"If one objects to building a true Single Market, to capital market integration, and objects to debt issuance, one objects to our EU objectives." Mario Draghi said that in Strasbourg on 17 September 2024, presenting his report on European competitiveness. It is a line about political will. The goods market suggests the problem is often something duller and harder to fix: not objection, but administration.
Mutual recognition is the oldest working idea in the single market and by some distance the cheapest. It says that a product lawfully marketed in one member state may in principle be sold in another, even where there is no common European standard, unless a government can justify blocking it on narrow grounds of public safety, health or the environment. The European Court of Justice established the principle in 1979. The Union gave it a dedicated regulation, in force since 2020, with procedural deadlines and an assessment template. It requires no new legislation, no budget line and no treaty change. It requires an official to apply it.
That last requirement turns out to be the binding constraint. The Commission's own analysis identifies two root causes of unnecessary barriers in the single market: regulatory choices by member states, and inadequate implementation of EU law. On mutual recognition specifically, the recurring complaint is the absence of a practical rule book for national inspecting agencies — with the predictable result that a civil servant faced with an unfamiliar product reaches for the national rule they know rather than the European principle they are supposed to apply.
This is not obstruction in the ordinary sense. It is the default behaviour of a competent official operating without guidance, and it produces the same outcome as obstruction while being much harder to litigate. A company that cannot afford to spend eighteen months on a complaint simply reformulates the product, or does not enter the market. Neither of those decisions appears in any dataset.
The pattern will be familiar to anyone who read EPIC's finding on procurement. Public contracts have been legally open across the Union for decades, and local firms are still over 900 times more likely to win them than foreign bidders — a figure from Herz and Varela-Irimia's 2020 study, cited in the July 2026 report by the European Policy Innovation Council, The Cost of Single Market Fragmentation: What We Know, What We Don't, and What We Need to Measure. The report's line for it is that legal integration is not market integration. Non-harmonised goods are the same finding, one aisle over — except that here there was never a harmonising law to point at in the first place.
EPIC's argument is not that fragmentation is unmeasured. It is that only three of the four possible measurements exist.
Europe knows roughly what the single market delivers: about €840 per citizen per year, some €427 billion across the Union, on Mion and Ponattu's 2019 estimate. It knows roughly what losing it would cost: EU GDP 8 to 9 per cent lower without it, on in 't Veld's 2019 modelling. It has scattered guesses at completion — the European Parliament's research service put broad gains at €651 billion to €1.1 trillion a year in 2014, €615 billion in 2017, and services alone at €297 billion in 2019. What it does not have is a current, methodologically coherent figure for what finishing the job would be worth, built the way the 1988 Cecchini report was built, when the cost of "non-Europe" was priced at around 200 billion ECU, roughly 5 per cent of Community GDP, and that number gave the 1992 programme its political force.
Goods and regulatory heterogeneity is one of the five workstreams EPIC proposes for a modern sequel, alongside services and regulated professions, procurement, territorial supply constraints and administrative burden. It is also the one where the measurement problem is worst, for a reason worth stating plainly: non-harmonised goods are a residual category. They are defined by what European law has not covered. There is no common taxonomy of them, which means there is no obvious denominator for the estimate — and no clean way to separate a genuine safety judgement from a protectionist reflex wearing one.
The political energy is real, and it is aimed elsewhere. EU leaders launched a "One Europe, One Market" agenda at the March 2026 European Council, with six priority measures to be delivered during 2026 where possible and by the end of 2027 at the latest — among them better mutual recognition of professional qualifications and stronger safeguards for placing products on the market. That builds on the Commission's 2026 Annual Single Market and Competitiveness Report of 30 January, which named the "Terrible Ten" barriers and launched an enforcement agenda.
All of that is enforcement and simplification. None of it is measurement. The barriers have been named repeatedly since the Letta report of 2024. What has not been produced is the figure that would tell a national regulator what its caution costs someone else.
Two objections deserve stating. The first runs against EPIC. A goods-fragmentation number would be genuinely hard to build and easy to attack — harder than services, where at least the sectors are defined. Produce a headline figure on a shaky base and the first credible rebuttal discredits the whole exercise, including the parts that are solid.
The second runs against the impatience. Some refusals of mutual recognition are correct. Member states retain real competence over public safety, health and the environment, and the exception exists because governments are sometimes right to use it. A framework that treats every blocked product as a barrier will overstate the prize, possibly by a lot.
What survives both objections is narrower and still uncomfortable. Cecchini did not work because its number was precise. It worked because a number existed, was contestable, and forced everyone arguing about the single market to argue about the same object. Thirty-eight years later, the cheapest instrument the Union owns is failing at the counter, and there is no agreed figure for what that failure costs.
Every other item on Europe's competitiveness agenda needs money, unanimity or a treaty. Mutual recognition needs neither. It needs guidance for inspectors, a functioning complaints route and a reason for a national official to prefer the European rule to the domestic one. That last item is the political problem, and it is the one a number would address, because the diffuse cross-border gains from integration always lose to the local, immediate and organised benefits of protection — unless someone can say how much is being given up.
The Union is about to spend two years on an agenda to remove single-market barriers without knowing what removing them is worth. It has been doing that since 1992. The instrument is not the missing piece. The arithmetic is.
