Tilburg University
Dimensions of technology regulation
Goodwin, M.E.A.; Koops, E.J.; Leenes, R.E.
Publication date:
2010
Document Version
Publisher's PDF, also known as Version of record Link to publication in Tilburg University Research Portal
Citation for published version (APA):
Goodwin, M. E. A., Koops, E. J., & Leenes, R. E. (Eds.) (2010). Dimensions of technology regulation. Wolf Legal Publishers (WLP).
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal Take down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
DIMENSIONS OF TECHNOLOGY REGULATION
Dimensions of Technology Regulation
Conference proceedings of TILTing Perspectives on Regulating Technologies
Morag Goodwin, Bert-Jaap Koops, Ronald Leenes (eds.) ISBN: 978-90-5850-511-8
Production: Simone Fennell
Published by:
aolf Legal Publishers (WLP) P.O. Box 31051 6503 CB Nijmegen The Netherlands Tel: +31 24 355 19 04 Fax: + 31 24 355 48 27 E-Mail: info@wolfpublishers.nl www.wolfpublishers.nl
Disclaimer: Whilst the authors, editors and publisher have tried to ensure the accuracy of this publication, the publisher, authors and editors cannot accept responsibility for any errors, omissions, misstatements, or mistakes and accept no responsibility for the use of the information presented in this work.
Table of Contents
1. Introduction. A Dimensions Approach to Technology Regulation ... 1
Morag Goodwin 1.1. Introduction: a Multi-dimensional Analysis ... 1
1.2. The Dimension of Technology Type ... 5
1.3. The Dimension of Innovation ... 8
1.4. The Dimension of Time ... 11
1.5. The Dimension of Regulation Type ... 13
1.6. The Multi-dimensional Discipline of Technology Regulation? .... 16
References ... 17
2. Calculable Risks? An Analysis of the European Seveso Regime ... 21
Esther Versluis, Marjolein van Asselt, Tessa Fox and Anique Hommels 2.1. Introduction ... 21
2.2. The Seveso Regime ... 22
2.3. The Seveso Regime and the Positivistic Risk Paradigm ... 24
2.4. The Dutch Regulatory Practice ... 29
2.4.1. Conceptualization of Risk ... 29
2.4.2 Transposition ... 30
2.4.3 Enforcement ... 31
2.4.4 Risk Calculations ... 32
2.5. Conclusions and Recommendations ... 33
Annex I ” Interviews ... 35
References ... 36
3. Facebook and the Commercialisation of Personal ... Information: Some Questions of Provider-to-User Privacy ... 39
Jennifer Hendry & Kay Goodall 3.1. Introduction ... 39
3.2. Access and Control ... 41
3.3. User Content and Intellectual Property ... 42
3.4. Facebook Photos: Why All the Fuss? ... 43
3.5. ‘Facebook Ads’ ” The Conundrum of Targeted Advertising ... 48
3.6. Cyberspace and Problems of Legal Challenge... 52
3.7. Limiting the Licence? ... 56
3.7.1. ‘You can check out any time you like…’ ... 56
3.7.2. Possible Alternatives ... 58
4. How Can Hybrid Nanomedical Products Regulation Cope with
Wicked Governability Problems? ... 63
Bärbel Dorbeck-Jung 4.1. Introduction ... 63
4.2. Specific Governability Problems of Nanomedical Products ... 64
4.2.1. Regulatory Gaps Related to Nanomedical Products ... 64
4.2.2. Problems of Effective and Legitimate Nanomedical Product Regulation ... 66
4.3. Responses to Governability Problems: Prudent Hybridisation ... 67
4.3.1. Expected Advantages of Hybrid Regulation ... 67
4.3.2. Prudent Hybridisation ... 68
4.4. Prudent Potential of the EU Advanced Therapies Medicinal Products Regulation ... 71
4.4.1. Brief History and Scope ... 71
4.4.2. Hybrid Regulation? ... 72
4.4.3. Questions 1 to 5 on the Legal Frame ... 73
4.4.4. Question 6 on the Vigilance Regulations ... 74
4.4.5. Questions 7 to 14 on the Quality of Regulation ... 75
4.5. Conclusions: Lessons to be Learned ... 77
References ... 78
5. Access to New Technology. In Defense of the Liberal Regime of Innovation ... 85
Wolfgang van den Daele 5.1. Introduction: Resistance against New Technology then and now... 85
5.2. The Liberal Regime of Innovation ... 86
5.3. A License to Expose the Society to ‘Creative Destruction’? ... 87
5.4. Precaution without a Principle ... 88
5.5. Protecting the Status Quo: Social Sustainability as a Criterion ... 91
5.6. The Return of the ‘sacred’ in the Regulation of Technology: Human Nature as a Holy Order... 93
5.7. Political and Moral Controls of Innovation: Towards a Proper Balance ... 95
5.7.1. Precaution with a Principle: Rules of Law for Administrative Agencies ... 95
5.7.2. Non-discrimination and Risk Comparison in Precautionary Law-making ... 96
5.7.3. New Technology and the Common Good ... 98
5.7.4. Proceed with Caution, but Proceed! ... 99
5.7.6. Dealing with Human Nature: Moral Rigor vs. Trust in
People ... 100
5.8. The Private Right to Innovate in a Reflexive Society ... 102
References ... 103
6. Patenting Nanotechnology: Are We on the Right Track?... 107
Maurice Schellekens 6.1. Introduction ... 107
6.2. The Uneasy Relationship between Nanotechnology and the Patent System ... 109
6.2.1. The Scientific Nature of Nanotechnology ... 109
6.2.2. The Interdisciplinary and Cross-industry Character of Nanotechnology ... 112
6.3. Addressing the Friction between Nanotechnology and the Patent System ... 113
6.3.1. Person Skilled in the Art ... 113
6.3.2. Novelty and Inventive Step ... 116
6.3.3. Industrial Applicability ... 118
6.4. Conclusion ... 120
References ... 121
7. A ‘Scanning Probe Agency’ as an Institution of Permanent Vigilance ... 125
Stefan Gammel, Andreas Lösch and Alfred Nordmann 7.1. Introduction ... 125
7.2. Nanotechnology and Existing Regulation ... 126
7.3. Outline of a Model Institution ... 129
7.3.1. Three Basic Functions ... 129
7.3.2. Two Modes of Working ... 130
7.3.3. The Point of Reference ... 131
7.4. Requirements and Objectives of the Model Institution ... 131
7.5. Type and Designation ... 133
7.6. Functions of the SPA ... 134
7.7. Operational Modes ... 136
7.7.1. Normal Case Mode ... 136
7.7.2. Incident Mode ... 138
7.8. Conclusion ... 140
8. Regulating after Parfit: Welfare, Identity and the UK Embryology
Law ... 147
Colin Gavaghan 8.1. Introduction ... 147
8.2. Background and Context ... 150
8.3. The Nature of the Problem ... 152
8.4. Person-affecting Arguments ... 155
8.5. Non-person Affecting Approaches ... 158
8.6. Conclusion ... 160
References ... 164
9. Law at a Crossroads: Losing the Thread or Regaining Control? The collapse of distance in real-time computing ... 167
Mireille Hildebrandt 9.1. Introduction ... 167
9.2. The Law and the Script: Control at a Distance ... 169
9.3. The Law and the Printing Press ” Authority and Contestation .. 172
9.4. Law in a Smart World ” The Collapse of Distance ... 174
9.4.1. Distantiation? ... 174
9.4.2. The New Brain: Digital Natives, Immigrants; Net Geners and Boomers ... 174
9.4.3. Distantiation and Virtualisation ... 177
9.4.4. Implosion of Distantiation ... 179
9.5. Regaining Control ” Distantiation in the Era of Real Time Profiling ... 180
9.5.1. Real Time Profiling and the Loss of Interpretation ... 180
9.5.2. The Contestation of Real Time Interceptions ... 181
9.5.3. Regaining Control? ... 182
9.6. Concluding Remarks ... 185
References ... 186
10. Pervasive Science. Challenges of Contemporary Technosciences for Governance and Self-Management ... 191
Hub Zwart 10.1. Introduction: Assessing the Present and Exploring the Future: the Basic Assignment of Philosophy... 191
10.2. Pervasive Science ... 192
10.3. Pervasiveness and Self-knowledge ... 194
10.4. Biomimesis as a Key Aspect of Pervasiveness ... 197
10.5. Pervasive Applications: Philosophical Reflections ... 200
11. 'Trust me, I’m a Regulator': the (In)adequacy of EU Legislative
Instruments for Three Nanotechnologies Categories ... 207
Geert van Calster, Diana Bowman and Joel D’Silva 11.1. Introduction ... 207
11.2. Regulatory Challenges Posed by Nanotechnologies: A Case Study Approach ... 210
11.2.1 Cosmetics ... 210
11.2.2 Foods ... 217
11.2.3 Medical Products and Devices ... 220
11.3. Initial Steps to Address the Regulatory Gaps ... 226
11.4. Regulatory Innovation and the Potential Role for the Precautionary Principle ... 228
11.5. Trust Me, I’m a Regulator ... 230
References ... 232
12. How Biological Science Got the Upper Hand in the Debate on Human Animal Hybrid Embryos in the UK ... 239
Nicolle Zeegers 12.1. Introduction ... 239
12.2. Biomedical Technologies and the Communicative Approach to Legislation ... 240
12.3. Power in Communicative Regulation ... 242
12.4. Introduction of the Issue of the Creation of Human Animal Hybrid Embryos ... 244
12.5. The Decision-making Process by the HFE Authority ... 246
12.6. The (Pre)-Legislative Process of Amending the 1990 HFE Act 251 12.7. Conclusion ... 256
References ... 257
13. Regulating Technologies and the Uncertainty Paradox ... 261
Marjolein van Asselt, Ellen Vos and Tessa Fox 13.1. Introduction ... 261
13.2. Uncertain Risks ... 264
13.3. Regulation in Situations of Uncertain Risks ... 267
13.3.1. The Pfizer Case: EU Regulation of Feed Additives and the Use of Antibiotics in Feedstuffs ... 267
13.3.2. EU Regulation of GMOs ... 270
13.3.3. Pfizer and GMO Cases Compared: Risk Aversion versus Risk Intolerance ... 274
References ... 281
14. De Facto Governance of Nanotechnologies ... 287
Arie Rip 14.1. Introduction ... 287
14.2. The Notion of De Facto Governance ... 289
14.3. De Facto Risk Governance in the Domain of Nanotechnology ... 292
14.4. Discourse and Practice of Responsible Development of Nanotechnology ... 297
14.5. An Overarching Pattern? ... 301
14.6. In Conclusion ... 304
References ... 307
15. Ten dimensions of technology regulation. Finding your bearings in the research space of an emerging discipline ... 311
Bert-Jaap Koops 15.1. Introduction ... 311 15.2. Ten Dimensions ... 314 15.2.1. Technology Type ... 314 15.2.2. Innovation ... 315 15.2.3. Place ... 316 15.2.4. Time ... 317 15.2.5. Regulation Type... 318 15.2.6. Normative Outlook ... 319 15.2.7. Knowledge ... 320 15.2.8. Discipline... 320 15.2.9. Problem ... 321 15.2.10. Frame ... 322
15.3. Finding Your Bearings in Research Space ... 322
15.4. To Boldly Go ... 325
References ... 326
Abbreviations
ALARP As Low As Reasonably Practicable AmI Ambient Intelligence
AmLaw Ambient Law
ART Assisted Reproductive Technologies ASTM American Society for Testing and Materials ATMP Advanced Therapy Medicinal Products
BMBF Federal Ministry of Education and Research [Germany] CNT carbon nanotubes
EC European Community / European Council EFSA European Food Safety Authority
EGE European Group on Ethics in Science and New Technologies
ELSA ethical, legal and social aspects EP European Parliament
EU European Union
EMeA European Medicines Agency EPO European Patent Office ESN external social networking GM genetically modified
GMO genetically modified organism
HFEA Human Fertilisation and Embryology Authority [UK] ICH International Conference on Harmonisation of Technical
Requirements for Registration of Pharmaceuticals for Human Use
ICT information and communication technologies ISO International Organisation for Standardization IVF In-vitro fertilisation
MEP Member of the European Parliament NGO Non-Governmental Organisation N&N nanoscience and nanotechnology
OECD Organisation on Economic Co-operation and Development
OJ Official Journal [EU] OSN online social networking
PGD Preimplantation Genetic Diagnosis
REACH Registration, Evaluation, Authorisation and Restriction of Chemical substances
RFID Radio Frequency IDentification
SCAN Scientific Committee on Animal Nutrition SI Statutory Instrument [UK]
SPA Scanning Probe Agency TR Technology Regulation
1
Chapter 1. Introduction. A Dimensions Approach to Technology Regulation
Morag Goodwin
Tilburg Institute for Law, Technology, and Society (TILT), Tilburg University
1.1. Introduction: a Multi-dimensional Analysis
From the first use of a stone as a hammer, the invention of the wheel or prehistoric man’s ability to create fire, the extraordinary creative abilities of our species have needed controlling. The use of tools to adapt and control our environment, while the driving force of our species, has always presented human beings with the possibility of harming others, ourselves, or the broader environment within which we live. This endless creative and simultaneously destructive curiosity has always needed some form of regulatory control for the safety and benefit of both individuals and the community as a whole. These regulatory efforts seek a balance between encouraging and curbing innovation, requiring periodic adjustment where that balance tips too far in the permissive direction and we are reminded again of the destructive potential of our inquisitiveness; or it tips too far the other way and our playful inventiveness is stifled by unchecked fears. In recent years, the pace and range of technological change ” a still accelerating phenomenon ” has constituted a sort of ‘permanent revolution’, in which constant technological innovation and conversion across a wide spectrum of technologies has left our ability to adapt always one step behind. Regulatory efforts thus face an enormous challenge in keeping pace with technological developments and in finding an optimal balance between protection and creativity. While the complexities of technology regulation have long been noted, research into in this area as a comprehensive field of study worthy of its own disciplinary tent is a recent phenomenon; and one we are, un-surprisingly, still struggling to fully grasp.
CHAPTER 1
2
regulation; it also conveyed the aim of the conference to address the challenges of technology regulation from alternative perspectives that have the potential to alter the angle from which we look at the field. These perspectives include those affected by disciplinary boundaries, by national background and by research interests.
The conference was organised as an open tent, in which speakers were invited from a range of different traditional disciplines ” including law, ethics, politics, sociology, philosophy, management, policy and communication studies, science and technology studies, science proper and information security ” from a range of national backgrounds, and without a preconceived overarching theme. What we most wanted to know from the conference participants was what their current preoccupations were; what aspect of the regulation of technology they found particularly challenging at present; and how they saw those challenges evolving as the technological revolution continues apace. The papers included here were selected to represent the range and depth of the themes that emerged from our discussions over those two days in Tilburg.
Through the open nature of the call for papers and the enthusiastic yet broad response, a number of themes emerged. These included a concern about the impact on individual fundamental values; a continuing preoccupation with risk and uncertainty, implicitly or explicitly paying heed to the well-known Collingridge dilemma (Collingridge, 1980);1 a focus on the particular regulatory challenges in the area of nanotechnology and varying calls for sophisticated regulatory responses; reflection upon the broader societal implications of new technologies; as well as the way in which the technology-human dynamic may be changing human nature, making it necessary to re-consider what it is to be human.
The nature and form of this wide range of overlapping concerns, reflections and analyses have led us to visualise the contours of this emerging research discipline through the imagery of multi-dimensional space. In the final paper of this volume, Bert-Jaap Koops uses a metaphor borrowed from theoretical physics to present the discipline of technology regulation as a space that contains ten dimensions (Chapter 15, in this volume). In doing so, he attempts to provide a field manual to help us find our bearings in a multi-dimensional field and to guide us through it. Even where a ten-dimensional space is beyond
CHAPTER 1
3
many of our imaginations, the analytical device of dimensions allows us, as Koops highlights, to gather our bearings and see more clearly where we are and what influences the determination of our positioning relative to others. Moreover, once we know better where we stand, we can move forward with more confidence into areas of the field that are yet to be explored, mapped and studied, and thus enhance our understanding of technology regulation qua discipline. As Brownsword and Somsen have recently noted, our regulatory intelligence remains primitive (Brownsword and Somsen, 2009:32); the final contribution of this conference, then, is to try and find new ways of understanding the emerging discipline of technology regulation as a way of beginning to think smart about regulatory possibilities.
Thus, in order to explore further this innovative approach to mapping the field of technology regulation research, we have decided to group the selected papers according to certain of the dimensions that Koops identifies. The aim is to better understand where we stand now, in order to be able to determine the research agendas of the future. For this purpose, we have been guided by the papers themselves and have identified the four dimensions of technology type, innovation, time and regulation type as best describing the authors’ preoccupations and the way in which they approach what they identify as the most pressing challenges. Of course, each contribution is firmly grounded in the overall space and therefore touches upon many other dimensions, but the categorisation here serves well to draw out the scope of particular dimensions. While much remains to be worked out in arguing for a multi-dimensional approach, the organisation of this volume represents a first attempt to see how a dimensional tilt may contribute to our understanding of the emerging discipline of technology regulation.
INTRODUCTION
4
feed the approach to nano-based medical applications. The increasing tendency towards technological convergence, rather than rendering the dimension of technology type redundant, in fact invests it with a new importance: to understand the challenges posed by convergence, we need to grasp the relevance of technology type in all its dimensionality.
The second section is devoted to the dimension of innovation. Efforts to define the regulatory challenge posed by technology frequently refer to ‘new technologies’. This is because new technology types, such as nanotechnologies or neuro-technologies, ask interesting new types of questions of regulators; however, ‘old technologies’ can also produce innovations that require regulatory intervention. Moreover, as suggested above, the regulator should not simply seek to respond to innovation but encourage it, or at the very least, seek not to stifle it. The challenge of creating a balance between stimulating creativity and controlling danger is located primarily within this dimension.
The third part of the book considers a perhaps underappreciated dimension of the regulatory landscape: time. Time, in relation to technology regulation, most obviously refers to the development cycle of technology: from fundamental science to applied science, and from product development to marketing and widespread usage. Different stages of the development cycle pose different regulatory challenges. Yet the time/knowledge disconnect coined by the Collingridge dilemma entails that technology and efforts to regulate it move at different speeds through time ” a problem that is becoming more acute as innovation cycles get shorter. Technological innovation is thus accelerating away from the regulator. However, beyond consideration of the speed of innovation, focusing on the multiple elements of time can help us to grasp better the nub of a regulatory problem or see more clearly the ethical issues at stake, such as with the future non-existence ” the missed life ” of an embryo not selected for implantation.
CHAPTER 1
5
characterisation of techno-regulation as ‘code as law’ (e.g. Dommering and Asscher, 2006; Brownsword and Yeung, 2008; for the original, Lessig, 1999a: chapter 7; Lessig, 1999b) ” however, the interaction between the multiple aspects of this dimension have not yet begun to be mapped adequately. Moreover, the further refinement of aspects of regulation type, as discussed by Koops, adds yet another layer to the complexity of our regulatory toolkit.
1.2. The Dimension of Technology Type
The three papers in the first section of this collection have been placed here in order to show different technology types and the way in which the properties of the technology itself determine the regulatory response. In the case of the first paper ” a joint contribution by Esther Versluis, Tessa Fox, Anique Hommels and Marjolein van Asselt ” the authors consider in detail the EU Seveso regime for the regulation of chemical hazard. The authors detail the particular hazards represented by industrial chemical production, both in terms of scale ” the potential, when something goes wrong, for large-scale catastrophes such as Bhopal ” and in relation to the nature of a chemical plant, in which the processing, storing and transport of dangerous substances works to inhibit risk, thus allowing risks to multiply. What these authors suggest is that the regulators in the form of the Seveso regime have misunderstood the nature of the risk at issue because they have misunderstood the nature of the technology. Whilst Seveso regulates on the basis of simple or singular risks (‘the positivistic risk paradigm’), the chemical industry involves an interplay of multiple risk factors that are heterogeneous in relation to both place and time. This creates a complexity that the current regulatory response is not well equipped to deal with.
INTRODUCTION
6
given the nature of the technology and thus the potential for accidents to cause great harm in a single moment.
In the following paper, Jen Hendry and Kay Goodall consider the regulatory implications of a different technology, that of ICT. By focusing on the global phenomenon of Facebook, they consider the regulatory difficulties of privacy protection in an online world. The unique aspect of ICT ” the fact that its most interesting properties reside not in the material aspect of the technology, as with chemicals, but in the virtual world they give rise to ” has created unique regulatory problems. Regulators are struggling to adapt their tools to the altered reality of time, place and society that marks the virtual world. Hendry and Goodall approach this issue from the relationship perspective not of peer-to-peer interaction but that of provider-to-user. They take as a case study the Terms of Use that every user must sign up to in order to create a profile on Facebook and, in doing so, bring to light the staggering license that this grants to the provider. While content ownership remains with the user, the sweeping scope of the license grants Facebook the right to do whatever they like with the content for any reason they chose, even where it belongs to a former user, i.e. someone that has since closed their profile.
In taking this focus, they draw out the acute problems of regulating for privacy by comparing the attempt by some users to achieve privacy-related goals under different legal systems. In particular, the Terms of Use raise interesting questions of jurisdiction, asserting as they do that, without regard to existing principles on the conflict of laws, the law of the State of California applies to any and all disputes arising from using the Facebook site. Hendry and Goodall examine whether such a clause and that granting the provider the licence to use all user content would stand up in court on either side of the Atlantic, particularly given such questions as the age at which one can consent to a binding agreement. What their analysis reveals is the difficulty of seeking to protect privacy with the traditional tools of intellectual property and contract. The particular problems that arise relate to the discrepancy between concepts such as ‘author’ or ‘owner’ within different jurisdictions; and the difficulty in selecting a legal location in which disgruntled cyberspace users can challenge the overreaching licence conditions of the networking site or appeal to privacy protection rights. As with the previous paper, Hendry and Goodall suggest that the regulatory response has not yet grasped the challenges the particular technology poses, in this case, to fundamental rights.
CHAPTER 1
7
uncertainty of the risks posed by nano-medical products, she reasons her way to regulatory suggestions by means of comparing technologies and their regulatory frameworks. Her starting point is the EU’s medical product regulatory regime and in particular the Advanced Therapy Medicinal Products Regulation. By comparing the governability problems of nano-medical products with the regime already in place, Dorbeck-Jung is able to highlight the regulatory gaps in relation to the particular problems of nano-based medicine; and, moreover, to suggest, where sufficient comparability exists, how the successes of regulation in relation to one technology can be applied to another. Moreover, the results of her analysis clearly influence her approach to bridging the uncertainty gap. Dorbeck-Jung argues that balancing the level of uncertainty inherent to nanotechnology with the extraordinarily high anticipation of the gains from nanotech in the medical field is only achievable by a form of hybrid regulation.
At the European level, hybrid regulation of nanotechnology takes the form of co-regulation and soft-law methods. These formats are understood as allowing for greater reflexivity and experimentation and thus as being better able to cope with uncertainty. Dorbeck-Jung goes on to lay out a number of questions that hybrid regulation must answer in order to be effective and legitimate, rotating around a number of core principles that she identifies as central to European regulatory ambitions, namely openness, accountability, proportionality, subsidiarity, coherence and vigilance (which could perhaps otherwise be expressed as precaution). The paper concludes, however, on an important cautionary note amongst the various calls for softer, de facto
or hybrid forms of regulation; her case study of EU regulation of nano-medical products strongly suggests the need for centralising product knowledge, particularly in situations of de-centralised or hybrid regulation. This warning about the weaknesses of regulatory de-centralisation is picked up by subsequent papers, notably that by Gammel, Lösch and Nordmann in the following section.
INTRODUCTION
8
technology type need not automatically lead to the uniqueness of different types of technology as the starting point of analysis.
1.3. The Dimension of Innovation
The second section of the proceedings opens with an impassioned plea from Wolfgang van den Daele to defend the liberal regime of innovation from populist fears of technological development. In the paper, Van den Daele charts the history of resistance to technological developments whilst making a strong case that it is a central tenet of liberalism that the benefits of technological progress outweigh the creative destruction that is inevitably part of conceptions of progress. In supporting liberalism’s belief in the benefits of science, Van den Daele takes on the challenge that the precautionary principle poses to this regime. The precautionary principle, he argues, is too often used to plug the regulatory gap created by a lack of full scientific certainty. For Van den Daele, such a radical interpretation of the precautionary principle shifts risk assessment away from experts and bases itself instead upon popular perceptions of risk, leaving little room, he argues, for the right to innovate or for the freedom to act as long as it does not harm others.
Van den Daele’s response to those who would wield human dignity as a trump card to prevent developments offending their moral sentiments, and to those who wish to see innovation made subject to public planning rather than left to market forces, is a plea for precaution with principle and for the recognition of the right of private innovation within a reflexive society. This proposal links back into an issue that was raised in the context of the conference, but that for reasons of space could not be developed in the present volume, concerning the responsibility of innovators and scientists for their own innovations.2 While Van den Daele notes that the quest for responsible innovation should not leave the task of regulation entirely to innovators, he advocates an approach that has much in common with the de facto hybrid regulatory regimes that have sprung up in relation to nanotechnology, in which public and private actors co-operate to avoid the worst of the destructive forces of technological creativity. He ends, however, with a useful corrective to the fears and the related
CHAPTER 1
9
demands for further regulation that dominates our present relationship with technology: whether a new technology is worthwhile pursuing is ultimately a question of private freedom.
The second paper in this section approaches the dimension of innovation from a very practical perspective. The contribution by Maurice Schellekens focuses on issues relating to the patenting of nanotechnology and in particular the need to find an intellectual property regime capable of balancing the uncertainty of risk with innovation in this field. In place of Van den Daele’s defence of a right to private innovation, Schellekens warns of the danger of over-patenting in the area of nano developments, reflecting upon the harms that would result and suggesting moderate reforms to the patent system in order to prevent this from occurring. Similar to Dorbeck-Jung’s reasoning, Schellekens bases his expectation that nanotechnology is likely to suffer from over-patenting on the parallels between this new technology and the somewhat older field of bio-technology. The evidence that patents are being granted on simple nano-particles, the building blocks of the technology, is similar to the conferring of patents upon individual genetic markers, according to Schellekens. Also similar is the way in which nanotechnology cuts across industries, with applications in a wide range of applied fields. As Schellekens nicely highlights, the cross-cutting nature of the technology is likely to entail not only that patent applicants are unaware of related developments in different fields, but also that patent assessors may be equally ignorant when deciding whether to grant or deny a patent, the likely result being patents that significantly overlap. Over-patenting in such a new and uncertain technology will stifle innovation as innovators struggle to pay the many multiple licenses that will be necessary to conduct basic research in the field. Given the anticipated benefits of nano-applications, avoiding such a scenario should be a top priority for regulators at every level. The call here, then, is for a balance between patents as a spur for innovation and the stifling of basic innovation from which we all stand to benefit, between private gains and broader public interest.
INTRODUCTION
10
regulatory possibility at present. As subsequent contributions detail (notably that by Van Calster, Bowman and D’Silva, and by Rip), the authors note that the classical regulatory model has been replaced by soft regulatory tools, such as observation, voluntary codes of conduct and stakeholder discussions. However, in contrast to the optimism of later contributions and of Dorbeck-Jung, the authors of this German-based study found that these softer forms of regulation not only could not meet the challenges of regulation alone ” a point unlikely to be disputed by Van Calster, Bowman and D’Silva or Rip and one also highlighted by Dorbeck-Jung ” but that, further, they do not satisfy the basic needs of regulation, which they understand to be public oversight, transparency and legal certainty.
In place of a drift towards softer regulatory forms in the absence of being able to opt for a more classical state-based model, Gammel, Lösch and Nordmann call for a reflexive adjudication procedure. This procedure is to take the form of a collective learning process, whereby existing regulation is subject to open and public critical scrutiny overseen by a special institution, which they provisionally term the ‘scanning probe agency’. This agency is conceived as a ‘learning community’, under the auspices of which experts from all relevant sections of society would be drawn and brought together. Their role would be to investigate and judge nanotech products and processes of development and to present those findings to the public. The remainder of the paper is dedicated to outlining the tasks and working methods of such an agency.
occupa-CHAPTER 1
11
tional health standpoint. A dedicated agency, so the authors argue, would be able to develop the necessary integrated and multi-dimen-sional approach to understanding the risks of innovation in this field. 1.4. The Dimension of Time
The third section focusing on the relevance of the time dimension to the relationship between technology and regulation opens with a contribution by Colin Gavaghan. In his paper, Gavaghan examines the 2008 UK Human Fertilisation and Embryology Act updating existing embryo regulation against a concern for the obligations we owe to future individuals and the issue of identity that such questions raise. While the legislation has produced a predictably polarised response thus far, Gavaghan ignores this debate to focus on the ethical concerns surrounding Section 14(4)(9) of the 2008 Act that requires that embryos known to have a significant risk of ‘serious physical or mental disability’ must not be selected over those that do not. While ostensibly straightforward ” what kind of parent would opt for a child with serious disabilities over one without when faced with the choice? ” Gavaghan shows that the intuitive response misses the ethical question that lies at the heart of this clause, that of the best interests of the not-selected embryo, the child not created. As he succinctly highlights, the issue is not one of improving the life of a child by opting not to allow it to be born with serious physical or mental disabilities but of replacing one potential child ” a disabled child ” with another potential, healthier child. Gavaghan’s paper examines philosophical efforts to justify this choice and shows up the ethical inconsistencies attached to the common focus on the obligations of parents towards a ‘generic child’. Instead, as Gavaghan notes, each embryo is a potenital individual and the parents’ decision of non-selection does not offer it a better, healthier future but destruction. In this way, by focusing on the issue of time and our (mis)understanding of future (non-)existence, Gavaghan brings the ethical dilemma of the Act’s disability clause sharply into focus and ultimately calls its legitimacy into question.
INTRODUCTION
12
instrumental to the development of rule by law and the rule of law. Central to Hildebrandt’s thesis is the work of Paul Ricoeur and his claim that distantiation is inherent to script, that is that the written word allows for distance in both time and space due to the need for interpretation. As the law is written down and disseminated, the author loses control over its meaning, opening up that meaning for discussion and alternative interpretations. The possibility of multiple interpre-tations builds hesitation and space for reflection into our appreciation and application of the law.
Hildebrandt’s view of law as being at a crossroads ” her concerns for the future ” relate to the increasing use of Web 2.0 technologies by ‘digital natives’, i.e. those who were born into a world of parallel processing that characterises the era of real-time computing. She fears that the speed and instant nature of communication are replacing delay, which, while positive in a number of aspects, simultaneously removes the opportunity for the critical assessment of information. As such, the younger generation are not developing the skills associated with linear sequencing and Hildebrandt sees in this development a dramatic challenge to law as we know it. In particular, the removal of distance and the necessary interpretation that accompanies it poses a very real threat to values fundamental to the ‘old’ or ‘current’ legal order, such as the protection of privacy, the principle of non-discrimination and the architecture of due process.
CHAPTER 1
13
consequence, Zwart suggests, of biomesis, in which man-made systems are inserted into ‘natural’ systems in such a way that the artificial elements become embedded, challenging our conceptions of artificial and natural. This Faustus-like ambition to transcend nature, to improve upon it, will take us down a path in which we begin to blur the divide between nature and man-made and in which the distinction between Self and technology itself becomes more than a little hazy. In this paper, Zwart presents concrete examples of fields and applications where this is likely to occur in the future and notes both the opportunities and dangers that these avenues present society and our attempt to make sense of the technological revolution. The tone of this paper suggests that there is no going backwards, no stopping the role of science in our lives and its potential to alter human nature itself. 1.5. The Dimension of Regulation Type
The four papers in the final section of the volume have been placed together to be viewed through the prism of regulation type because they represent different regulatory perspectives, providing alternate insights into the regulatory tools available, as well as the pros and cons of different approaches to technology problems.
The first paper is a joint effort by Maastricht-based colleagues, Marjolein van Asselt, Ellen Vos and Tessa Fox. It focuses on the process of regulation and the crucial importance of the influence and attitude of individual actors in determining a regulatory outcome, in particular the role of the European Court of Justice in interpreting and applying policy-type recommendations. This paper builds upon the earlier work of Van Asselt and Vos on what they have termed the ‘uncertainty paradox’ in the field of risk. They see this paradox of uncertainty ” a term for situations in which uncertainty is acknow-ledged but where science is presented, by policy-makers and judicial actors, as capable of providing certainty ” as particularly pertinent to new technologies.
INTRODUCTION
14
What the authors discover by contrasting these cases is that what is often considered to be risk aversion is, as in the case of GMOs, risk intolerance, i.e. an unwillingness to consider the possibility of uncertainty. Moreover, what their analysis shows is that an attitude of uncertainty intolerance held by a critical actor in the process of either risk assessment or risk management is what made uncertainty intolerance the dominant feature of regulatory efforts in all four cases. This study suggests that the psychological attitudes of the main actors towards uncertainty play a critical role in the design and implementation of technological regulation, and, as such, is an area in which much more research is needed.
The second paper reviews the suitability of EU legislative instruments in the area of nanotechnologies. The collaborative paper, based upon the first year of a four-year research project at the Catholic University of Leuven, examines the regulatory framework at the European level across three separate industries ” cosmetics, medical applications and food contact materials ” to determine the extent to which regulation is problematic and the ways in which the regulatory regimes in each case determine, assess and manage risk. In doing so, Geert van Calster, Diana Bowman and Joel D’Silva pick up on some of the concerns raised by Van Asselt, Vos and Fox, namely that the very novelty of this new technology and the considerable array of potential benefits it offers has persuaded governments and industry to allow experimentation across the manufacturing chain despite the new and unpredictable risks attached to such a new and untested technology.
CHAPTER 1
15
promising. The next phase of the Leuven project is to move beyond an academic consideration of the regulatory frameworks in the three areas of cosmetics, medical products and food contact materials and to identify benchmark criteria for regulatory nanotechnology ” a step further towards reducing our uncertainty about nanotechnology regulation and one that is therefore eagerly awaited.
A focus on the utility of self-regulation in areas of high uncertainty is picked up by the following paper in this section. The contribution by Arie Rip, a veteran of the field of Science and Technology Studies, highlights the extent to which, in the absence of legislators taking the lead, de facto governance mechanisms are in fact springing up in the area of nanotechnology at all levels of governance. Given the structural uncertainties highlighted by others, what surprises Rip is the extent to which nanotechnology is already being governed, albeit without any particular actor with a clear agenda being responsible for such activities. Rip’s contribution is thus concerned with charting the extent and nature of, and the interactions between, the various actions that are currently functioning as de facto governance arrangements in the field of nanotechnology. Here he provides necessary clarity to the emerging pathways of interactions of and between soft regulatory forms, and charts, by means of a helpful diagram, the development of these forms over time.
INTRODUCTION
16
the sphere of nanotechnology might contain the beginnings of an intelligent response to the Collingridge dilemma.
The final paper in this section moves away from consideration of self-regulation and presents instead an institutional angle. The contribution by Nicolle Zeegers presents a considered attempt to incorporate apparently incompatible moral and religious values within regulation ” in this case, legislation at the national level. As with Gavaghan’s contribution, Zeegers focuses on the 2008 UK Human Fertilisation and Embryology Act; however, Zeegers examines the processes and discussions leading up to the Act, and in particular to the debate within Parliament as to whether to prohibit human-animal hybrid embryos, and within the Human Fertilisation and Embryology Agency concerning the definition of an embryo, and whether human-animal hybrids fell within it.
Zeegers explicitly takes the communicative approach as her normative starting point, noting at the outset that according communicative processes a central position in the formulation of legislation in such a sensitive area would allow for a plurality of perspectives and understandings of what constitutes an embryo to be incorporated. However, by analysing the UK debate, she concludes that the integration of alternative perspectives concerning the use of hybrid embryos in research was prevented by various power relations in the communicative process, with concerns about human-animal hybrids simply being put aside at the crucial moment of decision. What this paper ultimately does is highlight the unwillingness of the dominant (scientific) perspective to take minority views in this debate seriously and raises again the issue of the legitimacy of regulation and of those values that we label as ‘fundamental’ in the face of the plurality of ethical standpoints in contemporary societies.
CHAPTER 1
17
findings to the fore, enriching our understanding of individual contributions within the broader field. Without ruling out the possibility of a better mapping device being possible tomorrow, for now we hope to show you what a map of the discipline based upon dimensionality might look like and why it may be a useful way in which to unravel and plot the present and the future of technology regulation.
References
Berman, H. (1983), Law and Revolution: the formation of the western legal tradition (Cambridge, M.A.: Harvard University Press).
Brownsword, R. and Somsen, H. (2009), ‘Law, Innovation and Technology: Before We Fast Forward ” A Forum for Debate’, 1(1)
Law, Innovation and Technology, 1-73.
Brownsword, R. and Yeung, K. (eds.) (2008), Regulating Technologies
(Oxford: Hart Publishing).
Collingridge, D. (1980), The Social Control of Technology (London: Frances Pinter).
Dommering, E. and Asscher, L. (eds.) (2006), Coding Regulation. Essays on the Normative Role of Information Technology (The Hague: T.M.C. Asser Press).
Lessig, L. (1999a), Code and Other Laws of Cyberspace (New York: Basic Books).
Lessig, L. (1999b), ‘The Law of the Horse: What Cyberlaw Might Teach’, 113 Harvard Law Review, 501-546.
Mandel, Gregory (2009), ‘Regulating Emerging Technologies’, 1(1)
20
21
Chapter 2. Calculable Risks? An Analysis of the European Seveso Regime
Esther Versluis, Marjolein van Asselt, Tessa Fox and Anique Hommels Maastricht University, Faculty of Arts and Social Sciences
Abstract
The chemical sector is confronted with risks pertaining to accidents involving dangerous substances. At the European level, a set of regulations ” the Seveso regime ” aims at controlling such risks. This paper explores how this regime is put into practice, by analyzing the local practices of enforcement by Dutch inspectors and compliance by Dutch chemical companies. These empirical insights demonstrate that the classical ‘positivistic risk paradigm’ ” which presents risks as calculable, controllable and reducible ” seems to dominate in the Seveso regime. The analysis in this paper shows that this can lead to ‘uncertainty blindness’ ” a regulatory regime where only yesterday’s accidents are managed and salient future risks are potentially overlooked. We suggest that both regulators and regulatees should start accepting the possibility of uncertain risks, which implies a cultural change in the current regulatory regime.
2.1. Introduction
Regulation plays an important role in controlling risks. This also holds for risks associated with the chemical industry. Industrial risks are complex because within a chemical plant, the processing, storing and transport of dangerous substances involve risks, hereby creating accumulation and interplay of many risk factors. The chemical sector is confronted with the possibility of accidents involving dangerous substances. Past accidents in the chemical industry in, amongst others, Bhopal, Mexico City and Seveso have led to attempts in the European Union (EU) to control such major-accident hazards. A set of two EU directives and three amendments ” together defined as the ‘Seveso regime’ ” aims to regulate the chemical industry in order to prevent accidents. In the history of the Seveso regime, it is clear that each time a major accident happened, the rules were redefined and sharpened. It is thus assumed by the regulators that tight(er) regulation is the best way to regulate risks in the chemical industry.
CALCULABLE RISKS?
22
two groups of actors involved in this regulatory regime ” the regulators and the regulatees ” perceive the risks at stake? Based on in-depth interviews with both parties, we will demonstrate that the classical ‘positivistic risk paradigm’ which presents risks as calculable, controllable and reducible seems to dominate within the Seveso regulatory regime. The Seveso definition of risk suggests a focus on simple, calculable risks at the expense of risks that are uncertain. This impression of ‘uncertainty intolerance’ is supported by the very detailed and complex character of the regulation, which breathes the pretence of full control and absolute safety. Two rounds of semi-structured interviews (n=17)1 with Dutch inspectors and chemical companies provide insight into the way in which risks are perceived and regulated. Our analysis shows that the current Seveso regime ” to put it somewhat provocatively ” ‘lags behind’ new academic insights, and could even be labeled as ‘uncertainty blind’. This paper concludes by stating that we are in need of a more reflexive regime in which regulators stimulate uncertainty tolerance.
2.2. The Seveso Regime
The explosion of a chemical plant in Flixborough (United Kingdom) in 1974 led to 28 fatalities. The next year, a naphtha cracker exploded in Beek (The Netherlands), killing 14 employees. A year later, two accidents occurred in Italy: one in Manfredonia and one in Seveso where a vapour cloud containing lethal dioxins escaped from a chemical plant and resulted in 2,000 people having to be treated for dioxin poisoning. While most member states of the European Union at that time had their national systems to regulate such risks, the quick succession of these accidents on European territory suggested a need for international action due to the ‘dread, novelty nature and uncontrollability of the hazard’ at stake (Arcuri, 2005: 207).
After three years of negotiations, the Seveso Directive was adopted in 1982. Further accidents led to amendments that broadened the scope of this first directive: the 1987 Bhopal (India) accident which caused more than 2,500 deaths led to a first amendment, and the 1988 accident in Basel (Switzerland) which caused major pollution of the Rhine triggered a second amendment. Research into registered accidents led to the ‘recognition that approximately 85% of over 300 accidents reported under Seveso I have shown some deficiencies in the management system’ (Porter and Wettig, 1999: 3). Therefore, a second
CHAPTER 2
23
(replacing) Seveso directive was introduced that changed the scope from identifying a list of named substances and regulating individual technical installations2 to focusing on the management systems of entire establishments.3 The latest amendment to the directive stems from 2003; it was introduced after the accident with the fireworks storage facility in Enschede (The Netherlands) in 2000 and the explosion of a fertilizer plant in Toulouse (France) in 2001.
Figure 2.1: Timeline of accidents and EU regulation
Every change in the Seveso regime is a response to a major accident (see Figure 2.1). Each time, the rules have been redefined and sharpened. It is thus assumed by the regulators that tight(er) regulation is the best way to regulate risks in the chemical industry. It is to be realized, however, that each accident revealed a new risk, i.e. a possible hazard neither considered nor known from previous experience. As a representative of a chemical company stated: ‘in case something new happens you get new insights’ (# 5). Instead of anticipating a broad range of both known and imaginable new risks, the
CALCULABLE RISKS?
24
regulators mainly seem to manage yesterday’s accidents rather than tomorrow’s risks.
Under this European Seveso regime, chemical companies that house a certain threshold of listed chemical substances (e.g. ammonium nitrate, hydrogen, or chlorine), are considered to be ‘Seveso establishments’. They are required, firstly, to prevent major accidents from happening and, secondly, in case accidents do happen, to control the consequences for man and the environment. Seveso companies have to draw up, among other things, a ‘major accident prevention policy’ via a ‘safety report’ and internal and external ‘emergency plans’. Furthermore, they have to control ‘domino effects’ in areas where Seveso companies are located close together, such as industry parks, and they have to alert all people liable to be affected by a major accident. Member state governments are to ensure that the risks regulated under this regime are taken into account in national land-use planning legislation, and they are to set up ‘competent authorities’ responsible for inspecting the regulatees. All in all, the directive asks for a considerable number of required activities from both regulators and regulatees. In this paper, we will especially analyze the regulation of risks via the safety report requirements (Directive 96/82/EC, article 9).
The Seveso regime makes use of European directives. As a directive is binding in the results to be achieved rather than in the means, all 27 EU member states first have to transpose such a directive into national legislation. Member states, in other words, have the opportunity to make the rather vague European requirements more explicit at the national level. As the Seveso directives are so-called minimum directives,4 they furthermore allow member states to add additional requirements while transposing. How the risks are regulated precisely, in other words, is to a large extent up to the member states. EU directives only come to life in the enforcement and compliance practices at the local level. Nevertheless, the Seveso regime obviously does set the context for the local practices.
2.3. The Seveso Regime and the Positivistic Risk Paradigm
In the Seveso directive, risk is defined as the ‘likelihood of a specific effect occurring within a specified period or in specified circumstances’ (Directive 96/82/EC, article 3). The definition of risk in the Seveso regime resonates with the classic definition of risk as a function of
CHAPTER 2
25
probability (= likelihood) and effect, which was inspired by the work of the economist Knight (1921). He argued that it is possible and necessary to distinguish sharply uncertainty from risk. He views risks as the calculable, hence controllable, islands in the sea of uncertainty (Nowotny et al., 2001). Langlois and Cosgel (1993: 3) argue that ‘Knight's distinction between risk and uncertainty has been taken to differentiate between the measurability/unmeasurability of probability’. In this paradigm ” also referred to as the ‘positivistic risk paradigm’ (Van Asselt, 2000; Krayer von Kraus et al., 2005) ” risk is used to refer to hazards that are known and calculable from previous experience. Risks are thus presented as calculable and controllable.
This dichotomy is still the dominant way of looking at risk. However, an increasing number of authors (see, for example, Vercelli, 1995; Gezondheidsraad 1995, 1996; Nowotny et al., 2001; Van Asselt and Vos, 2006; Renn 2006; Renn and Walker, 2008) argue that uncertainty and risk cannot as easily be distinguished as the positivistic risk paradigm assumes. Some risks are simple, in the sense of certain enough to be calculated as a function of probability and effect. In those cases, due to past experience and the associated availability of statistical data, probability can be estimated and a measure of effect can be derived. Simple risks are calculable and relatively easy to manage. Existing risk assessment tools and risk management approaches suffice.
However, many risks are not that simple. Risk refers to potential events with consequences that one or more actors evaluated as negative. In many cases such events and/or consequences are highly uncertain, because they consider new hazards or involve situations with structural changes compared to the past. In the latter case, the available statistics are of limited value to estimate probability and effect, as the historical data no longer do justice to current and future situations. Furthermore, many risks are complex, which also contributes to uncertainty. They involve a multitude of effects, of which some may extend into the long term, that cannot be easily assessed and compared; nor can measures of effect, if available, easily be added. Risks may also involve complex causalities, non-linear relationships as well as interactions between effects. Uncertainties about the relevant phenomena and the underlying multi-causal relationships may render it difficult, if not impossible, to determine what may happen. Such risks are thus not, or at best only partly, calculable, because the probability of occurrence or the damage cannot be estimated, and even the potential hazard and the relevant causalities may not be established, although there are suspicions of danger.
CALCULABLE RISKS?
26
uncertain risks pertain to uncertain situations, which may result in one or more effects that are valued negatively or considered unacceptable by at least one, but possibly more, societal actors. Renn (2006; see also Renn and Walker, 2008) proposes to further differentiate between uncertain risks, complex risks and ambiguous risks. However, as argued in Van Asselt (2000), complexity goes hand in glove with fundamental uncertainty. Although not all uncertain risks are necessarily complex, all complex risks involve uncertainty. Ambiguous risks refer to situations in which 'value judgements [about risks] (…) differ from one individual to another' (Renn and Walker, 2008: 38-40); this usually occurs in multi-actor settings. However, we would like to argue that this is not an independent category. Both simple risks and uncertain, complex risks may be laden with ambiguity, but ambiguous risks are either simple or uncertain. We would like to argue that the most important distinction is the difference between simple risks on the one hand, and uncertain risks on the other. The further differentiation into complex and ambiguous risks can be integrated in such a scheme (see figure 2.2).
Figure 2.2: Types of risk
In the positivistic Knightian risk paradigm, uncertain, complex and/or ambiguous risks are overlooked. However, the occurrence of new hazards demonstrates the need for expanding the conceptualization of risk. The most important reason to recognize different types of risk is that different types of risk require fundamentally different assessment, management and communication approaches (e.g. Wynne, 2001; Lofstedt, 2005; Van Asselt and Vos, 2006; Renn, 2006; WRR, 2008).
CHAPTER 2
27
However, we would like to argue that major accident risks are often uncertain, complex and ambiguous (see also WRR, 2008). Risks in the chemical industry do not concern singular risks, but involve accumulation and interplay of different but correlated risk factors, as well as multiple, heterogeneous and long-term effects. For example, within a chemical plant, not only the processing installations, but also the storing facilities and transport involve interdependent risks, which need to be addressed both separately and in relation to each other. Furthermore, uncertainties about the underlying processes and the complex multi-causal relationships between causes and effects may render it difficult, if not impossible, to determine what may happen. Many actors with different perspectives (for instance, emphasizing either environmental risks, health risks or economic benefits) have a stake in the regulation of major accident risks in the chemical industry. For example, one of the interviewed companies has a pit below a weighbridge that can accommodate the complete contents of a tank-lorry. Sometimes a bit of rainwater falls into the pit:
‘We have a level meter with a little pump connected to it, that automatically pumps away the water after it has reached a certain level. The environmental inspector argued that the pump should not work automatically because you cannot be certain what kind of liquid is in the pit. But the fire inspector said that it is important that the pump always works automatically as the risk of an overflow of liquids is too high. So these are contradictory advices’ (# 2).
The Seveso directive, furthermore, involves the local, regional, national and European level. Inspection teams illustrated this by saying:
‘You have to deal with politicians at the national and provincial level. However, my colleagues of the other inspection teams and industry also have their bosses at these levels. Thus overall, you have to deal with several political levels which complicates matters, especially when they have different priorities’ (# 3).
CALCULABLE RISKS?
28
Annex II of the Seveso II directive specifies the minimum data to be included in a safety report. Considering the fact that safety reports produced by the chemical industry are on average 400 pages long (Versluis, 2003: 130), this one-page annex is extremely ambiguous in defining the criteria. A safety report should, amongst others, contain a ‘detailed description of the possible major-accident scenarios and their probability or the conditions under which they occur including a summary of the events which may play a role in triggering each of these scenarios’ (Directive 96/82/EC, Annex II). It is not specified, however, what a scenario is, what is meant by a detailed description, or what types of events are referred to. A further problem is more fundamental. In the context of risk management, scenarios are coherent descriptions of alternative hypothetical futures as an effort to capture a wide range of possible future developments and circumstances. It is possible to reason about conditions under which they may occur or the events that may trigger a scenario to unfold; however, establishing a scenario’s probability, i.e. likelihood of occurrence, is difficult if not impossible. Generally speaking, there are two ways to arrive at probability estimates: 1) statistics about previous accidents are used to estimate how often such accidents occur, which is used as a basis to forecast the likelihood of such an accident in the future (this is referred to as the frequentist approach), and 2) probability is interpreted as a subjective degree of belief, which implies that expert judgements are used (this is referred to as the Bayesian approach). Especially when scenarios feature conditions or events not experienced before, the frequentist approach is not applicable; then, the question is how to value experts’ degrees of belief, as research on foresight (see, for example, Van Notten, 2005) has indicated that it is difficult, also for experts, to take unprecedented scenarios seriously. In such cases, probability is solely a reflection of our experience with past accidents, and does not necessarily inform about future risks. The series of accidents that inspired the Seveso regime demonstrate that unprecedented scenarios do happen in practice, notwithstanding low probabilities.
CHAPTER 2
29
2.4. The Dutch Regulatory Practice
So far, we have examined the Seveso regime from a risk perspective through close reading of the directives and amendments. In the second part of this paper, we will investigate the practice of the Seveso regime. Informed by two rounds of interviews, we discuss Dutch regulatory practice.
2.4.1. Conceptualization of Risk
As the Seveso directive is the basis for regulation in member states, it should not come as a surprise that we encountered the positivistic risk definition also in the interviews with both the regulators and the regulatees:
‘There are two sides to risk: chance and effect. In case of large effects within a scenario, you will work on chance calculations’ (# 3; regulator). ‘First they [risks] are quantified and then classified’ (# 4; regulatee).
Recently, a further distinction has been introduced in Dutch regulation and inspection procedures. Risks are classified as controllable risks or as so-called ‘leftover risks’ (In Dutch: rest-risico’s). ‘Leftover risks’ can be calculated, but they are either difficult to control pro-actively or pertain to too expensive, exceptional situations. One of the interviewees described leftover risks as follows:
‘Leftover risks are an articulation of the fact that you know upfront that a certain disaster with a very small calculated chance of occurring cannot be prevented. It is a theoretical model, but there is a risk that you tolerate’ (# 1).
What does this category of leftover risks, and the ways in which it is described, tell about the underlying risk assumptions? Different characteristics are ascribed to leftover risks. They are calculable, but difficult to manage or it is undesirable to manage them, because the costs of managing are either too high or disproportional compared to the (calculated) probability. So these risks are tolerated, in the sense that no risk management measures are taken to prevent or mitigate these leftover risks. With the notion ‘leftover’, it is communicated that these risks are unimportant from a risk-management point of view. The recognition of limits to controllability are also visible in the following interview quote:
‘We can control the normal risks that derive from operating a chemical factory. But we do not control all risks; when a plane crashes, there is nothing I can do’ (# 4; see also 5).
