Drake Insurance Plc v Provident Insurance Plc

The facts of this appeal might have been set by a committee of law professors with the express design of giving rise to points of interest and difficulty. Some at least of the points are of general importance and perhaps all the more so in that they concern a consumer contract of familiarity to all car owners and drivers, namely a motor insurance policy.

There is an added complication in that these proceedings do not arise between insurer and insured, but between two insurance companies. The claimant and in this court the appellant Drake Insurance plc (“Drake”) seeks to recover a contribution in equity from Provident Insurance plc (“Provident”, the defendant and in this court the respondent) on the basis that both companies were insurers of the same loss. A third party, Mr Beach, was injured by a car driven by Mrs Kamaljit Kaur. Mrs Kaur was driving her husband’s car. She was insured under her policy with Drake in respect of any car driven by her with its owner’s consent. Her husband, Dr Prim Singh, was insured under his policy with Provident in respect of the car which Mrs Kaur was driving at the time of the accident and Mrs Kaur was a named driver on his policy. However, Provident has purported to avoid its policy on the ground of Dr Singh’s non-disclosure. Drake disputes the validity of that avoidance. It is common ground that in the ordinary way double insurance of the same loss will entitle the insurer who pays the loss to recover in equity a rateable contribution from the other insurer. In this case when Mrs Kaur was sued by Mr Beach, Drake indemnified her by settling his claim. However, Provident disputes its liability to Drake to share in that loss for two reasons. First, it says that it validly avoided Dr Singh’s policy, so that there was in fact no double insurance. Secondly, it says that even if its avoidance was invalid, a special clause in Drake’s policy limited Drake’s liability in a case of double insurance to only half of the loss: so that by paying the loss in full, Drake was a volunteer and cannot recover. At trial Moore-Bick J agreed with both those defences. Drake appeals.

Dr Singh was in the habit of shopping around each year for low cost motor insurance. In February 1995 he made a proposal to Provident, a major provider of motor insurance. Provident offered a highly structured and automated service, under which approved brokers could accept proposals and fix the premium themselves. Such brokers were provided with detailed underwriting criteria. Using this information the brokers could open a proposal form on screen for completion electronically (followed by subsequent signature by the client of a hard copy, the only piece of paper in an essentially electronic process). The acceptability of a proposal and the rating of the premium were calculated by a rigid points system. Provident’s underwriting judgment entered into the formulation of that system and the setting of fixed premium rates for feeding into it: but the operation of the system itself was formulaic and Provident’s role was limited to one of monitoring the system to ensure that it was working properly, and no doubt to keeping the set premium rates under review.

A relevant example of the way in which the system worked is as follows. In order to determine whether a premium should be increased by reference to a proposer’s driving history a set number of points were awarded against convictions incurred and accidents suffered within the previous five years. A conviction for speeding known as SP30 carried 10 points. A “fault” accident carried 15 points. Below 17 points the proposer was charged a normal premium, at 17-29 points his premium would be increased by 25%, and so on. At 60 points or more the risk would be declined.

When in February 1995 Dr Singh approached his brokers, Hexagon Insurance Service Limited (“Hexagon”) for new insurance for his Renault car, he wanted third party, fire and theft cover for himself and in addition for his wife, Mrs Kaur, as a named driver. Neither of them had any convictions, but Mrs Kaur had suffered an accident just over a year before on 6 January 1994, when a third party had driven into the back of Dr Singh’s car when she was driving it. The accident was not her fault, but under Provident’s system it had to be recorded as a “fault” accident, despite its circumstances, until the matter had been settled by the third party in the insured’s favour. The accident had occurred when Dr Singh had been insured by another insurer (the Bradford & Pennine), and thus Provident was not at risk, but it was recorded by Hexagon on the proposal form as a “fault” accident in any event, as Provident’s system demanded. The following details were entered on the form: “Add[itional] Driver 1, Hit by T[hird].P[arty]. in rear, P[ersonal I[njury] No, Fault Yes, £2000 Own cost, £0 TP cost, NCB [no claims bonus] disallowed”. Thus Dr Singh rated on Provident’s points system at 15 points and was entitled to be charged premium at the normal level. Hexagon was itself authorised by Provident to accept the proposal and to set the premium by reference to Provident’s fixed underwriting criteria.

The proposal form was completed by Hexagon on screen, printed off and sent to Dr Singh for signature. The hard copy is dated 6 February 1995. In signing Dr Singh expressly confirmed that the broker was acting as his agent in providing information on the form and that that information was true to the best of his own information and belief. He also signed a direct debit form for payment of the premium in monthly instalments.

In February 1996 Dr Singh renewed his insurance with Provident. The certificate of insurance for the second year states that the insurance ran from 17 February 1996 to noon 17 February 1997. The policy number was 0ADC2195188. His premium remained at a normal rate.

Two relevant events had occurred in the previous year. The first was that Mrs Kaur’s January 1994 accident had been settled by the third party’s insurers entirely in the Singhs’ favour. Cheques in settlement were forwarded to them on 11 July 1995. The second was that on 18 December 1995 Dr Singh received an SP30 speeding ticket, which he paid, thus admitting the conviction, in January 1996. His licence was endorsed three points. When renewing his cover with Provident in February 1996 Dr Singh failed to disclose the conviction. He also failed to inform Hexagon that the January 1994 accident had been settled satisfactorily. Had he informed Hexagon of both those matters, the conviction would have counted 10 points against him, but the information about the settlement of the accident would have meant that that would have been reclassified as a “no fault” accident and thus would not have counted against him at all. In the circumstances he would still have been entitled to renewal at a normal rate. No documents recording the renewal have survived, other than the second year’s certificate of insurance, but it is accepted that Dr Singh was sent a renewal notice which, to quote from it, would have highlighted his –

“duty to disclose to the Company the existence of any factor which may affect the risk and to notify the Company of any driving convictions which may have arisen during the period of insurance now ending. Failure to do so may invalidate your policy.”

The first critical question raised on this appeal is thus whether Dr Singh’s non-disclosure of his speeding conviction entitled Provident to avoid his insurance, as it subsequently did.

On 20 May 1996 Dr Singh changed his Renault for a Mercedes and he was charged an additional premium. This was, I think, in part because he had now obtained comprehensive cover: his Mercedes was valued at £2000, whereas his Renault had only had a value of £500. He did not disclose his conviction. However, the papers do not disclose any paperwork or other evidence of a formal policy amendment, or indeed a new certificate.

On 2 July 1996 Mrs Kaur while driving the Mercedes collided with a motor-cyclist, Mr Beach. Unfortunately, it was a serious accident. Both of Mr Beach’s legs were broken, and ultimately he needed to have one leg amputated. Provident was immediately notified of a claim, first by telephone and then in writing. Dr Singh provided the telephone information: he was asked whether he had any convictions and said that he did not. The written claim form is dated 18 July 1996: this form merely asked whether the driver had any prosecutions pending, to which the correct answer given was no. As part of the claim procedures, Provident asked to be sent Dr Singh’s driving licence. This disclosed his SP30 conviction with its offence date of 18 December 1995. On 24 July 1996 Provident’s computer record shows a notation that “Non-disclosure of SP30 12/95, would have resulted in 25% load at renewal”. Although the record does not elucidate further, that was on the basis that the conviction together with the “fault” accident of January 1994 would have had that automatic effect. Without that fault accident, the non-disclosure of the conviction would have made no difference.

On 2 August 1996 Provident wrote to Dr Singh to avoid the policy. The letter said –

“The conviction is a material fact which was not disclosed to us. Had you informed us of the conviction we would not have insured you on the terms given.

“We regret to advise that we are unable to deal with any claims arising under the policy, which shall now be made void from inception, therefore no contract of insurance has been deemed to be existent since renewal 1996.”

Nothing was said, however, about the return of premiums paid or the cancellation of Dr Singh’s direct debit. The avoidance was not recorded on Provident’s computer record, which merely contains the following notation dated to 24 July 1996: “#41.45 back premium, and approx #148 RTA refund if we voided”, where # stands for £. The meaning of this notation, confirmed by another Provident document, was that in the light of the speeding conviction Dr Singh should have paid an additional £41.45 in premium, but that if Provident avoided the policy the premium refund should be £148. However, Provident continued to debit Dr Singh with monthly payments under his direct debit mandate and did not return any premiums to him (until much later).

On 19 August 1996 Dr Singh’s solicitors replied to Provident, taking an inaccurate point about the relative timing of the conviction and the renewal. Provident dealt with that point in its reply dated 3 September 1996, which also included the following:

“We are maintaining our decision to void the policy from inception due to the non-disclosure of material fact. Our Underwriting Department have been informed of the decision and require the current Certificate of Insurance in order to calculate the return.”

On 2 October 1996 Dr Singh’s solicitors put in issue whether his conviction would have altered the risk on renewal. They had obviously been investigating the way in which Provident’s rating system worked. Indeed, as their letter pointed out, Dr Singh had recently checked a quotation with Provident on the basis of his driving conviction and had been quoted the same premium as he had agreed on renewal. On 16 October Provident replied to confirm that “Had Dr Singh disclosed to us the speeding conviction, he would not have received insurance under the terms given”. Its letter stated that the decision to avoid the policy was maintained, but again failed to explain that it was the combination of the speeding conviction and the “fault” accident that would have increased Dr Singh’s premium.

On 5 February 1997 Dr Singh had a lengthy telephone conversation with a senior executive at Provident, Mr Robert Shaw, its technical claims manager. It subsequently emerged in Mr Shaw’s evidence at trial that Dr Singh’s claim had begun in Provident’s accidental damage claims department, apparently because that department had failed to observe that a serious personal injury was involved: in such a case Provident’s practice was to transfer the claim to the technical claims department. Thus at the time of Provident’s avoidance of 2 August 1996 Provident had not realised that the claim was potentially a very serious one, and it was then carrying a reserve of only [£6,000]. Thus Provident’s letter of 2 August 1996 and subsequent Provident correspondence had been signed by its accidental damage claims department. However, by February 1997 it had become more obvious that the claim was a serious one and it had been transferred to Mr Shaw’s department. On 6 February 1997 Mr Shaw wrote directly to Dr Singh to confirm their telephone conversation of the previous day. He now for the first time on Provident’s side explained (unless perhaps he had already done so the day before) that the avoidance had been triggered by the combination of the non-disclosure with the existing information regarding the “fault” accident. He said “Clearly therefore the non-disclosure was one of material fact which denied us the ability to collect the appropriate premium for the risk.” At trial Mr Shaw accepted that previous explanations by Provident of the reasoning for the avoidance had been inadequate. His letter also included a renewed request for the immediate return of Dr Singh’s certificate of insurance for cancellation “as previously requested”.

On 5 March 1997 Dr Singh responded in a lengthy letter sent directly to Provident’s chief executive. The first of a series of numbered points was that the accident disclosed in February 1995 was a no fault accident, although he did not at that time go on to explain that it had been settled entirely in his favour. On 10 March 1997 Provident’s deputy managing director replied, maintaining the essence of Mr Shaw’s letter of 6 February and saying that “the previously declared claim and the speeding conviction in combination cause the imposition of special terms”. On the same day an internal Provident memo raised a number of questions for action, among them whether the fault accident had subsequently been reversed by Dr Singh’s previous insurer so as to become no fault. I do not know what answer was ever given to this query, or whether it was ever progressed. It seems that it was not.

It is, however, now accepted by Provident and indeed was so accepted prior to trial that the January 1994 accident had been settled entirely in the Singhs’ favour and should be reclassified as a no fault accident. As stated above, that settlement had in fact occurred in July 1995, in other words prior to Dr Singh’s February 1996 renewal. Full details of that settlement were provided to Provident in June 1999, which was before Mr Beach’s claim had been agreed or these proceedings commenced. The facts of that settlement were “not denied” in Provident’s defence. Thus a “Schedule of Agreed Facts” was in due course drawn up by the parties which inter alia demonstrates Provident’s agreement that the January 1994 accident had by July 1995 become a no fault accident and that had Dr Singh notified Provident of its favourable outcome at the time of renewal of his policy his premium rating would have been unaffected by disclosure of his speeding conviction. At trial Mr Shaw agreed that if Provident had been satisfied at any time prior to avoidance on 2 August 1996 that the January 1994 accident was a no fault accident, then it would not have had grounds for repudiation.

A second issue, therefore, is whether the mutual duty of utmost good faith owed by an insurer to its insured operates in some way to limit the insurer’s right to avoid a policy where a fact critical to the insurer’s decision to avoid was not what the insurer thought it to be.

In the meantime, on 20 July 1996, after notification of the claim but before Provident had received his driving licence or avoided the policy, Dr Singh had replaced his Mercedes, which was a write-off, with a Volvo and on the same day obtained additional insurance from Provident under his existing policy for his new car. That was a few days before Provident learned from Dr Singh’s driving licence of the non-disclosure of his speeding conviction. A new certificate of insurance was subsequently sent to Dr Singh certifying insurance of the Volvo and that “the policy to which this certificate relates satisfies the requirements of the relevant law”. The certificate was part of a single sheet which also recorded the “Policy Amendment Schedule”. The policy number 00DC2195188 is the same as that of the original policy, save for the second digit (now 0 for A): although this difference is not explained, it is not suggested that this was a different policy. The effective date of the amendment was given as 20 July 1996; the expiry date was given as noon 17 February 1997, just like the original policy. Cover was again comprehensive. Provident’s computer record, under the number DC 2195188, does not refer to this policy amendment, nor to the one in May 1996 when the Mercedes replaced the Renault. The certificate for the Volvo cannot be precisely dated, save that it is on a form which itself demonstrates that it did not come into use until 1 September 1996, viz after the avoidance. It would seem that this fact, which is disclosed by small print on the form, was not brought to the attention of Moore-Bick J, who assumed that the certificate was contemporaneous with the agreed amendment.

On 2 November 1996 Provident’s direct debit pay plan unit wrote to Dr Singh under policy reference number 2195188 to thank him for his “recent instructions to amend your policy cover”. It is consistent with that letter that the certificate and amended schedule dealing with the Volvo had not yet been sent to Dr Singh. The letter listed all the direct debit payments which Dr Singh had made during 1996. These included monthly payments of £9.96 up to June 1996 (I would infer that these represented the premiums in respect of the Renault), then an additional premium of £110.50 in June 1996 (which I would infer was in respect of the change from Renault to Mercedes and the further change to comprehensive cover), then from July 1996 an enhanced monthly debit of £28.37, then in October 1996 a further additional premium of £93.59 (which I would infer was in respect of the addition of the Volvo). The August payment was made on 17 August. The judge found that payments continued to be made down to December 1996 on the basis of another, internal, Provident document: but it is extremely difficult to square that document (whose date is uncertain but cannot be prior to December 1997 and even so does not evidence the return of premium which it is common ground occurred in that month) with Provident’s letter to Dr Singh of November 1996. As far as I can see, direct debit payments may only have ceased with the expiry date of the policy in February 1997. It was only in December 1997 that there was any return of premium.

It would seem moreover that despite the series of letters to Dr Singh maintaining its avoidance of the policy, Provident did have some doubts about the correct response. Thus, on 22 January 1997 the technical claims department wrote internally to “Customer Service Group A” to say that “Please note we are currently in dispute over indemnity with both our policyholder and the Drake…we would ask that at this stage no further action is taken in connection with the policy without liaising with this department”. On 29 January 1997 a further note appeared on Provident’s computer record of Dr Singh’s policy, viz: “Do not process any amendments on this policy without prior referral to Tech Claims”. This note appears immediately after the note of 24 July 1996 cited above (see para 13) which said “if we voided”. At this time there was also communication between Provident and Mr Beach’s claim handlers in which the possibility of a review leading to a reversal of Provident’s decision to decline liability was discussed, even if such a reversal was thought to be unlikely “from evidence before us”. It was precisely at this time that the telephone conversation directly between Dr Singh and Mr Shaw was being planned, which led to Dr Singh being informed of the relevance of the “fault” accident to Provident’s decision to avoid and to Dr Singh’s subsequent response that the accident had been a no fault accident. There is at no stage any reference on the computer record to the express effect that the policy had been avoided. The final note, however, dated 10 March 1997 (and therefore after the expiry date of the policy) said: “Should bkr request RP [return premium], allow credit as per above memo”, sc the £148 RTA refund referred to under the note of 24 July 1996.

In cross-examination at trial Mr Shaw gave evidence that as of January 1997 the decision within Provident had been to keep Dr Singh’s policy running, since there was a dispute as to the validity of the avoidance. I will set out the passage in full:

“Q. So, you had taken the decision to continue this policy at this particular point in time?

A. Yes.

Q. So the repudiation was no longer effective at this particular point in time?

A. The repudiation was being discussed in correspondence at that time.

Q. Yes, but you had not actually rescinded the policy at this particular point of time?

A. No.

Mr Justice Moore-Bick: That was your understanding, was it?

A. Yes.

Q. The policy was still running?

A. Yes.

Counsel: And they were the instructions, you actually tell the underwriters to hold fire and not to cancel it for the time being?

A. Yes…

Mr Justice Moore-Bick: I am sorry, you will understand that I am finding this a bit puzzling. Letters were written in August 1996 to the insured telling him that the policy was void?

A. Yes…

Q. Is there any doubt about the position as you would understand it from the insured’s point of view, receiving your letter?

A. No.

Q. And did you know about this correspondence by January 1997?

A. Yes, we had the full file with us.

Q. It was your genuine understanding that you had not avoided the cover?

A. We hadn’t cancelled the policy at that point.

Q. I do not want there to be any misunderstanding. What do you mean by “cancelled the policy”?

A. Formally closed the policy down and given a returned premium, voided from inception.

Q. So you were administering this as if it had not been avoided?

A. Yes.

Q. Even though you knew that you had written to the insured telling him that it had been avoided?

A. Yes, but we hadn’t got the certificate back which we’d asked for; he still had that.

Q. I see. Thank you.

Counsel: So, the position is that you were deliberately keeping the policy running and you were continuing to seek the premiums and you were giving instructions, well, continuing instructions not to cancel the policy for the time being?

A.

Yes.”

Thus the third main issue is whether there had been a waiver by Provident of any right of avoidance, or a reinstatement of the policy.

On 25 March 1997 Dr Singh referred his claim against Provident to arbitration, as he had been threatening to do, under the inexpensive PIAS scheme (Personal Insurance Arbitration Service) run by the Chartered Institute of Arbitrators. He acted for himself without legal advice. His letter of claim dated 27 April 1997 did little more than submit the correspondence referred to above. It referred to the discrepancy between the attitude taken in the correspondence and the continued payments under the policy. Although it referred to Mr Shaw’s letter of 6 February 1997 and commented that “[it] brought into play other factors”, it did not focus on or refer expressly to the point about the no fault accident. The letter did, however, point out that in the meantime Drake had indicated that it would indemnify Mrs Kaur under her separate policy (albeit only for her third party liability, leaving him with the £2,000 loss of his car). It was presumably because of the limited exposure he and his wife faced in the light of Drake’s support (as to which see below) that he was prepared to approach his arbitration with Provident in the way he did.

Provident’s letter of defence dated 19 May 1997, signed by Mr Shaw, was an altogether more expert piece of advocacy, but it too did not refer to the question whether the fault accident was truly a no fault accident or to Dr Singh’s assertion that it was. However, the point did emerge in Dr Singh’s letter of reply dated 7 June 1997, where almost his first point (para 2) was that the January 1994 accident was a no fault accident “which the third party paid for”. He also observed (in his para 8) by reference to Provident’s underwriting criteria which were now in evidence that his conviction therefore counted for only 10 points and that the normal premium would have applied. He referred back to para 8 in a number of subsequent passages in his letter.

In a final award dated 20 June 1997 the arbitrator held that Provident had been entitled to avoid the policy on the grounds of non-disclosure of the conviction. The arbitrator did not deal at all with the significance, if any, of whether the previous accident was a fault or no fault accident. It is common ground that the award is binding between Dr Singh and Provident but not between Drake and Provident, and the matter was argued on this basis both before the judge and in this court.

Nevertheless, the arguments addressed on this appeal demonstrate that the significance of the award as it bears on Drake’s claim for a contribution in equity from Provident is obscure. Drake submits that because the award is not binding as between it and Provident, therefore it cannot be relied on by Provident to exclude the existence of double insurance and thus the doctrine of the equitable right to contribution between insurers of the same loss. Drake therefore submits that if it can show that Provident was never entitled to or waived the right to avoid its own policy, then Drake is entitled to contribution from it. Provident on the other hand submits that if the award can be ignored so that its policy in favour of Dr Singh is still to be regarded as in existence at the time of the motor-cycle accident, then Drake, under a term of its own policy, was only liable for a rateable proportion of the claim: and in as much as it indemnified Mrs Kaur for the whole claim, it was a volunteer. Drake, however, ripostes that to meet this defence that it acted as a volunteer it can rely on the award as demonstrating that no separate policy existed as between Provident and Dr Singh.

I must therefore say something further about Drake’s policy and the circumstances in which Provident raised the argument that Drake was a volunteer and thus was not entitled in equity to any contribution. As stated at the outset, Drake was Mrs Kaur’s insurer under a separate motor policy which insured her own car and extended cover to her while driving any other car with its owner’s consent (the Driving Other Vehicles or DOV extension). It therefore covered her for driving her husband’s Mercedes, but did not of course cover damage to the Mercedes.

Drake’s policy contained a “rateable proportion clause” as its general condition 3:

“If at any time any claim arises under this Policy there is any other existing insurance covering the same loss damage or liability the Company shall not be liable to pay or contribute more than its rateable proportion of such claim.”

The intended effect of the clause is to seek to supersede the doctrine of equitable contribution by limiting the primary liability of each insurer to the rateable proportion which the doctrine would otherwise achieve. Provident’s policy contained a clause in very similar terms.

Drake became involved as a result of Provident’s avoidance of its policy. Thus on about 6 September 1996 Mrs Kaur spoke to Drake on the telephone and sent it an accident report form which asked Drake to deal with any third party claims arising. In the telephone call she told Drake of the difficulties with her husband’s insurance with Provident. Drake wrote straightway on 9 September to Mr Beach advising him that it was Mrs Kaur’s insurer. On 13 September Mrs Kaur sent Drake copies of the correspondence with Provident. On 23 September Drake confirmed to Mrs Kaur that it would indemnify her under her policy if Provident did not change its mind. On 31 October 1996 Dr Singh’s solicitors informed Drake that Provident had not been persuaded to do so. Drake was interested to know whether Dr Singh would take the matter further, for instance by writing to Provident’s chief executive or involving the Insurance Ombudsman. It was of course in Drake’s interest for Provident to accept responsibility for Mr Beach’s claim. An internal Drake memo dated 31 December 1996 reveals that Drake was considering its own position in relation to Mrs Kaur’s insurance. It states that the SP30 conviction should have been disclosed to it as well on renewal in March 1996. (It is not clear to me why that is so, but it may be because Dr Singh was a named driver on Mrs Kaur’s policy.) The memo goes on to say that

“if the Provident are not to be persuaded to deal we will void our policy – this would create a ludicrous situation which would not reflect at all well upon us or the Provident or indeed the insurance industry as a whole. However, it would bring the Provident back into the equation.”

The logic behind that reasoning was, I think, that if because of the avoidance of both policies Mrs Kaur became uninsured, then the third party would be protected by the Motor Insurers Bureau agreement and by the arrangements made within the industry to the effect that total liability would fall upon Provident as the insurer which had issued a certificate of insurance in respect of the Mercedes. At that stage, therefore, Drake was thinking in tactical terms which would either force Provident to reconsider its position or else lead to Provident being left with the liability in any event, but at the same time realising that in strategic terms such behaviour on its part would bring the industry into disrepute. In the event Drake ultimately decided to act strategically rather than tactically and did not avoid its policy with Mrs Kaur.

On 2 July 1997 Provident wrote to Drake to inform it that the decision to avoid Dr Singh’s policy had been upheld in arbitration. Copies of the parties’ statements and of the final award were enclosed. An internal Drake memo dated 2 September 1997 records that “our underwriters will not support a decision to void our policy due to non-disclosure”.

On 15 September 1997 Mrs Kaur was convicted of careless driving and fined £400. Her licence was endorsed 6 points.

On 22 September 1997 Drake wrote to Mr Beach’s solicitors concerning Provident’s stance in the following terms:

“We neither agree with, nor accept, this decision, and it is our intention to further pursue the matter as we are of the opinion that the Provident’s actions are in breach of the market agreement, and it is in fact they who should be giving consideration to your client’s claim.

However, we do of course accept that we are providing cover to our Insured…and as we would not wish the unresolved dispute between ourselves and the Provident, to prejudice your client’s position, we confirm that until such time as our dispute has been resolved, we will be giving consideration to your client’s claims.”

A similar protest was sent by Drake directly to Provident, by letter dated 24 September 1997 which was in response to Provident’s letter of 2 July. Drake said:

“However, upon the basis of the information currently available to us, we must advise you that we do not accept or agree with your decision, and it is our intention to pursue the matter further.

Therefore, we are not prepared to provide you with the requested confirmation that we will deal with all the claims arising out of this accident, but in recognition of our involvement under the Driving other vehicles extension of our Policy, and in order to avoid any further adverse publicity to the industry, we have notified the Claimant’s Solicitors that until such time as the dispute between us is resolved, we will give Without Prejudice consideration to their client’s claim.”

It was therefore plain to Provident that Drake was protesting the former’s decision, that there was an unresolved dispute between the two companies and that in the meantime Provident would deal with Mr Beach’s claim on a without prejudice basis.

On 11 March 1998 Drake wrote to the Association of British Insurers (ABI). Its letter set out the situation and concluded:

“We did not wish Mrs Kaur to be in a position where two Insurers were refusing to provide her cover for legal representation against a serious charge which she was subsequently found guilty of. We are surprised and concerned at Provident’s handling of this matter, and perhaps you would be kind enough to look into this case for us, and provide us with your own opinions.”

The ABI replied on 17 March 1998 to the effect that the dispute appeared to be covered by the Dual Indemnity Undertaking. This is an undertaking to which members of the ABI Motor Conference are members. It is a market mechanism which would enable an insurer in Drake’s position to call upon an insurer in Provident’s position as direct insurer of the vehicle involved in the accident to bear the full amount of the claim. However, it presupposes that Provident has not validly avoided its policy and in any event the Notes to the Undertaking expressly recognise that it has no legal significance. Drake accepted at the trial below that it cannot give rise to an enforceable obligation at law.

On 27 August 1998 Drake wrote again to Provident to repeat that “While the decision you made to void your policy is noted, for reasons already expressed, we do not accept that and it is our intention to take this matter further.”

On 3 June 1999 Drake wrote directly to Mr Shaw at Provident. This letter clearly set out the argument that as the January 1994 accident had not been one of fault and a full recovery had been made, the non-disclosure of the speeding conviction could not have affected Dr Singh’s insurance on renewal. The letter enclosed relevant documentary material regarding the Singhs’ recovery. It continued:

“I felt it was only right to give you the opportunity to review the situation in light of those comments. I would ask you to confirm you will now take over the handling of the claim, failing which the matter will have to be referred to the ABI and/or for further litigation to be considered.”

The letter also pointed out that Mr Beach had now issued proceedings. On 8 June 1999 Mr Shaw replied, maintaining Provident’s position, stating that “Underwriters can only base their judgements on known or disclosed information”, and relying on the arbitration award.

On 21 September 1999 Drake’s chief executive wrote to Provident’s deputy managing director. His letter said “there clearly is a dispute to resolve” and he suggested confidential mediation as a process preferable to litigation. Provident’s reply was briefly to the effect that the arbitration award and Mr Shaw’s letter of 8 June 1999 had said it all. Mediation was refused, litigation was said to be a matter for Drake.

There were further discussions between Drake and Provident, but they did not come to anything. On 15 March 2000 Provident’s solicitors said that they were instructed to accept service of proceedings. By a consent order dated 24 May 2000 Mr Beach’s action was concluded by a settlement in the sum of £1.2 million plus costs. On 13 June 2000 Drake’s solicitors wrote to inform Provident of that compromise and to enclose a draft Part 20 claim and skeleton argument to explain the way in which Drake put its claim against Provident. On 29 August 2000 Provident’s solicitors replied to say that Mrs Kaur was bound by the arbitration award and therefore had no claim against Provident.

These proceedings were commenced on 10 October 2001. In its particulars of claim Drake pleaded that the settlement with Mr Beach had been properly and reasonably made. In its re-amended defence Provident did not deny that. It also pleaded that the arbitration award was binding as between Mrs Kaur, Dr Singh and Provident, but that if Provident was liable to Dr Singh, then Drake’s liability under its policy was limited by the rateable proportion clause, as was Provident’s under a similar clause in its policy.

The fourth issue is whether Provident is correct to say that Drake’s claim against it fails on the ground that Drake was a volunteer in as much as it indemnified Mrs Kaur for more than 50% of her liability to Mr Beach despite the existence in her policy of the rateable proportion clause. One aspect of that issue relates to the role of the arbitration award: is Drake entitled to say that Mrs Kaur could successfully raise against it the fact that Dr Singh had lost the arbitration?

The judgment of Moore-Bick J is reported at [2003] 1 All ER (Comm) 759. Although he made a number of findings of fact in favour of Drake, and more than once expressed some disquiet about Provident’s behaviour, his ultimate conclusions of law were in every case made in Provident’s favour.

Thus, as to the facts he said this at para 23 of his judgment:

“Mr Shaw accepted that if Provident had been told that Dr Singh and Mrs Kaur had received full compensation from the third party it would have reclassified the accident as a “no fault” accident. It would then have ceased to have any significance when it came to setting the premium on any subsequent renewal. He also accepted that if Dr Singh had given the same information to Provident shortly after its decision to avoid the policy it would probably have reversed its stance and accepted his claim. By the spring of 1997, however, things had moved on. It had become apparent that Mr Beach had sustained serious injuries and that his claim was likely to be substantial. Provident also had reason to think that if it maintained its refusal to pay, Drake would probably deal with the claim. Mr Shaw, who had by then assumed responsibility for handling this claim, confirmed that he had decided quite deliberately not to respond to the assertion made by Dr Singh in his letter of 5^th March 1997 that the accident in January 1994 had been a “no fault” accident because he did not think that of itself that amounted to sufficient proof. In his view it was up to Dr Singh at that stage to make out his case. I can only infer that Mr Shaw chose not to pursue the matter because he did not want to uncover information that might have made it commercially difficult for Provident to maintain its repudiation of the claim.”

On what I have described (at para 10 above) as the first issue, namely whether Provident was entitled to avoid the policy, the judge said this:

“17.

The leading case on the right to avoid a contract of insurance for non-disclosure is that of the House of Lords in Pan Atlantic Insurance Co Ltd v Pine Top Insurance Co Ltd [1994] 3 All ER 581, [1995] 1 AC 501. For an insurer to be entitled to avoid a policy on these grounds he must be able to show both that the fact which the insured failed to disclose was material and that the failure to disclose it induced him to accept the risk when he would have otherwise declined it or would have accepted it only on different terms. These are closely related, but distinct requirements.

“18.

In the present case it was conceded that the existence of a speeding conviction was a material fact. It follows, therefore, that it should have been disclosed, but it does not inevitably follow that Provident was entitled to avoid the policy because the conviction would not on its own have led to the imposition of an increased premium. The critical question is whether Dr Singh’s failure to disclose his conviction did in fact induce Provident to renew on more favourable terms than it would otherwise have been willing to agree…

“24.

[Counsel for Drake] submitted that an insurer is only entitled to avoid a policy if at the time he purports to do so there exist facts that justify his taking that step. The accident in January 1994 was in fact a “no fault” accident and therefore disclosure of Dr Singh’s conviction would not have led Provident to demand an increased premium. Accordingly, there was no basis for avoiding the policy.

“25.

This is an attractive argument, but it fails to distinguish between the nature of the previous accident and Provident’s state of mind at the time of renewal…No one had told Provident that the claim had been settled in full by the third party…In that context a speeding conviction would inevitably have led to an increase of premium and so, subject to one point to which I shall come in a moment, Dr Singh’s failure to disclose it did induce Provident to renew the policy on more favourable terms than would otherwise have been the case…

“27.

Having said that, I think one may legitimately ask what would have happened in this case if Dr Singh had disclosed his conviction at the time of renewal. Clearly Provident would have demanded a higher premium and, if Dr Singh had questioned it, the issue of the correct classification of the earlier accident might have surfaced. In that event I think he would probably have been able to persuade Provident that it should be treated as a ‘no fault’ accident and that he should therefore be charged the ‘normal’ premium. It might therefore be argued that his failure to disclose the conviction did not cause Provident to accept the risk on terms different from those that it would otherwise have been prepared to agree because the disclosure would of itself have led to the correction of his claims record and the imposition of the same terms.

“28.

The difficulty with this argument, however, is that one simply does not know what Dr Singh’s reaction to an increase in premium would have been. He might have questioned it, he might not. If he had questioned it, and if he had simply been told that it was due to his recent speeding conviction, he might have left it at that, or he might not. There is nothing in the evidence that points either way. Only if it could be shown that disclosure of the conviction would also have led to the disclosure of the information relating to the settlement of the earlier loss and to Provident’s accepting the risk on the same terms could one say that the non-disclosure had not caused it to accept on terms that it would not otherwise have agreed to. In the present case the evidence does not enable me to reach that conclusion.

“29.

On the face of it, therefore, Provident was entitled to avoid the policy…”

I would analyse the judge’s reasoning on this issue as follows. Only two questions arise – materiality of the non-disclosure and inducement. Those questions have to be looked at as at the time of renewal. It does not matter what the true facts at the time of renewal were, only what they appeared to Provident to be (“Provident’s state of mind”). It was admitted that the non-disclosure was material. The only remaining question was whether the non-disclosure induced Provident to contract on normal premium terms. It would not have done so if the disclosure of the conviction had led to the reclassification of the January 1994 accident. On that question the burden of proof lay on Dr Singh. Since there was nothing in the evidence that pointed either way, Dr Singh failed to meet that burden. It was therefore to be presumed that the non-disclosure did induce the contract, for that would have been the normal working of Provident’s underwriting criteria.

On what I have described as the second issue (see para 19 above), namely whether Provident’s right to avoid the policy was limited by the doctrine of good faith, the judge found and reasoned as follows:

“11.

When in August 1996 Provident gave notice to Dr Singh avoiding the policy it did so in perfectly good faith. It was still under the impression (as it had been in February) that the accident in January 1994 was, as had been stated in the original proposal form, a ‘fault’ accident and it knew that if the speeding conviction had been disclosed it would inevitably have led to the imposition of an increased premium…

“29…but [counsel for Drake] submitted that it could not [avoid the policy] in so far as its original decision to avoid or its refusal to investigate the true position and then reverse that decision involved a breach of its duty of good faith…

“30.

Although I too feel some unease at the prospect of an insurer’s avoiding the contract for non-disclosure in circumstances such as the present, I do not think that the solution is to be found in the exercise of the court’s equitable jurisdiction. Whatever the true origin of the insurer’s right to avoid for non-disclosure, it is in the nature of a right to rescind the contract and its exercise is subject to the same principles as apply to the exercise of the right to rescind generally…

“32…If grounds exist to justify avoidance, the insurer’s decision, once communicated to the insured, is effective immediately. The insurer does not need to invoke the assistance of the court, nor does the court have jurisdiction to declare that his right to avoid has been lost retrospectively by reason of subsequent events. This is quite distinct from the question whether the right to avoid has arisen in the first place. If it has, and if it has been exercised in good faith on sufficient grounds, I do not think that the insurer is precluded from maintaining his position by the subsequent receipt of further information or that he can then be required to reinstate the contract. To enable the insured at trial to defeat the insurer’s right to avoid, after it has been exercised, by establishing facts of which he was not aware when he wrote the risk would alter the whole basis of the underwriting exercise and introduce an additional and unwelcome element of uncertainty. For these reasons I am unable to accept that proof at trial of facts showing that the earlier accident should have been treated as a ‘no fault’ accident at the time of renewal can prevent Provident from relying on its avoidance of the policy.”

I would again seek respectfully to analyse the judge’s reasoning as follows. The exercise of an insurer’s right to avoid may indeed have to be made “in good faith on sufficient grounds”, but once again the review of that issue must be carried out as at the time of contract and only by reference to the facts known to the insurer at the time of contract. It is not open to attack the insurer’s good faith by reference to facts such as may have become known after the time of contract, let alone by reference to facts only proved at trial. Once an insurance contract has been properly avoided, there is no duty to reconsider or to reinstate the contract by reason of knowledge only acquired after avoidance. In any event an exercise of avoidance does not depend on the court’s assistance or declaration. In all these matters, the insurer’s right to avoid is like any other case of rescission of a contract.

The third issue raised (see at para 24 above) is whether Provident waived its right to avoid, or reinstated, Dr Singh’s policy. The judge found that Provident did neither. He said that avoidance is a matter of election, that election requires nothing more than the unequivocal communication by the insurer to the insured of his choice and that once made that choice becomes irrevocable. He held that Provident’s letter of 2 August 1996 was such an election and rescinded the contract with immediate effect. That left only the possibility of reinstatement. The judge considered the history of events following that letter and concluded as follows:

“40.

Despite the fact that Provident amended the policy terms to cover the new car and continued to collect premium from Dr Singh between July and December 1996, I am unable to accept that it entered into an agreement with Dr Singh to reinstate the original cover or to handle the accident claim as if it had not avoided the policy. The delivery of a certificate of insurance in relation to the new vehicle certainly evidenced an agreement to insure Dr Singh in relation to that vehicle. But that occurred before Provident received notification of the claim and before it became aware of its right to avoid the policy. In all its subsequent communications with Dr Singh and his solicitors Provident maintained its avoidance of the policy and no one seems to have been in any doubt as to its position. Certainly Dr Singh does not seem to have thought that Provident agreed to treat the original cover as reinstated or that it is precluded by estoppel or otherwise from relying on its original decision to avoid the policy.”

I am satisfied that there are certain errors of fact in that account. The insurance of the Volvo was made after notification of claim, and in particular the certificate of insurance in relation to the Volvo did not precede Provident’s learning of the non-disclosure but post-dated it. Moreover, it is not possible on the evidence to say that “no one seems to have been in any doubt about [Provident’s] position”. It is plain that Mr Shaw appears to have considered that in January 1997 the policy had not been finally avoided: at the very least he appears to have been in two minds about it. Dr Singh gave evidence that direct debit payments were taken from his account monthly down to 19 December 1996. The question remains whether the judge erred in finding no waiver or reinstatement.

The fourth issue relates to Provident’s allegation of voluntary payment on the part of Drake (see para 47 above). The judge held that Drake was bound by a decision of this court in Legal & General Assurance Society Ltd v. Drake Insurance Co Ltd [1992] QB 887 to the effect that it could not claim a contribution from Provident for moneys it had voluntarily paid in excess of its liability to Mrs Kaur, and that all efforts to distinguish that authority on the grounds, for instance, of Drake’s express reservation of its right to claim against Provident, failed.

On analysis this issue cannot be completely answered without taking notice of the question of good faith raised under the second issue: for if as a matter of good faith Provident was not entitled to avoid its policy, then that is an answer to this first issue. No authority, however, so far as is known, has ever decided that the insurer’s mutual obligation of good faith limits its rights of avoidance. The role of this first issue therefore is to ask whether, irrespective of any argument that can be raised as a matter of good faith, Provident was entitled to avoid the policy. If the answer is No, then the second issue does not matter. If the answer is Yes, then the second issue must be faced.

On behalf of Drake, Mr Roger Ter Haar QC submitted that there were two reasons (other than its obligation of good faith) why Provident was not entitled to avoid Dr Singh’s policy. The first was that Provident had to contrast the disclosure which did in fact take place at the time of renewal with the disclosure which ought to have taken place, and that involved taking account not only of the speeding conviction but also of the true facts about the January 1994 accident as a no fault accident. The second was that on the facts the non-disclosure of the speeding conviction did not induce Provident to accept the renewal at normal rather than increased rates because if the speeding conviction had been disclosed the probability was that the truth about the no fault accident would also have emerged. The judge was wrong to say that there was no evidence either way on that subject and in any event he applied the wrong burden of proof.

On behalf of Provident on the other hand Mr Robert Moxon-Browne QC submitted, as to the first point, that since there was no duty to correct the information about the earlier accident (see section 18(3)(a) of the Marine Insurance Act 1906 (“MIA”), which states that in the absence of inquiry “Any circumstance which diminishes the risk” need not be disclosed), the comparison was not with the true facts relating to the earlier accident but with the facts as Dr Singh had presented them to Provident. As to the second point he submitted that the judge had come to the right answer on the question of inducement, since it was simply uncertain how matters would have turned out on the hypothesis that the conviction had been disclosed. The broker had given no evidence, and Dr Singh had given no oral evidence.

I propose to deal with the second point first, because it is the more fact specific, whereas the proper analysis of the first point is elusive. In my judgment, the judge was in error in concluding that there was no evidence to suggest that the no fault status of the prior accident would have emerged. Until 7 February 1997 Provident had said nothing to suggest that the disclosure of the fault accident had anything to do with its claim to avoid the policy. When, however, the connection with the status of that accident was first made in Mr Shaw’s letter of that date, Dr Singh’s next communication with Provident, his letter to the chief executive dated 5 March 1997, straightway pointed out that the accident had been a no fault one. It was the point he made in his first numbered paragraph, after the words:

“Mr Shaw did not mention in our telephone conversation any of the other points raised in his letter dated 6^th February 1997, if he had I certainly would have corrected his misunderstandings…”

Dr Singh kept that point well in view, for instance in his letter in reply to the arbitrator. It was a great pity that in his award the arbitrator paid no attention whatsoever to this factor.

It was not in dispute that the accident was a no fault accident and was entitled to be reclassified as such by the time of the renewal. All that Provident can say is that its new status was not actually mentioned at that time. However, the question for present purposes is not what actually happened, but what would have happened if the speeding conviction had been declared. This is because Pan Atlantic Insurance Co Ltd v. Pine Top Insurance Co Ltd [1995] AC 501 says that an insurer who seeks to avoid for non-disclosure must show that he had actually been induced by the non-disclosure to enter into the policy on the relevant terms. Provident cannot do that without showing that if the conviction had been declared, it would have charged a higher premium. But it is common ground that it would not have charged a higher premium if the earlier accident had been a no fault accident. The judge found as a fact that if the issue of the correct classification of that accident had become a matter for discussion at the time of renewal, Dr Singh “would probably have been able to persuade Provident that it should be treated as a “no fault” accident” (at para 27 of his judgment). So the question resolves itself into this: if the conviction had been mentioned, would the question of the status of the accident have been discussed? It seems to me to be very likely that it would have been, first because it would have been Hexagon’s duty as Dr Singh’s broker to have raised the issue, and secondly because when the significance of the accident’s status was raised in correspondence Dr Singh addressed it, and kept on doing so. The judge mentioned neither factor in concluding that there was “nothing in the evidence that points either way”. As for there having been no evidence from the broker, there was nothing to prevent Provident from calling the broker: and although Hexagon was Dr Singh’s broker, it was also to some extent Provident’s agent, since it had authority to rate each proposal and to determine the premium on the basis of Provident’s underwriting criteria. As for Dr Singh giving no live evidence, this was because both parties were content to have his evidence in the form of his witness statement: Provident did not wish to cross-examine him. His witness statement lays stress inter alia on the fact that it was not until Mr Shaw’s letter of 6 February 1997 that he was made aware that Provident had recorded the accident as a “fault” accident and that in his own letter of 5 March 1997 he had informed Provident that it was not.

In my judgment these facts support the conclusion that even if the burden had been on Dr Singh to prove that his non-disclosure had not induced Provident to contract on the terms that it did rather on more advantageous terms, he could have met that burden. Mr Moxon-Browne accepted in the course of submission that, if the conviction had been disclosed, there would have been a discussion of its impact on premium in the light of the status of the earlier accident. If so, then the judge’s finding that such a discussion would have led to the premium remaining at the normal level is fatal to this part of his case. When this was pointed out to him, Mr Moxon-Browne accepted the court’s offer to him to withdraw his concession, if he so wished. He did: but in my judgment his one-time concession was rightly made on the facts of the case.

As it is, in my judgment it was not Dr Singh or Drake who bore the legal burden of proving that the non-disclosure of the speeding conviction induced the contract. That burden rested on Provident. But by stressing that there was nothing in the evidence “that points either way”, the judge was effectively placing the burden on Drake. If the case had simply been about an undisclosed conviction which in itself would have caused an increase in the premium, then the evidential burden might have shifted with or without the help of the presumption in Redgrave v. Hurd (1881) 20 Ch D 1. The issue, however, was more complex than that, and I am unpersuaded that Provident has managed to discharge the burden that rests on it. In any event, I have concluded above that Drake has proved on the particular facts of this case that the truth about the accident would have emerged, if the speeding conviction had been disclosed.

That is enough to invalidate Provident’s avoidance of the policy and I could proceed directly to the fourth issue of whether Drake’s payment was voluntary. However, because of the importance of the points raised, I will state my opinion, but, with the exception of the third issue, not my decision on them.

As for the first point made under the present issue, it seems to me that ultimately the question is whether the insurer’s right to avoid must be judged according to the information provided to the insurer at the time of contract (or what the judge called Provident’s “state of mind”) or according to the true facts as at that time. I would not have thought that the insurer’s state of mind was important, other perhaps than in the context of a plea of estoppel against the insured, and even such a plea would probably have to be judged objectively rather than subjectively. Provident did plead estoppel against Drake in these proceedings, but it does not seem to have been pursued, perhaps because it did not need to be: at any rate no estoppel has been relied on by Provident on appeal.

If consideration moves from Provident’s state of mind (which would in any event perhaps be a difficult subject to enquire into) to the information provided to the insurer at the time of contract: that was contained on the original proposal form. That information was already a year old at the time of renewal. It suggested an accident which in truth was not Dr Singh’s or rather Mrs Kaur’s fault, but which under Provident’s underwriting criteria had to be classified as a fault accident pending clarification. The description of it as a “fault” accident is in truth a description chosen by Provident (indeed, imposed by Provident) rather than provided by Dr Singh. I cannot accept that the “fault” description was proffered by Dr Singh, even though he signed the proposal form and thus vouched for the truth of the information provided by him.

As at the time of renewal, Dr Singh did not provide any new information regarding the accident, although he could of course have done so, for the accident had already been settled entirely in his and his wife’s favour. There was, however, no duty on the part of Dr Singh to provide this information, however much it may have been in his interest to do so, because section 18(3)(a) of the MIA states that in the absence of inquiry any circumstance which diminishes the risk need not be disclosed. Provident did not ask at the time of renewal for up to date information regarding the disclosed accident. In the absence of authority to the contrary, and none has been cited, I would not regard a renewal as incorporating automatically and implicitly all questions asked in the proposal form, nor a request for updates on all answers given in the proposal form. The renewal form asked specifically for notification of any driving convictions and referred generally to “your duty to disclose to the Company the existence of any factor which may affect the risk”: but again there would seem to me to be no duty to disclose a circumstance which diminishes the risk. The existing policy contained a standard term, headed “Material Alteration” which provided: “The Insured shall give prior notice of any material alteration in risk which may occur during the currency of this policy”. The court was referred to this term by Mr Ter Haar, but I would not be inclined to construe it as requiring disclosure of circumstances which diminish the risk: and the need for “prior notice” suggests a general problem about its ambit which need not be considered in this case.

I would return, therefore, to the question whether an insurer is entitled to avoid on the basis of the information which he has at the time of contract or on the basis of the true state of affairs at that time. This is a large question, and one on which no cases have been cited to us and on which the judge cited no authority. I would therefore wish to be cautious on this subject, and would emphasise that what I say is no more than opinion. In principle, however, I would have thought that a party which seeks to terminate a contract, whether under an express clause, or whether by way of rescission for misrepresentation or avoidance in insurance for non-disclosure, or whether for breach of condition, or by way of acceptance of a repudiation, must make good his ground for bringing the contract to an end. That does not mean to say that he is limited to the ground which he advances at the time of termination, for it has often been said that if a party has a good ground for termination, then it does not matter whether he gives it or any ground at all at the time of termination. That is of course subject to termination under an express power or option in a contract, in which case the clause may well define what has to be said or done for the valid invocation of the clause, and also subject to any case of estoppel. Generally speaking, however, subject to the rules of pleading and the just case management of proceedings, a party may justify a termination of a contract, if he can, by any means which do in fact justify it. In terminating, however, he takes the risk that he can justify bringing the contract to an end. That does not mean that the termination is only effective if sanctioned by the court. If the termination is justified at the time it occurs, then it operates validly and effectively from that time – however long a dispute about it may take to resolve. But if there is a dispute, and if at trial the party which purported to terminate cannot justify his termination, and if the court finds that the contract was not validly terminated but repudiated by the party which claimed to be entitled to terminate, then the consequences of what has latterly been shown to be an invalid termination lie at the risk of the terminating party.

If this is so generally, why should the insurer’s right to avoid a contract of insurance not similarly depend on the true facts of the case? Since the avoidance is a form of rescission from inception, the critical facts will be those in existence at the time of contract. But it is not impossible that other facts may have occurred between contract and time of avoidance which for some reason prevent rescission. For the purposes of the present question, however, I will limit myself to the time of contract. At that time, in this case, the true facts were, as has been common ground since even before trial, that Mrs Kaur’s January 1994 accident was a no fault accident. Why should it be treated as a fault accident? It is true that at the time of contract Provident did not know that it was a no fault accident: but it is often the case that contracts are terminated in circumstances where the true facts are obscure or mistaken, and it emerges that the termination is or is not justified contrary to the understanding or knowledge of one or other party.

Is there then something special about insurance? Insurance is about risk, which is difficult to evaluate. The evaluation of risk may change over time, as different information becomes available. The fact that later information may diminish or even increase the risk does not invalidate the disclosure or the assessment that had to be made at the time of contract. A good example is to be found in a case which is almost contemporaneous with these proceedings: Brotherton v. Aseguradora Colseguros SA [2003] EWCA Civ 705 (22 May 2003). In that case insurers gave banker’s bond and professional negligence indemnity cover to a Colombian bank and obtained reinsurance. At the time of the reinsurance contract the insurers failed to disclose that allegations of serious impropriety had been made in the Colombian media against the president and other senior officials of the bank. The reinsurers claimed to be entitled to avoid the reinsurance for non-disclosure. The insurers submitted that they were entitled to prove at trial that the allegations (and the arrest of and charges against the president of the bank) were entirely without foundation. The reinsurers sought an order barring the insurers from seeking to rely at trial on evidence not available at the time of contract and submitted that the materiality of any circumstances could be judged only by reference to the situation at the time of risk. This court held that proof at trial that the allegations were unfounded based on information that only came to light after the contract was made could not affect either the materiality of the non-disclosure or (which is relevant to an issue discussed below) the good faith of the avoidance. The only proper questions for decision at trial were whether there had been non-disclosure which was material and which had induced the reinsurance.

It is important, however, to underline the factual difference between that case and this, a difference which is also reflected in the order which the reinsurers sought from the court in order to pre-empt irrelevant evidence and issues at trial (see at [2003] 1 All ER (Comm) 774 at para 9)

“that the [insured] be debarred from adducing or relying on evidence of, or relating to, any matters that occurred after the contract of reinsurance had been written or evidence that was not available to them at that time, with a view to proving that the allegations against Mr Medina were without foundation.”

That issue does not therefore cover the present case, where the outcome of the January 1994 accident occurred and was known to Dr Singh prior to contract, even if then unknown to Provident. This case is also different from the facts in Lynch v. Hamilton (1810) 3 Taunt 15, Lynch v. Dunsford (1811) 14 East 494, 104 ER 691, which this court relied on in Brotherton as being the closest and most significant earlier authority. In that case a cargo owner failed to disclose a report that the ship carrying his cargo was unseaworthy. The report turned out afterwards to have been completely untrue, as was I think proved at trial, but the cargo owner did not know that at the time of placing his insurance. Lord Ellenborough CJ spoke of the report as one “which…afterwards turned out not to be true” (at 497, 692) and Bayley J said that the cargo owner was “willing to take that chance upon himself” (at 498, 692), responding to a submission put in argument on behalf of the insured that where he does not know the truth of a report, he may take his chance: “If it turned out to be true, the policy would be avoided, if false, it would be immaterial” (at 497, 692). Moreover, at para 28 of his judgment in Brotherton, Mance LJ emphasised that even a case where the new information was available by the time of avoidance might be significantly different from the facts in Brotherton. Nevertheless, there is in Brotherton repeated rejection of any element of hindsight in the analysis of materiality or inducement, and Buxton LJ stressed the importance of the requirement of certainty in insurance dealings (at para 41). Nor is it clear to me that the result of Lynch v. Dunsford would have been any different even if the cargo owner had known at the time of contracting that the vessel was seaworthy. For, even if the underwriter had been informed of the cargo owner’s knowledge, he would still have been in a position where he had two inconsistent reports about a vessel at sea with no way of testing between them until further information which he could consider wholly reliable had become available to him. In such circumstances the underwriter might have declined the risk or rated it more highly. If the risk had been declined, the loss would never have fallen on him. And if the risk had been accepted but rated more highly, the underwriter would have been entitled to keep his higher premium even if subsequent information had proved the cargo owner right.

The present case is unusual, however, in that although the speeding conviction had to be disclosed, for there was a specific request for information about it, the further information about the outcome of the accident did not have to be disclosed and yet, because the underwriting criteria were writ in stone, had that outcome been known to the broker (wearing his underwriter’s hat) it would have rendered the non-disclosure of the speeding conviction of no significance. If it mattered that Provident needed to see evidence of the settlement of the accident in the Singhs’ favour, that could have been provided and did not depend on further report nor evaluation at trial. When account has to be taken of a non-disclosure, the issue moves from the world of actual fact into the world of hypothesis. The non-disclosure is the actual fact, and the hypothesis is what effect disclosure would or might have had on a prudent underwriter (the issue of materiality) and what effect disclosure would have had on the actual insurer (the issue of inducement). I do not at present see why the hypothetical world is one in which the insured is assumed to have made the disclosure but not assumed to have provided true information about the settlement of the earlier accident as a no fault accident. On that basis, the non-disclosure of the speeding conviction could not have been material because it could not have induced a re-rating.

Mr Moxon-Browne submitted that materiality had been accepted by Drake before the judge. However, if upon the true facts the speeding conviction could not have led to a higher rating, I do not see why the speeding conviction would ultimately be material, even though its disclosure was required. In any event, it could not be said to have induced the contract for, on the true facts at the time of renewal, the contract could not have been made on terms other than the terms on which it was in fact made.

As for the argument from certainty, this is a case where, as of the time of renewal, there is no doubt about the facts of the no fault accident or Provident’s underwriting criteria. Moreover, the decision whether or not there is a right to avoid is not a decision that has to be made in speed as a matter of instant business: it is a decision made after the event in the light of new facts (the discovery of a non-disclosure) which have to be considered for their legal effect.

Ultimately the issue seems to be: who takes the risk that the true facts as of the time of contract, conclusively established by the time of contract, do not support the right to avoid? I do not see why, subject to estoppel or other such defences, the answer should not be in favour of the insured. However, because there was no citation of authority and I am not sure that there is in any event a case on all fours, I emphasise that my opinion is expressed with caution.

I would therefore be inclined to say, without deciding, that there was in any event no right to avoid in this case.

This is another important issue which I do not think it would be right to ignore, but which I would not wish to decide.

Mr Ter Haar submitted that the right to avoid was limited by the doctrine of good faith and that on the facts of this case Provident owed a duty in good faith to enquire whether the “fault” accident had, by the time of avoidance, been settled so as to become “no fault”. The duty existed because Provident knew in general that a “fault” classification included accidents which were so described merely on a holding basis, pending clarification, and also knew in particular from the details of the proposal form that the accident in question was likely to be reclassified as a no fault accident. It was after all a rear end shunt suffered by the insured’s car and the effect of avoidance was wholly disproportionate to the merits of the case.

In making this submission, Mr Ter Haar made it clear that, in the light of Brotherton, Drake no longer argued that a duty of good faith could survive termination of the contract of insurance following upon an effective avoidance, so as to require post-avoidance reconsideration in the light of further information, nor that avoidance in any way depended on the sanction of the court.

Mr Moxon-Browne on the other hand submitted that the judge had found that Provident had acted at the time of avoidance “in perfectly good faith”, for it was still under the impression that the January 1994 accident was a fault accident (para 11 of the judgment, cited above). It was impossible for this court, on any view of the scope of the duty of good faith during the period after the making of the contract and as it might affect an insurer, to conclude on the facts, in the face of that finding of the judge, that Provident had after all been acting in bad faith. It was only if the facts had been different and Provident had learned before avoidance about the outcome of the accident, or if it had shut its eyes to such knowledge, that then it might be possible that it would have been prevented by its duty of good faith from effecting a rescission.

It is clear that the duty of good faith is mutual and binds the insurer as well as the insured, for section 17 of the MIA provides that “if the utmost good faith be not observed by either party, the contract may be avoided by the other party”. The mutual aspect of the obligation has been referred to in a number of cases, see for example Banque Financière de la Cité (formerly named Banque Keyser Ullmann en Suisse SA) v. Westgate Insurance Co Ltd (formerly named Hodge General & Mercantile Insurance Co Ltd) [1990] 1 QB 665, [1991] 2 AC 249,although I am not aware that it has actually been successfully invoked to provide the insured with any remedy other than a return of premium. The only remedy spoken of by the statute is that of avoidance of the contract. Such a remedy is not likely to be often of much assistance to an insured. See in general Good Faith and Insurance Contracts, 1998, by Eggers and Foss.

Nevertheless, in Carter v. Behm (1766) 3 Burr 1905 Lord Mansfield said this (at 1918/9):

“The underwriter, here, knowing the governor to be acquainted with the state of the place; knowing that he apprehended danger, and must have some ground for his apprehension; being told nothing of either; signed this policy, without asking a question.

“If the objection “that he was not told” is sufficient to vacate it, he took the premium knowing the policy to be void; in order to gain, if the alternative turned out one way; and to make no satisfaction, if it turned out the other: he drew the governor into a false confidence…If he thought that omission an objection at the time, he ought not to have signed the policy with a secret reserve in his own mind to make it void; if he dispensed with the information, and did not think this silence an objection then; he cannot take it up now, after the event.”

I am not sure that this is an example of good faith acting as a remedy, as distinct from it helping to inform and define the extent of the insured’s obligation of disclosure or as supporting a doctrine of waiver of information (see now section 18(3)(b) and (c) of the MIA). However, in Pan Atlantic v. Pine Top at 555E Lord Lloyd of Berwick appears to have regarded this passage as using the doctrine of good faith to curtail the right to avoid, for he said:

“Nor is the obligation of good faith limited to one of disclosure. As Lord Mansfield said in Carter v. Boehm, at p.1918, there may be circumstances in which an insurer, by asserting a right to avoid for non-disclosure, would himself be guilty of want of good faith.”

Moreover, in Manifest Shipping Co Ltd v. Uni-Polaris Shipping Co Ltd (The “Star Sea”), where the ramifications of the doctrine of good faith in the post-contractual period were under consideration, Lord Hobhouse appears to have contemplated that its role in that context may be more versatile than being simply restricted to the statutory right to avoid (see at paras 51/52 and 57); and he also said (at page 497A):

“The courts have consistently set their face against allowing the assured’s duty of good faith to be used by the insurer as an instrument for enabling the insurer himself to act in bad faith.”

In his conclusion, Lord Hobhouse said (at para 79):

“Such authorities show that suitable caution should be exercised in making any extensions to the existing law of non-disclosure and that the courts should be on their guard against the use of the principle of good faith to achieve results which are only questionably capable of being reconciled with the mutual character of the obligation of good faith.”

In my opinion it would be consonant with these views that the doctrine of good faith should be capable of limiting the insurer’s right to avoid in circumstances where that remedy, which has been described in recent years as draconian, would operate unfairly. After all, Lord Hobhouse in The Star Sea described the doctrine of good faith as “a principle of fair dealing” (at para 48). It is true that Lord Mansfield’s conception that the doctrine of good faith would play a role across the law of contract generally has not borne fruit; and even within the insurance context where the doctrine has survived it has been used almost exclusively to protect the insurer against the insured, especially in the context of pre-contract disclosure and the requirement of a “fair presentation of the risk”. Nevertheless, more recently there appears to have been a new realisation that in certain respects English insurance law has developed too stringently or at any rate insufficiently flexibly: and leading cases of the last few years have shown the courts to be willing to find means to introduce safeguards and flexibilities which had not been appreciated before: see Pan Atlantic v. Pine Top itself, which discovered the requirement of inducement to be implicit in the MIA, The Star Sea, which discusses the role of good faith in the post-contractual situation, and K/S Merc-Scandia XXXXII v. Certain Lloyd’s Underwriters subscribing to Lloyd’s Policy No 25T 105487 (The “Mercadian Continent”) [2001] EWCA Civ 1275, [2001] 2 Lloyd’s Rep 563, which limits the operation of the right to avoid for want of good faith on the part of the insured in the post-contractual context to situations where the law of contract would justify termination on repudiatory grounds.

However, examples where Lord Hobhouse’s principle of fair dealing has been exercised to prevent an insurer from utilising a prima facie right to avoid are not to hand: and it may readily be appreciated that, if once an insured has been found wanting in good faith in the matter of pre-contractual non-disclosure, it is likely to be hard to conclude that the same doctrine of good faith itself prevents the insurer from exercising his right to avoid. On the whole English commercial law has not favoured the process of balancing rights and wrongs under a species of what I suppose would now be called a doctrine of proportionality. Instead it has sought for stricter and simpler tests and for certainty. As Lord Hobhouse again said in The Star Sea (at para 45):

“Lord Mansfield’s universal proposition did not survive. The commercial and mercantile law of England developed in a different direction preferring the benefits of simplicity and certainty which flow from requiring those engaging in commerce to look after their own interests.”

However, in insurance Lord Mansfield’s proposition did survive (in a form). Moreover, not all insurance contracts nowadays are made by those who engage in commerce. The existence of widespread insurance contracts of a consumer nature presents new problems. It may be necessary to give wider effect to the doctrine of good faith and recognise that its impact may demand that ultimately regard must be had to a concept of proportionality implicit in fair dealing.

The present case permits an opportunity to explore these considerations. At the end of the day, however, I do not think it is open to this court to go behind the finding of the judge that Provident acted in perfectly good faith in avoiding the contract. The issue that Provident could not validly rescind the policy where it would have known, if it had made any investigation as to the true facts, that the accident was a no fault accident and could not have contributed to an increase in premium was raised on Drake’s pleadings, but that went to a plea of immateriality and did not go expressly to Provident’s good faith. Even so, some of Mr Shaw’s evidence was directed in a general way to Provident’s good faith. Thus he dealt with the suggestion that Provident had avoided the policy in order to escape liability for what was known to be a large claim: he pointed out that the claim had not even been referred to the technical claims department, which is where more serious claims are dealt with, and had been given a modest reserving figure. He said that, upon the underwriting department informing the accidental damage claims department that the non-disclosure of the speeding conviction was material in that it would have led to a 25% increase in premium, the decision to avoid, although ultimately resting with the claims department, would invariably follow the underwriting department’s advice. As for the suggestion that Provident might have made further enquiries of Dr Singh before avoiding, he said that “this is just not what we do. It is not what happens and I would not accept that there was anything to suggest that this was appropriate or reasonably required”. This court was not taken to any passage in his cross-examination in which he was challenged on or departed from this. Nevertheless, Drake’s sole ground of appeal in this connection is that it was a breach of the duty of good faith for Provident to avoid before asking for further information on the status of the accident.

In the circumstances, I do not think that Drake can succeed on this issue, despite the brevity with which the judge deals with the matter, understandably in view of the state of the pleadings. The position obviously changed once Provident had been informed (in March 1997) that the accident was a no fault accident, but by that time, subject to the issue of waiver or reinstatement, the policy had already been avoided. If, however, the point were a live one, I would hazard the opinion that knowledge or shut-eye knowledge of the fact that the accident was a no fault accident would have made it a matter of bad faith to avoid the policy. Of course, if my opinion above that on the facts of this case Provident never had a right to avoid is correct, then this second issue would not add anything. It would be another example of how the rules relating to non-disclosure and materiality themselves define the opportunity for avoidance in good faith. But if my view of that matter is wrong, and Provident prima facie had a right to avoid, the question arises whether it could be done in good faith. In my opinion it could not where Provident knew or was shutting its eyes to the fact that the accident was a no fault accident. Mr Shaw himself acknowledged that if Provident had been notified prior to 2 August 1996 that the accident was a no fault accident and Provident had been satisfied that it would have been right to reclassify it, then Provident would not have been entitled to avoid (see para 18 above). After all, the risk which Provident had in fact been running throughout the period of the renewal was the standard risk, and Provident would have known that before avoiding.

The question then arises whether something less than such knowledge would have been enough to qualify an unrestricted right to avoid. Plainly, the description of the January 1994 accident on the proposal form could be said to put Provident on notice that by February 1996 the accident might have been settled in Dr Singh’s favour, entitling him to its reclassification as a no fault accident. Is it then enough that Provident was put on notice? In many areas of the law, it is sufficient to put a party on notice in order to affect his rights: but that is in the context of equity and rights in property. I can see scope for debate about this, but I would be inclined to say that notice was not enough unless there were to be a general principle that, at any rate where there is notice, it would not be in good faith to avoid a policy without first giving the insured an opportunity to address the reason for which the insurer is minded to avoid the policy. Despite Mr Shaw’s evidence about Provident’s practice, I suspect that, at any rate in matters of business, insurers often do raise an issue of non-disclosure in correspondence with the insured before avoiding the policy. It might be said that this would be a salutary rule especially given the difficulty of applying any test of good faith after avoidance. There are, however, counter-arguments along the lines that a rule of this kind would be a recipe for uncertainty and dispute, including issues of waiver. I do not think that this is the occasion to do more than expose these considerations.

I would add only this. If it is right to allow that circumstances could arise where an insurer would not be in good faith by acting on a prima facie right to avoid for non-disclosure, then the question would have to be faced as to the conceptual analysis whereby an exercise of a right to avoid could be invalidated by the insurer’s bad faith. This is not an easy question. It is evaded if the insurer’s bad faith is used to render a non-disclosure immaterial in the first place: because in that case no right to avoid ever arises. This technique is the subject matter of Lord Hobhouse’s discussion at paras 54/56 of his speech in The Star Sea, and it is probably the authorities and their like there alluded to that Lord Hobhouse primarily had in mind in saying (at para 57) that the courts had set their face against allowing the assured’s duty of good faith to be used by the insurer as an instrument for enabling the insurer himself to act in bad faith. That suggests that the answer to the special facts of this case might more properly or as properly belong under the first issue.

If I am wrong to have concluded that, on the point of inducement, Provident has failed to make good its entitlement to avoid the policy, the question arises whether nevertheless the right to avoid was waived or the policy was reinstated. The facts concerning this issue have been set out above.

If the policy was successfully avoided, I do not think that those facts make good a case for reinstatement. On that basis all that happened was that there was some uncertainty within Provident as to whether it would look again at the rescission, a willingness to go forward with the insurance of a new car, the Volvo, and administrative delay in closing down or altering the direct debit to reflect the avoidance and in returning premium. Although the insurance of the Volvo used the same policy number, nevertheless if the original policy had gone, I would agree with the judge that the evidence does not support reinstatement of that original policy.

The prior question, however, is whether the original policy had gone. On this question it seems to me that the material is ultimately equivocal. On the one hand there is Provident’s letter of 2 August 1996 and subsequent correspondence relating to Provident’s unwillingness to meet the claim and its avoidance of the cover. If all that had stood by itself, there would be every reason to agree with the judge that there was an unequivocal avoidance effected by the letter of 2 August 1996, an avoidance which was subsequently reconfirmed and never departed from. So far as the judge was concerned, the letter of 2 August was all that had to be looked at for the purposes of the argument. He found in it a “clear and unequivocal communication by Provident of its decision to avoid the policy. As a result the contract was rescinded with immediate effect” (para 36). He then went on to discuss the separate issue of reinstatement (at paras 37/40) and concluded that Provident had not agreed to reinstate the policy.

But the letter of 2 August did not stand by itself. The avoidance was disputed. At some time after 1 September, the certificate of insurance of the Volvo was sent to Dr Singh on a document which included an amendment to the original policy: ex hypothesi, on Provident’s case, that original policy, which had been amended as from 20 July, ie before the avoidance, had disappeared and there was nothing to attach the insurance of the Volvo to. Here, however, was a communication from Provident to Dr Singhtelling him that his policy was still in existence. Moreover, Provident continued to take money from Dr Singh’s account on a basis which was only consistent with the continuation of the policy: those direct debits were in their way also communications from Provident to Dr Singh. In November there was correspondence between Provident and Dr Singh regarding these direct debit payments, from which it is again unequivocably clear that Dr Singh’s policy was still in existence and being actively operated by Provident.

What are we to make of these inconsistent communications? All of them cross the line between the parties to the policy and therefore have to be considered for what they say on an objective basis. One possibility, I suppose, is to regard the continued operation of the policy and payments under it as merely accidental, mistaken, administrative acts by one part of Provident’s organisation which did not know that the policy had been cancelled – a case of the right hand not knowing what the left hand was doing. If that was the right way, objectively, to view those communications, then the inconsistency disappears and the road is open to say that a reasonable person in the position of Dr Singh would have understood that there was a single unequivocal line being adopted by Provident, to rescind his policy. However, I think it is rather hard so to regard Provident’s continuation with the policy. It would mean that Dr Singh should have known that a certificate which assured him that he was insured and that his policy “satisfies the requirements of the relevant law” was not worth the paper it was written on; and that an accompanying “Policy Amendment Schedule” misstated the position.

What is more, we know from Mr Shaw’s evidence that the continued operation of Dr Singh’s policy and the obtaining of payments from him were not merely mistaken, administrative acts by one part of an organisation which did not know that the policy had been cancelled, but were the deliberate acts of Provident pursuant to instructions to maintain and operate the policy pending further thoughts, and were regarded by Mr Shaw as such (see para 23 above). That evidence of course goes to matters internal to Provident and does not “cross the line”. Nevertheless, it does nothing to undermine the view that the proper objective consideration of these acts was that they evidence a policy which had not been avoided but was still in operation.

The judge at this stage of the argument looked only at the letter of 2 August, and one possibility, which somewhat reflects that approach, is to take that first letter all by itself: if it is unequivocal in one direction, then it provides the answer. That is, to my mind, however, an unnecessarily restricted approach. In the context of the opposite process, of making as distinct from rescinding a contract, there is high authority for the proposition that one should look at the whole series of communications to find whether a contract has been concluded, even if it might appear from a particular document that the parties had already then reached a binding consensus: see Hussey v. Horne-Payne (1879) 4 App Cas 311 at 316 where Earl Cairns LC said –

“You must not at one particular time draw a line and say, “We will look at the letters up to this point and find in them a contract or not, but we will look at nothing beyond.” In order fairly to estimate what was arranged and agreed, if anything was agreed between the parties, you must look at the whole of that which took place and passed between them.”

It might be said that such a rule or guideline only applies where the negotiation of a contract has to found in the course of correspondence: whereas here all that was required was communication of Provident’s unilateral decision or election to avoid the policy. The difference between such situations has to borne in mind: but they are not all that different and in any event the letter of 2 August is in truth part of a continuing relationship.

Thus, although Provident’s election is a unilateral act, it has to be communicated to Dr Singh to be effective; just as an acceptance, in one sense a unilateral choice, has to be communicated to be effective (subject to the posting rule). As for a continuing relationship, this case provides a good example of it. On 17 February the policy was renewed, on 20 May the Mercedes replaced the Renault and an additional premium was charged, on 2 July Provident was informed of the accident which had occurred that day, on 18 July the written claim form was signed and sent, on 20 July the Volvo was added to the policy and an additional premium was charged, on 24 July the speeding conviction was noted internally as a ground for avoiding the policy, on 2 August the letter of avoidance was sent, on 17 August a direct debit payment was made which covered premium in respect of both the Mercedes and the Volvo, on 19 August Dr Singh disputed the avoidance, on 3 September Provident replied maintaining its position, sometime after 1 September the Volvo certificate of insurance and the amended policy schedule were sent to Dr Singh, there was further correspondence in October, there were further payments, and so on. I do not think that it is correct to look solely at the letter of 2 August, however important or even determinative it might after all be judged to be, for to do so does not reflect the full range of the objective facts. Of course, just as in the context of making a contract, if on analysis an acceptance is binding but the parties continue negotiating, then it is only if those further negotiations themselves bear fruit (in the present case, if there is reinstatement) that the effect of the binding acceptance is altered. The question is whether the single document is determinative.

In my judgment, in this case it was not. On one view the policy was avoided from inception; on another possible view (that of Mr Shaw himself) the matter was unsettled while it was debated; on a third possible view (put forward by Mr Moxon-Browne in submissions) the policy was avoided so far as the Mercedes was concerned, but survived so far as the Volvo was concerned. It seems to me that, as soon as the focus is adjusted more widely so as to permit consideration of anything other than the letter of 2 August, the matter becomes equivocal. It is not possible to rescind a contract in part, just as it is not possible to accept a repudiation, another act of election, in part (unless a contract is divisible). In the present case, Provident put its case exclusively on the letter of 2 August, and yet accepted that a contract of insurance continued after 2 August in respect of the Volvo, even though that car was insured under the same policy, and even though at the time it was insured Provident had not yet been informed about the undisclosed speeding conviction.

I would therefore conclude, and decide, that for this reason, as well as on the previous ground of inducement, Provident remained bound by Dr Singh’s policy.

Despite its success so far, Drake will fail in this appeal unless this last issue is also answered in its favour. It faces an authority of this court which Provident submits, and the judge agreed, is conclusively against it, viz Legal & General v. Drake.

In that case Drake was in the position now occupied by Provident. The driver was insured both by Drake and by L&G. L&G settled the third party’s claim after discovering about Drake’s policy (in this respect the headnote is inaccurate, see the facts stated at 163E). It then sought a 50 per cent contribution from Drake. Drake had already repudiated liability under its policy with the driver on the ground that it had not been notified of the claim within time. For the same reason it said that there was no case of double insurance to give rise to a claim for contribution. This court disagreed (by a majority). The essential point was that at the date of the accident both insurers were liable. If that had been the only point in the case, L&G would have won its contribution.

However, a second point arose late in the day, so much so that Lloyd LJ said that it was “unfortunate that the point was not so fully argued as the other points in the case” (at 164C). This was that since L&G was only liable under the rateable proportion clause in its policy to pay half of its insured’s claim, in settling the whole claim it had paid as a volunteer and could not recover a contribution from Drake. Contribution in equity only arose where the claimant had been obliged to discharge an obligation which in equity rested on both. In accepting this new argument, Lloyd LJ said this (at 164):

“I find this new point a difficult one, the more so because of the impact of Part VI of the Road Traffic Act 1972, now re-enacted as section 151 of the Road Traffic Act 1988…Under section 149 of the Act of 1972, a third party who has obtained judgment against the insurer in respect of a liability required to be insured under the Act, can enforce the judgment against the insurer, notwithstanding any provision contained in the policy of insurance, such as the rateable proportion clause. Assuming the settlement of the third party’s claim, followed by a court order approving the settlement, is a “judgment” for the purposes of section 149 of the Act, it could be argued that the plaintiffs were compelled to pay the whole of the claim by force of law, in which case the excess over 50 per cent. was not a voluntary payment.

“The difficulty with that argument is that the plaintiffs, though obliged to pay the third party the whole of his claim, were entitled to recover the excess over 50 per cent. from Mr Arora [the insured] himself: see section 149(4). It follows that, so far as the defendants are concerned, the excess over 50 per cent. was a voluntary payment. I cannot see any answer to that reasoning…

“[Counsel] argued that the plaintiffs were acting very properly in not seeking to recover the excess over 50 per cent. from Mr Arora, and that it would be an unmerited consequence to deprive them of their right to contribution. Insurers should not be encouraged to take every legal defence, and pursue every legal remedy, which may be open to them against their assured. This is a valid point as far as it goes. But to allow a claim against the defendants based on such considerations would extend the equitable doctrine of contribution beyond any previous authority. I conclude, somewhat reluctantly, that the new argument must prevail.”

On this point both Nourse LJ and Ralph Gibson LJ agreed, albeit possibly stressing different reasons for doing so. Nourse LJ said (at 165F) that L&G’s rights to recover the excess over 50 per cent from the insured under section 149(4) “seems to be a conclusive objection” to a right of contribution. Ralph Gibson LJ said that there could be no right to contribution “in any case where there are effective rateable proportion clauses in the policies of each co-insurer” (at 173C).

There was no discussion in the judgments about the significance of a payment by the claimant insurer being made under reserves pending the resolution of a dispute between the two insurers. It would seem that there were no facts to support such a view of the situation in that case.

Legal & General v. Drake was considered by the Privy Council in Eagle Star Ltd v. Provincial Insurance plc [1994] 1 AC 130. In that case both insurers were liable under the Road Traffic Act (“RTA”), but not under their policies for each of them was entitled to repudiate contractual liability. Both policies contained rateable proportion clauses. Both insurers were entitled to cancel their policies for breach of condition and had done so: but in one case the policy had been cancelled already before the accident and in the other case only after the accident. The third party sued both insurers under the statute, and both were liable to him. However, the insurer who had cancelled before the accident submitted that the whole liability fell on the other, since its was the only policy in existence at what had been held in Legal & General v. Drake to be the relevant time. On this point, the Privy Council differed from the majority in this court. Lord Woolf said (at 141A/B):

“Approaching the issue as a matter of principle, in a case such as the present where both insurers are required to indemnify a third party by statute, there can only from a practical point of view be two solutions to the question of contribution: either the insurers should contribute in accordance with their respective statutory liabilities so that, if they are statutorily equally liable, they will so share the loss; or contribution is determined in accordance with the extent of their respective liabilities to the person insured under the separate contracts of insurance. Of these two alternatives, the contractual approach is the more appropriate since the extent of their respective liabilities to the person insured will indicate the scale of the double insurance.”

Lord Woolf went on to demonstrate that, looking at the matter contractually as between insurer and insured, what was important was the incidence of liability, not the date at which an insurer was discharged from liability. He gave examples as follows (at 141D):

“If both insurers would be under no liability to the person who would be insured, then they should share the statutory liability for loss equally irrespective of the date upon which they repudiated liability. If both insurers are liable at least in part to the person insured, then they should contribute to their statutory liability in accordance with their respective liability to the person insured for the loss.”

The reference to the insurers being liable “at least in part” is interesting, for this example covers the case where both policies, as there and as here, contain rateable proportion clauses. Albeit the Privy Council did not have to deal with the second point in Legal & General v. Drake, the voluntary payment point, it appears to be the logic of Lord Woolf’s analysis that where both insurers would be liable under their policies, but because of the clause liable for only 50 per cent of the loss, the right of contribution is not lost, but preserved – with the result that each should contribute to its statutory liability in the same proportion.

Three questions therefore arise in relation to this fourth issue. The first is whether, as Ralph Gibson LJ suggested, the rateable proportion clause excludes the right of contribution in general. The second, is whether Legal & General v. Drake can be distinguished on the ground that in this case Mrs Kaur could have said that, because of the arbitration award against Dr Singh, she could defeat any submission by Drake, if it had been minded to take the point, that it was liable only for 50 per cent of the loss on the ground that at the time the claim arose “there is any other existing insurance covering the same loss”. The third, is whether the presence of Drake’s protests is an alternative point to distinguish this case from the conclusion that Drake’s payment was voluntary.

As to the first question, it seems to me that Ralph Gibson LJ was in this respect going further than the logic of the decision of this court in Legal & General v. Drake required, as was demonstrated by the Privy Council’s decision and reasoning in Eagle Star. Subject to the question of voluntary payment, which to my mind is a separate matter, the Privy Council did not regard the mere existence of the rateable proportion clause as excluding the operation of the equitable rule of contribution. If it had done so, it could not have found in favour of contribution in that case, for it specifically rejected the statutory approach in favour of the contractual approach in the passage I have cited above. If I may respectfully say so, I would agree that the clause does not of itself exclude a right to contribution, seeing that the essence of the clause is to apply a rateable liability as a matter of contract. It would therefore be surprising if it had the effect of prejudicing one insurer who had paid too much. Each insurer has not ceased to insure for the same loss just because it cannot be forced by contract to pay more than its rateable proportion or because the clause in effect requires the insured to involve both his insurers at once in order to obtain a full indemnity. In Legal & General v. Drake at 161F Lloyd LJ asked:

“But what is a rateable proportion clause other than an attempt by insurers to exclude the equitable doctrine of contribution by a contractual provision intended to achieve the same effect?”

But he did not say that the attempt succeeded. Of course, if the insured is forced to involve both his insurers, then there will be no need for contribution and the need for the application of the doctrine is excluded as a matter of fact.

As to the second question, it is clear that in Legal & General v. Drake, because of the decision on the first point, the two insurers were treated as co-insurers, even though one of them had a complete defence to liability. Mr Ter Haar attempts to distinguish that authority by seeking to show that as between Drake and Mrs Kaur Drake was the only insurer, therefore it could not be a volunteer. However, in order to bring itself within the right of contribution it has to show that it was a co-insurer vis-à-vis Provident.

If, therefore, Drake is entitled to say in this case both that, having succeeded on the earlier issues, it can regard Provident as a co-insurer and that, because Mrs Kaur could on the basis of the arbitration award against Dr Singh have defeated any attempt by Drake by reference to Provident’s policy to limit its liability under its own policy, it can regard Provident as not being a co-insurer, then it will have successfully distinguished Legal & General v. Drake. For that case was decided on the straightforward basis that, at any rate at what was there regarded as the relevant time, the two insurers were co-insurers. In this case, however, because of the arbitration award, it is suggested that, as against Mrs Kaur, Drake was the only insurer.

In other words, in order to claim a contribution in equity from Provident, Drake must show that it was a co-insurer, but in order to escape from the decision in Legal & General v. Drake, Drake must show that it was not a co-insurer. Moreover, in order to show that it was a co-insurer, Drake must rely on the fact (which is common ground) that it is not bound by the arbitration award between Provident and Dr Singh; whereas in order to show that it was not a volunteer and was bound to indemnify Mrs Kaur against 100 per cent of her liability, Drake must rely on the fact (which is not common ground) that she would be defeated by the same arbitration award. It escapes being a volunteer on the ground that it is not a co-insurer and claims contribution on the ground that it is a co-insurer.

I am therefore perplexed by this submission. I am also concerned by its ramifications. I put on one side the distinction between Dr Singh and Mrs Kaur, for I am prepared to assume, as was common ground, that Mrs Kaur was bound by the arbitration award between Dr Singh and Provident as if it were in her own name: see section 58(1) of the Arbitration Act 1996. However, one link in the argument is that the award operates retrospectively by vindicating Provident’s claim to avoid its policy with Dr Singh from inception: therefore at any relevant time the Provident policy had, albeit retrospectively, ceased to exist and the rateable proportion clause was not engaged (“If at any time any claim arises under this Policy there is any other existing insurance…”). The court heard brief but conflicting submissions as to whether the “claim” spoken of in the clause is synonymous with the accident or at any rate its notification to the insurer for the purpose of making a claim on the one hand, or with the establishment of liability against the insured on the other; and also as to whether “existing” does or does not include a policy which is subsequently avoided.

Thus in the present case, if the initial factor for the application of this clause is the true state of affairs between Provident and Dr Singh as found by this court, there was double insurance at all times, and thus the rateable proportion clause applied. If, however, what is critical for that purpose is not the true state of affairs but the award of the arbitrators then that occurred after Drake was involved, albeit it related back to an avoidance on 2 August 1996 and in one sense related back to the inception of the policy. On that basis, was the Provident policy an “existing insurance” when the claim arose under the Drake policy? If one looks at the time of the accident or the time when Drake was notified, the Provident policy was in one sense still in existence, for the critical award had not yet been decided in Provident’s favour. In other words, prior to the award Drake could say against Mrs Kaur that there was double insurance and the clause applied, whereas after the award Mr Ter Haar submits that Drake’s position had altered. On the other hand, if the award is to be regarded as relating back to the inception of the Provident policy, then there was never double insurance. If it relates back to avoidance on 2 August, then it becomes important to decide whether the “time any claim arises” refers to the time of accident or the time the insurer is notified of a possible claim, for if the former the policy was still then in existence. In Legal & General v. Drake Ralph Gibson LJ referred to the date of the accident as the relevant time at 170F. Nourse LJ also referred to the date of the accident at 165C.

Whatever be the precise facts of this case, however, what of a case where both insurers are contacted immediately after the accident and one insurer then successfully avoids? If regard is had to the fact that at the time of claim both insurances were in existence (for an avoidable policy is only avoidable, not void), then the insured may be in a position where it can recover only half its claim. In Legal & General v. Drake Lloyd LJ said that the test of double insurance was one of potential liability at the date of loss and added (at 160E):

“But when I say potentially liable, there is a sharp distinction between steps required to enforce a valid claim under a policy in force at the time of loss, and a claim which was never valid, and never could be enforced. Thus if B has a good defence to the assured’s claim on the basis of misrepresentation or non-disclosure, there is no double insurance. Since the effect of the defence is that the contract is avoided ab initio, it is as if B had never been on risk at all.”

Lloyd LJ was there speaking generally about the problem of double insurance and was not concerned with the construction of the rateable proportion clause. It is a matter of concern that an insurer who takes a premium to cover 100 per cent of a risk may only be liable for 50 per cent of a loss, on the basis that the insured can obtain the other half elsewhere, in circumstances where the insured finds that he cannot contractually do so.

There is also a problem about the words “covering the same loss…”. Does this mean providing a remedy for it, or does it speak merely potentially, as Lloyd LJ considered was the position as a matter of the general philosophy of double insurance?

Having expressed these concerns, I am nevertheless inclined to agree with the approach of Clarke LJ on this issue. It is not so much a question of whether the arbitration award is binding on Drake: if it is not binding on it vis a vis Provident, it is not binding on it vis a vis Mrs Kaur. It is rather that the award is a fact in the world, binding as between her and Provident, to which, had Drake sought to enforce the clause against her, she was in a position to point on the issue of whether there was another existing insurance. The award was relied on as such by Provident in its correspondence with Drake. Thus it is common ground that the question whether there was double insurance to enable Drake to invoke the principle of contribution has to be determined on the position as found by this court, irrespective of an award which is not binding on the parties to this litigation. On the question, however, of whether Drake was a volunteer, Drake is entitled to point to the arbitration award as creating a grave if not insuperable difficulty in any dispute with Mrs Kaur as to the limit of its liability to her.

In these complex circumstances, I would nevertheless prefer to decide this fourth issue, as I think I can, on the third question, which is whether Drake’s protests prevent it from being a volunteer. This is a matter which was not discussed at all in Legal & General v. Drake. If, therefore, the protests are relevant, then there is here a potential ground for distinguishing that authority.

In my judgment, the protests are a relevant point of distinction and entitle Drake to recover against Provident.

Let me take a moment to remind my reader where the position stands. I have concluded that Drake was not entitled to avoid its policy with Dr Singh. Therefore there was in fact double insurance and prima facie Drake was entitled to recover a 50 per cent contribution from Provident in equity. The rateable proportion clause is not an agreement between Drake and Provident, but it is an attempt by contract with the insured to recreate the equity and forestall the need for its application. The only reason why it is said that Drake may not recover a contribution from Provident is that Drake was a volunteer in paying Mrs Kaur more than the sum to which she was entitled under Drake’s policy with her. However, if Drake was to enforce the rateable proportion clause against Mrs Kaur, it potentially faced grave difficulties, both of law and of business responsibility. In law, Mrs Kaur would have relied on the arbitration award against Dr Singh. There could also have been other difficult questions relating to the construction of the clause. In terms of business responsibility, Drake formed the view that it would have been contrary to business ethics and the good name of the industry either to refuse Mrs Kaur a full indemnity or to involve her in litigation between the two insurers potentially concerned. Therefore, Drake made clear to Provident its views on the matter, in particular that it was in dispute with Provident as to the latter’s liability to join in indemnifying Mrs Kaur, that it would pay under reserves and litigate the issue with Provident if it had to. Provident could have been under no illusions, and none are suggested, as to what Drake’s position was and that litigation was around the corner.

Is this the position of a volunteer? In my judgment, no. The question is not whether Drake could recover the money from Mrs Kaur, or Mr Beach. The question to my mind is whether the payment made in the circumstances in which it occurred was voluntary as against Provident, so as to remove the equity for Drake’s prima facie right to recover a contribution from it. If there had been no rateable proportion clause and the sole question was whether Drake had made a reasonable settlement with Mrs Kaur and Mr Beach, I cannot see what prospect Provident would have had for calling that settlement into question. Similarly here, in the presence of the clause.

Moreover, I do not think that anything can in the circumstances turn on the fact that the issue between Drake and Provident was litigated after the settlement with Mr Beach rather than before. If Drake had insisted that settlement with Mr Beach and the indemnification of Mrs Kaur should await the outcome of its litigation with Provident, or that Mrs Kaur (and/or Dr Singh) should sue Provident as part of that litigation, it would not only have brought the insurance industry into disrepute, but would also have relegated substance to mere form. If the litigation had preceded the payment, the voluntary payment point would not have existed. The mere fact, against the background of Drake’s reserves, that its litigation with Provident follows the payment cannot be critical.

In reaching this conclusion I have put on one side the fact that under section 151 of the RTA Drake could in any event have been compelled to make payment to Mr Beach once he had obtained an unsatisfied judgment against Mrs Kaur. I do so because that was not viewed as saving the situation in Legal & General v. Drake, even though it is not clear to me from the judgments in that case why the merely theoretical right to recover the excess of Drake’s liability under section 151(7) (there section 149(4) of the preceding Act) should make all the difference.

In conclusion, I would allow Drake’s appeal on the grounds that Provident was not entitled to avoid its policy and Drake was not a volunteer.

In preparing this judgment, I have had the advantage of reading the judgments of Rix LJ and Pill LJ in draft. I agree with them that the appeal should be allowed on the footing that Provident was not entitled to avoid the policy and that Drake was not a volunteer when it paid Mrs Kaur’s claim. I add a short judgment of my own because the reasons which have led Rix LJ and Pill LJ to those conclusions differ in some respects, because we have reached different conclusions from the judge and because the case raises or potentially raises a number of interesting and difficult aspects of the law of insurance. I shall adopt the same headings as Rix LJ but shall not repeat the facts, which are comprehensively set out by Rix LJ.

Like Rix LJ, I treat the role of this issue as to ask whether, irrespective of any argument that can be raised as a matter of good faith, Provident was entitled to avoid the policy. Also like Rix LJ, I would resolve this issue against Provident on the ground that it failed to show that it was induced to enter into the policy by Dr Singh’s non-disclosure of his speeding conviction. I recognise of course that Pill LJ has reached a different conclusion and for that reason, I will shortly give my own reasons for my opinion, which are I think essentially the same as those expressed by Rix LJ in paragraphs 60 to 64 above.

I assume for present purposes that Dr Singh’s failure to disclose his speeding conviction was non-disclosure of a material fact which ought to have been disclosed, as explained by the House of Lords in Pan Atlantic Insurance Ltd v Pine Top Insurance Co Ltd [1995] AC 501. The question, however, is whether Provident was induced to enter into the contract by that non-disclosure, again as explained in Pan Atlantic. On the facts of the instant case that depends upon Provident showing on the balance of probabilities that, if Dr Singh had disclosed the conviction, Provident would only have entered into the contract at an increased premium. That in turn depends upon whether Provident would have learned that Mrs Kaur’s accident on 6 January 1994 was a no fault accident in the sense explained by Rix LJ, because if it would have learned of that fact it is accepted that Provident would have been willing to enter into the contract without any additional premium in accordance with its underwriting policy.

We were told that at the trial no positive case was advanced on behalf of Drake that Provident was not induced to enter into the contract by what was accepted as a material non-disclosure on the part of Dr Singh. The point was raised by the judge, who proceeded to deal with it in his judgment. Before us, the point was squarely taken by Mr Ter Haar on behalf of Drake and Mr Moxon-Browne did not submit that we should not allow the point to be taken. It follows, to my mind, that we should consider it fully, as indeed the judge did.

Such a consideration involves considering what would have happened if Dr Singh had disclosed his conviction. There can be no doubt that, if he had disclosed it, he would have done so to his broker. Although the broker was his agent, he was also in some respects the agent of Provident, as Mr Moxon-Browne correctly accepted in the course of the argument. In one or other or both of his capacities he would, as I see it, have been bound to consider and calculate the appropriate premium under Provident’s system. On the information known to him, without further information from or discussion with Dr Singh, he would have calculated that the combined effect of the conviction and the ‘fault’ accident was that Dr Singh’s premium would have to be increased.

It seems to me that it was almost inevitable that the broker would have told Dr Singh that the premium would have to be increased and that it is much more probable than not that the broker would either have told Dr Singh the reason or, if he did not, that Dr Singh would have asked the reason. In that event, I entirely agree with Rix LJ that the status of the conviction would have been discussed and Dr Singh would have told the broker that liability for the 1994 accident had been settled on a no fault basis.

As the judge expressly found, once the correct classification of the accident had become a matter for discussion, Dr Singh would probably have been able to persuade Provident that it should be treated as a ‘no fault’ accident, which would have meant that it would have continued to insure Dr Singh without any increase in premium.

It follows from the above that I agree with Rix LJ that the non-disclosure did not induce the contract. I agree with him that it is wrong to say, as the judge did, that there was nothing in the evidence ‘that points either way’. On the contrary the evidence to my mind the evidence points to the fact that Provident would have entered into the contract on the same terms even if the speeding conviction had been disclosed. Like Rix LJ, I would hold that if the burden of establishing that fact was on Drake, whether the burden was legal or evidential, it discharged it. It follows that I would also hold that Provident failed to show that the non-disclosure induced the contract.

Those conclusions make it unnecessary to discuss the interesting questions touched on by Rix LJ in paragraphs 65 to 78 above, save to say that there seems to me to be much to be said for his views and I would be inclined to agree with them, although I have some reservations as to how far it is appropriate to compare avoidance of a contract of insurance for non-disclosure with acceptance of a repudiatory breach.

I wish to comment on only one point which particularly struck me in the course of the argument. It was common ground before the judge that Dr Singh’s conviction for speeding was a material fact which he should have disclosed to Provident. Section 18(1) of the Marine Insurance Act 1906 provides that the assured must disclose every material circumstance which is known to him. Section 18(2) provides:

“Every circumstance is material which would influence the judgment of a prudent insurer in fixing the premium or determining whether he will take the risk.”

In Pan Atlantic v Pine Top the House of Lords stressed that the question of materiality must be answered by reference to the position of the reasonably prudent insurer and not the particular insurer. It was no doubt for that reason that it was common ground that the speeding conviction was material, since the reasonably prudent motor insurer would regard speeding convictions as material. However, that depends upon what characteristics of the reasonably prudent insurer can be taken into account. It seems clear that some characteristics of the particular insurer should be taken into account in order to identify what the reasonably prudent insurer would regard as material. For example, in the case of a proposed motor insurance, it would surely be appropriate to ask what the reasonably prudent motor insurer would regard as material.

It struck me that there may be other characteristics of the particular insurer which might be relevant. So here, it seems to me that it might be relevant to ask whether a speeding conviction would be material to the reasonably prudent insurer with a ‘points’ system of underwriting. As I see it, there would be much to be said for such an approach, since on the facts of this case as they were known at the renewal date, the speeding conviction was not material to the reasonably prudent insurer with Provident’s ‘points’ system because the earlier accident had by then been agreed to be a no fault collision. Indeed, if the broker had been asked whether in such circumstances it was necessary to disclose the conviction to Provident, he might well have said that it was not.

However, I would not rest my decision on this consideration because materiality was common ground between the parties and the question what characteristics can or should be attributed to the reasonably prudent underwriter should perhaps be left for further consideration in a case in which the answer affects the actual decision.

Rix LJ has discussed a number of different aspects of this problem in paragraphs 79 to 93. Again, I am inclined to agree with the views which he has expressed but I too would prefer not to rest my decision on these considerations and would only add these brief comments.

It does seem to me that if, at the time of avoidance, Provident knew or turned a blind eye to the fact that the earlier accident was a no fault collision, it would not be acting in good faith if it avoided the policy because such avoidance would surely not be “fair dealing”, as Lord Hobhouse put it in paragraph 48 of The Star Sea. However, there is no suggestion that Provident in fact knew the true position and I agree with Pill LJ that it would not be appropriate to hold that it turned a blind eye in the relevant sense, especially in the light of the judge’s conclusion that Provident acted in perfectly good faith.

Pill LJ has concluded in effect that Provident was sufficiently put on notice of the true facts to make it a breach of its continuing duty of good faith to avoid the policy without making what he correctly says would have been a simple enquiry as to the true position. There is to my mind much to be said for such a conclusion, if only because of the draconian effect of avoidance by insurers. However, I prefer not to rest my decision on this point because of the difficulty of knowing where to draw the line. Indeed, in the absence of a general principle that insurers must always give the insured an opportunity to address the reason why they are considering avoidance, it is difficult to know what amounts to notice in the absence either of knowledge of the critical facts or of the insurers’ turning a blind eye. There is at present (so far as I am aware) no authority for the proposition that an insurer owes the insured a duty to take reasonable care to make appropriate enquiries before avoiding the policy.

I agree that there was no unequivocal avoidance of the policy for the reasons given by Rix LJ and Pill LJ.

I agree with Rix LJ and Pill LJ that the answer to this question is no for the reasons which they have given but wish to add an opinion on an aspect of the problem which struck me as of some significance in the course of the argument.

It does not follow from the fact that Provident was not entitled to avoid the policy with Dr Singh that Drake is not entitled to contribution from it in respect of its liability to Mrs Kaur. The judge held that Drake is not entitled to contribution because of two further conclusions as follows. The first was that Drake paid Mrs Kaur as a volunteer and the second was that it followed that he was bound by the decision of this court in Legal & General Assurance Society v Drake Insurance Co Ltd [1992] QB 887 to hold that Drake was entitled to contribution from Provident.

Mr Ter Haar submitted that both those conclusions were wrong. I shall consider them in turn. However, before doing so I should note that it was common ground before the judge that, as between Provident and Drake, it is irrelevant that Dr Singh’s claim to be indemnified by Provident failed and that, as between him and Provident, he was bound by the arbitration award to that effect. Equally it was not suggested that Mrs Kaur, as between her and Provident, was not bound by the award. It was, as I understand it, common ground that Mrs Kaur would have been a person claiming thorough or under Dr Singh within the meaning of section 58(1) of the Arbitration Act 1996 and that she was therefore bound by the award.

The judge held that Drake paid Mrs Kaur as a volunteer. He accepted Provident’s case that, on the true construction of the ‘rateable proportion’ clause in her policy with Drake, on the hypothesis that Provident was liable under its policy, she was entitled to only 50 per cent of what should have been the insurer’s full liability under the policy because she was entitled to the other 50 per cent from Provident. The clause in Drake’s policy (which was very similar to that in Provident’s policy) was part of the General Conditions and was in these terms, so far as relevant:

“3.If at any time any claim arises under this Policy there is any other existing insurance covering the same loss damage or liability the Company shall not be liable to pay or contribute more than its rateable proportion of such claim. …”

The judge held that the clause means what it says and that “if Provident was in fact liable to indemnify Mrs Kaur, Drake’s liability to indemnify her was limited to its own share of the loss”.

I would agree that that would be the meaning of the clause if it were the position, as between Mrs Kaur and Provident, that Provident was in fact liable to indemnify her, or, at any rate if there was (in the words of the clause) “any other existing insurance covering the same loss damage or liability”. However, at the time that Drake settled her claim, that was not the position. As between her and Provident, there was no other existing insurance covering the same loss damage or liability because Provident had avoided the policy and, in an arbitration award by which she was bound, the arbitrator had held that it was entitled to do so. The effect of that avoidance, as between Mrs Kaur and Provident, was that the insurance had been avoided from inception.

The position might have been different if Dr Singh had lost the arbitration because of some failure to give a contractual notice or, perhaps, because he had failed to take reasonable steps to pursue the arbitration, but nothing of that kind was suggested here. As I understood it, Mr Moxon-Browne accepted in argument (to my mind correctly) that the question whether Drake was a volunteer in settling Mrs Kaur’s claim on a 100 per cent basis was to be answered by asking whether, if Drake had taken this point against Mrs Kaur, she would have recovered against it on a 100 per cent basis or only on a 50 per cent basis.

In such postulated proceedings between Mrs Kaur and Drake, Drake would have said that there was existing insurance covering the same liability between her and Provident, whereas Mrs Kaur would have replied that there was no such existing insurance because Provident had avoided the policy and had been held to be entitled to do so. As I see it that defence would have succeeded. It would have been no defence for Drake to say that, as between it and Provident or indeed, as between it and Mrs Kaur, it was entitled to say that the policy had not been validly avoided because, in my opinion (as already stated), the question for decision was whether there was existing cover as between Mrs Kaur and Provident and Provident had persuaded an arbitrator that there was not.

For these reasons I have reached a different conclusion from the judge and would hold that, if proceedings between Mrs Kaur and Drake had been fought to a conclusion, she would have succeeded on a 100 per cent basis so that when Drake discharged that liability it was not acting as a volunteer. It follows that when Drake discharged the liability on a 100 per cent basis by way of settlement, contrary to the view expressed by the judge, it was not paying 50 per cent as a volunteer.

I am pleased to be able to reach this conclusion because it appears to me that the point taken by Provident in this litigation is very unattractive. When a claim was made against it by Dr Singh it avoided the policy on the basis of non-disclosure in circumstances in which it accepts that if it had known the true facts it would not have done so. It resisted Dr Singh’s claim and succeeded in doing so at an arbitration. For the reasons given earlier and by Rix LJ (with which I agree), Provident was not in fact entitled to avoid. Yet by its argument that Drake paid Mrs Kaur as a volunteer, it hopes to achieve a result by which Drake remains liable for the whole claim, which in equity should be shared between them. Such a result would show the insurance industry in a very poor light.

It is not, I think, in dispute that, if the construction of the ‘rateable proportion’ clause which I have identified is correct, so that Drake was liable to Mrs Kaur on a 100 per cent basis, and if Provident was not entitled to avoid the policy, Drake is entitled to a 50 per cent contribution from Provident. As the judge correctly put it in paragraph 44 of his judgment, the foundation of the insurer’s right to recover a contribution from others who have insured the same risk lies in the law’s recognition of the obvious justice of requiring that a common liability should be shared between those liable: see Royal Brompton Hospital National Health Service Trust v Hammond [2002] UKHL 14, [2002] 1 WLR 1397, per Lord Bingham at page 1399. It would not be equitable to leave Drake to carry 100 per cent of the liability because, as between Drake and Provident, Provident was not entitled to avoid the policy covering the same liability and should therefore bear 50 per cent of the liability.

That conclusion makes it unnecessary to consider the second of the points identified above, which would involve a detailed consideration of the decision in Legal & General v Drake, where ironically it was Drake which was taking the point which succeeded and which it would I think now like to say was wrong. It is not suggested that the decision affects the construction point which I have just considered. I would only say about the decision that it does seem to me to have unfortunate results if it means that an insurer, instead of making a sensible settlement with the insured and subsequently claiming contribution from a co-insurer, has to rely upon the ‘rateable proportion’ clause in order to protect its position against the co-insurer.

Perhaps the House of Lords will one day have an opportunity to reconsider Legal & General v Drake. However, given the conclusion which I have reached, I do not think that it is necessary to consider its reasoning on the facts of this case. In any event, I agree with Rix LJ and Pill LJ that there is a further ground upon which Legal & General v Drake can be distinguished on the facts of this case. It is that Drake made it clear to Provident (as Rix LJ put it at paragraph 125) that it was in dispute with Provident as to Provident’s liability to join in indemnifying Mrs Kaur, that it would pay under reserve or protest and litigate the issue with Provident if it had to. In these circumstances I agree that Drake is entitled to an equitable contribution of 50 per cent of its liability to Mrs Kaur from Provident.

For these reasons, I agree that the appeal should be allowed.

I gratefully adopt Rix LJ’s comprehensive statement of the facts, including the history of the dispute. I address each of the issues, referring on each issue to the number of the paragraph in which Rix LJ begins his analysis.

I would first venture a few observations on insurance contracts in the light of the stance adopted by the respondents in this case as described by Rix LJ. Many insurance contracts are made between large organisations with very substantial assets, a wealth of experience in business dealing and access to high-quality legal advice. Many others, numerically many more, are contracts by which members of the public insure property as personal to them as their houses and contents and their motor cars. These are of course of great importance to the members of the public who make contracts of insurance with large companies, without their having the resources or expertise of large organisations. The principles which underlie the protection of consumers in other contracts have some relevance in this area. There is also a public interest in the smooth and responsible operation of such contracts. In the case of motor vehicles, which can cause great bodily damage as well as damage to property, that is recognised by the requirement for compulsory insurance provided by European Directives and by statute and, by way of the MIB agreement between the Secretary of State and insurers, for compensation of victims of uninsured drivers. These are trite remarks but the nature and importance of the interests involved do, in my view, bear upon the insurer’s duty of utmost good faith in the present context.

I note Rix LJ’s expression of caution on this subject and also the view to which he is inclined.

There is no duty on the insured to disclose, when the contract is made, matters favourable to him upon the application of ratings adopted by the insurer for the assessment of risk. I am not persuaded that an insured who fails to disclose a factor in his favour can at a later stage defeat a purported avoidance by revealing a fact in his favour upon the avoidance occurring. Attractive though the proposition is that the right to avoid should be judged according to the true state of affairs, the result would, in my view, be to devalue unacceptably the principle that the terms of the contract are those agreed by the parties. Terms of insurance are agreed on the basis of what the insured has disclosed and a contract comes into being on that basis. Undisclosed matters cannot in my view be relied on to demonstrate that the effect of the contract was not what it appeared to be.

I do, however, consider the proposition that regard should be had to the true state of affairs as relevant to the next stage of a consideration of the avoidance: the application of the principle of utmost good faith to the insurer’s right to avoid. It is at that stage that any remedy based on the true state of affairs should arise.

Rix LJ has set out the relevant considerations under this heading and the lack of authority in relation to the insurer’s duty. He has concluded that Drake cannot succeed on this issue.

In Pan Atlantic Insurance & Co. Limited v Pine Top Insurance Co.Ltd [1995] A.C 501, Lord Lloyd of Berwick considered the consequences of the principle that “a contract of insurance is a contract of utmost good faith”. Having considered the duty on the assured to disclose every material circumstance known to him before a contract of insurance is concluded, Lord Lloyd stated, at page 555 D:

“Lastly, the duty of disclosure operates both ways. Although, in the usual case, it is the assured who knows everything, and the insurer who knows nothing, there may be special facts within the knowledge of the insurer which it is his duty to disclose, as where (to take the example given by Lord Mansfield in Carter v Boehm) the insurer knows at the time of entering into the contract that the vessel has already arrived. Thus the obligation of utmost good faith is reciprocal: see Banque KeyserUllmann S.A. v. Skandia (U.K.) Insurance Co.Ltd. [1991] 2 A.C. 249, per Lord Bridge of Harwich, at p. 268 and Lord Jauncey of Tullichettle, at p.281. Nor is the obligation of good faith limited to one of disclosure. As Lord Mansfield warned in Carter v. Boehm, at p.1918, there may be circumstances in which an insurer, by asserting a right to avoid for non-disclosure, would himself be guilty of want of utmost good faith.”

That the post-contractual duty of good faith is a very limited one was stated by Mance LJ in Brotherton v. Aseguradora Colseguros S A [2003] C.A.Civ.705 at paragraph 34.

In Manifest Co.Ltd v. Uni-Polaris Shipping Co.Ltd [2001] 2WLR 170, Lord Hobhouse of Woodborough affirmed the principle that utmost good faith is applicable to all forms of insurance and is mutual. He referred, at paragraph 47, to Lord Mansfield’s judgment in Carter v. Boehm 3 Burr 1905. Nor does it come to an end when the contract has been made. At paragraph 48, Lord Hobhouse stated that “there is a weight of dicta that the principle has a continuing relevance to the parties’ conduct after the contract has been made”. The content of the obligation to observe good faith has, however, “a different application and content in different situations”.

At paragraph 57, Lord Hobhouse stated that “the courts have consistently set their face against allowing the assured’s duty of good faith to be used by the insurer as an instrument for enabling the insurer himself to act in bad faith”. Lord Hobhouse went on to consider the remedy of avoidance:

“…. An inevitable consequence in the post-contract situation is that the remedy of avoidance of the contract is in practical terms wholly one-sided. It is a remedy of value to the insurer and, if the defendants’ argument is accepted, of disproportionate benefit to him; it enables him to escape retrospectively the liability to indemnify which he has previously and (on this hypothesis) validly undertaken…”

Lord Clyde considered, at paragraph 7, the position after the contract is entered into:

“Since even after the contract is entered into the relationship between the parties should in any event be coloured by considerations of good faith, the point is in some respects academic. But once it is recognised that in a contract of insurance, and indeed in certain other contracts, an element of good faith is to be observed, and that that element may impose certain duties particularly of disclosure between one party and the other, duties which may vary in their content and substance according to the circumstances, then a question may arise as to the utility of the concept of an utmost good faith or an uberrima fides. In my view the idea of good faith in the context of insurance contracts reflects the degrees of openness required of the parties in the various stages of their relationship. It is reasonable to expect a very high degree of openness at the stage of the formation of the contract, but there is no justification for requiring that degree necessarily to continue once the contract has been made.”

When the contract was made in February 1995, the application form was typed by the brokers Hexagon Insurance Services Limited and signed by Dr Singh. Under the heading “Claims within the last 5 years”, was entered “Add [additional] Driver 1, Hit by TP [third party] in rear, PI [personal injury] no, fault yes, £2000 own cost, £0 TP cost, NCB [no claims discount] disallowed”. No fresh form was filled in when the policy was renewed in February 1996 but the renewal notice had made clear the need for relevant disclosures and the speeding conviction on 18 December 1995 was not disclosed. By that time, however, the other potential claim had been resolved, the third party’s insurers, Norwich Union, meeting Dr Singh’s claim in full in July 1995. The word “fault” had been inserted in the form by the broker, notwithstanding that the circumstances indicated otherwise, because the terminology used by the respondents was that an accident is treated as a “fault” accident until the absence of fault in the insured is established to their satisfaction. As Mr Moxon-Browne QC put it, the word fault goes to the possibility of a claim for recovery under the policy rather than to blame. An accident becomes a “no fault” accident only when the respondents are satisfied that money will not be claimed against the relevant insurer under the policy. The word “claims” in the printed heading referred to, is thus taken by them to include potential claims.

Mr Moxon-Browne referred to the insurance policy as “quick, cheap, high street insurance”, circumstances in which “rough justice” is to be expected. The policy may have each of the three characteristics mentioned but the words do not convey the importance of the policy to an insured or to the public, given potential third party liability in road traffic claims recognised by the requirement to obtain cover. There is a duty of good faith on the insurer as well as the insured and that exists however “quick and cheap” the insurers may regard it. Avoidance of the policy may have drastic results for the insured, as it has in the present case, or would have had for Dr Singh and Mrs Kaur but for the serendipitous existence of the “driving other vehicles” extension on Mrs Kaur’s policy with Drake. Of course Dr Singh should, in his own interests, have told the respondents, prior to renewal in February 1996, that the consequences of the accident of 6^th January 1994 had been resolved in his favour. Because he did not, it remained registered with them as a fault accident notwithstanding the circumstances of the accident, of which they knew, and the fact that no claim against the relevant insurer had been pursued in the 2½ years between the accident and the purported avoidance.

On 24 July 1996, that is after the notification of the accident and before the purported avoidance, an employee of the respondents wrote to his section supervisor informing him that the SP30 had not been disclosed and that, if it had been disclosed, the company “would have invited renewal with a 25% premium loading”. The same document included a reference to “disclosed fault claim 1-94”.

In the course of argument, reference was made to the possibility of the respondents having blind-eye knowledge that the relevant accident was no longer a “fault” accident. The meaning of blind-eye knowledge was analysed by Lord Scott of Foscote in Manifest. Having referred to Nelson’s conduct at Copenhagen and to cases where the expression has been considered, Lord Scott, with whose speech Lord Steyn and Lord Hoffmann agreed, concluded, at paragraph 116:

“In summary, blind-eye knowledge requires, in my opinion, a suspicion that the relevant facts do exist and a deliberate decision to avoid confirming that they exist. But a warning should be sounded. Suspicion is a word that can be used to describe a state of mind that may, at one extreme, be no more than a vague feeling of unease and, at the other extreme, reflect a firm belief in the existence of the relevant facts. In my opinion, in order for there to be blind-eye knowledge, the suspicion must be firmly grounded and targeted on specific facts. The deliberate decision must be a decision to avoid obtaining confirmation of facts in whose existence the individual has good reason to believe. To allow blind-eye knowledge to be constituted by a decision not to inquire into an untargeted or speculative suspicion would be to allow negligence, albeit gross, to be the basis of a finding of privity. ” [The issue arose in the context of Section 39(5) of the Marine Insurance Act 1906 where the word privity appears].

Lord Clyde stated, at paragraph 3:

“…. Blind-eye knowledge in my judgment requires a conscious reason for blinding the eye. There must be at least a suspicion of a truth about which you do not want to know and which you refuse to investigate…”

Lord Hobhouse stated at paragraph 25:

“The illuminating question therefore becomes “why did he not inquire?” If the judge is satisfied that it was because he did not want to know for certain, then a finding of privity should be made. If, on the other hand, he did not inquire because he was too lazy or he was grossly negligent or believed that there was nothing wrong, then privity has not been made out. ”

The person who made the decision to send the letter of 2 August 1996 purporting to avoid the contract would, assuming as I do that he had read the relevant papers, have known:

In my judgment that knowledge creates more than a speculative suspicion that, at the time of avoidance, the January 1994 accident had no relevance to the level of premium chargeable. There was on those specific facts a firm basis for a conclusion that it was irrelevant. I would, however, be prepared to hold that it does not go quite far enough to establish in the respondents blind-eye knowledge of the true state of affairs.

However, a failure to make any enquiry of the insured before taking the drastic step of avoiding the policy was in my judgment a breach by the insurer of the duty of good faith. Assuming that blind-eye knowledge did not exist, the respondents, knowing what they did, (plus the fact that an insured is unlikely to be aware of the significance attributed by the respondents to the word “fault”) the duty of good faith required them at least to tell the insured what they had in mind and give him an opportunity to update them on the January 1994 accident. All that was required was a simple enquiry as to what had happened in relation to that accident. If more than lip service is to be paid to the principle that an insurer shall show the utmost good faith, the principle in my judgment required that enquiry to be made before the “wholly one-sided” remedy of avoidance was exercised. The respondents’ conduct subsequent to purporting to avoid, though giving rise to different issues, also demonstrates a willingness to continue to insure Dr Singh which may be admirable but is difficult to find consistent with the precipitate attempt to avoid.

I would hold, on this ground, that the respondents were not entitled to avoid the policy. The finding of the judge that the respondents acted in good faith is not a barrier to that conclusion. It was not a finding of primary fact by the judge, or a factual inference. It was a legal conclusion based on an assessment of facts which are not in dispute. The requirement of good faith, the judge in effect held, did not create an obligation to do more. While having every respect for that conclusion, it is the duty of the court to analyse and set the standard required and the court may, and in my view should, reach a different conclusion on the evidence.

A material misrepresentation will not entitle the insurer to avoid the policy unless the misrepresentation induced the making of the contract (Pan Atlantic per Lord Mustill at 549 E). No submission was addressed to the judge that the respondents were not so induced. The point first arose in the judgment, the judge stating, at paragraph 26, that “the important question for these purposes is whether the insurer was induced by the non-disclosure to accept the risk on terms that would not have otherwise been acceptable”. He added, at paragraph 27:

“I think one may legitimately ask what would have happened in this case if Dr Singh had disclosed his conviction at the time of renewal”.

The judge speculated on that question and found, in the appellants’ favour, that Dr Singh “would probably have been able to persuade Provident that it should be treated as a “no fault” accident”. In his continued speculation, the judge stated (paragraph 27) that “one simply does not know what Dr Singh’s reaction to an increase in premium would have been. He might have questioned it, or he might not. There is nothing in the evidence that points either way. Only if it could be shown that disclosure of the conviction would also have led to the disclosure of the information relating the settlement of the earlier loss and to Provident accepting the risk on the same terms could one say that the non-disclosure had not caused it to accept the risk on terms that it would not otherwise have agreed to. In the present case the evidence does not enable me to reach that conclusion”.

Thus the judge reached a conclusion adverse to the appellants on an issue that had not been mentioned until judgment. In my judgment, the judge ought not, in the circumstances, to have speculated in the way he did. I agree with Rix LJ that the manner in which he speculated is open to criticism but that defect does not open the way for this court to rely on parts of the judge’s speculation and build its own speculation upon it to reach a conclusion adverse to the respondents. While it must be shown that the misrepresentation induced the making of the contract, there was in my view an evidential burden on the appellant to demonstrate by evidence and argument the course events would have taken and it was a burden which the appellants made no attempt to satisfy.

When the judge used the expression “only if it could be shown”, he was doing no more than drawing attention to the paucity of evidence and commenting that it did not appear on the evidence that the chain of causation was broken. The respondents’ technical claims manager Mr R Shaw did refer to the inclination of customers to “shop around” if a loading is requested and added that Dr Singh “would undoubtedly have been able to get risk without loadings from another underwriter”. That the judge did not refer to the evidence is not surprising when the point that the chain of causation had been broken was not taken by the appellants. From the viewpoint of seeking a fair trial, I find unattractive the suggestion that the respondents should have called Dr Singh’s broker to deal with a point which was not taken against them. No indication had been given that it was in issue. I would not reverse the finding of the judge on causation so as to find that the respondents could not, on that basis, avoid.

I agree with the conclusion of Rix LJ. There was no effective avoidance. Writing a letter purporting to exercise a right to avoid has no effect in present circumstances when accompanied by conduct, described by Rix LJ, wholly inconsistent with avoiding the contract. Mr Shaw himself, in January 1997, believed the policy not to have been avoided. It had not been avoided before the true state of affairs emerged.

I repeat in this context my comment that the importance the respondents attached to continuing to insure Dr Singh is difficult to reconcile with their precipitate attempt, notwithstanding the duty of good faith, to avoid the contract. Their subsequent conduct at least throws light on what would have been good business practice when they learnt of the conviction.

But for the decision of this court in Legal and General AssuranceSociety Ltd v Drake Insurance Company Ltd [1992] 2 WLR 157, I would have had no difficulty in holding that the appellants were not volunteers in indemnifying Dr Singh fully. He had submitted his claim against the respondents to an arbitration which was binding upon both parties and, as between those parties, it was conclusively determined that the respondents were not liable to him. There being no other existing insurance covering Dr Singh for the liability, the appellants were required to pay the full amount of the claim. They might, however, have sought to recover a moiety from the respondents, the arbitration not being binding on the appellants.

In Legal and General Assurance Society Ltd v Drake Insurance Co Ltd [1992] 2 WLR 157, consideration was given to the predecessor of section 151 of the Road Traffic Act 1988 which imposes on insurers a duty to satisfy judgments against persons insured or secured against third party risks. Section 151(5) provides that in some circumstances the insurer must pay to the persons entitled to the benefit of a judgment relating to a liability, with respect to any matter where liability with respect to that matter is required to be covered by a policy of insurance under section 145 of the Act, certain damages. Section 151 (7) provides that an insurer who becomes liable under the section is entitled to recover from that person the excess. Lloyd LJ stated at page 164E that “it follows that, so far as the defendants are concerned, the excess over 50% was a voluntary payment”. Mr Moxon-Browne submits that the principle covers the present payment by the appellant. I have to confess that, with respect, I do find the concept that the right of recovery, which in most substantial cases will be of little, if any, value, renders the payer a volunteer to be a difficult one.

Rix LJ has demonstrated the complexity of the situation which arises. The appellants could have refused to indemnify, thus keeping Mr Beach out of his money. They would, when sued, have brought in the respondents as Part 20 defendants, and the litigation would have raised the issues considered in this action. The voluntary payment point would not then have arisen. Such conduct, would, as Rix LJ has observed, have brought the insurance industry into disrepute. The voluntary payment point would also not have arisen if the respondents had been prepared to agree a scheme for indemnity and payment without prejudice to the issues between themselves and appellants.

In my judgment, the absence of such an agreement is not fatal to the appellants’ case on this point. I agree with Rix LJ that the appellants, having made clear to the respondents their position, and acting responsibly as they did, should not be treated as volunteers. The protests of the appellants distinguish the case from Legal and General.

I would hold that in the circumstances of this case the respondents ought not be heard to rely on the statute, section 151 (7). Had the present issues between the present parties been resolved before indemnification of the insured by payment to Mr Beach, a wholly unsatisfactory sequence of events, the equitable right of contribution would, on the Court’s findings have arisen. Under protest, the appellants took the responsible course. Having declined to agree to a different procedure, the respondents should not, in the light of the court’s findings against them upon the issues of liability to Dr Singh under their policy, be permitted to defeat the equitable right of contribution by reliance on the provisions of the statute.

The sequence of events has been set out fully by Rix LJ. Payment was made by the appellants under protest. The respondents insisted on relying on the arbitration award, which between them and their insured was binding, as final and conclusive. Having adopted that stance they cannot in my view now be heard to rely on section 151 (7) to make the appellants volunteers and thereby to defect the appellants’ claim to a contribution.

In the result, I agree that the appeal should be allowed. The respondents were not entitled to avoid their policy and the appellants were not volunteers.