Swift v Wearing-Smith

1. [1]

   MEAGHER JA: I agree with Hoeben JA that this appeal should be allowed. For the following reasons and those given by his Honour, I consider the primary judge erred in his findings as to breach and as to the mechanism of failure of the balustrade. What follows assumes a familiarity with the factual background and evidence at trial as summarised by Hoeben JA.

2. [2]

   The primary judge held that in the circumstances which followed their receipt of the 2002 building inspection report, the exercise of reasonable care by the appellant home owners required that they take “advice and action to ensure the structural soundness of the balustrade before allowing guests to be in close proximity” to it: Judgment at [156], [159]. His Honour concluded that the fact that work had been carried out in 2006 by a painter “without structural qualifications of any kind” and “without any consideration of the state of the component parts of the supporting structures” ought to have raised in the minds of the appellants the question whether the rust identified in 2002 had been “adequately addressed from a structural and safety viewpoint”: Judgment at [153], [171]-175].

3. [3]

   In so concluding his Honour reasoned as follows: in 2002 the appellants were advised of the presence of rust in the steel components of the balustrade; that rust problem remained unaddressed until 2006 when Mr Mesker chipped scaled rust from the poles and lugs, applied rust converter, and then painted those structures; at that time the appellants took no active steps to satisfy themselves of the structural safety of the balustrade and its supporting components, and could not reasonably have assumed that the work undertaken by Mr Mesker, a painter, had the effect of ensuring the structural soundness of the areas he had treated or that he would have identified any structural problems in the balustrade if they existed. In particular, he was not in a position to determine whether the mounting bolts holding the glass in place were corroded: see Judgment at [145]-[162], [171]-[175].

4. [4]

   The mounting bolts holding the glass panel in place assumed particular significance, because the primary judge found that it was one or more of those four bolts which had failed due to corrosion: Judgment at [130], [131].

5. [5]

   In my view the primary judge’s conclusion that the appellants, acting reasonably, would have appreciated as a result of the 2002 inspection report that there was an underlying problem of rust in the steel components of the balustrade which remained unaddressed after the works undertaken by Mr Mesker in 2006, was not justified by the evidence or a consideration of what a reasonable person in the appellants’ position would have done in response to that report.

6. [6]

   The report described the “metal lugs and posts” of the balustrade as having “some corrosion” and recommended “rust proofing and repainting” or “simply replacing”. It did not, in terms or implicitly, recommend that the appellants also seek further advice from someone (other than a painter or rust proofer) as to the structural soundness of the component parts of the balustrade. The description of the corrosion was not such as to indicate that there was a need for such advice. By way of contrast, express advice to that effect was given in other parts of the report concerning items or sources of damage in respect of which it was considered that there should be an inspection by the council, a qualified engineer or a tradesperson with particular qualifications or experience. Nor did the report say that the balustrade was not or may not have been structurally sound at that time, or that it might become unsafe even if the work recommended was undertaken.

7. [7]

   In those circumstances there is no apparent basis for the primary judge’s conclusions that “in 2002, and thereafter, a reasonable person in the position of the [appellants] would have taken precautions by taking advice and action to ensure the structural soundness of the balustrade” beyond what was recommended by the report and that the appellants had been advised of the need for the balcony to be “appropriately treated”, it being suggested that the appropriate treatment extended beyond what was described in the report: Judgment at [156], [173]. It is correct, as the primary judge records at [157], that from 2002 the appellants were on notice that “proper remedial treatment of rusted components” was recommended. However that treatment did not extend beyond “rust proofing and repainting”.

8. [8]

   In 2006 the appellants engaged Mr Mesker to undertake work which answered the description of that which was recommended in the inspection report, albeit some three years after the recommendation was made. It was not suggested that he was not qualified to undertake the work described, or that he said as much to the appellants. On the contrary, his Specification and Tender dated 5 March 2006 described the rust proofing and repainting that he proposed to undertake in terms which might have given the appellants some confidence as to his ability to carry out the work properly.

9. [9]

   The primary judge considered that it was not sufficient for the appellants to rely on Mr Mesker to identify any structural or safety issues arising from the condition of the balustrade: Judgment at [148]. That observation is not to the point and, in any event, is contradicted at least to some extent by Mr Burn’s evidence. Mr Mesker was retained to undertake the work described in the inspection report. Mr Burn considered that a prudent painter in his position, who considered that a more thorough structural assessment was required, was likely to have raised the matter with the home owner. From the appellants’ perspective they might reasonably have proceeded on the basis that if the painter had observed something which he believed required further assessment or work, he would have raised that fact with them.

10. [10]

    That did not occur. Mr Mesker, who the primary judge accepted as a witness of truth, gave the following evidence:

11. [11]

    There was no reason following the completion of Mr Mesker’s work, for the appellants, acting reasonably, to believe that any further structural assessment of the balustrade was required. To the extent that a defect had been identified, that defect had been addressed in one of the ways recommended and the tradesperson who had been engaged to that end did not suggest that he thought that any further work was required to make the structure safe.

12. [12]

    I also agree with Hoeben JA’s reasoning and conclusion that the primary judge erred in finding the mechanism of failure of the glass panel was as suggested by Mr Burn (being the finding referred to in [4] above). Finally, I agree with Hoeben JA that the respondent’s application to rely upon his notice of contention should be refused.

13. [13]

    HOEBEN JA:

14. [14]

    The house, where the accident occurred, was built in 1952 and the appellants moved into the premises in February 2003. It was their place of residence. The balcony was about three metres above ground level and had glass panels around it. The balcony was constructed before 1992. The accident occurred when a glass panel gave way and the respondent fell to the ground.

15. [15]

    The respondent brought proceedings against the appellants in the District Court alleging that they were negligent in allowing him access to the balcony when they knew that the supports for the balcony were dangerously corroded, failed to warn him of that fact, failed to inspect the balcony and failed to rectify the corrosion of the metal lugs and posts. The proceedings were heard by Levy SC DCJ (the primary judge) on 11 – 13 August 2014, with judgment delivered on 9 October 2014. The respondent was successful in his claim.

16. [16]

    The appellants have appealed against the primary judge’s finding of negligence against them.

17. [17]

    The appellants commissioned a pre-purchase report in 2002 before buying the premises. The report was received in December 2002 and signed by a master builder. Of its 28 pages, 20 related to the house and pool, and the balance was a pest inspection report. The format of the report was to identify various aspects of the structures on the property and to comment on them, e.g. roof external/roof internal, external walls, footings, foundations, balcony, handrails, pool, each room.

18. [18]

    By way of an executive summary, there were two pages headed “Issues” and “Safety Concerns”. Under the heading “Issues” were the following matters – extensive termite damage, with a recommendation that this be treated, concrete cancer in the slab of the garage, with a recommendation that this be inspected, corrosion to the base of the water heater requiring its replacement, failure of the pool to comply with the safety code, with a recommendation for a Council inspection, damp in the wall of the rumpus room, with a recommendation for a damp course, covered edges of the concrete slab, with a recommendation that this be remedied.

19. [19]

    Under the heading “Safety Concerns” the following matters were listed. The pool fence was decayed and did not comply with the Council code. The electrical wiring was faulty, with a recommendation that a licensed electrical contractor inspect and remedy faulty or illegal electrical wiring. The paths were identified as constituting a slip hazard. The water heater had been incorrectly installed so that the control valve might malfunction and injury occur. There were no safety glass stickers on glass panes below one metre, e.g. on shower screens, windows, doors, built-in wardrobe doors there should be safety glass. The comment was “this is considered to be serious, have a glazier inspect the glass and upgrade as soon as practicable” (Blue 9K). Floor surfaces were identified as possibly being “slippery” when wet, with a recommendation that anti-slip material be placed on them.

20. [20]

    In another part of the report, the following comment and recommendations were made in relation to the balcony and handrails:

21. [21]

    Ms Swift, gave evidence that she would have read the pre-purchase building report at the time she and her husband were purchasing the property, but she did not specifically recall its content. She recalled retaining painters to do work on the house, approximately three years after they moved in. She said:

22. [22]

    In cross-examination, her evidence concerning the painting in 2006 was:

23. [23]

    Ms Swift gave evidence that in about 2010 more work was done on the property.

24. [24]

    Ms Swift said that she was otherwise unaware of any issue with the glass balustrade on the deck. She and her husband would frequently entertain on the deck, having family gatherings, people over for barbeques and a couple of parties, including New Years Eve parties.

25. [25]

    The person who carried out the painting in 2006 was Mr A Mesker. The quote which he provided for the work was dated 5 March 2006 and he appears to have signed it on 10 March 2006. Included in the quote was the following:

26. [26]

    Following the accident, the balustrade was replaced. Ms Swift’s evidence was:

27. [27]

    Mr Mesker gave evidence that he recalled doing painting work at the premises in early 2006. His evidence was:

28. [28]

    The first appellant, Mr Swift, gave evidence that he remembered reading the pre-purchase report and that he made some notes on the front of the document. He said that the arrangements to paint the balustrade and other parts of the property were made between his wife and the painter and he played no part in them. He said that after 2006 when the balustrade was painted, he did not notice any staining on the poles.

29. [29]

    Mr Livingstone was a fencing salesman, who attended the premises in 2010, to inspect the pool fence and quote for its replacement. At the same time he inspected the balustrade. He was asked to give an estimate of the potential cost to replace the existing tinted glass of the balustrade with clear glass. In a statement dated 8 July 2014 he set out his observation of the glass panels, which formed the balustrade:

30. [30]

    He set out his observations of the structure of the balustrade as follows:

31. [31]

    In his oral evidence, Mr Livingstone said in relation to the posts and flanges:

32. [32]

    I infer from that evidence (which was the only direct evidence on the subject) that each steel post had flanges (or lugs) welded to them. On the side of each post, attached to the welded flanges by bolts, were sheets of glass which were sandwiched between the two flanges at each point. This is consistent with the blurred photograph of the balustrade before the accident (Blue 52).

33. [33]

    Mr Livingstone explained what happened as a result of his inspection of the balustrade in 2010 as follows:

34. [34]

    In oral evidence, Mr Livingstone said that in 2010 the glass on the balustrade appeared “okay” to him. He tested the posts of the balustrade, but not the glass (Black 74L – T). The advice which he gave to Ms Swift was:

35. [35]

    Within a week following the accident, Mr Livingstone was asked to attend at the premises by Ms Swift and to provide a quote. His evidence in his statement as to what he observed on that occasion was:

36. [36]

    In oral evidence he said:

37. [37]

    When the evidence as to the bolts was tested under cross-examination, Mr Livingstone gave the following evidence:

38. [38]

    Mr Burn, a civil engineer, gave evidence on behalf of the respondent. His evidence comprised two reports, dated 25 November 2013 and 8 July 2014, together with oral evidence. While objection was initially taken to some parts of the reports, they eventually were received by the court without objection.

39. [39]

    When preparing his first report, Mr Burn did not perform a site inspection, did not inspect the component parts of the balustrade and was not provided with any photographs. He did have available the pre-purchase report obtained by the appellants. From that he noted “metal lugs and posts used to secure the handrails had visible corrosion with recommendation to rust proof and repaint and/or replace the fixings”. For the purposes of his report, he made an assumption that the handrail fastenings did not receive any attention between the date of the report and the date of the injury.

40. [40]

    On the basis of that material, and by reference to other instances which he had observed of the failure of a balcony handrail, Mr Burn concluded that inadequate handrail maintenance had caused the accident. He opined that what should have happened was for there to be a periodic inspection of handrails to determine if there were any inherent movement, the degree of movement and whether the handrail could be re-secured by tightening the existing fixtures. If the fixtures had corroded, they would have to be properly inspected, which would involve disassembling the fixture.

41. [41]

    In between the first report and the second report, Mr Burn became aware that painting of the balustrade had taken place in March of 2006. He assumed that no other maintenance had taken place between February 2003 and March 2006. On the assumption that painting of the balustrade had taken place, Mr Burn said:

42. [42]

    Mr Burn made the following observations concerning the painter:

43. [43]

    Mr Burn expressed an opinion as to the possible effect on the glass of the balustrade when somebody forcefully came in contact with it.

44. [44]

    Before giving his evidence, Mr Burn had the opportunity of reading the statement from Mr Livingstone of 8 July 2014. With the benefit of that information, he said:

45. [45]

    Mr Burn expressed an opinion as to the likely progression of any corrosion of the bolts on the balustrade as follows:

46. [46]

    Mr Burn estimated the likely load produced by the respondent leaning against the balustrade would be between 1 and 50 kgs (Black 62F). He described the likely mechanism of failure as follows:

47. [47]

    In cross-examination Mr Burn gave the following evidence on the assumption that the bolts had not failed.

48. [48]

    Mr Burn explained the basis for his opinion as follows:

49. [49]

    Mr Burn set out his experience in relation to how glass fractures as follows:

50. [50]

    Mr Burn gave evidence as to his understanding of how the glass on the balustrade may have reacted if it fractured:

51. [51]

    Mr Burn gave evidence about other modes of failure, apart from the bolts.

52. [52]

    Much of the evidence at trial related to how the accident occurred. The purpose of adducing that evidence was to determine how forcefully the respondent came in contact with the balustrade. The effect of the evidence, however, can be briefly stated.

53. [53]

    The respondent had no recollection of the mechanism by which he came to fall from the rear balcony. He had no clear recollection of how far from the balustrade he was seated, but thought it was about a metre to a metre and a half. He remembered reaching out with his left hand towards the balustrade but otherwise had no recollection of what happened. He had no recollection of putting his hand on the glass (Black 14).

54. [54]

    The respondent’s wife gave the following evidence:

55. [55]

    Mr Stobo was sitting near the respondent. It was his evidence that the respondent stood up and “seemed to get his foot caught perhaps, on the chair.” He thought that something made the respondent stumble. He observed the respondent to take a step and reach out with his left arm. He did not see the respondent fall.

56. [56]

    Mr Stobo thought that there was a railing on top of the balustrade and that the respondent was reaching for it. He was clearly wrong in that regard. It was common ground that there was no railing on top of the balustrade, just glass panels. Mr Stobo saw some glass where the respondent fell. He was not able to say whether there was glass on the balcony because he did not specifically look.

57. [57]

    Ms Swift did not see the respondent fall. She was not asked whether she saw any glass on the balcony. She did, however, see that one of the glass panels had disappeared. She observed glass where the respondent had fallen.

58. [58]

    Mr Swift gave evidence that he saw the respondent stand up, his left leg give way completely causing him to stumble to his left and crash into one of the glass panels with great force. He said that the respondent was heavier than he was at the trial because of his illness. He described the respondent as crashing through the glass panel. He said that he did not see the respondent put his hand out.

59. [59]

    When asked about the glass, he said:

60. [60]

    Mr Dobson gave evidence that the respondent got up, probably caught his foot on the chair or the table, stumbled towards the glass balustrade and hit it with his right shoulder and went straight through it, right in the middle of the glass. He described the respondent as hitting the glass hard. He did not see the respondent put his hand out.

61. [61]

    In relation to glass, his evidence was:

62. [62]

    The issue between the parties was the force with which the respondent came in contact with the glass panel.

63. [63]

    The primary judge made a number of credit findings. He accepted that the respondent, his wife and Mr Stobo were all witnesses of truth. He found that Mr Livingstone had an unreliable recollection of key events and that his evidence included elements of unwarranted assumption and reconstruction and also involved inaccuracies. His Honour concluded that caution should be exercised before his evidence could be accepted.

64. [64]

    He found Mr Swift to be an unreliable witness, who was not careful or considered in his responses. He found that Mr Dobson’s recollection had been affected by his ingestion of alcohol and that he was an unreliable witness because his evidence involved reconstruction and was in part based upon assumptions. The primary judge accepted Ms Swift as a witness of truth.

65. [65]

    The primary judge accepted Mr Mesker as a witness of truth. He concluded that because Mr Mesker did not have the skills of a structural engineer, some of his observations and conclusions had to be qualified by that fact. The primary judge found that no credit issues arose from Mr Burn’s evidence “which stood to be evaluated according to its content”.

66. [66]

    The primary judge made the following findings in relation to the premises. There was no maintenance carried out to the “corroded metal lugs and posts” between the purchase of the property in February 2003 and the painting carried out by Mr Mesker in March 2006. There was no evidence of the state of the securing bolts at the time that Mr Mesker treated and painted the posts and lugs on the balustrade. There was no evidence of any further maintenance inspection or repairs to the balustrade between 2006 and the date of the accident “… at the time of the accident, the underlying metal structure was not as thick, and its components were not as integral, compared to when it had been initially installed”. The extent to which the application of the rust converting compound and surface paint by Mr Mesker’s company in 2006 was successful in arresting the rusting process, especially the shafts of the bolts, was uncertain.

67. [67]

    When considering how the accident occurred, the primary judge observed that the description of what Mr Dobson saw was remarkably similar to that given by Mr Swift. He found that their evidence was “tailored” to exaggerate the force of the respondent’s contact with the glass panel. He preferred the evidence of Mr Stobo to that of Messrs Dobson and Swift.

68. [68]

    The primary judge found that the mechanism of the accident was that one or more of the four bolts, holding the glass panel in place at its base, failed due to corrosion having weakened those structures, as explained by Mr Burn in his evidence. He found Mr Burn’s expert explanation to be cogent and plausible and to provide a rational cause for why the panel gave way. He found that Mr Burn’s explanation combined a common sense understanding of the progression of the rusting process with his expert engineering understanding of the way in which the material in question was affected by that process. The primary judge inferred that the rusting process had continued from the time of acquisition of the house, so as to weaken the bolts to the point where they sheared or gave way when the respondent leaned upon the balustrade.

69. [69]

    The primary judge found:

70. [70]

    The primary judge articulated the relevant duty as:

71. [71]

    By reference to s 5B of the Civil Liability Act 2002 (CLA) the primary judge found that the number of guests in close proximity to the balustrade (which had earlier been the subject of a recommendation for replacement and which had not been the subject of an inspection for structural integrity in respect of any of its component parts) made it reasonably foreseeable that there was a risk of harm from a failure of the balustrade (Red 52F – J).

72. [72]

    The primary judge identified the risk of harm as:

73. [73]

    The primary judge found that because the appellants had not completely replaced the balustrade in 2010, reasonable maintenance required that care be taken in the form of dealing not only with surface rust, but also ensuring that the structural components were either not affected by rust or if they were, that those components were properly treated or replaced so as to maintain safety (Red 53E – H).

74. [74]

    The primary judge found that the appellants were in breach of the duty which they owed to the respondent in failing to obtain any structural advice, especially in the absence of the structure being disassembled for rust proofing and painting purposes.

75. [75]

    Put another way, the primary judge found:

76. [76]

    The primary judge found that while the posts and lugs on the balustrade may have been treated for rust by Mr Mesker in 2006, the shafts of the bolts were affected by corrosion at the time of the respondent’s fall because they had not been treated for rust, either in 2006 or in the intervening period. The primary judge found that a reasonable person in the position of the appellants would have taken precautions by taking advice and action to ensure the structural soundness of the balustrade before allowing guests to be in close proximity to the structure. His Honour reached that conclusion because the appellants were on notice that replacement of the structure, or proper remedial treatment of rusted components had been recommended as necessary.

77. [77]

    The primary judge found that an appropriate response to such a risk of harm was to ensure that all components of the supporting structure, which were likely to be subject to attack from rust, were adequately treated in such a way as to address the problem of corrosion or, absent any inquiry by the appellants along those lines, replacement of the structure. His Honour found that the relative burden on the appellants to take such precautions was not a great one, although it would have involved some cost.

78. [78]

    In summary, his Honour found that the exercise of reasonable care required that the appellants consider and respond to the risk of failure of the balustrade before allowing guests near it. His Honour found that the important difference between this case and other cases where homeowners had escaped liability was that the appellants in this case had notice of an underlying problem of rust in the steel components of the balustrade, as was identified in 2002 in the pre-purchase inspection report. His Honour found that the appellants were advised of the presence of rust and of the need for it to be appropriately treated or for the structure to be replaced, yet they took no active steps to satisfy themselves as to the structural safety of the balustrade and its supporting components in that there was no investigation into the safety of the structure.

79. [79]

    The primary judge found that the appellants could not have reasonably assumed that the work carried out by Mr Mesker as a painter had the effect of ensuring structural soundness of the areas which he had chipped free of rust, nor could they reasonably assume that Mr Mesker as a painter who had not dismantled the balustrade would have identified structural problems in it if they existed. His Honour found that the appellants failed to consider or recognise that corrosion of the supporting components of the balustrade that had first been identified in 2002 had weakened its structural soundness. He found that the appellants had failed to take reasonable and available safety precautions in conformity with the duty of care they owed to the respondent. Their breach of duty involved the failure to ensure the structural soundness of the balustrade and the failure to isolate the area from guest access.

80. [80]

    The primary judge noted that there was no real issue as to causation and that if the bolts which held the glass panel in place had not corroded, they would not have failed, thereby causing the glass panel to give way and the respondent to fall. He found that the failure of the bolts was a necessary condition for the harm to occur and that factual causation, pursuant to s 5D(1)(a) CLA had been established.

81. [81]

    The appellants relied upon the following grounds of appeal:

82. [82]

    The respondent relied upon a Notice of Contention as follows:

83. [83]

    There was no challenge to the evidence of Ms Swift that there was only surface rust on the posts and lugs in 2006 and no evidence of staining after painting (at [22] – [23] hereof). That evidence was supported by that of Mr Mesker as to the technique which he used to remove rust and to convert any rust affected surface to inert steel (at [26], [28] hereof). It was supported by the observations of Mr Livingstone when he inspected the balustrade in 2010 (at [31] hereof and Blue 60G). It was supported by the evidence of Mr Swift. The observations of Ms Swift were consistent with the pre-purchase report which referred to “the metal lugs and posts have some corrosion …”. That description did not suggest extensive corrosion or rust on those components.

84. [84]

    In other words, there was no evidence that the metal posts and lugs on the balustrade were extensively or dangerously rusted before Mr Mesker’s work in 2006. After that date, the only evidence was that there was no evidence of observable rust or corrosion. There was no evidence whatsoever as to the state of the bolts connecting the glass to the lugs, either before or after Mr Mesker performed his work. The pre-purchase report made no mention of them.

85. [85]

    The primary judge rejected in its entirety the evidence of Messrs Swift and Dobson on credit grounds. Rather than taking into account that their observations as to the movement of the respondent were similar, the primary judge found that they had deliberately “tailored” their evidence so as to exaggerate the force with which he came in contact with the glass panel. That was not a finding open to his Honour. This was a very serious adverse finding which required compelling evidence to justify it. A suggestion of concoction was never put to either witness, nor was such an issue raised by the parties at the trial. Moreover, the evidence of Mr Dobson was corroborated in an important respect by Mr Stobo in that he observed the respondent to trip, or catch his leg, when he got to his feet immediately before the accident.

86. [86]

    It was open to the primary judge to qualify the evidence of Messrs Swift and Dobson on demeanour grounds on the basis that they appeared to him to be describing what they saw in rather dramatic terms and to be exaggerating the force with which the respondent came in contact with the glass panel. That, however, did not involve the complete rejection of their observations. This is particularly so when the respondent had no recollection of how he came to fall and neither his wife nor Mr Stobo saw the fall. Given that Mr Stobo saw the respondent catch his leg or trip, which was consistent with the observation of Mr Dobson and to a lesser extent that of Mr Swift, it is clear that the force involved must have been substantially more than the respondent merely leaning on the glass panel.

87. [87]

    The primary judge was not entitled to reject the evidence of Mr Dobson on the basis that his observations were adversely affected by his ingestion of alcohol. While Mr Dobson volunteered that he had consumed some alcohol before the fall, it was never put to him that his ingestion of alcohol prevented him from making an accurate observation of how the respondent came to fall.

88. [88]

    The primary judge’s substantial rejection of the evidence of Mr Livingstone on credit grounds does not withstand scrutiny. The only issue on which it could be said that Mr Livingstone’s evidence might be unreliable was whether the “bolts were in order” when he made his observations of the balustrade within days of the accident having occurred. When that evidence was tested, it became clear that he did not have a recollection of actually having seen the bolts in position at that time.

89. [89]

    He did, however, explain the earlier evidence by reference to the configuration of the flanges or lugs on each post. It was his evidence as an eye-witness, not as an expert, that at each point where the glass was bolted to a post there was a flange on either side so that the glass was sandwiched between each flange. What is not altogether clear from his evidence was whether both flanges were welded onto the post or whether one of them was a floating flange, i.e. one which would be locked into position by the bolt. In his statement he said that each flange was fixed but in cross-examination his evidence was that there was one fixed and one floating flange (Black 80T, 81E). In re-examination he again said that both flanges were fixed (Black 81T). He explained his evidence by pointing out that because all of the flanges were in place on the post, the only reasonable inference was that the bolts were also in place. That explanation, however, was only available if there was one fixed and one floating flange in each corner. In that context, it is significant that Mr Livingstone never said that he actually saw the bolts.

90. [90]

    His Honour does not appear to have understood the effect of that evidence. In the circumstances, his Honour was not justified in dismissing the evidence of Mr Livinstone’s observations on the basis that they were “not based on any formal qualifications of an engineering kind” (Red 42J). This was eyewitness evidence of a person who had been selling pool fences and balustrades for 10 years. Such a person would know the significance of the presence of flanges or lugs and the function they performed. No engineering expertise was needed for that.

91. [91]

    Mr Livingstone said that he thought that he had prepared his first statement in about 2011. He was mistaken in that he had prepared it in 2013. With that as his only basis, his Honour characterised Mr Livingstone as a person with a faulty memory. This, with great respect to the primary judge, was also neither fair nor accurate.

92. [92]

    It was on the basis of that characterisation of his memory and his Honour’s apparent failure to understand Mr Livingstone’s evidence as to the flanges, that his Honour observed:

93. [93]

    It was not accurate to say that Mr Livingstone was mistaken about an important detail. At no point did he say that he actually saw the bolts. What he did say was that he observed the flanges/lugs to be in place and from that he inferred the bolts were also “in order”. Of course, that inference was only available to Mr Livingstone if one of the flanges in each corner was floating.

94. [94]

    His Honour also rejected the eyewitness evidence of Mr Livingstone on the basis that he had conducted many other inspections of pool and other fences between the date of the accident and when he prepared his statements and gave his evidence. Such a finding does not automatically follow from those facts. This is particularly so when Mr Livingstone’s unchallenged evidence was that he was interested to know what had caused the glass panel to fail and that is why he had carried out the particular inspection on his own initiative (Black 72M – N).

95. [95]

    It follows that there was no proper basis for the primary judge to reject the evidence of Mr Livingstone as to what he saw shortly after the accident and as to what he observed of the condition of the balustrade when making recommendations as to its replacement in 2010. This was the occasion when he was asked by Ms Swift to quote for the cost of replacing the tinted glass in the balustrade with plain glass. There was no basis for his Honour to reject Mr Livingstone’s evidence that he tested the posts of the balustrade in 2010 and found them to be solidly attached. This was one of the modes of testing recommended by Mr Burn for assessing the structural stability of a balustrade (Blue 36W).

96. [96]

    His Honour also ignored the corroboration of the evidence of Mr Livingstone by Ms Swift. Far from having a defective memory, Mr Livingstone was able to remember the reason why he had been asked to quote on replacing the glass in the balustrade in 2010 and why he refused to simply replace the glass and recommended that the whole balustrade should be replaced, i.e. because the regulations had changed in that the specifications for balustrades were different, requiring an increased height, and also a bar or railing to be placed on top. This was fully in line with the evidence of Ms Swift on that issue.

97. [97]

    I have concluded that it was not open to his Honour to reject the evidence of Mr Livingstone on the bases which he did. When assessing the circumstances of this accident, the evidence of Mr Livingstone should have been properly taken into account and given due weight.

98. [98]

    The appellants have challenged the evidence of Mr Burn, essentially on the basis that it was based on speculation not inference, and that it lacked a sound evidentiary basis. As already indicated, at the time that Mr Burn prepared his reports of 25 November 2013 and 8 July 2014 he had very little information available to him and was proceeding on assumptions he was asked to make. Most particularly, he had not attended the site for an inspection, had not seen any photographs and had not inspected any of the component parts of the balustrade. All that he had available was the pre-purchase report and the quotations and other documents produced by Mr Mesker. In his two reports, no mention was made by Mr Burn of the bolts connecting the glass panels to the lugs/flanges. The focus was upon the posts and the lugs/flanges. It was only in evidence at trial, after having read the statement of Mr Livingstone, that for the first time he focused on the lack of maintenance and suggested corrosion of the bolts as the cause of the accident. Even then he assumed (incorrectly as it turned out) that at each point of connection there was only a single metal lug which connected the glass to the bolt (see [45] hereof).

99. [99]

    What emerged from cross-examination was that Mr Burn had no knowledge of the length of the bolts, their thickness, what they were made of and the extent (if any) of any rust/corrosion of them between the date of the pre-purchase report and the accident. In his evidence he relied on the standard for bolts at that time (Black 68Q). That, however, may not have been relevant or applicable given the observation of Mr Livingstone that:

100. [100]

     Cross-examination also revealed that Mr Burn had no knowledge of the type of glass and its thickness used in the panels on the balustrade. His evidence was based on assumptions which were incapable of being verified. The potential problems with the glass and the likelihood that it was an older style of glass lacking the “toughness” of modern glass, was made clear by Mr Livingstone at Blue 59P – Q and Blue 60R – S. The limitations of a visual inspection of the glass (which Mr Livingstone was able to carry out but which Mr Burn could not) was exacerbated by the fact that the glass was tinted, not plain.

101. [101]

     Cross-examination revealed that whereas Mr Burn had expertise as a civil engineer, he had little knowledge of the shatter and/or fracture properties of glass of different types. His experience appears to have been limited to a period when he was working for a local council in a workshop which from time to time was required to repair windscreens on vehicles which had been damaged. The timeframe for that experience was never explored or established and the relationship between the glass used on the windscreens of cars and the older style of glass apparently used on the balustrade was never explored or established (Black 67S – V).

102. [102]

     Mr Burn said that the likely force exerted by the respondent when he “leaned” on the balustrade was between 1 and 50 kgs. How that figure was arrived at and what assumptions underpinned it, was never explained. How that figure might alter if more force were exerted than merely leaning on the glass, was also not explained or explored. Another difficulty with the evidence was that Mr Burn assumed that there was only one flange on one side of the glass at each point of connection. Mr Livingstone said that there was either a fixed flange on one side of the glass and a floating flange on the other or a fixed flange on either side at each point of connection. How the different configuration of the flanges to that assumed by Mr Burn might have or would have affected the mechanism by which the failure occurred was never explained.

103. [103]

     If the flanges were fixed, it is difficult to understand how the failure of a single bolt would have led to the glass panel failing in the manner suggested by Mr Burn. The flanges would have held the glass in position for at least some time. Alternatively, if there was one fixed flange and one floating flange in each corner, it is still difficult to understand how the mechanism of failure described by Mr Burn could occur. Even if it be surmised that the glass may have slipped downwards, rather than moved outwards, this would require all four bolts to have failed almost instantaneously.

104. [104]

     Alternatively, the glass could have fractured given that we know so little about it and what we do is based purely on the visual inspection of Mr Livingstone. The evidence as to where the glass was found does not greatly assist. There was no witness who said in terms that there was no glass left on the balustrade since no witness appears to have looked. Certainly some witnesses said that there was glass beneath the respondent when he fell with Ms Swift observing that some of the glass comprised large pieces. Little can be inferred from the sparse evidence as to where glass was observed after the accident.

105. [105]

     The problem for the respondent generally and by reference to Mr Burn’s evidence in particular, is the lack of evidence to establish on the balance of probabilities how the accident occurred. There is an underlying problem in that the post/flange structure was not a typical “off the shelf standard product but in fact had been constructed d.i.y”. This meant that assumptions could not be made that the products used were in accordance with the standard in place at the time. That problem was exacerbated by the fact that the date of the construction of the balustrade is not known except that it was probably before 1992 when the regulations changed and prescribed a height greater than 900mm with a railing or bar on top. Accordingly, to the extent that reliance was placed on relevant standards, we do not know the timeframe.

106. [106]

     With that limitation in mind, there was no information at all about the dimensions or makeup of the bolts. There was no information as to the extent of the rust/corrosion of the posts and lugs observed by the author of the pre-purchase report. What was clear from the report was that not all of the posts and lugs were so affected. To the extent that there was any evidence on the subject, it was that of Ms Swift that what she observed on the posts and lugs was surface rust only. There was some support for that observation in the evidence of Mr Mesker. The implicit assumption in the findings of the primary judge of extensive corrosion being present on the external fixtures which could be observed, i.e. the posts and lugs, was not supported by evidence. To the extent that there was any evidence, it was to the contrary.

107. [107]

     Putting the evidence of Mr Burn at its highest, it goes no further than identifying the failure of the bolts as a possible mechanism for the accident. For it to reach the level of being a probable cause, the assumptions upon which it was based had to be made out and they were not. There was no evidence that the bolts were in fact rusty. The evidence went no further than to establish surface rust on some posts and flanges. For Mr Burn’s theory to be made out, what then had to be inferred was that not only similar rust but more extensive rust was affecting the bolts (the dimensions and structure of which are unknown) which over time continued to progress thereby causing them to fail when a relatively modest pressure was imposed on them. That, of course, ignores Mr Burn’s misunderstanding of how the flanges securing the glass panels were configured. If the configuration was by fixed flange and a floating flange, the flanges simply could not have been in position, as Mr Livingstone stated, if the bolts had failed. If the two flanges were both fixed, it is difficult to envisage a mechanism which would have caused the glass panel to fail even if the bolts had done so unless it was the glass panel itself which failed by one of the mechanisms put to Mr Burn in cross-examination, i.e. the glass simply being exposed to a force which it could not withstand or a latent defect.

108. [108]

     The evidence in this case as to how the respondent came to fall is exiguous. To explain the cause of the accident, on the basis found by his Honour in reliance upon the opinion of Mr Burn, involved speculation not inference. It follows that Ground 6 has been made out and factual causation giving rise to liability on the part of the appellants has not been made out.

109. [109]

     There are real problems with his Honour’s articulation of the duty of care and its breach by the appellants.

110. [110]

     There was in fact no clear articulation of the duty of care, although the statement at Red 50F – G gets closest, i.e. “as occupiers of the premises the defendants owed the plaintiff a duty to take reasonable care, including to take precautions against the risk of injury occurring due to a condition of the premises which they knew or ought to have known to represent a foreseeable risk of injury to persons such as the plaintiff”. At no time did his Honour set out the content or scope of that duty of care.

111. [111]

     Having set out that statement of the duty of care, his Honour moved directly to breach of duty. The danger with such an approach is that the duty of care and its content are liable to be defined in terms of breach which is productive of error. In Graham Barclay Oysters Pty Ltd v Ryan [2002] HCA 54; 211 CLR 540 McHugh said at [106]:

112. [112]

     Gummow and Hayne JJ made observations to similar effect in that case.

113. [113]

     Observations to that effect were made by both Gummow and Hayne JJ in Vairy v Wyong Shire Council [2005] HCA 62; 223 CLR 422:

114. [114]

     When dealing with breach, the primary judge found that the injury was foreseeable because there were a number of guests in close proximity to the balustrade, which had earlier been the subject of a recommendation for replacement, and which had not been inspected for structural integrity. His Honour found that a reasonable response was to properly maintain the balustrade which involved ensuring that the structural components were not affected by rust, or if they were, that they be replaced. He found that if such structural components were not checked, it was necessary for the appellants to obtain expert structural advice (Red 51Q – 53L).

115. [115]

     The primary judge (see [76] hereof) found that in view of the contents of the pre-purchase report, retaining a painter without structural qualifications was not an adequate response to the foreseeable risk of injury posed by the balustrade. This was because although the lugs and posts may have been adequately treated by painting, nothing had been done about the bolts which would have become unsafe because of continuing corrosion/rust. His Honour found that it was not open to the appellants to assume that the work carried out by Mr Mesker as a painter had ensured the structural soundness of the balustrade (Red 54L – P).

116. [116]

     The effect of the reasoning of the primary judge was to define the scope of the duty of care as requiring the appellants to ensure the structural stability of the balustrade before allowing guests to position themselves upon it. The basis for that obligation was the pre-purchase report of 2002. The references to obtaining expert opinion and disassembling and inspecting components were merely ways of achieving that result.

117. [117]

     As can be seen, this reasoning gave rise to the very error which was identified in the statements of principle above. It, in effect, defined the duty by reference to breach of duty, i.e. by reference to what happened and ignored the proper inquiry which was a prospective one. It is the very error identified by the majority (Gummow, Hayne, Heydon and Callinan JJ) in Vairy. At first instance in Vairy the focus of the inquiry had been on the particular rock platform from which the plaintiff had dived and what, if any, remedial action should have been taken in relation to it. As the majority in the High Court made clear, the true inquiry was a consideration of the duty owed by the council, not just in respect of that rock platform but in respect of the whole 27kms of coastline for which it was responsible.

118. [118]

     This was explained as follows:

119. [119]

     It should be noted that both Gummow and Hayne JJ observed that in this respect, the position of the local council in Vairy was generally analogous to that of an occupier of land. Reference was made to what Hayne J said in Romeo v Conservation Commission of the Northern Territory [1998] HCA 5; 192 CLR 431 at [152]:

120. [120]

     A statement of the duty of care should be at a high level of generality with amplification being by way of a statement of its content or scope. The following articulation of the duty was given by Dawson J in Northern Sandblasting Pty Ltd v Harris [1997] HCA 39; 188 CLR 313. Although the members of the High Court gave different reasons in that case, Gummow J agreed with that statement of duty and Gleeson CJ, McHugh, Gummow and Hayne JJ cited it with approval in Jones v Bartlett [2000] HCA 56; 205 CLR 166:

121. [121]

     In Sakoua v Williams [2005] NSWCA 405; 64 NSWLR 588 at [9] Mason P quoted with approval a passage from the judgment of Hodgson JA (with whom Sheller and Bryson JJA agreed) in Ahluwalia and Ors v Robinson [2003] NSWCA 175, at [23] as follows:

122. [122]

     The start point for the inquiry as to the scope of the duty of care in this case, and whether the duty was breached, is that these were residential premises. The purpose of the pre-purchase report was to bring to the attention of the appellants defects in the premises which they were intending to purchase to put them on notice of likely or possible expenditure over and above the purchase price. Its purpose was not to place them on notice of possible dangers which visitors to the premises might encounter.

123. [123]

     The focus of the primary judge was on what the pre-purchase report said about the handrails. That recommendation, however, was one of many contained in the 20 pages of the report devoted to the premises. There was nothing to separate the recommendations made in respect of the handrails from other recommendations. Most particularly, the recommendations relating to the handrails did not appear under the headings “Issues” or “Safety Concerns” in the pre-purchase report and to which reference has been made at [19] – [20] hereof.

124. [124]

     There was nothing in the recommendation relating to the handrails which indicated a risk of danger or that the handrails or balustrade might fail. For example, there was no warning such as was given in relation to “Glass” under “Safety Concerns” i.e. “this is considered to be serious …”. There was nothing in the terms of the recommendation to indicate urgency or danger. The recommendation simply stated that there was “some” corrosion of the metal lugs and posts for which the remedy suggested was “rust proofing and repainting” such as was performed by Mr Mesker or “simply replacing”. I do not read that last recommendation as indicating a replacement of the whole of the balustrade but as replacing the particular post.

125. [125]

     When one looks at all of the recommendations in the pre-purchase report, there are a number which potentially could cause harm if not remedied and the likelihood of such harm was much greater than that indicated by the recommendation in respect of the handrails, e.g. in relation to “Lintels” (Blue 12N) the recommendation was:

126. [126]

     As indicated, the content of the duty found by his Honour (at least implicitly) was far too narrow in focusing upon the recommendation as to handrails, rather than the many other recommendations contained in the report. What the primary judge did was to telescope the concepts of duty and breach retrospectively by focusing on the accident which occurred and not prospectively looking at the position of the appellants as purchasers having before them a pre-purchase report with a large number of recommendations. Only a small number of the recommendations were identified as either “Issues” or as “Safety Concerns”. They did not include the recommendation in respect of the handrails.

127. [127]

     To impose a duty of care based on the pre-purchase report on the appellants in the way in which his Honour stated would involve an obligation to implement all the recommendations in the report which foreseeably might cause harm to a visitor. To separate and focus upon the recommendation for the handrails was to do the very thing identified in Vairy as impermissible when considering a duty of care and its breach. As was there made clear, the issue of duty and its breach involved not merely placing warning signs on the rock platform from which Mr Vairy dived, but on every rock platform from which a person could potentially dive over the 27 kilometres of coastline under the control of the Council.

128. [128]

     It follows that the appellants did not owe a duty to the respondent with the content and scope implicitly found by his Honour, i.e. not to allow visitors to enter upon the rear balcony until expert opinion had been obtained in relation to its integrity or until an inspection of the fixtures had taken place so as to ensure the absence of corrosion of the bolts. If the pre-purchase report had the effect sought to be given to it by the primary judge, the scope of the duty would have required the appellants to have refused to allow visitors onto the premises until every recommendation in the pre-purchase report, which might foreseeably cause injury, had been complied with. That is an unreasonable requirement. The duty owed by the appellants was to exercise reasonable care. The duty implicitly found to exist by the primary judge went well beyond that. A duty which focused only on the handrails invited a finding of breach.

129. [129]

     For similar reasons, breach of duty has not been established. When reasoning to that conclusion I have, contrary to my finding as to factual causation, accepted that the glass panel on the balustrade failed because of the corrosion of a bolt.

130. [130]

     By reference to s 5B CLA, the “risk of harm” would be that the failure of a bolt would occur. I am, however, not satisfied that the risk of such an event was reasonably foreseeable. There was no evidence that the appellants had actual knowledge concerning a corroded bolt, nor on the evidence should they have known.

131. [131]

     The pre-purchase report made no mention of bolts. It referred only to the posts and lugs. The extent of the corrosion, as of December 2002, was not established. It can, however, be inferred from the description that not every post and lug was so affected, otherwise one would have expected a greater qualification such as “extensive” or “widespread” rather than the use of the word “some”. The extent of the corrosion cannot have been very great if painting and rust proofing was identified as an adequate response. Importantly, there was no mention in the recommendation of the bolts securing the glass to the lugs. In the absence of some specialised knowledge on their part, there was no reason why the appellants should have been on notice that apart from some “surface rust” there could have been a corrosive process attacking the bolts. Similarly, there was no basis for finding that in the absence of some specialised knowledge, the appellants would have appreciated that since they could not visually observe the state of the bolts, they should obtain expert opinion on that issue. The identified risk of harm was not reasonably foreseeable.

132. [132]

     To the extent that there was a risk of harm known to the appellants, i.e. some corrosion of the lugs and posts, their response was that of a reasonable person in their position. What could be observed of the posts and lugs was surface rust, not deep seated corrosion. There was no urgency indicated in the pre-purchase report for the need to take remedial action. In the circumstances, retaining a painter to carry out one of the recommendations made in the pre-purchase report approximately three years after they had taken possession of the premises, was not unreasonable. Moreover, there was no evidence from the painter that the corrosion which he treated was such that “rust proofing and repainting” would not be adequate. There was no evidence of any rust observed on any part of the handrails after Mr Mesker’s painting in March 2006.

133. [133]

     Had there been significant corrosion giving rise to structural weakness, it was not unreasonable for the appellants to expect to be advised to that effect by Mr Mesker. Mr Mesker did not give evidence that he had observed significant corrosion and the appellants were not so advised by him. There was also the evidence of Mr Burn at Blue 45Q – T that if Mr Mesker had observed “the degraded state of the brackets and lugs”, he should have raised the condition of the handrail with the appellants and suggested a structural assessment be conducted before the metalwork was primed and painted.

134. [134]

     In the circumstances known to them, the response of the appellants was reasonable.

135. [135]

     On the basis of what the appellants were able to observe of the handrail, the information concerning it in the pre-purchase report and the absence of any notification by Mr Mesker of substantial corrosion, the probability that harm would occur and the likely seriousness of that harm was low.

136. [136]

     In relation to the burden of taking precautions to avoid the risk of harm (s 5B(2)(c) CLA) one of the findings of the primary judge, by way of response to what he identified as the foreseeable risk of harm, was for the rear balustrade to be completely replaced. When articulating the breach of duty by the appellants, his Honour on a number of occasions referred to there having been a recommendation for the replacement of the balustrade in 2010 but that the recommendation had not been implemented by them. The implication was that the recommendation to replace the balustrade was based on safety considerations.

137. [137]

     That is an incorrect statement of the evidence. The evidence on the subject came from Ms Swift and from Mr Livingstone. Its effect was this. Ms Swift was unhappy with having tinted glass panels on the balustrade. When Mr Livingstone attended to provide a quote to repair/replace the pool fencing he was also asked to provide a quote to replace the tinted glass with plain glass on the balustrade. Because the regulations as to balustrades had changed since the construction of this balustrade, Mr Livingstone was not prepared to replace the tinted glass panels with plain glass panels of the same size. His explanation was:

138. [138]

     Although not in the written submissions, there arose in the course of oral argument a discussion as to why the appellants completely replaced the balustrade following the accident instead of simply replacing the fractured glass panel. Neither appellant was asked about that issue and there was no other evidence directed to it. On that state of the evidence, it would be unsafe to infer that the whole balustrade was replaced because other parts of it were corroded and dangerous. To the extent that any inference is available, it can probably be drawn from the evidence of Mr Livingstone, to which reference has just been made. Given that an accident had occurred, Mr Livingstone almost certainly would have recommended that the whole balustrade be replaced with one which complied with the new rules.

139. [139]

     It follows from the above analysis that breach of duty was not established.

140. [140]

     For the reasons set out above, the appeal should be allowed.

141. [141]

     The notice contended that his Honour should have considered properly the alternative theory that the glass in the balcony had failed. The respondent should not be allowed to rely upon the Notice of Contention for a number of reasons.

142. [142]

     It was never part of the respondent’s case at trial that there had been a failure of the glass panel. That issue was raised in cross-examination by the appellants as an alternative theory to that put forward by Mr Burn.

143. [143]

     Had the respondent sought to rely upon the failure of the glass panel as an alternative basis for a finding of negligence against the appellants, it is the sort of matter on which evidence could and would have been given.

144. [144]

     The relevant statement of principle is set out in Coulton v Holcombe [1986] HCA 33; 162 CLR 1 where the plurality (Gibbs CJ, Wilson, Brennan and Dawson JJ) said at 7:

145. [145]

     The respondent’s application to rely upon his Notice of Contention should be refused.

146. [146]

     The orders which I propose are:

     1. (1)

        The appeal is allowed.

     2. (2)

        The verdict and judgment in favour of the respondent on 9 October 2014 is set aside.

     3. (3)

        In lieu thereof, there be verdict and judgment in favour of the appellants.

     4. (4)

        The order for costs of the trial in favour of the respondent is set aside.

     5. (5)

        In lieu thereof, the respondent is ordered to pay the appellants’ costs of the trial.

     6. (6)

        The respondent is to pay the appellants’ costs of the appeal.

147. [147]

     SIMPSON JA: The comprehensive statement of the relevant facts and circumstances contained in the judgment of Hoeben JA, which I have read in draft, relieves me of the need to undertake an extensive recitation. I will confine my references to the facts to those that are directly relevant to the issues raised on appeal, and my conclusions.

148. [148]

     The respondent particularised his case against the appellants as follows:

149. [149]

     The claim was governed by the provisions of the Civil Liability Act 2002 (NSW).

150. [150]

     The essential facts for the purpose of this appeal are as follows. In 2002, prior to purchase of the property, the appellants commissioned a building report which was provided by Mr Philip Connor, a building consultant. The report is lengthy (31 pages) and covers a range of issues. Relevantly for present purposes, the report made reference, under the heading “Safety Concerns”, to “Glass”, but this appears to be confined to shower screens, windows and “the like”, and not to include the glass constituting the balustrade on the balcony. After making some observations about the concrete and other aspects of the balcony, Mr Connor wrote:

151. [151]

     Following purchase, the appellants undertook some remedial work. That work included replacing glass fencing around the swimming pool (in 2010) and painting the steel poles supporting the upstairs balustrade (in 2006). The pool fences were replaced by a company called Abbey Fencing, of whom Mr Peter Livingstone was a sales consultant; the painting was done by Mr Alex Mesker.

152. [152]

     Mr Mesker gave evidence that he checked for rust on the “tabs” that were holding the glass between the posts. He observed no rust, and so found it unnecessary to dismantle the balustrade or remove the glass. He said that his practice was to clean or descale, and use a proprietary rust convertor which converts the rust to an inert steel; and then to apply an oil based primer that excludes oxygen, and finish with an oil based coat of paint over the final coat.

153. [153]

     Mr Livingstone’s evidence was given, in the first instance, by way of a statement made on 8 July 2014. He said that he had been asked to attend the premises to inspect, report on, and quote with respect to the pool fence, but that while he was there Ms Swift (the second appellant) asked him to estimate the cost of replacing the existing tinted glass on the balustrade with clear glass. In his statement he described the balustrade as follows:

154. [154]

     Because of the height of the balustrade, Mr Livingstone considered that replacing the glass alone would result in a balustrade that failed to meet then current building standards, and accordingly declined to take that course. In the process, however, he noted:

155. [155]

     Mr Livingstone was unable to locate any documentation that would enable him to pinpoint the date of his first visit to the premises, but he believed that it was sometime in 2010.

156. [156]

     Mr Livingstone next attended the premises on 12 November 2011, after the respondent’s fall. He said that there was no immediate or apparent damage to the balustrade and that the posts and flanges to which the glass would have been affixed appeared still to be sound and in good working order. All that was untoward that was observed by Mr Livingstone was the absence of the one sheet of glass that had been associated with the respondent’s fall.

157. [157]

     In oral evidence he said that, on his second visit, because he was interested to see why the accident might have happened, he had examined the site where the glass was missing; he observed that the posts were in good order, as were the brackets and bolts, which showed no damage. Only the glass was missing.

158. [158]

     This assertion was the subject of some cross-examination. Mr Livingstone acknowledged that he had made no mention in his statement of observing the intact bolts. He said, however, that he had a memory of seeing the bolts. He said that the recollection had come to him while thinking about what he had seen. He did not accept that his recollection was the result of reconstruction. It was put to him that he was guessing, to which he answered “possibly”. He agreed that he could have been mistaken.

159. [159]

     He agreed that he had no specific recollection of seeing the bolts, but that:

160. [160]

     It is of some interest that, although only one glass panel was broken, Mr Livingstone’s company replaced the whole balustrade. Why this was done in preference to repair of the one broken panel was not explored in the evidence. There was no evidence called from the contractor who replaced the balustrade, and who might have been expected to have had some knowledge of the state of the bolts on the broken glass, and also on the other panels.

161. [161]

     Accounts of how the respondent came to fall varied. The respondent said that he had consumed a small amount of alcohol, and had been sitting at a table with two of the guests (Dennis Stobo and his wife) when he suffered a cramp in his right leg. (This was a condition from which he suffered intermittently, and for which he was receiving treatment.) He stood up to steady himself, and put out his left hand against the balustrade. He was about a metre to a metre and a half away from the balustrade. He leaned against the balustrade. He had no further recollection of the fall; his next memory was of travelling in an ambulance to hospital.

162. [162]

     The respondent’s wife (Lyn Montgomery) gave evidence to similar effect. She said that she saw the respondent stand up; it seemed to her that he was having a problem with a cramp. He put out his left arm which landed on the top of the glass, although she was not sure whether contact was made. He then disappeared from view.

163. [163]

     Mr Dennis Stobo said that the respondent stood up, appeared to have his foot caught, perhaps in his chair, and stumbled. He reached out his arm (probably his left), reaching for the horizontal railing on top of the balustrade. (It is doubtful that there was a separate railing above the top of the glass.)

164. [164]

     The first appellant, Mr Anthony Swift, gave a different account. He said that he was on the balcony sitting at a table with other guests, including the respondent. He saw the respondent stand, his left leg gave way completely, he stumbled to his left and “crashed” into one of the glass panels with great force. His hip hit the glass panel. The glass panel gave way, and the respondent fell, with the glass panel, to the garden bed below. Mr Swift said that he had a very clear memory of the respondent “hitting that panel at considerable force”. He used language such as “smashing into the panel”; he repeated the word “crashed” on more than one occasion. He likened the event to “a rugby game or something”, and said that “it was a big impact, massive impact”.

165. [165]

     Mr Ian Dobson was another guest at the party. He was seated at a table on the balcony. He saw the respondent stand. Mr Dobson thought that he had probably caught his foot on the chair or the table. He stumbled towards where Mr Dobson was, corrected, and headed towards the balustrade, which he struck with his shoulder in the middle of the glass. Mr Dobson also used a rugby union analogy; he said that the respondent “hit it [the glass panel] fair and square with his shoulder, straight through it, right in the middle of the glass, straight through …”. He maintained that the respondent’s body struck the glass.

166. [166]

     In cross-examination, Mr Dobson volunteered some evidence about his alcohol consumption on the afternoon and evening. In answer to a question specifically challenging his recollection (about where the respondent was seated), he said (unresponsively):

167. [167]

     He said that they had begun drinking at about 3 or 4 o’clock in the afternoon, that he had been drinking light beer, and had probably had “a wine” with the meal, and possibly champagne for a toast, that he was not driving, and accordingly had no reason to count his drinks.

168. [168]

     There was also some evidence about the fate of the glass panel. Ms Montgomery said that there was no broken glass on the balcony, but there was a lot of glass on the ground under and around the respondent. Mr Stobo made no observations of glass on the balcony, but saw glass on the ground. Ms Swift said that the whole panel of glass had fallen to the ground, and was on the ground in some big pieces.

169. [169]

     The respondent called evidence from an expert, Mr Ian Burn, a consultant engineer, who provided two reports, dated 25 November 2013 and 8 July 2014. Because, by the time Mr Burn was retained, the balustrade had been replaced, he did not inspect the site. He had no details of the glazing work on the balcony. He assumed that the recommendation in Mr Connor’s building report had not been implemented. In this he was correct; Mr Mesker had painted, but not rust-proofed or replaced, the metal lugs and posts. He assumed therefore (and on the basis of past experience) that the collapse of the balustrade was caused by corrosion of the handrail fixings. In the second report he was asked to answer several specific questions.

170. [170]

     In his second report, Mr Burn drew attention to a discrepancy in a quotation given by Mr Mesker in March 2006, and a later quotation in August 2010, which was the quotation for the work actually undertaken. The earlier quotation specified remedial rustproofing to metalwork; this was not included in the later quotation. Mr Mesker expressed the opinion that, had the work quoted in 2006 been then carried out, the continued corrosion of the handrail and its fixtures may have been arrested, preventing the collapse in 2011.

171. [171]

     In oral evidence, Mr Burn was asked if he had reached a conclusion as to how the accident occurred. His answer is recorded in the transcript as follows:

172. [172]

     He expressed the opinion that:

173. [173]

     It was Mr Burn’s conclusion that it was the bolts that had failed, rather than the glass. He acknowledged that, if the bolts had still been in place after the respondent’s fall (as Mr Livingstone said they were), that would indicate that there was some other cause for the collapse of the balustrade.

The contentions of the parties at trial

1. [174]

   A significant factual issue, as to which the evidence was sparse, was the mechanism that caused the glass to give way: did the glass break, shatter, or crack before hitting the ground below, or did the whole of the glass panel separate from the uprights to which it was attached? There was, in fact, no evidence that the glass shattered before it hit the ground below. There was no direct evidence that it had separated from its metal anchorings (to the uprights); an inference that this is what happened emerged, principally from the evidence of Mr Burn.

2. [175]

   The transcript of 13 August 2014 indicates that both parties provided written submissions. Those written submissions were not included in the appeal papers before this Court.

3. [176]

   The case advanced on behalf of the respondent was that there was a defect in the balustrading and that this was known to the appellants, and had been known to them since 2002. The defect was caused by the corrosion of the metal lugs and posts by which the glass was attached to the posts, or uprights. As support for this contention, reliance was placed upon the evidence of the location of broken glass after the accident; there was no evidence of broken glass on the balcony; all the broken glass was in the garden bed below. This would indicate that the whole panel of glass had come apart from its anchoring bolts. On the respondent’s case, the painting by Mr Mesker did nothing to address the corrosion. Mr Mesker had no suitable qualifications to assess the integrity of the glass panel or the post to which it was attached. Nor did Mr Livingstone, on his inspection before or after the accident, have relevant qualifications; he was a sales consultant.

4. [177]

   On behalf of the appellants, issue was taken with the mechanism of the collapse of the balustrade. It was submitted that the evidence did not support a conclusion that the bolts were corroded; or, at least, corroded to a sufficient extent as to amount to a risk to safety.

5. [178]

   However, the principal argument advanced on behalf of the appellants was that the duty of care which an occupier of premises owes to an entrant was not such as to require the replacement of the glass balustrade. (Formulating the case in this way is of some significance, as will appear.)

The findings of the primary judge

1. [179]

   The primary judge commenced his judgment by identifying three issues for his determination. They were:

2. [180]

   He proceeded to make general credibility findings in respect of each of the witnesses. He found that the respondent:

3. [181]

   He made similar favourable findings about Ms Montgomery, Mr Stobo and Ms Swift. He considered that no credit issues arose with respect to Mr Burn. He made less favourable findings about the evidence of Mr Livingstone, whose recollection as to key events he said was unreliable, and included “elements of unwarranted assumption and reconstruction, and … inaccuracies”; about the evidence of Mr Swift, who he considered had exaggerated his evidence as to the manner in which the accident occurred, had engaged in a process of reconstruction and who he held to be unreliable; and Mr Dobson, who, he noted, had been consuming alcohol for 5-6 hours prior to the accident, such as to “have had some effect on the acuteness and accuracy of his observations”, and whose account, his Honour said, “varied materially at times”, “involved reconstruction, and was in part based on assumptions”. His ultimate conclusion in relation to Mr Dobson was that he was an unreliable witness. He described as “fanciful” the descriptions of the accident given by Mr Swift and Mr Dobson, and considered that their evidence had been “tailored” to suit an argument that the structure was sound, and that a great force had been applied to it to cause it to fall. He expressed no concerns as to Mr Mesker’s credibility, but noted that his recollection of detail was “understandably limited by the passage of time” and by the loss of computer data comprising his records.

4. [182]

   He found (at [39]) that, before December 2002:

5. [183]

   He then reviewed at some length the evidence of the various witnesses as to the circumstances of the accident, and to the hypothesis of Mr Burn. He stated his conclusion at [130], saying:

6. [184]

   This was his finding in respect of Issue 1. With respect to Issue 2, the primary judge set out s 5B of the Civil Liability Act, which is as follows:

7. [185]

   He said:

8. [186]

   The primary judge proceeded on the basis (unchallenged) that the appellants, as occupiers of the property, owed the respondent a duty of care. The relevant question was whether the appellants had failed to discharge that duty. He considered it critical ([145]) that the appellants had been advised, in 2002, to have the rust affected area treated, or the balustrade replaced. He made the following findings:

9. [187]

   He rejected a submission made on behalf of the appellants that they should not be held liable in negligence because they were not skilled in the building trades, they had engaged a professional painter to undertake remediation works identified in the 2002 pre-purchase inspection report, and that he had not identified any structural or safety issues. He rejected this submission on the basis of the four year delay between identification of the defects in the pre-purchase inspection report, and when the painting works were undertaken, a further four years thereafter, and because neither Mr Mesker, nor Mr Livingstone, was qualified to comment on structural matters ([148]-[150]).

10. [188]

    He went on to say:

11. [189]

    His ultimate conclusion was:

12. [190]

    Importantly, in my opinion, he also noted that an alternative precautionary measure was for the appellants to entertain guests in a different area of the premises that did not pose a risk to safety by reason of the unsafe balustrade.

13. [191]

    His Honour then dealt with causation in a manner that has not been challenged on appeal.

Contributory negligence

1. [192]

   Initially the appellants pleaded contributory negligence on the basis of asserted intoxication on the part of the respondent. That assertion was abandoned, and the appellants claimed that the respondent caused or materially contributed to his own injury by reason of:

2. [193]

   The primary judge rejected each of these, finding that the respondent had taken only one step when he experienced a momentary unsteadiness to which he responded reasonably by reaching out to steady himself on the nearby balustrade, and that it was the cramp that caused him to experience unsteadiness, that events transpired rapidly in such a way as not to sustain an assertion of contributory negligence. No ground of appeal challenges that conclusion.

The appeal

1. [194]

   The grounds of appeal are fully set out in the judgment Hoeben JA. It is unnecessary to repeat them.

2. [195]

   Grounds 3-7 concern the fact finding exercise. By Ground 3, the appellants complain that the primary judge failed to address “the unchallenged evidence of Mrs Swift that there was only light surface rust on the post, and no evidence of staining after [Mr Mesker’s] painting”.

3. [196]

   The whole of Ms Swift’s evidence in this respect was as follows:

4. [197]

   There was no suggestion that Ms Swift had any expertise in the assessment of rust penetration, nor, indeed, that she had made any attempt to examine the structure for rust penetration. In counsel’s final address in the trial, no reliance was placed on this evidence. During the course of counsel’s final address, the following exchange is recorded:

5. [198]

   No mention was made in this exchange, or anywhere else, of Ms Swift’s evidence that the rust was surface rust only. As mentioned above, the written submissions made at trial were not available to this Court, but one would have expected that if they contained anything additional they would have been included in the appeal books in order to support this ground which, it should be emphasised, complains of a failure by the primary judge to address that evidence.

6. [199]

   In the circumstances, it was entirely appropriate for the primary judge to attribute no weight to Ms Swift’s evidence in this respect. And the only mention made of this evidence in the written submissions to this Court was in the recitation of “background facts”. It was not further mentioned in that part of the written submissions that appeared under the heading “Appealable errors”.

7. [200]

   I would reject Ground 3 of the appeal.

8. [201]

   By Ground 5 the appellants complain of the treatment of the evidence of Mr Swift and Mr Dobson. The principal complaint is that the primary judge rejected their evidence “solely on the question of credit”, and failed to assess the extent to which the evidence of each corroborated the evidence of the other.

9. [202]

   One specific complaint made concerning the treatment of Mr Dobson’s evidence concerned the finding of his alcohol consumption, and the inference expressly drawn that this had impaired the accuracy of his observations and recollections. It was submitted that there had been no cross-examination to the effect that he was so intoxicated as to affect his perception or memory.

10. [203]

    I have set out above the passage of the evidence, in cross-examination, in which this subject arose. In fact, it was Mr Dobson himself who introduced the subject. In a passage of cross-examination in which his recollection was being tested, with specific reference to the identities of the persons with whom he was sitting, Mr Dobson said:

11. [204]

    The assertion that the evidence of Mr Swift was rejected “solely on the question of credit” is not correct. At [83] the primary judge set out seven instances of what he called “variations within the descriptions [Mr Swift] gave”. These led him to the conclusion that Mr Swift was “not a careful witness” ([84]).

12. [205]

    The primary judge subjected Mr Swift’s evidence to a lengthy analysis. He undertook the same exercise with respect to Mr Dobson’s evidence.

13. [206]

    Once that exercise had been undertaken, examination of the extent to which the two witnesses corroborated one another would have been of little value. It is true that Mr Swift and Mr Dobson both used (similar) vivid language to describe the force with which the respondent struck the balustrade, and, in that sense, could be said to have corroborated one another. But the search for corroboration would have required focus, not only on those parts of their evidence that coincided, but also on discrepancies. Importantly in this respect, Mr Dobson was positive that the respondent had hit the glass with his shoulder; Mr Swift that it was his hip. Moreover, had such an examination been undertaken, it would have been necessary to replicate it with respect to the evidence of the respondent, Ms Montgomery, and Mr Stobo, who gave essentially similar accounts of what happened immediately before the collapse of the balustrade. Those accounts did not include assertions that the respondent struck the glass with great force.

14. [207]

    I would reject Ground 5 of the appeal.

15. [208]

    By Ground 4 the appellants complain of the primary judge’s reliance on the evidence of Mr Burn as to the mode of failure of the balustrade and its cause, and, by Ground 6, they complain that the primary judge engaged in “impermissible speculation” as to these issues.

16. [209]

    Criticism was made of Mr Burn’s evidence. One such criticism was that he had reached his conclusions without examination of the balcony and without knowing any relevant specifications of the bolts or the glass. Given that, by the time Mr Burn became involved, the balustrade had been entirely replaced, and given that there appears to have been no information – certainly no evidence – about the specifications of the bolts and glass, these criticisms were somewhat hollow. Mr Burn had to do the best he could with the material available to him.

17. [210]

    With respect to his evidence that the accident had happened by reason of failure of the bolts, it was submitted that “the balustrade could have failed for a number of reasons, including an invisible defect in the glass”. This proposition was based upon one question and answer in the cross-examination of Mr Burn. The question and answer were:

18. [211]

    In any event, two responses may be made to the proposition. The first answer is to put the question and answer in the context of the whole of the cross-examination. It was preceded by this question and answer:

19. [212]

    The second answer to the proposition is that, had the glass had a defect and broken, it might reasonably be expected that pieces or shards of glass would have been on the balcony floor. But there was no evidence that there was, and some evidence that there was not.

20. [213]

    Moreover, in his second report, Mr Burn commented that:

21. [214]

    I do not consider that there was any error in the primary judge’s acceptance of Mr Burn’s evidence. He was the only witness with any expertise. True it was that, by reason of the absence of concrete facts, he was forced to engage in inferential reasoning. Experts are permitted, by reason of their expertise, to engage in a process of extrapolation from known facts: Hatziandoniou v Ruddy [2015] NSWCA 234; 72 MVR 171. That is what Mr Burn did, and it was no error by the primary judge to accept his process of reasoning. To do so is not to engage in speculation. I would reject these grounds.

22. [215]

    The final ground concerning the facts was Ground 7, by which the appellants complain that the primary judge erred in disregarding what was said to be the unchallenged evidence of Mr Livingstone, that on his inspection shortly after the accident, all fixed and floating flanges that would have restrained the glass panel were in place on the posts. The primary judge found to the contrary. Much of the relevant evidence is set out above, but I will, for convenience, repeat it here.

23. [216]

    The relevant evidence in chief was:

24. [217]

    The characterisation of Mr Livingstone’s evidence as “unchallenged” is incorrect. His evidence was challenged, so far as it could be in the absence of firm instructions as to the existence or otherwise of the bolts post-accident, by testing his memory, and by doing that by reference to his statement, which made no reference to the bolts. In any event, the cross-examination I have extracted shows that, in fact, Mr Livingstone was quite equivocal. His evidence was not “disregarded”, as asserted in the ground of appeal; it was, quite properly, treated as equivocal.

25. [218]

    I would reject Ground 7 of the appeal.

26. [219]

    Each of these grounds concerns one or other of the matters relevant to the finding of duty of care, the scope and content of the duty of care, and breach of duty.

27. [220]

    It is necessary to set out Ground 1 as pleaded. It is:

28. [221]

    The word “inferentially” was inserted for a reason. The reason is that the primary judge made no finding that the scope and content of the duty of care owed by the appellants to the respondent required them to have the balustrade inspected by a structural engineer, or required them to have the balustrade dismantled and inspected. Nor did the respondent advance any case to the effect.

29. [222]

    The fallacy was repeated in the written submissions. There it was asserted:

30. [223]

    Paragraph 153 of the judgment was cited as the source of the “finding”. Paragraph 153 reads in full:

31. [224]

    There was no evidence and no submission (at trial) that “taking advice and action to ensure the structural soundness of the balustrade” entailed retention of a structural engineer or dismantling the balustrade, or both. Ground 1 as pleaded overstates what the primary judge found. By s 5B(2)(c) of the Civil Liability Act, the burden of taking precautions to avoid the risk of harm is one of the considerations a court is required to take into account in determining whether a reasonable person would have taken such precautions. Neither party adduced evidence to establish what would have been involved in “taking advice and action to ensure the structural soundness of the balustrade”. Reasonable common sense dictates that a building contractor or consultant or the like could have undertaken the initial examination; what action needed to be taken would, of course, depend upon what that examination revealed.

32. [225]

    Against that must be balanced the magnitude of the risk if, as did happen, the balcony collapsed under the weight of a person who came in contact with it. Plainly, the likely seriousness of harm that would ensue is significant. Moreover, inspection and advice was not the sole precaution on which the primary judge relied. He made it plain that an alternative was simply to refrain from making the balcony available to guests until it had been rendered safe. Again, when seen in the light of the magnitude of the harm that could ensue, such a precaution could hardly be called burdensome.

33. [226]

    One of the submissions on behalf of the appellant was that:

34. [227]

    This is correct, but what might have been suggested can now not be known, given the total replacement of the balustrade. What Mr Burn suggested was that, had action been taken in 2002, or shortly thereafter, when the building report first signalled the problem, was that the continuing corrosion might have been arrested. What is known is that, nine years after the appellants were alerted to the existence of rust, the risk that was then identified eventuated.

35. [228]

    In my opinion the appellants have failed to make good any of the grounds of appeal concerning the duty of care owed by the appellants, and their breach of that duty.

36. [229]

    I would, accordingly, dismiss the appeal.

37. [230]

    In those circumstances it is unnecessary to consider the Notice of Contention filed on behalf of the respondent.

38. [231]

    Having regard to the conclusions of Hoeben JA, I would add the following.

39. [232]

    It was not, and could not have been, in issue that the appellants owed the respondent a duty of care. The key to the respondent’s case lay in the determination of what was encompassed in that duty. That is a question that could not, in this case, be answered until the cause of the balustrade collapse was determined. That having been identified as failure of rusty fittings (it scarcely matters whether it was lugs, posts, or bolts) – the questions that arose under s 5B of the Civil Liability Act were:

40. [233]

    The answer to the first question is that the risk of harm was that, if the fittings failed, a person might fall from the balcony. The second question is answered in the light of what the appellants knew. Once it was established that they had forewarning of a defect in the fittings, the risk that a person placing weight on the balcony might fall was foreseeable.

41. [234]

    The very purpose of a balustrade on a balcony is to prevent persons falling from the edge of the balcony. It must be in contemplation that some weight will be placed upon the balustrade and that it will be sufficiently robust to withstand that weight. Of course, it is not expected to cope with excessive pressure or weight. (At just what point pressure or weight becomes excessive need not and cannot be decided.) This is the significance of the evidence of Mr Swift and Mr Dobson concerning the manner in which the respondent struck the balustrade. If accepted, their evidence would have raised for consideration the question of what weight a balustrade might reasonably be expected to bear. Rejection of their evidence removes the need to come to a view about just what pressure would be seen as excessive. In Jones v Bartlett [2000] HCA 56; 205 CLR 166, at [176] and following, Gummow and Hayne JJ discussed (with respect to leased premises) what constitutes a dangerous defect. Their Honours drew a distinction between defects that are dangerous of themselves, and that will, or may, cause injury through normal use of the premises, and “defects” that arise from misuse.

42. [235]

    The risk that, with normal use of this balustrade – that is, the application of normal, not excessive, pressure – it might break, was not insignificant. Rejection of the evidence of Mr Swift and Mr Dobson meant that it was not necessary to consider what weight would have made the difference between a defect that would cause injury only through misuse, and a defect that could cause injury through normal use.

43. [236]

    For my part, I consider that a balustrade on a first floor balcony ought to be sufficiently robust to withstand the pressure of even a relatively heavy fall against it. It is not uncommon, after all, for a number of people to lean on a balustrade such as this.

44. [237]

    Moreover, given the well-known propensity of rust to progress, for as long as it was not attended to, the significance of the risk increased over time.

45. [238]

    What precautions a reasonable person in the position of the appellants would have taken is to be decided in the light of these considerations. That is what the primary judge did. It must be emphasised that the precautions upon which he relied did not necessarily extend to heroic engineering repairs; entertaining guests in another part of the premises was a viable alternative (although, perhaps, not so appealing to the owners of an attractive balcony).

46. [239]

    The orders I propose are: