Arbuthnott & Others v Feltrim Underwriting Agencies Limited & Others



HEARING-DATES: 10 March 1995

10 March 1995

This is a signed judgment handed down by the judge, with a direction that no further record or transcript need be made (RSC Ord 59, r9(1)(f), Ord 68, r1). See Practice Note dated 9 July 1990, [1990] 2 All ER 1024.

J Cooke QC, S Moriarty and M Smith for the Plaintiffs; .D Johnson QC, J Rowland and K Houghton for the First Defendants; B Eder QC, C Butcher and N Eaton for the Second Defendants.



PHILLIPS J: In October last year I gave Judgment in favour of some 3000 Names at Lloyd's who had sued their managing agents - the Gooda Walker agencies - and their members' agents for breaches of duty owed in contract and in tort. The breaches alleged and proved related to the negligent conduct of the business of writing excess of loss reinsurance on behalf of the Names.

In these Actions 1594 Names bring similar claims against their members' agents and their managing agents - Feltrim Underwriting Agencies Limited ("Feltrim"). Once again the gravamen of the Names' complaint is that their underwriters negligently left them exposed to the risk of huge losses in the event of one or more catastrophes occurring.


The claims relate to losses sustained by three syndicates managed by Feltrim for the underwriting years 1987, 1988 and 1989.

Syndicate 540

Syndicate 540 was a marine syndicate. In the three years with which I am concerned approximately 90% of its premium income was earned from excess of loss ("XL") business. The balance was marine direct and facultative business, including hull, cargo, marine liability and P&I Club reinsurances. The syndicate also wrote an aviation excess of loss account. The marine excess of loss business comprised whole account and specific protection. A significant proportion of this business was 'XL on XL'.

Syndicate 542

Syndicate 542 was, in 1987 and 1988, the incidental non-marine syndicate for Syndicate 540, and consisted of the same membership. In 1989 it became a full non-marine syndicate. It wrote almost exclusively excess of loss business. In 1987 and 1988 underwriting was carried on jointly on behalf of Syndicates 542 and 847 and shared on a split stamp basis.

Syndicate 847

Syndicate 847 was a non-marine syndicate writing in 1987 and 1988 almost exclusively excess of loss business. In 1989 it wrote a small direct and facultative property account and also wrote some marine business. Its membership was smaller than that of Syndicates 540 and 542, but nearly all its members were also on the other two syndicates.

Between 1983 and 1988 Syndicates 540/542 grew from a membership of 782 with a stamp capacity of approximately £15 million to a membership of 1,457 with a stamp capacity of approximately £37 million. During the same period Syndicate 847 grew from a membership of 150 with a stamp capacity of approximately £2 million to a membership of 1,174 with a stamp capacity of approximately £20 million. In the years with which these Actions are concerned, the stamp capacities of the syndicates were as follows:

Syndicate 1987 1988 1989

540/2 net £20.95m £37.1m £41.4m

gross £27.24m

847 net £8.9m £18.51m £24.5m

gross £11.6m

Mr Thompson, the Plaintiffs' expert witness, has extracted from the syndicates' reports and accounts the following sterling estimates of premium income for 1987, 1988 and 1989:

1987 1988 1989


Total Estimated Premium 24,558,300 21,515,957 25,651,455

XL Proportion 22,102,470 20,009,840 23,176,608

Percentage of XL : Total Est Prems 90.00% 93.00% 90.35%


Total Estimated Premium 2,728,700 2,399,605 6,300,000

XL Proportion 2,646,839 2,327,617 6,191,845

Percentage of XL : Total of Est Prems 97.00% 97.00% 98.28%


Total Estimated Premium 11,648,000 10,298,088 21,428,571

XL Proportion 11,648,000 10,298,088 21,428,571

Percentage of XL : Total Est Prems 100.00% 100.00% 100.00%

This gives some idea of the scale and composition of the business written by the syndicates although - in the event - these estimates were significantly exceeded, due in part to the effect of reinstatement premiums.


In 1974 WMD Underwriting Agencies Limited ("WMD") began managing two syndicates - Marine Syndicate 174 and its incidental Non-Marine Syndicate 175. The underwriter of both syndicates was Mr Colin Davies. In 1982 a new non-marine syndicate number 847 was formed. At the end of that year the propriety of certain reinsurance arrangements involving WMD and an associated agency PCW Underwriting Agencies Ltd ("PCW") was called in question. WMD was suspended by Lloyd's and only reinstated after Mr Davies had resigned, to be replaced as underwriter by his deputy Mr Fagan. In order to distinguish the new regime from the old, Syndicates 174 and 175 were re-numbered Syndicates 540 and 542.

In early 1986 it was decided that a new agency should be formed to take over from WMD the management of the syndicates. Mr Fagan and a number of his colleagues would continue to perform their existing functions in respect of the business of the syndicates, but as employees of the new agency. The new agency was Feltrim and Lloyd's approved that it should be registered to act as managing agent for Syndicates 540, 542 and 847 in December 1986.

The Board of Directors of Feltrim consisted of Mr Eric Andrew, the Chairman, who had previously been the Managing Director of Holmwoods Back & Manson (Underwriting Agencies) Ltd., and the underwriter of its Syndicate 144; Mr Jan Manning, the Finance Director, who had previously been finance director of Brown Shipley Insurance Services Ltd; Mr Andrew Drysdale, a Non-Executive Director, who had previously been the non-marine underwriter for Terra Nova Insurance Co. before becoming the Chairman of Andrew Drysdale Ltd and the underwriter of its Non-Marine Syndicate 43; Mr James Bazell, another Non-Executive Director who was also the Chairman of Stewart Wrightson Member's Agency Ltd and the two underwriters, whom I shall now introduce.


Mr Patrick Fagan's insurance career began in 1953 when he joined Home & Overseas Insurance Company Limited, a small company which wrote some excess of loss business. In 1956 he moved to English & American Insurance Company Limited and spent the next 17 years in their marine underwriting room. English & American was a member of the Institute of London Underwriters. It wrote a general marine insurance account and a book of foreign excess of loss business. In 1973 Mr Fagan left English & American to join WMD as deputy underwriter to Mr Davies. The business that Mr Davies wrote for Syndicates 174 and 175 was almost exclusively excess of loss. Mr Fagan became the active underwriter of these syndicates, under their new numbering, and also of Syndicate 847, when he superseded Mr Davies in 1982. At that time he was appointed a Director of WMD. He continued to underwrite for all three syndicates until the end of 1988. Mr Gofton-Salmond then took over as the active underwriter of Syndicate 847. Mr Fagan ceased to underwrite for Syndicates 540 and 542 on 29 June 1990 and resigned as a Director of Feltrim on 9 August of that year.

Mr Gofton-Salmond joined Norwich Union as a member of its graduate trainee scheme in 1972. There he worked in both the marine and the non-marine underwriting rooms in the London office before moving to the Liverpool office, where he was concerned with small craft business. In February 1974 he moved to WMD as underwriting assistant. At the end of 1974 he was involved in setting up a claims department, which he then took over, although most of his time was still occupied as an underwriting assistant. For six years he worked with Mr Davies, Mr Fagan and a third underwriter, Mr Jack Ritson, learning the underwriting trade. In 1980 he was given his pen - that is he was authorised to underwrite risks. Initially he was restricted to writing renewals but, from 1982, he was authorised to write new business under the direction of Mr Davies. In 1983, when Mr Fagan took over from Mr Davies, Mr Gofton-Salmond was appointed as his deputy. At the end of 1988 he resigned as deputy underwriter of the three syndicates in order to take over as underwriter of Syndicate 847.

Additional Underwriting Agencies (No.7) Limited ("AUA7")

On 8 May 1990 Syndicate 847 ceased underwriting as it had become apparent that premium income limits for 1990 were likely to be exceeded.

On 9 August 1990 Syndicates 540 and 542 ceased underwriting. On 12 December 1990 AUA7 was appointed Substitute Agent. The directors include Mr TR Berry, who was appointed Substitute Underwriter of Syndicates 540 and 542 for the years 1987-1990 on 14 December 1990 and Mr JA Beck who was appointed Substitute Underwriter for Syndicate 847 for the same years. All these syndicate years are in run-off under the management of AUA7.


As in Gooda Walker I am not concerned with quantification of loss but only with resolving certain issues of principle that arise in relation to damages. The size of the losses is only relevant insofar as it demonstrates the degree of exposure to which the Names were subject.

On 23 May 1994, in his Report on the Syndicate Accounts as at 31 December 1993, Mr Berry set out the following figures for Syndicates 540 and 542:

Year Loss % of Stamp Estimate of Ultimate Loss

1987 £79,095m 378% 375% - 425%

1988 £120,671m 326% 325% - 375%

1989 £225,604m 548% 550% - 575%

Mr Beck's Report of the same date sets out the following figures for Syndicate 847:

Year Loss % of Stamp Estimate of Ultimate Loss

1987 £9,068m 101% 100% - 200%

1989 £109,706m 448% 450% - 500%

(1988 showed a profit of 8% of stamp)

These figures are not agreed, but they give a broad picture of the overall results that are likely to be experienced by the syndicates in the relevant years.

The Catastrophes

AUA7's Accounts as at 31 December 1993 identify the catastrophes that are primarily responsible for the losses suffered by the syndicates and provide projections for the losses attributable to each catastrophe, after reinsurance recoveries:

87J Windstorms

1987 Syndicates 540/542 £40.8m 195% of stamp

1987 Syndicate 847 £13.8m 155% of stamp

Piper Alpha

1987 Syndicates 540/542 £ 8.3m 40% of stamp

1988 Syndicates 540/542 £55.8m 151% of stamp

Hurricane Hugo

1989 Syndicates 540/542 £84.9m 206% of stamp

1989 Syndicate 847 £77.1m 315% of stamp

Exxon Valdez

1988 Syndicates 540/542 £25.2m 68% of stamp

1989 Syndicates 540/54 £49.1m 119% of stamp

90A Windstorm

1989 Syndicate 847 £25m 102% of stamp

(1990 Syndicate 847 £33.5m 155% of stamp)

Once again these are not agreed figures, although it is common ground that these were the most significant catastrophes. Other catastrophes which resulted in significant losses before reinsurance recoveries were Enchova in 1988, Arco Baker (Atlantic Richfield) in 1989 and Phillips Petroleum in 1989.

The agreed FGO (from ground up) incurred losses from the eight catastrophes, as at 30 June 1994, were as follows:

DATE EVENT 540 542 847

16-17 October 1987 87J £52.3m £9.5m #39.6m

24 April 1988 Enchova $93.4m N/A N/A

6 July 1988 Piper Alpha $168.5m N/A N/A

19 March 1989 Arco Baker $61.4m N/A N/A

(Atlantic Richfield)

24 March 1989 Exxon Valdez $123.7m N/A N/A

15-22 September 1989 Hurricane Hugo $165.5m $93.8m $208m

23 October 1989 Phillips Petroleum $99.4m $6.8m N/A

25 January 1990 90A £41m £25.5m £101.5m

The Basic Issue

The evidence establishes and the Defendants accept that the Feltrim underwriters neither intended nor foresaw that individual catastrophes of the scale experienced would result in claims that outstripped the reinsurance cover that they had put in place. The Names allege that the exposure to which they have been subjected is attributable to a failure to apply, or to apply competently, established principles of excess of loss underwriting in relation to assessing and making provision for exposure to a single loss event. The Defendants contend that the underwriters adopted a competent approach to vertical exposure and that the losses that the Names have suffered are attributable to two factors that could not reasonably have been foreseen or provided for:

(i) The impact on marine cover of non-marine losses.

(ii) An unprecedented series of catastrophes.

I propose at this stage to deal with relevant aspects of the underwriting of excess of loss business and, before doing so, to introduce those who gave expert evidence in relation to underwriting.

Mr Thompson

The Plaintiffs only called one expert witness, Mr Hugh Thompson. Since 1991 he has practised as a full time consultant to the insurance and reinsurance markets. Before that he had had some thirty years experience in the London reinsurance market. This included two relatively short periods of employment with Lloyd's brokers. For most of this time, however, Mr Thompson worked as an underwriter for a number of different companies. This involved trading extensively with many Lloyd's syndicates and a period from 1979 to 1983 when he wrote a book of LMX business.

Mr Thompson's experience of LMX business did not match the experience of the Defendants' experts individually, let alone collectively, and his evidence at times was based more on theory than on practical experience. He was competent to express an opinion on all relevant aspects of excess of loss underwriting, but in some respects I found his evidence unrealistic and unconvincing.

Mr Emney

Mr Emney was the first Chairman of the Joint Excess Loss Committee, a body established by the Lloyd's Underwriting Association and the Institute of London Underwriters. He has the benefit of 24 years involvement in the LMX market, and during eleven years with the Merrett group he was one of the principal leaders in that sector. He has given many speeches and papers on the LMX market, and was highly qualified to give expert evidence in this case.

Mr Outhwaite

Mr Outhwaite gave evidence in Gooda Walker and gave me no reason to alter the impression I formed of his qualifications and ability, as expressed in my Judgment in that Action, but for one small matter. I had understood that Mr Outhwaite had written catastrophe business at the material time inasmuch as he wrote a book of back up business. He made it clear, however, that this business was written at working layer level. Mr Outhwaite kept prudently clear of the catastrophe market when the spiral was at its height.

Mr Fryer

Mr Fryer had also given evidence in Good Walker and once again I had no reason to revise my assessment of his expertise.

The London Excess of Loss Market ("LMX")

In my Judgment in Gooda Walker I dealt with the following matters:

1) the structure of Lloyds;

2) the duties owed to Names by Managing Agents and Members Agents;

3) the nature of LMX business;

4) the spiral;

5) the principles applicable to excess of loss underwriting;

6) the standard of skill and care to be expected of the excess of loss underwriter.

In these Actions my conclusions in relation to these matters have been largely accepted by Counsel and by the experts. In these circumstances I do not propose to repeat the detailed analysis that I carried out in Gooda Walker. This judgment should be read in conjunction with my judgment in that case. I do, however, propose to summarise certain conclusions that I reached in Gooda Walker that have been reinforced by the evidence that I have heard in these Actions. They set the context in which the issues that have been raised in this litigation fall to be resolved. In so far as issues of expert evidence have arisen, I shall deal with them later in my Judgment.

The Spiral

The writing of reciprocal excess of loss reinsurance, or tertiary business, which has been described as "the spiral", was an essential characteristic of the LMX market at the time with which these Actions are concerned. In their final written submissions, the Defendants observe:

"Without the LMX market, the Lloyd's (and indeed London) market would probably not have had the capacity to write all the excess of loss business which was actually placed, certainly not on the terms on which it was in fact written."

This is correct. Spiral business was an aberration. Many, if not most, of those who engaged in it did so in the belief that the reinsurance that they were buying from their colleagues was providing a protection from exposure when this was largely illusory. The capacity that was provided by the market was involuntary. Had all members of the LMX market correctly applied the agreed principles of competent excess of loss underwriting, the form and capacity of the market would have been radically different. The conduct of the individual excess of loss underwriter falls to be considered, however, having regard to the market that existed, even if this was an aberration. Some underwriters succeeded in conducting business in this market that did not result in heavy losses. Neither in Gooda Walker nor in these Actions have the Plaintiffs alleged that it was negligent per se to write spiral business. The allegations of negligence are freestanding charges of failure to follow fundamental principles of excess of loss reinsurance. The factual context in which those allegations have to be considered is one in which the buying of reinsurance cover to restrict exposure was a vital feature of the practice of most underwriters in the market.

Appreciation of the Spiral

I am satisfied that both Mr Fagan and Mr Gofton-Salmond appreciated the gearing effect of the spiral. Indeed, it transpired that Mr Gofton-Salmond was present when Mr Outhwaite delivered his now famous paper on "Mainspring or Vulnerability", which is extensively cited in my Gooda Walker Judgment. Whether the Feltrim underwriters paid proper heed to those implications is a matter to which I shall revert.

Rating and Vertical Exposure

Just as in any other kind of insurance, it is a fundamental principle of excess of loss underwriting that the premium received should be commensurate with the risk assumed. When considering a risk, the underwriter has to consider the odds on the risk written resulting in a claim. If the underwriter gets the odds right, a balance will be achieved whereby the premiums paid are sufficient to meet the claims made. In many fields the underwriter can aspire to achieve this balance year in year out. This will not necessarily be the case in respect of excess of loss underwriting. The excess of loss underwriter reinsures against the risk of an accumulation of losses resulting from a catastrophe. If cover is written at a level that will only be impacted by a serious catastrophe, it may only be possible to achieve a balance between premiums and claims over a period of a number of years. In some years there will be no such catastrophe, and thus no claims, but in a year where such a catastrophe occurs, the aggregate of claims will exceed the premiums received. The evidence in this case, and in Gooda Walker establishes that this feature of excess of loss underwriting poses particular problems for the Lloyds syndicate. Because underwriting at Lloyds is done on the basis of the syndicate year, the managing agent is not in a position to build up syndicate reserves in the good years in order to meet very heavy claims in a year where a severe catastrophe, or series of catastrophes, occurs. For this reason, the underwriter has to seek to restrict the exposure on the book of business written to a limit that his names can reasonably be expected to bear.

The data that the underwriter will need if he is to make an informed assessment of the appropriate rate for an individual risk will be similar, if not identical, to the data that he will need if he is to make an informed assessment of the extent to which the individual risks may accumulate, so as to arrive at a basis for planning his net exposure. This exercise is commonly described, at least now, as assessing the probable maximum loss, or PML, and is dealt with in detail in my Gooda Walker Judgment. In these proceedings the Defence experts tended to muddy the water somewhat by postulating that the casualty envisaged as the basis for the PML calculation should be an event that was "likely". On analysis it proved that this adjective was used to define the practical, as opposed to the theoretical, possibility, and the expert evidence gave me no cause to revise my findings on this topic in Gooda Walker.

The assessment of a PML is not an end in itself - it is a stepping stone to planning the reinsurance cover that will be required to reduce exposure to an acceptable limit. If no PML exercise is done, no sound basis will exist for calculating exposure. PML calculations will, however, be of no value if they are not allied to a sound reinsurance programme.

Horizontal Exposure

A further factor complicates the task of the excess of loss underwriter in deciding what is the appropriate level of exposure to run. It is possible that more than one severe catastrophe may occur in the same year. In considering the extent to which he will expose the Names of a syndicate year to loss, he has to make a judgment as to the possible effect of a series of catastrophes in that year. To cope with the effect of a series of catastrophes, or loss events, the underwriter has to consider both vertical exposure and horizontal exposure. If he writes ten covers, without reinstatements, all of which accumulate so that one catastrophe gives rise to ten claims, his only concern will be vertical exposure. Once that catastrophe has occurred, he will no longer be on risk. If none of the risks accumulate, he will have horizontal exposure to a possible ten successive loss events. It was the general practice of those writing marine catastrophe business to buy and sell cover that provided for two reinstatements. The non-marine catastrophe cover usually provided for one reinstatement. In selecting the appropriate level of vertical cover, the underwriter would secure protection against the risk of a series of two or three catastrophes exposing his Names to unacceptable losses from accumulating risks. He had, however to bear in mind that, to the extent that risks did not accumulate, a series of catastrophes would involve horizontal exposure, albeit to what would probably be lesser individual losses. Furthermore, depending on the terms of the cover, relatively modest losses could erode the vertical cover that would otherwise be available against the major catastrophe.

In these Actions the Names have restricted their attack to one aspect of the underwriting carried on by the Feltrim agency. The extent to which the Names were exposed to a single loss event. No allegations have been made in relation to the rating of risks. No allegations have been made in relation to the extent to which the Names may have been prejudiced by lack of horizontal cover. This restricts the factual issues that I have to resolve and carries implications in relation to the assessment of damages.

Unlimited Liability

The fact that a Name who joins Lloyd's deliberately agrees to expose himself to unlimited liability does not mean that he anticipates or accepts that when he joins a syndicate the active underwriter will deliberately expose him to the risk of such liability. On the contrary the Name will reasonably expect the underwriter to exercise due skill and care to prevent him from suffering losses. In many categories of insurance the Name will reasonably expect the underwriter to plan to procure profits, year in year out. The fact that syndicates are reconstituted each year does not, however, make it mandatory for an underwriter to conduct business in this way. Underwriting at Lloyd's must be conducted as an ongoing business. There is no reason in principle why an underwriter should not write business on the basis that net losses will be made in some years that are balanced by generous profit levels in the other years. If, however, an underwriter is deliberately to expose his Names to suffering losses from time to time, he must make sure that the Names are aware of this and of the scale of loss to which they will from time to time be exposed.

It was a fundamental principle of excess of loss underwriting that the underwriter should formulate and follow a plan as to the amount of exposure that his syndicate would run. During the period with which I am concerned such a plan would normally - if not inevitably - involve restricting the syndicate's gross exposure by reinsurance in order to attain the planned level of net exposure.

Aggregates and PMLs

I have already explained that in order to monitor the exposure that results from the business he writes, the excess of loss underwriter must be aware of his aggregates. He has the advantage that each piece of business he writes is subject to an express limit of liability. To calculate his exposure to a single event he needs to know how many of the covers that he has written are exposed to the risk of a claim should that event occur. He thus has to divide into different categories the covers that can accumulate. In practice this is normally done by a system of coding the different categories. The more carefully the business is recorded under appropriately chosen codes the more confident the underwriter will be able to be as to the limit of his exposure to a single event. There will be some categories where it is unlikely, or indeed inconceivable, that a single event will result in a claim on every cover. In respect of those categories the true exposure will be, not the aggregate, but the PML. The estimation of the PML has to be made by the application of judgment to the data available.


A competent excess of loss underwriter should adopt a reinsurance policy that includes the following elements:

(1) The underwriter should know the PML that his syndicate will be exposed to in the event of the worst catastrophe that is a practical possibility.

(2) He should then work out what net exposure he is prepared to run in that event.

(3) He should reinsure the balance.

(4) Where he retains his exposure should depend upon market considerations. Those layers of exposure which he protects by reinsurance should depend upon the view he takes of the terms of the reinsurance on offer.

The Detailed Issues

Neither the underwriters nor the Board of Feltrim intended or foresaw that the Names might suffer losses of the magnitude of those that they have sustained. The Defendants' case in relation to those losses can be summarised as follows:

1.The underwriters at all material times had in place a policy designed to leave the Names with an exposure, albeit a limited exposure ("the planned exposure") above the level of reinsurance protection.

2.The policy was a proper one.

3.The policy was approved by the Board of Feltrim.

4.The implementation of the policy was monitored by the Board of Feltrim.

5.Members' agents and, through them, Names should have appreciated that those on the syndicates would be subject to the planned exposure.

6.Insofar as losses exceeded the planned exposure, this was attributable to factors which the underwriters and the Board did not foresee and could not reasonably have been expected to foresee.

The Names join issue with each one of these contentions. Their case can be summarised as follows:

1.Underwriters had no policy of subjecting the Names to a planned exposure.

2.If such a policy was followed, it was not a proper one.

3.No such policy was approved by the Board of Feltrim.

4.The Board of Feltrim failed to monitor the exposure to which the Names were subject.

5.Neither members' agents nor Names knew or should have known that they were exposed to the extent of the allegedly planned exposure.

6.The underwriters should have foreseen and protected the Names against the exposure to which they were subject.

Before turning to issues that are controversial, I propose to summarise those areas of the relevant evidence that are not in dispute.

Records Kept by the Syndicates

1)When a risk was written a photocopy would be taken of the slip. On that photocopy a box code would be completed recording, inter alia, when the risk was written, the year of account and the category of business - eg. 'marine excess of loss'. Any written placing information that was considered particularly relevant would be photocopied and attached to the slip.

2)A further box code was then completed, recording information which included:

-the line size

-the anticipated signing

-an estimate of the premium income that would be generated by the risk

-the estimated net premium income of the account of the cedent

-the risks' statistical code.

3)The Statistical Code

In order to enable them to monitor the syndicates' business, the underwriters categorised the risks written using a statistical code ("the Stat Code"). Before 1989 the Stat Code consisted of four letters and a number. The first two letters represented the geographical location of the risk. For example, under the code used by Syndicate 847:



FJ=Japan (Nationwide)

NA =USA (Nationwide)


The third and fourth letters represented the class of business and the basis on which the risk was to be written. For example, the principal codes used for the Marine and Non-Marine accounts were as follows:

Marine Classifications

Accounts Reference

Whole Account including X/L HM

Whole Account excluding X/L HR


Hull HH

Building risk HC

Liability HL


Rig HO

War HW

Non-Marine Excess of Loss ("X/L") Classifications

Accounts Reference

Generals including XL

(ie. XL estimated net premium

income ("ENPI") > 10% NN

General including XL

(XL ENPI < 10% or specifically

protected NM

Generals excluding London

market Excess of Loss ("LMX") NJ

Generals excluding LMX & RIA NK


USA Reinsurance

(including casualty) NT

USA Reinsurance

(excluding casualty) NU

Generals Ex USA Losses NL

LMX excluding XL on XL NY

The final element of the Stat Code was a number which was intended to give an indication of what the underwriters described as the "volatility" of the risk -that is the likelihood that the risk would give rise to a loss. I shall describe this number as "the volatility factor". The volatility factor was determined in different ways, depending upon the nature of the risk, but in general the number would be determined by reference to a comparison between (a) the point in excess of which the risk attached and (b) the cedent's estimated net premium income. The excess point was expressed as a percentage of the estimated net premium income and the percentages figures were then graded into 7 categories or bands numbered 2, 3, 4, 6, 7, 8, 9 according to a matrix. 2 represented what was considered to be the most volatile risk and 9 the least volatile, or most remote. As the likelihood of a risk producing a loss was considered to depend to a considerable extent upon the type of business to which the risk related, each type of business had its own column in the matrix with the 2-9 bands corresponding to different percentages depending upon the perceived volatility of the different types of business.

In 1989 Syndicates 540 and 542 continued to use the code as described, but Mr Gofton-Salmond introduced some refinements to the code for Syndicate 847 which included a more detailed geographical code and a 'catastrophe perils code' which indicated the type of catastrophe to which a risk was most likely to be exposed.

4)The Register of Reassureds

An alphabetical register of cedants was kept. Risks written were entered against the appropriate cedant and a six year record was kept of each risk. This would include the information recorded on the slip and any other relevant information that the underwriter wished to record - such as details of the applicable rate on line.

5)The Accumulation Book

This book had a number of columns and separate sections for each category of risk written. Each of these sections was broken down further so that, for each category of business, there was a separate record of the risks written within each volatility band and each geographical area. For each risk written there was inserted details of the cedent, the aggregate exposure written, the premium and the expiry date of the risk.

Between 1988 and 1990 the syndicates were developing an alternative computerised record keeping system and, from 1989, Mr Gofton-Salmond ceased to keep a Register of Reassureds and an Accumulation Book in relation to Syndicate 847, relying instead on his computer records.

6)The Quarterly Accumulation Ledger

This was made up each quarter. It recorded the premiums written for each category of business and the unexpired aggregates. These were detailed for each volatility band of each category of business.

The Structure of Reinsurance Protection

The reinsurance programme effected for 1987 was, in effect, a renewal of the 1986 programme. This was a single common programme that was intended to cover all three syndicates. The Syndicates purchased Whole Account protection which was designed to provide vertical and horizontal cover for Marine Syndicate 540 and also to be available to protect Non-Marine Syndicate 847 and incidental Non-Marine Syndicate 542. The Syndicates then purchased War and Rig Specific cover for Syndicate 540 and additional cover for Non-Marine Syndicate 847 (and incidental Non-Marine Syndicate 542) to combine with the Whole Account cover to protect vertical and horizontal exposure of Syndicates 847 and 542.

The 1986 joint reinsurance programme that was renewed in 1987 was structured so that there was:

(a) A pillar of Aviation and Non-Marine protection which protected Syndicates 540, 542 and 847 in respect of Non-Marine and Aviation losses. (Syndicate 540 was covered because Aviation was written into that Syndicate):

(b) a pillar of Whole Account protection which0- excluded London market Excess of Loss business. (This was available to all three Syndicates and could provide substantial cover in the event of a US catastrophe causing losses to the US Catastrophe and Retrocession accounts written by the Syndicates):

(c) Rig Specific protection (which was very widely worded and provided protection in respect of all oil related losses) for Syndicate 540;

(d) War protection for Syndicate 540 which had the benefit of other protections without war exclusions or limitations; and

(e) a substantial Whole Account "umbrella" protection which protected all three Syndicates and had the benefit of all of the underlying policies (listed above). Within the retention of this programme, they also bought horizontal protection for up to twenty losses.

In 1988 this programme was repeated, albeit that the extent of cover was increased to a degree.

When Mr Gofton-Salmond took over responsibility for the underwriting of Syndicate 847 in 1989, it was decided that the reinsurance programme for that syndicate should be made independent of the programmes for Syndicates 540 and 542. Syndicates 540 and 542 retained the whole account protection which had previously provided cover for all three syndicates and some of the layers of the 1988 aviation and non-marine protection. The remaining layer of this protection was allocated to Syndicate 847. The major elements of Mr Gofton-Salmond's programme for Syndicate 847 consisted of:

-Marine and aviation protection

-USA specific protection

-Substantial general whole account protection which protected all accounts and which had the benefit of any underlying specific protection.

I shall, in due course, have to consider the adequacy of the protection afforded by the insurance programmes that I have described. My next task is, however, to consider the issues which arise in respect of the underwriting carried on by Mr Fagan.


Did Mr Fagan Calculate PMLs?

Mr Fagan was at all material times the underwriter for Syndicates 540 and 542. In 1987 and 1988 he was also the underwriter for Syndicate 847. Mr Gofton-Salmond, his deputy, took over as underwriter of 847 in 1989. Mr Gofton-Salmond told me that prior to this, Mr Fagan was responsible for the underwriting and reinsurance policy not merely of 540 and 542 but also of 847. Mr Gofton-Salmond was only involved in the planning to a limited extent.

So far as Mr Fagan is concerned, the Names contend that when underwriting in 1987, 1988 and 1989 Mr Fagan neither carried out PML calculations nor followed any policy of effecting reinsurance designed to result in a planned exposure for his Names. This case is one which has developed in the course of the hearing and has resulted in two significant amendments of the Points of Claim. I shall outline how this came about.

The Points of Claim originally alleged that PML calculations were carried out in respect of all syndicates by Mr Fagan on a basis that placed risks into different categories according to their perceived degree of remoteness ("volatility factors") and disregarded the more remote categories. The pleading further alleged that, in respect of the non-marine syndicates 542 and 847, percentages were applied to the aggregates for different classes of business in order to arrive at "wind factors" - that is the PML for the effects of a windstorm. This pleading was, I suspect, based largely upon the Report of a Lloyds Loss Review Committee chaired by Sir Patrick Neill ("the Neill Committee").

Discovery produced documents which appeared to confirm the use of wind factors when planning the underwriting for Syndicates 542 and 847 for all years. Also disclosed were what appeared to be contemporary documents in the hand of Mr Fagan evidencing PML calculations on the basis of volatility factors.

On the third day of the trial I gave the Plaintiffs leave to amend their pleading to advance a case that Mr Fagan carried out no PML calculations on behalf of Syndicate 540. Underlying this amendment was a contention that the manuscript PML calculations were not done as part of a contemporary planning exercise but retrospectively, after the business had been written and the reinsurance effected.

At this stage of the trial Mr Gofton-Salmond was called to give evidence. In the course of cross-examination, in a number of passages which were not without a degree of equivocation, he said that Mr Fagan's planning for Syndicates 542 and 847 had not been based upon wind factors, although he might have taken cognisance of these, but had followed the same approach as Mr Fagan had adopted for Syndicate 540. In the light of this evidence Mr Cooke, QC, for the Names, applied successfully further to amend his pleading to extend the allegations made in relation to the underwriting of Syndicate 540 to cover the underwriting for Syndicate 542 and for 847 up to the time that Mr Gofton Salmond took over responsibility for that syndicate.

In consequence of these amendments it is necessary to distinguish carefully between the underwriting for Syndicate 847 in 1989 and the underwriting for the other syndicate years.

The Evidence

The Plaintiffs have called no witness of fact. They have relied to prove their case entirely upon evidence emanating from the Defendants, and principally from Feltrim.

The principal witness of fact called by the Defendants was Mr Gofton-Salmond. As I shall indicate there were some respects in which I found his evidence unsatisfactory. The other witness of fact called by the Defendants was Mr Adamson. He was employed by Feltrim from 1 September 1988 as Deputy Underwriter for Mr Gofton-Salmond on Syndicate 874. His primary concern was with writing non-spiral business, but as Mr Gofton-Salmond's deputy he was involved in the entirety of the syndicate's underwriting, and he had a good understanding of the LMX market. He was a straightforward witness who gave his evidence with impressive confidence and objectivity.

Mr Fagan, who would naturally have been the principal witness for the defence is an invalid and has not been fit enough to give oral evidence. I have, however, had to evaluate a considerable body of evidence emanating from Mr Fagan: the evidence that he gave to the Neill Committee, an affidavit sworn in opposition to an application for summary judgment and three witness statements adduced under the Civil Evidence Acts. I regret that I have not had the opportunity to see Mr Fagan in the witness box, for his veracity has been put in issue by the Plaintiffs.

The Underwriting Policy for Syndicates 542 and 847 in 1987 and 1988

Although he did not have charge of the conduct of the underwriting for syndicates 542 and 847 in 1987 and 1988, Mr Gofton-Salmond prepared at the end of 1987 a policy document for 1988 which set out PML wind factors - that is the proportion of each category of business which would aggregate for PML purposes. These were the same for the two syndicates and, for syndicate 847, produced the following result:

Target Aggregate



GENERALS X/L > 10% 31.25M 80% $25.00M

GENERALS X/L < 10% 30.00M 50% $15.00M

GENERALS EX LMX 24.00M 25% $ 6.00M

GENERALS EX LMX & RIA 10.00M 10% $ 1.00M

LMX & RIA SPECIFIC 60.00M 100% $60.00M


NATIONWIDE 35.00M 100% $35.00M

LARGEST ZONE 5.00M 100% $ 5.00M


25.00M 100% $25.00M


It is the Defendants' case that, in 1987 and 1988 Mr Fagan carried out a similar exercise to that of Mr Gofton-Salmond. Mr Gofton-Salmond's wind factors were a little more pessimistic than Mr Fagan's and Mr Fagan was prepared to adopt Mr Gofton-Salmond's figures for planning purposes. This case is essentially based on a lengthy description in Mr Fagan's third witness statement of the exercise that he claims to have carried out. As this forms the foundation of the Defendants' case in relation to this issue, I propose to set it out verbatim:

"I always considered that the non-marine syndicates were more susceptible to major catastrophe losses than the marine syndicates over all classes of business. Accordingly I did not regard the division between categories 2-6 and 7-9 as a reliable indication of the exposure of the account to a major catastrophe but regarded the higher categories as likely to suffer some impairment on a PML basis. I therefore tended to investigate the aggregate written in each class of business in each category (i.e. 2-9). It was this assessment of the likely impact of a major non-marine catastrophe upon the aggregate in each of the categories 2-9 which was the basis of my percentage factors for the non-marine PMLs. Having formed this view, I would derive a composite percentage factor to be applied to the 100% aggregate figure for each class of business, taking into account my assessments of the percentage exposures in categories 2-9.

The major loss to which the non-marine syndicates were in my judgment practically exposed was a major US East coast windstorm. I considered the syndicates' exposure to earthquake, but I thought that, in general, the likelihood of a catastrophic earthquake giving rise to enormous insured losses was, in practical terms, very much less than that of a major US East Coast windstorm. For example, in California, the major claim arising out of the 1906 earthquake occurred because of the fires which destroyed the mainly wooden buildings in the city after the earthquake. By contrast, a major Californian earthquake in the 1980s would not, in my view, have caused such a significant insured property losses for a number of reasons, not least, the building regulations which had been introduced and the fact that very many property owners buy only limited earthquake cover (if any) due to its cost. Taking these factors into account I approached my PML as set out below.

For XL on XL (class NX), I generally regarded all categories (i.e. 2-9 - the entire aggregate excluding backups) as substantially exposed to such a windstorm which led me to include that class in the PML assessment at 100%.

With the other categories of clashing business I would carry out detailed reviews of the reinsureds, the structure of their reinsurances and the level of our participations, their particular exposures and my knowledge of the reinsured in order to assess what I believed to be the appropriate factor to be applied to those categories when assessing my PML.

Based on my analysis of the account, I would assess a PML for the non-marine Syndicates by applying percentage factors to the 100% aggregates in each class of business. These percentage discounts varied from time to time depending upon what view I formed as a result of my type of analysis. As I have said above, without my working papers, I cannot now recall the precise percentage factors which I applied to each category during 1987 and 1988 (apart from the fact that I applied a factor of 100% to the XL on XL aggregate); for 1989, the documents at pages 363 and 364 of the Underwriting Bundles are a copy of my workings as at 30th September 1989 which I would have carried out shortly after the quarterly accumulation ledger as at that date was prepared in, I expect, about early October 1989 (page 364 is a typed version of page 363) The percentage factors set out in that document for each of the classes of business would be similar to those I used during 1987 and 1988. I believe that, in the case of generals, the percentage factors used in 1989 may be lower because in 1989 we had consciously moved up the programmes of our reassureds and, accordingly, the layers being written by Syndicate 542 were at a higher level than previously. The figure of 50% applied to both Categories NN and NM and represented a composite figure derived from my analysis of the underlying account.

The division of responsibility between myself and Robert and myself was that, essentially, I dealt with all the underwriting policy decisions and reinsurance planning for the Marine Syndicate; Robert was primarily concerned with the non-marine side of the business and we did not, as a matter of course, discuss the Marine Syndicate's business in detail; as I have stated elsewhere, I did a lot of the analysis and planning in the evenings and weekends, and often at home. There was no need for Robert to be involved with the detail on the marine side and, in fact, he was not to any great extent. Robert concentrated on the non-marine side (although prior to 1989 the primary responsibility lay with me and I was in charge of reinsurance purchasing as this tied in with the marine syndicate) particularly when it was decided that the syndicates would split and that Syndicate 847 would become a truly independent unit. I was aware that Robert and I had slightly differing views as to the appropriate percentages to take for the wind factors. I recall that in 1987 he produced a set of wind PML factors which he applied to the non-marine aggregates during 1987 and 1988. I recall that I regarded Robert's factors as a little more pessimistic than mine, in particular his factors for US cat and US retro (classes NH, MG, NU and NT) and for whole accounts (Classes NN and NM).

I append to this judgment as Annexe 1, p.364 of the Underwriting Bundles to which Mr. Fagan refers. It is apparent that the exercise that Mr. Fagan alleges that he was accustomed to carry out was essentially the same as that carried out by Mr. Gofton-Salmond.

The Plaintiffs do not accept that the account given by Mr. Fagan in his Third Witness Statement is truthful. Nor do I. I have reached the firm conclusion that it is largely the invention of Mr. Fagan. This is not a conclusion that I have reached lightly and I propose to give my reasons for it in a little detail.

First there are some general comments that I have to make on the quality of the evidence.

There has been preserved a substantial body of contemporary Minutes of meetings of the Board of Feltrim. Inevitably the precise accuracy of such records must be open to question. Overall, however, they give a fairly clear picture of material events and I consider them to be evidence of high probative value.

Oral evidence given by Mr. Fagan, Mr. Gofton-Salmond and other witnesses to the Neill Committee was provided at relatively short and informal meetings where the questioning was not carefully structured and in circumstances where the witnesses were not able to refer to the relevant documentation to refresh their memories. The impression that comes across clearly from the transcripts is that the witnesses attempted, to the best of their recollections, to answer questions frankly and without reserve. In such circumstances it is not realistic or fair to expect that witnesses recollections will always have resulted in accurate answers, but where a clear picture is given, or witnesses corroborate one another, the evidence is cogent.

Mr. Fagan's Affidavit and his witness statements contain detailed evidence and have plainly been prepared with care. They represent an attempt, retrospectively, to demonstrate that Mr. Fagan followed competent practices in relation to the assessment of PML's and the formulation of a programme of reinsurance designed to place an appropriate limit on his Names' exposure. The attempt does not succeed, for the accounts that Mr. Fagan has chosen to give in his more recent statements cannot be reconciled with the contemporary documents and with the evidence that he gave to the Neill Committee. Nor has Mr. Gofton-Salmond been able to support them. I turn now to deal with the various respects in which I have found Mr. Fagan's statements unsatisfactory.

1) Mr. Fagan's Evidence is Incompatible with the Oral Evidence Given by Mr. Gofton-Salmond

I have already referred to equivocation on the part of Mr. Gofton-Salmond. He was in the difficult position of having to choose between loyalty to Mr. Fagan and frankness to the Court and the course that he adopted involved a degree of compromise. Nonetheless, his evidence when considered as a whole cannot possibly be reconciled with Mr. Fagan's statement.

He told me that he was responsible, during the summer of 1987, for devising the wind factors, adopting an approach of Mr. Robert Kiln, the marine underwriter and author of Reinsurance in Practice. He embarked on the exercise in order to get a better feel for the PML aggregates on the non-marine side - an improved system to that which was in use. Mr. Fagan knew what he was doing and found it interesting. This evidence makes no sense if Mr. Fagan was already operating a system using wind factors. In fact, Mr. Gofton-Salmond's oral evidence was that Mr. Fagan was using for the non-marine syndicates the same system that he was using for the marine syndicate. This was based on the clashing classes of business in volatility bands 2 to 6. Mr. Gofton-Salmond produced his wind factors as a cross-check. Mr. Fagan took cognisance of these wind factors, but continued to apply his own system. At a later stage of his evidence this theme developed to the extent that he suggested that in 1988 Mr. Fagan applied simultaneously two different systems of assessing PMLs. This contrasted with a clear and unequivocal passage in his earlier evidence:

"Q.Do you say that this [the system of volatility banding] was used by Mr. Fagan, the system in the context of assessing his reinsurance requirements?

A.As an element of it, yes, I do.

Q.It was not part of the system that you used in looking at the question of PML and reinsurance requirements for Syndicate 847 in 1989, was it?

A.No, it was not.

Q.What about its used in 1988 on the Non-Marine side? was it used by Mr. Fagan for that?

A.Yes, it was at that stage, yes.

Q.So when he was assessing reinsurance requirements for Syndicate 847 in 1988, he would have reference to this system rather than your system that you put into place for 1989?


Q.And 1987 likewise?


Q.So is it fair to say that the banding system, you say was a system used by Mr. Fagan, but you had a separate system that we have seen in the underwriting documents whereby you assessed various percentages for different classes of band as a wind aggregate or a wind-exposed factor?


Q.But Mr. Fagan did not adopt that sort of system himself?

A.No, he did not.

Q.Why did you not use the same sort of system in 1989 on the Non-Marine side as you think Mr. Fagan used in 1988 and 1987?

A.I think because we had set out in 1989 to expand the base of the account to encompass different classes of business within a Syndicate, and under that - under those circumstances we considered the, what I call new system, was more apposite for the type of account which we were writing or attempting to write."

It emerges clearly from Mr Gofton-Salmond's evidence that the system of wind factors which he devised was indeed a new system which contrasted with Mr Fagan's system which was based on volatility bands.

2)Mr Fagan's Third Witness Statement does not accord with his earlier evidence in this case

I have already explained how the Plaintiffs originally alleged that wind factors were used to calculate PMLs for Syndicates 542 and 847. Those wind factors were the ones prepared by Mr Gofton -Salmond. Mr Fagan dealt with the Points of Claim in an affidavit which he swore in resisting summary judgment and in his first witness statement. In the latter he simply stated in relation to the wind factors:

"it appears that these figures were taken from a document prepared for me by Robert Gofton-Salmond..."

No suggestion was made that Mr Fagan had his own, similar, system for calculating PMLs. In both that witness statement and his affidavit an essentially identical passage appeared dealing with "reinsurance methodology":

"I have already stated above that the reinsurance methodology applied on the Feltrim Syndicates was brought to WMD by my predecessor as chief underwriter at WMD: Mr. Colin Davies. Mr. Davies had used the methodology at the Indemnity Marine Insurance Company ("The Indemnity") - a respected company in the LMX market - and at the American Home Assurance Company ("The Home"), a very large US insurance company which is active in the London market. It is my understanding that this methodology (essentially a method of categorising risks according to their perceived remoteness from loss) as used at The Indemnity and The Home for many years, and as far as I know may still be in use. Mr. Davies introduced it to the WMD Syndicates and I learned this method from him. Mr. Davies was considered in his time to be one of the most astute marine underwriters in the market and one of the foremost leaders of LMX business.

This passage applied to all the syndicates, not just 540. It aptly applied to a system based on volatility bands, but not to a system based on wind factors, as exemplified by ANNEXE 1.

3) Mr. Fagan's Evidence to the Neill Committee

Mr. Fagan spoke to the Neill Committee about using volatility bands in relation to considering exposure, but made no mention of calculating wind factors. It is fair to say that the material passages might largely have related to the marine underwriting, but it is nonetheless surprising, if Mr. Fagan based non-marine PMLs calculations on wind factors no mention was made of this.

4) Inconsistency with Mr. Fagan's Evidence on Exposure

It was Mr. Fagan's evidence that he used his PML calculations as a basis for planning reinsurance cover that would restrict net exposure to 100% of stamp.

Mr. Fagan said in his Affidavit that he estimated that a truly catastrophic non-marine loss could impair 50% of the marine XL on XL aggregate. Had Mr. Fagan carried out the PML calculations for Syndicates 542 and 847 that the Defendants allege and envisaged that his whole account cover would be simultaneously exposed to claims on the marine and non-marine syndicates, he could not have believed that his overall exposure would be restricted to 100% of stamp. This point carries only limited weight in the present context, for I am sceptical as to the extent to which Mr. Fagan did in fact anticipate that one non-marine event would expose the whole account cover to claims from both the marine and non-marine syndicates.

5) Technical Implausibility

Mr. Fagan's description of how he arrived at wind factors was far from clear but, according to his statement, involved a detailed consideration of each risk written and an assessment of the exposure of each volatility band for each class of business. When I attempted to envisage the exercise described by Mr. Fagan I found this impossible. His description, superficially impressive, did not bear analysis. I sought assistance from Mr. Outhwaite, but he was as bemused as I was. Mr. Fagan had given, in his second witness statement, a similar description of the system that he adopted when assessing the PML for marine business. Neither Mr. Outhwaite nor Mr. Emney was able to help me to make sense of this.

6) Mr. Fagan's Evidence in Relation to Syndicate 540

The evidence to which I have just referred is one of the reasons why I also reject the account given by Mr. Fagan of his method of assessing the PML for Syndicate 540 - an area of the case that I shall turn to shortly. My inability to accept his evidence in relation to the marine syndicate reinforces the view that I have formed in respect of the reliability of his evidence in relation to the non-marine syndicates.

7) The Contemporary Documents

No contemporary documents have been disclosed which support Mr. Fagan's evidence that he was assessing PMLs by reference to wind factors in 1987 and 1988. Both he and Mr. Gofton-Salmond have referred to a large amount of missing documents. I have no doubt that they are correct, but many contemporary documents have survived and it is remarkable that none of them suggest that Mr. Fagan was carrying out wind factor calculations in these years. What have been disclosed are documents in which Mr. Fagan has carried out calculations of aggregates in which a distinction is drawn between volatility bands 2 - 6 and bands 7,8 and 9. These include comparisons between the position in 1983 and 1987 and consideration of the extent to which the aggregates in volatility bands 2 to 6 are covered by reinsurance. They support my conclusions as to Mr. Fagan's approach to exposure, both of the marine and the non-marine syndicates, which I shall develop in due course.

Only late in 1989 does one find manuscript calculations by Mr. Fagan showing wind factors for the non-marine cover and rig factors for the marine cover. I consider that Mr. Cooke is right to suggest that these were carried out retrospectively in an attempt to justify the exposure that had been left. The contemporary Minutes, to which I shall refer later, demonstrate repeated frustration on the part of the Board, and in particular Mr. Drysdale, at Mr. Fagan's failure to supply data which enabled a comparison to be made of gross exposure and reinsurance protection. This reached a head at a meeting on the 5th September 1989, when Mr. Fagan was away on holiday. The Minute records:

"Mr. J. Manning recommended that the formulation of Syndicate 540/542 Underwriting Policy for 1990 should commence immediately and that the Underwriter should seek the assistance of Messrs Andrew, Drysdale, Malim and Gofton-Salmond. The Chairman agreed to approach Mr P.F. Fagan on his return from holiday, he advised that we might need a number of different plans contingent on possible levels of stamp capacity.

Only after this meeting and, I suspect as a result of concerted pressure from his colleagues, did Mr. Fagan begin to carry out PML calculations that attributed wind or rig PML factors to the different classes of aggregates.

In answer to questions put in re-examination which were leading, in substance if not in form, Mr. Gofton-Salmond said that in the years 1987 to 1989 he had seen manuscript documents produced by Mr. Fagan of similar content and adopting a similar approach to Annexe 1. I do not consider this evidence reliable and it does not justify the conclusion that Mr. Fagan was assessing PMLs by the use of wind factors in those years.

8) Syndicate 542 in 1989

Mr. Fagan produced to the Board a draft underwriting policy for Syndicate 542 in 1989 which stated simply:

"The Non-Marine Excess of Loss account will be based on 40% of the entire account written in 1988 for both Syndicates 542 and 847.

The philosophy of 542 however, will be to write at a higher level than that which 847 intends to write, thereby targeting its writing very much towards being a major catastrophe account concentrating on Generals which have benefit of specifics."

The expected structure of the account was then set out showing premiums, rates on line and gross aggregates, but no wind factors.

This contrasts with Annexe 1. If Mr Fagan had devised a policy in accordance with Annexe 1 it is hard to see why he did not place this before the Board. I find that Mr Fagan's approach to exposure in 1989 was the same as in earlier years.

These are the reasons why I have been unable to accept Mr Fagan's evidence as to the basis upon which he effected PML calculations for the non-marine syndicates. It follows that he did not base the reinsurance that he purchased on such calculations. I shall in due course set out my findings on the approach that Mr Fagan did in fact adopt to reinsurance, but I must first turn to the evidence that he gave in relation to assessing the PML for Syndicate 540.

The Underwriting Policy for Syndicate 540 in 1987, 1988 and 1989

The disclosed documents included an analysis in Mr Fagan's hand comparing aggregate exposures and reinsurance protection for Syndicate 540 in 1987, 1988 and 1989. This document was generally referred to by its number in bundle E 2 as 418. I shall append it to this judgment as Annexe 2 and refer to it as such.

Annexe 2 applies PML factors based upon an assumed rig loss to the various categories of cover. For XL on XL the factor is 90%. For the whole account cover a different approach is adopted. Volatility bands 1-6 are adopted as aggregating for PML purposes. The PML consists of :

Bands 1 - 6 of the whole account

+ 90% of XL on XL

+ 100% of the rig covers.

It is the Defendants' case that Mr Fagan always assessed his marine PML in this way. Once again this case is largely based on evidence given by Mr Fagan in his second witness statement. The evidence in question is challenged by the Plaintiffs and I have concluded that it is untrue. Before explaining why I have reached this conclusion I shall set out the material passages from Mr Fagan's statement:

"Categories 2-6 gave me my sighting shot at the PML figure for the marine account but I would take my analysis further by looking at all of the other categories of business and in particular, rig exposure, liability exposure and war exposure. Thus, in addition to the whole accounts and XL on XL accounts, I would routinely consider the impact of a major rig or war loss on the account. Obviously in the case of a major rig loss, I could anticipate that my entire rig aggregate might be exhausted. In such a case however I had rig specific protections which would respond and my whole account reinsurances excluding LMX business would also be available to pay non-LMX claims. So far as a war loss was concerned, my view was that only 50% of my aggregate was likely to respond to such a loss (certainly the experience of the Iran-Iraq war and the Shatt Al Arab losses suggested that 50% was a reasonable estimate). I took a similar view on my liability exposure that 50% was a reasonable estimate.

It is appropriate at this point to comment on the document at E2 p.418. I understand that it is suggested that this document was prepared in late 1989 and was the first occasion on which I attempted to assess PMLs for the marine account. It is obvious that this particular document was indeed produced in late 1989 and I remember being requested by the board to produce the document. However it is not the case that the PML calculations for 1987 and 1988 shown on the document were first produced in late 1989. These calculations were transcribed onto this document from my working papers which had been prepared in 1987 and 1988 respectively. I can clearly recall turning up my working papers so as to produce these figures for the comparative exercise shown in E2 p.418.

I believe in the light of what I have said above that the only figure on E2 p.418 which calls for any additional comment is the estimated PML for the XL on XL account. Instead of restricting my XL on XL PML to categories 2-6 I have used (as recorded in E2 p.418) a figure of 90% of the 100% aggregate. This figure of 90% was my estimate of PML for the XL on XL account following upon a detailed exercise carried out by me in late 1987 which lead me to conclude (at that time) that there was little or no likelihood of all XL on XL contracts being exhausted even in a major loss and that 90% was a conservative assessment of the PML.

Once the [quarterly accumulation] sheets had been prepared, I would look at the 2-6 aggregates in order to gain a general impression of the way in which the account had performed. Thereafter, I would study the sheets in some detail in order to assess, based on my judgment and experience, which particular items of the account I thought, based on my knowledge of the reassureds and my experience of past losses, might be exposed to the type of losses I was considering. This was a time consuming exercise. I would usually take the quarterly sheets home with me to work on in the evenings or at weekends. I would look at each of the individual figures (for example, Lloyd's XL accounts in category 6) and decide for each one what percentage of that aggregate might be affected by, for example, a rig loss. Eventually, I would be able to arrive at a composite figure for the whole class of business. I would pay particular attention to the way in which aggregates in various classes and volatility bands were accumulating in order to review the development of the account."

The reasons why I am unable to accept that this evidence is truthful are similar to those that have led me to reject Mr Fagan's evidence in relation to the non-marine business.

1) Mr Fagan's Evidence is Incompatible with the Oral Evidence Given by Mr Gofton-Salmond

Mr Gofton-Salmond was strenuously cross-examined about Annexe 2. Once again he indulged in considerable equivocation, but Mr Cooke succeeded in pinning him down to agreeing:

(i) that the exercise set out in Annexe 2 was done in September or October 1989;

(ii) that prior to that he had never seen a similar document;

(iii) That in 1987 and 1988 the marine PML was arrived at by a different method from that adopted in Annexe 2. It was a "different methodology".

(iv) that Annexe 2 did not correspond with his view of the way that Mr Fagan calculated his PML in the years 1987 and 1988.

Mr Gofton-Salmond did say, however, that Annexe 2 had "some common features" with the way in which Mr Fagan calculated his PML.

In re-examination, this area of the case was introduced by the following question:

"Q.Would you please look at E2, 418? Now can you help us about this Mr. Gofton-Salmond? Mr. Fagan tells us that this is a compilation in 1989 of workings which he duly did in 1987, 1988 and 1989. Can you help us as to whether or not you saw the result, in other words his calculation result, in 1987, 1988 or 1989, or only later when it was compiled, or do you not remember?

Led in this way, Mr. Gofton-Salmond, in manifest discomfort, went so far as to say that he believed he would have seen documents which "contained the information" - "figures which were consistent with" - Annexe 2.

These answers did nothing to persuade me that Annexe 2 represented the way in which Mr. Fagan had calculated his marine PML.

As to the positive oral evidence given by Mr. Gofton-Salmond as to how Mr. Fagan calculated his marine PML, this was confused. This was largely because of confusion between considering aggregates for the purpose of assessing a PML and considering the extent to which aggregates should, or would, be covered by reinsurance. The first exercise should be an essential step towards the second. Mr. Gofton-Salmond's evidence confused the two, reflecting, I believe, a confusion in the approach that had been adopted by Mr. Fagan. On one point, Mr. Gofton-Salmond was relatively clear, and that was that consideration of volatility bands 1-6 was an important part of the exercise.

2) Mr. Fagan's Evidence is Inconsistent with the Evidence which he gave to the Neill Committee

Mr. Fagan gave evidence to the Neill Committee on the 6th and 11th September and the 15th October 1991. The questions put to him ranged widely, but they included questions about PMLs and the extent of reinsurance cover. Mr. Fagan's answers suggested that he had some difficulty distinguishing between the two. Thus, when he was first asked if a PML figure was put before the Board, he replied:

"Yes. I think we were hoping that by and large a PML figure would be 100% of capacity".

In subsequent passages of his evidence, when the Committee sought clarification, Mr Fagan confirmed that, when considering reinsurance requirements he took volatility bands 1 to 6 of the XL on XL and whole account as PML figures - or at least that "this was one way of looking at it". He said that they had always thought that all the categories up to about category 6 might go in the event of a big loss.

Much of the evidence that Mr Fagan gave to the Neill Committee related to the degree to which the syndicates were exposed. When dealing with this Mr Fagan always related exposure to the XL on XL account or to volatility bands 2-6. At no stage did he suggest that exposure was related to a PML exercise such as that set out in Annexe 2.

3) Mr Fagan's Evidence is Inconsistent with that given by Mr Gofton-Salmond to the Neill Committee

Mr Gofton-Salmond first gave evidence to the Neill Committee on 28 August 1991. He had this to say about marine PML calculations:

"Q.There is a specific question which I would like to put to you, which, hopefully, you will be able to remember as deputy on the 540, because that is where it is. Looking at the marine XL whole accounts they were coded on a range of 2-9 based on the excess point as a percentage of income. -- A. Yes.

Q.Out of this coding of 2-9 only 2-6 were put into the PML calculation and the layers, which were the top layers, of 7, 8 and 9 were left out of the PML calculation. Why was that? -- A. I do not believe that they were left out of the PML calculation. What was the understanding was that the layers 2-6 were the PML and that was the aggregate which would be looked at for buying protection against. Maybe I am being finicky about these questions.

Q.No. - A. They were not considered, they were considered as the higher layers but layers 2-6 were considered to be the PML."

He gave further evidence to the Committee on 14 October 1991:

"Q.When you got the policy document how did that differ from what had actually been happening, or was it simply a formulation in writing with figures of the practice that you and Mr. Fagan had been following, I suppose, for some six or seven years by them? -- A. Yes, it was really the formulation of the continuing practice of what had been going on prior to there actually being a policy in force. There was no material change at the time of producing a policy.

Q.Did that embody this concept of drawing a distinction between categories 2 to 6 and categories 7 to 9, on the other hand, for the purposes of reinsurance and calculation of aggregates? -- A. No, I do not believe so. To the best of my recollection, the categories 2 to 6 were perceived to be the PML of the account. It was not that categories 7, 8 and 9 were ignored as far as computing the aggregates were concerned but categories 2 to 6 were considered to be the possible - I will not say "maximum foreseeable loss" but, the probable maximum loss.

Q.I think what it means is that PML is probable -- A. Probable maximum.

Q.-- and not possible. -- A. You can debate it, yes. But the probable maximum loss would be considered to be categories 2 to 6, and that, I believe -- and I cannot be specific about this -- was the aggregate against which Mr. Fagan looked to purchase his reinsurance protection.

Q. Then they have become the same. I thought you were drawing a distinction between the PML and the aggregate for reinsurance purposes. -- A. No, I was not. What I was trying to say was that categories 2 to 6 were the PML for the purposes of the reinsurance protection.

Q.Did the Board know that? -- A. Yes.

Q.You said that categories 2 to 6 were perceived to be the PML -- questioned by whom? -- A. Yes, I think the Board were aware that was the basis upon which the PML for the account was arrived at.

Q. And, therefore, 7, 8 and 9 were not -- A. Perceived to be PML exposed.

Q. And if not PML exposed, they could -- is this unfair -- be disregarded for the purposes of reinsuring? Are they too remote? -- A. They would be considered to be PML failure which was not something which a catastrophe syndicate could protect against.

He went on to elaborate that only those categories of risk in bands 2 to 6 which clashed were taken as the PML. This evidence cannot be reconciled with Annexe 2 and, indeed, Mr. Gofton-Salmond suggested that in 1989 the practice was changed to one which did accord with Annexe 2. If the policy was changed, this was prospective. I am satisfied that the practice adopted in 1987, 1988 and 1989 was not that depicted in Annexe 2.

4) Mr. Fagan's Evidence does not Accord with the Contemporary Documents

The contemporary documents give no hint of PML calculations of the type depicted in Annexe 2, before the date when that document was itself prepared. There are, however, just as in the case of the non-marine business, manuscript documents in the hand of Mr. Fagan showing the extent to which volatility bands 2-6, and 2-4 were covered by reinsurance in 1987 as compared with 1983.

5) Technical Implausibility

I have already referred to the difficulty that I and Mr. Outhwaite had in making sense of the description given by Mr. Fagan of how he arrived at his wind factors when assessing the non-marine PML. His description of how he arrived at his marine PML factors gives rise to even greater difficulty. Neither Mr. Emney nor Mr. Outhwaite were able to understand the process that he described.

6) Mr. Fagan's Evidence in Relation to Syndicate 542

Just as the shortcomings in Mr. Fagan's evidence in relation to syndicate 540 throw doubt on the reliability of his evidence in relation to Syndicates 542 and 847, so the shortcomings of his evidence in relation to those syndicates throws doubt on that which relates to the marine syndicate.

For the reasons which I have given at some length, I have rejected the Defendants' case as to the PML calculations carried out by Mr. Fagan. It does not follow from this, however, that he paid no regard to aggregates when deciding how much reinsurance to buy. For this reason it is necessary to consider two additional aspects in which the Plaintiffs have attacked Mr. Fagan's approach to aggregates.

Earthquake Risk

It is Mr. Fagan's evidence that, when considering the non-marine exposure against which he had to reinsure, he did not treat the possibility of an earthquake as constituting the most serious risk for which he had to make provision.

In his third witness statement Mr. Fagan explained that he considered that in the 1980s a major Californian earthquake would not result in such a high level of claims on the LMX market as a major windstorm. The Defendants rely on this evidence and submit that the underwriters were right not to regard an earthquake as posing the most serious risk. In this respect they are supported by the evidence of Mr. Outhwaite. I asked him whether it was a general view in the market that windstorm posed a greater PML threat than an earthquake. He answered:

"In terms of reinsurance losses into London, I would say that probably was the case. I cannot say it was universally the case. Again, it would rather depend on the book of business that an individual would write, particularly if writing US catastrophe business, but in practice it certainly proved the case that a large earthquake in California has caused a very small reinsurance loss in London. The two earthquakes that have occurred in recent times, and one is, in fact, I believe, the second largest worldwide catastrophe ever, has actually caused virtually no losses into London, or very small losses.

I treat Mr. Outhwaite's tentative view on this question with scepticism. Mr. Thompson has collated a huge amount of contemporary articles and addresses on excess of loss underwriting and these tend to treat an earthquake as posing the most severe threat. I quote by way of example from a paper of Mr. Emney:

"Should the dreaded Californian earthquake occur, and the estimates of the likely cost of that event vary from $10 billion to $70 billion, depending on which expert you talk to..."

I treat with greater scepticism Mr Fagan's statement that he thought that a major windstorm would give rise to greater losses than a major earthquake. This was certainly not the view of Mr Gofton-Salmond. He stated in his half-yearly report for 1989:

"undoubtedly a large American earthquake would produce the largest loss to the Syndicate"

The Board Minutes contain entries which indicate that earthquakes were perceived as threatening the most severe losses and there is no contemporary document which gives a counter indication.

It may be that the losses which have resulted from the earthquake that in fact occurred in San Francisco in October 1989 lead to the conclusion, with hindsight, that earthquake exposure was less significant than windstorm exposure - though I do not believe that this case is made out. But at the time I do not believe that Mr Fagan had any reason to form this view and I doubt whether he did. Having regard to the findings that I have made and those that I will be making, this is not a matter of significance.

Exceeding Target Aggregates

The underwriting policies for 1988 and 1989 set targets for the gross aggregates for those years. The evidence, both documentary and oral, gave rise to considerable confusion as to whether these targets were for written or for in force aggregates, and this may well have reflected a degree of confusion that prevailed at the time. It is common ground that the underwriters and the Board should have monitored in force aggregates and in force reinsurance cover as each year progressed in order to ensure that Names were never over exposed. In the first half of 1989 records were not available to enable in force aggregates to be monitored. This was less than competent, but I do not think that it has causative significance. By the end of August each year written and in force aggregates were virtually the same, and that is the appropriate stage for considering the extent to which targets were achieved. The significance of these targets was that the size of the aggregates governed the extent to which the Names were exposed. If the targets were exceeded, and additional reinsurance was not purchased, the exposure would be increased. In both 1988 and 1989 the aggregate targets were exceeded for all three syndicates. I need not refer to the figures for Syndicate 847 in 1988, for no claim is made in relation to that year. For the other syndicate years for which Mr Fagan was responsible, the figures are as follows:

In the 1988 year:

Syndicate No: Class of Business Target Aggregates Actual Aggregates

Syndicate 540 XL $100 million $110 million

Whole Account $125 million $157 million

Rig Account $9 million $15.2 million

Liability Account $16 million $18 million

War Aggregate $30 million $38 million

Back-ups $20 million $26.2 million

Syndicate 542 Generals $8.75 million $12 million

LMX/RIA Specific $10.5 million $12 million

US Retro $4.4 million $6.4 million

In the 1989 year:

Syndicate No: Class of Business Anticipated Actual Aggregates


Syndicate 540 XL $50/66 million $88.7 million

Whole Account $120 million $164.7 million

Rig Account $5 million $7.5 million

It is the Plaintiffs' case that the underwriters' failure to keep to their aggregate targets was a further manifestation of incompetence. The Defendants contend that target aggregates were not limits which were set in stone, but had to be approached flexibly. As the underwriting year progressed, circumstances might call for departure from the targets, and there are sound commercial explanations for the fact that the underwriters exceeded their targets.

A competent underwriter when planning his next year's policy will not merely plan his gross exposure, involving target aggregates, and the extent to which he will cover this by reinsurance, but he will also project the premium that he anticipates he will receive and the premium he intends to spend on reinsurance. Should rates alter from those that he has assumed when making his plan, the competent underwriter will consider the effect of the change and adjust his plans accordingly. In particular if, as it is common ground he should, the underwriter has a policy as to the extent to which he will expose his Names, he will have to have regard to the restraint that that policy imposes, or decide whether the circumstances justify a departure from that policy. Mr Emney said that, in principle, he would expect an underwriter who went beyond his aggregate targets to make a commensurate increase of his reinsurance cover. Mr Fryer said that an underwriter who exceeded target aggregates without first getting the agreement of the Board to this would not be acting in accordance with good underwriting principles.

It is clear from the contemporary documents that the Board had well in mind the implications of exceeding aggregate targets. Thus, by way of example, on 2 March 1988 a Board minute records:

"The Chairman stated that as a Board we must monitor the aggregates, if we reached them we must then decide if we could increase our reinsurance programme or accept restrictions."

It is submitted on behalf of the Defendants that Mr Fagan exceeded his aggregate targets because his premium income fell below that which he anticipated. I find this hard to reconcile with reports made by Mr Fagan to the Board in 1988 that the proportion of premium spent on reinsurance was down because of increase in premium.

Had Mr Fagan been carrying out PML calculations, he would have been in a position to make a quantative assessment of the significance of exceeding aggregate targets. As it is, he was aware of the obvious fact that an increase in aggregates would result in an increase in exposure. I am about to turn to Mr Fagan's approach to exposure. As I shall demonstrate, Mr Fagan was not in a position where he could afford to exceed his aggregates. By so doing he made a highly unsatisfactory position even worse.

What Approach did Mr Fagan Adopt to Exposure?

I have rejected the Defendants's case that Mr Fagan carried out PML calculations on the basis of wind or rig PML factors in order to assess the exposure against which he needed to purchase reinsurance cover. It does not follow from this that he had no policy in relation to exposure. I now turn to consider the evidence in relation to the approach of Mr Fagan to the extent of the syndicates' exposure.

It is the Plaintiffs' case that this evidence demonstrates that Mr Fagan had no policy as to the extent to which exposure should be reduced by reinsurance but that he adopted the pragmatic approach of spending as much of his premium income on reinsurance as he thought he could afford.

It is the Defendants' case that Mr Fagan's paramount policy was to ensure that a single loss event could not expose a syndicate to an unreinsured loss that exceeded 100% of stamp capacity. He would, however, buy reinsurance that would reduce exposure below this if it was available on attractive terms.

The Origin of Mr. Fagan's Approach to Reinsurance

Mr Gofton-Salmond confirmed Mr Fagan's evidence that he inherited his approach to reinsurance from Mr Davies. Mr Fagan described this as "a method of categorising risks according to their perceived remoteness from loss." It seems clear that the practice of grading risks in volatility bands dates back to Mr Davies. How he made use of these bands is less clear. When giving evidence to the Neill Committee Mr Gofton-Salmond said of the syndicates business:

"In the 70's and 80's the basis behind it, I believe, was that there was an arbitrage on the grounds that they are writing retrocession business and are actually able to buy cheaper reinsurance than that which they are writing it at.

In a less than satisfactory passage of cross-examination Mr. Gofton-Salmond sought to explain that this comment was not intended to convey the meaning which the words clearly bear, but referred, at least in part, to the fact that in the 70s and 80s it was possible to place a large part of the outward business overseas, so that it did not come back into the spiral.

Financial Constraints

A policy of restricting the amount of premium income spent on reinsurance is perhaps the one consistent theme that one finds in the contemporary documents. In November 1986 a schedule prepared by Mr. Fagan showed that reinsurance costs for the three syndicates amounted to a little less that 50% of gross premium income. This formed the basis for his expenditure on reinsurance in 1987. In November 1987 Mr. Fagan informed the Board that his intention was to target reinsurance costs at 50% or lower for the 1988 underwriting account, and this formed part of the underwriting policy for all three syndicates in 1988.

In July 1988 Mr. Fagan reported that the overall reinsurance costs for 1988 were down to 45%, due in part to increased premium income. In August he reported that these costs were further reduced to 43% of premium income. In October 1988 Mr. Fagan reported that he expected the cost of whole account cover for syndicates 540 and 542 to increase in 1989 but that he would try to keep his reinsurance costs to within 50% of premium income.

In his evidence to the Neill Committee, Mr. Fagan said that he could not afford to pay more than 55% to 60% of premiums on reinsurance and his policy was that they should not pay more than that. Constraints of reinsurance cost prevented him from covering volatility bands 7 to 9.

He said that his general philosophy with regard to the design of the reinsurance programme "was obviously to get as much cover as we could for the money that we could afford to spend".

In his second witness statement he defended this philosophy:

"of course the programme was based upon buying as much cover as we could for the money that we could afford to spend. Any underwriter's reinsurance programme would be, and rightly so."

He went on to add that this statement should not be taken out of context and did not imply that cost was the only or the dominant factor.

The Extent of Unreinsured Exposure

Mr Fagan made the following statements to the Neill Committee as to the extent of the reinsurance cover that he purchased:

- He had told the Board that there was unlikely to be a loss from the non-marine aggregate, assuming that the whole account cover was not impaired from the marine side.

- He had told the Board that if the loss was big enough the Names could lose as much as 100% of stamp capacity.

- The marine excess of loss programme covered the whole of the XL on XL category.

-He concentrated on covering the XL on XL account, the total aggregate.

-One way of looking at it was that bands 2-6 of XL on XL and whole account were taken as the PML figure for reinsurance purposes, but in later years, when more aggregates moved into the 2-6 columns, he looked at the 2,3,4 categories much more.

- They had indicated to the members' agents that possibly on a worst case basis a major loss would probably produce a 100% loss to names, or thereabouts.

In his O14 Affidavit Mr Fagan deposed:

" relation to XL on XL business the entire account (ie each of the categories 2 to 9) was 100% protected as it was understood that, though more remote than the lower layers, the higher layers of LMX business would be exposed in the event of a major catastrophe. The PML figures given as examples in the Neill Report demonstrated that we always took the 100% aggregates over all categories 2 to 9 and applied a PML to each part of the account and a 100% expected loss (PML) for XL on XL business."

In his first witness Statement Mr Fagan said:

" relation to XL on XL business, a benchmark of 100% of the aggregate was used as a starting point in the planning of the reinsurance programme so that the entire account (i.e. each of the categories 2 to 9) was 100% covered...

In his second witness statement Mr. Fagan was at pains to explain that he had erroneously been understood to assess his XL on XL account as 100% for PML purposes - in his Affidavit he had expressly stated that this was the case, but this was a mistake on his part. The true position was that he assessed XL on XL at 90% for PML purposes, but always ensured that his reinsurance protection covered 100% of his XL on XL aggregate.

Later he emphasised that he planned the reinsurance cover so as to try to limit the potential loss on a PML basis to the names to no more than about 100% of stamp. If asked, he would always tell agents that there was potential for a 100% loss.

In his third witness statement, Mr. Fagan said:

"As with the marine syndicate, I sought to limit the loss on the non-marine syndicate on a PML basis by reference to 100% stamp capacity. I should point out that the figure of 100% was merely a figure used for calculation purposes to limit the maximum size of the gap. It did not take account of any net premiums which might be available to pay claims.

In his first witness statement Mr. Gofton-Salmond said:

"As far as I am aware, when determining how much vertical exposure should be written by Syndicate 540, Patrick would aim to ensure that the vertical reinsurance protection that he planned to purchase would cover 100% of the aggregate value of the "XL on XL" contracts written...

As well as aiming to ensure that the vertical reinsurance protection would cover 100% of the aggregate value of the XL on XL contracts written by Syndicate 540, Patrick also reviewed all of the accounts that he proposed to write on behalf of Syndicate 540 (including the XL on XL account) and attempted to identify the maximum number of risks which would clash; i.e. all risks that would be triggered by the worst probable loss event. Patrick then attempted to ensure that his vertical reinsurance protection would cover at least such of those clashing risks as fell within "volatility" bands 2-6 on the basis that it was his belief, based on prior experience, that risks within categories 7-9 were more remote. The risks falling within categories 2-6 usually represented about 60% of the clashing aggregates of the Marine account. However, if he found that categories 2-6 would actually represent less than 60% of the clashing aggregate, he would adjust the underwriting policy and/or reinsurance programme so that 60% of the clashing aggregate would be protected...

When planning his 1989 reinsurance programme, Patrick's overall objectives were the same as they had been in 1987 and 1988 (i.e. to ensure that:

(a)the programme covered 100% of the XL on XL aggregate;

(b)the programme covered all clashing aggregates falling within categories 2-6; and

(c)the programme was sufficient to cover 60% of the clashing aggregates).

In his second witness statement Mr. Gofton-Salmond sought to correct what he said were errors or misunderstandings in respect of his evidence to the Neill Committee and in his first witness statement:

"To the extent that, in my statement, I said at paragraph 206 that Patrick attempted to ensure that his vertical reinsurance protection would cover such of those clashing risks as fell within "volatility" bands 2-6, that is incorrect.

I did in fact know that Patrick planned his reinsurance protection in relation to his PML exposure by purchasing reinsurance cover to a level which when added to the 100% stamp capacity would cover his PML.

He referred to a report prepared by Ernst & Young in October 1990 as giving an adequate and concise summary of the basis upon which the account was underwritten:

"Underwriting Philosophy

The underwriting philosophy of the syndicate was to write a broadly based account including high layers that were intended to be in excess of the expected losses of the reinsured. Records are kept by book of business and rank all risks accepted by reference to the excess point of the layer expressed as a percentage of the insured's estimated premium income for the relevant class. Category 2 being "lowest" layers and category 9 being "high". The percentage involved varied from class to class... Losses up to and including category 6 were considered to form the basis of the PML calculation and the reinsurance arrangements were put in place on that basis.

The intention of the underwriter is currently to ensure that no name is exposed to an unreinsured loss in excess of 100% of their allocated stamp capacity for any type of catastrophe. Loss is defined as the aggregate PML net of available reinsurance protection."

He had never intended to say that Mr Fagan set out to buy reinsurance that covered his volatility bands 1-6. Merely that he used those bands as a basis for his reinsurance protection. Codes 2-6 were always "the broad categorisation which was regarded as the PML for the marine syndicate". The Board "fully understood that the syndicate carried a significant unreinsured exposure in the event of a PML loss, in the region of 100% or possibly greater".

In the course of his oral evidence Mr Gofton-Salmond said on a number of occasions that Mr Fagan deliberately exposed his Names to a loss of 100% of stamp capacity in the event of there being a major catastrophe.

Other Evidence Given to the Neill Committee

Mr Malim, Mr Fagan's deputy underwriter, gave evidence on 11 October 1991. His evidence was that he did not discuss PML's with Mr Fagan before it became apparent that the Names were exposed to serious loss in respect of Piper Alpha. His understanding was that PMLs were based upon the volatility factors with the high bands being left out of account because it was perceived that there would not be a claim at that level. So far as reinsurance is concerned, he described the position as follows:

"The general view was that reinsurance should exceed the excess of loss account aggregate and, without overspending, protect as much of the whole account, including excess of loss account, aggregate, as it was possible to purchase, on the understanding that it was not probably possible to cover the whole of one's aggregate across the book."

Mr Manning gave evidence on 5 September 1991. He said that he was not aware of any Board policy on net exposure before the quantification of the exposure in the underwriting Policy for Syndicate 847 - and that was not a Board policy.

Mr Andrew, giving his evidence the same day, stated that the Board was aware of the volatility bands and that the underwriters used them in calculating their PMLs, but was unable to be precise as to how they did so. He said that there was a policy of the Board with regard to running net amounts of exposure and that agents would know that Names were exposed to an unreinsured loss of 100% of stamp.


Plainly Mr Fagan purchased reinsurance cover in order to restrict his syndicates' exposure to loss. The evidence is unclear as to the principles, if any, by which he did so. On balance of probabilities I make the following findings:

1) Mr Fagan planned his reinsurance on the basis that he would restrict his expenditure to 50% of gross premium income, subject to any additional expenditure that might be incurred in the event of having to reinstate cover.

2) Mr Fagan considered his aggregates in an attempt to satisfy himself that they were adequately covered by reinsurance.

3) Mr Fagan derived comfort from the fact that he considered that his XL on XL aggregates were fully covered by reinsurance.

4) Mr Fagan also derived comfort from the extent to which the aggregates in the lower volatility bands were covered by reinsurance.

5) Mr Fagan believed that only the most severe catastrophe risked exposing his syndicates beyond the limit of their reinsurance cover and that such exposure was in any event limited to 100% of stamp.

I now turn to the respects in which Mr Fagan's approach to exposure is open to criticism.

Reliance on Past Experience

It was Mr Fagan's evidence that his underwriting practices were those he inherited from Mr Davies. I have already described the confusion engendered by Mr Gofton-Salmond's evidence as to Mr Davies approach to excess of loss underwriting, and it is not clear how far Mr Fagan's practices were developed from those adopted by Mr Davies. It is possible that Mr Fagan was, to an extent, proceeding on a basis of following past precedent - concluding that an approach which had worked successfully in past years was likely to continue to do so. That is not a satisfactory basis on which to conduct the business of catastrophe reinsurance.

Counsel for the Defendants have emphasised more than once in their final submissions the "high degree of stability" in Feltrim's book of business. I do not find the description appropriate. The LMX market was growing rapidly during the years after which Mr Davies resigned.

A passage in Mr Emney's Report gives one excellent summary of this phenomenon:

"Whilst the 'spiral effect' has been appreciated for many years, it only began to have a significant impact from perhaps the mid-seventies; from that time onwards it grew rapidly to a peak in the late eighties. (Whilst the notorious hurricane "Betsy" in 1965 produced some 'spiral effect' within the LMX market of the day, that market in itself was so tiny when compared with that which existed in the eighties that a comparison would be valueless.) The underlying reason for that stems undoubtedly from the rapid expansion in Lloyd's capacity which occurred at that same time. That expansion coincided with the beginning of a marked decline in the volume of traditional business coming into Lloyd's. That decline was brought about by the increasing strength of Lloyd's overseas competitors, coupled with the rise in 'nationalism' in many countries. That latter factor resulted in a reduced volume of business coming into Lloyd's by way of reinsurance of local companies rather than directly as in the past. That decline affected not only Lloyd's but also the company market which had traditionally operated alongside Lloyd's (colloquially known as the 'fringe' market), including of course the marine companies belonging to the Institute of London Underwriters. It was not only Lloyd's that was expanding at that time; the late seventies and early eighties saw a considerable expansion in the 'fringe' market, due principally to the establishment of many U.K. subsidiary operations of overseas insurers. The upshot of all that resulted in many many more participants frequently with considerably increased capacity, chasing less and less traditional business.

That was markedly so in the Marine sector, although it was offset to some extent by the enormous growth in the 'Drilling Rig/Energy' market occasioned by the development of the North Sea oil and gas fields. That development required the installation of massive production platforms, whose values simply dwarfed anything previously known in that area. By the eighties values of $1 billion for a single risk became commonplace, with the major platforms commanding values of at least double that amount. That situation in itself produced a massive increase in the demand for 'LMX' reinsurance as the primary underwriters sought increasingly to lay off the enormous liabilities they were now expected to assume.

Notwithstanding the spectacular growth in that one area, nevertheless there were still many syndicates and companies, whose capacity was far greater than the volume of available original business. That situation was exacerbated by the mere presence of over-capacity, which had the inevitable effect of driving down the original premium levels.

In order to fill the ever-widening gap between capacity and income, many syndicates and companies looked for new areas of business; the one that was most readily to hand was the 'LMX' sector...."

While the evidence suggests that Feltrim were not writing directly for new entrants to the market, Feltrim's aggregates were growing because those with whom they did business were expanding their layers. New business in one form or another must have been coming to Feltrim indirectly. At the beginning of this Judgment I outlined the growth of the Feltrim syndicates from 1983. The marine market was writing an increasing amount of non-marine business. Mr Fagan's approach cannot be justified on the basis that he was following reliable precedent.

Limit on Expenditure

I do not find it negligent per se that Mr Fagan budgetted to spend not more than 50% of his premium income on reinsurance. On the contrary, any competent underwriter would need to decide the maximum that he could afford to spend if the account were to remain viable. What is not satisfactory is Mr Fagan's statement that his programme was based on buying as much cover as he could for the money that he could afford to spend. The correct approach should have been to work out, on the basis of PML calculations, how much reinsurance he needed and then to calculate whether he could afford to buy it. If he could not, he should have concluded that his programme was not viable and attempted to redesign it accordingly.

Covering the XL on XL Aggregates

Mr Fagan could properly have concluded that if his reinsurance did not cover 100% of his XL on XL aggregates, it would not be adequate. To this extent it made sense to check that those aggregates were fully covered. Where Mr Fagan erred was in assuming that if he covered his XL on XL aggregates he was likely to have adequate reinsurance protection. This was something that he could not assess in the absence of PML calculations.

Reliance on the Volatility Bands

Some of the evidence given to the Neill Committee by Mr Fagan and Mr Gofton-Salmond suggests that Mr Fagan treated volatility bands 2-6 as constituting his PML. While it seems clear that Mr Fagan used the lower volatility bands as some measure of the exposure for which he had to cater by reinsurance, I am not persuaded that he did anything as precise as treating bands 2-6 as a PML.

The volatility bands, broadly speaking, banded risks according to the perceived likelihood of a catastrophe occurring of sufficient severity to result in claims. The lower bands were considered to present a greater risk to the syndicates than the higher bands. Mr Fagan appears to have considered that the extent to which the aggregates in the lower bands were covered was a valid guide to the overall adequacy of his reinsurance cover. I do not believe, however, that he applied any specific formula to the volatility bands in order to decide upon the amount of reinsurance that he needed to buy. When giving evidence to the Neill Committee on this topic, he gave some answers to which I attach significance:

"Q.Am I right in saying that on the XL on XL 2 to 6 would have been taken for the PML and also on the whole account, 2 to 6? Is that a correct statement? -- A. That would have been true for...

Q.We are talking about 88 and that account was written in 88. -- A. As the premiums went up or the accounts expanded the way we ran the aggregates everything tended to move to the left. We had the columns 2, 3, 4, 6, 7, 8 and 9. There was a tendency all the time for the aggregates to move to the left.

Q. More came into the 2 to 6. -- A. Yes, more came into the 2 to 6 category. I think in later years I was looking at the 2, 3 and 4 categories much more."

Mr Fagan was there speaking of Syndicate 540 and his answers seem to accord with certain contemporary exposure calculations for 1987 that have survived. They are in Mr Fagan's handwriting and I append them to the Judgment as Annexes 3(a) and 3(b). They show that there had indeed been a shift of aggregates to the left and that in 1987 the percentage of bands 2-4 of the marine risks that were covered by reinsurance was almost precisely the same as the percentage of bands 2-6 that were covered in 1983.

This evidence suggests that, while Mr Fagan gave consideration to the extent to which his marine volatility bands were covered, he only did so in the most general fashion and was not treating bands 2 to 6 as representing his PML.

I find that Mr Fagan adopted a similar approach to the exposure resulting from his non-marine underwriting as he did in the case of his marine underwriting. When Mr Gofton-Salmond produced his calculations of PMLs based on wind factors, Mr Fagan expressed interest and was content that they should be presented to the Board as the planning documents providing the aggregate targets for 1988, but did not use the windstorm PML as the basis for his purchase of reinsurance.

The Defence experts sought to persuade me that considering the lower volatility bands was an exercise that was at least relevant to the assessment of a PML. In my judgment any relevance was marginal. The risks which fall into the lower volatility bands are, prima facie, the risks that are more likely to come to pass. Those in the higher bands are more remote. They are nonetheless risks that the reassured has considered sufficiently real to justify the purchase of reinsurance cover. The more ordinary catastrophe may result in claims in the lower volatility bands and not the higher bands. The ordinary catastrophe is not, however, the catastrophe that properly forms the basis of the PML. It is the remote catastrophe, which is nonetheless a practical possibility, which forms the basis of the PML. The effects of such a catastrophe will not be restricted to the lower bands.

In so far as Mr Fagan was treating the aggregates in the lower bands as an indication of the degree of exposure for which he had to make provision he displayed a lack of competence. He was no more competent if he treated his XL on XL aggregates alone as a measure of the exposure that he had to protect by reinsurance.


It is the Defendants' case that Mr Fagan followed a policy of restricting net exposure to 100% of stamp. Mr Fagan was unaware of his gross exposure and thus not in any position to follow such a policy. The underwriting policy for 1988 recorded:

"Aggregates will be maintained by class of business as above and further broken down to relate US Exposure and Quake Exposure by zones. It is NOT intended that the Syndicate will purchase Re-insurance to cover its total aggregates, but will cover the maximum probable loss potential from US Wind."

Mr Fagan told the Neill Committee that he believed that the addition of the whole account cover to the specific cover was likely to provide full protection against a non-marine risk.

So far as non-marine exposure was concerned, Mr Fagan's evidence to the Neill Committee was that there might be a loss of as much as 100% if the loss was big enough. The Defendants have adduced under the Civil Evidence Act the following report by an agent of a comment by Mr Fagan in 1984:

" always had to be borne in mind that these syndicates were writing catastrophe type business on which it would be possible to envisage losses of 100% on lines written."

This contrasts with reports from other agents of Mr Fagan advising that exposure was less than this.

I find that Mr Fagan appreciated that his reinsurance cover would not be adequate in the event of the most severe catastrophe that was a practical possibility. He considered, however, that his cover would be adequate to cope with catastrophes of the severity of those with which these Actions are concerned and that, at the worst, exposure would not exceed about 100% of stamp. He had no valid basis for these assumptions, and they were unsound. He had failed to take the steps that a competent underwriter should have taken to ensure that his Names were not exposed to greater losses than they could reasonably be expected to bear. His shortcomings were compounded by his exceeding his target aggregates, which he should not have contemplated had he appreciated the true extent of his exposure. They were further compounded by a policy of jointly reinsuring marine and non-marine syndicates, which the Plaintiffs allege was a further manifestation of incompetence.

Joint Reinsurance

It was a feature of Mr Fagan's reinsurance programme that syndicates should share whole account cover. In 1987 and 1988 whole account cover was shared by all three syndicates. In 1989 it was shared by syndicates 540 and 542. It is not unusual for a marine syndicate and its incidental non-marine syndicate to share the same reinsurance cover. What was unusual was for a marine syndicate to share cover with an independent non-marine syndicate. Because of the possible conflict of interest that this involved, the approval of Lloyd's had to be obtained. In writing to Lloyd's Regulatory Service Group on 4 March 1987 Mr Fagan pointed out that if Syndicate 847 were independently reinsured, the programme would have to be significantly larger. He stated that this would increase the cost of the programme by 50% for both syndicates. He continued:

"This would not only render the current operation of these Syndicates impractical on grounds of Reinsurance Cost; but would place ourselves and other Lloyd's Syndicates at a very severe competitive disadvantage to the Company Market, where this type of Reinsurance Programme is used extensively to cover more than one class of Business written in the Excess of Loss Market. I and my predecessor spent twenty years in the Company Market writing this business before coming into Lloyd's, and therefore can appreciate the advantages of this 'structured' programme to the account protected."

Both Mr Emney and Mr Outhwaite challenged the statement that a common insurance programme of this type was used extensively in the Company Market. Certainly the programme was unusual in the Lloyd's market, and Mr Fagan should have given careful consideration to its implications.

Those implications were that, if a catastrophe led to claims on the non-marine syndicates that exhausted the specific non-marine cover, and at the same time led to claims on the marine syndicate, the whole account cover would be called on by both the marine and the non-marine syndicates. This is precisely what occurred, with particularly disastrous effect in the case of Windstorms 87J and Hurricane Hugo, though in the latter case joint cover was restricted to Syndicates 540 and 542. As at 30 June 1994 incurred losses from these catastrophes were as follows:

540 542 847

£52.3m £9.5m £39.6m 87J

$165.5m $93.8m [$208m] Hurricane Hugo

Mr Fagan had not catered for the possibility that a single catastrophe would impact both non-marine and marine syndicates in this way - a phenomenon loosely and misleadingly described as cross-over. The Plaintiffs say that he should have foreseen this possibility and designed his reinsurance cover to cater for it. The Defendants contend that the problem caused by cross-over was not reasonably foreseeable and that Mr Fagan was not at fault in failing to cater for it. In their final written submissions they put forward figures which they suggest show that Mr Fagan had made substantial provision for cross-over and that the 87J losses occurred because he failed "fully to appreciate the extent to which a non-marine loss might cross-over into the marine market". A footnote draws attention to the fact that the actual losses were all below, and in some cases significantly below, what were alleged to be Mr Fagan's PML figures. It seems to me that the footnote makes the point of just how inadequate Mr Fagan's reinsurance cover was and, in particular, the devastating effect of his failure to make provision for cross-over.

The figures that I have just quoted demonstrate that Syndicate 540 had written, directly or indirectly, very considerable aggregate cover for non-marine risks. If the Defendants are justified in their contention that Mr Fagan could not reasonably have been expected to cater for cross-over, this can only be on the basis that he could not reasonably have been expected to appreciate the nature of the business that he was writing for Syndicate 540.

The amount of non-marine business written by Syndicate 540 reflected a market trend. Marine rates had slumped and marine underwriters were keen to write non-marine business, either under their incidental non-marine facility or as an element in whole account cover. The Defence experts all gave evidence to like effect: the fact that non-marine business was being written in the marine market was appreciated; what was not generally appreciated was the extent of this. Mr Emney commented in his report:

" the light of the marked contraction of the direct marine market, the majority of direct marine underwriters within Lloyd's were tempted to, and did, join in the LMX market in order to utilise more fully their premium income capacity and in order to take advantage of the perceived profitability of this type of account for the benefit of their Names. Not only did these underwriters enter the marine LMX market in force but in the later 1980's marine syndicates wrote increasing quantities of non-marine LMX business.

During the 1980's I was writing substantial amounts of LMX business, firstly at Merrett and thereafter at Charter Re. Whilst I was aware of the increase in the marine LMX market, I did not fully appreciate towards the end of that decade the increasing extent to which marine underwriters were taking on greater quantities of non-marine LMX business. It is probably true to say the entire LMX market underestimated the extent of that latter factor.

.....what was not recognised, partly because it had not happened before, was the extent to which that phenomenon had increased in a very short space of time in the late eighties...

This echoed the following passages in Mr. Fagan's affidavit:

"....most underwriters in the mainstream of the marine catastrophe LMX market, like myself, under-estimated the impact of one market on the other. It is certainly the case that whilst we were aware of marine/non-marine cross-over to a limited extent based on past experience we did not fully appreciate how great it had become at the time in question."

He went on to explain that because Feltrim tended to lead risks, they were unaware of the nature of the syndicates who were following on the slip.

I do not find the plea that Feltrim were one of the leaders in the LMX market a convincing explanation of their ignorance as to what was taking place in that market. Mr Emney commented in his Report:

"The total number of recognised major underwriters in the LMX market was no more than 10-20 at Lloyd's and much the same number in the companies market. Being such a small and tightknit community, all the major participants tended to know a great deal about the business each of them underwrote because they were all writing the same slips either as reinsurances of each other or as reinsurances of other members of the London reinsurance community."

This mirrored a passage in the evidence of Mr Gofton-Salmond:

"This is a very small market. There are not that many players. People who are writing XL business, or the core market of XL underwriters is a small market and it is known, well known within the market, as to who is doing what within the market place."

Mr Thompson expressed the view that:

"A competent underwriter in 1987, 1988 and 1989 would not have put in place a reinsurance programme the adequacy of which was premised upon there being no risk of substantial accumulation between marine and non-marine excess of loss reinsurance exposures."

Counsel for the Defendants condemn this as hindsight, but there is considerable evidence that the crossover phenomenon was appreciated as being significant at the time. In a paper delivered in 1988 Mr Emney warned of the danger of catastrophe reinsurance being placed as part of whole account cover. He observed:

"Whilst talk of catastrophe insurance inevitably concentrates upon the non-marine field, it should not be forgotten that there exists equal potential for this type of loss in two other markets...

With the volume of non-marine business now being written in the Lloyd's marine market, it would probably be more appropriate to describe its inhabitants as incidental marine syndicates."

I suspect that Mr Outhwaite also had a clear perception of the extent to which non-marine excess of loss underwriting had invaded the marine market. In a paper that he delivered in April 1988 Mr Outhwaite referred to:

"Many examples where non-marine XL is being included with marine exposure".

When asked in evidence about the general perception of this in the market in the years 1987, 1988 and 1989 he said:

"Well, my general perception is that different underwriters will, of course, have different appreciations of this. I myself wrote quite a significant book of Non-Marine business and my appreciation may have been somewhat different to other people's, but the general appreciation was that although there was obviously some crossover and there was, therefore, some risk of substantial Non-Marine catastrophes impinging on the Marine market, that was not regarded as the main problem of the Marine market; in other words, it was not likely to be a very large loss that would develop from that source.

Subsequently Mr. Outhwaite made the point that the marine underwriter, when assessing his PML, might disregard the risk of cross-over on the basis that he would not expect his biggest loss to come from the non-marine cross-over, but from something like a drilling rig catastrophe in the North Sea. He accepted, however, that if you were writing a joint insurance programme you would have to have regard to cross-over when assessing your PML.

Mr. Gofton-Salmond was cross-examined at some length about Feltrim's approach to cross-over, and made some significant concessions:

"Q. Now the LMX market was a fairly small community where people knew one another, as I think you have told us?


Q. And I think you told us that it was known what type of business people were writing?

A. Yes.

Q. So was it not fairly obvious that Marine syndicates were writing Non-Marine excess of loss covers and this effect would occur?

A. It was well known there was quite a high degree of Non-Marine business written in the Marine market.

Q. If by the very nature of tertiary reinsurance you do not know the details of the underlying covers, then, of course, you cannot know in any real sense what the Non-Marine content of the covers are which you are writing in the Marine market, can you?

A.I think you will get a feel for it from the amount of income which is Non-Marine, which would be something which would be declared at the time.

Q. If you have a combined reinsurance programme of this type, would you agree the right thing to do is to have a combined PML?

A. In theory, yes.

Q. Well, not just in theory; it must be the right thing to do, must it not? If you have a joint reinsurance programme, you want to know what the joint exposure is, do you not?

A. In practice, it is very difficult to determine, just as we have been discussing this afternoon, how much of the loss is going to spill over into the Marine account, and therefore what the impact of the Non-Marine loss in the Marine XL market would be.

Q. So you would need details of the accounts written to decide if you could eliminate or discount from the 100 per cent aggregates that appear both on the Marine and Non-Marine side, would you not?


Mr Johnson asked him to elaborate on these answers in re-examination:

"Q .....look at the position at the end of, or towards the end of 1986, when you are looking towards the 1987 underwriting year. Now, at that time did you envisage the possibility of crossover between the Non-Marine and the marine markets?

A. Yes.

Q.Did you consider at all how likely or unlikely that spill-over might be and the extent of any spill-over?

A. We considered it, yes. I do not think we came to any finite conclusions.

Q. Who did you consider it with?

A. With fellow underwriters within the market. It was obviously an area which was of concern.

Q. Well, what was your understanding at the time, and again I am looking towards the end of 1986, looking ahead to the 1987 underwriting year?

A. I think, trying to put myself back into the time, to the end of 1986, I believe that at that time I would have considered that it was a feature of the marine market that was not, especially at that time, of immediate concern.

Q. Why was that?

A. I think at that stage we were able to look at what was happening with Alicia within the marine market. The loss was now, sort of, two and a half to three years matured. It was not causing - the Alicia loss was not causing a great deal of crossover in the marine market at that stage. I think it was still considered that there was a differential between the two markets."

Ultimately the question must be whether past experience, and in particular the claims experience in relation to Alicia, justified Mr Fagan in disregarding the risk that a single non-marine catastrophe might make heavy demands on both the marine and non-marine syndicates. In my judgment it did not. The decision to have a joint reinsurance programme made the risk of cross-over particularly significant. If anyone was in a position to appreciate the changes in the market that had occurred since Alicia, and which resulted in Syndicate 540 being exposed, as it was, to 87J and Hurricane Hugo, it was Mr Fagan, who was a leader at the heart of the market and who wrote the business that gave rise to those losses. The joint reinsurance cover required a joint PML exercise for the syndicates that were to rely on that cover for their protection. That exercise was not done. Insofar as consideration was given to the exposure of the marine and non-marine syndicates, they were considered individually and not jointly. This was a further demonstration of incompetence on the part of Mr Fagan.

It does not follow from this that I accept as valid Mr Thompson's opinion that the appropriate PML would have applied a factor of 100% to both the XL on XL and the whole account aggregates on the marine and the non-marine accounts. This is a matter to which I shall revert in due course.

There is one further discrete head of negligence alleged against Mr Fagan with which I must deal before turning to Mr Gofton-Salmond.

The Exchange Rate for Reinsurance Cover

Reinsurance was purchased at a fixed rate of exchange of US$2.8:£1. This was an artificial rate of exchange which significantly overvalued the £ in dollar terms. It is the Plaintiffs' case that the underwriters could, without extra premium, have negotiated a realistic rate of exchange, thereby increasing the limit of sterling cover, and that they were negligent not to do so.

If it was the case, and the evidence was not clear on this, that the sterling limit could have been raised without additional premium, there was a reason for this. Adopting a realistic exchange rate would have raised not merely the sterling limit but the sterling excess point also. In my judgment Mr Fagan was justified in regarding dollar exposure as more significant than sterling exposure. Indeed, in their final submissions the Plaintiffs have conceded that it is common ground that the predominant exposure of an account like Feltrim's would have been perceived as being to Dollar losses. In those circumstances, provided that Mr Fagan achieved an appropriate dollar limit of cover, it made sense to settle for a rate of exchange which resulted at once in a lower sterling excess point and a lower sterling limit of cover. For this reason I find that this allegation of negligence is not made out.


The 1989 Underwriting Policy for Syndicate 847

The proposals that Mr Gofton-Salmond ultimately put before the Board in relation to the business that he intended to write in 1989 was as follows:

Proposed Excess of Loss Account

1988 Projected US Projected Estimate Estimate

100% 1989 Windstorm Windstorm ROL Premium

Agg. 100% Agg. PML% PML (Approx)

London market Excess of Loss Accounts (ie. XL reinsurance of London companies & Lloyd's syndicates

Generals XL > 10% $34m $25m 80% $20m 11% $3m

Generals XL < 10% $34m $30m 50% $15m 11% $3m

Generals EX LMX $26m $30m 50% $15m 10% $3m

Generals EX LMX & RIA $11m $30m 50% $15m 7.5% $2m

LMX & RIA Spec $63m $50m 80% $40 12.5% $6m

Generals EX USA Losses $- $20 10% $2m 5% $1m

USA CAT Account

Nationwide Companies $32m $30m 50% $15m) 10% $4m

Single Zone $9m $10m 50% $5m)

US RETRO Account $35m $35m 80% $28m 8.5% $3m


Inc USA - $20m 25% $ 5m) 3.5% $1m

Ex USA - $10m - -)


Inc USA $9.6m $20m 25% $ 5m) 7.5% $2m

Ex USA - $20m - -)

Risk X/S - $20m

(ie any 10% $ 2m - $2m



Total $253.5m $350m - $167m - $30m

So far as reinsurance was concerned, he gave this account of his plans in his witness statement:

"With regard to vertical protection, my overall objective was to buy reinsurance which would provide cover of up to $100m in respect of a loss event and for as much of that cover as possible to be Whole Account protection (which would protect all of the accounts that I planned to write). To the extent that I was unable to buy Whole Account protection, I intended to fill in any gaps by buying appropriate Specific protection."

This policy was approved by the Board. Its effect was that Mr Gofton-Salmond proposed to buy sufficient reinsurance cover to limit exposure to the consequences of a single loss event to 150% of stamp.

The Plaintiffs allege that Mr Gofton-Salmond's underwriting programme was negligent in the following respects:

(i) He failed to treat an earthquake as the relevant catastrophe for the purposes of assessing his PML.

(ii) There was no justification for the discounts that he applied when arriving at some of his wind factors.

(iii) He subjected his Names to excessive exposure.

The Plaintiffs further contend that Mr Gofton-Salmond was negligent in that the business that he wrote exceeded his targeted aggregates.

Earthquake Risk

I have already referred to the fact that Mr Gofton-Salmond thought that an earthquake was the type of catastrophe which would cause the largest loss to his syndicate. He told me that the reason why he did not take an earthquake as the relevant catastrophe for PML purposes was that there was insufficient data to put a realistic figure on the syndicate's earthquake exposure. It is significant that for 1990 Mr Gofton-Salmond bought specific cover against earthquake for a layer of $10m X/S $130m when this was offered at a particularly attractive rate.

In my judgment the risk of a major earthquake was a practical possibility which Mr Gofton-Salmond should have had regard when assessing his PML and there was no basis upon which he could properly conclude, nor did he conclude, that the insured losses that would result from a major earthquake would be less than those that would result from a major windstorm. The relevance of this finding will fall to be considered when I deal with questions of causation.

The Wind Factors

LMX & RIA Spec

Mr Gofton-Salmond discounted this category to 80%, thereby reducing the PML by $10m. A great deal of time was spent with the experts in debating whether it was proper to make a discount in respect of XL on XL business. Mr Thompson's evidence was that it was never proper to take anything less than 100% of the aggregate of any class of business that included a significant element of tertiary business. This was particularly true of an XL on XL account. Both Mr Emney and Mr Outhwaite are on record as having said in the past that an XL on XL account should be given a PML factor of 100%. In their evidence in this case they said that an underwriter might have knowledge of particular features of an LMX account which would justify a degree of discount from 100% when assessing the PML. Mr Fryer thought that a discount could be justified, but that this would not be significant.

I am satisfied that, in the LMX market at this time, a PML catastrophe was likely to come so close to producing 100% claims on business in this category that it should have been assessed at 100% for PML purposes. Any discount would be based, not on a reasoned assessment, but on guess-work, for the amount of tertiary business involved would render it impossible to make reasoned assessment of the very limited extent to which a discount might be appropriate. The explanation that Mr Gofton-Salmond gave for his discount was not persuasive and I think it significant that in 1987 and 1988 his calculations had always attributed a 100% PML to this category, and that he reverted to 100% for his 1990 programme. Furthermore the original draft of the 1989 programme had had a 100% PML figure for this category. According to Mr Adamson, changes were made to this draft after intervention by Mr Drysdale and the Board. In his witness statement, he said:

"It is clear that "pure" LMX/RIA and tertiary XL on XL business should be aggregated at something near 100% PML factor: in the event of a major market loss, virtually all such contracts would be expected to respond to the loss event. 847 actually used 80% PML for this business, contrary to Robert's own (100%) view, but in deference to the overall Board view that not all tertiary XL on XL contracts would be a total loss in such a scenario."

Mr Gofton-Salmond told me that his judgment was swayed by the arguments of the Board. I think that Mr Gofton-Salmond agreed to this discount against his better judgment.

A further feature of the relevant evidence is the disclosure of a series of calculations in Mr Gofton-Salmond's handwriting showing different wind factors for some of the 1989 categories. His evidence about when these were produced was not consistent, but the wind factors are those which he subsequently suggested should be adopted for the 1990 programme and it seems to me that they probably reflected his considered opinion as to the appropriate wind factors. For the LMX category, the figure was 100%.

For these reasons I find that Mr Gofton-Salmond's discount of this PML figure to 80% was one that a competent underwriter should not have made.

Whole account covers of which less than 10% of the premium was attributable to XL on XL business ('whole account XL < 10%')

Mr Gofton-Salmond discounted this category to 50%, thereby reducing the PML by $15 million. Mr Thompson's view was that, because this category contained tertiary business, no discount should have been made from 100%. Mr Thompson's reasoning was as follows. It is impossible to know what primary risks are involved in tertiary business. It follows that there is always the possibility that all tertiary business will clash. Therefore, any class of business that embraces XL on XL cover must be assessed at 100% for PML purposes.

The Defence experts vehemently disagreed with this approach. Their view was that the possibility of 100% of the whole account business clashing was purely theoretical and it was inconceivable that this would happen in practice. The extent of the risk of clashing would depend upon the nature of the account. The percentage of premium income attributable to XL on XL business was a significant feature as it was possible for the underwriter to base a view on this of the XL on XL aggregate. Only the underwriter with knowledge of the account was in a position to decide upon the appropriate discount.

On this issue, I prefer the view of the Defendants' experts. I consider that Mr Thompson was unrealistic to suggest that the amount of premium income attributable to XL on XL business is irrelevant. It is true that the rate on line of XL on XL business can vary widely, so that premium income is an imperfect guide to aggregate exposure. Nonetheless, if the premium income attributable to XL on XL business is less than 10% of the whole in the case of each cover in the category, it is not realistic to postulate that a single event may give rise to a loss of 100% of each cover. Mr Thompson's approach is a 'safety first' approach, but it is the practical realities, not the theoretical possibilities, to which the underwriter must have regard.

The Defendants have prepared bar charts for most of the major catastrophes in this case which demonstrate the relationship of incurred losses to aggregates in respect of the various categories. One must remember that the loss figures used are not the ultimate losses and that the catastrophes do not represent the most severe that have to be considered as practical possibilities. Furthermore, this exercise has not been possible for Syndicate 847 in 1989, a year in which incurred losses from Hurricane Hugo have somewhat exceeded the PML produced by Mr Gofton-Salmond's factors. Nonetheless, I consider that these bar charts are of assistance when considering whether a PML factor is one which no competent underwriter could have reached. They show that, relatively, XL on XL is more susceptible to loss than whole account cover, and that whole account XL > 10% is more susceptible to loss than whole account XL < 10%. The highest percentage of loss actually incurred in this last category was a little under 40% caused to Syndicate 847 by the 87J windstorm.

Having regard to these matters, the Plaintiffs have not persuaded me that Mr Gofton-Salmond acted incompetently in adopting a factor of 50% for whole account XL < 10%.

Whole account covers of which more than 10% of the premium was attributable to XL on XL business ('whole account XL > 10%')

It was Mr Gofton-Salmond's evidence that any risk that included a very high proportion of XL on XL business would be put into the XL on XL category rather than the category that I am now considering. Mr Fagan gave similar evidence in his third witness statement. I found this evidence surprising, as there was not a hint of it in any of the prior witness statements or Affidavits. A single slip was produced by the defence where this practice had been followed, but I was left sceptical as to whether this was a regular practice. In the event, I do not think that it matters. The important question is whether a competent underwriter could give a whole account XL > 10% a PML factor of less than 100%

When giving evidence in Gooda Walker Mr Outhwaite said that where "a significant book of excess of loss business" is being protected under a whole account cover it should be treated as XL on XL for PML purposes. When asked in this action what he meant by significant, he said that this was subjective - perhaps 40% or 50% of the book. Subject to this, an underwriter could make a discount on the basis of his knowledge of the account and his own judgement.

Mr Fryer also considered that an underwriter could make a discount in relation to a whole account covering a proportion of XL on XL business. There was some area for the exercise of a considered judgment, based on past experience.

Mr Emney said that if there was a high proportion of XL on XL in the account it would be safe to treat it as if it were an XL on XL account, but that if the proportion was small, it would not be unreasonable to make a small discount. When I asked him how much, he said:

"I think if one went beyond 10% without very good reason, one might not be erring on the side of caution."

It seems plain that once a whole account contains significantly more than 10% of XL business, the likelihood must be that this is the area that most concerns the cedent in seeking reinsurance. I do not believe it follows, however, that the competent underwriter must assume that this will be the case in respect of every book where the proportion of XL on XL premium exceeds 10%.

I am impressed by the evidence of the Defence experts that, dependant upon the make-up of the book, there is some scope for making a reasoned, albeit limited, discount when assessing this category for PML purposes. In none of the bar charts does a catastrophe impact more than about 60% of the whole account XL>10% category.

Mr Gofton-Salmond told me that he based his discount on his feel for the overall account, having regard to his knowledge of its constituents. Applying a discount when assessing PML's because of a 'feel' for an account of this nature can only be an inexact science, but I am not persuaded that it is something no competent underwriter could do, making due allowance for the uncertainty inherent in the exercise. With the benefit of hindsight I think that Mr Fryer was right to suggest that 10% is the maximum safe discount, but I am not persuaded that Mr Gofton-Salmond's decision to take a factor of 80% was one that no competent underwriter could have taken and I reject this allegation of negligence.

Whole Account Excluding LMX

Mr Gofton-Salmond discounted this category to 50%, thereby reducing the PML by $15m. Mr Thompson's view was that, because this did not exclude RIA, tertiary business would be included and it was inappropriate to make any discount. Mr Gofton-Salmond had, in 1987 and 1988, and in his draft policy for 1989, given this category a factor of only 25%. In my judgment that factor was inadequate for an account containing tertiary business but I am not persuaded that Mr Gofton-Salmond was negligent in failing to increase the factor to more than 50%.

USA Business

It was the Defendants' case that Mr Gofton-Salmond had available sufficient particulars of the primary cover that fell into the different categories of USA business, to enable him to make informed discounts when assessing his PML. Discounts were possible because the business was essentially primary or secondary reinsurance and, to a degree, would cover different zones, even where the cedant was a Nationwide company. Mr Emney, whose experience of this area was impressive, gave evidence to this effect. I accept his evidence that these considerations could justify a degree of discount. With the exception of the single zone category, however, Mr Gofton-Salmond was not attempting a zoning exercise. Once again he was proceeding on a basis of feel. Nor, it seems to me, did he have a very certain feel for these categories of exposure. Had he felt confident that his discounts could be justified on a zoning basis, he would have been in a position to reach a reasoned earthquake PML - something he said that he was unable to do for want of data. I turn to consider the extent to which the Plaintiffs have established that the USA discounts were such as no competent underwriter could have made.

USA Retro

Mr Gofton-Salmond discounted this category to 80%, thereby reducing the PML by $7m. Mr Thompson said that because this category included tertiary business no discount from 100% should have been made. Mr Emney said that this demonstrated a lack of familiarity with this type of business on the part of Mr Thompson as it was "made up of the reinsurances of primary direct-writing companies, totally free of any reinsurance involvement whatsoever". Mr Gofton-Salmond told me, however, that this category included a certain amount of tertiary business and I suspect that it may have sent back to London and into the spiral some LMX business which had been reinsured out of London. This was a category that Mr Gofton-Salmond had given a factor of a 100% in his 1987 and 1988 calculations. He told me that he had reconsidered this category, and I note that he did not revert to 100% in his 1990 programme. I am not persuaded that Mr Gofton-Salmond was negligent to apply a factor of 80% to this category.

US Catastrophe Nationwide

Mr Gofton-Salmond discounted this category to 50%, thereby reducing the PML by $15m. Mr Thompson said that, in the absence of information to the contrary, the competent underwriter would have to assume that all the business in this category could


Mr Emney said that in practice nationwide US companies did not all write business right across the country, and there would not be a 100% clash. Furthermore, the reinsurer would be given a lot of information about the business written under this category.

It is plain that Mr Gofton-Salmond was in some doubt about the appropriate factor for this category. In 1987 and 1988 his calculations made no discount from 100%. In 1990 he raised the factor to 70%. According to Mr Adamson the PML for this category of cover proved inadequate and I do not believe that Mr Gofton-Salmond can have had information as to the business which he was reinsuring which justified reducing the wind factor to as little as 50%. At the same time Mr Thompson did not persuade me that the competent underwriter could make no discount from 100% for this business. I have concluded that 70% was the lowest factor that was consistent with competent underwriting and that Mr Gofton-Salmond was at fault in adopting a factor as low as 50%. An appropriate factor would have added $6m to his PML.


The only charges of negligence I find made out in respect of Mr Gofton-Salmond's wind factors was his failure to PML the XL on XL at 100% and his adoption of a factor below 70% for US Catastrophe Nationwide. This meant that his planned PML for 1989 was $16m less than it should have been. The consequence assumed greater significance, however, by reason of the fact that Mr Gofton Salmond exceeded his targeted aggregates. Mr Gofton-Salmond exceeded his target wind aggregates as follows:

Target Wind Actual Wind

Aggregates Aggregates

at 30/9/89

Syndicate 847 LMX/RIA $40 million $50 million

US CAT $15 million $23.5 million

Nationwide companies

US Retro $28 million $40 million

Total Wind $167 million $194.6 million Exposed Aggregates

Marine100% $20 million $39.5 million Aggregates

On 8 August 1989 the minute records that Mr Drysdale drew attention to the fact that Mr Gofton-Salmond had exceeded his aggregate targets and asked if these were merely desirable or fixed limits. Mr Gofton-Salmond replied that he believed that the targets were the Board's authority to him and that he had exceeded his authority. Aggregates would not be exceeded in the coming year.

Mr Gofton-Salmond gave a number of reasons why he exceeded his aggregate targets. These included the demands for reciprocity on the part of those from whom he was seeking to place reinsurance and the commercial desirability of meeting unexpected demand for back-up cover at attractive rates after Hurricane Hugo in September 1989. In the event Mr Gofton- Salmond exceeded his target wind aggregate by nearly $28m giving rise to a net exposure, according to his own wind factors, of some 205% of stamp. For reasons that I shall develop shortly I consider that the exposure that Mr Gofton-Salmond had intended to run of 150% was excessive. His paramount concern should have been to avoid increasing this. In these circumstances he has not persuaded me that he was justified in increasing his aggregates - the more so as he never sought the approval of the Board to this course. I find that the Plaintiffs are justified in alleging that he was negligent to exceed his aggregate targets.

Approach to Exposure

The contemporary documentary evidence discloses the following picture in relation to Mr Gofton-Salmond's approach to exposure.

On 2 March 1988 he put before the Board for the first time his Underwriting Projection for Syndicate 847 in 1989. This showed total USA wind exposed aggregates of $166.5m. The minutes record that when asked what level of reinsurance he envisaged he replied that in an ideal world he would want to place $150m of protection, but the problem was availability and cost. Mr Gofton-Salmond's ideal position would have limited his exposure to about 40% of his #25m stamp capacity. He estimated that his expenditure on reinsurance would be $12.5m out of written premium of $34m.

On 11 October 1988 Mr Gofton-Salmond produced to the Board a draft Underwriting Policy for 1989 which showed total USA wind exposed aggregates of $164.5m and reinsurance protection of $148.4m at a cost of $18.5m, out of written premiums of $35m. His planned exposure had not altered, although its cost had increased.

By 8 November 1988 the Policy had changed significantly. The wind exposed aggregates were the same, but the reinsurance cover at a cost of $18.15m had decreased to $109.5m, leaving an exposure of about 125% of stamp.

This Policy was again amended before it received Board approval, so as to show target wind exposed aggregates of $167, reinsurance cover of $100m, giving an exposure of $100m, stated to be 150% of a stamp of £26m.

In July 1989 Mr Gofton-Salmond reported that he had exceeded his target wind aggregates which stood at $183m, as against reinsurance cover of $106.5m, leaving exposure of $76.5, or 175% of stamp.

As at 30 September, the wind aggregates were recorded as $194.6m - erroneously stated to be 175% of stamp. Had the stamp been, as recorded #26m, the correct percentage would have been 194%. In fact the stamp was #24.505m and the correct percentage 205%.

These figures focus simply on the gap above the top layer of reinsurance and do not have regard to retentions or co-insurance.

This evidence persuades me that the net exposure of 150% that formed part of Mr Gofton-Salmond's Underwriting Policy, as approved by the Board, was not arrived at on a basis of a calculation of the maximum that the Syndicate could afford to run, but pragmatically having regard to what he felt he could afford to spend. The absence of a reasoned policy of maximum exposure was reflected by the extent to which he was prepared to increase his exposure beyond 150%.

My conclusion that the net exposure target for 1989 of 150% of stamp was pragmatic, rather than based on policy, gains support from the subsequent approach of Mr Gofton-Salmond to exposure.

On the 3rd/4 October 1989 the Minute records Mr Gofton-Salmond stating that in 1990 it was his intention to cover his first wind loss. Mr Gofton-Salmond's policy for Syndicate 847 as at 14 November 1989 planned that his target wind aggregates should be 100% covered by reinsurance, subject to a $3m excess, and this remained his policy well into 1990. In these proceedings the circumstances in which that policy was not achieved have not been explored.

The history of Mr Gofton-Salmond's underwriting suggests that at all times his wish was to come close to covering his wind aggregates by reinsurance and that, insofar as he adjusted his policy to a planned exposure of 150% of stamp, this was not on the basis of a considered assessment that his Names would be able to sustain losses of this dimension.

As Mr Gofton-Salmond has accepted, and as contemporary calculations, to which I shall refer shortly demonstrate, he was proceeding on the false assumption that only the most extreme wind storm would exhaust his reinsurance protection. There was no valid basis for this assumption. Even if the assumption had been correct, Mr Gofton-Salmond still had to consider the extent to which he could afford to expose his Names to such a catastrophe. I find that he did not do so, and that his approach to exposure fell below the standard to be expected of the competent underwriter.

The PML Factors for Syndicates 542 and 847 in 1987 and 1988

My findings in relation to Mr Gofton-Salmond's PML calculations for Syndicate 847 in 1989 have implications for the two earlier years. I can see no basis for distinguishing between those years and 1989. The PML calculations that Mr Gofton-Salmond presented to Mr Fagan in 1987 applied a factor of 25% to Generals Ex LMX when this should have been at least 50%. On the other hand, a factor of 100% was applied to both US Nationwide and US Retro, whereas I have accepted that a competent underwriter could not have been held at fault for discounting these to 70% and 80% respectively.

The PML Factors for Syndicate 540

I now revert to Syndicate 540. It is the Plaintiffs' case, on the basis of Mr Thompson's evidence, that to calculate the PML for Syndicate 540, a competent underwriter would apply a factor of 100% to both the whole account categories and to the XL on XL categories. The Defendants sought to support the PMLs, set out in Annexe 2, on the basis that these represented Mr Fagan's contemporary calculations. I have not accepted Mr Fagan's evidence on this. Had they represented Mr Fagan's contemporary calculations, I would have found them defective on two counts:

1)The XL on XL category was discounted by 10% when, for the reasons given when dealing with Syndicate 847, a factor of 100% should have been adopted.

2)The PML for the whole account categories consisted of the aggregates falling in volatility bands 2-6. This had the following effect in percentage terms:

In 1987 the global whole account category had a PML factor of about 60%.

In 1988 the whole account XL > 10% had a PML factor of about 46%.

In 1989 the whole account XL > 10% had a PML factor of about 43%.

These figures demonstrate that volatility factors were not a valid basis for PML calculations.

For the reasons given when considering Mr Gofton-Salmond's PMLs, I do not accept Mr Thompson's evidence that the whole account categories had to be given a PML factor of 100%. I can see no basis for adopting a different approach to the PML factors for Syndicate 540's whole account categories from that which I have applied when considering Syndicate 847. Accordingly I find that a competent underwriter should not have applied a PML factor of less than 80% to the 1987 whole account category, or to the 1988 and 1989 whole account XL > 10% categories. The lowest PML factor that a competent underwriter could have applied to the whole account XL < 10% categories was 50%.


The Feltrim Agency was created at a time of change in the relationship of a Managing Agency and its underwriters. Hitherto it had been the practice for the Board of a Lloyd's Managing Agent to delegate matters of underwriting policy and practice to its underwriters. By the time that Feltrim took over responsibility for the management of the syndicates it had become recognised that the Board of a Managing Agency had a duty to consider and approve underwriting policy and to monitor its implementation.

It is apparent that the Board of Feltrim appreciated the nature of their duties and attempted to perform them. From the outset they requested the underwriters to present underwriting policies and to set aggregate targets, and at Board meetings they were concerned, at least, to monitor these targets. They did not, however, receive from the underwriters the data that they needed if they were to perform their duties. Mr Drysdale seems to have been the Director most concerned about these matters, and most keenly aware of the deficiencies in the data that the underwriters were providing. He repeatedly sought from Mr Fagan information that would enable the Board to identify the extent of the syndicates' exposure, but he did not receive this information - no doubt because Mr Fagan was not in a position to supply it.

Not until Mr Gofton-Salmond put before the Board his underwriting policy for Syndicate 847 in 1989 did the Board receive proper particulars of planned exposure. In March 1989 Mr Drysdale asked for the opportunity to work with Mr Fagan to produce a similar document.

At the time that Mr Gofton-Salmond sought the approval of the Board to a policy that involved net exposure to the extent of 150% of stamp, he placed before them his calculations of the effect on his syndicate of windstorms of varying degrees of severity. These indicated that an original insured loss of both $5 billion and $7.5 billion would be contained within reinsurance protection, coupled with net premium income, whereas an original insured loss of $20 billion would result in a call of 164% of stamp. Those calculations postulated relationships between original insured losses and claims on the syndicate which were not consistent with Mr Gofton-Salmond's PML calculations and which Mr Gofton-Salmond was unable to justify when he gave evidence. I think that the Board must have accepted these at their face value and proceeded on the basis that only the most extreme catastrophe would expose the syndicate to the PML. Thus on the 3rd/4 October 1989 the Minute records Mr Gofton-Salmond stating that in 1990 it was his intention to cover his first wind loss. A week later the Minute records Mr Manning stating that Names on the marine syndicate should not be exposed beyond the $75m cover purchased and a discussion ensuing on the desirability of covering the syndicate's PML at all times.

On 14 November 1989 there was discussion as to the problems posed by the fact that the writing of business could not be synchronised with the purchase of reinsurance cover, and it was agreed that the underwriters were authorised to accept business in accordance with their underwriting policies for 1990 provided that aggregates at no time exceed reinsurance in force by more than stamp.

In 1987 and 1988 I believe that the Board accepted Mr Fagan's assurances that he had adequate reinsurance. Attempts to obtain the data to check this were unsuccessful. Their approval of an exposure of 150% for Syndicate 847 in 1989 did not reflect a considered policy and was given on the false assumption that cover would be adequate to protect against only the most severe catastrophe. Only when the effects of Piper Alpha began to become apparent did the Board give detailed consideration to the appropriate policy on exposure. I have some sympathy with the problems that they faced in attempting to monitor the underwriting of Mr Fagan, but I consider that the Plaintiffs have made out their case that supervision by the Board was less than adequate.


As Mr Emney confirmed, if an underwriter plans to expose his Names to the risk of a periodic loss making year, he must make sure that his Names are aware of the extent of their exposure. In his final speech Mr Johnson submitted that, as the Plaintiffs had failed to allege as a head of negligence any failure to inform Names of the risk that they were running, the question of the perception of the Names did not arise in this case. I do not agree. The Plaintiffs pleaded that the exposure to which the Feltrim underwriters subjected their Names was excessive. This plea raised the issue of the level of exposure that could properly be run, and that necessarily involved consideration of the perception that the Names had, or should have had, of the level of their exposure. That this was so, is apparent from para 35 of Feltrim's Amended Points of Defence, which alleges:

At all material times from at least June 1986 the Names (including all the Plaintiffs) knew or ought to have known (either themselves or by their Member's Agents):

(i) that the Feltrim Syndicates specialised in the writing of excess of loss reinsurance, with a substantial US$ content;

(ii) that, as such, claims on the Feltrim Syndicates

tended to result from major market losses, such as the loss of supertankers, drilling rigs and natural disasters such as hurricanes;

(iii) that a part of the risks underwritten by the Feltrim

Syndicates would be retained for their net account;

(iv) that the ability of excess of loss syndicates such as

the Feltrim Syndicates to obtain reinsurance was limited by the availability of capacity in the reinsurance market;

(v)that as a result of this limit on capacity, the Feltrim Syndicates were having to retain a larger percentage of risks for their net account.

The pleading alleges that the Names should have had this knowledge:

(i)Because it was apparent from Feltrim's Annual Reports and Accounts.

(ii) Because Feltrim were well known in the market as LMX specialists.

(iii) Because Mr Fagan and Mr Gofton-Salmond and other

members of Feltrim's staff informed Members' Agents of these matters at regular meetings.

A logical starting point for considering what others appreciated as to the extent of exposure run by the Feltrim underwriters must be to consider the appreciation of the underwriters themselves. This is a matter I have already touched on. Both Mr Fagan and Mr Gofton-Salmond appreciated that they were running a degree of exposure, but believed that it would require a catastrophe of significantly greater severity than any of those actually experienced to exhaust the level of reinsurance cover. This belief was evidenced by Mr Gofton-Salmond's loss calculations. Mr Adamson said that to the extent that he was involved in discussions, he understood that a loss of $4.5 to $5 billion might just be contained within the reinsurance protection, or just exceed it. His evidence was that the syndicate worked on the basis of one large catastrophe every twenty years that would probably be contained within the reinsurance protection and a number of smaller catastrophes that would certainly be contained within the reinsurance protection on a more frequent occurrence.

Reports and Accounts

It would be surprising if, when talking to Agents, or when drafting the annual Reports and Accounts the underwriters painted a more gloomy picture than their own appreciation of the position. Certainly they did not do so in the Underwriting Report for the syndicates for 1986, published on 2 June 1987. This included the following passage:

"The three Syndicates are protected by a substantial combined reinsurance programme designed to give both coverage for the worst conceived largest loss such as an earthquake, windstorm or natural disaster; and also for a multitude of smaller losses wherever they may arise throughout the insurance market."

The Underwriter's Report for 1987, published on 20 May 1988, commented in slightly more guarded terms:

"The Syndicates are protected by a substantial combined reinsurance programme designed to give both maximum affordable coverage for a major loss such as an earthquake, windstorm or natural disaster; and also for a multitude of smaller losses wherever they may arise throughout the insurance market."

The Report went on to express confidence that reinsurance would contain Syndicate 847's losses from the 87J windstorm.

The Underwriter's Report for 1988, published on 26 May 1989 contained a passage in identical terms and still predicted that losses from 87J would be contained.

The Annual Reports and Accounts gave no warning to the Names on the 1987, 1988 or 1989 syndicates of the degree of exposure to which the syndicates were subject.

Exposure Implicit in LMX Business

Just as in Gooda Walker, I have to approach the perception of the Names in these Actions on the premise that they were being properly advised by competent Members' Agents. In Gooda Walker Lord Strathalmond was called by the Defendants to give expert evidence in relation to the viewpoint of a Members' Agent. In these Actions I have had no expert evidence of that nature, although the Plaintiffs have adduced under the Civil Evidence Act the evidence given in Gooda Walker by Lord Strathalmond as evidence of the viewpoint of an individual Members' Agent. The Defendants have referred me to the agreed returns made by Syndicates 540/2 from 1980 to 1986 and Syndicate 847 from 1983 to 1986, being 16.6% and 25.96% on line respectively. Lloyd's regulations in force in 1987 permitted the writing of a line of twice a Name's proven wealth and required only 40% of proven wealth to be put up by way of a deposit. For the Defendants it was submitted that the return enjoyed by the Names up to 1987 had been truly spectacular, particularly if viewed by reference to the deposit which they had had to put up. In these circumstances they must reasonably have anticipated the possibility of significant inroads into their unlimited liability in bad years.

In these Actions I have not had evidence of the returns enjoyed by Lloyd's Names from other classes of business, but the Plaintiffs' Counsel did not challenge the assertion that the returns, as reported, were relatively generous and I so find. I also find that the competent Members' Agent would conclude that the higher degree of reward carried with it a higher degree of risk, and that in a bad year Names might be subject to a net loss. To this extent the Feltrim syndicates should have been known as 'high risk, high reward'. On Mr Outhwaite's evidence the acceptable amount of exposure would vary from syndicate to syndicate: "one could not say 50% is alright or 100% is alright". This evidence I accept. The Feltrim underwriters had regular meetings with Members' Agents and I find that the competent Members' Agent should have sought information from the underwriters as to the amount of exposure they were running. In the absence of such enquiries I do not consider that a Members' Agent would be justified in assuming that, in the event of the most severe catastrophe that could be contemplated as a practical possibility, the Names would be exposed to losses of significantly less than 100% of stamp. At the same time losses of this dimension would be considered "almost unthinkable", as would the catastrophe giving rise to such losses.

Information Given to Agents

Some of the contemporary documents refer to statements made by the underwriters as to the extent of the exposure that they were running. Some of this constitutes double hearsay, and the Defendants have stood on their right to object to the admissibility of such evidence. In these circumstances I do not have a reliable evidential basis for making findings in respect of this area of the case. I anticipate that it will receive more satisfactory scrutiny in the pending Action between Names on the 1990 Feltrim Syndicates and their Members' Agents. Such evidence as is admissible is not consistent, but there is at least some evidence of Agents being told of the possibility of losses of 100% of stamp in the worst possible scenario. I am, however, satisfied that the Agents were not at any material time told that the exposure was as much as 150% of stamp. Before the Neill Committee Mr Gofton-Salmond admitted that if the Members' Agents had been told this there would have been a flight of Names from the syndicates.


The burden of proof on this, as on other issues, is on the Plaintiffs. Having regard to this and on the unsatisfactory state of the evidence I find that Names properly advised by competent Members' Agents should have understood that in the unlikely event of the worst catastrophe occurring that was a practicable possibility, losses of as much as 100% of stamp might be suffered. I see no basis, however, upon which the Defendants can contend that Names should have anticipated that lesser catastrophes would cause them significant net losses. The underwriters themselves did not anticipate this and the Defendants are not in a position to suggest that Members' Agents or Names should have done so.

What was the level of Net loss beyond which it would have been unreasonable for the underwriters to expose the Names?

In Gooda Walker the Defendants contended that there was no principle that governed the extent of the net exposure that an underwriter could properly run. The degree of exposure run was a matter of underwriting judgment. The Plaintiffs for their part did not allege that any particular level of exposure was negligent per se. They alleged that the underwriters had failed to apply proper judgment and established principles to their underwriting.

In these Actions the bedrock of the Plaintiffs' case has been an allegation that the level of exposure to which they were subjected was negligent per se, and much of the Plaintiffs' evidence and argument has focused on the appropriate test for ascertaining the permissible level of exposure.

The experts were agreed that a, if not the, principle reason why an excess of loss underwriter buys reinsurance cover is to ensure that the syndicate, or company, for which he is underwriting will be able to weather the adverse effects of any catastrophe for which it is necessary to make provision as a practicable possibility. It was also agreed that, in order to assess the consequences of such a catastrophe, the underwriter has to carry out a PML exercise. Having ascertained his PML, the underwriter will be able to calculate how much reinsurance cover he needs to buy in order to limit his net exposure to the extent that he considers appropriate. It is common ground that the competent underwriter can plan on the basis that he will on occasion have a loss making year, that will be covered by the profits made in the good years. The extent of the loss that can be carried in this way, and the length of the period over which the underwriter will plan to balance profit and loss will depend upon the circumstances. In the case of a company, this is likely to depend upon the reserves that are maintained and upon considerations of cash flow. In the case of a Syndicate, it will depend upon the extent to which Names will find it acceptable to fund losses out of their own resources in recognition of profits that they have made in the past and profits that they anticipate making in the future.

Mr Emney expressed the view that a Lloyd's underwriter might reasonably aim to achieve a balance of profit and loss over a period of five years. This was something that he would have to plan. He accepted that if an LMX underwriter did not make calculations to assess his potential profits in good years against potential losses in bad years in order to achieve that balance over time, he would be ignoring the cornerstone of the LMX market's philosophy. He also agreed that part of the equation was that those being exposed to the risk of losses in such circumstances should agree knowingly to run that risk.

Mr Outhwaite said that there was no objective standard at all as to the level of exposure that was acceptable, but added:

"Obviously one could get to the point where one could say this is totally unreasonable. I have already said that if we are talking about a net exposure on a PML basis of 250%, I would regard that as excessive, and I think that would be echoed by virtually every underwriter. One could not say 50% is all right or 100% is all right. There is no standard of that kind that one can adopt."

In one important respect there was a degree of contradiction between the experts. Should one, when deciding how much net exposure to run, have regard to the likelihood that the casualty chosen as the basis for the PML will occur? Mr Thompson said that one should not. Mr Outhwaite disagreed. So did Mr Fryer. He summarised his approach thus:

"Because it is going to happen very infrequently I am prepared to take a bigger risk than I am if it is something that is going to happen every five or ten years."

I asked Mr Outhwaite whether, when considering what exposure to run, one did not have to have regard not only to the likelihood of the occurrence of the catastrophe that one has selected as the basis for ones PML, but also to the likelihood that a lesser catastrophe might exhaust the reinsurance cover. He replied that he did not think so. If a PML was based on a catastrophe that was only likely to happen once every thirty years, the assumption would be that smaller catastrophes would be unlikely to get to the top of the reinsurance protection.

This answer conflicted with other evidence given by the Defence experts. They pointed out that a catastrophe may result in claims to the limit of the exposure of a syndicate, even though it falls short of the most severe catastrophe that is a practical possibility. I do not see how a competent underwriter could automatically make the assumption suggested by Mr Outhwaite. An underwriter considering how much net exposure to run would have at least to consider the possibility that this exposure would render the syndicate vulnerable to less extreme catastrophes than that which had formed the basis of the PML calculation.

The reality is that the opacity inherent in the gearing effect of the spiral made it impossible for any particular underwriter to make an informed quantative prediction of the level of original insured loss that would exhaust his reinsurance protection, or approach his PML. The Defendants implicitly concede this in the following passage of their written final submissions:

"The Plaintiffs seek to criticise the Feltrim Agency for not appreciating that losses of the particular magnitude which occurred in 1987-1990 would cause net losses of the size which were actually sustained and for considering that it would only be larger losses that would have had this effect.

This criticism and the prominence now given to it by the Plaintiffs are surprising in light of the fact that it is a criticism which is not (or is wholly inadequately) made in the Plaintiffs' Points of Claim and is not supported by expert evidence. The experts agreed that the proper approach to underwriting LMX business involved the monitoring of aggregates, the assessment of a PML and the purchase of reinsurance in relation thereto. There was indeed no other way in which it could be done. But what the expert evidence did not support was the suggestion that, if these things were done properly so that (for example) the PML was accurate for the largest ordinarily foreseeable loss, the underwriter could be criticised for failing to appreciate that a smaller loss might come closer to his PML than he had anticipated.

It is, in fact, a very difficult criticism to follow. It is not suggested how the underwriter was supposed to assess exactly how far up towards the PML a smaller loss would reach, and indeed the Plaintiffs assert that this was impossible for any participant in the LMX market."

Those, such as Feltrim, who wrote high layers would know qualitatively that they were less exposed than those who wrote lower layers. What they could not do was assume, because their PML was based on a catastrophe with a statistical rate of occurrence of once in thirty years, that they would not be exposed to less infrequent catastrophes. This has particular relevance to the question of the frequency of occurrence of catastrophes for which the competent underwriter should seek to make provision when deciding how much exposure to run to a single loss event.

Mr Thompson's opinion was that an underwriter would have to make provision for the possibility of three severe catastrophes in a year. In advancing this view he relied upon the fact that it was market practice for non-marine XL policies to make provision for one reinstatement and for marine policies to provide for two reinstatements.

Mr Outhwaite's evidence was that reinstatements did not reflect the desire of the cedant to secure cover against a series of catastrophes, but were a requirement introduced by the reinsurer in order to procure that in the event of a claim he would receive some additional premium to offset his loss.

Mr Fryer did not agree with Mr Outhwaite about this. His view was that reinstatements reflected a desire on the part of the cedant to obtain cover. His opinion was that the market contemplated the risk of more than one loss making catastrophe in the course of a year, although two catastrophes were far less likely than one. Whatever the origin of the provision for reinstatement, I am satisfied that, even though Feltrim wrote high layers, the underwriters had to make provision for the possibility that at least two loss making catastrophes might occur within the same year.

Mr Thompson's opinion was that as a 'rule of thumb' the maximum net exposure that a competent underwriter could permit to a single loss event was 40% of stamp, thereby making provision for three possible loss producing catastrophes in the course of a year producing a total net loss of 150% of stamp.

Counsel for the Defendants ridiculed this evidence as being unsupported by any relevant experience on the part of Mr Thompson and being a figure plucked out of the air. They submitted on the basis of the evidence given by Mr Outhwaite that no underwriter could be criticised for deciding to run an exposure between 100% and 200%. Mr Outhwaite made it plain that in his view the appropriate degree of exposure depended upon the particular circumstances. I have to decide the maximum level of exposure to which competent underwriters should have exposed the Names having regard not merely to general considerations but to the particular features of this case. The relevant considerations include the following:

1) The fact that Lloyd's required that Names should not write more than 250% of 'proven wealth'.

2) The fact that Lloyd's considered a loss of 100% of stamp justified an enquiry by a Loss Review Committee.

3) The level of profitability enjoyed by Names in previous years.

4) The perception of the Names as to their exposure.

5) The intention of the underwriters to protect the Names from all except the most severe catastrophes.

I accept Mr Emney's view that a reasonable period over which a Lloyd's syndicate would need to balance profits and losses would be about five years. I also have regard to the fact that a severe catastrophe would result in a significant increase in rates. As against this, the underwriter would have to cater for the possibility of more than one loss making catastrophe in the period.

My conclusion is that the maximum net loss that the Feltrim underwriters should have been prepared to risk from a single severe catastrophe was 100% of stamp. As Mr Eder has pointed out more than once, this would not represent the overall loss for the year for, putting the effect of the severe catastrophe on one side, the balance of the underwriting, including investment income, could be expected to be profitable. Thus the Feltrim underwriters should have purchased sufficient reinsurance to reduce their Names' net exposure on a first loss basis to not more than 100% of stamp. The extent to which such exposure was represented by retention at the bottom, co-insurance, or exposure at the top would have been a matter for underwriting judgment. In the event the Names were left exposed far in excess of such a limit.


Mr Cooke submitted that the question of the maximum exposure that a competent underwriter could run was irrelevant. The reality was that it was impossible for the underwriters to purchase sufficient reinsurance - at least on terms that would leave the account viable. The business should not have been written and the Names should be put in the same position as if the underwriters had written a different, properly balanced book of business.

I think that Mr Cooke's appraisal of the position is probably correct. In the course of his evidence Mr Gofton-Salmond conceded that the upper layers of business written were grossly underrated. He also said that at no time did he or Mr Fagan do any calculations of anticipated profits that they expected to make on good years to compare with any loss potential on bad years.

When giving evidence to the Neill Committee Mr Fagan said that with hindsight he did not think a Lloyd's syndicate was the right vehicle to write an XL account.

There are, however, cogent objections to the approach which Mr Cooke urges me to adopt to the assessment of damages. The first objection is that I do not consider that it is a practical exercise. In Gooda Walker an attempt was made to identify a paradigm syndicate for the purpose of assessing damages. I held that this was not a realistic exercise. Mr Eder submitted that it would be no more realistic in this case to attempt a similar exercise and I agree with him. It would no doubt be possible to construct a model of a book of business that would not have resulted in the losses sustained by the Feltrim syndicates - indeed it would be possible to construct a model of a profitable book of business. Mr Adamson said he knew of syndicates that, so long as rates remained low, were able to secure profits by resorting to arbitrage. But such a book of business was not the kind of business that Feltrim set out to write.

The second objection is that it is not the Plaintiffs' pleaded case that the Feltrim underwriters were negligent in writing an account that was not viable. The underwriting business carried on by the Feltrim underwriters was that of writing catastrophe excess of loss reinsurance. The Feltrim Names might have alleged that it was negligent of the underwriters to write this class of business, but that would have been to advance a very different case to that which they pleaded, and would have been likely to evoke a very different response from the Defendants.

The Points of Claim allege under the heading Breaches of Duty that the underwriters were at fault in various respects for failing to ascertain or appreciate the extent of the risk implicit in the business being written. It is, however, acts and omissions, not mental attitudes alone, that give rise to legal liability. The only acts or omissions alleged by the Names to constitute breach of duty are failures to put in place adequate vertical reinsurance protection. I have held that these allegations are made out. In my judgment the correct approach to damages is to attempt, insofar as this is possible, to place the Names in the same position as if those breaches of duty had not occurred.

This is essentially the same approach to damages that I adopted in Gooda Walker. In that Action, however, in dealing with liability I did not have to determine either the PMLs for the various syndicate years or the maximum exposure that the underwriters could competently have run. In these Actions I have determined both the factors that should have been applied to aggregates in order to determine PMLs and the maximum net exposure that could properly have been run on the basis of those PMLs.

In the result the assessment of damages will involve a more sophisticated, and more complex, exercise than that on which I am about to embark in Gooda Walker. In preparing for the assessment of damages in Gooda Walker a number of points have been raised which I had not anticipated when ruling on the principles to be applied in the assessment of damages. These have persuaded me that in these Actions I should do no more than formulate the basic principle to be applied, leaving open for later argument the manner of application of that principle. That principle is simply that the damages awarded should place the Names on each syndicate year in the same position as if the underwriters had purchased reinsurance protection sufficient to restrict the Names' net exposure to the PML to 100% of stamp. In calculating net exposure regard should be had not merely to the gap above the top layer of reinsurance cover, but to retentions and co-insurance.

Had these Actions been concerned with loss flowing from a single catastrophe that was protected on a first loss basis by the reinsurance in place, assessment of damages in accordance with this principle would, I think, have been relatively simple. The assessment would have involved the following calculations.

1) The loss actually sustained as a result of the catastrophe by reason of there being inadequate vertical reinsurance.

2) The PML, applying the factors that I have held should have been applied to the aggregates actually written.

3) The additional costs that would have been incurred in buying reinsurance cover sufficient to reduce net exposure to 100% of stamp.

4) The losses that would have been sustained as a result of the catastrophe had such reinsurance been in place.

Prima facie the damages would have been represented by the differences between 1) and 3) + 4). Complications may arise by reason of the fact that the Plaintiffs are claiming in respect of losses caused by a number of catastrophes and that it is common ground that the reinsurance cover in force should be deemed to be available to meet each catastrophe on a first loss basis. The possibility of such complications merely underlines the desirability of making no rulings at this stage which may restrict the arguments that the parties may wish to advance as to the manner of application of the basic principle to the task of assessing damages. There are, however, a number of discrete issues which can properly be dealt with at this stage.

Earthquake Risk

I have held that, having regard to the perception at the time, the underwriters should have planned their reinsurance to cater for the risk of a severe earthquake. I have not attempted the exercise of assessing the PMLs that they should have adopted in order to do this, and I would have found this exercise no easier than did Mr Gofton-Salmond. I have, however, throughout these Actions been unable to see how failure to cater for the consequences of an earthquake can constitute causative fault on the facts of this case. The underwriters set out to protect the Names against catastrophes of the types that have led to the losses that they have suffered. Their liability arises from failure to buy adequate reinsurance cover to cater for the risk of these catastrophes. Had they restricted exposure to these catastrophes to an appropriate degree, I do not consider that they could be held to be causatively at fault because they had failed to buy additional cover, necessary only because of the risk of a different type of catastrophe.

Unprecedented Catastrophes

The Defendants allege that four of the catastrophes with which these Actions are concerned had features which rendered them unprecedented:


Piper Alpha

Hurricane Hugo

Windstorm 90A

The evidence of their experts, Mr Hopkins and Mr Clarke, supports this submission. It does not follow, however, that the underwriters cannot be held at fault for failing properly to protect their Names against these catastrophes. Any severe catastrophe is likely to have features which render it unique. Past catastrophes are no sure guide as to those which may happen in the future. No underwriter can expect to foresee, or to attempt to foresee, the infinite combination of circumstances which may give rise to a loss.

By basing reinsurance on the PML, the underwriter sets out to cover the worst catastrophes that he can foresee as a practical possibility on the basis that this will protect him against a wide variety of lesser catastrophes, whether or not the features or the circumstances of the particular catastrophe are unprecedented. Each of these catastrophes was of a type which the underwriters intended to guard against by reinsurance and insofar as inadequacy of reinsurance cover has increased the losses resulting from those catastrophes, I can see no reason to hold the losses too remote to be recoverable.

Concatenation of Catastrophes

The sequence of severe catastrophes with which these Actions are concerned was without precedent and, arguably, not reasonably foreseeable. I have already had to consider the frequency of catastrophes that should have been anticipated in the context of the maximum permissible exposure, and I have held that the underwriters should have contemplated the possibility of at least two severe catastrophes. It does not follow from this that the underwriters were under no obligation to protect their Names against a sequence of more than two catastrophes. Once one catastrophe has occurred the odds do not lengthen against a second following. These, however, are considerations which go to horizontal rather than to vertical protection. No complaint has been made in these Actions of lack of horizontal cover - hence the agreement that each catastrophe is to be deemed to have been protected on a first loss basis. The Defendants are confident that only one catastrophe in each calendar year - 87J, Piper Alpha, Hurricane Hugo and, possibly, 90A, resulted in losses which outstripped the first loss outwards cover. In these circumstances I do not see the relevance of the arguments addressed to the lack of foreseeability of the concatenation of catastrophes.

I now turn to a number of issues which I ordered should be determined at this hearing on 20 July 1994.

Other Heads of Loss

The Plaintiffs seek to recover other heads of loss which they contend were caused by the inadequacy of vertical cover, such as currency losses, run-off costs and syndicate expenses. I propose to make no ruling on these until I see detailed particulars of these heads of claim.

Stop Loss Protection and Voluntary Assumption of Risk

The Members' Agents have pleaded that those Names who took out stop loss protection voluntarily accepted the risk of losses up to the limit of their stop loss policies, or alternatively up to the excess. I ruled against a similar submission in Gooda Walker and I do so in these Actions for the same reason.

Failure to Take Out Stop Loss Insurance

I again rule against the contention, made in Gooda Walker and advanced in these Actions, that Names who were advised to take out stop loss insurance, but who declined to do so, are precluded from recovering any losses that would have been protected by their stop loss cover.

Stop Loss Recoveries

Here again, for the reasons given in Gooda Walker, I rule against the contention that the Names must give credit for stop loss recoveries.

Credit for Liabilities Remaining with Other Syndicates

This credit was claimed on the premise that I might hold that the Feltrim syndicates should have written less LMX business, and cannot be advanced having regard to my approach to the Actions.

Credit for Profits on Prior Years

My judgment leaves the Plaintiffs exposed to the possibility of significant losses, reflecting the fact that the Feltrim syndicates were 'high risk, high reward'. This of itself precludes the contention that credit should be given for the 'high rewards' of previous years. Accordingly I need not consider the other points made by the Plaintiffs in answer to this contention.


For the reasons given by Potter J in Gooda Walker the damages will be taxable and no credit falls to be given in respect of tax relief.


By an amendment to their Points of Defence the Members' Agents alleged that some of the Plaintiffs' claims were time-barred by reason of the provisions of the Limitation Act 1980. Potter J has recently held in PJ Aiken & Ors v Stewart Wrightson Members' Agency Ltd & Ors that claims for damages for breach of agreements between Names and Members' Agents, which by virtue of the requirements of Lloyd's Agency Agreement Byelaw No.1 of 1985 must be under seal, are subject to a 12 year time limit. The Members' Agents accept that if this point has been correctly decided, they have no defence of limitation. They do not seek to challenge Mr Justice Potter's decision before me, but reserve the right to do so in the event of an appeal.


The Feltrim Agency has not challenged the claim by the Members' Agents to an indemnity in respect of any liability the Members' Agents may be under to the Names. That claim is plainly well-founded and I hold that the Members' Agents are entitled to recover an indemnity from Feltrim in respect of 100% of their liability to the Plaintiffs. This right is likely to prove of little comfort to the Members' Agents.

Judgment accordingly

Richards Butler; Clifford Chance; Elborne Mitchell