PROVING THINGS 260: COURT OF APPEAL JUDGMENT ON ADVERSE INFERENCES: “WE REJECT THIS NEW WAY OF PUTTING THE ADVERSE INFERENCE CASE”
I am grateful to David Platt KC for drawing my attention to the Court of Appeal decision in Alexander Johnstone v Fawcett’s Garage (Newbury) Limited [2025] EWCA Civ 467, in particular to the judgment in relation to adverse inferences. The Court of Appeal rejected the claimant’s submission that the absence of air monitoring reports meant that the trial judge should have drawn adverse inferences against the defendant.
“We reject this new way of putting the adverse inference case. It would have required the judge to make findings of fact that he did not accept on the evidence, and prevented him from making findings of fact which he thought were appropriate and sensible. We consider that Mr Platt was right to say that it was less an adverse inference argument, and more an attempt at predetermination which would have rendered a large part of the trial (the factual evidence and the oral evidence of Mr Stear and Mr Chambers) wholly otiose. That goes far beyond the principles established by the case law and would, in our view, lead to an illogical and unsound result.”
KEY PRACTICE POINT
This case highlights the limitations of the “adverse inferences” argument. It can, and will, still apply in some circumstances. Where it does apply it leads to a party’s evidence being judged “benevolently”. It will not lead to a judge overriding key findings of fact. The safest assumption is that a party still needs to prove their case.
THE CASE
The claimant’s wife died from malignant mesothelioma. He brought an action stating that her exposure had occurred when she worked for the defendant and that, as a result of her employment, there was a material increase in the risk of her developing mesothelioma. Her office had been near the garage workshop.
THE JUDGMENT AT TRIAL
The claimant’s action failed at trial The judge concluded that there was an increase of 0.1% or less in the risk of Mrs Johnstone developing mesothelioma caused by the respondent. He said that that was “such a small increase” that it did “not satisfy the test of materiality”
THE CLAIMANT’S APPEAL TO THE COURT OF APPEAL
The claimant appealed to the Court of Appeal
THE CLAIMANT’S ARGUMENT ON THE ISSUE OF ADVERSE INFERENCES
One of the grounds of appeal was that the defendant had failed to keep and disclose air monitoring records. Consequently, it was argued, the trial judge should have drawn adverse inferences against the defendant.
THE JUDGMENT ON ADVERSE INFERENCES
The Court of Appeal outlined this ground of appeal.
“Ground 5, however, is an oblique attack on those first two stages of the judge’s analysis. It is said that the judge erred in refusing to draw an adverse inference against the garage because of its failure to keep proper air monitoring records. Although not entirely clear from the grounds of appeal, it was the appellant’s case in the skeleton argument, and repeated orally, that the judge should have used the adverse inference to dismiss the expert evidence of Mr Stear, and to find as proved the evidence of Mr Chambers. In that way, therefore, because Ground 5 went to the first two stages of the judge’s analysis, Mr Snowden KC dealt with Ground 5 first. We shall follow the same course.”
Ground 5: The Adverse Inference
20.The modern law relating to adverse inferences can be traced from British Railways Board v Herrington [1972] AC 877, where a defendant’s failure to call evidence about the state of the fence meant that it could not complain if the court drew from the facts which had been disclosed “all reasonable inferences as to what are the facts that the defendant has chosen to withhold” Many of the subsequent cases, such as Wiszniewski v Central Manchester Health Authority [1998] PIQR P324 concern the silence or absence of a witness.
21. In Keefe v The Isle of Man Steam Packet Co. Limited [2010] EWCA Civ 683 the Court of Appeal held that the defendant could not assert that noise levels on its boats were not excessive when it was in breach of the duty to measure those levels. There had been contrasting evidence of fact: the claimant had given quite detailed evidence about excessive noise on the boats; a captain had said rather more generally that the boats were not that noisy. The claim failed at first instance but the appeal was allowed. That was because of this court’s “serious reservations” about the judge’s reasons for rejecting the claimant’s evidence, and the fact that the defendant had “made it difficult or impossible for a claimant to adduce relevant evidence” about noise levels, so had to run the risk of adverse factual findings. Longmore LJ said at [19] that “in such circumstances the court should judge a claimant’s evidence benevolently, and the defendant’s evidence critically”.
22. Keefe was distinguished by Lloyd-Jones J (as he then was) in Shawe-Lincoln v Neelakandan [2012] EWHC 1150 (QB), noting that the reference to a benevolent/critical approach fitted well with a situation where a defendant made it difficult or impossible for a claimant to adduce relevant evidence. More generally, Lloyd-Jones J said:
“82.
Whether it is appropriate to draw an inference at all and, if so, the precise nature and extent of such an inference will depend on the particular circumstances of each case. Relevant considerations will include the proximity between a breach of duty and the non-available evidence, the effect of the other evidence before the court and what other evidence might have been available but which is not before the court.”
23. The most recent case in this line is Mackenzie v Alcoa Manufacturing (GB) Limited [2019] EWCA Civ 2110 in which Dingemans LJ referred to the principle that whether it is appropriate to draw an inference, and if so the nature and extent of the inference, will depend on the facts of the particular case; and that a failure to adduce relevant documents may convert evidence on the other side into proof, but that may depend on the explanation given for the absence of the documents.
24. As we shall see, the basis on which the appellant sought an adverse inference here, and the precise nature of the adverse inference sought, varied during both the trial and the appeal. But it seems tolerably clear that, by the end of the trial, the appellant’s case was that, because there should have been air monitoring records in the workshop but were not, the judge should have accepted Mr Chambers’ much higher exposure figures rather than those calculated by Mr Stear. The judge rejected that submission, saying at [100]:
“100. I have also considered carefully the adverse inference point made by C, but I do not find it helpful in this case. The difference between the calculations of the occupational hygienists rested in large measure on their very different interpretation of RR’s evidence, which required judicial determination. Any measuring of levels could only have assisted in relation to RR’s exposure, but would then have been the subject of judicial determination as to what such figures represented. Any measurement would not have assisted in relation to identifying the reduction factors to be applied to EJ’s exposure, which also required judicial determination. It does not seem to me be a sustainable argument that the lack of monitoring should lead the Court to adopt uncritically Mr Chambers’ calculations in circumstances where I have concluded that they do not reflect the factual matrix I have found to exist at the relevant time.”
25. On appeal, it seemed to the court that the appellant needed to demonstrate that that passage was, in some way, wrong in principle. After all, the extent to which a trial judge accedes (or otherwise) to a submission that he or she should draw an adverse inference is well within his or her remit as the fact-finding tribunal. But, beyond saying that the judge should not have found that factual matrix, Mr Snowden did not approach his task that way. Instead, he largely repeated the argument that had failed before the judge to the effect that, if the respondent had kept the records that they were obliged to, it would have been clear just how much asbestos RR was exposed to, and therefore a much more accurate picture would have been provided of Mrs Johnstone’s exposure. He submitted that the respondent should not be able to take advantage of its own wrong in failing to keep those records, and that therefore the judge should have used Mr Chambers’ exposure figures, and not those of Mr Stear. If that submission was correct, and Mr Chambers’ figures were used, then the effect on the final stage of the calculation, namely that of the material increase in risk, would mean that the appellant’s case was proved.
26. It seems to us, for the reasons noted below, that this ambitious submission, attractively though it was argued, must fail for a variety of reasons.
27. First, the argument about the failure to keep air monitoring records in breach of Regulation 15 of The Control of Asbestos at Work Regulations 1987, was not part of the appellant’s opening at trial: that had identified a different alleged breach (of Regulation 16, in relation to health records) which has since been abandoned. It is therefore unsurprising that the issue of air monitoring records was not addressed in detail at the trial or in the judgment. It is always difficult, and often impossible, for this court to decide points which were not the subject of detailed evidence below. Here, there might have been all sort of issues that could have been explored. When was the duty to keep such records in play, relative to Mrs Johnson’s employment? What would the records have shown, particularly given the reduction in the use of asbestos in brake linings during the 1980’s? Since the records would have related to the workshop only, what was their relevance, if any, to the position in the office? How long should such records have been kept? And so on.
28.The time that such records had to be retained turned on whether the levels of asbestos went beyond an “action level”. The relevant parts of the 1987 Regulations are in these terms:
“2.—(1) In these Regulations, unless the context otherwise requires—
“action level” means one of the following cumulative exposures to asbestos over a continuous 12-week period when measured or calculated by a method approved by the Health and Safety Commission, namely—
(a)where the exposure is to asbestos consisting of or containing any crocidolite or amosite, 48 fibre-hours per millilitre of air; or
(b)where the exposure is to asbestos consisting of or containing any other types of asbestos but not crocidolite or amosite, 120 fibre-hours per millilitre of air; or
(c)where both types of exposure are concerned, a proportionate number of fibre-hours per millilitre of air;
“adequate” means adequate having regard only to the nature and degree of exposure to asbestos and “adequately” shall be construed accordingly;
“approved” means approved for the time being in writing by the Health and Safety Commission or the Health and Safety Executive as the case may be;
“asbestos” means any of the following minerals, that is to say, crocidolite, amosite, chrysotile, fibrous actinolite, fibrous anthophyllite, fibrous tremolite and any mixture containing any of those minerals;
“asbestos area” and “respirator zone” shall be construed in accordance with regulation 14;
“control limit” means one of the following concentrations of asbestos in the atmosphere when measured or calculated by a method approved by the Health and Safety Commission, namely—
(a)for asbestos consisting of or containing any crocidolite or amosite—
(i)0.2 fibres per millilitre of air averaged over any continuous period of 4 hours,
(ii)0.6 fibres per millilitre of air averaged over any continuous period of 10 minutes;
(b)for asbestos consisting of or containing other types of asbestos but not crocidolite or amosite—
(i)0.5 fibres per millilitre of air averaged over any continuous period of 4 hours,
(ii)1.5 fibres per millilitre of air averaged over any continuous period of 10 minutes;
“the Executive” means the Health and Safety Executive.
…
15.—(2) The employer shall keep a suitable record of any monitoring carried out in accordance with paragraph (1) and that record or a suitable summary thereof shall be kept—
(a)in a case where exposure is such that a health record is required to be kept under regulation 16, for at least 30 years [amended to 40 years by paragraph 6 of the Schedule to the Control of Asbestos at Work (Amendment) Regulations 1992];
(b)in any other case, for at least five years.”
29.There was no evidence that, for example, RR was exposed to a level that exceeded the action level. In those circumstances, there is nothing to say that the records should have been kept for 30 years (later amended to 40 years). Moreover, in our view, this cannot be met by the argument that, because there were no records, the appellant does not know if the action levels were exceeded: there was evidence from RR about his exposure and the experts on both sides were able to calculate that exposure in figures. If it was said that the action levels were exceeded, that case needed to be pleaded and proved by way of factual and expert evidence. It was not.
30. Secondly, on the basis that it has not been shown that RR was exposed to a level beyond the action level, any air monitoring records would only have been kept for 5 years. In those circumstances, since Mrs Johnstone left her employment at the garage in 1989, there would have been no records relevant to her employment that would and should have been kept by the garage beyond 1994/1995. So any such documents would have been destroyed more than two decades before the trial.
31. To get round this obstacle, the appellant’s argument has to be that, since RR left the garage in full time employment in 1995, the 30 year rule (later amended to the 40 year rule) was triggered in his case, and that therefore there would have been documents relating to him at the trial. On that basis, therefore, the adverse inference is said to arise for a period when Mrs Johnstone was no longer employed by the garage, arising out of hypothetical records in respect of a third party (RR). That seems to us to be too tenuous an argument on which to base the draconian adverse inference now sought.
32. Thirdly, the argument about the records is very limited in time. The 1987 Regulations came into force on 1 March 1988. That was towards the end of Mrs Johnstone’s employment at the garage: she left between 12 and 18 months thereafter. That doubtless explains why, although this limit had not been put clearly to the judge, on appeal Mr Snowden was careful to refer only to the last 12-18 months of Mrs Johnstone’s employment. In that way, the obligation in respect of records only existed for 20% of the overall period of her employment. It is impossible to see how that would again justify the drastic consequences for which the appellant now contends. This is particularly because the evidence showed that, as the 1980’s wore on, asbestos-lined brakes were replaced by disc brakes, which in all probability would have reduced RR’s (and therefore Mrs Johnstone’s) exposure to asbestos anyway.
33.Fourthly, precisely what adverse inference he should draw – how it would work and what effect it would have – was not made clear to the judge. It was not explained in the appellant’s lengthy closing submissions at trial. As set out at paragraph 25 above, the judge understood that he was being asked generally to use the adverse inference to prefer the appellant’s expert evidence over that of the garage. He declined to do that for the reasons I have cited. In my view, no criticism can be made of that passage in the judgment.
34. Fifthly, we consider that the argument in the present case seeks to put the judge into a straitjacket, in that it seems to require him to make findings of fact which were contrary to the facts as he found them to be. That was apparent from Mr Snowden’s submissions in reply, when he put the point in a new way, arguing that what the judge should have done was to take RR’s witness statement at face value, and not subject it to any kind of scrutiny or interpretation at all. This would have meant that, for example, when RR said that he worked with asbestos one day a week, the judge would have been obliged to find that he worked with asbestos for 8 hours solidly on that day. The judge had of course rejected that, finding instead that it would only have been a percentage of that day in which he was exposed to asbestos dust. In this way, the inference sought runs completely counter to the findings of fact made by the judge.
35. We reject this new way of putting the adverse inference case. It would have required the judge to make findings of fact that he did not accept on the evidence, and prevented him from making findings of fact which he thought were appropriate and sensible. We consider that Mr Platt was right to say that it was less an adverse inference argument, and more an attempt at predetermination which would have rendered a large part of the trial (the factual evidence and the oral evidence of Mr Stear and Mr Chambers) wholly otiose. That goes far beyond the principles established by the case law and would, in our view, lead to an illogical and unsound result.
36. The expert evidence of Mr Stear, which the judge preferred, was (as the judge rightly noted at [85]) closely aligned to the judge’s own findings of fact. It would have been wrong in principle, and contrary to common sense, for the judge to have made those detailed findings, only then to ignore them and accept the calculations put forward by Mr Chambers which, as the judge said, were calculated on the wrong factual assumptions.
37.We do not suggest that adverse inferences can never be applied to expert evidence. But for the reasons that we have given, the adverse inference submission in this case makes a leap of logic which is unsustainable. The judge was obliged to do his best on the material that he had and to calculate the exposure to RR, and therefore the exposure to Mrs Johnstone. He did that on the basis of a number of findings of fact. Mr Chambers’ calculations were far removed from those findings; Mr Stear’s calculations were much closer. In those circumstances, the adverse inference principle cannot extend to an effective deletion of those findings of fact and an uncritical acceptance of Mr Chambers’ calculations.
38. Finally, we note that, in any event, this is not a case in which there were two different versions of the facts put forward by the opposing parties, with the judge having to choose between them, and where one side had been hampered by the other’s failure to adduce oral or written evidence. It is in that situation – as occurred in Keefe – where the adverse inference principle is of most relevance. It is in that situation where the court can look benevolently on the claimant’s evidence and critically on the defendant’s evidence. Here, there was no factual evidence from the garage at all so, unlike in Keefe, there was no conflicting evidence of fact on which the court could look critically. The main evidence of fact came from RR, the appellant’s own witness. What the judge had to do was to translate RR’s general evidence into realistic findings as to his daily/weekly exposure to asbestos. That is what he did.
39. For all those reasons therefore, we reject Ground 5 of the appeal. That means that the remainder of this judgment is concerned only with the last stage of the judge’s analysis, namely the method he adopted to establish that the respondent’s breach of duty did not cause Mrs Johnstone to develop mesothelioma.