorities.Accordingly, while we agree that the alignment of the semantic level (data definitions), the syntactic level (Metadata Model and Logical Data Model) and the infrastructure level (Database) is important, we believe that the starting point of the data dictionary is the alignment of the concepts and definitions used in of all the regulatory frameworks, i.e. prudential, resolution and statistics reporting and by all authorities at UE and national levels.
Besides, when a new data dictionary is to be implemented, it should be flexible enough to be implemented by all banks, easy to update, taking into consideration the costs of IT systems and internal processing updates when it comes to deciding on the frequency of data developments.
Furthermore, if reporting institutions could “report once”, harmonised and standardised data, i.e., produce a single report from which the relevant authorities could extract the data they require, it would be of great benefit to reporting institutions and authorities alike.
10) What is the role you think the data dictionary can have in regulatory compliance and reporting?
A unique data dictionary should be the “one language” used and shared by authorities and credit institutions. At the level of credit institutions, it will contribute to produce all regulatory requests in less burdensome and more efficient manner. By using uniform definitions, it will improve data quality, contribute to a better and common understanding of the reporting requirements and avoiding multiple interpretations of the same item. At the level of authorities, it will improve comparability.Please explain how would a standard data dictionary help institutions to improve the listed processes.
Data dictionary should be used as a unique reference of common definitions and concepts for all regulatory frameworksProcessing data will facilitate reconciliation and controls, reducing data transformation
12) How important is it for institutions to have a unique and standard data dictionary for all regulatory data with the aim of ensuring consistent use across the supervisory, resolution and statistical reporting?
Highly importantplease specify "other" from question 12 and include here any additional comment
It is of utmost importance to have a unique and standard data dictionary for all regulatory purpose under every approach (template driven or granular driven) and at EU or national level.There is a strong need of consistent and homogeneous definitions that are clearly understandable to ensure the coherence of the reporting framework and to contribute to reduce the reporting requirement overlaps.
13) How much would it cost to move to a unique regulatory data dictionary?
Highly costlyPlease include any other comment to question 13
It would be highly costly to implement a unique regulatory data dictionary, but the benefits would outweigh the costs as it will provide homogeneous definitions of items used in all reporting.14) How much cost reduction is expected by integrating the national regulatory reporting together with the harmonised reporting regulation into a unique data dictionary?
High cost reductionsPlease include any other comment to question 14
Integrating national regulatory reporting would potentially highly reduce costs to the extent that such those reporting will rely on common notions and definitions even if the data are used for national purposes.15) How much cost reduction is expected by integrating ad hoc regulatory reporting with harmonised regulation into a unique data dictionary?
High cost reductionsPlease include any other comment to question 15
A unique data dictionary will contribute to reduce costs provided that ad hoc requests will rely on definitions and concepts of the unique data dictionary and that they will be limited to specific purposes or specific situations and based on the reuse of data already reported regularly.18) For which reporting areas (prudential, statistical and resolution or modules/parts of these areas) may the use of granular data present a solution? (multiple choices)
prudentialPlease include any other comment to question 18
We consider that for all the reporting purposes the path for more granular data must be explored. But the decision of which templates or requests can be replaced by granular datasets must be conducted very carefully to assess their relevance and avoid any unintended consequences. Indeed, for the prudential reporting, legal constraints must be carefully assessed notably calculation of ratios: more relevant at consolidation level on an aggregated basis.Feasable and preferable
option 2Please specify "other" from question 19
Option 2 would be the preferred option.However, not all balance sheet items are relevant to be reported or can be explained at a granular level. Reporting on a granular basis is more appropriate for risk items, such as credit risk items. Other balance sheet items should remain at aggregated level within rules defined in accordance with the authorities.
22) What possible aspects related to the design of the option (Question 19) would make the costs for this option higher than the benefits and therefore not worth implementing?
The elements that would contribute to the success of the implementation of the option 2 would not notably be an appropriate and well-defined transition to the data driven approach from the current template approach23) If transformations are to be defined (as depicted in Option 2 or Option 3), who should be responsible for their definition (e.g. who takes responsibility for their correctness) and their execution?
Authorities and reporting institutions jointlyPlease include any comment to question 24
The responsibility of defining the transformation rules must be exercised jointly by the authorities and the reporting institutions. Thus, it is important that reporting institutions participate to the definitions of the transformation rules alongside with the authorities under a Joint Committee. Indeed, reporting institutions need to be able to know what type of queries or aggregations have been extracted, to replicate what supervisors have done and to be in capacity to answer to supervisors to any questions raised.The responsibility of the correctness of the data reported remains to the credit institutions as the data are derived from their systems.
The execution of the transformation rules should be under the responsibility of the authorities, as reporting institutions could not be responsible of transformation of data that are performed by authorities.
29) Is your institution reporting to different authorities in your home country?
Not applicable30) Is your institution reporting to other authorities in host countries?
Not applicable40) Would you prefer the future integrated reporting system to be based on:
A push approachpush
Under a push approach, the reporting institutions will remain responsible for the data they sent, the explanation they provide regarding these data, the monitoring and updating of their own internal reporting process. This will contribute to the quality of the dataWhile under the push approach, the reporting institutions will oversee the transformation of the data they sent, transformation rules should be clearly defined jointly between authorities and reporting institutions.
Feasibility of the central data collection point
Harmonization of the reporting requirementsCollection and access to granular data by authorities
Compliance with the General Data Protection Regulation (GDPR)Definition of protocols between authorities to enhance sharing data among themselves to contribute to the reuse of data and avoid duplication of same data.
The responsibility of the reporting institutions for the reported data, if the transformation of granular data is conducted by authorities (i.e. not the reporting institution)
as supervisors will be able to do all sorts of aggregations and drill downs with the extracted data, reporting institutions will need to be able to know what type of queries or aggregations have been extracted, for example to replicate what supervisors have done and to be in capacity to answer to supervisors to any questions raised. Therefore, the responsibility of defining the transformation rules from granular data to templates must be exercised jointly by the authorities and the reporting institutions under a Joint Committee where reporting institutions could bring their technical expertise.43) Do you agree with the suggested coordination mechanism for data requests? Do you see any benefits or disadvantages in this approach?
While the suggested coordination mechanism for data requests has many benefits, we would like to highlight the following points:• A common data dictionary with consistent definitions applicable to a large scope of data requested, i.e. statistical, prudential, resolution should be a prerequisite.
• A joint committee as a forum of authorities should be completed with the participation of the reporting institutions, in particular for the definition of the transformation rules from granular data to templates where the participation of reporting institutions would provide technical expertise. It is important as the responsibility of defining the transformation rules from granular data to templates must be exercised jointly by the authorities and the reporting institutions.
• Authorities must enhance data sharing among themselves.
• National and ad hoc requests should be limited to specific reporting to treat specific situations or activities. To avoid additional costs for reporting institutions to collect more granular data with persistent ad hoc requests and no burden reporting relief, we believe that a clear commitment of the authorities is needed to issue fewer ad-hoc requests and to reuse more the granular data they already receive.