Filippo Fontanelli, Lecturer in International Economic Law, University of Edinburgh
On 11 September 2014, the Court of Justice of the EU delivered the judgment in the Essent case (Joined Cases C-204 to C-208/12, not to be confused with Case -91/13, also named Essent, of the same day). The judgment contains a delicate assessment of proportionality involving several interlaced interests: promotion of free energy trade, protection of the environment and human health, compliance with international commitments on emission-reduction, promotion of local employment, and incentives to achieve energy self-sufficiency.
If this maelstrom of values were insufficient to raise the interest of the readership, the judgment has also a purely doctrinal undertone. Indeed, the Court appears to ignore – deliberately – the summae divisiones which it itself has maintained for decades between distinctly and indistinctly applicable measures, as well as between justifications under Article 36 TFEU (which set the explicit Treaty-based grounds for restricting the free movement of goods) and other mandatory requirements (the other grounds for restricting such free movement). Thousands of EU law academics might now need to finally drop the standard exam-questions on “Measures equivalent to a quantitative restriction (MEQRs) from Cassis to Keck and beyond” yearly rephrased and administered since time immemorial: it might turn out that, after all, Cassis is no longer good law.
The facts of the main proceedings relate to a complex domestic regulatory regime which, in turn, implements devilishly detailed EU acts. However, the crux of the dispute is easily summarised: a Flemish Regional scheme requires energy suppliers connected to the electricity grid to demonstrate the use of a certain amount of renewable energy, on pain of a hefty penalty. Only locally produced green energy counts toward that quota, hence suppliers cannot use green energy produced abroad to prove compliance.
Essent could have met or at least approached the quota had it been allowed to factor in the calculation renewable energy that it had sourced from Norway, Denmark, Sweden and the Netherlands, but it was not. It then incurred a series of yearly sanctions, and challenged them repeatedly before the Belgian courts. The Belgian judges raised a preliminary question to the Court of Justice asking, essentially, whether the discriminatory treatment of foreign green energy towards the fulfilment of the compulsory quota is in violation of the treaty obligations regarding freedom of movement of goods and the prohibition of discrimination based on nationality.
[Skip to the next section if you arrived to this post by Googling “mandatory requirements” and you must submit your essay three hours from now. For the others, a few words on the legal context of the dispute.]
Directive 2001/77 (governing the facts of the main proceedings, now replaced by Directive 2009/28) recognizes that renewable energies are underused in the Union, an unfortunate circumstance given that the development of a green energy industry could, at once, alleviate the dependence of EU countries from foreign supply, contribute to the fulfilment of the climate-change commitments entered into under the Kyoto protocol, and foster local employment policies. Therefore, the EU encourages national schemes supporting the consumption of renewable energy, without harmonising them. Better, “until a [Union] framework is put into operation”, which is presumably in the works (see recital 14 of the Preamble of Directive 2001/77).
In order to facilitate the trade in green energy, the Directive established a compulsory system of certification, whereby energy derived from renewable sources is entitled to a “guarantee of origin”. Guarantees of origin issued in EU members are mutually recognized across the Union.
National support schemes can provide various incentives to facilitate the production and consumption of green energy. Typically, benefits are available only to producers and suppliers handling green energy, which necessarily hold the relative guarantees of origin. However, member states are not required to make the specific benefits dependent on the possession of the guarantees. The Preamble of the Directive is clear in this sense: “[s]chemes for the guarantee of origin do not by themselves imply a right to benefit from national support mechanisms established in different Member States”. Guarantees of origin are therefore necessary but possibly insufficient conditions to enjoy national incentives.
The Flemish scheme implementing the Directive, for instance, is premised on the possession (and periodic surrendering) of “green certificates”. Green certificates are granted to producers of green energy based in the Flemish region, which sell them alongside the energy. Suppliers are under an obligation to surrender a certain quota of green certificates to avoid sanctions. At the stage of surrendering, guarantees of origin attached to green energy purchased abroad cannot substitute for the missing green certificates to fill the quota.
A provision of the relevant Flemish Decree envisages the possibility for the Flemish government to accept non-Flemish certificates for the purpose of filling the quota, but no implementing act has been adopted to regulate this possibility, which therefore remained wishful thinking.
The Flemish scheme was challenged under the Directive, under Article 34 TFEU prohibiting measures equivalent to quantitative restrictions (MEQRs), and under Article 18 TFEU prohibiting discrimination based on nationality.
The Directive clearly does not require Member States to link support schemes to guarantees of origin. It simply requires States to issue them and recognize them across the Union, so that consumers are always informed of the green-ness (or lack thereof) of the energy they purchase.
The real challenge to the Flemish scheme was brought under Article 34 TFEU. The domestic measure encourages suppliers operating in the Flemish regions to buy renewable energy from local producers, as equally green foreign energy is unusable towards fulfilment of the quota. Local green energy is artificially made more precious, because of the green certificates it is associated with, which rational producers must crave, to avoid fines. The restrictiveness of the scheme was plainly acknowledged by the Court (para. 83), although it focused on the trade of electricity rather than the trade of guarantees of origin, an impossibly murky issue that would have required determining whether intangible guarantees qualify as “goods”.
The Court therefore examined whether the scheme was justifiable “on one of the public interest grounds listed in Article [34 TFEU] or by overriding requirements” and whether it satisfied the test of proportionality (para. 89).
The purpose of the scheme was found in the promotion of the use of renewable sources for producing energy (para. 95). This legitimate purpose does not feature in Art. 36 TFEU and therefore, by exclusion, was provisionally accepted by the Court under the doctrine of the rule of reason. Under this principle, domestic measures setting mandatory requirements on goods, which restrict inter-state trade, are justified if they pursue a reasonable policy objective and the restriction entailed is not disproportionate to the contribution made towards that objective. Mandatory (or “overriding”) requirements are not limited to the values listed in Art. 36 TFEU, and routinely include objectives such as consumer interests and environmental protection. In its customary incarnation, the rule of reason can only be used to justify rules that do not discriminate according to the origin of the goods.
The reasoning of the Court on the proportionality of the Flemish scheme is not linear, but proceeds by accumulation. It essentially lists the reasons why the scheme is not as vicious as Essent would claim, invoking contextual elements, policy considerations and semantic tricks. In short, the Courts makes the following remarks, which I briefly assess in the brackets.
Essent claimed that the Flemish scheme does not encourage the consumption of green energy, but only its production. The Court objected (para. 98) that this is normal, as the greenness of energy relates precisely to its method of production (This is a massive non sequitur. If taken seriously, this would mean that green energy might as well be wasted after production, as the benefit to the environment was done already. To the contrary, the only raison d’être of renewable energy is that it is supposed to supplant the consumption of non-renewable energy).
The Court also noted that Member States are allowed by the Directive to set national targets of production of green energy (para. 99), suggesting perhaps that the discrimination inherent in the support scheme was dictated by the concern to achieve these targets (This suggestion is misleading. Whether its motivation is allowed or not by EU law has nothing to do with the EU-compliance of the Flemish measure).
The Court answered also the most difficult question. If the Flemish scheme genuinely aims to increase the production of green energy for environmental purposes, how come it does in fact discourage the purchase of green energy produced abroad, by refusing to accept foreign guarantees of origin as substitutes of Flemish green certificates? The brief answer is worthy of a Sibyl: “it should be observed that the starting points, the renewable energy potential and the energy mix of each Member State vary” (para. 100) (This is too obscure to attempt a reading. Suffice it to say that because guarantees of origin are subject to mutual recognition under the Directive, one would presume that these alleged variations are capable of being taken into account to certify the green nature of the energy, therefore an equivalence between Flemish certificates and any other guarantee is arguably possible to concoct).
Finally, the Court noted (para. 101) that national support schemes are important to achieve the objectives of the Directive (irrelevant per se) and that “it is essential that Member States be able to control the effect and costs of their national support schemes according to their potential, whilst maintaining investor confidence” (para. 102) (With a bit of effort, one can see why discrimination is justified: Member States do not want to make available to all EU producers subsidies whose cost is “ultimately [borne] by the [local] consumers” (para. 107). Also, they cannot suddenly withdraw support schemes when it turns out that they operate mainly for the benefits of foreign producers: investor claims would follow and hit hard).
“In the light of the foregoing,” the Flemish scheme was found to be proportionate, and therefore compliant with Article 34 TFEU (para. 103).
The Court then extended the proportionality analysis to other accessory elements of the schemes. It considered the penalties imposed, leaving it to the national judge to assess their proportionality to the sought objective (para. 114). It also speculated on the fairness of the market of green certificates that can fulfil the compulsory quotas, whose supply is artificially restricted. The Court therefore advocated the establishment of mechanisms under which these certificates can be traded “under fair terms” (para. 111). The irony of the last remark is highlighted in the comments below.
The Court nonchalantly applied a mandatory requirement to a facially discriminatory measure, sending Cassis to the museum of judge-made law no longer in force, and tacitly ignoring the exhaustiveness of Article 36 TFEU. It conflated overriding requirements and values listed in Article 36 TFEU into a single pool of excuses.
The passing reference to the contribution of reduced emissions to human health (para. 93) should not mislead: if human health hat were indeed the relevant objective of the Flemish scheme, it would have been at the centre of the ensuing proportionality test (just to mention one element, there should have been an analysis of the Flemish scheme’s contribution to increased health protection). That the quota has discriminatory effects which hinder trade was also undoubtable.
Therefore, a distinctly applicable measure was justified on the basis of a public interest other than those listed in Article 36 TFEU. This had also been the case in the parallel case of Ålands Vindkraft, C‑573/12, on which the Court delivered its judgment on 1 July 2014; see paras. 76-77 (the Opinion of this case was also prepared by Bot, and delivered shortly after his Opinion in Essent). Indeed, the Court has seemingly lost interest in making a big deal of the difference between discrimination de jure and discrimination de facto.
Advocate Bot had suggested the Court to abandon the distinction and expressly overturn its previous case-law. The Court did the former but not the latter, suddenly dropping a carefully constructed judicial test that it itself had devised in the wild years of Article 34 TFEU to accommodate reasonable policies which fell outside the scope of Article 36 TFEU.
That direct and indirect discrimination should be treated alike makes perfect sense. This is, for instance, the practice in the GATT, the agreement regulating the removal of trade barriers for trade in goods in the WTO regime. Treating them differently, for instance allowing one to be justified by excuses that do not apply to the other, is irrational: the hindering effect of de jure or de facto discriminatory measures can be identical in the facts, and a harsher treatment of direct discrimination serves only as an incentive for Member States to disguise discrimination on the basis of nationality within the folds of maliciously designed neutral measures.
However, there was some intuitive value to the bygone distinction: direct discrimination is prima facie illegal and must be hard(er) to justify. Think of the judgment in Test-Achats, which does not even attempt to look at a justification for gender-based direct discrimination. In fact, as registered by AG Kokott in the Opinion to that case (para. 59), justifications based on statistical data can support a proportionality assessment of indirect discrimination but cannot suffice to excuse direct discrimination.
In his proposal, AG Bot had reflected this concern, calling for a reinforced test of proportionality applicable to distinctly applicable measures: the reasons advanced to justify direct discrimination must be of particular weight. See at para. 94: “I think that discriminatory measures, particularly those which infringe a principle as fundamental as that of the prohibition of direct discrimination on grounds of nationality, ought to be subject to a strict requirement of proportionality” (emphasis added).
The Court did not deem it necessary to devise a reinforced (“strict”) proportionality test and, in my view, this might have contributed to a puzzling proportionality analysis, which features several controversial passages.
In short, by applying the routine test of proportionality the Court focused on the pursued objective (protection of the environment through increased production of green energy). In so doing, the discriminatory (and trade-restrictive) nature of the Flemish scheme was side-lined in the analysis. Once a decent contribution to the environment was somewhat found, it was downhill from there. The happy run of proportionality apparently skipped the stop of necessity: was there a less restrictive measure available for Flemish authorities to encourage the consumption and production of green energy? Yes there was: the interchangeability of green certificates and foreign guarantees of origin would have averted trade-restriction and promoted environmental protection even more than the Flemish scheme could possibly do.
The measure can only appear proportionate if the elective purpose is another unconfessed one, which is itself somewhat discriminatory. For instance, it could be argued that the growth and operation of a local green-energy industry is the real purpose, and therefore no alternative policy is available: any infant industry policy inherently involves some deal of temporary protectionism, but is for a good cause. The Court, however, did not resort to this “lesser-evil” reasoning, and pretended that discrimination is fully justified by the environmental purpose of the measure at large.
The irony of the finding is the after-thought of the Court with respect to the fairness of the market for green certificates. By virtually banning substitutable foreign certificates, the Flemish government is providing absolute protection to local producers, which might be few in number and, by definition, have the luxury of setting a higher price than they could if an undistorted transnational market were in function. Formally, there is no ban, but the value of local certificates is artificially inflated by the Flemish schemes, and makes it inconvenient for energy-suppliers to purchase green energy from abroad. This is a textbook example of a market distorted by protectionism, where demand is forced to opt for the local product, and foreign competitors are treated much less favourably. The Court concedes that buyers of green certificates, which need to comply with a compulsory quota, might find it unfair that the supply of certificates is artificially restricted to the collective monopoly of local producers, which can charge above-market prices. Local content requirements such as those envisaged by the Flemish scheme constitute a prototypical instrument of protectionism, and not surprisingly they are included, for instance, among the measures prohibited in the WTO Agreement on Trade-Related Investment Measures. It does not get any more protective and trade-distorting than that.
However, the Court limits itself to elegantly call for some “mechanisms” to ensure that suppliers can purchase green certificates “under fair terms” (para. 111), a puzzling statement in light of the previous validation of a protectionist scheme which is trade unfairness incarnated.
Ultimately, this is a memorable case, because it seemingly puts to bed the rule of reason as we knew it. Also, the reasoning on the merits is very hard to decode because the dropping of direct discrimination as a crucial element of the analysis resulted indeed in a confusing assessment of proportionality, which often loses sight of the differential treatment at stake and forgets about the necessity test altogether.
Barnard & Peers: chapter 12, chapter 16, chapter 22