Mario Draghi's sharpest line to the European Parliament on 17 September 2024 was not about paperwork. “There is no EU company with a market capitalisation over EUR 100 billion that has been set up from scratch in the last fifty years,” he told MEPs. He was describing an absence — of scale, and of the market conditions that produce it.
Two years on, the Commission's most measurable answer to that absence is arithmetic of a different kind. Package by package, it counts the recurring cost of its own rules that it has agreed to stop charging.
EPIC, the European Policy Innovation Council, published a report in July 2026 called The Cost of Single Market Fragmentation: What We Know, What We Don't, and What We Need to Measure. Its central recommendation is a research programme built around five defined workstreams: services and regulated professions, goods and regulatory heterogeneity, public procurement, territorial supply constraints, and administrative burden.
Four of those five have no live measurement anywhere in the European system. The fifth has an entire institutional apparatus.
The Commission has committed to cutting recurring administrative costs for business by at least 25 per cent, and by at least 35 per cent for small and medium-sized firms, by the end of the 2024–2029 mandate — a headline target of EUR 37.5 billion a year. Through 2025 it tabled ten omnibus simplification packages worth, on its own accounting, EUR 11.9 billion in recurrent savings, rising to about EUR 15 billion once other simplification measures were added, plus EUR 5.6 billion of one-off costs removed. By August 2026 the running total stood at twelve omnibus proposals and roughly EUR 18 billion a year — the Commission's own reckoning that it is about halfway there. Half of the proposals in its 2026 work programme are simplification proposals, spanning sustainability reporting, the common agricultural policy, chemicals, defence procurement, digital rules, environmental permitting, automotive, taxation and energy labelling.
Read it carefully and the whole exercise is subtraction. Every euro on that scoreboard is a compliance cost that Brussels first created and has now agreed to stop imposing. It measures how much of its own weight the EU can take off. It does not measure what the market underneath would produce if it worked.
That distinction matters, because deregulation and integration are not the same operation. Withdrawing a European reporting duty leaves twenty-seven national regimes exactly where they were, and in some files it leaves them more visible, not less. EPIC's evidence review makes the point with the bluntest figure in it: public procurement has been legally integrated for decades, and local firms remain over 900 times more likely to win a contract than a foreign bidder (Herz and Varela-Irimia, 2020). Legal integration is not market integration. On the same logic, neither is legal subtraction.
EPIC's core finding is an absence. Europe has measured what the single market delivers: roughly EUR 840 per citizen per year, about EUR 427 billion across the member states (Mion and Ponattu, 2019). It has measured what dismantling it would cost: EU GDP would be 8 to 9 per cent lower without it (in 't Veld, 2019). It has never measured what completing it would gain.
The estimates that do exist are scattered and rest on different baselines. The European Parliament's research service put broad completion gains at EUR 651 billion to EUR 1.1 trillion a year in 2014, EUR 615 billion in 2017, and services alone at EUR 297 billion in 2019. EPIC infers a plausible central range of 4 to 5 per cent of EU GDP, and says in terms that this is an inference rather than a measurement.
The historical comparison is the argument. In 1988 the Cecchini Report priced the cost of “non-Europe” at around 200 billion ECU, roughly 5 per cent of Community GDP. That number is a large part of why the 1992 programme happened at all: it gave a diffuse, technical project a figure that a finance minister could put in a speech. Nothing equivalent has been produced since, and the report EPIC wants — a modern sequel, The Benefits of Completing the Single Market — does not exist.
EPIC offers a political-economy explanation that applies neatly to its own list: the benefits of integration are diffuse, cross-border and long-term, while the benefits of national protection are local, immediate and politically organised.
Administrative burden is the exception that proves the rule. Its beneficiaries are identifiable and its effects are immediate. A firm that stops filing a report knows it has stopped filing a report. Nobody has to be persuaded that the gain is real, which is precisely why the target could be set, the baseline built and the running total published. The Commission's 2026 Annual Single Market and Competitiveness Report, adopted on 30 January 2026, named a “Terrible Ten” list of remaining barriers and launched a first annual enforcement agenda. The burden target is the piece of that agenda with a scoreboard attached — and it is the piece that asks least of member states.
The simplification programme deserves better than to be treated as a distraction. It is auditable, it is dated, it creates a form of accountability that most EU competitiveness rhetoric never acquires, and firms report the effects. A target with a number is harder to abandon quietly than a strategy without one.
It also has real limits, and they should be stated. The savings are estimated by the same institution proposing the cuts, with no independent audit of the arithmetic. A proposal is not a law: several of the twelve omnibus packages are still in negotiation, so the running total counts intentions as well as outcomes. And the honest objection runs the other way too. A Cecchini-style number would itself be a modelled estimate, contestable in the same way. Cecchini worked because a political programme was already in motion and needed a figure; a figure produced into a vacuum does not create the programme.
The asymmetry is not that Europe cannot measure. It is that Europe measures what it controls and not what it wants. The administrative-burden workstream is a demonstration that the machinery exists — targets, baselines, published totals — and that it gets built when the political demand is loud enough. The other four workstreams on EPIC's list are harder only because the gains land somewhere diffuse and the costs land on someone specific. Until one of them acquires a number, the honest position is that the EU knows the price of its own paperwork to the nearest hundred million and does not know, within a factor of two, what its unfinished market is worth. Draghi's missing hundred-billion-euro company will not be conjured by removing a reporting obligation. Whether it would be conjured by a number is a fair question — but Europe has not run the experiment since 1988.
