England and Wales Court of Appeal (Civil Division) Decisions

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE QUEEN'S BENCH
DIVISION (LONGMORE J.)

		Royal Courts of Justice Strand, London, WC2A 2LL Tuesday 3rd July 2001

LORD JUSTICE MANCE:

Introduction

1. The issues before us arise from a facultative reinsurance placed with Gan Insurance Company Limited of London ("Gan") by The Tai Ping Insurance Company Limited ("Tai Ping") in March 1996. The reinsurance was for 2%, part of a line of 35%, underwritten by Tai Ping on an erection all risks and third party liability insurance taken out by Winbond Insurance Corporation ("Winbond"). This liability insurance covered machinery whilst in the process of erection, installation and commissioning from 15^th February 1996 in a building called Fab 3 at a science park in Taiwan. The sum reinsured in respect of the machinery was expressed as "up to" a little over NT$ 6.5725 billion (on a 100% basis). Gan for its 2% line and another reinsurer, Eagle Star, for a further 2%, were approached through brokers and wrote their lines on a separate broker's slip. Tai Ping had previously reinsured other parts (in total 31%) of its line of 35% without the intervention of brokers, in particular 3.1% with treaty reinsurers, 12.5% with Cologne Re, 10% with Allianz, 5% with Munich Re and 0.4% with Central Re.
2. The reinsurance underwritten by Gan and Eagle Star was in slip form, providing inter alia:

"CONDITIONS: Full Reinsurance Clause NMA 416

Claims Coop Clause (Wording T.B.AGD)

Cancellation Clause as original plus 30 days.

….

All amendments, additions, deletions or endorsements hereto to be agreed by the Leading Reinsurance Underwriter only, such agreement to be binding on all other Underwriters subscribing hereto.

…."

In the slip, the second condition had originally read "Claims Control Clause (Wording T.B. AGD)", but the word "Control" was altered to "Co-operation" at Gan's request.

3. The Full Reinsurance Clause NMA 416 provides as follows:

"REINSURANCE WARRANTY CLAUSE (FULL R/I. CLAUSE No. 1)

(Approved by Lloyd's Underwriters' Non-Marine Association)

Being a Reinsurance of and warranted same gross rate, terms and conditions as and to follow the settlements of the ..........................................................................Company and that said Company retains during the currency of this Policy at least.......................................on the identical subject matter and risk and in identically the same proportion on each separate part thereof, but in the event of the retained line being less than as above, Underwriters' lines to be proportionately reduced."

Gan required and on 21^st March 1996 initialled the following Claims Co-operation Clause:

"CLAIMS CO-OPERATION CLAUSE

Notwithstanding anything contained in the Reinsurance Agreement and/or Policy wording to the contrary, it is a condition precedent to any liability under this Policy that:-

a) The Reinsured shall, upon knowledge of any circumstances which may give rise to a claim against them, advise the Reinsurers immediately and in any event not later than 30 days;

b) The Reinsured shall co-operate with Reinsurers and/or their Appointed Representatives subscribing to this Policy in the investigation and assessment of any loss and/or circumstances giving rise to a loss;

c) No settlement and/or compromise shall be made and liability admitted without the prior approval of Reinsurers.

All other terms and conditions of this Policy remain unchanged.

SCOR (UK) Clause 012 4/83"

4. A fire occurred in the building, damaging the machinery during the process of its erection, installation and commissioning. Winbond pursued a claim. Tai Ping in December 1996 claimed to avoid the original insurance for misrepresentation. Winbond reacted strongly, issuing proceedings on 6^th May 1997, which led to two court appearances and a site visit. From July 1997 negotiations took place, leading to a settlement or compromise in writing made between Winbond and Tai Ping on 30^th July 1997. In the present pleadings, Gan alleges, amongst other things, that Tai Ping failed to co-operate in the investigation and assessment of loss and agreed a settlement of Winbond's claim and admitted policy liability without Gan's approval.

The Issues Appealed

5. On 21^st June 2000 Longmore J. decided a number of preliminary issues, including issues of alleged misrepresentation and non-disclosure relating to the reinsurance. Following his decision, Tai Ping made an application for summary judgment on three further points, which were determined by Andrew Smith J. on 8^th February 2001. We are at present only concerned with (a) appeals (with permission of Longmore J) from his decision on four issues of law arising under the Claims Co-operation Clause and (b) an appeal by Gan (for which Andrew Smith J gave permission) on one of the three points that he decided. The remainder of Longmore J's decision is not under appeal, and the appeal by Gan against the other two points determined by Andrew Smith J. (for which permission has now also been granted) remains for hearing on a date to be fixed.
6. The four preliminary issues that Longmore J decided were:

"(ii) Whether, as a matter of construction, compliance by Tai Ping with the provisions of the claims co-operation clause was a condition precedent to liability on the part of Gan.

(iii) Whether, as a matter of construction, breach of subparagraph (c) of the claims co-operation clause is only established by showing that Tai Ping both settled and/or compromised Winbond's claim and admitted liability.

(iv) Whether there are to be implied into the slip policy the following terms: (a) that reinsurers may not withhold approval of a settlement unless there are reasonable grounds for withholding approval; (b) that reinsurers will respond with reasonable promptness to a request for approval of a settlement.

(v) Whether, as a matter of construction of the slip policy, Tai Ping, even if shown to have been in breach of the claims co- operation clause, is entitled to recover under the slip policy, if and to the extent that it is able to show that it was in fact and in law liable to Winbond in Taiwan, and under Taiwanese law for at least its proportion of the settlement figures of Taiwanese $2.65 million."

7. Longmore J answered these issues as follows:

(ii) Yes

(iii) Yes

(iv) (a) Yes

(b) Longmore J. did not think it useful to answer this issue, "because, if a request goes unanswered, it will be tantamount to a refusal".

(v) No

8. Gan appeals the judge's decisions on issues (iii) and (iv)(a), while Tai Ping (contingent upon Gan pursuing its appeal) appeals the decisions on issues (ii) and (v). The issue before us decided by Andrew Smith J. is whether the settlement agreement dated 30^th July 1997 contained an admission in writing. This turned upon whether such an admission was constituted by a recital in the settlement agreement that "a fire within the scope of coverage under [the policy] occurred on 14 October 1996 at the Winbond FAB III" and upon Tai Ping's agreement under the agreement to pay Winbond NT$2.65 billion "as the total amount due under the Policy" (in circumstances where the claim had been for some NT$3.70 billion).
9. The four issues decided by Longmore J. are in certain respects inter-dependant. The issue decided by Andrew Smith J. depends on Longmore J's affirmative answer to issue (iii). The arguments open on both issues (iv) and (v) depend very much on the correct answer to issue (iii). The ambit of the Claims Co-operation Clause (cf. issues (ii) and (iii)) may also be material when considering whether both issues (iv) and (v) can be answered against Tai Ping. And an answer to issue (v) in Gan's favour may suggest the answer to issue (iv) in Tai Ping's favour given by the judge, and vice versa.
10. Longmore J. observed that, since Gan both put forward the Claims Co-operation Clause, and seeks to rely on it, it should in case of any real ambiguity be construed against Gan. However, he felt able to come to clear conclusions on each of the issues of construction without invoking that approach.

Issue (iii) – Claims Co-operation Clause, sub-clause (c): one or two subjects?

11. I start, like the parties before us, with issue (iii). The judge considered that the operation of sub-clause (c) of the Claims Co-operation Clause requires both a settlement and/or compromise and "an additional element before the condition precedent is to bite, namely, an admission of liability". He regarded it as "idle to speculate why reinsurers might only wish to invoke the clause if there is not only a settlement or compromise, but also an admission of liability", since "That is what the clause says, and that is the end of the matter".
12. I start with some basic principles governing interpretation of documents such as the present, as stated by Lord Hoffmann in Investors Compensation Scheme Ltd. v. West Bromwich B.S. [1998] 1 WLR 896, 912H-913C:

"(1) Interpretation is the ascertainment of the meaning which the document would convey to a reasonable person having all the background knowledge which would reasonably have been available to the parties in the situation in which they were at the time of the contract.

(2) The background was famously referred to by Lord Wilberforce as the "matrix of fact", but this phrase is, if anything, an understated description of what the background may include. Subject to the requirement that it should be reasonably available to the parties and to the exception mentioned next, it includes absolutely anything which would have affected the way in which the language of the document would have been understood by a reasonable man.

….

(4) The meaning which a document (or any other utterance) would convey to a reasonable man is not the same thing as the meaning of its words. The meaning of words is a matter of dictionaries and grammars; the meaning of a document is what the parties using those words against the relevant background would reasonably have been understood to mean.

(5) The 'rule' that words should be given their 'natural and ordinary meaning' reflects the common sense proposition that we do not easily accept that people have made linguistic mistakes, particularly in formal documents. On the other hand, if one would nevertheless conclude from the background that something must have gone wrong with the language, the law does not require judges to attribute to the parties an intention they plainly could not have had. Lord Diplock made this point more vigorously when he said in Antaios Naviera S.A. v. Salen Rederierna A.B. [1985] A.C. 191, 201:

"if detailed semantic and syntactical analysis of words in a commercial document is going to lead to a conclusion that flouts business commonsense, it must be made to yield to business commonsense".

13. Construction, as Sir Thomas Bingham MR (as he then was) said in Arbuthnott v. Fagan (30^th July 1993) [1996] LRLR 143 is thus "a composite exercise, neither uncompromisingly literal nor unswervingly purposive". To paragraph (5), one may add as a coda words of Lord Bridge in Mitsui Construction Co. Ltd. v. A-G of Hong Kong (1986) 33 B.L.R. 14, cited in my judgment in Sinochem International Oil (London) Ltd. v. Mobil Sales and Supply Corp. [2000] 1 Ll. R. 339, 344. Speaking of a poorly drafted and ambiguous contract, Lord Bridge said that poor drafting itself provides:

"no reason to depart from the fundamental rule of construction of contractual documents that the intention of the parties must be ascertained from the language that they have used interpreted in the light of the relevant factual situation in which the contract was made. But the poorer the quality of the drafting, the less willing the court should be to be driven by semantic niceties to attribute to the parties an improbable and unbusinesslike intention, if the language used, whatever it may lack in precision, is reasonably capable of an interpretation which attributes to the parties an intention to make provision for contingencies inherent in the work contracted for on a sensible and businesslike basis."

14. In the present case, I consider that Longmore J. stopped short too early in the process of interpretation. First, sub-clause (c) does not, on any view, read naturally or ordinarily. The judge's interpretation rests on the grammatical or dictionary view that the use of the word "and" introduces an additional element. But, even as a matter of simple linguistics, if an additional element was intended, this would be an unnatural way to express it. One would expect the sub-clause to have read: "No settlement and/or compromise shall be made in which [or whereby] liability is admitted" or possibly "No settlement and/or compromise shall be made and liability thereby admitted".
15. In sub-clause (c), there is on any view an element of ellipsis. The words "liability admitted" are, tacitly, subject to the initial word of the sub-phrase, "No", and the verb "shall be" is tacitly understood between them. If the initial negative had been re-expressed before the words "liability admitted", the effect would, I think, have been clear. The word "and" would then have introduced not an additional element of one topic, but a separate topic. The sub-clause would have two separate subjects: one settlement and/or compromise, and the other admission; neither could occur without reinsurers' prior approval. As Mr Edelman QC pointed out, it is not uncommon for "and" to introduce not an additional element, but a separate subject – take the phrase "Smoking and spitting prohibited" or "No beverage shall be drunk and [no] food eaten in this park except in the cafeteria". In the latter prohibition, I do not think that omission of the bracketed "no" affects even the ordinary understanding of the sense.
16. In the formulation that we have to consider, the omission to re-express "no" immediately before the words "liability admitted" gives the whole sub-clause, for reasons which are not easy to define, a different and less natural feel. I accept that, if a separate topic was intended, without re-expressing the word "no", a more grammatical method of achieving this effect would have been to use the word "or", rather than the word "and". Nonetheless, in my judgment the sub-clause has no very natural meaning and is, at the least, open to two possible meanings or interpretations - one the judge's, the other that it addresses two separate subject-matters. In these circumstances, it is especially important to undertake the exercise on which the judge declined to embark, that is to consider the implications of each interpretation. In my opinion, a court when construing any document should always have an eye to the consequences of a particular construction, even if they often only serve as a check on an obvious meaning or a restraint upon adoption of a conceivable but unbusinesslike meaning. In intermediate situations, as Professor Guest wisely observes in Chitty on Contracts 28^th ed. Vol. 1, para. 12-049, a "balance has to be struck" through the exercise of sound judicial discretion.
17. In the present case, any consideration of the implications of the rival constructions leads, in my judgment, both immediately and inevitably to the conclusion that the interpretation for which Tai Ping contends and which the judge adopted cannot have been intended or understood. There is no commercial sense in a clause having the meaning and effect for which Tai Ping contends. There is no sensible basis on which either party can have meant it to have, or thought that it had, such a meaning. This is not a matter of speculation, but in my judgment one of obvious business commonsense in a reinsurance context.
18. Firstly, Tai Ping's interpretation involves a conjunction of concepts which is not just unfamiliar and obscure, but, so far as one can detect its probable effect, of most curious and limited scope. What is meant by an additional element, consisting of "liability admitted"? The words "settlement" and "compromise" may well be coterminous. But, if and so far as they, or settlement alone, refer to paying a claim without disputing or reserving liability, on one view every such settlement might then be regarded as involving an admission. That can hardly have been intended, if "liability admitted" involves an additional element. So the sub-clause would, for some unspecified reason, appear to be looking for an express admission. And, if and so far as compromise (or settlement) refers to making, or agreeing to make, a payment in respect of a claim which has been subject of dispute, whether as to the existence or amount of liability, the additional element could seldom exist. The mere agreement to compromise could not (presumably) suffice. The concept of compromise of disputes involving liability (as distinct from quantum) with the additional element of an admission of liability is not easy to pin down (as the issue litigated before Andrew Smith J. illustrates) and seems likely to be rare in practice. On Tai Ping's interpretation, only a compromise of quantum, with (or perhaps following) an admission of liability, would seem clearly within sub-clause (c). But, even then, the whole impact of sub-clause (c) could always be avoided if the cedant's claims department was careful to avoid making any formal admission of any liability.
19. Looking at the general scheme of the sub-clause, Tai Ping's interpretation thus raises fundamental incongruities. What sense could there be in omitting to cover admissions simpliciter? What sense could there be in addressing only those settlements and/or compromises that were combined with admissions (whatever that concept may here involve)? What sense could there be in omitting to cover any settlement or compromise not involving any admission? Why should the somewhat quibbling issue that Andrew Smith J. had to address be critical to reinsurers' liability? No satisfactory answers are possible, on Tai Ping's interpretation.
20. In contrast, on Gan's interpretation, the sub-clause addresses two separate and familiar topics, each of great potential relevance to reinsurers: first, any settlement and/or compromise; and, second, any admission of liability. The former would actually commit insurers vis-à-vis their original insureds (and reinsurers vis-a-vis insurers, under the "follow the settlements" provision in the Full Reinsurance Clause). The latter would potentially commit insurers vis-à-vis their insureds, and make it more likely that the insurers would have to enter into a settlement or compromise or have judgment given against them.
21. I do not think that this is a case where it is necessary to resort to any principle of last resort in cases of real ambiguity, such as construction against the person putting forward the sub-clause for incorporation into the contract. In my judgment, only one of the two possible interpretations makes any commerial sense, and this should be adopted. I would add that there is anyway some room for doubt what, if any weight, could, even as a point of last resort, attach to the fact that Gan, as reinsurers, put forward this particular Claims Co-operation Clause. This Clause appears to have been required by Gan in lieu of a "Claims Control Clause" referred to in the broker's slip. It might be of interest to compare the two, to see if the Claims Co-operation Clause was, in a material respect, more stringent than a claims control clause – though I appreciate that that observation assumes that the characteristics of the latter type of clause can be identified with some specificity. A second point that might have some materiality is that clauses such as the Claims Co-operation Clause are standard clauses, used in a range of reinsurances, where one might expect them to receive a uniform construction, whoever proposed them: cf Pioneer Shipping Ltd. v. B.T.P. Tioxide Ltd. (The "Nema") [1982] A.C. 724, 737F-H, per Lord Diplock; and Miramar Maritime Corp. v. Holborn Oil Trading Ltd. [1984] A.C. 676, 682C-F, per Lord Diplock. However that may be, I consider, as I have said, that there is no need or basis to invoke the principle of construction against the profferor in this case.
22. I do however find further confirmation as to my conclusion on the meaning of sub-clause (c), in the evident origin and rationale of the present form of Claims Co-operation Clause. The case of Insurance Company of Africa v. Scor (U.K.) Ltd. [1983] 1 Ll.R. 541; aff'd [1985] 1 Ll.R. 312 was decided at first instance by Leggatt J. on 21^st December 1982. The last line of the present clause, and a comparison between the present clause and that considered by Leggatt J., indicate that the present clause was drafted in April 1983 in direct response to Leggatt J's reasoning. The clause before him read:

"It is a condition precedent to liability under this Insurance that all claims be notified immediately to the Underwriters subscribing to this Policy and the Reassured hereby undertake in arriving at the settlement of any claim, that they will co-operate with the reassured Underwriters and that no settlement will be made without the approval of the Underwriters subscribing to this Policy".

23. Leggatt J. held that this clause consisted of a condition precedent concerned only with notification, followed by a simple two-fold undertaking by the reassured to co-operate and not to make any settlement without the reinsurers' approval (p.553). There was no suggestion that, but for the breach of the latter undertaking, the claim would not have been settled or would have been settled for a smaller sum, so any breach was immaterial. The Court of Appeal in November 1984 upheld Leggatt J's analysis of the nature of the obligations contained in the clause; it held further that, while failure to obtain reinsurers' approval to a settlement would prevent the reinsured from relying on the "follow the settlements" clause that was also incorporated, it remained open to the reinsured to prove that it was in fact liable in the amount settled.
24. Comparing the clause before Leggatt J. with the present, it is apparent that the draftsmen have separated out the three parts of the clause identified by Leggatt J. and have determined to make each into a condition precedent. In the process of separating out each, each has been expanded. In sub-clause (c), the concept of "compromise" has been (perhaps unnecessarily) covered expressly. There is so far no hint of any intention to limit or weaken the protection afforded to reinsurers. Yet, if Tai Ping is right, that is what the insertion of the further words "and liability admitted" must have intended - and intended pursuant to a policy that remains inexplicable. It seems to me obvious that what has, in reality, happened is that the draftsmen conceived that it would be desirable to cover not merely settlement and, by addition, compromise, but also admission as (separate) subjects. The precise phrasing, the use of "and/or" and the elliptical addition of "and liability admitted", were less than ideal, but the true sense is confirmed by this examination of the history. There should be limits to one's enthusiasm for interpretation by historical exegesis, but Scor (UK) and the principles for which it stands remain the stuff of insurers' and reinsurers' everyday life, as witnessed to some extent by the last line of the present clause, and in this case the process seems to me eminently justified.
25. I therefore disagree with the judge, and answer issue (iii): No.

Issue (ii) - condition precedent?

26. Before Longmore J, Tai Ping did not contest the affirmative answer that the judge gave on issue (ii), and, although the answer was challenged in the appeal documents, the challenge was not maintained before us orally. The clause is emphatic in stating that "Notwithstanding anything contained in the Reinsurance Agreement and/or Policy working to the contrary, it is a condition precedent to any liability under this Policy that ….". In my judgment, the judge was right on this issue. The clause introduces a "condition precedent" to "any liability under this Policy". It goes on in the ensuing sub-clauses to specify three matters. The first is clearly capable of being a condition precedent. The last likewise. The second is now submitted by Tai Ping to be "plainly too vague to qualify as a condition precedent". But the fact that it may require a degree of judgment to decide whether there has been a failure to "co-operate" cannot, in my judgment, preclude a conclusion that co-operation was a condition precedent. It can often occur that the continued operation of contractual obligations depends upon a careful evaluation and judgment of the significance of complex factual circumstances, e.g. when deciding whether there has been a repudiatory breach (cf also the cancellation clause in The Antaios, cited above). Whether or not sub-clause (b) here refers to any or only significant or serious failures to co-operate raises a potential separate issue which cannot impact on the question whether the Claims Co-operation Clause is capable of being read as a condition precedent. Whatever its scope, the Claims Co-operation Clause operates in my judgment as a condition precedent.

Issue (v) – condition precedent to what?

27. Issue (v) raises for consideration whether the Claims Co-operation Clause, or at any event sub-clause (c), intends a condition precedent to liability of any sort under the Policy, or precedent only to liability that would otherwise arise under the "follow the settlements" provision in the Full Reinsurance Clause. The judge's answer to this issue also appears to me to follow from the express language of the clause; the clause introduces a condition precedent to "any liability under this Policy" and says expressly that its provisions are "[n]otwithstanding anything contained in the Reinsurance Agreement and/or Policy working to the contrary". The judge distinguished the reasoning and conclusions in Scor (U.K.) [1985] 1 Ll.R. 312, 331 (left) per Robert Goff L.J. and 332 (right) per Fox L.J. and at first instance in Vesta v. Butcher [1986] 2 Ll.R. 179, 187 by Hobhouse J. In the former case, as I have already noted, the requirement of the claims control clause was "that no settlement shall be made without the approval" of the reinsurers, and the Court of Appeal was able to say that the only relevant effect was to circumscribe the power of insurers to make settlements binding upon reinsurers, so that reinsurers would only be bound to follow a settlement when it had received their approval. In the latter case, the clause provided that "In the event of loss hereunder, no payment, offer or compromise shall be made without the consent of underwriters who shall have sole control of all negotiations". Hobhouse J. described the clause as a qualification of the full reinsurance clause. In neither case was the clause expressed as a condition precedent or warranty, and in neither was it suggested that the reinsurers had suffered any recoverable loss arising from failure to comply with it.
28. Here, by contrast, the wording is very specific. Its apparent intention is clear. Nothing in the nature of any of its three sub-clauses suggests to me that it was or can have been intended to be directed simply to situations in which insurers might subsequently invoke the Full Reinsurance Clause. Prompt advice (sub-clause (a)) and co-operation in the investigation and assessment of any loss (sub-clause (b)) have evident importance for reinsurers in relation to any claim - whether or not insurers subsequently can or do rely as against reinsurers on a "follow the settlements" provision. Merely because insurers might ultimately be able to prove that the claim was one for which they were liable in law does not mean that prompt advice to and co-operation with reinsurers might not have led to a more beneficial outcome.
29. Mr Railton submitted that, since sub-clause (c) deals with settlements and compromises, the condition precedent that it involves can readily be limited to the application of the "follow the settlements" provision in the Full Reinsurance Clause. That is an ingenious argument, but, even on Tai Ping's interpretation of sub-clause (c), it would be very difficult to reconcile with the opening words of the Claims Co-operation Clause. The present case, unlike Scor (U.K) and Vesta v. Butcher, is one where the words really brook no doubt. More fundamentally, however, Mr Railton's argument depends upon Tai Ping's interpretation of sub-clause (c), which I have now rejected. If the phrase "liability admitted" is a separate subject, its purpose can only have been to preclude insurers, who have admitted liability, from recovering from reinsurers by proving that the admission was actually correct. Correct or not, any admission is likely to weaken insurers' position, and prospects of achieving any sort of favourable settlement. The "follow the settlements" provision can have nothing to do with simple admissions of liability. That being so, it would also be unrealistic to try to limit the significance of the reference to "settlement and/or compromise" to circumstances where the "follow the settlements" provision is subsequently sought to be relied upon.
30. In reaching a final conclusion on issue (v), I have borne in mind the conclusions which I express below on issue (iv)(a). The judge considered that his answer to issue (v) supported his answer to issue (iv)(a). The answer that I would give to issue (iv)(a) below is more qualified than the judge's, although the qualifications probably result from hearing fuller argument and feeling it necessary to look into the practical impact of the proposed implication; and so I am not convinced that the judge would disagree with it. Be that as it may, I do not think that my answer to issue (iv)(a) can alter the conclusions expressed regarding the scope of the condition precedent created by sub-clause (c). The power to withhold approval is granted to reinsurers for their protection; and the remedy, if the stringency of its operation creates practical difficulty where there is disagreement between insurers and reinsurers, or between different groups of reinsurers, lies, I think, to a large extent in insurers' hands, when they negotiate their reinsurances.
31. My conclusion is therefore that the judge gave the right answer to issue (v). Breach of the condition precedent created by the Claims Co-operation Clause, including sub-clause (c), has the effect that Tai Ping cannot recover under the slip policy, by showing that it was in fact and in law liable to Winbond in Taiwan under Taiwanese law.

Issue (iv)(a) – must reinsurers have reasonable grounds for withholding approval?

(1) The judgment below and the general issue

32. In Gan's submission, the judge was right on issue (v), as I have held, but wrong on issue (iv). In short, Gan submits that the Claims Co-operation Clause introduces strict conditions precedent to any liability under the reinsurance, but that reinsurers are under no constraint whatever in deciding whether or not to approve a settlement and/or compromise.
33. The judge thought that this would make the reinsurance effectively optional and be surprising and unbusinesslike. He said:

"Suppose Tai Ping were in truth liable to Winbond. Is it supposed to be the party's [sic] intention that Gan need not pay? Suppose all the reinsurers of the remaining 98% other than Gan insist that as part of Tai Ping's promised cooperation in the earlier part of the clause, Tai Ping should take advantage of what appears to be a reasonable offer of settlement. Is it to be supposed that the 2% reinsurer can insist on the claim not being settled, however capriciously he withholds his approval? Mr. Wynter [counsel for Gan] had no answer to these problems. His undaunted submission was that, however arbitrary, however capricious, however unreasonable Gan were, they were entitled to decline to approve any settlement, and if a settlement were made without their approval, they had no liability."

34. The judge then reviewed authorities from other areas of the law, recognising that a term that consent would not unreasonably be withheld should be implied when necessary to uphold the purpose (or efficacy) of the contract under which a requirement for consent arose: see Price v. Bouch (1986) 53 Property and Compensation Reports 257; Clerical and General Life Assurance Society v. Fanfare Properties Ltd.; and Cryer v. Scott Brothers Sunbury Ltd. (1986) 55 Property and Compensation Reports 183. He concluded that, in view of his answer to issue (v), the implication of a term, that reinsurers' consent would not be unreasonably withheld to any settlement and/or compromise falling potentially within sub-clause (c), was necessary to give the reinsurance business efficacy.
35. A potential problem about the judge's approach to issue (iv)(a) is that it does not take account of the very limited scope and application which – on the judge's interpretation - sub-clause (c) had. Since I have given sub-clause (c) a more ample interpretation, this problem is less evident. There are, however, still situations where sub-clause (c) cannot make recovery by insurers under the reinsurance dependant on reinsurers' approval. Insurers may in particular be held liable by judgment without having made any prior admission. Either (as I think probable in the context of the present clause) the words "settlement and/or compromise" do not cover a judgment at all, or (as Leggatt J. and Stephenson L.J. thought in the Scor (U.K.) case: see pp. 557 and 323 respectively) the reinsurers must be "taken to approve" of any settlement involved in, or in meeting, a judgment.
36. So it is submitted by Mr Edelman for Gan that, although the requirement of reinsurers' approval is absolute in relation to settlements, compromises and admissions, it is no hardship for insurers to (have to allow themselves to) be sued to judgment before they have to pay their original insured or are able to recover any indemnity from their reinsurers. In response, Mr Railton for Tai Ping submits that insurers' normal aim and practice is to settle or compromise claims and to look to reinsurers to meet such settlements or compromises; that the presence of the "follow the settlements" provision, qualified though it is by the Claims Co-operation Clause, recognises that underlying desire; and that it would deprive the "follow the settlements" provision of all or much if its purpose if insurers had to submit to being sued to judgment, before they could be sure that their reinsurers would indemnify them.
37. In Scor (U.K.) at p.332 right, Robert Goff described the dilemma faced by the insurers, in the light of the combined effect of the full reinsurance clause and a claims co-operation clause in the form there present, if reinsurers refused to approve a settlement. He said:

"The insurers have therefore to decide what to do in the circumstances. In effect, they have got to prove a liability under the policy with the assured, in respect of which the reinsurers have agreed to indemnify them. They can attempt to prove this, without going to the expense of defending the claim by the assured in legal proceedings; or they may feel it more prudent to defend the claim, on the basis that they may defeat it or, if they do not, they will be in a better position to establish their claim against reinsurers. They well be in a very difficult position."

38. Gan submits that the only respect in which the dilemma faced by Tai Ping differs at this point from that described in Scor (U.K.) is that, if Tai Ping chose to enter into any unapproved settlement or compromise, it would deprive itself (by reason of the condition precedent) of the alternative course of establishing liability (viz. establishing that it was in fact liable to the original insured).

(2) The need to establish liability to the original insured

39. The submission seemed to imply that, without any judgment, award or settlement establishing its liability to its original insured, Tai Ping could still pursue proceedings against Gan, in which it could aim to prove its liability to the original insured. During argument, I suggested to Mr Edelman that, as a matter of general law, insurers' rights could only ever arise under a reinsurance such as the present if and when insurers' liability to their original insured had been ascertained as to its existence and amount, by virtue of a judgment, award or settlement/compromise; and that, if this was so, inability (under the Claims Co-operation Clause) to settle or compromise without approval would mean that reinsurers could oblige insurers to allow themselves to be sued to judgment, before reinsurers could become liable. Mr Edelman replied that it was an open question whether insurers' rights under such a reinsurance arose at the same time as the insured loss gave the original insured a right to claim on insurers, or whether they were postponed until insurers had ascertained their liability to the original insured in respect of such loss.
40. Mr Edelman associated this question with the issue left open by Lord Mustill in Charter Reinsurance Co. Ltd. v. Fagan [1997] AC 313, 385B, "whether there can be cases where a contract of reinsurance is an insurance of the reinsurer's [sic] liability under the inward policy or whether it is always an insurance on the original subject matter, the liability of the reinsured serving merely to give him an insurable interest" (and cf, on the nature of the particular reinsurance in that case, Lord Mustill at p. 387B-D and Lord Hoffmann at p.392E-H). But the issue identified by Lord Mustill, however it may be resolved in any particular case, does not determine the question whether insurers must ascertain their liability, by judgment, award or settlement/compromise, before they can claim recourse from reinsurers. Nor does this court's decision in Skandia International Corp. v. NRG Victory Reinsurance Ltd. [1998] Ll.R.I.R. 439 bear on this question. In that case, there had been a settlement. There was no "follow the settlements" clause and the issue was whether the insurers had established that they would have been liable under Texan law (in which connection the prediction of a Texan lawyer that a Texan jury would ignore Texan law was insufficient).
41. The current state of authority in this court appears to me to indicate that ascertainment of liability is a pre-condition to the accrual of insurers' rights against reinsurers - in whichever of the alternative ways, identified by Lord Mustill, liability is relevant: see e.g. Versicherungs und Transport A.G. Daugava v. Henderson (1934) 49 Ll.L.R.252, 253-4, per Scrutton and Maugham L.JJ, a fire reinsurance case (which I cited on this point in Charter Reinsurance Co. Ltd. v. Fagan at pp.333C and 341H-342A). A proposition that insurers' rights against reinsurers (or presumably reinsurers' rights against retrocessionaires under a retrocession) arise at the earlier stage of any original loss would also create very great potential difficulties, in the operation of reinsurances and retrocessions and in matters of limitation. It follows, under the present Claims Co-operation Clause, that the effect of Gan's case (combined with my answer to issue (iv)(a)) is that the only way in which insurers could ascertain their liability to their original insured, if reinsurers do not approve a settlement/compromise, is by allowing themselves to be sued to judgment.
42. This represents a more stringent regime than that existing both in Scor (U.K.) and at common law in the absence of any "follow the settlements" clause. Both under the reinsurance in Scor (U.K.) and at common law, insurers could, if they wished to take the risk, still settle, despite the lack of approval. This was not a breach of any condition precedent to reinsurance liability generally. It merely deprived insurers of the benefit of the "follow the settlements" provision present in the reinsurance in Scor (U.K.). Having thus crystallised their exposure by the settlement, insurers could still go on to prove that it related to liability that had actually existed apart from the settlement. Here, under the more stringent modern Claims Co-operation Clause, that is not possible.

(3) The suggested implication

43. Is the result so uncommercial as to make it necessary to imply into sub-clause (c) the qualification that approval should not be unreasonably withheld? Such difficulty as I have in answering this question arises, I think, largely because the implications of the suggested qualification are uncertain and the submissions before us touched on a number of potential consequences without differentiating between them. What is meant by a proviso, apparently placing an onus on reinsurers, that approval of a settlement should not be withheld "unless there are reasonable grounds for withholding approval"? The uncertainty arises from the general nature of the suggested qualification and is accentuated by the relevant factual pleading. Tai Ping simply alleges in paragraph 8(5) of its re-re-re-amended defence as follows:

"In the present circumstances:

(A) The proposed settlement was reasonable and businesslike and was supported by other reinsurers of the Defendant.

(B) The Claimant had the opportunity to respond to the proposed levels of settlement in the defendant's letters of 16 and 24 July 1997 and chose not to respond.

(C) The Claimant had in fact no reason to refuse approval of the settlement at NT $2.65 billion.

(D) In the premises, it was not necessary for the Defendant to obtain the prior approval of the Claimant, and/or the Claimant was in breach of one or both of the implied terms pleaded in sub-paragraph (4) above and cannot now rely on their own breach of the slip policy to defeat the Defendant's claim."

44. As to (A), the allegations that the proposed settlement would have been "reasonable and businesslike" and was supported by other reinsurers suggest that Tai Ping intends to argue that, once it and/or reinsurers of other parts of its interest, took a particular view, that was a view that Gan should either adopt or at least take into account. That raises both the question whose interests Gan is intended or required to take into account under sub-clause (c) of the Claims Co-operation Clause, and in what way. There may be a number of different views and approaches, all reasonable and businesslike, that may be taken in relation to a claim such as Winbond's. Is the small size of Gan's interest as reinsurers material? If, so, how should it be weighed against other interests? As to (B), the relevance of Gan's alleged failure to respond (so withholding approval) is not obvious. In (C), the bare allegation is made that Gan "in fact had no reason to refuse approval of the settlement at NT$2.65 billion". What that means also calls out for particularisation. It is unclear whether it intends (a) to allege that Gan refused approval arbitrarily, without regard to the prospects of any successful defence of the claim, and the merits of the proposed settlement or (b) to lead to some more wide-ranging review. Taking the pleaded implication and facts together with some of the submissions made before us, the latter appears to be intended.
45. I do not, however, consider that we can or should resolve issue (iv) by simply declaring that the particular implication pleaded is inappropriately formulated, so leaving the parties in this or a future case to argue whether some more precisely worded implication exists. First, that would itself be confusing, in this and future cases, for the very reason that the implication pleaded is imprecise, and a positive declaration that it is inappropriate could lead to future argument about what we intended to exclude. Secondly, both Longmore J's reasoning (above) and counsel's submissions have sought to identify the practical effects of a conclusion that the alleged implication exists or does not exist, and both have referred to authorities which in other contexts recognise more specifically formulated implications (e.g. not to act arbitrarily or for some improper or ulterior purpose, and to form a genuine view on the relevant issue). It seems to me therefore necessary to address the judge's judgment and the submissions presented to us in these respects. After the close of oral argument, I located the further case of Distillers Co. Bio-Chemicals (Aust.) Pty. Ltd. v. Ajax Insurance Co. Ltd. (1974) 130 CLR 1 (High Court of Australia), and drew the parties' attention to it as containing dicta by Stephen J. relevant to the points that had been argued. We received written submissions on it, and, as will appear, I have not in the event been able to agree with Stephen J's approach.

(4) The law and the previous authorities

46. Any implication could only be justified if it represents the obvious though unexpressed intention of the parties or is necessary for the business efficacy of the contract containing the Claims Co-operation Clause: The Moorcock (1889) 14 PD 64; Liverpool C.C. v. Irwin [1977] AC 239; and cf Chitty on Contracts Vol. 1 General Principles, paras.13-04 to 13-09. Gan's case is, in substance, that there is no call for any implication, since Gan is the party affected by any admission or settlement or by any judgment that may flow from any refusal to admit or settle any claim; and that, on this basis, it is also difficult to identify criteria by which one could anyway judge whether or not a reinsurer acted on "reasonable grounds" (or in a way that a reasonable reinsurer might have acted) in withholding approval.
47. To consider whether, as a matter of necessity or business efficacy, sub-clause (c) calls for the alleged implication, a prior exercise of interpretation is required. One must identify the context and purpose of the right to withhold approval. Subject to its terms and limits the reinsurance transfers financial risk from insurers to reinsurers. The risk requires to be established by judgment, award or settlement, before reinsurers can be called upon to indemnify in respect of it. The immediate context of sub-clause (c) consists in the "follow the settlements" provision in the Full Reinsurance Clause. Apart from sub-clause (c), reinsurers would be bound by settlements and compromises provided they were made bona fide and in a proper and businesslike manner by insurers: see Scor (U.K.). Sub-clauses (b) and (c) give reinsurers further protection. However, the further protection is not limited to the operation of the "follow the settlements" provision. It is by way of a condition precedent to any form of policy liability. The stringency of this effect is relied upon by Mr Railton for Tai Ping.
48. Before analysing the matter further, it is of interest to identify certain situations that are covered by authority or have been considered in dicta:
49. The first is the basic situation identified in Scor (U.K.) as existing under a "follow the settlements" provision, apart from any claims co-operation clause. The obligation to follow the settlements of a reinsured is subject to the implied proviso that the insurers must have acted both honestly and in a proper and businesslike manner in the settlement: see pp. 326, 330 and 334. In this situation, the insurer is responsible for the whole handling of the settlement, but the settlement imposes liability on the reinsurer. The rationale of the proviso is to protect B in circumstances where A is able to bind B's financial interest. We did not hear detailed argument on what might be involved in a "proper and businesslike" settlement. However, I think that it must contemplate that insurers will have regard to the prospects and risks attaching to the claim as a whole, whether it is wholly or only partly re-insured. It is that claim that has to be met by the insurers, whether by settlement, compromise or judgment, and whether it is wholly or only partly reinsured. I can leave aside any issues that might in some cases arise of alleged non-disclosure or breach of warranty, if for example a claim proved to exceed a value or maximum probable exposure that had been represented or promised to reinsurers. They would raise quite separate considerations.
50. The present situation differs from the basic situation considered in Scor (U.K.). The requirement of reinsurers' approval does not appear to me necessarily to render the proviso regarding insurers' conduct of any settlement either inapplicable or irrelevant. One can, for example, conceive of "unbusinesslike" conduct by insurers or matters within the knowledge of insurers, which was or were unknown to reinsurers when reinsurers' approval was given. However, sub-clauses (b) and (c) do aim to offer reinsurers for their own protection the opportunity to evaluate claims as well as to withhold approval of settlements/compromises; and Scor (U.K.) does not touch on the question whether a right given to B (Gan) for B's own protection to withhold approval of a settlement by A which would otherwise bind B's financial interest is subject to an implied proviso that the power should only be exercised "on reasonable grounds".
51. A second situation was considered in Groom v. Crocker [1939] 1 K.B. 194 (C.A.). There a motor insurance provided that the insurer (National Farmers' Union – "NFU")) "shall, so long as it desires, have absolute conduct and control of all or any proceedings against the assured". The assured driver, Mr William Groom, was, as everyone knew, innocent of any negligence in respect of an accident in which he was involved with a lorry belonging to Tear Bros. and in which his brother, Mr Aubrey Groom, as his passenger, was seriously injured. But Tear Bros.' insurers (Motor Union) were concerned with the extent of damages that a jury might award against Tear Bros. and NFU had another insured who had injured another insured of Motor Union by negligence in a quite separate accident. Motor Union in these circumstances proposed and NFU agreed a scheme whereby they would, as between themselves, split any financial cost of the two accidents, and in order to keep the costs as low as possible, Tear Bros. would deny liability while Mr William Groom would admit liability to Mr Aubrey Groom. They believed that Mr Aubrey Groom could thus (as happened) be induced to pursue only Mr William Groom and that the jury would be more merciful when it came to damages in an action between brothers than in an action against Tear Bros. Messrs. Crockers, solicitors, implemented this scheme on NFU's instructions. Mr William Groom only then learned what had been done in his name. Messrs. Crockers defended his claim on, inter alia, the basis that the policy allowed NFU, and so Messrs. Crockers acting on NFU's instructions, to devise and implement such a scheme. The defence failed. Messrs. Crockers were acting as solicitors for Mr William Crocker, as well as NFU, and were liable to Mr William Crocker for breach of contract, since the admission of negligence was not within the power to act on Mr William Crocker's behalf conferred on NFU by the policy, properly construed. The matter is put in somewhat different terms in the three judgments. Sir Wilfred Greene MR said at p.203:

"The question therefore is, Was the admission of negligence one which in the circumstances the insurers were entitled under the policy to require the assured to make? In my opinion clearly not. The right given to the insurers is to have control of proceedings in which they and the assured have a common interest - the assured because he is the defendant and the insurers because they are contractually bound to indemnify him. Each is interested in seeing that any judgment to be recovered against the assured shall be for as small a sum as possible. It is the assured upon whom the burden of the judgment will fall if the insurers are insolvent. The effect of the provisions in question is, I think, to give to the insurers the right to decide upon the proper tactics to pursue in the conduct of the action, provided that they do so in what they bona fide consider to be the common interest of themselves and their assured. But the insurers are in my opinion clearly not entitled to allow their judgment as to the best tactics to pursue to be influenced by the desire to obtain for themselves some advantage altogether outside the litigation in question with which the assured has no concern. This is what was done in the present case. I do not wish to use hard words, but I should have thought that a moment's consideration would have shown that the course adopted was as unjustified in point of law as it is assuredly repellent to the sense of business decency. I am not surprised that the jury should have wished to express their emphatic disapproval."

52. Scott L.J. said at p.223:

"Under condition 2 the society were given, in consideration of their indemnity, an absolute right to control their assured's defence; but the scope of this right was, in my view, subject to certain implied boundaries and limitations. It was not one which they would be entitled to exercise arbitrarily. They were bound to exercise a real discretion upon each question as it arose in the conduct of the defence, making each decision after due consideration of the circumstances of the particular case; not, of course, consulting the wishes of the assured as if he were an uninsured person, but taking their decisions with their minds on the facts of the particular allegations made against him, whilst not forgetting their own rights arising from the bargain expressed in the policy - namely, that in return for his indemnity their assured allowed them freedom to deal with the pecuniary risk to which they were exposed as economically for themselves as they could without bringing into the account extraneous considerations wholly foreign to the subject-matter of the insurance between him and them. It is easy to see, even within a field of discretion so defined, that a great sense of moral grievance might still be caused to an assured, as, indeed, the learned judge felt. Let me suppose the case either of a private chauffeur, or a lorry driver engaged continually in commercial transport on the roads, in either case a man of unblemished road reputation. The insurance company settles the case without consulting the employer. His driver knows he was guilty of no negligence, and the employer knows it, too. The driver feels that his good reputation has been sacrificed. Very hard cases happen in this way sometimes. It is, however, not necessary for this appeal to express any opinion on the difficult question of whether on such facts the assured would have legal ground for complaint, for the facts of the present case do not leave room for doubt; they fall wholly outside the field of discretion entrusted to the society by condition 2."

53. MacKinnon L.J. was the only member of the court to express the limitation in terms of reasonableness, saying this at pp.226 and 227-8:

"I think that the second condition in the policy, which provides that "The Society shall have "absolute conduct and control of all or any proceedings against "the assured" is subject to an implied term that the solicitor who is selected by the society shall act reasonably in the interests both of the assured and the society. As the assured is the litigant, the solicitor is his solicitor on the record and owes him the duty of a solicitor to his client.

…..

This means that the solicitor, nominated by the society, is the solicitor for the assured, who is his client. But he is also appointed by the society to protect its interests. If in regard to any question of tactics in conducting the litigation the solicitor has reason to discern a conflict, or possible conflict, of interest between the society and the assured, it is the duty of the solicitor to inform the assured of the matter. If the assured then insists on a course that the society disapproves, it can refuse to conduct or control the proceedings any longer, and leave the assured to do so at his own cost, and at the risk, if the society are right in their view, of not being able to recover that cost under his policy. If people act reasonably, it is obvious that the business method contemplated by the policy can be carried out with perfect smoothness."

54. Groom v. Crocker concerned a direct insurance. The harshness of the language used in the judgments owes something, doubtless, to the impact that an unjustified admission of negligence might have on a personal motorist. But, if a reinsurance were to contain a claims control provision in similar form to that in Groom v. Crocker, I consider that reinsurers' power to act on behalf of and to bind insurers would be subject to similar limitations: it should, at the least, be exercised in good faith and in the common interest on the basis of the facts giving rise to the particular claim and not arbitrarily or with reference to considerations wholly extraneous to the subject-matter of the particular claim. The present situation is, however, again different. Gan was given no power to act on behalf of or to bind Tai Ping and no power to instruct solicitors to act in their common interest. What Gan was given, for its protection, were rights to co-operation and to withhold approval of any settlement/compromise proposed by Tai Ping.
55. Thirdly, I note that liability insurances commonly contain conditions requiring the insured to refrain from making any admission or settlement without insurers' approval (as well as usually entitling insurers to take over the defence of any third party claim). Such conditions represent a closer parallel to the present reinsurance clause. In both contexts they appear to have given rise to remarkably little litigation. In the Law of Insurance Contracts, Prof. Malcolm Clarke refers at para. 27-4A to a case in the High Court of Australia, Distillers Co. Bio-Chemicals (Aust.) Pty. Ltd. v. Ajax Insurance Co. Ltd. (1974) 130 CLR 1, where one member of the court expressed views which deserve particular note. The prime issue was whether the condition barring admissions or settlements only applied when the insurer chose to take over the defence of the claim. The majority held not. Menzies J. (one of the majority) considered that the condition would no longer preclude a settlement without consent if and after the insurer repudiated his obligation to indemnify, but said as to other situations:

"In these proceedings there has been no argument upon the question whether in the circumstances the refusal of the insurer to give its consent in writing to the insured making any admission, offer, promise or payment in connection with any claim lacks bona fides or is unreasonable and, if so, whether the insurer is under an obligation to consent. These are important questions of law depending upon findings of fact which have not been made and which require full consideration after argument. In my opinion, the declaration sought does not raise these questions and upon them I express no opinion."

56. Gibbs J. (dissenting on the main point) referred in passing to "the principle that the insurer could not refuse his consent arbitrarily and in complete disregard of the interests of the insured". Stephen J. (the other member of the majority) considered it necessary to go into this aspect more fully, although it had not been argued. He upheld the judge's second declaration that the making of a settlement without consent would constitute a breach, even though insurers had not taken over the defence of the claim. But he focused both on the possibility that (wrongful) repudiation or other breaches by insurers of other clauses of the insurance might discharge any need for approval and, more materially in this case, upon situations where an insured could himself assert a direct interest in settlement, in particular because of his exposure over and above the policy limits. Thus, Stephen J. said:

"35. In these circumstances it is not irrelevant to note some features peculiar to insurance against risks of third party liability and in relation to which condition 2(a) is of special significance. Policies insuring against such risks and which combine an upper limit of indemnity with a prohibition upon settlement of claims by the insured without the insured's concurrence, as does the present policy, are very likely to give rise to conflicts of interest as between insurer and insured whenever a claim is made against the insured in excess of that upper limit of indemnity. The insured will be anxious to settle the claim at a figure within that limit; the insurer, however, will gain little from a settlement close to the limit and may prefer to have the case fought out rather than have it settled on such terms. An immediate conflict of interests then arises.

36. Where settlement for less than the upper limit of indemnity appears unattainable and the third party's case is a strong one the insurer's immediate financial interests would be better served if he could successfully resist the insured's claim to indemnity instead of accepting liability to indemnify and conducting an inauspicious defence. Condition 2(a), as interpreted in the present declaration, may afford to an insurer a means of attaining this result; if the insured is anxious to settle and the insurer, while refusing to take over the conduct of the defence, withholds its consent to a settlement, a breach of the condition against settlement without consent may be procured and subsequent liability to indemnify may be avoided. An insured's anxiety to settle will, of course, be acute if his entitlement to indemnity is in doubt and will in any event be substantial if the likely amount of any judgment against him far exceeds the limit of his entitlement to indemnity under the policy.

….

37. …. The terms of the declaration would permit it [the insurer] to refuse consent to any proposed settlement, while warning the assured that to settle would constitute a breach of condition.

….

50. However it is important that the making of this second declaration should not be thought to confer arbitrary power of refusal of consent upon the insurer. On the contrary I regard the power of the insurer as so hedged around with safeguards for the legitimate interests of the insured as to result in a situation in which the insured may be little worse off as a result of the declaration than it would have been had it been held that it might settle claims without the insurer's consent.

51. Where conflicts of interest arise as between an insured and an insurer, as they frequently will where an indemnity against liability to third parties is limited to a maximum amount, the insurer must exercise its powers under the policy with due regard for the interest of the insured. Perhaps because of the rather different way in which the granting of cover for third party automobile claims has developed in the United States as compared with the United Kingdom or Australia, there appear to be very few instances in the latter jurisdictions in which the courts have had to consider this aspect of liability insurance. One is, however, provided by Groom v. Crocker [1939] 1 K.B. 194. ….

….

55. The nature of the proceedings before Helsham J., concerned as they were only with a point of construction and not involving any investigation of the merits of the particular settlement which the insured was anxious should be effected with infant plaintiffs, makes it quite inappropriate on this appeal to seek to judge the insurer's refusal of consent to settlement. It would I think be improper for the insurer to refuse its consent to an otherwise proper and reasonable settlement if it did so not because there were reasonable prospects of a defence to claims succeeding but rather because, remote as those prospects might appear, they seemed to it preferable to a settlement involving it in payment up to the full amount of its limit of liability. To refuse consent in such a case, thereby exposing the insured to the probability of a judgment against it for an amount much in excess of the limit of liability, would be substantially to disregard the proper interests of the insured. On the other hand the insurer might quite properly refuse consent to a settlement which was attractive to the insured not because of an objective evaluation of the prospects of a successful defence but rather because it would avoid the adverse publicity of a trial or would enable the insured to share in, and require its insurer to share in, a liability which properly should be wholly borne by its parent company, the supplier of the allegedly harmful drug.

56. No doubt the limit of liability to indemnify and its relationship to the amount likely to be awarded against an insured will be an important factor in considering the propriety of any refusal of consent to a settlement. The insurer necessarily ceases to have any interest in that part of an award in excess of the limit of liability but the insured is vitally interested in that excess. ….

57. The second declaration produces, I have said, a result little less favourable to the assured than would a declaration that the consent of the insurer to any settlement was unnecessary; in the latter event the insured would nevertheless be required to have regard to the proper interests of the insurer and could not claim indemnity under the policy in respect of amounts payable under a settlement which did not reflect, by its terms, a reasonable evaluation of the prospects of a successful defence to a third party's claim. …."

57. I return later in this judgment to Stephen J's approach, and to reasons why I cannot associate myself with it.
58. A fourth situation, arising from that on which Stephen J. was focusing in the context of direct insurance, was also considered in Scor (U.K.). Reinsurers' refusal to approve a proposed settlement had there led to the insurers being sued to judgment and being held liable by a Liberian jury not just for the insurance policy limit of $3,500,000, but also for general damages of $600,000 and costs. In consideration of abandoning an appeal, insurers were able to obtain some time in which to pay the $4.1 million and to limit their costs liability to $58,000. The reinsurance was for 98% of the insured limit of $3.5 million, and reinsurance cover did not therefore extend - as such - to the general damages and costs awards. But insurers claimed that it was implicit in the reinsurance that, if reinsurers withheld approval of a settlement and as a result insurers suffered loss or expense which they would not otherwise have suffered, reinsurers would indemnify insurers against such loss and expense, or at least do so in the percentage (98%) for which insurers were otherwise reinsured. The court by a majority rejected any such implication. There was, Robert Goff LJ said, no such agreement and it could not be predicated that reinsurers would have assented to it, if it had been proposed by an officious bystander. It would have been open-ended, when the reinsurance limit was expressly $3.5 million, and would thus have been inconsistent with an express term. It would have unpredictable consequences, since it would presumably embrace any extra expenses, for example further legal or adjusters' costs incurred after an initial refusal of approval in persuading reinsurers (perhaps successfully) to change their minds. Fox LJ agreed, while adding that refusal of approval did not prevent insurers from simply paying, and then proving as against reinsurers that they had been liable to pay. That course is not open to Tai Ping, on the answer that I have given to issue (v).
59. This fourth situation does not offer any direct parallel to the present. Insurers' submission in Scor (U.K.) was that reinsurers should incur automatic responsibility for any extra loss and expense, resulting in the event from any refusal to approve a proposed settlement. This would have had the effect of passing additional responsibility to reinsurers, irrespective of the reinsurance policy limit and irrespective how reasonable reinsurers' refusal may at the time have been. In this connection, it is of interest to note that Stephenson LJ, who dissented, appears not to have based himself on the mere fact of withholding of consent, but on reinsurers' conduct which in his view constituted a breach of a mutual duty to co-operate provided by the claims co-operation clause and effectively compelled insurers to act as they did. Thus at p.324 (bottom left to middle of right column), he said:

"Here Scor not merely withheld approval; they withheld co-operation also, except to a detrimentally limited extent. By withholding both approval and co-operation, Scor forced ICA into a settlement by adjudication, which cost much more than settlement by accepting ATC's claim; yet Scor now contend that the language of one of these two clauses entitles them to withhold payment of what their refusal to approve or to co-operate has compelled ICA to pay.

I do not feel driven by the claims co-operation clause to attribute to the parties who contracted to be bound by it any such unjust and improbable intention. By their conduct rather than by any implied approval or request, Scor took themselves and ICA's resulting settlement by adjudication outside the scope of the claims co-operation clause, as the Judge was willing to hold alternatively, and so brought or kept that forced settlement within those settlements which Scor had contracted to follow."

60. I can envisage, in the reinsurance context, breaches so serious or perhaps of such a kind as to deprive reinsurers of the right to continue to expect their approval to be sought. Whether such breaches would always have to be, and be accepted as, repudiatory is not a matter that has been argued before us, or which I think it appropriate to consider further in this judgment. There has been no suggestion before us that Gan was here in breach of any implied duty (if any) that may have existed on their part to co-operate or engage in investigations, arising out of sub-clause (b) of the Claims Co-operation Clause. When a case comes before the courts, where the requirement of approval is said to have been abrogated or overtaken by reinsurers' breach, argument will have to be directed to the extent to which assistance can be obtained from passages to which I have referred in the judgments of Menzies and Stephen JJ. in Distillers Co. Bio-Chemicals (Aust.) Pty. Ltd. v. Ajax Insurance Co. Ltd. and Stephenson L.J. in Scor (U.K.).
61. Mr Edelman submits that the majority's refusal to recognise any implied obligation to indemnify in Scor (U.K.) indicates a clear answer to issue (iv)(a) in the present case. I cannot agree. There are important differences: Scor (U.K.) did not address situations of unreasonable refusal of consent – there does not appear to have been any submission that reinsurers' refusal of approval had been unreasonable; and insurers' case in Scor (U.K.) was aimed at recovering an indemnity irrespective of the reinsurance limit – whereas here Tai Ping's case aims simply at dispensing with the requirement of approval and binding reinsurers, up to the reinsurance limit, by a settlement made without approval. Therefore, I think that Tai Ping's case faces lower apparent hurdles than those faced by insurers in Scor (U.K.).
62. Finally, there are the property cases on which the judge relied and a number of other cases to which we were referred dealing with the exercise of contractual discretions. The judge recognised that the authorities do not justify any automatic implication, whenever a contractual provision exists putting one party at the mercy of another's exercise of discretion. It all depends on the circumstances: see Price v. Bouch, per Millett J. at p.258; Cryer, per Waite J. at p.202. In Price, power to approve building plans had been passed to a committee of all estate owners. But, although Millett J. did not accept that the court could review the reasonableness of the committee's decision, it was conceded that the committee had a duty to inspect and consider any application submitted "and to act honestly and in good faith and not for some improper or ulterior purpose", and also that, "if the committee took into account irrelevant considerations or failed to take into account relevant considerations or reached a perverse decision that no reasonable committee could possibly reach, then their decision could be impugned, for it would be ultra vires" (p.261). In Cryer a covenant taken on sale of building land required building plans to be submitted to the transferors for their approval before building work was commenced. The implication was accepted that the transferors would not withhold approval unreasonably, in which context the members of the court referred to withholding approval arbitrarily or capriciously. The courts have also read similar implied limitations into contractual discretions and powers, such as the discretion of directors to require such further proof as they should think necessary to establish a death or accident claim under an insurance (Braunstein v. The Accidental Death Insurance Co. (1861) 1 B. & S. 762); the discretion of a Protection and Indemnity Club to reject or reduce the claim of a member who fails to notify the claim timeously to the Club or to reduce his claim for failure to take such steps to protect his interests as he would have done if uninsured (The Vainqueur Jose [1979] 1 Ll.R. 557); the decision of a master – on the fifth voyage of a time charter for six months with an option to extend for a further six months ... to discharge at a port other than that ordered by the charterers, relying on a contractual discretion to do so, if he considered the stipulated discharge port dangerous (The Product Star [1993] 1 Ll.R. 397); and the power of a building employer to make decisions about delays, unforeseen events, extra expenses and changes in the works (Balfour Beatty v. Docklands Light Railway (1996) 78 BLR 49). In The Glacier Bay [1995] 1 Ll.R. 560, at p.566 Waller J. categorised various situations where a court may uphold a contract, including the following:

"(a) A contract may not on its true construction give a party any contractual right on the fulfillment of conditions, but may only provide one or other of the following:

(i) A right to have discretion exercised in their favour in which event in order to provide something enforceable, the Court construes the contract as imposing an obligation to act reasonably (see The Product Star [1993] 1 Ll.R. 397); …"

Later, at p.567 he also identified

"…. a class of case in which rights having been granted by the contract, and there then is granted to one contracting party a discretion which will enable that party to reduce his own liability; The Vainqueur Jose, [1979] 1 Ll.R. 557 is an example of such case. Once again in this area, as in the cases where the only contractual liability is to exercise a discretion, the Court imposes an obligation on the contracting party to exercise that discretion reasonably."

63. A similar limitation was recognised on the power of a reassured to define what should be regarded as "each and every loss and/or one event" for the purposes of a reinsurance (Brown v. GIO Insurance Ltd. [1998] LRLR 201). The view that Mr Brown was "entitled to take any decision that favoured [his] syndicate, goes too far if it was being suggested that even an unreasonable view was permissible".
64. I gain some assistance by analogy from these cases. In all of them, it seems to me that what was proscribed was unreasonableness in the sense of conduct or a decision to which no reasonable person having the relevant discretion could have subscribed. Mr Edelman submits that all these authorities are distinguishable, principally on the grounds that they concerned situations where (i) one party was placed in the hands of another in relation to a matter central to the contractual purpose, and was without any alternative recourse; here, in contrast, the insurers could always refrain from making any settlement and simply allow themselves to be sued to judgment, if they wished to preserve their right to a reinsurance recovery; and where also (ii) the party exercising the discretion or power was depriving the other party of a benefit or causing it loss, whereas here reinsurers are given the right to withhold approval because they are themselves directly exposed to loss.
65. As to the first consideration, what is sufficiently central to the contractual purpose to require an implication is, ultimately, a matter of degree. The cases do not limit the possibility of such an implication to situations where the other party is at risk of losing the entire benefit of the contract. Further, insurers' suggested alternative recourse (allowing themselves to be sued to judgment) may, in comparison with the position both at common law and in Scor (U.K.), have a rigour which neither party can sensibly have intended should be imposed in bad faith, capriciously or arbitrarily.
66. As to the second consideration, it identifies a distinction, which I did not understand Mr Railton to question as such. But, although the intended effect of the reinsurance was (at least within its terms and limits) to pass financial risk from insurers to reinsurers, insurers remained legally liable to the original insured and continued to be responsible for claims handling (in co-operation with reinsurers as required by sub-clause (b)). To be unable to settle and to have to be sued to judgment, as a condition of retaining any right to a reinsurance recovery, may again be acceptable at the insistence of a reinsurer who genuinely concludes that this is the appropriate course, but not at the insistence of a reinsurer who has reached no such conclusion or is acting capriciously or arbitrarily. Further, the meaning of sub-clause (c) can hardly depend upon an assessment whether or not all financial exposure had been transferred to reinsurers, something which may depend upon the precise terms and limits of the particular reinsurance, or the nature and size of a particular claim.

(5) Analysis of present reinsurance

67. Against this background, it is necessary to decide whether there are any and, if so, what limitations, under the present reinsurance, on the withholding by reinsurers of their approval of a settlement. I consider that at least one general limitation falls to be acknowledged. The withholding of approval must reflect the context in which and purpose for which sub-clause (c) gives reinsurers the right to withhold approval. The context includes, as I have said, the fact that Tai Ping as insurer retains a legal interest, and is the company through whom the financial burden passes to reinsurers, and the general purpose is the protection of Gan's interests as reinsurer. Whether or not Tai Ping might be financially affected at the end of the day, I do not think that it can have been contemplated that Tai Ping should be exposed to having judgments given against it at the instance of third parties, unless Gan genuinely determined this to be appropriate having regard to the characteristics of the particular claim. I would therefore accept as a general qualification, that any withholding of approval by reinsurers should take place in good faith after consideration of and on the basis of the facts giving rise to the particular claim and not with reference to considerations wholly extraneous to the subject-matter of the particular reinsurance.
68. Contrary to Mr Edelman's submission, this conclusion does not involve an inadmissible extension of the duty of good faith in insurance law or of the consequences of breach of any such duty. The qualification that I have identified does not arise from any principles or considerations special to the law of insurance. It arises from the nature and purpose of the relevant contractual provisions. It is thus in no way inconsistent with the approach of the House of Lords in Banque Financière de la Cité v. Westgate Insurance Co. Ltd. [1991] 1 A.C. 249 and Manifest Shipping Co. Ltd. v. Uni-Polaris Shipping Co. Ltd. [2001] 1 A.E.R. 193. It is possible to envisage circumstances in which reinsurers' refusal of approval could take place for reasons extraneous to the claim and prejudice insurers. A clear case would be refusal of approval, not for any reason connected with the merits of the claim, but as part of an attempt to influence an insurer's attitude in relation to a matter arising under another quite separate reinsurance or to harm an insurer as a competitor in respect of other business or in the eyes of a local regulator. Another example might be that of a reinsurer who withheld approval because it had decided, for reasons unrelated to the particular claim, that it wished as reinsurer to prolong payment of any claims for as long as possible, however obvious it might be that they would have to be met in full and should as claims be settled on the best terms possible. However, we did not hear detailed argument on what cases might fall within such a category, and I do not intend to express final views on particular situations such as the latter.
69. This leads me to the considerations and interests which reinsurers may and should take into account when deciding whether or not to withhold approval of a proposed settlement. Reinsurance policy limits or terms may (as Stephen J. pointed out) result in cases in which reinsurers' are not the only interests potentially affected by the grant or refusal of approval to an admission or settlement. Tai Ping does not plead that this represented a problem in the present case. Reference was however made in evidence and submissions before us to threats by Winbond to claim damages for bad faith or triple damages under a local statute.
70. The right to withhold approval is provided for reinsurers' protection, but that does not necessarily mean that they can exercise it in or solely with their own sectional interests in mind. The reinsurance cannot contemplate either that insurers would look only at their retained interest, when deciding what settlement or compromise to propose, or that reinsurers would look only at their own financial exposure, when deciding what settlement or compromise to approve. Stephen J in the Distillers Co. Bio-Chemicals case appears to have treated the situation as one of conflicting interests and to have considered that liability insurers would have to take into account, and (presumably) balance against their own interests, the interests of their insured. That approach, as well as the sweeping summary of its effect in paragraphs 50 and 57 of Stephen J's judgment, appears to me to present problems. Suppose that a reinsurer takes the view that there is a 50:50 chance of defeating liability altogether, and is quite prepared to carry the whole risk up to the reinsurance limit, if liability is established. Why should he not, if he views his own interests, withhold approval? Suppose, however, that, if liability is established, then it will very greatly exceed the reinsurance limit. What weight must the reinsurer give to that when deciding whether or not to withhold his consent to the settlement proposed by the insurer? Is it material whether the insurer has further reinsurance for a layer above the reinsurer's layer, and must the reinsurer then consider the significance of the exposure for those excess layer reinsurers (if he knows of them)? If there is no such further reinsurance, the whole problem could be regarded as arising from a failure by the insurer to identify correctly, or take out a large enough reinsurance to cover, his maximum exposure. In any event, it could be regarded as arising from a failure to qualify expressly the requirement of approval of settlements by each reinsurer for its separate interest.
71. So far, this also approaches the problem as if it concerned a reinsurance of the whole of Tai Ping's interest for whatever was the relevant layer. In fact, of course, the present reinsurance with Gan was for no more than 2%, part of 35% of the whole Winbond risk (in other words of less than 6% of Tai Ping's interest). If reinsurers must balance their own against insurers' interests, when deciding whether to approve a settlement proposed by insurers, would this be significant? The judge (although he thought that Gan were only insuring as little as 2% of Tai Ping's interest, rather than 2%, part of Tai Ping's 35%) appears to have thought that this could be important. I do not think that it could be. One must ask: at what point would a reinsurer have to give weight to a retained interest? And what weight? If Tai Ping had reinsured three-quarters or half its 35% line with Gan, how much weight should Gan give to that, compared with the weight it is suggested that it should have given to the fact that it took less than 6% of Tai Ping's line? And would it make any, and if so what, difference if the interest is not retained, but reinsured (as here) with other reinsurers? Again, any problem could be said to arise from the way in which Tai Ping had arranged its reinsurances. If (as here) an insurer places a separate reinsurance, on a separate slip, without obtaining the benefit of provisions binding reinsurers to follow the lead reinsurer on the main slip, the insurer has, on the face of it, to treat each reinsurance separately, and, as regards each reinsurance, to obtain the approval of the reinsurer to any admission or settlement. There is an inevitable risk that reinsurers on different slips may take different approaches.
72. Because of the difficulties involved in treating the situation as one where competing interests fall to be balanced, I believe that a different analysis is required. Sub-clauses (b) and (c) address co-operation, settlement or compromise and admissions made in respect of claims made against the insurer by original insureds. Insurers report to and co-operate with their reinsurers in respect of such claims, viewing them as a whole. The "follow the settlements" provision applies to honest, proper and businesslike settlements and compromises in respect of such claims. Sub-clause (c) gives reinsurers the further protection that they will only be obliged to follow or provide any indemnity after a settlement or compromise, or provide any indemnity after any admission, made in respect of such a claim, if it has been made with their approval. What is, I think, contemplated, as regards both insurers and reinsurers, is that each should, in performing his role under sub-clause (b) and (c), view the original claim, so far as possible objectively and as a whole, without regard to his narrower sectional exposure as insurer or reinsurer. Just as it would be wrong for insurers to settle a claim without real regard to its merits, because the burden would fall on their reinsurers, so it would, for this reason, be wrong for reinsurers to insist on a claim being fought regardless of its real merits, e.g. because the reinsurance cover was very limited and they hoped that some complete defence might emerge. This follows not from any requirement to balance competing interests, but from a recognition that both insurers in proposing and reinsurers in deciding whether or not to approve any proposed settlement, compromise or admission should look at and seek to resolve one and the same question: should there be such a settlement, compromise or admission in respect of the original claim, viewing that as a whole? That is the common interest to which they must be taken to have subscribed.
73. If there is any further implication, it is along the lines that the reinsurer will not withhold approval arbitrarily, or (to use what I see as no more than an expanded expression of the same concept) will not do so in circumstances so extreme that no reasonable company in its position could possibly withhold approval. This will not ordinarily add materially to the requirement that the reinsurer should form a genuine view as to the appropriateness of settlement or compromise without taking into account considerations extraneous to the subject-matter of the reinsurance. But I would for my part be prepared to accept this further limited qualification – since the stringency of sub-clause (c) in relation to the position at common law and under Scor (U.K.) seems to me to require it, both as a matter of necessity and to give the sub-clause business efficacy in the rare cases in which it could be material.
74. On this basis, I repeat that I do not consider that there can be any materiality in the extent of any interest retained or reinsured elsewhere by insurers. There is no question of balancing the views or financial interests of a small reinsurer against those of an insurer or those of the reinsurers of other parts of the insurer's exposure. If, as here, it is a condition precedent to liability under a reinsurance that the particular reinsurer has the right to withhold approval of any settlement or compromise, then that reinsurer is, in respect of his interest and whatever its size, entitled to exercise its own judgment as to the appropriateness of any settlement with respect to the claim, regardless of what others may think.
75. The possibility that disagreement may arise as to the merits of a proposed settlement, compromise or admission is inherent in sub-clause (c). If an insurer places a 100% reinsurance, with a claims co-operation clause in present form, he accepts the risk that his reinsurer may, when evaluating the claim, reach a different conclusion to his own on the subject of settlement or compromise. If he places only a 50% reinsurance, with such a clause, he accepts the risk that his reinsurer may insist on its own view of the merits of a proposed settlement. If he places ten reinsurances, each for 10%, and fails to insist on a leading underwriter clause, whereby all reinsurers must follow a leader, he risks, at least in theory, being presented with up to ten different views as to what would constitute a reasonable settlement, and all of them may be reasonable. The situation in this respect differs only in degree from that where reinsurances are entered into on differing terms, for example one with a claims control and another with no more than a claims co-operation clause. There is an inherent risk of conflict, which cannot be resolved by requiring one or the other reinsurer to forego the rights granted to it. Further, in the absence of appropriate mechanisms, like a leading underwriter clause in all reinsurances, for ensuring unanimity of view, problems are just as, if not more, likely to arise from the holding of different but still reasonable views, as opposed to wholly unreasonable views.
76. In summary, the right to withhold approval was, here, Gan's, and no-one else's. It is a right to be exercised in good faith after consideration of and on the basis of the facts giving rise to the particular claim, and not with reference to considerations wholly extraneous to the subject-matter of the particular reinsurance or arbitrarily. It is to be exercised by considering the claim as a whole. The court cannot substitute its own view of the reasonableness of a reinsurer's decision to withhold approval under sub-clause (c).
77. Further, I see no basis for requiring a reinsurer, before withholding approval, to be able to establish positively that there are reasonable prospects of defeating a claim or bettering the proposed settlement. It is insurers, and not reinsurers, who receive, investigate and have the handling of any claim, although sub-clause (b) requires them to co-operate with reinsurers. The right to withhold approval under sub-clause (c) must allow a reinsurer to take the view that a particular claim is one that should be strictly proved by the original insured, or to take a view as to an appropriate level of settlement prior to its being so proved. The level of appropriate settlement, and the point at which insurers should dig their heels in, in relation to a disputed insurance claim are often difficult matters of judgment or feel, on which different people may well hold different views. A reinsurer is entitled under sub-clause (c) to impose his own judgment and policy on such matters. If and so far as Stephen J. in the Distillers Co. Bio-Chemicals case thought that the right to withhold approval there was likely to make "little" difference, it seems to me that, under the present reinsurance, it (or its exercise) can in some circumstances make a considerable difference.
78. In the circumstances, I would not consider it appropriate to grant the declaration sought by issue (iv)(a). Against the background of the pleading and the submissions presented to us, the making of a declaration in such unspecific terms could give rise to misunderstanding and argument, both in this case and in future. There is one respect, at least, in which I cannot agree with the interpretation that the judge evidently put on the declaration sought (see paragraphs 33, 71 and 74-75 above). However, this judgment will speak for itself as to the limited qualifications on the right to withhold approval that I, for my part, consider appropriate.

Conclusion

79. I would therefore uphold Longmore J's answers to issues (ii) and (v). I would answer "no" instead of "yes" to issue (iii). This makes it both unnecessary, and indeed inappropriate, to attempt to address the issue before us arising out of Andrew Smith J's judgment. A negative answer to issue (iii) means that the issue which it fell to Andrew Smith J. to answer was based on a false premise. Lastly, I would set aside Longmore J's affirmative answer to issue (iv)(a) and make no specific declaration on that issue, leaving the terms of this judgment to stand as guidance.

LORD JUSTICE LATHAM:

80. I agree. As to issue (iii) I have read the judgment of Sir Christopher Staughton in draft. I would respectfully adopt his straightforward and robust approach. Read literally the clause does not make grammatical sense. Clearly, however, some meaning must be attributed to the words used. In deciding what that meaning should be, it is inevitable that the overall context in which the words appear must be considered. This necessarily involves some analysis of the extent to which any business or commercial purpose can be discerned in the rival arguments. I can see none in the argument that the words should be construed conjunctively. Conversely, there is clear sense in construing them disjunctively.
81. As to issue (iv)(a), I have come to the conclusion that, for the reasons given by Mance LJ, the right to withhold approval for a settlement or compromise is not unqualified, for the reasons he has given. But, like him, I would decline to grant a declaration.

Sir Christopher Staughton:

82. On three of the four preliminary issues which we have to consider I have reached the same conclusions as Mance L.J., although not wholly by the same route in the case of issue (iii). Issue (iv)(a) I would answer, No, like issue (iii). The result is that I would allow the appeal.

Issue (iii) – "And liability admitted".

83. It is apparent in my view that this part of the Claims Co-operation Clause does not have an ordinary English meaning. Something has gone wrong. In those circumstances we are entitled and bound to see if we can divine what the parties intended to say. That is in my view perfectly obvious: no settlement or compromise shall be made and no liability shall be admitted.
84. The alternative possibility is that they meant to say, there shall be no settlement and/or compromise with, in either case, an admission of liability. If they had said that in plain terms I would have given effect to it. But they did not. I decline to infer that it is what they meant. It is highly improbable that they wished to allow a settlement or compromise if it stood on its own, but to ban a settlement and/or compromise together with an admission of liability. It would make no sense.
85. In reaching that conclusion I have not had recourse to what is "unbusinesslike", or to "commercial sense" or even to "business commonsense", which are all frequently invoked since the decision in Antaios Naviera S.A. v. Salen Redierna A.B. (1985) A.C. 191. Those expressions are a temptation for the court to make a contract for the parties; and it has been said many times that the court should not do that. See for example, Lord Mustill in Charter Reinsurance Co. Ltd v. Fagan (1997) AC 313 at p.388C.
86. If necessary I would refer, as Lord Mustill did, to what Lord Reid said in Wickman Machine Tool Sales v. S. Schuler A.G. (1974) AC 235 at p. 251:

"The fact that a particular construction leads to a very unreasonable result must be a relevant consideration.The more unreasonable the result the more unlikely it is that the parties can have intended it, and if theydo intend it the more necessary it is that they shall make that intention abundantly clear."

So far as I am aware that sentiment has not been criticised or reversed.

Issue (ii) – Condition precedent?

Issue (v) - Condition precedent to what?

87. On these issues I agree with the reasoning and conclusions of Mance LJ.

Issue (iv)(a) – must reinsurers have reasonable grounds for withholding approval?

88. Fifty years ago the conscience of the judges was troubled by the growing tide of exemption clauses in consumer contracts. So there grew up the doctrine of fundamental breach. Broadly speaking this was a rule that some breaches of contract were so serious that no exemption would excuse the contract-breaker from liability. At first sight it had all the appearance of a rule of law, rather than interpretation of the contract.
89. Then the doctrine was invoked in a commercial case, where the contract was between two powerful corporations which were presumably of equal bargaining power: Suisse Atlantique Societe d'Armement Maritime SA v. NV Rotterdamsche Kolen Centrale (1967) A.C. 361. That was a case of deliberate breach of contract. The ship had been chartered for as many transatlantic voyages as she could perform in two years at a fixed rate of freight, and also a fixed rate of demurrage if she failed to load or discharge in the time allowed. The charterers deliberately delayed the loading and discharging, as it was cheaper for them to pay demurrage than to carry coal across the Atlantic when there was no market for it in Europe.
90. In the House of Lords Mr Henry Brandon QC, Mr RA MacCrindle QC and Mr Anthony Evans did not argue for a rule of law that some contractual obligations could not be excluded by exemptions; they relied on interpretation of the contract. The House did not demur from that course, as it might well have done in a case of general importance if it thought otherwise. So it was that the doctrine of fundamental breach was held to be a product of interpretation – and soon it went into decline. But the conscience of the judges was eventually salved, in the case of consumer contracts, by the Unfair Contract Terms Act 1977.
91. We thus have the position, unlike some countries of continental Europe, that there is no general rule of law that contracts must be performed in good faith, or that the terms of mercantile contracts must be fair. How could the charterers in the Suisse Atlantique case be said to have performed in good faith, or to have considered the interests of the shipowners as well as their own interests? It is not difficult to think of other cases where a contractual claim has been upheld although the claimant had suffered no discernible loss and wanted to bring the contract to an end for his own purposes. See for example in the House of Lords A/S Awilco v. Fulvia SpA di Navigazione (1981) 1 WLR 314 and Bunge v. Tradax (1981) 1 WLR 711 (where the delayed notice caused no loss to the sellers of itself).
92. There is also no general rule of law by which the courts can override what the parties have said in their contract, except as provided in the Unfair Contract Terms Act so far as applicable. I do not suppose that commercial men are dismayed by this situation. They are generally able to assess what risks they are undertaking and what risks they do not accept. A contract of reinsurance is surely a prize example of that. I would not suppose that they prefer a court to tell them that their contract means something different, even if it is the House of Lords – except of course when the loser finds that the risk has turned out to be unattractive.
93. There are two routes by which a court may arrive at a different meaning for a contract than that which it appears to bear at first sight. They are interpretation and implication. I do not see how interpretation can help on this issue (iv)(a); it was relevant to issue (iii), which I have already considered. As to implication, it is hornbook law that, as Mance LJ states, a term may be implied if it is so obvious that the parties did not think it worth saying (the officious bystander), or if it is necessary to give the contract business efficacy.
94. The preliminary issue ordered to be tried under this head was as follows:

"Whether there are to be implied into the slip policy the following terms: (a) that reinsurers may not withhold approval of a settlement unless there are reasonable grounds for withholding approval …"

95. That was the only term considered in argument before Longmore J, or in the hearing before this court. But subsequently Mance LJ invited written submissions from counsel on the case of Distillers Co. Bio-Chemicals (Aust.) Pty Ltd v. Ajax Insurance Co. Ltd (1974) 130 CLR 1. There is much material in that case for a wider consideration of this topic. But I share the view of Menzies J cited by Mance LJ:

"These are important questions of law depending upon findings of fact which have not been made and which require full consideration after argument."

96. In the first place, an implied term has to be capable of being formulated, although that can be done in any language that is appropriate. Stephen J at paragraph 50 of his judgment said:

"I regard the power of the insurer as so hedged around with safeguards for the legitimate interests of the insured as to result in a situation in which the insured may be little worse off as a result of the declaration than it would have been had it been held that it might settle claims without the insurer's consent."

That does , if I may say so, seem to me to envisage a remarkable implied term.

97. I am tempted to elaborate my concern about the efforts to rewrite the contract by way of an implied term in the judgment of Stephen J and, if I may respectfully say so, in the judgment of Mance LJ in this appeal. But no useful purpose would be served by my doing so. My answer to preliminary issue (iv)(a) is No. I do not reject the term because it is imprecise, although it is. I reject it because I do not think the term proposed necessary for business efficacy, and because I am confident that the parties would not have regarded it as so obvious that it goes without saying. Tai Ping were reinsured for 100 per cent of the 35 per cent of the risk that they wrote (though the reinsurers may or may not have known that); so it is not surprising if the reinsurers required a high degree of control of settlements. That an awkward situation might arise if other reinsurers took a different view (as they did) is a consequence of Tai Ping not requiring all to be bound by decisions of one leading reinsurer. Furthermore it would normally be difficult to determine whether the reinsurers' refusal to approve a settlement was or was not unreasonable. Is a reinsurer not entitled to require the insured to go into the witness box and give evidence of his loss on oath? Is that unreasonable?
98. Like Menzies J I express no view as to some other implied term which might be but has not been put forward. If a different term is proposed it should be formulated with clarity, and included in a pleading so that counsel can consider with their clients in detail what its implications may be. It should not emerge fully armed after the hearings both at first instance and in this court have been concluded. I would allow this appeal, with the consequences indicated at the beginning of this judgment.
99. If the parties believe that the market would welcome the term proposed in this case, or any other term on the same topic, they have only to persuade the market to adopt it for the future. I wonder whether they would succeed.
ORDER: Defendants' appeal allowed in part as per paragraph 79 of the Judgment; cross appeal dismissed; permission to appeal to the House of Lords refused. Gan Insurance to have 75% of costs before the Court of Appeal; costs before Longmore J to remain as they are; costs before Andrew Smith J to be reserved to the Court of Appeal, or until further order. Detailed assessment. (Order does not form part of approved Judgment)