Why Does the EU Lose Billions in VAT?

What is the VAT gap, how is it measured, and what does the European Commission’s latest report actually tell us?

Paweł Mikuła talks to Grzegorz Poniatowski – lead researcher behind the EU VAT Gap Report 2025 – about the difference between compliance and policy gap, why the gap dropped for years and started rising again after COVID, and which tools like SAF-T and e-invoicing are actually making a difference.

This article summarizes a podcast conversation between Paweł Mikuła, partner and co-founder of “Halcyon | Tax. Customs. Legal., and Grzegorz Poniatowski, founding partner of Syntesia Policy and Economics and lead researcher behind the European Commission’s VAT Gap Report 2025. The discussion explores what the VAT gap actually is, how it is measured, what the latest report reveals about trends across Europe, and what conclusions can be drawn for tax administrations and policymakers.

The VAT Gap in Europe: Methodology, Findings, and What the EC Report 2025 Actually Tells Us

The Foundation: What Is the VAT Gap?

VAT as a Broad-Based Consumption Tax

To understand the VAT gap, one must start with the design of VAT itself. By assumption, VAT is a broad-based tax on consumption — or, in IMF classification, a tax on goods and services. From a purely fiscal perspective, in an ideal world, VAT would be levied at the standard rate on the entire tax base, which in economic terms means all final consumption, primarily consisting of household expenditure.

This design creates a natural counterfactual: a benchmark against which the current state of revenue collection can be compared. The distance between that benchmark and reality is, in broad terms, the VAT gap.

Compliance Gap and Policy Gap

The VAT gap is composed of two distinct components, each with a different nature and different drivers.

The compliance gap captures the difference between what taxpayers should pay under existing law and what they actually pay. It encompasses fraud, evasion, avoidance, but also unintentional errors and omissions. It reflects the degree to which taxpayers fail — for whatever reason — to meet their legal obligations.

The policy gap, by contrast, captures the difference between what would be collected if the entire consumption base were taxed at the standard rate, and what is actually collected under the existing system of reduced rates, exemptions, and other policy choices. It reflects deliberate political and legislative decisions, not taxpayer behaviour.

Together, these two components form the total VAT gap. A third concept worth noting is collection efficiency, sometimes called C-efficiency, which shows, as a percentage, how much revenue the system actually collects compared to what would be collected in the theoretical ideal scenario. On average across EU member states, this figure sits slightly below 50%.

Methodology: How Is the VAT Gap Calculated?

The Top-Down Consumption Approach

The methodology used in the EC VAT Gap Report is known as the top-down consumption approach. It begins with aggregate figures from national accounts — statistical data compiled by national agencies using standardised national accounting principles — and works downward toward a granular representation of who consumed what, and at which VAT rate.

The starting point is provided by statistical agencies such as Eurostat, but significant additional work is required. The data must be rescaled, realigned, and broken down into more granular categories to reflect the complexity of VAT rules. In practice, this involves calculating over 10,000 parameters per year across all covered countries.

A key clarification concerns why the consumption side is used as the basis for calculation. In theory, if all VAT in the value-added chain were fully recovered, the entire liability could be read simply from final consumption, since intermediate VAT transfers cancel out. In practice, however, exemptions without the right to deduct — particularly in sectors such as healthcare, financial services, and the public sector — mean that some VAT becomes embedded and irrecoverable within the chain. Between 60% and 70% of the calculated theoretical liability is attributable to household final consumption; the remaining 30–40% reflects VAT hidden in investment and intermediate transactions in exempt sectors.

Alignment with VAT Law

National accounts data does not map directly onto tax law. Consumption figures include components that have no equivalent in the VAT system — such as imputed rents (the notional value of housing consumed by owner-occupiers) and home production. These must be identified and excluded or adjusted before any comparison with actual VAT revenues is meaningful.

To calculate the compliance gap specifically, researchers apply the actual legal VAT rates to each component of consumption, determining what revenue should have been collected if all taxpayers were fully compliant. To calculate the policy gap, the same consumption base is instead taxed entirely at the standard rate, revealing how much is foregone through policy choices.

Alternative Methodologies

The top-down approach is not the only methodology available. An alternative is based on audit results — either risk-based audits, which target specific high-risk taxpayers, or random audit programmes, which allow more statistically representative extrapolation. Some countries with significant resources run random audit programmes precisely to complement or cross-check the national accounts approach.

Each method has advantages and limitations. The top-down approach based on national accounts was chosen for the EC study primarily because it allows consistent coverage across all EU member states — something that audit-based methods, which depend on national programme availability, cannot easily provide.

Key Findings of the 2025 Report

The Long Decline — and Its Reversal

One of the most significant findings concerns the long-term trend in the compliance gap. From 2012 onwards, the gap declined steadily for nine consecutive years — a drop that was both steep and remarkably consistent across the EU. The study team came to regard this downward trend as almost a baseline expectation.

The COVID-19 pandemic, however, produced a surprising result. Despite severe economic disruption — and contrary to the conventional expectation that economic hardship increases non-compliance — the compliance gap continued to fall in 2020 and 2021. This counterintuitive outcome prompted significant analytical attention.

The most plausible explanations identified relate to structural changes in consumption patterns during the pandemic. Sectors traditionally associated with higher non-compliance — hospitality, restaurants, hotels — were effectively shut down, reducing the opportunity for unreported transactions. Simultaneously, the share of electronic payments increased sharply in many countries, and those countries where this increase was largest also showed the greatest drops in the gap.

Following the pandemic period, the gap has since increased slightly in 2022 and 2023 compared to the preceding years. Whether this represents a sustained reversal or a temporary adjustment remains to be confirmed by future reports. Grzegorz Poniatowski expresses cautious optimism that the ongoing efforts of tax administrations will prevent further deterioration.

The Role of Digital Reporting Obligations

A central analytical finding of the report concerns the measurable impact of electronic reporting tools on the compliance gap. The study team examined the introduction of various digital obligations — VAT listings, SAF-T reporting, mandatory e-invoicing, and real-time reporting — and tested their effect on the gap using econometric methods, controlling for economic factors.

The results were statistically significant and consistent across countries. The introduction of SAF-T reporting was associated with an average reduction in the compliance gap of approximately three percentage points. Mandatory e-invoicing and real-time reporting showed an even stronger effect.

Countries in Central and Eastern Europe — including Poland, Hungary, Latvia, Slovakia, Czechia, and Lithuania — showed particularly pronounced declines in their compliance gaps in periods coinciding with the introduction of these tools. This consistency across multiple countries strengthens confidence in the causal interpretation.

A notable exception was Romania, which introduced SAF-T reporting in 2022 but did not show a corresponding decline in its compliance gap — a result that the study team found unexpected and that warrants further investigation.

Italy provided an interesting case study of a different kind. The introduction of mandatory e-invoicing contributed to reducing the gap, but a more dramatic effect came from the Superbonus 110% programme — a large tax credit scheme for building renovation. Because claiming the credit required documented invoices, the construction sector became significantly more transparent almost overnight. The compliance gap in Italy dropped by approximately five percentage points following the programme’s introduction, and has remained at that lower level.

Denmark as a Reference Point

Denmark consistently stands out in the report as the country with the lowest and most stable compliance gap. The reasons are instructive. Denmark applies a single VAT rate of 25% to virtually all goods and services, with only the mandatory exemptions required under the VAT Directive. This flat structure eliminates the need for complex rate calculations and significantly reduces the administrative burden for both taxpayers and researchers.

From a methodological standpoint, calculating the theoretical liability for Denmark requires a fraction of the effort needed for other countries. The compliance gap has remained essentially flat since 2000 — the earliest year covered by the series of studies — suggesting that a combination of simple tax design, high tax morale, and effective administration can produce a system that functions with very little leakage.

The Policy Gap: Structure and Actionability

The 2025 report introduced an important methodological development in the analysis of the policy gap: the concept of the actionable policy gap. Previous studies calculated the policy gap as a broad theoretical measure — the full distance between actual collections and what would be collected at the standard rate on all consumption. This figure, while informative, includes many components that cannot in practice be taxed.

The new analysis disaggregates the policy gap into three parts. The first is the non-actionable portion — consumption that falls outside the reach of the tax system regardless of political will, such as private rental income from non-VAT-registered landlords. The second is the portion under member state discretion — reduced rates and exemptions that individual countries have chosen to apply and could in principle remove. The third is the portion mandated by the VAT Directive itself, which member states have no authority to change unilaterally.

The findings show that more than 50% of the total policy gap falls into the non-actionable category. Approximately 12% of the notional theoretical liability is attributable to national discretionary choices — the area where member states could act. Only around 4% is embedded in the Directive itself.

This decomposition has significant implications. It suggests that while the policy gap appears large in absolute terms, the scope for meaningful action by individual member states is more limited than the headline figures might imply. At the same time, 12% of notional liability still represents a substantial sum across the EU.

The concept of the actionable standard rate illustrates the combined significance of both gaps. On average across EU member states, a flat rate of 15% applied to all actionable consumption would be broadly revenue-neutral compared to the current system — meaning that if all policy choices and all non-compliance were eliminated simultaneously, a 15% rate could replace the existing patchwork of standard and reduced rates while generating equivalent revenue.

Conclusions and Implications for Tax Policy

Measurement as a Precondition for Action

A consistent theme throughout the conversation is the importance of measuring the VAT gap in order to act on it. Without reliable estimates, tax administrations are effectively operating without a thermometer — unable to assess whether their interventions are working or where the greatest losses are occurring.

The 13 consecutive updates of the EC VAT Gap Study have created a dataset of rare depth and consistency, allowing the identification of long-term trends, cross-country comparisons, and the evaluation of specific policy interventions. Grzegorz Poniatowski notes that several national administrations — which had no gap estimates of their own when the study began in 2012 — have since developed their own methodologies, often drawing on the approaches pioneered by the EC study.

A Holistic Approach to Non-Compliance

The compliance gap is not a single phenomenon. It encompasses deliberate fraud, organised evasion schemes such as carousel fraud, opportunistic avoidance, and unintentional errors. Each of these requires different responses. Fraud requires targeted enforcement and international cooperation. Errors require simplification and guidance. Tax morale — the underlying willingness of taxpayers to comply — changes slowly and requires sustained investment in education, communication, and trust.

The report demonstrates that action works. Countries that implemented digital reporting tools saw measurable reductions in their gaps. The correlation is consistent enough across multiple countries and multiple tools to support a causal interpretation. This should encourage administrations to continue investing in these measures while recognising that compliance enforcement is an ongoing and evolving challenge — as the methods of evasion adapt to the tools deployed against them.

The Limits of Progress

No VAT gap can realistically reach zero. There will always be some level of errors, omissions, and residual non-compliance that cannot be fully eliminated. Some of the lowest national gaps in the study raise methodological questions about whether the figures are being slightly underestimated — a practical lower bound exists below which further reduction becomes implausible.

The policy gap, meanwhile, is primarily a product of deliberate choices about the structure of the VAT system. Those choices have social and distributional objectives — taxing food at reduced rates, exempting healthcare — that reflect political priorities beyond revenue maximisation. The report’s contribution is not to argue that these choices are wrong, but to make their revenue cost transparent and to distinguish the actionable from the non-actionable.

Final Observations

The EC VAT Gap Report 2025 is a document of considerable analytical depth, covering 27 EU member states plus candidate countries across more than two decades of data. Its headline findings — a compliance gap that fell for nearly a decade, held steady through the pandemic, and has since risen modestly — reflect the complex interplay of economic conditions, structural change, and administrative effort.

Its more granular findings — the measurable impact of SAF-T and e-invoicing, the dramatic effect of Italy’s Superbonus programme, the limited but real scope for member states to reduce the policy gap — offer concrete guidance for those designing and evaluating VAT policy.

As Grzegorz Poniatowski observes, the value of measuring the gap lies not only in the numbers themselves, but in the discipline it imposes: a tax administration that monitors its gap is one that takes seriously its responsibility to protect public revenue.

---