Risk assessment

Tax risk assessment is a key element of modern tax administration. It allows tax authorities to identify indicators that suggest specific taxpayers or arrangements may pose an increased risk to their jurisdiction and require further actions in terms of compliance. In general, EU tax authorities use automated methods based on domestic data and information received from other jurisdictions. Yet, a manual element may remain, as (i) tax authorities vary in terms of whether tax risk assessment is conducted centrally by a specialist risk assessment team incorporating input from the compliance function, or locally by the compliance team (or tax inspector); (ii) some data types remain challenging to be automatically processed, e.g. literal summaries.

Simultaneous Control

Tax controls carried out simultaneously by 2 or more tax authorities in their own territory, on one or more taxpayers of common interest.

Spontaneous Exchange of Information - SEOI

Unsystematic provision of information that the supplying tax authority deems to be of interest to the receiving tax authority.

Visits on DAC - VISDAC

Visits to Member States to support the effective implementation and functioning of the DAC, with a focus on quality of information exchanged and use of information.

European Union Advanced International Administrative Cooperation - EU -AIAC

Network of experts from 27 tax authorities which supports and promote the use of Advanced International Administrative Cooperation (AIAC) - i.e. multilateral controls (MLC, a term covering both simultaneous controls and joint audits) and Presence in Administrative Offices and Participation in Administrative Enquiries (PAOE) - on complex cross-border arrangements and complicated cases of tax fraud, tax evasion and tax avoidance in VAT, excise duties, and direct taxes. It is funded by FISCALIS.

Acronym	Meaning or definition
AML	Anti-Money Laundering
BEPS	Base Erosion and Profits Shifting
BO	Beneficial ownership
CCN/CSI	Common Communication Network/Common Systems Interface
CESOP	Central Electronic System of Payment
CRS	Common Reporting Standard
ECA	European Court of Auditors
EOI	Exchange of Information
EP	European Parliament
LPP	Legal professional privilege
MCAA	Multilateral Competent Authority Agreement
MNEs	Multinational Enterprises
OECD	Organisation for Economic Co-Operation and Development
TIN	Taxpayer Identification Number
TSI	Technical Support Instrument
VIES	VAT Information Exchange System

1.Introduction

Council Directive 2011/16/EU on administrative cooperation in the field of taxation ( DAC , DAC1 or the Directive) replaced the previous mutual assistance framework from 1977 and established rules and procedures for close cooperation between Member States’ tax authorities on direct taxation 1 matters.

The Directive aims to ensure the correct assessment of taxes in cross-border situations and to combat tax fraud, evasion and avoidance. To achieve this, the DAC sets up a common framework for cooperation between Member States, enabling them to assist each other by way of the exchange of information (on request, automatically or spontaneously) as well as other forms of cooperation such as participation in administrative enquiries, presence in administrative offices, simultaneous controls and joint audits.

Article 27(1) of the DAC requires the European Commission to submit a report on the application of the DAC to the European Parliament and the Council, every 5 years from 1 January 2013. The first evaluation of the DAC 2 covering the period from 2013 to 2017, was published in 2019 and was based on a study carried out by an external contractor. The present evaluation is the second iteration and covers the period from 2018 to 2023 3 .

The Directive and its main cooperation tool automatic exchange of information (AEOI) is relatively recent. Although mechanisms to facilitate bilateral exchange upon request featured in the previous Directive from 1977, the EU introduced AEOI in 2005 through Council Directive 2003/48/EC ( Savings Tax Directive ), however with limited scope. In 2011, DAC1 expanded AEOI to cover a broad category of income and capital. Since then, the framework has been constantly evolving to remain in line with international developments which have resulted in the expansion of the information to be reported and exchanged on: financial account information ( DAC2 ), cross-border tax rulings and advance pricing agreements ( DAC3 ); country-by-country reports (CbCR) ( DAC4 ); reportable cross-border arrangements (i.e. arrangement presenting an indication of a potential risk of tax avoidance) ( DAC6 ); income received from platforms ( DAC7 ); and crypto assets and e-money ( DAC8 ). Amendments also allowed tax authorities to have access to certain anti-money laundering (AML) data ( DAC5 ) and, more recently, extended the scope of AEOI on royalties, non-custodial dividends and cross-border rulings related to individuals and the possibility to carry out joint audits ( DAC7 and DAC8 ). The most recent amendment ( DAC9 ) provides for the exchange of information required by the Council Directive on ensuring a global minimum level of taxation for multinational enterprise groups and large-scale domestic groups in the Union .

The scope of the current evaluation includes all the amendments to DAC up to DAC5 which were already covered in the first evaluation and goes further by including DAC6 (which was previously excluded in the 2019 evaluation as it had not yet been transposed and implemented by Member States). Regarding DAC7, only qualitative information is provided, as the first relevant exchanges took place in February 2024. DAC8, on the other hand, is not covered in this evaluation, as the first relevant exchanges will commence in September 2027. However, as a result of the amendments to Article 27(2) of the DAC (by way of the amendment to DAC8), the scope of the current evaluation includes the hallmarks for the exchange of information on potentially harmful cross-border arrangements provided for by DAC6.

The key objectives of this evaluation reflect those of the first evaluation by aiming to evaluate the DAC and its amendments based on the 5 better regulation criteria – effectiveness, efficiency, coherence, EU added value and relevance 4 .

The evaluation firstly examines whether the DAC is effective in providing the desired outcomes and impacts. It assesses whether the information exchange is usable in terms of completeness, quality and timeliness, and whether it is possible to match the information received with specific taxpayers in the receiving Member State. It also examines whether the DAC has sufficient deterrent effect and assesses its impact on the twin objectives of safeguarding Member States’ tax revenues and improving the functioning of the single market.

Secondly, the evaluation looks at the efficiency of the exchange of information and other tools of cooperation provided for under the DAC, in cost-benefit terms. In line with the Commission’s ongoing action to rationalise and simplify reporting requirements for companies and administrations, specific attention has been given to identify areas where there is potential to reduce unnecessary regulatory costs and simplify the design and implementation of the Directive.

Thirdly, the evaluation assesses whether the DAC’s scope and purpose remain relevant, and if it addresses challenges that Member States are facing in terms of correct assessment of taxes in cross-border situations whilst fighting tax avoidance and evasion. Furthermore, especially given the successive amendments to the DAC, it is necessary to evaluate its internal and external coherence as well as its consistency with other relevant EU and international initiatives, mainly with respect to standards developed at OECD.

Lastly, the EU added value is assessed, especially in comparison to other internationally available tools to facilitate the exchange of information.

1.1. The evaluation methodology and sources

The methodology 5 used in this evaluation is in line with that of the first evaluation to ensure the comparability of results.

The evaluation methodology consisted of 3 main phases: (i) a structuring phase, which included preparatory activities for designing and developing the evaluation – such as selecting an external contractor and identifying relevant stakeholders to be consulted; (ii) data collection and analysis, involving stakeholder consultations, legal and desk research, and analytical work to support the assessment; and (iii) evaluation and insights, where the collected information was analysed to produce the evaluation report by examining the 5 better regulation criteria and subsequently drawing conclusions and lessons learned.

The evaluation is primarily based on the findings of the external study carried out in 2023 by a consortium led by Ramboll. Other sources include the results of the public consultation conducted by the Commission services in 2024, and a collection of data sources, such as feedback provided by Member States and private stakeholders 6 . Data from Member States includes their replies to the yearly questionnaires (on the functioning of the DAC and on the assessment of AEOI) as well as their replies to specific Commission led questionnaires (e.g. 2 surveys on the relevance of Annex IV on DAC6 hallmarks). The outcome of relevant work under the Fiscalis programme has also been taken into account (i.e. FPG/037, FPG/038, VISDAC Expert Team, FWS/022, FWS/153).

2.What was the expected outcome of the intervention?

2.1. Description of the intervention and its objectives

An increasingly globalised economy characterised by cross-border transactions, growing taxpayer mobility and the expanding digital landscape, presents constant challenges to the national application of direct taxation. Member States struggle to protect their national tax bases and therefore require greater access to information across jurisdictions to ensure accurate tax assessments, particularly for multinational enterprises (MNEs). As the economic environment evolves, so does the need for more harmonised policy approaches on the exchange of information to address emerging and evolving challenges, as described below 7 .

2.1.1.DAC1: Early Development and Introduction of Administrative Cooperation (2011)

In 2009, the Commission proposed a new framework for enhanced administrative cooperation. It became increasingly apparent that the model developed in 1977, was no longer fit for purpose in addressing the modern challenges faced by national tax authorities. DAC1 was adopted by the Council in 2011 and transposed by Member States by 2013, covering all taxes expect VAT and excise duties.

The core issue driving the need for reform was the increased level of cross-border activity in the EU, meaning Member States required greater levels of multilateral information for the purposes of direct taxation. Therefore, the Directive sought to promote better cooperation, establish common rules, and enhance trust between tax authorities, through a more efficient AEOI, improved communication channels, and the creation of common procedural standards.

DAC1 laid the foundations for large-scale AEOI on 5 categories of income and assets, namely (i) employment income, (ii) director’s fees, (iii) life insurance products, (iv) pension income, and (v) ownership of immovable property. It also provided Member States with a harmonised framework for cooperation instruments such as (i) exchange of information on request (EOIR) used when a Member State seeks additional information from another Member State for tax purposes and that the requested information is relevant to the tax affairs of one or several taxpayers; (ii) spontaneous exchange of information (SEOI), which takes place if a Member State discovers any information relevant or useful to another Member State on taxation matters, and sends it to that Member State; (iii) the presence of officials of a Member State in the offices of the tax authorities of another Member State (PAOE); (iv) simultaneous controls allowing 2 or more Member States to conduct controls of taxpayers of common or complementary interest; and (v) requests for notifying tax instruments and decisions issued by the authority of another Member State. 2 additional categories of income (royalties and non-custodial dividends) were introduced for exchange under the AEOI framework through the DAC7 and DAC8 amendments to DAC1.

In 2012, the Commission published an Action Plan to strengthen the fight against tax fraud and tax evasion. The Recommendation on aggressive tax planning highlighted the need to promote the AEOI as the future European and international standard for transparency and exchange of information in tax matters.

In its conclusions of 22 May 2013, the European Council called for the extension of AEOI at both EU and global level with a view to combating tax fraud, tax evasion and aggressive tax planning, which led to several amendments to the DAC.

2.1.2.DAC2: Expansion of Automatic Exchange of Information on financial assets and income (2014)

To address the tax transparency challenges posed by bank secrecy and hidden wealth in foreign accounts, the OECD developed the Common Reporting Standard ( CRS ). In response, the Council adopted DAC2 in 2014, expanding the automatic exchange framework to financial account information and marking the end of the bank secrecy in the EU. DAC2 aligned with OECD CRS to reduce administrative costs and burdens for tax authorities and businesses by preventing multiple reporting and exchange standards. In addition, a subsequent amendment to DAC2 introduced enhanced due diligence requirements for financial institutions, aiming to strengthen the accuracy and completeness of information reported (DAC8).

2.1.3.DAC3: Ruling Transparency and Information Exchange on Tax Rulings (2015)

Advance tax rulings (ATRs) and advance pricing arrangements (APAs) promote tax transparency and legal certainty, fostering investment in the single market. However, some rulings have been used to enable artificially low taxation and profit shifting. To address this, the Council adopted DAC3, building on OECD’s Base Erosion and Profit Shifting BEPS Action 5 , to mandate the automatic exchange of information on cross-border tax rulings and transfer pricing arrangements amongst Member States with a view to increasing transparency, curbing harmful tax practices and ultimately reducing the opportunities for tax avoidance. It also established a central directory to enhance coordination and reduce administrative burdens. By way of the DAC8 amendment to DAC3, the automatic exchange of ATRs was extended to cover rulings involving the tax affairs of individuals when either (i) the amount of the transaction or series of transactions covered by the ruling exceeds EUR 1.5 million, or (ii) the ruling determines, or is intended to determine, the tax residence of the individual.

2.1.4.DAC4: Addressing Multinational Enterprises and BEPS (2016)

The financial crisis and its consequences brought greater public attention to the tax practices of MNEs exploiting loopholes in tax laws to shift profits to low-tax jurisdictions. DAC4, aligned with the OECD’s BEPS Action 13 and the Multilateral Competent Authority Agreement on exchange of country-by-country reports (“ CbCR MCAA ”), imposed new transparency requirements on MNEs as a tool intended to allow tax authorities to perform high-level transfer pricing risk assessments, or to evaluate other BEPS-related risks. MNEs are required to report annually on their operations in each tax jurisdiction, including financial data such as revenue, profits, and taxes paid, by each constituent entity.

2.1.5.DAC5: Enhancing Access to Beneficial Ownership Information (2016)

Following several media leaks which exposed widespread tax evasion by way of offshore structures, DAC5 implemented a framework to grant tax authorities access to AML information. By enabling tax authorities to access this data, thereby increasing transparency of beneficial ownership information, the Directive aimed to enhance the effectiveness of tax monitoring and enforcement mechanisms.

2.1.6.DAC6: Combatting Aggressive Tax Planning (2018)

DAC6 was a response to the Panama Papers and built upon BEPS Action 12 on mandatory disclosure rules. The Directive aimed to increase the transparency of cross-border tax arrangements and to prevent the implementation of tax schemes used to shift taxable income to low-tax jurisdictions. DAC6 imposes reporting requirements on intermediaries on aggressive tax planning schemes and is intended to create a deterrent effect by disincentivising intermediaries in creating such schemes. The DAC8 amendment to DAC6 introduced an exemption for lawyers bound by legal professional privilege (LPP) from the obligation to notify other intermediaries of their reporting duties under the mandatory disclosure rules, in order to comply with the Court of Justice of the European Union (CJEU) ruling in Case C-694/20 8 .

2.1.7.DAC7: Sales of services and goods via electronic platforms (2021)

DAC7 emerged because of the rapid digitalisation in the economy and the significant expansion of digital platforms services, and cross-border transactions, which led to increasing opportunities for tax fraud, evasion and avoidance. It is noted that national tax authorities often lack sufficient information to correctly assess and monitor the gross income earned by sellers from commercial activities undertaken on digital platforms. In response to this need, the Council adopted DAC7 in 2021. The Directive primarily reflects the work undertaken at OECD on the Model Reporting Rules for Digital Platforms which provides for a report and exchange of information framework for platforms operators on their users’ income. DAC7 provides for, inter alia, AEOI on data reported by platforms operators regarding persons selling through their intermediation and the consideration obtained therefrom.

2.1.8.Evolution of DAC

The evolution of the administrative cooperation legal framework reflects 3 overarching general objectives which are reflected in each of the iterations of the Directive: (i) safeguarding Member States’ tax base and revenues, (ii) increasing the transparency and fairness of tax systems in the EU, and (iii) contributing to the smooth functioning of the single market. More specifically, the amendments to DAC aim to provide Member States with the tools to combat tax fraud, evasion and avoidance; foster voluntary tax compliance through a deterrent effect; and ensure the uniform implementation of OECD measures to prevent a fragmented approach in the EU.

In terms of effects, the amendments to the DAC have resulted in outputs that serve as indicators of its achievements, as analysed further in Section 4 below.

2.2. Points of comparison

The first evaluation of the DAC covered the period from 2013 to 2017, including all amendments to the Directive up to DAC5 and will serve as the point of comparison for assessing the intervention of DAC1 to DAC5 from 2018 to 2022.

Since DAC6 came into force in 2018 and was excluded from the scope of the previous evaluation, the point of comparison for assessing the intervention of DAC6 will be the situation as assessed in the impact assessment that accompanied the publication of the proposal in 2017.

2.2.1. DAC1 to DAC5 – a summary of the conclusion of the previous evaluation

Although the previous evaluation lacked detailed evidence and data on exchanges due to the short period between implementation and evaluation, the overall assessment of the Directive was positive. The conclusions highlighted several important findings across the 5 better regulation criteria 9 .

In terms of effectiveness, the evaluation found that, to a certain extent, the DAC had been effective in helping tax authorities combat tax fraud, evasion, and avoidance, with some monetary benefits noted. However, there was limited evidence of a deterrent effect. Nevertheless, the evaluation highlighted the importance of high-quality information exchanges and recommended better review processes, improved feedback among Member States, and enhanced data collection to strengthen future assessments.

Regarding efficiency, the Directive’s implementation cost was estimated at approximately EUR 230 million, with benefits estimated at around EUR 620 million, suggesting a positive cost-benefit ratio. However, the results were inconsistent across Member States. In some cases, the benefits of the intervention outweighed the reported costs, nevertheless more concrete data was required. The evaluation recommended that Member States share best practices for estimating the benefits of administrative cooperation and suggested efforts be made to generate better quantitative evidence to assess the true benefits and cost-effectiveness of the Directive in the future.

The DAC was considered a highly relevant and appropriate tool for tax authorities, offering a set of complementary mechanisms to address their needs. However, the evaluation noted that the Directive may need to be adjusted over time to respond to the new challenges posed by an evolving economy, such as digitalisation, the growing importance of intangible assets, and new forms of employment such as the gig economy, which in turn influence taxpayers’ behaviour.

In terms of coherence, the evaluation found no major issues with the Directive or in its interaction with other relevant legal frameworks. Despite frequent amendments, the internal consistency of the Directive had been largely preserved. However, there were some inconsistencies that could hinder the efficient use of the information exchanged, particularly in cases where strict interpretations of terms and national legal restrictions apply. For example, aligning deadlines for information exchanges could improve internal coherence and make the outcomes more efficient.

Regarding EU added value, the evaluation concluded that the Directive had added value by establishing a common EU framework for administrative cooperation, but not all Member States fully utilise the tools to the same extent. Therefore, improving data collection tracking how exchanged information is used are essential to maximising the Directive’s impact.

Overall, the evaluation emphasised the need for continuous improvement in the Directive’s implementation, focusing on better quality and tracking of information exchanges, more effective use of tools, and enhanced data collection practices.

2.2.2. DAC6 Impact Assessment

The DAC6 proposal was accompanied by an impact assessment published in 2017, which should therefore serve as the point of comparison for evaluating the amendment introduced by DAC6. However, it is important to note that the Commission’s proposal was significantly amended during Council negotiations and so the impact assessment which accompanied the proposal does not fully reflect the final legislative act adopted.

Nevertheless, the objectives of DAC6 remain relevant: (i) lack of timely information on aggressive tax planning schemes for Member States’ tax authorities; (ii) differences among national frameworks on disclosure requirements; (iii) ineffective and non-transparent monitoring of compliance with existing law; (iv) existing tax legislation may facilitate aggressive tax planning; (v) insufficient beneficial ownership information.

3.How has the situation evolved over the evaluation period?

3.1. Current state of play

3.1.1.Implementation

DAC 6 represents the primary focus of implementation work during the evaluation period. The Directive had a transposition deadline of 31 December 2019, and first exchanges took place in July 2020. The Commission services assisted the Member States in the legal and the technical implementation of the Directive by holding several dedicated meetings with Member States in 2018 and 2019.

3.1.1.1. Deferral of DAC6 reporting

Considering the challenges presented by COVID-19 pandemic, many MNEs and financial institutions sought a delay to the reporting deadlines set in the Directive. Council Directive (EU) 2020/876 adopted on 24 June 2020 provided Member States with the option to defer the deadlines for AEOI under DAC2 and for filing and exchanging information on reportable cross-border arrangements under DAC6.

As regards DAC6: (i) the deadline for the reporting historical cross-border arrangements (i.e. cross-border arrangements the first step of which was implemented between 25 June 2018 and 30 June 2020) was deferred from 31 August 2020 to 28 February 2021; and (ii) the date for the beginning of the period of 30 days for filing information on reportable cross-border arrangements whose triggering events occur between 1 July 2020 and 31 December 2020, was postponed to 1 January 2021. 24 Member States opted for a 6-month deferral, 1 opted for a 3-month deferral, and only 2 proceeded with the date initially provided for in the Directive.

3.1.1.2. Technical and logistic support

Since 2018, IT tools and formats for the exchange of information have been continuously upgraded. In particular, the Electronic Forms Central Application (eFCA), a centralised web-based application for non-AEOI exchanges, was updated to streamline the exchange process, reduce operational risks, and facilitate the collection of statistics. Furthermore, the alignment of DAC2 message size and format with the CRS standard has been adopted. The Commission has increased its technical and logistical support to Member States by implementing and managing a secure central directory for the exchange of data reported under DAC6.

3.1.2.Infringement procedures

In the period under evaluation, the focus of implementation was on DAC6, where 15 infringement procedures were open due to late or non-notification of transposition measures by Member States. However, each case was closed in advance of referral to the Court 10 . This is consistent with the previous experiences in implementation, where many Member States face challenges in transposing the Directive into the national legal order on time, which may be due to the fast-paced adoption of amendments to the Directive. By contrast, no infringements were open regarding the substantive implementation of amendments to the DAC during this period.

3.1.3.Case law

During the evaluation period and thereafter, the Court of Justice of the European Union ruled on several aspects of the DAC, particularly DAC6. More specifically, between 2022 and 2024, the Court issued 3 judgements regarding the DAC’s compatibility with EU fundamental rights 11 .

While the Court has found the DAC framework to be generally coherent with fundamental rights, it overturned one provision of DAC6 to strengthen protections against reporting obligations for legal professional privilege (LPP) 12 . It introduced a more restrictive interpretation of this concept 13 , by establishing a strict distinction between attorneys-at-law and other professionals and ruled that the LPP waiver applies only to lawyers authorised to ensure legal representation of their clients and not to other professionals who provide advice.

With regard to the principles of non-discrimination, equal treatment, legal certainty, legality in criminal matters and the right to respect for private life, the CJEU ruled that the Directive, and more specifically, the provisions of DAC6, does not infringe any of the fundamental rights as included in the Charter . The Court also upheld that the provisions are sufficiently clear, justified and proportionate to the Directive’s objectives.

More recently, the CJEU concluded that the validity of the DAC vis-à-vis Articles 7 (respect for private and family life) and 52 (scope and interpretation) of the Charter is not affected and, as a result, is not open to question 14 .

3.1.4.Volumes of exchange

The Directive provides for a significant level of information exchange.

The following paragraph presents figures related to the volume of such exchanges. First, it outlines the number of exchanges linked to AEOI in various categories, including income (DAC1), financial accounts (DAC2), rulings (DAC3), CbCR (DAC4) and potentially harmful cross-border arrangements (DAC6). Secondly, it examines the volume of exchanges related to information obtained through other forms of cooperation between tax authorities, such as EOIR, SEOI, administrative inquiries, presence in administrative offices and participation in administrative enquiries, simultaneous controls, and administrative notifications 15 .

3.1.4.1. AEOI

3.1.4.1.1.DAC1

Between 2018 and mid-2023, Member States exchanged information under DAC1 concerning approximately 55.5 million taxpayers and incomes worth over EUR 334 billion. On an annual basis, from 2018 to 2023, Member States exchanged, on average, more than 4,500 messages per year, covering over 9.2 million taxpayers and an average value of approximately EUR 55.7 billion. The volume of exchanges per year has shown a substantial increase since the last evaluation study 16 , both in terms of the value of information exchanged and the number of taxpayers (see Figure 1 below).

Consistent with the previous evaluation, over the period from 2018 to 2023 the exchanges relating to income from employment (IE) and pensions (PEN) were found to be the most important, accounting for 70% of taxpayers concerned by DAC1 exchanges and 98% of the value subject to those exchanges. Exchanges related to immovable property (IP) accounted for 29% of taxpayers, but for less than 3% of value in 2023. Exchanges related to director’s fees (DF) and life insurance products (LIP) accounted only for a very small proportion (1% or less) of both taxpayers and values.

Figure 1: Changes in levels of DAC1 exchanges by Message year (2015 - 2023) 17

DAC1 IE DAC1 PEN DAC1 DF DAC1 IP
Source: Ramboll study, Commission statistics

In 2023, nearly all Member States are able to exchange data on income from employment (IE) (26), pensions (PEN) (25), immovable property (IP) (24) and director’s fees (DF) (22). This has not changed significantly over the evaluation periods, with only 3 countries increasing the number of income categories available (Slovenia for DF, Cyprus for IP and Finland for LIP). One third of Member States already hold information on royalties (ROY) which is the newest category of income to be exchanged as of 2026 (Tax year 2025).

3.1.4.1.2.DAC2

Between 2018 and early 2023, Member States exchanged information concerning some 166 million accounts relating to financial information with a total value of EUR 10,546 billion. This marks a substantial increase in the volume of information exchanged, with the value of DAC2 information exchanged in 2023 being 75% higher than in 2018. This trend was confirmed in 2023: exchanges concerned 43,5 million of accounts, for a value of EUR 2,074 billion (see Figure 2 below).

DAC2 exchanges cover several categories of financial information. In line with findings in the first evaluation study, exchanges relating to gross proceeds from redemptions and sales of financial assets was, in terms of value, the category of financial information exchanged corresponding to the largest amount apart from the value of the accounts, corresponding to EUR 4,061 billion since 2018. Other financial information on which messages were exchanged include interest (amounting to some EUR 223 billion since 2018), dividends (EUR 55 billion), and other unspecified payments related to the accounts (EUR 856 billion).

Figure 2: Changes in levels of DAC2 exchanges by message year, 2017 – 2023

Source: Ramboll study; Commission statistics

3.1.4.1.3.DAC3

Between 2017 and 2023, 27,298 ATRs/APAs were uploaded in the EU dedicated central directory, 8,743 of which were still valid as of June 2023. The number of rulings shared under DAC3 was in the 4,000 – 5,000 range in 2018 and 2019, and then decreased during the 2020 - 2023 period, ranging around 1,600 – 2,000 rulings per year (see Figure 3 below). The number of rulings uploaded during the first year of mandatory reporting (over 11,000) is considerably higher, as it included the backlog of ATRs/APAs issued from 2012 onwards still in force at that time.

Figure 3: Number of ATRs/APAs uploaded per year

Source: Ramboll study, European Commission statistics

A closer analysis of the distribution of the origin of rulings available in the central directory in 2023 reveals the prevalence of a few Member States, namely the Netherlands, Luxembourg, Poland and Belgium (see Figure 4 below).

Figure 4: DAC3 Central Directory - Distribution by originating Member State - 2023

Source: European Commission statistics

3.1.4.1.4.DAC4

DAC4 statistics are collected yearly by Member States; they cover a 12-month period between July of the previous year and June of the current year. The information concerns both the number of reports received by national tax authorities from an Ultimate Parent Entity (UPE) and the exchanges of CbCRs between Member States. When the UPE is resident in a third country or is not obliged to file a CbCR in its jurisdiction of tax residence, the obligation to file the CbCR falls to those MNE’s entities resident for tax purposes in a Member State as a surrogate parent entity. This is to facilitate the collection of CbCRs for MNEs not headquartered in the EU.

As shown in Figures 5 and 6 below, Member States collected an average of approximately 5,200 CbCRs per year, corresponding to a total of just over 30,000 CbCRs between mid-2017 and mid-2023. The number of reports received by tax authorities by their UPE based in their countries declined between 2018 and 2020 and has remained in the 4,000-4,700 range since (see Figure 5 below).

Figure 5: Number of CbCRs collected yearly by Member States

Source: Ramboll study and Commission statistics

Each CbCR collected by tax authorities can be shared with one or more of the other Member States, depending on which countries are concerned by the reporting entity’s activities.

Since the first exchange in mid-2017, more than 113,000 exchanges of CbCRs have taken place among EU Member States (see Figure 6 below). The number of exchanges has shown an increasing trend over the years. During the first 12 months of DAC4 implementation, Member States exchanged just over 12,000 CbCR in 2018, while more than 25,000 reports were shared in 2023.

Figure 6: Total number of CbCRs sent to another Member State

Source: Ramboll study and Commission statistics

3.1.4.1.5.DAC6

From the implementation date of DAC6 in mid-2020 and the end of 2023, around 54,000 arrangements have been uploaded on the central directory by Member States. In 2020, only 3 Member States uploaded information subject to DAC6 exchanges, since most Member States applied the option for deferral introduced due to the COVID-19 pandemic. However, starting from 2021, all Member States uploaded reportable arrangements in the central directory. In subsequent years, the number of arrangements reported has reduced (the backlog of arrangements corresponding to the years 2018-2020 having been essentially cleared between 2021 and 2023) and is noted at approximately 6,600 per year in 2024 (see Figure 7 below).

Figure 7: Number of arrangements reported every year and total available in the Central Directory from 31/12/2020 to 30/06/2024

Source: European Commission statistics (2024)

The study commissioned on the evaluation of the DAC shows that around half of all arrangements reported solely concern non-EU jurisdictions (i.e., arrangements concerning only the reporting Member State and jurisdictions outside the EU, but no other Member States). The UK is the primary third country identified in the arrangements uploaded in the dedicated central directory, with over 9,500 arrangements identified. This includes approximately 1,000 arrangements each concerning the British Virgin Islands and the Cayman Islands. The UK is followed by the US (some 3,400 arrangements, approximately 7% of the total) and Switzerland (close to 2,400 arrangements, 5% of the total). Another 4 jurisdictions (Russia, the UAE, Canada, and Panama) are concerned by 700-1,100 arrangements each (approximately 2% of the arrangements exchanged).

3.1.4.2. Exchange of information on request (EOIR)

EOIR is a tool that is significantly utilised by Member States. EOIR is used to request “foreseeably relevant” 18 information that may concern data already available to requested tax authority or obtained following administrative enquiries. According to statistics 19 , between 2018 and 2021, nearly 46,400 requests for information were sent and followed up by some 45,350 replies. The yearly number of requests sent to other Member States varied between a minimum of just above 7,500 in 2020 and over 10,000 in 2018 and 2022. The 2018-2022 period demonstrates an irregular trend. Over the period under review, the number of requests initiated peaked in 2018, then decreased to a minimum in 2020, most likely due to a reduction of tax administration activity during COVID-19.

EOIRs generally result in full rates of reply. On average, during 2018-2022, the replies sent correspond to 98% of the requests. The option to refuse a request for information based on Article 17 of the DAC (on the ground of legal reasons including the disclosure of commercial, industrial or professional secret, or of information whose disclosure would be contrary to public policy) was rarely used from 2018 to 2020, with an average of 3 refusals per year. The number of refusals rose in 2021 and 2022 to 42 and 51 respectively. The reason given by the requesting country for refusing a request is generally that the request does not relate to the right tax head (wrong legal grounds) or does not fulfil the conditions to be foreseeably relevant (exhaustion of domestic resources not demonstrated by the requesting country).

3.1.4.3. Spontaneous Exchanges of Information (SEOI)

As regards SEOI under Article 9 of DAC, a total of over 160,000 SEOI were sent by Member States between 2018 and 2022. The yearly trend in SEOI remained highly variable, which is consistent with the previous evaluation, ranging from 6,300 SEOI in 2020, to over 70,000 in 2018.

Information received through SEOI mostly concerns employment income, business transactions, and other categories of income, as well as payments or other tax and banking information. SEOI are most often sent when a country receives information from another Member State that could in turn be relevant for another country, or because the sending Member State has reasons to suspect that there may be a loss of tax in the receiving Member States. Less often, SEOI are sent on the grounds of a tax break in one Member State possibly leading to increased tax liability in the receiving country, or on the detection of a cross-border business resulting in tax saving.

Member States use SEOI also to exchange information on cross-border ATR/APA involving third-country jurisdictions, when the agreement does not permit the disclosure to third parties under DAC3 AEOI, provided that the international tax agreement under which it was negotiated permits its disclosure, and the competent authority of the third country gives permission for the information to be disclosed.

Overall, this type of information constitutes a negligible part of the exchanges under SEOI and are a fraction (8% on average) of the cross-border ATRs/APAs concerning only EU Member States, shared under DAC3 AEOI.

3.1.4.4. Other forms of administrative cooperation

3.1.4.4.1.Presence in administrative office and participation in administrative enquires (PAOE)

The Directive allows Member States to request either to be present in the offices where the hosting tax authorities carry out their duties (presence in administrative offices), or to be present during controls and checks carried out in the territory of the requested Member State (presence in administrative enquiries), or both. The evidence collected suggests that, in most cases, Member States tend to combine the 2 aspects of a PAOE during the same visit.

Over the 2018-2024 period a total of 222 PAOEs occurred. The number of PAOEs carried out yearly decreased since 2018. More specifically, the figures are lower starting from 2020, also reflecting the effect of COVID-19, which hindered the possibility to travel and to be present in offices, and digitalisation (in some cases, PAOEs took place virtually because of the impossibility of in-person attendance). In the past 3 years, 10 Member States initiated a PAOE, leading to the participation of 20 Member States.

3.1.4.4.2.Simultaneous controls

Over 2018-2024, more than 370 simultaneous controls were initiated (see Figure 8 below). During the period covered by the evaluation, the number of simultaneous controls carried out across the EU continuously increased since 2018, except for years 2020-2022, reflecting an overall slowdown of audit activities and international cooperation due to COVID-19. As the use of this instrument requires specific resources (trained auditors, experts in cross-border tax issues, fluent in English) and procedures, Member States do not have the capacity to initiate large numbers of such audits. It is often concentrated in a few Member States: in the past 3 years, 15 Member States initiated at least one simultaneous control, 7 out of them initiating more than 5. In total, more than 400 simultaneous controls took place, with very significant impacts in terms of tax base adjustments, representing several billion euros 20 , 60% of them addressing direct taxation topics, mainly transfer pricing issues.

Figure 8: Number of simultaneous controls started and finished from 2018 to 2024

Source: MLC platform, EU AIAC and European Community statistics

3.1.4.4.3.Requests for notification

Article 13 of DAC provides Member States with the possibility to ask for assistance to the notification of tax-related instruments and decisions applying to taxpayers located in another Member State. Between 2018 and 2022 Member States sent between 230 and 340 requests per year.

3.1.5.External recommendations and Commission’s responding initiatives

In recent years, DAC has been assessed by the European Court of Auditors (ECA) and the European Parliament (EP). In particular,

·In 2021, ECA published a special report on DAC 21 following its audit of the EU’s tax information exchange system, covering the period from 2014 to 2019. The report concluded that “the system for exchange of tax information has been well established, but more needs to be done in terms of monitoring, ensuring data quality and using the information received”. It highlighted that the information collected by Member States is often incomplete and inaccurate, partly due to the lack of audits or checks on reported data and reporting entities. Additionally, the report noted the absence of sufficient guidelines on data quality and usage, as well as a common framework for monitoring the system’s performance and achievements.

·Also in 2021, the EP issued a report on DAC 22 , followed by a legislative resolution 23 . While acknowledging limitations in the available evidence, the EP identified several shortcomings in the current framework. In particular, it noted that certain income and assets remain outside the scope of DAC and that the information provided by private operators and exchanged by Member States is often of poor quality. The EP welcomed recent amendments that expanded DAC’s scope and strengthened its mechanisms, while emphasising the need for thorough implementation of existing rules, including those related to AML.

·In 2024, the ECA assessed the measures taken by the EU to tackle harmful tax regimes and corporate tax avoidance 24 and made several recommendations in relation to DAC6. The ECA suggested that the Commission (i) clarifies the EU legislative framework; (ii) develops its guidance in cooperation with Member States; (iii) assesses the need for possible amendments to DAC6 based on the results of the evaluation; (iv) improves the quality of DAC6 reports; (v) makes the data field for non-EU countries involved in cross-border arrangements mandatory in the reporting schema; and (vi) updates the DAC6 central directory architecture to make this information available to Member States for every arrangement concerned. The ECA stressed also that the Member States should perform risk analysis using the tax information received while the Commission should encourage and support Member States to adopt a common performance monitoring framework, including performance indicators and quantitative targets, to measure the achievement of specific objectives in the fight against harmful tax regimes and corporate tax avoidance. The Commission endorsed the recommendations on DAC6, with limited reservations. 25 .

3.1.5.1. Commission’s initiatives to address the ECA and EP reports

Legislative proposals and outcome

In response to the recommendations of ECA and the EP, the Commission adopted the DAC8 proposal, which increased the number of mandatory categories of income to be exchanged under DAC1 and expanded the scope of automatic exchange of advance cross-border rulings to high net-worth individuals. It also made the inclusion of TIN mandatory for certain AEOI exchanges (namely DAC1 and DAC3) and empowered the Commission to create a TIN verification tool. Furthermore, it allowed Member States to use the information received for other purposes than tax matters, i.e. customs duties, AML, and enforcement of international sanctions. It also included 2 provisions of principle supporting 2 recommendations targeted to the Member States: ensuring the effective use of data received from AEOI and monitoring the effectiveness (i.e. benefits) of DAC. The Directive was adopted by Council in October 2023.

Non-legislative actions

The Commission services also carried out supportive actions to assist Member States in meeting their respective recommendations.

·Fiscalis expert team on quality and use of data - VISDAC

The ambition of VISDAC was to identify best practices that can be shared across the EU and identify specific areas for improvement related to DAC1 to DAC4 for each Member State. The expert team undertook visits to all Member States and engaged in in-depth dialogue with their respective tax authorities regarding their practices and procedures to ensure the highest quality of data to be exchanged and to ensure that the data received is effectively used. Based on the outcome of the visits, the expert team issued specific recommendations to the individual Member States with a view to improving internal processes regarding data quality and use. Furthermore, based on the information gathered from all Member States, the expert team produced a general report on best practices on the quality and use of data that is exchanged under DAC1 to DAC4.

The Commission services highlighted areas that required improvements and produced a set of general recommendations for ensuring the highest quality of data and the use of the data exchanged under DAC1 to DAC4.

·Technical support instrument (TSI)

DG REFORM presented a flagship Technical Support Instrument (TSI) in 2023. The scope of the project covers the enhancement of DAC data quality, data usage (including risk analysis), and performance monitoring. The support consists of a review by external consultants of national AEOI systems, including IT systems and data flows, identification and sharing of best practices, and delivery of recommendations for improvement.

·Fiscalis Expert team on Data analysis

The expert team 26 developed a prototype IT tool to assist with the matching of data. The results of the group were presented to the Member States in June 2024.

·Fiscalis project groups on DAC performance measurement and on measuring the impact of the DAC

The work of 2 groups led to the development of a set of metrics to better estimate the outcome of administrative cooperation, which were approved by the heads of Member States’ tax administrations in December 2024. Further guidance and instructions in this respect was also provided to Member States in the various reporting tools, starting from the Yearly Assessment questionnaire.

4.Evaluation findings (analytical part)

The first evaluation of the Directive was based on a study published in 2019. While the findings from this first evaluation were largely positive, they identified the need for improvements in certain areas. These areas related to the practices concerning the collection of information, the quality of information exchanged and the need to ensure the more effective use of administrative cooperation tools.

The first evaluation of the Directive was also limited by the recent entry into force of the provisions concerning administrative cooperation and automatic exchange of information. Consequentially the findings were informed by the limited evidence available which included data from the initial exchanges.

5 years later these limitations no longer apply, and the situation has progressed considerably with new dedicated instruments to foster administrative cooperation and advanced functional information systems to support the exchange of information. These advances have improved the overall functioning of the DAC.

4.1. To what extent was the intervention successful and why?

4.1.1. Effectiveness

AEOI, which is the core instrument of administrative cooperation under the DAC, enables tax administrations to obtain information about income and capital earned in other Member States by taxpayers who are resident for tax purposes in their territory and, therefore, subject to taxation therein. The increased tax transparency which results from AEOI, increases the ability of Member States’ tax administrations to tackle cross-border tax fraud, evasion and avoidance, thereby contributing to safeguarding Member States’ tax bases and tax revenues and, overall, to the fairness of EU tax systems.

The success of AEOI depends on the capacity of Member States’ tax administration to use the data received from other Member States to assess the taxes to be paid by the taxpayers concerned and, if necessary, make tax adjustments which result in an increased tax base or additional taxes assessed. This, in turns, depends on: (1) the usability of the information received (i.e. the capacity of the receiving Member States’ tax administrations to process the data in order to achieve the above-mentioned objectives) and (2) the ability to match it (i.e. the ability to link the data to the correct taxpayers).

4.1.1.1. Usability of information

The assessment of the usability of the information received by Member States requires an analysis of two interconnected factors: (i) timeliness of the information exchanged (timely transmission/reception of the information ensures that information is available when needed for compliance and enforcement actions); (ii) completeness and quality of the information exchanged (complete and good quality information enable accurate interpretation of the data received by tax administrations).

Timeliness of information exchanged

No major issues were identified regarding the timeliness of exchanges. However, in the context of DAC1, some Member States expressed their need to receive the information earlier in the year to facilitate the pre-population of tax returns.

Completeness and quality of information exchanged

As regards completeness, DAC1 and DAC2 exchanges present a high degree of completeness, as the information exchanged includes granular data concerning the income or capital received and the identification information of the concerned person or entity. The only exception is the percentage rate of inclusion of the TIN (Tax Identification Number) of the receiving country in DAC1 exchanges. This remained low compared to the number of records exchanged (covering about 2% of the taxpayers concerned). Simultaneously, only 13 Member States carry out activities to verify that the TIN of the receiving country is correct 27 .

For DAC3 and DAC4, several tax authorities reported concerns regarding the completeness of the information exchanged. These concerns relate to the fact that the identification of the individual or entity concerned is inhibited by the non-reporting of the TIN.

In addition, some Member States stated that the summary of the content of the advance cross-border ruling, or advance pricing arrangement is not sufficiently detailed to readily assess and use the information received. This specific issue has been the subject of work initiated by the Commission services in 2023 with a group established to improve the quality and content of the DAC3 summaries.

Concerning DAC4, it was found that the completeness of the information may potentially be affected by Member States that do not carry out adequate compliance checks of local MNEs 28 . Simultaneously, the usability of the CbCR information is impacted by data quality issues related to missing, incorrect and inconsistent information, 29 which emphasises the importance of regular data quality checks.

Concerning DAC6, 22 out of 27 Member States have processes and procedures in place to assess the completeness of disclosures submitted by the intermediaries or the taxpayers concerned. While the scope of these processes and procedures varies across Member States, more than 50% of Member States carry out validation checks on the reported TIN’s before exchanging information to the central directory. However, other mandatory informational fields are not subject to such systematic assessment, which may affect the quality and, therefore, the usability of the information exchanged. In this respect the abstract summary of the cross-border reportable arrangement is affected by the absence of detailed facts and a mismatch between the reported hallmarks and the content of the summary 30 .

Figure 9: Member States carrying out activities or automatised controls to verify that the following collected information is correct before such information is sent to DAC6 central directory

Source: Yearly Assessment 2024 (data 2023)

Furthermore, the completeness of the information reported in respect of cross-border arrangements with non-EU countries was raised as an issue by both the ECA 31 and private stakeholders 32 .

4.1.1.2. Use of information

Data matching

A critical element of AEOI is the ability of the tax administration that receive the information to match this information to the taxpayer(s) concerned in the national database and, therefore, to identify the taxpayer. This process, which is referred to as data matching is necessary to ensure that the data can be used for tax compliance purposes (Figure 10 below). Income or capital that is likely to be subject to taxation in the Member State is systematically used by almost all Member States for DAC1 and DAC2. Some of these Member States apply specific processes or treatments before matching. Only 2 Member States do not have a systematic approach toward matching of DAC1 or 2 data.

In the context of DAC4, 20 Member States systematically match information received, while for DAC3 and DAC6, the approach is less systematic, however most of the Member States (15 to 16) process the data immediately, and match the taxpayers concerned.

Figure 10: Processing the data received

Source: Ramboll study and EC data

This systematic matching has led to an overall improvement in taxpayer identification compared to the previous evaluation. The 2019 evaluation concluded that only 7 Member States were able to successfully match 80% or more of the DAC1 and DAC2 information received, with another 2 successfully matching 75 % of the information received.

This evaluation concludes that in 2023, 17 Member States were able to successfully match 80% or more of DAC1 information with the rate of unmatched information across Member States averaging between 15% to 20%. More specifically the matching rates for the DAC1 categories of IP, DF and PEN is estimated at 85% and at 80% for the other income categories (LIP, IP)

As regards DAC2, 18 Member States had a matching rate above 80%. Simultaneously more than half of the Member States have an unmatched rate of 10% or less (see Figure 11 and 12).

Figure 11: DAC1 (IE) unmatching rate – 2023

Source: European Commission –AEOI yearly assessment 2023

Figure 12: DAC2 Unmatching rate – 2023

The overall unmatched rate has decreased over the evaluation period for both DAC 1 and DAC2. This trend is particularly relevant for DAC2, with the unmatched declining from 15% to 8% over the period covered by the evaluation (see Figure 13 below). However, wide variations exist between Member States.

Figure 13: DAC2 Unmatched data rate after automatic matching, fuzzy matching and manual matching

Source: European Commission – Annual report on AEOI yearly assessment 2023

As regards DAC3 and DAC4, only 5% of reports remain unmatched. The matching rate for DAC3 is very high (95%), as many Member States (20) resort to manual matching on a case-by-case basis for the remaining data that could not be matched automatically. It is efficient but relatively slow and time consuming. As regards DAC4, two-thirds automatically match the reports to their databases 33 .

The data available for DAC6 matching rates was limited to 19 Member States for the period until 31 December 2022. However, out of these 19 Member States, 16 had a matching rate of 90% or more, while further 3 had matching above 50% 34 . This trend was confirmed in 2023 by 17 Member States.

Despite the deployment of extra IT resources, Member States still cite a lack of domestic resources to systematically process the information received, as one of the main reasons for the lack of a systematic approach to data matching. The second reason identified is the quality of the information received.

Use of information after data matching

Information may be used for the administration and enforcement of the domestic laws of the Member States concerning the taxes referred to in Article 2 of DAC.

All Member States use AEOI information to carry out assessments, collections and enforcements of direct taxes. Simultaneously, AEOI information also supports mutual assistance in the recovery of tax claims and duties and is used in connection with judicial proceedings such as tax evasion.

While this evaluation finds that not all Member States systematically use information received from all DAC’s, the systematic use of information has steadily improved as compared with the findings from the previous evaluation. In reply to the Commission’s questionnaire, 10 out of the 25 Member States that responded stated that they require more personnel resources, while 6 stated that they require better IT tools to make improvements in the systematic use of the information 35 .

The most common use of DAC data is for risk assessment and compliance interventions e.g. desk controls and tax audits, but DAC data could also be useful to promote voluntary tax compliance.

Use of information received under DAC1

DAC1 data (see Figure 14) is often compared directly with the taxpayer’s tax returns. This correlation supports actions to address non-compliance e.g. tax assessment, audits, and support voluntary compliance e.g. pre-population tax returns, awareness campaigns and reminder letters to taxpayers.

Figure 14: Use of DAC1 data in the reporting period – type of actions carried out (multiple answers possible)

Source: Ramboll study and EC data

Use of information received under DAC2

The use of DAC2 information (Figure 15) is relatively similar to DAC1, as it mainly targets individuals. General audits are carried out more often as the potential tax yields may be higher than DAC1, owing to the higher value of income and assets reported. Half of the Member States use compliance measures e.g. awareness campaigns and, reminder letters to support those taxpayers who have not yet met their obligations to declare income from financial accounts held abroad, as opposed to imposing sanctions immediately.

Figure 15: Use of DAC2 data in the reporting period: actions carried out (multiple answers possible)

Source: Ramboll study and EC data

More than half the Member States also use DAC1 and DAC2 data for tax recovery purposes 36 .

Use of information received under DAC3, DAC4 and DAC6

DAC3, DAC4 and DAC6 data (Figure 16) is mainly used for risk assessment (21 Member States), and follow-up compliance actions i.e. tax audits, limited to specific issues (such as transfer pricing) or general audits. DAC4 information is also used to support further enquiries into the DAC4 information is also used to support further enquiries into the MNE groups transfer pricing arrangements or into other tax matters during a tax audit (15 Member States). DAC4 information is also used for economic and statistical analysis (13 Member States).

Despite its relatively recent entry into force, DAC6 information is used by the majority of Member States. In January 2023, 21 Member States had already integrated DAC6 data into their risk assessment process, and 16 Member States were using the data received for specific tax inspections or extended audits.

Figure 16: Use of DAC3, DAC4, DAC6 data in the reporting period – type of actions carried out (multiple answers possible)

Source: Ramboll study and EC data - Yearly assessment questionnaire on the AEOI (2023)

Use of information on request and spontaneous exchange of information

Following the initial automatic exchange of information 12 Member States report proactively engaging with other Member States on the country-by-country reports received or on information made available in the DAC3 and DAC6 central directories. These contacts involved specific requests for additional information and seeking clarifications on issues of interpretation and missing data points. In this context the provisions for exchange of information on request are systematically used by the Member States.

The type of use varies according to the nature of the information requested with all Member States using this information to carry out the assessment, management or enforcement of direct taxes. Furthermore, it is apparent that this information is not only useful for direct taxation, but also for other forms of taxation, revenue collection, and judicial proceedings e.g. tax evasion.

Figure 17: Use of information on request

Source: Ramboll study and EC data

A majority of Member States also confirmed the usefulness of the provisions for spontaneous exchange of information. In this context, Member States noted that they had received very relevant information on specific cases including income from construction work; income from crypto-assets users, income from foreign virtual currency providers; and income earned from social media activities 37 .

4.1.1.3. The deterrent effect of the DAC

Several factors explain the deterrent effect of the DAC: increased transparency, risk of detection and risk to be subject to penalties.

AEOI under DAC1 and DAC2 enable tax authorities to have access to data on income and assets held abroad, including foreign bank accounts and other financial assets, which were previously hard to detect. In most Member States, taxpayers are informed by their tax administration that foreign income or assets are likely to be reported to the Member State where they are tax resident, thereby reducing the incentive to hide them. Knowing that tax authorities can cross-check foreign income or assets with domestic tax declarations encourages accurate reporting and increases voluntary tax compliance.

The reporting of certain arrangements that may be tax harmful (under DAC6) or the distribution of the tax paid in other countries (under DAC4) enhance the ability of the receiving tax administration to detect possible loopholes or aggressive tax planning situations. This is particularly true as far as DAC6 is concerned, as this Directive has an early warning effect, as tax authorities receive timely information about potentially harmful schemes, reducing secrecy on arrangements that may have been structured for tax avoidance purposes. Hence, the risk of detection by tax authorities and the possibility to be subject to tax audits, discourage the use of aggressive tax planning schemes.

Moreover, the implementation of national provisions related to the transposition of DACs is accompanied by financial penalties that target both taxpayers and reporting entities. For example, EU Member States have implemented DAC2 into national law, and failure to comply with the national provisions can result in fines for individuals who fail to provide accurate self-certification of tax residency, and fines for financial institutions that fail to report or misreport account information. The same goes true for intermediaries and taxpayers who fail to comply with the obligation to disclose reportable cross-border arrangements under DAC6. These penalties can be substantial 38 , especially if the non-compliance is systemic or intentional, which has a strong deterrent effect.

Member States, stakeholders, as well as academics, lawyers and economists agree on the deterrent effect of the DAC for DAC1, DAC2 and to a certain extent DAC6. Although difficult to measure objectively, it appears that the impact has been very important for DAC2, and to a lesser extent DAC1. The rationale for this difference is that tax liabilities arising from DAC2 information are generally greater than those arising from DAC1.

With regard to DAC1, the findings from the VISDAC Expert Team confirmed that when Member States informed taxpayers of the presence of AEOI information for example by pre-populating tax returns or by way of reminder campaigns (letters or emails), taxpayers tended to declare their foreign income more spontaneously in subsequent years 39 . Moreover, the targeted consultation 40 with lawyers and the evidence gathered in the public consultation 41 confirmed the deterrent effect of the AEOI as regards DAC2 and, to a lesser extent, DAC6, and its consequences on taxpayers and intermediaries. In particular, the experts emphasised the effectiveness of DAC2 in disclosing financial assets abroad. These effects are considered persistent and well assimilated by taxpayers.

With regard to DAC3, the main effect observed was a change in legislation on rulings in certain Member States after the adoption in 2015 (a factor already mentioned in the previous evaluation). DAC4 was not intended to act as a deterrent, as the CbCR provides additional information to help tax authorities better understand transfer pricing structures without necessarily aiming to prevent the structures or related transactions. However, there are some scholars who stated that public CbCR has proven to have such a deterrent effect 42 .

Experts also indicated that they had perceived a change in the attitude of taxpayers regarding DAC6 arrangements. More specifically, it was confirmed that it had an impact on large companies, leading some of them to set up very elaborate systems of internal due diligence given the broad scope of the Directive. To ensure consistent reporting, some intermediaries have also developed elaborate internal guidelines, to take account of the differences in application between Member States.

4.1.1.4. Benefit analysis

The increased tax transparency brough about by the exchange of information under DAC and the consequent increase of tax authorities’ ability to detect and fight cross-border tax fraud, evasion and avoidance contributes to safeguarding Member States’ tax revenues.

In this respect, the actual benefits of the DAC can be measured in terms of:

-additional tax base i.e. increased source of taxable income, assets, or transactions that are brought under the scope of taxation by the tax authority receiving the information under the DAC.

-additional tax assessed i.e. an increase in the amount of tax liability determined by a tax authority, typically due to a re-assessment of the tax base, as a consequence of the use of the data received under the DAC. For personal or corporate taxes, reassessments may occur after tax audits or corrections.

The estimate of benefits generated by DAC in terms of safeguarding Member States’ tax revenues are consistent over time. However, the quantitative estimation of such effects is subject to uncertainties.

The method for calculating the benefits of information exchanges could be explained as follows: for records related to DAC1 and DAC2, which contain potentially taxable amounts (such as amounts of income or assets), it would be relevant to know the amount of taxes paid by taxpayers or the tax bases modified following the receipt of the information.

With respect to the use of DAC3 and DAC6 information it is not possible to establish a direct causal relationship with additional taxes collected and the modification of the tax base. The rationale for this is that DAC3 and DAC6 information must first be analysed, and thereafter it may be combined with other relevant information for the purposes of further investigations or audits by the tax authorities. As for data exchanged under DAC4, the use of CbC reports is limited by the Directive 43 , primarily to risk-assessments for transfer pricing purposes.

However, even if this analysis was systematic (which is not the case in all the Member States), an immediate fiscal benefit could not always be identified. Currently, Member States submit an annual report to the Commission’ services that estimates the benefits arising from the use of DAC information. However, not all Member States have the capacity to trace and monitor the use of DAC data. Therefore, caution should be exercised in interpreting the following elements.

For DAC1, DAC2 and non-AEOI, increasingly more Member States track and monitor the tax benefits derived from these instruments. Yet, in the absence of common, structured, fully integrated indicators, there is no independent and comparative means of verifying all the data submitted by Member States. Moreover, for DAC4, the main limitation concerns the fact that the estimates carried out are only based on public CbCR data, as no dedicated literature exists on private CbCR (DAC4).

Overall benefits

Taking these limitations into account, and based on 2021 – 2023 data, various scenarios have been established. The first is based on the strict use of the data available for the years concerned, with total DAC benefits estimated at EUR 1,2 billion per year. The other scenarios are based on models 44 that allow an estimation of higher benefits: applying extrapolations and taking into account possible variations in the available data and the random benefits of using certain instruments 45 , DAC benefits are estimated to fall within the range of EUR 5 – 10 billion, with a central estimate of EUR 6.8 billion per year 46 .

While significant in absolute terms, this represents only about 0.3% of EU direct taxation revenue, or 0.1% of total EU tax revenue in 2021 47 .

Specifically, considering the central estimates, AEOI under DAC1 and DAC2 is estimated to generate around EUR 1.2 billion per year in EU tax benefits, while DAC4 is estimated to contribute to an additional EUR 5.6 billion per year in tax revenues.

However, the situation in different Member States remains disparate with regard to how the data is used. The approach seems generally dictated by the capacity to integrate data with pre-existing systems for risk analysis.

For DAC3 and DAC6, the Commission services could only examine the impacts on taxpayers, as it is not possible to establish a causal link between the information received and tax revenues.

It is also difficult to directly attribute a specific tax increase/additional tax assessed to the use of exchange of information and/or spontaneous exchange of information. This difficulty arises from the fact that this would require a disaggregating of the information from domestic sources of information for the purposes of calculating a specific tax increase/additional tax assessed.

The importance of the impact of the use of other advanced cooperation instruments should also be stressed. These tools are coordinated and monitored by the Member States with the support of the Commission Fiscalis programme in particular by a community of experts (EU-AIAC), allowing reliable statistical evidence to be obtained.

In this context it has been establish that the yield from simultaneous controls in the years 2018 to 2023 was significant. The use of this instrument resulted in tax base adjustments amounting to more than EUR 5,5 billion and shifted profit of nearly EUR 100 million over that period 48 .

Figure 18: Tax adjustments and shifted profit resulted from simultaneous audits from 2018 to 2023

Source: EU AIAC platform and EC statistics

Benefits of automatic exchange of information per categories of DAC

·DAC1 and DAC2

As the information exchanged under DAC1 and DAC2 are designed to link certain income or assets with a specific taxpayer., it is possible to establish a direct link between the use of information and the benefits in terms of increased tax revenues. Findings from the VISDAC expert team and answers provided by Member States show that more and more tax authorities are developing monitoring systems to this purpose, leading to better monitoring of this outcome.

Some Member States shared voluntarily with the Commission information on the tax benefits generated by DAC1 and DAC2 over the period 2018-2023. Benefits are estimated as additional tax assessed or additional tax base. For DAC1, estimates of tax benefits were provided by 8 Member States in 2021, 11 Member States in 2022 and 12 in 2023. For DAC2, about half of the Member States provided estimates for 2021, 18 Member States for 2022 and 17 for 2023.

Figure 19 shows the reported tax benefits in terms of additional tax assessed, for both DAC1 and DAC2. For DAC1, reported tax benefits reached about EUR 150 million in 2021, due to the higher data coverage. In 2022, they are estimated at about EUR 50 million; data for previous years are lower, around EUR 10 million. Additional taxes assessed for DAC2 are higher compared to DAC1. This is in line with the qualitative insights from the questionnaire and the targeted consultation. While in 2018 and 2019 the additional taxes assessed were worth about EUR 40 million, this figure increased to more than EUR 200 million in 2020, up to EUR 340 million in 2021 and more than EUR 680 million in 2022.

Figure 19: Reported tax benefits from DAC1 and DAC2, additional tax assessed (EUR million)

Source: Ramboll study and DG TAXUD data

Tax benefits in terms of additional tax base are shown in Figure 20 below. A lower number of Member States, accounting for a significantly smaller part of the sample, could report this data. Hence, estimation of potential benefits at EU level is difficult to be provided.

Figure 20: Reported tax benefits from DAC1 and DAC2, additional tax base (EUR million)

Source: Ramboll study and DG TAXUD data

Data submitted by Member States for 2023 confirms the benefits of automatic exchanges of data on income and assets provided for in DAC1 and DAC2, with Member States reporting additional revenues of 522 million EUR in tax base assessed and 273.5 million EUR in tax assessed 49 .

Based on the data reported by Member States, estimated EU tax benefits have been extrapolated: the simulation is what total benefits would be with a consistent use of the data across the Member States. Estimated EU tax benefits are robust for 2021 and 2022, when the data cover a considerable share of the EU. For DAC1, EU benefits would amount to about EUR 200 - 400 million. For DAC2, estimated EU tax benefits have grown significantly in the last two to three years. These are estimated at about EUR 700 million in 2021 and about EUR 1,100 million in 2022. For the previous years, potential benefits are significantly lower, in the EUR 40 – 130 million range for 2018 and 2019. The increase is likely to reflect a better recording of benefit data, as well an increase in the use of DAC2 for enforcement actions.

·DAC3 and DAC4

As explained before, there is no specific data from tax authorities on additional tax base or additional tax assessed from DAC3 and DAC4 as the information which is shared under DAC3 and DAC4 does not provide amounts with are directly subject to taxation.

The Intra-European Organisation of Tax Administrations (IOTA) made an assessment the possible benefits from AEOI of cross-border tax rulings and advance pricing agreements (subject to DAC3 obligations), that could be measured by the high-risk transactions and schemes identified, and by the amount of additional taxes identified or collected on auditing such transactions. The outcome of this assessment was relatively limited, as only one of the countries participating to the assessment could provide estimates on the additional taxes assessed.

Tax benefits of Country-by-Country reporting have been documented in the literature which has focused on both public CbCR 50 and CbCR between tax administrations (which is covered by DAC4). 51

• On the period subject to the evaluation, the available economic analysis mostly focuses on banks, and for which a longer data series on the impacts of public CbCR exists:

• Overesch and Wolff 52 show that the effective tax rates increased for EU banks subject to public CbCR compared to domestic EU banks outside the scope of the obligation and their US peers. Taxes paid by banks subject to CbCR increased from 28.3% to 31.2% of their profits (+2.9 p.p.). The tax increasing was even more for banks with higher-than-average activities in tax havens (+3.7 p.p.).

• Joshi et al. 53 demonstrate that CbCR reduces profit shifting, i.e. the amount of profits allocated to foreign low-tax jurisdictions. At the same time, the effective tax rate borne by banks covered by this obligation did not increase, possibly because other tax evasion strategies escape CbCR scrutiny. Hence, albeit total tax burden remains similar, tax revenues shift towards high-tax jurisdictions. The numerical estimates suggest that a standard (inter-quartile) improvement of disclosure decreases profits in low-cost jurisdictions by 12.8%.

To estimate the impacts of DAC4, which concern all EU MNEs, the analysis expands, with some caveats, on the results from Overesch and Wolff. Here, the major concern is that – unlike large banks and few other companies – most MNEs are subject to private, not public, CbCR. Hence, impacts are likely smaller. Three scenarios are considered: in the central scenario, the effects of private CbCR are one third of those estimated for public CbCR; sensitivity values are set at half and one quarter. This means that the effective tax rate is expected to increase by 1.0% (range: 0.7 – 1.5%). Considering the yearly profits generated by EU MNEs in domestic jurisdictions (EUR 585 billion), and under the assumption the increase in the effective tax rates is homogeneous across jurisdictions, the effect of CbCR can be estimated in a range of EUR 4.2 – 8.5 billion (central estimate at 5.6 billion). In relative terms, the changes are a fraction of existing revenues, representing 1.2% of current revenues from corporate income taxes, and 0.1% of total tax revenue in the EU.

The absolute figures are important, as it is related to the large number of profits generated by EU MNEs domestically. As a consequence, even a limited change in the effective tax rate results in a large impact on revenues.

4.1.2. Efficiency

The analysis of efficiency examines the costs associated with the DAC, as reported by tax authorities, private stakeholders, and the academic literature.

4.1.2.1. Cost analysis

The cost analysis provides an overview of the costs associated with implementing and maintaining the functioning of the DAC both for the private sector and tax authorities. The costs considered in this analysis include annual recurring adjustment costs related to IT systems and administrative burdens due to periodic reporting, which affect both public and private stakeholders such as reporting financial institutions, intermediaries and companies planning cross-border tax arrangements, and MNEs subject to CbCR obligations. Set-up costs are excluded 54 .

This analysis is complicated by the lack of objective traceability on the part of private stakeholders 55 surveyed, both in the study and in the public consultation, and the relatively imprecise information obtained. As a result, the report is essentially based on the findings of the study commissioned by DG TAXUD 56 , which cannot be corroborated by other work or data made available to the Commission services. Nevertheless, tax authorities are best placed to identify spending on investments to establish new exchanges (e.g. DAC6), and those relating to the operational costs of current exchanges.

4.1.2.2. Implementation costs

Based on the external study findings for the period from 2018 to 2022, annual recurring adjustment costs and administrative cost from DAC are estimated at approximately EUR 646 million for all concerned stakeholders (i.e. private stakeholders and tax authorities).

Figure 21: Implementation costs

Costs	Total
AEOI adjustments costs for tax authorities	37
Non-AEOI adjustment costs for tax authorities	4,9
·EOIR	4.3
·PAOEs and SCs	0.6
Reporting administrative burdens for tax authorities	0.1
Total costs for tax authorities	42
Administrative costs (DAC2) for private stakeholders	550
Administrative costs (DAC4) for private stakeholders	54
Total administrative costs for private stakeholders	604
Total costs	646

Source: Ramboll study and DG TAXUD data

A large part of the administrative costs (EUR 604 million per year) accrues to private stakeholders who are subject to reporting obligations, i.e. financial institutions subject to DAC2, MNEs subject to DAC4, intermediaries and taxpayers reporting under DAC6. 91% of the recurring expense is linked to the collection of DAC2 data (EUR 550 million), which in 2023 would correspond to around EUR 12 per account declared under DAC2 57 .

The costs borne by tax authorities amount to about EUR 42 million per year, the greater part of which is linked to AEOI 58 This amount covers two types of costs: (1) adjustment costs linked to the operation of IT systems for AEOI and non-AEOI activities, as well as (2) administrative burden stemming from the periodical reporting obligations. It should be noted that the period from 2018 to 2023 was marked by significant one-off costs for setting-up the IT systems to enable the collection and use of DAC4 and DAC6 data. Moreover, recurring costs for DAC1 and DAC2 increased due to the increase in the volume of data exchanged.

More specifically, DAC1 and DAC2 generated the highest costs for tax authorities 59 . Tax authorities acknowledge that the set-up costs are significant when new DACs are introduced but then decrease significantly and stabilise over time, while recurring costs tend to increase 60 . This confirms the finding of the previous evaluation.

This increase in recurring costs over time is in line with the increasing use of information for risk analysis, tax determinations, or other compliance purposes. Recurring costs tend to stabilise in a narrow range (e.g. EUR 10-12 million for DAC2, EUR 7-10 million for DAC1); yearly variations may be due to national factors, such as changes in the number of resources dedicated to DAC, special campaigns or enforcement actions.

The costs associated with non-AEOI provisions are relatively low for the Member States, amounting to around EUR 1 million per year, insofar as the resources necessary to exchange information do not require significant IT infrastructure. Exchanges on request and spontaneous exchanges are carried out using a dedicated “CCN-mail” messaging system, the costs of which are partially covered by the FISCALIS budget and are included in the DAC1 costs in Figure 19 below.

It should be stressed indeed that the Commission provides the IT infrastructure necessary to ensure the exchange of information via the Common Communication Network (CCN). The CCN is a secure, private network information channel between national customs and tax authorities within the EU. For example, exchange of information on request and spontaneous exchange of information are exchanged via CCN/Mail, while DAC1, DAC2 and DAC4 data are exchanged via CCN/CSI. DAC3 and DAC6 are supported by CCN-based central directories. Regarding the costs, the Commission covers the central infrastructure costs of the CCN via the FISCALIS budget (Figure 9 below), while Member States are responsible for their own connection and operational costs 61 .

Figure 22: Financing of Fiscalis IT systems (2014-2020)

DAC	Total development cost for the Fiscalis 2020 Programme
DAC1	EUR 436,000
DAC2	EUR 587,000
DAC3	EUR 715,000
DAC4	EUR 300,000
DAC6	EUR 1,146,000

Source: Data provided by DG TAXUD in the context of the Study on the final evaluation of the Fiscalis 2020 programme

Finally, costs for reporting by Member States to the Commission are in the range of EUR 140,000 per year, half of these costs relate to the compilation of the data necessary for the Yearly Assessment survey.

Effects of differences in implementation

With respect to DAC6 the results from the public consultation , and the consultation of professional organisations in the framework of the Platform for Tax Good Governance 62 showed that there are some concerns with respect to the costs of reporting for intermediaries and taxpayers.

In particular, tax law experts and businesses highlighted the differences across Member States in the interpretation of the hallmarks, which increases the compliance costs and creates an unnecessary administrative burden. While it was not possible to precisely assess the costs generated by these differences, some respondents to the public consultation indicated that the cost of analysing a transaction potentially subject to reporting under DAC6 could exceed the amount of the transaction itself, creating an obvious cost/benefit distortion from the perspective of the taxpayers involved.

4.1.2.3. Costs vs benefits

According to the cost-benefit analysis, the recurring burdens due to DAC are commensurate with the benefits generated.

Under various sensitivity scenarios, annual net benefits range between EUR 500 million and 6.1 billion 63 .

Net benefits remain positive and significant even when considering the uncertainties linked to the quantification of benefits deriving from the impact of DAC3 and DAC4. However, DAC4 economic literature 64 indicates that Public CbCR has the potential to increase the effective tax rate for MNEs and reduce profit shifting.

Even under the most conservative estimates, which are based on primary tax benefits data provided by a growing number of Member States, the benefits from the DAC are estimated at between 8 to 10 times more than the costs incurred 65 .

Figure 23: Costs vs benefits

Scenario	Costs (€mln)	Benefits (€mln)	Net (Benefits – Costs)	Cost-Effectiveness Ratio (Benefits/Costs)
Comprehensive scenario	650	6780	6130	10,43
Low-range scenario	650	4940	4290	7,60
Data-based scenario	650	1180	530	1,82

Source: Ramboll study based on Member States and DG TAXUD data

However, certain factors raised by stakeholders must be considered even if it is not possible to assess their impact on the cost/benefit ratio.

First, the frequency of DAC amendments has been significant, expanding the scope of the Directive almost yearly, and recent provisions may not yet have reached their full expected impact. This has also resulted in the creation of a lengthy text with several annexes and frequently updated implementing acts. This increases complexity and makes it more difficult to understand the legal framework. The results of the consultations conducted with both tax authorities and the business sector confirm that the evolving legal framework is increasingly complicating the implementation for public and private stakeholders.

The time left for implementation after the adoption of DAC amendments is often considered by stakeholders to be too short. This increases pressure on both tax administrations and private stakeholders. This assessment is supported by the fact that most DAC-related infringement procedures relate to delays in implementation.

Moreover, the evolution of the DAC has led to a significant increase in information systems developments, increasing the costs of implementing and operating those systems, and the associated human resources required for their design and maintenance. This applies to the public sector as well as private stakeholders. In particular, some comments received during the public consultation from the financial sector identified the verification of the Tax Identification Numbers (TINs) in DAC2 exchanges, as an area for improvement. In particular, it was suggested to improve technical interoperability by investing in IT systems to reduce administrative burdens.

4.1.3. Coherence

The Directive is coherent with various EU and non-EU legal provisions.

4.1.3.1. Coherence with international initiatives

The degree of coherence of DAC provisions is high. In many cases, DAC provisions directly incorporate definitions and operational processes from OECD initiatives (e.g., XML schemas). Notable examples include DAC2 / Common Reporting Standard (CRS); DAC4 / BEPS Action 13 and CbCR; DAC7 / Model Reporting Rules for Digital Platforms (MRDP); and DAC8 / Crypto-Asset Reporting Framework (CARF). However, this alignment is not uniform. For example, DAC3 and DAC6 diverge from OECD framework or recommendations (BEPS 5 minimum standard is based on spontaneous information while DAC3 is AEOI; MDR report covers a limited range of reportable arrangements, while DAC6 hallmarks have a broader scope).

A number of contributions to the open public consultation 66 refer to the issue of coherence of the DAC with the international tax framework. In particular, 2 initiatives were cited: the US Foreign Account Tax Compliance Act (FATCA) and OECD Pillar 2. In this respect, 2 stakeholders argued for clearer self-certification procedures and better alignment between FATCA and DAC2 classifications. They argue that these elements would streamline the reporting obligations and improve the overall efficiency of tax compliance frameworks.

8 associations representing both tax advisors and businesses mentioned the need to enhance synergies among OECD Pillar II and DAC4 and simplify the reporting requirements under the latter. This would avoid duplication of effort and significantly reduce compliance costs. 5 of them mentioned the OECD temporary CbCR Safe Harbour as a possible common ground as it would involve less extensive calculations based on a smaller pool of already available data from the CbCR report.

4.1.3.2. Coherence with EU acquis

There is a notable coherence with other relevant EU initiatives particularly those with similar instruments for administrative cooperation in different tax areas (e.g., VAT, excise, asset recovery). However, small inconsistencies exist, mostly stemming from the differing objectives pursued by other EU Directives, such as the AML Directive (AMLD) and the Anti-Tax Avoidance Directive (ATAD). Nevertheless, significant synergies have been identified, such as those between DAC2 and the AMLD.

With regard to other EU legislation stakeholders participating in the public consultations highlighted that the use of consistent definitions across different pieces of EU legislation (e.g., ATAD, VAT administrative cooperation regulation, and the Recovery Directive) would greatly enhance clarity and efficiency. One example mentioned was the DAC6 hallmark B2 covering inter alia hybrid instruments, addressed also by ATAD.

Furthermore, 5 associations and business communities representing the banking and financial sector stressed the need for greater coherence between DAC6 obligations and the AMLD. With regard to the latter, one stakeholder argued that while the AMLD prohibits the disclosure of information to customers in cases of suspicion, DAC6 requires the exchange of information on tax arrangements potentially used to circumvent automatic exchange of information.

4.1.3.3. National implementation

Several private stakeholders raised the issue of the different implementation of the DAC across Member States. Differences in implementation may be linked to various factors. In some cases, the EU framework is adapted to domestic constraints, which may lead to slightly divergent approaches for example, this is the case with the concept of legal and professional privilege in DAC6.

Similarly, while the broad scope of DAC6 and certain definitions therein does not fundamentally conflict with the principles of legal certainty and legality in criminal matters, as noted by the CJEU, it may lead to differences in interpretation. Consequentially this results in a more complex application, which ultimately increases the administrative burden.

Overall, the issue raised by private stakeholders concerning the different implementation of the DAC across Member States in many cases seems to reflect more a need for clearer and common guidance on the interpretation of DAC provisions rather than a fundamental problem of coherence.

Another area where divergences have been identified by several stakeholders concerns penalties imposed on intermediaries for non-reporting, incomplete reporting or incorrect reporting in respect of DAC2, DAC4 and DAC6. According to the findings of the study, criticism has been raised by stakeholders on the differing levels of proportionality of penalties in some Member States. In relation to DAC6, the variation of penalties adopted by Member States are perceived by tax advisors as affecting the level-playing field within the EU. Overall, the study concluded that penalties related to the non-reporting, incorrect and incomplete reporting under DAC6 have been rarely or never applied.

The impact assessment for the DAC8 proposal as well as the 2021 ECA report confirmed that there were significant differences in the levels and range of penalties that may be applied by Member States. This disparity may then affect to a greater or lesser extent, the effectiveness and efficiency of compliance with the rules of the DAC. The VISDAC project also confirmed these findings in that only a limited number of Member States have applied penalties in practice, with most Member States applying few penalties or having not yet applied penalties. VISDAC found that only 10 Member States have applied penalties related to DAC2 while 17 either did not or could not confirm that they have applied penalties. For DAC4 3 Member States had applied penalties while 24 either did not or could not confirm whether they had applied penalties.

For some Member States it should be considered that there are comprehensive cooperative compliance measures in place that for the most part target the MNEs that are subject to DAC4. This would have the effect of reducing the number of cases where penalties are applied.

4.1.3.4. Common administrative guidance

According to the 2023 Yearly Assessment 24 Member States published administrative guidance on DAC6, to support intermediaries and taxpayers in complying with their reporting requirements. 18 Member States provided administrative guidance for assessing arrangements in accordance with the hallmark criteria in Annex IV.

There is no common administrative guidance for DAC1, DAC3 or DAC6. With respect to DAC2 and DAC4, the OECD has published guidelines for the equivalent CRS and CbCR. The absence of common administrative guidance was highlighted by the ECA audit on harmful tax regimes and is the subject of one of its recommendations. Member States raised the same point and stressed that efficiency would be enhanced if tax authorities had common guidance, which would also be available to intermediaries and taxpayers. This is somehow contradictory with the conclusions reached by Member States at the end of the 3 Fiscalis workshops held in 2020 and 2021 67 which focused on the implementation of DAC6 across the different Member States. On that occasion Member States concluded that the content of the final reports should not be taken as binding interpretations of the Directive and not be published.

4.2.How did the EU intervention make a difference and to whom (EU added value)?

The EU, through the DAC, has been a frontrunner in adopting measures to increase tax transparency and tackle tax fraud, evasion and avoidance, including by implementing standards that have been developed at international level but not yet globally applied.

All EU tax authorities agree that DAC adds value compared to international, bilateral, or national solutions in terms of efficiency gains and additional benefits. The main efficiency gains stem from the fact that the DAC ensure that information is exchanged in line with EU rules on data protection, from the binding nature of DAC under EU law, its underlying IT architecture, and the uniform reporting and exchange requirements across the EU, notably when implementing standards agreed at international level.

Firstly, the DAC provides a solid legal framework that facilitates comprehensive data sharing among all Member States, in a way that is fully aligned with the provisions of the General Data Protection Regulation (GDPR). Secondly, by providing for mandatory automatic exchange of information (in addition to exchange of information on request and spontaneous exchange of information), the DAC ensures that all Member States have access to critical tax information. The increased tax transparency which results from AEOI significantly reduces opportunities for tax fraud, tax evasion as well as tax avoidance strategies that exploit disparities between national tax systems and contributes to the fairness of EU tax systems. Thirdly, the common IT infrastructures established under the DAC and financed through the Fiscalis programme streamline the exchange processes, reducing administrative burdens, costs and inefficiencies associated with bilateral agreements.

This setup also strengthens mutual trust and cooperation between tax authorities. The significance of EU funding in supporting the set-up of the IT infrastructure and resource-intensive forms of administrative cooperation (e.g., PAOE and simultaneous controls) has been highlighted by all tax administrations. Additional benefits vary and are higher for the provisions without an equivalent at the OECD level, such as DAC1, non-AEOI instruments, and, to some extent, DAC6. Finally, in relation to international standards, the DAC ensures a coordinated approach to implementation and application which is crucial to safeguarding the functioning of the Internal Market.

This last aspect is also very relevant and has been highlighted by private stakeholders. From their perspective the DAC usually provides added value by avoiding the costs associated with a potential fragmentated implementation of the standards developed at international level which would result in a variety of reporting requirements across the EU. Thanks to the DAC, financial institutions and intermediaries operating in several EU Member States as well as MNEs covered by DAC4, benefit from a single set of reporting requirements, including standardised technical specifications. This eliminates the need to adapt to different local implementations and prevents unnecessary duplicate costs related to tailoring IT systems, data collection, and reporting procedures, thereby contributing the smooth functioning of the internal market. Nevertheless, some critical voices have been raised regarding certain provisions of the Directive, notably DAC6, insofar as its scope is very broad and does not match the standard developed at international level.

4.3.Is the intervention still relevant (relevance)?

4.3.1. Combating base erosion and aggressive tax planning

The erosion of tax bases and aggressive tax planning, resulting from the increasing globalisation and mobility of people, capital, and companies, is one of the main challenges that the DAC aims to address. Overall, about half of the 83 respondents from the targeted consultations conducted for this evaluation in the context of the Study believed that the scope of this issue has been reduced, although tax authorities hold a much more positive view compared to intermediaries.

Respondents often highlighted that the improvements stem from a global shift in public awareness and sensitivity on the topic, as well as from various international initiatives addressing tax avoidance and evasion, undertaken both at the OECD and EU levels, including DAC.

Regarding the DAC’s fitness for purpose in addressing the erosion of the tax bases of Member States and aggressive tax planning, stakeholders hold a broadly positive opinion, in line with the findings of the previous evaluation. Tax authorities generally consider DAC appropriate for its objectives, as reflected in annual questionnaires and targeted consultations. This is also thanks to the fact that the DAC is a flexible tool that allows for integration of new areas for cooperation and new exchanges of information when the need emerges. As economies become increasingly interconnected, new business models emerge and digital platforms proliferate, traditional tax systems face new obstacles in tracking income and ensuring compliance. By constantly evolving to include new components such as DAC7, which targets income from digital platforms, or DAC8 that ensures transparency in the area of crypto assets, the directive has remained relevant and effective in addressing the new challenges brough about by modern economic realities.

However, private stakeholders, while mostly positive (60% of around 73 respondents), expressed less consistent feedback. Above all, while it appears that the fight against tax evasion is still of paramount interest to many stakeholders, some of the MNEs’ professional organisations wish to deregulate some parts of the DAC, with two business associations even suggesting that they be abolished, e.g. DAC6. However, this very critical stance reflects only a minority of views, with the other business stakeholders calling instead for simplification and clarification of some provisions of the DAC (notably DAC6).

4.3.2. Improving data quality

The relevance of information is indicated by the ever-increasing volumes of exchange; however, the quality of the information is critically important to achieve the benefits of the exchange.

The higher the quality of the information the greater the possibility that it can be used by tax administrations. In the 2018-2023 period, the use of the information has significantly increased as compared to the previous 5 years. This increase can be attributed to the progress made in terms of the identification of taxpayers. Nearly all tax authorities use DAC information extensively with effective results.

This suggests that the issue of matching of information, which was very present in previous reports (previous evaluation and ECA report), is less problematic than before. This improvement can be attributed to the increase in human and technical resources used by tax administrations, owing to the fact that a significant part of the information is still manually matched when the usual matching techniques (automated and fuzzy matching) are insufficient to identify the taxpayer 68 . However, the findings of the VISDAC Expert Team and FPG 037 show that there is still room for improvement in data quality.

Bilateral feedback of a qualitative and quantitative nature is an integral element to improving data quality as it enables sending Member States to address quality issues for future exchange periods. This assessment is also supported by the findings of the VISDAC Expert Team and the evaluation.

Unfortunately, the current bilateral feedback system is archaic and not fit for purpose, based on a table-type file drawn up manually, which makes it time-consuming and consequently the feedback mechanism is not utilised to its full potential.

4.3.3. DAC6 reporting requirements

According to both public and, even more, private stakeholders, some aspects of DAC6 are neither clearly defined nor easily applicable. On the one hand, given that DAC6 was introduced only recently, some difficulties in the application may reflect typical initial implementation challenges. On the other hand, stakeholders claim that certain provisions are inherently problematic, particularly those concerning the definition of hallmarks, the main benefit test, and the intermediaries subject to reporting obligations.

The generic nature of some of these hallmarks has at least initially resulted in a very large number of reports, the relevance of which is not always established by the receiving tax authorities. In response reporting persons, especially MNEs, have voiced criticisms, going so far as to call for the removal of certain hallmarks deemed to be of little relevance.

Moreover, contributions received from tax advisors and legal professions point to inconsistencies in how DAC6 is applied among Member States, due to varied interpretations of hallmarks and differing legal protections, which increased administrative burdens and legal uncertainty. To avoid these issues, stakeholders suggest the Commission and the Member States to work on common guidance to clarify definitions and avoid discrepancies in application. 69

5.What are the conclusions and lessons learned?

5.1.Conclusions

Overall, DAC provides a robust legal framework that facilitates close administrative cooperation between Member States tax authorities with a view to fighting against tax fraud, evasion and avoidance. The Directive and its various amendments have been transposed without major problems and did not lead to infringements on the substance of the transposition.

DAC is an evolving, adaptative and relatively agile instrument. It has continued to evolve regularly between 2018 and 2023 to adapt to the challenges of a complex economic environment, and the ever-increasing sophistication of tax evasion and avoidance in the context of globalisation, in line with international developments.

The DAC led to the exchange of substantial volumes of information, which is growing over time especially for DAC1, DAC2 and DAC4. The data of automatic exchange is increasingly matched and used by tax authorities, both for risk assessment and control purposes and to foster voluntary tax compliance thanks to greater awareness of enhanced transparency. The evidence, as supported by the view of tax authorities and private stakeholders, suggests that the Directive is effective in contributing to safeguarding Member States tax base and tax revenue. Total tax benefits are estimated at about EUR 6.8 billion per year (about 0.3% of EU direct tax revenue).

There is a consensus, supported by evidence, that DAC contributed to increasing Member States ability to fight tax fraud, evasion and avoidance, thereby contributing to ensure that taxes are paid in the Member State where they are due. AEOI, particularly DAC1 and DAC2, allows to identify undeclared tax income and assets and to increase compliance, via both enforcement and deterrence.

The causal link between additional income and DAC3, DAC4 and DAC6 is more complex to establish, insofar as, for DAC3 and DAC6, the data is mainly descriptive, and for DAC4, it is not allowed to make direct use of the data exchanged to establish tax adjustments. While there is no specific data for DAC4, the economic literature suggests that CbCRs has the capacity to increase the effective tax rate for multinationals.

AEOI, particularly DAC1 and DAC2, also have a deterrent effect on individual taxpayers, though unquantified, and encourage voluntary tax compliance: DAC2 influenced the disclosure of financial assets abroad and related taxes; DAC3, and to a lesser extent DAC4 and DAC6, contributed to transparency in tax arrangements that were known to be particularly harmful for Member States revenue. The deterrent effect of DACs on corporations’ profit-shifting was less significant, stakeholders mentioning in the OPC that the influence of DAC4 or DAC6 on their decisions had a relatively limited impact.

This relatively positive outcome is the result of massive use of DAC1 and DAC2 data among the Member States, and the gradual integration of DAC3, DAC4 and, more recently, DAC6 data into the risk analysis models used to assess the tax risks associated with the taxpayers concerned. The VISDAC expert team was able to observe the extent of these developments, but also the incremental nature of the Member States’ approach. Each new type of information requires an update of the risk analysis system that requires mobilisation of substantial financial and human resources. As the study has shown, these investments are generally recovered over time by the improved accuracy of the analyses once they are in place. But they are nevertheless significant and cannot always be deployed immediately, due to the financial constraints that may limit their ability to deploy the appropriate tools rapidly.

The Directive has 2 further objectives, that is contributing to the smooth functioning of the single market and increasing the fairness of the EU tax system. They are fulfilled to the extent that the Directive safeguards Member States’ tax revenues. DAC positively contributes to the functioning of the single market through 2 channels. First, it limits the opportunities for tax evasion and avoidance for companies and individuals operating cross-border, thus removing or limiting tax-reducing distortions. Second, it has obtained positive results in harmonising reporting requirements for private sector operators, with respect to DAC2, DAC4 and DAC7, as well. Due to increased transparency, DAC also increases the fairness of the EU and national tax systems.

The system works efficiently at the cost of major efforts and investments on the part of the various stakeholders, notably Member States’ tax authorities and the business, who seem sometimes to be struggling to keep pace with the numerous changes.

A comparison of costs and benefits suggests that the costs associated with DAC are commensurate with the benefits generated.

The overall coherence of the DAC is relatively strong. It is aligned with the OECD principles for most of the measures that stem from it, such as DAC2 and DAC4 and, to a lesser extent, DAC3 and DAC6. As regards DAC6, the broad scope of the directive and the lack of clarity of some concepts have undermined the consistency of its application within the EU, leading to divergent interpretations among Member States. It is also complementary in its objectives with the measures to combat money laundering and terrorist financing (AML Directive), and those to combat VAT fraud (Regulation 904/2010). While the differences that may exist between the AML framework or the VAT framework and the concepts and definitions in the DAC may be justified insofar as the scope of each piece of European legislation is different, it would make sense to explore the possibility to better aligning terms or principles wherever justified and possible, with a view to ensure simplification.

As regards the EU added value, all tax authorities agree that DAC is more efficient than international or national alternatives. The main efficiency gains have to do with the binding nature of DAC under EU law, the underlying IT architecture and the uniform reporting requirements across the EU. DAC is believed to be a better framework, which is conducive to bigger exchanges compared with international peers, which in turn increases the awareness of and familiarity with DAC. More exchanges mean more experience and hence learning economies in data processing and usage and actual impact. Furthermore, the harmonisation also ensures the coherence with the rest of the EU acquis.

As regards the relevance, the DAC keeps being very relevant and crucial in the fight against tax evasion and aggressive planning. The volume of exchanges and data exchanged is constantly increasing, the use of data is virtually systematic for certain areas of the Directive, and the instruments of cooperation and control are also being used effectively, leading to a balance costs/benefits in favour of additional revenues for Member States. However, this generally positive picture is nuanced by more problematic aspects: as outlined above, the complexity of certain parts of the directive such as DAC6 and the relatively unharmonised way in which they are implemented and applied can weaken the impact of the Directive.

5.2.Lessons learned

5.2.1 The DAC legal framework is robust, but fragmentation of application across the EU increases administrative burden on business

While the DAC is overall perceived as providing a robust legal framework for administrative cooperation some concerns were raised regarding the lack of consistency in the application of the DAC provisions across Member States and its consequences, particularly in terms of increased administrative burden for the business. While the overall picture is coherent, there have been calls for greater accuracy in some concepts and definitions, which should be better aligned across all the tax initiatives proposed at the EU level. Stakeholders also referred to the need to better align with and exploit synergies with other legal provisions that share the same or similar purposes, like Regulation 904/2010 on cooperation in the field of VAT and the AML framework.

The concerns focus on DAC6 which is perceived as complex, to some extent unclear and subject to more fragmentation in application, increasing the risk of legal uncertainty and, consequently, the burden on businesses. According to private stakeholders and tax authorities, it is the most challenging legislation to apply. Some of these challenges are likely to dissipate in the light of the judgements of the CJEU, which clarify some definitions and the scope of the legal professional privilege, but they are unlikely to disappear completely due to the lack of harmonisation outside of tax policy.

One of the perceived problems of DAC6 is that some of its provisions (e.g., the main benefit test, certain hallmarks) are deliberately vague, as the legislator wanted the net to be broad enough to catch all forms of potentially tax harmful cross-border arrangements. Consequently, the legislation is posing interpretative challenges for the Member States, leading to different interpretations and application of the DAC6 provisions across them, which create uncertainty among intermediaries and increase the burden and the associated costs 70 .

5.2.2 The penalties framework for non-compliance with reporting obligations under the DAC varies considerably between Member States

The DAC requires Member States to implement effective, proportionate and dissuasive penalties for non-compliance by reporting legal entities and natural persons under DAC2, DAC4 and DAC6.

In its special reports on “Exchanging tax information in the EU” which was published in January 2021 71 , the ECA highlighted the risk that penalties that Member States apply are not deterrent enough to ensure full compliance with the DAC2 reporting requirements.

In the impact assessment for DAC8, the Commission analysed Member States’ penalties regimes and concluded that there were significant differences between the range of penalty amounts that Member States can impose, which affect their effectiveness and hinder the level playing field. Based on that analysis, the Commission proposed in 2022 to introduce minimum penalties for reportable offenses. However, this amendment was ultimately not accepted by the Council. The issue remains, and it is highlighted again by the ECA in the special report published in November 2024. The ECA recommended to the Commission to initiate infringement proceedings in those cases where there is sufficient evidence that Member States are implementing a manifestly inadequate penalty system for breaches of the DAC6.

5.2.3 The quality of data has improved, but identifying taxpayers is still an issue for some exchanges

The quality of the data has undeniably improved between 2018 and 2022, making it possible to significantly improve the matching of DAC1 and DAC2 data. Given the smaller populations involved, the issue is less pressing for DAC4, DAC3 and DAC6. The latter 2 nevertheless display recurring problems with the abstract data resulting from the provision of a substantial part of the summary information in free text format.

Nevertheless, as far as DAC1 and DAC2 are concerned, since Member States exchange hundreds of thousands of records each year, even a very small percentage of unmatched data can represent thousands of unidentified taxpayers and therefore cause tax losses due to the inability to match. In addition, this unmatched data generates a need for providing manual feedback which is administratively cumbersome.

Furthermore, the feedback process used to identify and solve matching problems at bilateral level is based on a communication model that is no longer adapted to the issues at stake: in other terms, the current model of feedback is not satisfactory for Member States, insofar as it is based on an outdated process that does not facilitate widespread handling of issues encountered.

Beside identification issues, there is also a general observation from tax authorities about the quality of data that depends on a literal approach, which is the case for the summaries of DAC3 and DAC6. This relatively poor quality of the DAC3 and DAC6 summaries calls for improvement.

Another aspect is that it appears that one of the major weaknesses in the reporting schema of DAC6 is the fact that the data field to indicate the names of non-EU countries involved in a cross-border arrangement is not mandatory. As a result, it is difficult for the Member States to identify whether/which non-EU jurisdictions are involved.

5.2.4 The use of data is widespread (while not systematic), but administrations should be made more accountable for how DAC data is used

Along with the constant improvement in the quality of the information exchanged, the increase in the use of data by the Member States is one of the most significant developments since the first evaluation. However, while DAC1 and DAC2 are now used almost systematically, this is not yet the case for DAC3, DAC4 and especially not for DAC6.

Use of data had developed massively, including use for other purposes than tax bases-related only (e.g. recovery of tax claims). However, the current situation cannot yet be generalised across the DAC, particularly insofar as the descriptive data present in DAC3 and DAC6 are not as easy to integrate as the digital data provided by DAC1 and DAC2.

Member States find it easier to integrate quantitative data (income or amount of assets) exchanged under DAC1 and DAC2 than data requiring more in-depth analysis as to their significance, such as DAC3, DAC4 and DAC6. While tax authorities make extensive use of DAC1 and DAC2 information to establish, re-assess tax bases or corroborate the tax returns of the taxpayers concerned, DAC3, DAC4 and DAC6 data are more easily included in a risk analysis cycle which, for example, makes it possible to assess the possibility of a tax evasion scheme. There is still room for improvement to ensure that a holistic approach to this data is implemented in certain Member States, moving from a case-by-case analysis to a more systematic approach.

Despite these very concrete results, some private stakeholders seem to be sceptical about the evidence, addressing tax authorities with calls for greater transparency when it comes to the use of data. The European Court of Auditors is mentioned several times in replies to the public consultation as the body which is best placed to make a ‘real assessment’ of the implementation of the DAC, DAC6 in particular.

This is one of the few points where there is a real gap between reality (widespread use of data, which continues to progress, allows tax authorities to better detect situations at risk or dubious taxpayers and procure revenues to Member States) and the perception from certain stakeholders. This is especially surprising since at least some Member States do not hesitate to communicate publicly on the use of data resulting from automatic exchanges of information. This shows that there is a grey area regarding awareness of the use of data, which should be clarified.

While the Member States report annually to the Commission under the obligations laid down in the DAC (Yearly assessment and annual questionnaire on the functioning of the DAC, statistics) on different aspects of the DAC, including quality and use of data, they do not necessarily do so vis-à-vis the public. Moreover, not all the Member States have tools to measure the effectiveness of administrative cooperation in the form of metrics or indicators. Where these indicators are available, they are not necessarily harmonised, and they often remain essentially internal.

In light of the comments from some stakeholders on the use of the DAC data, it seems obvious that the increasing development of use of information by tax authorities has not been reflected in terms of monitoring the outcome, or at least not sufficiently publicised.

In 2021, the European Court of Auditors requested more transparency and visibility of the achievement and results of Member States and called for a common EU framework for monitoring administrative cooperation activities and measuring their outcome better. In its conclusion the Court stated, “There is no common EU framework for monitoring the system’s performance and achievements, and the individual assessments carried out by some Member States are short on clarity and transparency”.

Member States have initiated workstreams, together with the Commission services, to address the recommendations of the ECA. A set of harmonised metrics and indicators was endorsed by the heads of Tax administrations in December 2024. The set covers DAC1, DAC2, DAC3 and DAC4, including non-automatic exchange of information. It was recommended that the Member States should implement the metrics and report the results on an annual basis by the Central liaison offices of the Member States to the relevant Commission services.

The collective efforts of the Commission and Member States must be consolidated and strengthened. At the same time there is a need for increased focus on publishing the results obtained, so that the public is better informed about efforts and successes in the fight against tax evasion.

5.2.5. The IT information exchange systems work, but require significant allocation of resources to operate them

Overall, the information systems made available by the European Commission and the Member States are not the subject of direct criticism, but the significance of the costs and resources deployed to make it work is an issue given the current state-of-play: the information exchanges provided for in the DAC operate in parallel, autonomously, for each type of exchange. In addition, they are handled by the Member States, which develop their own interfaces to make them available to stakeholders subject to reporting obligations.

Within the framework of the EU, this differentiated approach for the 27 Member States can be justified by the fiscal sovereignty that is applied, including in the implementation of the DAC. However, there is a strong shared interest among the various stakeholders, both in the private sector and on the part of the Member States, for enhancing tax compliance and efficiency through digital transformation.

This calls for a more uniform approach to achieve economies of scale, and improve the system on the basis of simplification, rationalisation of investments, agility and synchronisation of evolution among Member States.

Annex I: Procedural Information

1.Lead DG, Decide reference and, if relevant, Work Programme reference. Derogations granted and justification

DG TAXUD; PLAN/2022/2895 included in Annex II to the 2024 Commission work programme. No derogations granted.

2.Organisation and timing

The chronology of the evaluation can be summarised as follows:

-Interservice steering group: created on 10 January 2023, with the participation of DG COMP, DG FISMA, DG ECFIN, DG TAXUD.

-Meetings of the interservice steering group: the interservice steering group met 3 times on 9 February, 2 March 2023 and 15 April 2024.

-Terms of reference for the external study: finalised September 2022.

-External study: carried out between December 2022 and December 2023 by Ramboll in collaboration with Syntesia Policy and Economics and London Economics.

-Questionnaire on the functioning of the DAC to Member States: from May 2024 to July 2024.

-Questionnaire on the relevance of annex IV of DAC6 to Member States: from May 2024 to July 2024.

-Public consultation (questionnaire online): from 7 May 2024 to 30 July 2024.

-Call for Evidence: from 7 May 2024 to 30 July 2024.

-Targeted consultations: 25 and 26 September 2024.

3.Evidence used together with sources and any issues regarding its quality (i.e. has the information been quality assured?)

The evidence used in this evaluation collected through desk research was derived from the following sources:

-Commission documents:

o Commission Staff Working Document Impact Assessment Accompanying the document Proposal for a Council Directive amending Directive 2011/16/EU as regards mandatory automatic exchange of information in the field of taxation in relation to reportable cross-border arrangements , June 2017

o Commission Staff Working Document Evaluation of the Council Directive 2011/16/EU on administrative cooperation in the field of taxation and repealing Directive 77/799/EEC , September 2019

o Replies of the European Commission to the European Court of Auditors Special Report , November 2024

-Reports from the Court of Auditors:

o Special Report No 03/2021: Exchanging tax information in the EU: solid foundation, cracks in the implementation , March 2021

o Special Report 27/2024: Combatting harmful tax regimes and corporate tax avoidance – The EU has established a first line of defence, but there are shortcomings in the way measures are implemented and monitored , November 2024

-Report and resolution from the European Parliament:

o European Parliament legislative resolution of 10 March 2021 on the proposal for a Council Directive amending Directive 2011/16/EU on administrative cooperation in the field of taxation (COM(2020)0314 – C9-0213/2020 – 2020/0148(CNS)) , March 2021

o European Parliament resolution of 16 September 2021 on the implementation of the EU requirements for exchange of tax information: progress, lessons learnt and obstacles to overcome (2020/2046(INI)) , September 2021

-Case law from the Court of Justice of the European Union:

o C-694/20 - 08/12/2022 - Orde van Vlaamse Balies e.a.

o C-623/22 – 29/07/2024 - Belgian Association of Tax Lawyers and Others v Premier ministre/ Eerste Minister

o C-432/23 – 26/09/2024 - F and Ordre des avocats du barreau de Luxembourg v Administration des contributions directes

o Berlioz Investment Fund SA v Directeur de l'administration des contributions directes (Case C-682/15)

o B v État Luxembourgeois and État Luxembourgeois v B, C, D and FC (Cases C-245/19 and C-246/19)

The other sources of evidence used in the evaluation include:

-Evidence from Member States: replies to written questionnaires in the context of the external study and submitted to the Commission;

-Evidence from public consultation and call for evidence: contributions from business associations, individual companies, EU citizens and non-governmental organisation (NGO).

-Evidence from economic operators (targeted consultations):

The external study by Ramboll in collaboration with Syntesia Policy and Economics and London Economics is based on a comprehensive collection of information collected through desk research, stakeholder interaction, and data review, with a focus on both quantitative and qualitative analysis. The contractor cross-checked evidence from multiple sources available to the Commission, ensuring reliability.

Additional consultation activities carried out by the Commission helped refine the quantitative data, enhancing the robustness of the findings.

Overall, the present evaluation benefited from improved data availability compared to the previous evaluation and the quality of evidence is considered satisfactory.

4.Use of external expertise

The evaluation is supported by an external study conducted by a consortium of consulting firms and research institutions, led by Ramboll in collaboration with Syntesia Policy and Economics and London Economics (Specific Contract No. TAXUD/2022/AO-12 under Framework Contract TAXUD/2019/CC/148-149-150 for a “Study on evaluation of the Directive 2011/16 EU and its amendments”). This study provided evidence and findings on the implementation of the DAC, forming the basis of the assessment in this Staff Working Document.

The methodology for data collection was discussed and agreed upon with the Commission. Additionally, this Staff Working Document considers the perspectives of EU-wide trade representative organizations regarding the practical implementation of DAC provisions and related processes.

Annex II. Methodology and Analytical models used

Description of methodology

The Commission conducted this evaluation in line with better regulations principles, ensuring consistency with the 2019 evaluation and integrating findings from the 2021 and 2024 ECA reports and the 2021 EP resolution 72 .

Specifically, the evaluation consisted of 3 main phases:

1.Structuring phase: this phase involved preliminary activities for structuring and developing the evaluation, including the identification of the external contractor and other stakeholders to be consulted;

2.Data gathering and analysis: this phase encompassed data analysis, consultation activities, and legal and desk research, providing the necessary data and information for the subsequent assessment;

3.Evaluation and lesson-learning: in this final phase, the collected data and information were analysed and used to produce the evaluation report, covering the 5 evaluation criteria, as well as to formulate conclusions and lessons learned.

Figure 24. Evaluation methodology and its phases

Structuring phase

The first phase of the evaluation began in November 2022 with the signing of a contract for an external study. Specifically, the first phase encompassed the inception phase, which concluded with the submission of the revised inception report in March 2023. Key activities from the external contractor’s side included:

·Data mapping: identifying available data from both public and non-public sources. Public data mapping involved reviewing legal information and literature, including policy documents and academic contributions. Simultaneously, internal Commission data sources - such as EU and national statistics, DAC tool data, and Member States’ questionnaires - were organized and reviewed.

·Familiarization interviews: conducting 16 interviews to identify key issues, gather insights, and secure support from EU organizations for stakeholder engagement.

·Refinement and operationalization of methodology: fine-tuning the evaluation methodology and developing tools for subsequent fact-finding, including the Directive’s Intervention Logic (see Annex VI), Evaluation Matrix (see Annex III), and the final consultation strategy and materials, such as sample strategies and consultation questionnaires.

Data gathering and analysis phase

The data gathering and analysis phase lasted from March to June 2023. Specifically, this phase included:

·Implementation analysis: assessing the transposition of DAC provisions into national legislation.

·AEOI and administrative cooperation analysis: conducting a quantitative analysis of AEOI statistics to examine data availability, geographical exchange patterns, and the completeness and timeliness of exchanged information. The study also reviewed non-AEOI tools through CACT statistics.

·Member States questionnaires: evaluating responses to 3 key surveys:

1.The Yearly Assessment Questionnaire (covering AEOI provisions).

2.The Questionnaire on the Functioning of the Directive (assessing Member States’ perspectives on all DAC provisions, including administrative cooperation).

3.The Questionnaire on the Relevance of DAC6 Annex IV.

·Desk research and legal analysis: reviewing academic literature and other qualitative sources relevant to the Directive’s assessment.

·Targeted consultation: conducting over 50 interviews with private stakeholders and 14 interviews with tax authorities. Additionally, 3 surveys were carried out:

oAn email survey of the remaining 13 tax authorities.

oAn online survey of financial institutions (11 responses).

oAn online survey of intermediaries (60 responses).

Evaluation and lesson-learning phase

The evaluation and lesson learning lasted from June to September 2023 and is primarily based on the findings of the external study carried out by Ramboll in collaboration with Syntesia Policy and Economics and London Economics. During this phase, the final evaluation of data was performed. All collected data, information, and viewpoints were systematized, triangulated, and analysed to answer the evaluation questions, ensuring a comprehensive assessment.

Annex III. Evaluation matrix

Evaluation questions matrix – Efficiency

Question	Sub-questions	Indicators
# To what extent was the intervention successful and why	# Level of costs and cost savings borne by tax authorities	·Adjustment costs (investment, operating) due to AEOI ·Adjustment costs due to other forms of ACDT ·Share of expenditures financed by EU budget (particularly Fiscalis) ·Cost savings compared to pre-Directive situation ·Administrative burdens due reporting obligations
	# Level of costs and cost savings borne by the private sector	·Administrative burdens for financial institutions due to DAC2 AEOI ·Administrative burdens for MNE due to DAC4 AEOI ·Administrative costs and burdens for intermediaries/companies due to DAC6 AEOI ·Administrative burdens for online platform operators due to DAC7 AEOI
	# Extent to which costs are commensurate with the outcome achieved	·Comparison of costs and benefits ·Stakeholders’ views on whether the additional costs were justified based on the results achieved

Evaluation questions matrix – Effectiveness

Question	Sub-questions	Indicators
# To what extent was the intervention successful and why	# Extent to which the Directive provided an effective framework for Member States to cooperate	·Member States degree of satisfaction as reported in expert group meetings and/or DAC-related Fiscalis events ·TAs reporting good experience in the interaction with other Member States (by provision) ·TAs reporting good experience with sharing of best practices
	# Extent to which the Directive contributed to an increased capacity to fight tax fraud, evasion and avoidance	·Extent to which the information received via AEOI, EOIR, and SEOI is complete, of sufficient quality, and is received within the established time limits ·Extent to which data received via AEOI and SEOI is processed and used for tax control ·Effect of PAOE, simultaneous controls, and joint audits on the effectiveness of the fight against tax fraud
	# Extent to which the Directive resulted in an increased spontaneous tax compliance	·Extent to which AEOI data (particularly DAC1 and DAC2) are used to encourage spontaneous tax compliance (e.g. use of DAC1 and DAC2 data for pre-filling of tax returns, informal letters) ·Evidence of increased spontaneous tax compliance (deterrent effect)
	# Extent to which Fiscalis projects contributed to the achievement of DAC specific objectives	·Contribution of Fiscalis actions to the achievement of DAC specific objectives
	# Extent to which the Directive contributed to safeguard Member States’ tax revenues	·Additional tax assessed and/or collected as a result of the provisions of the Directive (by provision) ·Stakeholders’ perception on the extent to which the Directive contributed to safeguard Member States tax revenues
	# Extent to which the Directive contributed to the smooth functioning of the Internal Market and to a perceived fairer tax system	·Increased perceived fairness of the tax system ·Increased perceived functioning of the Internal Market

Evaluation questions matrix – Coherence

Question

Sub-questions

Indicators

# To what extent was the intervention successful and why

# Extent to which the Directive’s provision are consistent with other related EU policies

·Existence of conflicting or synergies

·Existence of gaps and overlaps

# Extent to which the Directive’s provision are consistent with OECD other international policies

·Existence of conflicts or synergies

·Existence of gaps and overlaps

·Existence of operational and technical inconsistencies

Evaluation questions matrix – EU added value

Question

Sub-questions

Indicators

# How did the EU intervention make a difference and to whom

# Extent to which the results generated are additional compared to national/international initiatives

·Additionality of the Directive with respect to effectiveness

·Additionality of the Directive in guaranteeing a homogeneous implementation of international provisions

·Additionality of EU financial support

# Extent to which the efficiency of the tools and mechanisms of the Directive is superior to possible alternatives

·Stakeholders’ perception of the advantages of having EU-wide systems, procedures, and tools compared to national or international alternatives

·Additional costs and cost savings generated by the systems and provisions connected to the Directive

Evaluation questions matrix – Relevance

Question

Sub-questions

Indicators

# Is the intervention still relevant

# Extent to which the objectives of the Directive and of its amendments are aligned to Member States needs

·Absolute magnitude of the issues at stake (e.g. estimates of incomes and assets generated in Member States other than the one of residence; estimates of cross-border tax evasion and avoidance)

·Relative magnitude of the issue at stake (e.g. ratio between country nationals’ income and incomes exchanged under DAC1 AEOI; comparison between DAC4 exchanges and MNE headquarters in Europe)

·Stakeholder perception of the alignment between the Directive objectives and the needs of EU tax systems

# Extent to which the tools and provisions of the Directive are appropriate and in line with its objectives

·Stakeholder perception of the fitness of the Directive to its objectives

·Stakeholder perception of the appropriateness of the main features of the Directive (scope, definitions, etc.)

·Evidence of issues in the design of the tools and mechanisms of the Directive

Annex IV. Overview of benefits and costs

Methodologies used in the cost-benefit analysis

Details on estimated costs for private stakeholders – Source: Ramboll Study

18 financial institutions or federations, via interviews and the survey carried out by Ramboll for the study provided elements to quantify the administrative burdens from DAC2. These include data from 6 very large institutions, as well as from 9 small and medium financial institutions, originating from six Member States across the various EU regions. In addition, to the data provided by single institutions, three federations provided aggregated or specific data about their members or whole Member States. Data include the number of accounts and clients reported, for both new and existing customers, and cover about 2.4 million reportable accounts (that is more than 10% of the total). Financial institutions and their federations were asked to estimate the additional time spent for onboarding clients, the set-up costs, as well as the ongoing costs. Between 11 and 13 institutions could provide full or partial estimates for the various cost categories. Eurostat data on earning statistics per occupation have been used to monetise the time spent by the internal personnel.

Data on the population were estimated based on the 2022 Facts and Figures report by the European Banking Federation and the European Central Bank statistical warehouse. The population includes all EU banks (5,263 in 2021) and an estimate of other financial institutions covered by DAC2. As for other financial institutions (e.g. custodian banks, certain investment entities, specialised insurance companies), no detailed information exists on the number of those covered by DAC2 reporting duties.

It has been considered that this population corresponds to half of the bank population (i.e. about 2,600 other financial institutions). Significant difference in cost estimates exist within the population, which is hence segmented between large banks on one side, and medium and small banks and other financial institutions on the other side. The banks directly supervised by the ECB (99) are considered to be large banks, while the rest is considered to be medium or small banks. Within these two groups, cost estimates are consistent.

For the quantification of administrative burdens from DAC4, data have been provided by 11 entities, including 4 business federations (and members thereof) and 7 MNEs, via interviews and the survey carried out by Ramboll for the study. Data include the set-up and recurring costs incurred by the companies, distinguishing among IT costs, external consultants’ fees, personnel costs and training costs.

The population of MNEs concerned can be inferred from DAC4 statistics, and it is estimated at 4,750. As it is often the case for obligations consisting of or requiring the collection of data from the MNEs’ internal systems, costs per company vary between large-scale and small-scale entities. Large-scale MNEs are very complex entities, with a significant presence in more than 5 Member States and a turnover in excess of EUR 10 billion; they are estimated to represent 10% of MNEs. Small-scale SMEs have a lower turnover and operate in fewer Member States; they represent the vast majority (90%) of MNEs.

Costs also depend on the extent to which CbCR data could be extracted from the internal systems of the companies or not. In the former case, initial compliance efforts are limited to familiarisation and mapping. Otherwise, the efforts could become significant, as they may require adaptation of the accounting and enterprise resource planning systems (with the ensuing IT costs), as well as additional efforts to collect and produce the information required. Based on the targeted consultation, the set-up efforts were limited for the majority of companies, since all or most of the information was already available or could be easily adjusted to match CbCR requirements. Hence, it is assumed that the process was more complex for 20% of MNEs (uniformly across both large- and small-scale entities) reporting, data handling and processing, as well as of obtaining additional data from customers compared to what is necessary for business purposes (or for compliance with other regulation).

Estimation of tax benefits from DAC1 and DAC2 AEOI, as well as those from non-AEOI activities – Source: Ramboll Study

First, data on tax benefits reported in the Yearly Assessment and CACT statistics were retrieved from the annual submissions. Data were first checked for consistency. Consistency checks consisted in: (i) identifying outliers, i.e. countries that generated a disproportionate amount of benefits for a given tool in a given year; (ii) identifying geographical outliers, i.e. countries which generated an amount of benefits significantly higher controlling for their economic size (in terms of tax revenues, GDP); (iii) identifying break in the time series, i.e., for any given country, years in which tax benefits where significantly different from the average of the period. Outliers were discussed with tax authorities during the targeted consultations carried out by the contractor in charge of the study, clarifying whether they resulted from mistakes in the calculation of benefits, compilation of the reporting tools, or reflected their actual estimates. When no sufficient information could be found to justify outliers, they were removed from the calculation or replaced with average values reflecting normal trends. When the above steps were completed, reported tax benefits, i.e. the benefits as reported by the national tax authorities, were estimated.

Consequently, for each DAC tool, the data coverage was then determined. This required considering the number of reporting Member States, as well as the weighted EU coverage. Weights per each country were calculated by considering GDP in PPS, direct tax revenue (as a proxy of economic size), and operating expenditures by tax administrations (as a proxy of compliance effort). Based on these weights, estimated EU tax benefits were extrapolated. Estimates are considered partly robust when the data coverage is higher than 25%. However, since this remains a low threshold, EU benefits are always presented as ranges. When data series do not achieve such a coverage, the extrapolation is not performed. When data series do not achieve such a coverage for a number of years, the extrapolation is presented as tentative estimates for these years. Based on the above, no disaggregation of estimates (e.g. in terms of DAC1 incomes, between ex-ante and ex-post activities, or per non-AEOI tools) is possible.

The following methodology applies to the estimates of tax benefits from DAC4 AEOI: given the nature of data exchanged under DAC3 and DAC4, tax authorities use them in their tax risk assessment analysis but do not link the data directly with tax benefits. The collection of such data was attempted between 2018 and 2020, and then discontinued, since no Member State could provide estimates. This is due to the more tenuous link between the information received and tax assessments. Indeed, the information from DAC3 and DAC4 do not directly lead to the identification of unreported taxable basis. Rather, they typically help identifying taxpayers at risk or suspicious behaviours and are one of the sources that contribute to risk analysis. Also, it is more difficult to attribute the benefits, in terms of additional taxes assessed for or reported by MNEs, to each of the various tools, i.e., DAC3, DAC4 and, in perspective, DAC6. Estimates on the impact of public CbCR can be retrieved from the economic literature.

The parameters from this analysis have been adapted to estimate the impacts of DAC4. DAC4 do not require tax authorities or MNEs to publicly disclose their CbCR (differently from the requirements for large banks or the extractive industries). Hence, the likely effect of DAC4 is a fraction of those estimated for public CbCR. This fraction is captured by the conversion factor. The central value of the conversion factor is 0.33, meaning that the effect of DAC4 is one third of that of public CbCR. Given the uncertainty on the magnitude of the effect of DAC4 compared to public CbCR, a sensitivity analysis of the conversion factor is carried out, and benefits are presented in a range. The sensitivity values are set at 0.25 (one quarter) and 0.5 (half). This corresponds to an increase of the MNEs’ effective tax rate of 1.0%, with sensitivity values at 0.7% and 1.5%.

Cost vs Benefits (net benefit) – Source: Ramboll Study

The estimate of benefits is subject to larger uncertainties. On the one side, only a minority of Member States is able to track and monitor the tax benefits from DAC, and the accuracy of the data could not be independently verified. Also, the largest share of benefits is attributed to DAC4, based on a conservative application of the impacts of public CbCR, but as discussed above, a number of caveats apply. Considering these limitations, total DAC benefits fall in the EUR 5 – 10 billion range, with a central estimate of EUR 6.8 billion. Considering only tax benefit estimates based on data from tax authorities (DAC1, DAC2 and non-AEOI), DAC benefits amount to about EUR 1.2 billion.

Figure 24: Benefits of DAC (EUR million) – Estimation with central estimate / low range – high range approach

Benefits per DAC	Central estimate	Low Range	High Range
DAC1	240	50	390
DAC2	880	680	1,070
DAC4	5,600	4,200	8,230
Non-AEOI	60	10	110
Total Benefits	6,780	4,940	9,800

Note. Non-AEOI benefits include those due to EOIR, PAOE and simultaneous controls. Source: Ramboll study

Based on this analysis, the recurring burdens due to DAC are more than commensurate with the benefits generated. Under the various sensitivity scenarios (see below), the annual net benefits are in between 500 and 6,100 billion. Net benefits remain positive and significant even considering the uncertainties linked to the quantification of DAC4 impacts. Even under the most conservative estimates, i.e. those carried out based on primary data on tax benefits provided by a growing number of Member States, benefits are at least two times the burdens incurred. In the other scenarios, they are between eight to 10 times larger than the costs incurred.

Figure 25: Net benefits and cost-effectiveness ratio

Scenario	Costs (€Mln)	Benefits(€Mln)	Annual net benefit (Benefits – Costs) (€Mln)	Cost-Effectiveness Ratio (Benefits/Costs)
Comprehensive scenario	650	6780	6130	10,43	~10
Low-range scenario	650	4940	4290	7,6	~8
Data-based scenario	650	1180	530	1,82	~2

Cost and benefit figures rounded up to the tens of millions. Costs include recurring adjustment costs and administrative burdens from DAC provisions for which benefits could be quantified. These estimates do not represent net costs, as the associated cost savings could not be quantified.

Sensitivity scenarios for cost-effectiveness analysis – Source: Ramboll Study

Three scenarios are considered for the cost-effectiveness analysis:

·The comprehensive scenario, based on the central estimates of all costs and benefits;

·The low-range scenario, based on the low-range estimates of all costs and benefits;

·The data-based scenario, that consider only data collected from stakeholders and tax authorities and excludes the literature-based estimates of DAC4 benefits.

As shown above, the cost-effectiveness analysis is robust across the three scenarios:

·In the comprehensive scenario, benefits are significantly higher than costs, with a cost-effectiveness ratio of about 10 and annual net benefits of about 6.1 billion.

·In the low-range scenario, benefits remain higher than costs, with a cost-effectiveness ratio of about 8 and annual net benefits of about 4.3 billion.

·In the data-based scenario, DAC4 benefits are excluded. In this case, net benefits amount to EUR 500 million, with a cost-effectiveness ratio of about 2.

Table 1. Overview of costs and benefits identified in the evaluation

Citizens/Consumers

Businesses

Administrations

[Other…] _ specify

Quantitative

Comment

Quantitative

Comment

Quantitative

Comment

Quantitative

Comment

Costs:

Direct compliance costs (adjustment costs, administrative costs, regulatory charges)

Type: recurrent

Irrelevant

EUR 604 million

The costs presented are based on projections using limited data shared mainly by financial institutions. These relate almost exclusively to mandatory due diligence procedures to get information from customers of these financial institutions within the framework of the common reporting standards.

This includes thousands of entities across the 27 Member States, but no official consolidated figures were found in the public sources consulted.

The costs are electronic infrastructure costs, but also human resources costs for tracking files. However, there is no solid documentation from reporting third parties illustrating the breakdown of these costs.

The costs of other exchanges are relatively limited.

EUR 42 million

The costs presented are better monitored by administrations. They do not follow the precise nomenclature proposed by the Commission in its assessments.

The costs of developing and maintaining the necessary IT architectures are particularly monitored. This leads to the following observation: while development costs are particularly significant for administrations, they tend to decrease over time, with recurrent costs being relatively limited.

Irrelevant

Benefits:

Direct benefits (such as improved well being: changes in pollution levels, safety, health, employment; market efficiency)

Indirect benefits (such as wider economic benefits, macroeconomic benefits, social impacts, environmental impacts)

Between EUR 1180 -6700 million

Under the various sensitivity scenarios, the annual net benefits are in a range between 500 and 6,100 billion

TABLE 2: Simplification and burden reduction (savings already achieved)

Report any simplification, burden reduction and cost savings achieved already by the intervention evaluated, including the points of comparison/ where available (e.g. REFIT savings predicted in the IA or other sources).

Citizens/Consumers/Workers

Businesses

Administrations

[Other…] _ specify

Quantitative

Comment

Quantitative

Comment

Quantitative

Comment

Quantitative

Comment

Title

Type: One-off / recurrent (select)

N/A

PART II: II Potential simplification and burden reduction (savings)

Identify further potential simplification and savings that could be achieved with a view to make the initiative more effective and efficient without prejudice to its policy objectives 73 .

Citizens/Consumers/Workers

Businesses

Administrations

[Other…] _ specify

Quantitative

Comment

Quantitative

Comment

Quantitative

Comment

Quantitative

Comment

Description:

Type: One-off / recurrent (select)

N/A

Annex V. Stakeholders consultation - Synopsis report

Timing of consultation activities

Date	Actions
December 2022 – December 2023	External study on the evaluation of the DAC for the period 2018 – 2022
May 2024 – July 2024	Call for evidence and consultations: -call for evidence (feedback online via “Have your say” platform) -open public consultation (questionnaire online via “Have your say” platform) -questionnaire on the functioning of the DAC to Member States -questionnaire on the relevance of DAC6 Annex IV to Member States
September 2024	Working Party IV on the evaluation of the DAC for the period 2018 – 2022, and DAC6 Annex IV
September 2024	Targeted consultations of academics, lawyers and economists
December 2024	Discussions in the Platform for tax good governance

The stakeholder consultation activities carried out by the Commission, as shown in the table above, involved various actions and the engagement of multiple stakeholders to cover the widest possible range of interests, priorities, and experiences. These activities aimed to gather additional information to complement what the Commission already possessed regarding the DAC for the period 2018 – 2022, excluding the most recent amendments of DAC7 and DAC8.

In the following paragraphs, key highlights of each action will be outlined. However, it should be noted that the contributions received during the stakeholder consultation, as described hereafter, cannot be considered the official position of the Commission and its services. Therefore, these contributions do not bind the Commission, nor can they be regarded as a representative sample of the EU population.

Stakeholders identified

·Tax authorities: EU tax authorities are the primary entities responsible for exchanging information under the DAC. They also play a significant role in policy development. The evaluation not only considered the general opinions of EU tax authorities on each evaluation criterion but also gathered detailed input on specific elements of the DAC and potential improvements. This input was further refined through tax authorities’ participation in various fora including regular meetings with experts and the Commission 74 , and expert project groups supported by the Fiscalis programme (e.g., the VISDAC project, which identified best practices to help Member States enhance data matching through an ongoing process, even when the data was not initially identified).

·Economic operators: the exchange of information system involves thousands of economic operators who must comply with DAC rules by submitting information to tax authorities. These operators directly experience whether the proposed amendments to the DAC achieve the desired effects or whether changes to the rules have unintended negative impacts. The sectors covered include the ones impacted by the DAC provisions, such as the financial institutions (DAC2), MNEs (DAC4, DAC6), intermediaries, especially lawyers and tax advisers (DAC6) and digital platforms (DAC7). Their interests vary significantly depending on the stakeholders who are concerned by specific reporting obligations, and size of business (large firms or SMEs). Representative organisations also play a crucial role due to their specialized knowledge of the technical issues at stake.

·Commission services: officials from DG TAXUD and other relevant DGs were invited to participate in an inter-service steering group (ISSG), which provided ongoing feedback on the study and facilitated consultation activities. Since the Commission played a key role in steering the study, its views are not summarized or presented in this synopsis report.

·EU citizens: EU citizens were also identified as relevant stakeholders and were invited to provide feedback through the “Have Your Say” platform on the Commission’s website. However, very few citizens participated, making it impossible to draw any general conclusions from this group.

Consultation methods and tools

Consultations with stakeholders are just one part of the overall evaluation methodology and tools described in Annex II.

Stakeholders were consulted through several complementary methods designed to balance the breadth and representativeness of different interests and priorities with the need for highly detailed information.

The main tools used during the consultation activities included interviews and online questionnaires. In particular, questionnaires were sent to all EU tax authorities and made available to economic operators on the Commission’s website through the “Have your say” platform. Among economic operators, a diverse range of sectors and profiles were represented, covering industry groups, business associations, tax firms and individual companies.

This comprehensive approach enhances confidence in the findings. However, it should be noted that most stakeholder input is not based on a statistically representative sample.

External study on the evaluation of the DAC for the period 2018 – 2022

A study on the evaluation of the DAC for the period 2018 – 2022 was conducted by an external contractor from December 2022 to December 2023.

In particular, a total of 162 stakeholders participated in the consultation by the contractor:

-16 during the familiarization interviews;

-75 in interviews during the targeted consultation (including both private stakeholders and tax authorities);

-71 in the online survey.

The targeted consultation covered various EU regions through a sampling strategy. The sample included a “core” group of 8 Member States 75 , in which the external contractor consulted both local tax authorities and private stakeholders, and an “extended” group of 6 additional Member States 76 , where interviews were conducted only with local tax authorities.

As for the private stakeholders, the categories concerned reflected the sectors covered by the various DAC provisions, namely:

-financial institutions, mostly relevant for DAC2;

-MNEs and large enterprises, primarily affected by DAC provisions on aggressive tax planning, specifically DAC3/4/6;

-intermediaries, relevant to DAC overall in relation to its impact on compliance and taxpayer behaviour.

Call for evidence and public consultations

As part of the evaluation and to reach a broader public audience, the Commission conducted a call for evidence and an open public consultation (OPC) 77 via the “Have your Say” platform on its website, running from 7 May 2024 to 30 July 2024.

In particular, the call for evidence allowed stakeholders to submit a written document in a free format, while the OPC consisted in the submission of a close-ended questionnaire structured in the following 3 main parts:

-Part 1 – overall assessment of DAC;

-Part 2 – foreign income and assets;

-Part 3 – tax transparency.

A total of 55 stakeholders participated in the consultation process. More specifically, 28 stakeholders and 2 citizens answered the OPC questionnaire focusing on AEOI that addressed the different instruments provided for in the DAC, while 25 stakeholders submitted their written contributions to the call for evidence 78 . 24 respondents identified themselves as “business associations” representing specific economic interests, whether in the legal and accountancy professions or in other areas of activity; 14 identified as “company or business”; 10 identified as “other” including tax intermediaries, accountants, or advisors. 7 participants identified themselves as 3 EU citizens, 2 NGOs and 2 academic research and institutions.

Figure 26. Distribution by categories (OPC and call for evidence)

The distribution by category of activity shows that tax intermediaries, accountants or consultants are the most common category of respondents, with the highest representation (23) followed by large companies and businesses (10), banks, and other financial services (10).

Figure 27. Distribution by categories and area of activity (OPC and call for evidence)

In terms of geographic distribution, Belgium had the highest participation rate, with 13 respondents. One explanation for this is the concentration of business associations and lobby groups based in Brussels. 9 respondents originated in France, while Germany, the Netherlands and Ireland were all represented by 5 respondents each. Other EU countries (Austria, Czechia, Finland, Luxembourg, Malta, Sweden) are represented by smaller percentages of respondents.

Questionnaire on the functioning of the DAC and the relevance of Annex IV of DAC6 to Member States

From May to July 2024, Member States were asked to respond to an updated questionnaire and invited to assess different aspects of Annex IV concerning the concept of the main benefit test (MBT) and the hallmarks provided for in DAC6, particularly in terms of clarity, practical application, and efficiency. In addition, Member States were given the opportunity to suggest new categories of hallmarks if they had observed relevant cross-border arrangements that are not currently covered in Annex IV since the adoption of DAC6.

A total of 26 Member States provided their feedback.

Working Party IV on the evaluation of the DAC and DAC6 Annex IV

On 21 September 2024, during a Working Party IV meeting, the Commission services presented to Member States the outcome of the external study supporting the DAC Evaluation, as well as the results of the OPC, the call for evidence, and the responses to the 2 questionnaires sent to Member States. The meeting allowed the Commission services to gather Member States’ comments on the results and address a few other points of interest for the Commission.

The analysis of the responses was presented in detail, highlighting the elements of Annex IV that some Member States believe should be improved.

This analysis was supplemented by a presentation on the CJEU judgment on DAC6 (Case C-623/22) to inform Member States of the potential consequences of this decision.

Targeted consultations

In September 2024, the Commission conducted a few additional targeted consultations with academics, tax lawyers and economists to gather their comments on specific aspects of the DAC, in order to complement the analysis on its effectiveness, efficiency, relevance, coherence, and EU added value.

A questionnaire 79 was distributed beforehand and during the discussions the experts provided their views and comments on the points of interests of the Commission.

Meeting with academics and tax law experts

On 25 September 2024, the Commission met with the following experts: Daniel Gutmann, Co-director of the Department of Sorbonne Fiscalité Finances Publiques, École de droit de la Sorbonne (France), avocat au Barreau des Hauts-de-Seine; and Denis Emmanuel Philippe, Lecturer at UL Liège; avocat au Barreau de Bruxelles.

Meeting with economists

On 26 September 2024, the Commission met with the following experts: Panayiotis Nicolaides, Director of Research at the EU Tax Observatory; and Jeanne Bomare, researcher at the EU Tax Observatory.

Discussions in the Platform for tax good governance

On 17 December 2024, the Commission held the meeting 80 of the Platform for tax good governance with the participation of representatives from different EU business and professional associations. The aim of the meeting, among others, was to discuss and gather stakeholders’ views on the output of the DAC evaluation and the open public consultation.

Results of consultation activities

A summary of the consultation activities is presented below, focusing on the main criteria assessed for the evaluation: effectiveness, efficiency, coherence, relevance and EU added value.

·Efficiency

Under the efficiency criterion, the study estimates both set-up and recurring adjustment costs and burdens for private and public stakeholders. Annual recurring adjustment costs and administrative burdens from DAC are estimated at approximately EUR 650 million.

From a distributional perspective, the majority (93%) of these burdens fall on private stakeholders. In 2018–2019, before the implementation of DAC6, recurring administrative burdens hovered around EUR 550 million. Following the introduction of DAC6, they increased to EUR 800–900 million. The costliest provision during this period was DAC2; however, costs tend to decrease over time. Estimates indicate that DAC6 was also relatively expensive in the reference period, mainly due to the initial compliance efforts. As for DAC4, it accounted for approximately 6% of total costs over the past 2 years.

Recurring costs and burdens for tax authorities represent less than 7% of total costs. DAC1 and DAC2 generate the highest costs, while DAC3 is less costly, with DAC4 falling somewhere in between. Overall, the findings of the previous evaluation have been confirmed: implementation costs are significant when new DACs are introduced but tend to decrease and stabilize over time, whereas recurring costs increase. The costs associated with non-AEOI provisions are relatively low, primarily due to their limited use, amounting to around EUR 1 million per year. These costs have further declined since 2020 due to reduced reliance on advanced cooperation mechanisms (PAOE and simultaneous controls), partly as a result of the COVID-19 pandemic. Finally, reporting costs for Member States to the Commission amount to approximately EUR 140,000 per year, half of which is allocated to compiling the Yearly Assessment.

Given the uncertainties in estimating benefits, the cost-benefit comparison is conducted across 3 scenarios:

-Comprehensive scenario: Based on the central estimates of all costs and benefits.

-Low-range scenario: Based on the lower-end estimates of all costs and benefits.

-Data-based scenario: Based only on data collected from stakeholders and tax authorities, excluding literature-based estimates of DAC4 benefits.

According to the cost-effectiveness analysis, the recurring burdens associated with DAC are more than justified by the benefits generated. Across the different sensitivity scenarios, annual net benefits range between EUR 500 million and EUR 6.1 billion. Even when accounting for uncertainties related to DAC4’s impact, net benefits remain positive and substantial. Under the most conservative estimates—based solely on primary data from tax authorities—the benefits are at least twice the incurred burdens. In other scenarios, benefits exceed costs by a factor of 8 to 10.

·Effectiveness

Tax authorities largely agree that the framework established by the Directive facilitates administrative cooperation and appreciate the operational tools and systems it entails. They also confirm that the interaction among different provisions and tools is one of the Directive’s strengths.

The assessment of the Directive’s specific objectives begins with an analysis of the usability of exchanged information in terms of completeness, quality, and timeliness. DAC1 and DAC2 exchanges demonstrate a high degree of completeness, as the information shared includes detailed and granular data on both content and identifiers. The only notable exception is the inclusion of the TIN of the receiving country in DAC1 exchanges, which remains very low. Regarding DAC3 and DAC4, tax authorities have raised concerns about the varying quality of exchanged information. In particular, for DAC4, not all Member States are able to ensure that CbCRs are collected from local MNEs, which may affect the completeness of the data. No significant issues have emerged regarding the timeliness of exchanges.

Between 23 and 25 countries systematically match DAC1 and DAC2 data, and most also match DAC3, DAC4, and DAC6 information. Matching rates are relatively high for DAC1 and DAC2, typically covering about 85–90% of received information, marking an increase from the previous evaluation, particularly in terms of automatic matching. Matching rates are also high and increasing for DAC3 and DAC4, and matching worked well for DAC6 in its first year. Once matched, tax authorities use the information for various purposes. In recent years, all tax authorities have reported using DAC1 and DAC2 information for tax control purposes, particularly for risk assessments and audits. DAC3 and DAC4 information is also used by most Member States, although not systematically by all. Non-AEOI tools are also considered valuable for identifying tax fraud and evasion, particularly EOIR.

AEOI, particularly DAC1 and DAC2, is also used to promote spontaneous tax compliance through notifications, disclosures, and awareness campaigns. The evaluation also examines the extent to which DAC has influenced taxpayer behaviour. Feedback suggests that DAC has had a deterrent effect on individual taxpayers, although quantifying this impact with precision remains challenging. Specifically, for DAC2, tax authorities report an increase in the disclosure of financial assets abroad and the taxes paid on them. However, the deterrent effect on corporations is perceived by stakeholders as less significant so far. The impact of DAC3 and DAC4 is often seen as part of broader international tax transparency efforts. Finally, DAC - particularly DAC3 and, to a lesser extent, DAC4 and DAC6 - has the potential to enhance tax rule transparency.

The Directive targets 3 interrelated general objectives. The first is safeguarding Member States’ tax revenues by increasing tax collection or recovery from taxpayers with income, assets, or operations in multiple Member States - either through tax control measures or increased voluntary tax compliance. Despite the limitations of quantitative analysis, available evidence, supported by tax authorities and private stakeholders, suggests that the Directive effectively contributes to safeguarding tax revenues. Total tax benefits are estimated at approximately EUR 6.8 billion per year. Although this figure is significant in absolute terms, it corresponds to only about 0.3% of EU direct tax revenue. DAC1 and DAC2 are widely credited with uncovering hidden tax bases and improving compliance through enforcement and deterrence. Together with non-AEOI tools, these mechanisms generate approximately EUR 1.2 billion per year in additional tax assessments, based on data from tax authorities. While no data are available for DAC4, economic research indicates that CbCRs can increase MNEs’ effective tax rates. Although this effect is likely smaller for private CbCRs than for public ones, the scale of the issue suggests benefits in the range of several billion euros.

The Directive also pursues 2 additional objectives: contributing to the smooth functioning of the Internal Market and increasing the fairness of the EU tax system. Both objectives are linked to the Directive’s role in safeguarding tax revenues. DAC enhances the Internal Market by limiting opportunities for tax evasion and avoidance among cross-border companies and individuals, thereby reducing tax distortions. Additionally, it has helped harmonize reporting requirements for private sector operators, particularly under DAC2, DAC4, and – prospectively - DAC7. DAC also enhances the fairness of both EU and national tax systems. Although this impact is difficult to measure, there is a consensus, supported by evidence, that DAC has reduced opportunities for individuals and corporations to conceal tax bases abroad and avoid fair taxation. However, DAC alone cannot resolve all tax-related challenges. If certain tax practices or loopholes are not addressed by substantive tax laws, information exchange alone cannot fully mitigate the issue, regardless of the volume of data exchanged. In this context, DAC functions as one of several tools, alongside other EU and international initiatives, that contribute to ensuring a fairer tax system and appropriate revenue distribution among Member States.

·Coherence

DAC interacts with several EU and non-EU legal provisions. The Directive and its provisions align with other relevant EU initiatives, particularly those that establish similar instruments for administrative cooperation in different tax areas (e.g., VAT, excises, and asset recovery). However, minor inconsistencies exist, primarily due to the differing objectives of other EU directives, such as the AML Directive and the Anti-Tax Avoidance Directive (ATAD). Nonetheless, these differences have not resulted in serious conflicts. Instead, significant synergies exist, such as between DAC2 and the AMLD.

In some cases, the legal framework is criticised for appearing fragmented. For instance, digital platforms are subject to several existing and upcoming reporting requirements (e.g., Article 242a of the VAT Directive and the Central Electronic System of Payment Information). Certain datapoints required under these frameworks may overlap with those in DAC7, potentially leading to duplicated efforts and increased implementation burdens. This occurs despite the distinct aims and scopes of the legislations - while DAC7 enables Member States to assess platform users’ income for direct taxation purposes, the VAT Directive and CESOP are primarily designed to combat VAT fraud. Additionally, the introduction of national reporting rules may contribute to inconsistencies in the implementation of EU Directives, creating multiple layers of obligations.

With regard to international initiatives, DAC also demonstrates a relatively high degree of coherence. In most cases, DAC provisions were adopted following corresponding OECD initiatives, incorporating definitions and operational processes directly. The best examples include:

-DAC2 / Common Reporting Standard (CRS);

-DAC4 / Base Erosion and Profit Shifting (BEPS) Action 13 and Country-by-Country Reporting (CbCR);

-DAC7 / Model Reporting Rules for Digital Platforms (MRDP);

-DAC8 / Crypto-Asset Reporting Framework (CARF).

In other instances, EU provisions diverge from OECD model rules, though to varying degrees. For example, DAC3 differs slightly, whereas DAC6 presents more significant differences. In the case of DAC6, the lack of harmonization at the EU level, combined with deviations from the OECD Mandatory Disclosure Rules (MDR), has led to uncertainty among private stakeholders and inconsistencies in national transposition.

·Relevance

The Directive aims to address 2 main challenges: the erosion of Member States’ tax bases due to increased globalization and the risk of an unharmonized policy response. Public and private stakeholders agree that these issues pose a problem for society.

The problems covered by DAC provisions are not all equally significant in quantitative terms. Financial assets held by EU citizens in Member States other than their country of residence are substantial: deposits in financial institutions and investments in other EU countries amount to 9% and 18%, respectively, of the gross financial assets held by EU residents. Profit shifting by multinational enterprises (MNEs) is also a major phenomenon. In 2018, MNEs with a parent jurisdiction in the EU accumulated nearly EUR 1,600 billion in profit before income taxes. Estimates suggest that close to EUR 150 billion - approximately one-tenth of total profits and 14% of foreign profits - were recorded in parent jurisdictions classified as offshore financial centres. According to the economics from the targeted consultations, these numbers are even bigger. As a matter of fact, 7 years after the launch of the BEPS actions and the introduction of CbCR, they have evidence that in 2022 the profits being shifted to offshore financial centres is of approximately USD 1 trillion, equivalent of 35% of profits booked by MNEs outside their headquarters’ jurisdiction. Corporate income tax losses remain significant, reaching up to 10% of global corporate tax revenues. However, economists provided evidence that without these measures, the situation would likely be worse, with even higher levels of BEPS activity.

A broader concern relates to the use of information exchanged under DAC. While Member States have provided proof or evidence of increasing data usage, stakeholders continue to express concerns about the lack of transparency regarding how Member States utilize the data. This discrepancy suggests that while DAC information exchange remains relevant, greater publicity and monitoring of its use are needed.

Lastly, a key issue with DAC4 is the ongoing difficulty tax authorities face in correctly identifying and monitoring reporting entities.

In contrast, the relative significance of the categories of income covered by DAC1 appears much lower. Cross-border employees account for between 1% and 4% of the total workforce, with their income representing an average of 2% of total workforce compensation. Similarly, cross-border pensioners constitute 2% of the total pensioner population, and the amount paid to them represents 1.5% of total pension payments.

Stakeholders have a generally positive view of the Directive’s fitness for purpose. The fitness analysis has been conducted for the various DAC tools, with very few issues have been reported concerning non-AEOI tools. However, regarding the AEOI, the most discussed issue is the DAC6 reporting requirements. According to both public and private stakeholders, several aspects of DAC6 are neither clearly defined nor easily applicable. The lack of clarity, exacerbated by the absence of guidelines endorsed by all Member States, allows for different interpretations and applications at the national level. On one hand, given that the provision was only recently introduced, some difficulties may be considered typical implementation challenges. On the other hand, some provisions have emerged as problematic in themselves, particularly concerning the definition of certain hallmarks, the main benefit test, and the classification of intermediaries subject to reporting.

Another issue relates to the difficulties financial institutions face in collecting and verifying clients’ foreign TINs for DAC2, particularly for pre-existing clients. More generally, identifying the tax residency of taxpayers remains challenging for both DAC1 and DAC2, as taxpayers may struggle to report it correctly. Regarding DAC3, issues persist with the summaries uploaded to the Central Directory, which, even when properly compiled, are often insufficient for risk assessment. To address this, a new template has been agreed upon with Member States and will be available as of 2024.

Lastly, a key issue with DAC4 is the ongoing difficulty tax authorities face in correctly identifying and monitoring reporting entities.

·EU added value

The assessment of the EU added value evaluates the extent to which the results achieved under DAC are additional to what would have occurred without the Directive. Specifically, it examines whether the same level of cooperation could have been achieved solely through bilateral agreements and existing OECD standards and actions.

All tax authorities agree that DAC is superior to international or national alternatives due to its binding nature under EU law, its underlying IT infrastructure, and its uniform reporting requirements across the EU. This enables larger-scale exchanges compared to international counterparts. Additionally, EU funding - through the Fiscalis program - has played a crucial role in supporting the development of IT systems and facilitating more resource-intensive forms of administrative cooperation, such as PAOE and simultaneous controls.

Private stakeholders share a similar view. However, there are some reservations where implementation is not fully aligned among Member States, which could lead to fragmentation and, consequently, higher compliance costs (e.g., DAC6).

Annex VI. The Intervention Logic Diagram

Annex VII. Legal developments from 2018

Transposition: overview of infringement procedures

As summarized in the table below, during the period of the present evaluation, the DAC transposition process has led to a total of, respectively, 88 infringement procedures for non-communication, and 8 infringement procedures for non- or partial compliance. At the time of publication of this document, all infringement procedures have been closed, except for 5 related to DAC7. Notably, none of these cases has (yet) been referred to the Court.

Figure 28. Overview of infringement procedures

Source: Ramboll study and DG TAXUD data

Although most non-communication cases were closed shortly after the procedures were opened, delays in communicating successful transposition often reflect actual delays in adopting national laws. In particular, the majority of cases (15) involved late notification or non-notification by Member States regarding DAC6 transposition.

This pattern reflects the fact that some Member States struggle to keep pace with frequent Directive amendments and face technical challenges linked to the expansion of AEOI, further exacerbated by resource constraints.

CJUE Cases on DAC

·Legal professional privilege

There have been CJUE Cases concerning legal professional privilege, mostly from the point of view of DAC6 reporting requirements.

The Court has first discussed the issue in the case C-694/20 81 and ruled against the obligation set out in Article 8ab(5) of DAC for a lawyer acting as intermediary and exempt from the reporting obligation to notify without delay any other intermediary who is not their client of that intermediary’s reporting obligations. The Court held that this provision should be invalid in the light of Article 7 of the Charter of Fundamental Rights of the European Union, in so far as the Member States’ application of that provision had the effect of requiring a lawyer acting as an intermediary to breach the confidentiality obligations to which it was legally bound.

This was later addressed in legislation, and an amendment was adopted by the Council in the context of DAC8 to ensure that the notification system initially provided for in the Directive would be modified according to the decision of the Court. It introduced the concept of lawyer’s “client”, who is entitled to receive a notification from a lawyer bound by legal professional privilege, thereby making Article 8ab(5) compatible with the Charter 82 .

A second case, where legal professional privilege was discussed was case C-623/22 83 , where the Court ruled that the invalidity of Article 8ab(5) declared by the Court in the ruling of 8 December 2022, applies only to persons who pursue their professional activities under one of the professional titles referred to in Article 1(2)(a) of Directive 98/5 84 . It confirmed the previous ruling and specified that the privilege only applies to lawyers authorised under national law to ensure legal representation of their clients.

The third time the Court was faced with the question of the extent of legal professional privilege in the context of DAC was case C-432/23, which did not relate to DAC6 but rather to exchange of information on request (EOIR). This judgment reconfirmed that lawyers enjoy strengthened protections of communication with their clients under Article 7 of the Charter of Fundamental Rights of the European Union, also regarding documentation that may be requested under DAC by national tax authorities.

The Court did not find any inconsistencies between the DAC and the Charter of Fundamental Rights but stressed that a decision requiring a lawyer to provide the authorities of the requested Member State – for the purposes of an exchange of information on request in accordance with DAC – with all the documentation and information relating to his or her relations with his or her client, concerning such legal advice, constitutes an interference with the right to respect for communications between lawyers and their clients guaranteed by Article 7 of the Charter.

·Infringement of equal treatment and legal certainty

In its ruling of 29 July 2024 (C-623/22), the court also ruled on issues other than professional secrecy and legal professional privilege. In practice, the questions of i) possible infringement of equal treatment (in so far as DAC6 does not limit the reporting obligation to corporation tax, but makes it applicable to all taxes falling within its scope) and ii) possible infringement of legal certainty (due to the terminology and concepts used in DAC6 e.g., “arrangement”, “intermediary”, “participant”, and starting point prescribed for fulfilling reporting obligations) were examined by the Court. On the first point, the CJEU found that the different tax types subject to the reporting obligations fall within comparable situations in the light of the objectives of the DAC, and the legislation was not inappropriate. On the second point, by analysing the various aspects of the concepts used in DAC6, the Court held that the degree of precision and clarity did not call into question the validity of DAC6.

·Conclusions

While the Court has generally found the system to be coherent with fundamental rights, it has overturned one provision in order to provide greater protection for legal professional privilege (LPP) 85 and then introduced a more restrictive interpretation of that privilege 86 . It appears that in the latter decision (C-623/22, July 2024), the analysis of the application of legal professional privilege under DAC6 by the Court seems to create 3 different categories of intermediaries, with rights that are not similar: the non-lawyer intermediary (who cannot ensure legal representation of their client in court) subject to reporting obligations, no waiver applicable; the non-lawyer intermediary bound by LPP who has mandate to ensure legal representation of their client in court, who benefits from the waiver, and must notify other intermediaries and/or their client; the lawyer intermediary who ensures legal representation of their client and who benefits from the waiver due to LPP, and has no obligation in terms of notification, except to their client.

The restrictive interpretation of the LPP seems to dismiss the use of the concept of ‘professional secrecy’ for the application of the waiver in article 8ab(5), which may, depending on the Member State, create a difference in treatment between lawyers and other intermediaries concerned by DAC6 obligations. The conclusions of the CJEU highlight the importance of ensuring legal certainty and show that the system in place is not flawless. It also notes the disparate interpretations of the concept among Member States.

In fine, the legal soundness of the framework has proved particularly relevant and robust over time despite attacks on the Directive's lawfulness, as confirmed by the decisions of the Court of Justice of the European Union which have stated the relevance of the general approaches of the DAC, upholding its legality 87 . However, it shows a recurring point of vulnerability on the LPP issue.

Annex VIII. DAC6 Hallmarks

1.Background

DAC6: Council Directive (EU) 2018/822 of 25 May 2018 amending Directive 2011/16/EU as regards mandatory automatic exchange of information in the field of taxation in relation to reportable cross-border arrangements

In order to tackle aggressive tax planning and help the Member States to identify possible loopholes that are exploited by taxpayers to avoid taxes, DAC6 establishes mandatory disclosures rules that apply on “reportable cross-border arrangement”, i.e. any cross-border arrangement containing at least one of the hallmarks set out in Annex IV of the DAC.

The EU nexus in DAC6

The arrangements concerned by DAC6 mandatory disclosures rules must have a link with the EU

They may concern either more than one Member State or a Member State and a third country where at least one of the following conditions is met: (a) not all of the participants in the arrangement are resident for tax purposes in the same jurisdiction; (b) one or more of the participants in the arrangement is simultaneously resident for tax purposes in more than one jurisdiction; (c) one or more of the participants in the arrangement carries on a business in another jurisdiction through a permanent establishment situated in that jurisdiction and the arrangement forms part or the whole of the business of that permanent establishment; (d) one or more of the participants in the arrangement carries on an activity in another jurisdiction without being resident for tax purposes or creating a permanent establishment situated in that jurisdiction; (e) such arrangement has a possible impact on the automatic exchange of information or the identification of beneficial ownership.

The reporting person, called ‘intermediary’, must also have a link with the EU

It is any person that designs, markets, organises or makes available for implementation or manages the implementation of a reportable cross-border arrangement, including any person that provides aid, assistance or advice in the matter. Intermediaries must be tax resident in the EU or have a permanent establishment in the EU that is providing the services linked with the arrangement or be incorporated by the laws of one of the Member States of the EU, or registered with a professional association related to legal, taxation or consultancy services in the EU. On this category of professionals also known as “lawyers”, DAC6 provides for a waiver from filing information on a reportable cross-border arrangement where the reporting obligation would breach the legal professional privilege under the national law of that Member State.

Finally, the EU nexus applies to the taxpayer as well, as any person to whom a reportable cross-border arrangement is made available for implementation (“relevant taxpayer”).

While the reporting obligation lies in principle on the intermediary or on the chain of intermediaries involved in the arrangement, the application of the waiver may lead to shift this obligation to the taxpayer. Such information shall then be filed by the taxpayer under certain conditions, i.e. only with the competent authorities of the Member State that features first in the list following: (a) the Member State where the relevant taxpayer is resident for tax purposes; (b) the Member State where the relevant taxpayer has a permanent establishment benefiting from the arrangement; (c) the Member State where the relevant taxpayer receives income or generates profits, although the relevant taxpayer is not resident for tax purposes and has no permanent establishment in any Member State; (d) the Member State where the relevant taxpayer carries on an activity, although the relevant taxpayer is not resident for tax purposes and has no permanent establishment in any Member State.

Annex IV and hallmarks in DAC6: the key factor to identify a reportable cross-border arrangement

The hallmarks are the characteristics or features of a cross-border arrangement that presents an indication of a potential risk of tax avoidance, as listed in Annex IV.

As the hallmarks are part of the set of information to be communicated by the competent authority of Member States, they are paramount to determine whether an arrangement is reportable or not. In this regard, the intermediaries or the taxpayers have to use the hallmarks to differentiate the nature of the arrangements that may be concerned by the reporting obligation.

Some of these hallmarks are combined with a complementary factor, known as the ‘main benefit test’ to characterise certain generic hallmarks, or certain specific categories of hallmarks corresponding to different types of transaction.

Annex IV to amended Directive 2011/16 (‘Annex IV’), entitled ‘Hallmarks’, provides for a main benefit test and lists categories of hallmarks as follows:

Part I. Main benefit test

That test will be satisfied if it can be established that the main benefit or one of the main benefits which, having regard to all relevant facts and circumstances, a person may reasonably expect to derive from an arrangement is the obtaining of a tax advantage.

In the context of hallmark under paragraph 1 of category C, the presence of conditions set out in points (b)(i), (c) or (d) of paragraph 1 of category C cannot alone be a reason for concluding that an arrangement satisfies the main benefit test.

Part II. Categories of hallmarks

A. Generic hallmarks linked to the main benefit test

1. An arrangement where the relevant taxpayer or a participant in the arrangement undertakes to comply with a condition of confidentiality which may require them not to disclose how the arrangement could secure a tax advantage vis-à-vis other intermediaries or the tax authorities.

2. An arrangement where the intermediary is entitled to receive a fee (or interest, remuneration for finance costs and other charges) for the arrangement and that fee is fixed by reference to:

(a) the amount of the tax advantage derived from the arrangement; or

(b) whether or not a tax advantage is actually derived from the arrangement. This would include an obligation on the intermediary to partially or fully refund the fees where the intended tax advantage derived from the arrangement was not partially or fully achieved.

3. An arrangement that has substantially standardised documentation and/or structure and is available to more than one relevant taxpayer without a need to be substantially customised for implementation.

B. Specific hallmarks linked to the main benefit test

1. An arrangement whereby a participant in the arrangement takes contrived steps which consist in acquiring a loss-making company, discontinuing the main activity of such company and using its losses in order to reduce its tax liability, including through a transfer of those losses to another jurisdiction or by the acceleration of the use of those losses.

2. An arrangement that has the effect of converting income into capital, gifts or other categories of revenue which are taxed at a lower level or exempt from tax.

3. An arrangement which includes circular transactions resulting in the round-tripping of funds, namely through involving interposed entities without other primary commercial function or transactions that offset or cancel each other or that have other similar features.

C. Specific hallmarks related to cross-border transactions

1. An arrangement that involves deductible cross-border payments made between two or more associated enterprises where at least one of the following conditions occurs:

(a) the recipient is not resident for tax purposes in any tax jurisdiction;

(b) although the recipient is resident for tax purposes in a jurisdiction, that jurisdiction either:

(i) does not impose any corporate tax or imposes corporate tax at the rate of zero or almost zero; or

(ii) is included in a list of third-country jurisdictions which have been assessed by Member States collectively or within the framework of the [Organisation for Economic Cooperation and Development (OECD)] as being non-cooperative;

(c) the payment benefits from a full exemption from tax in the jurisdiction where the recipient is resident for tax purposes;

(d) the payment benefits from a preferential tax regime in the jurisdiction where the recipient is resident for tax purposes;

2. Deductions for the same depreciation on the asset are claimed in more than one jurisdiction.

3. Relief from double taxation in respect of the same item of income or capital is claimed in more than one jurisdiction.

4. There is an arrangement that includes transfers of assets and where there is a material difference in the amount being treated as payable in consideration for the assets in those jurisdictions involved.

D. Specific hallmarks concerning automatic exchange of information and beneficial ownership

1. An arrangement which may have the effect of undermining the reporting obligation under the laws implementing Union legislation or any equivalent agreements on the automatic exchange of Financial Account information, including agreements with third countries, or which takes advantage of the absence of such legislation or agreements. Such arrangements include at least the following:

(a) the use of an account, product or investment that is not, or purports not to be, a Financial Account, but has features that are substantially similar to those of a Financial Account;

(b) the transfer of Financial Accounts or assets to, or the use of jurisdictions that are not bound by the automatic exchange of Financial Account information with the State of residence of the relevant taxpayer;

(c) the reclassification of income and capital into products or payments that are not subject to the automatic exchange of Financial Account information;

(d) the transfer or conversion of a Financial Institution or a Financial Account or the assets therein into a Financial Institution or a Financial Account or assets not subject to reporting under the automatic exchange of Financial Account information;

(e) the use of legal entities, arrangements or structures that eliminate or purport to eliminate reporting of one or more Account Holders or Controlling Persons under the automatic exchange of Financial Account information;

(f) arrangements that undermine, or exploit weaknesses in, the due diligence procedures used by Financial Institutions to comply with their obligations to report Financial Account information, including the use of jurisdictions with inadequate or weak regimes of enforcement of anti-money-laundering legislation or with weak transparency requirements for legal persons or legal arrangements.

2. An arrangement involving a non-transparent legal or beneficial ownership chain with the use of persons, legal arrangements or structures:

(a) that do not carry on a substantive economic activity supported by adequate staff, equipment, assets and premises; and

(b) that are incorporated, managed, resident, controlled or established in any jurisdiction other than the jurisdiction of residence of one or more of the beneficial owners of the assets held by such persons, legal arrangements or structures; and

(c) where the beneficial owners of such persons, legal arrangements or structures, as defined in Directive (EU) 2015/849 [of the European Parliament and of the Council of 20 May 2015 on the prevention of the use of the financial system for the purposes of money laundering or terrorist financing, amending Regulation (EU) No 648/2012 of the European Parliament and of the Council, and repealing Directive 2005/60/EC of the European Parliament and of the Council and Commission Directive 2006/70/EC (OJ 2015 L 141, p. 73)], are made unidentifiable.

E. Specific hallmarks concerning transfer pricing

1. An arrangement which involves the use of unilateral safe harbour rules.

2. An arrangement involving the transfer of hard-to-value intangibles. The term “hard-to-value intangibles” covers intangibles or rights in intangibles for which, at the time of their transfer between associated enterprises:

(a) no reliable comparable exist; and

(b) at the time the transaction was entered into, the projections of future cash flows or income expected to be derived from the transferred intangible, or the assumptions used in valuing the intangible are highly uncertain, making it difficult to predict the level of ultimate success of the intangible at the time of the transfer.

3. An arrangement involving an intragroup cross-border transfer of functions and/or risks and/or assets, if the projected annual earnings before interest and taxes (EBIT), during the three-year period after the transfer, of the transferor or transferors, are less than 50% of the projected annual EBIT of such transferor or transferors if the transfer had not been made.’

Evaluation of the DAC and the Annex IV

Legal Framework

According to Article 27(2) of DAC6 “every two years after 1 July 2020, the Member States and the Commission shall evaluate the relevance of Annex IV and the Commission shall present a report to the Council. That report shall, where appropriate, be accompanied by a legislative proposal”. In view of that evaluation, the Commission consulted the Member States on Annex IV through a questionnaire in 2022. Member States’ feedback have been processed in the context of the study carried out by an external contractor for the evaluation of the DAC.

In 2023, the Council decided to align the timeline of the assessment of Annex IV with the DAC evaluation which led to the deletion of Article 27(2). Consequently, not only the Annex IV but all the provisions of DAC6 are subject to the evaluation of the DAC which is currently underway.

Key figures on DAC6 reporting and distribution of hallmarks

The vast majority of potentially tax-harmful cross-border arrangements reported between 2020 and 2023 in the dedicated central directory originates from a few Member States: Germany, the Netherlands, Cyprus, Luxembourg, Ireland and Poland.

Figure 29: Total number of DAC6 arrangements per Member State (2020-2023)

Source: EC statistics

The distribution of hallmark by arrangement reported in the central directory is fairly uneven over time.

Hallmarks A (generic hallmarks linked to the main benefit test, 35%) and B (specific hallmarks linked to MBT, 34%) are very widely mentioned. In comparison, hallmarks C (linked to cross-border transactions, 15%) and E (transfer pricing, 12%) are used less often. It should be noted that hallmarks D (linked to arrangements aimed at avoiding AEOI obligations and concealing the BO), which also appear in the OECD’s MDR, account for only 4% of the hallmarks mentioned.

Figure 30: Distribution of hallmarks identified in arrangements reported

Source: EC statistics (2021-2024)

Methodology and fact-finding

The replies to the 2022 questionnaire to the Member States on Annex IV were considered in the context of the study commissioned by the Commission.

The subsequent public consultation and call for evidence that run between May and July 2024 included specific questions relating to the DAC6 and, more specifically, to hallmarks. Stakeholders and professional organisations were also given the opportunity to submit contributions in the light of the evidence gathered since the beginning of the exchange of information in July 2020 for 3 Member States, and April 2021 for the remainders 88 .

In addition, in May 2024, Member States were asked to reply to an updated questionnaire and to assess different features of Annex IV related to the concept of main benefit test and to hallmarks, in terms of clarity, actual application, and efficiency.

·Opinion of private stakeholders: results of the study, public consultation/call for evidence and targeted consultations

Tax intermediaries (business associations, lawyers, or companies) which are subject to the disclosure rules under DAC6 were critical about the impact that DAC6 had in terms of increased burden on reporting entities. They pointed out that DAC6 led to the analysis of substantial volumes of information, with significant impact in terms of human resources and IT developments for reporting entities.

They also complained about the lack of harmonised guidance on definitions, the ‘broad scope’ of the Directive, the ‘lack of precision’ and ‘clarity’ of certain hallmarks.

As showed in the tables below, the DAC6 is generally perceived as complex or very complex, with differences in implementation across Member States and the definition of the “main benefit test” and hallmarks (especially some of them, e.g. hallmark A1 regarding confidentiality clauses; hallmark A3 regarding standardised documentation; hallmark B2 regarding the conversion of income; hallmark C1b(i) regarding the “subject to zero or almost zero tax rate” condition) being particularly challenging for stakeholders.

Figure 31: Elements of DAC6 generating complexity

Respondents: 7 (Difficulties in obtaining advice from tax advisor); 9 (number of arrangements to be screened; description of hallmarks; determination of reportable arrangements; 8 (differences in implementation across EU Member States)
Source: open public consultation, 2024.

Moreover, although the wording of the hallmarks is the same in all Member States, private stakeholders noted that their interpretation may diverge from one tax authority to another, which may result in a different application of the Directive within the EU. This further adds administrative burden to intermediaries and taxpayers as well as increases legal uncertainty.

Figure 32: Hallmarks’ clarity and applicability

Respondents: 10. Source: open public consultation, 2024.

These findings are consistent with the replies provided by respondent to the public consultation questions on possible proposals to improve the functioning of DAC6. Respondents generally stressed the need for more detailed and practical guidance (including operational examples) on each hallmark to be endorsed by Member States and published by the Commission. Guidance should provide a detailed description and clarify the conditions of application of each hallmark. Stakeholders also strongly supported the systematic collection and provision of participants’ tax identification numbers, with the inclusion of mandatory elements in the DAC6 summary, reflecting a call for a more harmonised reporting. Harmonising penalties to ensure proportionality and effectiveness was moderately supported, underscoring the importance of fair and effective enforcement. On the contrary, there was less support for adding further detailed hallmarks, suggesting that stakeholders favour a balanced approach that avoids excessive complexity and administrative burden.

Figure 33: Relevance of possible proposals concerning DAC6

Source: open public consultation, 2024.

In addition, during targeted consultations, several private stakeholders stressed the need to reassess the hallmarks in the light of the current international tax framework as well as of the concrete application of DAC6, to establish if they are still relevant and fit-for purposes. They advised the Commission services to prioritise the elimination of certain hallmarks that no longer serve a clear purpose, rather than redesigning the existing hallmarks, with a view to achieving simplification and burden reduction.

It must be noted that these comments were gathered before the judgement of the Court of Justice of the European Union in the case C-623/22 Belgian Association of Tax Lawyers and Others (July 2024) which upheld DAC6 and did not judge that a possible lack of clarity in the definitions of hallmarks was an obstacle to its application. They were also collected before the ECA report on the tax harmful regimes (November 2024), that called for more guidance from the Commission, especially on Annex IV and hallmarks.

·Opinion of Member States: results of questionnaire and interviews

A thorough analysis of the hallmarks was carried out by the Member States in reply to the 2 questionnaires sent by the Commission services. Moreover, DAC6 and, more specifically the hallmarks, were discussed with the Member States during a number of interviews. The main conclusions are summarised below.

As regards clarity, most of the concepts and hallmarks used in Annex IV are considered to be sufficiently clear by most Member States. In addition, most of the Member States have not identified major issues in terms of application. However, between a quarter and a third of the Member States is of the view that, in some circumstances, the concept of main benefit test, as well as some hallmarks (A3, B2, B3, C4, E2 and E3) are not sufficiently clear. They may be too broad or being subject to divergent interpretations which raise questions about their application in terms of efficiency.

If most of the concepts and hallmarks of DAC6 are seen as useful both for tax risk analysis and for the detection of legal loopholes, their effectiveness may be questioned. However, most Member States do not wish to see them removed from Annex IV.

The large majority of the Member States has established guidance for cases deemed the most complicated (e.g. main benefit test), yet not built on common guidelines, which makes differences of interpretation of hallmarks possible.

Member States, therefore, generally uphold the relevance of DAC6 Annex IV as a tool aimed at identifying potentially aggressive cross-border arrangements, but ask for improvements in terms of the definitions, examples and concrete application of hallmarks to make it more efficient. As regards possible future action with respect to cases presenting possible issues, the option of amending the Directive received limited support by Member States. On the other hand, Member States favoured close cooperation with the Commission, in order to establish common guidance.

Finally, Member States made limited but targeted proposals for new hallmarks, targeting hybrid arrangements, structured real-estate arrangements using opaque entities, or potentially harmful arrangements based on the use of crypto assets.

Impact of the judgement of the CJEU in case C-623/22 (30 July 2024)

Clarity of the concepts: overall conclusions

Overall, the Court concluded that the so-called ‘broad scope’ of the reporting obligations under DAC6 is appropriate, given the objectives of DAC6 and the EU legislator's broad discretion in these matters. Consequently, the Court ruled that its scope does not affect the legality of DAC6.

The Court also assessed that the terms (e.g. ‘arrangements’) and deadlines challenged by the lawyers ‘association. In each instance, the CJEU determined that the provisions met the requirements of clarity and precision, referring to the definitions provided in DAC6, and suggested interpretations based on the common usage of the terms, internationally accepted practices, or the OECD’s Model Mandatory Disclosure Rules for CRS Avoidance Arrangements and Opaque Offshore Structures (2018) 89 .

Annex IV and the CJEU’s comments:

On hallmarks, the Court concluded that they are not drafted in such a way as to make the application of the reporting obligation unforeseeable for the persons subject to this obligation.

Regarding the question of the subjectivity of the main benefit test, that applies with certain hallmarks, the CJEU stated that it does not appear particularly difficult for an intermediary and, in the absence of an intermediary bound by the reporting obligation, for the relevant taxpayer, to decide whether the main benefit or one of the main benefits that can reasonably be expected of the arrangement they design and/or use is fiscal in nature.

The Court referenced the BEPS Action 12 Report, which indicates that the main benefit test involves comparing the expected tax advantage with other potential benefits of the transaction, concluding that it requires an objective assessment of the tax benefits.

The most relevant parts of the Judgement in that respect are paragraphs 53, 56, 57, 59, 64, 75 and 86.

Conclusions of the European court of auditors (28 November 2024)

The special report published by the European Court of Auditors (ECA) on 28 November 2024 assessed the appropriateness of measures and mechanisms employed in the EU by both the Commission and 5 EU Member States (Ireland, Cyprus, Luxembourg, Malta, and the Netherlands).

In particular, the ECA focused on the design and implementation of the following 3 directives that seek to curb harmful tax practices: the Anti-Tax Avoidance Directive, the 5th amendment to the Directive on administrative cooperation in the field of taxation (DAC 6) and the Directive on Tax Dispute Resolution Mechanisms. The audit covered the period from 2019 to 2023.

Overall, the ECA found that the established EU framework serves as a necessary first line of defence to support the fight against harmful tax regimes and corporate tax avoidance within the limited scope of the EU’s competences in matters of direct taxation. However, ECA criticised several points in terms of implementation by the Member States and monitoring by the European Commission, stressing the lack of guidance from the part of the Commission which would clarify the application of the EU rules in practice.

Although Member States have tools available for exchanging information on potentially harmful cross-border tax arrangements, they carry out few quality checks on reported information and make little use of the information received, which makes the fight against tax avoidance and tax evasion less effective; more specifically, in some Member States, the penalty systems for not complying with reporting obligations may not have a dissuasive effect due to the manifestly low level of the related penalties;

Although the Commission’s powers in the audited field is limited, the ECA recommend that the Commission clarify the EU legislative framework; improve the quality of DAC 6 reports; ensure that the impact of penalties is adequate, enhance its support to the Code of Conduct Group (the EU’s specialised body for business taxation); monitor the results and impact of the fight against harmful tax regimes and corporate tax avoidance.

In response to the ECA’s special report, the Commission accepted most of the recommendations, that takes into account the need to clarify the hallmarks, and the definitions required for their operation, including the possibility of examining a revision of the legal framework.

Annex IX. Replies from Member States on DAC6 Hallmarks

In the framework of the evaluation of the relevance of Annex IV of the Directive, the Commission services sent a questionnaire to the Member States in May 2024 to assess the concept of main benefit test and to the DAC6 hallmarks, in terms of clarity, actual application, and efficiency. The questionnaire is an updated version of a previous one, which was sent to the Member States in 2022.

The questionnaire is structured in 6 sections as follows:

–Section 1: Introduction

Section 2: Identification of the Member State

–Section 3: Evaluation of the concept of the Main Benefit Test

–Section 4: Assessment of existing hallmarks

–Section 5: Proposals for new hallmarks

–Section 6: Annex I – List of hallmarks provided for in Annex IV of DAC6

A summary of the relevant replies related to Section 2, 3, 4 and 5 is presented in the following subsections.

Member States replying to the questionnaire

26 Member States submitted their replies to the questionnaire under analysis.

One Member State specified that they have not replied because they did have nothing new to add to the answers sent in 2022.

Evaluation of the concept of the Main Benefit Test

1.Definition

Generic hallmarks under category A and specific hallmarks under category B and under points (b)(i), (c) and (d) of paragraph 1 of category C may only be taken into account where they fulfil the “main benefit test”. That test will be satisfied if it can be established that the main benefit or one of the main benefits which, having regard to all relevant facts and circumstances, a person may reasonably expect to derive from an arrangement is the obtaining of a tax advantage.

In the context of hallmark under paragraph 1 of category C, the presence of conditions set out in points (b) (i), (c) or (d) of paragraph 1 of category C cannot alone be a reason for concluding that an arrangement satisfies the main benefit test.

a)Questions on the MBT concept

Is the concept of MBT clear enough?

	Answers
Agree	11
Partially agree	13
Disagree	2

If partially agree / disagree, please choose which of the following concepts should be clarified:

	Answers
“main benefit”	8
“main benefit test”	8
Condition “may reasonably expect” fulfilled	6
Other / none of these	2
No answer	11

The majority of the replying Member States partially agreed that the concept of MBT is clear enough. Clarifications were deemed necessary for the concepts indicated in the above table.

b)Summary of the replies from Member States on how the concept of “main benefit”, “main benefit test” and the concept of the fulfilled condition could be improved

According to several Member States, the concept of “main benefit”, including that of “tax advantage”, is not clear enough and should be better defined at the EU level also to avoid misinterpretations. There is a common view to support additional guidance and practical examples provided by the Commission to tax administrations and taxpayers in order to understand if the “main benefit test” is satisfied.

Member States shared examples of guidance to the stakeholders in conducting this test, for example through the formulation of the following questions:

–Is there a tax benefit expected from the arrangement? Yes/No.

–Is the tax benefit substantial compared to the other impacts of the arrangement? Yes/No.

–Does the tax benefit fall specifically within the scope of a provision of the national law or jurisprudence of the taxpayers country of residence? Yes/No.

–Does the arrangement breach the purpose of the provisions applied or is the situation ambiguous or unclear? Yes/No.

Another Member State indicates that in order to clarify the concept of “main benefit” and the “main benefit test”, it would be useful to answer the following questions:

–Whether and to what extent are non-EU tax advantages relevant, as the scope of the DAC is limited to EU taxes (article 2 DAC)?

–How should “one of the main benefits” be interpreted?

–Whether and to what extent is deferral of taxation a tax advantage?

–Whether and to what extent is avoiding a fiscal disadvantage (e.g. avoiding legal or economic double taxation) a tax advantage for the application of the MBT?

–What is a tax advantage?

–How can an intermediary or relevant taxpayer substantiate that the main benefit or one of the main benefits is not a tax advantage?

–How should the following phrase be interpreted? “In the context of hallmark under paragraph 1 of category C, the presence of conditions set out in points (b)(i), (c) or (d) of paragraph 1 of category C cannot alone be a reason for concluding that an arrangement satisfies the main benefit test”.

–To what extent is “policy intent” relevant for the application of the MBT? And if “policy intent” is relevant, how should this be assessed when applying the MBT?

One Member State suggests improving the “main benefit test” including the term “effect” and reformulate it in the following way:

–Part I.Main benefit test Generic hallmarks under category A and specific hallmarks under category B and under points (b)(i), (c) and (d) of paragraph 1 of category C may only be taken into account where they fulfil the “main benefit test”.

–That test will be satisfied if :

–1. It can be established that the main effect or one of the main effects which, having regard to all relevant facts and circumstances, a person may reasonably expect to achieve from an arrangement is the obtaining of a tax advantage or,

–2. Regardless of paragraph 1, having regard to all relevant facts and circumstances, a person acting reasonably with the objective different than obtaining of a tax advantage could to choose other way of acting not related to obtaining of a tax advantage and the tax advantage is the main reason or one of the main reasons to choose way of acting determined by an arrangement.

–In the context of hallmark under paragraph 1 of category C, the presence of conditions set out in points (b)(i), (c) or (d) of paragraph 1 of category C can not alone be a reason for concluding that an arrangement satisfies the main benefit test.

Another Member State considers that a counterfactual test could be provided to satisfy the “main benefit test”: this would involve comparing the taxes resulting from the implementation of the cross-border arrangement with those that would have been due in its absence.

One Member State suggests supplementing the MBT with a threshold.

Furthermore, comments call for clarifying whether the assessment regarding the “main benefit” should be objective or subjective, in order to avoid the risk of having different interpretations by Member States.

The nuances between these examples and proposals tend to show that there is a difference in approach between Member States and confirms that there is no fully harmonised application of the concept at the moment.

2.Application

a)Questions on the application of the MBT

Does the MBT concept generate specific difficulties in its application?

	Answers
Yes	16
No	10

The majority of Member States agreed that the MBT may generate specific difficulties in practice because of the demonstration of the tax advantage:

One of the behaviours adopted by stakeholders and therefore risky for the purposes of DAC6 would be to adopt one of the schemes covered by DAC6 but to find many other reasons beyond the fiscal ones for the scheme, thus not satisfying the “main benefit test”. Instead, in some other cases, the arrangements are reported by taxpayers/intermediaries even though in their view the “main benefit test” is not met, because they prefer reporting than taking a risk on being subject to penalties.

The risk raised here is that these situations could lead to over-reporting, collection of information that is not necessarily relevant, and therefore of little use to Member States.

Several Member States reported receiving a significant number of requests for guidance on the assessment of the MBT, without quantifying them.

Do you have indications that would raise issues concerning the correct application of MBT in DAC6 (i.e. articles, conferences, informal exchanges between the stakeholders – e.g. banks, lawyers, assets managers… - and the administration)?

	Answers
Yes	10
No	16

Some Member States specified that the main issue concerning the correct application of MBT is the lack of clarifications and misinterpretations, sometimes intentional from intermediaries. According to one Member State, “some advisors build ways around [the concept of main benefit] so that to avoid being in scope. Therefore, for example, for those Member States who have chosen to use a ratio 90 to determine the main characteristic, it may be possible that tax advisors "play" with this ratio so that a particular arrangement be out of scope”.

Do you have indications that the MBT is a source of excessive administrative burden for intermediaries or taxpayers subject to the reporting obligations? (i.e. requests for guidance or for rulings, closed or ongoing administrative procedures, including legal challenges)

	Answers
Yes	8
No	18

In some cases (8 Member States), administrations have observed that the MBT concept may be a source of excessive administrative burden. According to these, excessive burden may arise from the time required to analyse all the benefits obtained from each arrangement 91 .

Has your Member State issued administrative guidance on the MBT?

	Answers
Yes	19
No	7

If no, please choose an answer

	Answers
DAC6 is clear enough	3
We provided additional explanations in the legal implementation	1
Other	3

It is notable that most Member States (20) have established ad hoc administrative guidelines or other legal framework in the context of the main benefit test. Only a very few (3) have not done so, considering that the DAC6 was sufficiently clear. However, many Member States stated that the Commission should issue common guidance to mitigate the risk of introducing inconsistent practices.

Assessment of existing hallmarks

Category A – Generic Hallmarks

a)Definition

The definition of Hallmark A in DAC6 Annex IV is clear

	Agree	Partially agree	Disagree	N/A
Hallmark A1	18	4	0	4
Hallmark A2a	21	1	0	4
Hallmark A2b	19	2	0	5
Hallmark A3	11	10	3	2

In principle, hallmarks A are considered to be fairly clear, except for hallmark A3, which appears to be more problematic, and to a more limited extent, hallmark A1.

Those Member States who replied “N/A” specified that this is due to the fact that the relevant hallmark A is not used by the intermediaries / taxpayers in the arrangements (2) either (i) reported by their Member State to the Central Directory (1), or (ii) related to their Member State reported by other Member States to the Central Directory. The same explanation applies below.

b)Application

Does Hallmark A generate specific difficulties in its application?

	Yes	No	N/A
Hallmark A1	1	16	9
Hallmark A2a	1	16	9
Hallmark A2b	1	13	12
Hallmark A3	5	15	6

Do you have indicators that would raise issues concerning the correct application of Hallmark A? (i.e. articles, conferences, informal exchanges between the stakeholders – e.g. banks, lawyers, assets managers, etc.– and the administration)

	Yes	No	N/A
Hallmark A1	3	17	6
Hallmark A2a	1	17	8
Hallmark A2b	1	15	10
Hallmark A3	5	18	3

Do you have indicators that would raise issues concerning Hallmark A being a source of excessive administrative burden for intermediaries or taxpayers subject to the reporting obligations? (i.e. requests for guidance or for rulings, closed or ongoing administrative procedures, incl. legal challenges)

	Yes	No	N/A
Hallmark A1	2	19	5
Hallmark A2a	2	19	5
Hallmark A2b	2	17	7
Hallmark A3	5	19	1

It should be emphasised that among the Member States responding to the questions above, the hallmarks A are not reported as having given rise to any specific difficulties, with the exception of hallmark A3. A few Member States mention elements that could possibly raise a problem in their correct application, particularly as regards hallmarks A1 (3) and A3 (5). Regarding excessive administrative burden, hallmark A3 concentrates most criticisms.

c)Efficiency

Hallmark A is sufficiently relevant for risk assessment

	Agree	Partially agree	Disagree	N/A
Hallmark A1	14	5	1	6
Hallmark A2a	17	2	1	6
Hallmark A2b	15	2	1	8
Hallmark A3	14	9	0	3

Hallmark A is sufficiently relevant for identification of loopholes in the domestic legislation or in the double taxation conventions

	Agree	Partially agree	Disagree	N/A
Hallmark A1	14	5	1	6
Hallmark A2a	14	4	1	6
Hallmark A2b	13	4	1	8
Hallmark A3	14	9	0	3

Of the Member States responding, most consider that hallmarks A are relevant for identifying tax issues, either in terms of risk analysis or in terms of possible legal loopholes. Hallmark A3 raises more specific issues described in the following paragraphs.

d)Summary of the replies from Member States on the definition of Category A Generic hallmarks

Some Member States have encountered practical issues and highlight ambiguities in the application of hallmarks A. A few Member States highlight the administrative burden imposed on taxpayers and intermediaries by DAC6 reporting.

Some Member States express the need for more detailed guidance on hallmark A1, especially concerning confidentiality obligations. They point out that hallmark A1 is frequently cited as a precaution with standard confidentiality clauses not specifically related to tax arrangements. This practice has led to significant administrative workload without necessarily increasing the value of the reports in certain Member States.

Although the wording of the directive makes it possible to differentiate between situations in a way that is generally clear for A1 (as the majority of Member States emphasise), one Member State consider that there is a possible overlap between hallmarks A1 and A3, suggesting further analysis to address the risk of redundancy.

2 Member States raised the possible issue of the application of the MBT to hallmark A2a, asking for more guidance in this respect.

According to 5 Member States, hallmark A3 has been more problematic due to its broad and ambiguous language. They emphasize the need for a clearer definition of the terms “standardized documentation” and “substantial customization” and note a possible confusion over what constitutes a significant change or major modification within the meaning of standardized. In 2 Member States, the broad application of this hallmark has led to excessive reporting, reducing the efficiency and quality of the reports.

Category B – Specific Hallmarks

a)Definition

The definition of Hallmark B in DAC6 Annex IV is clear

	Agree	Partially agree	Disagree	N/A
Hallmark B1	20	2	0	4
Hallmark B2	14	9	2	1
Hallmark B3	16	8	1	1

In principle, hallmarks B are considered by the Member States to be rather clear. However, hallmarks B2 and B3 concentrate criticisms from one third of the Member States, mainly on certain aspects of their definition.

Those Member States who replied “N/A” specified that this is due to the fact that the relevant Hallmark B is not used by the intermediaries / taxpayers in the arrangements either (i) reported by their Member State to the Central Directory, or (ii) related to their Member State reported by other Member States to the Central Directory. The same explanation applies below.

b)Application

Does Hallmark B generate specific difficulties in its application?

	Yes	No	N/A
Hallmark B1	0	19	7
Hallmark B2	6	18	2
Hallmark B3	7	17	2

Do you have indicators that would raise issues concerning the correct application of Hallmark B? (i.e. articles, conferences, informal exchanges between the stakeholders – e.g. banks, lawyers, assets managers, etc.– and the administration)

	Yes	No	N/A
Hallmark B1	1	17	8
Hallmark B2	4	20	2
Hallmark B3	5	19	2

Do you have indicators that would raise issues concerning Hallmark B being a source of excessive administrative burden for intermediaries or taxpayers subject to the reporting obligations? (i.e. requests for guidance or for rulings, closed or ongoing administrative procedures, incl. legal challenges)

	Yes	No	N/A
Hallmark B1	1	20	5
Hallmark B2	4	21	1
Hallmark B3	5	20	1

Among the Member States responding to the questions above, the hallmark B1 did not give rise to any specific difficulties. A handful of Member States reported possible issues with correct application and situations of excessive administrative burden with hallmarks of B2 (4 Member States) and B3 (5 Member States).

c)Efficiency

Hallmark B is sufficiently relevant for risk assessment

	Agree	Partially agree	Disagree	N/A
Hallmark B1	16	3	2	5
Hallmark B2	16	6	2	2
Hallmark B3	16	6	2	2

Hallmark B is sufficiently relevant for identification of loopholes in the domestic legislation or in the double taxation conventions

	Agree	Partially agree	Disagree	N/A
Hallmark B1	16	3	1	6
Hallmark B2	18	4	2	2
Hallmark B3	19	3	2	2

Despite hallmarks B2 and B3 are considered by 6 Member States as less relevant for risk assessment in comparison with hallmark B1, 2/3 of the responding Member States report that the hallmarks B are relevant for identifying tax issues, either in terms of risk analysis or in terms of detection possible legal loopholes.

d)Summary of the replies from Member States on Category B Specific hallmarks, namely on (i) which concept should be specified and (ii) concrete examples of issues concerning the correct application of such Hallmarks

In regard to hallmark B1, some Member States indicate that the concepts of “arrangement”, “participant”, “artificial steps” and “discontinuing its main activity” should be clarified.

In regard to hallmark B2, some Member States raise examples of transactions that could be possibly in scope, like debt-equity-swaps or waiver of claims. The terms “converting income” and “other categories of revenue” raise also specific questions in some Member States: scope of the concepts in case of the conversion of income into another category of income, or the conversion of a genuine loan into a hybrid loan, or the “disappearance” of income (e.g. before the arrangement there was a certain type of income, after the arrangement there is no (taxable) income (anymore)).

In case of newly set up arrangements, applicability of hallmark B2 to situations in which income does not yet exist or in which future income is uncertain for one Member State.

Another Member State expresses the need for clarifications on which level it should be determined whether there is a conversion of income, and on whether the requirement of the taxation at a lower level or the exemption from tax only refers to “other categories of revenue” or also to “capital” and “gifts”.

In regard to hallmark B3, according to some Member States, the following terms and concepts should be specified:

–“circularity” of a transaction (i.e. back to back loan).

–“roundtripping” (e.g. funds return to the same taxpayer or the same jurisdiction), “primary commercial purpose”(of interposed entities).

A Member State wonders whether the requirement “without primary commercial function” should be interpreted as the same of another requirement in hallmark D2a concerning entities “that do not carry on a substantive economic activity”.

The scope of hallmark B3 also raises a comment on whether hallmark B3 only applies in case of the interposing of entities and/or the offsetting of transactions (i.e. the last sentence of hallmark B3), or whether these situations are only examples of the application of this Hallmark, and if conditions.

In terms of the use of information made available, regarding Hallmark B3 arrangements the incompleteness of information on the whole arrangement may constitute a challenge, mainly because it may involve jurisdiction outside the EU, and EU intermediaries may not know the whole plan of transactions.

According to some Member States, common guidance with examples might be useful to understand how Hallmarks under Category B should work.

Category C – Specific Hallmarks related to deductible cross-border transactions

a)Definition

The definition of Hallmark C in DAC6 Annex IV is clear

	Agree	Partially agree	Disagree	N/A
Hallmark C1a	19	4	1	2
Hallmark C1b(i)	18	5	1	2
Hallmark C1b(ii)	20	3	1	2
Hallmark C1c	18	6	1	1
Hallmark C1d	18	6	1	1
Hallmark C2	22	1	1	2
Hallmark C3	22	2	1	1
Hallmark C4	16	7	2	1

In principle, hallmarks C are considered by the Member States to be rather clear. However, hallmarks C1c, C1d, and C4 raise some mitigated comments from certain Member States.

Those Member States who replied “N/A” specified that this is due to the fact that the relevant hallmark C is not used by the intermediaries / taxpayers in the arrangements either (i) reported by their Member State to the Central Directory, or (ii) related to their Member State reported by other Member States to the Central Directory. The same explanation applies below.

b)Application

Does Hallmark C generate specific difficulties in its application?

	Yes	No	N/A
Hallmark C1a	2	22	2
Hallmark C1b(i)	2	22	2
Hallmark C1b(ii)	2	22	2
Hallmark C1c	2	22	2
Hallmark C1d	5	20	1
Hallmark C2	2	21	3
Hallmark C3	1	23	2
Hallmark C4	5	19	2

Do you have indicators that would raise issues concerning the correct application of Hallmark C? (i.e. articles, conferences, informal exchanges between the stakeholders – e.g. banks, lawyers, assets managers, etc.– and the administration)

	Yes	No	N/A
Hallmark C1a	2	22	2
Hallmark C1b(i)	0	24	2
Hallmark C1b(ii)	0	24	2
Hallmark C1c	2	23	1
Hallmark C1d	1	23	2
Hallmark C2	1	22	1
Hallmark C3	0	25	1
Hallmark C4	4	20	2

Do you have indicators that would raise issues concerning Hallmark C being a source of excessive administrative burden for intermediaries or taxpayers subject to the reporting obligations? (i.e. requests for guidance or for rulings, closed or ongoing administrative procedures, incl. legal challenges)

	Yes	No	N/A
Hallmark C1a	1	22	3
Hallmark C1b(i)	2	22	2
Hallmark C1b(ii)	3	21	2
Hallmark C1c	4	21	1
Hallmark C1d	4	21	1
Hallmark C2	3	20	3
Hallmark C3	2	22	2
Hallmark C4	5	19	2

The hallmarks C generally do not raise major difficulties in their application. A handful of Member States reported specific issues with C1d and C4 (5 Member States), issues in terms of correct application with C4 (4 Member States), and situations of excessive administrative burden with C1c (4 Member States), C1d (4 Member States) and C4 (5 Member States).

c)Efficiency

Hallmark C is sufficiently relevant for risk assessment

	Agree	Partially agree	Disagree	N/A
Hallmark C1a	22	1	1	2
Hallmark C1b(i)	22	3	0	1
Hallmark C1b(ii)	23	2	0	1
Hallmark C1c	22	2	1	1
Hallmark C1d	20	4	1	1
Hallmark C2	20	2	1	3
Hallmark C3	22	2	0	2
Hallmark C4	22	3	0	1

Hallmark C is sufficiently relevant for identification of loopholes in the domestic legislation or in the double taxation conventions

	Agree	Partially agree	Disagree	N/A
Hallmark C1a	21	1	1	3
Hallmark C1b(i)	22	2	0	2
Hallmark C1b(ii)	21	3	0	2
Hallmark C1c	21	2	1	2
Hallmark C1d	19	4	1	2
Hallmark C2	20	2	0	4
Hallmark C3	21	2	0	3
Hallmark C4	20	4	0	2

The responses from Member States show that there is a consensus on the usefulness and efficiency of C hallmarks. Member States agree – to a large extent – on their relevance to tax risk analysis or the detection of loopholes, with minor reservations on hallmarks C1d (4 Member States) or C4 (4 Member States).

d)Summary of the replies from Member States on Category C Specific hallmarks related to deductible cross-border transactions, namely on (i) which concept should be specified and (ii) concrete examples of issues concerning the correct application of such Hallmarks

Some Member States ask for further clarification on the concepts used in hallmarks C as follows:

–One Member State mentions that regarding hallmark C1a, the means to prove tax residency, for example, through a tax residency certificate or further investigation, should be detailed.

–A Member State do not consider the concept of “deductible payment” clear enough and wonders whether it includes deemed payments and accruals, and whether payments for asset acquisition are deductible.

–Regarding hallmark C1b(i), the terms “almost zero” tax rate raise some questions among some Member States.

–Regarding hallmarks C1b(ii) and C1c, one Member State suggests clarification on how to apply these hallmarks to permanent establishments.

–With reference to hallmark C1c, a Member State mentions that it is not clear whether arrangements involving exempt entities should be reported.

–Concerning hallmark C1d, a Member State suggests a clarification of the concept of “preferential tax regime”, and the causality between payments and preferential tax regimes.

–Concerning hallmark C2, one Member State report concerns about cross-border leasing and the possibility of multiple depreciations and investment allowances in different Member States. According to this Member State, leasing contracts are often reported, and they suggest common guidance or an “angel list” (i.e. white list) to prevent irrelevant reporting. On the other hand, another Member State seeks clarification on hybrid mismatch arrangements and double deductions.

–Regarding hallmark C3, the claim for “relief from double taxation” did not seem clear to one Member State.

–As regard hallmark C4, some Member State express the need for clarification on exit taxation, the relevance of introducing a differentiation between beneficial ownership versus civil law ownership, as well as a definition of “material difference” in payable amounts to ensure harmonized application.

Category D – Specific Hallmarks concerning AEOI and beneficial ownership.

a)Definition

The definition of Hallmark D in DAC6 Annex IV is clear

	Agree	Partially agree	Disagree	N/A
Hallmark D1a	18	4	0	4
Hallmark D1b	20	2	0	4
Hallmark D1c	18	4	0	4
Hallmark D1d	20	3	0	3
Hallmark D1e	17	4	0	5
Hallmark D1f	18	3	0	5
Hallmark D2	20	3	0	3

In principle, hallmarks D are considered to be clear by 2/3 of the Member States. However, hallmarks D1a, D1c, D1e and D1f raise some mitigated comments from a minority of Member States.

b)Application

Does Hallmark D generate specific difficulties in its application?

	Yes	No	N/A
Hallmark D1a	1	20	5
Hallmark D1b	0	21	5
Hallmark D1c	1	19	6
Hallmark D1d	0	21	5
Hallmark D1e	1	19	6
Hallmark D1f	2	18	6
Hallmark D2	3	19	4

Do you have indicators that would raise issues concerning the correct application of Hallmark D? (i.e. articles, conferences, informal exchanges between the stakeholders – e.g. banks, lawyers, assets managers, etc.– and the administration)

	Yes	No	N/A
Hallmark D1a	2	21	3
Hallmark D1b	1	22	3
Hallmark D1c	2	20	4
Hallmark D1d	2	21	3
Hallmark D1e	2	19	5
Hallmark D1f	1	20	5
Hallmark D2	1	22	3

Do you have indicators that would raise issues concerning Hallmark D being a source of excessive administrative burden for intermediaries or taxpayers subject to the reporting obligations? (i.e. requests for guidance or for rulings, closed or ongoing administrative procedures, incl. legal challenges)

	Yes	No	N/A
Hallmark D1a	2	21	3
Hallmark D1b	2	20	4
Hallmark D1c	2	20	4
Hallmark D1d	2	21	3
Hallmark D1e	2	19	5
Hallmark D1f	2	19	5
Hallmark D2	3	21	2

90% of the Member States responding consider that the hallmarks D generally do not raise major difficulties in their application, and do not create excessive burden for intermediaries.

c)Efficiency

Hallmark D is sufficiently relevant for risk assessment

	Agree	Partially agree	Disagree	N/A
Hallmark D1a	20	2	0	4
Hallmark D1b	20	2	0	4
Hallmark D1c	20	2	0	4
Hallmark D1d	20	2	0	4
Hallmark D1e	19	2	0	5
Hallmark D1f	19	2	0	5
Hallmark D2	18	4	0	4

Hallmark D is sufficiently relevant for identification of loopholes in the domestic legislation or in the double taxation conventions

	Agree	Partially agree	Disagree	N/A
Hallmark D1a	20	2	0	4
Hallmark D1b	20	2	0	4
Hallmark D1c	19	3	0	4
Hallmark D1d	20	2	0	4
Hallmark D1e	18	3	0	5
Hallmark D1f	17	5	0	4
Hallmark D2	17	5	0	4

The responses from Member States show that there is a consensus on the usefulness of hallmarks D in terms of efficiency. All Member States agree – to a large extent – on their relevance to tax risk analysis or the detection of loopholes, with minor reservations on hallmarks D1f (5 Member States) or D2 (5 Member States). Unlike the other hallmarks, where there are always 1 or 2 Member States questioning the relevance of certain sub-categories, the D hallmarks are unanimously supported by the Member States.

d)Summary of the replies from Member States on Category D Specific hallmarks, namely on AEOI and beneficial ownership

According to a couple of Member States, there was no reporting of many cases under these hallmarks. One reason put forward by these Member States could be the difficulty in interpretation.

Regarding hallmark D1, some Member States raise the need for clarification on the question of whether the requirement is satisfied merely when the rules of AEOI are undermined, or when the participants have intended that outcome. One Member State points out that in the 2018 OECD report “Model Mandatory Disclosure Rules for CRS Avoidance Arrangements and Opaque Offshore Structures” mandatory disclosure rules in relation to CRS avoidance are described. Part of the text of Hallmark D1 is aligned with the OECD mandatory disclosure rules on CRS avoidance but partly deviates from it, causing a risk of unclarity to what extent the work of the OECD can be used for the interpretation of Hallmark D1.

According to certain Member States, the following concepts should be specified with examples in regard to:

–Hallmark D1a: “any equivalent agreements on the automatic exchange of Financial Account Information”

–Hallmark D1c: “reclassification of income and capital into products or payments that are not subject to the automatic exchange of financial account information”.

–Hallmark D2: “substantive economic activity supported by adequate staff, equipment, assets and premises”; “where the beneficial owners (…) are made unidentifiable”.

Category E – Specific Hallmarks concerning transfer pricing

a)Definition

The definition of Hallmark E in DAC6 Annex IV is clear

	Agree	Partially agree	Disagree	N/A
Hallmark E1	21	2	2	1
Hallmark E2	15	9	2	0
Hallmark E3	12	12	2	0

While Member States consider that the E1 hallmark is generally sufficiently clear, several Member States have identified areas for improvement regarding the significance of the E2 and E3 hallmarks.

b)Application

Does Hallmark E generate specific difficulties in its application?

	Yes	No	N/A
Hallmark E1	1	23	2
Hallmark E2	7	18	0
Hallmark E3	11	15	0

Do you have information that would raise concerns for the correct application of Hallmark E? (i.e. articles, conferences, informal exchanges between the stakeholders – e.g. banks, lawyers, assets managers, etc.– and the administration)

	Yes	No	N/A
Hallmark E1	1	23	2
Hallmark E2	3	23	0
Hallmark E3	5	21	0

Is Hallmark E a source of excessive administrative burden for intermediaries or taxpayers subject to the reporting obligations? (i.e. requests for guidance or for rulings, closed or ongoing administrative procedures, incl. legal challenges)

	Yes	No	N/A
Hallmark E1	3	21	2
Hallmark E2	6	20	0
Hallmark E3	8	18	0

With regard to the application of hallmarks E, the opinion of the Member States is divided: while hallmark E1 does not seem to pose a significant problem, hallmarks E2 and, even more, E3 raise concerns about their correct application and are likely to cause an excessive administrative burden to some of the Member States.

c)Efficiency

Hallmark E is sufficiently relevant for risk assessment

	Agree	Partially agree	Disagree	N/A
Hallmark E1	22	1	1	2
Hallmark E2	23	2	1	1
Hallmark E3	20	5	1	0

Hallmark E is sufficiently relevant for identification of loopholes in the domestic legislation or in the double taxation conventions

	Agree	Partially agree	Disagree	N/A
Hallmark E1	22	2	0	2
Hallmark E2	20	4	0	2
Hallmark E3	19	5	0	2

Despite certain reservations expressed by some Member States about the application of hallmarks E2 and E3, it must be emphasised that their relevance is not questioned in terms of effectiveness. Member States (with the exception of one) all agree that the hallmarks are relevant enough, both for risk analysis and for identifying loopholes.

d)Summary of the replies from Member States on Category E Specific hallmarks concerning transfer pricing

The comments on hallmark E mainly concern the terms used, which for some Member States seem to give rise to some confusion as to their exact definition. The risk of inconsistent reporting is raised to explain their concern.

For example, the terms “unilateral safe harbour rules” is open to interpretation and could benefit from explicit references to OECD Transfer Pricing Guidelines.

Several Member States report difficulties with Hallmark E2 due to the lack of a more precise definition for “hard-to-value intangibles”. They stress that it could lead to compliance burden imposed on companies, requiring extensive valuation efforts. These Member States suggest aligning the terms with the OECD Transfer Pricing Guidelines to mitigate these challenges. They also note that licensing arrangements and risk transfer mechanisms also complicate the application of this hallmark. Only one Member State highlights the need for more explicit criteria to determine reportability.

Hallmark E3 is the most frequently discussed, with 10 Member States raising concerns. The definition for “EBIT” and the risk for inconsistent application of the three-year period following a transfer are mentioned as possibly creating issues. Some Member States call for more explicit guidelines on calculating EBIT, including whether to follow IFRS, GAAP, or national accounting standards. Some other question how to handle situations involving prolonged losses or fluctuating EBIT, suggesting that the current 50% threshold and three-year period might be too arbitrary and lenient. 2 Member States point out ambiguities in what constitutes an “intragroup” transfer and how EBIT reductions should be measured, particularly for relocations or transfers involving permanent establishments, emphasizing the need for further clarification from the Commission.

Proposals for new hallmarks

To improve the functioning of DAC6 as an “alert system” monitoring the trends of possibly tax-harmful tax planification, some Member States called for revisiting certain Hallmarks and add new features to better capture relevant arrangements that may potentially harmful:

–In category B: a new hallmark targeting the dividend stripping schemes 92 ;

–In category C: a new hallmark targeting conduit arrangement for treaty shopping or Directive shopping, exemplified by non-EU investors using shell companies in EU member states for tax benefits;

–In category C, new hallmarks covering hybrid mismatches, referencing OECD reports . They should focus on financial instruments, hybrid entities, and permanent establishments leading to deduction without inclusion and double deduction scenarios;

–In category E, new hallmarks for arrangements based on transfer of intellectual property, transfer-pricing mismatches 93 , and profit allocation to permanent establishments.

New Hallmarks targeting arrangements with the following features are also proposed:

–Property transfer of asset within a short time;

–Differences between actual and legal places of business;

Limited substance in group companies, and shell entities;

Payments to entities with low effective tax rates;

Use of crypto assets in transactions;

–Use of hybrid entities not neutralized by ATAD;

–Real estate-based arrangements in relation with low effective tax rates.

A few other Member States believe that for the time being, no additional hallmarks are needed, as they still need time to become familiar with the existing hallmarks of DAC6.

Annex X. Summary of the VISDAC Recommendation Report

INTRODUCTION

In order to respond to the key recommendations in the Reports of the European Court of Auditors and the Committee on Economic and Monetary Affairs, the Commission in collaboration with Member States launched a programme of visits to all Member States (the ‘VISDAC project’). The aim of the VISDAC project was to improve the quality and use of DAC1-4 AEOI data exchanged between Member States. Accordingly, Member States received recommendations on quality and use issues. The VISDAC project took place between 2022-2024. This Annex summarises the VISDAC Recommendation Report (“Report”) that has been issued to Member States and is broken down by recommendations related to DAC data quality, data use and horizontal recommendations.

DAC1-4 Data Quality Recommendations

There were a number of quality recommendations that were common among Member States and will require ongoing monitoring by the Commission. There were many differences between Member States in the extent of systematic controls in place to check data collected at national for DAC1 purposes. The Commission will follow-up with Member States through the Fiscalis program to promote best practices.

The completeness and correctness of the non-resident TIN collected for DAC purposes remains an ongoing challenge for Member States. Member States were issued with recommendations to improve their processes, in particular with regard to TIN validation. The Commission will continue to promote tools like the TIN-on-the Web module 94 . Further, the Commission will launch a study in 2025 which will include the validation of the TIN as a topic as well as examining possible options for implementing an EU-TIN for natural persons.

With respect to CRS (Common Reporting Standard), which is closely aligned with DAC2, the OECD Peer Review has issued Member States a number of recommendations on the effectiveness of their exchanges with other jurisdictions 95 . During the visits, the VISDAC experts monitored how these recommendations are being addressed by Member States. Member States should continue their efforts to address these recommendations.

Although Member States have established systems for checking that MNEs correctly report data according to format of the DAC4 requirements for country-by-country reporting, some Member States have not included checking compliance with the underlying framework of country-by-country reporting in their national audits. This issue is of particular relevance for related information that will be collected under DAC9 which incorporates the OECD Pillar 2 exchange of information. Member States should ensure that checking DAC4 compliance is an integral part of their national audit processes for MNEs.

DAC1-4 DAC Data Use Recommendations

To make effective use of DAC AEOI data, efficient processes are necessary for matching DAC data received with taxpayer data in the national tax databases of the receiving Member State. The Report proposes a number of best practices to improve the matching processes. Essential elements of such matching processes are automation and the use of new technology plus a few other measures. Recommendations have been issued to Member States to implement actions to improve compliance of their taxpayers in declaring DAC1 and DAC2 related income and assets. The Report proposes a wide range of compliance activities, some of which have already been implemented by several Member States. These include Voluntary Compliance Programs (‘VCPs’), information notifications through taxpayer web-portals, the pre-filling of tax returns as well as nudge campaigns. Adequate digitalisation is a key requirement to use such tools.

The systematic use of DAC data is performed through the use of risk assessment processes. The Report proposes a number of actions to improve Member State DAC risk assessment processes such as integrating various data streams related to the taxpayer obtained through the DAC. While all Member States have some form of targeted risk assessment processes, the scope of these risk assessment models differs significantly among the Member States, including the frequency and the extent to which all DAC1 and DAC 2 data is used in the risk assessment processes of Member States. To be effective, risk assessment processes should use the data from most recent exchanges, be applied systematically, and be evaluated on a regular basis.

Further, Member States could further explore the use of DAC data in their wider risk assessment models, for example, combining all received DAC data with national tax data, for assessing overall tax liability. DAC data should also be used for other purposes beyond direct taxation as permitted by the DAC, like tax recovery and combatting Money Laundering 96 .

Horizontal Recommendations

The Report identified a number of recommendations that were common to DAC1-4 which are outlined below.

Underlying all the VISDAC recommendations is the need for Member States to ensure that adequate human and IT resources are provided, and adapted to the continuous development of EU legislation, to ensure the high quality of DAC data collected at national level and sent to other Member States. Further, adequate human and IT resources are necessary for the effective use of DAC data for the compliance activities of tax administrations.

Member State experts stressed the importance of the bilateral feedback process between Member States which is reflected in the number of recommendations related to the annual feedback process. The annual feedback process, revised in January 2024, involves providing other Member States with feedback on the quality of data they send, as well as quantitative data like the sending Member State’s matching rates. Unmatched data can also be notified to the sending Member States for follow-up. The Report includes a number of proposed actions that Member States could implement to improve their current feedback processes. Member States received recommendations in relation to measuring the effectiveness and efficiency of the DAC. Together with Member States, the Commission has established a KPI (Key Performance Indicator) framework which can be used to monitor the objectives of DAC1-4, including their benefits. Publication by Member States of these benefits increases transparency on how DAC data is used by tax administrations.

The DAC requires that Member States have sanction regimes that ensure compliance of RFIs (Reporting Financial Institutions) with respect to their obligations under DAC2, and MNEs (Multinational Enterprises) with respect to their obligations under DAC4. Recommendations in respect of sanctions were issued to Member States for the non-application of sanctions, increase the sanction fine, and application of sanctions that go beyond deliberate breach or gross negligence. Member States should review their existing provisions and take measures to ensure that their DAC sanction regimes have effective, proportionate and dissuasive penalties for non-compliance.

Conclusions

The main conclusion of the Report is that Member States can generally still make progress to ensure a better quality of DAC1-4 data exchanged, to fulfil their obligations under the DAC, and to better use this data to meet the objectives of ensuring tax fairness and reducing the tax gap.

The Fiscalis programme, through the use of expert groups or dedicated workshops, will continue to be used to address the issues identified in the Report as well as to promote best practices. Further, the Commission Services will use the YA AEOI Survey to monitor how the recommendations are being implemented by Member States.

The Commission will consider how best to move forward with the VISDAC project, for example, by having a targeted approach on the issues that are regarded as priority.

Annex XI - Targeted consultation with Economists

26 September 2024 European Commission – Brussels

Minutes

This document should be regarded solely as a summary of the contributions from the experts interviewed for the targeted consultation in the framework of the Evaluation of the DAC. It cannot in any circumstances be regarded as the official position of the Commission or its services. Responses to the interviews cannot be considered as a representative sample of the views of the professionals or the EU population.

Objectives

·A targeted panel of economists was invited to comment specific points on the DAC to assess its effectiveness, its efficiency; the relevance and the coherence of its provisions; and its EU added value.

Summary of the discussions

·Efficiency: impact of the AEOI on tax evasion and profit shifting

The panel of experts confirmed the deterrent effect of the AEOI as regards to DAC2. They confirmed that the impact measured led to the conclusions that DAC2 / CRS may be considered a major success, to a certain extent.

Economists found that rapid progress can be made against tax evasion when global agreements are enforced: the creation of AEOI CRS/DAC2 in force since 2017 (> 100 countries participating in 2023), CbCR/DAC4, and eventually Pillar 2 (endorsed by > 140 countries in 2021) led to a global effect: offshore tax evasion has declined by a factor of about three over the last 10 years.

Before 2013, households owned the equivalent of 10% of world GDP in financial wealth offshore, the bulk of which was undeclared to tax authorities; these assets were held mostly by HNWIs (or structures shadowing them). The situation regarding the assets has changed. The equivalent of 10% of world GDP is still held in financial wealth offshore, but only about 25% of it evades taxation: economic models and scenarios show a decrease in non-compliance.

However, new trends in concealing assets from tax administration have been detected and the observations of economists confirmed that several leverages were used:

i) shift to real estate investments: the trend is very clear even before the entry into force of DAC2/CRS. This was done in favour of a place like London in the years 2015-2020. After Brexit, there is a trend towards Dubai, where real estate investment has boomed (till effective nowadays).

ii) Use of ‘golden visas’ based on ‘residency by investment programs schemes’: evidence from outside and inside the EU market e.g. Portugal. This has also affected the property market, which has been completely disrupted in certain places.

·Relevance of DAC4: transparency in transfer pricing policies, change in the behaviour of multinational enterprises

Economists found the effects of DAC4 / CRS questionable, considering that 7 years after the start of the BEPS process (and CbCR), profit shifting appears to have changed only marginally. According to economists, a relatively large amount of profits is still shifted to offshore jurisdictions (~$1 trillion in 2022), equivalent of 35% of profits booked by MNEs outside the HQ’s jurisdiction. CIT losses are still significant, i.e. up to 10% of corporate tax revenues collected globally.

However, it doesn’t mean that the policy initiatives had no effect: economists had evidence that without them, situation would be probably worst and BEPS higher.

·EU added value and impact on EU competitiveness

Asked about the risk for EU businesses of being impacted by mandatory disclosure rules on potential aggressive tax planning or DAC4, economists had mitigated analysis, and seemed rather dubious about these hypotheses.

They indicated that they have strong evidence that DAC2/CRS has an economic impact on financial institutions at a global level, and that in some cases, the weight seems to be poorly distributed between large companies and SMEs. On the other hand, they questioned such an effect in relation to DAC6 and DAC4, especially for the latter given that CbCR are global.

When asked the relevance of relaxing obligations on MNEs to ensure a better level playing field for businesses, the economists recalled that the shift in profits from MNEs to offshore jurisdictions had been very little affected by the previous measures, and that the promise of a minimum of 15 % overall taxation had not yet been achieved, despite the entry into force this year of Pillar 2.

They also consider that it would be very premature to call into question the CbCR. They also referred to the public CbCR as an instrument of economic interest which makes it possible to have up-to-date data and to compare it with what administrations have received with the CbCR / DAC4 reports.

They also considered that the reduction in obligations would send a contradictory signal, if not negative, in so far as it could be perceived as an incentive to continue introducing aggressive tax schemes, to the detriment of Member States’ budgets.

·Efficiency and coherence of the DAC and possible challenges

In terms of individual tax evasion, recent reports tend to show that practices related to wealth management and capital concealment have evolved.

These have concentrated in new types of assets (e.g. crypto assets) or more traditional ones (real-estate investments through shell entities or opaque arrangements).

In terms of information exchange, the EU has instruments targeting income from real estate properties (DAC1 and, to some extent, some beneficiaries through DAC7) and more recently cryptos (DAC8). The issue has been raised from several angles.

The most important is the identification of the beneficial owner. The EU AML context was recalled, as well as the progress foreseen in the most recent AML package. This point has made it possible to identify the relative European harmonisation on the subject, whereas this is not necessarily the case in other non-EU jurisdictions.

An exchange also took place on the WP10 initiative at the OECD. It was indicated that the EU is ahead of the rest of the countries, as Member States already exchange real estate data and see the interest of it. This sequence of exchanges showed that, from this point of view, the European Union is a frontrunner.

However, access to beneficial ownership information for tax purposes seems crucial for the future, particularly in the case of a global tax on High net-worth individuals (HNWIs). In that regard, asked whether an automatic exchange of tax information on immovable property and related beneficial owners should be carried out, to ensure the success of a comprehensive tax on HNWIs, the economists agreed that it would be particularly complex to introduce taxation without the information making it possible to establish taxable bases, for example by relying on the market value of the property owned, collected by the competent authorities.

·Simplification

The exchange continued on the possible basis of the real estate thresholds to explore whether such an approach would be relevant and facilitate the exchange of quality data. This did not lead to firm conclusions but rather to different hypothesis regarding how to understand a taxpayer’s total capital and value, depending on the nature of his assets.

Related documents – List of questions sent to the experts

Efficiency: impact of the AEOI on tax evasion and profit shifting

Does administrative cooperation and the exchange of tax information have a general impact on the fight against tax evasion and tax havens?

After 10 years of CRS / DAC2, what lessons can be learned from this system, especially in terms of their impact on tax transparency?

Have you evidence that it led to better compliance?

Did it lead to new trends in concealing assets from the tax administrations?

Relevance of DAC4: transparency in transfer pricing policies, change in the behaviour of multinational enterprises

Do you think, or do you have any evidence to suggest that the CbCR (DAC4) declaration has promoted greater transparency in transfer pricing policies, or even a change in the behaviour of multinationals in tax management through profit shifting?

EU added value and impact on EU competitiveness

Could mandatory disclosure rules on potential aggressive tax planning affect the competitiveness of consulting firms that offer these arrangements?

Could this affect the competitiveness of economic agents who have to comply with it?

If so, is there an economic benefit that justifies this loss of competitiveness?

Will Pillar 2 put an end to aggressive tax planning practices by multinationals so that Automatic exchange of information would be not necessary anymore?

Efficiency and coherence of the DAC and possible challenges

Is there a need for strengthening current provisions, particularly regarding the exchange of information on real estate owned by non-residents of other Member States

Would be a European register of beneficial owners of real estate necessary?

Do you think such a measure would have a positive impact on the fiscal resources of Member States?

Annex XII

Annex XII – Targeted consultation with tax law experts

25 September 2024

Minutes

OBJECTIVES

·A targeted panel of tax law experts was invited to comment specific points on the DAC to assess its effectiveness, its efficiency; the relevance and the coherence of its provisions; and its EU added value.

SUMMARY OF THE DISCUSSIONS

·Efficiency: on deterrence effect of Automatic exchange of information (AEOI) provided for in the DAC

I)The panel of experts confirmed the deterrent effect of the AEOI as regards to DAC2 and DAC6, and its consequences on intermediaries and businesses.

In particular, the experts emphasised the effectiveness of DAC2 / CRS in disclosing financial assets abroad. These effects are considered persistent and well assimilated by taxpayers.

Experts indicated that they had also perceived a change in the attitude of taxpayers regarding DAC6 arrangements in the latest years. More specifically, it was confirmed that it had an impact on large companies, leading some of them to set up very elaborated systems of internal due diligence given the broad scope of the Directive.

To ensure consistent reporting, some legal firms have also developed important internal guidelines, to take account of the differences in application between Member States.

However, it was stressed that the dissemination of knowledge about DAC6 is not homogeneous, including among legal experts.

Furthermore, the fact that DAC6 has been challenged 97 in terms of its compatibility with fundamental rights, its legal foundations and its complexity, has not made it easy to share best practices on the subject.

II)On a more forward-looking level, the possible impact of the Public CbCR Directive 98 on tax behaviour was also mentioned.

The potential reputational damage or the public reaction, especially from civil society groups (e.g. NGOs), media or policymakers, are factors that may have an impact on the tax behaviour of MNEs. It may happen if the public CbCR data reveals a (perceived) mismatch between the profits and taxes paid in different jurisdictions, or a low overall effective tax rate (regardless of the legal and economic reasons behind it).

The relevance of combining the CbCR (DAC4) and the public data for compliance aspects was stressed, both for companies (which prepare the CbCR reporting and use it for the public CbCR), and for tax authorities (who use data for risk assessment).

·Relevance of DAC6: scope, and analysis of CJEU decisions

On the question of legal professional privilege (LPP), experts agreed that lawyers benefit from a favourable approach in the EU Charter of Human Rights, on which the CJEU took its decisions about DAC6 on the matter in 2022 and 2024.

More specifically, it appears that in the latter decision of the CJEU (Belgian Association of Tax Lawyers and Others (see C-623/22 ), the analysis of the application of legal professional privilege 99 under DAC6 by the Court seems to create three different categories of intermediaries, with rights that are not similar:

-non-lawyer intermediary (who cannot ensure legal representation of their client in court) subject to DAC6 obligations (subject to reporting obligations / no waiver);

-non-lawyer intermediary bound by LPP who has mandate to ensure legal representation of their clients in court, who benefits from the waiver, and must notify other intermediaries and/or their client;

-lawyer intermediary who ensures legal representation of their client and who benefits from the waiver due to LPP and has no obligation in terms of notification, except to their client).

The experts stressed that the Court chose a restrictive interpretation of the LPP. Hence, the concept of ‘professional secrecy’ (“secret professionnel”), wider than LPP per se, does not necessarily apply, which may, depending on the Member State, create a difference in treatment between lawyers bound by the LPP and other intermediaries concerned by DAC6 obligations.

In conclusion, the experts stressed the importance of ensuring legal certainty in line with the Court's decisions.

·Relevance and coherence of DAC6: main benefit test and hallmarks

I)The definition of the main benefit test (MBT) and some specific hallmarks were discussed, and their flaws or inconstancies described in detail by the experts.

As far as the main benefit test is concerned, the question of its definition is the most open to criticism, mainly because of the different interpretations within the E.U. The lack of harmonisation in the definition of the Main Benefit Test (MBT) in DAC6 (Directive on Administrative Cooperation) is considered a shortcoming for several reasons:

§Inconsistent application: different EU Member States may interpret and apply the MBT differently, leading to inconsistencies in what is considered a reportable arrangement. This can create confusion for multinational companies operating across multiple jurisdictions.

§Increased compliance burden: as a consequence, companies may face increased compliance costs as they need to understand and comply with varying interpretations of the MBT in each Member State. This can be particularly challenging for businesses with complex cross-border arrangements.

§Legal uncertainty: the lack of a uniform definition can lead to legal uncertainty for taxpayers and intermediaries. They may struggle to determine whether a particular arrangement needs to be reported, potentially leading to either over-reporting or under-reporting.

§Ineffective enforcement: without harmonisation, the effectiveness of DAC6 in combating tax avoidance and ensuring tax transparency may be compromised. Different interpretations can lead to loopholes that can be exploited by those seeking to avoid taxes and more specifically DAC6 reporting obligations.

§Competitive disadvantages: intermediaries and businesses operating in Member States with stricter interpretations of the MBT may find themselves at a competitive disadvantage compared to those in Member States with more lenient interpretations

Overall, harmonising the definition of the MBT across the EU would help ensure a more consistent, fair, and effective application of DAC6, possibly reducing compliance burdens and legal uncertainties for businesses.

While setting a threshold in terms of the tax advantage obtained or a ratio could provide a quantitative measure, these solutions might also be circumvented by structuring transactions just below the threshold. Additionally, such measures might not capture the qualitative aspects of tax planning that the MBT aims to address.

The concept of MBT has also been rendered complex to apply in certain Member States insofar as it seems to imply a form of intentionality, a factor which may trigger criminal proceedings in the event of a tax audit, if the arrangement declared proves to be possibly illegal. This assumption tends to create a form of reluctance from intermediaries, with regard to the concept as such.

II)Some hallmarks appear to be more problematic than others.

The comments focused on certain B (e.g. B2 and the definition of ‘conversion of income’), C and E (e.g. E3 and the concept of EBIT) hallmarks, confirming what has been reported by some Member States and private stakeholders previously. Experts also insisted that at this stage it would be preferable to clarify the existing hallmarks to improve their efficiency, and that this should be one of the Commission’s priorities on DAC6.

When asked whether arrangements involving countries / territories outside the EU were to be included in the DAC6 report, it was noted that the lack of mention to these countries outside the EU may be a drafting problem of the DAC and an editorial oversight in Article 8ab (14) (g) 100 .

·Improving clarity and coherence: need for harmonised guidance

The idea of common and harmonised guidance to establish a body of doctrine on DAC (and more specifically DAC6) was favourably received.

Concerning its legal value, although this cannot be enforced (for example during a tax audit), the experts believe that it would nevertheless provide clarification of problematic terms and definitions, particularly regarding hallmarks, since concrete examples could be shared with the public.

This would also allow Member States’ administrative instructions to be brought more consistent together and would help to clear up questions which at this stage lead to an excess of administrative burden.

It would still be up to the CJEU to have the last word if an interpretation based on this guidance were challenged.

·Simplification

Both experts stressed that the scope and objectives of recent legislations like Pillar 2 are different from those of DAC in general, and DAC6 in particular.

Overall, the experts were very cautious about any additional amendment or waiver to specific categories or taxpayers in DAC, stressing the need for legal certainty and coherence over time.

·Use of automatic exchange of information

The participants did not provide any specific comment to this point, leaving it to the Member States to define the appropriate policies in this area.

Related documents – List of questions sent to the experts

On deterrence effect of Automatic exchange of information provided for in the DAC

Do you think that CRS standards (DAC2), the obligation to report advance rulings (DAC3), the country-by-country reports (DAC4), and the obligation to report aggressive cross-border arrangements (DAC6) are likely to deter certain taxpayers from using tax schemes that may affect the Member States' tax bases?

Do you have any comments on taxpayers’ compliance behaviour, and the interaction between taxpayers’ behaviour and the reporting obligations under DAC?

Relevance of DAC6: scope and analysis of CJEU decisions

The European Union has a special reporting obligation for aggressive arrangements (DAC6). However, this exchange of information has been regularly challenged by certain tax lawyers, the professions of advice, law and number, and certain professional organisations. The European Court of Justice has recently upheld the principles and the robustness of the mechanism in place, while protecting clearly (but only) the legal professional privilege - professional secrecy - of lawyers (C623/22).

What is your analysis about the CJEU decisions on DAC6 (C-694/20, C623/22)?

The Directive was amended in 2023 (to limit notification obligations), do you think it needs to be amended again?

What points in DAC6 could be improved in view of these decisions?

In particular, are there any hallmarks that could be reviewed to make the legislation more specific?

Relevance of DAC6: Annex IV of DAC; MBT - Hallmarks

Have you comments on hallmarks?

Should arrangements involving countries / territories outside the EU make it mandatory for those jurisdictions to be included in the DAC6 declaration?

Should there be new hallmarks in DAC6, and if so, which ones?

Use of automatic exchange of information

Questions have been raised by the stakeholders about the use of automatic exchange data by administrations, particularly DAC6.

What is your view?

Should the use of DAC6 data by Member States be made more explicitly mandatory in the Directive?

Ways of improving clarity and need for guidance

Would it be relevant for DAC6 (e.g interpretation of hallmarks)?

What could be the legal value of such a document?

Simplification

Would be legally relevant to provide for an exemption of DAC obligations for MNEs affected by Pillar 2 reporting?

(1) The scope of the DAC encompasses all taxes of any kind with the exception of value added tax, customs duties, excise duties and compulsory social contributions because these are already covered by other EU legislation on administrative cooperation.

(2) Commission staff working document: Evaluation of the Council Directive 2011/16/EU, 12 February 2019 .

(3) For 2024 only very limited data is available and is therefore not covered in the Evaluation. However, where data is available it is presented in the analysis.

(4) The present evaluation follows and fulfils the Commission’s Better Regulation requirements as updated in July 2023. See “Better regulation” guidelines (2021) and toolbox (2023): Better regulation: guidelines and toolbox .

(5) Further details on the evaluation methodology are provided in Annex II.

(6) Including financial institutions, multinational enterprises (MNEs), tax advisors, law firms and lawyers, accountants, professional organisations, lobbies registered in the transparency register.

(7) A graph depicting the intervention logic of the Directive described in this subsection is provided in Annex VI.

(8) For further details, please see subsection 3.1.3.

(9) The report on the first evaluation of the Directive was published in December 2017, accompanied by a staff working document . A second report under Article 8b of the Directive, which covers only the automatic exchange of information mechanisms, was published in December 2018.

(10) See details in Annex VII: Legal developments from 2018.

(11) Further details are provided in Annex VII.

(12) C-694/20 - 08/12/2022 - Orde van Vlaamse Balies e.a.

(13) C-623/22 – 29/07/2024 - Belgian Association of Tax Lawyers and Others v Premier ministre/ Eerste Minister .

(14) C-432/23 – 26/09/2024 - F and Ordre des avocats du barreau de Luxembourg v Administration des contributions directes .

(15) Quantitative information on the volume of exchanges derived from joint audits will not be presented as the latter were introduced by DAC7 and started as of 1 January 2024.

(16) The 2019 evaluation study found that Member States had exchanged, between 2015 and mid-2017, around 11,000 messages referring to nearly 16 million taxpayers and to incomes/assets worth more than EUR 120 billion.

(17) The message year is the year in which information is exchanged between Member States. It is different from the tax year, which is the year to which the information refers.

(18) Cf. Article 5 of DAC (Procedure for the exchange of information on request) and 5a (Foreseeable relevance). To align with international standards for the EOIR and the conclusions of the CJEU in Berlioz Investment Fund SA v Directeur de l'administration des contributions directes (Case C-682/15) and B v État Luxembourgeois and État Luxembourgeois v B, C, D and FC (Cases C-245/19 and C-246/19), allowing tax authorities in one country to obtain information from another country’s tax authorities upon request, the concept of foreseeable relevance was clearly codified in DAC7 amendment (2021).

(19) Commission Implementing Regulation (EU) 2022/1467 of 5 September 2022 amending Implementing Regulation (EU) 2015/2378 as regards the standard forms and computerised formats to be used in relation to Council Directive 2011/16/EU and the list of statistical data to be provided by Member States for the purposes of evaluating that Directive.

(20) See chapter 4.1.1.3 on the benefits.

(21) Special Report No 03/2021: Exchanging tax information in the EU: solid foundation, cracks in the implementation .

(22) Report on the implementation of the EU requirements for exchange of tax information: progress, lessons learnt and obstacles to overcome (2020/2046(INI)), Committee on Economic and Monetary Affairs, A9-0193/2021, 3.6.2021.

(23) European Parliament legislative resolution of 10 March 2021 on the proposal for a Council Directive amending Directive 2011/16/EU on administrative cooperation in the field of taxation ( COM(2020)0314 – C9-0213/2020 – 2020/0148(CNS) ).

(24) Special Report 27/2024: Combatting harmful tax regimes and corporate tax avoidance – The EU has established a first line of defence, but there are shortcomings in the way measures are implemented and monitored , November 2024.

(25) Replies of the European Commission to the European Court of Auditors Special Report , November 2024.

(26) Minutes WG ACDT, 51st Meeting – Virtual Meeting, 3rd March 2022,

(27) Yearly Assessment 2024 (year 2023)

(28) According to the Yearly Assessment 2023 shared with the Commission in 2024, 21 out of 27 Member States had instruments to assess whether they received all due reports from other jurisdictions, including other Member States and third country jurisdictions.

(29) In particular, following issues have been raised by tax authorities:
missing information (particularly regarding the TIN);
missing or unclear indication about or inconsistent use of currency;
missing or unclear indication of the start and end date of the reporting period;
notation mistakes (e.g. data included as thousands or millions instead of units; inconsistent use of negative sign);
clerical mistakes (e.g. negative number of employees or share of capital).
To address this issue, several Member States highlighted the need for an automatic validation of data (e.g. rejecting CbCRs if certain information is missing or if the formats do not correspond). It is however worth noting that modifications on the CbCR format must be channelled through an OECD format update.

(30) FPG037: Improving the use of DAC data.

(31) ECA special report 27/2024 , recommendation 2: To maximise the benefits of the automatic exchange of DAC 6 information with other Member States, the Commission should make the data field for non-EU countries involved in cross-border arrangements mandatory in the reporting schema (if needed by a legislative proposal). Moreover, the Commission should update the DAC 6 central directory architecture to make this information available to Member States for every arrangement concerned.

(32) See Annex XII - Targeted consultations with academics – tax lawyers, September 2024.

(33) Source: Yearly Assessment 2024 (year 2023).

(34) Source: Yearly assessment questionnaire on the AEOI, year 2022, 15 April 2023.

(35) Questionnaire to the Member States on the functioning of the Directive (2024).

(36) Source: AEOI Yearly assessment 2023

(37) Questionnaire to the Member States on the functioning of the Directive (2024).

(38) For example, in the Netherlands, the sanction, where the reporting of a DAC6 arrangement is not timely, complete or correct, is a maximum penalty of EUR 870,000.

(39) Expert team on DAC, visits in the Member States (2023-2024).

(40) Annex XII - Targeted consultation with Tax law academics, September 2024.

(41) Summary report available here: https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/13678-Cooperation-on-direct-taxation-evaluation/public-consultation_en .

(42) See Joshi, P., Outslay, E., Persson, A., Shevlin, T. and Venkat, A. (2020), Does Public Country-by-Country Reporting Deter Tax Avoidance and Income Shifting? Evidence from the European Banking Industry. Contemp Account Res, 37: 2357-2397 ( https://doi.org/10.1111/1911-3846.12601 ); Müller, R., Voget, J. & Zental, J. The Effects of Mandatory Private Disclosure On Public Disclosure—Evidence from CbCR. Schmalenbach Journal of Business Research 76, 533–571 (2024), https://doi.org/10.1007/s41471-024-00194-2 .

(43) Information communicated between Member States pursuant to Article 8aa of the Directive shall be used for the purposes of assessing high-level transfer-pricing risks and other risks related to base erosion and profit shifting, including assessing the risk of non-compliance by members of the MNE Group with applicable transfer-pricing rules, and where appropriate for economic and statistical analysis. Transfer-pricing adjustments by the tax authorities of the receiving Member State shall not be based on the information exchanged pursuant to Article 8aa.

(44) The methodology used to estimate tax benefits from automatic exchange of information, as well as non-AEOI activities, includes validating and cleaning reported data through consistency checks and adjusting or excluding outliers. It also involves assessing data coverage across Member States using weighted economic indicators and extrapolating EU-wide benefit ranges only when sufficient coverage is achieved. However, the methodology acknowledges limitations in disaggregating results and estimating impacts for DAC3–DAC6 due to their indirect links with tax assessments. For more details, see Annex IV: Overview of benefits and costs, based on the study carried out for the evaluation of the DAC.

(45) E.g the use of simultaneous controls, which last an average of 18 months but may have an impact that exceeds tens of millions of euros for a single Member State

(46) The central estimate was chosen as the most plausible, given that Member States' use of data is constantly increasing and therefore tends to give rise to greater fiscal impacts. This is particularly true for DAC4 data, the use of which is limited to risk assessment by the Directive and consequently requires a process lasting more than one year, involving risk analysis and the implementation of tax audits, followed by the establishment of new tax bases in the event of anomalies being identified, the consequences of which can amount to several hundred million euros in the case of transfer pricing adjustments.

(47) See 2021 data from DG TAXUD, Annual report on taxation and Eurostat.

(48) See Study for the Evaluation of the Directive 2011/16, 18 December 2023, Annex 2 , pp. 60-61

(49) Sample of 17 Member States responding to the Yearly Assessment 2024 (data 2023).

(50) Under public disclosure, both tax authorities and the public at large are better informed about taxes paid across jurisdictions, as MNEs make their CbCRs available in the public domain. Over the period of the evaluation, this was only required for extractive industries and large banks in the EU. As of 2025, other MNEs are subject to public disclosure rules. Under public CbCR, MNEs thus face both possible enforcement actions and reputational effects.

(51) Under private disclosure (i.e. non-public, only to tax authorities and exchanged between relevant tax authorities), as provided by DAC4, CbCRs are directly transmitted to the tax authority of the Member State of main establishment. The tax authority then share the report to the other Member States concerned. This may lead to further investigations and possible tax audits and reassessments, depending on the result of risk analysis and administrative enquiries.

(52) Michael Overesch, Hubertus Wolff: Financial Transparency to the Rescue: Effects of Public Country-by-Country Reporting in the European Union Banking Sector on Tax Avoidance, https://doi.org/10.1111/1911-3846.12669 .

(53) Preetika Joshi , Edmund Outslay , Anh Persson , Does Public Country-by-Country Reporting Deter Tax Avoidance and Income Shifting? Evidence from the European Banking Industry, https://doi.org/10.1111/1911-3846.12601.

(54) In particular, the cost-benefits comparison is based on the following principles: (i) the scope of costs and benefits is homogeneous, i.e. the costs of the various DAC provisions are considered only when the benefits could be quantified; (ii) recurring burdens are compared to recurring benefits, excluding set-up costs.

(55) In particular, financial institutions as regards DAC2 and MNEs (turnover > €750 million) concerning DAC4.

(56) See: Ramboll study, Volume 2 – Annexes, 9-4 Costs for private stakeholders.

(57) The methodology used in the study combines targeted surveys and interviews with financial institutions and federations to gather quantitative and qualitative data on administrative burdens from DAC2 and DAC4. For DAC2, data from 18 entities—including large and small institutions across six EU Member States—were used to estimate onboarding time, set-up, and ongoing costs, monetised using Eurostat wage statistics. For DAC4, input from 11 entities (including MNEs and federations) helped assess cost categories like IT, consultancy, and training, with segmentation by company size and complexity to reflect varying compliance efforts. See more details on methodology applied in Annex IV: Overview of benefits and costs.

(58) The estimate of burdens on public authorities should be considered as very robust, since all or nearly all Member States are able to provide detailed estimates of these costs. As for private stakeholders, data originate from dedicated interviews and surveys; however, for certain DACs, the number of interviewees is limited, at about 10 to 15 entities.

(59) According to the Yearly Assessment survey of 2024, the overall estimated recurring costs for AEOI were below EUR 30 million.

(60) During the targeted consultation in the context of the Study, 20 out of 23 tax authorities agreed or partly agreed to the following statement: “for each DAC, significant costs are incurred for implementation and set-up, then total costs decrease in the subsequent years”.

(61) In 2023, 89% of the Fiscalis program budget was allocated to the development and operation of the common components of the European electronic systems, which includes the Common Communication Network (CCN) .

(62) Please consult Annex V for further details.

(63) See Fig. 23 and more details on methodology applied in Annex IV: Overview of benefits and costs.

(64) See Overesch M. and H. Wolff, Financial Transparency to the Rescue: Effects of Country-by-Country Reporting in the EU Banking Sector on Tax Avoidance Contemporary Accounting Research, Vol. 38 No. 3, Fall 2021; Joshi P., Outslay E. and A. Personn, Does Public Country-by-Country Reporting Deter Tax Avoidance and Income Shifting? Evidence from the European Banking Industry, Contemporary Accounting Research, Vol. 37 No. 4, Winter 2020.

(65) The low-range scenario shows net benefits of around €4.2 billion and a cost-effectiveness ratio close to 8. The comprehensive scenario indicates that the net benefits amount approximately 6.1 billion, with a high cost-effectiveness ratio of 10. See fig. 23. and details on methodology applied in Annex IV: Overview of benefits and costs.

(66) https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/13678-Cooperation-on-direct-taxation-evaluation/public-consultation_en .

(67) Fiscalis Workshop ‘DAC6 implementation differences’ FWS/153/001-003, 26 July 2021, Fiscalis Workshop ‘DAC6 implementation differences, part two’, FWS/153/004-005, 2-4 and 9 March 2021, Fiscalis 2021 Digital Workshop ‘DAC 6 implementation differences, part three’ FWS/153/006-007 24, 25 and 29 November 2021.

(68) Depending on the DAC exchange of information, between 5% and 15% of data cannot be identified automatically, which creates administrative burden for the tax authorities.

(69) For a full analysis of DAC6 and the hallmarks, see the summary of the questionnaire to Member States, WPIV - September 2024 (Annex VIII)) and the elements from the OPC (Annex IX), as well as the targeted consultations (Annex XII).

(70) EUROPEAN COURT OF AUDITORS, Special Report 27/2024: Combatting harmful tax regimes and corporate tax avoidance – The EU has established a first line of defense, but there are shortcomings in the way measures are implemented and monitored.

(71) EUROPEAN COURT OF AUDITORS, https://www.eca.europa.eu/Lists/ECADocuments/SR21_03/SR_Exchange_tax_inform_EN.pdf .

(72) See subsection 3.1.5.

(73) This assessment is without prejudice to a possible future Impact Assessment.

(74) Working Group ACDT and Subgroup-AEOI.

(75) France, Germany, Ireland, Italy, Spain, the Netherlands, Poland and Romania.

(76) Austria, Cyprus, Finland, Latvia, Luxembourg and Slovenia.

(77) The integral report of the call for evidence and the public consultation is published at this link: Cooperation on direct taxation - evaluation .

(78) The number of contributions as feedback differs from the statistics of the feedback online. As a matter of fact, 13 contributions were not included because they were the same contributions sent through the OPC Annex, and therefore are counted only once under OPC.

(79) The list of questions addressed to the experts with the integral minutes can be found in Annexes XI and XII.

(80) The integral report of the meeting of the Platform for good governance held on 17 December 2024 is available at this link: Platform for Tax Good Governance - Commission .

(81) C-694/20 - 08/12/2022 - Orde van Vlaamse Balies e.a.

(82) Council Directive (EU) 2023/2226 of 17 October 2023 amending Directive 2011/16/EU on administrative cooperation in the field of taxation: Article 8ab is amended as follows: (a) in paragraph 5, the first subparagraph is replaced by the following: ‘5. Each Member State may take the necessary measures to give intermediaries the right to a waiver from filing information on a reportable cross-border arrangement where the reporting obligation would breach the legal professional privilege under the national law of that Member State (unchanged) In such circumstances, each Member State shall take the necessary measures to require any intermediaries that have been granted a waiver to notify, without delay, their client, if that client is an intermediary or, where there is no such intermediary, that client is the relevant taxpayer, of that client’s reporting obligations under paragraph 6.’

(83) Belgian Association of Tax Lawyers, C-623/22

(84) Directive 98/5/EC of the European Parliament and of the Council to facilitate practice of the profession of lawyer on a permanent basis in a Member State other than that in which the qualification was obtained.

(85) Orde Van Vlaamse Balies e.a., C-694/20

(86) Belgian Association of Tax Lawyers, C-623/22

(87) In Case C‑623/22, the Court of Justice of the European Union considered DAC valid considering the principles of legal certainty, legality in criminal matters and the right to respect for private life as provided for in the article 7 of the Charter (EU charter of fundamental rights). It also considered that DAC does not infringe the principles of equal treatment and non-discrimination in extending the reporting obligations of certain provisions to any type of tax subject to potentially aggressive tax planning. It has been confirmed in Case C-432/23 DAC, the Court considering that DAC is compatible with articles 7 and 52(1) of the EU charter.

(88) A total of 55 stakeholders participated in the consultation process. More specifically, 28 stakeholders and 2 NGOs answered the public consultation based on a questionnaire focusing on AEOI that addressed the different instruments provided for in the DAC. The Commission also received 25 written feedback contributions to the call for evidence.

(89) Best practices recommended by the BEPS action 12 led to the Model Mandatory Disclosure Rules for Addressing CRS Avoidance Arrangements and Opaque Offshore Structures | OECD iLibrary (oecd-ilibrary.org)

(90) Commission view: it should be noted that the Directive does not provide for the application of a ratio.

(91) Commission view: This point may be relevant. However, the analysis of a risky transaction where there is a potentially tax harmful cross-border arrangement results from the intention of the Council, which with DAC6 established the framework to protect Member States against such actions.

(92) Dividend stripping is a series of transactions based on a complex mechanism of trading, selling and repurchasing shares over a certain period to avoid payment of dividend taxes, or to claim tax reimbursement (OECD (2023), Dividend Tax Fraud: Raising Awareness of Dividend Stripping Schemes, OECD Publishing, Paris, https://doi.org/10.1787/70ee934c-en .

(93) In transfer pricing policies, the arm’s length principle determines that the conditions agreed upon for inter-co transactions between entities that are part of a group must be similar to ones conducted between companies that are not under common ownership or control. If these conditions are not at arm’s length, a profit correction may be applied to take arm’s length conditions into account. The arm’s length principle is not necessarily applied in exactly the same way in different countries. In the case of internationally operating groups, this may lead to mismatches.

(94) https://taxation-customs.ec.europa.eu/online-services/online-services-and-databases-taxation/taxpayer-identification-number-tin_en.

(95) OECD Peer Review of the Automatic Exchange of financial account information (2022).

(96) DAC data can be used for assessment, administration and enforcement for the following: direct taxes, tax recovery, VAT, other indirect taxes, compulsory social security contributions, customs and other duties. As from January 2026, DAC AEOI can also be used for countering Money Laundering/Terrorist Financing, and for any act covered by Article 215 of the EU Treaty, including sanctions.

(97) DAC6 has been the subject of two lawsuits before the CJEU. ( C-694/20 , C623/22 ).

(98) Directive (EU) 2021/2101 of the European Parliament and of the Council of 24 November 2021 amending Directive 2013/34/EU as regards disclosure of income tax information by certain undertakings and branches. Public CbCR is linked with the DAC as it provides for that Member States shall permit the information to be reported on the basis of the reporting instructions referred to in Section III, Parts B and C, of Annex III to Directive 2011/16/EU.

(99) “intermediary bound by LPP subject to criminal sanctions under the national law of that Member State”

(100) “The information to be communicated by the competent authority of a Member State under paragraph 13 shall contain the following, as applicable: (…) the identification of the Member State of the relevant taxpayer(s) and any other Member States which are likely to be concerned by the reportable cross-border arrangement”.