The Commercial Court has refused to allow a reinsurer to avoid two treaties because the reinsured’s failure to disclose past loss statistics did not induce it to enter the treaties.


Axa Versicherung AG (“Axa”) sought to avoid two “first loss” facultative/obligatory reinsurance treaties entered into with Arab Insurance Group (“Arig”) (the “Treaties”) and recover sums paid to Arig.

The first treaty covered the first US$500,000 of losses for any one accident or occurrence on Arig’s book of inwards marine energy construction risks attaching between 1 January 1996 and 30 June 1997 (the “First Treaty”). The second treaty was a renewal of the First Treaty and covered risks attaching between 1 July 1997 and 30 June 1998 (the “Second Treaty”).

Axa sought to avoid the Treaties on the basis that Arig did not disclose, and misrepresented that there were no, loss statistics relating to its book of inwards marine energy construction risks written between 1989 and 1995. Axa also sought to avoid the Second Treaty for Arig’s alleged non-disclosure of its claims experience under the First Treaty.



The proposal which Arig’s broker sent to Axa stated “This is a new Treaty for the Reassured and as such does not have a corresponding loss record.” Axa alleged this was a representation that there were no loss statistics for energy construction risks of the kind which would be declared to the First Treaty. Arig in fact had such statistics. The judge accepted Arig’s argument that the statement simply meant “This is a new Treaty for the Reassured and as a new treaty does not have a corresponding loss record” (emphasis in original).

Materiality of the past loss statistics

Arig adduced evidence that the loss statistics were not material for three reasons. First, energy construction risks are unique and past risks are actuarially irrelevant to later underwriting decisions. Secondly, Arig had hired a new underwriter who had adopted a more conservative underwriting strategy. Thirdly, the new underwriter had underwritten too few risks for insurers to draw conclusions.

The judge rejected these arguments. The evidence of Axa’s expert that past loss experience was relevant accorded with commercial sense. Moreover, the broking experts said past loss records would always be part of the placement of reinsurance, and one of Arig’s previous brokers had presented the risk on this basis. Finally, Arig’s new underwriter had written a memo which stated that (a) Arig expected claims to arise after the first half of the year which it would not be able to recover unless the relevant policies were ceded to the First Treaty; and (b) it was difficult to find cover for the risks. These reasons were “overwhelming” evidence of the materiality of the loss statistics.


The judge rejected Arig’s submission that Axa had waived the requirement to disclose past loss statistics by failing to request them. Axa did not and had no reason to know that Arig had written energy construction risks in the past.


Mr Holzapfel, the head of treaty insurance at Axa at the relevant time, gave evidence that if he had seen the past loss statistics, he probably would have declined to write the risk irrespective of Arig’s explanation that it had changed underwriter.

The judge concluded that Axa had not established on the balance of probabilities that it would not have written the risk on the same terms, or at all, had the loss statistics been disclosed:

  1. Mr Holzapfel’s evidence about whether he would have written the risk was necessarily hypothetical. Although he was an honest witness, such evidence should be evaluated with “a healthy scepticism”, especially where (as here) the underwriter has no recollection of the transaction.
  2. Axa wrote the risk notwithstanding the absence of any information on deductibles, although this had been described as absolutely “essential” on another transaction. This demonstrated “the danger of focusing on an isolated fact ([…] in this case loss records)” when considering inducement.
  3. Axa had a significant line on a quota share treaty with Arig. Mr Holzapfel considered Arig to be a high quality reinsured which he was keen to support.
  4. The underwriter would have listened to and considered the broker’s explanation of the poor loss records.
  5. The broker could have explained that the poor results were largely on risks written by a previous underwriter who adopted a different underwriting approach.
  6. The losses occurred in years characterised by bad market conditions.
  7. Mr Holzapfel had written a number of first loss reinsurance treaties on energy construction, notwithstanding that other underwriters were unwilling to do so.

In the circumstances, there was “real doubt” regarding what Mr Holzapfel would have done if there had been a fair presentation of the risk.

Non-disclosure and misrepresentation – claims experience

Axa sought to avoid the Second Treaty on the additional ground that Arig failed to disclose, or made misrepresentations in respect of, its claims experience in relation to three incidents. Although Arig materially misrepresented that one claim had been made in relation to an incident (there were two), Mr Holzapfel’s evidence was that “[Axa] might have renewed”, without suggesting that Axa would have required different terms. The judge accordingly found there was no inducement. The other two incidents were not material in themselves. The judge held that fair presentation of the three incidents in combination would not have induced the underwriter to write the policy on different terms. Axa was therefore unable to avoid the Second Treaty on this basis.


While the case turned on the factual issue of inducement, a number of interesting points arose:

  1. There was potentially, but not ultimately, an issue as to the identity of the underwriter in circumstances where one candidate had failed to give evidence. The judge opined that there was no rule of law that a presumption of inducement could only come into play if the underwriter failed to appear for good reason.
  2. More generally on inducement, the judge noted that an underwriter’s evidence needed to be treated with caution given its hypothetical nature lended itself to exaggeration and embellishment in the interests of the party on whose behalf it was given, and where the remedy for breach of the duty of good faith was avoidance. That was all the more so where, as here, the underwriter had no recollection of the risk and was reconstructing what he would have done.
  3. Had Axa been entitled to avoid, the issue of limitation would have needed to be addressed. There is no direct authority in English law on when, or if at all, a limitation period starts to run for an avoidance – it being generally thought that there is no limitation period as avoidance is a self-help remedy. A limitation issue would also have arisen in respect of the recovery of claims previously paid by Axa – whether there was no limitation period or whether a six year period applied as a starting point. Unfortunately, and aside from reciting the arguments, the judge declined the opportunity to opine on these issues.
  4. As regards the reinsured’s own claims, there is an issue in English law as to when the limitation period starts to run. There is a chain of reinsurance authority finding reinsurance to be a type of insurance of the underlying risk – which should mean time starts to run on the date of the underlying property loss or establishment of the insured’s liability. There is a separate chain of authority finding reinsurance to be a type of liability insurance, which should mean time starts to run when the reinsured’s liability to its insured is established. There is however no direct authority on when the limitation period starts to run.

Both parties accepted without argument that time would run from the date when the reinsured’s liability was established. That was based on obiter comments of the first instance judge inTeal Assurance Co Ltd v W R Berkley Insurance (Europe) Ltd[2011] Lloyd’s Rep IR 285, which comments perhaps somewhat inconsistently analysed reinsurance as an insurance of the underlying risk but with time starting to run from the establishment of the reinsured’s own liability.

  1. Finally, there was a subsidiary issue as to whether one particular risk had been properly ceded. That turned on “NLOW” (“Neither limitation nor warranty”) being written beside the potential term, which the judge found to result in an (accurate) statement of intent, rather than a term of the reinsurance.