When Internal Politics Meet External Duties: Domestic Law in the Trade and Sustainable Development Chapters and How the Retreat of EU Environmental Law Is Breaching Its Bilateral Obligations

Table of Contents: 1. Introduction. – 2. Domestic Law in TSD Chapters – 2.1. Right to regulate. – 2.2. Upholding levels of protection. – 3. Farewell Green Deal: The retreat of EU environmental law. – 4. How the EU is in breach of its bilateral obligations under the TSD chapters. – 4.1. Obligations under the Right to regulate article. – 4.2. Obligations under Upholding levels of protection article. – 5. Conclusions.

Abstract: This article critically examines the European Union’s (EU) Trade and Sustainable Development (TSD) chapters in its Free Trade Agreements (FTAs), focusing on obligations related to domestic environmental and labour protection laws. While the EU presents itself as a global leader in sustainable development, recent regulatory regression in EU environmental legislation, such as delays to the Regulation on Deforestation-free products (EUDR), the watering down of the Nature Restoration Regulation, and changes introduced through the Omnibus Package to boost competitiveness, casts doubt on the coherence between its internal policies and external commitments. This tension raises legal questions concerning the EU’s compliance with its bilateral TSD obligations. Notably, sustainability obligations under FTAs concluded since 2011 have been largely absent from internal debates on these developments. Central to the analysis are two core provisions present in all EU TSD chapters to date: ‘Right to regulate and levels of protection’ and ‘Upholding levels of protection’. Although these provisions aim to safeguard domestic environmental and labour standards, their content and legal implications vary across FTAs and remain underexplored. Through comparative textual analysis grounded in the Vienna Convention on the Law of Treaties, the article categorises and clarifies these obligations and examines their relationship. It then assesses whether recent EU legislative and policy regressions may amount to breaches of its own TSD commitments. The analysis reveals a structural disconnect between the EU’s external normative ambitions and its evolving internal policy landscape, highlighting the need to reconsider the EU’s dual role as both norm-setter and norm-follower in the global sustainability regime.

Keywords: European Union – trade and sustainable development – external relations – domestic law – free trade agreements – regression.

1.   Introduction

In parallel to the efforts of the European Union (EU) to tackle climate change and transition to a sustainable society within the EU borders, the external relations of the EU aim to promote the same in its relations with the rest of the world. Not only does this stem from the EU being seen as a global norm exporter[1] but also it is a mandate given to the EU by its own treaties, particularly following the Lisbon Treaty.[2]

Recent developments in EU law, particularly the growing trend to relax or delay the implementation of key Green Deal legislation, risk undermining the EU’s carefully cultivated image as a global leader in sustainable development. This shift, often framed as pragmatic or responsive to economic pressures, not only challenges the internal coherence of the EU’s climate and environmental ambitions but also raises serious concerns about the Union’s compliance with its international obligations. Notably absent from the current political and public discourse is the fact that such legislative backtracking may constitute a breach of the EU’s bilateral commitments under its trade and sustainable development (TSD) chapters. This dimension, largely overlooked, highlights a worrying disconnect between the EU’s external promises and internal policy choices.

Against the background of the crisis the multilateral trade system is facing,[3], an enshrinement efforts of the EU to promote sustainable development in its bilateral relations is the TSD chapters. Since the South Korea Free Trade Agreement (FTA) of 2011,[4] every new generation EU FTA has a TSD chapter that establishes obligations regarding environmental protection and labour rights.

TSD enforcement or lack thereof has dominated the academic debate ever since, due to the institutional construction surrounding these chapters.[5] Following several rounds of dialogue with stakeholders over the years[6] some of these shortcomings seem to be answered with the latest communication from the European Commission.[7] For the first time, the Commission accepted the possibility of trade sanctions for certain grave breaches of TSD obligations and announced that it would expand the compliance stage of the general dispute settlement mechanism (DSM) of the FTAs to cover the TSD chapters. In all the FTAs prior to the recent EU-New Zealand FTA,[8] the TSD chapters are excluded from the general DSM[9] and subject to a separate regime where a dispute can be taken to a Panel of Experts (PoE) after the government consultations have failed. The PoE can only give recommendations, and there is no follow-up on compliance.[10]

Typically, the TSD chapters are seen as using non-binding and unclear language and consequently are commented to be ‘toothless tigers’.[11] Therefore, these developments on the procedural side bring to mind what substantive obligations the TSD chapters contain which has not received the same level of attention. Answering this question requires a close look at the substantive articles of the TSD chapters, the content of which varies from FTA to FTA.[12] Two articles of the TSD chapters establish obligations regarding the parties’ domestic laws on environmental and labour protection, entitled ‘Right to regulate and levels of protection’ and ‘Upholding levels of protection’. These obligations regard the levels of protection in domestic law and how these should be safeguarded and enforced. The content of the obligations the parties undertake hereunder, and how these two articles relate to each other, is, however, unclear. 

This article aims to clarify the content of and the interplay between these articles. Several versions of each obligation they establish are present under different FTAs, which require an in-depth and word-for-word analysis. This analysis finds its theoretical framework in the Vienna Convention on the Law of the Treaties (VCLT) and the customary rules of treaty interpretation reflected in its Articles 31–33. Because there is little jurisprudence or doctrinal guidance on the interpretation of TSD chapters, additional sources of interpretation are necessary. 

In doing so, the study takes a EU-centric approach. Many concepts used in these chapters are undefined in FTAs but are well established in EU law. Treaty negotiators tend to rely on their own legal traditions. This is also relevant for the Commission in its role as the negotiator of FTAs on behalf of the EU. Consequently, the sophisticated body of EU law, supported by well-established jurisprudence, provides a useful reference point for interpreting concepts that appear both in EU law and in the TSD chapters.

Secondly, the CJEU possesses specific competences with regard to the EU’s international agreements.[13] The Court treats international treaties concluded by the EU as acts of the EU institutions,[14] to which the full range of judicial procedures for the enforcement of EU law applies. First, the CJEU may review the compatibility of international agreements with EU law, both ex anteand ex post. Second, EU international agreements are subject to interpretation by the CJEU.[15] Consequently, the institutional bodies relevant to the TSD chapters are not limited to those expressly provided for therein. Ultimately, the CJEU acts as the final interpreter of all EU law, including international agreements that automatically form part of the EU legal order upon entry into force, such as FTAs. Should a case arise, the TSD chapters will therefore be interpreted on the EU side by the CJEU in light of EU law. It is thus appropriate to analyze the substantive provisions of the TSD chapters through an EU law lens, as they must in any event comply with EU law. 

The study at hand has two main sections. It will first look at the categories of obligations under these two articles and examine their enforceability. Secondly, it will analyse and clarify the complementary relationship between the obligations the two articles establish, aiming to cast a broad net to catch different types of regression and promote high levels of protection. Lastly, it analyses latest news from Brussels in how the EU is relaxing its laws and policies regarding the environmental protection in different areas as case studies and showcases the often-forgotten face of the TSD medallion that the EU itself is just as bound with these rules in its FTAs and is not compliant itself. 

2.   Domestic law in TSD chapters[16]

2.1.  Right to regulate

The first article of the TSD chapters regarding the domestic law of the parties is entitled ‘Right to regulate and levels of protection’. This provision, which appears in all TSD chapters from the very first example in the EU-South Korea FTA of 2011, regulates the level of environmental and labour protection and its improvement. A close reading of the article in different FTAs reveals three main versions, each with distinct grammatical constructions and legal implications for the parties.

First, all versions state that the parties have the right to determine their own level of protection for labour and environment and to adopt or modify their laws and policies accordingly. Such a level, however, should be consistent with ‘internationally recognized standards and agreements’ referred in the TSD chapter. They then prescribe high levels of protection and require the parties to continue to improve their laws and policies.

2.1.1.   Three main versions

Despite their similar coverage, the three versions differ in their formulation and, consequently, the enforceability of the obligations they impose. Table 1 below gives an example of each version. This section will firstly look at the content of the three versions of the article that appear under different FTAs. It will then analyse the content of the obligations and what is expected of the parties in detail.

Table 1: Versions of right to regulate and levels of protection article.

The first version, used in FTAs such as South Korea, Andean,[20] Armenia,[21] and Canada,[22] recognizes the right of each party to establish its own levels of environmental and labour protection and to adopt or modify its relevant laws and policies. It then states that each party ‘shall seek to ensure’ that its laws and policies provide for and encourage high levels of environmental and labour protection, consistent with internationally recognized standards or agreements referred to in the relevant TSD articles, and ‘shall strive to continue to improve those laws and policies’. The use of ‘seek to ensure’ and ‘strive to continue to improve’ signals an obligation of effort rather than result, requiring parties to demonstrate genuine attempts to uphold high levels of protection, but not holding them strictly liable for the outcome.

Found in for instance FTAs with Central America,[23] Moldova,[24] Georgia,[25] Singapore, and Ukraine,[26] the second version is structured in two paragraphs. It recognizes the right to establish protection levels ‘consistent with the principles of the internationally recognized standards or agreements to which it is party’ and states that the parties ‘shall continue to improve’ their laws and ‘strive towards providing and encouraging high levels of environmental and labour protection’. This version separates the obligation to maintain consistency with international standards from the duty to strive for high protection, placing greater emphasis on consistency with international standards and mandating continuous improvement of domestic laws. A clearer and more binding language that establishes result obligations is used in this version. Vietnam FTA[27]on the other hand brings neither obligation to be consistent with international standards nor an obligation to improve domestic law; only an obligation of effort to ensure a high level of protection.

Adopted in (draft) agreements with New Zealand, Kenya,[28] Mexico,[29] Australia, Indonesia,[30] India,[31] Chile,[32] and Mercosur,[33] the third version also uses two paragraphs. It affirms the right to determine sustainable development policies and protection levels, and to adopt or modify laws and states that such levels ‘shall be consistent with each Party’s commitment to the internationally recognized standards and agreements referred to in this Chapter’. Parties ‘shall strive to ensure’ high levels of protection and ‘shall strive to improve’ these levels. The use of ‘strive’ is generally interpreted as requiring more serious and demonstrable efforts than ‘seek to ensure’, raising the threshold for compliance and reflecting a more robust effort-based obligation.

The Kazakhstan FTA stands out for its brevity and lack of reference to high levels of protection, only requiring parties to establish their own levels of protection and adopt or modify laws in line with the international standards and agreements referred to in the TSD. This limited approach reflects the narrower scope of the agreement, or the specific context of the parties involved.

2.1.2.   Content of the obligations

In general, the right to regulate articles establish three obligations, each of which will be analysed now:

1) Domestic law being consistent with international agreements and standards;

2) Domestic law striving for high levels of protection;

3) Improvement of domestic law.

A key divergence among the versions is the placement and meaning of the requirement for domestic law to be ‘consistent with internationally recognized standards or agreements referred to in [TSD]’. In the first version, this phrase modifies the obligation for high levels of protection: parties ‘shall seek to ensure’ that their laws and policies provide for and encourage high levels of protection, consistent with the internationally recognized standards or agreements referred to in the relevant TSD articles. This means that the parties can adopt their own levels of protection, but they must show efforts to ensure that the level enacted in their laws and policies is consistent with these international standards and agreements.

In the second and third versions, the requirement for consistency is separated from the obligation to strive for high levels of protection. Here, the right to regulate itself is conditioned on consistency with the principles of the internationally recognized standards or agreements to which the party is a signatory, as referred to in the TSD. This creates a result obligation: the adoption or modification of laws and regulations must be consistent with these standards and agreements. Parties are free to choose their level of protection, but only if it conforms with the said standards, meaning domestic law must be at least at the minimum level required by these standards or higher.

There are two possible readings of this version. One is that it merely recognizes the sovereignty of states and their self-binding upon accession to an international agreement; the other is that it brings an actual limitation to the right to regulate, requiring conformity with the listed standards and agreements. Calling such clauses ‘minimum-level clauses’, Durán notes the limited additional value as it merely incorporates what is already binding upon the parties under the relevant international agreement.[34]

However, such incorporation under this article still brings added value; firstly a breach of the relevant international agreement will now also constitute a breach of the TSD and thus the FTA, creating a complementary jurisdiction.[35] This is particularly important for the multilateral agreement itself lacks a binding enforcement mechanism which are strengthened through the possibility of trade sanctions following the European Commission’s recent assertive approach to the TSD chapters. This covers serious breaches of the Paris Agreements and the fundamental ILO conventions. 

Similar obligations are already regulated elsewhere in the TSD chapter, both in articles on multilateral agreements[36] and in those on specific environmental areas. A critical point the Right to regulate articles differ from others also incorporating multilateral agreements and standards is the absence of a trade or investment linkage. As explained below under Upholding levels of protectionarticles, many TSD obligations are linked to whether the breach has been either ‘in a manner affecting trade/investment’ or committed‘as an encouragement for trade/investment’. This linkage limits the breaches that can be addressed through the TSD system. A breach of Right to regulate article, however, can still be established without having to show such a linkage.

The scope of ‘internationally recognized standards or agreements referred to in [TSD]’ must also be clarified. If ‘referred to’applies only to agreements, the standards are undefined and potentially limitless, undermining foreseeability. It would be unreasonable to expect the parties to follow and comply with all the international standards that are out there. If it applies to both, only those standards and agreements specifically mentioned in the TSD chapter are covered, ensuring legal certainty. This interpretation is preferable, as it aligns with the principle of foreseeability and the apparent intent of the parties.

One should first address what ‘striving for a high level of protection’ entails. All three versions require parties to seek or strive for high levels of protection in domestic law and policy, consistent with international agreements and standards. The first and second versions use ‘shall seek to ensure’, establishing an effort obligation: parties must demonstrate genuine attempts to achieve high levels of protection, but are not strictly liable for the outcome. The third version, using ‘shall strive’, raises the threshold for compliance by requiring serious, demonstrable efforts, mere token actions are insufficient.

The difference between ‘seek to ensure’ and ‘strive’ is significant. ‘Seek’[37] is defined as to try, attempt, or aim at something, while ‘strive’[38] means to devote serious effort or energy, or to try very hard to achieve something. Thus, the third version requires parties to do more than simply try: they must make strong, conscientious efforts, potentially devoting significant resources, time, and energy. The legal threshold for compliance is higher in the third version, and the threshold for establishing a breach is therefore lower.

This reading is supported by the principle of effectiveness in treaty interpretation.[39] The fact that the drafters of recent FTAs chose to deviate from previous practice and use ‘strive’ should be read as a conscious decision to impose a stronger obligation. The general rule of interpretation under Article 31 VCLT dictates that the ordinary meaning of words should be given in light of the object and purpose of the treaty and the context. The most recent FTAs, which use the ‘strive’ version, also have more extensive lists of sustainability objectives in their TSD chapters, reflecting a greater importance attached to sustainable development by the parties.

Under the ‘to seek to ensure’ version, showing of any efforts can satisfy the obligation and only a complete inaction or going in the opposite direction can give rise to a breach. In this sense, this version brings a ‘non-regression’ obligation, prohibiting sliding backs in the domestic level of protection. 

The evidentiary burden for effort obligations, or so-called ‘best-endeavour obligations’[40] is inherently challenging. It is difficult to objectively prove the presence or absence of efforts at any level. Result obligations are always more easily enforceable, as the achievement of a specific result can be objectively established. The lack of a clear definition of what qualifies as ‘striving for’ or what constitutes serious efforts creates uncertainty regarding enforcement, with different interpretations possible.

A persistent ambiguity in these provisions is the definition of ‘high level of protection’. Drawing inspiration from EU law, EU treaties and case law refer to this concept, especially in the environmental context, but do not provide a precise legal threshold. Article 3(3) TEU and Article 191 TFEU refer to a high level of environmental protection as a policy objective, but under EU law it has also been interpreted as a binding qualitative criterion that determines the content of legislation.[41] However, the Court has also affirmed that it does not require the highest technically possible level of protection.[42]

Following the general rule of interpretation of Article 31 VCLT, we should look at the ordinary meaning of ‘high’ which appears in the dictionary as means greater than the usual or average level,[43] but not necessarily the maximum. Thus, to decide what constitutes a high level of protection, an average level protection should be established. In the context of labour protection, the international standards referenced in TSD chapters, mainly the ILO core conventions, represent minimum, not high, levels of protection. These conventions establish a social minimum at the global level, not a high standard. The European Economic and Social Committee and the ILO itself have recognized these as minimum requirements, not high standards.

Establishing what constitutes an average level of protection in labour and environmental standards under TSD chapters is inherently difficult to determine across nearly 200 countries, efforts to collect data on the protection levels remain limited. Some tools, like the Sustainable Development Report (SDR)[44] and the ITUC Global Rights Index,[45] attempt to offer benchmarks, but their reliability and methodology vary, and they only partially capture the standards relevant under TSD chapters. The available data suggests that many countries perform below what could be considered a high level, and that expectations should differ based on countries’ income levels.

Environmental protection proves even more complex to assess due to the many factors and areas falling under the scope of ‘environmental protection’. Initiatives exist regarding certain aspects, such as environmental democracy or climate change performance of the countries which depend on publicly available date. In both areas, there is no clear or universally accepted definition of what an average level of protection entails. Without agreed-upon indicators or a global data infrastructure, the phrase remains vague and difficult to operationalize in legal or policy terms. As a result, unless treaty-makers clarify what ‘high level’means, this obligation risks remaining an empty commitment rather than an enforceable standard.

We are now turning our attention to what ‘improving domestic legislation’ means. All versions of the right to regulate article impose an obligation to improve domestic legislation. The first and third versions use ‘strive to continue to improve’ or ‘strive to improve’, creating an effort obligation: parties must show serious, ongoing attempts, but actual improvement is not required. Interestingly, as seen in Table 1 above, the second version uses ‘shall continue to improve’, creating a result obligation that is more easily enforceable: parties must demonstrate ongoing improvement, making it the strongest formulation in terms of legal effect.

The improvement obligation limits the right to adopt or modify domestic law: parties may do so only if it results in better protection for labour and the environment. However, the obligation is not without complications. There is no clarity on the intervals of improvement or when the obligation is satisfied. Can it be concluded that the improvement obligation is satisfied when the levels dictated by international agreements and standards are reached? If so, the separate formulation of the obligation to improve domestic legislation would be superfluous, as the same article already states that domestic levels of protection have to conform with international agreements and standards. It is unrealistic to expect endless improvement from the parties. Can the economic or technical capacity of a party at a given moment be considered while determining the limit of improvement attainable? This is yet another TSD obligation that needs concretization.

2.1.3.   Analysis

In general, clearer and more binding language is observed in the second version of the Right to regulate articles. Firstly, consistency with international standards and agreements is a condition of right to regulate; the parties can regulate the level of protection in their domestic laws as long as such levels are consistent with the international standards and agreements. Secondly, the parties are under the obligation to ‘continue to improve’ the domestic levels of protection. 

Recent FTAs increasingly favour the ‘strive’ language, reflecting a trend toward more robust effort obligations. This shift is motivated by the growing importance of sustainable development in trade policy, as evidenced by more extensive sustainability objectives in recent agreements. The third version, which appears to be the new standard, imposes a higher threshold for compliance for the effort obligations it established, requiring serious efforts rather than minimal attempts. This is consistent with the principle of effectiveness in treaty interpretation: different wording should reflect different obligations. It is noteworthy that the third version, which appears in all most recent FTAs, did not adopt the continuous improvement result obligation of the second version. This may be due to the practical challenges of defining and measuring continuous improvement, as well as the need to balance ambition with feasibility.

The practical effect of the obligation for domestic law to be consistent with international standards and agreements is significant: by incorporating international agreements and standards into the TSD chapters, a breach can constitute a violation of both the international agreement and the FTA itself, creating what the author terms ‘complementary jurisdiction’. Unlike other TSD obligations, the Right to regulate article does not require a trade or investment linkage, so a breach is covered even if it does not affect trade or investment between the parties.

The differences in formulation, especially the verbs used and the placement of key phrases, significantly affect the legal character of these obligations and what the parties undertake. The evolution of these versions possibly reflects the parties’ shifting priorities and the lessons learned from earlier agreements, as well as the influence of international legal developments and the growing emphasis on sustainable development in global trade governance.

Despite the evolution toward stronger language, the effectiveness of these provisions remains to be tested in dispute settlement, as the evidentiary burden for effort obligations is inherently challenging. The lack of clear standards for what constitutes ‘serious efforts’ or ‘striving’ leaves room for divergent interpretations, which may undermine the enforceability of the obligations.

The Right to regulate and levels of protection clauses in EU TSD chapters have developed into a complex set of obligations that balance regulatory autonomy with commitments to maintain and enhance labour and environmental standards. In general, these articles bring an obligation for the parties to ensure their domestic levels of protection are higher than the minimum standards set in international agreements and standards, to show efforts to improve such levels and to strive for high levels of protection. The specific formulation of these obligations, particularly the verbs used and the structure of the provisions, significantly influences their legal weight and practical enforceability. The trend in recent agreements toward stronger, more explicit effort obligations reflects a growing prioritization of sustainable development objectives in EU trade policy, even as interpretive uncertainties persist. 

2.2.  Upholding levels of protection

All EU Free Trade Agreements (FTAs) include provisions requiring the parties to uphold the level of protection in their domestic labour and environmental laws.

As an example, the relevant article in the South Korea FTA reads as follows: 

‘Upholding levels of protection in the application and enforcement of laws, regulations or standards 

1. A Party shall not fail to effectively enforce its environmental and labour laws, through a sustained or recurring course of action or inaction, in a manner affecting trade or investment between the Parties. 

2. A Party shall not weaken or reduce the environmental or labour protections afforded in its laws to encourage trade or investment, by waiving or otherwise derogating from, or offering to waive or otherwise derogate from, its laws, regulations or standards, in a manner affecting trade or investment between the Parties.[46]

These obligations, mostly articulated as separate articles entitled ‘Upholding levels of protection’, and sometimes combined with the Right to regulate article, are a core element of TSD chapters. The levelling of the playing field function of these articles is perhaps the most obvious among all the TSD articles. Regardless of their placement, the substance of these provisions is consistent: to prevent parties from lowering their standards or failing to enforce them as a means of gaining a competitive advantage in trade or investment. 

To do so, the Upholding levels of protection articles typically impose three main obligations:

1) Not to fail to effectively enforce domestic law (the effective enforcement clause);

2) Not to weaken levels of protection in domestic law (the non-regression clause);

3) In some FTAs, a separate obligation not to waive or derogate from domestic law.

At a prima facie reading, these obligations use some of the most binding language in the TSD chapters with the usage of ‘shall’.However, many other phrases and also the relationship between these obligations are in need of clarification to unveil what the content of the obligations is.

The intent is to preserve the protection already provided by domestic law: parties must not lower these standards and must ensure effective enforcement of the laws providing such protection. Both obligations are conditional on trade/investment linkage: they are triggered only when there is a connection to trade or investment, defined either as ‘in a manner affecting trade and/or investment’ or as ‘encouragement of trade and/or investment’. The implications of these two types of linkage are significant and will be analysed below.

2.2.1.   Obligation not to fail to effectively enforce (effective enforcement clause)

The effective enforcement clause typically reads as ‘A Party shall not fail to effectively enforce its environmental or labour laws, through a sustained or recurring course of action or inaction […]’.[47] This is then connected to a trade/investment linkage, which is the main point of divergence between the different FTAs of the EU. We will now analyse this obligation in detail. 

The issue to address first is how this obligation is named in the doctrine. The effective enforcement clause in EU FTAs obliges parties to effectively enforce their domestic law in labour and environmental matters, with varying trade/investment linkage. The terminology used to describe this obligation in doctrine varies. Some scholars refer to it as a ‘non-enforcement clause’.[48] Despite the appeal of its rhyme with ‘non-regression’, this is counterintuitive. Although the clause is phrased negatively (‘shall not fail to effectively enforce’), its content is a positive obligation: parties must ensure effective enforcement. The double negation (‘not fail to’)logically yields an affirmative duty. Thus, ‘effective enforcement clause’ better matches the substance of the obligation. This is contrary to the negative nature of the ‘non-regression’ obligation, where the parties promise not to do something; they shall refrain from lowering standards in their domestic laws. For a clause bringing an obligation not to regress, the name non-regression is logical. For an obligation bringing a positive obligation to do something, a name including ‘non’ is not logical. 

The effective enforcement obligation is always conditional upon a linkage to trade or investment, limiting its material scope. It is therefore crucial to analyze what this linkage entails. 

The related paragraph in the South Korea FTA, demonstrating an example of the first version, reads as follows: ‘A Party shall not fail to effectively enforce its environmental and labour laws, through a sustained or recurring course of action or inaction, in a manner affecting trade or investment between the Parties’.[49]

Demonstrating an example of the second version of the effective enforcement clause, the related paragraph of CETA reads as follows: ‘A Party shall not, through a sustained or recurring course of action or inaction, fail to I effectively enforce its labour law and standards to encourage trade or investment’.[50]

The two versions of linkage, ‘in a manner affecting trade or investment’ and ‘in order to encourage trade or investment’, differ in the requirements to establish a breach. The former requires proof of actual effects, while the latter focuses on the intent of the parties. All the recent (draft) FTAs, for instance with Kenya and Mexico, favour the ‘encouragement’ formulation, which is arguably harder to prove as it necessitates demonstrating intent. 

The meaning of ‘in a manner affecting trade’ has never been addressed in an EU FTA context, which is necessary to analyse to understand the content of the obligation. The only dispute this date arising from a TSD chapter, the EU-South Korea dispute,[51]does not regard this matter. The phrase was, however, analysed in the US-Guatemala dispute under CAFTA-DR with the identical effective enforcement clause.[52] This is the only other global example of a dispute arising from social standards in a trade agreement and the only example of such a dispute being brought before a panel with dispute settlement powers.

The US alleged that, by not following court orders regarding the eight employees’ rights for taking part in union activities, Guatemala failed to effectively enforce its domestic labour law which put US workers at a competitive disadvantage.[53] While agreeing with the US on the substance of the violation, the panel held that it did not satisfy the legal requirements set by the effective enforcement clause. 

The panel held that a breach requires a ‘sustained or recurring course of action or inaction’ that confers a competitive advantage in trade, and insisted on proof of actual trade impact, proof it found the United States had not provided. Applying the VCLT rules of interpretation, the panel examined the ordinary meaning of the relevant terms and rejected the view that any failure to enforce in a traded sector would suffice, as that would render the trade nexus meaningless. In light of the agreement’s objective of promoting fair competition within the free trade area, the panel concluded that a measure affects trade only where its effects on competitive conditions can be shown to ‘confer some competitive advantage to an employer engaged in the trade between the parties’.[54] In doing so, it rejected the broader WTO Appellate Body interpretation of ‘affecting’, emphasizing the term’s different function in this context. To establish a breach, the panel identified three factors:[55]

1) Whether the enterprise in question exports to or competes with imports from the other party

2) The effects of the enforcement failure and whether these effects confer a competitive advantage to the enterprise

The panel’s narrow interpretation has been criticized for setting a high bar, making it difficult to address even systematic violations.[56] This approach contrasts with the broader interpretation of the similar phrase of ‘trade-related’ in a different article in the EU-South Korea dispute, where the Panel of Experts (PoE) held that national measures implementing labour rights are inherently trade-related unless the FTA explicitly limits their application.[57] Morris stated that difference in approach between the two panels, which has been commented as being in stark contrast, demonstrates the two ends of interpretive spectrum available to future panels.[58] However, the PoE recognized that the Upholding levels of protection article explicitly includes a trade linkage, meaning its broader interpretation does not apply there.[59] Thus, future PoEs dealing with a dispute arising from this clause may be more likely to follow the narrow approach of the US-Guatemala panel, also considering that ‘affecting’ is by nature a more limited concept than ‘being related to’.

The other important point regarding enforceability is that ‘in a manner affecting trade’ demands proof of actual, tangible effects. This is also in line with the CAFTA-DR report. In regards to the obligation at hand, the panel concluded that, intentionality cannot be read into the provision. The letter of the clause does not suggest that the parties must have intended the failure to enforce to affect trade or investment. An effect can arise without the party’s knowledge or intent. Reading intentionality into this version would erase the key distinction between the two formulations and undermine the treaty-makers’ choices. This is also in line with VCLT rules of interpretation; object or purpose of a treaty cannot be used to rewrite the letter of a treaty rule.Once again, meaning of certain phrases used in the clause must be addressed to understand the content of the obligation. We will first look at ‘in a manner affecting investment between the parties’. Unlike CAFTA-DR, EU FTAs often require an effect on both trade and investment. The choice between ‘or’ and ‘and’ as a connector is critical: ‘or’ requires an effect on either, while ‘and’ requires both. The latter, as used exceptionally in the Vietnam FTA,[60] sets a much higher bar for establishing a breach. This is point where one wonders where this difference comes from if the treaty-makers are making a conscious choice of word. 

Trade and investment differ in aims, scope, and legal regimes. The exporting country aims to protect its investors, which is not a consideration in trade agreements. The asymmetrical relationship they enter into the host-state expose investors to greater risks,[61]including the power discrepancy involved in regulatory risks.[62] Many arbitral panels have awarded compensation to investors for changes to domestic law that were not anticipated at the time of the investment.[63]

As DiMascio and Pauwelyn put it ‘foreign products remain foreign, whereas foreign investments become local’.[64] If a trade deal fails, the goods can likely be sold elsewhere. If investment fails, the capital is most likely gone. Consequently, decision to make an investment involves many factors such as a good business environment, low level of corruption, political stability, market uncertainty. A failure to enforce domestic law ‘in a manner affecting investment’ can include actions that affect one of these factors. A failure to effectively enforce the domestic law can lower investor confidence, result in expropriation of investments, violate investor rights, or introduce regulatory barriers. 

Unlike trade, potential effects on investment even before the investment takes place, such as deterring future investment may suffice to establish a breach, depending on the investment relationship between the host and origin state. This also aligns with EU law on the free movement of capital. Measures having potential effects rendering the movement of capital less attractive or can make potential investors change their minds on investing constitute restrictions on the free movement of capital and, as such contrary to EU law.[65] The series of golden share cases and the CJEU’s interpretations can be shown as examples of this.[66]

Due to the nature of investment and the inherent differences between trade and investment, the two should be handled differently. In light of the discussion above, unlike ‘in a manner affecting trade’ requiring showing of actual effects; ‘in a manner affecting investment’ does not require showing of actual effects, potential effects should suffice for the linkage.

This renders the versions of effective enforcement clause with ‘trade or investment’ in the required linkage are easier to enforce. Where showing of actual effects on trade is difficult, showing of potential effects on investment can be an avenue to enforce this obligation. 

Recent FTAs often use the phrase ‘as an encouragement for trade or investment’. An example of this version can be found in Mexico FTA which reads as follows: ‘A Party shall not, through sustained or recurring course of action or inaction, fail to effectively enforce its environmental or labour laws in order to encourage trade or investment’.[67]

This version focuses on intent:[68] a breach occurs only if the failure to enforce is intended to stimulate trade or investment. Showing of actual effects is not necessary to establish a breach.[69]

The dictionary meaning of encouragement involves intentionality. Thus, only deliberate failures to enforce, aimed at making trade or investment more appealing. This is another key difference between the two versions of trade/investment linkage. Intentionality is not required in the ‘in a manner affecting’ version, thus incidental failures to effectively enforce occurring outside of one party’s intention or knowledge are also covered. 

While some FTAs use the formulation ‘in order to encourage’, the meaning is similar. The prepositions regard the relationship between the encouragement to trade/investment and the failure to effectively enforce domestic law. The usage of ‘as’ and ‘to/ in order to’, suggests that the failure to effectively enforce should be the way the encouragement for trade or investment is given. Thus, sustained, and recurring action or inaction that leads to not effectively enforcing domestic law should have the intent to make trade and/or investment more appealing. This means that failure to effectively enforce will not fall within the scope of this article unless it is intended to encourage trade or investment.

The formulation of ‘as an encouragement’ instead of ‘the encouragement’ suggests that not effectively enforcing the domestic laws can be a part of a series of actions to stimulate trade and/or investment.[70] Additionally, the intention to encourage does not need to be aimed at the bilateral relation between the EU and the FTA partner but can be directed at a general stipulation of trade and investment to the country in general.[71] This version has drawn criticism for excluding situations where domestic law is not effectively enforced for reasons unrelated to trade or investment, even if the lack of enforcement ultimately impacts them.[72]

This standard makes a breach more difficult to establish as it required demonstrating the parties’ intent to encourage certain conduct. In this sense, it is unfortunate that the encouragement version has become the standard in the recent FTAs. The effective enforcement obligation is triggered only by a ‘sustained or recurring course of action or inaction’ which satisfies the trade/investment linkage. The interpretation of this requirement, particularly as developed by the CAFTA-DR panel in the US–Guatemala dispute, is useful to understanding the threshold for a violation and the evidentiary burden in practice, as a possible interpretation.

The CAFTA-DR panel provided a detailed analysis of what constitutes a ‘sustained or recurring course of action or inaction’.The panel began by clarifying that both ‘sustained’ and ‘recurring’ qualify the phrase ‘course of action or inaction’. Thus, a breach requires not just any failure to enforce, but one that is either sustained, i.e. continuous, prolonged, maintained without disruption or recurring, i.e. periodic, repeated, occurring again. The use of ‘or’ means that either condition suffices.[73]

Importantly, the panel emphasized that ‘course’ must mean more than just a collection of actions or inactions.[74] The behaviour must be related, forming a pattern or line of conduct, rather than a series of isolated incidents. The similarity must be non-random and connected, whether by time, place, or type of conduct. The panel concluded that ‘sustained or recurring course of action or inaction’ consists of either:[75]

a) repeated behaviour with sufficient similarity (recurring) or 

b) prolonged behaviour with enough consistency in acts or omissions to form a connected pattern (sustained).

Isolated failures to enforce, even if repeated, are not enough unless they are linked by similarity or continuity and remain mere instances of sustained or recurring failure to enforce but do not constitute a ‘course of action’.

A significant issue addressed by the panel was whether the term ‘course’ implies deliberateness or intentionality. Guatemala argued that a ‘course’ of action or inaction should be understood as a deliberate policy. The panel rejected this, holding that such an interpretation would add words which are not present in the text and would conflict with the treaty’s rules of interpretation.[76] The panel’s reasoning relied on a specific paragraph in the effective enforcement clause of CAFTA-DR, which allows parties to exercise discretion in enforcement and resource allocation. If intentionality were required, it would be impossible for a party to justify its actions under this discretion clause, defeating its purpose.

The CAFTA-DR panel’s interpretation of ‘sustained or recurring’ brings the EU’s general and persistent breach (GAP) doctrine to mind. Traditionally, the European Commission brought infringement proceedings for individual breaches of EU law, which has been criticized.[77] In the Irish Waste case, it bundled twelve incidents of non-compliance with the Waste Directive into a single proceeding for general and persistent infringement. The CJEU accepted this approach, holding that a general practice of non-compliance, illustrated by specific cases, could establish a breach.[78]

The CJEU applies a higher threshold of proof in GAP cases than for individual infringements.[79] To prove a general and persistent breach, certain conditions must be met regarding duration, scale, and seriousness. The Advocate General has clarified that ‘general’ does not require nationwide scope; a regional pattern suffices.[80] Duration is also flexible: even four years have been found sufficient at times, though longer periods are more persuasive.[81] The number of incidents, i.e. scale, matters, but the key is that they are related and form a consistent practice. In the Irish Waste case, twelve incidents over twenty-five years were deemed sufficient to show a large-scale administrative problem, while in Finnish Wolfs case two incidents were not.[82]

In addition to being reluctant to bring GAP cases, the Commission has sometimes also struggled to prove the required pattern, especially where incidents are isolated or already remedied. Wennerås has warned that a preference for pursuing individual breaches over general patterns may undermine the effectiveness of the GAP instrument.[83]

While the GAP doctrine and the TSD effective enforcement clause use different terminology (‘general and persistent infringement’ versus ‘sustained or recurring action or inaction’), they share key features: both require a pattern of related, prolonged, or repeated conduct, not isolated incidents. Both also require a temporal aspect, prolongation or repetition over time.

Applying this reasoning of EU law to TSD chapters, ‘sustained or recurring action or inaction’ should be understood as requiring a pattern of failure to enforce domestic law on labour or environmental standards over a period of time, in addition to the necessary linkage with trade or investment. The requirement of a pattern ensures that only serious, systemic failures are actionable, not isolated lapses.

The last phrase that needs clarification is effective enforcement itself. In the labour dispute between the US and Guatemala, the CAFTA-DR panel also clarified what it means to ‘effectively enforce’ domestic law. Both parties to the dispute agreed that effective enforcement means being ‘productive of results’ and achieving compliance. The US argued that effective enforcement requires authorities to end conduct contrary to law and attain compliance; any non-compliance implies failure. The panel, however, rejected the idea that every instance of non-compliance is a breach, recognizing that perfect compliance is unrealistic.

The panel explained that enforcement is inherently discretionary, with varying levels of possible enforcement. The drafters of CAFTA-DR used ‘effectively’ to indicate that mere formal enforcement is insufficient to achieve the treaty’s objectives.[84] The key is the production of results, as evidenced by general compliance with the law.[85] Authorities must act to remedy non-compliance, and there should be an expectation among employers that the law will be enforced.[86] If such an expectation is absent, this suggests a failure to enforce effectively.

The panel summarized the obligation as one ‘to compel compliance with labour laws in a manner that is sufficiently certain to achieve compliance that it may reasonably be expected that employers will generally comply with those laws, and employers may reasonably expect that other employers will comply with them as well’.[87] This interpretation aligns with the ordinary meaning of ‘effective’ as producing the intended result. In legal contexts, this naturally relates to compliance.[88]

This interpretation of the CAFTA-DR panel might be followed by future panels. It is important to note, however, that the CAFTA-DR panel’s analysis was specific to labour law. In TSD chapters, effective enforcement obligations apply to both labour standards and environmental protection, which differ in important ways. Labor standards typically involve a narrower set of actors, employers and regulators. Environmental protection, on the other hand, encompasses a broader range of actors across many sectors on both addressees and the regulator's side, from ports and airlines to farmers and factories, and multiple ministries or government agencies. It is unclear whose compliance or whose ensuring of compliance should be looked at.

This diversity means that the CAFTA-DR panel’s approach to labour enforcement cannot be taken directly to environmental law and requires a mutatis mutandis application. For environmental protection, effective enforcement should still be measured by whether authorities ensure compliance and whether there is a general expectation among regulated parties that the law will be enforced. However, the relevant actors and expectations will vary by sector and context and should be looked at separately for each particular area of the environment.

In practice, effective enforcement of environmental law requires the private sector to comply with applicable regulations and public authorities to monitor, inspect, and sanction non-compliance. Mechanisms may include permits, inspections, fines, or other enforcement tools. The expectation among regulated parties should be that most actors comply and that authorities will act against non-compliance.

In conclusion, the obligation not to fail to effectively enforce domestic law in TSD chapters is triggered only by a sustained or recurring course of action or inaction, linked to trade or investment. The CAFTA-DR panel’s analysis provides a clear framework: the pattern must be connected and non-random, and the linkage to trade or investment may or may not require intentionality, depending on the treaty language. The EU’s GAP doctrine offers a useful parallel in understanding the required pattern and duration. Effective enforcement, whether in labour or environmental law, is ultimately about achieving compliance and creating an expectation of compliance among regulated parties. This nuanced and demanding standard ensures that only serious and systemic failures are actionable, balancing the need for robust enforcement with the realities of regulatory discretion and complexity.

2.2.2. Obligation not to reduce levels of protection (non-regression clause)

This section analyses the non-regression clause in the TSD chapters of EU FTAs. The discussion begins with the general concept and rationale behind non-regression clauses, then examines their specific operation in TSD chapters, including interpretive questions and their relationship with the obligation not to waive or derogate from domestic law.

Non-regression clauses, now recognized as an emerging principle of environmental law[89], prohibit parties from lowering their domestic protection levels. One of the earliest examples in a trade context is NAFTA of 1992. Since then, such clauses have become increasingly common in trade and investment agreements, with varied formulations.[90] The EU has embedded non-regression principle in many of its secondary environmental and labour legislation.[91]

The main aim of non-regression clauses in the trade agreements is to prevent parties from lowering environmental and labour standards to gain a competitive edge in trade. The EU’s rationale for including them in the TSD chapters is often explained as the use of its market power to promote global sustainable development, in line with its Lisbon Treaty mandate. Scholars such as Mitchell and Munro highlight the EU’s normative motivation, contrasting it with the US approach, which is more focused on preventing competitive disadvantage.[92] Nevertheless, maintaining a level playing field is a still key driver behind TSD chapters: trade agreements should not create competitive disadvantages through divergent sustainability standards.[93] The non-regression clause thus acts as a safeguard against a ‘race to the bottom’ in environmental and labour regulation.

This dual objective is especially clear in the non-regression clause: the EU seeks to avoid ‘a compromises on core principles like human rights and sustainable development around the world’,[94] while also ensuring that domestic producers are not unfairly disadvantaged by higher standards than those of their trading partners.[95] The latter was particularly emphasized in the EU-UK negotiations,[96] where the TSD chapters have been called the ‘level playing field’ provisions in the Trade and Cooperation Agreement (TCA).[97] Durán notes that the consistent linkage to trade in these clauses signals their main purpose: preventing competitive pressure from regulatory races to the bottom.[98]

The European Commission’s 2006 Global Europe strategy announced the inclusion of non-regression clauses in future FTAs to address environmental and social issues alongside economic ones.[99] Since the South Korea FTA, non-regression clauses have become standard in EU FTAs, and are also used by other trading actors in their agreements such as Brazil, Japan, and others.[100]Non-regression clauses complement other TSD chapter provisions, such as the Right to Regulate and multilateral standards articles, by prohibiting actions that, while consistent with international obligations, still lower domestic protection levels. In this way, non-regression establishes a higher, autonomous standard. Thus, even if the level of protection in domestic law remains higher than the one established in international agreements, regression is still prohibited. 

Now we turn our gaze to the content of the obligation. The non-regression obligation prohibits parties from lowering their domestic environmental and labour protection levels, provided the required trade or investment linkage is met. The first step is to clarify the scope of domestic law covered. 

Under VCLT, ‘the natural and ordinary meaning of terms is not to be determined in the abstract but by reference to the context in which they occur’.[101] The immediate context relates to the ‘grammatical construction of the provision’ that the term appears, [102]which also include the headings of articles.[103] In the South Korea FTA, the article is titled ‘Upholding levels of protection in the application and enforcement of laws, regulations or standards’, and uses these terms interchangeably. This suggests that the scope is not limited to ‘laws’ in the strict sense, i.e., statutes passed by parliament, but also includes regulations and standards, which may be set by the executive.

Other FTAs use varying formulations: some refer to ‘laws and policies’,[104] others to ‘laws and standards’,[105] or simply ‘laws’.[106] CETA, is the only FTA that provides a definition of ‘environmental law’ to include both laws and regulations.[107]Where ‘standards’ or ‘policies’ are mentioned, the scope extends to non-binding instruments as well. The rationale for these differences is unclear. 

The non-regression clause is among the strongest in TSD chapters, using ‘shall’ and imposing a result obligation: parties must not lower protection levels. This signals the importance of preserving domestic standards and can support a broad interpretation of what falls under domestic law. Nevertheless, the scope of the covered domestic law will need to be determined on a case-by-case analysis by a future PoE.

Non-regression clauses appear in different forms across FTAs, all using binding language. The main difference is whether the obligation is to ‘not weaken or reduce’ protection, ‘not waive or derogate from’ domestic law, or a combination. These appear in four main versions:

a) Only an obligation not to weaken or reduce levels of protection.

b) Separate obligations not to weaken or reduce protection and not waive or derogate from domestic law.

c) Only an obligation not to waive or derogate from domestic law.

d) An obligation not to weaken or reduce protection by waiving or derogating from domestic law.

We will first look at option (a). Here, parties undertake not to lower protection levels in domestic law. Changes can only maintain or improve standards. This limits regulatory autonomy and interacts with the Right to Regulate article, which sets minimum standards and encourages improvement. The non-regression clause prevents backsliding below established levels.

An example of the first version reads as follows: 

‘A party shall not weaken or reduce the levels of protection afforded in its environmental or labour law in order to encourage trade or investment’.[108]

To determine if a change constitutes regression, one must identify the what the level of protection afforded is at the first place. This is straightforward for quantitative standards, for instance with emission targets, bringing the famous Urgenda case to mind. but more complex for qualitative ones. 

This version’s scope is narrower than those that include a separate obligation not to waive or derogate from domestic law, as it covers only legal changes, not deviations in practice, unless ‘policies’ are included, which is rare. As only ‘laws’ are covered, the breach shall take place once the legislation is adopted according to the procedures defined under the national law. 

The version of option (b) adds a distinct obligation not to waive or derogate from domestic law next to the obligation to not to weaken or reduce the level of protection, covering both legal regression (changes to law) and regression in practice (deviations from law without formal change). 

In the most recent, FTAs this relationship is quite straightforward: these two concepts are regulated in separate articles as two separate obligations: 

3. A Party shall not weaken or reduce the levels of protection afforded in its environmental or labour law in order to encourage trade or investment.

4. A Party shall not waive or otherwise derogate from, or offer to waive or otherwise derogate from, its environmental or labour law in order to encourage trade or investment’.[109]

‘To waive’ means voluntarily relinquishing a legal right or refraining from enforcing a rule, to refrain from pressing or enforcing a claim or a rule,[110]‘to derogate’ means ignoring[111] or limiting the application of a law. [112] These terms, used as verbs, are not as strictly defined as their legal noun counterparts.

Waiving covers situations where authorities do not enforce a rule in a particular case by for instance not requiring an environmental impact assessment where the law prescribes it. Derogating means ignoring or limiting the application of a rule, not repealing it. Thus, the obligation not to weaken or reduce the level of protection prohibits legal regression, while the obligation not to waive or derogate prohibits practical regression, deviations from law without formal change.

This is closely related to the obligation of effective enforcement, under which states commit to ensuring compliance with the law and holding private parties accountable for breaches, thereby acting as intermediaries in achieving compliance. The present clause, by contrast, concerns the conduct of the state itself, which must comply with the law and refrain from deviations unrelated to enforcement. The fourth paragraph functions as a catch-all provision, intended to cover all forms of legal deviation, regardless of their nature. Accordingly, this version, now standard in recent EU FTAs, offers the broadest safeguard against regression.

Following option (c), Central America, Vietnam and Singapore FTAs and CETA only have an obligation to waive or derogate from domestic law regulated and not an obligation not to reduce the levels of protection. That version reads as follows:

A Party shall not waive or otherwise derogate from, or offer to waive or otherwise derogate from, its environmental and labour laws, in a manner affecting trade or investment between the Parties’.[113]

This raises the question: are changes to law that lower protection levels allowed under this version? As explained above, the ordinary meaning of ‘to waive’ relates to non-application of law, not its amendment or repeal.

One interpretation is that ‘otherwise derogate from’ could cover legal changes, but this would firstly mistakenly depend on the meaning of the legal concept ‘derogation’ and not the ordinary meaning of the word ‘to derogate’. Secondly, accepting such an interpretation would make the explicit ‘not to weaken or reduce levels of protection’ obligations redundant. 

The more logical view is that this version covers only practical regression (deviation from law as it stands), while legal regression is not prohibited unless explicitly regulated. This is also supported by the fact that this version always starts with stating that it is inappropriate to lower levels of protection. If the parties wanted to make this into a binding obligation, they would have done so. Thus, under this version, changes to domestic law that reduce protection are allowed, provided they do not fall below international standards which would be caught by the Right to Regulate article.

The version of option (d) links weakening protection to waiving or derogating, making the scope of waivers conditional on their effect on protection levels.[114] The non-regression clause in this form reads as follows: 

‘Party shall not weaken or reduce the environmental or labour protections afforded in its laws to encourage trade or investment, by waiving or otherwise derogating from, or offering to waive or otherwise derogate from, its laws, regulations or standards […]’.[115]

If waiving or derogating does not reduce protection, it is not covered. This limitation is unique to this version.

All versions of the non-regression clause include a linkage to trade or investment. Not all lowering of the levels of protection or waiving or derogating will constitute a breach, only those that ‘affect trade or investment’ or are done ‘as an encouragement for trade or investment’ are covered.

The analysis made above on what to understand from or ‘as an encouragement for trade or investment’ under the effective enforcement section applies here as well and will not be repeated. Accordingly, the ‘in a manner affecting trade or investment’formulation requires proof of actual effects. The ‘as an encouragement for trade or investment’ version focuses on intent, requiring proof that the action was motivated by a desire to foster trade or investment, but not necessarily actual effects. The intention aspect is the most obvious in the text of the non-regression clause in Kenya FTA which reads as ‘with the intention to encourage trade or investment’.[116]

All the recent (draft) FTAs, for instance with Zealand, Kenya, Australia, Mexico, India, Indonesia and Chile, have moved towards the ‘encouragement’ formulation, possibly in response to enforcement difficulties highlighted by the US-Guatemala dispute, which interpreted ‘affecting trade’ narrowly. The shift suggests a preference for a standard that is easier to enforce.

Concluding this section, it is important to underline that there is a close, two-fold relationship between these articles, both of which address domestic labour and environmental protection. The Right to regulate article focuses on the substance of domestic law, requiring consistency with international agreements, high levels of protection, and ongoing improvement. In contrast, the non-regression clause prevents lowering established domestic standards, even those exceeding international requirements, provided a trade or investment linkage exists. Notably, the Right to regulate article is not subject to this linkage, making enforcement of minimum international standards more straightforward.

The non-regression clause is most significant when domestic standards surpass international ones, as these higher standards are not otherwise protected. While the Right to regulate article sets and encourages minimum standards, the non-regression clause prohibits backsliding, and the effective enforcement obligation ensures these standards are actively maintained. Together, these provisions create a comprehensive framework for upholding and enforcing domestic protection levels.

3.   Farewell Green Deal: The retreat of EU environmental law

There is an upwards trend in including commitments regarding non-trade-related values such as environmental protection in free trade agreements. While not alone, the European Union is arguably leading in this endeavour, having included Trade and Sustainable Development chapters in all its FTAs since 2011, without exception, with an ever-widening scope. This can be shown as yet another example of the EU ‘exporting its norms’,[117] using its ‘normative power’[118] or exercising the ‘Brussels effect’.[119] Truly, the EU is still seen as a pioneer in the fight against climate change worldwide with its ambitious internal legislation and policies, such as the European Green Deal and its high ambition to become the first climate neutral continent by 2050. 

The last few years of the EU climate policy have been, however, anything but easy or free of conflicts. Many of the reforms initially planned or adopted have been weakened in the face of the shift in the political winds in Europe and intense backlash on certain pieces, such as the farmer protests across the EU.[120] This context led to the weakening of long-awaited proposals for new pieces of legislation, relaxation of some rules that are already in force or postponing the entry into force of others. Some examples are worth mentioning: 

a) The revised Renewable Energy Directive (RED III), adopted in October 2023, introduced new ‘renewables acceleration areas’and other special zones where renewable energy and related infrastructure projects are exempted from environmental impact assessments, provided certain mitigation measures are in place. This change prioritizes rapid deployment of renewables, but at the cost of bypassing some environmental safeguards.[121]

b) In February 2024, the European Commission announced it would withdraw its proposal for the Sustainable Use Regulation (SUR).[122]

c) In May 2024, the EU institutions agreed to relax the CAP rules on environmental protection, through an Implementing Regulation for a derogation on the application of the standard for Good Agricultural and Environmental Conditions (GAECs) of land, bringing changes to six of the nine GAECs standards. For instance, it removes the minimum 4% allocation of arable land at farm level to non-productive areas in order to receive subsidies, relaxes the monitoring and controls and exempt some farmers.[123]

d) In July 2024, the ‘watered-down’ Nature Restoration Regulation[124] was adopted amid intense criticism.

e) In October 2024, The Commission announced its proposal to delay the implementation of EUDR for 12 months.[125] The European Parliament accepted the proposal on the 14^th of November 2024. Additionally, an amendment introduced a ‘no-risk’country category, allowing the European Commission to exempt certain countries from strict checks if they demonstrate a net increase in forest cover, even if that increase is due to monoculture plantations, which can harm biodiversity, which has been criticized as adding a loophole.[126] On December 19^th 2025, the Regulation that delayed the application of EUDR for a second year in a row was published in the Official Journal.[127] The Commission had already eased some reporting requirements and plans to exempt most countries from the most stringent inspections.[128]

f) The Omnibus Simplification Package, introduced by the European Commission in February 2025,[129] brings substantial changes to previously adopted EU environmental and sustainability legislation. Its primary aim is to reduce regulatory burdens and reporting requirements for companies, especially in the context of sustainability reporting and due diligence.

g) The Omnibus Package, adopted by the European Commission in February 2025, introduces significant changes to the EU’s Carbon Border Adjustment Mechanism (CBAM) aimed at simplifying compliance, reducing administrative burdens, and enhancing competitiveness. In addition to delaying the full operationalization of CBAM with sales of certificates,[130] Omnibus Package exempts approximately 90% of current importers from CBAM obligations. While the Commission asserts that 99% of embedded emissions will still be covered, this change means that a vast majority of companies, especially small and medium-sized enterprises (SMEs), are no longer subject to CBAM’s carbon pricing and reporting requirements.[131]

h) Under the Omnibus Package, The European Parliament approved the ‘stop-the-clock’ Directive on April 3, 2025, which delays the application of the Corporate Sustainability Reporting Directive (CSRD) and the Corporate Sustainability Due Diligence Directive (CSDDD, also known as CS3D) for most companies.[132]

i) On the 16^th of December 2025, the European Parliament approved the changes that the Omnibus Package proposed to narrow the scope of these directives. While maintaining the original proposal’s 1,000-employee threshold for the CSRD, the agreement introduced an additional criterion excluding companies with annual revenues below €450 million, thereby removing an estimated 90% of firms from the sustainability reporting obligations.[133] Similarly, the threshold for the CSDDD has been increased to 5,000 employees and €1.5 billion in revenue, thereby excluding the vast majority of companies. 

l) On 27 October 2022, the European Union had reached an agreement in Brussels on legislation effectively banning the sale of new petrol and diesel cars from 2035, with the aim of accelerating the transition to electric vehicles and addressing climate change. In December 2025, The European Commission has formally proposed reversing the planned ban on the sale of new petrol and diesel cars from 2035, replacing it with a 90% CO₂ reduction target for new vehicles by that year rather than a full 100% ban, meaning some ICE vehicles could still be sold after 2035.[134]Under the revised plan, plug-in hybrids, mild hybrids and vehicles using low-carbon or synthetic fuels could continue to be sold beyond 2035 alongside EVs.

Debated heavily within the EU, the external aspect of these actions is absent in the discussions. As the EU is often seen as the ‘exporter’ of the TSD chapters insisting that they are included in all of its FTAs following its template,[135] the other side of the medallion is often forgotten: the TSD chapters are just as binding on the EU as they are on the FTA partner. Besides possible breaches to its multilateral obligations, the EU has bilateral obligations regarding its domestic law on sustainable development, arising from the TSD chapters. This article has shown that under its FTAs the EU has undertaken several obligations regarding its domestic law, ranging from upholding its level of protection to improving its domestic law under the Right to regulate articles and obligations to effectively enforce, not to regress and not to waive or derogate from its domestic law regarding the environment. Despite presenting itself as an ambitious global leader with high standards in labour and environmental protection, the EU’s compliance with these obligations is not guaranteed. Particularly, the change in the political situation in the EU and consequent recent trend in either delaying or weakening legislation adopted under the European Green Deal brings this compliance in doubt. 

It is worth noting that, the FTAs until the Singapore FTA of 2019 are signed as mixed agreements; trading partners on one side and the EU and its member states on the other side. Following the Opinion 2/15 of the CJEU,[136] the FTAs, including and after the EU-Singapore FTA, are signed solely by the EU based on its exclusive competence of the Common Commercial Policy. For the mixed agreements, as they are also parties, questions also arise regarding the compliance of individual member states and how the responsibility is shared between the EU and its member states. This aspect is outside the scope of this article which focuses on the compliance of the EU alone and left for future research.

4.   How the EU is in breach of its bilateral obligations under the TSD chapters

With this in mind, this section will analyse the recent changes made in the EU environmental legislation and policies in light of the TSD obligations regarding the parties’ domestic laws. Obligations under the two articles of the TSD chapters dealing with domestic law will be looked at separately: Right to regulate and Upholding levels of protection.

4.1.  Obligations under the Right to regulate article

The EU is seen as a pioneer in global sustainability governance, with ambitious legislation within its borders and strong advocacy for higher standards globally.[137] As the multilateral standards and agreements set minimum levels which the EU legislation typically goes beyond, it is unlikely that the recent actions of the EU will be caught by the first obligation of domestic law being consistent with international agreements and standards. This is also the case for second obligation of striving for a high level of protection under the Right to regulate articles, if one can assume that the EU standards go beyond the ‘average’ level of protection globally, even in their watered down form. The limitations in establishing what is a high level, as discussed above, remains valid.

A case can, however, be made under the third obligation to improve the levels of protection in domestic law. This is definitely the case under the second version, which states that the parties ‘shall improve’ the domestic levels, obliging them to a clear, binding result obligation. Delaying an already adopted piece of domestic law which would have improved environmental governance, such as due diligence directives or would have applied stricter norms such as the ones under EUDR, is certainly not improving such levels. Accordingly, delaying the implementation of previously adopted measures aimed at strengthening environmental protection constitutes a breach of the obligation to improve domestic levels of protection.

Even under the other versions of this obligation, where the parties undertake only an effort obligation, the EU’s recent streak of postponing or watering down its environmental legislation should be considered a breach of this obligation. Relaxing or postponing measures that would raise levels of environmental protection not only fails to advance improvement but actively undermines it, and therefore cannot satisfy, even at a minimum, the effort required under either the ‘seek to ensure’ or the ‘strive for’ versions of the obligation. 

Thus, a trading partner could trigger the TSD dispute settlement mechanism and bring a claim against the EU in front of a PoE for breaching its obligation to improve domestic law regarding the labour and environment, under any of the FTAs with a TSD chapter. 

4.2.  Obligations under upholding levels of protection article

Before analysing the three different obligations under the Upholding levels of protection articles separately, a common element should be addressed first.

As explained above, unlike the ones under the Right to regulate articles, all the obligations under the Upholding levels of protection articles require a trade or investment linkage. Under different FTAs, such linkage takes the form of the breach being either‘in a manner affecting trade or investment’ or ‘as an encouragement for trade or investment’. In the context of the EU’s recent legislative changes, this trade or investment linkage is not only present but also central to the rationale and public justification for the reforms.

The EU’s own framing of the Omnibus Package,[138] the Competitiveness Compass[139] and related legislative amendments repeatedly emphasizes the need to reduce administrative burdens and enhance competitiveness for European businesses, particularly SMEs.[140] For example, the introduction of a mass-based de minimis threshold under the CBAM, which exempts approximately 90% of importers from compliance obligations, is justified on the grounds of alleviating regulatory pressure on SMEs and facilitating their participation in global markets. Similarly, the relaxation of reporting and due diligence requirements under the CSRD and CSDDD is presented as a measure to ensure that European companies remain competitive in international trade, avoiding what is described as a disproportionate compliance burden that could disadvantage them vis-à-vis non-EU competitors. 

This logic is evident in the communication for the European Commission, as it states, ‘In order to boost competitiveness and unleash growth, the EU needs to foster a favourable business environment and ensure that companies are not stifled by an excessive regulatory burden. This, in turn, will unlock investments and enable companies to embrace the transition to a sustainable economy in a more effective and pragmatic way, ultimately meeting our climate and other sustainability goals’.[141] The legislative delays and scope reductions are not presented as isolated technical adjustments but as necessary responses to industry and state concerns about the impact of environmental regulation on trade flows and investment attractiveness.

The Commission sees regulatory obstacles as ‘a significant burden on the competitiveness of the European economy’, and the Omnibus Package is designed to ‘strengthen competitiveness and reduce bureaucratic hurdles, through greater efficiency, less regulatory complexity and accelerated administrative processes’.[142] The Commission further states that this will ‘unlock opportunities in European investment programmes, which will be vital for long-term growth and competitiveness’.[143]

The Commission’s willingness to postpone implementation deadlines, raise reporting thresholds, and allow for standard emissions values all serve to lower the cost and complexity of exporting to and importing from the EU, directly addressing the concerns of economic actors engaged in cross-border trade. The explicit goal of ’enhancing competitiveness while maintaining environmental objectives’ and to ‘level the playing field’ further underscores the trade linkage: the reforms are intended to make the EU market more attractive for investment and to prevent the relocation of businesses or supply chains to jurisdictions with weaker standards.

In sum, the trade or investment linkage requirement of TSD non-regression clauses is clearly satisfied in the case of the EU’s recent legislative actions. The reforms are motivated by, and directly aimed at, facilitating trade and investment by lowering regulatory barriers and compliance costs. This is not only evident in the substance of the legislative changes but also in their stated purpose and the political context in which they were adopted. 

Reactions from a wide range of stakeholders, governments, civil society, and industry underscore that the EU’s recent environmental legislative changes have immediate and significant effects on trade, further reinforcing the trade/investment linkage required by TSD non-regression clauses. Brazil has been outspoken in warning that the EU Deforestation Regulation (EUDR) and related measures could jeopardize billions of dollars in agricultural exports to the EU. Brazilian officials have argued that the EUDR’s requirements, deadlines, and compliance costs threaten to exclude Brazilian products, particularly beef, soy, and coffee, from the EU market, and have repeatedly called for delays and a reassessment of the law’s impact on trade relations.[144] This concern is not isolated: it is echoed by other Mercosur countries[145] and the USA;[146] and has become a central issue in the protracted negotiations over the EU-Mercosur Association Agreement, with environmental safeguards and their trade impacts at the heart of the debate.[147] Civil society organizations and producer groups have also highlighted the trade effects of the EU’s new rules, particularly on small scale farmers.[148]

Within the EU, some stakeholders have argued that the Omnibus Package’s weakening of sustainability reporting and due diligence rules could actually disadvantage companies that have already invested in sustainability. By lowering the bar for compliance, less ambitious firms may gain a competitive edge, while those that have prioritized sustainability could find themselves at a disadvantage, thus distorting competition within the single market and in global trade.[149] This ‘race to the bottom’ effect is cited as a risk not only to environmental integrity but also to the EU’s reputation as a leader in sustainable trade.

NGOs such as Human Rights Watch have also criticized the delay of the EUDR, arguing that it ‘punishes all the companies and EU trading partners who deployed efforts and resources to comply with the EUDR on time’, and sends a signal of regulatory unpredictability that can disrupt established trade flows and investment decisions.[150] There is no doubt that the changes to EUDR scope and delaying its application have effects on the trade between the EU and the trading partners with whom the EU has an FTA.

The ‘as an encouragement for trade’ version of the trade linkage, is also satisfied by the EU’s recent legislative changes, based on the EU’s own reasoning and public communications. As stated above, the EU’s Omnibus reforms and related legislative changes are explicitly justified as measures to ‘cut red tape’, ‘simplify business environment’, and ‘regain competitiveness’ for EU companies in global markets.[151] The focus on reducing administrative burdens for SMEs, raising thresholds for sustainability reporting, and delaying or weakening due diligence and environmental obligations are all presented as ways to make the EU a more attractive place for business and to prevent companies from being disadvantaged in global trade.[152] Keeping this in mind, we now move to analyse each type of obligation present under the Upholding levels of protection article. 

4.2.1.   Effective enforcement

The first category of obligations under the Upholding levels of protection articles in the TSD chapters is the effective enforcement obligation. The effective enforcement obligation in TSD chapters requires that parties not only maintain environmental laws on paper but also enforce them in practice. Importantly, for a breach to occur, there typically must be a ’sustained or recurring course of action or inaction’ that demonstrates a pattern of non-enforcement, rather than a single isolated incident.  

Based on the developments described above, there is strong evidence that the EU’s recent actions meet this threshold. The Omnibus Package and related reforms are not one-off adjustments, but rather part of a systematic and ongoing process of weakening and delaying environmental obligations. First, the delays and scope reductions are not limited to a single piece of legislation but are applied across multiple major environmental laws: the EUDR, vehicle CO₂ standards, and CBAM. The postponement of implementation dates, the introduction of broad exemptions (such as the ‘no-risk’ country category and the CBAM’s mass-based threshold), and the narrowing of reporting and due diligence requirements all reflect a recurring legislative pattern. Third, the effects of these changes are ongoing. Delays in implementation mean that, for extended periods, key environmental protections are not being enforced. The introduction of permanent exemptions (such as higher thresholds for reporting and due diligence) ensures that large segments of the market will remain outside the scope of enforcement indefinitely. The weakening of enforcement mechanisms, such as shifting the burden of proof from importers to the Commission or allowing standard values for emissions reporting, further entrenches this pattern.

The EU’s legislative changes, as described above, clearly amount to a ‘sustained or recurring course’. The pattern of delaying, exempting, and weakening enforcement across multiple environmental laws, coupled with the ongoing justification of these actions as necessary for trade and competitiveness, satisfies the requirement for a breach of the effective enforcement obligation under TSD chapters. This is not a case of sporadic or accidental non-enforcement, but a deliberate and repeated policy approach that undermines the effective implementation of environmental protections.

In conclusion, the question of whether delaying enforcement amounts to a failure to effectively enforce environmental laws is central to assessing compliance with TSD obligations. While changing the entry into force of a law is traditionally seen as part of the lawmaking process, the sustained or recurring postponement of implementation, as seen in the EU’s recent legislative changes, blurs the line between legislative timing and practical enforcement. By repeatedly deferring the application of key environmental protections across multiple regulatory regimes, the EU risks creating a situation where, in practice, enforcement is not taking place for extended periods. This pattern of delay, particularly when justified by ongoing competitiveness concerns, can be interpreted as a form of sustained or recurring inaction that undermines the effective enforcement obligation. Ultimately, while the distinction between legislative timing and enforcement practice is not always clear-cut. However, considering the consistent pattern of postponement and exemption across several laws certainly makes it a reasonable claim to bring to a Panel of Experts. 

4.2.2. Non-regression

Under the non-regression obligation where the parties commit not to lower or weaken domestic protection levels, changes must maintain or improve standards, The non-regression clause blocks backsliding below set levels. Within the context of the non-regression clauses, the obligation not to waive or derogate from domestic law will be analysed separately below. 

As stated above, identifying regression requires first determining the existing protection level, which is clear for quantitative standards, like emissions targets, but can be more complex for qualitative ones. Often in the TSD chapters, the scope of this obligation covers only ‘laws’, thus only when a legislation is adopted according to the procedures prescribed, regression from the standards established by this legislation will fall under the scope of the obligation and constitute a breach. Changes made to a proposal before its adoption will not be covered. In some FTAs, however, for instance CETA, not only ‘laws’ but also ‘policies’ or ‘regulations’ or ‘standards’ covered under the obligation. This extends the scope of the obligation; back up from previously announced policies will also constitute a breach.

Another question is whether regression made to proposed but not yet adopted legislation also fall within the scope. Of the developments listed above, this would regard to changes to the proposed CSRD and CSDDD. These proposed pieces of legislation have not completed the legislative procedures yet and as such are not pieces of law, covered under ‘laws’ term used in the TSD clauses. Although it is as not straightforward as in the case of already adopted legislation, one can argue that the regression made in the proposed legislation is necessarily a backsliding in the policy underneath these draft legislations and as such can still be seen to constitute a breach of the non-regression clause. 

Watering down of the to the substantive content of EU environmental law, a few examples of which are listed above, can mainly be pursued under this obligation. For instance, the introduction of a ‘no-risk’ country category under the EUDR exemplifies regression. By exempting nations with net forest gain, even if achieved through monoculture plantations, the EU has weakened the protection of the legislation compared to the original version, Monocultures, while increasing tree cover, often lack biodiversity and fail to replicate the ecological functions of natural forests. This exemption shifts the burden of proof from importers to the European Commission, reducing importer accountability and enabling continued trade in commodities linked to ecosystem degradation. Coupled with the 12-month delay in implementation, these highly criticized changes signal a deliberate dilution of standards, incentivizing business-as-usual practices in high-risk supply chains. Such measures directly conflict with TSD non-regression clauses, which prohibit parties from lowering environmental protections to facilitate trade. 

The flexibility mechanism introduced for vehicle CO₂ standards is another example of regression. By permitting automakers to average emissions over the 2025–2027 period rather than meeting annual targets, the EU enables short-term emissions increases, delaying progress toward climate goals and leads to more emissions.[153] Similarly, changes made under CBAM can also constitute regression and a subsequent breach of the TSD chapters. Exempting importers of less than 50 tons a year per product removes around 90% of companies from CBAM obligations.[154]

As mentioned, whether these regressions in EU’s environmental standards in its internal legislation are covered by the non-regression obligation in the TSD chapter varies per FTA, depending on what of the domestic law is covered thereunder and where at the life cycle of a legislation the changes to content have taken place. It suffices to say that a breach of the non-regression obligation under the TSD chapters through the recent changes to the EU legislation can be claimed in front of a Panel of Experts. Whether such a breach is established and upheld by the PoE will depend on the case at hand and the interpretations of the PoE itself. It is, however, certainly possible and plausible. The mere trigger of such proceedings against the EU, in itself, can have influential effects on the EU’s image, addressed further below. 

4.2.3. Not to waive or derogate from

As explained above, coupled often with the obligation to not to weaken or reduce the levels of protection in domestic law which prohibits regression through changes to domestic law, the obligation not to waive or derogate from domestic law on sustainability covers regression in practice, i.e. deviations from law without formal change. As the recent developments in the EU law as explained above all relate to formal changes and not deviating from the law in practice, this obligation is not relevant to the analysis at this moment, which could of course change in the future once the mentioned legislation comes into force.

5.   Conclusions

The EU has long been positioned as a global leader in promoting sustainable development through trade. By embedding ambitious environmental and labour commitments in its TSD chapters, the EU has sought to export its values and set global benchmarks. However, recent developments in the EU’s internal environmental agenda, such as the watering down of the Nature Restoration Law, the rollback of pesticide reduction targets, or delays in implementing the EUDR, pose serious risks not only to the legal coherence of its external action, but also to its normative authority and credibility as a global actor.

This article first clarified the intricate relationship between the Right to Regulate Levels of Protection and Upholding Levels of Protection articles within TSD chapters of EU FTAs. It delineated the obligations these articles impose on parties for their domestic laws regarding environmental and labour standards, emphasizing their complementary nature in promoting higher protection levels and preventing regression. The article Right to Regulate ensures domestic laws align with international standards and advocates continuous improvement and higher protection levels. The article Upholding Levels of Protection guards against any backsliding from these standards, and ensures effective enforcement, complementing the standards set through the Right to Regulate. 

Under the non-regression clause of the Upholding Levels of Protection, two types of obligations appear in different combinations. The interpretative exercise undertaken reveals a complementary relationship between these two obligations, as ‘waiving or otherwise derogating from domestic law’ addresses the changes to the domestic law as it is now while ‘lowering the levels of protection’ addresses changes to the domestic law and as such the domestic law in the future. Only such an interpretation renders both these clauses meaningful. Together, they provide a complete system.

Having analysed the recent developments and the wind of change regarding the EU’s environmental legislation, the article has shown that a many of these developments can be seen as a breach of these obligations. A partner country can trigger the DSM available for the TSD chapters of its FTA with the EU and bring the claim that the EU is breaching its obligations regarding the domestic law in the TSD chapters that the EU arguably drafted itself. 

All the FTAs up to the EU-Singapore FTA of 2019 are signed by both the EU and the individual member states. The question of who on the EU side would be responsible for such a breach in these mixed agreements is left outside the scope of this article. Yet, the stakes are not merely legal. The reputational cost of these developments may be even more damaging. For years, the EU has cultivated its reputation as a normative power,[155] an actor that upholds high social and environmental standards and globally pioneers them. TSD chapters are often commented as being the EU’s own norms that are being exported through the FTAs.[156]This image has lent credibility to its policy initiatives and has played a key role in encouraging partner countries to adopt higher sustainability standards. If the EU is now seen as retreating from those values when it becomes politically or economically inconvenient, it risks being labelled hypocritical. Such a perception would not only erode the EU’s soft power and diplomatic capital but also undermine the credibility of its ongoing and future negotiations with third countries.

More broadly, the EU's internal political dynamics, particularly the growing influence of agricultural and industrial interests pushing back against green regulation, are starting to visibly shape its external posture. This internal-external disconnect has been largely neglected in both scholarly and policy debates, but it demands urgent attention. If the EU wishes to maintain the integrity and enforceability of TSD chapters, it must ensure that its internal legislative decisions do not directly or indirectly contradict its international obligations.

The consequences of failing to do so are far-reaching. Beyond opening the door to legal disputes, however limited their effects might be in the TSD context, , the EU may gradually lose its position as a credible standard-setter in global sustainability governance due to the damage to its reputation. At a time when the EU seeks to assert itself geopolitically and build alliances based on shared values, coherence between its internal policies and external commitments is not a luxury, but a strategic necessity.