"Europe faces a choice between paralysis, exit or integration," Mario Draghi told the European Parliament in Strasbourg on 17 September 2024, presenting his report on the future of European competitiveness. "Exit has been tried and has not delivered… Paralysis is becoming untenable… So, integration is our only hope left."
Two years on, the European Policy Innovation Council runs two separate indexes that both try to answer a version of the same question: when someone is told to do something, does it get done? One of them points outward, at the ten countries queuing to join. The other points inward, at the institution doing the telling. Read side by side — which is not how either was designed to be used — they produce an uncomfortable symmetry.
EPIC's Integration Readiness Index, published in June 2026 with the Friedrich Naumann Foundation and written by Antonios Nestoras, Emil Kirjas and Bernada Cunj, scores ten candidates from 0 to 10 across the years 2021 to 2025. Its raw material is narrow by design: European Commission progress reports, and nothing else. Five indicators feed the score, and one of them is Recommendation Delivery — the share of what the Commission formally asked a country to do that the country subsequently did.
The 2025 field runs Montenegro 6.8, Albania 6.7, Moldova 5.87, Ukraine 5.85, Serbia and North Macedonia both 5.5, Kosovo 3.61, Georgia 3.26, Bosnia and Herzegovina 2.91, Türkiye 2.81. The benchmark is Croatia in 2011, the last country to make it all the way through. Nobody in the current field is close to it.
The index is unsparing where it needs to be. Türkiye scored 3.25 in 2021 and 2.81 in 2025: a candidate moving away from the door it is nominally standing at. Bosnia fell from 4.14 to 2.91 in a single year. Georgia slid from 4.06 in 2023 to 3.26. These are not gentle numbers, and the countries they describe are told, in plain terms, that they have not delivered.
EPIC's other tracker, the Draghi Implementation Index at draghiwatch.eu, counts how many of the Draghi Report's 383 recommendations have become binding EU law. Its July 2026 preliminary update puts strict implementation at 15.7 per cent — 60 of 383 — and strict plus partial at 41.3 per cent, or 158. In January 2026 those figures were 15.1 and 38.9 per cent.
The direction of travel is the part worth pausing on. Between September 2025 and January 2026 the index gained 7.5 points on the strict-plus-partial measure. Between February and June 2026 it gained 2.4. The delivery curve is flattening, and it is flattening in the second year, when the easy legislative wins have already been banked.
The July update also produced the index's first ranking by directorate-general. DG TRADE sits at 41.7 per cent strict. DG ENER is at 2.7. DG EMPL is at 0.0. A single institution contains a forty-point spread in its own delivery rate depending on which corridor you walk down.
There is a real methodological difference here, and it should be said before the comparison is pushed any further. A candidate country is asked to transpose an acquis that already exists: the law is written, the benchmark is fixed, and delivery is a matter of adoption and enforcement. The Commission, measured against Draghi, is being asked to invent law that does not exist yet, negotiate it through twenty-seven governments and a Parliament, and survive the resulting compromise. Those are not equivalent tasks, and a 15.7 per cent conversion rate on 383 open-ended recommendations is not the same failure as a candidate that has stopped filing reports.
EPIC does not claim the two indexes are interchangeable, and neither should anyone reading them. But the difference cuts both ways. The candidate is scored on a task the EU has already completed for itself; the EU is scored on a task it set for itself voluntarily, in a report it commissioned, whose central warning was that the cost of not acting is decline. If the harder job explains the lower number, it does not make the lower number less consequential.
Because enlargement has stopped being sold as a values project and started being sold as a competitiveness one. The argument advanced in Brussels since 2024 is that absorbing Ukraine, Moldova and the Western Balkans expands the single market, deepens the labour pool and buys strategic depth — in other words, that enlargement is one of the answers to the problem Draghi diagnosed. Targets have been set on that basis: Montenegro 2028, Albania 2029, Ukraine and Moldova 2030.
That argument only works if the market being joined is worth joining. The Draghi index is a running estimate of how much of the promised upgrade has actually been built. The pattern it keeps finding is that the EU legislates quickly where competitiveness fuses with security — defence moved from 35.7 to 78.6 per cent between September 2025 and January 2026 — and slowly on the structural single-market and capital-market reforms that would make accession economically meaningful. Those are among the weakest categories on delivery.
So a candidate country in 2026 is being asked to hit a moving target, in exchange for entry into a market whose promised improvements are 15.7 per cent complete, on the strict count, and decelerating.
Nothing in either index says candidates should be graded more gently. Montenegro's 6.8 and Türkiye's 2.81 are honest readings of very different records, and the accession process would be worse, not better, if the scoring softened. The point is narrower and more awkward: the EU has built a precise, public, indicator-level apparatus for measuring whether other governments do what they are told, and it has taken an outside think tank to build anything comparable for itself. One of those two scoreboards is a condition of membership. The other is voluntary, and the institution it measures is under no obligation to look at it. Draghi's line was that integration is the only hope left. Integration works in both directions — and only one direction is currently being marked.
