Chapter 4
International Air Carrier Liability and Social Perceptions of Risk: Passengers – the Drafting of the Conventions

I. A Brief Overview

4.1 The legal response to changing social perceptions of risk has been fragmented, haphazard, and in many ways inadequate. With respect to aviation, this general trend is discernible in the various treaties, conventions, and instruments that seek to govern the international liability of air carriers, as well as in the case law that these legal instruments have produced.

4.2 Currently, the international air transportation of passengers is governed by two normative frameworks: the Warsaw Convention “System” (WCS), which is an amalgam of International Conventions and Protocols,1 also encompassing “collective” special agreements, regional groupings, and domestic legislation,2 and the Montreal Convention for the Unification of Certain Rules for International Carriage by Air 1999 (MC). Although the MC was drafted in order to modernise and replace the WCS, it essentially consolidates the WCS into a single treaty.3

4.3 This analysis of social risk perceptions begins with the Warsaw Convention 1929 (WC), the first attempt to provide a uniform international liability scheme. As the WC became outmoded, the Hague Protocol (HP) and the Montreal Agreement (MA) were initiated as half-hearted attempts to respond to changing social perceptions of risk and liability. In these efforts, the Guatemala City Protocol (GCP) also stands out. Finally, this analysis will critically examine the MC and its shortcomings. Against this background of inadequacy, which the provisions of the various international legal instruments show to respond effectively to the needs and changes of post-industrial societies, the advantages of a no-fault liability scheme become evident.

II. The Warsaw Convention “System”

Introduction: Warsaw Convention 1929

4.4 The WC is a product of its time. During the romantic period of aviation, a first attempt was made to provide an international, comprehensive, and uniform liability scheme, designed to protect the nascent aviation industry. The allocation of risks achieved in the WC through Article 17 reflects this aim.

The liability scheme of the Warsaw Convention

4.5 The WC was conceived to protect the “fledgling and still fragile aviation industry”,4 as well as to unify private international air law, especially in the areas of air transport documentation,5 jurisdiction,6 and liability.7 These two aims were achieved by:

  • (i) providing a uniform international code which could be applied by the courts of all the high contracting parties. In those areas with which it deals the rules which it lays down were intended to be uniform and exclusive of resort to domestic law;8 and
  • (ii) setting out the limits of liability and the conditions under which claims to establish liability if disputed, are to be made.9

4.6 The WC established a system of liability which changed the common law rules on the liability of air carriers towards passengers. Drafted during the romantic period of aviation, the WC put the emphasis on individualistic risk management:

“One must indeed admit that those who use aircraft are not ignorant of the risks inherent in a mode of transportation which has not yet attained the point of perfection that one hundred years has given to railroads. It is therefore just not to impose absolute liability upon the carrier, but to relieve him of all liability when he has taken reasonable and normal measures to avoid damage.”10

4.7 Although the industry-protective views of the British participant were not shared by all of the delegates at the conference which led to the adoption of the WC’s final text, they set the tone for its provisions. The carriers were the focus of attention, and the main task was to design a control system for personal injuries that would provide the legal basis for carrier operations without stunting the industry’s growth.11 The need to protect passengers, though taken into account, was of peripheral consideration, as the remarks of one conference delegate clearly indicate:

4.8 The French delegate’s off-colour bon mots – which likely would not make it past the stenographer nowadays – are a telling indicator of the attitudes dominating the conference, at least among countries with high stakes in the developing aviation industry.

4.9 In this context, a suggestion was put forward by an Italian jurist to base the liability of air carriers on a system of strict liability, without fault-based defence. This system of liability was considered for a while but finally dropped after pressure from air carriers and their insurers.13 The adoption of strict liability as the overriding principle of the WC would have placed liability on the consequences of the carriers’ conduct, rather than their actual behaviour, and the creation of risks as distinguished from fault would have defined the core and borders of their liability. At the same time, it would have subjected aviation to a rough-and-ready cost-benefit test:

“Strict liability forces the innovator to internalize all the costs of his activity. If the activity still flourishes in spite of a strict liability standard, we can be confident that its benefits exceed its full costs…. At this point the argument in favour of strict liability weakens. Experience already has demonstrated that the activity’s benefits exceed its full costs, and society is now being burdened with the greater administrative costs associated with an increased number of claims brought about by the growth of activity.”14

4.10 In an era when flying was shedding its ultra-hazardous image, and when air carriers were presenting themselves as reliable yet financially fragile entities, strict liability was deemed unfair.15 The prevailing rationale was that aviation, an infant but highly promising industry, did not need to prove its utility, but instead required legal subsidisation similar to a number of other industries during the Industrial Revolution of the nineteenth century.

4.11 By the beginning of the twentieth century, however, social conditions had changed, and aviation could no longer receive a carte blanche in the form of fault liability. What could be achieved instead was a compromise between the industry’s preferred position, which was advocating for conventional forms of liability where the passenger would be required to prove fault as part of the cause of action, and consumer-protection issues.

4.12 As a result, the following liability scheme was devised: air carriers’ liability was based upon presumption of fault16 with a reverse burden of proof on the carrier to establish available defences.17 As a quid pro quo for the aggravated regime of the carriers’ liability, their liability was limited by fixed amounts for death, wounding, or other bodily injury of the passenger,18 unless the passenger could prove that the damage resulted from an act or omission of the carrier, its servants, or agents “done with intent to cause damage or recklessly and with knowledge that damage would probably result”.19 In addition, the scheme precluded carriers from excluding their liability or from fixing lower limits.20

4.13 This way the carriers were:

  • (i) prima facie aware of the conditions under which claims to establish liability were to be made;
  • (ii) able to calculate their exposures due to the limitation limits, safeguarding thus the financial stability of their operations

while the passengers were:

  • (i) relieved of the requirement to prove fault, which would have been extremely difficult due to the novelty and highly technical nature of the evidence required;
  • (ii) protected against provisions relieving the carrier of liability or restricting the amount of its liability.

4.14 The presumption of fault and the reversed burden of proof constituted a significant initiative, departing from the fault principle as embodied in the concept of negligence. Bearing in mind that the doctrine of res ipsa loquitur – which indirectly achieves similar results by permitting the claimant to rely upon an inference of negligence which the defendant must rebut – was not uniformly applied by common law courts in domestic aviation-related cases, the WC made a step forward by easing the task of the claimants in establishing liability.

4.15 At the same time, the shift of the burden of proof can be interpreted as adopting a wait-and-see approach; i.e. wait and see how the industry develops, give it space to breathe and innovate rather than asphyxiate it by regulating austerely an activity whose full potential was still unknown.21 In an era where engineers were incrementally improving the safety of operations, the drafters of the Convention rejected a more radical step to strict liability. The established view is that this reversal favoured solely the claimants. Yet carriers also benefited, in the sense that they could influence the conditions of their responsibility imposed by the Convention, since the exact level of care required of them was still open to determination through a trial-and-error process. One might argue that by shifting the burden of proof, the Convention provided, in practice, a form of strict liability. Although this suggestion contains elements of truth – which will be analysed further – the shifting “does not alter the basis of the liability, which remains that the harm-doer should have behaved differently”.22

4.16 In this sense, the presumption of fault and the paramount importance of establishing whether an “accident” had occurred set the tone for the legal regime of liability set down by the WC. However, these two elements proved to have other, indirect consequences from a social perspective: the term and content of “accident” gradually became open to changes in social perception.

The Warsaw Convention as maker of social risk perceptions

4.17 If it is true in general that the office of law is to find and reinforce norms that exist in communities – and not to light out in search of “just or efficient” norms and then impose them on a community23 – it is a truth that requires qualification in the context of aviation in the early twentieth century.

4.18 Flying was an activity mostly enjoyed by a small part of the population – the rich and famous who could afford it and the thrill-seekers who found reward in the risks. Trying to form legal rules based on aviation practices and perceptions at the time would have resulted in a legal system of very limited potential. Instead, the drafters of the WC attempted to look into the future and craft legal rules that would facilitate the development of norms on reasonable risk allocation and industrial practices worldwide. This does not necessarily mean that the rules achieved a just or efficient regulation of the risks between carriers and passengers. Rather they provided the industry with some leeway to shape the hypothetical model carrier by adopting efficient risk management measures and industry standards.

4.19 This risk-balancing process in infant industries normally comes through “observing the attitudes and practices of laymen and the expectations they generate about legal responsibility as the starting point for adjudicating disputes”.24 How ever, leaving this risk balancing to be performed in a fragmented manner by domestic legislatures and courts would have been a lengthy and cumbersome process for both carriers and passengers.

4.20 For this reason, the drafters of the WC went a step further than simply transforming or applying traditional notions of domestic law in an aviation context. They created a new liability system which viewed aviation risk in as objective a light as possible. The WC tried to contain the impact of what they considered diverse and irrational lay perceptions by creating a unified legal regime that had the contract of carriage as its basis and was influenced predominantly by French contractual legal thinking. This regime would insulate the aviation applications of the quantitative risk analysis against the fears, demands, and risk perceptions of the public as revealed in the tort law litigation process.

4.21 Certainly, one might argue that the scope of international conventions regulating the liability of transportation means already contains elements of artificiality in the sense that they are the result of compromises between the competing interests of carriers and consumers/shippers as well as between different legal cultures. What makes the WC distinct, though, is the dearth of references to domestic legislation and past case law with the exception of the US. Adrift in sparsely charted territory, WC drafters had wide berth to set rules that were not a direct reflection of the prevailing perceptions and practices of aviation. Rather, they were at liberty to “play God” and to shape a liability system that would fit into a predetermined policy objective that at the time was considered of the utmost importance, namely the unification of the legal framework for the protection of the industry and the balancing of interests of an infant industry and of passengers/shippers.25 This balancing of interests constitutes the cornerstone of the WCS/MC and has been a principal consideration in their drafting history.

4.22 The construction of this liability system meant that the drafters attempted – albeit unintentionally – to influence “the very perceptions that … are supposed to determine legal doctrine”.26 In that sense, the regulation of risks achieved in the WC did not result solely in regulating the relation between passengers and carriers, but went a step further and attempted to influence social perceptions of risk acceptability on an international level.

4.23 In an era where aviation was gaining impetus but was still perceived as an extremely hazardous activity, the WC played a trust-building role. On the one hand, carriers could manage the financial impact of the risks associated with flying and as a result projected an image of certainty, an air of confidence. On the other hand, passengers were relieved of the requirement to prove fault and enjoyed immunity from unfair clauses excluding liability. A protective legal cocoon was formed, which facilitated the interaction of the general public with carriers, and at the same time permitted the industry to improve its safety records through a trial-and-error process. Latent in this suggestion is also the intention to grant priority to quantitative systems of risk management, and most importantly to convince the public that the technocratic approach to aviation risk was the way to go despite frequent safety mishaps.

4.24 In a nutshell, then, the WC formed a legal shield for the nascent aviation industry. The allocation of risks achieved therein meant the Convention imposed a liability arrangement which went beyond the established norms of domestic law while retaining a certain similarity to its concepts. By so doing, the WC also imposed a mode of thinking about risk.

The aims of the Warsaw Convention: risk education and risk allocation

4.25 The adoption of the WC’s system of liability, with fault as its basis, hoped to satisfy a number of aims resembling but not mirroring the aims of the system of negligence. Peter Cane argues that the main aim of the law of negligence is “to provide compensation for loss and injury suffered as a result of negligent conduct”;27 deterrence and corrective justice play supplementary roles.

4.26 In the WC liability system, the compensatory function was intentionally downplayed by limiting the amounts recoverable in the event of death or injury to protect the infant industry. At the same time, the deterrent effect of the system was also compromised by the imposition of liability limits. Furthermore, the wilful misconduct provision, with its punitive and subjective characteristics, arguably influenced investments in risk management, albeit in a subsidiary role.

4.27 Most importantly, the WC played a dual risk-educational and risk-allocation role: in an era where the causes of aviation accidents were shrouded in mist, it provided the ground through the use of novel terms such as “accident” and “all necessary measures” to allocate risks between passengers and carriers in an innovative way, and facilitated the formation of standards of behaviour unique to aviation. By rejecting the traditional duty-of-care language mainly used in international conventions regulating the carriage of goods by sea and in domestic aviation cases, the WC broke loose with maritime law and the negligence system. At the same time it highlighted the need to look into the carriers’ organisations at various levels: “accident” provided an opportunity to examine the risks involved in frontline operations and interactions with passengers. The “all necessary measures” defence invited observation of the carriers’ risk management procedures, and to an extent influenced industrial practices. Finally, the “wilful misconduct” provision and its replacement in the HP of “done with intent to cause damage or recklessly and with knowledge that damage would probably result” introduced a subjective element in the risk analysis, which required an examination of the defendants’ employees’ states of mind.

4.28 Courts have found difficulties assessing this state of mind under both versions of Article 25. Much of this confusion in the unamended version of the Convention stems from the fact that French courts expanded the definition of the term dol to cover cases of faute lourde, which was treated as an equivalent to dol for the purposes of the WC.28

4.29 UK and US courts, however, did not extend the term wilful misconduct to cover cases of gross negligence. Instead they accepted a subjective interpretation of the term which requires that the perpetrator “must appreciate that he is acting wrongfully, or is wrongfully omitting to act, and yet persists in so acting or omitting to act regardless of the consequences, or acts or omits to act with reckless indifference as to what the result may be”.29 Similarly, in the US, wilful misconduct can be defined as “a conscious intent to do or omit doing an act from which harm results to another, or an intentional omission of a manifest duty. There must be a realization of the probability of injury from the conduct, and a disregard of the probable consequences of such conduct”.30

4.30 As Georgette Miller points out, French courts provided a broader interpretation of Article 25 where objective recklessness was sufficient to break the liability limits; that is, the perpetrator was not necessarily required to know of the risk of damage.31

4.31 This lack of uniformity led to the redrafting of Article 25 in the HP to bring it closer to the interpretation of the courts in the UK and the US.32 Furthermore, the French Act of 2 March 1957, that was enacted to unify the domestic and international legal regime for air transportation, introduced the concept of faute inexcusable and defined it to mirror Article 25 of WC/HP: the deliberate fault which implies knowledge of the probability of damage and its reckless acceptance without valid reason.

4.32 One would expect that the convergence of texts would put an end to the conflicting interpretations. Yet this was not the case, because the question arose whether the knowledge of the perpetrator that damage would probably result shall be subjective or objective: in general terms, the objective approach requires a court to examine whether in the circumstances of the case the air carrier ought to have realised that its conduct would result in damage. On the other, the subjective approach requires a court to examine whether there was actual knowledge on the part of the air carrier that its conduct would result in damage.

4.33 English case law did not divert from the interpretation of the unamended version and demonstrated a clear preference for the subjective test. Eveleigh LJ in Goldman v. Thai Airways International Ltd demonstrated in simple terms how the subjective test differs from the objective one and made clear that Article 25 WC/HP calls for the former to be applied:

“An act may be reckless when it involves a risk, even though it cannot be said that the danger envisaged is a probable consequence. It is enough that it is a possible consequence, although of course there comes a point where the risk is so remote that it would not be considered reckless to take it. We look for an element of rashness which is perhaps more clearly indicated in the French text ‘temerairement’. Article 25 however refers not to possibility, but to the probability of resulting damage. Thus something more than a possibility is required. The word ‘probable’ is a common enough word. I understand it to mean that something is likely to happen. I think that is what is meant in Article 25. In other words, one anticipates damage from the act or omission.”33

4.34 His opinion was followed by Auld LJ in Nugent and Killick v. Michael Goss Aviation Ltd and Others: “[t]he key … is the state of mind of the pilot at the time of the material conduct. In my judgment, the additional ingredient is actual knowledge, in the sense of appreciation or awareness at the time of the conduct in question, that it will probably result in the type of damage caused. Nothing less will do”.34

4.35 Georgette Miller notes that the issue of how to interpret the term knowledge in the amended version was never resolved during the travaux préparatoires .35 Bin Cheng disagrees and, following extensive reference to the travaux préparatoires , he concludes that “[t]he discussions at The Hague in 1955 made it clear that it was not merely knowledge of the probability of damage that was required, but actual, as opposed to simply imputed, knowledge”.36

4.36 Cheng’s reasoning has been persuasive, and the majority of courts worldwide apply a subjective test to Article 25 WC/HP. Molloy J in Connaught Laboratories Limited v. British Airways provided a detailed and clear description of the current state of the law:

“[T]he overwhelming body of international jurisprudence favours the application of a subjective test. The subjective test has been applied, for example, in England (Goldman v. Thai …); Australia (S.S. Pharmaceutical Co Ltd v. Qantas Airways Ltd …); Belgium (Tondriau v. Air India); Switzerland (Lacroix Baartmans, Callens, und, Van Tichelen v. Swissair …) and the United States of America (Saba v. Compagnie Internationale Air France … and Piamba Cortes v. American Airlines Inc)…. It would appear that the one exception to date has been France, the courts of which have consistently applied an objective test notwithstanding the jurisprudence from other countries. This has been the subject of comment and criticism amongst academics…. Other courts have also made reference to the divergent case law from France, but it would appear that at this point, no other country is following the French approach. In my opinion, the relevant authorities overwhelmingly lead to the conclusion that a subjective test must be applied in determining the applicability of Article 25. Regardless of whether the result of such a test may seem unfair or illogical in any given case, it is clear from the working papers and debates at the time that this was the intent of the drafters. Further, this approach has been adopted by virtually all courts of nations that are parties to the Convention.”37

4.37 Georgette Miller acknowledges the different approach of French courts and argues that the 1964 judgment of the Cour de Cassation in Rioult c. Mutualle d’assurances aeriennes marked the beginning of this divergence.38 Bin Cheng is very critical of the French approach:

4.38 Whatever the reasons are behind this difference of opinion over the appropriate test in Article 25, the fact remains that the subjective test has received worldwide acceptance. This way courts could scrutinise the carriers’ behaviour on an international level, and could get acquainted with the peculiarities of air transport and the human-technical risks involved.

4.39 One might argue that the traditional system of negligence, which required carriers to exercise due diligence to make the aircraft airworthy and to properly and carefully navigate, would have reached similar results in terms of risk education and allocation. This is arguably true as witnessed by the development of domestic US carriage by air which was subject to common law principles. Yet it would have meant that the formation of behavioural standards would have been slower and fragmented, and risk information would not have been easily disseminated between jurisdictions and used accordingly. Furthermore, traditional notions of negligence would not have been successful in shaping perceptions of aviation risk worldwide as effectively as the WC. Similarly to the limitation limits which attempted “to ‘unify’ the cost of living in the international community by pretending that the lowest denominator was a ‘common denominator’ under all circumstances”,40 the WC attempted to unify the way people perceive aviation litigation, and to an extent direct laypersons’ understanding of carriers’ technological systems and blameworthiness.

4.40 The dissemination of risk information was mainly achieved by providing “a self-contained code within the limits of which the plaintiff must found his jurisdiction”.41 The four fora provided were carefully selected to channel litigation to aviation-advanced jurisdictions, which were more likely to have adequate risk education.

4.41 Thus settlement negotiations and litigation would be conducted not necessarily upon principles of fairness or efficiency, but upon the risk-allocation principles established in the WC. It was expected that lawyers, courts, and consumers would gradually get acquainted with the norms of the WC. Eventually an informal global system of aviation litigation would be established, which would provide common litigation procedures and similar treatment to litigants irrespective of their itinerary, the place of the accident, or the venue of litigation. It is hardly surprising that any reference to the courts of the place of the accident was omitted, since “such a place may be quite fortuitous and unpredictable and also absolutely unrelated to the centre of gravity of the case”.42

4.42 The unification of liability perceptions and the allocation of risks – which was obviously key to the success of a convention unifying rules of liability – was predominantly achieved through the use of the term “accident” in Article 17 as the basis of the carriers’ liability. Article 17, which creates an exclusive cause of action in respect of claims for passenger injury or death in the course of international carriage by air,43 provides:

“The carrier is liable for damage sustained in the event of the death or wounding of a passenger or any other bodily injury suffered by a passenger, if the accident which caused the damage so sustained took place on board the aircraft or in the course of any of the operations of embarking or disembarking.”

4.43 Article 17 is in exactly the same form in all versions of the WCS (the Guatemala City Protocol of 1971 replaced “accident” with “event”; however, it has never come into force) and thus it has shaped aviation liability through the twentieth century. Even the corresponding article in the MC, Article 17(1), is worded almost identically:

“The carrier is liable for damage sustained in case of death or bodily injury of a passenger upon condition only that the accident which caused the death or injury took place on board the aircraft or in the course of any of the operations of embarking or disembarking.”

4.44 Therefore, a passenger seeking to hold a carrier liable for injury or death has the burden of establishing that: (i) there has been an “accident”; (ii) that the “accident” took place on board the aircraft or in the course of any of the operations of embarking or disembarking; and (iii) that the death, wounding, or bodily injury was caused by the “accident”.

4.45 These three prerequisites for receiving compensation are of equal importance, yet the requirement to prove “accident” stands out, as no definition or qualification of the term is included in the WC, despite it being the first condition for carrier liability under the Convention.44 Tony Weigand states that there is no plausible explanation for the absence of any discussion on the meaning of the term and concludes that it “simply appeared in final form as revised by the drafting committee at the Convention”.45

4.46 Article 21 of the draft Convention submitted by the CITEJA to the Warsaw International Conference reads as follows:

le transporteur est responsable du dommage survenu pendant le transport:

  • (a) en cas de mort, de blessure ou de toute autre lésion corporelle subie par un voyageur;
  • (b) en cas de destruction, perte ou avarie de marchandises ou de bagages;
  • (c) en cas de retard subi par un voyageur, des merchandises ou des bagages.”46

which in English reads as follows:

“The carrier shall be liable for damage sustained during carriage:

4.47 As one can clearly see from these draft articles, the term “l’accident” or its equivalent term in English, “the accident”, cannot be found in either version. Furthermore, the minutes of the Warsaw Conference, French and English, make no mention of discussion about the meaning of the term. It first appears during the sixth session of the Conference, where the Rapporteur mentions that “the accident can occur during the carriage and the death happen afterwards” without elaborating any further.48 Then, during the final reading in the seventh session of the Conference, Article 17 appears in its current format, yet the president of the drafting committee does not explain the reasons for using the term “accident” in it.49

4.48 Although the last-minute use of the term “accident” in the WC appears to be one of the great mysteries of international air law, it is less of a mystery when one considers the era’s growing faith in the risk management capacity of technology. Judith Green provides an interesting perspective on the reasons for undermining the term “accident” during the drafting of the WC:

“[I]n the first half of the twentieth century … the acceptance of the limits of rational explanation [was a defining feature]: That is, that there can be ‘accidents’, events whose causation is coincidental and thus could not be predicted…. That accidents happened was taken for granted, uninteresting and hardly the subject of scholarly enquiry.”50

4.49 This finding suggests that, for the drafters of the Convention, the word “accident” did not require any elaboration, since its meaning was straightforward and denoted events that occur outside the boundaries of rationality. Similarly, in the UK the term accident was used in the Workmen’s Compensation Acts of 1897, 1906, and 1925.51 The Acts did not provide any guidance as to its meaning, yet the courts drew on a widely held and equally broad definition of “accident”. Lord Lindley in Fenton v. J. Thorley & Co Ltd described the nature and the meaning of the term succinctly:

“The word ‘accident’ is not a technical legal term with a clearly defined meaning. Speaking generally, but with reference to legal liabilities, an accident means any unintended and unexpected occurrence which produces hurt or loss. But it is often used to denote any unintended and unexpected loss or hurt apart from its cause; and if the cause is not known the loss or hurt itself would certainly be called an accident. The word ‘accident’ is also often used to denote both the cause and the effect, no attempt being made to discriminate between them.”52

4.50 If these propositions are valid, the determination of whether there is an “accident” for the purposes of the WC did not necessarily entail considerations of fault and reasonable behaviour on the carriers’ part. In an era when many causes of injuries/deaths remained inexplicable, linking “accident” with fault would require passengers to delve into a prolonged and detailed consideration of all surrounding circumstances. Furthermore, a qualitative assessment of the carriers’ behaviour would be needed in order to judge it against prevailing industry standards and contemporary knowledge before liability could be established. This conclusion, though, would run against the reversal of the burden of proof and would make the “all necessary measures” defence obsolete. In the drafter’s mind, the use of the term “accident” as the basis of recovery would involve “an inquiry into the nature of the event which caused the injury rather than the care taken by the airline to avert the injury/death”.53

4.51 The prominence given to the term “accident” throughout Article 17 would provide the necessary link for the evolution of the concept of “accident” and fault, as these terms are prone to changes in social understanding. Indeed, these concepts were used as notions of blame control, which eventually eroded the very usefulness of the WC, as national courts began to interpret “accident” so as to reflect changes in social perceptions.

“Accident” and “all necessary measures” as blame control

4.52 In this sense the concept of “accident” dissociates accountability from culpability, and arguably creates a system which resembles strict liability within Article 17.54 Carriers are presumed liable regardless of the presence or absence of fault. The claimants would not be required to “inquire into whether the defendant’s behaviour was lacking in some respect. No assessment [would need to be] made of the mental state of the defendant, no comparison [to be] drawn between the defendant’s conduct and the usual conduct of others …”.55 The first instance judge in Fulop v. Malev Hungarian Airlines expressly demonstrated the fault-irrelevant character of the term “accident” when he argued that “an unusual and unexpected happening arising in the course of air travel need not rest on any notion of negligence or fault to be actionable, as long as the element of abnormality relates in some discernible … way to the inappropriate or unintended happening arising in connection with or during the course of the operation of the aircraft or airline”.56

4.53 It is apparent that the use of the term “accident” as a determinant of liability in the WC and the quasi-strict liability regime of Article 17 eased the position of claimants. At the same time, these legal provisions benefited carriers in that they excluded recovery for the usual and expected, which is used in Article 18 as the basis of liability for loss of, or damage to, baggage or cargo. This dual role of the term “accident” and the liability scheme of Article 17 were clearly described by Lord Scott:

“The omission from these conditions of any requirement of negligence on the part of the carrier is double-edged so far as an injured passenger is concerned. It is to the passenger’s advantage that negligence on the part of the carrier needs to be neither alleged nor proved. It is to the passenger’s disadvantage, however, that even clear causative negligence on the part of the carrier will not entitle the passenger to a remedy if the Article 17 conditions cannot be satisfied.”57

4.54 This demonstrates that the liability system of the WC was the unintended result of the drafters’ effort “to allocate the risk of the occurrence of certain events between various parties rather than to decide whether one person has caused injury to another by fault or whether one person should compensate another for a wrong”.58

4.55 Peter Cane argues further that the risk allocation objective was firstly used in industrial injuries schemes in order to relieve injured workers from bearing the whole burden of accidents.59 In the aviation industry, the allocation of risks and the quasi-strict liability regime of Article 17 came from the opposite side: the need to protect the future development of the industry in an increasingly paternalistic social environment. In that sense, notions of due care and causality as understood and applied in tort law not only would create a disparity of liability regimes around the world, they would also tilt the balance in the passengers’ favour once aviation romanticism faded away, since they are more sensitive to changes in social perceptions. To postpone or better guide the inevitable “erosion” of the carriers’ position and their technocratic approaches to aviation risk through notions of fairness and/or efficiency, the WC’s drafters created a risk management code that could be used as the carriers’ safety net:

air carriers:

“bore the risk of death, wounding, or bodily injury sustained by a passenger as a result of an accident which takes place whilst he is on board the aircraft or in the course of any of the operations of embarking or disembarking;”

while passengers:

4.56 As such, the allocation of risks achieved in the Convention was not necessarily intended to benefit either carriers or passengers, nor did it always provide a passenger with a remedy:

“It is tempting to give way to the argument that where there is a wrong there must be a remedy. That indeed is the foundation upon which much of our own common law has been built up…. [The Convention] was not designed to provide remedies against the carrier to enable all losses to be compensated. It was designed instead to define those situations in which compensation was to be available. So it set out the limits of liability and the conditions under which claims to establish that liability, if disputed, were to be made.”61

4.57 This further suggests that the way risks were allocated in 1929 was not based upon modern notions of enterprise liability or corrective justice, but rather on the need to create a balance between the conflicting interests. In that respect, no proof of negligence or fault was required when determining the liability of the carrier under Article 17, yet “accident” was not automatically proved as of right, a suggestion that would have brought about notions of control.

4.58 Article 17 cannot be viewed as a victory of carriers over passengers or vice versa, but rather as establishing a predictable rule, one which could resist changes in social perceptions by providing a predetermined allocation of risks:

4.59 The words of Lord Hobhouse confirm that risk allocation as understood by the drafters in 1929 should not be associated with modern notions of negligence, enterprise liability, or consumer protection. Rather, it purported to regulate the relation between carriers and passengers in what the drafters perceived to be a balanced way – a balance that tipped the scales somewhat in favour of carriers, and supported objective notions of risk because they were in need of assistance at the time. However, this balance, which was meant to regulate the aviation industry for the years to come, was based on a term – “accident” – that was equally sensitive to changes in social perceptions as the tortious ones it was replacing.

4.60 In that respect, the “all necessary measures” defence, which introduced the element of fault into the Convention, and the limitation of liability were used as instruments to control how “accident” was perceived. By providing the carriers with the opportunity to escape liability by proving that they took all precautions that were reasonably available to them,63 the WC opened the door to notions of negligence and enabled carriers to formulate industrial practices and to establish levels of socially acceptable risk.

4.61 Article 20 was playing a far greater role than merely exonerating carriers from liability. It gave them the opportunity to demonstrate how much control they had over their operations and thus guide notions of social blameworthiness. In an era where consumer interests were not presenting a unified front, and the media were more interested in the record-breaking achievements of the industry, carriers could control attributions of liability through Article 20. At the same time they could provide a set of expectations for how they could – as distinguished from should – behave; expectations that were going hand in hand with aviation technology.

4.62 In this way, the introduction of fault in Article 20 had the potential to indirectly influence the selection process of what constitutes “accident” by shaping the boundaries of rationality and drawing the line between accidental and non-accidental. This is not to suggest that fault considerations were intended to be actively involved when analysing the term “accident” for the purposes of the WC. It rather means that the term “accident” and the “all necessary measures” defence were, unbeknownst to the Convention’s drafters, interrelated in a rather ironic way: the more risk management measures the carriers would take and the safer their operations would get, the easier it would be for the claimants to prove the existence of an “accident”.

4.63 The increasing supremacy of quantitative risk analysis and the detachment from notions of fate and luck by the middle of the twentieth century transformed the notion of accident. No longer random and authorless misfortunes, accidents were the result of predictable risks, a shift that correlated with changing perceptions of blameworthiness.64 In that respect, the term “accident” which was used by the drafters of the WC to protect carriers from subjective ideas of risk and fault gradually boomeranged on them: accidents were gradually perceived to arise from knowable and calculable risks and were thus transformed into potentially preventable events.65 This way, the carriers’ behaviour and risk management policies influenced the determination of “accident”.

4.64 As such, notions of risk and control that pertained to negligence gained ground when determining whether there was an “accident” for the purposes of the WC. At the same time, the paternalistic shift in the social environment, together with the transition to the corporate era of aviation, established a pro-claimant atmosphere. In turn, the socio-legal environment induced legislatures and courts to inject more subjective elements into their reasoning and to view aviation accidents as a social problem. As prevailing attitudes and industrial behaviour changed, a reallocation of risks between carriers and passengers was required in order to adjust the WC’s liability system to the new social conditions.

4.65 This reallocation, though, did not occur overnight by means of a brand new Convention, but took the form of sporadic, ad hoc compromises between the “clashing interests” of carriers and passengers.66 During this lengthy legislative rebalancing exercise, courts, especially in the US and to a lesser extent in the UK, regularly broke out of the narrow boundaries of the WC. Acting as societal risk managers, they attempted to reallocate risks between carriers and passengers in a manner not intended by the WC drafters. In the process of resolving controversies over the interpretation and application of the WC, courts often were able to perform a creative regulatory function in moving the law as stated in the WC towards negligence standards, which were more flexible and responsive to social perceptions. This way, they were quick to fill the gaps that were intentionally or unintentionally left by the drafters, and also quick to disturb the balance of interests achieved in the WC on the basis of policy considerations.

The amendments of the Warsaw Convention


4.66 As changes in risk perception applied pressure for change in the legal regime of international liability for air carriers, the once-uniform legal structure of the WC was being amended in a piecemeal fashion, so that the aviation industry, and the countries which supported aviation, would be able to assuage the growing dissatisfaction of the public. These attempts at providing a new bargain regarding the allocation of risk were ultimately unsuccessful, as they did not alter the basic features of the WC – notably Article 17 – but rather focused on expanding liability limits.

From the Warsaw Convention to the Hague Protocol

The protective cocoon is ruptured: is it the result of low liability limits?

4.67 After WWII, with the invention of bigger and safer aircraft, a new era of aviation was under way. It was a time of great technological progress in the industry, indeed, but one of pessimism among the public. Dissatisfaction with air carriers was increasing, and the legal regulations of the WC were part of the problem. This is clearly reflected in the words of Sune Wetter, the Chairman of the Legal Committee of IATA, who noted in 1948 that:

“Customers who have good reason to be discontented are apt to become short-term customers, and would-be customers are likely to hold off when they hear of those experiences. The public cannot be expected to understand that airlines are not to be blamed or held responsible for strange Warsaw Convention rules.”67

4.68 Aviation was, in effect, going mainstream, and the cocoon provided by the WC was wearing thin. Air carriers were promising to improve the lives of customers – “plan now to escape from winter to these sunlit isles” was the motto used in an advertisement of Pan American World Airways in 194768 – and customers expected air carriers to make good on them. No longer an unfortunate part of the learning process, a broken promise was a dent in the trust that carriers were vigorously trying to build with the middle classes. In this effort, the WC turned from a “futuristic” legal instrument into an impediment to the build-up of trust between carriers and passengers, as social norms over the role of aviation and the blameworthiness of carriers superseded its liability rules. Sune Wetter, again, argues that “[c]hanges in the scope of and the extent of air transportation may justify changes in rules of liability”.69

4.69 But technical innovation per se is just the beginning of the story, and not necessarily the only guide to legal change. Air transportation in the postwar era was a totally new animal because it was attracting a different kind of passenger, one who held different ideas about individual and social risk acceptability. In that respect, the WC required amendment not just because jet aircraft could carry more passengers faster, longer, and arguably safer; but also and especially because expectations had changed, affecting the balance between carriers and passengers. So while the WC was drafted with a view to the future, the development of the industry and its social utility outpaced its liability scheme.

4.70 Rene Mankiewicz notes in 1955 that “the present weakness of the original Convention lies primarily in the relatively low limits of liability in the cases of death or bodily injury to passengers”.70 He is right in saying that the compensatory element of the original Convention was not sufficient to satisfy the ever-improving living standards of consumers in the Western world. But he fails to appreciate that the need to revise the limits was one of the symptoms of changing social perceptions: the increasing number of states ratifying the Convention constitutes proof that it was a handy tool for unifying liability rules. It was also a very useful blueprint for domestic legislation, especially in countries which did not have the expertise to develop their own rules. It is not a testimony, though, that the Convention still reflected aviation’s social norms.

The cocoon is ruptured: the risk society takes its toll

4.71 Post-war discussions of carriers’ liability were focusing on one problem, but they were ignoring the greater need to update the provisions of the Convention to accommodate paternalism and the diminishing faith in technology. Surely the expansion or outright elimination of liability limits was a step in the right direction. It was merely a patchwork measure, however, which resulted in a number of amendments in a short time. These reflected the social urgency to reallocate the risks between passengers and carriers in a way that favours the former.

4.72 In essence, the Hague Protocol was a missed chance to draft a convention that would replace the original WC and guide the aviation industry through its adolescence, just as the original WC guided the industry in its infancy. One might argue that the rapid technological development of aviation in the years following WWII, coupled with the social changes in technological risk perceptions that took place during the second half of the twentieth century, justified this multiplicity of legal instruments. But it does not fully explain why, within a period of twenty years (from 1955 to 1976), four conventions/protocols were enacted amending/supplementing the WC. More importantly, two unilateral agreements – the Montreal Agreement (MA) and the initiative of the Malta Group – were devised that used certain loopholes in the Conventions to completely change the liability scheme as applied to the US and certain countries in the EU.

4.73 Owing to the social and technological transformation of the risk environment during this period, a major overhaul of the original WC was required, an overhaul based on broad social considerations and not just economic ones, an overhaul which would result in “a single modern scheme which could be reasonably readily comprehended by the travelling public”.71

4.74 The reluctance to take a brave step towards modernisation in the 1950s, however, resulted in what Bin Cheng described in a relatively ironic way as “18 separate Warsaw regimes on international carriage by air [that are] capable of being affected by one or more of at least six schemes established separately by governments, regional economic integration organisations or the airlines …”.72

4.75 Richard Gardiner goes a step further and explains the practical consequences of this reluctance, describing the proliferation of instruments as self-defeating: “It works against the generally acknowledged aim of uniformity and must baffle even the most intelligent member of the public”.73 In other words, the legal uncertainty caused by the increasing complexity and quantity of rules undermined the trust-building role of the original WC. This added to the inadequacy of its provisions to keep up with the social norms. In this way, the amendments to the WC pierced the legal cocoon they were supposed to preserve, leaving carriers in a state of ambiguity regarding the management of their legal exposures, and leaving passengers with a feeling of discontent.

4.76 While it has been convincingly argued that the WCS’s successes in the areas of documentation, “uniformity of law” and simplified jurisdiction have been overshadowed by its limitations on liability – the WCS’s “most notorious aspect”, in Gardiner’s view74 – it is not clear that from a passenger’s perspective there was much to celebrate at all. For passengers, the liability limitation was as frustrating as the jurisdictional limitations, especially when the passenger could not bring a claim before the courts of their residence; as were the limitations of Article 17, which establishes the need to prove “accident” and “bodily injury”.

4.77 Although the various amendments were, in theory at least, intended to ameliorate the position of passengers, their fragmented implementation and the incremental rather than radical changes they introduced only amplified blame: in the aftermath of accidents claimants are not particularly interested in the uniformity of law or the fact that they can travel on e-tickets, but are mostly concerned with the procedure that needs to be followed in order to get the maximum compensation possible. Sometimes they are even puzzled why they cannot use the more familiar tort principles and civil procedure, even though the WC liability scheme arguably simplifies the compensation procedure.


Only gold members can continue reading. Log In or Register to continue