Proof of military compliance as a condition for temporary protection in the Netherlands

This article states the position as at 7 August 2026. Council Implementing Decision (EU) 2026/1912 was published in the Official Journal on 4 August 2026, the IND published its notice on 5 August 2026, and the Commission’s operational guidelines had not been adopted at the time of writing. The subject moves quickly: check the date above before relying on anything below.

If you already held temporary protection in the European Union before 5 August 2026, nothing changes for you. Your protection continues automatically until 4 March 2028, and the new military condition does not apply to you.

If you apply in the Netherlands for the first time on or after 5 August 2026 and you are subject to conscription in Ukraine, you must now show that you have satisfied your military obligations, or that you are exempt from them. The IND has published one further exception: the restriction does not apply to people joining family members who already hold temporary protection here.

What follows is what Council Implementing Decision (EU) 2026/1912 actually says, what it means in practice at the IND, and where the realistic points of legal attack lie. It is written for practitioners and for the people affected. It is not individual legal advice.

1. What changed on 5 August 2026

For four and a half years, the Temporary Protection Directive (Council Directive 2001/55/EC, “TPD”) operated in relation to Ukraine on a single, deliberately crude premise: you are Ukrainian, you fled the war, you are protected. No individual assessment, no merits test, no interrogation of why you left rather than stayed. That was the entire point of the instrument: Article 2(a) TPD defines temporary protection as an “exceptional character” procedure designed to give immediate collective protection while reducing formalities to a minimum.

On 4 August 2026 that premise was abandoned for a defined class of people. Council Implementing Decision (EU) 2026/1912 of 30 July 2026 extends temporary protection until 4 March 2028 and, in the same breath, makes it conditional, for the first time, on an applicant proving that he has complied with the conscription laws of the country he has fled.

The Union has, in substance, made access to European protection contingent on compliance with a third state’s mobilisation regime. Whatever one thinks of the policy, and the policy rationale is not frivolous, the legal construction used to achieve it is fragile, internally inconsistent, and in at least three respects almost certainly unlawful.

2. What the Decision actually says

The operative part is short. It is worth reading precisely, because most of the reporting has been imprecise.

Article 1 extends temporary protection for the categories of displaced persons identified in Article 2 of Implementing Decision (EU) 2022/382 for a further year, until 4 March 2028.

Article 2, first paragraph provides that, without prejudice to Union law and fundamental rights, temporary protection “shall only be granted to those who satisfy their military obligations in Ukraine, upon presenting, where applicable, proof thereof.”

Article 2, second paragraph carves out those already protected: the Article does not apply to persons enjoying temporary protection in a given Member State before or on 30 July 2026 and continuously retaining that status in that given Member State thereafter.

Article 3 provides that the Decision enters into force the day after publication, that is 5 August 2026, and applies from 5 March 2027, “with the exception of Article 2 which shall apply from 31 July 2026.”

The legal basis invoked is Articles 4(2) and 5(3)(a) TPD. Denmark is not bound; Ireland is.

Three features of the drafting deserve immediate attention.

2.1 The Decision contradicts itself on the start date

Article 3 says Article 2 applies from 31 July 2026. Recital 17 says the requirement “should apply from the entry into force of this Decision, which should be the day after its publication in the Official Journal,” that is, 5 August 2026. Recital 17 also frames the grandfather clause by reference to “the date of the entry into force of this Decision,” whereas Article 2 fixes it at 30 July 2026.

These cannot both be right. The operative text produces retroactive effect: a rule restricting access to protection, published on 4 August, purporting to govern applications made from 31 July. The recital explicitly disclaims exactly that outcome.

The Netherlands has resolved the contradiction in the applicant’s favour. The IND’s notice of 5 August 2026 states throughout that the new condition applies to applications from 5 August 2026, and that those who already held temporary protection before 5 August 2026 are unaffected. Czechia has taken the same operative date. Practitioners in other Member States should check: an applicant who registered between 31 July and 4 August 2026 sits precisely in the gap, and in the Netherlands he is protected.

2.2 The grandfather clause is narrower in Brussels than in The Hague

Article 2 protects those who held temporary protection “in a given Member State” and continuously retained it “in that given Member State.” On the face of the EU text, protection acquired in Poland does not immunise a subsequent application in the Netherlands: moving triggers a fresh grant, and with it the military-obligations condition.

The IND’s published notice is materially more generous: it refers to persons who already held temporary protection in the European Union before 5 August 2026. On the IND’s own wording, prior protection anywhere in the Union removes the military condition.

That divergence is significant, but it should not be over-read. Since 15 June 2026, IND policy IB 2026/20 independently bars applicants who hold or held temporary protection in another Member State, a policy which raises its own serious questions under Article 8(1) TPD and Krasiliva (C-753/23, EU:C:2025:133), and which recitals 5 and 6 of the new Decision now attempt to bolster. The practical effect for a client who moves is therefore unchanged: he is refused, only on a different ground.

There is a real trap here, and it should be stated plainly to clients. Renouncing or allowing the lapse of temporary protection in one Member State in order to re-apply in another is now, in most cases, an irreversible mistake. Article 2 requires continuous retention. Break the continuity and the military-obligations condition attaches to everything that follows.

2.3 The Dutch family carve-out

The IND notice adds a limitation found nowhere in the EU text: the restriction does not apply to persons joining family members who already hold temporary protection under the RTB.

This matters enormously in practice, and it is the single most useful line in the IND’s publication. It reflects Article 15 TPD, which imposes binding family-reunification duties, and Article 2 of Decision 2022/382, which brings family members within the protected category in their own right. Recital 15 of the new Decision gestures at this, the criteria are to be applied “in compliance with Article 15 of [the TPD], where applicable,” but does so obliquely.

For a Ukrainian man whose wife, partner or minor children already hold temporary protection in the Netherlands, this carve-out should be invoked expressly and in terms at the point of application, with documentary proof of the family link and of the family member’s status. Do not assume the IND will apply it ex officio.

3. What “military obligations” actually means, and the gap in the drafting

Here the Decision is at its weakest, and practitioners should exploit the weakness.

The operative provision conditions protection on satisfying military obligations in Ukraine. Recital 18 conditions it on something quite different: that the individual “was authorised to leave the territory of Ukraine legally,” evidenced by lawful border crossing or an exit stamp, failing which “an easily verifiable official document, in paper or electronic form, such as in the Reserv+ application,” confirming exemption from or compliance with military obligations. The burden of proof rests on the applicant.

Compliance with military obligations and authorisation to exit are not the same thing under Ukrainian law, and the difference is not academic.

• Men aged 18–22. Since 28 August 2025, following the Cabinet of Ministers’ amendment to the border-crossing rules, they may leave Ukraine freely during martial law, subject to holding a valid passport and military registration document. They should ordinarily be able to satisfy the condition outright.
• Men aged 23–24. They are barred from leaving, but they are not subject to mobilisation, which begins at 25. Their military obligations consist of registration (військовий облік), not service. A 23-year-old fully compliant with his registration obligations, with a clean Reserv+ record, satisfies the operative text of Article 2 even though he left irregularly and has no exit stamp.
• Men aged 25–60. Mobilisation-liable, barred from exit save on statutory grounds (deferment, exemption, disability, multiple children, guardianship, reservation / бронювання, humanitarian transport, certain students).
• Men over 60, and women. Generally no military obligations. Recital 20’s attempt to gender-neutralise the measure by noting that obligations may exist “irrespective of age or gender” through enlistment or “responsibilities related to military activities” is, on any realistic view, a drafting fig leaf.

The consequence is straightforward and should be argued in every case: Article 2 imposes a compliance test, not an exit-authorisation test. Recital 18 supplies an evidential proxy, not a legal condition. It is settled law that the preamble to a Union act has no binding force and cannot be relied on to derogate from its operative provisions (Nilsson, C-162/97; Tyson Parketthandel, C-134/08). An applicant who can show compliance with, or exemption from, his obligations, a deferment decision, a VLK unfitness ruling, a reservation certificate, disability documentation, a Reserv+ extract showing no outstanding liability, satisfies Article 2 whether or not he crossed the border lawfully.

Any decision refusing protection solely on the absence of an exit stamp is therefore, in my view, taken on a wrong legal test and is reviewable on that ground alone.

4. The practical position in the Netherlands

A man refused under the new condition is not left without any route, but the routes are thin.

Article 17 TPD guarantees the right to lodge an application for international protection at any time, and Article 18 of the Charter guarantees the right to asylum. Recital 8 of the Decision expressly preserves this, while adding the pointed observation that “desertion and draft evasion are not considered, by themselves, to be grounds for international protection.” That statement is accurate as far as it goes, and materially incomplete, as section 5.5 explains. The Dutch asylum procedure itself was rebuilt on 12 June 2026, and the changes affect how such an application is handled.

The realistic outcome is a hard one. EUAA figures put the recognition rate for Ukrainian applicants at approximately 45% in May 2026. A refused applicant faces a return decision, subject to individual non-refoulement assessment under Article 5 of the Return Directive, Article 19(2) of the Charter and Article 3 ECHR. Meanwhile Dutch reception capacity for this group is, on the government’s own account, effectively exhausted. Applications made from within temporary protection carry their own difficulty, which we have described separately in the treatment of asylum applications by temporarily protected persons from Ukraine.

The procedural route against an IND refusal is the ordinary one: bezwaar within the statutory period, a verzoek om voorlopige voorziening where removal or loss of provisions is imminent, then beroep to the rechtbank and hoger beroep to the Afdeling. That is also the only route by which the Decision itself can realistically be challenged, for the reasons set out below.

5. Where the Decision is vulnerable

I set out the arguments in descending order of strength. I do so without optimism about the short term: none of these will be resolved quickly, and a practitioner who tells a client otherwise is doing him a disservice.

5.1 Legal basis and vires: the strongest argument

Articles 4(2) and 5(3)(a) TPD are thin foundations for what has been built on them.

Article 4(2) confers a power to extend the duration of temporary protection where the reasons for it persist. It says nothing about redefining who benefits. Article 5(3)(a) requires the establishing decision to include a description of the groups covered, a provision about the content of the initial activation, not a rolling licence to re-draw the protected class annually.

More fundamentally: removing a defined group from the scope of protection is, functionally, a partial termination of protection for that group. Article 6(2) TPD subjects termination to a substantive condition, the Council decision must be based on the establishment that the situation in the country of origin permits safe and durable return, with due respect for human rights and Member States’ non-refoulement obligations.

The Council cannot possibly satisfy that test here, and it says so in its own recitals. Recital 9 states that the situation in Ukraine does not allow the majority of displaced people to return safely and durably. Recital 10 records continuing large-scale air attacks on civilians and a persisting risk of escalation. The Decision simultaneously asserts that Ukraine is unsafe for return and removes protection from the group most exposed to the conflict. That contradiction sits on the face of the instrument.

There is a serious counter-argument, and it should be anticipated. The scope of the current temporary protection regime is defined not by the Directive but by Decision 2022/382, which the Council has amended repeatedly. If the Council may widen scope, why not narrow it? And the Court in Kaduna (C-244/24 and C-290/24) accepted that Member States may terminate optional temporary protection they had granted beyond the EU minimum. That reasoning might be extended by analogy to the Union level.

The answer is that Kaduna concerned protection which was optional in the first place, granted under Article 7 TPD; the group affected here falls within the mandatory core of Article 2 of Decision 2022/382. And whichever route is taken, the Article 6(2) safe-return test is not met.

There is also a structural argument available under Article 291 TFEU and the Court’s case law on the boundary between implementing and delegated measures: an implementing act may not supplement or amend the essential elements of the basic act. Who is entitled to protection is about as essential an element as exists. Combined with Article 28 TPD, which sets out an exhaustive list of exclusion grounds framed on Refugee Convention lines and containing nothing resembling draft evasion, the vires case is real.

5.2 Discrimination

The measure is facially neutral and, in operation, applies overwhelmingly to men aged 18–60. That engages Article 21 of the Charter and Article 14 ECHR read with Articles 3 and 8. The Council of Europe Commissioner for Human Rights flagged precisely this risk before adoption.

Recital 22’s assertion that the Decision “respects fundamental rights,” and Article 2’s “without prejudice to Union law and fundamental rights” formula, are not analysis. They are the absence of analysis. No proportionality assessment appears anywhere in the instrument.

The State’s answer will run through Article 52(1) of the Charter: supporting Ukraine’s capacity for self-defence is an objective of general interest, and the measure is proportionate provided that individual non-refoulement assessment survives. That is a respectable argument, and it may well prevail. The counter is that indirect sex discrimination in access to protection from armed conflict requires the most compelling justification; that no assessment of less restrictive means was carried out; and that the Union has effectively delegated the definition of its own protection scope to a third state’s evolving mobilisation policy, a policy over which neither the Union nor the individual has any control, and which the Union cannot review.

5.3 Retroactivity and legal certainty

As set out at 2.1, Article 3 gives Article 2 effect from 31 July 2026, five days before publication. The principle of legal certainty precludes a Union measure taking effect before publication save exceptionally, where the objective pursued so demands and legitimate expectations are duly respected (Racke, 98/78). Neither condition is arguable here: recital 17 states that the intended date was entry into force, and no operational necessity for backdating is identified anywhere in the instrument.

This is a narrow point, affecting a small cohort. It is also the cleanest available point, and in a well-chosen case it could be the vehicle that gets the whole instrument in front of the Court.

5.4 Evidential impossibility, effectiveness and Article 47

A person who fled irregularly is, by definition, unable to produce evidence of regular departure. Placing the burden of proof on him while confining the acceptable evidence to documents issued by the authorities of the state he has fled comes close to conditioning protection on the cooperation of the very authority whose demands he is escaping. This sits uneasily with the confidentiality architecture of the asylum acquis and with Article 8 of the Charter, given that verification through Reserv+ necessarily involves interaction with Ukrainian state data systems.

The Decision provides no procedural framework at all: no definition of “easily verifiable,” no designated verifying authority, no evidential standard, no rule on what happens where a document cannot be obtained, no benefit-of-the-doubt provision, no stated right to a reasoned decision. Recital 19 defers all of this to future Commission Operational Guidelines: soft law, adopted after the condition took effect.

Article 47 of the Charter requires an effective remedy, and Krasiliva confirms its application in this field. A condition whose content is undefined and whose satisfaction is unverifiable cannot be meaningfully appealed. That is an effectiveness argument with real teeth.

5.5 The residual asylum route is narrower than recital 8 implies, and wider than the IND may assume

Recital 8 is correct that draft evasion is not, per se, a ground for international protection. It is not the whole law.

Article 9(2)(e) of the Qualification Directive treats prosecution or punishment for refusal to perform military service in a conflict where service would entail commission of war crimes as an act of persecution (Shepherd, C-472/13). In EZ (C-238/19) the Court held that, in certain contexts, there is a strong presumption that refusal of military service connects to a Convention ground, in particular political opinion. Conscientious objection engages Article 9 ECHR (Bayatyan v Armenia, GC, 2011). And separately from the refugee definition, conditions of forced mobilisation, treatment of those recorded as СЗЧ, and detention conditions may found subsidiary protection or an Article 3 ECHR bar to removal on the specific facts.

Recital 8 must not be permitted to operate as a shortcut past individual assessment. Where it is invoked as one, that is itself a reviewable error.

5.6 Article 7 TPD survives, and that is a political route, not only a judicial one

Recital 15 expressly preserves Article 7 TPD, which permits Member States to extend temporary protection to additional categories of displaced persons.

The Decision therefore does not prohibit the Netherlands from continuing to protect this group. It merely ceases to require it. That converts part of the question from a legal one into a political one, and it is worth saying so to clients, to Parliament, and in submissions: the State cannot hide behind Brussels. If men aged 25–60 fleeing Ukraine are refused in the Netherlands, that will be a Dutch choice made under a Union permission, not a Union command. Kaduna sets out the consequences that follow from exercising the Article 7 option, including the conditions for withdrawing it, which is a reason for care, not a reason for inaction.

6. How a challenge would actually be brought

Realism is required about the procedural route.

A direct action for annulment under Article 263 TFEU by an affected individual will almost certainly fail on admissibility. The Plaumann test for individual concern is not satisfied by membership of an open, objectively defined class, and a Council implementing decision addressed to Member States is not a regulatory act not entailing implementing measures.

The route is therefore national. A refusal by the IND, challenged through bezwaar and beroep, in which the validity of Article 2 of Decision 2026/1912 is put in issue, by direct challenge or by way of the plea of illegality under Article 277 TFEU, with a request for a preliminary reference under Article 267 TFEU. National courts have no power to declare a Union act invalid themselves (Foto-Frost, 314/85); a reference is mandatory if the court considers the ground of invalidity well founded. The Afdeling, as a court of last instance, is under a CILFIT obligation.

The narrower questions, the 31 July / 5 August retroactivity conflict, and the correct legal test under Article 2 as against recital 18, can and should be litigated as matters of interpretation without waiting for Luxembourg. They are available now, in the rechtbank, and they are winnable.

Timescales for the vires and discrimination challenges are realistic only if stated honestly: a preliminary reference filed in late 2026 will not produce a judgment before 2028, by which time the Decision will have expired and the question may be moot, a structural feature of annual instruments that insulates them from review, and one that itself deserves comment.

Strategic litigation, coordinated with NGOs and with the Dutch Council for Refugees, is a better vehicle than an individual case for the systemic points. Individual cases should be run on the narrow, fast arguments.

7. Practical guidance for a Ukrainian man seeking protection in the Netherlands now

1. If you registered before 5 August 2026, you are outside the new condition. Your protection continues automatically until 4 March 2028. Do nothing that interrupts it, and read what the rules say about leaving the Netherlands temporarily before you travel.
2. Do not give up status in one Member State to move to another. Continuity of status in the same Member State is what the exemption protects. Breaking it exposes you to both the military-obligations condition and, in the Netherlands, IB 2026/20.
3. If you are joining family who already hold temporary protection here, say so explicitly, in writing, at the point of application, with proof of the relationship and of their status. The IND has published a carve-out for this. Invoke it, and invoke Article 15 TPD alongside it.
4. If you are 18–22, gather your passport and military registration document. Lawful exit for your age group has been permitted since 28 August 2025 and should ordinarily satisfy the condition.
5. Assemble compliance evidence, not just exit evidence. Deferment decisions (відстрочка), VLK unfitness rulings, reservation (бронювання) certificates, disability documentation, proof of three or more children, guardianship orders, student status, a Reserv+ extract showing no outstanding liability. Any of these goes to the statutory test in Article 2 even if you have no exit stamp.
6. Keep proof of the date you applied. If you applied between 31 July and 4 August 2026, the retroactivity point is available to you.
7. If refused, lodge bezwaar in time and seek interim relief. Do not let the deadline pass while considering options.
8. Consider an asylum application in parallel, but do not treat it as a safety net. Recognition rates are far from assured, and refusal brings a return decision. Take advice on whether your individual circumstances engage Article 9(2)(e) QD, EZ, conscientious objection, or an Article 3 ECHR risk before framing the claim.
9. Instruct a lawyer before you register, not after you are refused. The framing of the initial application now determines the legal test the IND applies to it. Our immigration law practice handles exactly these applications and appeals.

8. Conclusion

Decision 2026/1912 is a political instrument wearing legal clothing that does not fit it.

It asserts in recitals 9 and 10 that Ukraine is not safe for return, and in Article 2 removes protection from those most exposed to the war. It states in recital 17 that the new condition takes effect on publication, and in Article 3 backdates it. It conditions protection on satisfying military obligations, then supplies an evidential test measuring something else entirely. It creates a category of person who may be refused temporary protection, refused asylum, and yet not lawfully removable, and provides no answer to what becomes of him. It defers the entire operational framework to guidelines that did not exist when the rule took effect.

The policy objective is legitimate, and it deserves to be stated fairly: Ukraine’s survival depends on its capacity to field an army, and the Union has an interest in not undermining that. But a Union which conditions its own protection standards on a third state’s mobilisation policy has given away something it may find difficult to reclaim, and it has done so through an implementing act, without the scrutiny that amending the Directive would have required. That, more than any individual refusal, is what should trouble lawyers.

For now, the litigation that matters is narrow and immediate: the start date, and the correct legal test. Those arguments are available today, in a Dutch courtroom, and they are strong. The larger questions of vires and discrimination will take years, and may well be overtaken by the expiry of the instrument that raises them.

Sources

• Council Implementing Decision (EU) 2026/1912 of 30 July 2026, OJ L, 4.8.2026 (data.europa.eu/eli/dec_impl/2026/1912/oj)
• Council Directive 2001/55/EC of 20 July 2001 (Temporary Protection Directive)
• Council Implementing Decision (EU) 2022/382 of 4 March 2022
• IND, “Richtlijn Tijdelijke Bescherming voor Oekraïne verlengd, met nieuwe voorwaarde voor nieuwe aanvragen”, 5 August 2026
• Council of the EU, press release, 15 July 2026
• CJEU: C-753/23 A.N. v Ministerstvo vnitra [Krasiliva], EU:C:2025:133; C-244/24 and C-290/24 Kaduna; C-472/13 Shepherd; C-238/19 EZ; C-297/17 etc. Ibrahim; 314/85 Foto-Frost; 25/62 Plaumann; 98/78 Racke; C-162/97 Nilsson
• ECtHR: Bayatyan v Armenia [GC], no. 23459/03, 7 July 2011
• M. İneli Ciğer and S. Peers, “What is wrong with the decision to exclude some Ukrainians from temporary protection?”, EU Law Analysis, 5 August 2026
• EUAA, Latest Asylum Trends, recognition rates, May 2026