HKSAR v. Tsui Tack Kong and Others

HCMA 289/2018

[2019] HKCFI 2925

IN THE HIGH COURT OF THE

HONG KONG SPECIAL ADMINISTRATIVE REGION

COURT OF FIRST INSTANCE

MAGISTRACY APPEAL NO 289 OF 2018

(ON APPEAL FROM ESS 39227-39230 OF 2015)

________________

________________

____________________

JUDGMENT

____________________

1.On 10 April 2018, all four appellants (“A1‑4”) were convicted after a 29‑day trial by Deputy Magistrate Winston Leung, of offences contrary to section 40(2B)(b) of the Buildings Ordinance, Cap 123 (“the BO”) and received fines of between HK$30,000 and HK$100,000. A1‑4 seek to appeal their convictions only. Additionally, A3 and A4 seek leave to adduce fresh evidence and to amend their Perfected Grounds of Appeal.

2.All four appellants were charged, in their various capacities, under section 40(2B)(b) of the BO, which reads as follows:

“Any person… directly concerned with any prescribed inspection, site formation works, piling works, foundation works or other form of building works who —

(b) carries out or has carried out such inspection or works, or authorizes or permits or has authorized or permitted such inspection or works to be carried out, in such manner as is likely to cause a risk of injury to any person or damage to any property,

shall be guilty of an offence and shall be liable on conviction.”

BACKGROUND

3.On 17 November 2014, Chan Shu‑moon was walking along the pavement on the northern side of Jardine’s Bazaar, Causeway Bay, when the pavement gave way and he fell into a large hole (“the Sinkhole”).  He suffered multiple injuries including three broken ribs.

4.The area where the pavement gave way was located outside of Prosperous Commercial Building at Nos 54‑58 Jardine’s Bazaar. There was an active construction site abutting that building to the east, which was situated at Nos 60‑66 Jardine’s Bazaar (“the Site”).  The sinkhole into which Mr Chan fell was just 2.3m from the Site.  The construction works upon the Site involved the demolition of the original building and the construction of a new commercial building with a basement.

Safety Supervision of Building Works

5.All building works in Hong Kong are controlled by the Building Authority under the BO.  The BO governs the supervision of buildings works in Hong Kong, which are ultimately supervised by the Buildings Department (“the BD”) in collaboration inter alia with the Geotechnical Engineering Office of the Civil Engineering and Development Department.  The Technical Memorandum for Supervision Plans [P8], supplements the provisions of the BO and it sets out the principles, requirements and operation of supervision plans.  The Code of Practice for Site Supervision [P7], which covers both quality supervision and site safety supervision, provides detailed guidance on the application of the Technical Memorandum.  Within the safety management area, duties and responsibilities are divided into four streams under the Authorised Person (“AP”), Registered Structural Engineer (“RSE”), Registered Geotechnical Engineer (“RGE”) and Authorised Signatory (“AS”) of the Registered Contractor (“RC”), respectively.  Within each stream, Technically Competent Persons (“TCPs”) would be identified who were responsible for routine safety supervision (T1‑3) and engineering safety supervision (T4‑5).

The Appellants’ Role in the Project

6.The project consultant was TK Tsui ‑ Gabriel Yu Limited.  Tsui Tack Kong (“A1”) was a director of that company.  He was also both the RSE and RGE for the Project.  He was the TCP‑T5 under the RSE and RGE streams as well as being the Directorate Site Supervisor (“DSS”) under the RGE stream. Yuen Ting Fung (“A2”) was the TCP‑3 under the RSE, RGE and AP streams.  He was A1’s representative on site.  Ching Lee Engineering Limited (“A3”) was the Registered Specialist Contractor (“RSC”).  Lee Kwong Kong (“A4”) fell under the RSC stream and was the representative of the Authorised Signatory of the Registered Contractor as well as TCP‑T2 and TCP‑T4. He was A3’s representative on site.

7.A1, A2 and A4 were charged with authorising or permitting building works to be carried out on the Site between 6 and 17 November 2014.  A3 was charged with carrying out those works during the same period. There was no issue before the learned magistrate that if the manner in which the works were carried out on the Site were found to be the cause of the Sinkhole, then A1-4 would be guilty by virtue of their various roles.  The trial therefore turned upon the issue of causation, as does this appeal.

The Building Works

8.Building works on the Site commenced in August 2013. The contractor had to demolish the existing building; build a cofferdam; construct a bored pile foundation comprising 4 x 2m diameter bored piles; excavate the soil for the basement and, lastly, construct the superstructure.

9.In simple terms, in order to construct a basement within the new building, the contractor had to first construct what is known as a “cofferdam”.  A cofferdam is a watertight structure designed to permit construction below the ground water level (“GWL”).  The cofferdam in question had four sides, which ran parallel with the perimeter of the Site, but had no top or bottom.  It was constructed first, by driving closely aligned pipe piles into the ground to form four walls to make a square.  These pipe piles, which were 18m in length, were driven into the ground to a depth of -14mPD[1] (18.5m below ground level (“GL”)), which was lower than the desired basement depth of 15m below GL.  Because the basement was below GWL, water would inevitably seep into the Site.  The depth of the pipe pile was calculated by the RSE/RGE (A1), to take account of the water coming into the Site.  The deeper the pipe pile, the less water which should be able to flow under the pipe pile and make its way back up to the surface of the ground within the Site.  This calculation had to take account of the lowering of the GWL outside of the Site and consequent associated ground settlement because of the loss of water into the Site, as well as the natural recharge of the water level outside of the Site and the additional recharge by taking water from the Site and pumping it back outside of the Site.  The aim on this Site, was to ensure that the GWL outside of the Site, would be maintained within a range of up to ‑1m.   The pipe piles were only the first part of the construction of the cofferdam.  Thereafter, grout tubes were sunk to a similar depth to the rear of, and in between, the pipe piles.  Two types of grout were then injected into the tubes to form an impermeable (or less permeable) layer/coating.  This was the Grout Curtain.  The pipe piles themselves were then joined by welded steel plate planking (“the lagging panel”) along the whole length of the pipes as the ground was excavated and steel lateral supports were then installed to stop movement.  The excavation was supposed to be carried out in 0.5m drops for each phase of lagging.  Whilst such a structure was not 100% impermeable, it was sufficiently impermeable to allow for excavation within the Site subject to some allowable water from above and from below the toe (or base) of the Grout Curtain and, hopefully, with only minimal seepage through the Grout Curtain itself.  Seepage through the Grout Curtain was not an unknown phenomenon in building works of this kind.  Given the structure of the cofferdam, it would be difficult to see any damage to the pipe piles or grout tubes behind the lagging panel once it was in place.

10.This case involves the phase of construction known as the excavation and lateral support works (“the ELS works”), which were required in order to construct the basement.  The Grout Curtain works commenced on 1 October 2013.  They were finished on 7 January 2014.  Between 30 December 2013 and 3 January 2014, pumping tests were carried out in order to ascertain the effectiveness of the impermeability of the Grout Curtain as designed by A1 and approved by the BD.  It was found to be effective at that time.  Within the Site, four Bore Piles were also drilled and constructed during this period, between 1 November 2013 and 12 June 2014.  They were required to form the support for the main building superstructure.  The ELS works proper were therefore not commenced until 17 June 2014.  They had to be carried out in stages as the name suggests.  First the ground was excavated, then the lagging panels were welded onto the pipe piles and then structural supports were put in place to prevent collapse of the cofferdam.  The ELS works were continuing up to the time of the creation of the Sinkhole on 17 November 2014.  At the stage that this accident occurred, the ELS works had successfully excavated to layer 4 and were in the process of excavating to layer 5.  They were approximately 9m below GL.

THE CASE BEFORE THE MAGISTRATE

11.There were a significant number of areas of dispute between the parties during trial, some of which advanced none of the party’s cases but took up a considerable amount of court time and intellectual energy.  I have synthesised the actual central arguments as follows.

Prosecution Case

12.In summary, the prosecution asserted that the manner in which the works were carried out caused the Sinkhole into which Chan Shu‑moon fell.  Specifically, they alleged that the Sinkhole (referred to in trial as “Sinkhole 3”), was directly related to the ELS works.  They alleged inter alia that the Grout Curtain was defective and/or its workmanship was poor; the cofferdam design was inherently risky as it did not allow sufficiently for the inflow of water into the Site and site supervision was not functioning as the monitoring equipment was either not working or non‑existent.  The prosecution relied primarily upon the evidence of KK Pang, who opined that the Sinkhole was created because:

(1)  the Grout Curtain was ineffective;

(2)  as a result, excessive amounts of water were allowed to flow into the Site through the Grout Curtain;

(3)  that water leakage gave rise to a “piping effect” outside of the Grout Curtain;

(4)  that “piping effect” caused erosion and loss of ground which in turn caused the formation of cavities outside of the Grout Curtain; and

(5)  soil above the cavities collapsed into the cavities and the Sinkhole was formed into which Chan Shu‑moon fell.

13.The prosecution case was based upon two simple facts i.e. the observed excessive inflow of water into the Site and the fact that the Sinkhole was dry at a depth of -3.7m below GL.  Given the known GWL before construction, the GWL had dropped some -1.27m to -1.82m, or more, as a result of the building works.  The excessive amounts of water flowing into the Site, which was drawing water from outside of the Site, should have been detected by the GWL monitoring equipment and the Ground Settlement Level (“GSL”) monitoring equipment required on and around the Site.  Records recovered do not show any such noticeable change of GWL or GSL.  The prosecution say that was because that monitoring equipment was incorrectly positioned and either faulty or not in existence.

Defence Case

14.In summary, the main thrust of the appellant’s case was that there was no leakage through the Grout Curtain itself.  There was no proof of any damage to the Grout Curtain.  They accepted there was water coming into the Site but from underneath the base of the Grout Curtain and not through it.  The water which could be seen to flow into the Site was not excessive.  Whilst the monitoring equipment was not in its originally designated position, it was in a position to monitor the GWL and GSL satisfactorily.  The prosecution had not proved that the equipment was either not present or faulty at the time of the formation of the Sinkhole.  They had not proved the falsity of what records there were.  The formation of other sinkholes (Sinkholes 1 and 2) was not relevant and the appellants were not acting in concert to ignore or falsify monitoring records.

15.Significantly, the formation of the Sinkhole was caused by a means other than the manner in which the works were carried out at the Site.  In essence, by a combination of two factors i.e. inherent GW movement in the area from the hills to the harbour and/or prolonged water leakage from various water carrying facilities.  If accepted as possible causes of the Sinkhole, the Court cannot draw an irresistible inference that it must have been caused by “piping effect”, which hypothesis, in any event, has been undermined.

PERFECTED GROUNDS OF APPEAL

16.The Grounds of Appeal from all of the appellants were somewhat over lengthy when their argument actually boiled down to a criticism that the learned magistrate should not have rejected the opinion evidence of the defence expert witnesses and that some of the underlying facts from which an inference was drawn, were not proved beyond reasonable doubt.  There was over reliance upon a multiplicity of small points which were not determinative of the prosecution.  Much of their argument, especially that of A3 and A4, was based upon expert opinion, some of which had not been advanced in the same manner before the magistrate.  However, to ensure clarity, they were as follows:

A1 and A2

Ground 1

17.The learned magistrate erred in accepting the evidence of KK Pang.  In particular, the learned magistrate erred in respect of the following matters:

(a)  his failure to take into account the impossibility of soil above groundwater level being carried into the Site;

(b)  his finding of “excessive” water was without sufficient evidential basis;

(c)  incorrect conclusions drawn by him from the fact that the Sinkhole was dry;

(d)  incorrect conclusions drawn by him from “post incident irregularities”;

(e)  improper finding that A3 had fabricated the water level and ground settlement records;

(f)  improper reliance by him on KK Pang’s explanation on the lack of conduit between Sinkhole 3 and the Site;

(g)  his taking into account irrelevant considerations relating to the lagging plates and the removal of grout;

(h)  his improper finding that welding could be carried out in wet conditions;

(i)  his failure to take into account a number of relevant considerations, e.g.:

(1)  Video evidence taken on 17 and 21 November 2014, showing clear water running into the Site.

(2)  The lack of any trace of concrete used for the filling‑in of Sinkhole 3 within the Site.

(3)  That sinkholes similar to Sinkhole 3 are not uncommonly found in Hong Kong at locations apart from nearby construction activity.

(4)  Pumping tests prior to the construction works showed that the Grout Curtain was constructed properly and there was no evidence or explanation as to how the Grout Curtain might have been damaged afterwards.

(5)  A material inconsistency between the two experts called by the prosecution.  According to KK Pang, the soil previously in the space occupied by the Sinkhole leaked through the Grout Curtain into the Site.  According to Dr Eric Li, the majority of the soil moved downwards round the toe of the Grout Curtain and then upwards through the base of the excavation.

(6)  The lack of any evidence of defect in the Grout Curtain to allow water and some 42m³ of soil to enter the cofferdam.

(j)  improper findings by him in respect of rain water from mountains/hills;

(k)  his overlooking/disregarding of the significance of the length of leakage and material non‑disclosure by the prosecution;

(l)  his improper finding on the seawall and culvert; and

(m)  his unjustified criticisms of the defence expert witnesses.

Ground 2

18.The learned magistrate erred in that:

With regard to A1 and A2 they: 

(a) were in default by disregarding monitoring systems concerning ground water level and ground settlement;

(b) had not checked the monitoring systems as they knew that there was nothing to check;

(c) had “Special Tasks specified in Form A to check everything provided by the monitoring systems”;

(d) had “first-hand knowledge” of excessive water leaking on 6, 13, 17 and 21 November 2014;

(e) ignored “signs of distress” by virtue of Sinkholes 1 and 2 occurring on the boundaries of the Site on 6 and 8 November 2014.

With regard to A1:

(f) He knew “the manner by which A3 carried out the building works”, “the flooding condition of the Site which was to be dealt with by way of four pumps continuously” and had failed to intervene,having “full knowledge of what was going on in the Site”.

With regard to A2:

(g) He “should have known the manner by which A3 carried out the building works and the fact that the Site was always flooded with water, came from the leakage on the grout curtain”.

(h) The duties assigned to him required him to “check a lot of things and to ensure they were in order”but he “never checked anything that might be inconsistent with the manner with which A3 carried out the building works”.

A3 and A4

Ground 1

19.The learned magistrate erred in accepting the evidence of KK Pang.  In particular, the learned magistrate erred in respect of the following matters:

(a) He wrongly assumed that the piping phenomenon had occurred without first ascertaining the existence of the conditions precedent to its application (such as the presence of fast moving underground water, the relative size of the soil particles vis‑a‑vis the size of the opening(s) and the location of the opening(s) etc, and consequently failed to notice the absence of these conditions precedent;

(b) He failed to take into account the inherent improbability of having 42m³ of soil above ground water level leaking into the Site by the piping phenomenon (see e.g. the evidence of Dr James Lau);

(c) His findings that over a period of three months (from 17 August 2014 to 17 November 2014) the Site was flooded with excessive water was without sufficient evidential foundation;

(d) There was no proper factual foundation or expert evidence to support the finding that the underground water leading to the alleged flooding primarily came through the cofferdam instead of the base of the excavation (which would be an innocent cause for the offences charged);

(e) There was no proper factual foundation or expert evidence to support the prosecution theory as to when and where the leakage occurred, in particular when there was no forensic examination as to whether there was in fact leakage in the cofferdam, and, if so, whether the soil from the outside was carried into the Site, and its extent;

(f) Incorrect conclusions or improper inferences drawn by him from the fact that Sinkhole 3 was dry on 17 November 2014 when there were other innocent competing inferences that could equally be drawn;

(g) Incorrect conclusions or improper inferences drawn by him from post-incident evidence of irregularities when there were other innocent competing inferences that could equally be drawn;

(h) Wrong or improper finding that A3 fabricated or authorised the fabrication of the water level records and ground settlement records;

(i) His finding that “Sinkholes 1 and 2 were caused by the similar way that caused Sinkhole 3 through excessive ingress of water allowed by the leaking grout curtain into the Site”, was without proper evidential basis.  In particular, this was inconsistent with the evidence of KK Pang and the results of the post-incident investigations carried out under the supervision of government officials from the Geotechnical Engineering Office (“GEO”) and/or Buildings Department (“BD”).  The prosecution had not established the nexus between Sinkholes 1 and 2 to the Sinkhole.  The circumstances of the present case were not such that the prosecution could irresistibly infer such nexus in the light of the (1) contradictory evidence given by KK Pang as to the probable cause of Sinkhole 1 (namely the result of heavy traffic at that location) and (2) the discovery of a pre-existing buried cavity at the northern side of the Site that is similar to Sinkhole 2 in size and depth, which could provide an alternative explanation for the occurrence of Sinkhole 2;

(j) He wrongly took into account and/or gave undue weight to the single incidence of temporary seepage of water caused by the removal of a portion of grout as a common, necessary and proper process of carrying out the ELS works;

(k) He failed to take into account a number of relevant matters, including:

(1)  the contemporaneous documentary evidence adduced by the prosecution which is contradictory to the evidence given by the prosecution witnesses or the prosecution case, namely, (i) the GEO or the BD officials did not identify nor express any concern that there was excessive water prior to 17 November 2014, or even shortly after the formation of the Sinkhole; (ii) the occasions on which the GEO or BD considered the ELS works were carried out in accordance with the approved plans; and (iii) prior to the incident the GEO and BD officials never related the incidents of Sinkhole 1 and Sinkhole 2 to the need to backfill sandbags or install spot grout behind the cofferdam;

(2)  video evidence taken on 17 and 21 November 2014 showing clear water (as opposed to muddy water) running within the Site;

(3)  that sinkholes similar to Sinkhole 1, Sinkhole 2 and the Sinkhole are not uncommonly found in Hong Kong at locations whether or not they are near to any construction activities;

(4)  material inconsistencies in evidence between the two prosecution expert witnesses.  According to KK Pang, the soil previously occupied by the Sinkhole leaked through the cracks in the Grout Curtain into the Site.  According to Dr Eric Li, the majority of the soil moved downwards round the toes of the Grout Curtain, then upwards through the base of the excavation site; and

(5)  the existence of unknown cause(s) of the two buried cavities found (at S12 and DH‑S12) in the northern side of the Site at different depths (one of which was found to be an old void) which undermined the prosecution case that the Sinkhole must have been caused by the piping phenomenon;

(l) His improper finding that there was excessive water in the Site based on the use of four pumps when there was no evidence as to when, why and for how long the pumps had been used.  See in particular, the evidence of Henry Ho, the structural engineer of BD who did not relate the pumping of the water at the Site to the previous ground subsidence event and did not expect sandbags to be stockpiled and grouting works being carried out at the scene prior to the formation of the Sinkhole;

(m) His wrong finding that the pumping test results were unreliable when they showed that the extent of the overall seepage via the cofferdam was within the acceptable parameters set out in the ELS plans and approved by the GEO and BD;

(n) The prosecution’s failure to prove that the cofferdam became ineffective causing the subsequent alleged abnormal seepage and the formation of the Sinkhole; and

(o) His failure to conclude that the prosecution was not entitled to draw the irresistible inference that the piping phenomenon must have been the cause of the Sinkhole when the prosecution failed to eliminate other possible cause(s) for the formation of sinkholes/buried cavities.

Ground 2

20.The learned magistrate erred in concluding that A3 and A4:

(a) were in default by disregarding the monitoring systems concerning groundwater level and ground settlement;

(b) ignored signs of distress revealed by Sinkholes 1 and 2 occurring on the boundaries of the Site on 6 and 8 November 2014 when they were not perceived by either the GEO or BD to be indicative of the existence of abnormal seepage via the cofferdam;

(c) they had “first-hand knowledge” of excessive water leaking from 17 August 2014 to 17 November 2014; and

(d) that the fact that nothing was done on the water leakage from the Grout Curtain on or before 17 November 2014, as no spot grout or sandbags were seen in the Site, constituted as a failure to carry out the precautionary measures when neither the spot grout nor sandbags were expected to be seen by the BD.

With regard to A3:

(e) that it “failed to install and maintain ground water monitoring systems properly and according to approved plans” and had “carried out such building works in such manner as was likely to cause a risk of injury to person or damage to property”.

With regard to A4:

(f) that he “should have known about the flooding condition of the Site”;

(g) that he “delegated such important engineering duty to TCP‑T1 who was not qualified to do so” and the “wrong delegation might have resulted in the problem of excessive ingress of water into the Site and also resulted in the neglect of signs of distress revealed by Sinkholes 1 and 2”; and

(h) that he “turned a blind eye to the manner of which D3 carried out the building works of ELS”.

21.There was also an application to amend the Perfected Grounds of Appeal of A3 and A4 dated 26 July 2019, to add a further ground in respect of non-disclosure of documents by the prosecution.  This was not pursued.

PROCEDURE

Appeals under s 113 of the Magistrates Ordinance, Cap 227

22.Unfortunately, the extensive grounds outlined above are framed in a manner more suited to an appeal to the Court of Appeal.  The parties seek, through their respective PGAs, to attack the conclusions of the learned magistrate in his Statement of Findings and the errors which they say he made in coming to those conclusions, most of which were based upon the expert evidence.  If this were the Court of Appeal, it would be required to find an error on the part of the magistrate before it could act.  However, this court is not the Court of Appeal and it carries out its role in a different manner.  The key distinguishing feature being that this court is required to make up its own mind on contested issues of fact and law, having regard to the evidence given in the court below.  This is an important distinction which needs to be recognised when drafting grounds of appeal.  Grounds which are drafted to indicate errors made by the magistrate are of less assistance to this court in carrying out its function.  This is because, following the decision of the Court of Final Appeal in Chou Shih Bin v HKSAR (2005) 8 HKCFAR 70, it is clearly established that an appeal under section 113 of the Magistrates Ordinance, is “by way of rehearing on the evidence before the trial court, supplemented by such further evidence as the intermediate court may admit under its statutory power to do so”.

23.On a rehearing this court is required to assess the evidence and come to its own conclusions as to whether that evidence proves the essential elements of the offence beyond reasonable doubt.  I can do no better than rely, as McWalters J (as he then was) did, in Histollo Pty Ltd v Director‑General of National Parks & Wildlife (1998) 45 NSWLR 661 at page 678 B – D:

“It is not incumbent upon the appellant to show that such conclusion was not open to the trial court in the sense of being a conclusion to which on the evidence, it should not have come. It is merely necessary to persuade this court on a review of the evidence that guilt is not, to the necessary standard, proved by the prosecutor. If, as at trial, on the rehearing there is a reasonable doubt of guilt, the appeal must be upheld.”

Limitations to appeal by way of rehearing

24.There are limitations to this form of hearing.  One such limitation is the content of the appeal bundle.  Given that it is as much for the prosecution to maintain the verdict of guilty as it is for the defence to seek to disturb the conviction below, the appeal bundle should contain that which the parties believe will meet that aim and enable the court to reach a verdict of its own.  A rehearing is therefore not a rehearing of all the evidence but only so much of the evidence as the parties determine is necessary for them and for the court to discharge their respective burdens.

25.There are other limitations.  There is normally no live evidence before the appellate court.  There was no live evidence before this court.  That limitation is particularly important in the present case where the magistrate rejected the evidence of the two expert witnesses for the defence, not simply on the basis of preferring the expert opinion of the prosecution expert KK Pang, but also on the basis of bias.  Appellate courts generally are exhorted to exercise caution before concluding that the trial court formed an erroneous view of the credibility of a witness[2]. To that extent it is important to consider the transcript carefully but also to give due credit to the Statement of Findings where it may assist with any concerns arising from the transcript and any supporting documentary evidence.

The Role of this Court

26.There is no dispute as to the relevant legal principles to be applied.

The Law

27.Neither party disputes that this is an offence of strict liability.  According to Lord Scarman in Gammon (Hong Kong) Ltd v Attorney General of Hong Kong [1985] AC 1 at p 19:

“The wording of s40(2B)(a) points to strict liability, once injury or damage has in fact been caused. Anyone who has carried out, authorised or permitted works to be carried out in a manner which has in fact caused injury or damage is caught”.

28.However, the prosecution must still prove:

(a) the defendants were persons directly concerned with the building works; and

(b) they carried out or authorised or permitted the carrying out of the building works; and

(c) the building works have been carried out in such a way that damage is done to property or injury caused to any person.

29.There is no dispute as to (a) or (b) above.  The issue is as to (c).  The issue is whether the “manner” in which the works were carried out “caused” the injury in question.

30.On the issue of causation, Patrick Chan J, as he then was, in R v Anwell Building Construction Co Ltd & Others, unreported, HCMA 590/1996, said:

“If the prosecution can show that the damage to property was caused by some negligence or deliberate act or omission on the part of the person who was concerned with the building works in the carrying out of such work, that would clearly be evidence sufficient to approve the causal link.”

He went on to say:

“However, I do not think that such proof is absolutely necessary. In fact, in many cases, proof of negligence and/or default would not be easy. This is because the accident must have already happened, the injury must have already been caused to some person or the damage must have been done to some property. In my view, it is not fatal if the prosecution cannot prove or adduce evidence of the cause of the collapse. The exact cause may never be known.”

And lastly and most importantly he said:

“But the circumstances may be such that the court can draw the irresistible inference that there must be some default on the part of those directly concerned in the carrying out of the building work which leads to the injury or damage to person or property.”

31.As a consequence, in analysing the facts before him he posed the question:

“In the absence of evidence showing the exact cause of the collapse, the question the court should ask is: were the circumstances of the case such that the court can draw the irresistible inference that the collapse of the formwork with its associated reinforcement bars and scaffoldings was caused by a default in the building works rather than some other reasons?”

32.That question, suitably reformulated, is at the core of the issue before the Court on this appeal i.e.,

In the absence of evidence showing the exact cause of the collapse of the Sinkhole, were the circumstances of the case, such that the Court can draw the irresistible inference that its collapse was caused by the manner in which the building works were carried out rather than some other reason or reasons?

Facts Not in Dispute

33.There were six sets of Admitted Facts and several lay witnesses.  It is not necessary to rehearse them here as they are not directly relevant to the arguments before this court on appeal.

34.There were a number of facts relevant to the geology of the Site and surrounding area, which were important to an understanding of the issue of how the Sinkhole was, or could have been, created.  In this respect, the magistrate referred to and relied upon a number of them, as do I.

(i)      Prior to building works, Inter Pacific Limited had recorded a GWL of between ‑1.88m and ‑2.43m (below GL).

(ii)     Building works began on 16 August 2013 and the GWL monitoring records for August to November 2013, showed a GWL of between ‑1.82m to ‑1.45m below GL.  In December it was between ‑1.73m and ‑1.19m below GL.  The highest pre‑accident GWL was therefore -1.19m and the lowest GWL was -2.43m below GL.

(iii)    The area in which the Sinkhole was situated, comprised of approximately 6m of fill overlaying 18‑20m of alluvium and then completely decomposed granite.

(iv)    The fill and alluvium layers were permeable and allowed for hydraulic conductivity whilst the decomposed granite layer was not permeable and did not allow for hydraulic conductivity.

(v)     Ground water travels faster horizontally than vertically in such conditions and water moving horizontally has driving force which is reduced if moving vertically or over a longer distance.  Hence the deeper the base of the cofferdam, the greater the reduction in the flow of any water seeking to enter the cofferdam via the base.

(vi)    There had to be a balance in the design of the cofferdam between effectiveness and cost.

(viii)  The design had been approved by the appropriate authorities but this would necessarily be subject to its being shown to be effective once built.

(ix)    If the Grout Curtain was effective during the ELS works, then the drop of GWL outside of the Grout Curtain should be small but it would not be zero.

(x)     The accepted level of drop was up to ‑1m.

(xi)    Some water would enter the cofferdam but would be pumped out again in order for works to be carried out in dry conditions within the cofferdam.

(xii)   If the GWL drops below the allowable amount because more water was allowed into the Site, further works were required to the Grout Curtain as it was the Grout Curtain which was designed to prevent excessive water ingress.

(xiii)  It was possible for water to go into the cofferdam through the Grout Curtain or underneath the base/toe of the Grout Curtain.

(xiv)  Pumping tests carried out after the construction of the Grout Curtain and before the excavation works began indicated the Grout Curtain was effective.

(xv)   The ELS works had reached -9m below GL at the time of the accident.  This was below GWL and therefore the pressure of the water outside of the Grout Curtain/cofferdam would increase the deeper the excavation.

(xvi)  The Sinkhole was 3.5m (w) x 2.5m (l) x 3.7m (d) and was dry.

Discussion

35.A lowering of the GWL outside of the Site could be indicative of ground settlement as the water level lowers.  Ground settlement may indicate a risk of the creation of a sinkhole.  Equally, water moving away from an area may have sufficient driving force to carry with it fine or lighter soil particles.  This too creates a risk of a cavity or cavities forming.  These were not the only ways in which a sinkhole could be created but it was accepted that a significant lowering of water levels or ground settlement indicated a risk against which building works of this kind seek to protect themselves.

36.One can see from this, that the task of the Court is to be satisfied that the Sinkhole was created by the manner in which the works were carried out rather than by some other plausible means unrelated to those works, given that the Court is being asked to draw an irresistible inference as to how it came to be created.

37.This Court, as did the magistrate, had to find certain facts proven in order to be able to draw the irresistible inference that the Sinkhole was created by the manner of the building works.  These facts centred around:

(a) the geology of the area;

(b) the nature of the Sinkhole;

(c) the nature of the building works;

(d) the amount of water on the Site; and

(e) the reliability of the GWL and GSL monitoring records.

The Geology of the Area

38.It was accepted that the ground upon which the Site was situated was made up of approximately[3] 6m deep layer of fill.  There was a 18‑20m layer of alluvium below that and thereafter decomposed granite.  The 6m layer of fill is the stratum of most interest to the Court as it is within this layer that the Sinkhole was formed.  The layer of fill consisted of further layers of materials.  It was said to contain a top layer of just over a metre of “fine to coarse sand with some angular to sub angular, fine to coarse gravel sized rock fragments”[4].  This layer went to a depth of +3.54mPD and was therefore approximately 1.1m in depth.  The whole of this layer must have gone into the Sinkhole, or moved out of the Sinkhole altogether, as it was the layer closest to the ground surface.  Under this layer was 1.28m of broken concrete debris down to a depth of +2.26mPD.  It is accepted that large and heavier debris would not move in water in the same way fine sand or even gravel might move.  This layer would have to fall within the Sinkhole or into any large adjoining cavities.  Below the 1m layer of concrete debris, was approximately 1.6m of loose, reddish brown clay e.g. silty fine to coarse sand with some angular to sub angular, fine gravel sized rock fragments down to a level of ‑0.76mPD.  The remainder of what should have been in the Sinkhole will come from this layer.  Some of this layer would have been heavier and larger and therefore less likely to move in water and some of it was silty fine sand and clay which could do so.  Thereafter there was a 4.5m layer of alluvium described as medium dense, pale brown, clayey, silty, fine to coarse sand.  This permeable layer, may or may not have been the means through which the water and/or any soil reached the Site.  This would depend upon whether the water passed into the Site via the toe/base of the Grout Curtain or through the Grout Curtain itself.

39.Surprisingly, there were no post-accident bore hole tests carried out by Victor Li & Associates immediately adjacent to the Sinkhole.  There was therefore no direct evidence before the Court of the ratio of silty sand, fine sand, or clay to gravel in the immediate vicinity.  Despite that, Dr Lau for A3 and A4, opined that 90% would be rock fragment or concrete debris (presumably referring to the 1.28m concrete layer), which would affect its ability to be washed away by flow of water.  This was a core part of the case of A3 and A4 as adopted by A1 and A2 in the appeal.  In simple terms they say, the constituents of the soil were too large to pass into the Site via piping.  As Miss Wong maintained, “the conditions precedent were not met”.  I am afraid I disagree.  If that was the case, the opinion of Dr Lau, would have been  equally applicable to the prosecution’s reliance on the “piping effect” as it was to the defence case that the soil in the Sinkhole moved because of leakage from drains and was deposited elsewhere, alternatively that it moved by reason of the natural flow of water from the hills to the harbour and was deposited against old sea walls.  Both of these defence hypotheses (as did that of KK Pang) required the lost soil to be “carried” by water.  If 90% was rock fragment or concrete debris, it was going nowhere.  Whereas, had the constituency of the fill layer at the site of the Sinkhole been proved to contain a greater percentage of the finer sand or finer gravel, its movement away from the Site would be accounted for under all of the hypotheses put forward.  Given that a cavity was created by some means or other, it follows that the soil moved somewhere and I am satisfied beyond reasonable doubt that the soil which moved was of such a consistency as to be able to do so.  In my view, the Inter Pacific Report supports this finding.

The Sinkhole

40.The Sinkhole into which Chan Shu‑moon fell was one of three in close proximity to the Site.  It was the last of the three to occur.  The other two have been referred to above.  They have no real significance to this rehearing save to indicate that sinkholes can be formed in and around a building site.  This was accepted.  These were two shallow sinkholes appearing on 6 November and 8 November respectively, shortly before Sinkhole 3 on 17 November, and are unlikely, in my view, to have been caused by the same phenomenon as the Sinkhole. There was a reference to other sinkholes in the area (in the Victor Li report relied upon by Professor Yeung) but these were not as close to the Site; were considerably smaller/shallower; above the GWL and the manner of their creation was more readily identified.

41.The Sinkhole in question was significantly different to any other sinkhole relied upon by any party.  The opening to the Sinkhole measured 1m x 1m, but the void within was approximately 3.5m x 2.5m and 3.7m deep (below GL).  It was by far the largest sinkhole referred to by any party during the trial.  It had the capacity to hold at least 42m³ of fast setting concrete, which was poured into it immediately after the accident.  As a result, it was not possible to ascertain whether there was one or more than one cavity into which the concrete was poured and how deep it was.  The evidence prior to filling it with concrete was that it appeared to be 3.7m deep immediately below the pavement.  It was just 2.3m away from the closest boundary of the Site.  Of particular significance was the fact that the Sinkhole was found to be dry even though it was 3.7m deep and the GWL should have been ‑1.73m below GL[5] at this point in time.  The question was therefore, where had the water gone, which should have been present between approximately ‑1.19 and ‑1.73m below GL and where had approximately 42m³, or more, of soil gone?  For that the Court considered the opinions of the four expert witnesses.

The Building Works

42.The ELS works were designed to take into account the inherent risk of water ingress.  There was no dispute that the Grout Curtain was effective within the design margins during the pumping tests of January 2014 prior to works commencing.  Thereafter, four bore piles had to be drilled and constructed to take the superstructure in due course.  That work finished in early June 2014.  The ELS works commenced on 17 June 2014 and were intended to be carried out in six layers.  Work was taking place on layer 5 at a depth of 9m below GL when the accident occurred.  At that stage there was an expanse of soil and Grout Curtain, which had not yet had the benefit of the lagging panels welded onto it.  It was accepted that in November 2014, water was coming into the Site from somewhere.

Water Ingress

43.There was a substantial dispute about water ingress.  It is clear that no one disputes there would be some ingress.  The cofferdam design allows for an acceptable drop of GWL of up to 1m and for the pumping out of water and re-charging of that water to the outside of the Site.  There was a possibility that water would enter the Site via the base/toe of the Grout Curtain.  There was also a possibility that the manner in which the pipe piles were drilled and the grout tubes were positioned allowed for some space between the soil filled with grout and the lagging plates.  If the Grout Curtain was not effective this would allow water to come through the Grout Curtain and then flow down along the plates to the bottom of the last layer of welding as opposed to coming up through the ground at the base/toe of the Grout Curtain.  There was a further possibility that as the excavation proceeded at 0.5m intervals, the face of the Grout Curtain (without the lagging plates) would be exposed between and behind the pipe piles and leakage could occur at this stage as well.  If any such leakage was not dealt with at that stage, then it would continue despite the fact that the lagging plates had been welded into place.

44.It was also accepted that water pressure would build up the deeper the excavation went below GWL.  Layer 5 was well below GWL.  The excavation was taking place within the alluvial layer, which was permeable and more water could be forced through it and into the cofferdam at that level.  This was alleviated by the depth to which the Grout Curtain had been constructed, for the deeper the water had to travel vertically to the base of the Grout Curtain the less force it could maintain.  However, only a grout curtain down to the decomposed granite level would eliminate all water seepage.  This was not the case here.  The Grout Curtain went down to ‑14mPD, which was considered by all to have been an acceptable balance between effectiveness and cost at the time of design.

45.On 17 November, there was a significant amount of exposed soil observed by PW 2 and 3.  They maintained that the Site had water problems when they visited on 17 November immediately after the accident.  The ingress was sufficient to require four pumps working continuously to dewater the Site.  Their evidence was accepted by the magistrate and I accept it.

Monitoring Records

46.Appendix A (p 61) of KK Pang’s first report includes a graph taken from the various monitoring records which takes a GL of +4.421mPD.

47.Monitoring equipment was required to ensure that the GWL and GSL did not fall below levels referred to as Alert, Alarm and Action.  Those levels were as follows:

	Alert	Alarm	Action
GWL PP1-3	0.5m	0.75m	1m
GSL GS3-6	12mm	18mm	25mm

It follows that there is no dispute that the appellants had a duty to monitor and act once the “Action” level had been reached.

48.PP1‑3 and GS3‑6 were the means by which the appellants could monitor the GWL and GSL around the Site.  PP1 and GS 5 and 6 were closest to the Sinkhole.  PP1 should have consisted of a standpipe (PP1‑S) to monitor water level and a piezometer (PP1‑P) to measure water pressure[6]. GS 5 and 6 were positioned close to PP1.

49.The graph indicates that the GWL appears to be consistent as of 10 March 2014, at just under approximately +2.8 to +3.2m mPD i.e. ‑1.2m and ‑1.6m below GL.  There are no records from 31 March to 16 June 2014.  The records recommence on 16 June immediately before the ELS works.  The variation thereafter is negligible and appears to be identical as between monitors on different parts of the Site.  Records from after the accident up to the end of January 2015 indicate a similar small variation.

50.On 11 March 2015, PW 2 and 3 visited the Site and noted that PP1 had been re-positioned west and inside the Site.  PP1 was found to contain only one pipe and not two, yet the records for the preceding week indicated readings from two pipes.  According to Wong Yu Che, PW4, there had been only one pipe in PP1 in January 2015, when he visited the Site.  However, there were also two sets of records for January 2015 from PP1.

51.The GS records indicated a settlement of just 2mm between August 2013 and November 2014, despite the fact that there was obvious ground settlement in the adjacent Sinkhole at some stage during this period.

Expert Evidence

52.The opinion of an expert witness is admissible to furnish the court with information which is likely to be out‑with the experience and knowledge of the court.  The fact that someone is called and accepted by the court as an expert witness does not give that evidence any enhanced status.  The court may accept it or reject it.  It may accept part only where it sees fit, so long as there is good reason for so doing.  The magistrate went to some lengths to give due credit to the expertise of all four expert witnesses.

53.It is widely accepted that expert evidence presented to the court should be and should be seen to be the independent product of the expert uninfluenced as to form or content by the exigencies of litigation[7]. Cresswell J in the Ikarian Reefer went on to say inter alia that an expert should provide independent assistance to the court by way of objective unbiased opinion in relation to matters within his expertise.  An expert witness should never assume the role of advocate.  The magistrate took the view, despite accepting their expertise, that Professor Yeung and Dr Lau did just that.

54.The prosecution called two expert witnesses as to the formation of the Sinkhole, Pang Kwok Kay (“KK Pang”), PW15 and Dr Eric Li, PW16 (“Dr Li”). KK Pang’s first report [P81] is dated 14 October 2015 in respect of his investigations carried out between January and May 2015.  He also submitted a response dated 3 October 2016 [P87].  Dr Li’s report [P83], upon the papers, is dated 13 November 2015 and his response is dated 12 December 2016 [P85].  There is no necessity here to explain their different hypotheses save to say that KK Pang’s opinion was preferred by the magistrate and is preferred by this Court.

55.A1 and A2 called one expert witness Professor Albert Yeung, DW1, with reports dated 2 August 2016 [D1/2-A] and 5 June 2017 [D1/2‑B].  A3 and A4 called one expert witness, Dr C W Lau, DW2, with a report dated 2 August 2016 [D3/4-A] and response dated 10 June 2017 [D3/4-B1].  All of the expert reports were admitted under section 65B Criminal Procedure Ordinance, Cap 221.

56.There were several further reports which were adduced during the trial, three of which are of particular significance to this Court’s findings.  These were an Inter Pacific Limited report of January 2013; a report and supplemental report by Victor Li & Associates dated January 2015 and March 2016, and what is known as the BUDA report dated November 2014.  These three sets of reports have been adduced to support the opinions of the defence experts.  They are not expert reports in and of themselves, and were not adduced nor accepted as such by the magistrate.  Accordingly, where those reports have been relied upon for opinion evidence rather than the results of tests undertaken, that evidence is inadmissible as it was not adduced for that purpose or accepted as expert opinion.  Nevertheless, this Court notes that a significant amount of the factual content of these reports were relied upon by all parties, as were a number of reports by various government agencies.  In summary, the Inter Pacific Report of 21 February 2013 covered the state of the Site pre‑works.  The Victor Li Report covered ground investigations post‑accident and the BUDA report looked at local utilities and possible leakage.

57.There were a number of hypotheses from all four gentlemen, many of which were ruled out on the facts by the experts themselves or by other experts; some of which were still relied upon by the defence on appeal but which could not be maintained if the defence case in respect of KK Pang was correct.  Nothing will be served by explaining that in detail here.  They were rightly rejected.   The magistrate listened very carefully to them all and was at pains to ensure that he had understood each theory and its premise. It was not necessary for him, in rejecting any one (or several) of them to explain his reasoning in detail in his Statement of Findings.  The magistrate made it clear that whilst he accepted some of the evidence of three of the expert witnesses, he had chosen to rely upon the opinion of KK Pang.  It was open to him to do so, as it is for this Court.  An expert witness has no special status.  In this instance the magistrate listened attentively to all that was said and clearly analysed each case put to him and sort to engage in that case with the expert witness with varying degrees of success.  His frustration in not receiving a satisfactory answer to some of his questions is evidenced in the transcript.  Not providing a sufficient explanation in answer to a question from the bench must be laid at the door of the witness and at the door of the relevant counsel.  This Court notes that the appellants’ argument often relied upon the phraseology of the magistrate in his Statement of Findings which, upon closer inspection of the transcript, was not made out when taken in context of the whole of the evidence given and the discourse between witnesses, bench and bar, which is the inevitable consequence of cases involving expert witnesses.  In any event, the exercise before the Court now is not one of a minute analysis of the Statement of Findings.  This Court must draw its own conclusions about the evidence.  A perusal of the transcript immediately indicates why certain theories did not hold good after examination.  I was greatly assisted by the interchange between the bench and the expert witnesses in this regard.  In this case, the magistrate rejected the opinion evidence of both defence experts and, to some extent, also that of Dr Li. He did so after a considerable period of time listening to them and seeking clarification, which at times he did not receive.

58.Despite his apparent total rejection of the evidence of Professor Yeung and Dr Lau for other reasons, there was, in fact, a considerable amount of expert evidence from both Professor Yeung and Dr Lau, (as well as Dr Li), which was relied upon by the magistrate even though he favoured the hypothesis put forward by KK Pang.  There was consensus on a number of issues e.g. that it is not uncommon to have leakage through a Grout Curtain.  On more than one occasion, the magistrate remarked upon their expertise in a positive manner and expressly relied upon it.  I see no reason why this Court should not do the same with one caveat.

59.The magistrate’s criticism was not as to expertise, but as to bias when it came to their opinion as to how the Sinkhole was created.  Given that he had the benefit of listening to and engaging all of the witnesses first hand and this Court having had the benefit of the transcript thereon, I am not prepared to interfere with that view[8].  It necessarily means that this Court will be cautious of that evidence but it has nevertheless taken it into consideration in coming to its own conclusion on the facts as to the issue of causation.

60.For my part, taking all of the evidence into consideration, I do not accept the alternative explanations for the creation of the Sinkhole advanced by either Professor Yeung or Dr Lau, or for that matter, Dr Li.  I do not find them mutually exclusive either.  Of all of the explanations advanced, I, like the magistrate, find the argument as to “piping effect” the only reasonable cause of the creation of the Sinkhole on 17 November 2014.  That is because I, like the magistrate, am satisfied beyond reasonable doubt that the water which appears to be flowing into the Site was excessive.  For that I rely upon the evidence of Ivy Chung when she stated “...there was much, much, much water”.  Although the flow rate of the pumps is not known, there were four of them and they were all in operation when the Site was visited by PW 2 and 3 on more than one occasion.  No matter where the water was coming from, or how it was coming into the Site, there was a consistent flow of water requiring four pumps to dewater the Site.  If that was the case, one would expect the GWL outside of the Site to lower to some degree and the records do not indicate that.

61.It is common ground that there was no water in the Sinkhole and so the court can find as proved that the GWL was at, or below, ‑3.7m.  If it was indeed below -3.7m, that would be a drop of GWL in the region of 2m.  Where had that water gone?  What had caused the water level to drop?  Why was it not recorded?  There was no other obvious and irresistible answer for a drop in GWL as great as that which was found in the Sinkhole.  All of the alternative options had inherent inconsistencies and were not supported by either the Victor Li or BUDA reports.

62.Even if the drop in GWL in the Sinkhole was not caused solely by the loss of water into the Site (catastrophically or over a period of time), the amount of water actually entering the Site created a risk of ground settlement.  Dr Lau was adamant about the consequences of the lowering of ground water levels upon ground settlement.  It does not of itself create a sinkhole but it does increase the risk of formation of a sinkhole because it increases the risk of the creation of a cavity.  That risk should have been evident from the excessive inflow of water and steps should have been taken to ascertain why and from where.

63.It is not necessary, nor is it possible, for this Court to come to any finding on exactly how that inflow of water occurred or in exactly what manner it entered the Site.  Was it all via the toe/base of the Grout Curtain; through the Grout Curtain and down behind the lagging panels or through the expanse of exposed soil in layer 5?  This Court cannot say whether the Sinkhole was caused wholly by the works on the Site or by any combination of factors including the works on the Site, which led to the final collapse of the pavement.  It cannot say whether there was one or more than one cavity into which the concrete debris fell.  But the only reasonable inference which can be drawn is that it was the manner in which the works were being carried out just 2.3m away that caused the final collapse of the pavement and the injury to Mr Chan.  It also finds that those working on the Site, as well as the monitoring equipment, should have been alert to and picked up on the excessive rate of water ingress at such a deep level.  There should have been a correlative and noticeable draw down and re-charge of GWL outside of the Site and yet there was no such record.  If the equipment was positioned as the appellant’s suggest and if it was working satisfactorily, the lowering of the GWL to the east of the Site should have been recorded.  Any lowering of the GWL, may very well have affected ground settlement and if it did, that too should have been identified by the monitoring equipment if it was in place and working.  The records suggest that readings were taken yet there is no indication of either the lowering of the GWL (which there must have been) or the GSL changing (which, again there must have been).  Why not?  Even if this had been a single catastrophic incident and not one which developed over time, the monitoring equipment should have picked it up on 17 November 2014.  What else was it there for?  It did not.  This Court does not know why it did not do so and is not going to jump to any conclusions as to why it did not do so.  The fact is that it did not and it should have and it was the responsibility of the appellants to ensure that it could do so.  There is little point of having a warning system which does not warn.

64.No one takes issue with the fact that all of the appellants bear some responsibility for that fact.  Even if the equipment was not working, the amount of water entering the Site was significant.  It should have put the appellants on notice that there was an increased risk of ground settlement and steps should have been taken to investigate it and to reduce the inflow of water and thereby reduce the risk of ground settlement.  That was not done.

CONCLUSION

65.This court poses a question similar to that of Patrick Chan J in Anwell supra:

In the absence of evidence showing the exact cause of collapse of the soil underneath the pavement and subsequent formation of the Sinkhole: were the circumstances of the case such that this Court can draw the irresistible inference that the formation of Sinkhole 3 was caused by a default in the building works on the Site rather than some other reason(s)?”

66.Having considered all of the evidence from both lay and expert witnesses and having taken into account the view of the learned magistrate as to the quality of that evidence, I am of the view that there are sufficient proven facts from which the Court can draw the irresistible inference that the Sinkhole was formed by a default in the building works and therefore that the prosecution have proved beyond reasonable doubt that the creation of the Sinkhole on 17 November 2014, was caused by the manner in which the ELS works were carried out on the Site.  I, like the learned magistrate, do not accept the alternative hypotheses put before the court by the defence and valiantly improved upon by Miss Wong in her submissions before me.  I do not have to be sure of the exact cause of the formation of the Sinkhole but there is sufficient evidence before the court from which this court can infer the causal link between that formation and the Site.  It is not necessary for this court to form a view as to how it was that excessive water came to be flowing into the Site but it does find that there was excessive water and that there was significant dewatering.  This should have given cause for concern about the GWL and GSL outside of the Site, which in turn should have given cause for concern about the apparent lack of indication of any significant drop in GWL or GSL on the relevant records, if they existed, and if they were checked.  It is not necessary for this court to go as far as to suggest some form of conspiracy as between A1‑4 in order to do so.  The manner in which the works were carried out and supervised, lead the Court to the overwhelming and irresistible conclusion that monitoring equipment was not working properly and that excessive amounts of water were coming into the Site, which required significant pumping out and that action alone should have put the parties on notice if they were carrying out their duties in a satisfactory manner.  They were in default by not taking action to stem the flow and ascertain any damage caused thereby.

67.I am satisfied beyond reasonable doubt of the guilt of each of the appellants and therefore dismiss their appeals and confirm the magistrate’s decision.

68.There be a costs order nisi that the costs of and occasioned by this appeal and the trial below be paid by the appellants to the respondent to be taxed if not agreed with certificate for two counsel.

Ms Alice Chan SADPP and Mr Pierre Lui SPP, of the Department of Justice, for the respondent

Mr Peter Duncan SC, Mr Calvin Cheuk and Mr Ernie Tung, instructed by MinterEllison LLP, for the 1^st and 2^nd appellants

Ms Maggie Wong SC, Mr Eric Chung and Mr Lawrence Pang, instructed by Deacons, for the 3^rd and 4^th appellants