Skip to content

HFW successfully upholds in Court of Appeal important first instance judgment of Males J on inducement, March 2017

2 March 2017

AXA Versicherung AG (successor to Albingia) v Arab Insurance Group (B.S.C.)

HFW represented Arig in the appeal of a 7 July 2015 High Court judgment of Mr Justice Males, in which AXA, as successor in title to German reinsurer, Albingia, was unsuccessful in its attempts to avoid two first loss energy construction treaties by which the reinsurer had reinsured Arig in 1996 and 1997.

Arig successfully defended the appeal on all grounds, in what was, and remains, the leading case on inducement in the context of avoiding a re/insurance policy.

Summary of decision at first instance

By way of background, the main issue in the case at first instance was whether Albingia was entitled to avoid the two treaties for non-disclosure of or misrepresentation as to the existence of loss statistics relating to Arig’s historical book of inwards marine energy construction risks. It was common ground that such loss statistics had not been disclosed but there was a dispute as to whether they were material, given the nature of the portfolio and the unusual circumstances of the placement. There was also a dispute as to whether Albingia’s underwriter, Mr. Holzapfel, would have underwritten the treaties even if loss statistics had been disclosed.

The reinsurer’s case on misrepresentation pertained to the broker’s offer fax of 17 July 1996. This fax stated in part that “This is a new Treaty for the Reassured and as such does not have a corresponding loss record.” The reinsurer argued that this statement constituted a representation to the effect that Arig had no loss statistics for energy construction risks of the kind which would be declared to the 1996 treaty. This argument was rejected by the court because the interpretation contended for by the reinsurer ignored the effect of the words “as such” and that the statement was more readily understood as meaning that there were no loss statistics for the treaty, which as it was a new treaty, was obviously true. Thus the judge concluded that there had been no misrepresentation and the reinsurer’s case on misrepresentation therefore failed. He noted, however, that the reinsurer’s misrepresentation case “would have added little or nothing to its case based on non-disclosure [because]… the question would still arise whether Arig was under a duty to disclose its past loss statistics.”

As regards this question – that is materiality – the Court heard expert evidence from Richard Outhwaite, for Arig and Stephen Hartigan, on behalf of the reinsurer. Mr. Outhwaite told the court that energy construction risks are unlike many other types of risk (including energy operating risks) in that they are very much “one-off” projects. He also downplayed the relevance of Arig’s loss statistics in particular insofar as they related to a period when Arig’s underwriting strategy had been different. This was because, in the wake of Iraq’s invasion of Kuwait, Arig’s former head of energy underwriting became unwilling to work in the Middle East and had returned to Europe. This left Arig’s energy department without any real direction or leadership until, in mid-1991, it recruited a new underwriter who began gradually to implement a different underwriting strategy, which was more conservative than that of his predecessor. As regards the relatively few risks that had been accepted thereafter, Mr. Outhwaite suggested that there were simply too few of them to draw any meaningful inferences.

Males J., however, preferred Mr. Hartigan’s view that, as a matter of principle, historical loss statistics were material to be disclosed, albeit that the 1989 and 1990 statistics were only “on balance” material. He also rejected Arig’s argument that Albingia knew or should have known that such loss statistics would have been available and that by not asking for them Albingia had waived disclosure of them. He therefore concluded that Albingia had not been given a fair presentation of the risk. This was not, however, “game over”, as counsel for the reinsurer submitted. It remained incumbent on the reinsurer to prove that Mr. Holzapfel would not have underwritten the treaty if a fair presentation had been made.

In this regard, it was held that Albingia was not entitled to avoid the treaties because it failed to prove that on the balance of probabilities Mr. Holzapfel was induced to enter into the reinsurance contracts by the non-disclosure by Arig and/or its broker of loss statistics. In other words, based on his assessment of all the factual evidence, the judge found that, more probably than not, Mr. Holzapfel would have entered into the treaties on the same terms even if loss statistics had been disclosed.

This conclusion was based on a myriad of factors that came out in oral evidence and in various documents, which gave clues as to Mr. Holzapfel’s motivations for underwriting Arig’s first loss treaty and indeed first loss treaties in general which, as was common ground between the experts, were particularly risky propositions in principle, especially in a sub-class of marine energy business which was well-known to be particularly hazardous.

Grounds of appeal

That the reinsurer had received an unfair presentation of the risk was not appealed by Arig. The subject of the appeal, therefore, was whether it was open to Males J. to conclude that the reinsurer to whom an unfair presentation of the risk was made had failed to discharge its burden to prove, on the balance of probabilities, that the risk would have been declined if a fair presentation had been made to him.

Thus, the appeal itself was focussed on whether the judge:

  1. Applied the correct test for inducement.
  2. Had properly assessed the content of the hypothetical fair presentation, such that he was entitled to conclude that that reinsurer would have accepted the risk if it had been made.


The first ground of appeal was that the judge had applied the wrong test by deciding what could have been said if the material loss statistics had been presented, when the relevant test was what would have happened. Indeed, in the first instance judgment, the judge had used both the terms, “could” and “would”, in relation to the hypothetical fair presentation, somewhat loosely.

This question led Clarke LJ. to unpack in some detail the appropriate test for determining the content of the hypothetical fair presentation, against which the question of inducement falls to be determined. He said that there were three issues to which the “could/would” argument may pertain:

  1. One needs to determine what needed to be said in order for a presentation to be fair. In Clarke LJ’s judgment, this was said to be an objective question and, as such, falls to be assessed from the perspective of a reasonable and prudent underwriter.
  2. One then needs to decide what, if there had been a fair presentation, the insured or his broker would have said (in addition to (i) above), in order to make the presentation more attractive. It is not relevant that the broker could have said something if in fact he would not have done so.
  3. One may also have to decide what other, subjective, considerations may have affected the mind of the particular underwriter to whom the hypothetical fair presentation was made. Again, what matters is what would, rather than merely could, have influenced his decision.

In the round, however, Clarke LJ. recognised that even judges “do not always speak with precision in the use of the modal verbs to which the submissions of this point have been directed”. He also acknowledged Males J’s own recognition, after he had handed down his decision in draft, that he had intended to apply the test of what would have happened, rather than what could have happened.

The hypothetical fair presentation

The second ground of appeal was that there was no proper basis for the judge’s assessment of what the hypothetical broke would have entailed and, therefore, that the judge’s findings on inducement were wrong. It is significant to note that the reinsurer stressed, in its skeleton arguments and oral submissions, that it did not seek to appeal the primary facts found by the judge, but rather the inferences that he drew from them.

This was a difficult matter for the judge to decide at first instance because, among other things, there was a dispute then as to how far back, in a fair presentation, the loss statistics should go, and what else, beyond the bare statistics, would have been included in a fair presentation of the risk (whether negative or positive). In particular, Arig contended at first instance that if the loss statistics were material at all, notwithstanding that Arig had changed its underwriter and underwriting philosophy, it was not market practice to provide statistics for more than the previous five years. “As-if” loss statistics for 1989 and 1990, therefore, were not put to the reinsurer’s underwriter, Mr. Holzapfel, by Arig and the reinsurer was therefore precluded from doing so in re-examination.

Amongst the potential candidates for what would have been included in the hypothetical fair presentation were certain “as-if” statistics that were adduced by the reinsurer by way of voluntary particulars. These “as-if” statistics presented Arig’s loss history in an especially unpromising light and, according to Mr. Holzapfel’s witness statement, were what he would have required to see, assuming that he had not declined the risk outright based on the primary loss statistics alone.

In contrast to the reinsurer’s “as-if” statistics, Arig’s counsel put alternative “as-if” statistics to Mr. Holzapfel, which purported to show a far more favourable outlook on the basis that they excluded risks of a type that would not be ceded, in the light of Arig’s change of underwriter and underwriting philosophy.

Furthermore, Males J. put his own questions to Mr. Holzapfel, based on yet another set of loss statistics, which had been used for the purposes of an abortive earlier attempt to place similar cover, through brokers Swire Fraser.

In addition, Males J. was presented with extensive evidence, in documents and by witnesses, as regards what other considerations would have influenced Mr. Holzapfel. These included his willingness to underwrite similar treaties for other cedants, his apparent support of Arig in general, as evidenced by his participation on a separate quota share treaty, the attractiveness of the risk from a cash-flow perspective (since a number of risks could be ceded to it immediately), his willingness to listen to what the broker had to say in favour of the risk, and so on.

The judge also had to weigh up the tensions between the documentary evidence, written witness statements and oral evidence, while also taking into account that even where a witness was giving evidence in good faith (as he considered Mr. Holzapfel to be), such evidence had to be treated with caution where, as here, he had no recollection whatsoever of the relevant events. His view in this regard was influenced by the fact that Mr. Holzapfel’s evidence in another reported case, that he would have always insisted on minimum deductibles for ceded business, was contradicted by the documents in this case, which showed this to be untrue.

Despite the inherent difficulty in assessing what the hypothetical fair presentation would comprise, which is an objective question, Males J. was able to reach the conclusion that had such presentation been made, Mr. Holzapfel would have accepted the risk nonetheless. Accordingly, it fell to the Court of Appeal to analyse forensically Males J.’s findings as regards what a fair presentation would have entailed, by reference to the first-instance evidence and transcripts.

In the light of this lengthy analysis, Clarke LJ. considered that the judge “… was not disabled from reaching a conclusion as to what would be a fair presentation… by the absence of direct evidence from the broker about what he would have said or disabled from inferring what the broker would have said (in hypothetical circumstances) in the absence of such evidence.” and, having taken into account a myriad of factors that may have influenced Mr. Holzapfel, that it was open to the judge to be left in a state of doubt as to what Mr. Holzapfel would have done if there had been a different presentation, including such explanatory points as could have been made. As such, he was entitled to hold that the reinsurer had not discharged its burden to prove inducement.

Procedural unfairness

The reinsurer raised a further point, which is closely connected with its second ground of appeal above. That is, that the judge’s finding was a product of procedural unfairness since the hypothetical broke put by the judge to Mr. Holzapfel when he was giving evidence was not pleaded and was not supported by evidence from the brokers that that is what they would have presented to him. These circumstances followed from the fact that Arig had contested that the actual presentation was not unfair at all and that it had simply put the reinsurer to proof on inducement, rather than pleading any positive case itself.

The nub of the reinsurer’s argument on procedural unfairness, however, was that important parts of Mr. Holzapfel’s written evidence stood unchallenged (having not been put to him in cross-examination). In particular, Mr. Holzapfel’s witness statement said that if he had not declined the risk outright based on the primary loss statistics alone he would have insisted on “as-if” statistics in the form provided in the reinsurer’s voluntary particulars.

Clarke LJ. recorded that the reinsurer’s case on procedural unfairness was “forcefully and attractively presented”. He also acknowledged that Males J. had not expressly dealt with this point in his judgment. However, Males J. had recorded that categorical statements by Mr. Holzapfel could not safely be relied upon and, having regard that the onus of proof lay on the reinsurer, Clarke LJ., concluded that Males J. was not “bound to accept that revelation of the 1989 and 1990 statistics would have made all the difference or to accept [Mr. Holzapfel’s written evidence] at face value.” He further noted that although Males J. did not refer to the possibility of Mr. Holzapfel asking for “as-if” to the treaty statistics, this was not entirely surprising because his doubt as regards his evidence had arisen in relation to an earlier stage in the analysis, which was whether Mr. Holzapfel would have examined and made a careful assessment of the loss statistics at all.


Notwithstanding that the decision that was the subject of this appeal was decided before the Insurance Act 2015 came into force, the judgment of Clarke LJ. nevertheless sheds light on how the test for inducement may apply under the new Insurance Act 2015, which allows for remedies other than “all or nothing” avoidance. In particular, the new Act permits remedies such as additional premium and/or revised terms. Since these remedies also depend on subjective hypothetical questions as to what the underwriter would have done had he received a fair presentation, it is suggested that Courts considering inducement in future cases will, like Males J., look very carefully at the subjective considerations that would have influenced the underwriter, as well as what additional positive gloss or explanations the broker would have offered in the context of such hypothetical broke.

For the run-off market, this judgment demonstrates the risks and difficulties inherent in seeking to avoid policies in an acquired portfolio, many years after the events complained of took place, where documentary evidence may be incomplete and where key witnesses, such as the underwriters who were given an unfair risk presentation, may have scant recollection of it. It is a stark reminder that the burden of proof to show inducement falls on the re/insurer seeking to avoid and that a finding of inducement does not necessarily follow from a finding that the risk presentation was unfair. This is of course because there are all manner of considerations that may affect an underwriter’s decision, even if the risk assessed objectively, may appear to have been unattractive.

More generally, the decision also illustrates the difficulty of challenging a first-instance judge’s findings of fact, even where they were arguably mixed findings of fact and law because of the inferences drawn therefrom. Indeed, this decision underlines the reluctance of appellate Courts to second-guess a judge’s finding, based on live witness evidence, where it only has access to the transcripts, which of course cannot reveal the witness’s demeanour when giving evidence.