EBA ¦ Peer Review on Tax Integrity and Dividend Arbitrage Schemes

Tax Integrity and Dividend Arbitrage: How AML/CFT Supervision Is Adapting

Why dividend arbitrage matters for AML/CFT supervision

Dividend arbitrage schemes (for example, so‑called cum‑ex and cum‑cum transactions) exploit dividend taxation and refund mechanisms to create unjustified tax refunds. Even when a scheme’s criminal classification varies across jurisdictions, its use of the financial system can generate large illicit proceeds and create material legal, operational and reputational risks for banks. That makes tax integrity — the obligation of firms to prevent involvement in tax fraud and to avoid becoming channels for laundering proceeds of tax crimes — a legitimate and important focus for both prudential and AML/CFT supervisors.

Following the EBA action plan of April 2020, EU prudential and AML/CFT guidance was updated to require supervisors to consider tax issues in governance, suitability assessments, Supervisory Review and Evaluation Process (SREP), ML/TF risk factors and risk‑based AML/CFT supervision. In 2024 the EBA carried out a targeted peer review across six Member States (Denmark, France, Germany, Ireland, the Netherlands and Spain) to assess how supervisors implemented these changes and whether their supervisory approaches are proportionate to the national ML/TF risk posed by tax crimes.

What the peer review checked

The peer review examined four supervisory benchmarks:

• Whether AML/CFT authorities integrate tax integrity into risk‑based supervisory work and cooperate with tax authorities.
• Whether tax integrity is embedded into sectoral and institution‑specific ML/TF risk assessments, and whether supervisors collect and use tax‑related data.
• Whether supervisors assess how institutions incorporate tax integrity into internal governance, tone from the top, codes of conduct and risk management.
• Whether tax integrity is considered in fit‑and‑proper (F&P) checks of management bodies and key function holders (KFHs), and how supervisors cooperate with other authorities during those checks.

Key findings — overall picture

Supervisors generally adjusted frameworks and practices after the EBA action plan; most authorities were graded as largely or fully meeting the benchmarks. However, the peer review also identified gaps and unevenness across jurisdictions and across the four benchmarks. Important patterns emerged:

• Legal and institutional variation matters. Supervisory outcomes depend on national laws, organisational set‑ups and the scope for information exchange with tax authorities, FIUs and prosecutorial bodies. No single approach fits every Member State.
• Information sharing and practical cooperation are essential. Where supervisors had regular, operational cooperation with tax authorities, FIUs and multi‑agency forums — and where those forums included private sector partners — supervisory assessments and follow‑up actions were stronger and more consistent.
• Data collection from supervised firms strengthens risk assessment. Supervisors that systematically ask banks for tax‑related data (e.g., non‑resident client volumes, cross‑border flows, STR counts with tax suspicions, controls for CRS/withholding tax) can better calibrate sectoral and bank‑level ML/TF risk ratings.
• Governance and suitability checks have improved but can be more targeted. Most supervisors assess governance, codes of conduct and F&P criteria for ties to tax misconduct, but methodology and depth vary. In particular, relying only on absence of criminal convictions risks missing serious integrity concerns that have not led to conviction.

Bastian Schwind-Wagner_Follow "Banks must treat tax integrity as a core element of their AML/CFT and governance frameworks, ensuring management‑level ownership, robust controls and timely reporting of tax‑related suspicions. Supervisors expect proportional, data‑driven oversight and clear policies that distinguish legitimate tax planning from abusive schemes."

Country contrasts and practical examples

• France: The Autorité de Contrôle Prudentiel et de Résolution (ACPR) scored among the strongest performers. It collects tax‑related data, leverages FIU and tax authority inputs, runs targeted on‑site checks (notably CRS controls), and used ECB‑led horizontal surveys on dividend arbitrage to trigger supervisory action. The ACPR also has concrete SREP practices that incorporate tax risks in risk appetite and governance checks.
• Netherlands: De Nederlandsche Bank (DNB) combines an AML/CFT questionnaire with inputs from the Financial Expertise Centre (a multi‑agency partnership) and uses a tax‑related knowledge document for supervisors. DNB’s approach is data‑driven and strongly integrated into risk assessments, although some interactions remain signal‑driven rather than proactive.
• Spain: SEPBLAC (the FIU/AML supervisor) built a dedicated tax unit and seconded staff from the tax authority, which enhanced tax expertise in supervision. That unit supports data collection and knowledge sharing; SEPBLAC and the Banco de España (BdE) cooperate in the national AML committee. Supervisory activity could be expanded in light of rising STRs with tax components.
• Germany: BaFin has institutional structures (single points of contact) and legacy lessons from past cum‑ex cases, and it actively coordinates with prosecutors and AFCA (a PPP). Nonetheless, BaFin collects less structured tax data directly from banks than some peers, and its approach often treats these matters as prudential rather than AML first, which limits consistent AML/CFT data collection and preventive outreach.
• Denmark and Ireland: Both authorities participate in multi‑agency cooperation. Denmark’s DFSA benefits from operational forums (e.g., ODIN) and legal pathways for exchanges but still needs to scale supervisory activity to the high tax‑crime risk identified in its national risk assessment. Ireland’s CBI maintains strong bilateral ties with Revenue and the FIU and will expand tax‑data collection in its redesigned AML questionnaire; its current assessment is that dividend arbitrage presents low local ML/TF risk, but it must validate that assessment with new data.
• Cross‑cutting note on STRs: All jurisdictions require banks to file STRs/SARs for suspected tax crimes; supervisors use STR statistics in different ways. Where statistics are disaggregated (e.g., STRs with an explicit tax suspicion) and shared with supervisors, that improves the identification of sectoral and bank‑level risk.

Practical supervisory strengths and weaknesses

Strengths

• Use of multi‑agency public‑private partnerships (AFCA, FEC, ODIN) to share typologies and operational information.
• Concrete guidance from some supervisors (ACPR, DNB) clarifying expected controls and red flags for banks.
• Secondment of tax experts to AML/CFT supervisory bodies (Spain) and tax training of supervisors (France).
• Inclusion of tax factors in SREP governance modules and in AML/CFT risk questionnaires by several supervisors.

Weaknesses and gaps

• Not all supervisors collect structured tax‑integrity data from banks (notably Germany and until recently Ireland), limiting quantitative risk calibration.
• Some supervisory work remains reactive, triggered by leaks, litigation or foreign investigations, rather than preventive and proportionate to national risk.
• Fit‑and‑proper assessments often focus on convictions or formal proceedings; they may miss integrity concerns where no conviction exists but where conduct indicates serious tax‑integrity risk.
• A lack of common, practical guidance across jurisdictions on “acceptable versus unacceptable” tax behaviour makes supervisors’ expectations less predictable for banks.

What supervisors should do next (recommended supervisory steps)

• Ensure operational, timely information exchange with tax authorities and FIUs. Supervisors should establish clear, documented channels for receiving and acting on tax‑related intelligence and for sharing supervisory findings that could support tax or criminal investigations, respecting legal confidentiality constraints.
• Collect tax‑relevant data from supervised institutions on a risk‑sensitive basis. Questionnaires should gather information that helps reveal tax‑integrity exposure: non‑resident client volumes, jurisdictions of exposure, cross‑border flows above thresholds, counts and quality indicators of STRs with tax suspicions, CRS compliance controls and relevant product lines (private banking, wealth management). Use this data in sectoral and institution‑specific ML/TF risk assessments.
• Calibrate supervisory intensity to national risk. Where national risk assessments or FIU data show significant tax‑crime exposure, supervisors must allocate staff, run thematic reviews or on‑site inspections and publish targeted guidance (including red flags) to raise banks’ detection and reporting capabilities.
• Strengthen in‑house tax expertise. Supervisors should develop or retain tax specialists, run regular cross‑training with tax authorities, and — where appropriate — second tax officials into supervisory teams to bridge knowledge gaps.
• Make governance reviews more explicit on tax conduct. SREP and governance assessments should explicitly verify whether boards and senior management define, approve and monitor policies that identify acceptable and unacceptable tax behaviours, integrate tax risk metrics and ensure escalation and remediation when issues arise.
• Improve F&P processes for integrity risks. Fit‑and‑proper checks should go beyond advertising criminal records; they should systematically screen public and administrative databases (EBA/ESMA registers), request criminal records from foreign jurisdictions where relevant, query tax authorities when permitted by law, and apply proportional scrutiny when allegations exist even without conviction.
• Issue practical guidance and red flag lists. Where possible, publish concrete indicators and case studies (anonymised) so banks understand what behavior will attract supervisory attention: unusual refund claims, complex temporary securities transfers around dividend dates, rapid use of non‑resident vehicles without economic rationale, and weak CRS/withholding tax controls.
• Leverage multi‑agency partnerships. Use established forums and PPPs to exchange typologies, signal emerging schemes, and to coordinate supervisory follow‑up and outreach to industry.

Why this matters for banks and compliance teams

Banks should expect supervisors to ask for more granular tax‑integrity data and to evaluate governance, controls and the role of management in preventing tax‑related misuse. Compliance teams need to ensure customer due diligence, transaction monitoring and STR reporting capture tax‑related red flags and that senior management understands tax‑integrity risk and sets a clear zero‑tolerance tone where misconduct is suspected. Firms should also review code‑of‑conduct language, ensure adequate training and document how the institution distinguishes acceptable tax planning from schemes that are abusive, opaque or criminal.

Concluding practical takeaway

Tax integrity is no longer only a matter for tax authorities; it is an integral supervisory concern for both AML/CFT and prudential supervisors because tax‑fraud schemes can create significant ML/TF, operational and reputational risk for banks. Supervisors are moving toward more structured data collection, closer cooperation with tax authorities and clearer expectations for governance and F&P checks. Banks should anticipate more explicit supervisory scrutiny, adopt clearer internal tax‑conduct policies, and strengthen systems for identifying, escalating and reporting tax‑related suspicions.

Dive deeper

• EBA ¦ Reviewed supervisors overall applied the EBA’s recommendations on tax integrity and dividend arbitrage trading schemes, the EBA Report finds. ¦ Link