I. Introduction
Between 2019 and 2024, the European Union adopted an expansive suite of environmental legislation under the umbrella of the European Green Deal.Footnote 1 The Corporate Sustainability Reporting Directive (CSRD) and the Corporate Sustainability Due Diligence Directive (CS3D) were crucial pillars in this architecture, introducing informational, reporting and governance obligations intended to align corporate conduct with environmental objectives and human rights standards.Footnote 2 Yet by 2024, the political and economic context had shifted. The post-pandemic recovery agenda was overshadowed by inflationary pressures, energy insecurity and geopolitical turmoil. Against this backdrop, short-term competitiveness became a central concern of EU leaders.Footnote 3 The 2024 report prepared by Mario Draghi, which diagnosed a competitiveness deficit and called for regulatory simplification, became a focal text in this narrative.Footnote 4
This shift towards competitiveness gave rise to a broader legislative trend exemplified by the increasing use of “omnibus” acts as vehicles for regulatory consolidation and recalibration. Omnibus legislation refers to the bundling of multiple amendments into a single act.Footnote 5 Rather than proposing stand-alone amendments to individual instruments, the Commission began bundling revisions across multiple legislative files into single, cross-cutting packages.Footnote 6
It is in this political moment that the Commission tabled its Sustainability Omnibus in February 2025. Formally presented as a simplification exercise, the Sustainability Omnibus proposal packaged amendments to the CSRD and CS3D with the stated aim of reducing burdens, clarifying obligations, and focusing regulation on large undertakings.Footnote 7 Only a year later, Directive (EU) 2026/470 (Omnibus Directive) was adopted, which is notably fast compared to the average legislative adoption time of 17 months.Footnote 8
Second, and separately, the package included the so-called “stop-the-clock” proposal, designed to temporarily delay the application of certain CSRD and CS3D obligations in order to avoid companies having to comply with requirements that were still under political renegotiation.Footnote 9 The stop-the-clock Directive was adopted at an early stage of the legislative process and entered into force in 2025.Footnote 10
The Omnibus proposal was framed in technocratic terms,Footnote 11 yet it introduced changes that materially narrowed the scope and density of corporate sustainability obligations.Footnote 12 At the same time, the way the package was prepared raised procedural concerns. The Commission neither produced a dedicated impact assessment nor launched an open public consultation, relying instead on selective stakeholder contacts and invoking “urgency” to justify procedural shortcuts.Footnote 13 These concerns, which were already present at the time the Commission’s proposal was introduced,Footnote 14 were also echoed by the European Central Bank, which cautioned against the potential loss of crucial sustainability data resulting from the proposed changes.Footnote 15 They ultimately culminated in a complaint to the European Ombudsman,Footnote 16 who found maladministration on the grounds that the Commission had failed to ensure that the proposal was prepared in a manner that was transparent, inclusive and evidence-based.Footnote 17
In this context, the Sustainability Omnibus provides a particularly illuminating case for examining how the emerging principle of non-regression might be operationalised in EU law. Non-regression, as articulated by the International Union for Conservation of Nature (IUCN) in 2015, expresses the idea that “substantive and procedural rules for environmental conservation shall be maintained without regression … unless compelling reasons of public interest require otherwise.”Footnote 18 While the principle has not yet crystallised into a fully binding rule of EU or international law, it finds support in developments in international environmental and human rights law as well as in EU law.Footnote 19
The principle of non-regression has emerged as a theme in environmental law scholarship in response to concerns about environmental backsliding through legislative reforms. The concept is most closely associated with Prieur, who developed the first systematic account of non-regression as a principle of environmental law, grounded in sustainable development, intergenerational equity, human rights and the objective of achieving a high level of environmental protection.Footnote 20 However, its legal status as a principle remains contested. Vordermayer-Riemer argues that the available evidence is insufficient to establish non-regression as a rule of customary international law or a general principle of international law, suggesting instead that it is best understood as a principle of environmental soft law.Footnote 21 More recent scholarship has shifted attention from the existence of non-regression as an autonomous legal principle to its practical function. In particular, Bryner argues that non-regression should be understood as a safeguard against unjustified deregulation, requiring evidence-based reasons before reducing environmental protection standards.Footnote 22 This procedural perspective provides the conceptual foundation for the present analysis.
The article argues that, in EU law, non-regression is best understood not as an absolute prohibition on lowering environmental standards, but as an emerging principle establishing a presumption against environmental backsliding. Where environmental protection is reduced, that presumption should trigger heightened requirements of evidence, participation, transparency and justification. In that sense, the Sustainability Omnibus allows an inquiry into non-regression in a double sense: as a matter of substance, through the narrowing of environmental obligations, and as a matter of procedure, through the weakening of the safeguards that ordinarily structure EU lawmaking.
Against that background, the central research question of this article is: to what extent does the Sustainability Omnibus constitute a form of substantive and procedural regression in EU law, and how can the emerging principle of non-regression be operationalised in EU law as a procedural constraint on the adoption of regressive measures?
The article proceeds in four steps. First, it identifies the substantive changes introduced by the Sustainability Omnibus and explains why, taken cumulatively, they may be characterised as regressive in light of the level of environmental protection established by the original CSRD and CS3D. Second, it analyses the procedural deficiencies surrounding the preparation and adoption of the Omnibus, with particular attention to the absence of a dedicated impact assessment, the curtailment of public participation, and the Commission’s reliance on a broad and undefined notion of urgency. Third, it examines the emergence of non-regression in international and EU environmental law and argues that, while non-regression has not yet developed into a fully-fledged principle, it is best understood as an emerging principle expressing a presumption against backsliding that triggers a justification requirement. Fourth, the article applies that understanding to the Sustainability Omnibus and argues that, in the EU context, non-regression should be operationalised primarily in an ex ante way through heightened procedural requirements whenever environmental protection is reduced.
II. Substantive regressive elements of the Sustainability Omnibus
The Omnibus Directive introduces significant modifications to both the CSRD and the CS3D, many of which narrow the scope of the original legislation and reduce the intensity of regulatory obligations imposed on companies. Although the Commission presented these amendments as measures intended to simplify compliance and enhance competitivenessFootnote 23 the cumulative effect of these changes raises questions as to whether the reform constitutes a form of regulatory regression within EU environmental law.Footnote 24 Against this background, a closer examination of the specific amendments to each directive is necessary to assess the extent to which the Omnibus Directive alters the original regulatory framework.
1. Amendments to the CSRD
One of the most significant changes introduced by the Omnibus Directive concerns the scope of sustainability reporting obligations under the Accounting Directive, The Omnibus Directive limits mandatory sustainability reporting to undertakings that exceed EUR 450 million in net turnover and have an average of more than 1,000 employees during the financial year.Footnote 25 This amendment substantially narrows the personal scope of the CSRD compared with the original legislative framework, which had been intended to extend sustainability reporting obligations to a broader group of undertakings, including listed SMEs.Footnote 26
The narrowing of scope is reinforced by the deletion of the provision concerning the development of sustainability reporting standards for SMEs. More specifically, the Omnibus Directive removes the Commission’s power to adopt sustainability reporting standards for listed SMEs and eliminates the legislative mechanism through which such undertakings would have progressively been integrated into the sustainability reporting framework.Footnote 27 As a result, a segment of companies that may nevertheless generate significant environmental or social impacts is no longer subject to mandatory reporting obligations.
A similar limitation of scope can be observed in relation to third-country undertakings, as the EU turnover threshold for non-EU parent undertakings is increased from EUR 150 million to EUR 450 million and a threshold of EUR 200 million for EU subsidiaries or branches is introduced.Footnote 28 Raising these thresholds limits the extraterritorial reach of the EU sustainability disclosure framework.
The Omnibus Directive also introduces new constraints on the ability of reporting undertakings to obtain sustainability information from actors in their value chains with the creation of the concept of “protected undertaking,”Footnote 29 Under this mechanism, companies with fewer than 1,000 employees that form part of the value chain of a reporting undertaking may refuse to provide information exceeding the content required by the voluntary sustainability reporting standards, and contractual clauses requiring additional disclosures are rendered non-binding.Footnote 30 While intended to reduce administrative burdens for smaller companies, this mechanism may also limit the ability of reporting undertakings to obtain detailed information about environmental and human rights impacts occurring within their supply chains.
The Omnibus Directive also affects the content of the European Sustainability Reporting Standards (ESRS). It removes the Commission’s power to adopt sector-specific sustainability reporting standards, which were intended to capture industry-specific environmental and social risks.Footnote 31 As a result, sustainability reporting will rely exclusively on cross-sectoral standards, supplemented only by non-binding guidance.
The Directive further modifies the assurance framework applicable to sustainability reporting. The Commission is now required to adopt limited assurance standards for sustainability reporting by 1 July 2027, while the possibility of subsequently moving to a reasonable assurance standard has been removed.Footnote 32 Under a limited assurance engagement, the auditor expresses a negative form of assurance, indicating that the procedures performed have not revealed any material misstatements.Footnote 33 By contrast, reasonable assurance results in a positive opinion, whereby the auditor concludes, on the basis of more extensive verification and testing, that the information is free from material misstatement.Footnote 34 Since limited assurance provides a lower level of verification than reasonable assurance, the removal of the possibility of transitioning to reasonable assurance reduces the level of scrutiny applied to sustainability disclosures.
Overall, these amendments significantly reduce the scope and robustness of the CSRD framework by narrowing the number of companies subject to reporting obligations and weakening elements of the disclosure architecture. Thus, the Omnibus Directive reduces the ability of the EU sustainability reporting regime to generate comprehensive and comparable sustainability information across the internal market.
2. Amendments to the CS3D
The Omnibus Directive also introduces substantial modifications to the CSRD. Most notably, it significantly raises the thresholds determining which companies fall within the scope of the Directive.Footnote 35 The revised framework now applies only to EU companies with more than 5,000 employees and net worldwide turnover exceeding EUR 1.5 billion. Non-EU companies fall within the scope of the Directive if they reach EUR 1.5 billion in turnover generated within the EU.Footnote 36 These thresholds represent a considerable increase compared with the original framework of the CS3D and substantially reduce the number of companies subject to mandatory due diligence obligations.Footnote 37
The Omnibus Directive also modifies the structure of the due diligence process. Companies are now required to carry out a scoping exercise based solely on reasonably available information to identify general areas where adverse impacts are most likely to occur or be most severe, followed by an in-depth assessment only in those areas.Footnote 38 This modification, replaces the broader mapping expectations associated with the original due diligence framework and reduces the intensity of investigative obligations imposed on companies.Footnote 39 Although this approach may reduce compliance and administrative burdens, it also narrows the Directive’s investigative reach and increases the risk that certain adverse environmental or human rights impacts remain unidentified, particularly if they occur in opaque value chains.
Additional limitations are placed on the ability of companies to request information from business partners during the due diligence process. Companies may request information from business partners only where necessary and, in the case of partners employing fewer than 5,000 employees, only where such information cannot reasonably be obtained by other means.Footnote 40 These restrictions may reduce the depth of due diligence investigations and limit companies’ ability to obtain information concerning environmental or human rights risks within their value chains.
One of the most significant changes concerns the removal of the climate transition plans requirement.Footnote 41 Article 22 required companies to adopt and implement climate transition plans aligned with the objectives of the Paris Agreement. Its deletion removes a central mechanism through which the Directive sought to align corporate strategies with EU climate policy. The significance of this amendment extends beyond the removal of a single compliance obligation. Climate transition plans were intended to serve as a forward-looking mechanism linking corporate decision-making to the Union’s long-term climate objectives. By requiring companies to assess and plan the compatibility of their business models with the transition to a climate neutral economy, Article 22 sought to operationalise the objectives of the Paris Agreement within corporate governance structures. Its repeal therefore weakens one of the most direct connections between the CS3D and the broader framework of international and EU climate law by eliminating a mechanism specifically designed to facilitate progressive alignment with climate objectives.Footnote 42 It is therefore difficult to characterise the amendment as purely technical or administrative in nature, as it directly affects the capacity of the Directive to promote the transition towards climate neutrality. It is worth noting that the Commission’s original proposal had already attracted criticism for removing the requirement that companies “put into effect” their climate transition plans.Footnote 43 The subsequent deletion of the provision in its entirety during the legislative process only reinforces such concerns.
The Omnibus Directive also modifies the enforcement architecture of the CS3D. The specific EU-wide civil liability regime is removed and that provisions concerning representative actions and the overriding mandatory application of civil liability rules in cross-border cases are deleted.Footnote 44 These changes leave the conditions governing civil liability largely to national legal systems, potentially resulting in a fragmented enforcement landscape across Member States.Footnote 45
The amendments introduced by the Omnibus Directive substantially reshape the EU’s corporate sustainability framework. Many of the changes narrow the scope of the legislation, reduce the intensity of due diligence obligations, or weaken enforcement mechanisms. Although the Omnibus Directive is presented as a simplification exercise aimed at reducing administrative burdens, the cumulative effect of these modifications may be to reduce the level of environmental and human rights protection originally envisaged under the CSRD and the CS3D. For these reasons, the reform has been widely discussed as a potential example of regulatory regression in EU sustainability law.Footnote 46 More specifically, Alemanno has argued that the proposal departs from the traditional function of omnibus legislation as a technical simplification tool and instead reopens substantive legislative compromises that had only recently been negotiated and adopted.Footnote 47 In his view, the Omnibus risks transforming simplification into a mechanism for reducing regulatory ambition, particularly where competitiveness and burden reduction are prioritised over previously agreed sustainability commitments.Footnote 48
III. Procedural deficiencies in the adoption of the Sustainability Omnibus
A key feature of the Sustainability Omnibus is the legislative technique through which it was adopted, namely the use of an omnibus instrument. This technique warrants closer attention, as it shapes the procedural context in which the identified substantive changes were introduced. Concerns about bundling heterogeneous legislative measures are not new. Already in Roman law, so-called leges saturae (“stuffed laws”) combined multiple unrelated provisions within a single act, a practice that was eventually restricted by the lex Caecilia Didia, which prohibited the inclusion of unrelated matters within a single legislative proposal.Footnote 49
While the use of omnibus legislation is not per se unlawful under EU law,Footnote 50 the technique has historically been used primarily for technical purposes, such as aligning terminology, correcting cross-references, or consolidating amendments across several legal acts. In some areas of EU law, such as consumer protection,Footnote 51 omnibus legislation has been used to update and align several directives simultaneously in order to improve coherence within the regulatory framework.
Yet omnibus instruments may also obscure substantive choices under the guise of technicality or simplification.Footnote 52 They reduce legislative scrutiny and may enable deregulatory measures to be embedded within broader packages that receive less attention than standalone proposals. As a result, greater reliance on omnibus bills raises concerns about transparency, democratic oversight, and meaningful stakeholder participation.
It is precisely these concerns that come to the fore in the case of the Sustainability Omnibus. Its adoption raises significant procedural issues, as the expedited process followed by the Commission departed from safeguards designed to secure transparency, evidence-based decision-making, and meaningful public participation. These safeguards are embedded within the broader framework of EU law designed to promote transparency, participation, reasoned and evidence-based decision-making and good administration. Articles 10 and 11 TEU require decision-making to be conducted as openly as possible and oblige the Commission to maintain broad consultations with parties concerned. Moreover, Article 296 TFEU requires EU institutions to state the reasons on which their acts are based and Article 298 TFEU establishes the principle of open, efficient and independent European administration. These requirements are further reflected in Article 41 of the Charter, which guarantees the right to good administration, including the obligation of the administration to give reasons for its decisions.
These principles are further operationalised through the Interinstitutional Agreement on Better Law-Making and the Commission’s Better Regulation Guidelines, which emphasise the importance of impact assessments and stakeholder participation in the preparation of legislative proposals.Footnote 53 In environmental matters, these requirements are reinforced by sector-specific obligations, including Article 6(4) of the European Climate Law, which requires the Commission to assess the consistency of legislative proposals with the Union’s climate neutrality objective and climate targets, as well as by the participatory logic underpinning the Aarhus Convention.Footnote 54
The procedural shortcomings surrounding the adoption of the Omnibus Directive can be grouped around two closely connected issues: the absence of an impact assessment and the curtailment of public participation. Both stem from the Commission’s reliance on an expansive and weakly substantiated notion of “urgency.” The notion of “urgency” is not defined in the Better Regulation framework. In the Omnibus context, urgency was linked to competitiveness concerns and the objective of reducing regulatory burdens.Footnote 55 Yet those considerations, while politically salient, do not amount to a sudden or unforeseeable emergency.
Under the Commission’s Better Regulation Guidelines and the Interinstitutional Agreement on Better Law-Making, legislative proposals likely to have significant economic, environmental or social effects should, as a general rule, be accompanied by an impact assessment.Footnote 56 The Sustainability Omnibus, which amends sustainability reporting and due diligence obligations with significant implications for environmental protection and fundamental rights, was not supported by such an assessment.Footnote 57 Instead, the Commission invoked “political urgency” and replaced the impact assessment with a Staff Working Document, claiming that it would rely on the impact assessments prepared for the directives that were being amended.Footnote 58
The recourse to urgency by the EU legislator is not without precedent. For instance, the REPowerEU amendments to the Renewable Energy Directive were introduced without an impact assessment due to the urgency of the situation.Footnote 59 Those amendments were adopted in the context of a genuine geopolitical emergency, i.e., Russia’s war of aggression against Ukraine and the resulting energy crisis. Even in that situation of real and immediate disruption, the decision to omit a full impact assessment attracted criticism given the potentially significant environmental implications of the measures.Footnote 60
Unlike the REPowerEU, the Sustainability Omnibus does not respond to an acute crisis of similar magnitude but to a change of political priorities. In the Sustainability Omnibus, urgency was grounded primarily in competitiveness concerns and the perceived need to provide regulatory relief before certain obligations became applicable.Footnote 61 However, competitiveness is a long-standing policy objective rather than an unforeseen event. The legislation was only recently adopted, its entry into force and application timeline were known in advance and no sudden legal vacuum or external crisis made an impact assessment practically impossible. Thus, the invocation of urgency appears to reflect a change in political priorities rather than an objective impossibility of preparing an impact assessment within the available timeframe and risks normalising what was intended to remain exceptional.Footnote 62
Moreover, the Commission’s reliance on the impact assessments prepared for the original directives is methodologically problematic. If circumstances have changed so fundamentally as to require urgent amendment, it is difficult to assume that the earlier assessments remain reliable. That concern is amplified by the scale of the proposed reforms. The Omnibus substantially narrows the scope of both the CSRD and the CS3D.Footnote 63 Where amendments significantly alter the personal and material scope of legislation, the underlying assumptions concerning economic, environmental and human-rights effects necessarily change as well. In those circumstances, an evidence-based approach would ordinarily require at least a targeted reassessment of the likely consequences of the new measure. The analytical document does not offer an equivalent substitute as it is not subject to the same methodological discipline as a full impact assessment, nor to review by the Regulatory Scrutiny Board.Footnote 64
The result is a weakening of the evidence-based dimension of EU law-making. Impact assessments are not merely technocratic exercises but accountability instruments that discipline political discretion and provide a transparent basis for public scrutiny.Footnote 65 It is true that EU law does not impose an autonomous obligation to prepare an impact assessment in every case. However, the Court of Justice has made clear that the exercise of legislative discretion must rest on a sufficiently robust evidentiary basis. That requirement can be seen as the procedural dimension of the principle of proportionality.Footnote 66
The link between impact assessment and proportionality has been made more explicit in the caselaw of the CJEU. In Czech Republic v European Parliament and Council, the Court confirmed that the absence of an impact assessment does not automatically invalidate a legislative act.Footnote 67 However, it also held that such an omission must be justified by a “particular situation” and that, absent such justification, the lack of sufficient information enabling the legislature to assess the proportionality of the measure may render the act unlawful.Footnote 68 The Court developed this procedural dimension of proportionality review further in Lithuania and Others v European Parliament and Council.Footnote 69 There, it required the EU legislature to show that it had actually exercised its discretion on the basis of sufficiently solid and detailed evidence, and it annulled part of the contested legislation because the institutions lacked sufficient information to assess proportionality.Footnote 70
The derogation from the impact-assessment requirement also produced spill-over effects on the compliance with the European Climate Law. Article 6(4) of that Regulation imposes a legally binding obligation on the Commission to assess the consistency of draft legislative proposals with the climate-neutrality objective and the Union climate targets, and to make the result of that assessment publicly available at the time of adoption.Footnote 71 Where an impact assessment exists, Article 6(4) expressly requires that this consistency assessment be included in it. The legislation is less clear as to the form that assessment must take where no impact assessment has been prepared. Yet, that uncertainty cannot obscure the minimum requirement laid down in Article 6(4) regarding the publication of the assessment when the proposal is adopted. In the case of the Omnibus, the Commission referred only to general statements concerning consistency with the European Green Deal while no clearly identifiable climate-consistency assessment was published, nor was its result distinctly made public at the time of adoption.Footnote 72 The Ombudsman considered that such a lack of transparency does not constitute good administration.Footnote 73
The Better Regulation Guidelines further underline the importance of wide consultations and require a12-week consultation period as a minimum standard.Footnote 74 However, for the preparation of the Omnibus proposal, the Commission relied on earlier calls for evidence on reporting rationalisation, hybrid stakeholder events, bilateral meetings, and a “reality check” workshop involving selected stakeholders.Footnote 75
This is problematic in two respects. First, meaningful participation presupposes knowledge of the options actually under consideration.Footnote 76 General exchanges about simplification or reporting burdens cannot substitute for consultation on concrete amendments that alter the scope and substance of EU legislation. Second, questions arise as to representativeness and balance. The criteria used to select participants for targeted events were not transparently explained, and the available material suggests a strong presence of industry actors.Footnote 77 While targeted consultations may be appropriate in technical matters, Articles 10 and 11 TEU require decision-making to be conducted as openly as possible and oblige the Commission to maintain broad consultation with parties concerned in accordance to the principle of transparency. In Scheke, the Court explained that transparency enables citizens to participate more closely in the decision-making process and guarantees that the administration enjoys greater legitimacy, effectiveness and accountability in a democratic system.Footnote 78 Curtailing participation in a file with substantial environmental and human-rights implications therefore weakens the inclusiveness and democratic quality of the legislative process. Therefore, the procedure followed for the Omnibus proposal seems to depart from the Commission’s Better Regulation Guidelines.
Although the Better Regulation Guidelines themselves expressly state that they are not legally binding,Footnote 79 the Court held in Dansk Rørindustri that an institution which adopts rules to govern the exercise of its discretion cannot depart from them without regard to the principles of equal treatment and legitimate expectations.Footnote 80 Such self-imposed rules therefore structure discretion and constrain arbitrary departure from announced practice.
These considerations also bear on the duty to state reasons. Article 296 TFEU requires legal acts to state the reasons on which they are based. According to settled case-law, the statement of reasons must be appropriate to the act at issue and must disclose in a clear and unequivocal fashion the reasoning followed by the institution, so as to enable the persons concerned to understand the measure and the Court to exercise its power of review.Footnote 81 The duty to state reasons is an essential procedural requirement in the context of a ground a review for an action for annulment.Footnote 82
Moreover, the reasons given must still be sufficiently clear and specific to make review possible. The Court has held that insufficient reasons are tantamount to an absence of reasons where the factual and legal basis of the decision is not disclosed in a way that permits judicial review.Footnote 83 In the case of the Omnibus, the reasons offered by the Commission remain thinly substantiated. The proposal’s explanatory memorandum acknowledges that no ex post evaluation or fitness check of either the CSRD or the CS3D had been possible.Footnote 84 That does not by itself establish an infringement of Article 296 TFEU, but it does weaken the capacity of the statement of reasons to explain, in a clear way, why the existing legislation had to be reopened so quickly and so extensively.
Finally, these shortcomings should also be viewed through the lens of good administration. Article 41 of the Charter of Fundamental Rights and Article 298 TFEU reflect a broader requirement that EU administration be open, diligent and reviewable. The CJEU has recognised good administration as a general principle of EU law in Burban v Parliament.Footnote 85 In this context, good administration is not limited to courtesy or procedural neatness but imposes a duty of diligence, careful and impartial examination of the relevant factors, and sufficient transparency. The failure to document and publish a clearly identifiable climate-consistency assessment, the absence of an impact assessment, the compressed interservice consultation, and the limited and potentially unbalanced stakeholder process all point in the opposite direction.
The European Ombudsman, following a complaint by eight civil society organisations,Footnote 86 examined these procedural choices in detail and concluded that the cumulative shortcomings amounted to maladministration, as part of a broader failure to ensure a transparent, evidence-based and inclusive preparation of urgent legislative proposals. The Ombudsman criticised, inter alia, the broad and undefined interpretation of “political urgency,” the insufficient justification provided in the explanatory memorandum, the absence of clear internal records documenting the derogation, the limited and potentially unbalanced stakeholder consultation, and shortcomings in recording and publishing climate consistency assessments.Footnote 87 The Commission, by contrast, argued that the proposal responded to an urgent need to provide legal certainty to companies and Member States before the application of the CSRD and CS3D obligations and that the broader competitiveness concerns identified in the Draghi Report justified swift regulatory action. It further maintained that sufficient evidence already existed in the form of previous impact assessments, stakeholder consultations and implementation experience, making a new impact assessment unnecessary.Footnote 88 However, the Ombudsman did not object to the possibility of derogating from the Better Regulation requirements in urgent situations. Rather, she found that the Commission had failed to explain and document adequately why such derogations were justified in the present case and had not ensured that the resulting procedure remained sufficiently transparent, evidence-based and inclusive.
Against this background, the procedural shortcomings identified above are not merely of administrative relevance. They also bear directly on the substantive direction of EU environmental law. If legislative amendments are adopted without adequate evidence, scrutiny, or justification, there is an increased risk that existing levels of environmental protection may be weakened without sufficient safeguards. This concern brings into focus the concept of non-regression, which seeks to limit unjustified reductions in environmental standards and provides a useful analytical lens through which to assess the broader implications of the Sustainability Omnibus.
IV. The emerging principle of non-regression in environmental law
Non-regression in environmental law expresses the idea that once a certain level of environmental protection has been attained, it should not be reduced without compelling justification.Footnote 89 Although the concept has gained doctrinal, political, and judicial attention, it has not yet matured into a universally binding rule of international or EU law.Footnote 90
Its conceptual foundations lie first in the principle of sustainable development. Under its classic formulation, i.e., development that meets the needs of the present without compromising the ability of future generations to meet their own needs, sustainable development contains an inherently forward-looking logic.Footnote 91 If environmental standards are weakened, the ecological conditions necessary for future generations to satisfy their needs are correspondingly diminished.Footnote 92 Non-regression thus appears as a corollary of sustainable development’s intergenerational dimension.Footnote 93 This reasoning is closely linked to the principle of intergenerational equity, which demands that present generations manage natural resources in a manner that preserves comparable environmental opportunities for those yet to come.Footnote 94
International human rights law provides a more doctrinally developed articulation of the concept through the doctrine of non-retrogression. Under the International Covenant on Economic, Social and Cultural Rights, states are required to progressively realise rights such as health and an adequate standard of living.Footnote 95 The UN Committee on Economic, Social and Cultural Rights clarified in General Comment No. 3 that any deliberately retrogressive measures are prima facie incompatible with the Covenant, “would require the most careful consideration and would need to be fully justified.”Footnote 96 Importantly, this doctrine has been extended explicitly to environmental protection. In its General Comment No. 27 on the right to a clean, healthy and sustainable environment, the Committee affirmed that weakening environmental regulations may constitute impermissible retrogression unless strictly justified.Footnote 97 Environmental degradation that undermines Covenant rights may therefore breach the obligation of non-retrogression. Although treaty body interpretations are not formally binding, they significantly shape the normative landscape and strengthen the conceptual foundations of this presumption against environmental backsliding.
Within international environmental law, the outcome document of the 2012 Rio+20 Conference, “The Future We Want,” affirmed that states must not “backtrack” on prior environmental commitments, a formulation understood as an endorsement of non-regression, even if the term itself was politically contested.Footnote 98 The logic of progression is also embedded in several multilateral environmental agreements. The Escazú Agreement explicitly recognises both non-regression and progressive realisation as guiding principles.Footnote 99 Similarly, the Paris Agreement requires in Article 4(3) that each successive nationally determined contribution represent a progression beyond the previous one, thereby institutionalising an expectation of increasing ambition rather than retrenchment.Footnote 100 This dynamic understanding of obligations is further reinforced by the recent advisory opinion of the International Court of Justice on climate change, which emphasises that the content of States’ due diligence obligations must be interpreted in light of the urgency and seriousness of the climate crisis.Footnote 101 Thus, these developments suggest that while non-regression may not yet constitute a settled rule of customary international law, it can be considered an emerging principle of environmental law.Footnote 102
In the EU legal order, non-regression is not explicitly codified as a principle of environmental law. Nevertheless, it finds substantial support in the teleology of the Treaties. Article 3(3) TEU commits the Union to strive for a “high level of protection and improvement of the quality of the environment,” while Article 191(1) TFEU requires Union environmental policy to pursue preservation, protection and improvement, aiming at a high level of protection. The inclusion of “improvement” signals a dynamic orientation implying that environmental protection is not merely to be maintained but enhanced over time and that, a contrario, backsliding is prohibited. This reading is further reinforced by Article 11 TFEU, which requires environmental protection requirements to be integrated into the definition and implementation of all Union policies and activities. By constitutionalising environmental integration across the Union legal order, Article 11 TFEU supports an understanding of EU law as oriented towards the progressive advancement of environmental protection.Footnote 103 Although these provisions do not expressly prohibit regression, their combined reading suggests that deliberate backsliding sits uneasily with the progressive nature of EU environmental law.Footnote 104
EU secondary legislation provides more concrete expressions of non-regression obligations. The Water Framework Directive obliges Member States to prevent deterioration of water bodies, and the Nature Directives require the avoidance of habitat deterioration or any disturbances affecting the birds.Footnote 105 The CJEU has interpreted these non-deterioration clauses as general obligation of protection that contribute to the “fundamental objective of preservation and protection of the quality of the environment.”Footnote 106 Moreover, the Court has recognised a principle of non-regression in the field of the rule of law. In Repubblika, it held that Member States may not reduce the level of protection of the rule of law guaranteed under Article 2 TEU.Footnote 107 This reasoning was subsequently extended in Commission v Hungary, where the Court clarified that a Member State may not amend its legislation so as to bring about a reduction in the protection of the values enshrined in Article 2 TEU and that Member States are required to ensure that any regression is prevented., extending non-regression to all values protected by Article 2 TEU.Footnote 108
In addition, non-regression clauses have become a standard feature of trade agreements and are now common in EU FTAs, committing the EU and its partners to avoid lowering environmental and labour standards in ways that distort trade, as seen for instance in the Comprehensive Economic and Trade Agreement (CETA) between the EU and Canada, the EU-Korea Free Trade Agreement and the EU UK Trade and Cooperation Agreement (TCA).Footnote 109
What distinguishes the TCA is the strength of the enforcement machinery attached to non-regression. Article 391 TCA sets out the non-regression obligation for environmental policy, prohibiting either party from weakening or reducing its levels of environmental protection “in a manner affecting trade or investment between the Parties” below the standards in place on the withdrawal date of the United Kingdom (1 January 2021). An equivalent obligation applies to labour and social standards.Footnote 110 Crucially, both obligations are conditional since a regression is only actionable if it can be shown to affect trade or investment, a threshold that significantly narrows the scope of the clauses and leaves purely domestic erosions of standards potentially beyond challenge. Where a breach is alleged, Articles 408 to 410 TCA provide the enforcement mechanism, involving consultations followed, if necessary, by referral to a specialist Panel of Experts. Non-compliance with the panel’s findings can then trigger suspension of TCA obligations.Footnote 111 Separately, a “rebalancing mechanism” is established allowing either party to take proportionate measures where significant divergences between the parties’ standards create “material impacts on trade or investment.”Footnote 112 Any rebalancing measures must be restricted to what is “strictly necessary and proportionate.”Footnote 113 It should be noted, however, that these mechanisms remain untested, and key concepts such as “significant divergence” and “material impact” have not yet been concretised.
From the preceding analysis, two observations emerge. First, the legal nature of non-regression remains indeterminate. It cannot yet be characterised as a fully-fledged general principle of international or EU law, but neither is it merely rhetorical. Rather, it occupies an intermediate normative position, functioning as an emerging principle.Footnote 114 However, the legal status and conceptual clarity of non-regression remain uncertain. Moreover, different formulations and terminologies are used across legal regimes, such as “non-regression,” “non-retrogression,” “standstill” or “non-deterioration,,” giving rise to conceptual ambiguity.Footnote 115 These variations are further compounded by differences in scope, including whether the concept applies to legal standards, factual environmental outcomes, or both.Footnote 116 These issues make it difficult to characterise non-regression as a settled legal principle. Instead, it is more accurately understood as an emerging normative concept that is progressively gaining legal relevance.
Secondly, and more importantly, non-regression consistently manifests as a qualified constraint rather than an absolute prohibition.Footnote 117 Despite the conceptual fragmentation and variability in its expression, the qualified nature of non-regression emerges as a common denominator. Across the different legal regimes examined, the concept operates through a structured presumption against backsliding and not as a categorical ban. This is logical since there must be some flexibility for legitimate adjustments, since environmental science and policy evolve and thus consequently laws might need updating. In certain circumstances, urgency or new evidence may justify a recalibration of regulatory approaches.
This qualified nature is particularly evident when examining how the principle is defined and operates in practice in different legal contexts in international and EU law. For instance, the IUCN defines non-regression as a principle according to which environmental law should not be revised in a manner that reduces the level of environmental protection achieved, except where such changes are duly justified.Footnote 118 This formulation makes clear that non-regression does not impose an absolute ban on backsliding but instead establishes a presumption against it.
In international human rights law, retrogressive measures are not per se unlawful but must be justified.Footnote 119 The UN Committee on Economic, Social and Cultural Rights articulated a “strong presumption against retrogressive measures”Footnote 120 rather than an outright prohibition. Thus, regressive steps are not per se unlawful, but they require particularly careful justification, must pursue a legitimate aim, and must be proportionate and consistent. Similarly, in EU trade agreements such as the EU–UK Trade and Cooperation Agreement, non-regression clauses are limited to situations where reductions in environmental protection affect trade or investment, meaning changes that do not impact trade or investment likely fall outside its scope’.Footnote 121
In EU environmental legislation, non-regression clauses are often framed by thresholds of significance and allow for certain derogations under strictly defined conditions. For instance, Article 6(2) of the Habitats Directive and Article 4(4) of the Birds Directive impose obligations that are not absolute, operating within a significance threshold and providing exceptions for cases of overriding public interest.Footnote 122 The 2019 UN Environmental Rule of Law Report confirms this approach, emphasising that regressive measures may be permissible where they are strongly justified.Footnote 123
Against this doctrinal background, the following section applies non-regression to the Sustainability Omnibus, assessing whether and to what extent the proposed reforms are compatible with this emerging principle and proposing ways that it could be operationalised to prevent environmental backsliding.
V. Applying non-regression to the Sustainability Omnibus
The Sustainability Omnibus provides a particularly useful test case for the emerging principle of non-regression in EU environmental law. Having identified the substantive changes introduced by the Sustainability Omnibus and the manner in which it was prepared, the issue is now whether, and in what sense, those features amount to regression, and how the emerging principle of non-regression could be operationalised in order to prevent environmental backsliding without adequate justification.
1. Why the Sustainability Omnibus is substantively regressive
A measure is regarded as substantively regressive where it lowers the level of protection previously secured by the legal framework.Footnote 124 This may occur in several ways such as by reducing the number of actors subject to obligations, by lowering the intensity of those obligations, by weakening monitoring, reporting or review mechanisms, by removing enforcement tools; or by transforming binding duties into discretionary ones.Footnote 125 On that basis, the Sustainability Omnibus contains clear substantive regressive elements. The fact that the directives formally remain in force does not alter that conclusion. Non-regression is not confined to outright repeal.Footnote 126 A legal regime may regress even when its basic structure survives, provided that the amended framework no longer secures an equivalent level of environmental protection. In the Omnibus case, the original legislation sought to operationalise environmental and human rights protection through information, traceability, due diligence and accountability. The Omnibus leaves the architecture standing but hollows out several of its central pillars.
At the same time, not every amendment qualifies as regression. Rules may legitimately be updated in light of new scientific evidence, implementation experience or demonstrated regulatory failure. However, this is not the case with the Sustainability Omnibus since it was introduced before the original framework was fully implemented or properly evaluated as acknowledged by the Commission itself.Footnote 127 In those circumstances, the lowering of scope and intensity constitutes a regulatory rollback rather than an evidence-based recalibration.
2. From substantive regression to heightened procedural justification
The identification of substantive regression cannot, in itself, determine the legality of the measure. As mentioned above, non-regression, as it currently stands in EU law, cannot operate as an absolute prohibition on lowering environmental protection. Rather, it could function as a qualified constraint, giving rise to a presumption against backsliding.
Therefore, the critical question is how that presumption should operate in practice. The argument advanced in this article is that non-regression should be understood primarily as triggering heightened procedural requirements of justification. Once a measure is substantively regressive, it must be adopted under conditions that ensure that the retreat from previously attained levels of protection is informed and convincingly justified. As Bryner argues, non-regression operates as a principle of administrative discipline requiring reasoned decision-making for measures that weaken existing environmental protections.Footnote 128 Therefore, it serves as a safeguard against unjustified environmental backsliding rather than a mechanism for freezing environmental law in its existing form.Footnote 129
In that sense, the procedural shortcomings identified in the previous section acquire a different significance as they are not simply defects in administrative practice. They are problematic because they fall short of the level of justification that should be required where environmental protection is being reduced.
3. Operationalising non-regression: ex post and ex ante
Once understood in those terms, non-regression can be operationalised in two complementary ways.
The first is ex post, through judicial review. In principle, a legislative act that reduces environmental protection may be successfully challenged after adoption, especially where the reduction is insufficiently justified, unsupported by adequate evidence, or adopted in breach of procedural requirements.Footnote 130 In such cases, non-regression would not necessarily operate as an autonomous cause of action, but rather as an interpretative principle informing the application of existing grounds of review, especially proportionality, the duty to state reasons, legal certainty and good administration or substantive requirements imposed by the duty to uphold a high level of environmental protection.
However, the practical limits of this ex post pathway should be acknowledged. Direct actions for annulment under Article 263 TFEU remain constrained by restrictive standing requirements, making it difficult for NGOs or affected companies and individuals to challenge EU legislative acts directly.Footnote 131 It is therefore not particularly likely that non-regression will be operationalised primarily through direct annulment actions against legislative rollback. Judicial review may nevertheless arise indirectly through preliminary references under Article 267 TFEU, where national courts are asked to assess the validity of EU measures in the context of domestic proceedings. This may be particularly the case in Member States such as France and Belgium,Footnote 132 where the principle of non-regression has been recognised in domestic law, potentially making national courts more receptive to arguments invoking it and, in turn, more inclined to refer questions on the validity of EU measures under Article 267 TFEU.
For that reason, the more effective avenue lies ex ante, in the structuring of the legislative process itself. If EU law has not yet recognised a fully autonomous prohibition of environmental backsliding, it can nevertheless require that regressive measures be adopted only through procedures that are transparent, participatory and evidence-based.
From this perspective, non-regression operates as a principle of heightened procedural discipline. Once a proposal is identified as lowering the level of environmental protection, the Commission should be required to meet stricter justificatory standards. This does not mean that every regulatory amendment must be accompanied by an exceptional procedural burden. Rather, the relevant trigger is substantive regression, namely situations in which a measure narrows the scope of existing environmental obligations or reduces their intensity, enforceability, or effectiveness. In such circumstances, ordinary lawmaking safeguards should be strengthened.
Existing procedural safeguards provide an important starting point. However, as the Sustainability Omnibus illustrates, they do not adequately address the specific challenges posed by measures that reduce previously attained levels of environmental protection. Impact assessments may be bypassed, and public participation may be curtailed where urgency is invoked. Crucially, the Better Regulation framework does not clearly define the circumstances in which departures from ordinary procedural safeguards are justified.
A procedural understanding of non-regression does not necessarily require the creation of new procedural mechanisms but calls for the strengthening of existing safeguards whenever a proposal entails a reduction in environmental protection. In such cases, impact assessments and climate consistency assessments should not be bypassed and should include a specific statement of reasons identifying the extent of the reduction of protection, assessing its likely environmental consequences, and examining whether less regressive alternatives could achieve the same policy objective. In this way, non-regression could be operationalised as a procedural safeguard ensuring that any departure from previously attained levels of protection is transparent, evidence-based, and subject to enhanced scrutiny.
The Sustainability Omnibus illustrates why that ex ante approach is necessary. If non-regression is left to operate only after adoption, its practical force will be limited by barriers to judicial access and by the absence of an explicit doctrinal basis. By contrast, if it is understood as requiring heightened procedural discipline when environmental protection risks to be reduced, it can shape legislative behaviour before rollback is entrenched. This is especially important in a context in which crises are increasingly invoked to justify procedural shortcuts.
VI. Conclusion
To conclude, the Sustainability Omnibus marks more than a set of amendments to existing sustainability legislation and appears to signal the emergence of a new phase in EU lawmaking. The growing recourse to Omnibus packages reflects a broader transformation in the Union’s regulatory approach, increasingly shaped by concerns over competitiveness, flexibility and speed. Yet this legislative style also exposes a deeper tension within the EU legal order between these priorities and its constitutional commitment to principle of good administration.
Both substantively and procedurally, the Sustainability Omnibus departs from established trajectories. It narrows the scope and intensity of the CSRD and CS3D in ways that cannot be characterised as merely technical, while its preparation relied on procedural shortcuts that weakened evidence-based and inclusive law-making. These features make the Omnibus a compelling example of regression within the EU sustainability acquis, and an illustration of the risks associated with this emerging legislative paradigm.
Against this backdrop, non-regression emerges as a key normative framework for assessing the legality and legitimacy of such legislative change. The article argued that the significance of non-regression does not lie in operating as an absolute substantive prohibition but in structuring the conditions under which regressive change may occur. Since non-regression has not yet crystallised as a principle of EU law, its most immediate practical relevance is procedural. Once a measure is identified as substantively regressive, this finding should trigger a heightened duty of justification. The deeper the rollback, the stricter the requirements of evidence, transparency, participation and reasoning should become.
In that respect, non-regression can be operationalised in two complementary ways. Ex post, it may guide judicial review through existing grounds such as proportionality and the duty to state reasons. In practice, however, this route is limited by restrictive standing under Article 263 TFEU.
Its primary role is therefore preventive. When environmental protection is reduced, non-regression should trigger heightened procedural scrutiny, requiring a compete impact assessment, a clear climate-consistency assessment, meaningful consultation, consideration of alternatives, and specific justification. In this way, it does not freeze environmental law in its existing state, but it acts as a safeguard against weakly justified backsliding.
As the Commission announced that the 2025 omnibus packages were “only the start” and that the Better Regulation Guidelines will be amended,Footnote 133 the Sustainability Omnibus operates as a turning point. This forthcoming revision of the Better Regulation Guidelines provides a concrete opportunity to address the structural shortcomings revealed by the Omnibus adoption process and to embed safeguards capable of governing this emerging legislative technique. In particular, it should be used to articulate more clearly the conditions under which substantively regressive measures may be adopted. This includes, first, the need to define more precisely the notion of urgency. As crises become a recurring feature, the risk increases that urgency is invoked in a broad and politically malleable way. A revised framework should therefore distinguish between genuine emergencies and longer-term policy objectives and should specify the procedural consequences that follow from invoking urgency. Second, the revision offers an opportunity to integrate the logic of non-regression into the Better Regulation framework itself, by recognising that measures which lower the level of protection trigger heightened requirements of evidence, transparency, participation and justification. In this way, non-regression could evolve from an abstract emerging principle to a concrete structuring element of EU lawmaking.
As the EU continues to reconcile environmental ambition with economic competitiveness, the central challenge is not whether the sustainability acquis can evolve, but how such evolution is governed. A procedurally robust understanding of non-regression offers the means of ensuring that this new wave of legislative reform remains anchored in the Union’s foundational commitments to environmental protection, democratic accountability and the rule of law.
Funding Statement
Open access funding provided by Wageningen University & Research.
Open access
