Brussels, 14.4.2016

COM(2016) 204 final


Report from the Commission to the Council and the European Parliament under Article 18(2) of Directive 2004/35/EC on environmental liability with regard to the prevention and remedying of environmental damage

{SWD(2016) 121 final}
{SWD(2016) 122 final}

1.    Introduction

This second implementation report 1 on Directive 2004/35/EC on environmental liability with regard to the prevention and remedying of environmental damage ('the Environmental Liability Directive' or 'the Directive' or 'ELD') 2 presents the experience gained in applying the Directive between 2007 and 2013 and must be read in conjunction with a Staff Working Document 3 . The report is based on Article 18(2) ELD. It contains conclusions and recommendations on how to improve the implementation on the ground, based on the REFIT evaluation undertaken over the past two years. The evaluation assessed how the Directive works and whether it is “fit for purpose” 4 , backed up by two studies done in 2012 and three from 2013 5 .

The ELD addresses cases of significant environmental damage. It implements the 'polluter pays' principle which is set out in Article 191 of the Treaty on the Functioning of the European Union, meaning that the public should not pay if an industrial operation causes significant environmental damage. Operators carrying out occupational dangerous activities as listed in Annex III ELD (see section 2. below) are strictly liable for the environmental damage they caused, i.e. there is no need to prove fault (intent or negligence). Operators carrying out other than dangerous occupational activities are liable on the basis of fault.

The main objectives of the Directive are to prevent environmental damage if there is an imminent threat and to remedy it if it has already occurred. In line with the polluter-pays principle, the liable operator must take the necessary preventive or remedial action and must bear all costs. Damage is considered to be remedied once the environment has been returned to its pre-damage state. The ELD covers damage to biodiversity (protected species and natural habitats), water and land. Traditional damage (damage to property, loss of life and bodily injury or economic loss) is not covered by the Directive.

2.    Transposition and implementation

Several Member States failed to meet the deadline for transposing the Directive into national law (30 April 2007) and it has only been fully transposed by all 27 Member States 6 since mid-2010. The first Commission report outlined these delays and other shortcomings. In addition, the Commission analysed in detail the complex situation that was created by the introduction of a rather flexible framework EU instrument on environmental liability in addition to pre-existing national legislation in most Member States. In summary, this led to a certain degree of better coherence from a legal point of view in comparison to the situation before the ELD was transposed (e.g. some MS did not have any environment liability regime in place). However, the current situation remains in legal and practical terms diversified and further steps would be needed to establish a European level playing field 7 .

The Commission has taken action concerning late transposition and issues relating to non-conformity (seven Member States still need to resolve some non-conformity issues). These actions have slightly improved the coherence of the Member State legal systems in place. However, the diversity continues due to the framework character of the ELD, offering a considerable degree of flexibility to the Member States.

As regards the implementation, between April 2007 and April 2013, Member States reported approximately 1 245 confirmed incidents of environmental damage which triggered the application of the ELD. However, the number of cases varies greatly between Member States. Two Member States account for more than 86% of all reported damage cases (Hungary: 563 cases, Poland: 506) and six Member States reported most of the remaining cases (Germany (60), Greece (40), Italy (17) 8 , Latvia, Spain and the United Kingdom). Eleven Member States have reported no ELD damage incidents since 2007, possibly because they deal with cases exclusively under their national system.

The number of annual ELD cases per Member State varies considerably, from 95 to less than 1. A high number of cases in a Member State does not necessarily indicate that this Member State applies the ELD more strictly than required. The evaluation and the work on conformity-checking demonstrates that this divergence can be explained by different legal frameworks and traditions (in particular whether pre-existing legislation was repealed or not), possible differences in the state of the environment and different interpretations of key terms and concepts, as presented further below. Higher use of the Directive is often due to:

use of registers of ELD cases;

better opportunities for interested parties to submit comments and to engage with competent authorities;

competent authorities being obliged on a secondary basis to carry out preventive and remedial action if operators fail to do so; and

a higher level of awareness of the ELD of the public and among stakeholders, in particular among operators.

In contrast, the scope of the ELD has not been identified as a major reason for the differences in implementation (see section 4).


Around 50 % of the reported cases of environmental damage concern damage to land. Damage to water accounts for 30 % and damage to biodiversity for around 20 %.

Figure 1: Category of environmental damage based on 1450 cases 9

The dangerous occupational activities (linked to strict liability) causing environmental damage are mostly:

waste management activities;

treatment of dangerous substances, preparations, plant protection products or biocidal products;

activities under the Industrial Emissions Directive 10 and

transport of dangerous or polluting goods by road, rail, inland waterways, sea or air.

Other occupational activities (linked to fault-based liability) also caused environmental damage, but cause according to the ELD exclusively biodiversity damage.

 Figure 2: Reported ELD cases by number of cases according to the type of damaging activity

Relatively few requests for action were reported as having been initiated by people affected by environmental damage or by environmental NGOs 11 . Judicial review was sought in about 60 cases, 44 of these in Poland.

The average time from start to completion of remediation was 12 months. However, in some cases, remediation took longer than six years.

3. Evaluation and regulatory fitness 12

3.1. Relevance

The objectives of the Directive, namely to prevent and remedy environmental damage, are still relevant and match current needs because there are always risks of these types of damage and accidents. The evaluation shows that the Directive has helped improve the level of environmental protection in the EU to a limited degree. However, it has not yet fulfilled its potential. Moreover, as the ELD is more relevant in some Member States than in others, its relevance needs to be assessed against the diverse national legal frameworks and traditions already in place.

3.2. Effectiveness

The effectiveness of the Directive varies significantly because it has been implemented very differently across the Member States. This is partially caused by the framework character of the ELD, which provides for many exceptions, options and flexibility. In particular, the different interpretations and application of the 'significance threshold' for environmental damage has been identified as main reason for the uneven application of the Directive. Competent authorities, economic operators and insurers have frequently called for more clarity and guidance on this issue. The effectiveness of the Directive is best reflected in the amount of remedied environmental damage (turning around EUR 6 million without five major losses, and EUR 180 million if the five major instances are included). The incentivising effect, materialising in enhanced precautionary measures and better financial security, and the prevented damage due to immediate action in case of imminent damage, is largely unknown due to yet non-existing data.

As regards the scope of environmental damage, the evaluation confirmed that it is sufficient to cover damage to the most important natural resources. The scope of strict liability for hazardous activities (Annex III) also appears to be up-to-date, with the possible exception of off-site pipeline transport of dangerous substances. However, stakeholders have indicated issues with third party liability in this regard.

3.3. Efficiency

The efficiency assessment looked at the principal cost categories: remediation cost, administrative cost and financial security cost.

The remediation costs for the restoring of the damaged natural resources are to be borne by the person liable, in line with the polluter-pays principle. The available evidence shows that the cost of remedial action averages around EUR 42 000 13 . Greece reported a mean value of EUR 60 000. However, the remediation costs of individual cases range from a few thousand EUR to more than EUR 50 million for large-scale losses due to major accidents (such as in Kolontár, Hungary or in Moerdijk in the Netherlands).

The administrative costs for public authorities are the continuous costs which cannot be recovered from liable operators. Only three Member States provided precise data on administrative costs, ranging from EUR 55 000 (in the Flemish Region of Belgium) to EUR 2 million (in some of the autonomous communities of Spain). No data were provided as regards the administrative costs for businesses. Due to the limited information on administrative costs for authorities and the private sector, it is not possible to draw sound conclusions on the administrative costs. However, no stakeholders have complained about the administrative burden.

The costs of environmental damage for liable operators can be reduced through the use of financial security instruments (covering insurance and alternative instruments, such as bank guarantees, bonds or funds). Most markets provide sufficient cover for all ELD risks, but demand is low due to the insufficient number of cases in many Member States, lack of clarity regarding some ELD concepts 14 and slower emerging insurance markets 15 . A recent report suggests an increase in the amount of environmental liability insurance purchased by European companies (average annual increase of 13.6 % since 2006 particularly in the high risk sectors) 16 . Despite progress in financial security developments, problems persist regarding the application of the Directive to large-scale accidents and insolvency among liable economic operators.

It is not possible to make a meaningful overall assessment of the efficiency of the Directive based on the available information. More work will be needed including collection of additional data to overcome the shortcomings of this evaluation.

3.4. Coherence

The evaluation found that the Directive is generally consistent with other parts of EU environmental law and with the relevant international conventions listed in its Annexes IV and V. The Directive is satisfactorily integrated in the EU acquis, supplementing other environmental legislation, 17 in particular the Environmental Impact Assessment Directive, 18 the Industrial Emission Directive 19 and legislation on waste. A more detailed appraisal of some specific issues can be found in section 4 where there are also indications on the need of some possible further work, as well as in the annexed Staff Working Document.

3.5. EU added value

The assessment demonstrated again that there is some EU added value as regards setting a minimum level of environmental protection, in particular for cross-border circumstances (e.g. transboundary water pollution). The Directive also aims at ensuring a better level playing field in the internal market, but has not yet achieved that primary objective of EU harmonisation. While the Directive has increased the use of precautionary measures, for example by increasing the availability of environmental insurance cover on the market (see point 3.3), in practice, the shortcomings identified earlier prevent a higher EU added value.

4.    Specific points for review

This report includes a review of four specific areas applied under the Directive in accordance with Article 18(3) ELD.

4.1. Exception to the ELD scope of some international conventions (Annexes IV and V)

Some conventions of the International Maritime Organisation (IMO) and related conventions 20 are exempt from the scope of the ELD. These conventions apply world-wide to ship-owners whose vessels are registered in a State Party to the conventions, provide for strict liability for ship-owners, for mandatory financial security, 21 and, in the case of specific types of damage, for special compensation funds which pay any claims in excess of the ship-owners' liability. In particular, in the case of oil pollution from tankers, a three-tier system is in place to provide compensation for victims of oil pollution incidents. These international conventions prohibit Parties from introducing additional compensation claims.

The ELD and these international conventions apply different remediation standards, as the ELD effectiveness study for the Commission has shown. 22 This difference is potentially important depending on the individual case. The International Oil Pollution Compensation Funds (IOPC Funds) and the International Group of P&I Clubs also provided reports on the costs of preventive measures and environmental damage following all oil pollution incidents from tankers since 2002, which did not identify evidence of insufficient compensation of environmental damage.

Based on this information, the Commission will consider further exploring whether the different remediation standards can be addressed by non-legislative means, in particular by working towards a common understanding of concepts, for example through interpretation within the 'Claims Manual' of the IOPC Funds and/or within fora composed of the Parties to the Conventions.

The Commission has not encountered new circumstances which would call into question the exclusion of nuclear damage.

4.2. Application to genetically modified organisms (GMOs)

There were no incidents of environmental damage caused by GMOs 23 in the EU within the reporting period. The acquis is fully consistent with the requirements of the Nagoya-Kuala Lumpur Supplementary Protocol on Liability and Redress to the Cartagena Protocol on Biosafety of October 2010, so no further action at EU level was necessary to conclude that Protocol.

4.3. Application in relation to protected species and natural habitats

The study on biodiversity damage 24 looked into the potential for increasing the level playing field by extending the scope of the ELD to cover all 'protected species and natural habitats' in the Member States. This is because half of the Member States apply this broader scope 25 and the others limit their application to the scope defined by the Habitats Directive. However, extending the scope of the application of the ELD to include purely nationally protected natural habitats and species would pose legal difficulties (due to national competence) and may not improve harmonisation as national scopes may also differ widely.

The study found that:

the definitions of 'significant damage' and 'favourable conservation status' in both directives deserved more attention in the face of potentially different meanings and/or uses of the threshold of significant damage in the ELD and in Article 6(2) of the Habitats Directive;

there is a need for more consistent application and better clarification of the threshold of 'significant damage'. This was also pointed out by several competent authorities and by the insurance sector;

the geographical reference of the 'favourable conservation status' in the ELD (EU territory, national territory, natural range) has proved problematic. It should therefore be clarified that a site-related approach is required to ensure correct and effective implementation;

measures preventing biodiversity damage from becoming significant are often not taken on the incorrect assumption that preventive action can only be taken if it is known that the damage will be significant.

The Commission is currently analysing a policy option for a possible initiative on "no net loss" 26 which could also address these issues. The study is under finalisation and will be considered as input for the Commission when determining whether, and if so how, to address the above-mentioned aspects.

4.4. Possible inclusion of other instruments into Annexes III, IV and V 27

The Commission looked at the potential candidates for inclusion. However, the evaluation has not shown any need to extend Annexes III (see above under 3.2.), IV or V. 28

5.    Conclusions

The implementation of the ELD has improved the prevention and remediation of environmental damage to a limited extent in comparison to the situation before the transposition of the ELD. In particular, the Directive strengthened the polluter pays principle (thus avoiding significant costs for the public purse), implementing strict liability across the EU for environmental damage and raising the remediation standards for restoring damaged natural resources, in particular for damage to biodiversity. However, the degree of variability among the Member States regulations may be a cause for limited concern, although few concerns have been communicated to the Commission, and the evaluation has helped identify the main shortcomings.

As for the evaluation, one of the major challenges is the lack of data on ELD incidents and comparable incidents treated under national legislation. Other issues are the lack of awareness and information of stakeholders on the ELD, insufficient resources and expertise to implement the ELD, uncertainties and ambiguities about key concepts and definitions such as the 'significant' threshold, under-use of complementary and compensatory remediation and insufficient data about environmental damage, remediation cases and costs. While some of these aspects of the Directive and its context help explain the large variation in the number of reported cases and their content, the differences in the use of the 'significance threshold' together with pre-existing national frameworks is identified as the main reason for the significant divergences. Other factors contributing to the variable effect of the Directive include varying use of ELD registers, varying degrees of public participation, varying application of the subsidiary obligation of competent authorities to take action in the absence, or due to inaction, of liable operators, and varying levels of stakeholders' awareness.

On the positive side, implementation of the Directive continues to improve. The industry and other stakeholders contributing to the evaluation are largely satisfied with the current legal framework. Together with the Member State experts, they have argued strongly for continuity, predictability and legislative stability in relation to environmental liability. The Directive also stimulates preventive action and precautionary approaches, and thus guards against high remediation costs which often exceed the costs of preventive measures. However, it is difficult to quantify the benefits gained through prevention, in particular due to the lack of complete information on the total of preventive actions and other precautionary measures taken under the ELD.

Any future evaluation should assess the ELD together with the pre-existing national legislation to determine the extent to which a level playing field has been created. The playing field could be levelled further by clarifying some key terms (in particular 'significant damage'). To this end, the Commission has identified supportive actions at EU level which, combined with a number of recommendations to Member States, would constitute a proportionate response to help make the Directive more 'fit for purpose' than this evaluation has found it to be.

6.    Recommendations and next steps

Over the next years, the key priority for the Commission will be to promote alignment of national solutions and practices within the framework provided by the Directive and to identify on how this can be applied effectively and coherently in a wider legal liability framework. Efforts also need to be made to improve the evidence base on the actual impacts of the ELD, both on the environment and on relevant stakeholders. Doing so will require improvements in the collection of data on ELD cases in those Member States that have not yet established registers. Aligned national solutions (on such matters as remediation models, risk analysis, insurance calculation, etc.) could also produce useful results since available databases can be used to step up financial security (offer better-targeted products), enhance companies' risk management schemes and improve the knowledge of all practitioners and stakeholders, including the authorities responsible for damage abatement.

More information is needed to identify the precise reasons and impacts of the divergent solutions in Member States. More importantly, steps could be taken to build more administrative capacity and boost supporting implementation instruments. Furthermore, developing effective regulatory monitoring as set out in the Better Regulation Guidelines (SWD(2015)110) should ensure a more robust and conclusive evaluation of the Directive in the future. 29  It will therefore be necessary to discuss with all relevant stakeholders and the Member States how this evidence base can be improved.

The Commission services will therefore propose a multi-annual rolling work programme to Member States’ experts and stakeholders in order to improve the evidence base and help align national solutions. Moreover, the Commission will continue to provide administrative support measures, such as

(a) guidance or interpretative notices on key issues ('significance');

(b) training programmes; and

(c) helpdesks for practitioners (covering competent authorities, operators, loss adjusters, financial security providers, affected individuals, NGOs, etc.) providing information, assistance and assessment support for risk and damage evaluations.

To complement these efforts, the Commission recommends that all Member States undertake to:

support their implementation efforts with proactive initiatives (such as guidance documents, training, electronic tools for risk analysis, baseline setting, financial security models etc.) as some Member States have done already;

exchange administrative experiences and best practices and support each other in capacity-building efforts;

review their interpretation of key provisions of the Directive, in particular in relation to 'significance';

record data on ELD incidents and publish ELD registers if they have not done so already;

systematically gather the necessary data that can document that the application of the Directive in their country is effective, efficient and in line with the overall situation in the EU.

All this will help prepare for the next evaluation of the Directive. The Commission considers that these steps constitute a proportionate approach to improving the intended positive impacts of the ELD and will also help to produce the information that can be used in the next evaluation to then better demonstrate the scale of EU added value, efficiency and effectiveness of the Directive.


 The first implementation report was presented on 12 October 2010 (COM(2010) 581).


OJ L 143, 30.4.2004, p. 56 (as amended by Directive 2006/21/EC, Directive 2009/31/EC and Directive 2013/30/EU)


SWD (2016)121


See COM(2013) 685, COM(2014) 368, SWD(2014) 192 final/2 and COM(2015) 215, SWD(2015) 110


All studies available at:


Croatia, which notified the transposition of the Directive at the time of its accession to the EU, was not obliged to submit a national report on the ELD application by 30 April 2013 as it acceded the EU only on 1 July 2013.


More details of this analysis and the underlying studies is available in the ELD Implementation Study, Annex Part A – Legal Analysisi of the national transposing legislation, 2013, and in the Legal Analysis Study, 2014


However, Italy accounts for a further potential of 1000 incidents, many may become later ELD cases.


Some cases appear in more than one damage category


2010/75/EU (OJ L 334, 17.12.2010, p. 17)


132, but 93 of which were requests for action in Italy alone.


The Commission decided to carry out an evaluation under its REFIT programme in 2013 well before the guidelines for evaluation were adopted in May 2015.


Calculated on the basis of 137 cases representing just over 10% of all reported ELD cases by Member State and without considering in particular the three largest losses in Kolontár (Hungary), Moerdijk (Netherlands) and the Greek Assopos case (since they were considered as outliers).


In particular regarding the 'significance threshold'.


For instance the Baltic republics and smaller island Member States.


Stuart Collins, EIL market grows post ELD, in: Commercial Risk Europe, 27 November 2015.


 How the ELD tiers in with the Habitats Directive is summarised under point 4.3.


2011/92/EU (OJ L 26, 28.1.2012, p.1)


See footnote 10.


i.e. the Liability Convention or CLC, the 1992 Fund Convention, the Bunker Oil Convention, the Hazardous and Noxious Substances by Sea or HNS Convention and the Dangerous Goods Convention.


Having become part of EU law by means of Directive 2009/20/EC on the insurance of shipowners for maritime claims, OJ L 131, 28.5.2009, p. 128.


 See footnote 5.


Annex III.10 and 11 ELD.


 See footnote 5.


Belgium, Cyprus, Czech Republic, Estonia, Greece, Hungary, Latvia, Lithuania, Luxemburg, Poland, Portugal, Slovenia, Spain, Sweden, United Kingdom (England, Wales, Northern Ireland).



 'The report … shall include a review of: (d) the instruments that may be eligible for incorporation into Annexes III, IV and V.' (Article 18(3)(d) ELD).


These international instruments are either not likely to ever enter into force or are complementary to the ELD (see for further details section 6.4 of the SWD.


Which according to the Commission Communication 'Better regulation for better results – An EU agenda' (COM(2015) 215) should take place every 5 to 7 years.