Brussels Approved Balkan Transport Grants Before Projects Were Ready. Auditors Found €52m Prepaid Before a Works Contract.

Factrail investigation | EU and Western Balkans | Core audit published 9 June 2026 | Research checked 31 October 2026

Finding. The European Commission's Western Balkans transport-grant system approved some projects before detailed designs or cost evidence were ready, then relied on international lenders for supervision it could not fully verify. In a June 2026 European Court of Auditors investigation (opens in a new tab), all 12 sampled projects suffered delays; six had implementation delays exceeding two years. For one unnamed project, the Commission transferred €52 million in pre-financing to a pooled fund in October 2024 although no construction costs had been incurred by the auditors' June 2025 visit, because the works contract had not been signed. That is a documented timing and oversight failure, not proof that €52 million was stolen or that a politician personally received it. The larger political question is why officials approved projects whose promised connectivity could not yet be delivered, and who could have withheld approval.

A financial promise is not a finished railway

The Western Balkans Investment Framework, or WBIF, mixes EU grants with loans from institutions such as the European Investment Bank (EIB) and European Bank for Reconstruction and Development (EBRD). Its stated purpose is to close costly gaps in the transport network connecting Serbia, North Macedonia, Kosovo, Bosnia and Herzegovina, Montenegro and Albania to European markets. The policy has an intelligible economic logic: a cross-border railway may generate public benefits that cannot be captured as fare revenue, leaving a financing gap that a commercial loan alone cannot cover.

The auditors' 58-page report (opens in a new tab) is more precise than the convenient headline that the EU spent €2.7 billion on failed roads. From 2015 to mid-2025, the Commission committed €2.665 billion for 55 transport projects. The report says €527 million had been paid into the Joint Fund for transport over that period. The auditors examined 12 projects with €341.6 million in approved grants in four countries. Commitments, transfers into a pooled fund, loans, construction invoices and completed assets are different accounting stages. None of those numbers can be used interchangeably.

The Commission contributed roughly 86% of all donor money transferred into the Joint Fund since 2015, across sectors, according to the audit. This makes Brussels the dominant grant financier, but not the contractor signing each construction contract. National public project developers procure works. International financial institutions arrange lending, supervise procurement and implementation, and validate expenditure. The Commission screens and approves grants and remains responsible for the EU budget. Responsibility is distributed, but not absent.

The finding is also not that the selected roads and railways were useless. All 12 sampled projects aligned with the region's designated core European transport network. The failure was in the distance between an eligible policy objective and an executable construction package. Auditors concluded the region was unlikely to finish the core network to the promised standards by 2030. Euronews's independent report (opens in a new tab) separately emphasized the gap between headline finance and delivered infrastructure.

The documentary timeline: from the Berlin Process to the June audit

In 2014, the Commission asked Western Balkan governments to build national priority lists of infrastructure investments. The WBIF then used these pipelines when considering grant applications. The aim was national ownership, not an externally imposed wish list. In 2015, the initiative required projects to contribute to the core European transport network. Through the EU's pre-accession instruments, grants were intended to make otherwise difficult infrastructure financeable.

The 2015–2025 audit chronology (opens in a new tab) shows how the original filter evolved. The operational board initially had no sufficiently specific definition of a mature, construction-ready project. From 2020 it expected more developed designs; from 2022 it required environmental and social assessments and, depending on contract type, more complete design documentation before approval. Those changes are relevant mitigating facts: the Commission did not leave its selection method entirely static.

Nevertheless, auditors found eight of the 12 sampled applications premature. Five grants were approved without a validated detailed design, and one without a cost-benefit analysis. In three countries the operational board approved projects with low national priority rankings because they were considered ready; two of the three then started more than a year late. The median start-up delay across ten projects that started after grant approval was 17 months. A delay alone does not prove the original decision was irrational: land acquisition, permits, seismic standards and political developments can change after approval. But repeatedly accepting incomplete preparation makes those later problems more likely to be costly.

In 2020, the Commission sent letters seeking updated schedules for three audited projects and later made similar requests for two others. According to the auditors, none of those letters prevented subsequent delays. In 2021, the monitoring system gained a field explaining slippage, yet auditors found that deadlines for resolving problems were too optimistic in five cases. This is a record of incremental intervention that did not solve the underlying problem, not evidence of total institutional inaction.

In October 2024, one project received the striking €52 million pooled-fund pre-financing described in the audit. By the June 2025 field visit, no works costs had been incurred because the construction contract was unsigned. On 9 June 2026 the Court published its findings and recommended tighter maturity, risk, payment and sustainability controls with a 2027 target. The report evaluates evidence principally through mid-2025. It does not establish that the same project still lacked a contract, or that the money remained idle, in October 2026.

Who held the pen, the vote and the veto?

The ECA's approval flow diagram (opens in a new tab) separates at least five institutions. National ministries and project developers choose priorities and prepare the application. A national coordinator submits it. The lead international lender supplies due diligence and signals willingness to finance. Commission officials and other experts screen the proposal; a group of financiers reviews it; the WBIF operational board gives its approval opinion; and the assembly of contributors approves grant financing. The lead lender and the beneficiary then enter agreements to implement and supervise the work.

That structure creates multiple genuine veto points. A national government can decline to sponsor a project or insist on a better design. A lender can withhold financing or its no-objection to procurement. Commission screeners can challenge costs and readiness; the operational board can reject or defer an application; contributors can decline a grant. A later contractor can be required to correct defective work under its contract. The ECA found that this chain did not consistently stop underprepared projects, but did not identify a single commissioner personally ordering an improper payment.

The political leadership matters at the level of policy and accountability, not presumed authorship of every technical approval. Commission President Ursula von der Leyen has publicly promoted the Western Balkans growth agenda. Former enlargement commissioner Olivér Várhelyi joined Serbian President Aleksandar Vučić in announcing the Belgrade–Niš rail finance package in 2023 (opens in a new tab). Their visible sponsorship shows the political value placed on connectivity. It does not establish that either personally approved any of the 12 anonymous audited grants, much less knew of the €52 million pre-financing case. The relevant approvals require grant-by-grant records and signatures.

For the Commission, the essential question is whether delegation to international financial institutions came with enforceable evidence obligations. The auditors say Brussels did not systematically require lenders to show that risk-mitigation measures were actually implemented. Annual reports were often outdated or thin. Commission delegations had broad awareness of delays but lacked a formal project-monitoring role. In effect, the principal financier could know that a programme was behind schedule without seeing enough current technical evidence to determine why.

Who received an advantage—and who bore the risk?

The most direct named institutional beneficiaries of financing are public railway and road project developers, and the countries obtaining grants and loans on terms that would otherwise be harder to arrange. For example, the EBRD and EIB's 2023 Belgrade–Niš package (opens in a new tab) described up to €598 million in EU grants, a €1.1 billion EIB loan and a €550 million EBRD loan. Serbian Railways Infrastructure is the named public-sector beneficiary of the EBRD's loan project 53136 (opens in a new tab), which the bank lists as disbursing. These are publicly documented beneficiaries of a named financing package, not proof that the Belgrade–Niš scheme was the unnamed €52 million case or one of the 12 audited grants.

The lender also benefits institutionally from a larger project portfolio and from repayment rights under signed loans, but its fees, net earnings and project-specific risk exposure cannot be inferred from headline loan ceilings. Private construction and engineering firms can win contracts as projects proceed, yet the audit anonymises sampled projects. Assigning the problematic advance to a named contractor without the original award, payment information and contract would manufacture an accusation. This investigation therefore names the confirmed public project beneficiary, not an unverified private payee.

The financial mechanism is more subtle than a simple transfer from taxpayers to a builder. In three sampled cases, the auditors found that loan and national finance had already been sufficient when the grant was requested. Lenders agreed to reduce their loan commitments by 30% to 52% if the EU grant was approved. A grant in that situation can reduce a sovereign borrower's debt burden or interest exposure, a genuine public benefit, while failing the narrower test of whether the grant was necessary to unlock the loan. The ECA called this a deadweight effect. A fiscal benefit for the borrower and weak additionality for the EU donor can both be true.

In six cases, according to the audit, Commission pre-financing forecasts exceeded what actual financial and physical progress justified. Money was transferred to the EBRD-managed Joint Fund too early. The €52 million example is a particularly clear control test: at the point of the auditors' visit, the works contract was not signed. But pre-financing into a pooled account is not equivalent to an invoice paid to a contractor. Whether any amount was later clawed back, earned interest, remained reserved or financed eligible work is not established by this report.

The public bears opportunity costs when EU grant resources are committed to projects that cannot quickly proceed, and local users wait for faster, safer transport. Yet the record does not quantify a net loss of €52 million or a specific additional tax burden. It also does not identify personal enrichment by Vučić, von der Leyen, a bank executive or a contractor. A serious financial investigation must not convert delayed execution into theft.

The procurement and environmental control gaps

The Court found that ten of the 12 applications presented works costs as one undifferentiated figure rather than a usable itemised breakdown. In one application the screeners failed to spot ineligible costs, leading to a €0.9 million over-calculation, or 3.6% of that grant. Another project developer raised an estimate by 50% without reliable current market analysis. These findings describe weaknesses in estimating and review; they do not, on their own, prove that €0.9 million was improperly paid out.

Risk assessments also deserve careful wording. The audit says no sampled application explicitly identified corruption risks. Two of the three lead financial institutions had been notified of four reported prohibited practices. One institution investigated three allegations and found them not credible; the fourth was still under investigation when the auditors examined it. These are allegations with different procedural outcomes, not four established offences. An investigation that suppresses the exculpatory findings would be as misleading as one that ignores the unresolved allegation.

The auditors' fieldwork supplied a tangible environmental example. One road project budgeted €1.05 million for tree planting and revegetation. Auditors saw invoices for about €35,000 of that purpose and were told the remainder was used for slope-stabilisation nets. The developer said more plants than planned had been installed, albeit over a smaller area; auditors questioned their sustainability. That is an observable change in what an environmental budget delivered, not proof that a contractor personally pocketed the difference. It raises a contract-change and monitoring question requiring invoices, site records and environmental performance evidence.

Another sampled project had work performed with unapproved concrete that was to be de-certified; auditors found that the problem had not appeared in the regular reports to Brussels. The concern is not only whether money was paid, but whether the lender, supervising engineer and Commission were seeing the same quality warnings soon enough to intervene. The most consequential information may be the missing engineering alert, rather than a spectacular headline contract value.

The strongest counter-case: public infrastructure cannot wait for perfection

The defence of the WBIF model is substantive. Western Balkan governments need cross-border transport capacity, and the private revenue from an isolated railway section may be too low to support a commercial loan. Grants lower debt pressure on countries with constrained budgets; multilateral lenders bring engineering, procurement and environmental safeguards that national agencies may not have. The ECA itself concluded that every sampled project aligned with strategic connectivity needs, and information improvements to selection rules since 2020 and 2022.

The EBRD's North Macedonia–Serbia border station announcement (opens in a new tab) illustrates the plausible benefit: a €2.8 million EU grant paired with a €5 million loan to support a joint rail station intended to reduce border waiting times. A 2026 WBIF account of Montenegro's Bar–Golubovci upgrade (opens in a new tab) describes a €112 million EU grant alongside a €63 million EIB loan. These public project announcements are evidence of financing arrangements and intended outcomes, not audited proof of finished travel-time savings.

National permitting problems, political blockages and engineering revisions are real constraints. In North Macedonia, a rail link needed cross-border agreement to become fully useful. In Bosnia and Herzegovina, the Commission paused two initially approved projects in 2022 because of the political situation in Republika Srpska. Some completed roads were already open to traffic. A blanket verdict that the entire programme produced nothing would contradict the audit.

The Commission's position, as described in the audit, is that projects met network objectives, had funding gaps and could create social benefits; it also says it largely relied on lenders' forecasts before authorising Joint Fund transfers. That is a reasoned explanation of the mechanism, although it does not answer why a €52 million advance preceded a works contract. A complete, separately authenticated Commission response to each 2026 recommendation was not established during this research; this limitation should remain visible in independent review.

Factrail context and separate accountability mechanisms

The Commission presidency of Ursula von der Leyen provides political context, while the actual grant decisions were handled through the documented WBIF and lender chain. The concern belongs in the wider institutional-corruption (opens in a new tab) accountability category without assuming criminal wrongdoing. Rail electrification also depends on the infrastructure and supply-continuity issues described in Factrail's energy-security analysis (opens in a new tab); these contextual pages are not evidence that any specific grant altered those outcomes.

What it moves: four measurable accountability indicators

First is the share of approved grants supported by validated detailed designs. The audit identified five of 12 without such validation. A future Commission report could disclose that percentage for all new awards, rather than leave readers to extrapolate from a nonrepresentative sample. The desired direction is upward.

Second is elapsed months from grant approval to works-contract signature. The median start-up delay was 17 months in the ten sampled projects that began after approval. Requiring mature designs should reduce this measure, although the direction must be verified in later project cohorts.

Third is pre-financing outstanding against actual incurred eligible costs. The six over-forecast cases and the €52 million example make the risk concrete. A twice-yearly reconciliation of amounts transferred to the fund, amounts paid to final beneficiaries, certified costs and physical progress would show whether money is parked ahead of need. A lower unjustified gap would indicate better control, not necessarily less useful public investment.

Fourth is operational corridor completion to EU standards, not GDP growth or aggregate trade. By mid-2025 only six of 43 approved transport investments in the four visited countries had been completed, according to the audit. The Transport Community's 2024 indicators showed 65% of the core rail network electrified, compared with 73% in 2021, and 60% meeting the minimum freight-speed requirement, compared with 72% in 2021. Changes in definitions, network scope and measurement should be checked before treating those comparisons as a pure deterioration caused by WBIF grants. The right outcome is completed, connected, operational infrastructure, not a higher number of approved grant announcements.

What the documents cannot yet prove

The 12-project audit sample is not a census of 55 transport projects and cannot support a universal misconduct rate. The report did not verify final expenditure accuracy for the sampled projects because the Commission had not financially closed them. It does not name the €52 million project, establish a criminal case, identify a private beneficiary of improper payment, or prove any political quid pro quo. Nor does a 2025 field visit establish project status as of October 2026.

The unresolved documentary requests are specific: the grant approval and screening minutes for the advance-financed project; the lender's forecast and Joint Fund transfer record; the construction-contract signature and certified-cost timeline; the latest Commission response and implementation status for all four ECA recommendations; and named procurement awards for each sampled project. These records could materially strengthen or overturn the present judgment.

The evidence supports a narrower, consequential conclusion: Brussels financed a strategically plausible transport network through a system whose approvals and payment forecasts sometimes ran ahead of construction readiness, while the body accountable for the EU budget lacked sufficiently current evidence to challenge its intermediaries. The political accountability test for the next round is not another investment information. It is whether the Commission, national sponsors and lenders can demonstrate that approved designs, contracts, physical works and payments now line up.

Documentary source record

1. European Court of Auditors, Special Report 16/2026: Western Balkans Investment Framework, 9 June 2026 (opens in a new tab) — primary audit, findings, limitations, sample and recommendations.
2. ECA press release, Western Balkan transport network's integration into the EU is too slow, 9 June 2026 (opens in a new tab) — auditor's public explanation.
3. Euronews, EU auditors flag weak oversight of €527 million Balkans transport fund, 9 June 2026 (opens in a new tab) — independent reporting.
4. Agence Europe, European Court of Auditors highlights shortcomings, 10 June 2026 (opens in a new tab) — separate specialist reporting.
5. EBRD, EU and EIB support high-speed rail in Serbia with €2.2 billion, 2023 (opens in a new tab) — named political announcement and finance.
6. EIB, Team Europe announces Belgrade–Niš railway package, 2023 (opens in a new tab) — financing structure.
7. EBRD, High Speed Rail Belgrade to Niš, project 53136 (opens in a new tab) — named borrower and project stage.
8. EBRD, North Macedonia–Serbia rail crossing grant, 2025 (opens in a new tab) — comparative example.
9. WBIF, EU and EIB back Montenegro rail upgrade, 15 May 2026 (opens in a new tab) — comparative grant/loan project.
10. EUROSAI, Special Report 16/2026 audit registry (opens in a new tab) — independent audit registry.

Editorial limits: This is an original document-based synthesis, not a criminal accusation or a completed forensic payment audit. Contextual Factrail links and Commons file rights were examined separately on 1 November 2026; the direct image response remains to be confirmed. Publication and legal review are separate.