Friday 24 March 2023

Anti-discrimination and labour rights: CJEU confirms protection from discrimination (including on grounds of sexual orientation) covers self-employed workers


Chiara De Capitani, linguist agent at the European Commission and PhD candidate and honorary fellow in European law at the University of Naples “L'Orientale”. The information and views set out in this article are those of the author and do not necessarily reflect the official opinion of the European Commission.

Photo credit: Silar, via Wikimedia Commons



In TP « (Monteur audiovisuel pour la télévision publique)», Case C-356/21 (the present case) the EU Court of Justice (ECJ) analyses a case dealing with the protection of self-employed individuals providing goods and services, discrimination on the basis of sexual orientation and its balancing with the right to freedom of contract.

Building on the previous HK v Danmark and HK/Privat case (Case C-587/20), this landmark ruling has been celebrated (Covington & Burling LLP; Countouris, Freedland and Stefano; Lasek-Markey) for its progressive interpretation of the protection provided by EU anti-discrimination law but, most of all, for its broad definition of which ‘workers’ are protected by it, “making labour law fit for all those who labour” (Countouris, Freedland and Stefano). Indeed, this case covers the situation of self-employed workers (whose rights lack clarity at EU level) but also of workers that may not fall into that category but ‘provide goods and services’.

Facts of the case

From 2010 to 2017, J.K. (the applicant) entered into a series of consecutive short-term contracts with TP, a public television channel in Poland, as a self-employed individual. During this time, J.K. worked on editing material for trailers and features that were later used in TP's promotional materials. He worked under the supervision of W.S. He worked two one-week shifts per month with another journalist, as assigned by W.S.

In August 2017, TP was planning to reorganize its structure and transfer J.K.'s tasks to a new unit. Two new employees were appointed to handle the reorganization and to assess the associates who would be transferred.

At meetings in October and November of that year, J.K. received a positive evaluation and was listed among the associates who passed the assessment. On November 20th, J.K. and TP entered into a one-month work contract. On November 29th, J.K. received his work schedule for December, which included two weeks of service.

However, on December 4th, J.K. and his partner published a Christmas music video promoting tolerance towards same-sex couples on their YouTube channel. Two days later, on December 6th, TP cancelled J.K.'s first week of service, and on December 20th, J.K. was informed that he would not be required for the second week as well. As a result, J.K. did not perform any service in December and was not paid for it. Subsequently no new contract for specific work was established between him and TP, and the decision to end the work collaboration was made by the employee(s) responsible for carrying out the reorganization.

J.K. filed a case at the District Court for the Capital City of Warsaw (the referring Court) seeking compensation, claiming that he was the victim of direct discrimination by TP due to his sexual orientation. He alleges that the probable reason for the cancellation of his work periods and the termination of his employment with TP was the publication of the previously mentioned Christmas music video on YouTube. TP argues that the case should be dismissed, as there is no guarantee of contract renewal in its practice or law.

The referring Court is uncertain to which extent self-employed workers are covered by the scope of the ‘Equality Framework’ Directive and has doubts about the compatibility of Polish Equality Law with the directive.


Absence of analysis as to whether there was discrimination

Recital (15) of the ‘Equality Framework’ Directive states that:

The appreciation of the facts from which it may be inferred that there has been direct or indirect discrimination is a matter for national judicial or other competent bodies, in accordance with rules of national law or practice […].

Following this and since the referring court based its request for a preliminary ruling on the premise that J.K.'s sexual orientation was the reason behind the refusal by TP to conclude the contract, neither the opinion of Advocate General Tamara Ćapeta (the AG) nor the ECJ analyse whether the applicant has actually been discriminated against or not.

This is regrettable as the case itself raised several interesting elements previously not addressed by the ECJ.

For one, the fact that the applicant’s shift was cancelled (and subsequently his contract not renewed) merely two days after his publication of a video aimed at promoting tolerance towards same-sex couples is quite suspicious and bears some resemblance with the Hakelbracht ruling (Case C‑404/18). In that ruling, the ECJ clarified that the protection against retaliation afforded by the ‘Gender Equality Directive’ applies to all employees who have informally supported a person who has been discriminated against. In that case, a company had dismissed an employee only eight months after she had objected to their refusal to hire a woman based on the latter’s pregnancy, a fact that the referring court in that case presumed to be retaliatory behaviour and that the ECJ duly noted in its ruling (para 24).

This presumption is all the more legitimate as J.K had worked regularly for seven years for the public television channel, TP.

Additionally, the decision to end the work collaboration was made by the newly recruited employee(s) responsible for carrying out the reorganization of the television channels’ structure only a couple of months after the applicant had received a positive evaluation and was listed among the associates who passed the assessment for said reorganization.

Furthermore, in its request for a preliminary ruling, the referring court notes that one or two days after the publication of the Christmas video a meeting took place which was attended by – among others – the applicant’s immediate supervisor W.S. and the employees responsible for carrying out the reorganization. While the meeting focused on the creation of content for Christmas to be broadcasted, one of the employees responsible for carrying out the reorganization made a sarcastic joke about the fact that “(Channel 1) already had a spot and its own Santas”. After the meeting, one of the employees responsible for carrying out the reorganization (possibly the same who made the sarcastic joke) gave instructions to W.S. to suspend the applicant and assign his shifts to another person.

The joke during a meeting expressly referring to the applicant’s video promoting tolerance towards same-sex couples (where a man dressed as Santa kisses another man) and following request to suspend him might hint to a correlation between the suspension (and following non-renewal) of the applicant’s contract and his sexual orientation.

So far the ECJ currently has ruled on three cases (Feryn, Case C-54/07; Asociaţia Accept, Case C‑507/18Associazione Avvocatura per i diritti LGBTI, Case C-507/18 – which I discussed here)  where an employer or a person perceived as being capable of exerting a decisive influence on the recruitment policy of an employer stated that they would not hire a person from a protected category. In these three cases, these statements were always public. In this specific case, it would have been interesting for the referring Court to ask the ECJ clarifications as how to assess situations where statements made privately implying prejudice are followed by a termination decision.

Finally, the referring court reported that the employees and associates of the editorial office where the applicant performed his tasks within the defendant’s organisation were aware of the applicant’s sexual orientation and that the defendant (TP) has argued that “the applicant’s sexual orientation was common knowledge”. This argument resonates in part with the Lee v. the United Kingdom European Court of Human Rights ruling (Application no. 18860/19) where the defendants (a bakery) had argued that they had employed and served gay people in the past and that their refusal to provide a good was not based on the applicant’s sexual orientation but on their refusal to put a pro-same sex marriage slogan on a cake. In the author’s opinion, this could have led to interesting reflections not only as to the discrimination of the present cases’ applicant on matters of employment, but also its implications as a possible retaliation to his right to freedom of expression in producing a video aiming at tolerance towards same-sex couples.

Scope of the protection afforded against discrimination in relation to access to employment

The first question the ECJ and the AG try to answer is whether Article 3 (1)(a) of the ‘Equality Framework’ Directive covers situations such as the one in the present case. According to said provision, the directive applies to ‘conditions for access to employment, to self-employment or to occupation’. Both the ECJ and AG proceed with examining what ‘conditions for access to employment, self-employment or to occupation’ entail: since the directive does not refer to national law to define this concept, it must be given an autonomous and uniform interpretation across the European Union.

The Court starts by noting, following its previous reasoning in HK v Danmark and HK/Privat, that the usual meaning in everyday language of ‘employment’, ‘self-employment’ and ‘occupation’ must be construed broadly and “cover conditions for access to any occupational activity, whatever the nature and characteristics of such activity” (para 36). In HK v Danmark and HK/Privat the Court had found that the post of sector convector of an organisation of workers constituted a real and genuine professional activity (para 35). Therefore, the applicant in that case was protected by the ‘Equality Framework’ Directive against discrimination on grounds of her age, even though her post was a political post (decided with elections by the members of that organisation) (paras 36–39).

The Court reaches a similar conclusion in the present case, noting that it was not the intention of the EU legislature to restrict the application of the ‘Equality Framework’ Directive solely to positions held by individuals classified as ‘workers’ according to Article 45 of the Treaty on the Functioning of the European Union (TFEU). As a matter of fact, whereas the ‘Equality Framework’ Directive was adopted on the basis of the (current) Article 19 (1) TFEU conferring the EU the power to combat discrimination, Article 45 TFEU only protects workers “as the weaker party in an employment relationship” (HK v Danmark and HK/Privat para 34, present case paras 40-43). In fact, the Court reiterates, the ‘Equality Framework’ Directive applies to “all persons […] whatever the branch of the activity and at all levels of the professional hierarchy” and was adopted to eliminate all discriminatory obstacles in the field of employment” (HK v Danmark and HK/Privat paras 29, 34, present case paras 38, 43).

What constitutes ‘work’ and its intersection with ‘provision of goods and services’

The Court, while noting that a wide range of occupational activities are protected by the ‘Equality Framework’ Directive, proceeds to highlight activities falling out of the scope of that directive which are “the mere provision of goods and services to one or more recipients” (para 44).

Furthermore, the Court adds that, to fall under the scope of the directive, occupational activities must:

be genuine,

be pursued in the context of a legal relationship that is

characterised by a degree of stability (para 45).

The Court does not further define these three requirements but applies them to the present case (while leaving it for the referring court to decide whether the applicant satisfies these criteria): J.K prepared personally specific work for the public television, depended on the assignment of his shifts and had received a positive evaluation (para 46). In other words, the Court finds, J.K pursued a genuine and effective occupational activity on a personal and regular basis for the same recipient, which enabled him to earn (at least in part) his livelihood (para 47).

The fact that the three criteria above seem to apply to the present case creates a situation where his ‘occupational activity’ with the company does not even need to be classified as ‘employment’ or ‘self-employment’ to fall under the scope of the directive (para 47).

This broad interpretation of the scope of the directive is highly welcome and will likely protect a lot of self-employed individuals and providers of goods and services that are currently left out in an overly de-regularised and flexible labour market. The shift of focus from the AG and the Court from the conditions of employment decided by an employer to, rather, the personal aspects of the work provided by the worker have been greeted favourably by academics (Countouris, Freedland and Stefano).

Nevertheless, the exclusion by the court of “the mere provision of goods and services” from the scope of the directive is not very clear and seems to rebut its previous finding that “any occupational activity, whatever the nature and characteristics of such activity” is covered by the scope of the directive. Coupled with the requirements that an occupational activity “be genuine”, “pursued in the context of a legal relationship” and that said relationship must be “characterised by a degree of stability” seem to exclude from the scope of the directive several occupational activities whereas currently, as the AG finds, “non-standard forms of work have increased, causing fragmentation in the labour market” (para 63).

Firstly, the definition of ‘services’ in Article 57 TFEU provides that they “shall be considered to be “services” within the meaning of the Treaties where they are normally provided for remuneration”. This creates – for the author – a presumption that the payment of these services implies the existence of an occupational activity. The AG underlines in her opinion that a person may be simultaneously self-employed and a provider of goods and services since “as recipients of their goods or services, we ‘buy’ their work and the end product of that work at the same time” (para 41).

Secondly, the Court does not further define in particular what constitutes a “genuine” occupational activity and while the author agrees that the existence of a legal relationship and its level of stability do create a presumption of an existing occupational activity, it is unclear why the lack thereof should be excluded from its scope. With regards to the “level of stability” required of the legal relationship, it is interesting to note that the AG believes that in the present situation the “continuity of work” of the applicant (who had worked for seven years for TP) does not make any difference: “previous working relationships are unrelated to applying for a job and succeeding in concluding a contract” (paras 95-97).

Conversely, the AG opinion provided a far-reaching definition of what may constitutes “work”. She argues that the ‘Equality Framework’ Directive aims at protecting access to work from discrimination “in all the different forms in which work can be offered” (para 61). For her, work refers both to the activity and the result of that activity and the way in which someone approaches the same occupational task vary widely:

A person can earn his or her living by working for only one or for multiple ‘employers’; for longer or shorter periods of time; part-time or only seasonally; at one place, or at different places; using his or her own tools or somebody else’s. Likewise, work can be agreed on the –basis of time (for example, 20 hours per month), or on the basis of the tasks to be performed (for example, painting six walls white) (para 64).

Therefore, she adds, different legal frameworks should not be relevant for the application of the directive as long as the person engages in “personal work” (para 66). She further adds that whether a potential employer ‘buys’ the work or the ‘goods and services’ provided by a person, a company’s refusal to conclude a contract because of a discrimination ground unrelated to the worker’s capacity to perform the work prevents their access to that particular job and, therefore, limits their access to work (paras 79-80):

There is no problem accepting that such discrimination should not be allowed if [the person providing the work was] seeking traditional employment. Why should the same not apply in all other situations where [that person] was offering [his/her] work on the basis of contracts for goods or services concluded with [them] as a person, or on the basis of contracts for goods or services concluded with [their] company, but promising [their] personal work? (para 81)

Finally, she highlights that exempting the personal provision of goods and services from the purview of the ‘Equality Framework’ Directive could create a loophole that would allow companies or individuals to bypass anti-discrimination laws by opting to purchase goods or services instead of employing a service provider (para 85). This would be contrary to the useful effect of that directive, she finds.

Both the Court and the AG come to the conclusion that the ability to enter into a contract for specific work may be an essential factor for someone like the applicant to effectively pursue their professional activities and that, thus, such ability may fall under the scope of ‘conditions for access’ to self-employment (present case para 50, AG opinion paras 80, 88).


The Court and AG proceed to examine whether Article 3 (1)(c) applying the protection conferred by the ‘Equality Framework’ Directive to “employment and working conditions, including dismissals” applies to the present case.

That provision does not explicitly mention ‘self-employment’ but the AG and the Court’s findings described above on Article 3 (1)(a) apply here as well (present case paras 53,54, AG opinion para 101): since the ‘Equality Framework’ Directive was adopted to remove all discriminatory obstacles, its terms must be construed broadly:

It follows that the protection conferred by [the directive] cannot depend on the formal categorisation of an employment relationship under national law or on the choice made at the time of the appointment of the person concerned between one type of contract and another (present case para 55).

While the Court recognises that the concept of ‘dismissal’ is usually applied to the termination of an employment contract it agrees with the AG that Article 3 (1)(c) also covers the unilateral termination of any activity covered by Article (1)(a) of that directive (present case paras 60-62, AG opinion para 102).

Exceptions on grounds of public security, public order, prevention of criminal offences and protection of the health and rights and freedoms of others

Article 2 (5) of the ‘Equality Framework’ Directive lays down exceptions to the application of the directive which “must be interpreted strictly” (present case para 71, AG opinion para 105): where national measures are necessary for public security, for the maintenance of public order and the prevention of criminal offences, for the protection of health and for the protection of the rights and freedoms of others.

Furthermore, while the Directive mandates that discrimination on the grounds of sexual orientation in employment is prohibited, Poland's transposition of the directive only extends this prohibition to sex, race, ethnic origin, and nationality with respect to freedom of contract.

The AG and the Court proceed to balance the rights enshrined in the ‘Equality Framework’ Directive with the freedom of contract granted by the Polish domestic law and Article 16 of the Charter of Fundamental Rights of the European Union (‘Freedom to conduct a business’). An analysis of the aims of said Polish law (protection against discrimination) coupled with the fact that freedom of contract is not an absolute right leads the Court and AG to the conclusion that freedom of contract has not been disproportionately limited by the ‘Equality Framework’ Directive.


This landmark ruling will likely have a significant impact on EU equality and labour law alike.

Firstly, all grounds of discrimination currently recognised by secondary EU law will benefit from this ruling. Indeed, the scope of application of all EU Equality Directives that have similar wording to that of the ‘Equality Framework’ Directive will be impacted (Countouris, Freedland and Stefano) by the Court’s definitions of ‘conditions for access to employment, to self-employment or to occupation’, ‘self-employment’ and ‘occupational activity’.

With regards to LGBTIQ+ rights specifically, while the ruling will have a clear-cut impact on bisexual and homosexual individuals, it is not clear what its impact could be as regards other members of the LGBTIQ+ community. Indeed, existing EU gender equality legal framework anchors transgender and intersex equality within the binary concept of sex, likely excluding non-binary persons (European Commission, Directorate-General for Justice and Consumers, Brink, M., Timmer, A., Dunne, P., et al., pp. 53-54). Furthermore, existing provisions on gender identity have covered only individuals who intend to undergo or have undergone gender reassignment surgery (European Union Agency for Fundamental Rights, p.46).

It also remains to be seen whether categories of individuals that are protected by Article 21 of the Charter of Fundamental Rights of the European Union but not by the scope of the directive (discriminations based on social origin, genetic features, language, political or any other opinion, property, birth) can be considered to be afforded the same protection.

Secondly, with regards to labour law, the focus on the person doing the work and the ‘personal’ work they provide instead of the contract the employer has negotiated with them will hopefully be further analysed by the AG and the Court in future rulings.

Furthermore, as noted by the Court, the protection afforded by the directive “extends to the professional relationship concerned in its entirety”. While the present case concerned the conditions for pursuing and terminating an activity, the author believes it safe to assume it affords protection to workers also while performing their ‘occupational activities’ under the scope of the directive. For instance, other sections of Article 3 (1), may afford protection against discriminatory retaliation with regards to promotions, vocational guidance and training and pay.

Sunday 5 March 2023

Just Say No? The new ‘Stormont Brake’ in the Windsor Framework (part 2 of the analysis of the framework)



Professor Steve Peers, University of Essex


Photo credit: Dom0803, via Wikimedia Commons 




The first part of this analysis looked as a whole at the new Windsor Framework, which amends the controversial Northern Ireland protocol and includes a number of other legal texts. The second part of the analysis, set out in this blog, looks in more detail at a specific issue: the ‘Stormont Brake’ on amendments or replacements to EU legislation within the scope of the protocol. It first explains the text, and then discusses issues of interpretation. For reasons of (relative) brevity, I’ve left out interesting issues of comparisons with similar brakes in other EU treaties (the EEA), or within EU law itself, but may come back to these issues another time. It's been argued that the Brake is an 'ornament' that may never be used in practice (see David Allen Green here), and this may well prove to be the case. However, in my view it's still useful to examine the text of the Brake to see how it would work if it is ever used.   


The Stormont Brake: text


A key point overlooked in some of the initial discussion of the new Stormont brake is that it builds upon an existing provision in the current Northern Ireland protocol. Article 13(3) of the Protocol already states that in principle, Northern Ireland is bound by amended or replaced versions of the EU law which applies to it already. Article 13(4) then provides for a specific rule as regards new legislation within the scope of the protocol, in effect a different form of brake (I’ll call this the ‘Westminster Brake’ for simplicity’s sake, in particular because it doesn’t explicitly involve Northern Ireland, although the UK could unilaterally decide to take account of Northern Irish views). The new Stormont Brake, concerning amended or replaced EU legislation, fits in between them, as a new Article 13(3a), and cross-refers to both. So it’s better to read all three provisions together (I’ve underlined the new text which the Windsor Framework would add).  

3. Notwithstanding Article 6(1) of the Withdrawal Agreement, and unless otherwise provided, where this Protocol makes reference to a Union act, that reference shall be read as referring to that Union act as amended or replaced.

3a. By derogation from paragraph 3, and subject to the fourth subparagraph of this paragraph, a Union act covered by this paragraph that has been amended or replaced by a specific Union act (hereinafter: “specific Union act”) shall not apply as amended or replaced by the specific Union act as from two weeks after the day on which the United Kingdom has notified the Union in writing through the Joint Committee that the procedure set out in the unilateral declaration on involvement of the institutions of the 1998 Agreement made by the United Kingdom, as annexed as Annex I to Joint Committee Decision [XX]/2023, has been followed.

Such notification shall be made within two months of the publication of the specific Union act and shall include a detailed explanation of the United Kingdom’s assessment as regards the conditions referred to in the third subparagraph of this paragraph, as well as of the procedural steps taken within the United Kingdom prior to the notification. If the Union considers that the United Kingdom’s explanation is insufficient as regards the circumstances referred to in the third subparagraph of this paragraph, it may request further explanation within two weeks as of the date of notification and the United Kingdom shall provide that further explanation within two weeks as of the date of the request. In that case the Union act covered by this paragraph shall not apply as amended or replaced by the specific Union act as from the third day after the day on which the United Kingdom has provided that further explanation.

The United Kingdom shall make the notification referred to in the first subparagraph of this paragraph only where:

(a) the content or scope of the Union act as amended or replaced by the specific Union act significantly differs, in whole or in part, from the content or scope of the Union act as applicable before being amended or replaced; and

(b) the application in Northern Ireland of the Union act as amended or replaced by the specific Union act, or of the relevant part thereof as the case may be, would have a significant impact specific to everyday life of communities in Northern Ireland in a way that is liable to persist.

Where the conditions set out in points (a) and (b) are met in relation only to a part of the Union act as amended or replaced by the specific Union act, the notification shall be made only in respect of that part, provided that the latter is severable from the other parts of the Union act as amended or replaced by the specific Union act. If the latter is not severable, the notification shall be made in respect of the smallest severable element of the Union act as amended or replaced by the specific Union act containing the part in question.

Where the notification is made in respect of a part of the Union act as amended or replaced by the specific Union act, in accordance with the second sentence of the previous subparagraph, the Union act shall not apply as amended or replaced by the specific Union act only in respect of that part.

Where the notification referred to in the first subparagraph of this paragraph has been made, paragraph 4 shall apply with regard to the Union act as amended or replaced by the specific Union act; in case the Union act as amended or replaced by the specific Union act is added to this Protocol, this shall be in lieu of the Union act before being amended or replaced.

This paragraph covers Union acts referred to in the first indent of heading 1 and headings 7 to 47 of Annex 2 to this Protocol, and the third subparagraph of Article 5(1) thereof.

4. Where the Union adopts a new act that falls within the scope of this Protocol, but which neither amends nor replaces a Union act listed in the Annexes to this Protocol, the Union shall inform the United Kingdom of the adoption of that act in the Joint Committee. Upon the request of the Union or the United Kingdom, the Joint Committee shall hold an exchange of views on the implications of the newly adopted act for the proper functioning of this Protocol, within 6 weeks after the request.

As soon as reasonably practical after the Union has informed the United Kingdom in the Joint Committee, the Joint Committee shall either:

(a) adopt a decision adding the newly adopted act to the relevant Annex to this Protocol; or

(b) where an agreement on adding the newly adopted act to the relevant Annex to this Protocol cannot be reached, examine all further possibilities to maintain the good functioning of this Protocol and take any decision necessary to this effect.

If the Joint Committee has not taken a decision referred to in the second subparagraph within a reasonable time, the Union shall be entitled, after giving notice to the United Kingdom, to take appropriate remedial measures. Such measures shall take effect at the earliest 6 months after the Union informed the United Kingdom in accordance with the first subparagraph, but in no event shall such measures take effect before the date on which the newly adopted act is implemented in the Union.

As noted in the new text, the UK’s unilateral declaration on the application of the Stormont Brake is attached as an Annex to the new draft Joint Committee decision, which will (among other things) amend the protocol to add the new Article 13(3a). That declaration is as follows:

1. The United Kingdom will adopt the following procedure to operate the emergency brake mechanism in Article 13(3a) of the Windsor Framework. This mechanism will apply in the unique circumstances of this Declaration and is without prejudice to the status of cross-community voting and safeguards in the 1998 Agreement, which apply solely and exclusively to devolved matters.

a. The mechanism will operate solely and exclusively in the event that after the date of this declaration, the Northern Ireland Executive has been restored and become operational, including with a First Minister and deputy First Minister in post, and the Northern Ireland Assembly has been in regular session. Thereafter, Members of the Legislative Assembly (‘MLAs’) wishing to operate the mechanism must be individually and collectively seeking in good faith to fully operate the institutions, including through the nomination of Ministers and support for the normal operation of the Assembly.

b. The minimum threshold for the mechanism will operate on the same basis as the separate ‘Petition of Concern’ process within the 1998 Agreement, as updated through the New Decade, New Approach Agreement in 2020. This means 30 MLAs from at least two parties (and excluding the Speaker and Deputy Speakers) will need to notify the UK Government of their wish that the emergency brake mechanism should be applied.

c. When providing notification to the UK Government, MLAs will need to demonstrate, in a detailed and publicly available written explanation:

i. that they have met the same requirements as those set out in Annex B of Part 2 of the New Decade, New Approach Agreement, namely that the notification is only being made in the most exceptional circumstances and as a last resort, having used every other available mechanism;

ii. that the conditions set out in the third subparagraph of Article 13(3a) of the Windsor Framework are met; and

iii. that MLAs have sought prior substantive discussion with the UK Government and within the Northern Ireland Executive to examine all possibilities in relation to the Union act; taken steps to consult businesses, other traders and civic society affected by the relevant Union act; and made all reasonable use of applicable consultation processes provided by the European Union for new Union acts relevant to Northern Ireland.

2. If it accepts that the conditions in paragraph 1(a) and (b) have been met and that the explanation provided under paragraph 1(c) is satisfactory, the United Kingdom will notify the Union in accordance with the first subparagraph of Article 13(3a) of the Windsor Framework.

3. The United Kingdom, following a notification by MLAs, commits to informing the Union without delay.

4. The United Kingdom, following a notification to the Union that the emergency brake has been triggered, commits to intensive consultations in the Joint Committee on the relevant Union act as provided for by Article 13(4) of the Windsor Framework

In addition, there are further non-binding measures relating to the Stormont Brake: a Joint Committee recommendation and a Joint Declaration on what happens if the UK pulls the Stormont brake and arbitrators rule against it. The first measure states that:

In case the arbitration panel has ruled, in accordance with Article 175 of the Withdrawal Agreement, that the United Kingdom has failed to comply with the third sub-paragraph of Article 13(3a) of the Protocol, the Union and the United Kingdom agree no later than 30 days after such notification that in order to comply with the arbitration panel ruling, and as the case may be, to the extent set out therein, the Union act applies as amended or replaced by the specific Union act, as defined in Article 13(3a) of the Protocol, as from the first day of the second month following the notification of the arbitration panel ruling to the Union and the United Kingdom.

The Joint Declaration states that:

The Union and the United Kingdom recognise that for a notification under Article 13(3a) of the Windsor Framework to be made in good faith in accordance with Article 5 of the Agreement on the withdrawal of the United Kingdom of Great Britain and Northern Ireland from the European Union and the European Atomic Energy Community (‘the Withdrawal Agreement’), it needs to be made under each of the conditions set out in paragraph 1 of the Unilateral Declaration by the United Kingdom on involvement of the institutions of the 1998 Agreement, as annexed to [the draft Joint Committee Decision].

In case the arbitration panel has ruled, in accordance with Article 175 of the Withdrawal Agreement, that the United Kingdom has failed to comply with Article 5 of the Withdrawal Agreement in relation to a notification under Article 13(3a) of the Windsor Framework, swift compliance with the ruling of the arbitration panel should be achieved, as set out in [the draft Joitn Recommendation].

To explain the context of these two soft law measures, which obviously assume that the dispute settlement process of the withdrawal agreement might be used if the ‘Stormont Brake’ were pulled, it’s necessary to summarise that process. (For a full annotation of the dispute settlement rules in that treaty, see my blog post here; the issue is also discussed in my working paper on the withdrawal agreement, and the Yearbook of European Law article based on that working paper).

Article 175 of the withdrawal agreement, referred to in the ‘soft law’ measures, simply requires the parties to comply with a ruling of the arbitrators, once a ruling has been handed down. (To date, the withdrawal agreement dispute settlement rules have not been used in practice). But it does not set a precise time period for compliance; rather, Article 176 of the agreement goes on to set out a process to determine what a reasonable period for compliance is, potentially asking the arbitrators to rule again on that point. (This is based on the rules on WTO dispute settlement, where arbitrators are frequently asked to rule on this issue).

But when it comes to arbitrators’ rulings on the use of the Stormont Brake, the parties have agreed a fast track: if the EU successfully challenges the UK’s use of the brake (and/or successfully argues a breach of the ‘good faith’ rule in the withdrawal agreement as regards use of the brake), the EU law in question will apply from a specific date, which probably falls well before the date that would apply if the usual process to set a time for compliance set out in Article 176 ran its course (up to 30 days for the losing party to suggest a date for compliance, 40 further days for the winning party to contest it, and 40 or 60 further days for arbitrators to rule on the point). The agreement to give effect to the arbitration ruling quickly will also, if applied in practice, avoid the messy consequences of failing to apply an arbitrators’ ruling, as set out in the agreement – namely possible eventual fines or trade sanctions.

But note that these soft law parts of the Windsor Framework do not actually amend the withdrawal agreement. So the UK would not be legally obliged to give effect to the EU law in question within 30 days, if it lost a case. Legally, it could still choose to drag the process out longer or even refuse to comply and face the risk of penalties, as set out in the agreement.

“Why didn’t the UK and EU just agree to amend Article 176 of the withdrawal agreement?” I hear you ask. Well, one or both of the parties may have objected to that idea for political reasons, but in any event there’s a simple legal reason: Joint Committee decisions can amend parts of the withdrawal agreement under certain conditions, but they cannot be used to amend Part Six of the agreement; and the dispute settlement rules are in Part Six.

Legal issues of the Stormont Brake

Leaving aside the purely soft law measures discussed just above, the Stormont Brake raises a number of legal issues.

Scope of the Stormont Brake

First of all, issues arise as regards the scope of the brake. As noted already, the new rules concern objections to amended or replaced EU laws, and will sit in parallel – but partly overlap – with the provisions on objections to new EU laws, ie the Westminster Brake. It is necessary to make this distinction because the grounds and process of applying the Stormont Brake are mostly different to the grounds and process for applying the objection to new EU laws in the Westminster Brake.

The Westminster Brake allows the UK to object to the application of new EU laws for any reason at all – there is no need to show a ‘significant impact specific to everyday life’ of Northern Irish communities which is liable to persist, and no link to the criteria set out in the UK’s unilateral declaration on the Stormont Brake which is annexed to the protocol, or any other criteria. What the two Brakes have in common is what happens after the use of each Brake (Article 13(4) of the protocol). First, the parties must examine the possibilities to keep the protocol functioning by other measures, and may ‘take any decision necessary’ to this end. (In practice, the Westminster Brake has not been pulled to date, and the Joint Committee has amended the protocol to add some new EU laws).  

After a ‘reasonable time’ (not further defined), if there is no such Joint Committee decision, the EU can adopt ‘appropriate remedial measures’ (not further defined), which apply six months later at the earliest, but no sooner than the new EU laws take effect.

So it will be necessary, for at least some purposes, to distinguish between the scope of the two Brakes, ie what is ‘new’ on the one hand, and what is ‘amended or replaced’ on the other. Imagine, for instance, an EU law which is largely new but which amends a couple of provisions of existing legislation to update cross-references to include the new law, or perhaps to make some modest amendments to existing law such as adding tasks relating to the new law to the role of an EU agency. (This is a common EU legislative technique). Is it new, so the Westminster Brake applies, and the UK can reject it for any reason at all? Or is it an amendment, so the Stormont Brake applies, with more stringent criteria to reject it? 

Another issue of scope is tucked away in the final sub-paragraph of the new Stormont Brake clause. The Stormont Brake only applies to parts of the Protocol (the Westminster Brake has no such limitation). It applies to the first indent of heading 1 and headings 7 to 47 of Annex 2, and to the third subparagraph of Article 5(1) of the protocol. Decoding this legalese, Annex 2 sets out EU laws on customs and regulation of goods which apply to Northern Ireland, and the third subparagraph of Article 5(1) refers to EU law on reliefs from customs duty for personal property. But the Stormont Protocol only applies to some of Annex 2: the provisions on the EU customs code and general and specific rules regulating goods, but not to other customs laws listed in heading 1, or headings 2 to 6 of the Annex (fraud against the EU, trade statistics, and various laws on international trade in goods). Nor does the Stormont Protocol apply to laws within the scope of Annex 1 (equality law), Annex 3 (VAT and excise tax law, although note the Windsor Framework amends Annex 3 to provide some exceptions for Northern Ireland), Annex 4 (electricity), or Annex 5 (State aid). Articles 5, 7 and 13 of the Protocol also make further references to EU law in the main text (this includes references to EU Treaty articles).    

So, while the Stormont Brake will apply to most of the EU laws referred to in the Protocol, it does not apply to all of them; which means that there is no process to object to amending or replacing those EU laws outside the scope of the Brake. For instance, the Stormont Brake would not apply to the current proposals to amend EU equality law as regards the power of equality bodies – although it might be argued, as discussed above, that the Westminster Brake applies to those proposals, on the grounds that these are ‘new’ measures, only making minor changes to the text of existing EU legislation and mostly creating free-standing legislative rules.

Substance and procedure of the Stormont Brake

Procedurally, to use the Stormont Brake, the UK must comply with time limits and information requirements: a notification within two months of publication of the EU law; an explanation regarding the substantive conditions; the previous ‘procedural steps’ taken by the UK (not further defined); and the time limits for the EU to ask for more explanations, and for the UK to provide them. These points should be straightforward, other than the definition of ‘procedural steps’ the UK has taken; this may beg the question as to how binding the UK side of the Stormont Brake process is, at least as between the EU and UK (more on that below).

As for the substance, at least two legal issues arise: the interpretation of the conditions referred to in the new Article 13(3a), namely a) a ‘significant’ difference, ‘in whole or in part’, from the ‘content or scope’ of the EU act being amended or replaced; and b) the requirement that that the whole or part of the amended or replaced act ‘would have a significant impact specific to everyday life of communities in Northern Ireland in a way that is liable to persist’. Both requirements must be met for the Stormont Brake to apply (‘and’). Whether a difference is ‘significant’ and whether an impact on everyday life is ‘significant’ could both, within reason, be legitimately the subject of differing views, although at least some cases should be obvious: a mere codification of existing law (ie, a replacement without any amendments to that existing law) could not seriously be regarded as a ‘significant difference’ from the existing law, and so the codification process could not be a valid opportunity to exercise objections to that existing law via means of the Stormont Brake process (no matter how sincere or well-founded those objections might be). Nor does it necessarily follow that a significant difference has a significant impact (or vice versa); both criteria must be met independently.  

The application of the second criterion entails some degree of future prediction (‘would have’ and ‘liable to persist’), but it should not require that evidence to that end must be supplied by a Terminator or the passengers in a de Lorean coming back from the future. It should be sufficient that there are cogent and plausible reasons, backed up by some credible indications, about what the impact of the law might be. (It might have been better if the Stormont Brake had included a review clause, including an assessment of the impact of the amended or replaced law in the EU in practice – which may have some bearing on considering whether the concerns in Northern Ireland have turned out to be well-founded).

An interesting question is the definition of ‘communities’ here. Is there a numerical threshold of the number of communities, or the number of people, who have to be affected? The most obvious question is whether an impact on the unionist community (or some of it, perhaps) is sufficient to trigger the Brake; but there have also been questions about the effect of the protocol on the supply of kosher food to Northern Ireland’s Jewish community. That specific issue may have been dealt with by the Windsor Framework provisions which simplify the movement of food from Great Britain to Northern Ireland; but the underlying question about the threshold remains.  

The unilateral declaration

This brings us to what exactly to make of the cross-reference to the UK’s unilateral declaration on the Stormont Brake, referred to in Article 13(3a) and annexed to the Joint Committee decision. First of all, the cross-reference in the main text of the decision, and the status of the unilateral declaration as an annex to it, must mean that the UK cannot unilaterally alter it. Secondly, the procedural obligations include an express requirement that the UK notify that the procedure in the declaration has been followed; the requirement for an explanation of ‘procedural steps’ may be a reference to the declaration too. This approach to the unilateral declaration means that it is a hybrid of soft and hard law, with the consequence that it is neither fully unilateral nor fully a declaration – it has at least some hard law impact, and if the UK were to change it or withdraw it unilaterally, attempts to use the Stormont Brake afterwards would be invalid, due to the entrenchment of the current version of the declaration within the Joint Committee decision.

Secondly, how much impact does the unilateral declaration have between the parties – as distinct from within the UK? At the very least, a notification of use of the Stormont Brake must include a statement that the procedure in the declaration has been followed; and it is arguable that there is also a requirement for a ‘detailed explanation’ of the ‘procedural steps’ taken by the UK as regards the declaration. So the use of the Stormont Brake is invalid if the UK fails to state that the procedure in the declaration had been followed, and (arguably) if the UK fails to supply a detailed explanation of the procedural steps taken in that process. It should also be invalid if the UK falsely states that the procedure in the declaration has been followed, as this would be a breach of the ‘good faith’ requirement in the withdrawal agreement (see the soft law on that point, discussed above).

But is the substance of the declaration a matter for dispute between the parties? The wording of Article 13(3a) suggests not. The references to ‘procedure’ and ‘procedural steps’ suggest that the UK does not have to justify anything related to the substance of the declaration. This is bolstered by the requirements for the UK to explain the substance of its objections to the significant change/significant impact criteria in Article 13(3a) – and to explain those objections further if the EU requests it – which is not matched by requirements to explain the how the substantive requirements in the declaration are satisfied.

But this is not the end of the story, for the declaration may be relevant within the UK, especially if it is implemented in some way into UK primary or secondary legislation. As recent (and earlier) judgments have shown, litigation about the protocol itself, Brexit and Northern Ireland, or the status of Northern Ireland in general is not uncommon. So disputes about application of the Stormont Brake may reach the courts, within the UK at least.  

The UK legislation giving effect to the declaration will be relevant in that respect, and we don’t know yet what that legislation will say. But we can, for now, identify legal issues in the text of the declaration. First of all, whether the Assembly is in regular session, the Northern Ireland executive is operational, and there is a First Minister and Deputy First Minister should be simple questions of fact. But arguably there could be room for dispute whether MLAs who want to trigger the Stormont Brake are ‘individually and collectively seeking in good faith’ to operate the Northern Ireland institutions? What if it might be argued that some of those seeking to trigger the Brake are acting in good faith in that sense, but others are not? It cannot seriously be argued that any attempt to use the Brake must necessarily be regarded as an act of bad faith, for the whole Brake would then disappear in a puff of logic. But equally the existence of the good faith test must mean something – otherwise why not refer only to the setting up of the institutions?

Next, there might be some questions about how the voting rules work (on the current ‘petition of concern’ process, as amended, in the context of the recent Supreme Court judgment on the protocol, see the analysis of Anurag Deb on this blog).

As for issues of substance, the objectors must show, in a detailed explanation, that a) their notification is ‘most exceptional’, ‘a last resort’, and they have ‘used every other available mechanism’; b) the conditions in Article 13(3a) of the protocol are met; and c) they have consulted within UK, EU, and Northern Ireland processes, as well as with business and civic society.

The UK government ‘will’ use the Brake if it ‘accepts’ that the conditions are met and the substantive explanation is satisfactory; this suggests a degree of obligation but also some degree of independent assessment of the objectors’ position. But the wording does not suggest that the UK government decision to use (or not use) the Brake, having considered the objections made, is wholly discretionary. It follows that the UK’s decision to trigger (or not trigger) the Brake might be challenged by the objectors (if the Brake is not pulled), or by those opposed to using the Brake (if it is pulled). The Brexity opponents of parliamentary control of the executive’s international relations prerogatives have manoeuvered themselves into a position where they have agreed to at least consider the views of a minority in a devolved legislature as to what the UK government should do when exercising that prerogative; and the courts might have a say too. If the opponents of the protocol really want to maintain their opposition to it, they could try objecting to every amended or replaced EU law on the grounds that it inherently has a significant impact on those in Northern Ireland because of the iniquitous nature of the protocol (in their view) – although remember that objections can only be made if the EU law also is a significant change from the status quo.

This discussion shows the importance of the question discussed above – whether the criteria in the unilateral declaration might be invoked by the EU too. Can the EU argue that the objectors were not acting in good faith, that they did not consult widely enough, and that the objection is not ‘most exceptional’ or ‘a last resort’, and the objectors have not ‘used every other available mechanism’? (Interestingly, a ‘last resort’ test is one of the conditions for the EU to use ‘enhanced cooperation’, ie to adopt some EU law with only some Member States participating) If the UK’s argument, when pulling the Brake, about the substantive requirements of Article 13(3a) is different from the arguments of the objectors, can the objectors and/or the EU challenge that? If there is a judicial review pending against the UK decision to pull the Brake, what happens to the EU/UK discussions (or dispute settlement) in the meantime?


It’s sufficient to note briefly that there is a ‘severability’ requirement in the Stormont Brake. If the conditions for using the Brake only relate to part of the amended or replaced EU acts, and the offending part of that act is severable from the rest of it, then the Brake should only be pulled as regards that severable part of the EU act. Whether severability is possible, and whether the objection in fact relates to only some rather than all of the act, might well be disputed.

EU retaliation

As noted already, the EU can take ‘appropriate remedial measures’ if either the Stormont Brake or the Westminster Brake is pulled, under Article 13(4) of the protocol, on the same conditions. Whether the measures are ‘appropriate’ might be contested; it is not clear whether appropriateness is an objective standard, or a subjective assessment by the EU, but in any event proportionality is a general principle of EU law, and the word ‘remedial’ suggests that the measures must be limited to remedying the effect of the use of the Brake. What is a ‘reasonable time’ before the EU can adopt the acts might also be contested.

Dispute settlement

Can the UK’s use of the Stormont Brake be subject to dispute settlement under the withdrawal agreement – and could the EU or UK courts get involved? On the first point, the soft law discussed above suggests that the parties anticipate the dispute settlement process might apply. In fact, the preamble to the Joint Committee Decision also refers to the prospect. This is surely correct, because there is no exclusion from the scope of the dispute settlement process (the Theresa May version of the protocol had some exclusions from dispute settlement, but the Boris Johnson version dropped them, because it dropped the parts of the protocol which they related to). Nevertheless, the question could arise whether the arbitrators could look at issues purely related to the UK unilateral declaration, as discussed above.

The dispute settlement process could also be relevant to a UK challenge to EU retaliation against the UK’s use of either Brake, which could raise the proportionality issues discussed above, as well as questions of the ‘reasonable time’ to wait to respond.

It should be noted that it’s the UK government that would be in control of the dispute settlement process – not those who objected to the amended or replaced EU law becoming applicable in Northern Ireland (or, for that matter, those in Northern Ireland who agreed with the new law). The UK could, however, choose to consult with them informally.

As noted above, it is possible that the UK’s use (or non-use) of the Brake could be disputed in the UK courts, in particular as regards interpretation of the unilateral declaration, as implemented in UK law. It’s possible that the UK courts could be asked about Article 13(3a) of the protocol as such too – especially since its substantive criteria are referred to in the unilateral declaration.

Finally, could the UK’s use of the Brake be subject to the jurisdiction of the CJEU? The UK government claims not, because the use of the Brake is not an issue of EU law, and the CJEU can only be involved where there is a question about the interpretation or validity of EU law (the protocol gives the CJEU its usual jurisdiction over EU law as regards parts of the protocol, which does not include Article 13 but does include the laws that the Brake could apply to; and the arbitrators must send the CJEU any questions that arise about EU law as part of the arbitration process).

However, the issues of severability of the EU law, of significant change from existing EU law, and even (arguably) about the significant impact of the EU law are questions of EU law; so the potential role of the CJEU cannot be entirely excluded. In any event, a challenge to the validity of any EU retaliation would fall within the Court’s jurisdiction, as would questions about the interpretation or validity of EU legislation that was the subject of the Brake (leaving aside the Brake’s use as such).

Whether the courts or the arbitrators become involved, there may be questions about the intensity of review, especially of the UK government’s or objectors’ arguments about the impact of EU law. Frankly, both the critics and advocates of EU law sometimes say silly things about it. To come back to the kosher food example discussed above, a fact check suggests that it was hyperbole for the UK government to suggest that the Jewish community in Northern Ireland would be wiped out due to the difficulty obtaining kosher food from Great Britain, given that kosher food was available from non-British sources. On the other hand, the fact check also notes that those other sources of kosher food are more expensive; and to apply the wording of the Brake, in my view it would be perfectly reasonable for the Jewish community to argue that an increase in the price of kosher food would have a significant impact on their everyday life. But what if an amended or replaced EU law had the effect of blocking imports of kosher food from Britain again, and the UK government made its argument using hyperbole instead? Should the EU or the arbitrators overlook the government’s pound shop Godwin’s Law rhetoric, and look at the underlying facts, which nevertheless still make a case for the Brake to be used?   



The tension between powers exercised in the central and sub-central levels of governance is common in federal States and other systems of multi-level governance – including the non-federal UK and the EU, an international organisation with far more powers than most such organisations have. But Northern Ireland is distinct in that it is subject to two overlapping and competing such multi-level legal orders – and its population is furthermore internally divided between groups who feel stronger connections to one legal order or the other. These tensions could more easily be reconciled, at least after the Good Friday Agreement, when the UK was a Member State of the EU; but the Northern Ireland protocol, as constantly tested and amended, is an attempt to reconcile them after Brexit. The protocol is, in effect, one of the most contested ‘middle bits’ of a Venn diagram in human history.

The Stormont Brake tries to address these overlapping multi-level tensions by providing both a mechanism to address disputes between the UK government and (parts of) the Northern Ireland polity – in the form of the unilateral declaration – and a mechanism for the UK to convey those concerns within the UK/EU framework. These mechanisms necessarily have to take account of the parallel legal and political constraints on both sides: the UK is not a federal State, and the executive controls international relations, not the minority in a devolved assembly; while the EU no longer includes the UK as a Member State, and there are political and legal limits on the impact which a non-Member State can have. But equally both parties are aware of the importance of ensuring the stability and prosperity of Northern Ireland – which cannot be secured by satisfying the demands of only one community or the other, yet it may be difficult or impossible to reconcile those demands when they conflict. The drafters of the Stormont Brake have made some efforts to address these tensions; time will tell how effective their efforts are.

See also: flow chart on the Stormont Brake, by Simon Usherwood 

*This blog post was amended on 6 March 2023, to add a link to the flow chart, the point that it may only be used as an 'ornament', and a clarification of the CJEU's jurisdiction.