State aid rules apply to measures which are either harmonised at EU level or are left to the discretion of Member States. The protection of public health is no sufficient reason for removing a public measure from the scope of Article 107(1). The fact that undertakings are obliged by law to comply with certain tests does not justify the subsidisation of their costs with public money. A measure is selective when it does not apply to all undertakings.
This article tackles an issue that often puzzles public officials. It appears reasonable and fair that undertakings, which are compelled by law to follow a certain procedure or undertake a certain test, are also reimbursed for the cost of the compulsory procedure or test. Since they are legally obliged to do so, they are not freely choosing to carry out the procedure or test. And since advantage in the meaning of Article 107(1) TFEU means the reduction of the normal costs of undertakings, the reimbursement for costs of compulsory tests should not be considered as State aid because those costs are “abnormal”. This issue was at the heart of the judgment of the General Court of 25 March 2015 in a case brought by Belgium against the Commission.
This article also comments briefly on another recent judgment which did not advance our understanding of State aid for the simple reason that the national court which submitted the request for a preliminary ruling put questions to the Court of Justice that were too general and consequently received answers that were also too general and of little usefulness.
In case T-538/11, Belgium v European Commission, Belgium requested the General Court to annul Commission decision 2011/678 concerning State aid for the application of tests for the BSE disease in cattle. In that decision, the Commission found that the public funding of some tests contained compatible aid while for some other tests the State aid was incompatible and had to be recovered. The Commission considered that even though the tests were compulsory, their funding should have been wholly borne by farmers and that their compulsory nature was not sufficient to remove them from the scope of Article 107(1) TFEU.
Belgium put forth a single plea with six different parts. The gist of its arguments was that since there was no EU legislation harmonising tests and since the tests were obligatory, the state could choose to bear their costs and offer them for free to farmers. In addition, farmers obtained no advantage as the tests aimed to protect public health. Belgium also claimed that the funding of the tests was a general measure since it was made available to all cattle farmers.
i) Absence of harmonising legislation
The General Court first explained that the absence of harmonising legislation at EU level was irrelevant to the application of State aid rules. These rules apply even when there is no EU competence in a specific policy field, such as direct taxation. [Paragraph 66]
The Court recalled that State aid is any advantage that is not available under normal market conditions. It then observed that the Commission considered that the cost of compulsory tests concerning the production or marketing of any product had to be borne by the budgets of undertakings [paragraph 74]. In response to the argument of Belgium that compulsory tests were not normal market conditions, the Court explained that normal costs also include those which result from legal obligations, regulations or other requirements concerning economic activities. [Paragraph 76]
This is an important statement. There is a widespread but mistaken view that normal market conditions correspond to the market without any state presence at all. On the contrary, the normal market conditions do include the regulations and the obligations imposed by the state on market participants.
ii) Protection of public health
Then the Court turned its attention to the second Belgian argument that the farmers obtained no advantage because the tests were intended to protect public health. On this point, the Court reminded us of the established principle in the case law that State aid rules apply to public measures on the basis of their effects rather than their intentions and that the protection of public health was not sufficient to remove a measure from the scope of Article 107(1). [Paragraphs 80-81]
iii) Polluter pays
Third, Belgium also argued that farmers were not responsible for the BSE disease, unlike other diseases for which polluters were liable because, for example, they could be spread by failure of farmers to dispose properly dead animals. The Court repeated that the costs which must be included in the budget of undertakings are not only those which are caused by the “polluter pays” principle. [Paragraph 85]
iv) Exercise of official authority
Fourth, Belgium claimed that the compulsory tests were connected to the exercise of official authority and therefore were not economic in nature. The Court found that the tests were obligatory for farmers in order to be able to market their products and so they could not be regarded as exercising official authority. Cattle farmers had to bear the costs of the test in their budgets. The fact that a public authority imposes a financial burden when it exercises its official authority does not mean that that burden should not be borne by undertakings. [Paragraphs 94-96]
To understand why bearing costs is unavoidably linked to the exercise of authority it is sufficient to consider that no undertaking would voluntarily accept to pay anything or bear any costs unless it is obliged by law to do so. If the exercise of official authority would be sufficient to remove a cost item from the scope of State aid and the normal costs of market operators, then a large chunk of corporate expenses could be absorbed by public authorities without being counted as State aid. That would be absurd.
Do you know we also publish a journal on State aid?
The European State Aid Law Quarterly is available online and in print, and our subscribers benefit from a reduced price for our events.
Fifth, Belgium alleged that the Commission was wrong to classify the funding of the tests as a selective measure. In its view, it was a general measure because it was open to all farmers putting on the market bovine products. The Court first recalled that in order to prove the selectivity of a measure, the Commission is required to show that the measure introduces differentiation between undertakings which are in a similar situation. [Paragraph 102]
The Commission confirmed that the measure applied to all producers of bovine products and who were subject to compulsory tests. At first glance, the measure appeared to be general since it applied to all undertakings that were under the obligation to undertake BSE tests. However, the Court agreed with the Commission that the correct framework of reference was not the bovine sector but all the sectors which had to comply with compulsory tests. In this reference framework, the measure was selective because it relieved only bovine producers form the burden of compulsory tests while producers in other sectors had to pay for compulsory tests themselves. [Paragraph 110]
This finding of the Court underlines how difficult it is to establish the right framework of reference in order to determine whether a measure is selective. Belgium thought that it was the sum of bovine producers. The Commission thought that it was the sum of producers subject to compulsory tests.
Belgium, therefore, counter-argued that the Commission’s chosen reference framework was wrong because the requirement for the BSE test created a liability only for bovine products. The Court replied that the selective character of a measure is determined in relation to the totality of undertakings [“la totalité des entreprises” ]and not in relation to the undertakings which within the same group, benefit from the same advantage. [Paragraph 114]
It is very difficult to know what to make of this statement. There are three problems with it. First, it appears to contradict the statement in paragraph 102 according to which it is first necessary to establish which undertakings are in a similar situation. For sure, the totality of undertakings in an economy are not all in similar situations. Second, what is the totality of undertakings and where are the boundaries of the set that defines that totality? Did the Court literally mean all undertakings in a Member State? Third, it goes contrary to the case law which has found that some measures, by their very nature, do not apply to the totality of undertakings [e.g. the British aggregates case, the Dutch NOx case, the German Lubeck airport case].
vi) Affectation of trade
Sixth, Belgium lastly argued that in other Member States similar measures of compulsory character were not considered to be selective. The Court reiterated the well-established principle that the selectivity of a measure can be assessed only in relation to the undertakings subject to the legislation of a particular Member State [paragraph 119]. Differences between Member States are irrelevant. Differences in the treatment of undertakings from different Member States are relevant only in relation to establishing that the measure in question affects trade between Member States. [Paragraph 124]
At this point the Court criticises the Commission for mixing in its assessment differences in payments for tests within Belgium and differences between Belgium and other Member States [paragraph 125]. However, the fact that the analysis of the Commission was faulty was not sufficient to warrant the annulment of its decision. Therefore, the General Court rejected Belgium’s appeal.
Thinking about selectivity
Apparently selectivity is the criterion of Article 107(1) that causes the most confusion and the most vehement discussions. It arouses passions like no other criterion. I believe that the best way to understand selectivity is to apply the following three-step test.
The reason that Article 107 does not prohibit general measures is that it allows competition between the economic systems of Member States. But for competition between these systems to reflect genuine national advantages, the allocation of resources within the systems must not be distorted. This allocation is much less likely to be distorted when national policies apply equally to all undertakings. Why is that? Because if a Member State can sustain a policy that applies to all undertakings, then that policy is more likely to reflect systemic advantages. The best exposition of this reasoning can be found in the judgment of the Court of Justice in the Azores case where it had to decide whether lower taxes in Azores constituted State aid. In that case the Court of Justice ruled that a regional measure was not State aid where it was adopted by a decision of the regional government without interference by the central government and, more importantly, where the fiscal consequences of the measure were borne entirely by the regional government. This latter point implies that the measure was sustainable without any cross-subsidisation. Similarly, where a national government pursues a policy that is open to all undertakings the measure can be sustained without cross-subsidisation between different classes of undertakings when it reflects a systemic advantage. But since there are many different public authorities in a country it first becomes necessary to identify the area to which their measures apply. Hence the three-step test runs as follows.
Step 1: What is the jurisdiction of the public authority that adopts the measure in question – whole country for national authorities and partial territory for sub-national authorities? Once the jurisdictional area is identified go to step 2.
Step 2: Does the measure in question apply to all undertakings in that jurisdiction? If yes, the measure is general. If not, go to step 3.
Step 3: Is there any objective reason why the measure applies only to some undertakings? If yes, the measure is general. If not, the measure is selective.
Now it is necessary to define the meaning of “objective”. Following the logic of why Article 107 allows general measures, which was explained above, an objective differentiation between undertakings can only occur when, first, a measure by its very nature can apply only to some undertakings and, second, any other undertaking would not want to benefit from it. For example, the Belgian government claimed that it subsidised the BSE tests because they were compulsory. But this implies that it should have subsidised every other compulsory test. If the objective reason was the obligatory nature of the tests, then it should not have limited them to the bovine sector. Moreover, every other undertaking subject to compulsory tests would want to benefit from a subsidy to cover the cost of the tests. Their exclusion made the measure selective.
This is a case that illustrates well the difficulty of designing truly general policies. One has to be sure that all aspects of such policies can apply to all undertakings that can fall within its potential scope of application.
Postscript: What not to ask the Court of Justice
On 16 April 2015, the Court of Justice rendered its judgment in case C-690/13, Trapeza Eurobank Ergasias v Agrotiki Trapeza tis Ellados. At issue was whether certain privileges that were granted to “Agrotiki Trapeza tis Ellados” [Agricultural Bank of Greece] when it was established in 1929, constituted State aid and, if they were State aid, whether the aid could be considered as new or existing aid. The judgment was in response to a request from a Greek court for a preliminary ruling [i.e. guidance on how to interpret EU law].
Instead of phrasing its questions as precisely and as narrowly as possible so the Court of Justice could provide answers that would be immediately applicable to the facts of the case, the referring national court framed its questions in very broad terms. Consequently, the Court of Justice could only explain the principles of State aid without being able to offer guidance as to how they would apply to that particular case. The lawyers on both sides did a pretty poor job by consenting to the general questions that were submitted by the national court.
This case also highlights the risks for Member States that grant privileges to certain undertakings and then forget about them. With respect to the Agricultural Bank of Greece, the Court of Justice stated clearly that if those privileges were state aid and if they were adjusted after Greece’s entry into the EU or at least during the past ten years, then they were illegal because they were new aid that was never notified to the European Commission. In 1929 there was no State aid law. But State aid rules apply to any law, irrespective of when it comes into force, that fulfils the criteria of Article 107(1) TFEU. A couple of years ago, the Commission demanded that Luxembourg abolished a law on holding companies that was also adopted in 1929 because it too was found to be State aid.
 The full text of the judgment can be accessed at:
 The full text of the judgment can be accessed at: