In the aftermath of the disaster: liability and compensation mechanisms as tools to reduce disaster risks.
In addition to analyzing the theoretical potential of various instruments to contribute to disaster risk reduction, this article provides many examples that show which instruments are used in practice. It also provides a critical analysis of international environmental agreements, arguing that the liability rules used in those agreements show particular features that may reduce their ability to contribute to disaster risk reduction. It therefore argues that there is substantial scope for policy change, more particularly in international environmental agreements. By making a smarter use of liability rules and having risk-dependent contributions to compensation mechanisms, the liability and compensation schemes in international environmental agreements could better contribute to disaster risk reduction than is currently the case.
INTRODUCTION I. STARTING POINTS AND BACKGROUND A. Technological Versus Natural Disasters B. Providing Incentives for Disaster Risk Mitigation C. Risk Differentiation D. Different Instruments for Different Disasters E. International Law Instruments II. LIABILITY RULES AND SOLVENCY GUARANTEES A. Potential, Conditions, and Limits 1. Potential 2. Conditions a. Strict Liability Versus Negligence b. Solvency Guarantees c. No Limits d. Attribution e. Summary 3. Limits of Liability Rules a. Priority of Regulation b. Public Authority Liability? c. Summary B. Liability Rules in International Environmental Agreements 1. Civil Aviation 2. Nuclear Liability 3. Marine Oil Pollution 4. Other Relevant Treaties a. HNS Convention b. Transboundary Civil Liability Protocol c. Space Liability C. Analysis 1. Liability Regimes Compared 2. Critical Analysis 3. Alternatives a. U.S. Price-Anderson Act b. Unlimited Liability c. U.S. Oil Pollution Act III. ADDITIONAL COMPENSATION MECHANISMS A. Potential and Models 1. Potential a. Technological Disasters b. Natural Catastrophes 2. Models a. Extra Layer Compensation b. Nuclear Liability Conventions c. Marine Oil Pollution d. Ad Hoc Charity e. Compensation Funds 3. Potential Disadvantages a. Lacking Incentives for Prevention b. Lacking Incentives to Purchase Insurance c. Positive Effects on Disaster Risk Mitigation Are Doubtful B. Analysis 1. Extra Layer Compensation a. Nuclear i. A Distortive Subsidy ii. Reduced Incentives for Disaster Risk Mitigation iii. Reduced Victim Compensation b. Marine Oil Pollution c. Comparison 2. Alternatives to Public Funding a. Risk-Sharing Agreements: P&I Clubs b. U.S. Price-Anderson Act 3. Government Compensation a. Lacking Incentives for Disaster Prevention b. Underinsurance 4. Alternatives IV. FIRST-PARTY INSURANCE A. Potential, Problems and Solutions 1. Potential 2. Problems a. Lacking Demand b. Lacking Supply 3. Solutions a. Comprehensive Disaster Insurance b. Government as Reinsurer of Last Resort B. Analysis 1. Positive Effect on Disaster Risk Mitigation 2. Limits V. OUTLOOK A. Status Quo: Limits of the Liability and Compensation Schemes in International Environmental Agreements B. Dynamic Evolution 1. Adaptation to Changing Circumstances 2. Interactions in a Multi-level Governance Setting a. International Versus Domestic b. International Versus Regional C. Scope for Policy Change CONCLUDING REMARKS
There is a strong demand for compensation in the aftermath of a disaster. The pressure on politicians to provide some form of compensation is especially strong. International environmental law and domestic law both include a variety of liability and compensation mechanisms for victims.
Liability rules are frequently used in the case of manmade or technological disasters. In those cases, a tortfeasor can usually be identified and, to the extent that tortfeasor is solvent, held liable to compensate the victims. Because liability rules will usually not be available in cases of natural disasters, other compensation mechanisms--such as insurance and compensation funds--are used in those instances. Moreover, many international conventions employ hybrid forms of compensation. (1)
This Article focuses on the extent to which these various liability and compensation schemes can reduce disaster risks. Economic analysis and environmental law conventions show that an ex post compensation mechanism may have positive or negative effects on the ex ante incentives to reduce disaster risks. Depending upon their specific design, liability rules may provide incentives to tortfeasors to invest in risk reduction mechanisms. However, those positive incentive effects are less likely in the case of government financed funds or compensation schemes.
This Article will therefore primarily explore the interrelationship between liability and compensation mechanisms, on the one hand, and their ability to reduce disaster risks on the other. A variety of available liability and compensation mechanisms will be reviewed--specifically liability rules, ex post government compensation, compensation funds, and insurance. Each of these mechanisms' ability to generate incentives for disaster reduction will be critically discussed. Moreover, examples of how these particular mechanisms are applied in international environmental law or in domestic law will be provided. The Article will argue that many of the international environmental agreements dealing with liability and compensation do not sufficiently contribute to disaster risk reduction. What is worse, some of those international conventions even create substantial perverse effects. As the Article will show, domestic law often provides better incentives for disaster risk reduction and better protection to victims. (2) Hence, it will also be shown that there is considerable scope for improving international conventions by learning from some examples of domestic law and from the economic approach to liability and compensation.
Following this introduction, Part II will sketch the starting points for the analysis. First, the Article will illustrate how liability and compensation mechanisms can reduce disaster risks in theory; next, it will discuss how incentives for risk mitigation can be provided. In addition, Part II will also outline which instruments of international environmental law will be further analyzed. In particular, this section will explain that the most prominent instruments regarding liability and compensation can be found in the nuclear accidents and marine oil pollution international conventions; these will therefore constitute the core of the analysis. Having outlined the starting points in Part II, Part III will focus on liability rules and solvency guarantees. The potential, conditions, and limits of liability rules, as well as their prevalence in international environmental law, will be outlined and followed by a critical analysis. Part IV will focus on additional compensation mechanisms that take place either in a rather informal way (generosity as a form of charity) or through more formal payments, via compensation funds or governments stepping in and providing additional layers of compensation. Different forms of the latter can be found in international environmental agreements. Part V will focus on an instrument which is not explicitly present in international environmental law, but which plays an important role in domestic law by mitigating risks particularly related to natural disasters: First-party insurance for victims. This Part of the Article will explore the extent to which experience with models of first-party insurance compensation can provide interesting lessons for liability and compensation regimes in future instruments of international environmental law. Part VI will compare the different instruments that have been observed and discussed and consider how liability and compensation instruments in international environmental law have been increasingly improved. This Part will also argue that there is significant potential to improve the contents of the international conventions so as to more effectively contribute to disaster risk reduction. Part VII concludes the analysis.
I. STARTING POINTS AND BACKGROUND
A wide variety of legal rules, both in domestic as well as international law, focus on regulating the different phases of a disaster. (3) Legal rules can focus on three different phases of a disaster: Prevention and precaution, relief efforts and recovery efforts. (4) Mileti refers to these three different phases of disaster efforts as "preparedness, response and recovery." This Article focuses exclusively on the efforts made in the aftermath of the disaster and, more specifically, on the instruments that are available to provide compensation to victims. The three phases of disaster efforts mutually influence each other. This is especially the case for compensation mechanisms, which may have a positive or negative influence on the incentives to invest ex ante in preventive efforts. (5) The crucial question in this Article is therefore how instruments applied in the aftermath of the disaster, in particular liability and other compensation mechanisms, can be constructed in such a way as to provide optimal incentives ex ante for disaster mitigation. The question of which instrument may be optimal under which circumstances depends to some extent on the nature of the disaster, but also on the ability of the instrument to expose those who create the risk of disasters (e.g., operators of hazardous facilities) to the damage in case disaster strikes. International environmental law offers relatively few compensation mechanisms. Those that it does offer focus on nuclear accidents and marine pollution. Most instruments in international environmental law, however, focus on technological risks rather than on natural catastrophes. This Article will follow a law and economics approach to determine optimal instruments of compensation in the view of their effects on disaster risk reduction, since law and economics has paid a lot of attention to the question of how different instruments can provide incentives for prevention.
A. Technological Versus Natural Disasters
The question of which instruments can be adequately applied to provide ex post compensation to victims as well as ex ante incentives for disaster mitigation depends to some extent on the cause of the disaster. A distinction is usually made between technological disasters, also referred to as man-made disasters, and natural catastrophes. (6) Examples of technological disasters are oil spills and nuclear accidents, as well as explosions in particular plants or fires in public buildings. By contrast, natural catastrophes include heavy rainfall, flooding, earthquakes, volcano eruptions, and tsunamis. Data show that while the insured losses resulting from man-made disasters seemed to remain constant in the period between 1970-2007, during that same period there was a substantial increase in the insured losses due to natural catastrophes. (7) A third type of catastrophe usually dealt with separately is the catastrophe caused by terrorism. Such catastrophes are treated separately because on the one hand they are obviously man-made, but on the other hand they share a similarity with natural catastrophes: The injurer (the terrorist) typically can either not be found or is insolvent. As a result, liability rules cannot apply.
In some cases it may be difficult to adequately distinguish between manmade disasters and natural catastrophes. For example, heavy rainfall could sometimes lead to flooding because infrastructural works have changed rivers and, as a result, the natural carrying capacity of waters has decreased and governments may have even promoted building in flood prone areas. In those cases, man-made activity can in fact encourage devastating consequences of natural catastrophes. (8) Another example of the hybrid nature of the distinction between technological and natural disasters relates to the fact that, just as private actions could trigger a natural disaster (e.g., drilling activities might trigger a mud flow), natural disasters could in turn trigger a technological disaster like in the case where a tsunami triggers a nuclear disaster, as was the case in Fukushima. (9)
B. Providing Incentives for Disaster Risk Mitigration
One of the basic starting points of the economic approach to law (10) is that people react to financial incentives. (11) For example, when the operator of a hazardous facility is exposed to the financial consequences of the potential disasters he may create, perhaps via liability rules, this will in turn give him ex ante incentives to invest in disaster mitigation. (12) Hence, an important starting point for analysis is that, to the extent possible, those who create particular disaster risks through their activities should be held liable to compensate the damage they cause. The ex post duty to compensate will provide ex ante incentives to invest in disaster mitigation. This general idea may be of utmost importance in the field of technological or so-called man-made disasters, where there is an operator (or, in terms of tort law, a tortfeasor or injurer) who can be held accountable for the consequences of the disaster.
The idea of providing incentives through exposure to financial consequences is not only important as far as potential risk creators are concerned. It may also be important to expose potential victims to the financial consequences of their decision. Exposing individuals to the risk with which they are confronted--of flooding, for example--will make them aware of their vulnerability to a natural hazard and may therefore have a positive impact on their behavior. This may in turn provide incentives for disaster mitigation to potential victims. (13) Further, failing to alert potential victims to the consequences of their choices--such as location in a hazard prone area--may negatively affect where they choose to live and reduce incentives to proactively mitigate disaster risks. Adopting this strategy will strongly support first-party insurance solutions (14) and weigh against charity payments to victims. The latter may lead to an underinvestment in disaster mitigation measures. (15)
C. Risk Differentiation
Exposing both potential risk creators and potential victims to the financial consequences of disasters is important in order to provide correct incentives for disaster mitigation to the stakeholders involved. However, it does not suffice to expose those who can take mitigation efforts; it is equally important that they are exposed to risk to the extent that they actually contributed through their behavior to that risk. This suggests that the exposure to financial consequences should be related to risk contribution--in other words, risk differentiation should be applied. (16) It further implies that risk creators contributing a larger amount of risk should be exposed to a larger extent of financial consequences. The degree to which a victim is exposed to risk should therefore be reflected in the price the victim pays for protection against damage resulting from the risk, e.g., through insurance. Risk differentiation is crucial because it positively affects incentives for prevention, risk reduction, and mitigation of damage.
Moreover, risk differentiation could also be defended on distributional grounds. A solidarity on the basis of which all taxpayers pay for those exposed to risk could result in a situation in which those who faced no risk and those who faced high risk are rewarded equally. The distributional problem can be made clear on the basis of the following example: Suppose that a particular individual purchases a high quality villa at a low cost in a flood prone area next to a river. If the individual could, when the flooding occurs, rely on compensation by government--and, necessarily, general taxpayers--a distributional problem might arise. The individual has, at least from an economic perspective, already been compensated for the loss by purchasing the house at a lower purchase price; he may not have taken out insurance and thus may not have paid insurance premiums or have taken the essential preventive measures, and can subsequently shift the risk to the collectivity. (17)
D. Different Instruments for Different Disasters
This Article will focus on the instruments that could be used in the aftermath of a disaster to compensate victims. In light of the aforementioned principles of risk differentiation, (18) a variety of instruments will be reviewed and their relative strengths and weaknesses will be discussed based on their capacity to provide incentives for disaster risk reduction. First, this Part will consider liability rules that are basically applicable in the case of technological disasters when there is an identifiable tortfeasor (III). Next, additional compensation mechanisms will come into the picture. Those play on liability rules by providing an extra layer of compensation beyond the amounts due by the tortfeasor, but also on natural catastrophes (IV). Finally, first-party insurance--the compensation mechanism appropriate for compensating victims of natural disasters--will be presented (V).
While insurance can play a role as liability insurance which companies may have to take to cover their liability risks for technological disasters (III), potential victims could also insure against natural disasters via first-party insurance (V).
E. International Law Instruments
There is an impressive amount of international environmental agreements, many of which have the goal of reducing disaster risk. In order to limit the analysis, this Article will focus on tools of international law that have explicit provisions with respect to liability and compensation. Hence, the article's focus specifically concerns instruments that are applied in the aftermath of a disaster to compensate victims. (19) As will be made clear below, (20) we distinguish between international regimes that can be categorized into four different groups: International conventions related to civil aviation, nuclear liability, marine oil pollution, and other international conventions that contain provisions on liability and compensation.
It should be recalled, however, that important provisions can also be found in the work of the International Law Commission (ILC) concerning the codification of international law. Interesting documents in that respect are on the one hand the 2001 Articles on State Responsibility for Internationally Wrongful Acts (ARSIWA) (21) and on the other hand the Principles on the Allocation of Loss in the Case of Transboundary Harm Arising out of Hazardous Activities of 1976. The latter points to the necessity of providing adequate and prompt compensation for transboundary damage, primarily by the operator, but otherwise by the source state. The degree to which those principles focus on transboundary damage will not be discussed, although a deeper examination of the relationship between ILC instruments and disaster risk reduction would undoubtedly be an interesting point of further research.
A variety of other legal instruments of international law may include some provisions regarding compensation for transboundary harm or may at least be indirectly relevant for the prevention of disaster with a transboundary character. In this respect, one can for example refer to the United Nations Convention on the Law of the Sea 1982 (UNCLOS), (22) which only imposes a vague obligation upon all states to protect and preserve the marine environment, but contains no specific liability rules or compensation mechanisms. (23) The same is the case for the Energy Charter Treaty, (24) which was established in 1994 to develop international cooperation in the energy sector, including trade, transit, investments, and energy efficiency. Again, Article 19 of the ECT contains the duties of the contracting parties to strive to minimize harmful environmental impacts occurring within the energy sector, but it lacks specific liability or compensation mechanisms. Another example is the Convention on Environmental Impact Assessment in a Transboundary Context, also referred to as the ESPOO Convention. (23) Again, this Convention sets up obligations of contracting parties to assess environmentally adverse impacts of certain hazardous activities and to consult affected states prior to decision-making, (26) but fails to offer specific liability rules or compensation mechanisms. For that reason, those international environmental agreements--and others like them--remain undiscussed here, although they can undoubtedly contribute to an important extent to disaster risk mitigation.
Finally, we could also mention the importance of principles of international environmental law that could either affect states' responsibilities or play an indirect role in domestic litigation concerning transboundary harm. (27) We therefore recognize that those principles may have an impact by shaping the scope of liability rules and consequently affecting the incentives of risk creators. However, they are beyond the scope of this discussion since the importance of environmental principles is only indirectly related to the scope of the operator's liability.
II. LIABILITY RULES AND SOLVENCY GUARANTEES
Liability rules can undoubtedly be considered instruments that can fulfill the goals of exposing operators to the costs related to the harmful consequences of disasters. In that way, the imposition of liability ex post could provide ex ante incentives for disaster mitigation. However, in order for liability rules to fulfill this preventive effect, particular conditions must be met. Moreover, liability rules may also contain limits (A). As previously mentioned, (28) liability rules found in particular international environmental agreements show interesting similarities (B). Nonetheless, the way in which liability rules are shaped in international environmental agreements is often problematic given the starting point of providing efficient incentives for disaster mitigation. An analysis will reveal that, in many cases, rules of domestic law (e.g., in the United States) serve the goal of providing incentives for disaster mitigation better than the liability rules contained in international environmental agreements (C).
A. Potential, Conditions, and Limits
The economic approach to liability rules reveals that they have the major advantage of providing incentives for hazard mitigation. (29) For more traditional tort lawyers, the main goal of tort law would be to compensate accident victims, and not to primarily provide deterrence. (30) By exposing them to the costs of their activities via liability rules, parties will be given appropriate incentives for taking optimal care to prevent accidents. Since it is the level of care that minimizes the costs of prevention and the expected damage costs, taking optimal care would reduce the total social costs of accidents. (31) This basic insight can apply to the damage resulting from disasters as well: The exposure of the risk taker to liability provides incentives for disaster mitigation. (32)
In 1961, Guido Calabresi stressed the deterrent effect of liability rules. (33) He suggested they would force a potential tortfeasor to take efficient care. The effect of a liability rule is that the social costs of the accident are allocated to the source of the risk that created the accident. Put more simply: The costs of the potential tortfeasor's activity will confront the enterprise causing the particular risk, which will, in turn, incentivize prevention. Moreover, the tortfeasor--the one who creates the risk--pays compensation. As a result, a perfect diversification of risk can take place.
Some have been critical of the assumption that tort law could have any deterrent effect and would therefore influence the behavior of individuals. (34) However, most of these doubts only relate to the behavior of individuals, and not of enterprises. Enterprises' decisions regarding how much to invest in disaster risk prevention may be the result of a conscious cost-benefit calculus more so than in. the case of individuals. There is, moreover, increasing empirical evidence of the deterrent effect of liability rules. (35)
Finally, it is important that in the economic approach prevention is considered to be the primary goal of liability rules. Victim compensation as such is not. Rather, the exposure of the risk taker to liability is considered to have the advantage of providing incentives for optimal prevention. Compensation is hence a means rather than a goal.
Although it was just discussed how exposing risk creators to liability rules may provide efficient incentives for disaster mitigation, it was also stressed that this effect will not be reached automatically and unconditionally. A careful design of liability rules is necessary in order to enable them to provide incentives for disaster risk mitigation.
a. Strict Liability Versus Negligence
A basic question, often addressed in the literature, (36) is whether the operator--the tortfeasor in a liability context--should be exposed to a strict liability or a negligence rule. A distinction is usually made between the case where only one party, the injurer, can influence the accident risk--this is referred to as a unilateral accident situation--and the situation where both the injurer and the victim can influence the accident risk; the latter is referred to as a bilateral accident situation. (37) The goal of tort law should be to provide parties that can influence the accident risk with incentives in order to reduce the total sum of accident costs. Those costs consist, on one hand, of the costs of prevention and, on the other hand, of the costs of the accident (the damage). Efficient care levels can be found where the marginal costs of care equal the marginal benefits in reduction of the expected loss. (38)
In a unilateral accident case, where only the injurer can influence the accident risk, both negligence and strict liability will provide incentives to follow an efficient care level. On the condition that the due care required in the legal system is equal to the efficient care, the injurer will take the efficient care level. The simple explanation is that the injurer will be found liable under the negligence rule if he spends less than the due care level required in the legal system. Hence, for the injurer following the due care level is a mechanism to reduce his expected costs. The negligence rule can therefore in principle give a tortfeasor an incentive to spend on care to reach the optimal standard. However, in order for the negligence rule to work optimally, the legal system must define the due care level as the efficient care level.
Additionally, a strict liability rule will lead to optimal incentives for care taking for the polluter, since taking efficient care will minimize the expected accident costs which the potential polluter has to bear under a strict liability system. (39) Therefore, the literature generally accepts that both a negligence rule and a strict liability rule will provide a potential polluter with incentives to take the efficient care level. However, this is only valid in a unilateral accident setting, in an accident wherein only the injurer can influence the accident risk.
In a bilateral accident situation the potential victim also needs to be provided with incentives for mitigation. In that case a contributory negligence defense should be added to the strict liability rule. Under a negligence regime, victims will always have an incentive to take efficient care as well, since they will--in principle--not be compensated by the injurer, who will take efficient care to avoid liability under a negligence rule.
In the joint care or bilateral case, strict liability with contributory negligence and a negligence rule--with or without contributory negligence--will incite parties to adopt efficient levels of care. However, the accident risk cannot be completely minimized by increasing the levels of care. Accident losses also depend on the extent to which parties participate in the activity that might cause the damage (e.g., the miles driven). Therefore, reducing the activity level will also reduce the accident risk. (40) The activity level can be interpreted as any control variable not taken into account in setting the optimal level of care. Under a negligence rule, an injurer has no incentive to adopt an optimal level of activity. This cannot be remedied, because judges cannot easily calculate the optimal activity level into the due care standard. (41) A strict liability rule has the advantage that the injurer will automatically adopt an optimal activity level. This is also a means to minimize his costs. If the victim's activity has no influence on the accident risk, strict liability might have a slight advantage, because it might also lead to an optimal activity level of the injurer. Nevertheless, in a joint care case this advantage is cancelled out by the fact that the victim will not adopt an optimal activity level. This is due to the impossibility of calculating the activity level into the due care standard when considering contributory negligence. (42)
What is the importance of the activity level for the choice between negligence and strict liability? In a unilateral accident model, wherein only the behavior of the injurer influences the accident risk, strict liability seems to be the preferred rule since it is the only liability rule which will lead both to efficient care and to an optimal activity level. In a bilateral case the answer is more balanced.
Since changes in activity level are not calculated into the due care standard, strict liability with a defense of contributory negligence will encourage activity level changes on the part of the injurer. On the other hand, a negligence rule will encourage activity level changes of the victim. Therefore, several authors suggest that in bilateral cases strict liability will be a superior device if giving injurers an incentive to change their activity level is more important than giving victims a similar incentive. (43) This implies that if the injurer's activity is very dangerous and creates a high accident risk, even if optimal care is taken, it will be more desirable to control the injurer's activity than it is to control the victim's.
Even though a clear-cut test is thus difficult to provide, Landes and Posner describe several factors that may lead to a preference for a strict liability rule. These elements are: (1) high expected accident costs; (2) the impossibility that more care by the injurer would reduce the accident risk; (3) the impracticability of constraining the victim's activity in favor of the injurer's; and (4) the desirability of reducing the risk by an activity level change of the injurer. (44)
Since tortfeasors such as petrochemical plant operators are undoubtedly much more influential than victims with respect to technological accident risks that could lead to man-made disasters, there are strong arguments in favor of applying a strict liability rule to such ultra-hazardous activities. The strong appeal of the strict liability rule (of course combined with a comparative negligence defense to account for the victim's influence on the risk as well) (45) is that it advantageously shifts all social costs of the accident to the operator. Thus, the operator will appropriately weigh costs and benefits concerning optimal preventive measures and activity levels. (46) For this reason, it is no surprise that the literature generally advocates strict liability both for environmental harm (47) and for technological disasters. (48)
b. Solvency Guarantees
Unfortunately, strict liability may only be efficient if the insolvency problem can be cured. Insolvency should be seen here as a situation where the amount of the damage is higher than the tortfeasor's wealth. This scenario is very likely in case of catastrophes. The literature has even indicated that, if an insolvency problem persists, strict liability may more easily lead to underdeterrence than negligence. (49) This is because strict liability leads to underdeterrence as soon as the total amount of the damage is lower than the injurer's wealth. Under a negligence regime, the tortfeasor still has an incentive to spend on care as long as the costs of due care required in the legal system are less than his wealth. Spending on due care will in that case still be attractive for the injurer since it implies that he will not be held liable. (50)
When strict liability is introduced, it should thus be accompanied by some guarantee against insolvency. This is, moreover, a more general point made in economic literature: Liability rules only work efficiently, in terms of both deterrence and compensation, if solvency of the tortfeasor can be guaranteed. Otherwise, a so-called "judgment proof' problem will arise whereby liability rules may generally fail to lead to a deterrent effect. (51) There is, therefore, a strong economic argument in favor of imposing a duty to purchase financial coverage (e.g., compulsory liability insurance) for disasters simply because the magnitude of the damage caused by a disaster can greatly outweigh the assets of an individual tortfeasor. (52)
c. No Limits
In sum, although strict liability can, in principle, be efficient for technological disasters, due to a potential insolvency problem it should be accompanied with solvency guarantees since it could otherwise lead to underdeterrence. Precisely for that reason there should likewise be no limit on the liability of the potential tortfeasor. A so called "financial cap," which limits the tortfeasor's liability to a particular amount, will have the same effect as insolvency, i.e., it will lead to underdeterrence and thus negatively affect incentives for prevention.
An obvious disadvantage of a system involving financial caps is the serious impairment of the victim's right to full compensation. However, if the cap is indeed set at a much lower amount than the expected damage, this would not only violate the victim's right to compensation, but would also prevent the full internalization of the externality mentioned above. From an economic point of view, a limitation on compensation therefore poses a serious problem, since the risky activity will not be internalized.
Indeed, if one believes that exposure to liability has a deterrent effect, limiting the amount of compensation owed to victims poses yet another problem. There is a direct linear relationship between the magnitude of the accident risk and the amount spent on care by the potential wrongdoer. Therefore, if the liability is limited to a certain amount, the potential injurer will assess the accident's magnitude in proportion to that limited amount. Hence, he will only spend on care to avoid causing an accident with a magnitude equal to his limited amount of liability, and he will not spend the care necessary to reduce the total accident costs. Obviously, the amount of care spent by the potential injurer will be lower, causing a problem of underdeterrence. The amount of optimal care reflected in the optimal standard--that is to say the care necessary to efficiently reduce the total accident costs--will exceed the amount the potential injurer will spend to avoid an accident equal to the statutory limited amount. (53) Thus, as a result of the cap too little care is taken. (54)
Moreover, another effect of imposing a financial limit on liability (in addition to victim undercompensation and operator underdeterrence) is that it would constitute an indirect subsidization of the industry that enjoys a particular limit on liability. (55)
Another point that remains important in the design of efficient liability rules is that liability should be attributed in such a way that all parties that contributed to the risk should be held liable to the extent that their actions actually affected the accident risk. When several tortfeasors have acted together, a joint and several liability rule may provide incentives to the joint tortfeasors for mutual monitoring. (56) In any case, liability of other actors who contributed equally to the loss should not be excluded by, for example, exclusively channeling the liability to one selected tortfeasor, e.g., the licensee of a particular plant. Doing so would negatively affect the incentives of others who could have equally contributed to the loss. (57)
However, inefficiencies may arise in case of insolvency of one of the actors, since recourse may become impossible. (58) Joint and several liability is debated since an injurer could, in principle, also be held liable for a part of the damage not caused by his activity, thus potentially increasing his liability exposure. (59)
The application of joint and several liability is a general rule in all cases where more than one tortfeasor is involved. Many legal systems adopt a joint and several liability rule where more than one party (e.g., an operator and a subcontractor) has contributed to the accident risk. (60) Joint and several liability has, however, been criticized from an insurance perspective, based on the argument that it increases the necessity to purchase insurance coverage for all parties involved. (61)
To summarize, liability rules can be used to provide incentives for disaster mitigation, but particular conditions must be met:
* A strict liability rather than a negligence rule should apply;
* solvency guarantees should exist;
* no statutory limit (a so-called cap) should be imposed on the operator's liability; and
* liability should be attributed to all actors who can influence the risk of a catastrophe. (62)
Under these conditions liability rules can have the desirable effect of incorporating the social costs of technological risks into the prices of the activity concerned and to provide incentives to operators for disaster risk reduction. The liability rules only provide incentives to take optimal care, (63) which can thus reduce the chilling effects that could follow from what the literature considers "crushing" liability. For example, the latter could occur when operators would also be held liable for harm that was not caused by their own activity. (64) However, the mere fact that the application of a liability rule will lead to increased prices of particular products or services (resulting from higher investments in preventive efforts) is from a social welfare perspective nothing but desirable. Efficient operators--those taking optimal care to reduce disaster risks--will be able to produce at lower prices than inefficient ones, who would not take optimal preventive measures and would thus be confronted with higher liability risks and higher insurance premiums. That is precisely the desirable result of the application of liability rules: Efficient operators will be rewarded with lower liability risks and, therefore, lower prices.
3. Limits of Liability Rules
a. Priority of Regulation
Although a strict liability rule could in theory provide incentives for disaster risk mitigation--provided that solvency guarantees are in place and that liability is effectively attributed to all those who contributed to the risk--in practice there can be many reasons why liability rules may generally not have a deterrent effect. Shavell's well-known work on the choice between liability rules and safety regulation identifies these limits of liability rules. (65) Shavell indicates that: (1) When information on preventive technology would be greater with government than with private parties and (2) when insolvency problems would arise and/or (3) for a number of reasons a liability suit would never be brought, regulation can be a more effective instrument to control externalities--like damage to critical infrastructure--than private law instruments such as liability rules. (66)
Turning to each of those criteria, it seems that these potential weaknesses of liability rules are all relevant to the potential damage resulting from technological disasters. Private parties may in some cases lack adequate information on preventive technology, whereas governments can use economies of scale and thus invest more efficiently in prevention. Moreover, regulation could pass on information regarding optimal preventive technologies to the parties in the market. The insolvency problem will obviously arise in all cases where smaller operators may also cause high damage whose potential magnitude may outweigh their personal assets. In addition, corporations' limited liability may support the externalization of harm onto third parties and society in general. (67) To the extent that solvency guarantees would be unavailable or unable to cure the underdeterrence that follows from insolvency, this may likewise be an argument in favor of ex ante safety regulation.
There can also be a number of reasons why victims never bring tort suits despite in theory meeting all the conditions. First, tortfeasors cannot be identified in some cases; second, there can be a long period between the accident and the damage (referred to as latency); third, problems of proof may arise as well as problems related to uncertainty over causation. Fourth, victims may face large hurdles such as going to court and effectively litigating. Legal aid, contingency fees, or other instruments to lower the barriers to access to justice are often insufficiently developed. (68) Precisely given these hurdles, using liability rules may also have a rather ad hoc character as far as compensating victims of catastrophes. In other words: Some victims may receive generous compensation if they are successful in the "tort law lottery," whereas others may receive no compensation whatsoever. (69) Compensation via liability rules therefore has no structural nature and may also come at odds with the equality principle.
b. Public Authority Liability?
At first blush it may be obvious that an important limit of tort liability in providing incentives for disaster mitigation is that liability rules can only work in the case of technological disasters where a liable tortfeasor can be identified and where the damage can be directly attributed to manmade causes. No liable tortfeasor can be found, for example, when an earthquake, a flooding, or a volcanic eruption occurs. The only possibility to apply tort law to those natural disasters is to argue that public authorities were at fault, e.g., by failing to prevent the disaster or to take adequate measures to mitigate the damage. This raises the question of the adequacy of public authority liability.
It is possible to theoretically imagine situations where public authorities would be at fault in case of a natural disaster. In fact, some scholars have held that there are no natural disasters, but only natural events that turn into disasters as a result of the intervention of men. (70) Indeed, there are many ways, particularly through the design of critical infrastructure, to ex ante reduce the probability of damage or mitigate the seriousness of the consequences. (71) Precautionary measures to reduce the likelihood of disasters can be taken by individuals but, especially where large-scale measures are concerned, by governments as well. Many disasters can be prevented, and a lack of precautionary measures is often the real reason why natural events have catastrophic consequences. (72) A failure to prevent the disaster or to take adequate measures to mitigate the damage can hence in some cases be attributed to a government. It could for example be held that the government failed to give adequate warning, e.g., in case of a flooding, or it could be questioned why governments provide building permits allowing house construction in flood prone areas or on the slopes of active volcanoes.
Critical questions concerning the role of public authorities are often asked after many natural disasters cause substantial damage. For example, in the case of Hurricane Katrina, Shughart showed that no effective precautionary measures had been taken before Katrina was announced because of bureaucratic myopia, bureaucratic inertia, and corruption. (73) As a consequence, the question of governmental responsibility was raised in the wake of Katrina. (74) However, none of those lawsuits were successful. (75)
One reason why lawsuits against public authorities are often not brought in cases of natural disasters is that governments generously intervene with public aid. For example, in the case of Katrina, a report of the U.S. Senate refers to a total amount of USD 88 billion that the U.S. Federal Government has committed as of March 8, 2006 to the response, recovery and rebuilding efforts. (76)
There is also a serious potential for public authority liability since public choice analysis shows that politicians tend to underinvest in precautionary efforts because these do not lead to substantial political gains during the term of office of the particular politician. (77) This not only plays out in the case of Katrina, as mentioned before, but more generally with all necessary precautionary efforts to mitigate damage. For example, scholars have held with respect to hurricanes that "a number of important potential precautionary strategies that are designed to minimize the expected costs or consequences associated with a natural disaster--but not the risk of its occurrence, which we are assuming to be exogenous--have many of the characteristics of public goods and if left purely to private markets are likely to be underdemanded and undersupplied as a result of collective action problems." (78)
Public authority liability could obviously also arise in the transboundary context, where the question of state responsibility could be raised, in case of breach of an international obligation, pursuant to the ARSIWA, or the question of state liability could arise based on the ILC principles on the allocation of loss. However, in practice one rarely sees cases of public authority liability for the simple reason that governments have a strong tendency to provide ex post compensation, in part because this generates high political rewards. The tendency of governments to provide this ex post compensation may to some extent even be triggered by the fear of public authority liability, thus avoiding the imposition of blame on governments.
Although we argued that liability rules--more particularly strict liability with solvency guarantees--can provide incentives for disaster risk mitigation, in practice the impact of liability rules may not be that large. Due to high barriers to entry, the liability regime may turn out to be merely an ad hoc system available to only a small percentage of accident victims. The primary role in preventing disasters will thus most likely be played by safety regulations, often also resulting from international environmental agreements directly aiming at disaster risk mitigation. As a result of many problems, such as causation, latency, proof, access to justice, among others, in practice liability rules often only play a limited or supplementary effect in preventing technological disasters.
Moreover, liability rules can basically only apply in cases of technological disasters where a liable tortfeasor can be identified. In the case of natural disasters, the only actor who could be held liable would be a public authority, but suits against public authorities, either based on domestic law or on rules of state responsibility or liability, are rare. This leads to the conclusion that the scope of liability rules in providing prevention and compensation will remain limited both for technological as well as for natural disasters. As a result, alternative rules will have to be called upon. For prevention purposes these rules may be ex ante government regulation, whereby compensation can be provided through alternative compensation mechanisms (79) or insurance. (80) In both cases it is obviously important to also determine how those mechanisms affect the ex ante incentives for disaster risk mitigation.
B. Liability Rules in International Environmental Agreements (81)
1. Civil Aviation
With the development of civil aviation, people started to realize the potential damage caused by aircrafts to the people and property on board, as well as to third parties on the ground. The Convention for the Unification of Certain Rules relating to International Carriage by Air was concluded in Warsaw in 1929 and applies to damage to international carriage of persons, baggage and cargo. Damage to third parties on the ground has a different characteristic, in that usually an ex ante contract cannot be reached between the potential injurers and victims. A separate liability regime was therefore established in 1952: The Rome Convention. This convention was later revised by the 1978 Protocol to Amend the Convention on Damage Caused by Foreign Aircraft to Third Parties on the Surface (1978 Protocol). Another attempt to further revise the convention resulted in the 2009 Convention on Compensation for Damage Caused by Aircraft to Third Parties. However, this new convention has not come into force yet.
The 1952 Rome Convention establishes strict liability by attributing liability to aircraft operators. (82) The Convention defines the term "operator" as the person who was making use of the aircraft at the time the damage was caused, provided that an individual who cedes the right to use the aircraft but directly or indirectly retains control of its navigation shall be considered the aircraft's operator. (83) With the exception of deliberate acts or omissions, Article 9 of the Rome Convention constitutes the only basis of liability for operators. (84) However, this Convention does not expressly exclude liability of other parties, and it does not prejudice the right of recourse against other parties. (85)
The Rome Convention establishes liability caps according to the weight of the aircraft. (86) In addition, it also establishes a mandatory financial security system. Any Contracting Parties may require the aircraft operator to maintain insurance coverage for damage up to the liability limitation in its territory. (87)
The liability regime for international carriage of persons, baggage, or cargo was established in 1929 in Warsaw, but substantial revisions were made from the 1950s to the 1970s. Under the auspices of the International Civil Aviation Organization, a new convention was concluded to modernize and consolidate the Warsaw Convention and related instruments in Montreal in 1999: The Convention for the Unification of Certain Rules for International Carriage by Air. (88) According to its preamble, the Montreal Convention has two aims: To ensure the protection of an equitable compensation for consumers of international carriage by air and to promote the development of international air transport operation. It applies to international carriage of persons, baggage, or cargo performed by aircraft. To satisfy the "international" requirement, the places of departure and destination must either be in different State Parties, or in the same State Party but with an agreed stopping place in another State. (89)
The Montreal Convention establishes three types of liability: Liability for death and injury of passengers, damage to baggage, and damage to cargo. (90)
If a passenger dies or suffers bodily injury "on board the aircraft or in the course ... of embarking or disembarking," strict liability will apply. (91) In case of destruction, loss, or damage to baggage, the basis of liability depends on the baggage's checking status. Strict liability applies for checked baggage. For unchecked baggage and personal items, the carrier is only held liable if its servants or agents can be found at fault. (92)
The Montreal Convention also applies to delays in the carriage. Liability caps are established respectively for those types of damage. According to Article 25 of the Montreal Convention, a higher liability cap or no liability cap is also allowed if the carriage contract so stipulates. (93) A mandatory insurance framework is established under Article 50 of the Montreal Convention: "State Parties shall require their carriers to maintain adequate insurance covering their liability ..." (94)
2. Nuclear Liability
Two separate international compensation regimes were established in the 1960s, and both were substantially revised after the Chernobyl accident of 1986. (95)
The Convention on Third Party Liability in the Field of Nuclear Energy of July 29, 1960 (Paris Convention) (96) and the Supplementary Convention to the Paris Convention on Third Party Liability in the Field of Nuclear Energy of January 31, 1963 (Brussels Supplementary Convention) (97) were developed under the auspices of the OECD Nuclear Energy Agency (NEA). The second regime was developed under the aegis of the International Atomic Energy Agency (IAEA): The Vienna Convention on Civil Liability for Nuclear Damage of May 21, 1963 (Vienna Convention). (98) These two regimes are usually referred to as the first generation of nuclear liability conventions. (99)
The 1986 Chernobyl accident triggered both intensive discussion about those limitations and an eventual revision process of the existing regimes. The so-called second generation of nuclear liability conventions was established thereafter. Those conventions include the Joint Protocol Relating to the Application of the Vienna Convention and the Paris Convention (Joint Protocol), (100) the Protocol to Amend the Vienna Convention on Civil Liability for Nuclear Damage (Protocol to the Vienna Convention), (101) the Convention on Supplementary Compensation for Nuclear Damage (CSC), (102) the Protocol to Amend the Convention on Third Party Liability in the Field of Nuclear Energy (Protocol to the Paris Convention), (103) and the Protocol to Amend the Convention of January 31, 1963 Supplementary to the Convention of July 29, 1960 on Third Party Liability in the Field of Nuclear Energy (Protocol to the Brussels Supplementary Convention). (104)
The Paris Convention establishes a system of absolute liability. (105) According to this system, the operator is liable for damage caused by a nuclear incident in a nuclear installation or involving nuclear substances coming from such installations. (106) Similar stipulations regarding absolute liability and exonerations can also be found under the Vienna Convention. (107) The conventions of the second generation have not changed the principle that strict liability applies to the operator of a nuclear power plant. However, an important change took place as far as the operator's available defenses are concerned: Natural disasters are no longer an applicable defense. (108)
Under the Paris Convention, liability is channeled to operators. No one else is liable for the damage caused by a nuclear incident. (109) The "operator" is defined as "the person designated or recognized by the competent public authority as the operator of that installation." (110) The Vienna Convention also has similar provisions. (111)
Under the Paris Convention and the Vienna Convention, the operator's liability is limited both in amount and in time. The Paris Convention sets the maximum liability of the operator at 15 million SDRs but allows the Contracting Party to establish a greater or lesser amount by legislation considering the capacity of insurance and financial security. The Contracting Party can also require a lower amount of liability according to the nature of the installation. The lower amount should be no less than 5 million SDRs. (112) By contrast, the Vienna Convention sets the cap of liability at no less than USD 5 million. (113)
The liability limitation has, however, been changed under the second-generation nuclear conventions. The Protocol to the Paris Convention increases the limit for nuclear operators to no less than EUR 700 million. The Contracting Party can reduce the liability to no less than EUR 70 million for an incident originating from a nuclear installation, or no less than EUR 80 million for the carriage of nuclear substances according to the reduced risks. (114) The Convention even allows for the adoption of unlimited liability by the Contracting Parties, as long as the financial security required is no less than the amount mentioned above. (115)
Seeking financial security coverage for the operator's liability is important for the international regimes on nuclear liability. Both conventions require the operator to have and maintain insurance or other financial security up to its liability cap. (116) In addition, it should be mentioned that the Brussels Supplementary Convention added two additional layers of compensation via public funds. They will be discussed below when discussing additional compensation mechanisms. (117)
3. Marine Oil Pollution
The international oil pollution compensation system consists of two important conventions that are interrelated. The first is the International Convention on Civil Liability for Oil Pollution Damage (CLC), (118) and the second is the International Convention on the Establishment of an International Fund for Compensation for Oil Pollution Damage (Fund Convention). (119)
The CLC adopts strict liability. (120) Hot debates took place during the negotiation of the convention with regard to whom the liability should rest on. Under the influence of the international regimes for nuclear liability, no doubts have been expressed on the reasonability of the channeling of liability. The debates focused on whether the shipowner or the oil industry should bear the liability. In the end, a compromise was made: Liability under the CLC fell on the shoulders of the shipowner. In return, the oil industry also needed to contribute to compensation through a compensation fund. At the conference on the passage of the 1969 CLC, parties agreed that an international compensation fund would be established in the near future. (121)
From what was just mentioned as far as the creation of the 1969 CLC and the 1971 Fund Convention is concerned, it is clear that the 1969 CLC channeled liability to the tanker owner. The ship-owner is defined as "the person or persons registered as the owner of the ship or, in the absence of registration, the person or persons owning the ship." (122) However, the 1969 CLC preempts other legislation: No other claims are eligible other than those under the Convention. It shows explicitly that no claims are made against the servants or agents of the owner. (123)
The liability established under the 1969 CLC was capped at FF 210 million or FF 2,000 for each ton of the ship's tonnage. Several serious oil spills that happened after the adoption of the 1969 CLC and the Fund Convention--for example, the Amoco Cadiz in 1978 and the Tanio in 1980--triggered the revisions to the original conventions. The first protocols to revise the conventions were drafted in 1984. Since the United States did not ratify the protocols, they could not come into force. Nevertheless, the changes in the 1984 protocols were largely incorporated in the 1992 conventions. (124)
In 1992, two protocols were adopted to revise the original conventions: The 1992 CLC and the 1992 Fund Convention. The 1992 CLC increased the liability limit to 4.51 million SDRs or 89.77 million SDRs, depending on the size of the ships. As a compromise for increasing the liability limit, the criteria preventing shipowners from limiting their liability were further constricted: Damage must result from their willful misconduct. (125)
The 1969 CLC requires the owner of a ship that is registered in a Contracting State and carries more than 2,000 tons of oil in bulk as cargo to maintain insurance or other financial security up to his liability limits. In addition to insurance, financial security can also be in the form of a bank guarantee or certificate delivered by an international compensation fund. (126) The 1992 CLC retained this requirement. (127)
4. Other Relevant Treaties
a. HNS Convention
Since the CLC only covers oil pollution damage, a separate convention was drafted to address liability and compensation for damage connected to the carriage of hazardous and noxious substances by sea: The HNS Convention. (128) This convention governs personal injuries, environmental damage, economic losses, and the costs of preventive measures which resulted from the carriage of hazardous and noxious substances by vessels. (129) The HNS Convention imposes a strict liability regime on the owner of the ship. According to Article 1, the shipowner is "the person or persons registered as the owner of the ship or, in the absence of registration, the person or persons owning the ship." (130)
The shipowner's liability under the HNS Convention is again capped. According to a calculation scheme provided in Article 9 of the Convention, the shipowner is entitled to limit his liability depending on the tonnage of the ship and on the form of the transported substances (bulk HNS or packaged HNS). (131) If the amount of the damage exceeds the owner's liability cap, the HNS Fund will step in.
Moreover, according to Article 12 of the HNS Convention, the owner of a ship carrying hazardous and noxious substances is required to take insurance or to maintain other acceptable financial securities to cover sums fixed by applying the limits of liability. (132) Since the HNS Convention is structured in the same way as the conventions on marine pollution discussed above, it will not be analyzed in further detail.
b. Transboundary Civil Liability Protocol
In execution of the relevant provisions of the Convention on the Protection and Use of Transboundary Watercourses and international Lakes and the Convention on the Transboundary Effects of Industrial Accidents, both signed in 1992, a transboundary civil liability protocol was adopted on May 21, 2003. (133) However, this protocol is not yet in force. (134) The preamble to the Transboundary Civil Liability Protocol refers to the "polluter pays" principle as the basis for the liability regime. (135) Based on the "polluter pays" principle, operators are strictly liable for the damage caused by industrial accidents. (136) Furthermore, fault-based liability might also apply to other persons such as servants or agents of the operator if they cause or contribute to damage through their "wrongful[,] intentional, reckless or negligent acts or omissions." (137)
The concept of "operator" has been defined in Article 1(e) of the 1992 Convention on the Transboundary Effects of Industrial Accidents as "any natural or legal person, including public authorities, in charge of an activity, e.g., supervising, planning to carry out or carrying out an activity." (138) The protocol includes financial limits to the amount of compensation. Operators must accordingly ensure that they have financial security covering at least the minimum limits specified in Annex 11 of the Protocol, namely, 2.5 million units of account for Category A hazardous activities and 10 million units of account for both Categories B and C. (139)
c. Space Liability
The 1971 Convention on the International Liability for Damage Caused by Space Objects (Space Liability Convention) provides that "[a] launching State shall be absolutely liable to pay compensation for damage caused by its space object on the surface of the earth or to aircraft...." (140) Furthermore, a launching State shall also be liable for damage due to its faults in space. The convention imposes absolute liability, which refers to "a liability system without any defense or exclusion." The Space Liability Convention refers to the concept of "launching States," as defined in Article 1(c): "(i) [A] state which launches or procures the launching of a space object; (ii) a State from whose territory or facility a space object is launched." (141) The term "launching" also includes attempted launching. (142)
In the preamble to the Convention, the need is recognized to ensure "prompt payment ... of a full and equitable measure of compensation to victims" of damage caused by space objects. (143) The term "damage" is defined in Article 1 of the Convention and refers to "loss of life, personal injury or other impairment of health; or loss of or damage to property of States or of persons, natural or juridical, or property of international intergovernmental organisations." (144)
Having briefly sketched the liability regimes adopted by international environmental agreements, we will now proceed (1) to a comparison of the liability regimes described in Subpart 1; and (2) to a critical analysis of those regimes in light of the conditions for designing optimal liability rules explored above in Subpart 2. (145) This will lead to the conclusion that the liability regimes incorporated in the international conventions show important limits, at least from the perspective of the optimal design. Interestingly, some domestic liability regimes, particularly in the United States, seem better able to provide incentives for disaster risk mitigation. That is why the United States did not join the international conventions with respect to marine oil pollution and nuclear liability, but rather drafted its own separate regime. Analyzing those regimes can provide interesting lessons for a better design of international conventions, discussed in Subpart 3.
1. Liability Regimes Compared
Surprisingly, the liability regimes in the international conventions just discussed show some striking similarities. For example, all conventions introduce strict liability. There are a few nuances in the sense that some conventions adopt a liability regime that is even stricter than the others. For example, the nuclear liability conventions refer to "absolute liability," as does the Space Liability Convention. Indeed, the latter seems to be "absolute" in the sense that it does not allow any defenses whatsoever. A second feature common to all conventions is a requirement for some form of mandatory financial security. An overview of the conventions reveals that the imposition of a duty to provide financial security in order to meet the strict liability threshold appears to be the state of the art at the international level. Moreover, the liability regimes discussed demonstrate that operators can have a broad set of options to provide proof of their solvency. Proving solvency should therefore not necessarily be limited to insurance. For example, the Rome Convention and the Transboundary Civil Liability Protocol refer to a variety of financial security mechanisms such as cash deposits, bank guarantees, and guarantees by the Contracting State. (146) These examples would be valuable tools to cover liability should one consider introducing a duty to provide financial security.
Liability was limited in six of the seven international conventions discussed above. The only convention that does not apply liability limitation is the Space Liability Convention. A liability cap can be limited in different ways; in some cases the convention itself establishes the cap (like in the marine pollution convention, the HNS Convention, and the conventions related to aviation), but the caps are dependent upon the nature and amount of either the cargo transported or the tonnage of the ship. In other cases, the conventions provide the scope for liability limitation, but the precise amount of the cap can be set through domestic law (that is more particularly the case in the conventions on nuclear liability).
Channeling liability to the operator is yet another of the conventions' interesting features. Again, although the formulation can be different, liability channeling can be found in all seven conventions that have provisions regarding liability. The difference in the way the channeling works relates more to whether liability of other agents is still possible in principle or whether it is totally excluded. For example, in the marine pollution conventions, others besides the ship owner are in principle excluded from liability, whereas in the aviation conventions the liability is channeled to the operator/air carrier, but the possibility of holding others liable as well is not completely excluded.
If one were to summarize these findings, one would notice that four important features appear in those international conventions: (1) strict liability, (2) compulsory financial guarantees, (3) a financial limit on the liability, and (4) a channeling of the liability. A fifth feature, present in some conventions--in fact only in the nuclear liability, marine pollution, and HNS Conventions--is the presence of an additional layer of compensation. (147)
2. Critical Analysis
It is interesting to note how those four characteristics of liability provisions in international environmental agreements compare to the theoretical starting points for the optimal design of a liability regime explored earlier in the article. (148) The first element, the introduction of strict liability, is clearly in line with the theoretical starting points. All activities regulated by the international conventions (nuclear liability, aviation, marine pollution) can certainly be considered ultra-hazardous activities, which clearly correspond with the economic criteria that favor choosing a strict liability rule. (149) Those ultra-hazardous activities are cases where the influence of the operator is much more important than the influence of the victim. (150)
Likewise, the introduction of financial guarantees is also clearly in line with the theoretical starting points. (151) Without guarantees, strict liability itself would be meaningless. A specific feature of catastrophes is precisely their ability to lead to damage that can easily outweigh the capacity of an individual operator. Financial security mechanisms are hence necessary to avoid operators externalizing risk to society as a result of their insolvency. (152) Through insurance, ex ante incentives for disaster risk mitigation could be provided via the control of moral hazard by the insurance company. (153)
The third characteristic, the limitation of liability in six of the seven international conventions, is more problematic. Indeed, most international conventions appear to pay an important price for the strict liability of the operator: A financial cap is imposed on his liability. (154) The literature with respect to the nuclear and marine pollution conventions is very critical with respect to the financial cap on liability. (155) The criticism is directed on one hand to the fact that the operator will only have incentives to prevent an accident up to the amount of the limited liability. A financial cap could therefore reduce the incentives for prevention. On the other hand, to the extent that the actual damage could be substantially larger than the capped liability, victims would remain uncompensated. This leads to a third type of criticism, arguing that a financial cap on liability allows operators to de facto externalize harm to society. In fact, from an economic perspective a financial cap constitutes a subsidy to an industry, which may be problematic in itself.
Channeling liability to the operator--the fourth feature of the liability. regime in the international conventions--is undoubtedly equally problematic from the point of view of deterrence. The simple reason is that channeling liability to the operator prevents holding liable others who could equally contribute to the risk of a disaster. That would consequently reduce their incentives for disaster risk mitigation. Such liability channeling is therefore highly criticized in the literature. (156)
In summary, the liability regimes in the international conventions introduce strict liability and compulsory financial guarantees. However, these conventions problematically limit liability and, moreover, channel liability to the operator. Both of these features are especially problematic when taken in combination. Liability limitations not only provide operators insufficient incentives for disaster risk mitigation, but may also cause victims to remain uncompensated. Moreover, channeling liability to the operator excludes the possibility for victims to bring a lawsuit against anyone other than the operator to whom liability was channeled. That not only highlights the potential undercompensation of victims, but also the lack of deterrent effect. Channeling de facto excludes anyone other than the operator from liability and thus does not provide incentives to others, like suppliers or other stakeholders involved in creating the risk. As a result, channeling can increase the danger of a disaster resulting from a technological risk.
The question obviously arises as to the extent to which these features of liability regimes will effectively lead to undercompensation and underdeterrence. This may to an important extent depend on the specific design features of the regime. For example, channeling is most problematic when it is exclusive (meaning that victims are not allowed to bring suits against anyone other than the channeled operator) and when it is absolute (in the sense of not allowing recourse against other stakeholders that equally contributed to the risk). As far as the financial cap is concerned, it may be clear that capping liability is most problematic where the financial cap is set at a much lower amount than the expected damage resulting from the disaster, which is arguably the case with nuclear liability in the international conventions. (157) If, to the contrary, limited liability may still be able to provide adequate compensation, and hence deterrence for most of the accidents, meaning that the total damage will likely not often be higher than the amount of the cap, capping liability is certainly less problematic. This situation is arguably the case in the marine pollution regime. Finally, the problematic nature of the liability regimes in the conventions also depends upon another feature yet to be discussed (158)--whether additional compensation mechanisms are available and how they are financed. If a substantial amount of compensation were available in situations where the operator's liability is capped, the regime would not necessarily lead to undercompensation. Moreover, depending on how the second layer is financed, either by stakeholders equally exposed to the risk or by the government, the second layer of compensation could still provide additional incentives for disaster risk mitigation. Hence, it is only possible to pass a final judgment on the effectiveness of the liability regimes in the multilateral environmental treaties when additional compensation regimes have also been analyzed.
We just showed that the liability regimes found in the international environmental agreements all display a similar pattern of strict liability, compulsory financial guarantees, financial caps, and liability channeling. It is no accident that the conventions show similar features. These features originated from the nuclear liability conventions in the 1960s and were introduced at the time because they clearly served the interests of industry. (159) The fact that those features were present in the nuclear liability conventions was subsequently also used as a justification to include them in the conventions with respect to marine oil pollution when these were created. (160) "Path dependency" apparently influences the fact that future conventions dealing with liability followed a similar pattern. However, that is surely not the only way in which a liability regime for catastrophic risk could be created. Domestic law provides interesting examples of different liability regimes that do not show the same problematic features we discovered in the international conventions and that may hence come closer to the optimal liability regime design aimed at disaster risk mitigation. A few alternatives from domestic law will be presented in order to show that alternatives are not only possible, but also already in place in several legal systems.
a. U.S. Price-Anderson Act
In the United States, nuclear liability is governed by the Price-Anderson Act, which was adopted in 1957. (161) One specific feature of the Price-Anderson Act is that it has been regularly revised at approximately ten years intervals. (162) Thanks to those revisions the Act has been able to adapt to changing circumstances much better than the international regime.
The initial Price-Anderson Act, just like the international compensation regime, wanted to spread the risk of nuclear activities between the private industry on the one hand, and the nation that benefits from the development of nuclear energy on the other. The nuclear operator needed to buy all the insurance coverage presently available, which at the time was USD 60 million. As will be explained below, initially government funds were made available to cover sums above that amount, but in 1975 the financing of the second layer shifted from public to private funding. (163)
Today the Price-Anderson Act has the practical effect of imposing strict liability for nuclear incidents. This is realized via the extraordinary nuclear occurrence (ENO), introduced in the Price-Anderson Act in 1966; (164) in case of a nuclear accident, the Nuclear Regulatory Commission is given the power to determine whether or not there is an extraordinary nuclear occurrence. (165) If that is the case, the nuclear operator will not be able to invoke certain defenses that are available under federal law or state tort law. (166)
There are a few differences between the Price-Anderson Act and the international regime. First, the liability of the power plant operator under the Act has been increased to USD 375 million. (167) That is substantially more than the amounts currently available for operator liability in the international regime. The Paris Convention sets the financial cap between USD 5 million and USD 15 million special drawing rights. (168) According to the Vienna Convention, the liability cap should be no less than USD 5 million. (169) After the 1986 Chernobyl accident the financial limit was increased so that the limit for nuclear operators would be no less than EUR 700 million. (170) Moreover, according to a Protocol to the Vienna Convention, the liability limit would be increased to no less than USD 300 million or no less than 150 million SDR provided the installation state makes public funds available to cover the amount between the limitation and USD 300 million. (171) However, these additional protocols which would substantially raise the operator's amount of financial liability have not yet entered into force. (172) Today the liability limit under the U.S. Price-Anderson Act is hence still substantially higher than the operator liability under the international conventions.
A second important difference will be highlighted below: Whereas the international conventions rely on a second tier of compensation (beyond the liability limit of the operator) stemming largely from public funding (by the operator's state and by all contracting states), the U.S. Price-Anderson Act relies on a system of private funding to provide an additional layer of compensation.
The third important difference is that the U.S. Price-Anderson Act did not incorporate the highly criticized liability channeling system, whereby any stakeholders other than the operator are not subject to liability. This is related to the fact that nuclear liability insurance under the Price-Anderson Act provides a system of so-called "omnibus"-coverage. This means that, per the Act, liability insurance provides cover for "anyone who may be liable" (173) for "public liability." The Act defines "public liability" as "any legal liability arising out of or resulting from a nuclear incident or precautionary evacuation...." (174) The only exceptions to that definition pertain to "claims arising out of an act of war," worker's compensation claims, and claims for damage to on-site property at a licensed nuclear facility. (175) As a result of this provision, everyone who can be held liable for the damage from a nuclear accident (including the supplier) can benefit from the liability insurance coverage of the nuclear operator. The mechanics of this system were demonstrated following the Three Mile Island accident, where a single law firm represented all defendants--the nuclear operator as well as the designer and constructor of the nuclear power plant. (176) Thus, unlike the international compensation regime, the Price-Anderson Act has a system of economic rather than legal channeling. (177)
It is interesting to note that the United States played a decisive role in the initiation and drafting of the specific text of the international nuclear liability conventions. (178) The U.S. interest in an international liability convention was due to the United States' essential monopoly on nuclear knowledge and technology in the 1950s. The developing Western-European economy was interested in acquiring the U.S. technology and nuclear material. U.S. suppliers protected themselves through so-called "hold-harmless" clauses. Under such clauses, the European nuclear operator who would have purchased nuclear material from a U.S. supplier would hold the U.S. supplier harmless for all claims that might arise. (179) However, there was uncertainty with respect to the ability of those clauses to provide sufficient protection to the U.S. suppliers.
Consequently, the U.S. "Atomic Industrial Forum" conducted two studies on the possible liability claims of nuclear accident victims against U.S. suppliers. The first study was the "Preliminary Report on Financial Protection against Atomic Hazards." (180) The second study was titled "International Problems of Financial Protection against Nuclear Risk." (181) The preliminary report basically dealt with the liability regime for domestic accidents. It concluded that the interests of both the industry and the public could be met by limiting the nuclear operator's liability to the amount of coverage available on the insurance market and by providing public funds for damage not covered by the operator or his insurer. Concerns regarding potential transboundary nuclear accidents were addressed in the Harvard Report. The Harvard Report suggested principles like liability channeling, strict liability, liability limitation, etc. that were later included in the international conventions. According to the drafters of the Report, those principles were an effective tool to protect U.S. "Atomic Suppliers" since channeling liability to the European power plant operators would preclude any liability claims against U.S. suppliers. (182) The draft of what later became the 1960 Paris Convention on Third Party Liability in the Field of Nuclear Energy was even literally annexed to the Harvard Report.
It is therefore striking that, on the one hand, the United States has insisted on introducing legal liability channeling in the nuclear liability conventions, but on the other hand, U.S. domestic nuclear liability law--the Price-Anderson Act--provides for a system of economic channeling. The United States has hence used its power to influence the nuclear liability conventions with the sole purpose of protecting its own companies, who are world leaders in the supply of goods and services to the nuclear industry. (183)
b. Unlimited Liability
The legislation of several E.U. Member States, moreover, shows that it is possible to apply both strict liability and unlimited liability to the activities of nuclear power plant operators. The European Commission provides an interesting overview of a nuclear power plant operator's liability in E.U. Member States. (184)
A few E.U. Member States impose unlimited liability upon operators, but some of those (like Austria) do not have nuclear power plants. Nonetheless, others who do have nuclear power plants still impose unlimited liability. A striking example is Germany, even though the country is currently undertaking a phase-out of nuclear energy. Germany is a party to the Paris Convention and to the Brussels Supplementary Convention and therefore requires financial security to cover the operator's liability up to EUR 2.5 billion. (186) As the table shows, this is the highest amount of financial security required amongst the European Member States. Therefore, in principle Germany provides for unlimited liability, but in practice it only imposes a duty to provide financial guarantees up to a maximum of EUR 2.5 billion. This constitutes yet another example of domestic law providing wider-reaching liability, and thus better incentives for disaster risk mitigation, than the international regime.
c. U.S. Oil Pollution Act
The relationship between the United States and the marine oil pollution conventions is to some extent comparable to the role the United States played with respect to the international nuclear liability conventions. (187) The United States was involved in the enactment of the International Convention on Civil Liability for Oil Pollution Damage 1969 (CLC) and the International Convention on the Establishment of an International Fund for Compensation for Oil Pollution Damage 1971 (Fund Convention). However, the United States did not join the international conventions. After new major incidents demonstrated the inadequacy of the 1969-1971 regime, (188) it was decided to draft protocols aimed, inter alia, at increasing the limits of liability and compensation and creating a new fund. (189) One of the legal conditions for the enactment of the 1984 Protocols was ratification by the United States and Japan since those were the two largest receivers of oil transported by ship. However, the United States did not ratify the 1984 Protocols, as a result of which they never entered into force. Moreover, since the United States was confronted with the Exxon Valdez incident in 1989, it decided to draft its own Oil Pollution Act, which came into effect in 1990. Hence, the International Maritime Organization continued international negotiations without the United States, as the enactment of the Oil Pollution Act 1990 had made clear that the United States would no longer join the international regime. The changes made by the 1984 Protocols were eventually largely implemented by the 1992 Protocols. (190)
The most important reason why the United States did not join the international conventions was related to their belief that liability limits in the conventions were too low. Moreover, they rejected the principle of channeling liability to ship owners. Adherence to the conventions would also preempt state laws which had unlimited liability. (191) In response to the 1989 Exxon Valdez accident, the U.S. Congress quickly passed the Oil Pollution Act 1990 (hereafter OP A). The OPA shares a few similarities with the international regime, such as strict liability, limited liability, and a compulsory financial guarantee. However, there are substantial differences as well: Liability is not channeled; under OPA the liability limits are much higher and, as will be explained below, allow potentially responsible parties to lose their right to limit liability. Moreover, the OPA does not preempt state laws, which means that states can still impose additional liability or financial responsibility. (192) It is important to mention that in the United States thirty of the fifty states have a coastline and, in fact, all of these states but six have legislation on vessels liability. (193)
Some states, such as Alaska, California, North Carolina, and Rhode Island, impose strict and unlimited civil liability for cleanup costs, natural resource damage, and private losses caused by oil pollution, including pure economic losses. In some other states, unlimited liability is only established for certain categories of damage, such as in Washington, for cleanup costs and damages to persons or property; Maryland, for cleanup costs, damage to real and personal property and natural resources damages; Massachusetts, for natural resources damages; and Florida, for natural resource damages, damage to real and personal property, and losses consequential upon property damage. (194)
In addition to being exposed to potentially unlimited liability under state law, the limits under OPA also involve conditions. OPA establishes limits on oil pollution liability for different types of facilities. A cap is established for the total sum of removal costs and damages. The cap is the greater of a per incident cap and a per gross ton cap. (195) In 2006, the Coast Guard and Maritime Transportation Act (CGMTA) increased the cap. (196) The Coast Guard and Maritime Transportation Act (CGMTA) also established different caps for single hull and double hull tankers. (197) Moreover, in 2009 the Coast Guard made a further increase to the cap. (198) The result can be summarized in the following table:
In 2009, the Coast Guard made its first consumer-price index (CPI) adjustment to the liability limits, increased the limits for double-hull tankers from 1900 to 2000 USD per gross ton, and increased single-hull tankers from 3000 to 3200 USD per gross ton. (199)
In spite of those caps, a responsible party can lose its right to limitation if the incident was proximately caused by "gross negligence or willful misconduct," or "the violation of an applicable Federal safety, construction, or operating regulation." (200) A responsible party may also face unlimited liability if it fails to report an incident, provide requested cooperation in connection with removal activities, or comply with an order of the President. (201) Hence, under the OPA, the amount of the financial cap is related to the preventive measures taken by the operator, and the limits are easily breakable.
Furthermore, unlike the international regime established through the CLC, OPA does not channel liability to one particular party but provides for joint and several liability. (202) OPA identified as responsible parties the ship owner, the operator, and the demise charterer. (203) All those features imply that OPA is better <6719028H_TB003> aligned with the economic principles of liability (204) and hence provides better incentives for disaster risk mitigation.
III. ADDITIONAL COMPENSATION MECHANISMS
As we outlined above, liability rules have a lot of potential to provide compensation to disaster victims and, when appropriately designed, can even provide adequate incentives for disaster risk mitigation. However, it was equally shown that liability rules have substantial limits as well. One problem is that their application is mostly limited to technological (man-made) disasters, which leaves victims of natural disasters uncompensated. Moreover, the design of the liability rules in international conventions is, as we showed, somewhat flawed due to its inclusion of channeling and low liability limits, as a result of which the positive incentive effects for operators will often not be attained.
In reality one can therefore observe a variety of alternative compensation mechanisms (205) which all usually aim at some form of recovery after a disaster. Those additional compensation mechanisms are in some cases unstructured and ad hoc (e.g., governments or other actors providing relief to victims); in other cases they involve structural compensation funds. Some mechanisms can be found in domestic law; others have a basis in international environmental agreements, usually in the conventions we discussed earlier that also introduce liability rules. The fund then constitutes an additional layer of compensation, supplementing the operators' liability.
In all those cases where additional compensation is provided, the crucial question again arises as to how those compensation mechanisms affect stakeholder incentives to take disaster risk mitigation measures. The incentives at stake in technological disasters are usually those of the operators. By contrast, the incentives that could be affected by ex post payments when natural disasters occur mostly belong to potential victims.
We first provide an inventory of the different models of additional compensation, showing their potential advantages and dangers. Next, we present examples of additional compensation mechanisms in international environmental agreements (A). Finally, we provide a critical analysis of these mechanisms (B).
A. Potential and Models
a. Technological Disasters
The main potential of additional compensation mechanisms lies in their ability to go beyond the limits of liability and insurance. One problem with seeking compensation for catastrophic risks via traditional liability rules is that the capacity of the operator and of traditional insurance markets may be limited, as a result of which operators may not be able to fully compensate the catastrophic risk they are causing. (206) Providing an additional compensation mechanism could advantageously generate higher capacity to deal with damage resulting from catastrophic risks. If additional capacity, for example of the state or of a pool of operators, (207) were enacted, risk spreading would generate higher capacity to deal with the catastrophic risk. Ultimately, an additional compensation mechanism could thus provide greater coverage, guaranteeing better compensation for victims.
A second potential advantage of using additional compensation mechanisms involves the ability to prevent operators from externalizing harm. It is well-known (208) that the deterrent effect of liability rules starts diminishing the moment the amount of the damage is substantially higher than the wealth of an operator. (209) Mandating solvency guarantees would solve that problem. Nevertheless, that solution would still leave another problem unresolved. (210) Capacity must also be available in order to provide security to the operator. To the extent that the additional compensation mechanism would charge risk-related premiums (and in that sense function as an insurer), the contributions to the additional compensation mechanism would provide operators incentives for disaster risk mitigation. Yet a crucial condition for the effectiveness of the additional compensation mechanism in providing incentives for disaster risk mitigation requires that contributions be related to the risk posed by the operator, as outlined above. (211) Incentives for prevention could only be generated via risk-related contributions. If, to the contrary, contributions were not risk-related (e.g., did involve a general tax), the additional compensation mechanism could even produce negative incentives, since safer operators would then de facto cross-subsidize riskier ones. This negative effect on incentives for disaster risk mitigation would certainly apply when the additional compensation mechanism could no longer be financed by the operators that create the risk but, for example, by the government. In that scenario, the additional compensation mechanism would constitute a subsidy to operators and would allow them to externalize the disaster risk to society. Based on that hypothesis, the additional compensation mechanism would even increase the disaster risk in lieu of providing incentives for disaster risk mitigation.
b. Natural Catastrophes
In the case of natural catastrophes no operator can be made liable for the consequences of the catastrophe. Hence the goal of an additional compensation mechanism is not related to providing incentives for operators. Rather, its primary goal is to generate financial capacity for recovery in the form of infrastructure reconstruction and victim compensation. Again, in the case of natural catastrophes, a compensation mechanism (212) could have positive effects on society. Leaving victims of catastrophes without any relief would probably be incompatible with the concept of the welfare state, at least as it is conceived in most E.U. Member States. (213) A disaster can potentially lead to a total disruption of society, and ex post recovery can therefore fulfill an important function by, for example, restoring infrastructure and, to the extent possible, bringing life after the disaster back to normality. There is also a strong belief that providing relief in the immediate aftermath of a disaster is one of the principal functions of government. (214)
Another potential rationale for ex post recovery would be that, although ex post payments in the case of a natural catastrophe cannot affect the incentives of potential tortfeasors, they might affect the incentives of government. The prospect of large-scale payments in the aftermath of a disaster might encourage the government to take cost-benefit justified precautions long before disasters strike. (215) This argument may be of particular strength in the case of terrorism. The goal of terrorist attacks is often to disrupt society. Providing ex post relief may then help to restore public trust. Moreover, terrorists may adopt adaptive strategies to which governments can potentially react better than individuals. Additionally, in the case of terrorist attacks there is little one can expect from civilians as far as preventive measures are concerned. The same is obviously true of large-scale infrastructural works which would be necessary to prevent catastrophes (e.g., building dykes against tsunamis). These are typically public goods that are not provided through private action and may therefore require government intervention.
A related argument in favor of government intervention is that government has the capacity to diversify the risks over the entire population and to allocate past losses to future generations, thereby creating a form of cross-time diversification, which the market could not achieve. (216) Likewise, the argument can be made that, as far as the terrorism risk is concerned, the risk of terrorist attacks is partly in the government's control and the government may have more information on ongoing terrorist groups' activities through intelligence services. (217) Hence, in some cases the government may be in the best position to prevent disasters. Government intervention would from this perspective provide incentives to politicians to invest in preventive measures. A few examples that come to mind are the building of dykes against the risk of a tsunami, taking measures to prohibit building residences on the slopes of volcanoes (like Mount Merapi in Indonesia), or effective zoning and planning decisions to avoid locating residences in flood-prone areas.
Moreover, one could also make the argument that government can in principle compensate without limits. If the damage were to exceed the current budgetary possibilities of the government, the aforementioned cross-diversification over time and future generations could in principle take place.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Abstract through III. Additional Compensation Mechanisms A. Potential and Models 1. Potential, p. 95-136|
|Author:||Faure, Michael G.|
|Publication:||Stanford Journal of International Law|
|Date:||Jan 1, 2016|
|Previous Article:||Shareholder protection reloaded: redesigning the matrix of shareholder claims for reflective loss.|
|Next Article:||In the aftermath of the disaster: liability and compensation mechanisms as tools to reduce disaster risks.|