A CONTEXTUAL ANALYSIS OF ARTICLE 16 OF THE IRELAND–NORTHERN IRELAND PROTOCOL

Abstract Article 16 of the Ireland–Northern Ireland Protocol annexed to the EU–UK Withdrawal Agreement is an escape clause which allows the parties to deviate from their obligations under certain conditions. This article maps out the main features of the safeguards provision in the Protocol in light of international trade law and international relations literature on treaty design. It provides a detailed examination of the safeguards provision in the Protocol and highlights the key design flaws associated with this regime as well as some potential solutions to such flaws.


I. INTRODUCTION
On 22 January 2021 a European Commission Implementing Regulation Proposal governing export restrictions on COVID-19 vaccines was leaked to the press. 1 The proposal stated that such restrictions should also apply to trade between the European Union (EU) and Northern Ireland and that these would be justified as 'a safeguard measure pursuant to Article 16 of that Protocol in order to avert serious societal difficulties due to a lack of supply threatening to disturb the orderly implementation of the vaccination campaigns in the Member States'. 2 The leak led to an uproar as such restrictions would have inevitably led to North-South border checks within the island of Ireland 3 -an outcome the Ireland-Northern Ireland Protocol (Protocol) annexed to the EU-UK Withdrawal Agreement was designed to avoid. 4 The European Commission quickly dismissed the notion that these export restrictions would be applied and the proposal was ultimately never adopted. 5 Nevertheless, this incident brought to the fore the importance of, as well as some of the potential dangers associated with, Article 16 of the Protocol.
The Protocol is a legal instrument whose primary purpose is to establish a trade regime for Northern Ireland that obviates the need for border checks on trade in goods between the North and the South of the island of Ireland. 6 The Protocol is a consequence of the view espoused by both parties during the negotiations of the Withdrawal Agreement that the Good Friday/Belfast Agreement 7 -that is, the international treaty concluded by the Republic of Ireland and the UK that underpins the peace process on the island of Ireland-prohibits the establishment of a hard border within the island of Ireland. 8 This view is expressly recognised in Article 1(3) of the Protocol. 9 Border checks would have been an inevitable consequence of the UK's decision to withdraw from the EU, and in particular, from the EU customs union and internal market. 10 To avoid this, the Protocol requires NI, in contrast to the rest of the UK, to remain subject to EU customs and internal market rules on goods.
Article 16 of the Protocol is an escape clause that allows the parties to deviate from their obligations under the Protocol under certain conditions. It has quickly become the most (in)famous and politically charged provision of the Protocol. Since the COVID-19 vaccine export restrictions debacle, there have been numerous calls to 'trigger' Article 16 of the Protocol from politicians and interest groups opposed to the continued operation of the Protocol. 11 There is a sense that the provision is increasingly being weaponised by politicians who 4 Agreement on the withdrawal of the United Kingdom of Great Britain and Northern Ireland from the European Union and the European Atomic Energy Community [2020] OJ L29/7. 5 K Adler, 'EU Vaccine Export Row: Bloc Backtracks on Controls for NI' (BBC, 30 January 2021) <https://www.bbc.co.uk/news/uk-55865539>. 6  International and Comparative Law Quarterly see it as offering a potential way out of commitments made by the EU and the UK under the Protocol. 12 In July 2021, the UK government released a Command Paper on the Northern Ireland Protocol 13 outlining its conviction that the conditions for the invocation of Protocol safeguards had been met and, since then, there have been regular reports that the UK could be willing to 'trigger' Article 16 of the Protocol. 14 This article seeks to go beyond the politics by carrying out a contextualised legal analysis of this provision, mapping out its main features in light of international trade law and international relations literature on treaty design. The article argues that the wide scope and broadly phrased conditions for the invocation of the Protocol extend a significant margin of discretion to the parties in deciding whether or not to apply safeguard measures. This, in the context of an escape clause, is highly problematic as past practice suggests that such loosely termed clauses can create a temptation for the parties to exercise their discretion abusively. This concern is further exacerbated by the unique nature of the Protocol as a trading arrangementone that must be seen primarily through the lens of the Northern Irish peace process. 15 Derogations from the Protocol, especially where politically motivated, could undermine the existence of the Protocol and the careful balancing act it seeks to achieve.
This article is subdivided into three substantive sections. Part II examines the international relations literature on the function and design of escape clauses included in trade agreements. Part III provides a brief descriptive overview of the trade-related components of the Protocol. More specifically, it highlights some of the unique features that distinguish it from other international legal instruments intended to remove barriers between States, and considers some of the challenges that may arise in relation to escape clauses as a consequence of such features. Part IV then provides a more detailed examination of the safeguards provision in the Protocol and highlights some of the design flaws associated with this regime, as well as some potential solutions to such flaws. G Parker, J Webb and A Bounds, 'UK to Make New Offer on Northern Ireland in Talks with EU' Financial Times (London, 11 February 2022). 12 C Rice and C Murray, 'Northern Ireland Protocol: The UK's updated approach' (UK in a Changing Europe, 4 August 2021) <https://ukandeu.ac.uk/northern-ireland-protocol-uk-updatedapproach/>. 13 HM Government, 'Northern Ireland Protocol: The Way Forward' (21 July 2021) 13-15 <https://www.gov.uk/government/publications/northern-ireland-protocol-next-steps>. 14 P Leahy and B Hutton, 'Irish and EU Prepare for UK to Trigger Article 16 of Protocol' The Irish Times (Dublin, 8 November 2021) <https://www.irishtimes.com/news/politics/irish-and-euprepare-for-uk-to-trigger-article-16-of-protocol-1. 4722094>. 15 B Rosher, '"And Now We're Facing That Reality Too": Brexit, Ontological Security, and Intergenerational Anxiety in the Irish Border Region' (2022) 31 European Security 21.
Article 16 of the Ireland-Northern Ireland Protocol 533

A. Flexibility in International Trade Agreements
When States negotiate international treaties, they inevitably weigh up the benefits of international cooperation against the risks associated with binding themselves to common international disciplines. States may be reluctant to bind themselves to strong commitments where, for example, there is uncertainty regarding the potential outcome of the treaty, the reception of the treaty by domestic constituents or even the behaviour of other States. 16 One way to alleviate concerns relating to potential uncertainty is to incorporate mechanisms that are intended to provide some flexibility for parties with respect to compliance with the treaty. Such flexibilities facilitate both the conclusion of treaties and the negotiation of more ambitious commitments, by providing governments with an insurance policy against potential risks associated with international cooperation. 17 The classic approach to conceptualising flexibilities in international relations theory is through the distinction between the policy space theory and the safety valve theory. 18 The policy space theory posits that flexibilities are intended to allow governments to pursue legitimate policy objectives that are not traderelated. The release valve theory, on the other hand, explains flexibilities as mechanisms that allow governments to alleviate political pressure from domestic industries adversely affected by trade liberalisation.
In his seminal work examining the function of exceptions in international trade law, Krzysztof Pelc adds nuance to the distinction, arguing that flexibilities in trade agreements typically address two types of uncertainties. 19 First, they provide an insurance against exogenous shocks, namely the risk that liberalisation mandated in trade agreements can expose domestic industries to adverse effects resulting from increased competition. 20 Here, the main purpose of an escape clause is to quell opposition from domestic industry to the signing of trade agreements by giving them the comfort of knowing that liberalisation commitments can be derogated from under certain circumstances. Secondly, escape clauses provide a form of political insurance to governments. 21 They are, in effect, a recognition that governments face significant internal pressures in implementing international commitments. 22

534
International and Comparative Law Quarterly groups, flexibilities provide governments with a means to reduce internal political pressure. One type of flexibility that has proved particularly popular in the context of international trade law are so-called 'escape clauses'. These allow parties to temporarily suspend or derogate from obligations under certain conditions. Most escape clauses included in contemporary trade agreements either replicate or take inspiration from those included in the General Agreement on Tariffs and Trade (GATT)-that is, the main multilateral agreement regulating trade in goods. 23 These include, for instance, policy space exceptions such as general exceptions provisions which enable States, under certain conditions, to derogate from trade law obligations where it is shown that derogations are intended to achieve legitimate public interest objections, 24 as well as national security exceptions authorising derogations for security purposes and to the extent that certain factual scenarios occur. 25 Another example is the WTO safeguards regime, which falls neatly within the category of release valve exceptions. Here, the goal is to allow WTO Members to temporarily derogate from international commitments in order to allow domestic industries to adapt to the unintended consequences of trade liberalisation. This then relieves the political pressure on governments, keen to be seen to be responsive to the demands of domestic industries, as well as on the trading system as a whole by legalising non-compliance in exceptional circumstances.

B. Designing Escape Clauses-the Case of the WTO Safeguards Regime
One challenge posed by escape clauses is how to design them in a manner that ensures they fulfil their functions whilst at the same time avoiding abuses (through overuse) by States. The theory of efficient breach puts forward one potential solution. 26 This theory recognises that there are instances where the cost of compliance outweighs those of non-compliance. 27 To reflect this, escape clauses should incorporate mechanisms that impose a cost on deviation, normally by requiring the party deviating from its commitments to provide compensation to those adversely affected by the breach. The cost of deviation should be high enough to dissuade States from frivolous and unnecessary non-compliance, but not so prohibitive that deviation becomes impossible.
There are, however, problems associated with the theory of efficient breach. By turning the decision to derogate into a cost-benefit analysis, it creates an incentive for domestic interest groups who are opposed to liberalisation to 23  make it politically costly for governments not to avail themselves of escape clauses. 28 The higher the political costs associated with compliance are, the more likely governments will be to accept the economic costs resulting from compensation. The decision then becomes a purely political one in nature. In short, efficient breach generates an environment where domestic interest groups can artificially fabricate the internal circumstances that will cause governments to pursue deviation from international commitments.
Another option, which has proved more successful in practice, is the 'contingent flexibility' 29 theory, where deviation is no longer merely a function of internal political imperatives but rather contingent upon the demonstration of 'observable, exogenous hard times'. 30 The requirement that exceptional exogenous events must occur to justifiably deviate from international obligations ensures that, on the one hand, escape clauses are not abused and, on the other hand, domestic constituents are not able to manufacture situations of political necessity which leave governments with little choice but to opt for non-compliance. 31 The WTO safeguard regime is a successful example of escape clauses based on contingent flexibility. It allows WTO Members temporarily to reimpose trade barriers in situations where sudden increases in imports resulting from unforeseen events cause or threaten to create injury to domestic industries. As a starting point, safeguards may only be applied in response to an increase in the quantity of imports into the WTO Member concerned, which can be either absolute or relative to domestic production. 32 But a mere increase in imports is not enough, on its own, to justify the application of safeguards. Article XIX GATT provides that the increase in imports must also be unforeseenthat is, the result of factors which could not have been reasonably foreseeable at the time when trade concessions were made. 33 In other words, increases in imports which are the consequences of trade liberalisation are not covered by the safeguards regime. 34 The increase must also be 'recent, sudden, sharp or significant enough to cause or threaten to cause serious injury'. 35 The suddenness of the increase reinforces the extraordinary and unexpected 28

536
International and Comparative Law Quarterly nature of the circumstances that must be in place to trigger the application of the safeguard. The requirements of exogeneity and unforeseeability are key features of the WTO safeguards system. If safeguards are to be imposed, one must not only demonstrate that the circumstances justifying it could not have been predicted at the time trade liberalisation commitments were made, but also that such extraordinary circumstances have caused serious harm to domestic industries. These stringent requirements remove the ability of WTO Members to manufacture circumstances to justify safeguards. By making the imposition of safeguards contingent upon a demonstration that external circumstances beyond the control of a WTO Member have occurred, the WTO safeguards system shields governments from political pressure from domestic interest groups to reimpose protectionist measures. 36 WTO law also subjects the imposition of safeguards to a number of procedural requirements. First, there is an obligation on all WTO Member States to ensure that a safeguard measure is applied on the basis of an investigation carried out at the domestic level by a competent authority. 37 Secondly, WTO Members are required to notify matters relating to safeguards measures to the WTO Committee on Safeguards-whether it be a decision to initiate an investigation, or a decision to apply or extend safeguards measures. 38 The notification should include all pertinent information concerning the safeguard measures and provide an adequate opportunity for the exchange of information and discussions relating to compensation of other affected WTO Members.
There is therefore an obligation to enter into consultations to agree on compensation due to affected Members as a consequence of the measures, and if no agreement is reached, the affected WTO Members can opt to withdraw substantially equivalent concessions or other obligations. 39 However, not all WTO Members are entitled to compensation: only those who have substantial interests in the products targeted by the safeguards. 40 Moreover, compensation is not due where the safeguard measures are applied for less than three years and the measure is taken in response to an absolute increase in imports. 41 In short, whilst a compensatory mechanism exists, WTO law significantly limits its potential scope. This reflects the general sense that WTO Members should not be punished for taking emergency measures which result from exceptional circumstances or factors that are beyond their control. 42  safeguards. 43 The WTO safeguards regime is an example of an escape clause whose operation is not reliant on the imposition of an economic cost on those that opt to derogate from obligations. Instead, it is one based on contingencybased flexibility where the legitimacy of the derogation is derived from the demonstration of the existence of extraordinary and unexpected events that have created the need for a deviation from international trade liberalisation commitments.

III. CONTEXTUALISING THE PROTOCOL IN THE WORLD OF TRADE AGREEMENTS
The Protocol achieves the goal of avoiding a hard border between Northern Ireland and Ireland by removing barriers to trade in goods between Northern Ireland and the EU. In this sense, although the Protocol is part of a wider international treaty between the EU and the UK governing the terms of the UK's exit from the EU, the Protocol can be viewed as a trade agreement in that it incorporates commitments between two separate customs territories to remove restrictions on goods traded between each other. It is, however, a peculiar trade agreement for a number of reasons.
It is not, as is the norm, a trade agreement that lowers trade barriers between two whole and separate customs territories. Rather, it only liberalises trade between a customs territory (the EU) and a component of another customs territory (Northern Ireland). By committing in the Withdrawal Agreement to ensure that Northern Ireland would remain subject to EU customs and internal market laws, the UK accepted that any decision to diverge from such rules would create trade barriers between Northern Ireland and the rest of the UK. The upshot is that while the Protocol ensures unrestricted trade within the island of Ireland, it creates the possibility of a scenario where complying with the Protocol means an increase in barriers to trade within the UK customs territory between Northern Ireland and the rest of the UK.
The Protocol is also different in that it belongs to a very select group of trade agreements that are intended to achieve market integration rather than trade liberalisation. 44 Most trade agreements-whether at WTO, regional or bilateral level-can be placed in the trade liberalisation paradigm in that they focus on the reduction or removal of border measures such as tariffs, but do little in terms of removing regulatory barriers to trade. 45 47 are good examples of this. Although they are generally described as ambitious and comprehensive trade agreements, they do not remove all tariffs, customs procedures and regulatory barriers to trade. By contrast, trade arrangements that pursue market integration, such as the European Union and the European Economic Area, not only remove border measures but also regulatory barriers to trade by mandating mutual recognition of rules across the board. And this is achieved thanks to the establishment of common institutions and rules that not only harmonise rules of parties involved but also enforce those rules. 48 The Protocol, by requiring that Northern Ireland remains part of the EU internal market for goods, is a trade agreement that can be placed in the market integration paradigm.
At a more fundamental level, the Protocol is unique in the world of international trade agreements because of the aim that it pursues. Trade agreements are typically concluded in the hope of reaping the economic benefits that classic economic theory associates with trade liberalisation (increased efficiencies leading to higher levels of productivity and living standards). 49 Trade agreements are also concluded to tie the hands of governments with respect to domestic industry representatives that are opposed to economic liberalisation reforms. 50 This is not to say that trade agreements do not also pursue non-economic objectives. For example, the EU's preferential trade agreements (PTA) with its neighbouring countries are just as much about geo-strategic goals of promoting peace and stability at its borders and expanding its sphere of influence as they are about economic growth. 51 Equally, the recently concluded Comprehensive and Progressive Trans-Pacific Partnership (CPTPP) is known to have been partly motivated by the United States' desire to counterbalance China's growing influence in the Asia-Pacific region. 52 But even in these cases, where economic concerns are perhaps secondary to political and security goals, the benefits of economic integration are seen as key in achieving the primary aims. In other words, creating closer economic links and increasing economic interdependence enhances political ties between countries, and increased economic welfare provides a platform for regional stability and security. 53 The Protocol is different in this respect in that the potential benefits of trade liberalisation were not a significant consideration in its creation, if at all. One of the central aims of the Protocol is to ensure compliance with the Good Friday Agreement by avoiding a hard border within the island of Ireland. 54 Peace and security, rather than economic welfare, are the driving forces behind the Protocol. This raises questions as to whether and how legal mechanisms typically included in trade agreements, such as escape clauses, can work. To be more specific, should parties to a trade agreement that is fundamentally about maintaining peace be given the right to escape, even temporarily, from their obligations? And, if so, how should such escape clauses be operationalised?

A. Architecture of Article 16 of the Protocol
Article 16 of the Protocol on safeguards is the main escape clause included in the Protocol. It allows either party to unilaterally apply safeguard measures where the application of the Protocol 'leads to serious economic, societal or environmental difficulties that are liable to persist, or to diversion of trade'. 55 Irrespective of the condition for invocation that is used, the safeguard measure must be proportionate and limited to what is strictly necessary to achieve its goal. 56 The other party may respond to the safeguard measure by adopting rebalancing measures where the safeguard measure has created an imbalance between rights and obligations under the Protocol. 57 Article 16 of the Protocol is also complemented by Annex 7 of the Protocol, which details the procedural rules that parties must respect when applying safeguards. 58 Although Article 16 of the Protocol refers to safeguard measures, the regime it establishes is very different from the WTO law regime. Firstly, unlike Article XIX GATT, there is no requirement that the circumstances triggering the safeguards should be unforeseen. As discussed, the requirement that an increase in imports be the result of unforeseen circumstances is crucial in ensuring that the WTO and standard PTA safeguards are not abused. Its noninclusion in the Protocol is therefore significant because it means that, potentially, parties are able to apply safeguards in response to outcomes that could not have been seen as unpredictable or exceptional at the time the Protocol was signed. 53

540
International and Comparative Law Quarterly Secondly, whereas Article XIX GATT limits the application of WTO safeguards to a very narrow set of circumstances, Article 16 of the Protocol casts the net more widely. There is no requirement to demonstrate that a particular domestic industry has been seriously injured, nor that the injury was caused by an increase in imports. Instead, safeguards can be envisaged where it is shown that the application of the Protocol is causing or threatens to cause serious difficulties of an economic, societal or environmental nature that are liable to persist, or diversion of trade. By outlining an exhaustive list of grounds that can be invoked to justify the adoption of safeguards, Article 16 of the Protocol combines both elements of the WTO safeguards regime and Article XX GATT on General Exceptions. The provisions recognise that the application of the Protocol may produce certain economic and noneconomic negative externalities and therefore the provisions establish the rights of the parties to derogate from their obligations under the Protocol under certain conditions.
Article 16 of the Protocol is inspired by certain safeguard mechanisms included in other international treaties signed by the EU. Those involved in the negotiations of the Withdrawal Agreement have acknowledged that Article 16 of the Protocol is modelled on Article 112 of the European Economic Area (EEA) Agreement. The EEA governs the EU's trading relationship with the three European Free Trade Association (EFTA) States (Norway, Liechtenstein and Iceland). Under the EEA Agreement, EFTA States are subject to most aspects of EU internal market law, including compliance with the four economic freedoms. However, the main safeguard clause, Article 112 EEA Agreement, permits safeguard measures '[i]f serious economic, societal or environmental difficulties of a sectorial or regional nature liable to persist are arising'. The similarities with Article 16 of the Protocol do not end there. Like Article 16 of the Protocol, Article 112(2) provides that safeguard measures must be 'restricted with regard to their scope and duration to what is strictly necessary in order to remedy the situation' with priority being given to measures that least 'disturb the functioning of this Agreement'. And like Article 16 of the Protocol, rebalancing measures can be adopted by other parties where the safeguard measure creates an imbalance between the rights and obligations under the EEA Agreement. 59 Article 112 was conceived as a compromise solution between, on the one hand, the EFTA States who wished to secure certain permanent derogations from EU internal market law and, on the other hand, the EU's reluctance to allow for permanent carve-outs. 60  regime can perhaps be explained by the fact that, like the EEA legal framework, the Protocol is a trade arrangement that is premised on market integration through regulatory convergence. As discussed in the previous section, the wide set of circumstances included in Article 112 reflects the fact that the EEA goes beyond the liberalisation of trade in goods by also requiring compliance with the four EU economic freedoms and EU internal market legislation. 61 EEA Agreement parties wanted the flexibility to deviate from the EU acquis communautaire under certain circumstances. Similarly, the negotiators of the Protocol may have considered that a traditional WTO-type safeguards system would be too narrow in its scope to reflect the breadth and depth of commitments under the Protocol.
However, Article 112 EEA Agreement has only been used once in the history of the EEA. In practice, it seems that the EFTA States have tended to view Article 112 mostly as a tool to derogate from rules on the free movement of persons. Indeed, to date, the only instance where Article 112 has been triggered relates to Liechtenstein's measures restricting the entry, residence and employment rights of EEA nationals. 62 Further, a number of EFTA States have issued unilateral declarations confirming that the use of safeguard measures would be justified if continued unrestricted migratory flows were to lead to disturbances in the labour market and access to housing. 63 The reasons behind the reluctance of EEA States to invoke Article 112 are not entirely clear. One theory, put forward by Glencross, is that the non-use of this provision reflects the 'absence of direct confrontation within the EEA framework, which itself is a reflection of the asymmetry of economic power whereby the EU can absorb the costs from any suspension far more easily than the smaller EFTA states'. 64 Another possibility, which was posited by Piris, is that 'maybe' 65 the threat of being subject to rebalancing measures has discouraged the use of safeguard measures. Under this view, it is the cost associated with escaping EEA obligations via Article 112 which reduces the attractiveness of this safeguard mechanism.
It is worth mentioning that both Article 112 of the EEA Agreement and Article 16 of the Protocol bear some similarities with the economic safeguard mechanism which was included in Article 226 EEC until 1969 and has since been included in accession treaties concluded by the EU and new Member  International and Comparative Law Quarterly States. 66 These provisions grant the European Commission, on application by a Member State, the power to allow the latter to derogate from EU Treaty obligations if difficulties arise which are serious and liable to persist in any sector of the economy or which could bring about serious deterioration in the economic situation of a given area. These provisions have, in many instances, served a very similar function to the WTO safeguards regime insofar as they have been used temporarily to protect domestic sectors-through the reimposition of tariffs and quotas-that have struggled to adjust to increased imports resulting from liberalisation. 67 However, these accession economic safeguards have a number of important features that distinguish them from the safeguard mechanism under Article 112 EEA Agreement. The first difference is that the scope of the accession safeguards is comparatively narrow as there is only one condition for the invocation of safeguards-the deterioration of an economic situation in a given area. The second difference is that Member States cannot unilaterally decide to apply economic safeguard measures. Rather, a request must be submitted to the European Commission, which will issue an authorisation following an assessment of whether or not the conditions for the imposition of safeguards were met. 68 The third significant difference is that EU Member States cannot respond to the application of safeguard measures by adopting rebalancing measures.
These differences have had a considerable impact in terms of the operation of economic safeguards. In the absence of a rebalancing measures-type mechanism that would impose a cost on the use of economic safeguard measures, these have been invoked on multiple occasions. However, on the whole, their use has been fairly uncontroversial, in large part because of the important 'gate-keeper' role the European Commission plays in the process. Although the European Commission has, historically, benefited from a considerable margin of discretion when deciding whether economic safeguards can be applied, 69 its position as an impartial third party entrusted with the task of verifying that the external circumstances for their invocation are present, ensures that such safeguards are not abused by EU  The first category of grounds for the invocation of Protocol safeguards relates to instances where the Protocol leads to 'serious difficulties of an economic, societal or environmental nature'. As stated above, these conditions are borrowed from Article 112 EEA Agreement. However, the exact meaning of the conditions for the invocation of Article 112 remains unclear. As discussed, what constitutes a 'serious difficulty' of either an 'economic, societal or environmental nature' is neither defined in the text of the EEA Agreement or in EEA case law. As such, EEA practice does not provide any useful insights on the definition and scope of the terms used under Article 16 of the Protocol. 71 The reference to 'economic' difficulties chimes with the WTO safeguards regime in that it recognises the possibility of derogating from the Protocol for economic reasons. The language is borrowed from Article 112 which allows safeguard measures where there are 'serious economic difficulties of a sectorial or regional nature' and is also reminiscent of 'economic safeguard' provisions typically included in the EU Accession Treaties which allow the adoption of protective measures where 'difficulties arise which are serious and liable to persist in any sector of the economy or which could bring about serious deterioration in the economic situation of a given area'. 72 As discussed, these economic safeguard provisions were mostly used to protect specific domestic industries from the increased competition resulting from liberalisation. However, the omission of any reference to the regional or sectorial nature of potential economic difficulties indicates that the scope of this condition under Article 16 of the Protocol will be wider than that of the EU economic safeguards provisions. This is certainly a view that was espoused by the UK in its Command Paper which identified high consumer prices, increased operating costs faced by businesses and disruptions to food and parcel supplies as evidence of difficulties of an economic nature resulting from the application of the Protocol. Under this view, any serious economic harm-whether at the national, regional or sectoral level-caused by the Protocol could potentially fall under the scope of Article 16 of the Protocol. However, as discussed further below, establishing a causal link between the application of the Protocol and the existence of serious economic difficulties may prove easier said than done.
The second category of grounds for invocation relates to non-economic goals, namely the existence of serious social and environmental difficulties. There are other legal mechanisms within the Protocol that can be used to 71

544
International and Comparative Law Quarterly derogate from the trade-related obligations to achieve non-economic goals. Under the Protocol, the lawfulness of the placing of goods in Northern Ireland is subject to TFEU provisions governing the free movement of goods, 73 which have to be interpreted in conformity with EU case law. 74 This is significant because whilst, in accordance with Article 34 TFEU, Member States are precluded from applying non-fiscal barriers to trade (namely quantitative restrictions and measures having an equivalent effect), Member States can apply restrictive measures at the national level if these are justified by reference to non-economic goals. These non-economic goals include those that are specifically listed under Article 36 of the TFEU (eg public morality, public policy or public security, the protection of health and life of human, animals or plants), as well as the so-called mandatory requirements-a non-exhaustive list of public interest goals which have been recognised by the CJEU as valid policy goals to justify derogations from the principle of free movement of goods. Mandatory requirements include a wide array of policy goals such as, for example, environmental protection, 75 the prevention of crime, 76 the protection of working conditions, 77 the protection of socio-cultural characteristics, 78 consumer protection, 79 the protection of human rights, 80 and road safety. 81 Thus, Member States are allowed to adopt national measures that constitute restrictions on the free movement of goods as long as they can justify the derogation by reference to a valid public interest objective and demonstrate the proportionality of the restrictive national measure.
The upshot of all this is that, under the Protocol, one can derogate from EU Treaty provisions on the free movement of goods by reference to a nonexhaustive and ever-evolving list of public interest goals that are recognised under EU case law and that go significantly beyond what is covered by Article 16 of the Protocol. However, Article 16 of the Protocol is limited to two specific policy goals (social and environmental), and there is also a requirement to show that a derogation is intended to address a 'serious difficulty'. Thus, the threshold for the invocation of Article 16 of the Protocol is certainly higher than that applied for free movement of goods derogations under EU law.
All this raises questions regarding the practical usefulness of invoking the non-economic justifications under Article 16 of the Protocol to justify a derogation, given that they only cover a narrow set of non-economic goals and that the threshold for their invocation appears significantly higher. There are two points to bear in mind in this respect. First, that Article 36 TFEU derogations and the mandatory requirements can only be invoked to justify the application of non-fiscal barriers to trade in goods. By contrast, Article 16 of the Protocol permits parties to derogate from any obligation derived from the Protocol, even those that do not relate specifically to trade in goods. Second, there is the fact that national restrictions on the free movement of goods cannot be justified in areas where national rules have been harmonised via EU secondary legislation. 82 Once the rules for an economic sector have been exhaustively harmonised at EU level, Member States are no longer permitted to adopt more restrictive measures in that area. This is where Article 16 of the Protocol also comes into play. It would allow a derogation from the obligation to comply with EU legislation, which has fully harmonised an economic activity where it is shown that compliance has led to serious societal or environmental difficulties.
Recent events suggest that the parties to the Protocol see the grounds of 'difficulties of a societal nature' as potentially encompassing an extremely wide variety of circumstances. The European Commission's leaked proposal for export restrictions on COVID-19 vaccines contended that the restrictions were justified under Article 16 of the Protocol 'in order to avert serious societal difficulties due to a lack of supply threatening to disturb the orderly implementation of the vaccination campaigns in the Member States'. 83 In other words, the European Commission sees the term 'societal difficulties' as encompassing measures that are intended to protect public health goals. The UK has also identified political and community instability as societal factors that could constitute sufficient grounds to apply safeguard measures. 84 In theory, the term 'societal' could cover any non-economic policy objective and may end up serving as a residual category of grounds that can be used by the parties for public interest goals which do not fall under the serious economic or environmental difficulties category.
The 'environmental' component has not yet been mentioned by either party and is less likely to be a relevant concern in the short to medium term given that the environmental protection standards maintained in the EU and the UK remain largely aligned. 85 To the extent that Northern Ireland is tied to EU environmental protection rules, environmental difficulties are only liable to arise from the application of the Protocol in cases where the EU has decided to significantly reduce its standards relative to those applicable in the UK. However, in practice, this situation is unlikely to occur as EU environmental

546
International and Comparative Law Quarterly legislation typically only establishes minimum standards. 86 Member States (and Northern Ireland under the Protocol) are therefore not precluded from applying higher standards than those applicable under EU law. Irrespective of the grounds invoked, a party wishing to apply safeguard measures must demonstrate that these are intended to address 'serious' difficulties that are 'liable to persist'. It is first worth noting in this regard that the Protocol does not contemplate the use of safeguards in instances where a party considers that its continued application may lead to serious difficulties in the future. To apply safeguards, a party should demonstrate the actual existence of serious difficulties that have materialised in practice-meaning that any assessment should be based on objectively verifiable facts rather than speculative projections. That being said, the determination of the existence of difficulties resulting from the Protocol may prove more straightforward in some cases than others.
Environmental difficulties that are due to differing environmental protection standards applicable in the UK and the EU would be fairly easy to spot. Economic difficulties caused by trade disruptions could be based on econometric analyses. As far as difficulties of a societal nature are concerned, in some cases, these may entail political, and therefore subjective, assessments which will be informed by complex social, historical and cultural factors. This will be the case when assessing the impact of the Protocol on community relations in Northern Ireland. The UK government has pointed to the anger felt towards the Protocol by certain elements of the unionist community as a reason for the potential suspension of aspects of the Protocol. A suspension of the Protocol may, however, further exacerbate the political instability by stoking anger from the nationalist community. Any decision in this area would be highly political and one where a large margin of discretion will be allowed.
It must also be determined whether the difficulties resulting from the Protocol are of a serious nature. Again, what constitutes a 'serious' difficulty is not clarified in the text of the Protocol, but the use of the adjective is clearly intended to underline the degree and extent of the difficulty experienced by the parties invoking Article 16 of the Protocol. The standard of 'serious' difficulty is a high one that should reflect the importance of the goals pursued by the Protocol and the need to ensure that derogations are reserved for exceptional circumstances. It is therefore not enough for a party to show that 86  a safeguard measure is intended to address one of the difficulties identified in the provision, it must also that show the gravity and the pressing nature of the difficulty is such that the suspension of elements of the Protocol is required.
The serious difficulties must also be shown to be 'liable to persist'-that is, that the difficulties are not limited to a short period of time and that there is a real possibility that they will continue into the future in the absence of some form of intervention. For example, short-term trade disruptions resulting from the Protocol that can be addressed through supply-chain adjustments would arguably not constitute difficulties that are liable to persist. By contrast, the condition is more easily fulfilled where a party can show that supply chain adjustments are not sufficient to compensate for the reduction in imports of key goods.
The text of the Protocol provides that safeguard measures are intended to address scenarios where its application has led to serious difficulties. A causal link between the Protocol and the identified serious difficulties must be established. WTO jurisprudence on safeguards has consistently held that the increase in imports can be one amongst many causal factors as long as there is a 'genuine and substantial relationship of cause and effect'. 87 There is no requirement to show that the increase in imports is the only cause or even the primary cause of the domestic injury, only that the causal relationship between the two is significant rather than merely marginal. Given that Article 16 of the Protocol covers scenarios where the Protocol has 'led' to difficulties, the causal standard used in relation to this provision is unlikely to be lower than that used in the WTO safeguards regime. A party invoking the provision would have to show that the difficulty it faces is the result of the application of the Protocol and that, notwithstanding the existence of other contributing factors, the application of the Protocol is, at least, a significant cause of such difficulties.
Establishing the existence of such a causal link may not be as simple as the EU and UK, based on their recent actions, appear to have assumed. As Howse notes, when the European Commission floated the argument that trade restrictions on COVID-19 vaccines could be justified by reference to serious difficulties of a societal nature, it ignored the fact that these societal difficulties were not the result of the application of the Protocol. 88

548
International and Comparative Law Quarterly direct result of the UK's own decision to leave the EU. 89 The same can be said in relation to the UK government's contention that rising costs faced by consumers and businesses demonstrate economic and societal difficulties caused by the application of the Protocol. The Protocol, on its own, does not require the imposition of barriers to trade in goods between Northern Ireland and Great Britain. It is the decision of the UK to leave the EU customs union and the internal market which has made such barriers a necessity.

Trade diversion
Article 16 of the Protocol also permits safeguards in cases where the application of the Protocol leads to trade diversion. Substantively speaking, this is the main deviation from Article 112 EEA Agreement and it is, to put it mildly, a rather baffling one. Trade diversion is defined as an 'increase in trade volume through the replacement of imports from third countries with low-priced imports from trading partners in the free-trade area'. 90 In simple terms, when countries enter into PTAs, trade with more efficient third countries may decrease as a consequence of the granting of preferential treatment, as it is replaced by trade with PTA parties. There is a wealth of economic literature showing that PTAs can have trade diversionary effects. Trade diversion, whilst not inevitable, is a predictable consequence of an international agreement removing barriers to trade between parties. 91 Indeed, three months into the application of the Protocol there were already anecdotal accounts that trade was being diverted to Ireland. 92 Three months later, the UK government published a paper arguing that the mounting evidence that the Protocol had disrupted trade flows between Northern Ireland and the rest of the UK could justify the exercise of Article 16 of the Protocol. 93 What is more, unlike the other circumstances, Article 16 of the Protocol does not require a determination of any significant or 'serious' level of trade diversion. In theory, any evidence of trade diversion, however minimal, could be invoked to justify a derogation from the Protocol.
At first sight, then, the trade diversion condition would appear to lack the requirements of exogeneity and unpredictability that the theory of contingent flexibility argues is key to avoiding abusive uses of exceptions. As discussed, 89  the practice at WTO level suggests that escape clauses that allow parties to derogate from any obligation subject to conditions that are relatively easy to satisfy are likely to prove ineffective. Any use of the safeguard is likely to lead to abuses as it will encourage domestic interest groups which have a vested interest in non-compliance to exert pressure on the parties to trigger the application of safeguards.
There are, however, a number of questions surrounding the condition of 'trade diversion'. The first question concerns the nature of trade diversion envisaged under Article 16 of the Protocol. The term 'trade diversion' is typically understood in relation to the reduction of trade between members of a trade agreement and non-members. However, trade agreements can also affect domestic sales within the territories of member countries.
In a paper assessing trade diversion effects of trade agreements, Dai, Yotov and Zylkin show that internal trade flows-that is, domestic sales of members of trade agreements-are particularly vulnerable to trade diversion. 94 They find that whilst trade agreements tend to reduce imports from non-member countries, trade diversion 'affected internal trade an additional … 21.1% more so than international trade'. 95 The distinction seems relevant given the negotiating history as well as the political discourse that has surrounded the concept of trade diversion in relation to Article 16 of the Protocol.
The first reference to trade diversion in the Protocol can be traced back to the Withdrawal Agreement draft negotiating texts released in November 2018. 96 In this first version of the Withdrawal Agreement, the EU and the UK would form a single customs territory, with Northern Ireland remaining subject to internal market legislation. 97 A number of observers have noted that the inclusion of the reference to trade diversion was made at the request of the EU and was born out of a concern that Northern Ireland would be used as a backdoor into the EU for UK exports. 98 If this is the case, then trade diversion was understood by the EU in terms of its effect on trade between the parties to the agreement.
In the final version of the Protocol, although the UK is no longer part of the EU customs territory, the reference to trade diversion was maintained and, following its entry into force, much of the discussion has focused on how the Protocol has potentially caused an increase in trade between Northern Ireland and the Republic

550
International and Comparative Law Quarterly Northern Ireland. In other words, the recent discourse on trade diversion has focused on how the Protocol may disrupt trade within the UK rather than its effect on EU or UK trade with third countries. There are two points that are worth bearing in mind in relation to the focus on internal trade diversion. The first one is that it reflects the unique nature of the Protocol as a trade arrangement-one that is about removing barriers to trade in goods between one party and a constituent part of the other party. The focus is on how the arrangement affects trade flows within the parties rather than with third-countries. The second point speaks to the very broad language used in Article 16 of the Protocol. The provision does not clarify whether it is seeking to provide parties with a tool to address international or internal trade diversion. That being so, while it seems that the parties have been mostly concerned by the potential impact of the various versions of the Protocol on internal trade, it would be possible, in theory at least, to argue that the effect of the Protocol on trade with non-Protocol parties could justify the invocation of safeguard measures under the Protocol. In practical terms, however, given that the UK's departure from the EU has not reduced but rather created barriers to trade between the UK and the EU, it is difficult to see how any international trade diversion could result from the application of the Protocol.
The second question concerns whether there is any clear evidence of trade diversion since the entry into force of the Protocol. The UK government argued that this was the case by identifying the increase in the value of exports from the Republic of Ireland to Northern Ireland in 2021. 99 However, the correlation of increased exports from the Republic of Ireland to Northern Ireland with the operation of the Protocol, or anecdotal evidence of Northern Ireland firms switching suppliers is not enough to determine the tradediverting or trade-creating effects of a trade agreement. In practice, isolating and quantifying ex-post effects of trade agreements on trade flows is not an easy task. 100 It is primarily an empirical question requiring the identification of the counterfactual and the results will vary from one case to the next depending on the initial structure of the economic relationship, the economic sectors and the level of integration pursued by the parties. 101 Some trade agreements have been found to cause significant trade diversion whilst others have led to none at all. 102 The magnitude of the effects can also vary depending on the statistical models used to capture the effects of trade 99 UK Command Paper (n 13) 13-14. agreements, with design choices often exaggerating the trade-diverting or tradecreating effects of trade agreements. 103 Because of its unique features, the Protocol will, of course, present its own challenges. It does not remove barriers to trade between two parties but maintains the status quo between the EU and one constituent part of the UK. The barriers to trade are also mostly regulatory in nature. This is potentially significant in that most studies on trade diversion have focused on trade agreements that seek to reduce or remove tariffs between members. However, recent research on so-called deep trade agreements (focusing on the removal of regulatory barriers to trade) indicates that such agreements tend to have a positive effect on trade between members, and between members and nonmembers. 104 In short, the assessment of trade diversion is a complex matter and it can neither simply be assumed that the Protocol will inevitably lead to trade diversion nor that evidencing diversion will be straightforward.
Where trade diversion is demonstrated, there is also a question of what level of trade diversion would be sufficient to justify the invocation of safeguards. Article 16 of the Protocol does not suggest a threshold above which trade diversion could be deemed sufficiently problematic to justify the temporary suspension of Protocol obligations. Nor does it specify whether parties invoking the provision must demonstrate the existence of net trade diversion (that is, that the overall value of trade diversion outweighs the value of trade creation).
The absence of any qualification as to the magnitude of trade diversion falling under the scope of Article 16 of the Protocol seems incongruous in light of the fact that safeguards are tools that have traditionally been reserved for exceptional circumstances in most trade agreements. It also stands in contrast to the other conditions for invocation, which require a demonstration of the seriousness of negative-economic, societal and environmentalexternalities associated with the Protocol. This terminological distinction between the first category of conditions and that of trade diversion suggests that the parties opted to subject the latter to a lower threshold. But should that mean that any trade diversion, however insignificant, can justify the application of safeguard measures? Such an interpretation would be highly problematic as it would potentially give the parties a free pass to deviate from the obligations under the Protocol and fundamentally undermine the purpose and the operation of the agreement in its entirety.
The third and final question relates to the establishment of a causal link between the Protocol and the diversion of trade. According to Article 16 of the Protocol, the application of safeguard measures is justified where the application of the Protocol leads to diversion of trade. On this, it is questionable whether the Protocol, on its own, has led to diversion of trade. As discussed, the Protocol did no more than maintain the status quo for Northern Ireland in terms of its trade relationship with the EU in relation to goods-that is, Northern Ireland continues to trade with the EU as if it were still part of the EU customs union and internal market.
When the Protocol was negotiated and agreed upon, the barriers to trade which currently exist between Northern Ireland and Great Britain were not an inevitability. Such barriers were a consequence, not of the Protocol but rather of the UK's decision to leave the EU 105 and the conclusion of the EU-UK TCA, which raised barriers to trade in goods between the EU and the UK. To what extent, then, can any potential decrease in trade between Great Britain and Northern Ireland and corresponding increase in trade between Northern Ireland and the Republic of Ireland be attributed to the Protocol? And, even if it is accepted that the Protocol has, at least partially, led to trade diversion, can such diversion constitute a valid reason to suspend Protocol obligations when it is entirely self-inflicted?
At this point, it is worth circling back to the lack of any requirement to demonstrate the exceptional or unforeseeable nature of diversion of trade in Article 16 of the Protocol. This would have surely assuaged the parties' concerns about being unable to suspend their obligations. However, as discussed, validly invoking this condition may not be as easy as some may assume. Once it is established what constitutes trade diversion for the purposes of the Protocol, demonstrating the existence of diversion and a causal link with the Protocol may prove problematic.

Good faith requirement
Until further clarification is given, either by the parties or through case law, the meaning of the substantive standards for the invocation of Article 16 of the Protocol will remain the subject of debate. Compared to the WTO safeguards regime, Article 16 of the Protocol is wider in scope and imposes less restrictive substantive requirements, potentially giving parties greater discretion in deciding whether to apply safeguards.
On the surface, this could be explained by the fact that the Protocol is far more ambitious than GATT in terms of the depth of its liberalisation commitments. Whilst both cover trade in goods, the former achieves market integration through the removal of regulatory barriers to trade in goods between Ireland and Northern Ireland. It should come as no surprise, then, that negotiators envisaged the application of safeguards in circumstances which did not relate exclusively to increase in imports in goods. It is also not particularly surprising that, in a trade arrangement requiring regulatory alignment, a safeguard system should allow for the suspension of obligations other than just tariffs and tariff-rate quotas. It is possible that the negotiators considered that such a far-reaching arrangement required a similarly far-reaching escape clause. This was, after all, one of the main concerns of the drafters of the EEA Agreement, another market integration arrangement, when they conceived Article 112 EEA Agreement, the main inspiration for Article 16 of the Protocol.
But such a broad escape clause also creates certain problems with which the parties to the Protocol must now contend. If the application of safeguards is made contingent upon the determination of circumstances that are, in some cases, neither exceptional nor unpredictable, the Protocol gives the parties carte blanche to deviate from their obligations whenever it suits them. There is a reason why similar provisions at WTO level, as well as Article 112 EEA Agreement, have rarely been used. Once a signal is sent to domestic interest groups that an escape clause allows parties to pick and choose when they should comply or not with their international commitments, the temptation will be there to impose political costs on governments to make use of such clause.
The EU's aborted attempt to invoke the safeguard provision to impose export restrictions on COVID vaccines in trade between the EU and Northern Ireland is a perfect illustration of how vulnerable Article 16 of the Protocol is to abuse. Whilst the export restriction was never applied in relation to Northern Ireland, the mere notion that the safeguard could be invoked in practice was enough to embolden domestic interest groups within the UK to call for the UK government to apply its own safeguard measures. 106 The issue is further exacerbated by the risk that domestic interest groups may instigate the very circumstances that could justify the invocation of Article 16 of the Protocol. For example, from the moment the Protocol entered into force, unionist political parties opposed to its operation have repeatedly called for the suspension of the Protocol on the basis that the border checks on goods from Great Britain destined for Northern Ireland could lead to anger within the unionist community and undermine stability in the region. 107 Such claims were followed by protests that have then been held up as evidence of the societal unrest caused by the Protocol and as reasons for the invocation of Article 16 of the Protocol.
A further fundamental issue arises from the very existence of Article 16 of the Protocol. In most trade agreements, when an escape clause is invoked, regardless of whether or not it is poorly designed, the worst that can happen is that barriers to trade will be reimposed. There will be adverse economic, perhaps even political and social, consequences for parties involved, but very rarely do such barriers pose existential threats to the arrangement. But, in the context of the Protocol, any deviation from international commitments which leads to the imposition of a barrier to trade challenges one of the central goals of the agreement-that is, to avoid border checks within the island of Ireland. Any imposition of any sort of border checks on goods would be extremely problematic and holds the potential to destabilise the entire agreement.
This poses the question of how to ensure that the discretionary powers under Article 16 of the Protocol are not used in an unreasonable or arbitrary manner. How is the abuse or disingenuous use of safeguards to further the interests of domestic interest groups to be avoided? On this matter, the concept of good faith may prove useful. The good faith obligation is a generally accepted principle of international law, which requires rationality and reasonableness in the application of international law. 108 Of particular relevance is the obligation on treaty parties not to exercise their rights in an abusive manner-that is, in a manner or for reasons that are contrary to the very purpose of the provision. 109 The misrepresentation of the true reasons for the use of a safeguard would certainly fall under banner of an abuse of rights.
It is worth noting in this regard that the Withdrawal Agreement also includes a duty of good faith, which applies to both the EU and the UK in relation to the Protocol. Article 5 of the Withdrawal Agreement requires that the parties 'shall, in full mutual respect and good faith, assist each other in carrying out tasks which flow from this Agreement'. Some have argued that the duty of good faith in the Withdrawal Agreement does not add much to the international law principle. 110 Others have posited that the interpretation of the good faith requirement should vary in accordance with the level of integration being pursued in the treaty, that 'the greater the depth of integration that an agreement is thought to require, the more far-reaching the principle of good faith is likely to be'. 111 To the extent that the Protocol maintains Northern Ireland in the EU's internal market for goods, one might wonder whether the EU law equivalent to the duty of good faith-the EU law principle of loyalty, or sincere cooperation 112could potentially reinforce the scope of the obligation. Such a conclusion is not as far-fetched as it may appear. The text of Article 5 of the Withdrawal Agreement replicates to a large extent both Article 4 TEU and Article 3 EEA on the principle of loyalty. 113 The exact impact of the good faith requirement in relation to the Protocol is yet to be defined and will undoubtedly play a significant role in disputes between the EU and the UK. However, irrespective of whether the duty of good faith is interpreted in the narrow sense of international law principles, or in its broader expression under EU law, this should be enough to preclude the unreasonable or abusive exercise of rights under Article 16 of the Protocol.

Scope and duration of safeguard measures
Whilst there may be some uncertainty as to the scope of the conditions for the invocation of safeguards under Article 16 of the Protocol, the following segment of the provision imposes strict limits in terms of the manner in which safeguard measures can be applied. Safeguard measures, it is stated, should be 'restricted with regard to their scope and duration to what is strictly necessary in order to remedy the situation' and that priority should be given to measures that 'least disturb the functioning of the Protocol'.
The limitation on both the scope and duration of safeguard measures is that they should be strictly necessary to rectify the difficulty. This alludes to the application of a proportionality test, akin to those developed in the context of the EU and WTO legal orders, 114 when assessing the legality of a trade restrictive measure. The practice at both EU and WTO level shows that the application of the proportionality test will often vary depending on the nature of the difficulty arising from the application of the Protocol, the restrictiveness of the measure and the intended aim of the derogation. The discretion accorded to parties in terms of the level of restrictiveness that could be applied to achieve a policy goal would fluctuate depending on the importance of said goal. One would expect, for example, that a party invoking a safeguard in order to address a public health or environmental concern would be given more leeway than one derogating from the Protocol to address trade diversion concerns.
However, the presence of the term 'strictly' suggests the application of a strict proportionality analysis where adjudicatory bodies will assess the extent to which a less restrictive measure could achieve the same objective. 115 Article 16 of the Protocol thus significantly limits the discretion of the parties in the determination of a measure's necessity. A party wishing to adopt a safeguard measure must demonstrate that said measure is the only one that could achieve its intended goal and that there are no less burdensome options available.
An assessment based on strict necessity makes sense in light of the importance of the Protocol's aims and the highly problematic and undesirable consequences which would arise from any derogation from the trade-related obligations included in the Protocol (notably the possibility of the introduction of border checks in the island of Ireland). It would also serve as a counterbalance to the very broad scope and nature of the conditions for the invocation of safeguards under Article 16 of the Protocol. Whilst it may in theory be relatively straightforward for a party to justify the imposition of safeguards under the Protocol, demonstrating that a measure is strictly necessary to achieve one of the goals listed in Article 16 of the Protocol will prove a more arduous task.
Beyond the requirement of proportionality, there are no further limits regarding the scope of safeguard measures. The upshot is that Article 16 of the Protocol is a horizontal exception that covers all obligations under the Protocol, meaning that parties are free to deviate from Protocol obligations whose purpose is to avoid border checks between Ireland and Northern Ireland.

Procedural requirements
Although safeguard measures can be adopted unilaterally, they are subject to the procedural requirements set out under Annex 7 of the Protocol. A party considering adopting safeguard measures must notify the other party of its intention to do so and provide all relevant information. 116 A formal notification would presumably include the nature of the serious difficulties, the provision of the Protocol with which compliance is causing the difficulties, and the proposed measures to address such difficulties. Upon notification, the parties must immediately engage in consultations in the Joint Committee 'with a view to finding a commonly acceptable solution'. 117 The role of the consultation process is to give the parties an opportunity to discuss solutions that may obviate the need for derogations from the Protocol or, if the latter is not possible, to agree on safeguard measures that would be least problematic in terms of the operation of the Protocol.
In any event, the safeguard measures can only be applied one month before the initiation of the consultation procedure unless either the consultations were terminated before the end of the time-period or where 'exceptional circumstances requiring immediate action exclude prior examination'. 118 In the latter scenario, safeguard measures should, again, be limited to measures that are 'strictly necessary to remedy the situation'. 119 Once adopted, safeguard measures must be formally notified to the other party with all relevant information. 120 These measures are subject to a review by the Joint Committee every three months to explore the possibility of either removing such measures prior to their expiry date or mitigating their impact. 121

Rebalancing measures and remedies for non-compliance
Article 16(2) provides that rebalancing measures may be taken against the party adopting the safeguard measures. Rebalancing measures can be adopted where the application of safeguard measures causes an imbalance between the rights and obligations of the parties under the Protocol. To the extent that any move to derogate from an obligation in the Protocol would arguably disturb the balance of the negotiated outcome, any use of a safeguard measure by one party could trigger the right for the other party to apply rebalancing measures. This means that rebalancing measures are permitted even if the safeguard measure is lawful.
The rules applicable to rebalancing measures broadly mirror those that apply for the safeguard measures to which they are they are responding. 122 Rebalancing measures must be adopted in accordance with the procedure set out under Annex 7 for safeguard measures. They must comply with the proportionality requirement-that is, they must be strictly necessary to remedy the imbalance and be the least restrictive measures possible for the functioning of the Protocol. And, as with safeguard measures, the Protocol does not clarify what form rebalancing measures should take-that is, whether they should be limited to the reimposition of tariffs or tariff-rate quotas-or whether parties are free to suspend any obligation derived from the Protocol. This mirrors the approach adopted in the EU-UK TCA where rebalancing measures are also not defined. 123 The inclusion of a mechanism allowing parties affected by a derogation to apply rebalancing measures is a standard feature of escape clauses based on the efficient breach theory. It is a feature which, one could argue, is all the more justifiable when dealing with an escape clause which, like Article 16 of the Protocol, allows a significant margin of discretion for the parties invoking it. It could, presumably, serve as a disincentive for countries that would seek to 118 Annex 7, para 3 of the Protocol. 119 ibid. 120 Annex 7, para 4 of the Protocol. 121 Annex 7, para 5 of the Protocol. 122 Art 16(3) of the Protocol. 123 E Lydgate, E Szyszczak, LA Winters and C Anthony, 'Taking Stock of the UK-EU Trade and Cooperation Agreement: Governance, State Subsidies and the Level Playing Field' (UKTPO Briefing Paper 54, January 2021) <https://blogs.sussex.ac.uk/uktpo/publications/taking-stock-ofthe-uk-eu-trade-and-cooperation-agreement-governance-state-subsidies-and-the-level-playingfield/>.

558
International and Comparative Law Quarterly abuse that discretion. This, as discussed in Part I, is the central feature of escape clauses that rely on the imposition of an economic cost as a means to deter parties from escaping their obligations. However, the option of applying rebalancing measures in response to safeguards is problematic. First, this is because, as discussed in Part I, the 'escape and compensate' 124 mechanism has been criticised, namely because the practice suggests that imposing an economic cost for the use of exceptions in trade agreements is not necessarily an effective means to disincentivise the abusive use of such exceptions. On the contrary, it may create an incentive for domestic actors that are opposed to the trade agreement to increase costs of compliance.
Second, this is problematic because it makes little sense to punish a party for adopting a safeguard measure that is genuinely intended to pursue a legitimate policy objective. In the context of general exceptions provisions under WTO law, for example, countries are allowed to adopt measures that would otherwise be incompatible with their obligations if they are able to show that those measures are necessary or relate to certain non-trade policy goals, and that such measures are not disguised forms of protectionism. In such instances, other WTO Members do not have the right to respond by suspending their commitments in relation to the WTO Member that has derogated from its obligations for valid reasons. Similarly, it is odd that either party to the Protocol should be entitled to derogate from their own obligations in response to safeguards adopted by the other party that are entirely justified by reference to an important public policy goal, such as the protection of the environment.
Third, rebalancing measures appear incongruous in the overall context of the Protocol. Trade agreements that pursue liberalisation as a means to achieve economic welfare are the result of negotiations based on reciprocity. The upshot is that where a party opts to suspend market access concessions for economic or political purposes, it is at least arguable that the other parties affected by such measures should be able to respond in kind, to redress the balance. However, it is debatable whether a similar logic should apply to a trading arrangement where one of the central objectives is to avoid border checks on traded goods. It is worth considering what an appropriate rebalancing measure from the UK would have looked like, in response to the EU's attempt to impose COVID-19 vaccine restrictions on EU-Northern Ireland trade. If, in this hypothetical scenario, the UK were to impose similar export restrictions on trade between the UK and the EU, then rebalancing measures would only further undermine one of the central goals of the Protocol by creating the need for more border checks.
The rebalancing measures mechanism provides the parties with a tool to respond to application of safeguards irrespective of whether the use of safeguards is considered lawful. The objective of rebalancing measures is not to punish a party for making use of Article 16 of the Protocol but rather to redress the balance by allowing the other party to withdraw equivalent concessions.
However, other remedies are available to encourage compliance where there is a determination that a safeguard under Article 16 of the Protocol has been applied unlawfully. To do this, a party must challenge the legality of the safeguard under the Withdrawal Agreement's arbitration mechanism. The arbitration panel must give its ruling within 12 months from its establishment. 125 If the ruling is delivered in favour of the party claiming the breach of Article 16 of the Protocol, that party (the complainant) must notify the other party (the respondent) within 30 days regarding the amount of time it considers necessary for the other party to comply with the ruling. 126 If there is a disagreement regarding the reasonable amount of time needed to ensure compliance, the complainant can request another panel to rule on the matter. If the respondent then fails to comply within a reasonable period of time, the complainant can request the original arbitration panel to rule on the matter. 127 In the event where the panel confirms the continued noncompliance with the original ruling, it can impose a lump sum or penalty to be paid by the respondent to the complainant. 128 In addition to financial sanctions, the Withdrawal Agreement allows for the application of cross-retaliatory measures. If the respondent fails to pay the financial sanctions imposed by the arbitration panel within a month after the ruling ordering such payment, or six months after the delivery of the original ruling, the complainant can suspend other parts of the Withdrawal Agreement or any other agreement concluded between the EU and the UK. 129 The latter category includes the EU-UK TCA, meaning that the complainant could, for example, reimpose tariffs and tariff-rate quotas in response to the unlawful application of safeguards under the Protocol. 130 Any such suspension of the TCA must, however, be 'proportionate to the breach of obligation concerned, taking into account the gravity of the breach and the rights in question'.
The threat of financial sanctions and cross-retaliatory measures could, in theory, discourage parties from using safeguards in circumstances where they are not justified. However, an obvious drawback of these remedies is that they can only be applied following time-consuming arbitral proceedings. By the time these remedies for non-compliance can be applied, the unlawful safeguard measures will have been in place for months, perhaps even longer, and the damage will have been done. It is therefore unsurprising that other options have been explored by the parties. The EU, for example, was said to be considering the possibility of 125

560
International and Comparative Law Quarterly using the threat of terminating or suspending the TCA in response to any use of the Protocol safeguards by the UK. Here, there are two options available. First, a party can terminate the TCA without justification with one year's notice. 131 Second, a party can terminate or suspend the TCA after a 30-day notice period if it can show that the other party has breached the 'essential elements' clause included in the TCA. Under this provision, a party would have to argue that the unlawful use of safeguards under the Protocol amounted to a serious and substantial failure by the other party to fulfil the rule of law. 132 Some have questioned whether an action under Article 16 of the Protocol could justify the use of the TCA's essential elements clause. 133 Indeed, it seems unlikely that a decision to derogate from a limited set of obligations under the Protocol could justify a suspension or termination of the TCA under the essential elements clause. This would hardly constitute a serious or substantial failure to uphold the rule of law and, in any case, would fall foul of the requirement that such a suspension or termination of the TCA should be proportionate and limited to measures that least disturb the functioning of the agreement. By contrast, an argument that the essential elements clause has been breached is likely to stand on a much surer footing in instances where Article 16 of the Protocol is used by a party as a pretext to avoid compliance with their obligations in the absence of any reasonable justification or where a safeguard measure is applied in a disproportionate manner (eg refusal to apply the Protocol in its entirety).
Ultimately, whether an abuse of Article 16 of the Protocol may or may not constitute a sufficient basis for suspension or termination of the TCA is perhaps not as important as the fact that the threat of such suspension or termination provides each party with a powerful tool to discourage the other party from invoking Article 16 in an abusive manner. In this regard, it is telling that when, in the autumn of 2021, the prospect of the UK invoking Article 16 of the Protocol became less remote, the EU reportedly considered making use of the essential elements clause in the TCA, as well as withholding equivalence decisions in the financial services sector and suspending adequacy decisions for the transfer of personal data. 134 Such actions could be applied almost immediately and would increase the EU's leverage in any discussions relating to Article 16 of the Protocol. This is in contrast to rebalancing measures, which are limited in their scope, and the remedies available for non-compliance under the Withdrawal Agreement, which can only be applied at the end of time-consuming arbitral proceedings.
The EU's willingness to signal its intent to ramp up the economic cost that would be incurred by the UK were it to derogate from its obligations on the basis 131  Article 16 of the Protocol, reinforces the notion that this is an escape clause that is perceived as being prone to abuse. The seemingly broadly-phrased external conditions for invocation, combined with the absence of a third-party gatekeeper that would verify the existence of such conditions and the validity of the safeguards prior to their application means that making escape costly is viewed as the only viable tool to deter abuse.

V. CONCLUSION
When the EU and the UK drafted Article 16 of the Protocol, they transposed a legal mechanism which perhaps might have made sense in the framework of trade liberalisation arrangements, with seemingly little thought being given to whether such mechanism could work in the context of a trade agreement where the removal of barriers to trade is not a means for economic prosperity but rather for the maintenance of peace and security. Perhaps the negotiators considered that including an escape clause would make selling the Protocol to those interest groups that were opposed to it easier. And perhaps the assumption was that, like the corresponding provision in the EEA, Article 16 of the Protocol would never be used in practice. If so, this was a considerable gamble, and one that may well still pay off. Still, one wonders whether the drafters of the Protocol may eventually rue not having opted for a more cautiously worded text which heeded the lessons of international relations literature and practice on escape clauses. In any event, whilst the conditions for the invocation of safeguard measures under Article 16 of the Protocol are broad and confer upon the parties a certain amount of discretion, there are also a number of requirements that will limit such discretionary powers. These include a requirement to demonstrate the existence of difficulties of a serious nature or trade diversion and the establishment of a causal link between the application of the Protocol and these external circumstances. The Protocol also includes the application of a good faith requirement, which should preclude the exercise of the provision in an abusive manner and a high threshold for the necessity test, which limits the scope of potential safeguard measures.
Further, where safeguards are adopted, this should occur in line with the procedural requirement set out in Article 16 of the Protocol and, preferably, with the agreement of both parties. If not, judicial review of such measures is possible, either in the context of the arbitration mechanism created by the EU-UK Withdrawal Agreement or before EU or UK courts. 135 Here, the onus will be on the relevant adjudicatory bodies to use the limited tools available in Article 16 of the Protocol and to interpret said provision in a manner that ensures that the use of safeguards does not undermine the delicate balance the Protocol seeks to achieve in light of the unique circumstances on the island of Ireland.