Uncertainty in EU offshore safety regulation : an emerging experimentalist system potentially more effective than its Norwegian counterpart?

(1)

Uncertainty in EU offshore safety regulation: An emerging

experimentalist system potentially more effective than its

Norwegian counterpart?

Written by: Filip Ask von Ubisch Supervisor: Prof. Dr. Jonathan Zeitlin

Second Reader: Dr. Rosa Sanchez Salgado Submitted: June 2017

Master Thesis Political Science: European Politics and External Relations University of Amsterdam

(2)

Acknowledgements

The thesis would not have reached the same result without the help and support of several important people. First, I would like to express my gratitude towards my family and friends, who in various ways have supported me since the beginning. I would like to thank Dr. Sorraperra for his valuable comments on my work. I would also like to thank my good friend Mr. Houghton for helping me with grammar, spelling and language. I would also like to thank all the interviewees for taking their valuable time to share their insightful knowledge and expertise from the field. And last but not least, special thanks go to my supervisor Prof. Dr. Zeitlin, who has given me a lot of support and provided me with thorough and extensive feedback throughout the entire process.

(4)

Abstract

Regulation under uncertainty calls for an experimentalist mode of governance involving event reporting procedures, but such a mode or procedures are only weakly reflected in the Norwegian offshore oil and gas safety regulation, according to Sabel et al. (2017).

Meanwhile, new offshore safety regulation has been enacted in the EU after the catastrophic Deepwater Horizon accident. Therefore, this thesis examines whether the current offshore oil and gas safety regulation in the EU have adopted an experimentalist mode of governance compared to other modes, and whether event reporting procedures have been established. On this basis, the event reporting procedures in the EU and Norway are compared to assess which one potentially has the most effective offshore safety system. The findings reveal that the EU offshore safety regulation can be characterized as

emergently experimentalist, and its event reporting procedures are relatively more well-developed compared to those in Norway. The latter finding is explained by the difference in the approach of the authorities from the EU and Norway towards offshore safety, and the difference in the ability to include the industry in offshore safety forums. Consequently, the thesis concludes that the EU offshore safety system is potentially more effective than its Norwegian counterpart.

(5)

List of Acronyms

ALARP As Low As Reasonably Practicable AOC Acknowledgement Of Compliance CD Commission Decision 2012/C 18/07

CIR Commission Implementing Regulation (EU) 1112/2014 DG ENER Directorate General Energy

DMF Drilling Managers Forum

DWEA Danish Working Environmental Agency EU European Union

HLP Hydrocarbon Leaks Project

HMG Hierarchical Modes of Governance NCS Norwegian Continental Shelf

NMESA Norwegian Ministry of Employment and Social Affairs NMGSH New modes of governance in the shadow of hierarchy NOG Norwegian Oil and Gas

NOGEPA Netherlands Oil and Gas Exploration and Production Association NMPE Norwegian Ministry of Petroleum and Energy

OSD Offshore Safety Directive 2013/30/EU OSS Offshore Safety System

PAF Plug and Abandonment Forum PSA Petroleum Safety Agency QRA Quantitative Risk Analysis RoMH Report on Major Hazards SSM State Supervision of Mines WIF Well Integrity Forum WITF Well Incident Task Force

(6)

Introduction

Over the last decades, the offshore oil and gas industry has caused several catastrophic accidents, which have led to fatalities, polluted oceans and damaged economies. Such accidents include the Norwegian Alexander L. Kielland accident in 1980, where 123 people lost their lives after a platform collapsed; the British Piper Alpha accident in 1988, with a death toll of 167 persons due to an explosion; the Norwegian Ekofisk B accident in 1977, which caused the greatest oil spill ever in the North Sea by releasing over 80.000 barrels of oil due to a blowout; the Australian Montara accident in 2009, in which oil and gas leaked into the Timor Sea for over 10 weeks; and finally the US Deepwater Horizon accident in 2010, where an explosion killed 11 men, caused an oil spill of four million barrels into the Mexican gulf, and damaged the regional economy (EUOAG, 2015b). In sum, offshore oil and gas production is dangerous, and can have a devastating impact on societies, economies and the environment. In many of these major accidents, the absence of well-functioning offshore safety systems contributed to their occurrence, of which the Deepwater Horizon accident is one (EUOAG, 2015b). Based on the US investigation report, the European Union Offshore Oil and Gas Authorities Group (EUOAG) concludes that “a series of organizational and safety management failures” caused that accident (EUOAG, 2015b). Therefore, in a hazardous industry like this, well-functioning safety systems are crucial. Given the terrifying social, economic and environmental consequences of major accidents caused by the offshore oil and gas industry, offshore safety is a highly relevant topic to study.

The Norwegian Offshore Safety System (OSS), which involves the regulatory arrangements in place to ensure health and safety in the offshore oil and gas industry, has recently been criticized in an article by Sabel et al. (2017) among others. First, problems regarding the Norwegian industry`s safety performance are increasing (Sabel et al., 2017). Twice in the last 15 years, Norway has experienced near misses with potentially disastrous consequences. At the same time, industrial production processes are changing in various sectors and across borders, resulting in increasing levels of uncertainty (Ibid.). Offshore safety regulation under

(7)

uncertainty, these authors argue, calls for an experimentalist approach: a governance mode, in which the regulatory system is continuously reviewed and revised in light of experience (Ibid.). In practice, this implies the institutionalization of “event reporting procedures”, which are systems that i) detect errors in products or procedures that could lead to

disasters; ii) identify and correct the initial cause of the problem; iii) notify actors who might be in a similar situation; iv) ensure that preventive measures are carried out to maintain safety on ongoing operations, and finally, v) that the design standards of the relevant

products are upgraded (Ibid.: 2). However, efforts to institutionalize such systems in Norway have so far remained “fragile, checked, and undercut”, partly because the Petroleum Safety Agency (PSA) has been reluctant to actively support and facilitate its development (Ibid.: 18). This reaction, or inaction, is associated with the PSA`s regulatory approach relying on the “internal control doctrine”, which stresses the importance of self-regulation and strict liability for operators in case of damage (Ibid.: 9). Thus, in Sabel et al.`s (2017) view, the PSA`s regulatory approach encourages the authority to refrain from active involvement in developing event reporting procedures.

At the same time, extensive regulatory developments have taken place in the European Union (EU) in recent years. The EU has a substantial offshore oil and gas industry, in which over 600 platforms currently operate, and in 2012, oil and natural gas corresponding to over 120 million tons of oil in total were produced (EUOAG, 2015c). In the aftermath of the

Deepwater Horizon accident in the Mexico gulf in 2010, the EU have adopted a new offshore safety system (OSS). This involve three pieces of legislation such as the Directive

2013/30/EU, which the thesis will call the Offshore Safety Directive (OSD) hereafter (EUOAG, 2015a; OSD, 2013). Moreover, since Sabel et al. (2017) argue that the production

procedures, which have resulted in increasing level of uncertainty, have changed not only in Norway, but across sectors and national borders, there is great reason to believe that this also includes the EU. Therefore, the thesis assumes that the current EU OSS has emerged under increasing levels of uncertainty. Finally, almost no studies have yet examined the current EU OSS, as we will see later. Thus, little is known about either the regulatory arrangements in place or their governance mode.

(8)

Given similar levels of uncertainty in the EU OSS as in the Norwegian OSS, it appears highly interesting to examine whether the EU has responded to increasing uncertainty by adopting an experimentalist approach compared to other governance modes. More specifically, in light of the shortcomings identified in the Norwegian OSS by Sabel et al. (2017), it is interesting to examine whether the EU have responded to increasing uncertainty by institutionalizing its own event reporting systems. Based on a comparison of the

developments of event reporting systems, the thesis would subsequently be able to provide an indication of the potential effectiveness of the EU OSS relative to its Norwegian

counterpart. Finally, applying the same lenses to the EU OSS as Sabel et al. (2017) applied to the Norwegian case may also shed new light on the Norwegian OSS itself.

On this basis, the thesis addresses the following research question and three sub-questions:

In terms of governance modes, how can the EU offshore safety system be characterized? How potentially effective is the EU offshore safety system compared to its Norwegian counterpart?

1. What are the core features of the EU offshore safety system?

2. In terms of experimentalism, hierarchical governance and new modes of governance in the shadow of hierarchy, which governance mode is most applicable to the EU offshore safety system?

3. To what extent has the EU offshore safety system developed event reporting

procedures that could help to overcome shortcomings identified by external critics in its Norwegian counterpart?

To answers these questions, the thesis proceeds through the seven steps. First a literature review is presented, in which scholars have mostly sought to describe, analyse and compare the developments of oil and gas producing nations. Second, the theoretical framework follows, in which the governance modes, in terms of experimentalism, Hierarchical modes of governance (HMG), and New Modes of Governance in the Shadow of Hierarchy (NMGSH), will be described. Third, the thesis proceeds to the methodology section, which mainly

(9)

describes how the analysis will be conducted via two case studies. Fourth, a case description of the EU OSS follows, which focusses on its core features by studying relevant legislation. The fifth step involve a two-folded analysis. The first case study analyses the EU OSS analyzed through the lenses the of experimentalism, HMG and NMGSH, while the second case study compares the EU OSS to its Norwegian counterpart in light of article by Sabel et al. (2017). Sixth, a discussion is presented, which aims to both situate the findings within the existing literature presented in the literature review, as well as discussing the generalizability of the findings in light of empirical, theoretical and methodological limitations of the thesis. Finally, a conclusion follows, which summarizes the major findings and highlight pathways for future research within the field offshore safety.

Chapter I: Literature Review

Introduction

The following section will demonstrate that in the field of offshore safety systems (OSS), most scholars have been studying the development of various offshore oil and gas producing nations. In this regard, three aspects have been especially emphasised, which involve i) legislative developments in light of major accidents; ii) the parallel development of

regulatory approaches; and iii) the practical implications of these developments in national OSSs. At the same time, the literature has not made any significant efforts to characterize any of these national offshore safety systems in governance theoretical terms, i.e.

experimentalism. In particular, regarding the EU OSS, neither its development nor its governance mode have yet been studied. Since chapter VI will closely examine the

Norwegian OSS in light of the EU OSS, the Norwegian OSS receives relatively little attention in the following section, despite its major role in offshore oil and gas production.

Major accidents and legislative developments

It is not only in the EU, that major accidents are followed by regulatory review and revision, several studies show (Sutton, 2013; Paterson, 2011; Paterson, 2014; Weaver 2014a; Crawley 1999; Wang, 2002; Morgan et al., 2010; Baram, 2014). According Paterson (2011), in the UK, two major accidents were followed by two regulatory changes, which altered the British

(10)

regulatory approach over time. The first occurred already in 1965, when the Sea Gem drilling rig collapsed, which took 13 lives (Ibid.). The accident was followed by an inquiry, and in 1971 the “Mineral Workings (Offshore Installations) Act” was adopted (Ibid.: 374). The second accident was the Piper Alpha accident in 1988 (Ibid.). Again, an inquiry was

conducted, which resulted in an extensive report offering 106 recommendations (Ibid.). Four years after the accident, these recommendations came into law, when the “Offshore Safety Act” and the “Offshore Installation (Safety Case) Regulations” were adopted (Ibid.: 381).

According to Sutton (2013), also the U.S. regulatory system changed after the Piper Alpha accident. Based on a shared responsibility for safety policies between firms and the state, these two parties developed the “Safety and Environmental Management Program (SEMP)” in the beginning of the 1990s (Sutton, 2013: 311; Falker & Nickerson, 1996). In 2011, one year after the Deepwater Horizon accident, a new regulatory body, named the “Bureau of Safety and Environmental Enforcement” (BSEE), was established (Weaver, 2014b: 496). Shortly after, the BSEE changed the SEMP into the “Safety and Environmental Management System” (SEMS) (Sutton, 2013: 311-312). In a long article by Weaver (2014a), she claims that that not only did the regulation change, the industry itself also changed substantially. Based on available reports published in the aftermath of the accident, Weaver (2014a) argues that three changes have been significant. First, the industry recognized that if the industry is overly satisfied with its own safety system, this may lead to inattention towards potential hazards (Ibid.). Secondly, the technology has developed, thereby improving safety conditions (Ibid.). And finally, the sharing of best practice, in terms of “technical standards, safety management practices and procedures, and training requirements” across the industry worldwide, have increased (Ibid.: 205). The Deepwater horizon accident also led other countries, such as Australia, Canada and the UK, to review their own OSSs (Baram, 2014).

In contrast to the UK and the U.S., the Norwegian regulatory system appears to have

undergone less revision in relation to major accidents, following an article by Kaasen (1991). According to Kaasen (1991), while the Piper Alpha accident led to extensive revision of the British OSS, the Norwegian OSS remained intact. However, Hale (2014) argues that also Norway have experienced major accidents which have led to revision of the Norwegian OSS, such as the regulation adopted in the 1980s after the Alexander L. Kielland accident.

(11)

Regulatory approach

Based on these regulatory developments, many scholars believe that the regulatory approaches of most countries have shifted from command-and-control towards more goal-oriented approaches (Crawley, 1999; Wang, 2002; Paterson, 2011; Paterson, 2014; Steinzor, 2011; Morgan et al., 2010; Dahle, 1994, Kaasen, 2014). Baram & Lindoe (2014) stress that this shift should be seen in light of the historic and global debate between command-and-control regimes and other modes of governance, which was accompanied by discussions concerning governmental deregulation and outsourcing of public services to private actors. Following Crawley (1999), the British OSS took part in this shift, in which the British OSS before the Piper Alpha accident was characterized by command-and-control, in the sense of having prescriptive rules and regulation, while afterwards implemented a goal-oriented approach in connection with the adoption of the offshore installation regulations in the early 1990s (see also Paterson, 2011; Paterson, 2014). According to Paterson (2011), these

regulations entailed a new approach, which deserted the idea that government officials can provide detailed rules to a diverse industry that requires flexibility. Instead, the new

regulations provided the operator with considerable autonomy to decide how to ensure safe operations, following Wang (2002). In his words “The regulations set forth high-level safety objectives, while leaving the selection of particular hazard arrangements in the hands of the operator” (Ibid.: 84). In other words, the approach entailed more self-regulating, according to Crawley (1999). Following Paterson (2011), the shift towards a goal-oriented approach reflects the criticism against the previous command-and-control regime levied by the commission established after the piper alpha accident, which had the task of reviewing the British OSS. It argued that a command-and-control approach were likely to foster a

“compliance mentality”, in which operators were more concerned about following requirements than ensuring that operations were actually safe (Paterson, 2011: 380). Moreover, Baram & Lindoe (2014) criticize the command-and-control approach for being unable to follow the rapid developments happening in the industry; for not being able to fit diverse types of operations; the implementation and enforcement of the approach involve high costs; and finally, governments lack specialized expertise to implement and enforce the approach.

(12)

Multiple studies have identified a similar development in Norway (Kaasen, 2014: 105-106; Dahle, 1994; Morgan et al., 2010). Dahle (1994) argues that the Norwegian competent authority`s “supervisory strategies have gradually developed from detailed specific government control into system-oriented supervision”, implying a move from command-and-control to goal-setting requirements (Dahle, 1994: 381). In a book chapter comparing the Norwegian, the British and the U.S. offshore safety systems (OSS), Hale (2014) even argues that the Norwegian system is even more goal-oriented than its British counterpart. Other studies show that also Canada, Denmark and Australia have moved in the same direction (Morgan et al., 2010, Hayes, 2014). However, despite many institutional changes after the Deepwater Horizon accident, the U.S. OSS has maintained a prescriptive mode of governance, Baram (2014) argues. Consequently, the U.S. “stands out” among other oil and gas producing nations (Ibid.: 154).

Practical implications of regulatory approach

As a practical implication of the new goal-oriented approach, the British OSS implemented the so-called “safety case approach” (Paterson, 2014: 132; see also Crawley, 1999; Wang, 2002; Paterson, 2011; Morgan et al., 2010, Steinzor, 2011). According to Paterson (2011), this approach was introduced together with the 106 recommendations presented in the inquiry report after the Piper Alpha accident. The safety case approach entail that a safety case report shall be prepared by the operator, and accepted by the regulatory agency before operations can be executed (Ibid.). According to an extensive report by Morgan et al. (2010), the report shall primarily demonstrate that the operator has detected all hazards; assessed their risk; reduced these “As Low As Reasonably Practicable (ALARP)”; besides describing management systems and the installations in place, and finally, add a summary of the five-yearly review of the report itself (Morgan et al., 2010: 11). Several scholars contend that the ALARP principle is also applied in most other OSSs, such as the Norwegian and the U.S. (Baram & Lindoe, 2014; Morgan et al., 2010). Apart from the UK, also Canada, Australia and Denmark later applied the safety case approach (Morgan et al., 2010; Hayes, 2014). In Norway after 1985, the new regulatory approach involved the introduction of the internal control system (Bang & Thuestad, 2014). According to Kaasen (2014), in general, the system implies that the regulator only defines general norms, while firms shall both define their own specific target within these norms, as well as designing the methods through which

(13)

those targets shall be reached. In return for this autonomy, ultimate liability rest on the operator in case of damage, following Sabel et al. (2017). Besides the internal control system, also “tripartite collaboration” have been established, involving labor unions,

employer unions and national authorities (Bang & Thuestad, 2014: 255). Following Bang and Thuestad (2014), through multiple forums, such collaboration discusses and contributes to the development of regulatory arrangements. In a comparative perspective, Hale (2014) argues that the tripartite collaboration is a key feature of the Norwegian OSS, while labor unions are less present in the UK OSS, and nearly non-existing in the US OSS.

As a result of the continuation of a command-and-control approach in the US OSS, it maintains a reliance on prescriptive and detailed regulation, combined with “hard law enforcement” to ensure operator compliance (Baram, 2014: 154). According to Baram (2014), such regulation implies the authorities prescribe detailed rules and standards, describing how firms shall design and conduct operations. Furthermore, to ensure

compliance among companies, the BSEE conducts inspections on installations based on a “national checklist”, which contain 160 detailed requirements, while non-compliance is fined (Ibid.: 173). Returning to Hale (2014), the frequency of these inspections of facilities and management systems is significantly higher than in Norway and the UK. Concerning labor unions, although these are nearly absent in the offshore oil and gas industry compared to its Norwegian and British counterparts, as mentioned earlier by Hale (2014), employees have recently been empowered, according to Baram (2014). For example, employees present on the offshore installation now are allowed to terminate specific operations in case they fear that the situation poses an “imminent risk or danger to a person, the public, or the

environment” (Baram, 2014: 185). Nevertheless, Baram (2014) points out that important requirements are still being enacted without stakeholder consultations.

Despite the considerable criticism towards command-and-control approaches to offshore safety, also goal-oriented approaches have been criticized by Watson (2015) and Steinzor (2011), who argues that the latter approach is not feasible under U.S. conditions. Following Watson (2015), the safety case approach suits markets with few operators or where national companies are major shareholders. However, this is not the case with the U.S. offshore industry, because it is both diverse an encompass a relatively high number of operators

(14)

(Ibid.). Steinzor (2011) adds three more reasons why the U.S. should not import safety case regulation. First, since the safety cases are kept in secret from the public and often also the worker representatives, in sharp contrast to the U.S. OSS, introducing them would

undermine transparency (Ibid.). Second, the acceptable level of risk in safety case regulation is much higher than in U.S. regulation, in which the U.S. OSS is characterized as more

protective (Ibid.). Finally, the U.S. regulatory system does not have the necessary resources or political support to successfully enforce a safety case regime (Ibid.). In sum, Watson (2015) and Steinzor (2011) argue that a goal-oriented approach, such as the safety case approach, does not correspond with the U.S. economic, regulatory and political context.

In conclusion, this literature review has demonstrated that the current literature in the field of OSSs mainly have sought to describe, analyze and compare the developments of major oil and gas producing nations worldwide. Scholars have focused on legislative processes, which are often linked to major accidents (Sutton, 2013; Paterson, 2011; Paterson, 2014; Weaver 2014a; Crawley 1999; Wang, 2002; Morgan et al., 2010; Baram, 2014). Based on these developments, regulatory approaches evolved, in which most countries have moved towards goal-oriented approaches, while the US has maintained a command-and-control approach to offshore safety (Crawley, 1999; Wang, 2002; Paterson, 2011; Paterson, 2014; Steinzor, 2011; Morgan et al., 2010; Dahle, 1994, Kaasen, 2014, Baram, 2014). At the same time, both goal-oriented and command-and-control approaches have been criticized, especially the latter (Paterson, 2011; Baram & Lindoe, 2014; Watson, 2015; Steinzor, 2011). Consequently, while safety case regulation has emerged in countries such as Canada, the UK and Denmark, and the internal control system was implemented in Norway, the U.S. has continued to adopt prescriptive legislation combined with strict law enforcement, even after the Deepwater Horizon accident (Paterson, 2014, Crawley, 1999; Wang, 2002; Paterson, 2011; Morgan et al., 2010, Steinzor, 2011; Hayes, 2014; Bang & Thuestad, 2014; Kaasen, 2014, Sabel et al., 2017; Baram, 2014, Hale, 2014). On the other hand, and finally, one commonality among most nations is the application of the ALARP-principle, in which risk shall be reduced as low as reasonably practicable (Morgan et al., 2010; Baram & Lindoe, 2014).

(15)

So far, however, this review demonstrates that the literature on offshore oil and gas safety has not examined the EU OSS. It also does not explicitly deal with governance theory. Although the review has shown that the literature concerns the continuum from command-and-control to goal-oriented approaches, scholars have not yet made efforts to place these various OSSs into governance modes, such as Hierarchal Modes of Governance (HMG) (See Craig & De Burca, 2011). This observation, furthermore, resembles Bartolini`s point

regarding governance literature in general, which he characterizes as “predominantly

descriptive and normative” (Bartolini, 2011: 2). Since this thesis seeks to examine the EU OSS from a governance theoretical perspective, an evident gap in the literature appears to be present. The lack of attention towards the EU OSS is, nevertheless, understandable

considering that the OSD was adopted in 2013, implemented into national law in 2015, and that existing offshore operations are exempted from these laws until 2018 (OSD, 2013). In other words, the OSD is not fully in force yet. Since this imply that certain aspects of the EU OSS are still not yet observable, this thesis takes into account this limitation in relation to its scientific scope.

Chapter II: Theoretical Framework

Introduction

In order to examine the EU OSS properly, the thesis applies three different theoretical approaches within governance theory. This include Hierarchical Modes of Governance (HMG) and New Modes of Governance in the Shadow of Hierarchy (NMGSH) and

experimentalism (See Craig & De Burca, 2011; Héritier & Lehmkuhl, 2011; Sabel & Zeitlin, 2008). The selection of these approaches rest on two considerations. First of all, governance theory in itself should be valued. According to Bartolini, “Governance theory has great potential for analyzing phenomena of interest to scholars and practitioners of the European Union, domestic-global political linkages, transnational cooperation and different forms of public-private domestic exchange” (Bartolini, 2011: 5). Second, although there exists a functional need for event reporting procedures following an experimentalist approach, this does not imply an automatic, deterministic development towards an experimentalist approach (Sabel et al., 2017). At the same time, the literature review revealed both that

(16)

OSSs worldwide are highly diverse, and that little is known about the governance modes of any OSS. In particular, scholars have neither examined the EU OSS in terms of governance theory, nor in terms of regulatory development. In sum, considerable theoretical uncertainty exists around the EU OSS, and given that this thesis primarily focus on governance theory, more theories than only experimentalism should be applied to ensure a theoretically robust analysis. Later, in the chapter VII, the regulatory development of the EU OSS will be briefly assessed in the light of the regulatory developments described the literature review.

Hierarchical Modes of Governance

Hierarchical Modes of Governance (HMG) consist of multiple elements (Craig & De Burca, 2011). Firstly, policies are enacted “top-down”, implying that they are created by the central authorities (Craig & De Burca, 2011: 160). Secondly, the policies prescribe highly detailed rules, which imply that those affected by the policies are left with little room for discretion (Craig & De Burca, 2011). In other words, such policies are “prescriptive” (Craig & De Burca, 2011: 160). This also mean that the policies are mandatory, in the sense that they must be complied with (Craig & De Burca, 2011). In other words, they are “binding” (Craig & De Burca, 2011: 160). Furthermore, to ensure compliance, HMG typically rely on “legal enforceability” (Craig & De Burca, 2011: 160) On a more general level, specific HMG are often also characterized as “command-and-control-type regulation” (Craig & De Burca, 2011: 160). Bartolini (2011) characterizes such regulation as formulated and implemented by the central government and administration, respectively, while binding requirements and the use of “hard legal instruments” shall ensure compliance (Bartolini, 2011: 7). The great importance of the central authorities rest on the assumption that the people, in terms of “citizens, voters, consumers and taxpayers”, channel their demands through the political system, which in turn respond by adopting policies (Bartolini, 2011: 7). Thus, Bartolini`s (2011) characterization contains to a large extent the same factors mentioned by Craig and De Burca (2011), thereby consolidating this thesis` understanding of HMG.

New Modes of Governance in the Shadow of Hierarchy

Following Héritier & Lehmkuhl (2011), New Modes of Governance in the Shadow of

Hierarchy (NMGSH) is characterized by two general tendencies. First, governmental bodies delegate competencies to lower-level units, in terms of “sectoral experts”, especially in

(17)

policy fields characterized as “highly complex” (Ibid.: 49). In practice, an example is the empowerment of independent regulatory agencies at the member state and EU level, who are dealing with an increasing number of regulatory tasks (Ibid.). Second, governmental bodies rely increasingly on cooperation with private actors regarding complicated regulatory activities (Ibid.). Consequently, industries have been subject to rising levels of

“self-regulation” at both member State and EU level (Ibid.: 49). In addition, private and public actors are increasingly engaged in “co-regulation or co-production” (Ibid.: 49). Based on these empirical observations, Héritier and Lehmkuhl (2011) offer a twofold definition of NMGSH. They understand it as “(1) public-policy-making under the inclusion of private actors and/or (ii) public-policy-making outside the traditional democratic-representative governmental arenas” (Ibid.: 50).

In order to reach defined policy targets, NMGSH involve several instruments (Héritier & Lehmkuhl, 2011). Héritier & Lehmkuhl (2011) differentiate between soft and hard instruments and if they are used directly or indirectly. In general, “incentivization,

bargaining, persuasion, information and ‘naming and shaming’” constitute soft instruments, while command-and-control constitute hard instruments (Ibid.: 58). Incentivization provide positive and negative incentives in order to direct behaviour in certain ways (Ibid.).

Bargaining aim to achieve a certain policy target, through the “exchange of resources and positions between actors” (Ibid.: 57-58). Persuasion rely on the “power of arguments and reasoning”, to shape the behaviour of actors (Ibid.: 58). By sharing and publishing

information, the instrument of information seeks to influence performance (Ibid.). An example of this is naming of shaming, which aim to lead actors in the preferred direction, by publishing their performances (Ibid.). The final instrument is closely linked to the instrument of monitoring, given that monitoring provides information (Ibid.). Command-and-control instruments are conceptualized in line with Bartolini (2011), as prescriptive rules formulated by central authorities, and enforced via the threat of sanctions (Héritier & Lehmkuhl, 2011).

Within NMGSH, they find that soft instruments are used directly to achieve policy targets, while hard instruments are used mostly indirectly in cases where the soft instruments produce unsatisfactory results (Héritier & Lehmkuhl, 2011). Therefore, NMGSH rely mostly on soft instruments (Ibid.). Compared to HMG, which mostly use hard instruments, this

(18)

illustrate a clear difference between the two approaches. However, although command-and-control instruments are being rarely deployed, they are still available for use, according to Héritier and Lehmkuhl (2011). For example, the “shadow of hierarchy”, which implies the threat of intervention by the central authorities if lower-level units are unable to reach desired outcomes, “frequently looms large over the new modes and their instruments” (Ibid.: 59).

Héritier and Lehmkuhl (2011) explain the emergence of NMGSH by pointing to four

elements, and the first consists of several factors. Policy fields, such as “market integration and market correction”, have become “highly complex”, and new problems evolve regularly (Ibid.: 54). At the same time, interests seek to ensure that the public-policy-making process proceed in a trustworthy and unwavering manner (Ibid.). Under these conditions,

governments have found themselves less effective than independent regulatory agencies, because the latter are believed to possess more expertise, flexibility and higher tempo, as well as greater stability, since they are not involved in “the mainstream of legislative political decision-making” (Ibid.: 54). In sum, the belief in regulatory agencies have increased (Ibid.). Second, NMGSH have also emerged due to the industry’s resistance against legislation (Ibid.). This resistance is in turn caused by the fact that they “shun public intervention into their economic activities” (Ibid.: 55). As a result, self-regulation operating under the shadow of hierarchy or legislative threats that appear credible have emerged (Ibid.). Third, since member states have become more reluctant to formally delegate power to the classic EU institutions, while the EC advocates for empowering policy-making at the EU level, the increase in new modes of governance constitutes a compromise (Ibid.). The EC would ideally prefer “the Community Method”, which implies the adoption of “hard legislation” (Ibid.: 56). Therefore, they consider NMGSH the “second-best solution” (Ibid.: 56). On the other hand, member states lose their previously exclusive competencies over policy areas (Ibid.). Finally, NMGSH have emerged to strengthen member states` compliance with EU rules, in addition to improving the implementation of EU policies (Ibid.). Since the EU lack resources,

delegating power to regulatory networks of national agencies is believed to improve both the harmonization and performance of EU policies across the member states (Ibid.).

(19)

Experimentalist Governance

An experimentalist architecture consists of four core features, which Sabel and Zeitlin (2008) understand as functions, that can be achieved via different administrative constellations. The first feature is the establishment of “framework goals”, in which an example could be the lowering of Co2 emissions (Sabel & Zeitlin, 2008: 273). These goals are produced in a collaboration between member states and EU institutions, who also establish metrics in order measure the actor`s performances (Ibid.). Second, “lower-level units”, which involve “national ministries or regulatory authorities and the actors with whom they collaborate” decide themselves how to pursue the framework goals, in terms of developing own framework rules (Ibid.: 273). Therefore, experimentalism is firmly characterized by

subsidiarity, which is turn provides lower-level units with a high degree of autonomy (Ibid.). Thirdly, in exchange for this, lower-level units must report their results, based on the metrics agreed upon (Ibid.). Moreover, lower-level units shall also take part in some form of peer or non-hierarchical review, which implies a comparison of the performance of actors, who seek the same general target, but who deploy different type of measures to reach it (Ibid.). In case a lower-level unit is performing poorly in light of the common goals, this unit is

expected to take corrective action, which takes into account the lessons learned by its peers (Sabel & Zeitlin, 2012). Finally, both the original actors, who created the system, and the new actors, who may have entered the system at a later stage, regularly revise the

overarching goals, metrics and procedures (Sabel & Zeitlin, 2008). In sum, experimentalism involves establishing framework goals jointly; providing lower-level units with autonomy; ensuring reporting and peer review; and finally, revising the entire system from time to time (Ibid.). Together, these features foster recursive learning, which is a corner stone in

experimentalist governance (Sabel & Zeitlin, 2008; Sabel & Zeitlin, 2012). The term recursive means that experience and learned lessons from processes are taken into account when similar processes take place again (Sabel & Zeitlin, 2012). This creates an “iterative cycle” (Ibid.: 169).

Based on these four features, experimentalism can be implemented in various ways (Sabel & Zeitlin, 2008). There is, thus, no “one-to-one mapping” between the core features and institutional arrangements and procedures (Ibid.: 274). Consequently, reporting and

(20)

monitoring can be organized by various actors, with various frequency, and on various issues (Ibid.).

Experimentalist architecture is more likely emerge under two conditions, which are strategic uncertainty and a polyarchic distribution of power (Sabel & Zeitlin, 2008). Strategic

uncertainty means that policy makers do not possess the necessary information to achieve their stated goals (Ibid.). In other words, they are not able “to guide action in a particular domain” (Ibid.: 280). The second condition, a polyarchic distribution of power, means that “no single actor has the capacity to impose her own preferred solution without taking into account the views of the others” (Ibid.: 280). Subsequently, with increasing strategic uncertainty, problems, goals and solutions are usually developed through joint action and not by one single actor (Ibid.). In addition, identifying problems, and finding solutions typically occur during process itself, rather than being clearly defined in advance (Ibid.). These issues are effectively addressed by experimentalism and its recursive learning, since this enables continuous revision of framework goals and metrics in light of experience when circumstances change (Ibid.).

While experimentalism addresses the needs of changing conditions in policy-making, HMG, which relies on command-and-control methods founded on principal-agent theory, become increasingly problematic (Sabel & Zeitlin, 2008). Based on the same changes resulting from strategic uncertainty, principal-agent relations become both blurred and ineffective (Ibid.). Since problems needs to be addressed by joint actions, it becomes difficult to identify who is the principal and who is the agent (Ibid.). Moreover, since responses to problems needs to be revised continuously, command-and-control methods becomes unworkable (Ibid.). Consequently, in stark contrast to experimentalism, hierarchical modes of governance based on principal-agent theory is becoming outdated, the authors claim (Ibid.).

These arguments also imply that the way principal-agent theory characterizes accountability is becoming less relevant (Sabel & Zeitlin, 2008). In principal-agent theory, accountability in rule-making is ensured through the establishment of goals by the principal, while agents have been delegated the responsibility for achieving them (Ibid.). Subsequently, the agent is

(21)

held accountable for complying with the rules and goals established by the principal, and to ensure compliance, the agent`s activities are subject to sanctions and rewards (Ibid.). Thus, the consequences of strategic uncertainty for rule-making are clearly not taken into account in principal-agent theory (Ibid.). Therefore, experimentalism argues that accountability instead can be ensured via peer review (Ibid.). Although decisions are taken and

continuously revised by lower-level units, accountability can be ensured if these are required to justify their action in light of experience to each other (Ibid.). This accountability is

defined by the authors “as the justifiable exercise of discretion subject to peer review” (Ibid.: 307). To ensure the effectiveness of such accountability via peer review, experimentalism has established a so-called “penalty default” (Ibid.: 306). A penalty default is a rule, which applies in case actors are not willing to cooperate, in terms of “warranting to one another the information they disclose, and then acting on what they currently know to regulate their interdependence fairly and efficiently” (Ibid.: 306). The rule should be designed to make actors worse off compared to the option of cooperation (Ibid.). Thus, a penalty default intends to incentivize actors to cooperate (Ibid.). At the same time, the penalty default also creates “destabilisation regimes”, in which forced exchange of information reveal new

possibilities and solutions that in effect needs be to be addressed by incumbents (Ibid.: 308).

Although the penalty default might appear to resemble the shadow of hierarchy, that is not the case (Sabel & Zeitlin, 2008). As mentioned earlier, NMGSH as presented by Héritier & Lehmkuhl (2011), see the shadow of hierarchy as an evident threat hanging over lower-level units. Following Sabel and Zeitlin (2008), in the shadow of hierarchy, the intervention of the authorities would almost produce an equally good result than if the agreement was reached by the lower-level actors. In addition, in the shadow of hierarchy, the lower-level actors can estimate their own benefits of negotiating themselves or leave it to the authorities (Ibid.). In experimentalism, by contrast, a solution by the authorities is far from able to match the quality of a solution reached by the lower-level units themselves (Ibid.). Since an alternative solution produced by the authorities themselves would imply traditional hierarchical rule-making, which does not take into account the described consequences of strategic uncertainty, the outcome would serve as a “draconian penalty and an incalculably costly disruption” for lower-level units (Ibid.: 308). Therefore, experimentalist architecture does not involve the shadow of hierarchy, but instead deliberation “when hierarchy has itself

(22)

become a shadow” (Ibid.: 309). In conclusion, experimentalism is a distinct theoretical approach that differs from both HMG and NMGSH.

Chapter III: Methodology

The thesis seeks to answer the research questions via two case studies. Before describing each of them, a few remarks should me made concerning the overall research design.

In general, this entire research “emphasizes words rather than quantification in the collection and analysis of data”, and it should therefore be characterized as qualitative (Bryman, 2012: 714). Its approach is “inductive”, in the sense that the aim is to generate theoretical insights concerning the governance of offshore safety (Ibid.: 380).

Epistemologically, the thesis emphasizes the importance of “understanding” how actors interpret the field of offshore safety governance (Ibid.: 380). Finally, ontologically, the thesis works from “constructivist” point of view, believing that it is the interaction between actors, which produces outcome, rather than external forces and ideas from the outside (Ibid.: 380).

Single Case Study

In order to be able to characterize the EU OSS in governance theoretical terms, the following research design will be applied.

To gain in-depth knowledge on the current EU OSS, the thesis studies relevant EU legislation adopted after 2010 (See CD, 2012; OSD, 2013; CIR, 2014). Regarding the EU legislation, its design and how it is understood and practiced by relevant actors is included. In sum, these elements represent the EU OSS. However, it should be noted that the current EU OSS is also shaped partly by legislation and general practice adopted before 2010. For example, this involves the directives 92/91/EEC and 89/391/EEC, concerning working environmental safety (OSD, 2013: 68). This also regards regional and international agreements, such as “the

Barcelona Convention”, “the Offshore Protocol”, and UN conventions (OSD, 2013: 72). Thus, the case does not cover all regulatory arrangements in the EU OSS. The EU OSS will be examined through a case study (Bryman, 2012). A case study can be defined as “the detailed

(23)

and intensive analysis of one single case” (Ibid.: 66). This allows an in-depth study of EU legislation and administrative act from 2010 until 2017, as well as both EU-level and national level activities. At the same time, this design reflects an “case-oriented approach”, in which the objective is to gather a rich understanding of the case, which can contribute to

generalize generated insights beyond the EU OSS (Della Porta, 2008: 206).

Following Della Porta (2008), an important question to ask is “what is this a case of?” (Della Porta, 2008: 209). Based on the theoretical and empirical focus of this thesis, the EU OSS is primarily understood as a case of regulatory modes of governance in the EU- potentially a case of experimentalist modes of governance in the EU. Secondary, the EU OSS is also understood as a case of OSSs worldwide.

This research design extensively use semi-structured interviews (Bryman, 2008). They base on a list of relatively open questions, but these are necessarily not followed in a slavish manner, thereby allowing the interview to adjust based on the replies from the interviewee (Ibid.). Since the thesis seeks to assess how relevant actors understand relevant EU

legislation, among other issues, the flexibility provided by semi-structured interviews entail a great advantage. The interviewees were selected based on “purposive sampling” (Ibid.: 458). Thus, in light of this research topic, interviewees were picked based on their perceived relevance. Interviewees were also selected to reflect various perspectives, in order to apply “triangulation” (Tansey, 2007: 766). Triangulation is a method, where information is

controlled through several sources, which aims to enhance the robustness of results (Ibid.). In other words, triangulation contributes to increase the credibility of the findings (Bryman, 2012).

Besides interviews, the case study also entails a form of “qualitative content analysis” of relevant documents, such as laws, regulations and official statements (Bryman, 2012: 557). Such an analysis seeks to extract general implication from text and documents, and is thus a method for interpreting complex material (Ibid.). How content analyses are conducted varies, but they usually involve quotes intended to reflect upon to larger theme (Ibid.). Given this flexibility, this case study closely examines documents, and illustrates findings by

(24)

through qualitative content analysis, triangulation becomes more extensive, and results should be considered even more robust (Tansey, 2007).

In total, 9 interviews were conducted. One interview was conducted over e-mail, three via phone, and five interviews were conducted through face-to-face in Oslo, Brussels,

Copenhagen, and The Hague. The interviewees were: the Danish Working Environment Agency (DWEA), where I met simultaneously with three officials (DWEA, 2017); the

Directorate General Energy (DG ENER), where the same official was interviewed twice (DG ENER, 2017a; DG ENER, 2017b); The Norwegian Petroleum Safety Agency (PSA),

represented by previous the Director on Legal Affairs, and currently Special Adviser on Legal Affairs, Anne Vatten (PSA, 2017); the Norwegian Ministry of Petroleum and Energy (NMPE), in which two separate interviews with two officials were conducted (NMPE, 2017a; NMPE 2017b); the Netherlands Oil and Gas Exploration and Production Association (NOGEPA), represented by Secretary Operations/Health and Safety and General Deputy, Gert-Jan Windhorst (NOGEPA, 2014; NOGEPA, 2017); and, finally, the Dutch State Supervision of Mines (SSM), in which two separate interviews with Senior Inspector, Anton van Gulik, and Specialist Inspector Vincent van Claessens were conducted (SSM, 2017a; SSM, 2017b).

The reasons for interviewing the national agencies, such as the DWEA and the SSM, was because these are responsible for implementing and enforcing the OSD, and besides are participating in the European Union Offshore Oil and Gas Authorities Group (EUOAG) and the North Sea Offshore Authorities Forum (NSOAF) (DWEA, 2017; SSM 2017a; SSM, 2017b; EUOAG, 2015g, PSA, 2013). The NMPE and the PSA were interviewed to shed light on Norwegian authorities` view of the OSD in relation to their own OSS. Since the NMPE has been responsible for the Norwegian official position towards the OSD, following an e-mail from the Norwegian Ministry of Employment and Social Affairs (NMESA), while the PSA is responsible for monitoring offshore and onshore operations, as well as participating in the EUOAG as an observer, they were considered as highly relevant for the thesis (NMESA, 2017; PSA, 2017, PSA, 2017b). Since the DG ENER is the branch of the European Commission (EC) responsible for drafting the OSD, and is now responsible for supervising the implementation of enforcement of the OSD, as well as organising the EUOAG, interviewing the DG ENER was of great relevance (DG ENER, 2017a; DG ENER, 2017b). Finally, by interviewing NOGEPA,

(25)

which represents oil and gas producers in the Netherlands, and who are directly affected by the OSD, the thesis also cover an important industry perspective (OSD, 2013; NOGEPA, 2017; NOGEPA, 2014).

The research design draws on information from all the above-mentioned interviews, except for those with the NMPE. Furthermore, the central documents subject to the qualitative content analysis are mostly EU legislation. This involve mainly the Offshore Safety Directive 2013/30/EU (OSD) (OSD, 2013); the Commission Decision 2012/C 18/07 (CD) (the

abbreviation ‘CD’ will be used onwards) (CD, 2012); and the Commission Implementing Regulation (EU) 1112/2014 (CIR) (the abbreviation ‘CIR’ will be used onwards) (CIR, 2014).

Comparative Case Study

Besides seeking to examine the EU OSS through the lenses of experimentalism, HMG and NMGSH, the second research design seeks to enable the examination of the EU OSS in relation to the Norwegian OSS. More specifically, it seeks to enable the assessment of whether and to what extent the EU OSS seems to address the weaknesses identified in the Norwegian OSS by Sabel et al. (2017). Therefore, the two systems are studied based on a comparative design, which is defined as a “comparison of two or more cases in order to illuminate or generate theoretical insights as a result of contrasting findings uncovered through the comparison” (Bryman, 2012: 711). To shed light on the relation between the two systems, semi-structured interviews have been used, in which the interviews with the NMPE, the PSA and the DG ENER have been especially relevant (Bryman, 2008; NMPE, 2017a; NMPE, 2017b; PSA, 2017; DG ENER, 2017a; DG ENER, 2017b). Furthermore, again qualitative content analysis has been used to interpret text and documents, which mainly involve the CD (2012), the OSD (2013) and various internet sources (Bryman, 2012).

Together, these two methods also seek to triangulate, and, thereby, increase the robustness and credibility of the subsequent findings (Tansey, 2007; Bryman, 2012).

After having described and discussed the upcoming two case studies, the thesis proceeds to the case description of the EU OSS.

(26)

Chapter IV: Core Features of the EU Offshore Safety System

Before entering the analysis, it is necessary to briefly describe the European offshore oil and gas industry, as well as the relevant legislation in the EU OSS, and their core features.

The production of offshore oil and gas in the EU is distributed highly uneven across member states. While over 80% of the production is located in the North Sea, involving countries such as the UK and Denmark, correspondingly less is produced in Mediterranean waters, the Baltic Sea and the Black Sea, and many member states do not have any production at all (EUOAG, 2015e). At the same time, Norway is a major offshore producer in Europe (Ibid.). For example, in 2012, Norway produced both more oil and gas than rest of the EU industry in total (Ibid.). To improve the safety of the EU industry in the aftermath of the Deepwater Horizon accident, the EU adopted the CD, the OSD and the CIR (CD, 2012; OSD, 2013; CIR; 2014).

Starting with the latter, the OSD primarily aims to prevent “major accidents” from occurring in the offshore industry (OSD, 2013: 73). Major accidents comprise incidents involving fatalities (or incidents which potentially could lead to fatalities), severe human injury, or major environmental accidents, resulting from serious damage to the installation, caused by incidents such as explosions or fires (Ibid.). Consequently, a “major hazard” entail a

situation, which could potentially lead to a major accident (Ibid.: 75). The OSD covers both single and combined operations, in which the latter means “an operation carried out from an installation with another installation or installations for purposes related to other installation(s)”, in addition to well operations defined as “any operation concerning a well that could result in the accidental release of materials that has the potential to lead to a major accident” (Ibid.: 75). In order to lower the risk major accidents, the OSD introduces the Report on Major Hazards (RoMH), which shall be prepared by the operator and then submitted to the competent authority (CA) (Ibid.). While the operator is defined as “the entity appointed by the licensee or licensing authority to conduct offshore oil and gas operations, including planning and executing a well operation or managing and controlling the functions of a production installation.”, the CA is a public authority on the national level,

(27)

which in part is responsible for examining and accepting the RoMH (Ibid.: 74). The CA is also responsible for supervision of operators, through irregular inspections, for example (Ibid.). Regarding the RoMH, the operator shall primarily demonstrate that potential major hazards have been detected, their likelihood and effects assessed, and risk reduced, following the ALARP-principle, and describe how the facilities and management systems are ensuring safe operations (Ibid.). The exact information requirements are listed in annex I, point 2 and 5 (Ibid.: 90-93). In sum, the RoMH concern both risk assessment and risk management (Ibid.). The same report shall also contain a “corporate major accident prevention policy”, a

description of the “safety and environmental management system”, a scheme concerning independent safety verification; and finally, an “internal emergency response plan”, following §11 (Ibid.: 80). The independent verification entail that an external body shall verify whether the installation is fit for production in accordance with regulation, pursuant §17 (Ibid.: 82-83). In total, the RoMH consists of five documents.

Apart from the OSD and the RoMH, the CIR has created what this thesis characterizes as a common Incident Reporting System (IRS) (CIR, 2014). In case of certain incidents, such as a “failure of a safety and environmental critical element”, operators shall report these to the CA, which, in turn, pass this information on to the EC (OSD, 2013: 105; CIR, 2014). Such elements are defined as “parts of an installation, including computer programmes, the purpose of which is to prevent or limit the consequences of a major accident, or the failure of which could cause or contribute substantially to a major accident” (OSD, 2013: 75). Furthermore, the IRS contains a common and mandatory reporting template for both operators and CAs, in which all data from the IRS is standardized (CIR, 2014). Hence, the IRS “shall make it possible to compare information from competent authorities”, and make it possible for competent authorities to compare performance between operators (OSD, 2013: 105).

Finally, the CD has established the European Union Offshore Oil and Gas Authorities Group (EUOAG), which is a multilateral forum, where interested member states, stakeholders and the EC discuss regulatory issues within the EU OSS and recent major accidents (CD, 2012; DG ENER, 2017a). These issues relate primarily to the OSD, following the officials from the DG ENER, the SSM and the DWEA (DG ENER, 2017a; SSM, 2017b; DWEA, 2017). The aim is both

(28)

to improve the overarching regulatory regime as well as helping member states to improve their application of the regulatory regime (CD, 2012). The EUOAG typically meets two or three times per year, depending on the need, and the DG ENER organizes three kind of meetings, in terms of ordinary and plenary sessions, as well as workshops (DG ENER, 2017a; EUOAG, 2015f). The ordinary sessions are only composed of the national authorities, where every member state can bring a maximum of two delegates (DG ENER, 2017a). So far, usually two delegates from 16 member states participate (DG ENER, 2017b). Second, the EUOAG also set up plenary meetings, which include relevant stakeholders, such as the industry, labour unions, and environmental organizations (DG ENER, 2017a). So far, the plenary sessions have included international organizations such as the International Organization of Oil and Gas Producers (IOGP), the International Association of Drilling Contractors (IADC), and the International Marine Contractors Association (IMCA), which often raise the total number of participants above 50 (DG ENER; 2017a; DG ENER, 2017b). The EC does not invite individual firms or national associations, which would have made the group even larger, for fear of reducing its effectiveness (DG ENER, 2017b). In both sessions, the dialogue is typically based on scientific reports and presentations from different member states and

stakeholders (DG ENER; 2017a). Finally, the EUOAG also organizes workshops, where member states with less experience can learn how to deal with more in-depth technical issues (Ibid.).

In sum, this section has shed light on the North Sea concentrated European offshore oil and gas production, as well as the OSD, the CIR, and the CD, in terms of the RoMH, the IRS, and the EUOAG.

Chapter V: Experimentalism in the EU Offshore Safety System?

Framework goals

The first element of experimentalist architecture entails the joint establishment of

framework goals established by central and lower-level units (Sabel & Zeitlin, 2012). In line with the examples of framework goals given by Sabel and Zeitlin (2012), the OSD contain a clear, but broadly defined main goal, which is to improve environmental protection in

(29)

offshore oil and gas production (OSD, 2013). As it states, “The objective of this Directive is to reduce as far possible the occurrence of major accidents” in order to increase “the

protection of the marine environment and coastal economies against pollution” (OSD, 2013: 66). The same goal is also stated nearly identically in the Commission Decision 2012/C 18/07 (CD), which says that “The Union policy aims at reducing the occurrence of major accidents related to offshore oil and gas activities and to limit their consequences, thus increasing the protection of the marine environment and coastal economies against pollution” (CD, 2012: 8). In this regard, both the OSD and the CD reflect upon experience by referring to the Macondo accident, and the OSD stresses that major accidents are “likely to have devastating and irreversible consequences on the marine and coastal environment as well as significant negative impacts on coastal economies” (OSD, 2013: 66; CD, 2012). Thus, EU legislation on offshore safety seems, at least, aware of the past, thereby already indicating a focus on learning from experience, which constitute a central aspect of experimentalism (Sabel & Zeitlin, 2008). Another noteworthy goal presented in the CD, is the importance of peer review to foster effective offshore safety governance (CD, 2012). As the CD states “the continuous exchange of experience, identification of best practices among regulatory and the industry and the improvement of implementation measures are being recognised as key aspects of a well functioning regulatory regime” (Ibid.: 8). From an experimentalist point of view, this goal is promising, since experimentalism also values the use of peer review to ensure accountability and recursive learning (Sabel & Zeitlin, 2008). Yet another important, but less explicit goal, is to provide the competent authorities and operators with

considerable autonomy. For example, the OSD notes that best practice in offshore safety systems rely on a “goal-setting approach”, and that the directive only seeks to establish “minimum requirements” (OSD, 2013: 69/73). Also, the national competent authorities see the OSD as aiming to give operators leeway in finding their own ways of ensuring safe operations (DWEA, 2017, SSM, 2017a). Therefore, providing autonomy seems to be a central aim in the OSD. This is relevant, since this goal not only indicates that the EU OSS establishes broad framework goals, but also that it seeks to follow the second experimentalist feature, which stresses the need for lower level units to be able to pursue framework goals as they prefer (Sabel & Zeitlin, 2012). Apart from these goals, the OSD also introduce several

principles concerning how the EU OSS should be governed. Regarding risk management, for example, the OSD require the operator to follow the ALARP-principle, which means that risk

(30)

should be reduced as low as reasonably practicable, as mentioned before (OSD, 2013; Morgan et al., 2010). And regarding monitoring and supervision, the OSD requires the actions by competent authority to reflect ”transparency, consistency, proportionality and objectivity” (OSD, 2013: 98). Based on the interview with the DG ENER, these goals and principles also seem to have been established by joint collaboration between the EC, European offshore oil and gas producing countries, and relevant stakeholders (DG ENER, 2017a). During the drafting of the OSD, the DG ENER consulted countries such as the Netherlands, Denmark, Norway, and last, but not least, the UK, in addition to the industry and labour unions, among others (Ibid.).

In sum, while referring to previous experience, The EU OSS set broad framework goals, thereby indicating an emphasis on the importance of learning from experience. Its main objective is to ensure the protection of the environment against major oil and gas accidents (OSD, 2013). Furthermore, these goals are likely to have derived from concerted deliberation between the EU, national and private institutions, through extensive consultations in the policy-making process. In conclusion, the first experimentalist feature is considered to be firmly present.

Lower-level autonomy

Although the OSD stresses the importance of autonomy for lower-level units, is this

objective also reflected in the regulatory arrangements in place? This question is essential in the second feature of experimentalism, which require that “local units are given broad discretion to pursue these goals in their own way” (Sabel & Zeitlin, 2012: 169-170).

Starting by considering the level of autonomy for the member states, the OSD`s content corresponds with the idea of only establishing minimum requirements (OSD, 2013). First, the member state appears to be free to select which public authority that should function as competent authority (Ibid.). This autonomy is reflected by the diverse group of appointed CAs. Whereas Denmark appointed the Danish Working Environment Agency (DWEA), which also covers sectors such as agriculture, transport and education, together with the Danish Environmental Protection Agency (EPA), the Netherlands only appointed the State

(31)

(b); SSM, 2017a; SSM, n.d. (e)). Moreover, when and how the required reviews of the CA`s effectiveness shall be conducted, or under what circumstances taking corrective action is deemed necessary, is left to the member state itself to decide when examining the OSD (OSD, 2013). Even in the case of non-compliance, the competent authority, formally, retains its autonomy, following the DG ENER official interviewed (DG ENER, 2017a). If the

competent authority does not implement or apply the OSD correctly, no EU institutions have the powers to take over its tasks (Ibid.). Instead, infringement procedures may be launched (Ibid.). Furthermore, given the highly uneven distribution of offshore oil and gas production in the EU, member states with minimal or no oil and gas production, are obliged to

implement only a limited part of the OSD (Ibid.). In sum, these examples indicate that the OSD constitute a framework, at least concerning central parts of the directive, in which member states have evident room for manoeuvring (OSD, 2013).

Shifting focus towards the Competent Authority (CA) and the Report on Major Hazards (RoMH), the latter seems to involve room for discretion on the part of the CA. Although the OSD establish several requirements regarding the content of the RoMH, and require the CA to assess it, how the competent authority is actually planning to assess the RoMH is

seemingly up to the authority itself to decide (OSD, 2013). This can be illustrated by comparing the practice of the SSM and the DWEA. In the Netherlands, the SSM partly assesses the RoMH with the help of a traffic light system (SSM, 2017a). For each operator requirement, the SSM gives a score (Ibid.). This implies that if the operator complies, partly complies or does not comply with the requirements, the SSM, gives the operator,

respectively, a green, orange or red light (Ibid.). Whether corrective action is needed by the operator, as well as assistance by the operator, depends on the color of the light (Ibid.). So far, this procedure has been applied to assess two companies, namely Seefox and Mearsk (Ibid.). Furthermore, the assessment period shall last for a maximum of 12 weeks (Ibid.). The first assessment by the SSM shall take only four weeks, in order to give both itself and operators the potentially necessary time to adjust and carry out a second assessment within the time limit (Ibid.). In contrast, the DWEA uses neither the same traffic light system nor the same time frame (DWEA, 2017). For some reason, the DWEA, would not disclose much information about their own assessment procedure, other than the fact that they manually assess the RoMH with a checklist approach, based on earlier practice (DWEA, 2017).

(32)

Moreover, the assessment period depends on the given case (Ibid.). The DWEA looks cursorily through the RoMH within 14 days, and in case of missing documents, a dialogue is initiated between the DWEA and the operator (Ibid.). As soon as all documents are in order, the assessment begins (Ibid.). In sum, parts of the assessment procedure relating to the important RoMH, and the assessment period vary between the two competent authorities, which demonstrate that competent authorities have substantial autonomy to pursue similar ends by different means. This finding is further supported when considering inspections by competent authorities, which seeks to ensure that operators have all the required

documents in order, and that their information correspond with the actual conditions on the installations (DWEA, 2017). According to the DWEA, they decide for themselves to what extent they should focus on the operator`s management system compared to visiting actual installations offshore, as long as these priorities are risk-based, following the OSD §21 (DWEA; 2017; OSD, 2013: 84). The DWEA interprets risk-based oversight in terms of focussing inspections towards those hazards, which represent the greatest risks (DWEA, 2017). Similarly, the DWEA also decides on the number of inspections, although they have a general practice, which entail that the DWEA visits all staffed installations at least once every year (Ibid.).

Finally, the industry also seems to enjoy a great deal of autonomy in the OSD. Concerning the RoMH, the OSD does not prescribe how the operators need to detect, assess and reduce risks (OSD, 2013). For example, following annex I in the OSD, the RoMH “shall include an assessment of oil spill response effectiveness”, without mentioning a single word about how to demonstrate this requirement (Ibid.: 90). Instead, the OSD, for example, encourage operators to “prepare and revise standards and guidance on best practice”, following §19, sub-paragraph 7 (Ibid.: 84). Therefore, the OSD provides the operator, at least in regard to the RoMH, with considerable autonomy. In accordance with the OSD, also the SSM and the DWEA give operators freedom in relation to the RoMH. For example, regarding risk

assessment, the DWEA leaves it open to operators to decide whether to assess risk with either quantitative or qualitative approach (DWEA, 2017). Furthermore, both the DWEA and the SSM suggest ways for operators to identify, assess and reduce risks, via providing

guidelines and standards (SSM, 2017a; DWEA, 2017). These documents explain how

Uncertainty in EU offshore safety regulation : an emerging experimentalist system potentially more effective than its Norwegian counterpart?