EBA欧洲银行-CRD4-DPM-Database-description2-_5页_380kb

Summary of ITS on Supervisory Reporting

Core Content

The European Banking Authority (EBA) has developed a Data Point Model (DPM) as part of the draft Implementing Technical Standards (ITS), initially outlined in consultation papers CP 50 and CP 51. The DPM aims to provide a structured and coherent framework for supervisory reporting, particularly for COREP and FINREP templates. To improve the DPM's methodology and usability, the EBA has introduced a relational database as the central repository for DPM metadata, moving away from traditional MS Excel structures. MS Access was selected for its simplicity and ease of use in this context.

Main Purposes of the Database

• Enhanced Coherence: The database introduces logical constraints through referential integrity, ensuring consistency and reducing errors in the DPM.
• Automatic Consistency Checks: It enables the execution of automatic consistency checks, which help validate the model and improve the quality of data points.
• Flexible Views: The database allows for the creation of multiple views of the same metadata, facilitating better interpretation and understanding of the reporting framework.
• Dimensional Categorisation: It explicitly defines dimensional data points, with clear links to business templates, eliminating the need for manual interpretation.

Key Features of the Meta-Model

The DPM meta-model is technology-neutral, not bound to any specific data exchange format like XBRL. However, an extra layer for XBRL properties (e.g., namespaces) will be added in future versions to streamline the translation process from DPM to XBRL taxonomies.

The meta-model is structured around three main groups of concepts:

1. Representation of Templates' Metadata
2. Dimensional Concepts for Categorisation
3. Links Between Them (Categorisation)

Dimensional Concepts

• Dimensions: Categories used to describe data points (e.g., "Counterparty").
• Members: Specific instances of dimensions (e.g., "Central banks").
• Domains: Grouping of members by semantic nature (e.g., "Geographical area" domain includes members like "Residence", "Location", and "Country").
• Hierarchies: Define parent-child relationships among members and enable aggregation at different levels. These are not yet included in the current database version.

Categorisation Process

• Each table cell (excluding grey-shaded ones) corresponds to a unique DataPoint.
• DataPoints are identified through the unique combinations of [Dimension].[Member] pairs.
• Grey-shaded cells indicate data not required or logically invalid combinations.
• Data type (e.g., monetary, percentage, integer) and period type (stock or flow) are automatically assigned during the modelling process.

Supporting the DPM Process

• Template Analysis: Business experts define dimensions and members for rows and columns.
• Header-Level Categorisation: Some templates require categorisation at the header level, which is then applied to all cells in the sheet.
• Iterative Process: The DPM process is iterative, involving:
  1. Template Analysis & Modelling
  2. Loading & Generating Metadata
  3. Running Consistency Checks & Analyzing Errors

Future Enhancements

The next version of the DPM database will include:

• Versioning of metadata to track changes over time.
• Validation Rules to ensure data integrity.
• XBRL Properties for easier integration with XBRL taxonomies.
• References to Legislation to provide legal context.

Database Access

The database is available on the EBA website under the file name:
"CRD4 DPM - Database.accdb"
The database diagram can be accessed via the menu:
"Database Tools/Relationships"

Contact for Feedback

General comments or questions regarding the database design or related topics not covered in the DPM consultation can be sent to:
carlos.martins@eba.europa.eu