The most quoted number in EPIC's Draghi Implementation Index is defence: 35.7 per cent of the report's defence recommendations implemented at the September 2025 baseline, 78.6 per cent by the January 2026 interim audit. It is the largest sectoral movement the index has recorded, and it is usually offered as proof of a rule — that the European Union delivers quickly when competitiveness and security point the same way.
The rule is real. It is also incomplete, and the evidence that it is incomplete is sitting in the same audit, two lines further down.
The Draghi Implementation Index, published by EPIC through its Draghi Observatory at draghiwatch.eu, does one narrow thing well. It takes the 383 recommendations in Mario Draghi's September 2024 report on European competitiveness and asks, recommendation by recommendation, whether each has become binding EU law. Nothing else counts: not a communication, not a strategy, not a work-programme entry. Strict implementation stood at 15.1 per cent of the 383 at the January 2026 audit and 15.7 per cent at the July 2026 preliminary update. Counting partial implementation as well, the figures were 38.9 per cent and 41.3 per cent. The next full review is due in September 2026 and has not been published.
Because the measure is binding law, the index is at its sharpest when comparing sectors inside the same window. Everything is scored by the same test, so a sector that moves and a sector that does not are being judged identically.
Between September 2025 and January 2026 the Union produced the instruments that its defence recommendations had asked for. Energy-intensive industries moved too, from 40.5 to 57.1 per cent. Energy remained the laggard among the movers at 22.9 per cent.
EPIC's own reading of this pattern, repeated in its July 2026 update and in its Single Market Fragmentation work, is that the EU acts fastest where competitiveness fuses with security, and slowest on structural reforms that force market outcomes. It is a good explanation. It explains defence and it explains capital markets. It does not explain what happened to the other two sectors in the audit.
Pharmaceuticals and space were both recorded at roughly 27.8 per cent in January 2026, and both were flat — unchanged from the September 2025 baseline in the same four months that took defence up by more than forty points.
Space is the difficult case for the security-badge theory. Nobody in Brussels argues that space is a civilian file. Satellite communications, earth observation and launch capacity are treated in every recent EU strategy as strategic infrastructure; the IRIS² constellation is justified in explicitly sovereign terms; the connection between launch autonomy and defence autonomy is made by the same officials making it about ammunition. If the rule were that a security argument unlocks delivery, space should have moved with defence. It did not move at all.
The distinction that survives is narrower and less flattering. Defence moved because the Commission could reach for instruments it already controls — a loan facility, joint procurement, an industrial programme, budget lines. Those are regulations and Council decisions, which is to say precisely the objects an index of binding law can count. Space policy in the same period ran mostly through governance arguments, agency relationships and a legislative proposal moving at ordinary speed. The security framing was identical. The available instruments were not.
Pharmaceuticals is flat for a different reason, and the contrast is useful. The EU's pharmaceutical legislation has been in negotiation between Parliament and Council for years, and the file is genuinely contested: regulatory data protection, orphan-drug incentives, environmental requirements and supply-chain obligations pull member states in opposite directions along lines that do not correspond to any competitiveness consensus. Nothing about the Draghi Report changed those positions.
So one sector is stuck because the instruments were not available and the other because agreement was not. Both register as 27.8 per cent. That is the limit of what a single implementation percentage can tell you, and it is worth stating plainly rather than reading the number as a verdict on effort.
There is a fair case against the argument made here. Four months is a short window, and a sector can be flat because its legislation is mid-procedure rather than because anything is wrong; the July 2026 update covers a longer span and shows the overall pace slowing from 7.5 percentage points in the first half-year to 2.4 in the second, which suggests the September–January period was unusually fast rather than that pharma and space were unusually slow. It is also true that EPIC does not claim the index measures policy quality. It measures whether one report's recommendations became law, and a recommendation can go unimplemented because it was wrong.
The point survives those objections in a modest form. Defence at 78.6 per cent is used as evidence of what Europe can do when it decides something matters. Space at 27.8 shows that deciding something matters is not the operative variable.
The useful lesson from the flat lines is about instruments, not will. Ask which Draghi recommendations became law in a given period and the answer tracks, almost exactly, which ones could be delivered with tools the Commission already holds. Where the tool existed, delivery followed within months. Where it had to be invented, negotiated with twenty-seven capitals, or fought through a contested legislative file, the number did not move — regardless of how strategic the sector was declared to be.
That has a practical implication for the autumn. The September 2026 review will land alongside a Commission work programme heavy on competitiveness initiatives, and the temptation will be to read a rising percentage as rising ambition. A more honest reading of the index is that it measures the match between an ambition and the instrument chosen to carry it. On that test, the interesting question about any new proposal is not whether Europe has decided the sector is strategic. It is whether anyone has checked that the Commission owns a lever long enough to move it.
.png)
