Crane v Hannover Ruckversicherungs- Aktiengesellschaft & Anor

Legion Insurance Company (“Legion”) provided casualty insurance to businesses in the United States of America (“U.S.”). Legion was a subsidiary of Mutual Risk Management Ltd of Bermuda (“MRM” – an abbreviation which I shall use to refer not only to Mutual Risk Management Ltd but also to its predecessor companies and to the Mutual Risk Management group of companies along with their predecessors). The policies written by Legion often included cover against liabilities under statutes passed by individual states for the benefit of those injured at work. This cover was known as Workers Compensation Act cover (“WCA cover”). It was standard for such cover to comprise two sections. Section A gave cover for statutory benefits in respect of death or bodily injury arising from an accident at the place of work and during working hours. Section B gave cover for payments in respect of an employer’s fault-based liability for an accident killing or injuring an employee. Legion’s policies also often provided, among other things, general and automobile liability cover.

Many of the policies written by Legion were classified by it as “traditional Insurance Profit Centre” business. The book of business comprising these policies was known as the “Mainframe Account,” and reinsurance arrangements designed for this book of business were known as the “Mainframe Treaty” or “Main Treaty.” I shall refer to policies falling within this book of business as “Mainframe policies”. Certain other policies were the subject of separate reinsurance arrangements known as the “Cessions Treaty.” The book of business comprising these policies was known as the “Cessions Account” and I shall refer to such policies as “Cessions policies.”

The defendant, which I shall refer to as “Hannover”, is a German reinsurance company. In 1998 and in previous years it underwrote a proportion of two excess of loss reinsurance policies giving cover to Legion for its liabilities in respect of business allocated to the Mainframe Account. I shall generally refer to such reinsurance policies, directly protecting the Mainframe Account, as “the Mainframe XL contracts” and to those who underwrote them as “the Mainframe XL reinsurers”. The policies underwritten in 1998, and those who underwrote them, I shall refer to as “the 1998 Mainframe XL contracts” and “the 1998 Mainframe XL reinsurers.” Other companies underwriting the 1998 Mainframe XL contracts included First Excess & Reinsurance Corporation (through Core Managers Incorporated) and TIG Reinsurance Company. The 1998 Mainframe XL contracts were in force from 1 October 1998 onwards and reinsured the net excess liability of Legion resulting from loss occurrences under Mainframe policies allocated to underwriting years commencing during that period. Each of them included both specific excess cover and aggregate excess cover:

MRM Hancock Ltd (“MRM Hancock”), a broking firm based in London, was – like Legion – a subsidiary of MRM. It acted as Legion’s broker in relation to the Mainframe XL contracts. It was instructed both by Legion directly and by Towers Perrin in the U.S. on Legion’s behalf. Hannover has made a contingent claim against MRM Hancock (under its current name, Park London Ltd) as a Part 20 defendant in these proceedings. That claim has been stayed by consent. On 2 August 2007 MRM Hancock went into administration. It has played no part in the proceedings before me. My findings in this judgment, and references below to matters being common ground, concern only the position in the proceedings before me.

In 1998 Syndicate 53 at Lloyd’s was an aviation syndicate. It wrote other books of business as well, among them personal accident (“PA”) business. This type of business was also known as accident and health (“A&H”). The syndicate’s active underwriter in 1998 and 1999 was, and had for some years been, Mr Ian Crane. He is the claimant in these proceedings and sues on behalf of himself and other members of Syndicate 53. I shall refer to it as “the Syndicate” or “Syndicate 53”.

It is common ground that by four excess of loss reinsurance contracts the Syndicate, for its appropriate proportion, agreed to participate in the reinsurance of some, but not all, of those who underwrote the 1998 Mainframe XL contracts. For three of the four contracts the reinsured included Hannover. The remaining proportion of the four excess of loss reinsurance contracts was underwritten by Reliastar Life Insurance Company (“Reliastar”). Each of these four contracts can be described as a retrocession – a reinsurance of a reinsurer. Each of the Syndicate and Reliastar was, for its own proportion, reinsuring aspects of potential liabilities which the reinsured companies had themselves underwritten as reinsurers under the 1998 Mainframe XL contracts.

The Syndicate’s involvement in Legion’s business in 1998 was not limited to the Mainframe Account. By three excess of loss reinsurance contracts the Syndicate, for its appropriate proportion, agreed to participate in the reinsurance of Legion’s Cessions Account. These three contracts were not retrocessions, for the Syndicate was reinsuring aspects of potential liabilities which Legion had underwritten as a direct insurer. The Syndicate was asked to quote rates for a fourth contract reinsuring the Cessions Account, and duly did so, but in the event this did not come to fruition.

A feature common to all seven of these contracts was that they did not apply to all classes of business. They extended to one class only, which was in essence section A of WCA cover. By 1998 underwriters in the A&H market in London had for some years been providing to American casualty underwriters reinsurance limited to this single class. The cover thus provided was effectively carved out of the wider cover in the casualty policies, and it was accordingly known as “WCA Carve-out”, or simply “Carve-out” reinsurance. I shall generally refer to it as “Carve-out cover.” I shall refer to the four retrocessions as “the 1998 Mainframe Carve-outs,” and to the three reinsurances of Legion’s Cessions Account as “the 1998 Cessions Carve-outs.” These contracts formed part of the wider reinsurance structures comprising the Main Treaty and the Cessions Treaty. In contemporary documents and in evidence before me, however, the expressions “Main Treaty” and “Cessions Treaty” were often loosely used to refer to particular contracts protecting the Mainframe Account and Cessions Account respectively.

The WCA Carve-out market was highly specialist. It was, however, well known to Mr Crane. He had been introduced to it by a specialist broking firm which by 1998 was known as Stirling Cooke Brown Insurance Brokers Limited (“SCB”). It is common ground that in 1998, as in previous years, MRM Hancock sought Carve-out cover as broker on behalf of Legion, which wanted Carve-out cover directly for its Cessions Account, and also as broker on behalf of some or all of the potential or actual Mainframe XL reinsurers. It is also common ground that in relation to Carve-out cover SCB acted as sub-broker for MRM Hancock.

The 1998 Mainframe Carve-outs applied to occurrences during the period of 12 months commencing on 1 October 1998. The contract wording described the reinsured as Legion “and/or all other Companies which are now or hereafter become part of the Mutual Risk Management, Limited Group... and/or various Reinsurers [of Legion]”. Despite this wording, there is no suggestion that Legion has any claim against the Syndicate, and it has not been suggested that I am concerned with any claim that any reinsurer of Legion other than Hannover may have against the Syndicate. Nor am I concerned with any claim that may be made against Reliastar.

Two of the 1998 Mainframe Carve-outs provided a first and second layer of specific cover, and the remaining two provided a first and second layer of aggregate cover. Hannover was reinsured only under the first specific layer and the two aggregate layers – having, after consideration of the rate offered on the second specific layer, decided that it was too costly. The parties agree that Hannover and the Syndicate became contractually bound to each other in respect of each of the first specific layer and the two aggregate layers simultaneously. However they have not been able to agree on the date when this happened. For convenience, I shall refer to this date as “the Hannover/Syndicate contract date.”

This is the judgment of the court following the trial of two main questions. The first is whether the Syndicate is not liable to Hannover under the relevant 1998 Mainframe Carve-outs because it has validly avoided them as against Hannover for misrepresentation. The second main question is whether the Syndicate is not liable to Hannover because it has validly avoided the relevant 1998 Mainframe Carve-outs as against Hannover for non-disclosure. The answers to these questions may have financial consequences. If these consequences cannot be agreed they will have to be the subject of a separate trial.

I heard oral evidence from some of those who were directly involved at relevant times. The Syndicate called as witnesses Mr Crane and Mr Hugo Morris of SCB. Hannover called as witnesses Mr Perry Crickmere of MRM Hancock, Mr Glenn Partridge and Ms Bevlyn Donohue of Legion, and Mr Jurgen Lang and Mr Uwe Kramp of Hannover. Each of these witnesses was asked to recall events going back to late 1998, and in some respects earlier. It is no surprise that their recollection of events more than 9 years ago was incomplete and sometimes inaccurate. Where witnesses could not recollect what happened they sought to assist me by explaining what they thought would have happened and why, a process which I shall refer to as “reconstruction.”

Hannover tendered in evidence a written statement of Mr Scott Barry of Legion. Mr David Ellis was called by the Syndicate to give evidence about an inspection of Legion’s books which he had conducted in his capacity as a reinsurance consultant employed by Cranmore Adjusters Ltd.

I had the benefit of written reports and oral evidence from Mr Stephen Wills, an expert underwriter called by the Syndicate and Mr Colin Owen, an expert underwriter called by Hannover. Mr Owen has been involved in the personal accident reinsurance market for 36 years. From 1984 to 1995 he was Active Underwriter of Lloyd's Syndicate 718 and, from 1996 to 2001 he was Active Underwriter of Lloyd's Syndicate 1206. During his time as Active Underwriter of Syndicate 718 and Syndicate 1206 he underwrote a broad spread of PA business that included Carve-out cover. Mr Wills was PA class underwriter for Syndicate 861 until 2002. He stated that his experience as a PA underwriter included the provision of London Market Excess of Loss reinsurance, Excess Employers’ indemnity insurance and Carve-out cover. He had been involved in the underwriting of Carve-out business since 1990 and was the leading underwriter on a number of Carve-out facilities and covers.

During the course of 2000 the Syndicate’s 1997 year of account was closed at a loss. It was decided that the 1998 year of account, in which the 1998 Mainframe Carve-outs had been written, should be kept open and that the Syndicate should cease underwriting altogether with effect from 31 December 2000. The Syndicate thus went into “run-off” as regards its 1998 and subsequent years of account, and a run-off manager was appointed. In November 2002 the Syndicate purported to avoid the relevant 1998 Mainframe Carve-outs as against Hannover on grounds of misrepresentation of the estimated premium. Mr Crane’s employment had in the meantime been terminated with effect from the end of 1999. He was not consulted about the decision to avoid the relevant 1998 Mainframe Carve-outs. This purported avoidance was, however, abandoned by the Syndicate in July 2003.

In April 2004 Mr Ellis was appointed to review the Syndicate’s exposures in respect of Legion. Following that review the Syndicate’s solicitors wrote to Hannover on 31 October 2006. Their letter advised it that the present proceedings had been issued and claimed to avoid the relevant 1998 Mainframe Carve-outs as against Hannover “for material non-disclosure and/or misrepresentation …”. At that stage the Syndicate’s solicitors, “without limiting the facts and matters upon which” the Syndicate relied, identified two specific heads of non-disclosure. The first was non-disclosure of “a series of grave concerns by [Hannover] regarding, among other things, Legion’s underwriting and claims handling.” The second was non-disclosure of the fact that “the renewal for 1998 was driven, at least in part, by political factors.” Mr Crane was, again, not consulted about the decision to avoid.

Particulars of claim served on 8 December 2006 repeated these two heads of alleged non-disclosure. In addition, they identified four particular alleged features of the underwriting requirements for the Mainframe Account. The particulars of claim said that false representations were made as to the existence of these features of the underwriting requirements. Further or alternatively they alleged that false representations had been made that Legion was ready and willing to comply with relevant underwriting requirements. The first alleged feature was that these underwriting requirements were “strict and were, in any event, stricter than the requirements for the Cessions Account…”. The second was that actual loss histories provided by prospective insureds were used by Legion in order to calculate expected losses. The third was that expected losses were calculated by Legion “in accordance with a prescribed methodology”. The fourth was that Legion imposed “a substantial retention upon prospective insureds such that they retained a significant portion of their own risk”.

The particulars of claim also included a new third head of alleged non-disclosure. The matters which it was said ought to have been disclosed mirrored the four features relied upon in the misrepresentation claim. As regards the first such feature it was said that Legion failed to disclose that the underwriting requirements for the Mainframe Account “were not strict and were not, in any event, stricter than the requirements of the Cessions Account”. As regards the second, third and fourth such features it was said that there was a failure to disclose that either:

(i)

[the underwriting requirements for the Mainframe Account] did not require Legion to use actual loss histories in order to calculate expected losses, and/or to use a prescribed methodology in order to calculate expected losses, and/or to impose a substantial retention on prospective insureds; or

(ii)

if they did, Legion was not ready and willing and did not intend (or did not intend invariably) to comply with such underwriting requirements.

The allegations in the particulars of claim as to misrepresentation, and as to the third head of non-disclosure, were supported by schedules identifying complaints about 20 risks allocated to the 1998 Mainframe Account. In relation to the second, third and fourth relevant features the schedules said that Legion “did not, (or did not invariably)” comply with the relevant alleged underwriting requirement, and relied on this to demonstrate that at the Hannover/Syndicate contract date Legion “did not intend (or did not intend invariably)” to comply with that requirement. The particulars of claim said that the assertions in the schedules were made “pending disclosure and/or the provision by [Hannover] of further information.”

Draft amended particulars of claim served in June 2007 added a fourth head of non-disclosure. This asserted that the Syndicate ought to have been informed, but was not informed, that Legion failed to issue underwriting guidelines to program administrators or that program administrators had failed to follow them.

Amended particulars of claim served by the Syndicate on 18 January 2008 incorporated the new fourth head of non-disclosure and gave particulars by reference to nineteen of the risks allocated to the 1998 Mainframe Account (some of them being risks identified in relation to earlier complaints). The amended particulars of claim also added to the particulars of the allegations made earlier.

Re-amended particulars of claim were served on 29 February 2008, just over a month before the trial was due to begin. They involved two changes to the Syndicate’s case on non-disclosure. One change involved a deletion: the second head of alleged non-disclosure (referring to a “political” motive for renewal on the part of Hannover) was abandoned. The other change involved an addition: the Syndicate’s fourth head of non-disclosure was expanded. The expansion picked up on references made in Hannover’s witness statements to the effect that “program business” could be allocated to the Mainframe Account in accordance with what was described as an ‘underwriting box’. The new part of the fourth head of non-disclosure said that if this was the case then there was a culpable failure to disclose “the existence of the ‘underwriting box’ approach and/or the manner in which that approach operated.” Those familiar with Lloyd’s may find the use of the term ‘underwriting box’ surprising, for active underwriters have boxes at Lloyd’s where brokers will present the business they seek to have underwritten. In the present context the term ‘underwriting box’ however has nothing to do with Lloyd’s and refers instead to a special set of criteria to be met before program business could be allocated to the Mainframe Account.

The parties undertook at two interlocutory hearings in the run-up to trial to seek to agree a List of Issues. Despite this no agreed final List of Issues was prepared. At the start of the trial I had a draft List of Issues which contained one suggested issue in square brackets and said of another that it was “the least developed” of the Syndicate’s allegations. Uncertainty as to revisions to the Syndicate’s case meant that in this respect, as in others, the position at trial was far from ideal.

At trial Mr David Edwards QC and Mr Jawdat Khurshid appeared on behalf of the Syndicate. Their written outline opening for the trial was served on 27 March 2008. This set out the Syndicate’s case on the four alleged misrepresentations. In relation to the alleged representation that underwriting requirements for the Mainframe Account required Legion to use actual loss histories the Syndicate added “for the avoidance of doubt” that this meant “the whole of the insured’s actual loss history and not just part of it.”

As to the Syndicate’s case on non-disclosure, the second head of non-disclosure had by this time been abandoned. The Syndicate’s case on the first and third heads of non-disclosure was set out in line with the re-amended particulars of claim. In relation to the fourth head of non-disclosure, the Syndicate’s outline opening advanced a case which made no reference to the original way in which that head had been put – there was no reference to Legion not issuing underwriting guidelines to program administrators or failure by program administrators to follow those guidelines. The only aspect of the fourth head of non-disclosure advanced by the Syndicate in its outline submissions concerned the ‘underwriting box’. Here it was positively asserted that Hannover had from 1996 onwards permitted Legion to use an ‘underwriting box’ for program business with less than 5 years of loss history, and that this was not disclosed to the Syndicate. This assertion was supported by a further witness statement from Mr Crane served the following day.

In relation to its particulars under the first head of misrepresentation and the third and fourth heads of non-disclosure the Syndicate acknowledged that many of the particulars concerned things which happened after the date when the 1998 Mainframe Carve-outs were agreed. It was suggested that the court should draw an inference from these later events that they reflected Legion’s underwriting intentions at that date. The Syndicate added that Mr Ellis had been able to inspect a “sample” of Legion’s files of risks forming part of the 1998 Mainframe Account, and contended that there was no reason to think that the sample was unrepresentative of that Account.

This contention, among others, was strongly countered in Hannover’s written outline opening, prepared on 28 March 2008 by trial counsel, Mr Mark Templeman QC and Mr Andrew Neish. They observed that there was no onus on Hannover to demonstrate that the “sample” was unrepresentative: on the contrary, it was for the Syndicate to prove its case.

In written closing submissions served on 18 April 2008 the Syndicate refined its case further. It said that only three complaints were pursued. I shall refer to these three complaints as “the Syndicate’s final complaints.” They can be summarised as follows:

The matters described below are, with some important exceptions, common ground. Where there are disputes between the parties I shall seek to summarise the rival contentions before explaining how I consider that they should be resolved.

Mr Partridge was an underwriter with experience of developing what were known as “Alternative Risk Transfer” (or “ART”) insurance products. In 1983 he joined MRM and was involved in the development by MRM’s marketing arm, Commonwealth Risk Services (“CRS”), of an off-shore ART insurance product trademarked by CRS as its “Insurance Profit Center” (or “IPC”) product.

The IPC product involved the use of a Bermuda based “rent-a-captive” structure. Original insurance policies were issued to insureds by fronting U.S. insurance companies which were external to the MRM group. Those companies paid premium to and were protected by an MRM reinsurance company, Mutual Indemnity. MRM invested the premium to create a fund from which losses up to specific and aggregate limits were paid. Losses above these limits were protected by external reinsurance. If the account was profitable, the original insured participated in the underwriting profit by payment of a dividend under the terms of a Preferred Shareholder Agreement between the insured and MRM. In a profitable year part of the premium was thus, in effect, returned to the insured. In this way an insured could gain some of the benefits available to those whose insurance business was sufficiently large to justify the expense of setting up their own “captive” insurance companies.

In 1987 MRM acquired Legion to act as MRM’s own “issuing carrier” – that is, to act as the fronting reinsurer in place of the external companies previously used. From 1990 until Legion’s liquidation in 2002 Mr Partridge was Executive Vice President and a director of both Legion and MRM with overall responsibility for underwriting. In 1990, Mr Barry joined Legion as Manager-Underwriting, and from 1996 until 2002 he was a director of Legion and Vice President, Manager-Underwriting.

Beneath Mr Barry in the Legion underwriting hierarchy were its day-to-day underwriters. At the time material to these proceedings, Legion’s underwriters included Ms Donohue, Mr Eric Brossard, Ms Karen Ridder, Mr Jonathan Isakoff and Mr Jeff Rudd.

Initially external reinsurance was provided on a facultative basis. In 1991 Legion’s then broker, EW Blanch (which was later replaced by Towers Perrin), suggested that treaty reinsurance be obtained. Skandia was the first reinsurer of what then came to be known as Legion’s traditional or Mainframe Account. Reinsurance Pricing Guidelines (“the RPG”) were an integral part of this reinsurance. In 1992 Hannover became involved, and Skandia ceased to be involved, in reinsurance of Legion’s Mainframe Account.

At this stage most risks in Legion’s Mainframe Account were single insureds. Typically they would be corporations big enough to have an annual premium spend of at least US$500,000 but not big enough to have their own captive insurance company.

From its earliest involvement Hannover, through MRM, purchased Carve-out cover in respect of its WCA exposures when reinsuring the Mainframe Account. By 1993, the market for Carve-out cover in London was such that MRM suggested that Legion seek its own Carve-out cover for a proposed new account which would give Legion greater underwriting flexibility than was permitted for the Mainframe Account. This new account became Legion’s Cessions Account.

In 1992 a single Carve-out retrocession was written by the Syndicate for reinsurers of Legion’s Mainframe Account. A similar Carve-out retrocession was again written in 1993, along with two Carve-out covers reinsuring Legion’s Cessions Account. In 1994 three Carve-out retrocessions were written for reinsurers of the Mainframe Account and a further three Carve-out reinsurance contracts for the Cessions Account. In some or all of the retrocessions the reinsured included Hannover. During this period it was Mr Crane who was responsible for the Syndicate’s participation. The Syndicate was not offered Legion business at the end of the 1994 year of account. Increased capacity in Bermuda had caused a drop in rates enabling the Carve-out covers to be placed at better terms in that market.

By mid 1996 Legion had become concerned that the market for workers compensation business had softened markedly. This made it difficult for Legion to retain existing business and even more difficult to secure new business. Like other U.S. casualty insurers, it was receiving fewer submissions for individual businesses, and more for program business, than in prior years. Mr Partridge’s witness statement described program business in broad terms as involving any of the following:

(i)

A book of business controlled by an agent which he chooses to put into a rent-a-captive so as to enjoy the potential underwriting and investment benefits of the arrangement (often described as an “agency captive program.”) Such a book of business could consist of either homogeneous or heterogeneous insureds. The risk taker under these arrangements would be the agent who would normally be responsible for collateralizing some or all of the “gap”, often by way of a Letter of Credit or deferred commission, and it is he who would receive the profits in the event of good results.

(ii)

An association, by contrast, is typically a group which has something in common (for instance, restaurants or truckers or schools). Often, but not always, this was homogeneous business, where it was the association which would put up the collateral and receive the profit.

(iii)

Other types of programs typically involved specific lines of business (for instance auto or workers compensation) and tended to involve an MGA, or “Managing General Agent,” which had, or was starting, a book of business. Such business would generally attach over the course of an underwriting year.

During the period prior to 1 October 1996 the RPG required that on Mainframe policies expected losses should be calculated by Legion using the Average Worst Average of previous years. Also during that period program business of the kind described above was not permitted to be allocated to the Mainframe Account. Legion sought Hannover’s permission to make changes in these and other respects when the Mainframe XL contracts came up for renewal on 1 October 1996. Hannover dealt with these requests in two stages. By a fax to MRM Hancock of 1 October 1996 Hannover agreed to various changes to the RPG, including as follows:

-Allow Scott Barry only to sign-off on the substitution of a Straight Average expected loss calculation rather than the Average Worst Average. Average Worst Average will continue as standard practice.

As regards program business, a fax from MRM Hancock to Hannover Re dated 20 September 1996 proposed that Hannover “allow Scott Barry only to sign-off on the creation of an ‘underwriting box’ for prospective program business…” This proposal was subject to a proviso that the individual risks attaching to the program were to be individually underwritten by Legion in accordance with the RPG. MRM Hancock added:

We trust you would agree that these proposals are not dramatic, but rather adjustments that will provide Mr Barry with the flexibility to put business into the Main Treaty that he feels would be attractive to Legion’s core reinsurers.

Hannover sought more information on the proposal to allow program business to be included in the Mainframe Account. Its fax of 1 October 1996 said this:

As regards the proposed ‘underwriting box’ for prospective program business we would be interested in receiving some additional information about what kind of accounts would be eligible for this “underwriting box.” We would especially like to learn whether these accounts would be truly new business for Legion or would it also be used to cede business to the Main Frame Program which now has to be ceded to the Cessions Treaty.

In order to answer Hannover’s queries MRM Hancock obtained assistance from Towers Perrin. In a fax dated 24 October 1996 to Mr Crickmere, Mr Hole of Towers Perrin stated:

On the issue of business moving from the Cessions Treaty to the Mainframe, Legion advises that this has and will continue to happen. The likely scenario in which business would move is in the event that the account, through additional loss experience, can now meet the Mainframe Treaty guidelines or has grown in payroll such that the minimums in the Mainframe are not a problem for the account. We believe that a movement of business from the Cessions Treaty to the Mainframe Treaty is to the Mainframe Treaty reinsurers advantage. Presumably, the account has proven successful and/or has grown such that it now qualifies for the Mainframe Treaty

As respect the “Underwriting Box.” Legion would typically use such a program structure when they encounter a controlled book of business, usually through an association or agency, of at least $1,500.000 in primary premium volume. Think of this ‘underwriting box’ as a facultative reinsurance facility within a treaty. Legion will develop certain underwriting parameters to price the ‘underwriting box’ that it will bind into the Mainframe Treaty. When the ‘underwriting box’ is bound into the treaty, typically the individual account loss experience is not immediately available for all business that would ultimately go into the facility, as these programs are set up to exploit market opportunities such as depopulating a residual market or selecting the “cream of the crop” from an agency’s book of business. Legion will set up an individual IPC program that will be funded by an association or agency who will bear risk in the program for the first $250,000 or $350,000… etc each and every loss with an aggregate attaching at, for example, 80% of reference premium. The following points indicate how an ‘underwriting box’ would be set up:

•

Broker would contact Legion with a legitimate opportunity to write a controlled book of business of at least $1,500,000 in projected or actual reference premium.

•

Book of business to have no OD, CT exposed class codes, and all classes must comply with Mainframe Treaty Underwriting Guidelines.

•

Business could be heterogeneous or homogeneous, single state or multi-state.

•

Working with the broker. Legion would generate a state by state premium estimate.

•

Legion would use the above information to calculate taxes and develop the expenses and aggregate attachment point from a chosen loss ratio.

Once the box is set up, Legion Underwriters will individually underwrite each new piece of business going into the facility based on not only the traditional Mainframe Treaty reinsurance pricing guidelines but also on the more strict ‘underwriting box’ guidelines. Thus, though the ‘underwriting box’ itself had no loss experience, the business ultimately bound into the box does have, in fact, such business will have to have enough experience to otherwise qualify it for Mainframe facility.

Mr Hole’s fax then gave an example of an ‘underwriting box’ criteria grid, and continued:

You may well ask why the ‘underwriting box’ is needed at all; why not put the individual accounts into the Mainframe Treaty on a stand-alone basis? There are a few reasons why this would not work. One, the individual accounts may be too small on their own and must be combined with like accounts to make the IPC work efficiently. Two, the concept for the program, whether stemming from an association or agency, is, in essence, owned by that association or agency. Since such association or agency controls the business, they will only structure an arrangement in which they will take underwriting risk and profit from the good business that they ask Legion to place into the facility.

On 25 October 1996 Mr Crickmere forwarded to Hannover a copy of Towers Perrin’s fax of the previous day. Mr Crickmere’s covering letter said in this regard:

As regards your ‘underwriting box’ question, please find enclosed a copy fax received from Towers Perrin which gives a full explanation of this. The comments in this fax regarding the movement of business from the Cessions Treaty to the Main Treaty are more of a general nature rather than specific to your underwriting box question, however they are still valid. For any existing account of this type under the Cessions Treaty it is unlikely that they will move at present. The situation when Legion would consider moving one to their accounts would be when they have matured enough to be able to underwrite this under the Main Treaty guidelines as a program with sufficient loss experience and income to justify this move. It therefore seems unlikely that this will happen, at least in the short term; additionally the greater flexibility offered by the Cessions Treaty will make a decision to move an account difficult to justify from both Legion and the clients’ point of view.

Also on 25 October 1996 Mr Crickmere sent to Hannover a formal letter setting out points that had been agreed. The points listed included the following:

•

Allow Scott Barry only to sign-off on the substitution of a Straight Average expected loss calculations rather than the Average Worst Average, Average Worst Average will continue as standard practice …

•

Allow Scott Barry only to sign-off on the creation of an ‘underwriting box’ for prospective program business provided that the individual risks attaching to the program are individually underwritten by Legion in accordance with the Pricing Guidelines. Primary focus to be new business.

This second letter of 25 October 1996 asked that Hannover indicate its formal agreement by signing and returning a copy of the letter to MRM Hancock. This was duly done by Hannover on 1 November 1996.

Mr Denis Burk conducted underwriting audits of Legion on behalf of Hannover. In September 1997 he did so with assistance from Ms Claudia Born. The executive summary of his report included the following:

There was not much changed from last year’s audit. They still do not do an adequate job of documenting the files to reflect the underwriter’s evaluation of the risk. 98% of the file documentation is directed toward the mathematical development of the needed premium based on loss information developed. There is no real evaluation of the exposures or their controls. While that approach is reasonable for large individual accounts that can be loss rated, it is not a good approach for smaller risks or small risks written under a program, because the loss information in those cases is generally not going to be credible. Since, however, most program business (which is the area they have been growing in) is not written in our mainframe (traditional) treaty this is not a critical concern.

For the same reason and for the same business, there may be a question of the adequacy of the rates. Again, however, for our treaty it is not critical as long as the program business does not get written in our treaty.

…

Because most of the problems I have with the operation is for things that do not have much impact on our treaty, I do not feel we need to take any action on our current participation unless they begin to write more program business in our treaty.

In the body of his 1998 report Mr Burk stated:

Underwriting Adequacy

I have become less and less comfortable with the lack of real underwriting on this treaty. It is still just number crunching based on prior year losses and, for a large individual account that might work out but for the program that is being written more and more and that is coming more and more under this treaty, I do not believe it does an adequate job. There is no actual exposure analysis being done. Therefore, if there is an exposure to a risk has that has not yet generated a loss, the Legion underwriter does not address it (and may not even know of it). This opens Hannover up to the infrequent but severe loss that will eventually come out of any multi-party program book. Evidence of severe losses occurring that were not anticipated by underwriting can be seen in a couple of the files reviewed above. Many of the program risks are not of the same type within one program and, therefore, the exposure analysis would be drastically different between them. Yet this is not even considered. Scott says that most of the program business written is of a homogeneous nature, so this should be less of a concern, however, he also admits that it is not watched as closely as it should be. Because he says it is mainly homogeneous, he still feels their loss rating approach will work. I would agree to a certain extent, if I truly felt the program really was or would remain homogeneous. I am not sure I buy into that.

I can no longer say that the underwriting of the business subject to this treaty is adequate and I do not foresee that changing.

Rating Adequacy

As the rating is very difficult to follow, this is an area that is difficult to assess. On a loss rating basis, it does appear to be adequate. However, as mentioned, I don’t believe this rating approach on the program business they are writing is adequate.

...

Overall Opinion

I do not believe they are doing the type of underwriting that they need to do on the program business that is making up an increasing percentage of their book. As long as they continue to loss rate on this business, they will continue to overlook basic underwriting risk specific exposure analysis that should be occurring on it.

Recommendation for Head Office

I can no longer say that we should remain on this program on the first excess of loss and because of the severity of potential of loss that could occur, we may consider retiring from the second excess of loss as well.

A claims handling audit of Legion was undertaken on behalf of Hannover by Mr Ryan on 22 and 23 July 1997. His report was generally positive, subject to a concern about the size of the workload and the burden on Legion’s staff. Legion, as was common practice in the US health insurance market, made arrangements for third party administrators to handle various aspects of claims processing. Mr Ryan’s report concluded that “Legion is doing an excellent job overseeing a TPA [third party administrator] program involving over 200 separate TPAs”. Mr Ryan’s principal reservation was about the workload compared to staff numbers. His recommendations in this regard involved raising claims settlement authority levels for TPAs and increasing Legion’s own claims staff.

A further claims audit of Legion was undertaken by Mr Ryan on 18 and 19 August 1998. In the executive summary to his 1998 report Mr Ryan stated:

This most recent audit looked only at workers compensation files. The results of this audit were not at all favourable. ... The companies and TPAs are supposed to be complying with published Legion claims guidelines and claims administrator guidelines. This is not happening in most cases. ...

File documentation was weak and sometimes non-existent. The workers compensation claims I reviewed showed little or no activity from the time the file was set up through closure. The condition of most of the files I reviewed looked like the claim handler/supervisor would simply wait for medical bills and reports and state reports to be received and then close the file. Some files sat without review for months, sometimes 12 to 18 months went by before status requests were sent out. Then, in many cases, the status requests were ignored by the TPAs. In most of the files it was impossible to determine a reserve history; what the reserve was initially set at and the subsequent changes and the present reserve amount. In many cases the closing of the file by payment or denial was not explained or even addressed in the file. There was, for the most part, no proactive involvement in the files on the part of Legion personnel. Status reports were lacking in the files, as were status requests. When status requests were sent out, they were not followed up on when the TPAs ignored them. ...

As was the case last year, I had some serious concerns and dissatisfaction with the claims workload as it now stands … As was the case last year, reserve and payment information and adjuster log notes were missing from most of the files. ...

The workers compensation files at Legion were very difficult to review, to say the least. They were poorly documented and poorly handled / supervised. In my opinion the condition of the files is a direct result of the unmanageable workloads, workloads that have increased since last year. I feel it is only a matter of time before one of these files becomes a very big problem (bad faith or punitive damages) due to Legion’s handling or should I say their lack of handling. ...

None of the recommendations I made after last years audit were implemented. The condition of the files reflects this situation. Based on the results of the recent underwriting review and unless some immediate changes are made in staffing, documentation, and reporting procedures and the new system is implemented, I feel that we should look very closely at this treaty as to whether or not we want to continue on it at the renewal date.

Following Mr Ryan’s 1998 claims handling audit Mr Klaus Kune, a senior claims counsel at Hannover, was asked to carry out his own review of Legion’s claims handling during a two day visit to Philadelphia in mid-September 1998. Mr Kune reported in a memo dated 14 September 1998:

I partially agree with Dennis Ryan’s assessment … Legion appear not to provide updates, evaluations or valuations or apply a proactive approach. This by itself should not prevent the renewal of the contract provided that it is profitable. …Legion cannot necessarily assume that the numerous TPAs work reliably. In this respect it is clearly imperative that the problems associated with claims handling are thoroughly brought home to Legion.

…

Finally, I should like to emphasise once again that the poor claims handling by Legion, in the absence of other reasons, ought not be the sole reason for a non-renewal. It has to be made clear to Legion, however, that the quality of their claims handling must be improved as quickly as possible.

Hannover’s documentary records include a note made by Mr Kramp concerning a meeting in Hanover between Mr Barry of Legion, Mr Hancock of MRM Hancock, Mr Hole of Towers Perrin, and Messrs Rohlf, Lang and Kramp in September 1998. The note does not record the precise date of the meeting, but witnesses told me that it must have been on either 15 or 16 September 1998. Mr Kramp’s note records among other things a continuing shift towards program business, and a continuing decline in new and renewal business in a still softening market. It ends with a reference to concerns held by Legion about the acquisition by Hannover of a US insurer, Clarendon Insurance Company (“Clarendon”). At that time Clarendon was one of Legion’s principal competitors. At the meeting Hannover assured Legion that their current business relationship would not be affected by the acquisition, with Legion continuing to be viewed as one of Hannover’s major partners. Concerns expressed by Legion about confidentiality were met by an assurance on the part of Hannover that no Legion business would be shown to Clarendon. Mr Kramp’s note made no reference to the concerns that had been expressed in the underwriting and claims handling audits.

On 18 September 1998 Mr Crickmere sent to SCB an email attaching files to enable SCB to produce what Mr Crickmere described as “the full Cessions Treaty information package.” This document was to comprise a cover page and index, along with four sections. Section 1 had three parts: “Overview,” “Operations” and “Management Profiles”. It included colour diagrams showing a “typical IPC structure”, the general structure of the Mainframe and Cessions Carve-out covers, and detailed structure of the arrangements for aggregate Carve-out cover. Section 2 gave an overall summary of premiums and claims, section 3 dealt with specific claims, and section 4 dealt with aggregate claims.

Later that day Mr Crickmere sent an email to SCB in order to enable it to produce a similar document for the Mainframe Carve-out covers. The attachments comprised a cover page and index, section 2 which gave a business profile, and material for section 3, “Loss Experience” comprising tables of aggregate incurred losses, aggregate paid losses, a specific incurred loss summary, a specific paid loss summary, and a specific loss listing. Section 1 was not attached to Mr Crickmere’s second email: he instructed SCB to use the material which he had attached in this regard when sending his earlier email about the Cessions Treaty.

I shall refer to the documents which were thus duly compiled as “the 1998 Carve-out renewal proposals”, and individually as “the 1998 Cessions Carve-out renewal proposal” and “the 1998 Mainframe Carve-out renewal proposal”. Reliance was placed by the parties on various parts of these documents. I shall set out relevant parts when necessary below.

An initial indication of terms for the Mainframe Carve-out specific first and second layers and aggregate first and second layers, and for four proposed Cessions Carve-out layers, was given by Mr Crane to Mr Hugo Morris of SCB on 25 September 1998. It is probable that Mr Morris had with him the 1998 Carve-out renewal proposals along with slips prepared by SCB using the format and rates applicable to similar contracts for the period from 1 October 1997 onwards. Neither Mr Crane nor Mr Morris had any clear recollection of this event. Nor has any note of it been found. Both, however, accepted that it had occurred as described in a fax sent later that day by Mr Morris to a colleague.

… we have been to see Underwriters … and they have indicated the following terms:

[Terms were then set out including rates which were quoted as a percentage of “applicable excess premium.” These rates were 74% on each of the Mainframe Carve-out specific layers and 75% on each of the Mainframe Carve-out aggregate layers, along with rates of 70% on the Cessions Carve-out first specific and first aggregate layers, 95% on the Cessions Carve-out second specific layer and 51% on the Cessions Carve-out second layer aggregate.]

…

All these indications are subject to:

1.

Underwriters receiving a satisfactory definition of Applicable Excess Premium.

2.

Underwriters receiving the Applicable Excess Premium Estimates for all layers for 1997/98 and 1998/99 years of account.

3.

Underwriters receiving a satisfactory answer as to how the Applicable Excess Premium is calculated? I.E: How do Legion allocate each dollar received and arrive at the Applicable Excess Premium for both the specific and aggregate layers for all parts of this reinsurance placement? Please can we have a detailed breakdown?”

There is a dispute as to the extent to which Mr Crane on 25 September 1998 actually read parts of the 1998 Carve-out renewal proposals. I shall deal with that dispute when examining the Syndicate’s third final complaint below.

In a fax sent by Mr Kramp to Mr Crickmere on 29 September 1998 Hannover advised MRM Hancock that it would renew its participation in the Mainframe XL contracts. Mr Kramp’s fax included the following:

After thorough review… we are pleased to advise that we will continue our participation with lines as expiring.

Perry, quite honestly this authorization was not as easy as we had hoped, as our underwriting and claim audits have raised serious concerns about the adequacy of Legion’s handling of the business. We owe you a report on these, to specifically outline our findings and recommendations for Legion, however in a nutshell they are as follows:

Claims: Legion relies heavily on TPAs such as Lindsey Morden and Gallagher and Basset however is not able to adequately monitor/control their claims handling, guidelines notwithstanding. The problem seems to be not the professional ability of Legion’s own claims department, but more Legion’s inadequate staffing. The guidelines are too frequently not complied with by the TPAs and Legion is not in a position to follow up or enforce lapses. We understand that even though there is significant growth in the book, there have been no additions to the claims department staff. To compensate for this, we also understand that Legion is contemplating increasing the authority of the TPAs, in order to reduce workloads. We are very concerned that it is only a matter of time before Legion starts to pick up punitive damages/ECO claims. We also understand that Legion’s system is not Year 2000 compliant.

Perry, these findings are very similar to what our auditor found last year. Notwithstanding his exit briefing and assurances given to him that these problems would be addressed, he found virtually no improvement this year.

Underwriting: We challenge Legion’s loss-rating approach for program business. This is reasonable for a large individual risk, however not for a program with non-homogeneous business, which is where they are showing the most growth. Although loss control reports on individual risks are good and useful, it seems that these are rarely available on programs and when they are the quality is not that good. In short, we are concerned that not enough attention is being paid to specific risk exposures within the programs. This is opening up a high probability that within a program there will be an individual risk which has not yet generated a claim, is therefore not priced for (or otherwise addressed) by Legion, but will have an impact through the random, but high severity, claim.

As already mentioned above, we will be sending a summary of findings to you, however want to make you aware of our concerns now. Please also note that we are in fact renewing our lines as expiring, which should be understood as a sign of confidence in Legion and our expectation that they will take appropriate action in the coming year.

Mr Barry was emailed a copy of Mr Kramp’s fax on the morning of 30 September 1998. He replied by return as follows:

Please forward the list of concerns when it becomes available. I will formally answer the concerns at that time. I believe we answered the Underwriting concerns while in Germany. What I believe would be best for Uwe would be to spend some time with us and learn how we do underwrite the Programs.

…

Mr Barry’s witness statement said that he could not now recall events at the time, but gave an explanation of how he thought he would have responded:

Loss rating is, in my view, a reasonable approach to underwriting program business. A program is a compilation of smaller (in premium terms) individual risks. When they are combined together for purposes of calculating expected loss, as was Legion’s usual practice, the expected loss of the group is a fair predictor of future losses. Legion’s underwriters, in the course of their analysis, would look at a spreadsheet showing the past experience of all of the accounts prior to the inception of a program, or for each account as it came up for a quotation during the course of the annual period. The underwriters would modify the losses by utilizing factors which developed the losses to their current dollar values, taking into account exposure growth. Legion’s underwriting model used factors which were actuarially derived (including loss development factors, exposure factors and claim development factors). The underwriters would then make sure that the risks quoted fit within the parameters of the loss ratios and exposures in the program underwriting guidelines. Such guidelines usually dictated a target loss ratio that each individual risk had to meet before it could be written into a program. In other words, business was priced and the aggregates set based on actual account loss history.

As regards whether Legion’s loss rating approach is appropriate when applied to program business, any large account, whether it is an individual corporate account or a program, contains a range of exposures depending on the classifications of employees and their functions. Under the workers compensation system in effect in the various states in the United States, such exposures are assigned rated class codes and Hazard group rankings generally based on data compiled by NCCI. There were four Hazard groups: I, II, III, and IV (although my recollection is that, for purposes of business ceded to the mainframe treaty Hazard group IV was excluded.) The premium for workers’ compensation insurance was based on rated class codes and Hazard groups and the class codes generate a higher rate for higher severity exposures.

Both individual corporate account and programs contain a range of exposures – from clerical on one hand, to trucking and contracting on the other. Therefore, in my view, there is the same chance for a high severity loss in a corporate account as there is in a non-homogeneous program.

For all these reasons, therefore, I do not agree with the conclusions expressed by Hannover Re’s auditor in 1998. The probability, I believe, is that my comments to Hannover Re at the meeting in September 1998 were along the lines which I have expressed above and were accepted by Hannover Re on that basis.

Following receipt of this email Hannover stated in the covering letter to MRM Hancock on 23 October 1998:

As previously mentioned, please find attached the recommendation letter from Dennis Ryan…

As for the record, we have renewed our expiring lines since we are confident in Legion’s abilities, its staff and that our recommendations will find the appropriate action.

Overall, our findings support our notion that Legion should not consider to defer any further supervisory responsibility to independent organisations. On the contrary, we believe that Legion should control these files and follow up on the findings of its audit team.

… we … stress that criticism was not directed against the professional ability of Legion’s claims department. The files we found showed a lack in attention which we believe is due to the workload each claims handler has to handle.

During September 1998 MRM Hancock approached other potential markets for the Mainframe and Cessions Carve-outs. Mr Alec Cramsie of MRM Hancock received queries in that regard from Mr Bryan Muraki of a Canadian underwriting agency called Insurance Services Associates Limited (“ISA”). Mr Muraki sought a copy of the RPG, and asked, “What differentiates the Main Treaty from the Cessions Treaty?”

Mr Crickmere sent a memo to Mr Cramsie answering Mr Muraki’s question. The answer included the following:

Nowadays, not a lot. Historically the Main Treaty needed the full five year loss history for the underwriting guidelines to operate. This precluded Legion from writing Association and program business. Start up operations could not be rated and any insureds with major changes (selling or buying subsidiaries, or mergers etc) would be difficult. The Cessions Treaty enabled Legion to write these types of business, by using some underwriting judgment. The Cessions also allowed some classes that were excluded from the Main Treaty. ... The Main Treaty now allows the creation of an ‘underwriting box’ for association/program business, so we have now come full circle with both books having similar content.

Mr Crickmere was surprised by the rates quoted on 25 September 1998 and queried them with SCB. He supplied further information, including as to losses and experience figures, in a series of emails through to the morning of 2 October 1998. By this time Mr Morris and Mr Crane were both in the United States, where they were participating in a joint visit to SCB clients. An email was sent at 8.59 p.m. that evening by Mr Michael Worsfold of SCB to Mr Crickmere as follows:

Many thanks [for] your earlier email.

I am still awaiting written confirmation from Hugo however it would seem that we have the following indications for the Main treaty Carve-outs:

[rates for the 1998 first specific, second specific, first aggregate and second aggregate Mainframe Carve-outs were set out as 57%, 67.5%, 50%, and 50% of applicable premium income.]…

Trust these to be of interest and look forward to your advices.

Cessions Treaty

As discussed we have some very rough indications from Cigna Re …

You already have the Lloyd’s prices for the Cessions Treaty.

The information required to firm up these indications is fairly onerous therefore could you please advise the customers interest prior to submitting the information requests.

Subsequent communications between MRM Hancock and SCB suggest that on the Mainframe Carve-outs Mr Crane’s “indications” were now subject only to one thing: confirmation that a premium figure of $3m odd was for premium ceded on an expiring 43% line. This confirmation was given to SCB by MRM Hancock in the second paragraph of a fax dated 9 October 1998. The documentary material included a copy of this fax bearing a large tick against that paragraph: Mr Crane accepted that this tick had been put there by him. The fax also gave information about the Cessions Account in three sub-paragraphs. Two of these sub-paragraphs bore similar ticks.

On 13 October 1998 Mr Crickmere emailed Mr Kramp the rates for the Mainframe Carve-outs set out in Mr Worsfold’s fax of 2 October 1998. Mr Crickmere noted that the minimum and deposit premiums had still to be finalised. However there was no suggestion that the rates were subject to any further confirmation.

A compliments slip from SCB dated 14 October 1998 was put in evidence. It stated:

Ian, please find attached the information you requested when providing indications on this account.

No witness was able to identify precisely what had been “attached”. Mr Crane commented in his witness statement that the information attached to this compliments slip must have been satisfactory, otherwise he would not have written the business.

Slips were put in evidence bearing Mr Crane’s initials with a date of 22 October 1998. These slips incorporated the rates recorded in Mr Worsley’s email of 2 October 1998. None of the witnesses was able to tell me how these slips came into being. However Mr Crane accepted that on 22 October 1998 he had indeed scratched these slips. He thought he had done so not in order to accept offers put forward by the prospective reinsured but in order to record offers on the part of the Syndicate which were open to be accepted by the prospective reinsured.

On 23 October 1998 Hannover advised Mr Crickmere that the rates quoted for the Mainframe Carve-out aggregate layers were acceptable, but those for the specific layers were too high. Hannover proposed rates of 54% and 57% for the first and second specific layers respectively.

Also on 23 October 1998 Mr Crickmere emailed Mr Morris expressing concern about the quoted rates for the Cessions Carve-outs. He said that if there were no improvement on pricing, Legion would be more likely to use a facultative route or other facilities, with the danger that this would introduce adverse selection.

On 30 October 1998 Mr Crickmere reported to Hannover that he was not having much luck in getting better pricing. The documentary material contains nothing further of significance until 13 November 1998.

It is common ground that on Monday 16 November 1998 Mr Crickmere and Mr Morris visited Mr Crane. The parties do not agree on what occurred. No note of the meeting was in evidence. There were, however, records of various communications on 13 and 16 November 1998 which may assist in working out what happened.

On 13 November 1998 a fax was sent by Mr Worsfold and Mr Morris of SCB to Mr Hancock of MRM Hancock. It included the following:

… We are pleased to confirm the rates as provided by Clarendon, which we are seeking to be agreed by Ian Crane. He has requested full details concerning Legion’s approach to allocating business between the Main Treaty and the Cessions Treaty.

[The fax then set out Clarendon’s rates. For the 1998 first specific, second specific, first aggregate and second aggregate Mainframe Carve-outs they were 55%, 67.5%, 50% and 50% respectively. For similar layers on the 1998 Cessions Account they were 60%, 80%, 75% and 55% respectively.]

Clarendon have agreed to support with a 33.33% share of the above layers … it appears almost certain Reliastar (Copenhagen) will support these terms. … We would anticipate having (subject to Ian Crane’s agreement) final support of 1/3^rd [Syndicate] 53 and 2/3rds Reliastar (Copenhagen).

At 11.41 on Monday 16 November 1998 Mr Crickmere faxed to Mr Morris a response to the request in the fax of 13 November for “full details concerning Legion’s approach to allocating business between the Main Treaty and the Cessions Treaty.” Mr Crickmere’s response included the following:

Re Legion’s approach to allocating business between the Main Treaty and the Cessions Treaty.

Firstly let us go back to basics. Legion is not a “Risk Taker” as such. Legion/Mutual provide risk management services to their insureds, including the buying of reinsurance for the excess and aggregate risk, with the insured keeping the retained risk in a Captive vehicle.

Any suggestion of selecting for or against any particular Reinsurance Treaty is simply wrong. As Legion is reliant on Reinsurance any action of this nature would be counterproductive, leading to a loss of credibility and ultimately, the loss of their reinsurance facilities.

I would guess that the question of allocating business falls into two parts, renewals and new business. I shall deal with renewals first.

Legion will try to keep all renewals as renewals into the same treaty that was used on the expiring. If they do move a renewal there would have to be a reason to do this. This will be a costs issue and as both Treaties run off the same basic pricing guidelines there [is] no advantage gained by switching treaties.

This is not entirely in Legion’s control. If Legion’s treaties are uncompetitive compared to other options available to the insured (other facilities, guaranteed cost market, facultative market etc.) Legion will explore the alternatives before their insureds do to avoid losing the account.

If an account is moved it will more likely be removed from the Treaties and placed facultatively.

Turning now to new business, simply, each piece of business will use whichever Treaty is the best fit.

The Main Treaty does provide coverage for non WC lines (not reinsured under the Carve Out), so if there is some [automotive/general liability cover] required the Main Treaty will likely be the best fit.

The Cessions Treaty provides more flexibility in the underwriting of a risk than the strict requirements of the Main Treaty. The Main Treaty has more excluded classes and does not cater well with abnormal changes in a risk (e.g. a company who has sold or purchased a large subsidiary changing the make up of the business going forward). The Cessions Treaty allows some underwriting judgment in these cases.

Hugo I trust the above is helpful. If you need additional help please call. As discussed I am available and do have an in depth knowledge of all Legion’s operations and Treaties. If you need me to sit in on any meetings let me know.

At 11.54 on 16 November Mr Crickmere faxed to Mr Hancock a copy of SCB’s fax of 13 November along with his reply. The covering fax from Mr Crickmere to Mr Hancock included the following:

It seems that when Hugo saw Ian Crane on Friday [13 November 1998], he told him that he had to go along with the Cessions pricing or Legion would move all the business to the Main Treaty.

Ian Crane was not impressed with this, as he does the Main Carve-outs at a cheaper rate!

This is why Ian has asked for how Legion allocates risks and, I am told, he has passed comment regarding being selected against.

I have sent Hugo the fax and will chase him up now, I want to go in with him when he next visits Crane (hopefully this afternoon). With any luck we will be able to rescue the situation.

There is then a further fax from Mr Crickmere to Mr Hancock, timed at 17.22 on 16 November 1998. It included the following:

Have just got back from Lloyd’s. Ian Crane has worries about risk selection/anti selection between the Cessions Treaty and the Main Treaty. He favours the Main Treaty as it has had better experience. The upshot was that after the meeting I had a long chat with Hugo. He had in his slip case a signed agreement to the latest Cessions pricing from Cigna (Alan Bird). This with Clarendon and Reliastar gives us 100% cover. ... We will deal with Ian Crane separately. He will be offered the [1998 Mainframe Carve-outs] only and I will go in again to satisfy him with the Main Treaty Underwriting Guidelines showing him the criteria needed for risks to go into the Main Treaty.

In two respects matters did not end up as this fax would suggest. In the event Mr Crane was not “offered the [1998 Mainframe Carve-outs] only” but on the contrary went on to write not only those retrocessions but three Cessions Carve-out covers as well. The other respect in which events did not turn out as the fax suggests is that Mr Crickmere did not in fact return to see Mr Crane and Mr Crane did not at any stage in 1998 see the RPG.

Mr Crickmere’s account of what happened at the meeting, as set out in his witness statement, was as follows:

To the best of my present recollection and as a matter of probability, I believe that I would have explained to Mr Crane the differences between the Mainframe and Cessions Treaties, e.g. the fact that the Mainframe Treaty had more excluded classes and was underwritten pursuant to Pricing Guidelines. I believe that we would also have discussed the make-up of the two books and their inclusion of program business. In any event, I was aware that Mr Crane had underwritten workers compensation business for a number of years. My recollection is that he seemed very familiar with such business, the state of the market in the US and the fact that the business being written by Legion included program and association business. Quite apart from anything else, that point is made in terms in the Underwriting Information [i.e. the 1998 Carve-out renewal proposals] I had prepared (see under “Overview”) a copy of which, I assume, Hugo Morris had provided to Mr Crane. Although I no longer remember the specifics of the discussion with Mr Crane, I believe it likely that I indicated that there was now more underwriting flexibility in the Mainframe Treaty because of the fact that program business was being written in the treaty, but that, otherwise, there were differences between the treaties and that there was no question of the Syndicate being selected against. If we discussed program business (which I feel sure we would have) it seems highly likely that I also mentioned the use of the ‘underwriting box’. Given the misinformation provided by Mr Morris, I needed to make sure that Mr Crane had a full and proper understanding of the position and, although a discussion at his box was not ideal, I believe that that is what I achieved.

This account was in substance maintained under cross-examination. During the course of cross-examination Mr Crickmere accepted that what he had said to Mr Crane would probably have included the points made in his fax to Mr Morris on the morning of 16 November 1998.

As against this, the Syndicate contends that it is “overwhelmingly likely” that Mr Crickmere said no more to Mr Crane than what he had set out in the fax earlier that day to Mr Morris, and in particular made no mention of the ‘underwriting box’ for program business. It relies on Mr Crane’s evidence that he would have recalled any mention of more flexibility, and in particular the ‘underwriting box’, having been introduced into the Mainframe Account. Mr Crane added that he had been under the specific understanding that the underwriting criteria for the Mainframe Account had not changed since he wrote the account in 1992 through 1994.The Syndicate also relies on positive assertions by Mr Morris that the ‘underwriting box’ was never mentioned and that Mr Crickmere was panicking because by this time there had been considerable delay since the start date for the proposed contracts of 1 October 1998.

Mr Crickmere did not have a clear recollection of the 16 November 1998 meeting. Nevertheless I consider that he is likely to be right in his description of what he believed happened, that Mr Morris is mistaken in the account that he gives, and that the belief held by Mr Crane as to what happened is unlikely to be correct. First, Mr Morris himself stated in cross-examination that the Legion Carve-outs were not actually “worth something” to SCB, by which he meant that they were not really important to SCB. He said this in order to rebut any suggestion that he would have threatened Mr Crane, but it seems to me also to make it unlikely that what happened at the meeting would remain clearly in his mind for more than 9 years. Mr Crane frankly acknowledged that he had no recollection of the meeting. It was clear from his evidence that to him the Legion Carve-outs would have been run of the mill. It was suggested that use of the term ‘underwriting box’ in a sense other than its meaning at Lloyd’s would have been so unusual as to make it particularly memorable. I do not believe that would have been so: the focus would have been on the content of the guidelines for underwriting program business, not on the terminology.

Second, it was simply unreal for Mr Crane to suggest in evidence that he had been under the specific understanding that the underwriting criteria for the Mainframe Account had not changed since he wrote the account in 1992 through 1994. November 1998 was four years down the line. There is no basis whatever for thinking that Mr Crickmere or anyone else had said that the RPG were static, and Mr Crane cannot seriously have expected the RPG to have remained unchanged during that period. Indeed, for reasons which I explain in the next main section of this judgment, I think it probable that Mr Crane read the “Overview” part of the 1998 Carve-out renewal proposals. This part included a statement that a growing number of association, agency captive and program business had been written. Turning to the “Business Profile” part of the 1998 Mainframe Carve-out renewal proposal, I consider Mr Crane is likely to have read the historical figures for class distribution. These clearly showed one aspect of the introduction of program business into the Mainframe Account: for the period 1 October 1996 onwards entries appeared under the rubric “Heterogeneous” which was plainly a reference to programs comprising different types of risk. These entries showed that reference premium under this rubric was substantial and growing: $6.9m for 1996/97 and $8m for 1997/98.

Third, there was a dispute as to whether Mr Crickmere’s covering fax to Mr Hancock accurately described how Mr Crane’s concern had arisen. I think it likely that one or both of Mr Morris and Mr Crane had misunderstood what had been said by Mr Crickmere on 23 October 1998. Whatever the cause, the meeting came about because there was a need to discuss Mr Crane’s concerns about being “selected against”. The starting point for any such discussion was that the 1998 Mainframe XL contracts governed what could go into the Mainframe Account, and that these in turn permitted only limited classes of business and in relation to those classes required compliance with the RPG. Mr Crane must have known all these things. He accepted that if he had looked at the RPG he would have seen that they specifically referred to the ‘underwriting box’, and that the ‘underwriting box’ laid down fixed criteria.

Fourth, it is clear from the fax sent by Mr Crickmere after the meeting that he intended to see Mr Crane again and show him the RPG. That would have been quite inconsistent with any intention to conceal from Mr Crane the arrangements under which program business could be allocated to the Mainframe account. In fact he did not see Mr Crane again. In my view the likely reason is that Mr Crane was no longer worried about the criteria for allocation of business to the Mainframe Account. I do not think that Mr Crickmere was panicking in mid-November – he had notified Hannover of Mr Crane’s rates on 13 October 1998, and it was Hannover which thereafter delayed matters by seeking better rates.

Fifth, it is inherently improbable that Mr Crickmere would simply have turned up at the meeting, recounted what was in his fax to Mr Morris that morning, and said nothing more. The obvious purpose for having a meeting was so that Mr Crickmere could give more detail. The fax to Mr Morris dealt with three main topics. The first concerned Legion’s role as a provider of risk management services, which Mr Crickmere said would make it counterproductive for Legion to select “for or against” any particular reinsurance. The second concerned renewals, where any move out of a particular Treaty was likely to be a move to facultative reinsurance. The third concerned new business. After pointing out that each piece of business would use whichever Treaty was the best fit, Mr Crickmere set out in two paragraphs some very general observations about the differences between the Main Treaty and the Cessions Treaty. A point was made that the Main Treaty, unlike the Cessions Treaty, covered automobile and general liability. An assertion was made that the Cessions Treaty provided “more flexibility in the underwriting of a risk than the strict requirements of the Main Treaty”. In support of this it was said that the Main Treaty had more excluded classes and did not cater well with abnormal changes in a risk, for example where there had been a change in the make up of the business going forward. These general assertions were not what Mr Crane had asked for. His request, as described on 13 November 1998, was for “full details concerning Legion’s approach to allocating business between the Main Treaty and the Cessions Treaty.” I think it highly unlikely that at a meeting a few days later Mr Crane would have been fobbed off with the broad general observations in Mr Crickmere’s fax to Mr Morris.

Sixth, I doubt whether the introduction of program business into the Mainframe Account would have come as any surprise to Mr Crane. He was thoroughly familiar with the market, and would have been aware of the way in which the market increasingly gave cover for program business.

For all these reasons I cannot accept the Syndicate’s contentions that at the meeting on 16 November 1998 Mr Crickmere told Mr Crane no more than what had been set out in the fax that Mr Crickmere had sent to Mr Morris earlier that day, and I am satisfied that he told Mr Crane about the arrangements enabling program business to be allocated to the Mainframe Account.

On 1 December 1998 Mr Crickmere emailed Mr Kramp setting out final terms for the 1998 Mainframe Carve-outs. It is common ground that these were rates proffered by Mr Crane on behalf of Syndicate 53, and that Mr Crane had by this time matched Clarendon’s rates. The documentary evidence does not reveal precisely when or how this came about.

In his first witness statement Mr Crane said that it was not clear what had happened after the meeting of 16 November 1998, but it seemed that he had agreed to the revised terms proposed by Clarendon. When asked about this under cross examination he gave more information about the rating system that he had used. He recalled that he would treat figures after 5 years as being “as good as done”. After 4 years he would add a figure, and similarly for intervening years down to the figures for the first year which he would multiply by about 10. That would inevitably produce much worse figures than those in the experience summary for the specific Carve-out provided as part of the 1998 Mainframe Carve-out renewal proposal. As to the passage in the renewal proposal referring to a significant shift from hazard group 2 to hazard group 3 for the 1997/98 year, if the premium was appropriate this would not matter. He was asked about the figures for reference premium as broken down by state in the 1998 Mainframe Carve-out renewal proposal. These showed a loss ratio against 43% of gross premium. If the losses were 20% of gross premium, then the loss ratio was 50% whereas he was charging 57% of gross. Even allowing for inflation and other factors, he would have expected a small profit on the rates that he quoted.

The Hannover/Syndicate contract date has only limited significance for present purposes. It arises in two ways. First, in seeking to demonstrate falsity of particular representations the Syndicate relies on what happened when Legion was considering certain of the risks forming part of the “sample” examined by Mr Ellis. The Syndicate’s case is that these events demonstrate that alleged representations did not describe the true position at the date when the contracts – now sought to be avoided – were made. It is suggested that an inference to that effect may be stronger or weaker depending on whether the event in question occurred before or after that date, and in the latter case depending on how long after. Second, in considering the Syndicate’s case on non-disclosure I must examine the position at the date when the relevant contracts were made. Both parties stress that the precise date may be of limited significance. However as they have been unable to reach agreement on it I am required to determine what the Hannover/Syndicate contract date was. I do so below, before turning to conclude this part of my judgment by making further observations in relation to Mr Ellis’s “sample.”

On 7 December 1998 Mr Kramp replied to Mr Crickmere’s email of 1 December. In his reply Mr Kramp said that Hannover was willing to purchase the 1998 first specific Mainframe Carve-out and the 1998 first aggregate and second aggregate Mainframe Carve-outs. However Hannover was not willing to purchase the 1998 second specific Mainframe Carve-out as the price was too high.

On Thursday 10 December 1998 Mr Crickmere faxed Mr Kramp confirming Hannover’s signed lines on the 1998 Mainframe XL contracts. His fax continued:

As regards the Carve-out protections we have confirmed to Reinsurers your orders on the above signed lines and will forward the cover notes for these as soon as finalised.

It is common ground that when Mr Crickmere faxed Mr Kramp on 10 December 1998 MRM Hancock was not in a position to advise those underwriting the 1998 Mainframe Carve-outs full details of the companies that were to be reinsured and the percentage of the risk that those companies represented. MRM Hancock knew that whatever the eventual order was, in combination the Syndicate and Reliastar would cover it. However, while it knew Hannover’s position, it did not as yet know precisely which other companies underwriting the 1998 Mainframe XL contracts would want to be protected under the 1998 Mainframe Carve-outs, nor did it know what proportion of the risk those companies would wish to have protected.

It is also common ground that by 7 January 1999 MRM Hancock had established the position for all the companies underwriting the 1998 Mainframe XL contracts. On 7 January 1999 Mr Crickmere sent Mr Worsfold an email headed “Legion Main Carve-out Covers.” The first five paragraphs of the email read as follows:

“We now have all the orders in on these.

You will recall that Odyssey Re ultimately said no …

We have increased orders from CORE, who only purchased 50% of their line last year and only purchased the first layer, they now want 100% all layers.

We have, as previously discussed, a new buyer on the second Layer – TIG Re.

So, overall we are down on the first layer, but with larger orders on the second layer.

The remainder of the email referred to attachments setting out the orders by layer and giving the precise names of each individual reinsured company, and reiterated the pricing that had been given in SCB’s fax of 13 November 1998. These were not in evidence in the proceedings before me. Mr Crickmere’s email of 7 January 1999 made no specific reference to Hannover, but no doubt Hannover would have been included in the two attachments.

As regards the three 1998 Mainframe Carve-outs which protected Hannover, it is common ground that the eventual position was as follows;

In each case the respective shares of the Syndicate and Reliastar amounted to 43% and 57% of the total cover given.

The records of the Syndicate include notes of a meeting held on Thursday 14 January 1999, item 2 of which was headed “General business review and large renewals during the last two weeks.” Under this heading an entry appeared:

Legion insurance – new PA Treaty – three layers $1,859,368 …

Mr Crane’s witness statement said that this concerned the “Main Treaty reinsurance business.” I consider that this is unlikely to be right. The reference to three layers strongly suggests, and I consider it likely, that the entry under item 2 concerned the 1998 Cessions Carve-outs.

The Syndicate’s notes of a meeting held on Thursday 28 January 1999 contain, under a similar heading for item 2, an entry as follows:

Legion Insurance - 4 layers - new - P.A. carve out deal ... $295,184

Mr Crane’s witness statement said that the entry on 28 January 1999 concerned the “Cessions reinsurance business.” I consider that this is unlikely to be right. The reference to four layers strongly suggests, and I consider it likely, that the entry under item 2 concerned the 1998 Mainframe Carve-outs.

Cover notes were issued by SCB for the 1998 Mainframe Carve-outs on 17 February 1999, and were received by Mr Crickmere on 8 March 1999.

Documentary material at trial included what appeared to be the final slips for the 1998 Mainframe Carve-outs. It was common ground that they had been scratched and dated by Mr Crane. At trial it was also common ground that the date written on the slips was “20.2.99.” However 20 February 1999 was a Saturday. Both sides now agree that the slips are unlikely to have been scratched on a Saturday. On the copies of the relevant slips it can be seen that the number written for the month is indistinct. I consider it probable that the date in fact written on these slips was “20.1.99” - 20 January 1999 having been a Wednesday. This is consistent with the 1998 Mainframe Carve-outs being noted in the Syndicate’s minutes of 28 January 1999 under the heading “General business review and large renewals during the last two weeks.”

In reliance on the general course of business at Lloyd’s – described in agreed propositions 1 and 2 in annex 1 to this judgment – the Syndicate submits that the Hannover/Syndicate contract date was the date when Mr Crane scratched the final slips. For the reasons given above this would be 20 January 1999.

I readily accept that the general practice is that, as stated in agreed proposition 2, the presentation of the slip by the broker constitutes the offer and the writing of a line by the underwriter constitutes an acceptance of this offer. However that general practice is not uniform – see agreed proposition 3. As noted earlier, Mr Crane when asked about the scratching of the slip on 22 October 1998 said he could not remember the occasion, but characterised it as a way in which he formally recorded that he was offering particular rates. I think he was right to characterise it in this way. As regards the 1998 Mainframe Carve-outs, at all material times what SCB and MRM wanted from Mr Crane were rates that they could put in the first instance to prospective participants in the 1998 Mainframe XL contracts, and eventually to actual participants. To do this they needed, and in my view secured, offers from Mr Crane. His initial offer was conditional in various respects, and it was then the subject of revisions until eventually he matched Clarendon’s rates. It would not have made commercial sense for his initial and revised quotations to have been merely invitations to treat. The Syndicate’s stance would entail that only some time after all orders were known would a contract arise. I do not believe that to be a fair inference from the scratching of final slips on 20 January 2008. In my view each of those slips is much more likely to be a formal record by the Syndicate, for the layer in question, of the names and percentage lines of those who were reinsured under that layer. When placing orders the reinsurers seeking Carve-out cover needed to know where they stood. It is not suggested by the Syndicate that there was any difficulty in working out what proportion of Hannover’s order would be borne by each of the Syndicate and Reliastar respectively. I cannot accept that after placing orders Hannover are to be regarded as in limbo until all other orders are known, still less until formal scratching of a slip after that time. On the contrary, consistently with the course of dealing, it seems to me clear that the orders placed by reinsurers, and by Hannover in particular, constituted acceptance of the Syndicate’s offers on the layers in question.

Mr Crane’s reconstruction of the matter was that initially he thought that he was being asked to provide Carve-out cover directly reinsuring Legion under its Mainframe Account. I do not think this is right, for I believe he would have recalled from what happened in 1992 to 1994 that the Mainframe Carve-outs had been partial protections for reinsurers of the Mainframe Account. The meeting on 16 November 1998 must have proceeded on this basis, for if at that time Mr Crane had thought he was reinsuring Legion directly then the obvious way of dealing with any concerns about undue flexibility as to what could be allocated to the Mainframe Account would have been for Mr Crane to say that he – as Legion’s reinsurer - would tighten the requirements for that account.

For all these reasons I consider that Hannover and the Syndicate were contractually bound on the date when the Syndicate was notified of Hannover’s order. There was no direct evidence as to when this happened: none of the witnesses could remember. However Mr Crickmere’s email to Mr Kramp on 10 December 1998 stated, “…we have confirmed to Reinsurers your orders.” The course of action thus described is exactly what MRM Hancock, through SCB, would have been expected to do. None of the documentary material gives any reason to doubt that it was in fact done. Accordingly I accept Hannover’s submission that the Hannover/Syndicate contract date was on or shortly before 10 December 1998.

The Syndicate relies on files for 27 risks. The Mainframe Account in 1998/99 comprised 71 relevant risks. The Syndicate’s contention is not that the 27 risks represent a random sample, but that they amount to a representative sample by virtue of the proportion of the overall Main Treaty book for 1998/9 which they cover.

As to this contention, the 27 files were not presented to Mr Ellis as “representative.” I am asked in effect to extrapolate from a finding that n out of 27 risks involved a certain event and to conclude, in broad terms, that the event is likely to have occurred in relation to the Mainframe Account in 1998/99 on n x 71/27 occasions. It is said by the Syndicate that the 27 files involved 60% of the original reference premium and 77% of the gross written premium for the Mainframe Account in 1998/99. This seems to me to be nothing to the point. If there were any valid basis for an inference it could only be by reference to the number of occasions rather than the value of the premium. I do not think there is any valid basis for an inference. First, the sample is accepted not to be random. Second, even if it were random, I have no information about how the type of risks comprising the “sample” compares with the type of risks making up the remainder of the 1998/99 Mainframe Account. For these reasons I do not accept the Syndicate’s contention.

There is a separate point which needs to be stressed in relation to Mr Ellis’s “sample.” Each file comprised largely unexplained documents. Some of these documents were put to Ms Donohue. Most of them were undated and unsigned. In large part her evidence was simply her best guess at what happened on a file that she had not had involvement in. She was provided for the purposes of cross examination with a document headed “account files”. It set out the alleged inception date for each of the 27 risks. It did not identify the documents relied upon to evidence the particular events that the Syndicate complained of. Nor did it identify an alleged date for that event. The general approach taken in cross-examination was to ask Ms Donohue to comment on an undated document, but did not usually involve an attempt to identify a precise date for the document. Neither orally nor in writing did the legal teams go into any great detail about the dates of the events complained of. I do not say that counsel ought to have done this – to do so would have greatly increased the costs of the trial with little real benefit. Moreover Ms Donohue had travelled from America and was available for a limited period of court time, all of which was used up in a cross-examination crammed with detail. What I wish to stress is that where the “sample” was relied on the material available was extremely limited, and that for the purposes of this judgment I have tried to identify a date for documents in the sample only to the extent that it seemed to me proportionate to do so.

Here I am concerned with whether the Syndicate can avoid by reference to what was or was not said to the Syndicate about the use by Legion of actual loss histories provided by prospective insureds in order to calculate expected losses, and the calculation by Legion of expected losses in accordance with a prescribed methodology – along with the impact on these matters of what was done by Legion under the ‘underwriting box’ arrangements.

The Syndicate’s primary case is that representations were made to the Syndicate because Mr Crane was shown the 1998 Carve-out renewal proposals, including both the “Overview” and “Operations” sections. In support of each of the relevant representations the Syndicate cites passages in the 1998 Carve-out renewal proposals. I set them out below in italics. In order to put them in context I have included some additional passages which are not in italics.

The first alleged representation in this regard (“the actual loss history representation”) is that the underwriting requirements for the Mainframe Account required Legion to use actual loss histories provided by prospective insureds in order to calculate expected losses, and Legion was ready and willing to comply with the requirement for actual loss histories and intended to do so. The relevant passages in the 1998 Carve-out renewal proposals are:

(1)

in the “Overview” section:

“... the key to the IPC approach is recognition of the insured’s own loss results. This begins with the use of the insured’s historical losses in determining the expected losses (and loss ratios) for the upcoming account year”; and

(2)

in the “Operations” section:

“The underwriting process essentially consists of screening interested prospects to determine whether the individual insured satisfies the various criteria which make a substantial retention feasible from an economic standpoint. …

Legion requires information from the prospective insured … This generally includes:

1.

Five years of exposure data and loss history, valued as of a recent date. Broker summaries of loss histories are not accepted - prior carrier “hard copy” loss runs and description of large losses over $50,000 are required.

…

Many other details are reviewed and evaluated in the underwriting process which are too numerous to describe here except in general terms.”

The second alleged representation in this regard (“the prescribed methodology representation”) is that the underwriting requirements for the Mainframe Account required Legion to calculate expected losses in accordance with a prescribed methodology, and Legion was ready and willing to comply with the requirement for the use of a prescribed methodology and intended to do so. The relevant passages in the 1998 Carve-out renewal proposals are:

(1)

in the “Overview” section:

“...the insured’s known occurrence losses ... are trended and developed using various factors to determine an “expected loss.” The factors are agreed with reinsurers and defined in the Reinsurance Pricing Guidelines booklet which is periodically reviewed and updated to stay abreast of changes in benefit levels, medical trends, etc.. The expected loss is the basis for setting the annual aggregate attachment point as well as determining the specific and aggregate reinsurance premium”; and

(2)

in the “Operations” section:

“The underwriting process essentially consists of screening interested prospects to determine whether the individual insured satisfies the various criteria which make a substantial retention feasible from an economic standpoint. …

…

In the evaluation format used by Legion both aggregate losses as well as individual large losses are summarised.

The losses are “normalized” using actuarially based exposure factors, loss development factors, and claims cost factors to evaluate “adjusted” losses for each year. Average, worst case and best case losses are then calculated and payout ratios are compared with industry averages. …

Also included is a calculation of “aggregate attachment point” which incorporates the preceding analysis and includes a “comfort factor”, in addition to expected losses,

…

Many other details are reviewed and evaluated in the underwriting process which are too numerous to describe here except in general terms.”

Hannover accepted that Mr Morris probably took the 1998 Carve-out renewal proposals with him to his meeting with Mr Crane on 25 September 1998 and gave them to Mr Crane. It contended, however, that there was no evidence that prior to the Hannover/Syndicate contract date Mr Crane read the passages now relied upon by the Syndicate. It is true that I have no direct evidence of this. Mr Crane could not recall actually reading specific passages in the “Overview” and “Operations” parts of the 1998 Carve-out renewal proposals. Hannover submitted that Mr Crane’s approach to underwriting was such that he would have had no interest in these parts. I shall examine that contention in detail when I consider below the question whether Mr Crane was induced to make the 1998 Mainframe Carve-outs by what was said in the relevant passages. In fact – see paragraph 59 above- the “Overview” and “Operations” parts of the 1998 Mainframe Carve-out renewal proposal and the 1998 Cessions Carve-out renewal proposal were identical. These parts of the renewal proposals would have taken no more than ten minutes to read. It seems to me that whatever his approach to underwriting Mr Crane would not have been so foolish as to assume Legion’s business to have remained unchanged since he was last involved with it in 1994. It seems to me to be inherently probable that Mr Crane wanted to know how Legion’s business had changed and for that purpose read these sections in one or other of the two documents which Mr Morris brought with him.

I can deal shortly with the next point taken by Hannover, which is that the representations were no more than statements of Hannover’s expectation and belief. This is simply not borne out by the format and structure of the 1998 Carve-out renewal proposals. As mentioned earlier, the “Overview” and “Operations” sections were identical in each of the two documents. So far as the Cessions Account was concerned, the document gave an account on behalf of Legion as to the nature of its business and the way in which underwriting was carried out. Exactly the same account was given in the document which MRM Hancock had prepared on behalf of the 1998 Mainframe XL reinsurers when seeking their Carve-out protections. No qualification was put forward to the effect that the reinsurers were simply reciting what they believed to be the case. In all the circumstances it appears to me plain that MRM Hancock simply adopted on behalf of the reinsurers what it had prepared on behalf of Legion.

This reasoning does not, however, carry with it any assumption that the 1998 Mainframe XL Reinsurers were representing that Legion was ready and willing to comply with the relevant requirements and intended to do so. There was no express representation in this regard. Nor can I see any basis to infer such a representation. The 1998 Mainframe Carve-out renewal proposal sought only to describe the underwriting requirements as they currently existed. There was no representation to the effect that the requirements as so described were immutable. On the contrary, it was said expressly that the RPG were periodically reviewed and updated. In these circumstances I do not think there was any implicit representation as to what would happen in the future. For these reasons I conclude that only two relevant representations were made on behalf of the Mainframe XL Reinsurers:

When considering whether these were misrepresentations entitling the Syndicate to avoid I shall adopt the structure used in the Syndicate’s written closing submissions. This started with particular instances said in Schedule 2 to the re-amended particulars of claim to demonstrate that Legion did not (or did not invariably) use actual loss histories provided by prospective insureds in order to calculate expected losses. It then turned to particular instances said in Schedule 3 to the re-amended particulars of claim to demonstrate that Legion did not (or did not invariably) calculate expected losses in accordance with a methodology prescribed by the RPG. Next it put forward a case, not expressly advanced in Schedules 2 or 3, that what happened in relation to the ‘underwriting box’ demonstrated the falsity of both the actual loss histories representation and the prescribed methodologies representation. I shall adopt the same structure, and then deal with questions of materiality of the representations, reliance on them by Mr Crane, and non-disclosure.

There is a short and complete answer to the Syndicate’s assertions in Schedule 2. Those asserted complaints are simply as to what allegedly happened. What allegedly happened is then said to show as a matter of logical inference that either (1) the underwriting requirements for the Mainframe Account did not require Legion to use actual loss histories in order to calculate expected losses or (2) Legion was not ready and willing and did not intend (or did not intend invariably) to comply with such underwriting requirements. However for the reasons given earlier I consider that no representation was made by Legion as to the second of these alternatives. It follows that the logical inference asserted in Schedule 2 does not demonstrate falsity of the limited actual loss history representation. The Syndicate’s final complaint under this particular head depends upon there having been a representation as to Legion being ready and willing to comply with the requirement for the use of actual loss histories and intending to do so. For the reasons I have given earlier there was no such representation, and for that reason alone this particular head of complaint fails.

Even if, however, the actual loss histories representation had been as broad as the Syndicate claims, I would not have held that Schedule 2 demonstrated that it was false. This is because the Syndicate’s allegations of falsity involved giving two special meanings to the actual loss histories representation, namely that for the purpose of calculating expected losses Legion would not make any adjustment to the actual loss history, nor would it use a loss history of less than 5 years. In my view it is not possible to derive these special meanings from the 1998 Carve-out renewal proposals. As regards actual loss histories, the “Overview” section simply said that recognition of the insured’s own loss results began “with the use of the insured’s historical losses in determining the expected losses…”. This said nothing about the precise way in which the insured’s historical losses would form part of the determination of the expected losses. All that was represented was that use of the insured’s historical losses was a starting point. Similarly, the “Operations” section did not have the special meaning contended by the Syndicate. The passage to the effect that “information from the prospective insured… generally includes … five years of … loss history” was plainly a representation only as to what “generally” happened. The use of the word “generally” connotes that there will be occasions when it does not happen. Nothing in the 1998 Carve-out renewal proposals warranted the assertion in the Syndicate’s written opening that what was said about the use of actual loss histories meant “the whole of the insured’s actual loss history and not just part of it.”

It follows from these conclusions that the Syndicate has failed to make good its contention that the actual loss histories representation was false. I take in turn each of the complaints of falsity relied upon by the Syndicate in its written closing argument.

It is common ground that in five cases expected loss was calculated using loss histories from which “non-renewing accounts” had been removed. In broad outline what happened was this. A program had been in existence for some years. The precise business making up the program changed from time to time. When calculating the expected loss on renewal of the program, Legion paid no regard to past losses on business which did not form part of the proposed renewal. If one treated the “prospective insured” as being the business which it was known would form part of the renewed program then the omission of loss histories for others involved no adjustment whatever to the actual loss histories of such “prospective insureds”. But even if one treats the words “prospective insureds” to be a reference to the program as a whole, it is clear that on any ordinary use of the English language the actual loss histories were used in order to calculate expected losses. The process involved first of all identifying what the actual loss histories were and secondly adjusting them by removing losses in respect of business which was not being renewed.

In the case of an account known as “Bus Comp Association” a worksheet prepared by Legion showed that deductions had been made from the actual loss history provided by the assured. The evidence was that in part these deductions concerned non-renewing accounts (as to which see the preceding paragraph) and that as to the remainder the deductions represented what were described as “shock losses”. In the event, Legion’s underwriter did not fix the “expected loss” at the level which use of the adjusted actual loss figures would have produced. Instead, expected losses were fixed at a higher level constituting 45% of the reference premium. The Syndicate suggests that the underwriter had concluded that an appropriate AAAP would be 65%, and had worked backwards from this in order to determine an expected loss of 45%. That may well be right. Nevertheless it appears to me to be a fair inference from the working papers that the underwriter’s method of proceeding was first to determine the actual loss history, second to adjust it by deducting non-renewing accounts and shock losses, and third to form the view that an “expected loss” calculated in this way might well be too low, and accordingly the underwriter made a further adjustment increasing the expected losses and consequently the AAAP to a percentage which he considered reasonable. Hannover’s written closing submissions added that the upward adjustment more than compensated for any effect caused by the removal of the shock losses, and that in any event the shock losses were irrelevant for the purposes of the 1998/1999 renewal because specific additional reinsurance covering such losses had been purchased for that year with the result that if such losses were to recur they would not give rise to a claim against Legion. Ms Donohue’s written and oral evidence bore out both these assertions. The consequence is that this particular complaint does not assist the Syndicate.

The final matter relied upon by the Syndicate under this head as demonstrating the falsity of the actual loss histories representation concerned a program known as Preferred Non-Profit. The Syndicate complains about an adjustment made to the incurred losses for the 1994 year when calculating the expected loss at the time of the 1998 renewal. It also complains about adjustments in March and June 1999. So far as what was done in 1998 is concerned, it was plain that Legion had available to it historical information about the 1994 incurred losses. The documentary material suggests that when calculating the expected loss for the 1998 renewal the 1994 figure was adjusted. For similar reasons to those that I have given earlier, however, this does not seem to me to be inconsistent with the actual loss histories representation. Legion adopted a procedure under which the actual loss history was considered as part of a process leading to determination of the expected loss. In any event actual loss histories for other years were not adjusted. So far as the further consideration given to the matter in March and June 1999 is concerned, similar considerations apply, and on any view the relevant calculations of expected loss took place after the Hannover/Syndicate contract date. These matters provide no justification for the Syndicate’s assertion that the actual loss histories representation was false.

As with the previous head of complaint there is a short and complete answer to the Syndicate’s assertions in Schedule 3. Those asserted complaints are simply as to what allegedly happened. What allegedly happened is then said to show as a matter of logical inference that either (1) the underwriting requirements for the Mainframe Account did not require Legion to calculate expected losses in accordance with a methodology prescribed by the RPG or (2) Legion was not ready and willing and did not intend (or did not intend invariably) to comply with such underwriting requirements. However for the reasons given earlier I consider that no representation was made by Legion as to the second of these alternatives. It follows that the logical inference asserted in Schedule 3 does not demonstrate falsity of the limited prescribed methodology representation. The Syndicate’s final complaint under this particular head depends upon there having been a representation as to Legion being ready and willing to comply with the requirement for the use of a prescribed methodology and intending to do so. For the reasons I have given earlier there was no such representation, and for that reason alone this particular head of complaint fails.

Even if, however, the prescribed methodology representation had been as broad as the Syndicate claims, I would not have held that Schedule 2 demonstrated that it was false. This is because the Syndicate’s allegations of falsity involved either (a) an assumption that an unintentional error by Legion’s employees would be inconsistent with the prescribed methodology representation or (b) an assertion that Legion’s employees had intentionally acted contrary to the RPG. The latter assertion would involve a grave allegation of impropriety, for which the skeletal documentary material before me offered no sound basis. The former assumption would make no allowance for human frailty. Those who were ready and willing to comply, and intended to comply, with the RPG might well make mistakes when putting that intention into effect. Indeed it would be surprising if they did not.

It follows from these conclusions that the Syndicate has failed to make good its contention that the prescribed methodologies representation was false. I take in turn each of the complaints of falsity relied upon by the Syndicate in its written closing argument.

The Syndicate relied firstly on three cases where Legion’s underwriter had departed from the norm of an “average worst average” calculation. Such a departure was permitted by the RPG, which gave Legion the option to use one of three modifications. The third of these was described as follows:

Calculation using an arithmetic average incurred adjusted losses of four full account years of the prior five account years, excluding the current account year. This method is allowed to be used by the vice president, manager of underwriting only.

In three instances, which seem to me likely to have been in late December 1998, (PA Restaurant program), early March 1999 (American Health Centers program), and early September 1999 (International Hockey League program), the Syndicate said that a Legion underwriter had used an arithmetic average but failed to comply with this specific modification in the RPG because (a) the method used had not in fact involved an average of incurred adjusted losses of four full account years of the prior five account years excluding the current account year, and (b) the method in question had not been used by Mr Scott Barry who was the vice president, manager of underwriting. It is convenient to take these complaints in reverse order. As to the role of Mr Barry, the Syndicate relied on the fact that such records as Legion had been able to produce did not include a “sign-off” by Mr Barry, whereas such a “sign-off” appeared in the records for other risks. I do not accept that the absence of Mr Barry’s signature on such copies of the relevant calculation as the Syndicate has been able to produce from Legion’s records warrants an adverse inference. The evidence of Mr Partridge and Ms Donohue was that Mr Barry sat within a few feet of all of Legion’s underwriters and knew of and approved all relevant underwriting decisions, that there was a general peer review culture within the Legion underwriting department encouraged by Mr Partridge and Mr Barry, and that it would not be uncommon for more than one print-out to be made of an underwriting checklist describing the calculation that had been made. In these circumstances the mere fact that Mr Barry’s signature appears on some checklists but not on others does not in my view show that Mr Barry had not approved the use of a straight average calculation in the cases cited by the Syndicate.

As regards point (a) above it was necessary for the Syndicate to assert that what happened in these three cases demonstrated that at the Hannover/Syndicate contract date Legion was not ready and willing and did not intend to comply with the RPG. The instances in question occurred after the Hannover/Syndicate contract date. I do not consider that they demonstrate any particular state of mind on the part of Legion at that date. Even if they had occurred earlier, I would not have regarded them as sufficient to demonstrate that Legion was not ready and willing to comply with the RPG and did not intend to do so. What occurred appears simply to have been an error on the part of the underwriter in question and Mr Barry. The Syndicate relied on evidence that each of the risks in question involved calculations made by different Legion underwriters (and in one case approved by a superior of the underwriter in question). It was said to be difficult to accept that four different underwriters could have coincidentally made the same error. In my view, however, the likely explanation is that the error was Mr Barry’s. There is no sufficient reason to infer that Mr Barry was guilty of anything other than an innocent error in this regard. Moreover, even if one puts aside the role of Mr Barry, all that the Syndicate has been able to establish is that in the case of three out of numerous risks underwriters made a straight average calculation which did not conform with the precise requirements of the RPG. That is a wholly insufficient basis for a conclusion that Legion was not ready and willing to comply with the RPG and did not intend to do so.

The second type of alleged failure relied upon by the Syndicate concerned what the RPG defined as the “Exposure Factor”. The RPG explained that this provided an adjustment to the losses for changes in exposure. It added:

The actual calculation of the exposure factor is the renewal payroll estimate divided by the trended payroll associated with each individual policy.

The Syndicate relied on what had happened in relation to three risks, Alliance, PA Restaurant, and Bus Comp. Of these, only the first pre-dated the Hannover/Syndicate contract date. Even if all three had pre-dated the Hannover/Syndicate contract date I would not have concluded that they assisted the Syndicate. In cross examination Ms Donohue accepted that the exposure factor for the program in question had been calculated, not by reference to change in payroll, but by reference to change in the number of participants in the program. Ms Donohue speculated that in the cases in question the brokers did not have historic payroll information. While it would have been possible for Legion to insist that such information be provided, it could have taken six or eight months to put it together, and in the market place at that time there probably was not the opportunity to delay for such a long period. Where the underwriter thought that the participants would have policies of the same size, then it would be appropriate for an underwriter to assess the changes in overall payroll by reference to the changes in the overall number of participants. Ms Donohue added that because Legion appeared to have underwritten in this way in these three cases she would assume that reinsurers were aware of what had been done.

There was no evidence that in relation to these three particular risks an underwriter could not reasonably have assumed that the policies of the different participants were going to be of a similar size. Accordingly I am not satisfied that in these three cases the RPG did not permit the underwriter to estimate the renewal payroll and the trended payroll by reference to the number of participants. Even if, however, on these three occasions Legion underwriters had done something which the RPG did not entitle them to do, this would not in my view warrant an inference that Legion was not ready and willing to comply, and did not in fact intend to comply, with the RPG. In these circumstances it is not strictly necessary for me to comment on Ms Donohue’s assumption that reinsurers would have been aware of what had been done. Hannover received a great deal of detailed information from Legion. There was, however, no obligation on Hannover to go through this information with a toothcomb. In my view, unless Legion specifically pointed out a failure to comply with the RPG, Hannover was entitled to assume that detailed calculations sent to it were calculations which had been made in accordance with the RPG.

The only remaining complaint under this head concerned a program known as “Preferred Utilities.” The complaint is that at some point after June 1999 Legion prepared a request for a special acceptance enabling it to include in this program business which altered the make up of the program and involved setting the expected loss by reference to the historic losses of a book which differed from the book that was about to be written. The Syndicate adds that what was sought in the special acceptance application was in fact put into effect in December 1999.

In my view there is nothing in these events in the second half of 1999 which justifies any adverse inference as to Legion’s state of mind on the Hannover/Syndicate contract date several months earlier. It is clear that Legion thought that what was proposed called for a special acceptance. Hannover suggested that there had indeed been agreement on the part of reinsurers to what Legion proposed, albeit that no documentary record of that agreement could now be found. The Syndicate countered that the application for special acceptance appeared to have been sent to the wrong reinsurer. At the end of the day, even if events in the second half of 1999 could be regarded as shedding light on Legion’s intentions much earlier, those events simply showed an intention on the part of Legion, where it wanted to do something which did not comply with the RPG, to seek a special acceptance. This complaint does not assist the Syndicate.

The Syndicate’s closing written submissions said that the use made by Legion of the ‘underwriting box’ was relevant both to the complaint about the use of actual loss histories and the complaint about the use of a prescribed methodology.

As regards actual loss histories, the Syndicate noted Mr Crickmere’s oral evidence that originally the ‘underwriting box’ had been intended to be used for “prospective program business in the sense of programs without loss histories.” By contrast, Mr Partridge said in oral evidence that even in a case where there was historical information for a program Legion could set the expected loss either by using that information or “by picking a target expected loss and working to that.” These criticisms, however, failed to have regard to the way in which the ‘underwriting box’ worked. As was pointed out on behalf of Hannover, the agreement it made for use of the ‘underwriting box’ was on the basis that Legion underwriters would individually underwrite each new piece of business going into the program in question, and that business had to have enough experience to qualify it for the Mainframe Account. Thus on each occasion when Legion agreed to provide cover for a particular program on the basis of the ‘underwriting box’ arrangement, it was providing cover to individual insureds by reference to their actual loss histories. It follows that the requirements for and use of the ‘underwriting box’ were inconsistent with neither the limited actual loss histories representation nor the broader representation alleged by the Syndicate. Moreover, it is not in my view open to the Syndicate to advance this particular complaint: for reasons given in the previous main section of this judgment, I consider that Mr Crane was told about the ‘underwriting box’ on 16 November 1998.

Turning to the prescribed methodologies representation, the syndicate’s final complaint only referred to the ‘underwriting box’ arrangements in this regard in order to assert that those arrangements were not in fact complied with. Any assertion that the underwriting requirements did not include the use of a prescribed methodology was – inevitably in the light of Mr Crane’s evidence – effectively abandoned in the Syndicate’s written closing submissions. It follows that this part of the Syndicate’s third final complaint fails, for the complaints made are not inconsistent with the limited prescribed methodology representation.

Even if, however, the prescribed methodology representation had been as broad as the Syndicate contends, I would not have held that it was false. The arrangements for the use of an ‘underwriting box’ stated that reinsurers would “allow Scott Barry only to sign-off on the creation of an ‘underwriting box’ for prospective program business”. The Syndicate’s first main complaint about the use of the ‘underwriting box’ was that Mr Barry had not signed off in the case of Lippa Insurance Services (April 1998 placement and April 1999 renewal), West Coast Workers’ Compensation (April 1999 renewal), Great Northern Underwriters (offer letter sent by Legion on 8 June 1999), United Messenger Courier Program (August 1999 renewal), Wateridge Compensation (September 1999 renewal) and Pacific Coast Workers’ Compensation (September 1999 renewal). Here similar considerations apply to those identified in relation to the “straight average” complaint discussed at paragraphs 135 to 137. For the reasons given in those paragraphs I do not consider that the Syndicate has demonstrated failure to comply with the relevant requirement. Moreover, only one of the cases relied on occurred before the Hannover/Syndicate contract date. These cases comprise a wholly insufficient basis for a conclusion that at that date Legion was not ready and willing to comply with the relevant requirement, or that Legion did not intend to comply with it.

There were cases where for a particular program five years of loss history was available, but nevertheless Legion adopted an ‘underwriting box’ approach. Those identified in the Syndicate’s written closing submission were West Coast Workers’ Compensation (April 1999 renewal), Preferred Construction Industries (July 1999 renewal) and Wateridge Compensation (September 1999 renewal). There was, however, nothing in the agreement about the use of an ‘underwriting box’ which precluded such use in a case where five years of loss history for a particular program were available. The Syndicate relied upon oral evidence from Mr Crickmere that he “expected the expected loss to be calculated in accordance with one of [the] methodologies [set out in the RPG].” A mere expectation on the part of Mr Crickmere, however, does not amount to a requirement. In any event the occasions in question occurred after the Hannover/Syndicate contract date, and are insufficient to warrant any conclusion that at that date Legion was not ready and willing to comply with relevant requirements or did not intend to do so.

The Syndicate’s next complaint concerned the proviso under which the individual risks attaching to the program were to be individually underwritten by Legion in accordance with the RPG. The Syndicate referred to underwriting guidelines for documents under which, subject to specified limitations, Legion delegated to particular companies authority to approve individual risks. Three of these documents evidenced delegation to CRS, this company being, like Legion, part of the MRM group. The programs involved were Wateridge Compensation (where the document referred to the period from 1 September 1997 to 1 September 1998), Lippa Insurance Services (referring to the period 1 April 1998 to 1 April 1999) and West Coast Workers’ Compensation (also 1 April 1998 to 1 April 1999). In a fourth case, Great Northern Underwriters, there was a lengthy debate between Legion and a company known as Granite as to whether Granite could have authority to approve individual insureds. The Syndicate accepts that when this was suggested in October 1998 it was accompanied by an observation, “reinsurance will not be in Legion’s treaty due to broker’s need to have ‘pen’.” It was only in the second half of 1999 that Legion eventually granted Granite limited underwriting authority. In my view what happened in the second half of 1999 cannot warrant an inference that at the Hannover/Syndicate contract date Legion was not ready and willing to comply with relevant underwriting requirements or that it did not intend to do so. Moreover, the fact is that shortly before that date the documentation reflects a belief that a broker’s need to have the underwriting pen would prevent relevant reinsurance. This plainly demonstrates a recognition of, and indeed an intention to comply with, the requirements for the Mainframe Account.

Thus in this regard the Syndicate is left with the three occasions on which limited authority was given to CRS. I do not consider that they warrant an inference that Legion was not ready and willing to comply with relevant requirements or that it did not intend to do so. In the case of Wateridge Compensation an underwriting guidelines document relied on by the Syndicate was for an initial period when this program formed part of the Cessions Account rather than the Mainframe Account. If this arrangement continued to apply when the program became part of the Mainframe Account, this may simply have been an error on the part of whoever was responsible for allocating the program to that account. Similarly, it seems to me that there may have been errors for Lippa Insurance Services and West Coast Workers’ Compensation for the period 1 April 1998 to 1 April 1999 either on the part of whoever allocated these programs to the Mainframe Account or on the part of the person approving the delegation of authority. An error of this kind appears to me to be understandable when the delegation was to an associated company of Legion. What happened in these instances does not, accordingly, warrant the inference which the Syndicate seeks to draw.

If my conclusions above are correct it is not necessary to consider whether the alleged representations were material, nor whether Mr Crane in fact relied on them. I shall accordingly deal with these questions only briefly.

Hannover submitted that the alleged representations were too vague and imprecise to have been material. As a matter of logic this seems to me to be obviously right. Representations about the use of actual loss histories, and about the existence of a prescribed methodology, are of no use for decision-making purposes unless one knows how the loss histories will be used and what the prescribed methodology is. The Syndicate relied on Mr Wills’s evidence that what was represented about the use of actual loss histories was material because Legion’s actual practice in relation to actual loss histories was contrary to market practice. It does not seem to me that the latter assertion, even if correct, can assist the Syndicate in its misrepresentation claim. For the reasons given earlier in this judgment the representation as to the use of actual loss histories said nothing about how they were used. There was accordingly no representation that Legion did not adopt the practices that the Syndicate complains of. This is the case whether the representation was merely, as I have held, that there were requirements to use actual loss histories or whether, as the Syndicate contends, it implicitly represented that Legion was ready and willing to comply with that requirement and intended to do so.

In these circumstances I do not need to resolve the question whether the practices adopted by Legion were consistent with market practice. For completeness, however, I record that Mr Owen on behalf of Hannover accepted that it would be inappropriate to remove a “shock loss” entirely from the loss history. What Legion did, however, on the occasions described earlier was not simply to remove the relevant “shock losses” entirely from history, but to remove it and then make an upward adjustment which had the effect of partially compensating for the removal of the shock loss.

Mr Owen considered it to be standard market practice to remove losses for accounts that were not being renewed and to set the expected loss not by reference to the actual loss history of a program but by reference to the insured book going forward. He explained that it made sense when considering the terms for the book going forward to adopt such an “as if” basis. The logic of his reasoning appears to me to be compelling, and I prefer his evidence to that of Mr Wills in that regard.

Turning to what was represented in relation to the use of a prescribed methodology, the Syndicate asserts that Mr Owen accepted that the representation in question “would have given comfort to the prudent reinsurer.” I disagree. In the passage cited by the Syndicate Mr Owen merely accepted that the 1998 Carve-out renewal proposals demonstrated that “they have a way of approaching this business.” Mr Owen firmly maintained that any reinsurer would want to know what was in the RPG. I am satisfied that for this reason the representation alleged by the Syndicate was not material.

Turning to reliance on the alleged representations, Mr Crane could not recall what it was that he in fact read, and thus could not specifically recall relying on any particular assertion in the 1998 Mainframe Carve-out renewal proposals. His evidence was that he would have read extracts from this document, and that he understood that Legion would calculate the expected loss using some sort of underwriting methodology or rational approach to underwriting. The syndicate relied upon his oral evidence that:

For a reinsurance treaty, you’ve got to believe in your client. The figures meant to an extent I believed in Legion, but only to the extent to which I believed that they were behaving in the same way they had previously. I would’ve read the information as to what they were doing going forward because it is highly material.

The Syndicate summarised Mr Crane’s evidence as being that the manner in which Legion underwrote the business went to the integrity of the figures presented to Mr Crane and was thus a necessary and important consideration in his underwriting decision making process. The Syndicate added:

275.

It was suggested to Mr Crane that it was fanciful to suggest that he was concerned to know whether Legion was underwriting the Mainframe Account in accordance with certain guidelines which had built-in safety features for reinsurers; that he was not in the least bit interested in the detail of Legion’s underwriting. Mr Crane rejected this. He said … :

“That is absolutely not true, because I was assuming that what they [Legion] had done in 1992 through 1994, which had enabled them and us to make a profit, they were going to continue to do through into 1998, and because the market in the US was somewhat more competitive by 1998, it was even more vital that they underwrote the business in a proper and orderly fashion, and it said in their presentation that that is exactly what they were going to do.”

276.

More specifically, it was suggested to Mr Crane that he would not have been in the slightest degree concerned if he had been told that Legion had removed from the actual loss history losses on particular accounts which were not being renewed into the coming year. Mr Crane rejected this. He stated … :

“The program is the program, and … you can’t arbitrarily add in and delete figures. … An account could have an unusually unlucky year, so therefore Legion could have chosen not to renew it, but they could add in an account that was going to forthcoming have an unusually unlucky year. So you take the whole of the figures, you take the whole of the figures as they are in fact and work from there. … you can’t fiddle with figures. If you fiddle with figures, you make them inaccurate. As soon as you make them inaccurate, they become meaningless.”

277.

Mr Crane was asked what the difference was between taking out risks which are not renewing and taking out risks on a class of business that you are no longer writing for the purpose of compiling as-if figures. He answered … :

“Because the accounts you are leaving behind and the accounts that you are adding to the program are the same sort of accounts as the accounts you are taking out. If you are taking out all of a class of an account, that is fair.”

278.

It was similarly suggested to Mr Crane that he would not have been troubled if he had been told that Legion removed shock losses from the loss history of a program. Again, Mr Crane disagreed … :

“A shock loss is what you put business into a program to cover. You have the program big enough ideally to cover shock losses but if an account within that program has a shock loss, what is to say another account within that program is not going to have a shock loss in the following year? You can exclude losses because you don’t like them.”

279.

As regards methodologies, it was suggested to Mr Crane that, because he did not ask to see a copy of the [RPG] or ask any questions about certain matters in the [1998 Carve-out renewal proposals], either he must not have read the [1998 Carve-out renewal proposals] or, if he did, he must have not been remotely interested in Legion’s underwriting. Mr Crane disagreed … .

280.

This line of questioning appears to have been directed to reasonableness of Mr Crane’s assumption, in absence of any statement to the contrary, that Legion’s methodology or rationale approach to underwriting had not changed since his last involvement in 1992 to 1994.

281.

Mr Crane’s evidence, however, was that he would have read extracts from the [1998 Carve-out renewal proposals], including the “underwriting methodology” …, and that he understood that Legion would calculate the expected loss using some such methodology or rational approach to underwriting. It is not clear whether and, if so, why Hannover Re contends that Mr Crane was not entitled to rely upon what he was told.

The Syndicate must at least demonstrate that but for the representations in question Mr Crane would not have written the 1998 Mainframe Carve-outs on the terms that he did: see agreed proposition 10. I consider that Mr Crane’s reconstructions in this regard are unreliable. The position in 1998 was that Mr Crane had written similar contracts before, held the view that the results were good, had that view confirmed by the experience summaries, and needed more premium income to meet his targets, particularly on Carve-out business. In order to secure the Legion Carve-out business he was prepared to abandon his “mathematical stabs” at rating to the extent necessary to match Clarendon’s rates. If he had attached any significance to the representations now relied upon by the Syndicate he would have asked for an explanation of the precise ways in which actual loss histories were required to be used by Legion and he would have insisted that he be supplied with the RPG, or at least be given more detailed information about their content. He did none of these things. I am satisfied that his decision to write the 1998 Mainframe Carve-outs was not influenced by the representations now relied on.

Here the Syndicate draws attention to the reporting documentation regularly sent by Legion to Hannover, including underwriting checklists and loss analyses, in accordance with article 9 of the 1998 Mainframe XL contracts. It is common ground that Hannover was sent such documents. The Syndicate goes on to submit that from this material Hannover knew or ought to have known “how Legion was going about its underwriting.” It is not asserted, and there could be no foundation for an assertion, that Hannover’s disclosure obligations extended so broadly as to require disclosure of “how Legion was going about its underwriting.” The specific matters put in cross examination to Mr Kramp were that the reader of the reporting documentation would have seen:

I do not accept that this material shows that on the Hannover/Syndicate contract date Hannover knew or ought to have known that Legion was acting in a manner contrary to Hannover’s requirements or expectations. The first difficulty facing the Syndicate in this regard is that the vast majority of the instances relied upon occurred after the Hannover/Syndicate contract date. The Syndicate’s closing submissions made no attempt to identify instances which occurred before that date. In my view there can have been no more than a handful.

The second difficulty facing the Syndicate is that no witness gave evidence that Hannover was aware at the relevant time that Legion was not doing what Hannover required or expected. When the detailed matters relied on by the Syndicate were put to Mr Kramp he said that he did not go into these details. He added:

What I presume from the underwriting process and my superiors, is that they knew about this, yes.

If this was intended to refer to knowledge that Legion was not complying with Hannover’s requirements or expectations then it was no more than speculation on Mr Kramp’s part, and it is speculation which I do not find convincing. I consider that if Hannover had appreciated that the material supplied to it showed any failure to comply with its requirements or expectations then it would have raised this with Legion. There is no evidence of it doing so. As indicated earlier, I do not consider that Hannover was under any obligation to go through the reporting documentation with a toothcomb. The fact that by doing so one could, in a handful of instances, discern the matters which the Syndicate now says were of concern does not warrant a conclusion that Hannover ought to have known, still less actually knew, the matters now relied upon.

Thus I conclude that Hannover was not aware of the matters complained of, and that the Syndicate is wrong to say that Hannover ought to have been aware of them. It follows that the claim for non-disclosure in this regard fails. As to materiality and inducement, it seems to me that for the reasons given earlier these specific matters of complaint advanced by the Syndicate do not in fact demonstrate any departure by Legion from recognised market practice. Mr Owen said that he would have wished to know how Legion dealt with shock losses. However he did not suggest that such knowledge would have influenced the judgment of a prudent reinsurer in fixing the premium or determining whether to take the risk. Nor, more generally, do I consider that the tiny number of cases relied on as had occurred prior to the Hannover/Syndicate contract date would have done so. Finally in this regard, for the reasons given earlier in relation to the alleged representations, I do not consider that the matters in question would have affected Mr Crane’s decision to write the 1998 Mainframe Carve-outs on the terms that he did.

The second final complaint was “misrepresentation and/or non-disclosure concerning the comparative strictness of the underwriting requirements [for the Mainframe Account and Cessions Account].” This complaint concerned the first matter relied upon by way of misrepresentation in the original particulars of claim of December 2006, along with the corresponding element in the third head of non-disclosure asserted in those particulars of claim. In support of the claim of misrepresentation in this regard concerns were expressed about the ‘underwriting box’. These concerns were a development of those in the expanded fourth head of non-disclosure advanced in the re-amended particulars of claim served on 29 February 2008.

Until closing submissions the Syndicate had advanced a case that what Mr Crickmere had said in his fax to Mr Morris was false. By the time that closing submissions came to be prepared the Syndicate had to cope with the difficulty that Mr Wills had accepted in cross-examination that what Mr Crickmere said in his fax to Mr Morris was true. The Syndicate’s final complaint ignored this evidence and instead asserted that at the meeting on 16 November 1998 Mr Crickmere, by failing to say more than was in his earlier fax to Mr Morris, misrepresented the position or told “at best only a half truth” because of three allegedly undisclosed matters. These were that Mr Crickmere:

mentioned neither the availability of the ‘underwriting box’ to write start-up programs in the Mainframe Treaty, nor that the Main Treaty was now more flexible than had historically been the case, nor the fact that the Mainframe Account and Cessions Account now had a similar content.

I do not agree. For the reasons given in my discussion of the meeting of 16 November 1998 I am satisfied that Mr Crickmere did not merely repeat what had been in the fax to Mr Morris. On the contrary, in order to give Mr Crane the “full explanation” he wanted, Mr Crickmere told Mr Crane about the arrangements for program business to be allocated to the Mainframe Account.

In these circumstances it is no more than speculation for the Syndicate to assert that no mention was made that the arrangements for program business included start-up programs. Further, even if Mr Crickmere had not specifically mentioned start-up programs, I do not accept that this would give Mr Crane an inaccurate picture. Once Mr Crickmere had mentioned programs there was no need for him to go into this level of detail unless Mr Crane asked for it. As to the Main Treaty now being more flexible, that inevitably followed from the introduction of programs and would have been perfectly clear to Mr Crane. The third allegedly undisclosed matter, that the two Accounts had a similar content, relies on the closing observation by Mr Crickmere in his fax of 30 September 1998 to Mr Cramsie. What Mr Crickmere explains in that fax, however, is that the similarity he is referring to is the ability of both Accounts to write program business. This third allegedly undisclosed matter thus adds nothing to the first two.

For these reasons the Syndicate cannot make good its revised misrepresentation claim, nor can it make good any separate claim of non-disclosure of the three allegedly undisclosed matters. Its closing submissions added that Mr Crickmere owed a duty as a broker to make sure that the RPG were communicated to Mr Crane. For the reasons given in my discussion of the 16 November 1998 meeting, however, I am satisfied that Mr Crickmere intended to give Mr Crane the RPG, and that the likely reason for his failure to do so is that Mr Crane was no longer worried about the criteria for allocation of business to the Mainframe Account.

In these circumstances I can deal with materiality and inducement briefly. Mr Owen’s view was that a prudent reinsurer could not have made any sense of the vague representation asserted by the Syndicate. As a matter of logic this seemed to me to be obviously right. Mr Wills advanced no coherent argument to the contrary. Turning to inducement, in a context where Mr Crane had sought a “full explanation” I think it highly unlikely that either the broad observations in Mr Crickmere’s fax to Mr Morris, or the alleged failure on the part of Mr Crickmere to supply information including the RPG, induced Mr Crane to do anything.

The first final complaint concerned non-disclosure by Hannover of underwriting and claims audits which it had conducted. This was, in substance, the first head of non-disclosure alleged in October 2006. The key elements of this complaint were summarised in the Syndicate’s closing written submissions. Both the underwriting and claims audits contained serious findings and raised serious concerns, to put the matter at its very lowest. These findings and concerns were known to Hannover. The underwriting and claims audits were taken into consideration by Hannover when making its decision to renew its participation on 1998 Mainframe XL contracts. The facts that there were underwriting and claims audit reports and that the auditors had made the findings and expressed the concerns that they had were not disclosed to Mr Crane. Hannover had been able to take such matters into account when making its underwriting decision, but Mr Crane had been unable to do so.

Hannover’s response acknowledged that it had the benefit of the audit reports and took them into account. That did not necessarily make them material. As to the underwriting audits, Hannover’s written closing submissions pointed out that the Syndicate’s pleaded case concerned Mr Burk’s criticism of the loss rating approach to program business. As to that Mr Crane had ample information to form his own judgement. As to the claims handling audits, Hannover said that they did not reveal any serious problems. They expressed concern about dependency on TPAs, but that dependency was apparent from the 1998 Mainframe Carve-out renewal proposal.

I start with factors which affect disclosure issues on both the underwriting and the claims handling audits. I must consider the position at the Hannover/Syndicate contract date. The obligation on that date is that by then, in relation to what it knows or ought to know, Hannover must have disclosed “every material circumstance”: see agreed proposition 13. Agreed proposition 14 confirms that “circumstance” in this context can include information received by Hannover. Agreed proposition 15 is that every circumstance is material which would influence the judgment of a prudent reinsurer in fixing the premium or determining whether he will take the risk. In that regard, however, agreed proposition 20 draws a dividing line: the reassured is not obliged to offer the reinsurer his own opinion on matters equally known to the reinsurer and which the reinsurer is in as good a position as the reinsured to assess.

The expert evidence on each side in my view was difficult to reconcile with agreed proposition 20. On the one hand Mr Wills did not adequately focus on what matters were known to Mr Crane and the extent to which he was in as good a position to form a judgment as Hannover. On the other hand Mr Owen seemed to think that a reinsurer could not expect to be told information gleaned by the reinsured. He said that this flowed from a “follow the fortunes” principle. However a “follow the fortunes” clause obliges a reinsurer to accept the reinsured’s decisions in various respects, usually as to the admissibility of claims. The 1998 Mainframe Carve-outs contained no “follow the fortunes” clause. If they had it seems to me that this would simply make proper disclosure all the more important.

The Syndicate asserted that Hannover was willing to participate in the 1998 Mainframe XL contracts despite the adverse audit comments because it would have the protection of the 1998 Mainframe Carve-outs. This cannot be right. Hannover committed to the 1998 Mainframe XL contracts on 29 September 1998. At that stage it was entirely unclear whether Carve-out cover would be offered at acceptable rates.

It is convenient at this stage to consider the underwriting and claims handling reports separately – I will return to consider them together in relation to inducement. In relation to the underwriting audits, in my view the evidence clearly establishes that the use of loss rating for programs was a matter on which there were differing views as a matter of underwriting judgement. Some, like Mr Burk, considered it inadequate. Others, like Mr Barry, thought it provided a means of assessing risk which was as valid for programs as it was for individual insureds. Hannover’s eventual position by the time of the Hannover/Syndicate contract date appears to have been that the use of loss rating did not warrant abandoning either of the Mainframe XL contracts, but that it wanted Legion to bear the dangers of loss-rating more fully in mind. It does not seem to me to matter for this purpose whether or not Mr Burk’s concerns were discussed at the meeting in Hannover in mid-September 1998.

In this regard the 1998 Carve-out renewal proposals described Legion’s loss rating approach. The major concern about use of such an approach would be in relation to heterogeneous business. As noted at paragraph 89 above, the “Business Profile” in the 1998 Mainframe Carve-out renewal proposal demonstrated the presence of a substantial amount of heterogeneous business from 1 October 1996 onwards. It seems to me that Mr Crane was in just as good a position as Hannover to form his own view about whether Legion’s use of loss rating should lead him to decline, or to seek higher rates.

In its closing submissions the Syndicate seeks to broaden its case on the underwriting audits to rely on matters other than loss-rating. The further matters relied on, were, however, not treated by Mr Burk as of substantial significance, and in the present proceedings were no more than afterthoughts on the part of the Syndicate’s legal team. Paragraph 17 (b) of the re-amended points of claim alleged a failure to disclose “concerns about the appropriateness to program business of the underwriting requirements for the Mainframe Account and the Main Treaty.” The only such concerns in the audits reports involved aspects of loss-rating. I do not permit the Syndicate to depart from its case so as to introduce other concerns: in the absence of some written attempt to define these additional matters before the witnesses gave oral evidence I do not consider that it would be fair to allow reliance on them. For present purposes I need only add that nothing in the evidence showed that in relation to these other concerns Legion’s failings were so serious as to cause a prudent underwriter to decline the risk or increase the rates which would otherwise apply.

Turning to the claims handling audits, the first was highly positive. The second involved an about turn on questions such as the extent to which it was desirable to use TPAs. Overall it seems to me that the most that can be derived by the Syndicate is that the claims handling was no better than average. Mr Wills said that by good and efficient administration the business could be improved by 2 or 3%. That may be so, but it does not follow that merely average claims handling is something so significant that it would affect the judgment of a prudent reinsurer. As to the use of TPAs, the “Operations” part of the 1998 Carve-out renewal proposals described the use of TPAs and explained that the insured was required to make a contract with the TPA. Here too it seems to me that Mr Crane was in just as good a position as Hannover to form his own view about whether Legion’s use of TPAs should lead him to decline, or to seek higher rates.

Accordingly I conclude that agreed proposition 20 is a complete answer to the Syndicate’s case on non-disclosure of the audit reports. In these circumstances the question of inducement does not arise. I can deal with it briefly. In essence it seems to me that the questions which arise are similar to those which arise under the Syndicate’s third final complaint. Mr Crane’s desire to write this business was very strong, so strong that he decided to bring his rates down in order to match Clarendon. I do not consider that anything in the audit reports would have prevented him from concluding that he should match Clarendon’s rates so as to secure this business.

For the reasons given above none of the Syndicate’s final complaints succeeds. I conclude that the Syndicate is not entitled, as against Hannover, to avoid the relevant 1998 Mainframe Carve-outs.

As a matter of practice, contracts of reinsurance are concluded in the Lloyd’s market when a slip is scratched on behalf of a syndicate: see Eagle Star v Spratt [1971] 2 Lloyd’s Rep. 116, 124.

The orthodox understanding in the Lloyd’s market is that the presentation of the slip by the broker constitutes the offer and the writing of a line by the underwriter constitutes an acceptance of this offer: see General Reinsurance Corporation v Fennia Patria [1983] QB 856, 866.

The scratching of a slip, however, is not the only method by which a contract of reinsurance can be concluded; how and when a contract of reinsurance is concluded depends on an objective analysis of the evidence in terms of offer and acceptance and intention to contract: see Assicurazioni Generali v ARIG [2003] Lloyd’s Rep IR 131, paragraphs [49]-[51], p.147.

Every material representation made by the reassured or his agent to the reinsurer during the negotiation, and prior to the conclusion, of a contract of reinsurance must be true (Marine Insurance Act 1906 (“MIA”), s.20(1)).

A representation is material if it is one which would influence the judgment of a prudent reinsurer in fixing the premium, or determining whether he will take the risk (MIA, s.20(2)). Whether a particular representation is material or not is, in each case, a question of fact (MIA, s.20(7)).

If a material representation made by the reassured or his agent as above is untrue, the reinsurer may avoid the contract of reinsurance if the representation induced him to contract on the terms on which he did (MIA, s.20(1); Pan Atlantic v. Pine Top [1995] AC 501, 549).

A representation may be either as to a matter of fact, or as to a matter of expectation or belief (MIA, s.20(3)). Strictly speaking, there can be no representation as to future facts; a statement relating to such facts (if not a promise) must either be a statement of present intention (or readiness or capacity) or a statement of opinion or belief (MacGillivray on Insurance Law, 10th Ed., paragraph 16-40, p.399; Spencer Bower, Actionable Misrepresentation, 4th Ed., paragraph 14, p.5).

A representation of fact is true if it is substantially correct, i.e. if the difference between what was represented and what is correct would not be considered material by a prudent underwriter (MIA s.20(4)).

A representation of expectation or belief is true if it is made in good faith (MIA s.20(5); Economides v. Commercial Union [1998] QB 587).

The requirement that the underwriter should have been induced to contract by the representation is the same as in the general law of contract. It requires that the representation should have been an effective cause (but not necessarily the sole cause) of the reinsurer entering into the contract on the terms that he did. Causation cannot exist unless (at the least) it can be said that the reinsurer would not have contracted on the terms that he did but for the representation (Pan Atlantic (above), p. 549; Assicurazioni Generali v ARIG (above) paragraphs [59], [187], pages 148, 170).

The question of whether any and if so what particular representation was made depends on an objective assessment of what was said or done and its likely effect on the alleged representee in the context in which the particular parties were concerned. In other words, in the present case, what would the documents and exchanges relied upon have conveyed to a prudent reinsurer in the position of the Syndicate (Sumitomo Bank v BBL [1997] 1 Lloyd’s Rep 487, 515).

A representation made by a reinsured can be a representation of fact notwithstanding that it relates to the underlying insured or matters which the reinsured has learnt from the insured: see Highlands Insurance Co v Continental Insurance Co [1987] 1 Lloyd’s Rep 109, 111-2; Sirius International Insurance Corporation v Oriental Assurance Corporation [1999] Lloyd’s Rep IR 343, 350-1.

The reassured is obliged to disclose to the reinsurer, before conclusion of the contract, every material circumstance which is known to the reassured, and the reassured is deemed to know every circumstance which, in the ordinary course of his business, ought to be known by him (MIA, s.18(1)).

The term “circumstance” includes any communication made to, or information received by, the reassured (MIA s. 18(5)); rumours, allegations, communications and intelligence can be material circumstances regardless of whether they are, or are believed by the reassured to be, true (McDonald Eggers, Picken & Foss, Good Faith and Insurance Contracts, second ed. paragraphs 7.70-7.74, pp. 151-3; Brotherton v Aseguradora Colseguros [2003] Lloyd’s Rep IR 746, 753-8).

Every circumstance is material which would influence the judgment of a prudent reinsurer in fixing the premium or determining whether he will take the risk (MIA, s.18(2).

In the absence of inquiry, the reassured is not obliged to disclose any circumstance which is known or presumed known to the reinsurer (the reinsurer is presumed to know matters of common notoriety and matters which a reinsurer ought, in the ordinary course of his business, to know) or as to which information is waived by the reinsurer (MIA, s.18(3)(b), (c)).

Where a contract of reinsurance is effected for the reinsured by an agent, the agent has an independent obligation to disclose every material circumstance which is known to it, and an agent to insure is deemed to know every circumstance which in the ordinary course of business ought to be known by, or to have been communicated to, it. (MIA, s.19(a)). The only agent affected by the s.19 obligation, is the agent who actually deals with the reinsurer; an intermediate agent is not subject to a s.19 obligation, however matters known to an intermediate agent may be matters which ought to have been communicated, and which are therefore deemed to be known, to the agent who actually deals with the reinsurer and which therefore this agent is obliged to disclose (PCW Syndicate v PCW Reinsurers [1996] 1 Lloyd’s Rep 241, 257, 258-59).

If a material circumstance which should have been disclosed is not disclosed by the reassured or its agent, the reinsurer may avoid the contract of reinsurance if the non-disclosure induced him to contract on the terms on which he did (MIA, s.20(1); Pan Atlantic v Pine Top (above)).

There is no distinction, for the purpose of establishing inducement, between the requirement of causation as it relates to misrepresentation and as it relates to non-disclosure (see Assicurazioni Generali (above)).

The reassured is not obliged to do the reinsurer’s job for him; that is to say, the reassured is not obliged to offer the reinsurer his own opinion on matters equally known to the reinsurer and which the reinsurer is in as good a position as the reinsured to assess for himself (Carter v Boehm 3 Burr 1905, 1910-1; Bates v Hewitt (1867) LR 2 QB 595, 605, 610; Glasgow Assurance v Symondson (1911) 16 Com. Cas. 109, 119-20; Aiken v. Stewart Wrightson [1995] 2 Lloyd’s Rep 618, 648-9; MacGillivray (above), paragraphs 17-79 and 17-80, pages 444 and 445).

A reinsurer who fails to make inquiry of a particular matter when on notice of matters which would prompt a reasonable reinsurer to make further inquiry waives disclosure of matters which such an inquiry would have revealed, provided the presentation made to him was, in all the circumstances, fair. The doctrine of waiver cannot be applied, however, to undisclosed facts which are special, such that their non-disclosure distorts the presentation of the risk. Any waiver, furthermore, must be clear (CTI v. Oceanus [1984] 1 Lloyd’s Rep 476, 497-9, 511-2, 529-30; Marc Rich v Portman [1997] 1 Lloyd’s Rep 225, 234; Wise (Underwriting Agency) Ltd v Grupo Nacional [2004] 2 Lloyd’s Rep 483, 505-7, 510).