Ask what the EU single market is worth and you get answers from €427 billion a year to several times that. The temptation is to stack the studies and read off a total. EPIC, the Brussels research initiative, says that would be wrong. In its July 2026 report The Cost of Single Market Fragmentation, it explains why the numbers do not add up, and what to do instead.
The first problem is the baseline. The 1988 Cecchini Report estimated what creating a single market would bring before the 1992 programme, putting the cost of "non-Europe" at around ECU 200 billion, about 5% of Community GDP. Mion and Ponattu (2019) estimate the value of the single market that exists: about €840 per EU citizen per year, roughly €427 billion in total. In 't Veld and Felbermayr estimate what would be lost if integration were reversed. The European Parliament's research service estimates the gain from completion. EPIC's point is blunt: these are different counterfactuals, so they answer different questions.
EPIC lists the problems in turn. Baselines differ. Outputs differ: some studies measure GDP, others welfare, trade shares, consumer surplus or administrative costs. In EPIC's words, a 13% trade-intensity gain is not the same as a 13% GDP gain. Sector coverage differs: goods, services, procurement, digital, energy and finance are not covered evenly, and some of the strongest model-based studies are narrow. Overlap is large: one rule can show up as a goods barrier, a services barrier and a procurement barrier at once, so adding the estimates at full value would double-count. And static and dynamic effects are treated inconsistently, even though the long-run gains from firms scaling up, innovating and investing may be the largest of all.
EPIC does not conclude that measuring is impossible. Its line is: "The conclusion is not that measurement is impossible. The conclusion is that measurement must be organised." The existing literature is enough to say that completion gains are large, it argues, but it is not yet structured enough to produce one policy-usable figure.
Its proposal is a research steering group to produce a new Cecchini-style study, The Benefits of Completing the Single Market, split into workstreams: services and regulated professions, goods and regulatory heterogeneity, public procurement, energy, and digital and consumer markets. Each would have its own barrier definition, method and data. The group would then integrate the results into a coherent range rather than a single total.
The case for finishing the single market keeps losing to local interests, in part because the cost of delay is invisible. The same week EU capitals argue over how to count the 2028-34 budget at all, that matters: numbers contested at the margins are easy to dismiss. EPIC's message is that a credible range, built on agreed methods, would do more than a headline figure that falls apart under scrutiny. The 1988 Cecchini number worked because people believed it. Whoever produces its successor has to make sure they can.
