In these proceedings, the Court is being asked to resolve a number of issues of construction which have arisen under insurance policies in connection with claims for business interruption losses resulting from the Hatfield rail disaster which occurred on 17^th October 2000.

The Claimants in both actions are train-operating companies (“TOCs”) owned by National Express Group plc. They each operate passenger rail services on various parts of the UK mainland railway network (“the network”). The network was owned and operated by Railtrack plc (“Railtrack”).

The Defendants are six insurance companies which insured the Claimants against, amongst other risks, loss of revenue and increased costs of working caused by interruptions to and/or interferences with the Claimants’ business, as follows: -

With the exception of the Third Defendant in Folio No. 1062, the Defendants are all represented by Messrs. Herbert Smith. They are referred to as the Herbert Smith Defendants. The Third Defendant in Folio No. 1062 is referred to as Eagle Star.

On 17^th October 2000 at about 12.23 pm, a Kings Cross to Leeds passenger train derailed on a curved section of the Down Fast line on the East Coast Main Line (“ECML”) between Welham Green and Hatfield, approximately half a mile south of Hatfield junction. Four passengers were killed and a number of others were seriously injured. The train, track and other infrastructure in the immediate vicinity of the derailment suffered extensive damage.

The derailment was caused by a broken rail, which in turn was caused by gauge corner cracking (“GCC”), a type of rolling contact fatigue (“RCF”) which can occur in the head section of rails. The existence (albeit not the extent) of the cracking was known to Railtrack and re-railing was scheduled to start only six weeks later.

A section of the ECML between Alexandra Palace and Stevenage was closed immediately following the derailment. Trains scheduled to run on that section of line were diverted via an alternative route, known as the Hertford Loop, and operated on a reduced service. The section of the line remained closed until mid-November 2000, while the damaged train and accident debris were removed, the accident was investigated and the damaged track and other infrastructure were repaired.

In the immediate aftermath of the derailment, on 17^th and 18^th October 2000 Railtrack imposed emergency speed restrictions (“ESR”s), or reduced the maximum permitted line speed, on all sites across the network where GCC was known to exist and which had previously been identified for renewal. On 19^th and/or 23^rd October Railtrack issued further instructions requiring inspections of all known GCC sites and specifying a number of mandatory actions to be taken where cracking or spalling was found. The particular actions, which depended generally upon the degree of cracking found, included speed restrictions, ultrasonic examinations, rail grinding and rail renewal.

In the weeks and months following Hatfield, Railtrack issued a number of further instructions relating to RCF. Pursuant to these instructions and the instructions referred to in the previous paragraph, a large number of speed restrictions were imposed on numerous sections of track across the network and a considerable amount of re-railing or other remedial work was carried out (which in some instances required temporary closures of sections of track, or “possessions”, while the work was carried out). The scale of the operation was such that it was later described as “a watershed for the industry” and “the biggest programme of work on the network for more than a century…”.

These measures (the collective content of which are referred to for convenience as “the Railtrack Programme”) caused disruption to the Claimants’ services. The resultant losses of revenue and increased working costs that some of the Claimants have suffered are the subject of claims under the Policies.

The terms of the CGNU Policy are contained in: (a) a slip stamped and initialled on behalf of each of the Defendants on various dates between 29^th October 1997 and 4^th November 1997; (b) a Commercial Union policy wording dated 19^th May 1997; (c) an endorsement applying to the period from 1^st November 1999 to 31^st October 2000 dated 16^th February 2000; and (d) a further endorsement dated 13^th May 2000.

The slip provides as follows: -

The Policy Wording provided as follows: -

inherent vice, latent defect, gradual deterioration, wear and tear, frost, change in water table level, its own faulty or defective design or materials

DENIAL OF ACCESS IN RESPECT OF TRAIN OPERATING COMPANIES

the action of the Insured or other competent authority for reasons of public safety other than disease hygiene or sanitation

extreme climactic conditions

The endorsement referred to in paragraph 12(c) above sets out the sums insured and the limits applicable to the third year of cover (1^st November 1999 to 31^st October 2000) and provides for a single sum insured (£50 million) for all the TOCs. The endorsement referred to at paragraph 12 (d) above varied the indemnity period under the business interruption section of the Policy from 24 months to 36 months.

The terms of the St Paul Policy are contained in the following documents: (a) various renewal schedules, endorsements and updated sections of the standard policy wording which the Defendant stamped and initialled on 21^st August 2000; and (b) the Defendant’s previous standard policy wording, except for those parts which were replaced by the updated sections referred to in (a).

The Renewal Schedule provides as follows: -

The Claimants, together with various other companies, are named as subsidiary companies of Prism Rail plc in general endorsement number 1.

The Business Interruption Schedule provides as follows: -

The standard policy wording provides, amongst other terms, as follows: -

The updated policy wordings referred to in paragraph 16(a) above provide, amongst other terms, as follows: -

Prevention of Access

the action of a competent authority for reasons of public safety other than disease hygiene or sanitation

extreme climactic conditions

landslip heave or subsidence of the site immediately surrounding adjoining or adjacent to Property used by the Insured

loss or destruction of or damage to property not being used by the Insured but in the immediate vicinity thereof

Finally, General Endorsement 3 provides as follows: -

Against this background, the following issues arise in the CGNU action: -

Whether, on its true construction, the occurrences (or deemed Incidents) referred to in the Denial of Access clause were,

any of the matters set out in sub-clauses (a) to (d) thereof, which cause the Insured to be prevented from or hindered in the use of or access to any station depot or track or other part of the rail network owned or operated by British Rail Board Railtrack Group plc or other CAHA parties, or

the Insured being prevented from or hindered in the use of or access to any station depot or track or other part of the rail network owned or operated by British Rail Board Railtrack Group plc or other CAHA parties, and caused by any of the matters set out in sub-clauses (a) to (d) thereof?

Whether, on whichever is the correct construction of the Denial of Access clause,

Railtrack's decision to implement the Railtrack Programme or the communication of that decision to the Zones or the implementation of the Programme constituted an occurrence (or deemed Incident) which occurred in the policy period; or

each speed restriction or line closure imposed by Railtrack constituted a separate occurrence (or deemed Incident) only some of which occurred during the policy period; or

each instruction issued by Railtrack following the Hatfield derailment constituted a separate occurrence (or deemed Incident), only some of which occurred during the policy period?

Whether the Claimants' business interruption claims are excluded by Exclusion 1.1, which involves the following sub issues:

whether Exclusion 1.1 is capable of applying at all to claims under sub-paragraph 1(a) of the Denial of Access Extension; if so

whether RCF of rails is a form of “gradual deterioration” or “wear tear”; and

whether “gradual deterioration” or “wear and tear” in the form of RCF of rails was a proximate cause of the Claimants’ losses?

Whether, if the Court concludes that each speed restriction or line closure imposed by Railtrack or instruction issued by Railtrack constituted a separate occurrence,

there is a separate deductible for each such occurrence; or

each individual TOC is entitled to aggregate its insured losses on the grounds that such losses arise out of one event, namely:

Most of the issues arising in the St Paul action are almost identical to the issues arising in the CGNU action save for the following additional issues: -

It was not suggested by any of the parties that the differences in wording between the two policies gave rise to any disparity in construction (save of course for the effect of Special Conditions 2(ii) and (iii) which only applied to the St. Paul policy). Accordingly I propose to try and avoid confusion by concentrating on the CGNU policy. The conclusions that I reach on the proper meaning and effect of that policy apply pari passu to the St.Paul Policy.

The threshold issue relates to the construction of the Denial of Access extension. It was Eagle Star’s case, later adopted by the Claimants, that the “occurrences” referred to in the first paragraph were the events specified in the lettered sub-paragraphs to the second paragraph. The Herbert Smith Defendants submitted that the “occurrences” in the first paragraph were the events specified in the entirety of the second paragraph.

It was common ground that the relevant initiating activity for the purpose of the present case was that specified in sub-paragraph (a), namely the action of a competent authority (Railtrack) for reasons of public safety. Thus the scope of this dispute on construction was as follows: -

If the Claimants and Eagle Star were correct, the relevant insured peril provided by the extension was the action or actions of Railtrack in that regard occurring before the expiry of cover on the 31^st October.

If the Herbert Smith Defendants were correct, the relevant insured peril provided by the extension was the prevention or hindrance in the use of track caused by Railtrack’s action or actions prior to the expiry of cover.

The impact of the resolution of this dispute was somewhat narrowed by the supplementary argument deployed by the Claimants and Eagle Star that there was a single action by Railtrack in the form of the Railtrack Programme or, alternatively, a single prevention or hindrance which incorporated all the ESRs issued under what was termed the Railtrack Programme, whereby in either case all losses sustained whether before or after the expiry of cover were recoverable.

That said, the determination of the threshold issue of construction has an important influence on this secondary issue as to whether there was a single insured peril or multiple perils. The fact remains that, as is explained in greater detail later in this judgment, it may be easier to identify the Railtrack Programme as a composite action of a competent authority than as both such a composite action giving rise in turn to a composite prevention of access.

Considerable oral and written argument has been devoted to the topic of the nature of the occurrence or deemed Incident. Nonetheless I believe I can take the matter very shortly. Reflection has not dispelled my initial impression that the Herbert Smith Defendants’ analysis is to be preferred: -

The structure of the extension clearly suggests that the “following occurrences” referred to in the first paragraph are what in fact follows in the second paragraph.

This is confirmed by the phrase “and caused by” which requires to my mind that the prevention, appropriately caused, is to be the occurrence.

The contrary argument gives rise to the surprising outcome that the assured will be enabled to claim under the policy in respect of an action occurring within the policy period even though by the time cover terminated, no prevention or hindrance had yet occurred: this outcome is particularly difficult to reconcile with a policy that otherwise only responds in the event of some occurrence which by definition has an impact on the insured’s business.

The Claimants placed considerable emphasis on the fact that “following occurrences” was expressed in the plural. I derive no assistance from that, not least because in the same breath the clause provides “such occurrences being deemed an Incident”. In short, on either view of the clause, four different classes of occurrence are provided for.

I conclude that the cover has been extended inter alia to encompass the peril of losses in consequence of an insured being hindered in the use of track which hindrance was caused by the action of a competent authority.

On the basis of their construction, which I have found to be sound, the Herbert Smith Defendants go on to contend that, in consequence, individual ESRs which came into effect prior to midnight on the 31^st October 2000 fall within the period of cover provided by the policies and (subject to exclusion, deductibles, aggregation etc.) losses attributable to each of them are insured under the policy. On the other hand ESRs coming into effect after that time fall outside the period of cover and consequent losses are recoverable (if at all) under any successor policy.

The Claimants and Eagle Star submit that, albeit there were numerous ESRs both before and after the 31^st October, they constitute only one prevention or hindrance of access. In short, it is submitted that there was only one occurrence. That occurrence is formulated in somewhat different ways: -

The Claimants say that the occurrence took the form of Railtrack’s decision to implement the Railtrack Programme or, alternatively, the communication of that decision to the Railtrack zones, or, in the further alternative, the implementation of the programme, whereafter the imposition of all the ESRs for incidents of GCC up to May 2001 and beyond was inevitable as the “die was cast”.

Eagle Star say that the prevention and hindrance of each Claimant’s use of the tracks was a single occurrence in the form of a “continuing state of affairs” caused by the adoption and the implementation of the Railtrack programme during the policy period.

A vast quantity of documentary material was placed before the court dealing with the inspection and maintenance regime pursued by Railtrack both prior to and after Hatfield, supplemented by experts’ reports put in by all three parties. To the extent that this material helped put into context the nature, scope and extent of the “Railtrack Programme” which followed upon the derailment, its presentation was legitimate and helpful. But, as I reiterate later, to the extent that it formed the basis of a submission by both the Claimants and Eagle Star that the pre-Hatfield procedures were “woefully inadequate”, it was both illegitimate and unhelpful. I should make it plain that in outlining the circumstances leading up to the derailment and the immediate response to it, nothing I set out should be perceived as constituting an express or even implicit criticism of the procedures adopted by Railtrack.

Before recounting some of the background, it is necessary to explain, albeit briefly, the phenomenon of rolling contact fatigue (“RCF”) and one of its specific manifestations, gauge corner cracking (“GCC”). This is rendered easy by virtue of the joint memorandum of the experts on this topic: -

RCF is a form of fatigue that occurs as a result of rolling and sliding of one component over another. The relevant example is a rolling-stock wheel on a railway rail, giving rise to cracks that begin at, or slightly beneath, the surface of the rail. It is caused by the high stresses close to the point of contact between the two components. Many wheel passes are required for the development of RCF.

GCC is a form of RCF in which cracks appear in the “gauge corner” of the rail. Head checking (“HC”) is a term variously used to refer to cracks on the top of the railhead as opposed to the gauge corner, or to cracks anywhere on the railhead (so that GCC is a subset of HC).

RCF is a complex phenomenon brought about by a number of factors including:

Track usage (number and the speed of trains).

Weight and type of rolling stock.

Acceleration and braking.

Type of steel used in the rail.

Imperfections in the wheel and/or the rail.

Wheel and rail geometry.

Maintenance procedures.

Presence of contaminants between the wheel and rail.

Railtrack’s inspection and maintenance requirements were inherited from British Rail in 1994. The standards at that time were set out in BR’s Track Maintenance Handbook published in September 1993. This handbook required routine visual and ultrasonic track inspections to be carried out at intervals dependant on the tonnage and speed of the traffic concerned. For the type of track thus identified as requiring the highest frequency of inspection, visual inspection was required twice per week and ultrasonic inspection once every six months. The ultrasonic testing involved used what was termed the U3 test which involved positioning the probe above the centre section of the railhead only.

An appendix to the handbook specified the minimum actions called for on discovery of a rail defect. The only action required for GCC was to report its extent and location with a view to its inclusion in the rail-grinding programme. The new Railway Group Standard published by Railtrack in September 1995 made no change to these requirements.

On the 2^nd of December, 1995 Tony Hind, Railtrack’s Permanent Way Maintenance Engineer, wrote to all zone engineers as follows: -

The letter went on to say that, pending the committee’s evaluation of the recommendations, where any “tache ovale type” defect was discovered during periodic ultrasonic testing with the U3 test, at a location where GCC was also present, a further ultrasonic test was to be carried out using the U14 procedure. (The U14 procedure uses the same probe as the U3 procedure but is positioned approximately 10mm closer to the gauge corner of the rail). The letter went on to state that where any defect was detected using the U14 procedure, a temporary speed restriction of 20 mph was to be imposed and the rail was to be changed within 36 hours (i.e. what was termed a 1A defect).

This reference to “tache ovale” type defects requires some explanation. Tache ovale defects are fatigue defects that develop internally. Strictly speaking the term should only be used to describe the kidney shaped cracks which develop from a hydrogen induced shatter crack created in the manufacturing process, but the term has often been used loosely in the UK railway industry to describe RCF defects in general. It is clear from the draft testing procedure that it was being used in the latter sense.

The use of the U14 procedure at GCC sites was incorporated into issue 2 of the Railtrack Line Standard on Track Inspection Requirements in April 1997 with the effect that the minimum action required on discovery of GCC, whether visually or ultrasonically, was amended so as to require not merely a report for inclusion in the rail grinding programme but also the application of U14 in the next testing cycle.

By means of a letter dated 15^th December 1997 from Dr Allen, Railtrack’s Permanent Way Development Engineer, an experimental change to the procedures for dealing with RCF on the West Coast Main Line (WCML) was introduced. The experimental procedure required sites with visible head checking to be inspected routinely using the U14 procedure every 3 months in addition to the routine U3 inspections which were to continue at the normal frequency (i.e. every 6 months). A range of mandatory actions was prescribed for defects discovered using the U14 procedure depending on the level of flaw indication produced by the U14 tests and on the extent to which the U3 procedure also resulted in a flaw indication. In particular it was only if a significant flaw indication was obtained using both U3 and U14 that 1A action was to be implemented. In April 1999 issue 3 of the Line Standard was promulgated. The effect of a new appendix D.4 was to adopt the experimental procedures introduced on the WCML for the whole network.

On the 5^th May 1999, a meeting took place between the Zonal Track Engineers. Their discussion focused on two relatively recent examples of RCF related rail failures. One example had been a broken rail at Hemel Hempstead in a switch diamond attributable to a tache ovale type defect emanating from gauge corner cracking. The formal investigation into the incident revealed that the track had been subjected to the appropriate weekly visual and 3 monthly U3 ultrasonic inspections thereby satisfying the requirements of the line standard. Since no GCC was visible at the site, no U14 inspections had been carried out.

The enquiry concluded that the cracks in the gauge corner area of the rail had not been identifiable visually and were only visible after treatment by magnetic particle inspection (MPI). The fractures had in fact occurred at locations where there was severe lipping to the bottom of the gauge face contact area. The board concluded that there was “no developed technique for examining rails with lipping to the base of the contact gauge face area to identify the presence of cracks”. The outcome was to embark upon examination by U14 and U3 testing of similar S & C layouts. It also became mandatory to remove all lipping in S & C layouts to be followed up by MPI.

On the 12^th December 1999 the head of Track Asset Management at Railtrack, David Ventry, wrote to all zone track engineers on the issue of management of head checking : -

A note was attached to the letter setting out what was Railtrack’s current best practice for management of RCF. This reflected the existing standards as supplemented by the letter itself but also recognised that shallow cracks underneath the surface of the railhead could deflect a signal from ultrasonic probes resulting in the loss of “bottom echo”. The attachment prescribed that the procedure identified in appendix D4 should be followed in these circumstances with modifications. In short, in the event of loss of bottom echo, such was to be treated as a defect and, if the defect was isolated, the rail was to be clamped and changed within 7 days (and, if it was not possible to clamp or if the defect was not isolated, then a 20 mph ESR was to be imposed and the rail changed within 36 hours).

At a meeting of the Zonal Track Engineers on the 7th March 2000 it was decided to postpone discussion of the management of head checking until the next meeting. Concerns were being expressed that the problem was getting out of control. Urgent comments from the zones were requested so that the matter could be considered in some detail.

The next meeting duly took place on the 19^th April 2000. At that meeting a paper by Simon Hardy, Railtrack’s Principal Asset Manager, on the management of head checking was handed out and discussed. The paper focused particularly on the prospect of management of head checking by grinding. David Ventry’s view as expressed in the minutes was that Simon Hardy’s work “was leading towards rewriting the guidelines and eventually, if necessary, changing the standards – at present the emphasis is on collecting and understanding of all the data.”

At the track engineers’ meeting on the 5^th July 2000, Simon Hardy circulated a further paper, this time relating to broken rails. Some 917 broken rails had been reported for 1999/2000 (an increase of about 20%) but there were doubts as to the root cause of the failures. The meeting recognised that the problem of head checking was a national one and it was decided that the whole question of ultrasonic testing needed to be looked at again.

A further paper was circulated by Simon Hardy at the meeting on the 26^th September 2000 (three weeks before the derailment) entitled “Inspections and Actions for Head Checking/Rolling Contact Fatigue”. The paper proposed the following: -

The derailment occurred at 12.23 pm on the 17^th October, 2000. It is the Claimants’ and Eagle Star’s submission that the Railtrack Programme began that afternoon or evening when the National Control Centre of Railtrack issued a letter of instruction marked for the urgent attention of all zone track engineers and entitled “Identification of gauge corner cracking sites with high risk potential”. The instruction provided (albeit amended slightly the following day) as follows: -

On the 19^th October 2000 Railtrack issued a memorandum to zone directors and track engineers entitled: -

Severe i.e. sites with spalling which precluded the ability to test the rail ultrasonically and sites with visible cracks 30mm or longer. In both cases a 20 mph ESR was required and the rail was to be changed as soon as reasonably practicable.

Heavy i.e. sites with visible cracking 20 – 29mm long a 20 mph ESR was required until the rail had been examined ultrasonically. If the ultrasonic testing detected a defect, the ESR was to remain in place and re-railing to take place within 3 months; if no defect was found by ultrasonic testing, periodic visual and ultrasonic inspections were required at the frequencies specified.

Moderate i.e. visible cracks 10 – 19 mm and light i.e. less than 10 mm long - no speed restrictions required but visual and ultrasonic inspections were to be carried out at specified frequencies. If any defects were detected by ultrasonic inspection they were to be treated as prescribed in an appendix.

On the 23^rd October 2000, the scope of the remedial activity was greatly enlarged by a further memorandum from David Ventry to all zone track engineers which called for all known GCC sites on the network, regardless of whether ESRs had already been imposed at those sites, to be subjected to the criteria applied by the 19^th October instructions.

On the 24^th October 2000, David Ventry issued a further letter of instruction, this time relaxing the requirements for sites with heavy RCF. It read as follows: -

The impact of these instructions as at Wednesday 24^th October is revealed by a briefing note prepared by Railtrack for the information of the train companies. At that time it was currently thought that there were some 1,800 sites where gauge corner cracking was suspected to be a problem. Of those, some 300 were severe sites which, following the grouping together of some sites, had produced 181 ESRs. It was anticipated at that stage that further inspections would lead, by the end of the month, to a further 150 – 200 ESRs.

Railtrack issued two further instructions on the 26^th and 27^th October putting back the date by which ultrasonic testing of the heavy sites was required to be completed to the 31^st October and 5^th November 2000 respectively. In addition a risk assessment procedure was introduced on the 27^th October to be performed by a multi-disciplinary team at sites where ESR’s had been imposed with a view to making recommendations to a central review panel for the prescribed speed of the ESRs to be increased.

However, it remained the case that, on all sites where the railhead had spalled, a 20 mph speed limit was mandatory. A further memorandum that same day emphasised that those locations where there was a loss of the rail bottom signal from the probe should be classified as severe due to the potential for the defects being hidden beneath the surface. This latter matter was varied by yet a further memorandum issued on the 30^th October 2000 which provided that the risk assessment procedure could nonetheless be applied to sites with spalling where the rail was ultrasonically testable and no defects had been found on testing.

The period of cover under both policies terminated on the 31^st October 2000. November saw the first references in any document to some form of “programme” or “plan”. For instance, on the 2^nd November, the Chief Executive of Railtrack, sharing a platform with John Prescott, then Transport Secretary, announced progress on the “National Track Recovery Plan”. (Mr Prescott had announced on the 27^th October that such a plan would be “up and running within two weeks”.) On the 3rd November, Railtrack duly wrote to Central Trains under the heading of “GCC Recovery Programme”.

On the 4^th November, Permanent Way Special Instruction (PWSI) No.1 was issued by Railtrack on the management of head checking. It had the express status of a Line Standard and it superseded all previous instructions, in particular those issued on or after the derailment. The instruction dealt separately with three different scenarios: -

Type A. Plain line and full section rails in switches and crossings where head checking either affects more than 200mm rail or was within 4.5 metres of any other defect. The steps to be taken varied according to the severity of the observed head checking and were substantially the same as those in the instructions dated 19^th October 2000 subject to some variations in periodic inspection frequencies and the actions required following ultrasonic inspection.

Type B. Plain line and full section rails in switches and crossings where head checking neither affects more than 200mm of rail nor is within 4.5 metres of any other defect. For moderate and light defects, the steps to be taken were substantially as for Type A. For either severe or heavy head checking, the actions were now the same in either case and required re-railing as soon as reasonably practicable. However in the meantime an immediate 20 mph ESR was only required if the defect either could not be clamped or was in any part of switch, stock or crossing rail.

Type C. Planed rails in either switches, and crossings or expansion switches. Broadly speaking the actions to be taken here were stricter than for Types A and B and in the case of severe and heavy checking required the imposition of an immediate 20 mph ESR and re-railing as soon as reasonably practicable.

The first amendment to PWSI 1 was made on the 7^th November. On the 9^th November, Railtrack made a presentation to both the Prime Minister and the deputy Prime Minister entitled National Recovery Programme. The press release following this meeting stated:

Thereafter on the 17^th November and 28^th November 2000, PWSIs 2 and 3 were issued which afforded degrees of relaxation of the procedures that had been put in place since the derailment.

In the meantime, a presentation had been made by Railtrack to the TOC’s and a further presentation to the Prime Minister of what had now been renamed the National Recovery Plan. The response of the Rail Regulator in a letter dated the 29^th November to Steven Marshall, the new chief executive of Railtrack was highly critical:-

On the 8^th December, Railtrack produced a document entitled “National Recovery Plan”. Its executive summary describes the “initial reaction” by Railtrack to the discovery that GCC was a significant factor in the Hatfield derailment as being to impose speed restrictions at sites thought to be at risk. Subsequent inspections revealed the extent of the problem albeit that ongoing research had allowed relaxation of some of the restrictions. “The National Recovery Programme (NRP) involves eliminating the risk associated with existing GCC sites and tackling the basic causes of the phenomenon.” The programme contemplated a return to normal working by Easter 2001.

On the 2 January 2001, the Rail Regulator, following a submission by the TOC’s, made an Order under Section 55 of the Railways Act 1993 requiring Railtrack to provide “a robust and fully risk-assessed national recovery plan before the 18^th January 2001” on the grounds that it was otherwise in contravention of its licence. An updated plan was duly produced by that date followed by further revisions in February and March. It was eventually wound up and replaced by an Infrastructure Improvement Programme in May by which time something approaching a normal timetable had been restored.

In April 2001, PWSI 4 was issued which replaced PWSIs 1 to 3. This update reflected, as its introduction stated, Railtrack’s developing knowledge of RCF. This in turn was replaced in December 2001 by PSWI 4 Issue 2 with its particular emphasis on grinding.

There is a wealth of evidence to the effect that the implementation of the instructions caused a massive degree of disruption across the whole rail network as a consequence of the imposition of ESRs and the work necessary for them to be removed. The problem is pictorially represented for the Scottish zone in a graph put in evidence showing speed restrictions applied in great numbers in the immediate aftermath of the accident but continuing to this day.

The scale of the problem can be comprehended by reference to the following table: -

These figures compare with the usual number of ESRs within any 4 week period of between 150 and 200. It is probably unnecessary to spend time on explaining the impact of the large number of ESRs being imposed over a 7 month period. But I heard evidence on this topic from a number of directors and other executives of the Claimant companies. The thrust of their evidence was as follows: -

Timetables, which are normally prepared 15 months in advance, incorporate an allowance for 1 or 2 ESRs that are likely to be present at any given time.

The effect of, say, half a dozen speed restrictions requiring deceleration and acceleration inevitably leads to significant delay, further exacerbated if the restriction itself covers a substantial length of line.

Late arrival can lead to delay or cancellation of succeeding services, with the added factor of trains thereby being put out of position and train crews out of hours.

The disruption to one train operating company (TOC) will affect another if trains use or cross the same routes.

A good example of the sort of evidence to this effect is contained in the statement of Mr Ian Kay the Performance Manager of Central Trains: -

The scale of the work needed to remove these speed restrictions (which of course needed to remain in force until the remedial work was completed) is reflected in the following table relating to re-railing as at the end of January 2001: -

In the event by May 2001 the re-railing requirements had risen to 500 miles with some 1150 S & C units requiring replacement.

It is also clear from the evidence that, where re-railing was required, ESRs remained in place for longer periods than might otherwise be expected on account of problems encountered by Railtrack in obtaining replacement rail, providing engineering trains and making competent staff available. Against that background a potential for anything near a normal network timetable capability was not restored, as I have already recorded, until about May 2001. Even then most TOCs found it impossible to meet timetables in the face of the continuing imposition of ESRs.

The issue here is whether there was one occurrence or event in the form of the overall impact of the implementation of the Railtrack programme as the Claimants and Eagle Star contend (and thus all ESR losses occasioned by virtue of the implementation of the programme over the entire 36 month extended period of indemnity are covered) or whether the imposition of each individual ESR was a separate occurrence or event (and thus only the losses resulting from ESRs imposed before midnight on the 31^st October 2000 are covered).

It has to be borne in mind that the question here is not one of the construction of an aggregation clause nor strictly speaking does it involve the determination of whether the losses were all in consequence of the same event. It is simply whether there was one hindrance to access consequent upon action of a competent authority (the Railtrack Programme and its implementation) or numerous hindrances consequent upon action of a competent authority (each and every ESR). Whether, if this latter construction is correct, those losses are to be aggregated is a different matter.

It follows that direct help may not be forthcoming from previous authority in determining the nature of what is in effect a unique factual situation. That said the present context makes it plain that the two words “occurrence” and “event” in so far as they define the scope of the peril insured against are to be regarded as synonymous: see also Kuwait Airways Corporation v Kuwait Insurance Co.[1996] 1 Lloyd’s Rep 664 at p.686, Murray Arnold Scott v The Copenhagen Reinsurance Co. (UK) Ltd, [2002] Lloyd’s Rep IR 775 at p. 784 and [2003] EWCA Civ 688 at para 54 & 55.

The conventional starting point of what constitutes an “occurrence” or “event” is the decision of Mr Michael Kerr QC (later Kerr LJ) sitting as an arbitrator in the Dawson’s Field Arbitration in 1972 (albeit his award did not enter the public arena until 1998). The arbitration arose out of the hijacking of four aircraft, three of which were flown to Dawson’s Field airstrip in Jordan and one of which was flown to Cairo. Mr Kerr QC held that the destruction of the aircraft at Dawson’s field, which were all blown up within a period of about 5 minutes, was one occurrence, but the loss of the other aircraft at Cairo was a separate occurrence. He put the matter in this way: -

This approach is not dissimilar to the observations of Lord Mustill in Axa Reinsurancev Field[1996] 1WLR 1026 at p.1025: -

In Kuwait Airways Corporation v Kuwait Insurance Co.[1996] 1 Lloyd’s Rep, 664, Rix J having cited Caudle v Sharp [1995]LRLR 433 and the Dawson’s Field Award went on: -

The issue of unity arose again in Mann v Lexington [2001] LR IR 179. A retrocession contained a limit of $5m “per occurrence” and applied a deductible “each location any one occurrence”, against the background of reinsurance which spoke of “each and every loss, each and every location”. The losses were at 67 supermarket stores each in different locations, which had been damaged over the course of two days by rioters in Indonesia. The retrocessionaire’s case, that the total riot damage constituted one occurrence, was based on the allegation that all of the rioting was deliberately orchestrated by the government then in power.

The Court of Appeal held however that, in its contractual context, an “occurrence” had to occur at a particular location and therefore all of the losses could not be aggregated. But Waller LJ went on to consider what would follow on the assumption that the word “occurrence” was not limited to an occurrence at a particular location; -

The most recent foray into the arena is to be found in the decision in the Court of Appeal in Scott v CopenhagenInsurance Co. (UK) Ltd. [2003] EWCA Civ.688 which was concerned with the loss of a BA aircraft at Kuwait International Airport at the beginning of the 1990 war and, in particular, whether its loss arose from the same event that had led to the loss of the KAC aircraft following the invasion and capture of the airport. Rix LJ gave the lead judgment. He disagreed with the submission made by the appellant that the test of the “unities” was inappropriate in a case where one event led to both immediate total losses and to other total losses more distant in time and went on to say: -

Against this background, the criteria for testing the existence or otherwise of one event were not controversial and can be summarised as follows: -

There must be a sufficient degree of unity to justify the label of an event.

The assessment of unity will be by reference to time, locality, cause and motive.

The matter is to be scrutinised from the perspective of an informed observer in the position of the assured.

The assessment is to be made both analytically and as a matter of intuition and common sense.

The Claimants’ submission was, first, that the measures taken by Railtrack in the wake of the Hatfield derailment were all part of a single and systematic programme, albeit adapted over time as the company’s knowledge and understanding of RCF developed. Put another way, it was said that “there was a sufficiently common thread running through all the instructions to justify them being regarded as a single programme”. Second, it was submitted, that, even if the relevant occurrences or events (as indeed I have found it to be) were the TOCs being prevented from or hindered in their use of or access to the network by a sequence of ESRs, they in turn could be properly regarded as one occurrence since the “die was cast” from the moment that Railtrack issued the first post Hatfield instruction on the 17^th October 2000. From that moment, the argument ran, the Claimants were effectively hindered in their use of or access to the network by a “logical, systematic and inevitable progression of restrictions of access” and that, accordingly, “the overwhelming impression to be gained from the facts” was that there was only one occurrence or event, namely Railtrack’s decision to implement the Railtrack Programme or the communication of that decision to the Zones or the ensuing implementation of the programme.

One of the features of the case is the difficulty in identifying precisely what the Claimants mean by the term “programme”. The pleaded case is as follows: -

This definition immediately reveals some of the tensions in the Claimants’ argument: -

The programme is said to encompass instructions issued (and action prompted thereby) over a period of 6 months, yet the implementation is said in the same breath to have taken place prior to the expiry of cover some two weeks after the derailment.

The programme is said to include the imposition of both the emergency speed restrictions and the re-railing, yet the Claimants in their closing submission were at pains to distinguish between the imposition of ESRs (which were said to have caused the disruption and consequent losses) and Railtrack’s attempts to remove the ESRs and restore access; the former it was said was the subject of the Railtrack Programme whilst the latter was the subject of the National Recovery Plan.

The Claimants’ case is further undermined by the rather different case as to the nature and content of the programme advanced by Eagle Star. It was their primary submission that the programme was made up of the three sets of instructions issued on the 17, 19 and 23^rd October. These were, it is true, issued during the currency of the policy but, on the other hand, they were almost immediately superseded by PWSI 1 on the 4^th November. Furthermore, it was to be noted that Eagle Star’s case was that the Railtrack Programme, far from being something separate from the National Recovery Programme, simply underwent a change of title.

It is not difficult to see why the Claimants at least were anxious to divorce the National Recovery Programme from the Railtrack Programme. The recovery programme was heavily criticised by the TOCs and the rail regulator as not constituting any form of robust and risk-assessed plan at all. The plan, having emerged in embryonic form in December, was updated and revised over the following months and never received any final approval.

I accept the Herbert Smith Defendants’ submission that the term “the Railtrack Programme” is in effect an invention for the purpose of these proceedings. There is no reference to such a programme in any of the contemporary documents prior to the expiry of the policies. Although the Network Management Statement in 2001 refers to “the biggest programme of work for more than a century”, this is a convenient label to encompass the overall response to the derailment.

The incremental development of inspection and repair is spelt out in the statement itself : -

Of course, the immediate impact of the derailment was that Railtrack came to appreciate that GCC was a problem that needed to be addressed as a matter of urgency. But first impressions do not begin to lead to the conclusion that there was a time (or a place) where Railtrack decided to issue and implement a Railtrack Programme. To the contrary the impression is of a series of instructions and measures undertaken with a view to understanding the scale of the problem and then to remedy it.

The Claimants relied on the scale and cost of the measures taken to deal with RCF after the derailment. This was said to reflect a fundamental change to Railtrack’s approach to the problem, in stark contrast with what would have happened, it was contended, if the derailment had not taken place. This was said to reinforce the conclusion that there had indeed been a Railtrack Programme.

But it has to be pointed out that at the time the initial instructions were issued in the wake of the derailment, it was anticipated that no more than 80 to 100 sites would prove to be in need of remedial work. Only later events proved that an enormous number of sites required urgent action, a process spread out over the following 7 months.

As to what would have happened if the Hatfield derailment had not taken place, there was considerable debate. I confess I was never persuaded that the exercise had any value. It simply restated the issue. It is certainly common ground that at the time of the Hatfield derailment there was a large amount of GCC affecting the entire rail network. A report dated the 25^th October 2000 (a week after the Hatfield accident) by Transportation Technology Centre Inc made an assessment of Railtrack’s previous methods and plans for managing broken and defective rails. In their report they indicated the proposed programme of action for dealing with the increase in rail breaks as from September 2000. These included: -

Briefing note on head checking: November 2000.

Good practice guide: December 2000.

Increase in rail grinding capacity: January 2001.

Reintroduction of Railtrack’s ultrasonic

To the extent that it is relevant, I think it is probable that the new criteria for dealing with RCF suggested in the run up to Hatfield would have been adopted within 6 or 12 months. This would in turn have led to the unearthing of many of the GCC sites albeit they would have been dealt with in a manner which would not have caused disruption on anything like the scale which in fact occurred.

I have not forgotten the evidence of the directors and executives of some of the TOCs. They spoke variously of “a rolling programme across the network”, “a single management response”, “a single unified programme” and so on. But I derive no assistance from this material since I am not persuaded that this evidence reflected views contemporary to the events. Further, and in any event, they were clearly trespassing on the decision that was for the court.

In summary, I am unable to accept that there was a “Railtrack Programme” in anything other than anecdotal sense of there being a range of measures brought into play in reaction and response to the derailment for the purpose of ascertaining the extent of the GCC “infection” and the best means of eliminating it.

A decision or a plan cannot constitute an event or occurrence. It is the promulgation and application of the programme that might. Since the insured peril is only brought about in the event of a prevention or hindrance, it is the ESRs that has to constitute the occurrence or event. It is in this context that the Claimants assert that the “die was cast” i.e. that there was a single but continuing prevented. But each ESR related to a specific incidence of GCC, prevented access to a specific length of track and prevented its unrestricted use for a specific period. No individual ESR was the natural or necessary consequence of any other. If, contrary to my view there was a Railtrack Programme in any singular and coherent sense, its existence simply led to the situation that a large number of ESRs were likely to be imposed across the network for many months.

Eagle Star’s approach was to assert that the imposition of the ESR’s was a “continuing state of affairs” caused by the adoption and the implementation of the Railtrack programme. But even on the assumption that there was a Railtrack Programme, a continuing state of affairs is not an “event”.

These impressions are confirmed when regard is had to the degree of unity of time, of place, of cause and of intent as it would appear to an informed observer. On the face of it:

There is no unity of time. The ESR’s were imposed over a long period running into months if not years. Indeed, on one view, they continue to be imposed today.

There is no unity of location. The ESRs were imposed at hundreds if not thousands of different locations over the entire network.

Whilst there is a degree of unity of cause and intent in the form of the adoption and implementation of instructions issued in the aftermath of the derailment, yet, as has been explained, the response to Hatfield was incremental and variable. The justification for the ESRs was the discovery of individual incidents of GCC of sufficient severity to call for action under the instruction then in force, not by way of imposition of a unitary programme issued on a particular day.

For all these reasons, I accept the submission of the Herbert Smith Defendants that only losses sustained as a result of ESRs imposed before the end of October are within the scope of the cover.

The threshold issue here is whether the general exclusion of damage caused by wear and tear is applicable to the denial of access provisions at all. The Claimants placed particular reliance upon the definition of CONSEQUENTIAL LOSS : -

It was accordingly submitted by the Claimants that, since cover under the extension was triggered by a deemed incident or damage rather than actual damage or loss, it followed that by definition there could be no CONSEQUENTIAL LOSS caused by or consisting of wear and tear.

I reject this submission primarily on the grounds that the denial of access clause expressly states that the extension is to be “subject to all the terms and conditions and provisions of the policy”. Furthermore, it is common ground that the philosophy of the extension to the basic cover was to substitute a different specified event as a deemed incident. It follows, in my judgment, that losses which are consequent upon the prevention of access are incorporated into the definition of CONSEQUENTIAL LOSS.

Although this matter was debated at some length, I have no doubt that RCF is a paradigm example of wear and tear. The general discussion earlier in this judgment reflects the agreement on the nature and cause of RCF. The important section of the experts’ joint memorandum is set out in the section headed “Conclusions” as follows: -

In reality there is in fact no disagreement between the experts. Their approach to the incidence of RCF is on all fours with the definition of wear and tear to be found in Arnould Law of Marine Average and Insurance 16^th edition para 780: -

The Claimants sought to persuade me that the evidence demonstrated that “the condition of the network was plainly a consequence of years of inadequate maintenance, under investment and neglect”. Quite apart from the fact as I have already intimated that it would be wholly improper to embark on a critique of Railtrack’s husbandry of the railway network on the limited material and argument before me, it would be a pointless task for present purposes. The fact that wear and tear can be controlled by proper maintenance does not change the nature of the deterioration: it remains wear and tear.

Of course, deterioration may not be attributable to wear and tear. One example is provided by the Canadian decision in The La Pointe [1991] 1 Lloyd’s Rep 89 where repairers negligently installed the wrong type of cap screws whereby a flange failed. But in contrast, RCF, as one commentator put it, “will always be present in one form or another”. Or again, Railtrack were advised that “no administration has reported a means to prevent the occurrence of head-checking while operating a commercially viable railway”. The only way of eliminating the incipient development of RCF is to replace the rail (albeit the process may start again). Grinding may slow its development but the failure to grind does not change the nature of the process – it is still wear and tear.

The wear and tear exclusion would only be operative if the excluded peril was the proximate cause of the loss. But I agree with the Claimants that wear and tear was not the proximate cause of the losses in the present claim. The proximate or effective cause of the Claimants’ losses was the imposition of the various ESRs. The issue is sometimes put forward as essentially one of common sense. What brought about the loss? RCF was a known feature of the network (albeit the scale of it was not). It was formerly dealt with as part of routine maintenance with little, if any, disruption to the timetabled services. The losses are brought about by the myriad of ESRs imposed to try and bring the network back to a safe shape as soon as possible. The gradual development of cracks in numerous RCF sites was no more than the underlying state of affairs providing the occasion for action. It was not the event that triggered the cover.

Accordingly, I conclude that the general exclusion does not provide a defence to the claim.

It is common ground that the special condition in the St Paul policy was introduced in response to the Ladbroke Grove disaster. The immediate cause of that crash was the fact that one of the trains had passed a red light which was ill-sited. Further, when Paddington Station reopened following repairs and temporary modifications, there were a number of long-term speed and route restrictions imposed pending permanent modifications.

Special condition 2(ii) provides: -

The Claimants’ submitted that this argument was not well founded for two principal reasons: -

A construction whereby the restriction must be imposed because of physical damage to the section of line in question would deprive the cover furnished by paragraph 1A of the extension of any meaningful effect since cover would always be available under the general insuring clause in sub-section B of the policy.

Accordingly on a proper construction of the clause, damage anywhere in the network leading to the imposition of ESRs would allow recovery. Here, damage caused by the Hatfield derailment itself led to the restrictions placed upon the network elsewhere.

In response St.Paul submitted as follows: -

There are limitations upon the extent to which the fact that a particular construction occasions redundancy furnishes assistance in the proper construction. In any event the extension would not be deprived of effect since it would make it clear that cover is provided where, in consequence of physical damage, a competent authority acts to restrict access since otherwise it might be argued that the act of the competent authority and not the damage was the cause of the loss.

More importantly, the ESRs were not consequent upon the Hatfield derailment. The derailment was merely the occasion for identifying other GCC sites. The requirement that the restriction be “consequent upon” damage means, in the insurance context, proximately caused by such damage.

At one stage, I was attracted to the view that a distinction should be drawn between wear and tear simpliciter (e.g. thinning) and wear and tear giving rise to damage (e.g. cracks) and that the clause only operated to exclude the impact of the former. But this is not tenable since it is common ground that GCC is not damage if, as I have held, it is not fortuitous but attributable to wear and tear.

It also occurred to me that, read in its special context of a response to the Ladbroke Grove incident, the clause should not be construed so as to enlarge the wear and tear exclusion to circumstances where wear and tear was not the proximate cause of the insured peril, but should be treated as exhibiting a common intention to exclude the impact of restrictions brought about by defects in track design or layout as opposed to physical damage. But this approach is presented with the same difficulty since the infrastructure in the first part of the condition is presumably the same as the infrastructure in the second part. Thus it is perhaps not surprising that it was not an analysis promoted by any of the parties.

Approaching the matter more conventionally, I accept that the expression “consequent upon” will, in the insurance context, usually import the application of the notion of proximate cause: see e.g. Lloyds TSB v. Lloyds Bank [2002] Lloyd’s Rep IR 113 at p.123. On this approach, the special condition raises the question: what was the proximate cause of the ESRs? The answer is not, in my judgment, the Hatfield derailment. The only restriction proximately caused by the derailment was the closure of the East Coast mainline. The proximate cause of the other closures or speed limits was the individual decision to impose them as an outcome of the re-assessment by Railtrack of its approach to GCC.

Accordingly I conclude that special condition (ii) furnishes a defence to St Paul.

This reads: -

It was the Claimants’ case that the phrase “continuing restrictions” following reinstatement applied only to the sort of on-going or long-term restriction that ensued after the Paddington repairs. This construction was said to be supported by the word “placed” where in its context it meant, “first imposed” after reinstatement.

St.Paul contended that the exclusion or condition applied to all continuing restrictions which remained in place after the physical damage had been reinstated and thus all losses incurred after the 10^th November (when the damage caused by the Hatfield derailment was reinstated) fall outside the policy.

I prefer the Claimants’ construction. Once there is damage to the relevant part of the infrastructure, the context clearly supports the construction that, once that damage is repaired, any restriction then imposed is outside the cover.

Again, there is no need to distinguish between the two policies. Aggregation is permitted for a series of insured losses arising from one (or a “single”) event. It is common ground that this event must be different from and wider than the “occurrence” or “event” under the denial or prevention of access extensions. The Claimants submitted that the event, out of which the series of losses attributable to ESRs arose, was the derailment. The Defendants contended that the ESRs and the losses sustained as a result did not “arise out of” the derailment: in short that they were not proximately caused by the derailment.

In Axa Reinsurance v. Field supra, Lord Mustill contrasted “event” with “originating cause”:-

This approach was followed by Hobhouse LJ in Municipal Mutual Ins Ltd v. Sea Ins Co Ltd [1998] Lloyd’s Rep IR & Reins 421 at p.433 by way of contrast between “any one occurrence” and a series of occurrences attributable to “one source or original cause”.

In the present case the right of aggregation can only arise if there was a single event from which a series of insured events arose. To restrict the qualifying single events to those which proximately cause the series of losses would render the aggregation clause nearly redundant. I prefer the Claimants’ argument. In my view the qualifying single event can properly include the original causative event: see Caudle v. Sharp [1995] Lloyd’s Rep IR 433. Here, the original cause of the ESRs was the derailment since it provided the unifying factor in the form of the occasion for a complete review of GCC on the network. In that sense, the losses all arose from one event and can be aggregated.