One point apart and three reasons no one can verify: the Council of State suspends the appointment of the auditor of the Belgian Foreign Trade Agency
The Belgian Foreign Trade Agency awarded the statutory-auditor mandate for financial years 2026-2028 to Baker Tilly with 86 points against 85 for L&S Réviseurs d’entreprises, but because neither the evaluation report nor the administrative file makes it possible to understand why L&S’s approach was ‘relatively general’ and its experience documentation ‘succinct’, the Council of State suspends the award under extreme urgency.
What happened?
In early 2026 the Belgian Foreign Trade Agency decided to tender the appointment of a statutory auditor for the financial years 2026, 2027 and 2028. It chose a negotiated procedure with prior publication; special specifications no. I 0002237 divided 100 points over two award criteria: ‘Price’ (70 points, awarded through the formula lowest price divided by the price offered, multiplied by the weight of the criterion) and ‘Approach plan’ (30 points). For that second criterion, article 1.10 of the specifications required bidders to attach a concise document of no more than five A4 pages, in which they could explain their firm, their own approach and the services included, mentioning their experience with other public institutions. The evaluation method was announced in advance: the extent to which the proposed programme is appropriate and realistic, the extent to which the approach is adapted to the client’s concrete needs and to the applicable regulatory framework, and the extent to which the bidder has relevant public-sector experience. Three bidders submitted an offer; all three were selected and all three offers were declared regular. On price, L&S Réviseurs d’entreprises obtained 70 points, Baker Tilly 57 and Ernst & Young 34. On ‘Approach plan’, L&S received 15 points, the other two 29 each. Final standing: Baker Tilly 86, L&S 85, Ernst & Young 63. On 26 June 2026 the director general awarded the contract to Baker Tilly; L&S was informed by registered letter of 26 June 2026 and email of 29 June 2026. On 10 July 2026 it filed two applications: a request for suspension under extreme urgency and an action for annulment. The Agency raised four objections of inadmissibility — the contested act was wrongly dated (29 instead of 26 June), the articles of association and the proof of the decision to litigate were missing, the pleas were not developed in the suspension application itself, and the application was obscure. The Council rejected all four: the identification of the contested act may not lapse into excessive formalism where the defending party, on the evidence of its own observations, could not have been mistaken; article 3bis of the Regent’s Decree does not apply under extreme urgency by virtue of article 8, § 1, of the Royal Decree of 19 November 2024, as the report to the King expressly confirms (Belgian Official Gazette, 2 December 2024, p. 130,598); article 4, § 1, first paragraph, 7° of that same Royal Decree allows the suspension application to refer to the simultaneously filed annulment application; and the Agency had, if only in the alternative, defended itself against each argument in its observations, which proves it had understood them. On the merits, everything turned on the three reasons underpinning the score of 15 out of 30: the offer allegedly did not contain the required concise note of no more than five pages, the proposed approach remained ‘relatively general’, and the documentation on relevant experience remained ‘succinct’. The Council recalled that the wide discretion in comparing offers has as its corollary an extensive duty to state reasons: the mere allocation of points does not suffice, the points must rest on reasons expressed as a descriptive assessment based on concrete references to the content of the offer. Because neither the contested decision nor the administrative file makes it possible to determine what weight each of the three reasons carried in the 15 points, it had to be accepted prima facie that any error in those reasons could have changed the score — and therefore the ranking, which was decided by a single point. On the second reason, the Council found that the Agency, having first considered that the required note was missing, nonetheless examined the offer, without indicating which passages it had taken into account; why the approach would then be ‘relatively general’ appears nowhere, nor why the assessments of the two other offers — ‘clear, structured and concise’, ‘coherent, professionally developed and adapted to the Agency’s needs’ — were justified. On the third reason, the comparison led to the same result: L&S stated on page 14 that it carries out external audit assignments for several public-sector and non-profit entities and displayed the logos of sixteen clients, and included on page 42 a reference list naming five of them, whereas the successful bidder listed five clients on page 5 as ‘relevant references’ and produced four client attestations, one of them from a client not otherwise appearing in the offer. Why the one counts as ‘succinct’ and the other as ‘concrete and convincing’ the file does not clarify. The explanation the Agency belatedly gave in its observations — about the fourteen pages of ‘Working methodology’ running from page 16 to 29 and about the comparison with the sections of Baker Tilly and Ernst & Young — the Council qualified as a formal statement of reasons a posteriori, which cannot be accepted, even supposing it to be accurate and relevant. Nor did the defence that L&S had no interest in its complaint because its offer could have been declared irregular succeed: the Agency had not itself declared the offer irregular when awarding the contract, and it is not for the Council of State to examine that regularity afterwards. In the balancing of interests, the Agency identified no disadvantages of a suspension outweighing its advantages, and the Council saw none either. The suspension of the award decision was ordered, the offers (documents A, B and C of the administrative file) remained confidential at this stage, and the costs were reserved.
Why does this matter?
This judgment sharply marks the boundary between discretion and arbitrariness in qualitative award criteria. An authority may weigh an approach plan, may attach thirty of a hundred points to it, and may even let the cheapest bidder lose — but it must afterwards be able to show where one offer fell short and the other convinced, with references to what those offers actually say. Qualifications such as ‘relatively general’ or ‘succinct’ are conclusions, not reasons. The judgment also illustrates two points that are often underestimated in practice. The first is the effect of a narrow result: because the points were not broken down per reason, the applicant need not prove that it would have won without the error — it suffices that this was prima facie possible, and with a one-point gap that threshold is quickly met. The second is the prohibition on repair after the fact: whatever the authority explains in its observations does not count, however convincing it sounds. The judgment is procedurally useful as well. It confirms that article 3bis of the Regent’s Decree — articles of association, instrument of appointment, proof of the decision to litigate — does not apply under extreme urgency, that a suspension application may refer to the simultaneously filed annulment application, and that a wrong date for the contested act does not sink the action as long as the authority could not have been mistaken. That a company appearing without a lawyer got through four objections of inadmissibility and obtained the suspension says something about how flexibly the Council reads those formalities. At the same time the nuance deserves mention: the Council does not find for L&S on the merits. It does not say that L&S’s approach was concrete or that its experience was amply documented, but that no one can verify why the Agency decided otherwise. A bidder who sees the specifications impose a five-page limit and delivers fourteen keeps a problem either way.
The lesson
If you are the contracting authority: do not let your evaluation report end at a verdict for each criterion, but at the source. Do not write that an approach is ‘relatively general’, but which passages you read, what was missing there and which concrete element the better offer did provide. Anchor the assessment for every offer, including the winning one — the Council compares your assessments with one another. If you believe an offer breaches a formal requirement of the specifications, such as a page limit, draw that conclusion at the moment of award and state the reasons; whoever assesses the offer anyway can no longer rely on the irregularity afterwards. And do not count on being able to explain in your observations what is missing from the file: that statement of reasons comes too late. If you are a bidder: read the reasons you receive not for their tone but for their verifiability. If they contain only qualifications without any reference to what your offer says, and without your being able to check why the winner did better, you have a serious plea — certainly where the points gap is small and the authority has nowhere broken down how many points each criticism cost. Meanwhile, respect the rules yourself: if the specifications cap a note at five pages, make it visible with a clear heading and pagination where that note begins and ends. A sound plea about the reasoning does not undo an easily avoidable vulnerability you created on that point.
Ask yourself
Could someone reading your evaluation report without the offers at hand understand why one bidder got 15 and another 29 out of 30? Does every qualitative verdict in that report refer to a concrete element of the offer assessed, and does that also hold for the offer you wish to select? Have you broken down how many points each criticism cost, so that it can afterwards be established whether a contested reason was decisive? If you consider a formal requirement such as a page limit to have been breached, did you actually qualify it as an irregularity in the award decision — or did you tacitly assess the offer anyway? And as a bidder: have you checked whether the reasons you received allow you to infer which passages of your offer were read, or do they rest entirely on qualifications you cannot verify?
About this database
The Council of State (Raad van State / Conseil d'État) is Belgium's supreme administrative court. In disputes over public procurement — from contract awards to tenderer exclusions — the Council of State is the final arbiter. The rulings in this database are summarised by TenderWolf in plain language, with practical lessons for tenderers and contracting authorities. View all rulings →