DPM data dictionary

The Data Point Model (DPM) is the data dictionary providing the structured representation of the data required for regulatory purposes by the EBA. It encompasses its regulatory processes and integrates the data of different EBA Technical Standards and Guidelines. The DPM may include also other data definitions relevant to the common understanding of the EBA regulatory data - business concepts and their relations, as well as their validation and calculation rules.

Since its inception, the EBA has been using the same DPM Standard 1.0. In 2020, the EBA and EIOPA launched the joint DPM Refit project to modernise and harmonise the Data Point Model methodology used by the two authorities. The project aimed to establish a common, technology-independent standard and supporting toolset capable of accommodating the growing volume, granularity and complexity of regulatory data, integrating different reporting frameworks and modelling approaches, and reducing complexity, duplication and maintenance costs. The resulting DPM Standard 2.0 was published in June 2023.

DPM Alliance

In March 2024, the EBA, EIOPA, and the ECB strengthened their cooperation through a Memorandum of Understanding, establishing a joint DPM Alliance to govern the standard going forward with the SRB participating as an observer.

More recently, the DPM Alliance has been working in close coordination with the JBRC—which is leading efforts on the semantic integration of statistical, supervisory, and resolution reporting - through the development of the DPM 2.1 metamodel. This new version introduces capabilities and model enhancements designed to support the integration of ECB statistical data, including the future Integrated Reporting Framework (IReF), with the existing EBA prudential and resolution reporting frameworks.  

Links

DPM 2.1 (future release)

The new features introduced in DPM 2.1 build upon and fully preserve all existing DPM 2.0 capabilities, ensuring complete continuity with the current framework. In addition, DPM 2.1 introduces eight key enhancements that have been identified as important enablers for the continued convergence of the EBA and ECB reporting ecosystems. These improvements support the broader objective of developing and maintaining a shared data dictionary while extending the DPM's ability to accommodate both supervisory and statistical reporting requirements.

The principal benefits of DPM 2.1 include:

  • Enhanced standardisation through the use of patterns (regular expressions) for property validation.

  • Greater flexibility in maintaining and versioning item names and descriptions.

  • Improved modelling and management of compound properties.

  • Enhanced organisational modelling through structured hierarchical organisation data.

  • Support for conceptual modelling through the explicit distinction between logical and physical frameworks.

  • Clear identification of the role and origin of modules within the data lifecycle (reported, calculated, or disseminated).

  • Improved consistency and harmonisation of naming and labelling conventions.

  • Better support for the convergence of supervisory, resolution, and statistical reporting through a common data dictionary.

A more comprehensive description of the new features can be found in the sections below:

1. Regex-Based Property Validation

Users can define regular expression (Regex) patterns directly on properties to enforce value constraints. This provides a simpler alternative to creating multiple validation rules, improves data quality, and aligns DPM with the SDMX Facet concept.

2. Item versioning on names and description 

Item names and descriptions become versioned alongside existing versioned attributes. This allows users to track the full history of item definitions and regulatory references across releases.

3. Compound Property Support

DPM can explicitly represent compound properties and their underlying base properties. This supports more sophisticated business modelling scenarios while preserving contextual information associated with specialized properties.

4. Organizational Hierarchies

DPM can explicitly represent different domains, departments, or units within the same organization and define their hierarchical relationships, improving organizational modelling and governance.

5. Logical and Physical Framework 

Frameworks can now be explicitly classified as Logical or Physical 

This helps users distinguish conceptual models from implementation models and relate them through the existing DPM structures.

6. Modules classification

Modules can be explicitly classified according to their origin or purpose as reported, calculated or disseminated.

This provides greater transparency regarding how each module is used and managed.

7. Compound Property Support

DPM can explicitly represent compound properties and their underlying base properties. This supports more sophisticated business modelling scenarios while preserving contextual information associated with specialized properties.

8. Naming Consistency

Field names are standardized by replacing "Label" with "Name" in selected entities, improving consistency and clarity throughout the model. 

9. Explicit JSON Taxonomy in module version

Each module version can include its JSON taxonomy directly in the DPM database. This makes glossary usage and taxonomy information readily available to users and supports a variety of downstream use cases.

The DPM data dictionary enables the integration of data from different banking regulatory frameworks, by providing a clear interpretation and identification of each data element to be shared by all relevant stakeholders.

The DPM data dictionary copes with change and evolutionary regulation requirements. Therefore, new data requirements or new definitions can be added, but all previous data requirement frameworks can still be tracked.

The DPM bridges the communication gap between policy experts, data analysts and IT by providing a common platform of understanding. The business concepts are specified in the DPM according to formal rules, as required by IT experts, while being still manageable by policy experts and other data users.

The DPM provides the metadata structure and contents that provide a machine-readable data dictionary to support the fully automation of regulatory processes of data exchange, analysis, and disclosure.

The DPM is a cornerstone of EBA metadata centred approach and is composed by three distinct facets: the structure (syntactic model), the process (methodology) and the content (semantic definitions).

The DPM contains information about what data should be reported and how it should be exchanged. It includes:

  • A common core glossary of financial and prudential concepts, arranged in different categories and ready to be used  in the definition of the regulatory data;
  • A complete set of data concepts uniquely identified as part of the regulatory frameworks, described by a distinct combination of glossary elements and tracked along time by a consistent versioning;
  • The reporting templates, their rendering information, the mapping of template cells with the data concepts, and the agreed validation rules and/or calculation rules. Additional metadata required for defining reporting obligations and to generate standards data exchange formats (XBRL taxonomies) is also included.

The DPM does not contain information about institutions which are the subject of a complementary Master Data Management system.

The DPM syntactic model is a stable metadata structure that includes all elements that fit the needs of data integration of different types of reporting.

The DPM contents are the result of a systematic data definition process with embedded automated checks, which translates each piece of regulation into a formal element in the data dictionary and ready to be used by electronic information systems, or directly accessed by people.

The core part of the process is performed by policy reporting experts who are involved in the design of reporting templates, instructions, and the legal texts, and produce a formal description of the complete set of the required data concepts. The process starts with the modelling of each template row and column (and sheet, if applicable), followed by the assignment of a specific dimensional categorisation – set of pairs of elementary concepts (Members/Items) of different categories (Dimensions/Properties) that fully describe the concept associated to the row/column/sheet.

The full categorisation of a data concept (data point / variable) is obtained by combining the categorisations of the cell position coordinates in an automated process that generates the dimensional categorisation for each individual cell and checks the consistency and uniqueness of each data concept definition. The modelling process must iterate until the DPM overall consistency is complete, taking in account all the frameworks and correspondent data definitions already included in past regulatory processes. The XBRL taxonomies or other outcomes are generated by IT processes, taking the DPM as an input.  

The results of the successive data definition processes are stored in the DPM database, which has a formal metamodel structure implemented as a relational database, agnostic to any particular financial or regulatory technology and available on the EBA website as a free public good.

The DPM database can be incorporated as a component of IT solutions, thus fully leveraging the possibilities of integration of digital processes across the different phases of regulatory life cycle, and between all the institutions involved in the reporting, validation, calculation, analysis and disclosure of financial data.

The DPM lifecycle is supported by a set of ongoing processes organised in projects to add or amend the data definitions of existing or new reporting frameworks. Each of these projects delivers a new DPM release which always add new versions of data definitions assuring the overall consistency and integration with other data already defined in the data dictionary .

The EBA staff has been developing the DPM data dictionary in cooperation with banking and policy experts appointed by national competent authorities of the EEA countries, who regularly participate in the EBA committees and working groups.

The development of the DPM started in 2012 with the first publication as a database format in 2013. Since then, the DPM contents have been included in the same model adding the successive versions of the regulatory data requirements, evolving the requirements and the data definitions amended by each new reporting framework decided and decisions of the Single Rulebook Q&A process, always with the systematic feedback received from the Authorities' experience on the data integration effort.

Regulators, supervisors, financial institutions, service providers, other organisations, and the general public have access to the DPM data dictionary, and can use it as the common repository of clear and structured specifications of the data referred in Regulation.

The DPM serves, in different ways, both policy and IT experts involved in the regulatory data processing chain, from data exchange harmonization to data analysis and dissemination, and streamlining the processes and promoting systems interconnection of those who are dealing with regulatory data within and across institutions.

The DPM is designed as a metamodel in order to ensure the necessary flexibility for extending the EBA Technical Standards and Guidelines to new requirements. The DPM new releases are implemented by adding new metadata content to the same database metamodel, without any change at the level of its structure. The metadata model also enables the set-up of logical constraints on metadata content, and the definition of relational rules to ensure the consistency of the data contents.

The complete structure of the DPM database includes a high number of tables and relationships, which are needed in different contexts. However, the initial approach to the DPM can be facilitated by focusing just on the core part of the metamodel, which includes only a relatively small number of concepts of the glossary, the templates structuring and the data concepts. And a part dedicated to define the validation rules and shared calculation rules.

Click here to find more details on the DPM Standard 1.0 metamodel

Click here to find more details on the DPM Standard 2.0 metamodel

Click here to find the different functionalities covered by both DPM Standard metamodels