Zagreb: The sums in the indictment filed this week are, by the standards of European fraud enforcement, unremarkable. Eligible procurement costs of €945,858, of which 45 percent was non-refundable grant. What makes the case worth attention is not its size but its shape, and the fact that a supranational prosecutor took it to court at all.
The European Public Prosecutor’s Office in Zagreb charged four individuals and two companies on 18 September with subsidy fraud connected to the construction of a briquette production plant. One of the indicted companies had been a beneficiary of support co-financed from the European Structural and Investment Funds and from REACT-EU, the recovery instrument created to push money into regions quickly during the pandemic. The investigation concerns what happened between the award of that support and the plant it was supposed to build.
REACT-EU is the detail that connects a mid-sized Croatian case to a much larger institutional problem. Crisis instruments are designed to disburse fast, and speed is bought by relaxing exactly the ex-ante checks that would otherwise catch a procurement file assembled to produce a predetermined winner. Everybody involved in designing those instruments knew this. The bargain was that controls would move downstream, from approval to audit to prosecution. Cases like this one are what downstream control looks like when it works, arriving years after the money left and requiring a criminal standard of proof to recover anything.
The prosecutor’s office was not idle elsewhere the same day. A separate operation in Italy involved searches and asset freezes in an investigation into suspected fraud against the recovery facility, the far larger successor to REACT-EU. Estimates circulating this year suggest prosecutors suspect several billion euro of irregularity across the pandemic recovery fund as a whole, a figure that is necessarily soft but not obviously exaggerated given the disbursement speed involved.
Behind the prosecutions sits an architecture that is still being assembled. The anti-fraud office and the prosecutor’s office have overlapping but distinct mandates: one conducts administrative investigations and recommends financial recovery, the other prosecutes crimes against the Union’s financial interests in the twenty-four participating member states. Coordination between them works, but it works through practice and memoranda rather than through a single coherent statute. A Commission communication expected during this year has been trailed as the basis for legislative proposals to tighten that cooperation, and the anti-fraud office signalled in July that strengthening protection of taxpayers’ money was moving up the agenda.
The numbers give a sense of the scale being managed. The office recommended recovery of close to €600 million across 2025, closed 209 investigations, opened 254 new ones from more than 1,140 incoming reports, and prevented a further €18 million from being spent improperly. Across the past decade it counts roughly €6.8 billion retrieved or protected. Recommendations, however, are not recoveries. Whether money comes back depends on national authorities acting on them, and the variation between member states on that point is one of the least discussed weaknesses in the system.
Which returns the argument to Zagreb. An indictment of six defendants over less than a million euro of eligible costs would, under the old arrangement, most likely have been a national administrative matter resolved by a repayment order and a quiet exclusion from future calls. Prosecuting it as a crime changes the incentive facing the next applicant, and it does so in a way an audit finding never has. The next budget cycle will move a great deal of money through instruments built for speed. The deterrent has to be built now, in cases this ordinary.
