INTERNATIONAL AIR CARRIER LIABILITY AND SOCIAL PERCEPTIONS OF RISK: PASSENGERS � THE CONVENTIONS BEFORE THE COURTS AND RISK PERCEPTIONS
I. Introduction
5.1 There are only a handful of cases which define the legal framework for international carriage by air when compared to general tort case law. Most importantly, there are even fewer judgments of courts of last resort. This means that the judicial delineation of the provisions of the WCS/HC rests predominantly with courts of first instance and appellate ones, which are more inclined to use notions of tort law when analysing them. If one considers that a number of countries apply the WCS/HC to their domestic carriage, courts are becoming increasingly accustomed to applying their domestic tort law to interpret them. As a result, it is becoming easier and more justifiable to blur – intentionally or unintentionally – the boundary between tort law and Convention-type principles and reasoning when dealing with international carriage by air.
5.2 A key element in this interplay is the principle of exclusivity, which both the WC/HC and the MC embody in Articles 24 and 29 respectively. This means that the Conventions are the exclusive cause of action for any claim by a passenger against a carrier for death, wounding, or bodily injury suffered during international carriage by air. In practical terms, it suggests that once a passenger is in the process of embarking, she/he has no resort to contractual or negligence-based recovery in the event of a claim. The only recovery will be as allowed by the Conventions.
II. The Exclusivity of the Conventions as a Liability Control Technique
5.3 On the face of it, Articles 24 and 29 safeguard the provisions of the WC/HP/MC against erosion from the influence of tort law, and as such convey the intention of the drafters to create a self-determining legal system. The independency of the WCS was not enjoyed automatically, but instead was a matter of great debate between 1991 and 1999. During this period, claimants in the UK and the US exerted pressure on the courts to allow contractual and negligence claims to exist side-by-side with the WCS. Until then, courts in the US were “split on the question of whether the Convention pre-empted state causes of action once they determine the facts fall within Warsaw”.1 Yet the 1990s saw a significant rise in endeavours to circumvent the WCS.2 It is not a coincidence that these attempts intensified in the 1990s, a period when the private international air law regime was in tatters and the socio-technological realities of the risk society started affecting the aviation industry.
5.4 Litigation was used as a means to rebuild the trust relation between passengers and carriers which was undermined by the fragmented legal framework. This way, the trust relation would be based on the social perceptions of risk and not on an “objective” legal instrument that intentionally sought to compromise the interests of passengers for the sake of carriers. In essence, claimants attacked the provision that detaches the WCS from tort law and expressed their dissatisfaction with the risk allocation achieved therein. Counter-arguments, such as the need to protect the industry and the difficulty to prove fault in the aftermath of accidents were not persuasive anymore. The perception that air carriers were “deep-pocketed” corporations that paid more attention to cost-benefit analyses than to the needs of their customers was already well-founded.
5.5 Furthermore, claimants and their lawyers had better access to medical and technological analysis, which in turn made the requirement to prove fault an easier hurdle to overcome. As such, it was thought preferable to resolve disputes by resorting to tort law, although it had already come under fire for being an expensive and unjust system, rather than use the WCS, which was perceived as anachronistic and protective of the industry.
5.6 It is true that the 1980s and early 1990s witnessed large damage awards against multi-national corporations which inevitably attracted media interest and, in the words of an English solicitor, “brought the courtroom into the living room”.3 The level and publicity of these awards concealed the deficiencies of the tort system and at the same time exercised more pressure upon the WCS, which imposed a number of substantive and procedural limitations upon recovery. Having levelled aviation with the rest of the corporate world, the average claimant in an aviation accident case governed by the WCS thought carriers were insulating themselves from accountability for the harms they were causing to society. As such, he/she wanted air carriers to be directly exposed to tort law which then seemed like a larger piece of the pie.
5.7 Especially in the US, where tort law was used extensively to decide cases arising from domestic carriage, air carriers were accused of having double standards in the way they were treating domestic and international passengers:4 the coexistence of state law causes of action with the Conventions would have meant the explicit expansion of carriers’ liability for modern policy reasons – “primarily that the defendants are well positioned both to protect the public from harm and, when harm does occur, to bear the burden of compensation by spreading losses through liability insurance and/or via the prices they charge for what they sell”.5 Although UK courts prima facie are less political than their US counterparts and juries play no role in civil trials, a potential dismantling of the WCS in favour of common law would have meant that the liability of carriers was purposefully increased to fit the socio-legal demands.
5.8 Much of the debate about the exclusivity of the WCS reflects the difficulty of US courts to conclude that the WC creates an independent cause of action. Up until 1978 the prevailing view in the US was that the WC does not create a cause of action but simply limits the liability of air carriers:
“the actual basis of liability [did not] arise from the WC but from the contract of carriage which had been entered into by the parties. According to this view, the purpose of the WC was therefore supposed to be merely to subject this liability in cases pursuant to Articles 17 to 19 to certain conditions and limitations. In order to prevent an avoidance of these conditions and limitations with the help of other bases of claim (particularly tortious) which were applicable at the same time, it was consequently necessary (by way of Article 24 of the WC) to expand them to these conditions and limitations.”6
This would have meant that the claimants “would have to seek some other legal basis for the claim (which might lead to difficult issues in the conflict of laws) on to which the Conventions would have to be grafted”.7
5.9 A long line of judgments starting in 1978 with the case of Benjamins v. British European Airways8 established beyond doubt that the WC itself shall be an independent cause of action for death or personal injury.9 This way, it aligned US with UK jurisprudence that had long ago treated the WC/HP as an independent legal basis on the basis of special legislation enacted.10 As to question what sort of action is created by the WCS/MC, Shawcross provide a plausible explanation:
“… there is merit in the view taken in a New South Wales court (Proctor v. Jet Aviation Pty Ltd [1982] 2 NSWLR 264) that although founded on a contract of carriage the action is simple ‘a cause of action created by statute’, as there is in the view of the Supreme Court of Ohio that, like other actions involving carriers, it incorporates elements of both contract and tort. The language of the Montreal Convention 1999 [in Article 29] would seem to suggest that an action under the Convention would be sui generis, neither in contract nor in tort.”11
5.10 However, this conclusion was just the beginning of the story, since having an independent cause of action did not necessarily mean that it was exclusive as well. Courts in the UK, US, and Germany did not face much difficulty in coming to the right conclusion that in cases where the carrier is liable under the WCS no other domestic cause of action is available to the claimants.12
5.11 Yet the issue whether the WCS applies to the exclusion of domestic law in cases which fall within its scope but the carrier is not found liable under Article 17 was not easy for the courts to resolve. For a long period of time the prevailing opinion in the US and a strong opinion in the UK was that such a claim can be brought under domestic laws, limiting thus substantially the exclusive characteristics of the WCS.13
5.12 Without doubt, the exclusivity of the WCS was a critical issue for air carriers.14 In that respect, air carriers heaved a sigh of relief when the House of Lords in the Sidhu case and the SCt in the Tseng case confirmed in unequivocal terms the exclusive character of the WCS. Lord Hope clarified the point that passengers do not have recourse to domestic law when the carrier is liable under Article 17:
“It would be largely destructive of the system which this chapter seems to have been designed to lay down if a passenger were to be able, for example, to maintain a claim of damages for non-bodily injury, for loss of or damage to the personal possessions which he had with him inside the aircraft or for economic loss, outside the conditions and limits set by the Convention while maintaining a claim under the Convention for the bodily injury.”15
5.13 Ginsburg J of the SCt, quoting O’Connor J in Saks, concurred with Lord Hope: “The text, drafting history, and underlying purpose of the Convention, in sum, counsel us to adhere to a view of the treaty’s exclusivity shared by our treaty partners.”16
5.14 Then, Lord Hope provided a very strong argument for the proposition that passengers do not have recourse to domestic law when the carrier is not liable under Article 17:
“The phrase ‘the cases covered by Article 17’ extends therefore to all claims made by the passenger against the carrier arising out of international carriage by air, other than claims for damage to his registered baggage which must be dealt with under Article 18 and claims for delay which must be dealt with under Article 19. The words ‘however founded’ which appear in Article 24(1) and are applied to passenger’s claims by Article 24(2) support this approach. The intention seems to be to provide a secure regime, within which the restriction on the carrier’s freedom of contract is to operate. Benefits are given to the passenger in return, but only in clearly defined circumstances to which the limits of liability set out by the Convention are to apply. To permit exceptions, whereby a passenger could sue outwith the Convention for losses sustained in the course of international carriage by air, would distort the whole system, even in cases for which the Convention did not create any liability on the part of the carrier. Thus the purpose is to ensure that, in all questions relating to the carrier’s liability, it is the provisions of the Convention which apply and that the passenger does not have access to any other remedies, whether under the common law or otherwise, which may be available within the particular country where he chooses to raise his action. The carrier does not need to make provision for the risk of being subjected to such remedies, because the whole matter is regulated by the Convention.”17
5.15 The SCt was again in agreement with Lord Hope: “… we hold that the Warsaw Convention precludes a passenger from maintaining an action for personal injury damages under local law when her claim does not satisfy the conditions for liability under the Convention”.18
5.16 These two judgments, in combination with the linguistic amendments in Article 24 of MP4 and Article 29 of MC, arguably make it difficult for claimants to circumvent the WCS/MC and undeniably constitute a victory for carriers. Yet it was a Pyrrhic victory because it brought a sea change in litigation strategy and in judicial perceptions:
“the abhorrence that many courts would have to leaving a claimant legally naked, without any remedy whatsoever, would encourage them to expand the circumstances under which an ‘accident’ and ‘injury’ were deemed to exist, bringing claims previously outside Warsaw, within it”.19
5.17 It is difficult to quantify the actual effect of the judgments on litigation rates and judicial behaviour. Yet one can safely argue that they gave to the Conventions the perception of autonomy and self-sufficiency that their drafters were aiming for. Following decades of uncertainty, the legal regime of the Conventions got judicial approval to regulate the liability of air carriers in an exclusive manner. However, the judicial reiteration of the exclusive and uniform application of the Conventions did not travel alone. It put a strain on their liability system and induced claimants to put forward a broad interpretation of the term “accident” through a tort law prism rather than a direct application of domestic law.20
5.18 This way, the Conventions became a field for experimentation, with both carriers and passengers instilling tort law concepts in their pleadings while at the same time the judiciary tried to come to terms with the disguised attempts to inject domestic law principles. It would be an overstatement to argue that the judgments in Sidhu and Tseng triggered the “tortisation” of the Conventions. This actually proceeded by small incremental advances from the aftermath of WWII, and can be attributed to the failure of the Conventions to correspond to the evolving social perceptions of risk. Rather, the judgments and the textual amendments acknowledged the trend and tried to preserve the allocation of risks achieved in the Conventions by declaring their independence from domestic law.
5.19 In that respect, it is more accurate to say that the judgments did prevent the direct “tortisation” of the Conventions by: (i) clarifying their external boundaries; and (ii) preserving prima facie their balance of interests in accordance with the spirit of the Conventions and the requests of carriers. The HL and the SCt in these two occasions did their best to keep lay risk perceptions away from the Conventions’ system of attributing liability, and thus maintained their integrity and “objective” character.
5.21 Instead it proceeded by small incremental advances that were either rejected by appellate courts, or the ones that stood the test of time were finally affirmed by the courts of last resort. This process did not take the same paths in the UK and the US, a diversity which is arguably attributed to the different individual and social perceptions of risk prevailing in these two jurisdictions as well as the different judicial attitudes towards them. Yet the end result is the same: principles and concepts of tort laws have been used in varying degrees to interpret the liability system of the Conventions in both jurisdictions. Most importantly, there are instances where the tortious analysis has been achieved in a quiet way which does not attract attention, and as a result it is more easily integrated into the mainstream of judicial interpretation of the Conventions.
5.22 The judgment of the European Court of Justice (ECJ) in R (on the Application of International Air Transport Association and European Low Fares Airline Association) v. Department for Transport21 is instructive here, because it highlights how creative a court can be in order to bypass the exclusivity of the MC.
5.23 One of the questions referred to the ECJ from the English High Court was “[w]hether Article 6 of Regulation No 261/2004 is invalid on grounds that it is inconsistent with the … Montreal Convention …, and in particular Articles 19, 22 and 29 [thereof], and whether this (in conjunction with any other relevant factors) affects the validity of the Regulation as a whole …”.22
5.24 The ECJ was quick to employ policy considerations and resorted to a purposeful interpretation of MC sensing that the “however founded, whether in contract or in tort or otherwise” limb in Article 29 creates a strong presumption that any claims arising from delay shall be filled in accordance with the Convention:
“[i]t is clear from the preamble to the Montreal Convention that the States party thereto recognised ‘the importance of ensuring protection of the interests of consumers in international carriage by air and the need for equitable compensation based on the principle of restitution’. It is therefore in the light of this objective that the scope which the authors of the Convention intended to give to Articles 19, 22 and 29 is to be assessed.”23
5.25 This extract from the judgment sets the tone for the court’s analysis and makes clear that the ECJ was not so much interested in examining the interrelation between MC and the Regulation. Instead, it had a predetermined aim – holding the Regulation 261/200424 valid for the sake of consumers’ protection – and tried to provide the legal underpinning of this conclusion, albeit in an unconvincing way. Interestingly, the ECJ followed the opinion of the Advocate General (AG) L.A. Geelhoed, but not his reasoning in its entirety.25
5.26 The AG argued that the obligations established in Article 6 of the Regulation 261/2004 (provision of meals, refreshments, hotel accommodation, etc) are of a public nature. As such they shall be distinguished from civil liability for delay on the basis that the enforcement mechanism of the Regulation is different than for ordinary claims.26 It is then paradoxical when the AG goes on to say that “the passenger can initiate court proceedings if the carrier did not perform its public-law obligations. Such a claim evidently is aimed at forcing air carriers to comply with their obligations, irrespective of whether a passenger has suffered damages as a result of this non-compliance”.27 Jorn Wegter points out that the AG makes a distinction “between an action for damages as a result of delay under the Convention and measures as contained in Regulation 261/2004 regarding delay …”.28 This distinction, though, is not significant for the MC since the different enforcement mechanism of what the AG characterises as public law obligations “is in part recourse by the passenger to the courts for compensation”.29
5.27 The words of Lord Hope in Sidhu provide support to the proposition that the distinction between private – and public – law obligations, apart from artificial, makes no difference to the Conventions’ exclusive character:
“Chapter III of WC/HP expresses its subject matter in the words ‘Liability of the Carrier’. In contrast to the title to the Convention itself, which uses the expression ‘Certain Rules’, we find here a phrase which is unqualified. My understanding of the purpose of this chapter therefore … is that it is designed to set out all the rules relating to the liability of the carrier which are to be applicable to all international carriage of persons, baggage or cargo by air to which the Convention applies.”30
5.28 The ECJ did not base its analysis on the distinction between public and private law obligations for delay. Instead, it focused on what kind of damages delay can cause to a passenger:
“First, excessive delay will cause damage that is almost identical for every passenger, redress for which may take the form of standardised and immediate assistance or care for everybody concerned, through the provision, for example, of refreshments, meals and accommodation and of the opportunity to make telephone calls. Second, passengers are liable to suffer individual damage, inherent in the reason for travelling, redress for which requires a case-by-case assessment of the extent of the damage caused and can consequently only be the subject of compensation granted subsequently on an individual basis.”31
5.29 The court then concludes that the latter damages are covered by the MC, whereas the former are not. The basic explanation that the court offers for this distinction is that “[t]he system prescribed in Article 6 simply operates at an earlier
stage than the system which results from the Montreal Convention”.32 As such, the ECJ argues that passengers affected by delay can claim compensation for damages under the MC and seek the assistance prescribed in the Regulation, conveniently ignoring both practical realities and legal principles.
5.30 With respect to practical realities, the flaws in the Court’s reasoning are succinctly identified: “[i]n practice, if an airline refuses to accommodate the passenger, the latter has no other alternative than to submit an official complaint and eventually – if the worst comes to the worst – to go to court … it is impossible to make a distinction between enforcement under the Regulation as opposed to enforcement under the Convention”.33 With respect to legal reasoning, the ECJ effectively redrafts Article 29 by importing into it a chronological element that was never intended to be included and contradicts its language. Damages in Article 19 encompass non-monetary awards, such as the ones provided under the Regulation.34
5.31 Yet the most problematic aspect of this judgment is that domestic legislators are given the green light to implement Regulations which run in parallel to the MC and to regulate matters which are in the core of the contractual relationship between the carrier and the passenger. If that was the intention of the drafters of the MC a national law reservation could have been inserted in either Article 19 or 29, which would expressly permit states to implement such regulations for delays. In the absence of such provision, any delay by the carrier in international air transport falls within the scope of the MC and the court would have to look into the contract of international carriage in question and apply the MC to the exclusion of any other legal framework.
5.32 In that respect, the judgment of the ECJ might be justified on the basis that it strengthens consumers’ protection – an aim that the drafters of the MC felt very strongly about. Yet this well-intended aim does not change the conclusion that the judgment seriously questions the exclusivity of the MC and has the potential to open up its interpretation to national laws.
III. The First Judicial Attempts to Define “Accident”
5.33 Ronald Schmid argues that courts in the aftermath of WWII “prefer a broad, instinctive approach. A situation which appears to be an ‘accident’ is treated accordingly and precise definitions are rarely considered”.35 This is not surprising if one adopts the view that the term “accident” is primarily a social construct and then a judicial one. In the period between WWII and the middle 1980s, accidents were transitioning from being unpredictable, exceptional events to being part and parcel of any entity’s operations. Especially during the period before the imposition of the MA, courts lacking any guidance from the Conventions as to the inter pretation of the term inevitably “searched for evidence of negligence”.36 The liability limits that protected the carriers and the unavoidable attempts of claimants to circumvent the WCS in order to secure greater awards gave an impression that the term “accident” does not carry much weight. As such, an authoritative judicial definition was not deemed necessary, but instead courts and litigants favoured an approach that would balance the technical and social aspects of aviation risks on a case-by-case basis.
5.34 In a world where the rationality of technology was creating high expectations for the management of risks and the behaviour of corporations, an open-ended definition of “accident” would have been troublesome for the liability exposures of carriers. Thus, it is not surprising that the first attempts to define the term “accident” in the Conventions’ context were made in the late 1970s. During this time, the legal framework regulating the liability of air carriers was in a state of disarray and the exclusion of the “all necessary measures” defence from the MA put a lot of strain on the term “accident”. Furthermore, the first signs of a risk society made their appearance, and corporate entities, mainly in the US but also in the UK, were facing increased levels of litigation that inevitably raised their liability exposures and contributed to the insurance crisis of the 1980s in the US.
5.35 As a result, the US courts attempted on three different occasions within seven years, from 1977 to 1984, to provide an accurate definition of the term “accident” for the purposes of the WC.37 In all three judgments, the reasoning of the courts was similar, and “accident” was interpreted as an unusual and unexpected event or happening that takes place without foresight.38 This line of reasoning set the basis for the judgment of the SCt in Air France v. Saks,39 which is still considered the starting point of any analysis of the term “accident”. The judgment of the SCt was the result of two conflicting decisions: (i) the first-instance judgment by the US District Court for the Northern District of California which granted motion for summary judgment to Air France on the basis that the change in cabin pressure was normal and did not constitute an “accident” within the meaning of Article 17; and (ii) the judgment of the US Court of Appeals for the Ninth Circuit which reversed the decision of the District Court on the grounds that the MA imposed absolute liability on air carriers; the showing of malfunction or abnormality in the operation of the aircraft was not a prerequisite for the claimants’ recovery as long as the injury was the result of an inherent risk in air travel.40
5.36 What is very interesting about the analysis of the Court of Appeals is the unconcealed attempt to inject tort law principles into the text of the Conventions:
“Absolute liability made sense on several levels. First, it guaranteed to passengers the prospect of quick, less-expensive settlements … Second, it was in harmony with theories of accident cost allocation, since airlines could distribute among all passengers what would otherwise be a crushing loss to victims and their families … Finally, absolute liability could be justified on the theory that while the passenger was in the airline’s control, the airline could best prevent injuries from occurring.”41
5.37 The court’s complete disregard for the text of the Conventions and the intrinsic characteristics of the term “accident” might be an outlandish methodology to an English lawyer, yet they were in harmony with the role of US courts as societal risk managers. The WC was not providing much liberty for achieving policy objectives other than the ones that its drafters intended. Yet the MA, which amended the liability system of the WC to protect the interests of passengers, was a first-rate chance in the eyes of US Courts to reallocate the risks of air carriage between carriers and passengers on the basis that carriers are in most cases better able and equipped to address or insure against risks associated with travel than the passengers are.
5.38 Courts in the US attempted to gain control of the allocation of aviation risks in this way for the sake of deterrence and efficient resource allocation. Especially in cases involving hijackings and terrorist attacks, courts used domestic principles of risk allocation with great ease. In Husserl v. Swiss Air Transport Co Ltd, the Court of Appeals expressly approved the application of modern theories of accident cost allocation to the interpretation of the WC as amended by the MA.42 Most importantly, in Day v. Trans World Airlines Inc, the court, in deciding the issue of whether the claimants had been in the process of embarking and disembarking when terrorists in the airport terminal attacked them made a full-blown analysis of the WC as amended by the MA through the prism of these theories:
“The airlines are in a position to distribute among all passengers what would otherwise be a crushing burden upon those few unfortunate enough to become “accident” victims … Equally important … [t]he airlines, in marked contrast to individual passengers, are in a better posture to persuade, pressure or, if need be, compensate airport managers to adopt more stringent security measures against terrorist attacks … If necessary, the airlines can hire their own security guards … Moreover, they can better assess the probabilities of accidents, and balance the reduction in risk to be gained by any given preventive measure against its cost. Finally, the administrative costs of the absolute liability system embodied in the Warsaw Convention, as modified by the Montreal Agreement, are dramatically lower than available alternatives. If Article 17 were not applicable, the passengers could recover – if at all – only by maintaining a costly suit in a foreign land against the operator of the airport. The expense and inconvenience of such litigation would be compounded by the need to prove fault and the requirements of extensive pretrial investigation, travel, and other factors too difficult to anticipate. Such litigation, moreover, would often unduly postpone payments urgently needed by the seriously injured victim or his surviving dependents.”43
5.39 These judgments tried to introduce a liability system in which the determinant factor was whether the carrier or the passenger is better able to bear or minimise the risk. If the carrier can properly insure against the risk, or is in better position to minimise or eliminate it, the passenger can avoid the difficulties of litigation. In essence, where injury or death occurs to passengers caused principally by something other than an internal condition of the passenger, the carrier is prima facie liable unless he can establish he is not at fault.
5.40 Bearing in mind that the “all necessary measures” defence had been eliminated since the MA, the only question for the courts was to establish a causative link between the “accident” which was equated to event and the injury or death. This analysis, however, is based upon a superficial reading of the MA and a simplistic application of the increasingly popular economic theories of Guido Calabresi in an instrument of international law that long preceded them. Furthermore, “one can hardly deny that the airport authorities were the much better and cheaper cost-avoider” than the carriers.44
5.41 This comment amply demonstrates the artificiality of the judicial attempts to stretch the liability rules in order to satisfy the prevailing social perceptions. It further proves that the MA was an incomplete attempt to give voice to public opinion, which demanded channelling liability through carriers and a simplified recovery process. This is so because the MA did not provide for a straightforward absolute liability legal framework as the aforementioned judgments stated and the US Administration publicised. Instead, the MA, by removing the “all necessary measures” defence, blurred the boundary between accidental and non-accidental for the purposes of the Conventions, and permitted a manipulation of the Conventions’ liability scheme on policy grounds.
IV. The Definition of “Accident” in Air France V. Saks: Offering a Protective Cocoon to Carriers, Opening the Door of the Conventions to Social Perceptions, or Both?
Safeguarding the risk allocation achieved in the Conventions
5.42 The foremost issue to be decided before the SCt in the Saks case was not the definition of the term “accident”, but the way of imposing liability upon carriers. On the one hand, the supporters of a system of so-called absolute liability made a liberal interpretation of the term, disregarding its natural meaning and the relationship between Articles 17 and 18. On the other hand, the advocates of a system of fault liability favoured a narrow interpretation which emphasised its fault connotations, arguing at the same time that a carrier cannot be held liable absent of fault. The SCt found in the middle, putting forward an interpretation of the term “accident” that created a quasi-strict liability regime within Article 17.
5.43 The court looked into the nature of the event and defined “accident” as something caused “by an unexpected or unusual event or happening that is external to the passenger”.45 Thus the court denied recovery to Ms Saks because her injury resulted from her own “internal reaction to the usual, normal and expected operation of the aircraft”.46 According to the SCt, the WC’s drafters understood the word “accident” to mean something different from occurrence in Article 18, and Article 17 to refer to an “accident” that caused the injury and not to an “accident” that is the claimant’s injury. This way, the court put principles of absolute liability to rest, but did not go all the way to absolve carriers from liability in all cases that they were not at fault.
5.44 Paul Dempsey and Michael Milde observe that “Saks … reject[ed] the limitation that the event must be fortuitous or unintentional, wholly dropping these requirements from its reformulated definition of ‘accident”’.47 Therefore, “the fault connotations of Warsaw were read out of the ‘accident’ language of Article 17, and vested instead in the ‘all necessary measures’ defense of Article 20”.48
5.45 Here the SCt injected into the WC elements of strict liability while retaining their fault liability foundations in an attempt to reconcile such different points of view. Having accepted the fault interpretation of the WC, the court would have demonstrated a complete disregard for the changing social perceptions of the time which demanded a liberal interpretation of the WC regardless of the drafters’ intentions and the aim of international uniformity. It is not surprising then that the SCt: (i) ignored the connotations of fortuity and lack of intention that the term “accident” carries; and as a result (ii) endorsed, albeit in dictum, judgments of “lower courts in this country [which] have interpreted Article 17 broadly enough to encompass torts committed by terrorists or fellow passengers”.49
5.46 If one considers the social pressures to expand the liability of organisations at the time, the Saks judgment was a blessing for air carriers. The SCt expressly rejected the use of principles external to the WC to interpret their terms, rejected the view that the MA provides for an absolute liability system, and followed the interpretative methods prescribed by the Vienna Convention on the Law of Treaties. Arguably it took advantage of the Conventions’ drafting flaws and allocated aviation risks between passengers and carriers in a way that was closer to the expectations of the public, but it did not distort in an obvious manner the structure and objectives of the Conventions.
5.47 Arguments that the SCt amended the WC “through active judicial interpretation” are not entirely correct.50 The SCt attempted to interpret a term that did not receive much attention in the drafting of the WC.51 In this endeavour, the court respected the objective notions of risk and the paternalistic spirit that characterised the drafting of the WC while acknowledging – but not necessarily accepting – elements of social risk sharing.
5.48 As such, the SCt attempted to protect the liability system of the Conventions from domestic law influences in the 1980s and formulated a definition that kept subjective notions of risk and volition away from Article 17. At the same time it linked the “accident” to the technical operation of the aircraft. It is true that the elements of fortuity and lack of intention would have provided a more narrow meaning to the term. To a certain extent they would have safeguarded the liability of carriers in the short term. The lack of intention requirement would have made it especially easier for carriers to defend cases where they were not the risk creators or had not actively caused the damage, such as aircraft hijackings or failure to assist cases.
5.49 However, it would have opened the doors for courts to look into the behaviour of air carriers, not in a defensive way as under the “all necessary measures” provision, but in an attempt to strengthen the imposition of liability. This way, the Convention would have been left unguarded to the direct application of tort law concepts. Furthermore, it would have directly linked Article 17 to the “wilful misconduct” provision which requires an analysis of the employees’ intentions and motives.
5.50 The SCt wrongly linked these two provisions in the Husain case almost twenty years later to justify the imposition of liability on the carrier in cases of failure of the cabin crew to assist a passenger in medical emergency: instructing lower courts to look into motivational issues while analysing Article 17 would have created this link earlier and would have put more pressure on the liability limits.
5.51 It is true that the SCt in the Saks case did not justify the omission of the fortuitous and non-intentional connotations of the dictionary meaning of the term “accident”. Scalia J, with whom O’Connor J concurred – two of the judges in the Saks case – while delivering the dissenting judgment in the Husain case gave a clear explanation:
“The word ‘accident’ is used in two distinct senses. One refers to something that is unintentional, not ‘on purpose’ – as in, ‘the hundred typing monkeys’ verbatim reproduction of War and Peace was an accident.’ The other refers to an unusual and unexpected event, intentional or not: One may say he has been involved in a ‘train accident,’ for example, whether or not the derailment was intentionally caused.”52
5.52 The SCt in the Saks case seems to have gone a step further than the French definition of the term “accident”.53 However, it did not exceed the plain meaning of the term and the French text of the Convention. Instead, the SCt moved within the confines of the dictionary definition and made a distinction that corresponds to the plain meaning of the word in English.
5.53 The definition of “accident” promulgated in Saks has received acceptance by Courts worldwide. Especially in the UK, it has been unanimously upheld by the English Court of Appeal in Chaudhari v. British Airways54 and in Barclay,55 and by the HL in Morris and in Re DVT. In Chaudhari, Leggatt LJ endorsed the “careful analysis” of O’Connor J in Saks and held that “in principle, ‘accident’ is not to be construed as including any injuries caused by the passenger’s particular, personal or peculiar reaction to the normal operation of the aircraft”.56 In Morris, Lord Hope held that the interpretation of Article 17 in Saks “gives the word ‘accident’ a natural and sensible meaning in the context of Article 17, and I too would adopt this approach”.57
5.54 In Re DVT, their Lordships were unanimous in holding that an “accident” must be an unexpected event, external to the passenger and occurring during carriage. Lord Scott noted that “Article 17 distinguishes between the bodily injury on the one hand and the ‘accident’ which was the cause of the bodily injury on the other”.58 His Lordship cited with approval Saks and further held that “an event or happening which is no more than the normal operation of the aircraft in normal conditions cannot constitute an Article 17 ‘accident’ and … the event or happening that has caused the damage of which complaint is made must be something external to the passenger”.59 Furthermore, Lord Mance supported the observation of counsel for the defendant that “the word ‘flexibly’ used by the Supreme Court [in Saks] does not … necessarily equate with ‘broadly”’.60
5.55 The agreement of the courts is to demonstrate that the definition in Saks created a balance between the interests of passengers and carriers in accordance with the drafting spirit of the WC and the plain meaning of the term. Most importantly, the SCt in Saks managed to create a protective legal cocoon around carriers which were facing a significant increase in their legal exposures at the beginning of the risk era. It is true that a narrower definition might have been in greater accordance with the risk environment in the 1920s when the original Convention was drafted.
5.56 At the same time, this narrow definition would have meant that social risk perceptions would be repressed. If one takes into consideration the disintegrated state of the WCS in the 1970s and 1980s, and the changing social conditions, it is easy to come to the conclusion that a narrow definition would have “exclude[d] the application of the Convention”, with domestic laws filling the gap.61
5.57 The threat that a narrow definition might have posed to the integrity of the WCS was counterbalanced by the more liberal yet pragmatic definition that the Saks judgment offered. In effect, the SCt saved the WCS from immediate collapse. At the same time, it tried to minimise the effects on aviation of the increasing judgments of blame against multinational corporations which were evident in the US during this period. This way the SCt formed a cocoon around air carriers – a cocoon that protected them from the adverse social conditions at the time, while giving claimants some leeway as to how to bring the definition of the term closer to the prevailing social perceptions of risk.
Inviting the judicial redrafting of the Conventions
5.58 Saks was decided at an era of change: the objective analysis of accidents which required “a causal connection between the accident and the operation of the aircraft”62 and limited recovery to “the inherent risks of air traffic”63 was giving priority to a subjective analysis. This new interpretative approach incorporated broad behavioural elements and expanded recovery to cover any situation that “the accident occur[ed] within the carrier’s sphere of control”.64
5.59 Saks tried to reconcile these opposing views and interpret the WC in a way that would enable objective and subjective interpretations of risk events to co-exist. Inevitably, it received criticism from both claimants and defendants because it did not provide clear-cut evidence of which side it took. Probably the most characteristic example is its endorsement on the one hand of the view that only flying-related risks can be classified as accidents, and on the other hand its acceptance that acts of terrorism are indeed such risks.
5.60 The reference of the SCt to terrorist acts and torts committed by fellow passengers is characteristic of the change in perceptions that was already under way. Although the court was deciding an equipment-related case, it felt the need to bring into the analysis cases that had human behaviour as the centre of attention. This way the SCt did not intend to throw light on the dispute in question, but to give an indication of how air carriers’ liability was going to evolve. Bearing this in mind, it is not surprising that Scalia J linked the flexible application of the term “accident” to deaths and injuries arising from both terrorist attacks and passenger-to-passenger assaults. If the case was decided in the 1960s or 1970s, it is highly unlikely that any reference to man-made aviation risks would have been made.
5.61 In the 1980s, though, the SCt, under pressure from changing social and corporate environments, could not limit its analysis by clarifying that technical risk can be allocated to air carriers only in case of a malfunction. It went a step further and put flesh on the social change by arguing that under conditions to be set by subsequent case law, behavioural and terrorist risks, although not risks inherent in aviation, can give rise to the liability of carriers.
5.62 However, the absence of any guidance by the SCt in setting these conditions proved to be a bone of contention. It fuelled judicial speculation as to the intentions of the SCt judges in Saks and the limits of their preferred definition of “accident”, ignoring sometimes the intention of the Conventions’ drafters: is it an unusual and unexpected event external to the passenger irrespective of its origin and causal connection to flying, or an unusual and unexpected event external to the passenger that also requires a unique association with air travel? In case a special risk inherent in air travel is required, are considerations of intention and behaviour sufficient to create a distinctive link between air travel and the risk-generating event?
5.63 These are hard questions to answer, especially when the Conventions do not give persuasive answers, and the SCt in Saks took notice of them but made a superfluous analysis at best. One might argue that the SCt laid itself open to criticism with such an unnecessary and short reference. However, it must be pointed out that the judgment of the SCt in Saks was a product of its time, i.e. a period when the distinction between unintentional, safety-related risks and intentional, man-made risks was becoming irrelevant as to liability considerations.
5.64 Theories and techniques of preventive and precautionary risk management were gaining impetus, and gradually moved attention from the risk event to the risk creator, and its potential to manage the technical and behavioural risks that it exposed. As aircraft accident investigators were beginning to pay more attention to latent organisational defects rather than active failures, courts, in the build-up to the SCt judgment in Saks, were increasingly looking into the organisation of air carriers rather than the injury- or death-causing event in an attempt to identify the deficiencies that caused or contributed to it.
5.65 However, this investigation was not performed in a systematic way, but instead was based on a sometimes out-of-context application of law and economics principles on the Conventions’ legal regime, especially in the area of terrorism-related risks. In this environment, the SCt judgment in Saks performed a risk mapping and risk management function: risk mapping because it pointed out in a clear albeit short way that the era when technical risks were the sole triggers for carriers’ liability was long gone; risk management because it clarified the way the Conventions deal with technical risks, and essentially triggered the examination of the relation between man-made risks and carriers’ liability by courts, aviation lawyers, and academics in a way that would be compatible with the Conventions.
5.66 It is true that the SCt did not make any decision as to whether air carriers’ liability would follow the general liability trend of proactively managing risks and thus equating technical to man-made risks. Doing so in the early 1980s would have been detrimental to the liability exposures of air carriers, since the social pressures to hold airlines liable for every mishap encountered by passengers were increasingly popular. The series of US judgments in the 1970s finding air carriers liable for the personal injury and death of passengers as a result of terrorist attacks in airport terminals, where they were not necessarily under the direct control of air carriers, is a testimony to this trend. In that respect, it would have been difficult for the SCt not to mention this trend, but it would have been equally difficult to give a persuasive reply to the aforementioned questions without opening the door to tort law considerations.
5.67 As a result, the SCt concluded that the definition in Saks should not be applied rigidly, and new forms of social relations developing in and outside an aircraft cabin should not be excluded from the scope of the term “accident” just because of their novelty. Sensing the evolution of social risk perceptions and at the same time trying to respect the rules of construction of international treaties, the SCt avoided laying down a general principle. Instead it preferred a case-by-case approach that would evaluate “all the circumstances surrounding a passenger’s injuries” and would apply the definition of “accident” flexibly. What was lost by many courts in the aftermath of Saks is that the proposed flexible application did not give them a carte blanche to take into consideration theories of domestic law.65
“… [The Conventions] are intended to provide an enduring uniform code which will govern contractual and, where relevant, delictual relationships not just for a finite time but for the future as the transactions to which they apply are entered into … Notionally to relate them back to a supposed state of affairs existing in 1929 is not only wrong but wholly impractical. It leads to … complication and confusion … It is also destructive of uniformity since when a convention has later been amended, logic would require that one starts the clock again and asks what was in the minds of the delegates at the later conference … The principle is … simple. Words have a meaning which does not change but the application of those words to the decision of any question depends on the facts and circumstances of the case in which that question arises. It is the facts and circumstances of the cases that change, not the meaning of the contractual words.”66
5.69 The approach of Lord Hobhouse is a pragmatic one, and attempts to bridge the gap between the objective risk reality of the Conventions and the evolving social perceptions of risks without disturbing the balance achieved in the Conventions or the rules of interpreting international treaties. In essence, Lord Hobhouse endorses the flexible application of the term “accident” in the Saks judgment and takes it a step further by acknowledging that the aims of the Conventions will be better served if courts pay attention to the evolution of risk and the changeable character of the term “accident”.
5.70 His Lordship is not arguing for the judicial redrafting of the Conventions based upon contemporary societal needs as perceived and defined by judges.67 Instead his Lordship argues for the application of the terms and the spirit of the Conventions in a post-industrial risk environment. This way, the Conventions retain their commitment to the objective interpretation of risk, and at the same time provide space to accommodate modern risk perceptions when they are in accordance with their object and purpose.
V. Morris V. KLM and Wallace V. Korean Airlines: Raising the Social Expectations
Transforming the “risk characteristic to air carriage” requirement from a liability control mechanism to a liability expansion tool
5.71 The practical application of the proposition of Lord Hobhouse has considerable side-effects, since the courts are not always able to discern the fine line between judicial redrafting and adjustment to the current social risk environment. The judgments of the UK HL in the Morris case and of the US Court of Appeals for the Second Circuit in Wallace v. Korean Air Lines68 constitute examples of such confusion.
5.72 At this point a number of significant points should de made. First, the judgments in Morris and in Wallace shall be distinguished on the basis of their ratio decidendi. Both decided that the sexual assault of a passenger by a fellow passenger constitutes an “accident” for the purposes of the WC and the HP despite the absence of a causal role on the part of the air carriers. Yet their respective justification has small, albeit important, differences. In the Morris case, Lord Phillips in the CA rejected unequivocally any link between accident and the operations of an aircraft for the purposes of Article 17: “There is nothing in Saks that justifies the requirement that an ‘accident’ must have some relationship with the operation of the aircraft or carriage by air. Nor do we consider that a purposive approach to interpretation requires that gloss on the word.”69
5.73 For Lord Phillips the liability scheme of the WC/HP is very clear: if an unexpected and unusual event that is external to the passenger and occurs in the cabin causes death or bodily injury to him/her, the carrier is prima facie liable unless he can invoke the “all necessary measures” defence. The text of the Conventions does not reveal any requirement to link the “accident” with the operation of the aircraft, and as such there is no further need to delve into the deliberations that led to the drafting of the Conventions. Faithful to the idiosyncrasies of the English judiciary, he refrains from directly invoking policy considerations and he takes little notice of the flexibility requirement of Saks.
5.74 Instead, his Lordship examined closely the structure of the SCt judgment in Saks, and essentially argues that the general principle laid out by O’Connor J as to the meaning of “accident” does not include any reference to the “characteristic of air travel” requirement.70 Further, he notes that the “characteristic of air travel” requirement in torts by fellow passengers was imposed by subsequent judgments in the US and as such does not stand the scrutiny of the court.71 What his Lordship arguably implies is that any reference to the operation of the aircraft was made by O’Connor J only when deciding upon the specific facts of the case before her.72 Inherent in this subtle distinction is the attempt to contain the effects of any generalised application of the term “accident”, and at the same time to urge courts to pay more attention to the special facts of each case.
5.75 For the UK CA, the SCt in Saks adjudicated a whole different set of facts, and as such the application of the definition of “accident” upon them should be approached with caution. Once again, the effort of the CA to create “islands of liability” within the Conventions rather than lay down general principles and manage social problems through the leeway given by legal rules rather than policy considerations is apparent. After all, the malleability of the term “accident” and the confusion over its exact meaning allows judges a lot of room to allocate the risks in the Conventions without resorting to policy considerations.
5.76 Lord Phillips remains faithful to the text of the WC/HP and the definition in Saks. He finds that the extra layer of protection that the “characteristic of air travel” requirement offers air carriers is artificial, at least in cases of air rage. At the same time, his Lordship provides a somewhat liberal, yet permissible, interpretation of the term “accident”. It is very important to note that his analysis so far is not result-oriented, but instead seems to be a delicate balancing exercise between the text of the WC/HP, the judgment in Saks, and the modern risk realities of carriage by air.
5.77 At this point, one would expect that Lord Phillips would examine whether sexual assault was an unusual and unexpected event external to the passenger so as to prove or not prove the occurrence of an “accident”. However, his Lordship did not go into that analysis, but instead chose to decide the case upon the assumption that the term “accident” incorporates the “characteristic of air travel” requirement.
5.78 It is difficult to ascertain what lies behind this decision, yet arguably the aim of uniformity and the contemporary risk environment must have played a dominant role. Under the guise of uniformity, the UK CA follows the reasoning of the US CA for the Second Circuit in Wallace, effectively complicates the search for an “accident”, and brings into its analysis policy considerations that do not necessarily fit into the text of the WC/HP. What the majority in Wallace argued was that the SCt did not make clear whether the term “accident” in Article 17 contains a “risk characteristic to air travel requirement”, citing Gezzi v. British Airways73 to that effect.74 As such, the US CA seems to take a different stance from the UK CA and admits that a weak link between the unusual and unexpected event that caused the injury and the operation of the aircraft should exist in order to have an “accident” under Article 17. Thus the majority of the court went on to examine whether this link existed in the case before it.
5.79 From the very beginning, the reference of the US CA to the “virtual strict liability” imposed by the MA and the “Saks Court’s admonition to interpret the term ‘accident’ both ‘flexibly’ and ‘broadly”’ brings into play contemporary risk perceptions and enables the court to achieve a result-oriented liability allocation. By invoking the strict liability of the MA, the US CA treats the term “accident” as synonymous to event and avoids examining the unexpected and unusual characteristics of the event that caused the injury in question.
5.80 Furthermore, by invoking the flexible interpretation of “accident” the US CA makes a broad interpretation of the “risk characteristic to air travel” requirement. As a result it injects notions of control and standards of conduct that resemble tort law principles. Malcolm Clarke concurs with the reasoning of the majority in Wallace and dismisses any arguments that the US CA introduces elements of control: the “risk characteristic to air carriage” requirement was satisfied by the fact that “[Ms Wallace] was cramped into a confined space beside two men she did not know, one of whom turned out to be a sexual predator. The lights were turned down and the sexual predator was left unsupervised in the dark”.75
5.81 This assumption, though, raises what the concurring opinion in Wallace described as an even more “‘Talmudic’ question …: how associated with air travel need a hazard be before it can fairly be described as ‘characteristic”’?76
“There should … be a close logical nexus between the injury and air travel per se. A fundamental premise of the argument for expanding carrier liability in this case is that the risk of death or injury in a terrorist attack is appropriately regarded as a characteristic risk of air travel … We do not think that this can be said of the sort of senseless act of violence involved in this case. The risk of violence at the hand of zealots is all too present in any public place whether it be a bank, courthouse, university campus, an Olympic village, or airport. Unlike the risk of hijacking … where the aircraft and the fact of air travel are prerequisites to the crime, we think the risk of a random attack such as that which gave rise to this litigation is not a risk characteristic of travel by aircraft ….”77
5.83 For Chief Judge Coffin the intention of the Conventions’ drafters was not to make air carriers insurers of the passengers’ security. For that reason the connection of “accident” with carriage by air requires a unique feature of air carriage to have facilitated or brought about the “accident”.
5.84 However, both the US and the UK CAs do not share the same view, and have decided that any feature of air transport – unique, special, or common to other means of transportation – can trigger an “accident” under Article 17. This way, the requirement of “a risk characteristic to air carriage” is transformed from a liability control mechanism to a tool that facilitates the expansion of the liability of air carriers. However, it is only fair to argue that the choice of both Appellate Courts to decide the case on the basis of this requirement was result-oriented, especially in the case of the UK CA, which noted the requirement and rejected it.
5.85 Both courts were aware that the rejection of the link most probably would have enabled them to reach the same result. Yet it would not have been a very effective method to contain the efforts of carriers to “exclude accidents that might occur in other walks of life”.78 Instead, expressing doubts about the “risk characteristic to air carriage” requirement, yet at the same time interpreting it broadly to include risks in the cabin, respects in a superficial way the intention of the drafters. It also creates an assumption that a passenger who suffers injury or death is prima facie entitled compensation from the carrier just for being in an aircraft cabin.
5.86 Most importantly, analysing the “risk characteristic to air carriage” requirement in a passenger-friendly way enabled the US CA, and to a lesser extent the UK CA, to equate the unusual and unexpected character of the event in question to the potential that the carrier had to supervise the behaviour of its passengers. It is true that the judgments in Wallace and Morris are not introducing elements of control and supervision as directly as the judgment in Maxwell v. Aer Lingus: “[w]hile passengers are permitted, and in most instances required, to place these items in the overhead bins, this is done under the supervision of the cabin crew who are responsible for securing them before take off”.79
5.87 However, it is a fallacy to argue that the Wallace case illustrates the correct interpretative approach, because (i) the majority opinion failed to examine whether the damage was accidental, in the sense of unusual and unexpected; but instead (ii) made explicit reference to the cabin crew’s failure to supervise a passenger who unbeknownst to them turned out to be a sexual predator. This way the US CA gave the strong impression that the search for an unusual and unexpected event that caused the injury must be influenced by, if not limited to, issues of the carriers’ fault and associated control. As such, a direct link was created between “accidents” under Article 17 and contemporary perceptions of risk. If one takes into consideration that according to Lord Phillips in Morris “the defendant accepts a degree of responsibility for the care of unaccompanied minors in flight”,80 it becomes apparent that both Appellate Courts avoided the Talmudic debate in order to use the “risk characteristic to air carriage” requirement as a Trojan horse that will introduce tortious notions of duty and responsibility without raising criticism.
The practical effect of Wallace and Morris: tort thinking in the liability scheme of the Conventions
5.88 The HL did not decide on the “risk characteristic to air carriage” aspect of the Morris case, as there was no appeal against this finding by the carrier. Lord Hope approved Lord Phillips’ findings in the CA, yet he did not shed light on the fine issues raised above. As a result, the definition of “accident” under Article 17 is characterised by a flexibility not so much of the unusual and unexpected character of the event that caused the injury/death, but by a broad interpretation of the peripheral “risk characteristic to air travel” requirement. This broad interpretation is targeted to introducing tort thinking into the Conventions. And as such aligns the standards of the Conventions with the prevailing ones in society: “‘Warsaw accidents’ include spillage of hot drinks and food poisoning. Even an injury sustained because the passenger did not have enough leg room has been held to be capable of being a Warsaw accident. The last and latest straw is Waxman, in which a hypodermic needle protruding from the seat in front stuck in the claimant’s leg. The airline’s ‘failure to remove the hypodermic needle’ was ‘an unusual, unexpected departure from ordinary proceedings’ and, therefore, the injury was an accident subject to Art.17”.81 Even when courts decide in favour of carriers, tort thinking is omnipresent:
“In Gotz … the carrier was not liable to a passenger who strained his shoulder seeking to place a heavy bag in the overhead locker, because of the sudden rise of another passenger from his seat. The reason was that the ‘event is not within the airline’s purview or control’. The decision was the same in Price, when one passenger assaulted another passenger but the cabin staff had no reason to anticipate the assault.”82
5.89 In that respect, it is not just Wallace and Morris that should give concern to air carriers, but the general judicial tendency to establish a new baseline of socio-legal expectations about risk allocation. This baseline does not refer to the level of damages awarded, but most importantly to what kind of accidents the general public expects to recover from air carriers.
5.91 This judicial choice might result in cases where Courts apply the Saks definition literally and impose liability on air carriers despite their limited causative role.83 Yet, at the time, it was thought to safeguard the uniform application of the WC and most importantly to protect the carriers from judicial decisions that “mirror … prevailing attitudes and give voice to emerging notions of justice as they dawn on society as a whole”.84
5.92 Tony Weigand, writing in 2000, provided concrete proof that overall the definition in Saks has managed to keep the interests of carriers protected: “A large portion of the decisional law to date is less literal in its application, and more in line with the intent behind the Convention.”85 However, the reluctance of the SCt in Saks to decide whether an “accident” for the purposes of the WC must involve “a risk characteristic to air travel” was eventually used as a means to broaden the scope of the term “accident”. Where some court judgments in the aftermath of Saks