Peter Van Elsuwege, Ghent European Law Institute

What was still a hypothetical scenario few months ago has now become reality: the Dutch citizens rejected the approval of the EU-Ukraine Association Agreement (AA) with a majority of 61 % and with a turnout of 32.28 %. As a result, the referendum is considered “valid” under the Dutch Advisory Referendum Act (DAR) and it is now for the Dutch government to decide “as soon as possible” upon the actual implications of this result. Before analyzing the potential legal options, some preliminary remarks are in place.

First, both chambers of the Dutch Parliament had already adopted the approval act, required for the Dutch ratification of the AA with Ukraine. The question is thus whether this parliamentary decision should be “corrected” due to the outcome of the popular vote. Since the referendum was only of a consultative nature, there is no legal obligation to repeal the approval act. However, in the days before the referendum, all political parties explicitly stated they would take the outcome very serious in case of at least 30 per cent turnout. Consequently, simply ignoring the public poll is politically speaking not an option. This was also made clear by Prime Minister Rutte when he declared that the agreement cannot be ratified under the present circumstances. The Dutch government, therefore, launched a “reflection period” to explore the possible options. According to Foreign Minister Koenders, this exercise should be finished by September at the latest. It seems no coincidence that also the EU-Ukraine Summit, planned to take place in May, has been postponed until September when also the outcome of the that other major referendum – on Brexit – will be known.

Second, the AA is a so-called “mixed agreement”, requiring a double ratification process: at the Member State level as far as the Member States’ competences of the Agreement are concerned and at the Union level with respect to the EU competences. Except for the Netherlands, where the ratification process is currently on hold due to the outcome of the referendum, all other Member States already finished the ratification process. At the EU level, the European Parliament already gave its consent but, in line with Article 218 (6) TFEU, the Council still needs to adopt a decision concluding the agreement. Such a decision is still pending and has to be adopted by unanimity. Hence, the two remaining steps in the ratification process are closely interconnected. The Council decision on concluding the agreement will not be adopted as long as there is no clarity concerning the implications of the Dutch referendum and the position of the Dutch government.

Third, a significant part of the AA already provisionally entered into force. A Dutch decision to reject the approval act will not automatically halt this practice, precisely because the provisional application only concerns those matters falling within the Union’s competence. This caveat is explicitly included in the relevant Council Decisions. In other words, the provisional application of the AA continues in anticipation of the finalisation of the ratification procedure, i.e. the adoption of a Council Decision concluding the agreement. There is no time limit for adopting the latter decision and it is not uncommon that the provisional application lasts for several years (with a remarkable record of 11 years for the agreement on cooperation and customs union with San Marino). In any event, termination of the provisional application on behalf of the EU is only possible upon a unanimous decision in the Council (see Art. 486 AA and Art. 218 (8) TFEU).

Of course, the question arises what kind of options exist to solve the existing conundrum. Re-opening negotiations on this comprehensive deal, which is the result of a long and careful process and which been signed by the Netherlands and the EU and has already been ratified by 27 EU Member States and Ukraine, is out of question. Broadly speaking, three more realistic options can be distinguished (See for further comments on the possible options also the analysis of Guillaume van Der Loo and Ramses Wessel/Adam Lazowski).

A first, and from a legal perspective, easiest solution would be the addition of interpretative declarations. One may, for instance, imagine a clarification that the AA does not include any commitment regarding the EU accession prospects of Ukraine or a statement that the AA will not lower animal welfare rights (as had been argued by the Dutch Party for the animals in the referendum campaign) This may be combined with other political initiatives to increase the democratic legitimacy of the EU decision-making process, since this was also one of the concerns of the “no-camp”. In fact, the possibilities for such creative political solutions are unlimited since they do not require a formal amendment of the Agreement. The main question, of course, is whether this will be sufficient to convince the Dutch ‘no’ voters that the referendum outcome was taken seriously.

A second option could be to add a so-called “adjusting protocol” to the agreement. This is precisely what happened after the Swiss Confederation was not in a position to ratify the European Economic Area (EEA) Agreement in the 1990s after a negative referendum. Of course, the situation is slightly different since the Netherlands is not just a third party but an EU Member State. However, the legal logic could be similar. It would imply that “the Netherlands” should be deleted as one of its contracting parties. As a result, those provisions of the AA belonging to Member State competences would not be applicable in the Netherlands. One could, for instance, think about the provision on mobility of workers (Art. 18 of the AA) whereas most other provisions belonging to EU competences would still be binding for the Netherlands in their capacity of EU Member State. This option presupposes that a distinction is made between the role of the Netherlands as a sovereign state, on the one hand, and its role as EU Member State in the Council, on the other hand. In other words, a potential repeal of the Dutch Approval Act and, as a consequence, the withdrawal of the Netherlands as a contracting party to the AA should not necessarily imply that the country blocks the adoption of the Council decision concluding the AA at EU level. As found by the Court of Justice in Case C-28/12, the procedures for concluding a mixed agreement at Member State level and EU level are separate and this, mutatis mutandis, also applies for the ratification process. Whereas this option would thus be legally sound, it requires a somewhat schizophrenic approach of the Dutch government (approving the agreement in the Council at EU level and opposing ratification at the national level). Moreover, it may lead to competence disputes since the agreement does not clearly distinguish between EU and Member State competences – even though the list of provisions falling under provisional application could help.

A third scenario is the status quo. It may be expected that the Council Decision concluding the AA will not be adopted as long as there is no clarity about the position of the Netherlands. At the same time, it is very unlikely that the Council will unanimously decide upon the withdrawal of the AA’s provisional application, even if the Dutch government would indicate its inability to ratify the agreement. If such a Dutch position would prevent the adoption of the Council Decision concluding the agreement, a legal deadlock may thus exist where the provisional application applies indefinitely. Arguably, such an outcome is difficult to justify from a legitimacy point of view. It would be against the very logic of the mechanism of “provisional” application, which anticipates upon the ratification of the agreement. It if turns out that ratification of an agreement will be impossible, it is hard to justify its continued provisional application. However, termination of the provisional application can only be expected when there is a viable alternative which is acceptable for all parties involved. The result may thus be a vicious circle where the status quo is for the time being the only option. This would not be too problematic since the most significant parts of the agreement already provisionally entered into force. Moreover, the planned abolition of the visa requirement for short term travels does not depend upon the full entry into force of the AA. This is part of a separate process which only requires the approval of the European Parliament and the Council, which decides by qualified majority on this matter (Art. 77 (2)(a) TFEU).

In sum, whereas the direct legal implications of the Dutch referendum may be limited, the search for an acceptable solution is a delicate political exercise. Whatever the final option will be, it will require political courage to explain the complexities of the EU’s system of multi-level governance. Moreover, the Dutch referendum saga rewards a deeper reflection upon the practice of mixed agreements as a tool of EU foreign policy. Opposition against the EU-Canada Comprehensive Economic and Trade Agreement (CETA) inside Belgium and in the Netherlands as well as the ongoing preparations for a Dutch referendum on the anticipated Transatlantic Trade and Investment Partnership (TTIP) illustrate that this discussion goes beyond the EU-Ukraine AA. In a context where the legitimacy of the European project is under pressure, the importance of this discussion cannot be underestimated.