The IReF/BIRD implementation represents a paradigm shift in banks' regulatory reporting

The Integrated Reporting Framework (IReF), together with the Banks' Integrated Reporting Dictionary (BIRD), will fundamentally change the future of statistical reporting. For the first time, reporting obligations will no longer be fulfilled by submitting forms or reporting forms, but by providing granular data points.

With this paradigm shift, IReF pursues the main goal of harmonising statistical reporting in the EU. This is achieved through a uniform definition of requirements, as well as harmonized frequencies and reference data for different reports. Currently, for example, cross-border institutions face challenges due to different definitions and reporting requirements of different national regulatory authorities. IReF will be mandatory for all institutions in the EU, applying the principle of proportionality. The implementation of the IReF provides for the provision of the necessary data points (rather than forms) in a central location from which the various national authorities can retrieve the necessary data to produce their own reports and analyses.

BIRD is an initiative to establish a data dictionary that will contain a uniform definition of reporting concepts ("Define and Report once"). BIRD is currently being developed by the European Central Bank (ECB) together with the banking industry, but in contrast to IReF, implementation is optional for the institutions, because BIRD is not a classic IT tool. In order to achieve the full degree of standardization, the BIRD definitions must be integrated into the Institute's reporting architecture. The latest version of BIRD, version 6.5, was released by the ECB in February 2025. In the future, the results of the BIRD initiative will be published under the Apache 2.0 open source license. The reason for this is that BIRD is treated as a public good, and should therefore be freely available to all banks and interested parties, such as software providers. This is intended to increase the adoption of BIRD while simplifying the creation of supporting tools by different vendors.

Possible approaches to implementing IReF/BIRD
Based on the ECB's current timeline for the start of IReF reporting in the fourth quarter of 2029, institutions should set the fourth quarter of 2028 as the internal deadline for implementation, as it should ideally be completed before the pilot phase. This leaves a good three years for full implementation. Even if the period until the pilot phase still seems comfortable, institutions should evaluate in a timely manner which implementation approach suits them best, as implementation can be complex and time-consuming. Based on this, a target operating model (TOM) must then be developed, an implementation plan drawn up and budget and resources allocated.

In the following, we present three possible implementation approaches for IReF and BIRD with their main advantages and disadvantages. The choice of the right approach depends in particular on the individual situation of the respective institution.
The following factors must be taken into account:

• Size of the Institute
• Complexity of the business model
• Status of the current IT and reporting environment
• Amount of statistical reporting obligations (e.g. cross-border business including branches)
• Degree of standardization that the institute wants to achieve
• Time and effort required
• Desired flexibility in the use of different tools and processes

The IReF-only approach
With this approach, only the requirements of the IReF are implemented and no implementation of BIRD is carried out. The focus here is essentially on minimizing the implementation effort and risk. Only what is necessary for compliance with the IReF Regulation is implemented.

The main advantage of this implementation approach is that the total cost of implementation, as well as the time required for it, and also the implementation risk can be significantly lower compared to other approaches. In this approach, institutions can use the data dictionary or model of their choice, such as their own or that of a software vendor. Alternatively, a combination of data models may be considered suitable. If the institution already uses reporting software from an established provider for regulatory reporting, it can probably continue to use the software (possibly supplemented by an IReF module) for its reporting purposes. A prerequisite for this is that the data structure of the reporting software is granular enough to integrate the IReF requirements, as there will usually be a mapping between the granular data structure of the reporting software and the IReF Collection Layer. In this case, institutions should obtain confirmation from the reporting software provider at an early stage that the data model and software currently in use will meet the requirements for IReF. Changes to the data source systems/databases, the interfaces to delivery or to the reporting processes (outside of the software itself) are determined based on the individual situation of the institute and the IReF requirements, as the end-to-end data delivery to the reporting software is already in place. However, minimal changes to the data structure cannot be ruled out, which are necessary to fully comply with the IReF requirements. A possible need for adjustment will only become apparent after the publication of further information on the implementation of IReF by the ECB.

One of the drawbacks of the IReF-only approach is that it does not implement the standardised data dictionary (BIRD) envisaged by the ECB. This may lead to increased reporting overhead in the long term, as new regulatory requirements are expected to be based on the BIRD data model.

In addition, if the institution chooses a proprietary data model, there is a certain degree of dependency on the software provider, which could lead to a reduction in flexibility compared to institutions using other approaches.

IREF Collection Layer is the reporting scheme of IReF and establishes the connection between BIRD and IReF. This was presented for the first time as part of BIRD version 6.0. An extended technical layer is also to be created for country-specific data requirements (source: Bundesbank)

The IReF-first-BIRD-later approach
In this approach, both IReF and BIRD are implemented in phases. In the first phase, only IReF is implemented (similar to the IReF-only approach). After the institute is IReF compliant, it can choose to implement BIRD (either fully or partially) to achieve a higher degree of standardization.

The advantage of this approach is the lower implementation or project risk, which is mainly due to lower complexity compared to a comprehensive BIRD and IReF implementation. In the first phase, the focus is on ensuring timely IReF reporting capability and taking advantage of the IReF-only approach. This also includes the possibility of using the existing reporting software with its advantages and disadvantages, as described in the paragraph above. The subsequent implementation of BIRD has the advantage that all participants (authorities, institutions) have gained sufficient practical experience with BIRD by the time it begins, and the quality and completeness of the uniform data model should be significantly improved. In addition, the BIRD implementation can also have the benefit of reducing the number of queries from the supervisory authority regarding the submitted data due to the uniform data language. In addition, there is also an opportunity to use multi-party service providers.

A disadvantage of this approach is the additional effort and time required to implement BIRD separately in the second phase. Some of the implementation steps (for example, data quality checking) may need to be done twice – first for IReF and later again for BIRD. In addition, a decision must be made as to which parts of the existing reporting architecture need to be changed in order to implement BIRD, or what degree of alignment with BIRD should be implemented in order to achieve better standardization and thus exploit automation potential.

Institutions that are already using a reporting software provider for regulatory reporting may choose to follow the IReF-first-bird-later implementation approach to first ensure IReF reporting capability based on the vendor's data model and current software, and then decide how much flexibility and standardization they need in the future. This would depend on how different the final version of BIRD will be compared to the software vendor's data model.

The IReF and BIRD approach
This approach implements BIRD and IReF, where BIRD is considered the basis for IReF. This requires an initial analysis to verify the compatibility of the existing data repository with the BIRD definitions. As both BIRD and IReF must be implemented by the respective launch or pilot phase, institutions adopting this approach should start the compatibility analysis at an early stage on the basis of the BIRD resources published so far on the ECB's website.

A major advantage of this approach is the high degree of standardization. Full alignment with the unified BIRD data dictionary means automation capabilities can be implemented more easily. It is possible to use artificial intelligence and cloud solutions to develop tools and integrate them into the reporting landscape. These can include data analysis tools, data quality verification tools, testing tools, and data aggregation tools, among others. The approach also makes it possible to outsource certain aspects of regulatory reporting with less effort in the future (use of multi-party service providers). A high degree of standardization can also lead to a reduction in ad hoc data requirements from supervisors, which can contribute to a significant reduction in resources in the long term. It can also be assumed that new regulatory requirements will be easier to implement, as new supervisory requirements are expected to be aligned with the BIRD standard for data definitions.

One disadvantage is the initial implementation costs and risks. The simultaneous implementation of BIRD and IReF requires detailed planning and monitoring to minimize problems and detect them at an early stage. The implementation costs are also high, as the data repository as well as existing reporting systems and processes may have to be revised in order to fully meet the new requirements (the degree of change depends on the previous compatibility analysis).

For institutions with cross-border operations and foreign branches, it may be beneficial to follow the IReF and BIRD implementation approach to maximize the benefits of standardized data requirements across borders. This is particularly relevant as institutions with cross-border activities and foreign branches are currently faced with different national reporting requirements, where they have to maintain different national data models and heterogeneous data definitions. This is inefficient and costly. In addition, there are sometimes different IT landscapes in these institutions, which hinder the granular storage and provision of data. In general, data is provided to headquarters in aggregate, for example via reporting packages.

Even institutions that do not use an established provider of regulatory reporting software, but have their own reporting solutions, could decide to follow the IReF and BIRD implementation approach to build their regulatory reporting with the standardized data model.

Preparing for the implementation of IReF/BIRD
IReF will fundamentally change the reporting system of banks and thus the selection of the right implementation approach is also of great importance. For this reason, institutions should allow sufficient time for analysis and decision-making. Only when a clear picture of the exact situation of the institute is available can the decision be made on a tailor-made implementation approach and concrete measures for action can be derived. Therefore, we recommend that you start as soon as possible with an analysis of the initial situation, in which we will be happy to support you with our experience.

The strategic decision on the right approach should also take into account that the long-term goal of supervision is the Integrated Reporting System (IRS). In addition to statistical reporting, the IRS is also to include prudential reporting (such as CoRep; FinRep) as well as resolution reporting, so that in the long term these are no longer to be reported on the basis of reporting forms, but are data-based. The first step towards this integration has already been taken with the inclusion of FinRep and Asset Encumbrance in BIRD. The integration of CoRep will start later this year, according to the BIRD 2025 work plan.

Institutions are required to consider their entire reporting landscape in light of the planned IRS in order to choose the right implementation approach that meets their vision for regulatory reporting in the future. We will be happy to assist you as a reliable partner.

Result
IReF and BIRD will fundamentally change regulatory reporting in the EU – away from reporting templates to granular data. While IReF will be mandatory from 2029, BIRD will remain optional but offer benefits through standardization. Institutions should choose the appropriate implementation approach at an early stage in order to control costs, effort and risks and to ensure more efficient processes in the long term.