Jakto Transport Ltd. v Derek Hall

This is an appeal against the decision of His Honour Judge Robinson, sitting at the Sheffield County Court on 17^th February 2005, in which he found the defendant, Jakto Transport Ltd, the appellant in these proceedings, liable to compensate the claimant/respondent, Derek Hall, for personal injuries sustained in an accident on 28^th September 2000, in the course of his employment as a HGV driver with the defendant.

The respondent alleged that the he had suffered injury while using a torque wrench to tighten wheel nuts on the vehicle he drove for the appellant. He contended that his accident had been caused by a breach of Regulation 5 of the Provision and Use of Work Equipment Regulations 1998, which provides:

“Every employer shall ensure that work equipment is maintained in an efficient state, in efficient working order and in good repair.”

At the trial, it was not disputed that this regulation imposes an absolute duty on an employer who provides equipment for use by his employee. Nor was it disputed that the respondent had had an accident while using a torque wrench. The dispute arose as to the how the accident had happened. The respondent contended that the torque wrench malfunctioned while he was using it in a proper manner. The malfunction must have been caused by a defect (which the respondent could not identify), for which the appellant was liable under Regulation 5. The appellant contended that the wrench had not malfunctioned; it was not defective and the accident must have happened in some other way. It contended that the most probable explanation for the accident was that the respondent himself had caused it by the manner in which he had used the wrench – in other words by operator error. The respondent denied that. It was common ground (and the judge so directed himself) that the burden lay on the respondent to prove, on the balance of probabilities, that the accident had happened in the way that he alleged and was due to a hidden defect in the wrench.

The judge accepted the respondent’s evidence as to the way in which the accident had happened. He found that the wrench had had a hidden defect, the nature of which could not be ascertained. He held that the defect constituted a breach of Regulation 5 and that the respondent’s injury had been caused thereby. He found that the respondent had not contributed to the accident in any way. Following the judge’s finding on liability, the parties agreed damages in the sum of £67,500. The appellant now contends that the judge’s findings were wrong.

In the course of his employment as a HGV driver, the respondent often had to change wagon wheels using a torque wrench provided by the appellant. He was an experienced HGV driver and had been using tools similar to this torque wrench throughout his career. He had worked for the appellant for three years at the time of this accident and was thoroughly familiar with this particular wrench which had been in use throughout that time.

The torque wrench and the manner of its operation are illustrated in a series of photographs to be found in the Court bundle. The wrench was about 43 inches long. At its head was a socket which fitted onto the wheel nut. Within the head was a ratchet mechanism. To tighten a nut, the operator would fit the socket onto the nut and then lever the handle downwards to apply torque. Then he would raise the handle (the ratchet allowed this to happen without undoing the nut) and repeat the process until the required tightness was achieved. In order to prevent over-tightening, the wrench was set to achieve a required degree of torque. When that was reached, the handle would ‘operate’ or ‘break’ to an angle of about 20 or 25 degrees at a hinge part way along its length. The hinge was spring loaded and, after operation, the handle would return to its usual straight position.

It was not disputed that the respondent had an accident while using the torque wrench and that he sustained injuries to the lower right leg including fractures of the tibia, fibula and the fibular and posterior malleoli and associated soft tissue injury. The agreed medical reports did not throw any light upon the way in which the accident had happened.

The respondent said that when the accident occurred, he had successfully tightened 7, 8 or 9 of the 10 nuts of the wheel he was changing. The nut on which he was working was towards the top of the wheel. He claimed that, as he applied downward pressure to the handle of the wrench, the handle suddenly gave way and ‘just went straight down’ from about the 3 o’clock position to about the 5 o’clock position. He said that the wrench struck his leg and he fell over.

He was in severe pain and could not get up. He was lying on the floor with his head close to the wheel. His first thought had been the nut had broken but that was not so. The wrench was still attached to the wheel. The handle was pointing downwards in the 5 o’clock position. Colleagues came to help him and an ambulance was summoned. There was some conversation while he was on the floor and he said that his accident had been caused by the wrench giving way. The respondent said that, because there was a danger that the wrench might become detached from the nut, one of the colleagues took it off the wheel. Later, when he was being put in the ambulance, Mr Robinson, a tyre fitter, put the wrench back onto one of the nuts and was using or testing it.

The respondent claimed that the giving way or slip of the wrench must have been due to a hidden defect. He suggested a way in which a defect might have arisen. Some years previously, he had been employed by the National Coal Board and had used similar wrenches. He said that he had known of occasions when an intermittent fault of this kind had been caused by the penetration of dust into the ratchet mechanism. The wrench would give way and yet afterwards there would appear to have been nothing wrong with it. It was put to him that the expert engineer whose evidence was before the judge had said that this was not a ‘reasonable’ explanation for the accident because the environment at the appellant’s premises was not dusty. The respondent said that it was quite dusty in the yard where his accident had happened. This was quite near a crusher plant. However, he agreed that it was nothing like as dusty as a coal mine.

It was put to the respondent that the accident had not happened in the way he claimed. Two propositions were put to him. It was suggested that there had been no abnormal slippage and the wrench had simply ‘broken’ at the hinge as it was intended to do when the appropriate torque had been reached. He had fallen over because he had been putting too much pressure on the handle. The respondent was adamant that this was not so. He said that he had not yet reached the stage in the tightening process when the handle would break. He could tell when the handle was about to break because the pressure required on it increased. Any experienced operator familiar with the wrench could tell when it was about to break. In any event, he was accustomed to the movement that occurred when the handle broke; the movement that caused his accident was different from that.

It was also suggested to the respondent that he had been using his foot to press down on the handle of the wrench and that his foot had slipped and caused him to lose his balance. The respondent denied that he had done any such thing.

The respondent’s credibility was challenged on several grounds. It was put to him that he had given different accounts of the accident when he went to the hospital. Reliance was placed on the hospital notes. Also, it was put to him that his account of how the handle came to strike his lower leg was impossible. In due course, the respondent appeared to accept that he might have been wrong about that; he had thought that the handle had struck his lower leg but admitted that the pain he experienced might have resulted simply from falling over.

There was no eye witness to the accident besides the respondent himself. The appellant called four witnesses. These were Messrs Bettinson, Hellewell, Robinson and Ramsden. The first three arrived on the scene, very shortly after the accident occurred. Mr Ramsden, the chargehand fitter, arrived a little later, shortly after the respondent had been taken to hospital.

Mr Bettinson, the Plant Manager, was the first to arrive. He agreed that the wrench was still attached to the wheel nut but said that it was not pointing downwards in the 5 o’clock position; it was horizontal in the 3 o’clock position. Moreover, he did not agree with the respondent that the wrench was removed from the nut for safety reasons. Mr Bettinson said that although the wrench was heavy and was a danger, it was left on because the appellant’s staff needed to know where the respondent had been up to in the process of tightening the nuts. He said also that it was left on so that the wrench could be checked. He said that Mr Robinson used the wrench to check the tightness of the nuts before Mr Ramsden arrived. He denied that the respondent had said anything about the wrench having been the cause of his accident.

Mr Bettinson was asked about the possibility of dust entering the ratchet. He said that the wrench would have to be immersed in dust for that to happen. He agreed that the respondent’s accident had happened in the yard (as opposed to the workshop) and that the yard could sometimes be quite dusty in the summer. He thought it had not been dusty in that September. He was of the view that the only possible explanation for the accident was that the respondent had been standing on the wrench, in effect operating it with his foot.

Mr Hellewell was the next to arrive. He is a director of the company and, in effect, the office manager. He agreed with Mr Bettinson that the wrench was still on the nut; he said that it was in a horizontal position and was not removed. He said that it was not removed because “nobody was that bothered about touching it and moving it because we did not know at that stage what had gone on what had happened and we didn’t want to disturb it in case there was a fault with it or for whatever reason”. He said that he did not think that the wrench presented any danger to the respondent lying on the floor. “It was not that heavy”. He said that the respondent said about his accident only that he had felt a pain in his leg and went down; the next thing he had been on the floor. In other words he had not blamed the wrench. Mr Hellewell claimed that it was not until 3 or 4 months after the accident that the respondent had made any mention of the wrench being the cause of his accident. The appellant’s contention was that the respondent had had no idea what had caused his accident and had deduced at some later time that it must have been the wrench that had caused it.

Mr Robinson was the third to arrive. He said that the wrench was still on the nut. He did not touch it. In so saying, he disagreed with Mr Bettinson. No one touched it, he said, until Mr Ramsden, the chargehand fitter, arrived after the respondent had been taken to hospital. While the respondent was on the floor, Mr Robinson asked him what had happened but he said that he could not remember what the respondent said in reply. He said that later he saw Mr Ramsden use the wrench to finish tightening the nut, that the wrench was on and then he checked several more and the wrench worked normally. He said that, in twenty years of experience, he had never known anyone fall and break a leg while using a wrench. The only explanation he could think of for the accident was that the respondent had been ‘standing on it’. However, he had never seen any one doing that. Nor was there any need to do so, as the amount of force needed to operate the wrench by hand in the usual way was not great.

Mr Ramsden said that he heard there had been an accident and went to the scene. The respondent had already been taken to hospital. Somebody told him, at that time, that it was thought that the accident had been caused by the wrench slipping. He could not now say who. He found that the wrench was on the wheel; the handle was roughly horizontal and he used it to make sure that all the nuts on the wheel were properly tightened. He found no problem with the wrench. He took it off the wheel and examined the ratchet, which was working normally. The wrench was kept in service and was used for about another three years when it became defective and was removed from use. He was of the view that it would be possible for an operator to fall and injure himself if he pressed down too hard on the handle.

A consultant engineer, Mr Graham Glenn, had been instructed as a joint expert. He produced one report and answered a number of questions by way of clarification. He did not give oral evidence.

He examined the wrench externally and tested its operation at an inspection which took place some 10 months after the accident. He found nothing wrong with it. He was of the opinion that a failure such as the respondent described would be expected to make the wrench unfit for use afterwards. He envisaged that such a failure might be caused by the breakage of one of the teeth of the ratchet or breakage of the ratchet pawl but such failure would not be self- rectifying.

However, at paragraph 6.04 of his report, he wrote:

“There is another possible explanation for the failure, but even that possibility seems to me to be unlikely.”

He then discussed the respondent’s suggested explanation for the accident, which was that the ratchet had malfunctioned due to the presence of dust. He confirmed that, if used in a dusty atmosphere, dust can accumulate in the ratchet mechanism and can hold one or more of the teeth out of engagement. He said that the shock of a sudden movement might disturb some of the dust so that the cause of the failure would not be apparent immediately afterwards. In other words, the sudden movement might correct the defect and allow the tooth or teeth to re-engage. It is not disputed that Mr Glenn was here accepting that there was a possible explanation in engineering terms for the way in which the respondent said the wrench had behaved. However, Mr Glenn went on to express the view that dust in the ratchet was not a ‘reasonable’ explanation for the accident. Torque wrenches of this type are commonly used in garages and engineering workshops and he was not aware of problems caused by dust in such places. In any event, he said, the kind of routine lubrication that Mr Ramsden had told him was carried out on this wrench should be sufficient to keep the ratchet clean and in good order. Moreover, if the working environment was very dusty, this problem might be expected to occur on more than one occasion. There was no history that it had.

His conclusion at the end of the report was that he could find no explanation for the accident.

In a letter dated 15^th August 2003, Mr Glenn responded to questions asked by the respondent’s solicitor. One question related to possible explanations for the accident. He said this:

“In my opinion there are two possible causes of the accident if the wrench remained on the nut. My view is that either there was a defect in the wrench causing an unexpected and uncontrolled movement or the wrench acted as it ought to and Mr Hall simply misjudged the point at which the wrench would operate.

The use of the torque wrench is to limit the torque applied to the nut. This is achieved by the use of the spring within the tool. When the set torque is applied, the handle will move relative to the head of the tool, indicating to the operator that the nut is correctly tightened. The movement is in the direction of the applied force, so the operator’s hands will inevitably move with the handle in the direction in which the operator is pushing. My understanding is that the movement that occurred at the material time was similar to the correct operation of the tool when the required torque is achieved. The handle moved downwards in accordance with the force applied by Mr Hall.

I understand that Mr Hall was an experienced fitter and that he was familiar with the use of the torque wrench. He told me that the movement of the wrench at the time of his accident was not like the movement that normally occurred. It was more sudden and involved a more substantial release under the applied force. His impression at the time was that the bolt had broken. My experience of using a torque wrench is that there is a difference between normal operation and the movement that will occur when a bolt breaks.

If it is accepted that there was a different type of movement at the material time to that normally expected when the required torque is achieved then, in my view, the only explanation is that there was a defect in the torque wrench. Such circumstances could be a breach of statutory duty as discussed in my report. However, if the torque wrench seen at the inspection is the same as was used at the material time, then I have difficulty in explaining what the defect might have been.

If there was no defect in the torque wrench at the material time then, in my view, the balance of probabilities is that Mr Hall misjudged the effort applied to the wrench and the stage at which the wrench would operate. Under those circumstances, Mr Hall’s description that the movement was as though a bolt had broken must be mistaken.

The description by Mr Hall is not consistent with my understanding of the Defendants’ evidence. If the wrench failed as described by Mr Hall, then I would expect the type of investigation described by Mr Ramsden to indicate a fault, even if there was insufficient means to accurately diagnose the fault. I find it difficult to see how a wrench that behaved as described by Mr Hall could remain in regular use for a period of several months without repair.

In my opinion, this case depends on whose evidence is accepted.”

In answer to yet further questions, Mr Glenn wrote:

“Normal operation of a torque wrench does not require the operator to judge the point at which the wrench will operate. An inexperienced man or a man carrying out unfamiliar work will have no indication that the torque is approaching the preset limit and so cannot prepare himself for operation. Such circumstances are not normally considered to be unsafe. However, over- enthusiastic use of any hand tool can lead to a loss of control. Providing Mr Hall operated the torque wrench in a calm and controlled manner, he should not have encountered any difficulty whether or not he expected the wrench to operate when it did.

I understand that Mr Hall was experienced in his work and familiar with the use of a torque wrench in the manner that it was used at the material time. Under such circumstances, Mr Hall may have been able to anticipate roughly when the torque wrench would operate. If he misjudged the operation, then he may have continued to apply force to the wrench with more enthusiasm than was appropriate, losing control of his movements when the wrench operated.”

The judge said that he had to decide which of the three possible mechanisms had caused the accident. Had the respondent’s foot slipped off the wrench? If so, he was the author of his own misfortune. Had the wrench made a sudden and unexpected movement? Or had the wrench simply ‘broken’ or ‘operated’ - as Mr Glenn called it? Was the cause of the accident an unexplained fault with the wrench or was it operator error?

The judge said that he found Mr Ramsden to be a truthful and convincing witness. However, he found the appellant’s other three witnesses to be unsatisfactory. He explained his reasons and, in this appeal, it is accepted that he was entitled to form that view of them. He said that he found the respondent honest and straightforward. He had watched him in the witness box for an hour and had observed him under searching cross examination. He said that he was ‘inclined’ to believe him. On the basis of that assessment of the lay witnesses and without making any reference to the expert evidence, he made some ‘preliminary’ findings of fact. He found that the respondent was not standing on the wrench at the time of the accident. That disposed of that issue. He found that the wrench did not break or operate at the moment of the accident. After the wrench had moved suddenly, in the manner described by the respondent, the handle was in about the 5 o’clock position and not at 3 o’clock. The claimant had said words to the effect that the wrench had given way on him. Someone removed the wrench from the wheel while the respondent was on the ground and someone had replaced it at about the 3 o’clock position before Mr Ramsden arrived at the scene.

The judge then turned to consider the evidence of Mr Glenn, saying that he would examine whether any of his preliminary findings of fact required alteration in the light of that evidence and the submissions of counsel. He summarised Mr Glenn’s evidence, noting that he could find no defect in the wrench. Mr Glenn had considered that the ‘dust’ theory did not provide a ‘reasonable’ explanation for the accident. He considered that either there was a defect in the wrench or the wrench acted as it ought and the respondent had misjudged the point at which it would operate. Mr Glenn had said that he found it difficult to see how a wrench that behaved as described by the respondent could remain in regular use for several months without repair. The judge acknowledged that Mr Murphy, counsel for the appellant, had ‘fairly observed’ that, faced with a choice between the two options, operator error was ‘more likely in the sense of being most probable’. He said that because there was no medical evidence relating to the manner in which the injuries had been caused, he was not prepared to accept the submission that the injuries could not have been caused as the respondent claimed by the handle striking his lower leg. It seems that the judge thought it likely that the respondent had stumbled into the handle rather than the handle striking his leg. But he thought that this did not mean that the accident had not happened as the claimant had described. Finally, he considered whether he should revise his preliminary findings and decided that he should not. Mr Glenn had not said that the accident could not have happened as the respondent said. The judge said that he believed the respondent and he found that the accident had happened as he alleged and was due to an unexplained defect in the wrench. Liability followed.

In this appeal, Mr Murphy contended that the judge had approached his findings of fact in the wrong way. He relied on a passage from the speech of Goff LJ, as he then was, in The Ocean Frost [1985] 1 Lloyds Law Reports 1 at page 57, where he said that it was important when testing the credibility of a witness to test his veracity by reference to the objective facts proved independently of the witness’s testimony. That is plainly right and a well-established approach to judicial fact-finding. In the present case, however, there were very few objective facts proved independently of the witnesses. There were some agreed facts; the wrench was still on the wheel after the accident and the respondent was on the floor injured. There was no medical evidence as to the mechanism by which the injuries had occurred as the result of a fall or as the result of the wrench coming into forcible contact with the respondent’s leg or a combination of the two. Mr Murphy’s complaint was that the judge had considered the respondent’s credibility in isolation from the expert evidence rather than in the light of it. In that respect, in my view, Mr Murphy’s submission has considerable force. The judge reached preliminary or provisional conclusions as to what had happened before he considered the evidence of Mr Glenn. Then, when he did consider it, he did so only to ascertain whether it caused him to change his mind. In my judgment, that was an error of approach.

Where, as here, the expert evidence was relevant to the way in which the accident could or might have happened, it was incumbent upon the judge to consider it at the time when he was reaching his conclusions on the credibility of the witnesses. That correct approach was recently confirmed by this Court in Mibanga v Secretary of State for the Home Department [2005] EWCA Civ 367 where Wilson J, giving the first judgment of the Court, with which Ward and Buxton LJJ agreed said:

“It seems to me to be axiomatic that a fact-finder must not reach his or her conclusion before surveying all the evidence relevant thereto. …. Mr Tam, for the Secretary of State, argues that decisions as to the credibility of an account are to be taken by the judicial fact-finder and that, in their reports, experts, whether in relation to medical matters or to in-country circumstances, cannot usurp the fact-finder’s function in assessing credibility. I agree. What, however, they can offer, is a factual context in which it may be necessary for the fact-finder to survey the allegations placed before him; and such context may prove a crucial aid to the decision whether to not to accept the truth of them. What the fact-finder does at his peril is to reach a conclusion by reference only to the appellant’s evidence and then, if it be negative, to ask whether the conclusion should be shifted by the expert evidence.”

That was an asylum case and the expert evidence related to the conditions in the country to which the asylum seeker contended he could not safely be returned. The circumstances were very different from the instant case but the principle is the same and the judge fell into a similar error as did the adjudicator in Mibanga.

In the present case, in my judgment, the judge should not have found the facts, even provisionally, without considering the evidence of Mr Glenn. Having said that, it does not necessarily follow that the judge reached the wrong conclusion. Because the judge’s approach was wrong, either the case must be remitted for a re-hearing or the evidence must be re-examined by this Court. In my judgment, the appropriate course is for this Court to re-examine the evidence, giving due weight to the judge’s impressions of the honesty of the witnesses and to reach its own conclusion as to how the accident probably happened. As the expert evidence was all in written form, this Court is as well placed as was the judge to assess its import.

I have already summarised Mr Glenn’s evidence and have quote a lengthy passage from the letter in which he addressed the question of how the accident might have happened. Mr Murphy submitted that the effect of Mr Glenn’s evidence was that operator error was a much more likely explanation than a defect in the wrench and that there was, in truth, no other feasible explanation; Mr Glenn had, to all intents and purposes, said that the dust theory could not provide an explanation. Therefore, one was left with the overwhelming likelihood that the accident had happened as the appellant contended.

Mr Williams for the respondent submitted that that was not the effect of Mr Glenn’s evidence. Mr Glenn had been careful not to usurp the function of the judge. He had said that, if the wrench had behaved as the respondent claimed it had, there must have been a defect which caused it to slip. The defect must then have corrected itself, as none was found afterwards. He could not think of any likely explanation for such a course of events. He accepted that the dust theory was a possible explanation but he thought it unlikely because the environment was not sufficiently dusty and there had been no other occasions on which the ratchet had malfunctioned. If the wrench had not malfunctioned, the most likely explanation was, in his view, over-enthusiastic use by the respondent. He said that that could happen but he did not express any view as to how likely it was to have happened. Mr Glenn’s summary of the situation was that the case depended upon what findings of fact were made. The judge was therefore left to make his findings of fact and Mr Glenn’s evidence did not point to one explanation being significantly more likely than the other.

Mr Williams pointed to the fact that there was no evidence that any other driver or tyre fitter at the appellant’s premises had ever fallen over during the normal operation of the wrench. If that explanation was much more likely than a fault in the wrench, as the appellant contended, one might have thought that it would have happened before. The evidence from Mr Robinson was that it had not and Mr Murphy accepted that there was no evidence of previous incidents. Moreover, there was evidence from Mr Glenn that the wrench was regarded as a safe piece of equipment even for inexperienced operators who would not recognise when it was about to ‘break’. Mr Williams reminded the Court that Mr Bettinson and Mr Robinson (both men with long experience of using wrenches) had not favoured the explanation advanced by the appellant; they thought that the only feasible explanation was that the respondent had been ‘standing’ on the handle. The evidence as a whole suggested that both of the explanations under active consideration by the judge were very unlikely to have happened. The judge had not found that one was more likely than the other although he had said that Mr Murphy’s submission to that effect was ‘fair’. However, one could not say that one explanation was much more likely than the other.

I, for my part, accept Mr Williams’ analysis of the effect of Mr Glenn’s evidence and his submission that its effect was that the outcome of the case depended upon what facts the judge found. Mr Glenn did not say which explanation was the more likely. He could explain the accident as being due to operator error. The only possible engineering explanation if the accident was due to a defect was the dust theory but he thought this was unlikely. However, he did not consider the likelihood of the accident being due to operator error. He said that it was possible. He did not consider the absence of any history of previous accidents due to operator error or the significance of that history. In effect, he confined himself very properly to matters of engineering expertise, implicitly recognising that it was a judicial function to assess which was the more likely explanation for the accident.

In my judgment, the effect of Mr Glenn’s evidence, taken together with the absence of any previous incidents due to operator error, was that there were two possible explanations for the accident, both of which were very unlikely to occur. To my mind, there was no clear indication from this evidence that one explanation was much more likely than the other.

On that analysis of the expert and ‘historical’ evidence, the judge’s impressions of the honesty of the lay witnesses becomes crucial. It is open to the fact-finder to reach a conclusion based upon the impressions made by the lay witnesses. I adopt the impressions of the judge, as he had the advantage of seeing them in the witness box. The judge said that he regarded the respondent as honest and straightforward and accepted his account.

The judge rejected the evidence of the respondent’s witnesses in respect of the post-accident events. In particular, he accepted that the respondent had immediately attributed his accident to some sort of slippage of the wrench. The appellant’s case, advanced through Mr Helliwell, was that the respondent had first blamed the wrench some months after the accident. Yet, Mr Ramsden’s evidence was that someone told him when he arrived on the scene that it was thought that the accident had been caused by slippage of the wrench. The judge concluded that the respondent had indeed said words to that effect to the people who came to his aid. In my judgment, that was a pure finding of fact, which there is no reason to reconsider.

Further, the appellant had argued (and the point was pursued at the appeal hearing) that the medical records suggested that the respondent did not know what had caused his accident and that he had given inconsistent accounts to the medical staff at hospital. This affected his credibility. The judge expressed the view that there was nothing in the medical notes inconsistent with his claim that he was blaming the wrench from the start. I agree with that conclusion. The first note, apparently made by a triage nurse, records:

“Right lower leg injury at work today. Using a torque wrench something hit him at force just above the boot.”

The second note records:

“Something flew out of torque wrench 1hr ago hit Right lower leg Too painful to weight bear since.”

The third note, written by a senior house officer records that he had been tightening wheel nuts on a lorry using a torque wrench. The doctor has then written:

“?what happened. Wrench exploded something hit him on right leg.”

In my judgment, there was nothing in the notes that cast real doubt on the veracity of the respondent’s claim that he had been injured by the wrench and that he believed that it or something had struck him on the leg. There is in fact support for the respondent’s contention that he had blamed the wrench at the time and not three or four months later.

Mr Murphy had another submission by which he sought to undermine the respondent’s credibility. At the trial, the appellant had challenged the respondent’s claim that the wrench had struck his lower leg. It was put to the respondent that, if he had been using the wrench in the manner shown in the photographs in the court bundle, (namely facing the end of the wrench handle and with his right shoulder towards the vehicle), and if the wrench had given way as he claimed, the wrench could not have struck him on the leg. It was put to him that he might well have lost his balance and stumbled forwards but not because the wrench struck his leg. In cross-examination, the respondent appeared to accept these propositions. It appears to me that, if the respondent was standing sideways on to the vehicle, the wrench could not have struck his lower leg. However, if he was facing the side of the vehicle or was at an angle to the side of the vehicle, it seems to me that it would have been possible for the wrench to strike his lower leg. The respondent’s exact position while using the wrench was not closely examined. Mr Murphy suggested that the hospital notes did not support the respondent’s claim that he had been struck on the leg. In my view, they clearly do support that claim. It appears that the judge thought it more likely that the respondent had stumbled against the wrench than that the wrench had struck his leg. He concluded that this did not affect his view of how the wrench had behaved. In my view, the evidence as to whether the wrench struck the respondent’s leg is very finely balanced. However, it does not appear to me that the issue is crucial. If the wrench did strike his leg, that would tend to support his claim that the wrench had made a sudden and uncovenanted movement due to a defect. But if the wrench did not strike his leg, and the injury was caused when he stumbled forwards on to the wrench and fell over, this would throw no light, one way or the other, on whether the accident was due to a defect or to operator error. In my view, the possibility that the respondent might have been wrong when he said that the wrench struck his lower leg does not undermine his credibility. Given that he suddenly found himself on the floor and in great pain, it is wholly understandable that he should not be entirely clear about how he came to be there.

The judge also found that, after the accident, the handle of the wrench was pointing downwards in the 5 o’clock position. He rejected the appellant’s case that it had been horizontal. Mr Murphy accepted that he could not challenge this finding of primary fact. During argument, it was put to Mr Murphy that this finding was strongly supportive of the appellant’s contention that the wrench had made an uncovenanted movement. If the accident had been caused by the respondent pushing too hard on the handle, one would not have expected it to finish up in such an abnormal position. It appeared to me, at least, that the appellant’s argument in the court below had been that the handle had been found in the horizontal position and that that was a strong indication that the wrench had not behaved abnormally. The judge’s finding had gone against the appellant and he had drawn the inference that the wrench had behaved abnormally. However, Mr Murphy contended that the finding that the handle was in the 5 o’clock position was equally consistent with the respondent’s theory of an uncovenanted movement as with the appellant’s theory of operator error. It seems to me that this represented a significant shift in the way in which the appellant’s case was being put. I cannot accept Mr Murphy’s submission. It appears to me that, if the handle was found at the 5 o’clock position, the likelihood is that there was an uncovenanted movement. It seems to me most unlikely that an experienced operator could, by over-enthusiasm, push the handle down so far against the resistance within the wrench.

After reviewing all the evidence, I have come to the conclusion that the appellant did prove, on the balance of probabilities that the accident was caused by a sudden and unexpected and abnormal downward movement of wrench, which caused him to lose his balance and fall forwards. As to whether or not the wrench struck the appellant’s leg, I would make no finding, as it is not a matter of importance. I would find that that the accident was probably due to a defect within the ratchet, probably caused by dust or grit. Although Mr Glenn had said that his was an unlikely explanation for the accident, it appears to me to be more likely than the alternative theory of operator error. I would hold that this respondent would have been very unlikely to overestimate the force necessary for the proper operation of this wrench to such an extent as to push the handle down far enough to make him lose his balance. The fact that there was no history of other operator making such errors supports that conclusion.

In summary, I accept that the judge did not approach the finding of facts as he should have done. He should have considered all the evidence, including that of the expert Mr Glenn before making any findings of fact, even provisional ones. However, on a proper analysis, the evidence of Mr Glenn, taken together with other relevant evidence about the absence of any accidents due to operator error, suggested that both potential explanations for the accident were unlikely, although both were possible. On a fresh analysis of all the evidence, giving due weight to the judge’s impression of the witnesses, I am of the view that the accident was caused by a defect in the wrench for which the appellant was strictly liable under Regulation 5. I would dismiss the appeal.

I have not found this an easy case for the reasons given by Pill LJ. However, in the end I am persuaded by Smith LJ's careful analysis of the judgment and evidence that the judge's overall conclusion can be supported. I therefore agree that the appeal should be dismissed.

I gratefully adopt Smith LJ’s statement of the facts and of the issues which arise.

I agree with Smith LJ that the judge did not approach the fact finding exercise as he should have done. I respectfully agree with the statement of principle by Wilson J, giving the first judgment in this court in Mibanga v Secretary of State for the Home Department [2005] EWCA Civ 367, cited by Smith LJ at paragraph 29.

I also agree with Smith LJ that there was a burden on the respondent to establish, on a balance of probabilities, a defect in the torque wrench. This is not a case, as the judge recognised, where the happening of the accident itself gives rise to an inference of breach of duty.

The claimant was an experienced workman doing a simple and routine task which he had done hundreds of time before. He was using a tool well suited to the task. The particular tool had been used without difficulty for months before the accident, when tightening other bolts immediately before the accident, when examined by Mr Ramsden, the chargehand fitter, shortly after the accident, and for months after the accident. This is a case where a court has to consider which of two unlikely possibilities is the less unlikely; a defect in the tool or a mistake by the operator.

I have serious doubts as to whether the claimant has established a defect in the tool:

(a)

On the basis of Mr Glenn’s evidence and that of the chargehand fitter, whom the judge found to be truthful and convincing, that the wrench operated satisfactorily immediately after the accident, and for months afterwards, the possibility of a defective function when the accident happened is very small.

(b)

Quite apart from his credibility, the reliability of the respondent’s account is suspect. He himself, in cross-examination as to how the accident happened, agreed that he thought he “fell over handle or whatever” rather than the account he had given and, as Smith LJ has stated, it seems that the judge thought it likely that the respondent had stumbled into the handle rather than the handle striking his leg. This must cast doubt on the reliability of his recollection.

(c)

While I would be slow to decide a case on the basis of what a claimant tells doctors after the accident, it is surprising in this case that the respondent did not refer to a fault in the wrench when asked the questions he was.

(d)

The claim he made in evidence that he said at the time that there was a fault was not made even in his witness statement.

(e)

I can place no reliance in the respondent’s favour on the acceptance by Mr Ramsden that somebody at the scene of the accident mentioned the possibility of a fault in the wrench. Given the respondent’s experience, it would be surprising if his workmates who arrived at the scene did not consider that possibility. Mr Ramsden found no fault or problem with the wrench.

(f)

I would accept Mr Murphy’s submission on behalf of the appellants that the position in which the judge found the wrench to have been after the accident is as consistent with over-enthusiastic use by the operator as with a defect.

(g)

The judge’s reservations about other evidence given by the appellants’ witnesses does not appear to me to affect the issue.

If I were deciding this case on paper (but with the benefit of the transcript) I would find for the appellants on the basis that a defect in the torque wrench had not, on a balance of probabilities, been established. However, in a case such as this, and notwithstanding the trial judge’s erroneous approach to the question of the credibility of the respondent, respect for the judge’s impression on credibility is due. The judge decided to hear evidence-in-chief orally, a good decision in my respectful view. The respondent was in the witness box for an hour. The judge found him to be “honest and straightforward” and was “inclined to believe [him]”. While I am a little surprised that there is no reference to his reliability which, on a consideration of alleged over-enthusiastic use, was the real issue, I am prepared to assume the judge had it in mind. After hesitation, I am not prepared to overrule the finding of the trial judge that there was an “unexplained mechanical defect”. I agree that the appeal should be dismissed.