Competition Commission v. W. Hing Construction Co Ltd and Others
Read the full judgment text of CTEA 2/2017 on BabelCite. This CTEA judgment was delivered on 17 May 2019.
1. These proceedings concern decoration works undertaken for the tenants of particular flats in three buildings in Phase 1 of On Tat Estate, Kwun Tong. The Competition Commission (“ Commission ”), the applicant herein, alleges that between June and November 2016, the 10 respondents made or gave effect to an agreement or engaged in a concerted practice among themselves, whereby each of them was allocated designated floors in each of the three buildings in the Estate to carry out decoration work
Cited by 1 case · Cites 12 cases
|
CTEA 2/2017 [2019] HKCT 3 IN THE COMPETITION TRIBUNAL OF THE HONG KONG SPECIAL ADMINISTRATIVE REGION COMPETITION TRIBUNAL ENFORCEMENT ACTION NO 2 OF 2017 ____________
____________
_____________________ J U D G M E N T _____________________ TABLE OF CONTENTS
A. INTRODUCTION 1.These proceedings concern decoration works undertaken for the tenants of particular flats in three buildings in Phase 1 of On Tat Estate, Kwun Tong. The Competition Commission (“Commission”), the applicant herein, alleges that between June and November 2016, the 10 respondents made or gave effect to an agreement or engaged in a concerted practice among themselves, whereby each of them was allocated designated floors in each of the three buildings in the Estate to carry out decoration work for individual tenants. It is also alleged that the respondents made or gave effect to an agreement or engaged in a concerted practice whereby they jointly produced a promotional flyer distributed to tenants setting out certain package prices used or offered by the respondents for the basic packages in the first instance. 2.The Commission contends that these arrangements contravened the first conduct rule in the Competition Ordinance (Cap 619) (“Ordinance”) in that they had the object of restricting competition and that, furthermore, they amounted to “serious anti‑competitive conduct” in the form of market sharing and price‑fixing respectively. The Commission seeks a declaration accordingly and an order for pecuniary penalty against each respondent. 3.The respondents, who were companies, partnerships or individual traders, were decoration contractors approved by the Hong Kong Housing Authority (“HKHA”) for Phase 1 of the Estate under its decoration contractors registration system, although, as explained below, tenants who wished to engage decoration contractors were not confined to the respondents. 4.The respondents are as follows:
5.The respondents’ stance may be summarised as follows:
6.Having regard to the parties’ contentions, the following broad issues arise for determination:
7.These issues will be considered in turn in the following sections of this judgment which deals with the question of liability only. B. BACKGROUND 8.This section sets out the relevant factual background as found on the evidence. B1. HKHA’s Decoration Contractor System 9.HKHA is a statutory body which seeks to provide affordable rental housing to low‑income families with housing needs, and help low-income to middle‑income families to gain access to subsidised home ownership. Eligibility for public rental housing is subject to the applicants’ income and assets being within specified limits. 10.At all material times since 1982, HKHA has operated a Decoration Contractor System (“DC System”) for new public housing buildings[2] aimed at protecting tenants by “preventing infiltration of illegal elements so that decoration works in new estates and courts can be carried out in an orderly manner”. Under the system:
11.Each Appointed DC has to enter into the Licence which is an agreement by deed with HKHA, with the title “Licence for Tenants’ Decoration Works”. By this, HKHA grants to the Appointed DC the right to enter the estate and carry out decoration works on behalf of tenants. While there are obligations attendant upon the appointment, the fact that a contractor is so appointed brings significant advantages, as described below. B2. On Tat Estate 12.On Tat Estate is a public rental housing estate in Kwun Tong comprising 11 residential blocks. The intake of tenants was carried out in four phases. Phase 1 includes 3 buildings, namely, Chun Tat House, Shing Tat House and Oi Tat House. Each of the 3 buildings has 40 floors above Ground Floor, and typically 22 flats on each floor. There are four types of flats of differing sizes: 1‑2 person, 2‑3 person, 1‑bedroom, and 2‑bedroom. The distribution of these types of flats on each floor varies and depends on the particular floor and the particular block. There are a total of 2,582 flats in the 3 buildings. Like other public housing estates in Hong Kong, the Estate is not enclosed by walls or fences and is therefore easily accessible from outside. 13.The flats are finished to a basic standard upon completion of construction. There is no partition of bedrooms. Unlike the kitchen and bathroom which are tiled, the floor in the living area is paved with concrete and not tiled. 14.The management of the Estate was outsourced by HKHA to Nice Property Management Co Ltd (“Nice Property”). B3. Appointment of Appointed DCs 15.On 3 August 2015, when Phase 1 was about to be completed, HKHA sent a letter to each of the 10 respondents (presumably because they topped the priority list) advising them that they had been selected to be Appointed DCs and inviting them to attend a briefing session to be held on 20 August 2015. Only 10 contractors were selected based on the ratio of one contractor to 250 flats. All the 10 respondents had confirmed their acceptance of the appointment in writing by 6 August 2015. B4. Briefing sessions 16.On 20 August 2015, HKHA held a briefing session at its headquarters for the respondents regarding On Tat Estate which was attended by the representatives of the 10 respondents. HKHA staff gave a presentation that set out certain principles and requirements. The respondents, as Appointed DCs, were informed that “any malpractice is subject to regulatory actions” and were reminded in particular that they should not “agree among themselves to allocate flats to a certain contractor, commonly referred to as ‘pie‑sharing’, but should allow the tenants to choose freely”. They were also told that during the course and upon the completion of decoration works, they should clear and remove at their own cost all the refuse and surplus materials and dispose of them at the refuse station designated by HKHA or the management company. 17.At the Police’s request, Nice Property invited the respondents to attend another meeting on 30 December 2015 (without any HKHA staff present). A manager from Nice Property reiterated the rules for the respondents including the preparation and wearing of identification tags and the report of illegal conduct. The Police’s Anti‑Triad Section also addressed the attendees on the prevention of triad activities and crimes. 18.At this meeting, at the request of Nice Property, the respondents chose two persons as their representatives for communication with Nice Property, being Mr Wong Fu San of Luen Hop (R10) and Mr Chan Yiu Kwai of Hip Yick (R6). B5. Site Office 19.In March 2016, Nice Property arranged for the respondents to make a site visit to the Estate to choose the location of their temporary office. On 30 May 2016, the respondents by a joint letter to HKHA and Nice Property applied to set up their temporary office outside Shing Tat House (“Site Office”). The site was provided by HKHA for a licence fee of $30,000 per month but the office itself and the facilities such as electrical fittings were arranged by the respondents themselves at their own expense. The use of a joint site office is not in any way criticised by the Commission in these proceedings. 20.Construction of the Site Office commenced on 7 June 2016 when Nice Property took over the Estate’s management, and was completed on 16 June 2016, one day before tenants intake began. There was a sign above its entrance which read 「安達邨一期裝修部」 (“Decoration Office of On Tat Estate Phase 1”). There were 10 desks inside the Site Office, in 2 rows of 5 desks on each side. Each desk was occupied by an Appointed DC with its business name displayed on the wall behind the desk. The respondents held a ceremony to worship deities at the Site Office on 16 June 2016, after which they drew lots to allocate the seats in the Site Office. As will be elaborated below, there is evidence that the lots drawn determined not only the seating but also the floors in each building allocated to each of the respondents. B6. Tenants intake 21.HKHA started issuing letters to public housing applicants in around November 2015 notifying them of the units in On Tat Estate that would be allocated to them, so that they could complete the process of accepting the offer by a specified date. Construction of Phase 1 was completed at around end of May or early June 2016. Nice Property began issuing letters to tenants, about 10 days prior to the intake registration process, with details of the relevant procedures. The intake registration was arranged in batches, with registration for around 100 units on each day. It was organised in the order of Chun Tat House, Shing Tat House and Oi Tat House, but otherwise not in any floor by floor sequence. The staff of Nice Property would call the respondents’ joint representative in advance to let him know the day on which and the building for which registration would take place. 22.Beginning from 17 June 2016, the tenants attended On Tat Estate on their respective appointed dates to complete the registration process, which involved the following steps:
23.The intake of tenants of Phase 1 was completed in August 2016. On 22 September 2016 the respondents gave notice to quit the Site Office. The lease for the Site Office was terminated and the Site Office was vacated and demolished by the respondents on 31 October 2016. B7. The Flyer 24.For the purposes of the decoration works in Phase 1 of On Tat Estate, a joint flyer was produced by the respondents (“Flyer”). It was headed “On Tat Estate (Phase 1) Decoration Contractor” and listed 10 items of work as follows, which appear to be the types of basic decoration works typically desired by tenants:
25.The Flyer then set out the following service packages and prices respectively for the 4 types of flats, with their floor plans shown (“Packages” and “Package Prices”):
26.Below the floor plans and Package Prices were the following words:
27.A copy of the Flyer together with its English translation are attached to this judgment as Appendix 1. B8. Decoration of the flats in the Estate 28.Between June and November 2016, out of the total of 2,582 flats in the 3 buildings, the respondents carried out decoration works for 867 flats. It is not entirely clear how many tenants moved in without decoration, or decorated the flats by themselves or their friends or relatives, or engaged commercial decoration contractors other than the Appointed DCs (“outside contractors”). 29.It has since been discovered that the distribution of the vast majority of the flats decorated by the 10 respondents fell into a strikingly regular pattern. According to this pattern, each respondent worked on 4 and only 4 floors in each of the 3 buildings. The 4 floors each respondent worked on in each building are each 10 floors apart, so that the last digit of the floor number is constant. The pattern is shown in the following table:
30.832 of the 867 flats worked by the respondents follow this pattern, while 35 flats do not. The number of flats worked by each respondent on each floor in each building and the number of flats that fall outside this pattern are shown in the 3 tables appended to this judgment as Appendix 2. C. LEGAL FRAMEWORK 31.The first conduct rule is found in s 6 of the Ordinance, which provides:
32.The first conduct rule focuses on agreements, concerted practices, and decisions of associations of undertakings. An “agreement”, as defined in s 2(1),
As the present case is in the ultimate analysis concerned with agreements only, it is unnecessary to deal with concerted practices or decisions of associations. The concept of “undertaking” is dealt with below in connection with the sub‑contractor defence raised by W Hing (R1) and Wide Project (R9) (and which Tai Dou (R4) wished to raise). There is no dispute that each of the other respondents is an undertaking. 33.The first conduct rule prohibits any agreement that has an anti‑competitive “object” or “effect”. It is common ground that these two elements are disjunctive. The Commission’s case is that the agreements had the object of preventing, restricting or distorting competition in Hong Kong, and that it is therefore unnecessary to investigate their concrete effects. 34.Section 7(1) and (2) supplements s 6 in relation to object by providing that:
35.Section 30 provides that the conduct rules do not apply where they are excluded by Schedule 1. Schedule 1 contains several exclusions. The relevant one in this case is s 1 of Schedule 1 (quoted in §152 below), which is concerned with agreements that contribute to the improvement of production or distribution or the promotion of technical or economic progress. This exclusion is relied upon by some of the respondents for their “efficiency defence”. 36.The Ordinance contains the concept of “serious anti-competitive conduct”, which is defined in s 2(1) to mean:
The significance of the classification of conduct as “serious anti-competitive conduct” is explained in §125 below. 37.The first conduct rule may be enforced by the Commission by bringing proceedings in the Tribunal. Section 93(1) provides:
The maximum penalty, by s 93(3), is 10% of the turnover of the undertaking concerned for each year in which the contravention occurred or, where the contravention occurred in more than 3 years, 10% of the turnover of the highest 3 years. 38.Because pecuniary penalties are sought in these proceedings, the Commission accepts that they involve the determination of a criminal charge within the meaning of Art 11 of the Hong Kong Bill of Rights (s 8 of the Hong Kong Bill of Rights Ordinance (Cap 383)). On that basis, the respondents submit that the Commission has to prove its case beyond reasonable doubt, whereas the Commission submits that the civil standard of proof (ie balance of probabilities) should be applied when the Commission seeks to discharge its burden to establish a contravention of the first conduct rule, subject to cogent evidence being required commensurately with the seriousness of the allegations. 39.The issue of the applicable standard of proof in proceedings in this Tribunal in which the Commission applies for an order for a pecuniary penalty under s 93 has been extensively discussed in the Tribunal’s judgment in Competition Commission v Nutanix Hong Kong Ltd and Others [2019] HKCT 2, which is handed down at the same time as this judgment. Essentially I take the view that this Tribunal is bound by the Court of Final Appeal’s decision in Koon Wing Yee v Insider Dealing Tribunal (2008) 11 HKCFAR 170 to proceed on the basis that, once the proceedings are properly regarded as involving the determination of a criminal charge for the purposes of Art 11 of the Bill of Rights, proof beyond reasonable doubt is applicable unless modified by proportionate legislation by express provision or necessary implication. Since the Ordinance does not expressly or implicitly mandate a different standard, the standard of proof beyond reasonable doubt should be applied. 40.The burden and standard of proof in relation to the “efficiency defence” in s 1 of Schedule 1 will be discussed in section G2 below. D. FLOOR ALLOCATION ARRANGEMENT D1. The pleas 41.The Commission alleges that from around June to November 2016, the 10 respondents made and/or gave effect to an agreement and/or engaged in a concerted practice, whereby the respondents were each allocated 4 floors in each of the 3 buildings of On Tat Estate in accordance with the pattern set out in §29 above, and the respondents (a) would not actively seek business from tenants on floors allocated to other respondents; (b) would not accept business from those tenants; and (c) would direct those tenants to the respondents who had been allocated their floors (“Floor Allocation Arrangement”). 42.In their pleadings, all the respondents, except W Hing (R1) and Wide Project (R9) who made no admission, denied there was any arrangement to allocate floors. By the time of the trial, they contended that there was an agreement among the respondents to allocate floors for the cleaning of which each respondent was responsible (“Cleaning Agreement”).[3] D2. The pattern 43.The stark pattern of the distribution of the flats decorated by the respondents, as set out in §29 above and in Appendix 2 below, is strong evidence that there had been deliberate design and concertation rather than free and normal competition. In each building, each respondent worked exclusively or almost exclusively on 4 and only 4 floors. The 4 floors that each respondent worked on in each building were each 10 floors apart so that their last digit was always the same. Over 95% of the 867 flats decorated by the respondents fell into the pattern; only 35 flats (less than 5%) were “out of pattern”. 44.Even though I take the view that the respondents would have achieved some concentration of flats on certain floors without any coordination among themselves (as to which see §§232-247 below), the floors on which they won their first deals respectively would not without prior design be exactly 10 floors apart for each respondent, and it is also highly improbable that they would each have concentrated on 4 and only 4 floors in each of the 3 buildings. It is in my judgment inconceivable that the distribution found here could have been arrived at by anything other than an agreement (which, for present purposes, need only be a mutual understanding). There is no plausible basis for this division other than a pre‑ordained arrangement amongst the respondents. 45.The “out‑of‑pattern flats” can be readily accounted for by the evidence that where the tenants themselves insisted, the respondents were permitted to take on work on floors not allocated to them. There may also be other special reasons such as, as mentioned in evidence, where a tenant was referred to a particular respondent by their common friend or relative. In any event, the fact that less than 5% of the flats fell outside the pattern is in my view manifestly insufficient to detract from the inexorable inference. D3. The tenants’ evidence 46.The Commission adduced evidence from 8 tenants which was largely unchallenged by the respondents.
D4. The respondents’ evidence 47.While many of the respondents’ witnesses denied the Floor Allocation Arrangement on the stand, some actually admitted its existence. Their admission is not confined to their own firms but extends to all the respondents without exception. Thus WC Wong, a witness and the alleged sub‑contractor of W Hing (R1), frankly admitted in cross‑examination that on the day of the worship ceremony, the respondents drew lots to allocate the floors they would be working on. They would each receive a piece of paper, which stated a number in relation to each building eg “1” for Chun Tat House, “4” for Oi Tat House, and “8” for Shing Tat House, and these would be the last digit of the floors on which the respondent would work in each building. 48.Likewise, Lo Yiu Ming, a witness as well as a director of Kam Kee (R5), had admitted in his voluntary interview with the Commission on 8 March 2017 that on the day of the worship ceremony, the respondents agreed to divide the market by floors and that they would not try to seek business on the floors not allocated to them. A statement by him was subsequently filed on behalf of Kam Kee (R5) which claimed that he did not understand what he was being asked during the interview, and that when he mentioned “distribution” of the floors at the interview he had meant to refer to the Cleaning Agreement. I am not impressed by this attempt to resile from his answers. It is clear from the record and the unchallenged evidence of the interviewer that Lo gave his answers well knowing what he was being asked and also entirely voluntarily. His answers would not make sense in the context of the alleged Cleaning Agreement. When cross‑examined at the trial, Lo effectively accepted there was an understanding among the respondents that they would each concentrate on the floors that they were allocated, for all purposes. 49.Both WC Wong and Lo Yiu Ming testified that there had been a long history of such “pie‑sharing” in the decoration of flats in public housing estates in Hong Kong. 50.Poon Wai Woon, a partner of Wai Sun (R8), also admitted in a cautioned interview by the Commission on 13 March 2017 that there was an agreement among the respondents to divide up the floors. When asked, as an example, why his company mainly did decoration works on 1/F, 11/F, 21/F and 31/F of Oi Tat House, he said there was an agreement among the respondents (known also to their staff) as to the floors they were to work on, so that they “shouldn’t compete with each other”. If Wai Sun (R8) was approached by a tenant of a floor not allocated to it, it should first tell him to go to the company which had been allocated that floor; but if the tenant did not wish to go to that other company and wanted to engage Wai Sun (R8) instead, that would be permissible. This applied also in the converse situation: where a tenant of a floor allocated to Wai Sun (R8) approached another respondent of his own volition, he would first be told to contact Wai Sun (R8). When asked about these statements in cross‑examination, Poon, who appeared quite nervous throughout his testimony, mainly evaded the questions and at times accepted there was a market allocation agreement and at times denied it. Unlike Lo Yiu Ming of Kam Kee (R5), Poon’s witness statements filed in the proceedings did not seek to cast any doubt on the accuracy of his answers in the interview. 51.Some other respondents’ witnesses denied the Floor Allocation Arrangement in their evidence, including Chim Lap Fan (R2), Chan Kam Shui (R3), Chan Yiu Kwai (R6), Cheung Ping San (R7), Wong Fu San (R10) and Ma Yick Yin (who worked in the name of Tai Dou (R4) in relation to the decoration works in On Tat Estate but was called to give evidence by the 2nd, 3rd, 5th, 6th, 7th, 8th and 10th respondents). They all said there was no agreement among the respondents to divide up the floors for the purposes of getting decoration business but instead there was a Cleaning Agreement. Their evidence is discussed below in connection with the alleged Cleaning Agreement. 52.I should mention that the supplemental statements of Chim Lap Fan (R2), Lo Yiu Ming (R5), Chan Yiu Kwai (R6), Cheung Ping San (R7), Poon Wai Woon (R8) and Wong Fu San (R10) and the statements of Chan Chun Wah (R3) and Ma Yick Yin all contain the following paragraph:[4]
53.The last sentence in this quotation seems to me to be a clear admission that the Floor Allocation Arrangement alleged by the Commission was indeed entered into by the respondents. However, for reasons explained in §§205-210 below and as the Commission has fairly submitted, these “admissions” should be accorded no weight. D5. Cleaning Agreement 54.Sun Spark (R2), Mau Hang (R3), Kam Kee (R5), Hip Yick (R6), Tai Wah (R7), Wai Sun (R8) and Luen Hop (R10) contend that there was no market sharing agreement among the respondents but a “Cleaning Agreement”. This was described in the witnesses’ evidence as an agreement whereby the respondents drew lots at the worship ceremony to determine both the seating in the Site Office and the designated floors they were each required to clean up. It was said that agreements like the Cleaning Agreement had been used in other public housing estates for many years. Under the Cleaning Agreement, each respondent, as an Appointed DC, would be responsible for cleaning the debris and construction waste on specific floors and would be accountable to the management office if cleaning problems arose. The purpose of the agreement was to assign responsibilities for cleaning up waste and debris left by contractors in the common areas such as corridors and stairways of the buildings. Without the Cleaning Agreement, no one would voluntarily clean up such waste and debris, which would cause obstruction and fire hazard. 55.It was further said that since the respondents were required under the Cleaning Agreement to clean up the respective designated floors, each of them independently and separately decided that it should target the floors allocated to it and focus on tenants of those floors for decoration business as well, for example by distributing flyers on those floors. This was done purely to save costs and that each of them did not object to the other respondents competing on its floors. 56.For a number of reasons explained below, I do not find the Cleaning Agreement a credible story. First, it is significant to see how the story evolved. It was not mentioned by many of the respondents’ personnel interviewed by the Commission in 2017, who simply denied that there was any agreement at all. Further, the reasons, where given, for not having mentioned the Cleaning Agreement in the interview, are not consistent. Chim Lap Fan (R2) said in his witness statement that he did not mention it in the interview because it was a common practice in the industry and he thought it was a trivial matter and he was not asked about it by the Commission. In oral evidence, however, he said he did not mention it in order to avoid suspicion on the part of the Commission and to avoid stepping on a “landmine”. When asked about this inconsistency, Chim claimed it was possible that his witness statement, which was made in English, was not sufficiently clearly explained to him by his former lawyers. 57.In his interview on 26 June 2017, when asked why Mau Hang (R3) only decorated flats in Chun Tat House on 3/F, 13/F, 23/F and 33/F, Chan Kam Shui said “[p]erhaps it’s just a coincidence”.[5] While he discussed cleaning obligations, he specifically denied there was any agreement to allocate floors for cleaning purposes.[6] 58.In his first interview by the Commission on 8 March 2017, Chan Yiu Kwai (R6) simply denied there was any agreed floor allocation and did not mention any Cleaning Agreement. In his second interview on 23 May 2017, he said there was an agreed division of the floors for cleaning waste and debris, but when asked why the floors allocated to each one were 10 floors apart, he said it would be fairer to allocate the floors in cycles rather than for consecutive floors to be allocated to a respondent because, for example, 1/F to 3/F are “not good” as there might be many elderly tenants there who would not spend much money on decoration. This is a telling answer because if the division was done to allocate responsibility for cleaning only, those floors should be very welcome for they should generate much less construction waste. In cross‑examination, Chan was unable to explain this answer except to say that it would be fairer to make it more even among the contractors. 59.Cheung Bing San (R7) never mentioned the Cleaning Agreement in his interview with the Commission on 9 March 2017, although he denied that in oral evidence without any basis. He mentioned in his witness statement the drawing of lots to determine seating arrangements in the Site Office but when asked why there was no mention of the Cleaning Agreement, he said he was not asked the question. It seems to me that in his oral evidence he tried to seize every opportunity to give a “practised” answer that there was a Cleaning Agreement involving allocation of floors for removing debris. 60.Wong Fu San (R10) said in his interview on 8 March 2017 that many meetings of all 10 respondents were convened with the management office when the latter complained about the existence of debris on a specific floor. Of course, if the Cleaning Agreement existed, such meetings would have been unnecessary for Wong would know which respondent was responsible for the cleaning of that floor and could immediately notify that respondent to remove the debris. Wong also said in his second interview that the respondent who happened to be doing more flats on a floor would be asked to clean up accumulated debris on that floor where the source was unidentifiable. He never mentioned either in his two interviews or in his first witness statement any agreement for the allocation of floors to the respondents for cleaning purposes. 61.It would be remarkable, if there was a Cleaning Agreement as alleged leading to the respondents taking up the decoration work on agreed allocated floors, that none of these interviewees had mentioned it to the Commission who was investigating the cause of the striking pattern of distribution of the flats decorated by the respondents. 62.Secondly, there was still no mention of the Cleaning Agreement in the pleadings, after the respondents had obtained legal representation. The agreement was not mentioned in the Response jointly filed by the 2nd, 3rd, 4th, 6th, 7th, 8th and 10th respondents[7] or the separate Response filed by the 5th respondent. The joint Response dated 26 September 2017 states:
63.It is pleaded that the drawing of lots was done for determining the seating arrangement in the Site Office (§26), and there is a mention of cleaning in §29(h) which pleads:
but this is plainly not the Cleaning Agreement subsequently contended for by the respondents. 64.Likewise, under the alternative plea of enhanced overall economic efficiency, §32(d) pleads “Enabling collective cleaning up debris and waste” but does not plead any Cleaning Agreement. 65.The Response of Kam Kee (R5), even more starkly, states (at §11(2), see also §13(6), which have remained the same after an amendment in February 2018):
66.The total lack of a pleaded case of the Cleaning Agreement is in itself a sufficient reason why the explanation for the floor distribution pattern now advanced by these respondents cannot be accepted. But just as importantly it also reveals the lack of credibility of the belated assertion of the Cleaning Agreement. 67.Thirdly, of the respondents’ witness statements produced in April 2018, only those of Chim Lap Fan (R2), Chan Kam Shui (R3), Ho Kit Chung (R4) and Lo Yiu Ming (R5) made reference to the Cleaning Agreement; none of Chan Yiu Kwai (R6), Cheung Ping San (R7), Poon Wai Woon (R8) or Wong Fu San (R10) mentioned it. For the latter group of witnesses, they raised the Cleaning Agreement for the first time only in their supplemental statements dated 4 October 2018 which, for reasons explained in §§205-210 below, can be given little weight. 68.Fourthly, the existence of the Cleaning Agreement is directly denied by WC Wong, the witness of W Hing (R1) and thus not part of the group of respondents who had raised that defence. Further, in cross‑examination Poon Wai Woon (R8) said there was no agreement between the respondents as to which floors they should clean up on — an answer he confirmed in re‑examination. 69.Fifthly, the property manager from Nice Property gave evidence that no such arrangement as the Cleaning Agreement was made known to it. His evidence was not challenged by any respondent; nor did any respondent give evidence that Nice Property was informed. It is difficult to comprehend why Nice Property did not know about it if there was a Cleaning Agreement, since its purpose was supposedly to meet the demands of the management office for the clearing up of accumulated debris in the common areas. 70.Sixthly, the lack of clarity in the parameters of the alleged Cleaning Agreement militates against its existence. For example, each respondent was subject to the requirement under clause 9(h) of the Licence not to leave any refuse of whatever kind in the common areas. Clause 5 of the standard terms and conditions which every Appointed DC was required to print on the back of its work orders also provided that the contractor was required to “clear and remove at its own costs all the rubbish, refuse and surplus materials relating to the decoration works as and when they arise during the course, and upon the completion, of the decoration works”. There was nevertheless some suggestion that under the Cleaning Agreement, the respondents would be responsible for clearing up other respondents’ waste on the same floor. How the Cleaning Agreement interacted with the obligations under the Licences and work orders was nowhere credibly explained. It was also wholly unclear whether the cleaning responsibilities continued after a respondent had finished all its decoration work on a floor. 71.Seventhly, the suggestion that each respondent assumed responsibility for cleaning up building waste on particular floors (including waste generated by outside contractors) even if the respondent did no work on those floors is, as accepted by Chim Lap Fan (R2), not rational or commercially sensible. In the case of Chim’s company, namely Sun Spark (R2), there were actually 3 such floors (2/F and 32/F Chun Tat House and 35/F Oi Tat House) as shown in Appendix 2. 72.Under the rules of the Estate, the tenants and their decoration contractors were responsible for taking their own construction waste to the refuse collection point at the ground floor of each building. They were prohibited from leaving any construction waste in the corridors or lift lobbies. It is not surprising that there might be breaches from time to time. Where that happened, if the source of the construction waste was identifiable, the tenant and decoration contractor responsible would be required to dispose of it. If the source could not be identified, Yee Hop, the cleaning company engaged by Nice Property for the Estate, would remove the waste to the collection point. From there lorries arranged by Nice Property would transport the waste to a landfill site. The cost of this service was covered by the “construction waste fee” that tenants were required to pay upon registration. 73.Accordingly, as submitted by the Commission, the contention that there was a free‑standing Cleaning Agreement is deeply implausible and is to be rejected. I accept that the respondents’ workers gave help to Yee Hop’s cleaners from time to time but this does not support the Cleaning Agreement. What is far more likely to have happened is that, because there was the Floor Allocation Arrangement, and pursuant to that arrangement a respondent was decorating a number of flats on a floor (to the exclusion of the other respondents), when there was debris in the common areas whose source was unidentified, it was understood that the respondent might on occasions be expected to help remove it. In other words, as was the evidence of Poon Wai Woon (R8), it was because a certain respondent was doing most of the work on a floor that he would be expected to clean up debris on that floor, not the other way round. 74.Finally, it should be added that even if the respondents expressly agreed to allocate floors only for cleaning work and not for decoration business, there could still have been an infringement. This is because on their own case it was to be expected that they would each concentrate on their allocated floors for the purpose of acquiring decoration business. As such, the Cleaning Agreement may arguably amount to a concerted practice among the respondents by which they knowingly reduced the uncertainty with regard to the conduct and strategy of one another. D6. Conclusions and findings 75.For all these reasons, I reject the respondents’ contentions. The denial of the Floor Allocation Arrangement and the assertion of the Cleaning Agreement are specifically rejected. The evidence in my view establishes beyond any doubt the factual case advanced by the Commission. 76.This Tribunal finds that:
E. PACKAGE PRICE ARRANGEMENT E1. The pleas 77.The Commission’s case is that from around June to November 2016, the respondents made and gave effect to an agreement, or engaged in a concerted practice, whereby (1) the respondents would jointly produce the Flyer where the Package Prices would be printed and which would be used by the respondents for advertisement to the tenants; and (2) the Package Prices would be the prices used or offered by the respondents for the Packages in the first instance. This is referred to as the “Package Price Arrangement”. 78.W Hing (R1) simply makes no admission to the allegation. 79.Wide Project (R9) denies that he authorised or participated in the production or use of the Flyer but this is based on the sub‑contractor defence. He adduced no positive evidence to the effect that his sub‑contractor did not agree with the production of the Flyer or did not use it. 80.In their joint Response, Sun Spark (R2), Mau Hang (R3), Tai Dou (R4), Hip Yick (R6), Tai Wah (R7), Wai Sun (R8) and Luen Hop (R10) admit that “the promotional Flyers were used by the Respondents” and plead that “[t]he use of the Flyers was for promotion only”; “[t]he prices stated in the Flyers were for reference only”; “[t]here was room for … price adjustments”; “[t]he use of the flyers showing the Packages Prices was intended as a general reference in negotiation with the Tenants … The aim of setting such reference was intended as a provision of basic information to the Tenants”; “[t]he reference pricing is an ancillary restriction which falls outside the scope of the First Conduct Rule …”; and “[s]uch arrangement was an objective necessity for initial negotiation between the Respondents and the tenants”. 81.In the particulars of their Response, given by their solicitors’ letter dated 2 March 2018 in relation to the suggestion that the “reference pricing” is an “ancillary restriction”, it was stated: “The main agreement is to facilitate the initial negotiation between the Rs and the tenants regarding price and scope of works. To [achieve] such objective, the Rs coordinate with each other to provide Package Prices to various types of works”. 82.Kam Kee (R5) denies the Package Price Arrangement and further pleads that “the Respondents only printed the Flyers for marketing purposes only”, and that it “never adopted the Basic Packages as a starting point and/or anchoring reference point for negotiations with Tenants”. E2. The evidence 83.There is evidence from Chim Lap Fan (R2), Chan Kam Shui (R3), Chan Yiu Kwai (R6) and Cheung Bing Sang (R7) in their witness statements, from WC Wong (R1) in his oral evidence, and from Wong Fu San (R10) in his interview by the Commission and his oral evidence at trial, that the respondents all agreed that the Flyer was to be produced jointly for use by all the respondents. Wong Fu San and Chan Yiu Kwai as the representatives were responsible for its production. Wong Fu San obtained sample flyers from other public housing estates for reference and adopted their design and the contents of package items. The two representatives set the Package Prices in the Flyer, without receiving any costs information from the respondents individually. All the respondents shared the cost of printing. Indeed, the Flyer itself plainly represented that it was published by all the respondents, as it stated “Our decoration division is one of the 10 decoration contractors on the register of the Housing Authority, all of which are governed by the Housing Authority” and gave the Site Office as their common address. 84.While there was some suggestion (eg from WC Wong (R1)) that the respondents were not consulted about the prices before printing, I think that was mistaken and the preferable evidence was that of Chan Yiu Kwai (R6) and Cheung Bing Sang (R7) who said there was a brief discussion among the respondents before they agreed to bulk‑print the Flyer. In particular, in his interview by the Commission on 8 March 2017, Wong Fu San (R10) recalled with some satisfaction that all the Appointed DCs agreed with the contents of the Flyer, which was not always the case in his experience with other estates. 85.The main dispute raised by the respondents was whether and the extent to which they actually used the Flyer in dealing with the tenants. 86.On this question, the Commission was able to point to 23 work orders in which the respondents applied the precise Package Prices to the Package items as shown in the Flyer. This is strong indication that the Flyer was used leading to the conclusion of those orders. These 23 work orders belonged to 8 of the respondents (except Mau Hang (R3) and Tai Wah (R7)).[8] Moreover, an analysis by Dr Sharon Pang, the respondents’ economics expert, of the 233 work orders performed by the respondents other than W Hing (R1) and Wide Project (R9) in Chun Tat House shows that around 59% have the same 10 Package items that were advertised on the Flyer and pre-printed on the work orders. 87.The witnesses’ evidence also shows that the Flyer was used by the respondents. WC Wong (R1)’s evidence was that where a tenant adopted one of the Packages, the Flyer price would be used as the starting point and would become the actual price if the tenant did not ask for a discount. If the tenant did ask, a discount would be given, but the records of the actual transactions show that the discount given was small, in the region of 2%. 88.In his witness statement, Chim Lap Fan (R2) said that the Flyer prices would be used as the starting point with the tenants but that his firm would try to give as much discount as possible. In his oral evidence, however, he tried to distance his firm from the Flyer, saying that it was rarely used by the staff and that his statement was written by his former lawyers. He also said that Sun Spark (R2) had its own prices for the Package items to use as a starting point ($6,000 in the case of 1‑2 person flat, which was higher than the Flyer price of $5,380). I did not find this evidence credible, which was contrary to his statement. As stated in his statement, “Sun Spark did distribute flyers to the 1st and 3rd floors of Chun Tat House after received [sic] orders from room 122 and 307. This was the main reason Sun Spark received another order from Room 304”. He also had to admit that Sun Spark (R2) had agreed to work orders for the Package items with precisely the Flyer prices and there were occasions when the customer himself drew attention to the Flyer prices. Indeed, in relation to 2 flats which were charged the Flyer prices, his statement stated “the tenants ordered a package only, therefore Sun Spark did not give the tenant any discount”. 89.Chan Kam Shui (R3) said in his statement that “[t]he two packages were only a starting point for the tenants who usually add or subtract items from them to meet the specific demand for their flats. I tried not to stick to the price listed in the flyer because the flyer was merely used as an advertisement to attract potential customers. I tried to give as much discount to my customers as possible …”. This was adopted as his evidence‑in‑chief. His bare denial in cross‑examination that the Flyer prices were used to start negotiations on the Package items does not seem credible to me. 90.Lo Yiu Ming (R5) said in his statement that the Flyer was printed “as a way of promotion only, as we know the tenants would get flyers and compare prices” and that “the package prices stated in the flyer are only a matter for promotional purposes”. His statement that “Kam Kee did not follow the price stated at all” is however hardly credible since, as he had to accept, there were several work orders of Kam Kee (R5) that matched the Flyer prices exactly. He accepted that in those cases the Flyer prices were taken as a reference point, though he suggested that the prices were proposed by the tenants themselves, having obtained the Flyer. 91.In his statement, Chan Yiu Kwai (R6) said that the Flyer prices “were for general reference purpose” but “the reference value of flyers was limited”. He did not dispute that Hip Yick (R6) distributed the Flyer for promotional purposes. Although he claimed in oral evidence that his company did not use the Flyer very often, he accepted there were flats done by Hip Yick (R6) where the prices charged matched precisely those quoted on the Flyer. He explained that he adopted the Flyer prices, but if the customer required additional work or change of materials, then the prices could be different. 92.In the case of Tai Wah (R7), while in his statement Cheung Bing Sang said that the Flyer prices “were just for general reference only”, he also admitted that his staff would distribute the Flyers to the tenants and if a tenant did not bargain or put forward other requirements, his firm “normally would charge the price shown on the flyers”. 93.In his statement Poon Wai Woon (R8) said that the Flyers were “distributed to customers as promotion and the prices on the Flyers were set for reference(s) and would be adjusted according to customers’ requests such as the materials used, works required”. He however accepted there that: “If a customer only wishes to do the work specified in a package, Wai Sun would charge only the package price”. In oral evidence he also accepted that if the customers wanted something different from the Flyer, Wai Sun (R8) would charge a different price, “but otherwise [Wai Sun] generally use[d] the flyer prices for those particular items”. 94.In the case of Luen Hop (R10), Wong Fu San said that the Flyer was used “to attract the customers and for general promotion”, and that there were “about less than 10 job contracts out of the 90 contracts … concluded simply on the stated package price(s)”. In oral evidence, he said that every respondent distributed the Flyer. 95.In oral evidence, when asked by Ms Ng on behalf of Tai Dou (R4), Ma Yick Yin, who was acting in the name of Tai Dou at the time, denied that he used the Flyer. This is of course contrary to Tai Dou’s pleaded case that the Flyer was used for promotion and the prices stated were used for reference in negotiations with the customers. Moreover, Ma’s bare denial cannot in my view stand against the weight of all the other evidence including the objective fact that there were 3 work orders of Tai Dou where the prices charged for Package items exactly matched the Flyer prices. 96.In the case of Wide Project (R9), although no one handling the day‑to‑day decoration business in its name in On Tat Estate was called to testify at the trial, the available evidence shows that the Flyer was also agreed to and distributed by those working in its name. In the interview by the Commission on 30 March 2017, Mak Sau Kwan, who was a decorator that worked in On Tat Estate in the name of Wide Project, said that when asked by the tenants, she would give them the Flyer. She said that the Flyer prices were only for reference, and there was no fixed rule that those prices had to be used. It is, however, indisputable that Wide Project did have 3 work orders with precisely the Flyer prices for the Package items. E3. Conclusions and findings 97.In my judgment, it has been established to the requisite standard that:
F. INFRINGEMENT OF FIRST CONDUCT RULE BY OBJECT F1. General principles 98.Various jurisdictions have seen the need to simplify the application of competition law and provide legal certainty by laying down bright‑line rules where appropriate. Thus certain “hard-core” anti‑competitive practices are regarded under United States law as “illegal per se” under the Sherman Act.[9] Collusion between competitors has been called by the US Supreme Court “the supreme evil of antitrust”.[10] In Australia also, the law is that certain practices (especially cartel conduct between competitors) are regarded as so inherently damaging to competition that they are treated as illegal without further inquiry. As the High Court of Australia stated in Rural Press Ltd and Others v Australian Competition and Consumer Commission (2003) 216 CLR 53, at §82:
99.In the competition law of the European Union (and in jurisdictions which have adopted the EU model), no conduct is classified as illegal per se. Instead, as far as Art 101 of the TFEU[12] is concerned, the distinction is drawn between agreements with the object and effect of preventing, restricting or distorting competition. The language in Art 101(1), in particular the phrase “object or effect”, has since the earliest days of EU competition law been held to mean that there is no need to examine the effects of an agreement if it is found to have the object of harming competition: Case 56/65 Société Technique Minière v Maschinenbau Ulm GmbH[1966] ECR 235, 249. 100.Several rationales for a rule prohibiting agreements that restrict competition by object have been articulated. First, such prohibition is justified by the fact that “certain forms of collusion between undertakings can be regarded, by their very nature, as being injurious to the proper functioning of normal competition”: Case C‑209/07 Competition Authority v Beef Industry Development Society Ltd, §17. Identifying such coordination for prohibition achieves a deterrent effect and helps to prevent such conduct: per AG Wahl in Cartes Bancaires at §35. 101.Another justification, related to the above, is that, as AG Kokott said in Case C‑8/08 T‑Mobile v Commission at §47, the prohibition of infringements by object:
102.Thirdly, such a rule creates legal certainty, provides predictability and allows all market participants to adapt their conduct accordingly: per AG Kokott in T‑Mobile at §43; AG Wahl in Cartes Bancaires at §35. Uncertainty in the law imposes direct costs on society through the need for complex legal advice and the increased likelihood of lengthy, complicated legal proceedings. 103.Fourthly, the rule also sensibly conserves resources of competition authorities and the justice system: per AG Kokott in T‑Mobile at §43; per AG Wahl in Cartes Bancaires at §35. As a corollary, without prohibition of infringement by object there will be a considerable increase in the burdens of investigation and proof on the regulator even in the case of such infringements, which are already serious in themselves. This will severely impair the effectiveness of Art 101: C‑407/04P Dalmine v Commission, §137. 104.In the first conduct rule in the Ordinance, which is modelled on EU law, the same bifurcation into object and effect is adopted. The purpose of adopting this language, and the mischief that would follow an alteration of this formulation, were explained to the Bills Committee during the legislative process as follows:
105.The meaning of restriction of competition by object has been much discussed in EU law which provides useful guidance in the application of the first conduct rule. There is no dispute between the parties that the current leading EU case in that regard is Case C‑67/13P Groupement des Cartes Bancaires v European Commission. The approach set out there has been confirmed by the CJEU (Grand Chamber) in Case C‑179/16 F. Hoffmann‑La Roche Ltd v Autorità Garante della Concorrenza e del Mercato. The position in EU Law displays these elements:
106.The parties do not dispute the applicability of these principles in connection with the first conduct rule. To them must be added the two statutory rules in s 7 of the Ordinance: first, for an agreement with more than one object, it is sufficient if one of its objects is to restrict competition; and secondly, the object of restriction of competition may be found even if it can be ascertained only by inference. These provisions reflect the established position in relation to Art 101 of the TFEU. 107.Because it potentially permits undertakings to be sanctioned for infringement of competition law without examination or proof of the concrete effects of the agreement, the concept of infringement by object needs to be interpreted restrictively: Cartes Bancaires, §58; F. Hoffmann‑La Roche, at §78. 108.Paying regard to the legal and economic context of the agreement as referred to in Cartes Bancaires at §53 is, however, to be distinguished from an effects analysis. To conflate the two would undermine the benefits and defeat the purpose of proscribing restriction of competition by object in the first place. As AG Wathelet stated in §67 of his opinion in Case C‑373/14P Toshiba Corporation v Commission, the economic and legal context “is there to assist the authority responsible for examining the alleged restriction by object to understand the economic function and the real significance of the agreement”. What is required is an examination of the agreement in its proper context in order to ascertain whether it amounts to a restriction by object. It does not involve a full examination of market effects or any balancing of pro‑ or anti‑competitive effects: Agents’ Mutual Ltd v Gaxcoigne Halman Ltd [2017] CAT 15, §149. 109.On behalf of Sun Spark (R2) and Mau Hang (R3), Mr Carter Chim submitted that what constitutes a restriction of competition by object remains a controversial topic in EU law and continues to be hotly debated. He submitted that this Tribunal should avoid the “mistakes” made in the EU and UK and avoid a “form based” approach to defining the “object box”. 110.It is true that there have been debates concerning the proper delineation of the category of infringement by object, but it seems to me, with respect, that agreements such as those for price fixing and market allocation between competitors are generally recognised to restrict competition by object. As Mr Beard submitted on behalf of the Commission, they are paradigmatic, not penumbral, cases. As horizontal pacts between competitors that directly exclude competition in respect of significant parameters in their operation, they are regarded as cartels and virtually universally condemned by every system of competition law. 111.It will moreover be noted that even in recent case‑law such as Cartes Bancaires and F. Hoffmann‑La Roche, the CJEU has continued to pronounce that there are certain types of coordination that are regarded, by their very nature, as being harmful to the proper functioning of normal competition. While the concept of infringement by object is to be narrowly interpreted, there are established categories of conduct that are generally expected to fall within it, of which price‑fixing and market sharing agreements are clearly prime examples. This is not to say that such agreements do not need to be examined in their context — they still do, but it would be fallacious to suggest that there is any great controversy and uncertainty in other jurisdictions as to whether they amount to cartel conduct between competitors to be condemned by competition law. 112.Mr Chim relied on, inter alia, AG Wahl’s opinion in Cartes Bancaires and suggested that where the undertakings concerned hold only a tiny share of the market, a horizontal price‑fixing agreement amongst them may not be highly harmful for competition. He appeared to suggest that the same observation would apply to all established hard‑core restrictions, including market sharing agreements. In reaching its judgment in Cartes Bancaires, however, the CJEU did not endorse or give effect to that opinion. Furthermore, in the EU, the Notice on agreements of minor importance which do not appreciably restrict competition under Article 101(1) of the Treaty on the Functioning of the European Union (2014/C 291/01), which covers situations where the aggregate market share held by the parties to the agreement does not exceed 10%, expressly does not cover agreements which have as their object the prevention, restriction or distortion of competition within the internal market;[14] see Case C‑226/11 Expedia Inc v Autorité de la concurrence and Others, §§35‑38. It is therefore clear that EU law has recognized that hard-core infringements are problematic even if the parties only have an aggregate market share of less than 10%. 113.Counsel further relied on Case C‑32/11 Allianz Hungária Biztosító Zrt and Others v Gazdasági Versenyhivatal where the CJEU’s judgment stated at §48 that the court should “in particular take into consideration the structure of that market, the existence of alternative distribution channels and their respective importance and the market power of the companies concerned”, and argued that there should be a detailed analysis in the present case that took into account, inter alia, the market power of the respondents. That was a case where car insurers entered into arrangements with car dealers on the rates payable by insurers for repairs to insured vehicles. Since the car dealers also acted as insurance brokers for the insurers, however, the dual nature of the arrangements could have effects injurious to competition on the car insurance market, particularly given that Hungarian domestic law required such brokers to be independent from the insurance companies. It is for the purpose of determining the likelihood of damage by the repairs arrangements to competition in the car insurance market that the CJEU made the statement in terms of §48 of the judgment quoted above. As AG Wathelet demonstrated in his opinion in Toshiba Corporation v Commission at §§84‑85, subsequent case‑law has confirmed “the specific and isolated nature of the judgment in Allianz Hungária Biztosító” and that those elements of assessment referred to in §48 of that judgment were “specific to that case and cannot be applied generally without giving rise to confusion between restrictions by object and restrictions by effect”. Similarly, it has been commented in Whish & Bailey, Competition Law (9th ed), p 130, that “the analysis suggested by [the CJEU in Allianz] seems to risk improperly intermingling object and effect analysis”; see also §51 of the opinion of AG Wahl in Cartes Bancaires. It has also been noted that the EU courts have not repeated that particular formulation in Allianz in subsequent cases: Bellamy & Child, European Union Law of Competition (8th ed), §2.151. 114.A further submission made, relying on academic writings,[15] was that once there is a plausible pro‑efficiency justification advanced for an impugned agreement, it should not be characterised as a restriction of competition by object and should only be condemned if found to be restrictive of competition after a full effects analysis. This approach, as acknowledged by Mr Chim, is demonstrably contrary to that hitherto adopted and consistently applied by the EU and UK courts. 115.In Case C‑209/07 Competition Authority v Beef Industry Development Society Ltd, which concerned the Irish beef processing industry (for slaughter and de‑boning of meat), studies carried out at the instance of the Irish Government and the industry concluded that it was necessary to reduce the number of beef processors. Pursuant to these conclusions, 10 principal processors formed a society to devise a scheme that involved reduction in processing capacity of about 25% and compensation to be paid by the undertakings which were to remain in the sector to those forced to withdraw. The Irish courts found that there was overcapacity in the industry for the processing of beef at 32%, and that this excess capacity would have very serious consequences for the profitability of the industry as a whole and would over time lead to very significant losses and ultimately to processors and plants leaving the industry. On a reference by the Irish Supreme Court to the CJEU, it was argued that the purpose of the arrangements was not adversely to affect competition or the welfare of consumers, but to rationalise the beef industry in order to make it more competitive by reducing, though not eliminating, production overcapacity. It was also argued that price increases would be ruled out on account of the economies of scale achieved through higher concentration and on account of the power on the purchasing side of the market. Nevertheless, the CJEU disagreed and held (at §21):
116.That an agreement may be held to have an anti‑competitive object even if it also pursues other, legitimate objectives is, of course, expressly recognised in s 7(1) of the Ordinance. 117.The CJEU stated further in the Beef Industry case:
118.As stated in Bellamy & Child, European Union Law of Competition (8th ed, 2018), §3‑012:
See also Case C‑382/129 Mastercard Inc v Commission, §§180‑181. 119.This is also the approach adopted by the Competition Appeal Tribunal in the UK. As stated by the Tribunal in Generics (UK) Ltd & Others v Competition and Markets Authority [2018] CAT 4, at §308:
120.Likewise, in Ping Europe Ltd v Competition and Markets Authority [2018] CAT 13 at §104, the Competition Appeal Tribunal stated:
121.I am not persuaded that this Tribunal should travel down “the road not taken” as suggested by Mr Chim, which, as the Commission submitted, will act against the wealth of jurisprudence which the legislature saw immense value in and which informs the approach to the first conduct rule. F2. Whether the Floor Allocation Arrangement was a restriction by object 122.It is plain that the terms of the Floor Allocation Arrangement, to which the respondents were all parties as found above, directly prevented competition between the respondents altogether. The flats in the 3 buildings, for which each respondent was entitled to solicit decoration business but for the agreement, were carved up by the floors, with each floor allocated to one and only one respondent. The aim of the agreement was evidently to prevent competition. 123.As regards the intentions of the respondents themselves, having regard to the analysis of the evidence above, there is no credible evidence that the agreement was entered into in order to pursue the alleged efficiencies. Nor is there any credible evidence that it was entered into in order to “counter the cost disadvantage the Appointed DCs were faced with vis‑a‑vis the outside contractors” as asserted in economists’ language in the respondents’ supplemental witness statements (quoted in §52 above). Had the respondents allocated floors with the object of achieving efficiencies, consecutive floors within a single building should have been allocated to each respondent as far as possible. This would not only avoid the need for workers to move between buildings and between floors served by different lifts, but would also facilitate the use of the stairs for vertical travel. The evidence is not consistent as to which floors were preferred by the Appointed DCs: there was a suggestion that they preferred the highest floors because the lifts were likely to be less crowded there; but Chim Lap Fan (R2) also said contractors preferred the lowest floors because it took less time to wait for the lifts there, though he later said what he meant was it was easier for the workers to walk down to the ground floor during lunchtime; a yet further suggestion from Chan Yiu Kwai (R6) was that the first few floors were not desirable because there were more elderly tenants there who were reluctant to spend money on decoration. Whatever the precise preferences might have been, the actual pattern of allocation reveals that the respondents’ concern was equality and fairness among themselves rather than efficiencies. 124.As submitted by the Commission, market sharing arrangements eliminate or reduce competition in respect of the aspect of the market being shared. It is a particularly pernicious form of anti‑competitive conduct: it removes or reduces choice for customers and insulates the suppliers from the dynamic effects of rivalry between themselves. 125.The serious nature of the distortion of competition by a market allocation arrangement has been recognised in the Ordinance by classifying such conduct (amongst others) as “serious anti‑competitive conduct”, with the result that (i) the exemption for agreements of lesser significance in s 5(1) of Schedule 1 is inapplicable; (ii) the requirement under s 82 for the Commission to issue a warning notice before bringing enforcement proceedings does not apply; and (iii) the Commission may, instead of bringing proceedings in the Tribunal in the first instance, issue an infringement notice to the person concerned under s 67 of the Ordinance. 126.The respondents have placed great store on the fact that out of 2,582 flats in the 3 buildings, they together secured contracts for “only” 867 flats. They have also argued that the outside contractors imposed such an effective competitive constraint upon the respondents that their agreement among themselves did not reveal a sufficient degree of harm to competition. 127.There are several points to observe in this connection. First, the respondents clearly took more than 33.58%[18] of the tenants that engaged decoration contractors, since there were tenants who moved in without decoration or decorated by themselves, friends or relatives. The precise percentage cannot be ascertained but given the level of finish in the flats upon completion (with kitchen and bathroom tiling, painted walls etc) it was entirely possible for some of the tenants, who by HKHA’s rules must be from the low‑income end of the population, to move in without decoration or with minimal works that could be carried out by themselves or friends or relatives. 128.There is limited contemporaneous evidence in the decoration contractors appraisal returns. One form recorded nil outside contractors as at 31 October 2016. That figure has been justifiably attacked by the respondents as likely to be inaccurate due to under‑reporting of the engagement of outside contractors, though it does throw doubt on whether their presence was anywhere near as prevalent as has been asserted on behalf of the respondents. What that form did clearly suggest, however, was that there was a significant degree of reported self‑decoration (542 flats). Notwithstanding this, for present purposes I am prepared to assume in favour of the respondents that outside contractors worked on around 55% of all flats (using the average of the range of 50% to 60% estimated by Chan Yiu Kwai (R6) even though in reality it might have been much lower), which would mean that the respondents, as Appointed DCs, had an aggregate 38% share by number of tenants who engaged decoration contractors in Phase 1 of On Tat Estate. 129.Secondly, the evidence has demonstrated that, as Appointed DCs, the respondents enjoyed clear advantages over outside contractors in attracting the patronage of the tenants:
130.The benefit of being an Appointed DC is also demonstrated by the respondents’ evidence that the status was something contractors “treasured” very much and that the sanction of removal from the DC System was perceived to be a real threat and penalty. 131.The respondents argued that as Appointed DCs they had to incur additional expenditure, which amounted to a cost disadvantage compared to outside contractors. Leaving aside the cost of the surety bond which would apparently be returned if the duties of the Appointed DCs were duly performed, the other costs for each respondent such as the rent for the Site Office ($3,000 per month for 5 months), the cost of setting up the Site Office ($18,000 to $20,000), air‑conditioning units ($2,400), signage at the entrance ($600), and the premium for the contractors all risks policy ($5,000) together amounted to $43,000 which is between 0.8% and 3.3% of each respondent’s revenues at On Tat Estate (Phase 1). For 8 out of the 10 respondents, this cost is less than 1.5% of revenues earned. This “cost disadvantage” is clearly outweighed by the attractions of being an Appointed DC. 132.The proof of the pudding is in the eating. It is significant that there was a “market price” of $200,000 for “sub‑letting” the Licence and that each of the two “sub‑contractors” in this case (for W Hing (R1) and Wide Project (R9) respectively) chose to take it up and pay $200,000 in addition to the various costs which they could otherwise avoid, even though there was nothing to prevent them from operating as an outside contractor to obtain business from tenants in the Estate. This underscores the immense value in being an Appointed DC as opposed to being an outside contractor, as WC Wong (R1) accepted in the following exchanges during cross‑examination:
133.On the available evidence, it is plain that the respondents would be the closest competitors of one another, as accepted by some of the respondents’ witnesses. The Floor Allocation Arrangement, by its very nature, had as its object precisely the elimination of such competition. The tenants were deprived of the benefit from competition between Appointed DCs to offer better prices, products or services. 134.While the outside contractors competed with the respondents in broad terms, they were in reality not considered by the respondents themselves to be their closest competitors. It cannot be assumed that simply because the outside contractors took 55% and the respondents 38% of the flats that engaged contractors, the outside contractors possessed market power and the respondents lacked it. No detailed and rigorous analysis of the market and market power has been conducted in the present case, taking into account, for example, capacity constraints and product differentiation. In the absence of a proper definition of the relevant market, crude “market shares” do not provide a good proxy of market power: British Telecommunications plc v Offers of Communications (Market Definition) [2017] CAT 25, §157. After all, the 10 respondents were able among themselves to secure 867 of the flats even when faced with competition from, on their own case, “hundreds” of outside contractors. In my opinion, the presence of outside contractors in this case does not exonerate the respondents’ collusion from being seriously injurious to the proper functioning of normal competition. 135.There is little doubt that market sharing agreements are generally regarded in EU law as restrictions of competition by object. In Case C‑373/14P Toshiba Corporation v Commission, decided after Cartes Bancaires, the CJEU stated:
136.It follows from the above analysis that the Floor Allocation Arrangement is properly to be characterised as an agreement that restricted competition by reason of its very object within the meaning of the first conduct rule. F3. Whether the Package Price Arrangement was a restriction by object 137.An agreement between competing suppliers to fix prices is of course a prime example of cartel conduct, and is generally considered in EU law as an agreement constituting, by its very nature, a restriction on competition by object: see eg Bellamy & Child, European Union Law of Competition (8th ed, 2018), §5.041. Price‑fixing is also classified within “serious anti‑competitive conduct” in the Ordinance: see §36 above. 138.By adopting the Flyer with the Package Prices, the respondents had plainly “made” an agreement to fix prices, even though their own evidence would suggest significant differences in unit costs. Arguments that the Flyer was not frequently used in practice may concern whether the agreement was “given effect to” and may raise grounds for mitigation but do not negative a contravention of the first conduct rule which prohibits undertakings from making or giving effect to anti‑competitive agreements. 139.In any event, the evidence shows that the Package Prices were reached by many respondents as the final contract prices in a number of transactions (see section E2 above). For tenants who did not bargain and simply agreed to pay the Flyer prices, the collusion among the respondents removed the benefit of price competition. An analysis by Dr Helen Jenkins (the Commission’s economics expert) of the respondents’ work orders for Chun Tat House showed that for the 10 identified work orders that consisted exactly of one of the Packages, the contract prices were indeed very close to the Flyer prices, and that the Package Prices did indeed serve as an anchor influencing the final prices charged to tenants. Dr Pang agreed that, at least in theory, the Flyer prices had the potential to influence final prices for tenants that ordered the Packages, although she emphasised that very few tenants ordered them without any modifications or add‑on items and that therefore the Flyer prices might have had a limited impact on the final prices offered to the tenants. 140.Even where the contract works varied from the Packages, the Flyer nevertheless acted as a starting or reference point for the negotiations, particularly since the Packages comprised basic items desired by many customers. An analysis by Dr Pang of the 233 work orders of the respondents for Chun Tat House showed that almost all of them (230) included at least one Package item and around 59% contained the same 10 Package items as printed on the Flyer. Although 192 included at least one “add‑on” item and for those that included some of the Package items there might be modifications to the details, this does not detract from the respondents’ evidence and indeed their pleaded case[19] that the Flyer provided “reference” prices. The inescapable conclusion is that a very significant proportion of all 867 tenants had been affected by the reference point set by the price fixing agreement, even if they had ordered modifications or additions to the Packages. 141.As Dr Jenkins explained, the use of common flyer prices among the respondents has the potential to raise final transaction prices by influencing tenants’ understanding of what the costs of and the prices should “within reason” be for decorative work items. Where consumers do not have complete information about producers’ costs, high list prices may signal to consumers that costs of operations are high, which has the potential to weaken consumers’ negotiation position and make them more willing to accept higher prices. The list prices may also serve as focal points by influencing buyers’ reference points or anchoring points before negotiation. 142.In any event, as the Commission submitted, it is generally regarded as unnecessary for proof of an infringement of competition law to show that the price‑fixing agreement has any influence on the actual selling price achieved. The reasons for this are twofold. 143.First, the agreed prices, even if they are only list prices subject to negotiations with the customers, may well serve as a starting point or an anchor for such negotiations. As a matter of common sense, the starting point will often affect the end point. Coordination of the list prices has therefore the potential of influencing the final transaction prices: Joined Cases T‑379&381/10Keramag Keramische Werke and Others v Commission, §§53‑59; Case T‑588/08 Dole Food Co Inc v European Commission, §551. This is also consistent with the approach taken by the United States Court of Appeal (9th Circuit) in Plymouth Dealers’ Association of Northern California v United States 279 F 2d 128 (9th Circuit, 1960) at 132 (footnotes omitted):
144.Secondly, agreement on list prices also allows members of the cartel to foresee with a reasonable degree of certainty what pricing policy will be pursued by their competitors: Case T‑410/03 Hoechst GmbH v Commission of the European Communities, §349; Case 8/72 Vereeniging van Cementhandelaren v Commission, §21. 145.Apart from trying to distance themselves from the Flyer and to minimise its significance and utility, the respondents have given no credible evidence of any legitimate objectives served by the Package Price Arrangement. The experts are agreed that it was simply a supportive mechanism for the Floor Allocation Arrangement. As explained in Dr Pang’s report, if the respondents had used their own individual flyers with different prices, they would not be able to maintain the Floor Allocation Arrangement. 146.It is therefore plain that the Package Price Arrangement constituted an agreement among the respondents that had the object of restricting competition in On Tat Estate (Phase 1). G. THE EFFICIENCY DEFENCE G1. Overview 147.Whilst the respondents have all denied being party to the Floor Allocation Arrangement or the Package Price Arrangement, a number of them have contended that, in the event the Tribunal finds these arrangements established, they should not be held liable for contravening the first conduct rule because the arrangements were economically efficient within the scope of s 1 of Schedule 1 to the Ordinance. This defence is pleaded in the joint Response of the 2nd, 3rd, 4th, 6th, 7th, 8th and 10th respondents and in the separate Response of the 5th respondent. However, the 4th respondent has since abandoned reliance on any efficiency defence in circumstances referred to in this Tribunal’s Reasons for Decision dated 31 October 2018.[20] References herein to the respondents who rely on this defence should accordingly be understood as references to the 2nd, 3rd, 5th, 6th, 7th, 8th and 10th respondents, though for convenience I shall simply refer to them as the respondents in this section. 148.Following EU and UK law, the Ordinance also provides for an “efficiency defence”. In the EU, this defence appears in Art 101(3) of the TFEU, which provides: “ The provisions of paragraph 1 may, however, be declared inapplicable in the case of:
149.A similar defence exists in UK law under s 9(1) of the Competition Act 1998 which provides that an agreement is “exempt from the Chapter I prohibition” if it satisfies certain conditions, which are very similarly worded to Art 101(3). 150.In the Ordinance, the defence is found in s 30 of and Schedule 1 to the Ordinance. Section 30 provides:
151.Schedule 1 contains a number of exclusions on different grounds. Sections 1, 2(1), 3, 4(1) and 5 are relevant to the first conduct rule. Section 1 concerns agreements enhancing overall economic efficiency; s 2(1) excludes agreements made for the purpose of complying with a legal requirement; s 3 concerns undertakings entrusted by the Government with services of general economic interest; s 4(1) excludes agreements resulting in a merger; and s 5 concerns agreements between undertakings with a combined turnover of no more than $200 million that does not involve serious anti‑competitive conduct. Only s 1 has been relied on by the respondents in this case. 152.Section 1 of Schedule 1 to the Ordinance, which is evidently modelled on Art 101(3) of the TFEU, provides:
153.It is common ground that this provision consists of four conditions which are cumulative so that each condition must be satisfied if the exclusion is to apply, namely, in broad terms:
154.By virtue of s 9(1)(a), an undertaking may apply to the Commission for a decision as to whether an agreement or proposed agreement is excluded from the first conduct rule as a result of Schedule 1. Sections 10‑14 provide for the machinery for the making and rescission of such decisions and their effect. Section 34 requires such decisions to be published on a register kept by the Commission. These decisions may be reviewed by this Tribunal on application under s 84 as “reviewable determinations”. Sections 15‑20 make provisions for the Commission to issue a block exemption order where a particular category of agreement is found to be excluded from the first conduct rule as a result of s 1 of Schedule 1. 155.The Commission does not dispute that an agreement may benefit from the efficiency defence even if it would otherwise infringe the first conduct rule by object, and even if it takes a form that falls within the meaning of “serious anti‑competitive conduct” as defined in s 2(1). The Commission submits, however, that a “hardcore” restriction of competition, such as market allocation or price fixing, would be unlikely to fulfil the conditions of the defence, as observed by the EU Commission in §46 of its Guidelines on the application of Article 81(3) (2004/C 101/08) (“Article 101(3) Guidelines”); see also Bellamy & Child, European Union Law of Competition (8th ed) §5.063 & §5.080 and the Commission’s Guideline on the First Conduct Rule, §4.4. G2. Burden and standard of proof G2a. The parties’ positions 156.The Commission’s position is that if a respondent seeks to rely on the exclusion from the first conduct rule provided for in Schedule 1 to the Ordinance, it is for the respondent to make out its case in that regard. Insofar as s 1 of Schedule 1 is concerned, the Commission contends that the burden lies on the respondents to show on the balance of probabilities that the four conditions are fulfilled. 157.In contrast, the respondents (with the 5th respondent taking the lead on this point) contend that the entire burden in relation to s 1 of Schedule 1 lies on the Commission as a matter of construction of the Ordinance; alternatively, if the burden is placed by the statute on the respondents, then it unjustifiably derogates from the presumption of innocence enshrined in Art 11(1) of the Hong Kong Bill of Rights[21] and should be read down as merely imposing an evidential burden on the respondents. 158.The issues raised may be divided as follows:
G2b. The EU and UK position 159.It is convenient to refer to the EU and UK positions before turning to the arguments. As regards burden and standard of proof, the position in the EU is that it is for the party claiming the benefit of Art 101(3) to demonstrate by means of “convincing arguments and evidence” that Art 101(3) applies to an agreement: Bellamy & Child, European Union Law of Competition (8th ed), §§3.014‑3.015. Art 2 of Regulation 1/2003 of the European Council provides:
160.It is of note that Regulation 1/2003 makes this express allocation of burden of proof even though, in line with respect for national procedural autonomy, the Regulation is largely silent on procedural matters to be observed by national competition authorities and national courts of the Member States of the EU.[22] 161.In the domestic regime of the UK, the undertaking which raises the defence has to show on a balance of probabilities that the requirements of Art 101(3) have been satisfied: Sainsbury’s Supermarkets Ltd v Mastercard Inc [2018] EWCA Civ 1536, §§77‑81. In particular, “robust analysis and cogent evidence would be required to establish, on the balance of probabilities, that the restrictive agreement in fact and in the real world (as opposed to in theory) gives rise to procompetitive effects” (at §79). 162.Prior to Regulation 1/2003, Art 101(3) was not directly applicable as a defence. Applications had to be made by undertakings concerned to the EU Commission for exemption under Art 101(3) (then Art 81(3)). The burden of proof was the same under settled case law before Regulation 1/2003: Unilever Bestfoods (Ireland) Ltd v Commission, Case C‑552/03P, at §§102‑103; GlaxoSmithKline Services v Commission, Joined Cases C‑501/06P, C‑513/06P, C‑515/06P and C‑519/06P, at §§82‑83; Remia BV v Commission, Case 42/84, at §45. A person who relies on that provision must demonstrate, by means of convincing arguments and evidence, that the conditions for obtaining an exemption are satisfied. The position was the same under the UK system prior to Regulation 1/2003: The Racecourse Association v Office of Fair Trading [2005] CAT 29, §134. 163.The respondents submit that this Tribunal should not follow EU jurisprudence on the question of burden of proof because it has not engaged with the fundamental protection afforded by the presumption of innocence which is applicable in Hong Kong. It is said that the European Convention on Human Rights does not apply directly in the application of the TFEU, and that the question is therefore not engaged. But this argument is flawed because following the adoption of the European Charter of Fundamental Rights, the EU courts and institutions are required to apply the Charter provisions with the same legal force as obligations under the founding EU Treaties.[23] In particular, Art 48(1) of the Charter makes provision for the presumption of innocence.[24] Even before the formal adoption of the Charter, when Regulation 1/2003 was promulgated the Charter had already been solemnly proclaimed by the European Parliament, Council and Commission in 2000. Recital (37) of Regulation 1/2003 specifically states:
164.It was also submitted that the burden of proof under the previous system where the undertaking concerned had to apply to the EU Commission for exemption had simply, without sufficient consideration, been carried over to the regime under Regulation 1/2003. This suggestion cannot be accepted as it ignores Recital (5) of the Regulation, which states:
165.The implications of human rights provisions in the context of burden of proof under Art 101(3) were considered by the UK Competition Appeal Tribunal in Ping Europe Ltd v Competition and Markets Authority [2018] CAT 13, at §§107‑108. There the undertaking accepted that it bore the burden of proving that the impugned restriction was exempted under Art 101(3), but argued that the Competition and Markets Authority had to prove that less intrusive measures could have been used without compromising the objective (ie the indispensability condition under Art 101(3)). The Tribunal, having expressly taken account of the presumption of innocence enshrined in Art 6(2) of the European Convention on Human Rights, held that the Authority only had an evidential burden to show there were alternative measures available, and that the undertaking had the burden of proving that the restriction was necessary and, as part of that exercise, of rebutting the Authority’s case as to possible alternatives. 166.The considered policy and prevailing experience in the EU and UK is that the imposition of the burden of proof on a balance of probabilities on those who rely on Art 101(3) has not been found to be incompatible with, or in any sense an unjustified derogation from, the presumption of innocence enshrined in Art 6(2) of the European Convention on Human Rights or Art 48(1) of the Charter of Fundamental Rights of the EU. G2c. Construction of the Ordinance 167.As the Court of Final Appeal stated in HKSAR v Lam Kwong Wai & Anor (2006) 9 HKCFAR 574 at §29 and HKSAR v Ng Po On (2008) 11 HKCFAR 91 at §30, the first step in the requisite analysis is to ascertain, by applying common law principles of construction as supplemented by any relevant statutory provisions, whether the statute imposes a persuasive burden of proof on the respondent. 168.The structure of the Ordinance is as follows. Part 2 makes provisions on the conduct rules. Within Part 2, Division 1 relates to the first conduct rule which is contained in s 6. Division 2 relates to the second conduct rule which is set out in s 21. The conduct rules as enacted are of general application, but Division 3 sets out exclusions and exemptions from them. In particular, s 30 provides that the conduct rules “do not apply in any of the cases in which they are excluded by or as a result of Schedule 1”. In this regard, the various expressions used in different jurisdictions, viz. “may be declared inapplicable” (in Art 101(3)), “exempt from the Chapter I prohibition” (in s 9(1) of the UK Act) and “does not apply” (in Schedule 1 to the Ordinance), in my view, merely represent different drafting styles without substantial divergence in meaning. 169.While criminal case‑law such as R v Edwards [1975] QB 27 and criminal statutes such as s 94A of the Criminal Procedure Ordinance (Cap 221) (discussed in Ng Po On at §§33‑34) may not be directly and fully applicable to these proceedings, applying analogous principles of construction and having regard to the structure and language of the relevant provisions, it seems to me to be the legislative intention that in proceedings brought by the Commission to enforce the first conduct rule, it is for the Commission to prove the elements of the alleged contravention, but not to prove the absence of an applicable exclusion. Though called an “exclusion” rather than “exception” or “exemption”, s 1 of Schedule 1 serves to identify a specific situation in which the commission of a generally prohibited act is rendered not unlawful. The first conduct rule is contravened by any agreement that has as its object or effect the prevention, restriction or distortion of competition in Hong Kong. The burden of the charge is restriction of competition, though the legislature has seen fit not to condemn agreements which restrict competition provided they promote efficiencies in the way prescribed by s 1 of Schedule 1. 170.It is of note that besides s 30, the Ordinance by virtue of s 9 also allows for the application for a decision from the Commission on whether or not an exclusion under Schedule 1 applies. Section 9 provides:
171.The Commission is in my view correct in submitting that this mechanism plainly suggests that it would be for the requestor, ie the undertaking, to present cogent and compelling evidence and arguments in support of its application. This is consistent with the pre‑2004 position in the EU where the undertaking that relied on Article 81(3) had to demonstrate that the conditions for obtaining an exemption were satisfied (see §162 above). In The Racecourse Association & Others v Office of Fair Trading [2005] CAT 29, the UK Competition Appeal Tribunal, referring to the equivalent regime for individual exemption under the then s 4 of the Competition Act 1998, stated (at §134) that they could see no reason why the same legal burden of proof as set out in Regulation 1/2003 should not apply and that they did not understand how it could be otherwise, specifically rejecting the submission that the undertakings were only subject to an evidential burden. 172.If, in contrast, in an enforcement action, the burden lies on the Commission to prove that the four conditions under s 1 of Schedule 1 are not satisfied, it would severely undermine the s 9 system for decision on exclusion. As the Commission submits, it would imply that no undertaking would care to seek a decision because its obligations under s 9 would be more onerous than its position in an enforcement action where the burden would lie on the Commission. 173.The respondents’ answer that a decision under s 9 does not involve any threat of penalty is not in point. By virtue of s 12(1), a decision that an agreement is excluded from the first conduct rule renders the relevant undertakings immune from any action under the Ordinance with regard to that agreement, including any proceedings for pecuniary penalties. It does not make it logical for the burden under s 9 and an enforcement action to be opposite. 174.It is also relevant to have regard to the structure of the first conduct rule itself and, in particular, the provision for infringement “by object”, which, as explained above, dispenses with the need to carry out an effects investigation and analysis. If in the first place the Commission, as an inherent part of its case of infringement of the first conduct rule, already has the onus of alleging and establishing that the rule has not been excluded under s 1 of Schedule 1, given the complexities of economic issues inherent in competition enforcement actions, the Commission will be forced to carry out what amounts effectively to an effects analysis even where it alleges only an infringement by object. This, in my opinion, could not have been the legislative intention. 175.The fact that the Ordinance does not expressly replicate Art 2 of Regulation 1/2003 cannot, in my view, be read as evincing an intention that the position on the burden of proof in Hong Kong is the opposite. Regulation 1/2003 is not part of Art 101 which the legislature adopted as a model; it did not change the position in the EU in relation to burden of proof. The burden was likewise on the undertaking under the previous individual exemption regime, which is similar to the individual decision regime under s 9 of the Ordinance. 176.In making a conspicuously deliberate decision to model the Ordinance so closely upon the EU provisions and, in particular, to adopt the language of Art 101(3), the legislature in my view envisaged and intended that Hong Kong should draw from the wealth of jurisprudence in the EU and the UK with reference to Art 101(3), of which the imposition on the undertaking of the burden of proving satisfaction of Art 101(3) has long been a fundamental and integral feature. 177.Reliance is placed by the respondents on s 8 of Schedule 7 to the Ordinance which provides:
It is argued that the absence of an equivalent provision to s 8(2) shows that there is no intention on the part of the legislature for the burden to be placed on the undertaking claiming the benefit of the efficiency defence. 178.To begin with, however, the merger regime in Schedule 7 is not a direct adoption of either the EU regime or the UK regime. The structure of Schedule 7 which expressly contains an efficiencies exclusion from the merger rule has similarities to the Singaporean regime (s 3 of the Fourth Schedule to Singapore’s Competition Act)[25] and the Canadian regime (s 96 of Canada’s Competition Act).[26] In Canada it has been established that under s 96 of the Competition Act, while the Commissioner of Competition has the burden of proving the anti‑competitive effects of a merger, the merging parties bear the onus of establishing all other elements of the defence, including that the efficiency gains exceed or offset the anti‑competitive effects: see Tervita Corporation & Others v Commissioner of Competition 2015 SCC 3, §122 and cases therein cited. Furthermore, in explicitly placing the burden of proof of economic efficiencies on the undertakings concerned, s 8(2) appears to have done no more than reflect the position in the previous regime under s 7P of the Telecommunications Ordinance (Cap 106).[27] 179.I have borne in mind the principle of legality relied on by the respondents (according to which human rights and fundamental principles of law, even where derogable, cannot be overridden except by express words or necessary implication: see A v Commissioner of Independent Commission Against Corruption (2012) 15 HKCFAR 362, §24, per Bokhary and Chan PJJ). It is common ground that the Ordinance does not expressly impose the burden of proof on a respondent in relation to the efficiency defence. I have also reminded myself that a necessary implication is not the same as a sensible or reasonable implication but is one that necessarily follows from the express provisions of the statute construed in their context: see R (Morgan Grenfell) v Special Commissioner of Income Tax[2003] 1 AC 563, §45. For the reasons I have endeavoured to set out above, however, I am of the view that the Ordinance by necessary implication imposes the persuasive burden on the undertaking that claims the benefit of the efficiency defence. It is not merely a reasonable and sensible result, but one that follows as a matter of logic from the structure and content of the statute. 180.For these reasons, I take the view that the Ordinance is to be construed as imposing a persuasive burden on the undertakings in question to bring themselves within the exclusion in the form of the efficiency defence. G2d. Whether the presumption of innocence is engaged 181.On the basis that the Ordinance places the persuasive burden on the respondents to establish the efficiency defence, is there a derogation from the presumption of innocence? 182.The Commission submits that the presumption is not engaged since there is no reversal of onus of proving the essential ingredients of the charge. The argument is that the gravamen of the relevant charge is a contravention of the first conduct rule, not a contravention of the first conduct rule without efficiencies justification. Reliance is placed on EU jurisprudence under which the balancing of efficiency gains and anti‑competitive effects is conducted exclusively within the framework of Art 101(3): Bellamy & Child, European Union Law of Competition (8th ed), §3.012; Article 101(3) Guidelines, §10. The absence of efficiency gains, the Commission submits, is not an ingredient of the infringement as such. 183.It is true that there are passages in earlier cases such as HKSAR v Lam Kwong Wai, at §41, and HKSAR v Hung Chan Wa (2006) 9 HKCFAR 614, at §74, which may perhaps be read as suggesting that it is relevant to characterise the “substance” of an offence, and that where the impugned provision requires the defence to negate a “critical element” of the offence, a reverse onus arises and the presumption of innocence is engaged. The Commission also relies on R v Foye (Lee Robert) [2013] EWCA Crim 475 in which the English Court of Appeal drew a distinction between requiring a defendant to disprove an element of the offence and requiring him to establish an exception or excuse and held (at §30), agreeing with what Lord Woolf said in R v Lambert, Ali and Jordan [2002] QB 1112 at §16, that the latter was less objectionable, though the court also recognised that even where the distinction could be discerned, “it does not universally provide an answer to the question whether the reverse onus provision is justified and that further enquiry into justification is needed”. 184.The difficulty with the Commission’s argument, however, is that it seems to me to run counter to the approach that has been adopted by the Court of Final Appeal in subsequent cases, which recognises the presumption of innocence as being engaged whenever a persuasive burden is placed on the defendant to prove “an ultimate fact which is necessary to the determination of his guilt or innocence”: Ng Po On, §38. In that case, Ribeiro PJ said at §39 (footnote omitted):
185.On this basis, it matters not how one characterises the matter that the defendant is required to prove. There is a derogation from the presumption of innocence whether the defendant has to prove an essential component of the charge or an element of what one would normally characterise as a defence. 186.It is also notable that earlier in his judgment, Ribeiro PJ stated (at §30) that the answer as to whether the enactment as a matter of construction imposes a persuasive burden of proof on the defendant effectively provides the answer to the second question of whether the enactment so construed derogates from the presumption of innocence (see also §35 which is to the same effect). The other members of the court, including Sir Anthony Mason NPJ who gave the leading judgment in Lam Kwong Wai, agreed with Ribeiro PJ. 187.It is true that in Lee To Nei v HKSAR (2012) 15 HKCFAR 162 at §§18‑19, Ribeiro PJ referred to matters such as knowledge that the goods bear forged trademarks, which the enactment in question required the defendant to disprove, as being part of the “gravamen of the offence”. I do not, however, read those passages in his Lordship’s judgment as saying that there is a reversal of onus engaging the presumption of innocence only if the matter the defendant needs to prove is an essential constituent of the charge. Paragraph 14 of the same judgment suggests that the presumption is engaged and abrogated whenever a statute reverses the burden by placing a persuasive burden on the defendant “to prove specified matters on a balance of probabilities as a condition of avoiding liability”. 188.I conclude therefore that the constitutionally protected presumption of innocence is engaged, so that it is necessary to proceed on to the 4‑stage proportionality assessment as set out in §158(3) above. G2e. Steps 1 and 2 of the proportionality analysis 189.There is no dispute that the presumption of innocence is not an absolute right. A reversal of onus is not unconstitutional if it passes the proportionality analysis. The burden lies on the Commission in this case to justify the derogation and the justification must clear the standard of intense scrutiny: Lee To Nei, §23. 190.As stated above, the first two steps of the analysis essentially ask whether the derogation is rationally connected with the pursuit of a legitimate societal aim. As I understand their submissions, the respondents — rightly, in my view — do not seriously contest these two steps. 191.As submitted by the Commission, the imposition of a burden on the undertakings concerned pursues the legitimate societal aim of facilitating the objective of the Ordinance as stated in its long title, namely, to “prohibit conduct that prevents, restricts or distorts competition in Hong Kong”. This is advanced by requiring an undertaking which asserts that an agreement, albeit shown to restrict competition, is justified by the efficiencies generated, to prove it. It is rationally connected with the pursuit of the legitimate aim because the information relevant to the defence is usually in the possession of the undertaking, not the Commission. G2f. Step 3 of the proportionality analysis 192.It is step 3 of the analysis that has attracted the most debate in this case. The respondents contend that it is for the Commission to show that imposing a persuasive or legal burden on the undertaking who invokes the defence does not derogate from the presumption of innocence more than necessary, and that the Commission has failed to do so because imposing an evidential burden on the undertaking is sufficient. 193.There are a number of matters relevant to the consideration of this question. First, as submitted by the Commission, in a case such as this the gravamen of the charge is the contravention of the first conduct rule by object. It is in relation to “object” cases that the burden of proof is most material, for in “effect” cases the Commission has the burden, in any event, of carrying out an effects investigation as the basis of its case, based on the same counterfactual as required for assessing the efficiency defence. For the reasons explained in section F above, the respondents’ arrangements had the object of restricting competition and were prohibited as such. It is to be recalled that this means their arrangements are properly viewed “by their very nature, as being harmful to the proper functioning of normal competition” and “reveal a sufficient degree of harm to competition” that there is no need to examine their effects (see §105 above). It is also well established in EU and UK case-law that proof of restriction of competition by object does not as such depend on an analysis of the efficiency benefits alleged (see §§114-120 above). The Commission’s position on the burden of proof would promote the legislative object of protecting Hong Kong from such conduct by requiring undertakings to exercise vigilance and refrain from such conduct unless they consider they can establish the conditions for the efficiency defence. It would also encourage undertakings to ensure that all information and evidence concerning their case of efficiencies is properly recorded and preserved. 194.Further, it is relevant to note that these proceedings, though conceded to involve the determination of a criminal charge for constitutional purposes, do not concern a paradigm criminal offence punishable by imprisonment. Whilst not providing a positive justification for encroachment into the presumption of innocence, this is a “negative factor” which means that “the objection against derogation from the presumption of innocence is not augmented by the prospect of severe punishment following a conviction” at least in comparison with offences that carry custodial penalty: Ng Po On at §58; see also R v Johnstone [2003] UKHL 28 at §50, cited in Lee To Nei v HKSAR,at §33. 195.The information required to assess the claimed efficiencies is often exclusively held by the undertakings concerned. While the defence is concerned with objective benefits rather than the undertakings’ subjective assessment of benefits, its focus falls on the operation of the undertakings. What their production costs are, how those costs will be affected with or without the impugned agreement, how much of the efficiency gains has been passed on to the consumers and in what ways, and whether the impugned arrangement is indispensable to the generation of efficiency gains in the context of the undertaking’s operations, are matters peculiarly within a respondent’s knowledge. Whether competition has been eliminated in respect of a substantial part of the products or services in question is also a matter of which the undertakings concerned should have intimate knowledge and which they are particularly well placed to advance. This can be a significant factor in support of a reverse onus, albeit not necessarily sufficient in itself: Ng Po On, §65; R v Johnstone, §§50, 52(5) & (6); Fu Kok Kuen Patrick v HKSAR (2012) 15 HKCFAR 524, §95; HKSAR v Choi Wai Lun, §67. 196.This is to be contrasted with the serious difficulty which would be faced by the Commission if it was required to prove — and a fortiori to prove beyond reasonable doubt — that the conditions of the defence are not satisfied, given the often secretive nature of anti‑competitive agreements and the evidence being confidential to the undertakings. If the respondent has merely an evidential burden, once the respondent does enough to enliven the issue, then it is not just a matter of the Commission dismantling the respondent’s evidence; the Commission will have to prove beyond reasonable doubt that the defence does not apply. 197.In support of their position, the respondents have placed reliance on a speech given by Rose J of the English High Court (as she then was) at a conference in 2011, in which her Ladyship commented on the reverse onus of proof in relation to Art 101(3) of the TFEU.[28] It seems to me notable, however, quite apart from the fact that those comments were made extra‑judicially without hearing counsel’s arguments, that her Ladyship was raising questions more than expressing any concluded views. It is true that it was suggested there are matters within Art 101(3) that are not particularly within the knowledge of the undertaking concerned or more difficult for the Commission to establish, and that they might be matters that the Commission has to tackle in other contexts without there being a reverse onus.[29] It is important, however, to note that (i) the comments are apt to deal with an “effects” case where the Commission would, for the purpose of establishing its own case, have to carry out a detailed effects investigation and analysis; (ii) the comments were made in the context of the EU and UK regimes where the Commission does not have to prove an infringement beyond reasonable doubt, in contrast to the position that obtains in Hong Kong (as stated in §39 above); and (iii) as Mr Beard submitted, the comments would not have been made with hard‑core cartels in mind, having regard to the general understanding that it would be rare for the efficiency defence to be applicable to them. In addition, Rose J also noted that one could expect that the Luxembourg courts would regard the seriousness of the social problem that the statute in question is designed to combat as a factor pointing in favour of the compatibility of the reverse onus for establishing the efficiency defence.[30] It should also be noted that at the end, the learned judge concluded by saying it may well be that the CJEU will conclude that the reverse onus is compatible with the presumption of innocence.[31] 198.The point has already been made above that the purpose of having the category of “object” infringement is to dispense with the need to carry out an effects analysis. Without imposing the burden on the undertaking, if all that is required is for them to show some evidence (as opposed to mere fanciful assertion) that there are efficiency justifications, given the complexities of economic issues inherent in competition enforcement actions, the Commission will be forced to carry out what is tantamount to an effects analysis for the purpose of establishing an object infringement. This would defeat the intention and choice of the legislature in adopting a system with carefully calibrated allocation of burdens against a very similar human rights backdrop. 199.There is plainly a difference between an evidential burden and a persuasive or legal burden. The respondents submit, not surprisingly in the present context, that an evidential burden is not an insubstantial one. Reliance is placed on Ng Po On where Ribeiro PJ said at §27 (footnote omitted):
200.It is undeniable, however, that the evidential burden is much lower than and indeed different in kind from the balance of probabilities. It only requires a respondent to raise a reasonable doubt, which is a relatively low threshold. Where assertions have been made without documentary evidence, it is not always easy to conclude they are so incredible that no issue is raised at all. Often the assertions will, before trial and judgment, have to be taken on their face. In practical terms whenever a respondent pleads the efficiency defence, since it could be difficult to strike it out at a pre‑trial stage given the many fact‑sensitive matters and complex economics involved, if the burden lies on the Commission, it would be obliged to conduct an investigation many months if not years after the event and in the course of the proceedings, and engage experts to try to “disprove” beyond reasonable doubt the efficiency justifications, simply because the respondent raises it as an issue. This would in my opinion severely undermine the effectiveness of the procedures of competition law in Hong Kong and could not have been the legislative intent. 201.It has been commented in relation to the EU regime that if the whole burden of proof is put on the regulatory authority including the burden to demonstrate that the conditions in Art 101(3) have not been met, “the system may here find its Achilles heel” as it would encumber and delay the enforcement procedures:
It seems to me the same consequences would infest the Hong Kong regime if the burden was placed on the Commission, a fortiori given the need to prove infringement beyond reasonable doubt. 202.The legislature is entitled to stipulate (albeit by implication), in circumstances where ex hypothesi the Tribunal is satisfied an undertaking has entered into an agreement that is restrictive of competition in Hong Kong, that contravention of the first conduct rule should be found if it is not shown that the efficiency defence more likely than not applies. In my view there are compelling reasons for concluding that the third stage of the proportionality analysis is satisfied. G2g. Step 4 of the proportionality analysis 203.As Ribeiro PJ pointed out in Hysan at §78, in the great majority of cases, the result arrived at after undertaking the first three inquiries is unlikely to be changed by the fourth stage of the analysis. In my judgment, there is a reasonable balance struck between the societal benefits (including effective enforcement of competition law and the prevention of anti‑competitive conduct) and the inroads made into the constitutionally protected rights of the undertaking (namely, the presumption of innocence in the context of an enforcement action seeking pecuniary penalties). Placing the onus on the undertaking does not, in my view, result in an unacceptably harsh burden on the undertaking. The undertaking itself is best placed to proffer evidence and argument in relation to its claimed efficiencies. G2h. Conclusion 204.For the above reasons, I conclude that it is for the respondents to establish the efficiency defence on the balance of probabilities. G3. General comments on the factual basis of the respondents’ case 205.A notable feature of the respondents’ case in these proceedings is that they have advanced two fundamentally inconsistent factual contentions. In their pleaded case, they deny the existence of any of the alleged agreements. As an alternative plea, and in witness statements submitted following the Tribunal’s ruling on 26 September 2018 and in the expert report of Dr Pang, however, an entirely different story was put forward, namely, that such agreement was necessary to deliver efficiencies from which their customers benefited. Quite apart from being objectionable from the point of view of the principles of pleading, this regrettable and opportunistic approach has not only undermined the credibility of their denial of the agreement, but has also haunted them in their attempt to invoke the efficiency defence. 206.Not surprisingly, the respondents have found it difficult coherently to give any details of their efficiency defence in the Responses or the particulars, given their primary stance that there was no Floor Allocation Arrangement in the first place. Even in their witness statements filed in April 2018, there was hardly any evidence which would be needed for the efficiency defence to get off the ground. 207.The more detailed factual assertions underlying the respondents’ contention first appeared in the report of their expert dated 30 July 2018 in the absence of any witness statements covering those factual matters. For reasons explained in this Tribunal’s Reasons for Decision dated 28 September 2018,[33] this was not acceptable. Upon the Commission’s objection that the expert’s opinions were not supported by any material factual evidence and were as such admissible and liable to be struck out, the respondents obtained leave to file supplemental witness statements to cover the relevant factual matters. 208.However, the supplemental statements subsequently filed in early October 2018 (signed by 6 existing witnesses and 2 new witnesses) turned out to be not only remarkably similar in content and form, but also expressed in highly technical language which one can hardly imagine to have originated from the witnesses themselves. Thus the statements refer to a “very important source of efficiencies”, make comparison with “the scenario in the ‘but for’ world”, assert that HKHA’s rules and restrictions created “an unequal playing field” against the Appointed DCs, and state that without the ability to “take advantage of the cost savings and efficiencies that arise from the Alleged Floor Allocation Arrangement, [the Appointed DCs] would likely pass on (at least partially) the increasing costs to the tenants in higher prices for decorative work in the ‘but for’ world”. It is clear that in practice, relevant passages in the expert report had simply been cut and pasted into the supplemental statements. 209.Further, these statements all contain the assertion (quoted in §52 above) that the Floor Allocation Arrangement was part of a broader cooperative arrangement among the respondents entered into, following customary practices, in order to counter the cost disadvantages that Appointed DCs faced. Upon cross‑examination, however, it transpired that all of the respondents’ witnesses who made that statement had great difficulty understanding that paragraph (among others) of their supplemental statements. 210.The supplemental statements thus appear to be words put into the witnesses’ mouth, taken from a report which contains the expert’s interpretations of her instructions, information collected by the expert from the clients, as well as her opinions. In these circumstances little weight can, in my view, be put on those statements. 211.The respondents’ witnesses were also constrained (by their primary case) to give answers in cross‑examination which are contrary to or at least undermine their efficiency defence. For example, one said they simply did not consider the matter in terms of efficiency or costs savings from the point of view of consumers. Another was compelled to say that the prices they charged did not reflect any efficiencies from the Floor Allocation Arrangement because it did not exist. On the basis of such evidence, Dr Pang’s report would appear to be primarily an attempt of ex‑post rationalisation. This severely undermines the credibility of the respondents’ case such as the assertion that the efficiencies and cost savings were passed on by the respondents to the tenants. 212.Another problem in the respondents’ evidence is that there was little critical assessment by the expert of the assertions made to her by the respondents, mostly without documentary basis. Whether this was due to the extreme haste in which the expert report had to be prepared (for reasons beyond the control of the current lawyers and expert of the respondents as described in the decision of the Tribunal dated 10 July 2018 [2018] HKCT 2) or other reasons, the consequence is that the reliability of the expert’s analysis and conclusions is seriously called into question. G4. Condition 1 — efficiencies 213.According to EU case law, only objective benefits are taken into account. The improvement must display appreciable objective advantages of such a characteristic to compensate for the disadvantages which the agreement entails for competition: Case 56/64 and 58/64 Consten and Grundig v Commission. It stands to reason that cost savings arising directly from the mere exercise of market power are not considered objective benefits so that, for example, reduction in sales and marketing expenses resulting from a market allocation agreement is not an improvement in the production or distribution of goods or services. G4a. The respondents’ case 214.The main plank of the respondents’ case is that the Floor Allocation Arrangement enabled them to concentrate work at any one time on flats located on the same floor. Having regard to the long lift-waiting time in the Estate, significant efficiencies arose from being able to perform work in multiple number of flats on the same floor on account of the time saved from not having to wait for the lifts to move tools, equipment, supplies or raw materials from one floor to another. 215.Each of the 3 buildings in Phase 1 of On Tat Estate is served by 5 lifts, each measuring about 1.2m by 1.2m as follows:
216.According to the respondents, the time spent waiting for the elevators can be substantial, especially if the workers have to change elevators in the lobby. Hip Yick (R6) says that the “normal” waiting time for a lift for work crew or a raw material supplier to go from one floor to another “can easily be 20 or more minutes”. Sun Spark (R2) says that the lift‑waiting time can be up to 2 hours during peak time, eg when many flats in the same building are being decorated and there are a lot of furniture deliveries as the tenants begin to move in. 217.Dr Pang’s opinion is that under the Floor Allocation Arrangement, each respondent focused their decoration work on a small number of specific floors in each building, aiming to generate work from tenants in multiple number of flats on the same floor. In contrast, in the “but for” world, each respondent would market their services to tenants from every floor in the building and would end up with their work spread across many different floors in the same building. 218.Dr Pang categorized these cost savings and efficiencies into 5 categories: (i) savings in labour costs in performing decorative work; (ii) savings in raw material delivery costs; (iii) savings from the reduction of raw material wastage; (iv) savings in construction waste and debris clean up and disposal costs; and (v) efficiencies from more focused sales and marketing efforts and management of decorative work progress. These categories are further elaborated as follows. 219.Another flaw in the respondents’ evidence is that it contains references to alleged substantial savings resulting from their sharing of overhead and operating costs such as the setting up and fitting out cost and rent of the Site Office, the cost of erecting signage for the Site Office, and the printing cost of work orders and invoices. Since these joint activities are not part of the agreements objected to as infringements of the first conduct rule, the alleged savings resulting from them are not to be taken into account for present purposes. They are not linked to the restraints that are being impugned and are not therefore excluded in the counterfactual. G4a(i). Savings in labour costs 220.It is said that, relative to the “but for” world, when a work crew perform decorative work in multiple number of flats on the same floor, once they complete a specific task in a flat and are ready to start work in another flat, a substantial amount of time can be saved for the work crew in moving their tools, equipment, supplies and raw materials from one flat to the nextif it is located on the same floor, especially for decorative work that requires a long preparation or set up time (eg carpentry related work) or decorative work that requires a very long wait time between sub‑tasks. Raw materials can be readily “shared” among the different flats on the same floor where an unforeseen need arises. The additional labour time that a work crew has to incur to perform the same tasks in the “but for” world would lead to increased labour costs, as the Respondents would either have to hire additional workers to perform the same amount of work, or to compensate the same work crew for theadditional time it took them to complete the work. 221.Dr Pang gave 5 examples of work from the Package items on the Flyer. For laying floor tiles and baseboards (items 1 and 2 of the Flyer), the work crew’s toolkit consists of a mason’s trowel, 3 buckets, 3 finishing trowels, a spirit level, 4 feather edges, chalk line, an angle grinder, a broom, mop, cleaning sponges and cloth, and a hawk. It takes about 30 minutes to pack up the tools and carry them to the next flat in a trolley cart, but substantial time can be saved if the work crew can simply carry the tools to another flat on the same floor without packing them neatly. According to some respondents, the work crew can usually finish the tiling work for a flat in less than a day. The work crew would start setting up in the next flat during the remainder of the day if it is on the same floor, but not if it is on a different floor. Certain estimates by some respondents have been provided: two 1‑2 person flats can be tiled in a day if they are on the same floor, but only 1.5 flats if they are on different floors; 9 man‑hours are needed to tile a 2‑3 person flat on average if the flats are on the same floor, but 12 man‑hours if they are on different floors. According to the respondents’ estimates, the increase in average labour cost for tiling a 2-3 person flat in the “but for” world ranges from 9.1% to 50%. 222.For partition of a room with a sand brick wall (items 6 and 7 on the Flyer), the work comprises erecting the wall with sand bricks, applying 3 to 4 layers of different kinds of plaster, and applying finishing paint. Each stage may take only a few hours, but one has to wait for the materials to dry and set after laying the bricks and each time after plastering. Each waiting period is said to be a few days or 2 to 3 days. It is said that if the flats to be worked on lie on the same floor, after (for instance) plastering the wall in one flat, the work crew can easily gather the necessary tools and supplies and start work in the next flat. But if the flats are on different floors, they will have to pack up their tools and supplies, put them in a cart, wait for the lift, and then unpack, before they can start work again. According to one respondent, Hip Yick (R6), the labour cost for a bedroom partition wall would be increased by 50% in this “but for” world. 223.For installing an internal door and door frames inside a flat (item 8 on the Flyer), a carpenter needs a table saw, 3 wood trimmers, 3 wood planes, an air nail gun and compressor, an electric drill, a battery operated drill, a variety of screwdrivers and a circular saw. If the flats to be worked on are on the same floor, substantial time in packing the tools and supplies, waiting for the lift, unpacking and setting up again, can be saved relative to the “but for” world. According to Hip Yick (R6), 6 door frames can be installed in a day for flats on the same floor, compared to only 4 door frames for flats on different floors, translating into a 50% increase in labour cost. According to Wai Sun (R8), 10 frames can be completed in a day on the same floor, but only 4 frames on different floors, translating into a 150% increase in labour cost. 224.A similar explanation is given in relation to wall painting (item 5 on the Flyer) which requires the application of 3 layers of paint, each of which takes about 2 to 3 days to dry. The toolkit includes a scaffolding platform, gloves, sandpaper, a hawk and a finishing trowel, buckets, paint brushes and rollers. If the flats to be painted are located on the same floor, after painting one layer, the work crew can easily go to paint the next flat, without having to pack up the tools and supplies, wait for the lift, and unpack. Based on the respondents’ various estimates, the increase in labour cost in the “but for” world is said to range from 7.7% to 66.7%. 225.As to the remaining items on the Flyer, one respondent says an electrician can install 6 exhaust fans (item 4) in flats on the same floor in a day, but only 4 fans in flats on different floors. Another respondent says their workmen can install 4 to 5 kitchen cabinets (item 9) in flats on the same floor in a day, but only 3 if the flats are spread across various floors. One respondent says that it takes their workmen 1 hour per flat on average to install marble windowsills on the same floor but double the time for flats on different floors, though the figures given by another respondent for this task are 2 hours on the same floor and 3 hours on different floors. G4a(ii). Savings in raw material delivery costs 226.The expert report states that the prices of the raw materials charged by the suppliers to the respondents included the cost of delivering the materials to the flats and, in the actual world, were based on delivery to multiple flats on the same floor for a given delivery trip. In the “but for” world, additional time and thus higher cost would be incurred by the suppliers for delivery to flats spread out across multiple floors in a single trip. The respondents believe that the suppliers would begin charging higher prices to reflect the increase in the cost of delivery. Various estimates of the likely increase have been stated by various respondents for different materials: eg one respondent has mentioned 4% for Portland cement, 14% for sand, and 12% to 20% for sand bricks. G4a(iii). Reduction of raw material wastage 227.It is said that another source of savings arises where there are surplus raw materials left from the work performed in one flat. For example, according to one respondent, laying the floor tiles in a 2‑3 person flat requires 5.5 bags of Portland cement, 7.5 buckets of sand, and half a bag of white cement, so that there will be half bags and half buckets left from single flats. If the flats being decorated are on the same floor, the workers are more willing to move the surplus materials from one flat and use them in another flat; but if the flats were on different floors, such materials would go to waste. Savings ranging from 7% to 15% have been asserted by various respondents for sand, cement and bricks. G4a(iv). Savings in construction waste and debris clean up and disposal costs 228.It is said that if each of the respondents works only on a number of floors, they will incur lower labour cost for cleaning up and disposal of the construction waste and debris generated by themselves. Thus for the cleaning up work in the Flyer (item 3), the debris from several flats on the same floor can be collected in one cart with a single trip to the dumping site on the G/F, but several roundtrips will be needed if the flats are on different floors. One respondent says that it employed one worker per day for cleaning up waste and debris in On Tat Estate but if the same number of flats worked on were spread out across the whole buildings it would have had to employ two workers for that purpose. Another respondent says that it pays its cleaning crew a fixed sum per flat per day in the expectation they would be working on multiple flats on the same floor, and that the sum will have to be increased by 66.7% if the flats are located on different floors. Overall increases in labour cost for the removal of waste and debris from floor tiling work have been asserted by various respondents in the range of 11.4% to 50%. G4a(v). Sales and marketing efforts and progress management 229.The respondents also contend that with the Floor Allocation Arrangement, they can achieve cost savings in their marketing and promotional efforts because they can concentrate those efforts on the 4 floors in each building they are each allocated instead of trying to cover substantially more floors. One respondent indicates that in the “but for” world, it would have to double the size of its sales staff from 3 to 6 men. Concentrating the work in a small number of floors also enhances the oversight and management of the work in different flats by the on‑site foremen. G4a(vi). Overall efficiencies and cost savings in relation to all 10 items 230.Based on information provided by the respondents, Dr Pang’s report set out the respective respondents’ aggregate labour and material cost for performing all 10 items on the Flyer in a 2‑3 person flat, and the estimated increase in cost in the “but for” world, in a table which is reproduced below:
G4b. Discussion 231.In addition to the general comments on the factual basis made in section G3 above, there are several major difficulties with the efficiencies analysis put forward on behalf of the respondents, including problems with (i) the proper conception of the counterfactual; (ii) reliability of the respondents’ data; (iii) discounts for out‑of‑pattern flats; (iv) assessment of the magnitude of efficiency gains; and (v) assessment of the adverse impact on competition. These are considered in turn below. G4b(i). The “but for” world 232.In the respondents’ analysis, the counterfactual — the “but for” world — is one where they would not be able to focus their sales and promotional efforts on tenants of targeted floors, and as such would be performing work in flats spread randomly across different floors. Dr Pang has calculated that in practice each respondent on average performed decorative work in 6 to 7 flats per floor in the Estate, amounting to an average of 24 to 28 flats in each building. According to her report, in the “but for” world:
This was the basis on which the respondents provided their data and estimates to Dr Pang. 233.The notion that without mutual coordination, the respondents would each be working on 24 to 28 floors in each of the 3 buildings, is an extreme and unjustified assumption. In my opinion, this construction of the counterfactual cannot be accepted to provide by comparison a correct and accurate assessment of the efficiencies generated by the Floor Allocation Arrangement. 234.First, this flies in the face of the evidence of the respondents’ witnesses, which was that in the absence of the Floor Allocation Arrangement, their strategy would be, if they won a single flat on a floor, they would make an effort to win the other flats on the same floor. This was the evidence‑in‑chief of Chan Yiu Kwai (R6), who stated: “After we have succeeded to solicit work contract from a certain flat on the floor, our salesman started to promote to other tenants on the same floor by showing these tenants the work(s) which we had done, and also distributing the flyers”. To similar effect was the statement of Cheung Bing Sang (R7). The statement of Wong Fu San (R10) also said that once they started having business on a certain floor, they would send more sales staff to that floor to get more business there. 235.The witnesses also referred to having an “advantage” there once they had secured a contract for a flat on a particular floor. This seems to me credible and readily understandable. Decorating a flat on a given floor would provide a valuable opportunity for additional business, as the contractor could showcase its work to other tenants on the floor. Tenants might well be attracted to a particular contractor being used by their neighbour on the same floor, whether from express recommendation or simply out of herd mentality. Further, the respondent in question could price itself so as to get more work on that floor, especially if the efficiencies alleged would actually result from working on a number of flats on the same floor. On the respondents’ case, the contractor would have a clear incentive to offer a lower price for subsequent flats on the same floor since the cost of decorating more flats would be lower once the first flat was won. Thus, for example, Chim Lap Fan (R2) responded as follows in cross‑examination:
236.Similar evidence was given by WC Wong (R1), Chan Chun Wah (R3), Chan Yiu Kwai (R6), Poon Wai Woon (R8) as well as Ma Yick Yin (R4). 237.This deliberate effort which made every sense on the respondents’ own case would plainly very significantly reduce the number of floors across which the flats were spread, allowing for a scenario where there would be concentration of flats on certain floors in the counterfactual world. 238.Against this direct factual evidence of her clients, Dr Pang effectively said that the respondents’ answers were naïve, that they lacked sophistication and were unable properly to assess what would really happen without the Floor Allocation Arrangement. Her own view was that because there would be uncertainty in the counterfactual world as to whether the Appointed DCs might or might not achieve the same concentration of flats on particular floors, it would be a “gamble” for them if they priced their services on the assumption that they would obtain concentration. Once their assumption turned out to be wrong and they got “burnt” (ie when their prices turned out to be unprofitable because they did not achieve sufficient concentration), they would simply not price at a lower level to try to win further flats on a floor. 239.With respect, I do not think that the respondents’ own evidence can be brushed aside in that way by their expert simply because they are not economists or competition lawyers. They are all businessmen, skilled in the trade, with many years of experience of the kind of work in question, who have actually worked in public housing estates including On Tat Estate. None of them has mentioned any worry about getting burnt in the scramble for work in the counterfactual world. 240.Of course, as Dr Jenkins observed, competition does involve uncertainty, and cartels do remove a lot of that uncertainty, which is very convenient for suppliers. But the uncertainty, rather than implying total randomness, produces a dynamic that incentivises competitive practices which ultimately enure to the benefit of consumers. As explained by Dr Jenkins, in the but‑for world the respondents would not be competing entirely in the dark. There were a variety of ways in which they could identify the tenants on the floors on which they had secured a contract. There were different ways for shrewd businessmen to match the demand to their selling points and advantages. The respondents could adopt pricing and discount strategies to try to win 3 or 4 flats on a floor close in time. A competitive market would bring benefits because the contractors would need to communicate to customers through price. In contrast, a market sharing arrangement removed the need to post prices that reflected the efficiency savings, since the respondents knew that their closest competitors would not be competing with them for the business on “their” floors. 241.The competitive process in the counterfactual would be facilitated by the intake of tenants in batches (from various floors) so that a respondent would have time to “take stock” of his position, devise his strategy, and target the floors he had in mind, for example, by offering discounts to flats on specified floors during particular time periods. The tenants’ units on the day of intake were identifiable by the Sticker worn by them. The respondent could also station salespersons on targeted floors. For their part, once the tenants saw there was price differentiation between the respondents, they would be much more eager to “shop around” and find the respondents who offered the best price. The respondents argued that a tenant would only randomly approach at most 3 respondents to discuss his decoration needs, giving a probability of ( 242.The respondents have referred to anecdotal information on Ying Tung Estate, a more recently completed public rental housing estate in Tung Chung, to assert that without an arrangement similar to the Floor Allocation Arrangement, there would be no floor‑concentration of flats for each contractor. As Dr Jenkins pointed out, however, even the most basic information regarding Ying Tung Estate is lacking so that no assistance at all can be derived from it. 243.The respondents’ contention that each floor (which typically has 22 flats) could not accommodate multiple concentrations of 4 flats each by more than one respondent is mathematically unsound. A glance at the tables in Appendix 2 will show that in the actual world there are many floors where 7, 8 or more flats were decorated by the respondents. It would not be unreasonable for a respondent to expect, if he was willing to compete with prices that passed on the alleged cost savings to the tenants, to win 3 more flats after it had secured one flat on a given floor, even if another respondent too had won a flat on the same floor. 244.Furthermore, in a competitive environment, there would be no reason to expect each respondent to win work on exactly 4 floors in each building. Competition would be expected to lead to the more efficient respondents winning work on more than 4 floors. For example, Kam Kee (R5) which, according to the table at §230 above, had the lowest costs (assuming the same quality), might win a large proportion of flats under the competitive counterfactual. This would in turn deliver additional benefits to tenants. There is nothing at all to suggest that this respondent produced work of lower quality. 245.Another difference in the competitive world is that if the alleged efficiencies did exist, firms would compete to get work on floors that form a cluster, such as 8/F to 15/F of one building, rather than on 4 floors, each 10 floors apart, in each of the 3 buildings. On the respondents’ case, having clusters of floors or consecutive floors would allow them to achieve even greater efficiencies which, in the competitive world, they would have an incentive to pass on to the customers. 246.It follows that if there are indeed efficiencies as alleged from concentrating work on a given floor, then the competitive process is likely to lead to substantial floor‑level concentration. In that but‑for world the more efficient suppliers might win work on more floors, and the floors each respondent concentrated on might be closer together. It may readily be accepted that the “pattern” would not be as neat and regular as under the Floor Allocation Agreement, and each respondent might have some “stray” flats, but it would plainly be a far cry from the assumed scenario where each respondent would be working on 24 to 28 flats spread randomly across 24 to 28 floors in each of the 3 buildings. 247.It follows that the cost estimates for the but‑for world advanced on behalf of the respondents are likely to have been seriously overstated and can hardly be relied on, because their conception of that world is simply unrealistic. G4b(ii). Unreliability in the data used by the respondents 248.As can be seen from the respondents’ contentions summarised above, of critical importance to their efficiency defence are the costs comparisons between the actual and the hypothetical worlds. In addition to the fundamental problem in conceiving the counterfactual discussed above, there are other serious shortcomings in the evidence put forward by the respondents, all of which tend to exaggerate either the inefficiencies in the counterfactual or the efficiencies attained under the Floor Allocation Arrangement. 249.The questionnaire filled in by the respondents after being interviewed by Dr Pang required them to list the total costs for each Package item based on a 2‑3 person flat in two different scenarios: Scenario I, where the respondents were effectively subject to both the Floor Allocation Arrangement and the Package Price Arrangement, and Scenario II, where no such cooperation agreement was in place. As explained by Dr Pang, the costs evidence for Scenario I elicited from the respondents was based on the assumption that each respondent was working on 4 flats on a single floor at the same time, rather than each respondent working on various flats on 4 floors in a building each 10 floors apart. In other words, the approach adopted took no account of any vertical travel between the allocated floors of a respondent when it was working on multiple flats on multiple floors at the same time. 250.At one stage Dr Pang sought to justify that approach by suggesting that the tenants came in batches by floor, but if she had been so informed by the respondents they certainly gave no evidence to that effect. On the contrary, the uncontradicted evidence of the Estate manager was that the intake registration was done in batches in sequence by building, but there was “no particular arrangement for floor‑by‑floor registration within each building”. Thus within each building each batch of newly registered tenants could comprise tenants randomly from any floor. 251.The respondents also tried to justify omitting all vertical travel in Scenario I by arguing that, so long as there were clusters of flats on each allocated floor, they could so organise their work among the flats on each floor as to take advantage of the efficiencies claimed. However, different tenants on the same floor could be admitted at quite different times. An example based on Luen Hop (R10)’s floors shows that the date of contract for flats on the same floor of a building could span over a wide period of time (eg from the second half of June to early October 2016) and be weeks or even months apart. Under the standard terms mandated by the Licence, the decoration work had to be completed within one month of the date of contract, thus imposing considerable constraints on the respondents’ ability to schedule their work. 252.Furthermore, as Mr Scott Adams, expert on construction planning and programming called by the Commission, testified, there are intrinsic obstacles to continuity of work even for flats on the same floor. Some trades are by nature slower than others so that faster trades may have to wait for preceding slower trades to be completed before they can start. Some trades have necessarily to be carried out in priority to other trades (eg brickwork has to be completed before plastering and then painting). In the absence of detailed as‑built records, he examined work continuity by reference to a generic example using 4 flats where the same activities in different flats are staggered but the activities are arranged to start as soon as possible. In this way he demonstrated that even with 4 flats on the same floor made available at once and with detailed planning using sophisticated project management software, there would be downtime when there was nothing that could be done on that floor. 253.In positing a situation where a respondent would only be working on the same floor without any need to travel between floors, Scenario I was in my view based on an unrealistic assumption of the degree of continuity possible. It follows that the costs figures produced by them for that scenario are significantly underestimated. 254.An area rife with exaggeration is the difficulties caused by having to travel between floors. First, there is the lift-waiting time. Although none of the respondents had actually timed it, they asserted that the normal waiting time for a lift for a work crew to go from one floor to another could “easily” be upwards of 20 minutes and could be up to 2 hours during peak times. There is a ring of truth in Mr Beard’s observation that it is easy to overestimate waiting times which can feel much longer than they in fact are. The Estate manager’s evidence in contrast is that he had not had to wait more than 10 minutes for the lifts at any of the 3 buildings even during peak times. Even in the worst‑case scenario which he had seen about 5 or 6 times in the 3 months from July to September 2016, he estimated the queue for the lifts would take up to 30 minutes to clear. He had never received any complaints from either the tenants or decoration contractors about excessive lift-waiting time. The experiments undertaken by Mr Adams together with several members of the Commission’s staff during the peak hours of early evening recorded waiting times of a far smaller order of magnitude. Admittedly such experiments could not replicate exactly the conditions during June to October 2016, when tenants were moving in and many flats were being decorated, but the tenants also started to reside there and the sort of waiting times claimed by the respondents would have been intolerable to all and yet no one mentioned anything to the Estate management. 255.Secondly, there is the additional time the respondents claim to be required for packing and unpacking tools when going from one floor to another. It was said that 8 or 10 minutes would be needed each for packing and unpacking. Whilst it might be more convenient to move the tools to another flat on the same floor (assuming it was nearby), I agree with Mr Adams who could not conceive that it took anything like the sort of time claimed to organise one’s tools in order to move to another floor and who therefore took a firm view that the respondents’ evidence was an exaggeration. 256.A significant plank of the respondents’ case is the reluctance of the skilled workers to go to a different floor during the day. It is also said that they would make use of surplus or leftover raw materials from one flat in another flat on the same floor but not in a flat on a different floor. Quite apart from the fact that there is no suggestion that this is an inalterable custom of the trade, there is likewise in my view exaggeration here. For example, Ma Yick Yin (R4) claimed that if a mason had completed floor tiling in a flat between 2 and 3pm, he would be prepared to install a threshold or windowsill in another flat during the remainder of the afternoon if the flat was on the same floor, but would simply “knock off” if the flat was on a different floor. I do not find that evidence credible. Mr Adams also confirmed it would be contrary to industry practice. 257.Then there is the actual increase in cost that working on different floors supposedly led to. Many of the cost estimates the respondents gave Dr Pang in the questionnaire seem to me arbitrary, self‑serving and exaggerated. For example, Mau Hang (R3) stated it cost $250 to install a ventilation fan in Scenario I, but $350 in Scenario II, and $180 to clear the debris in Scenario I, but $300 in Scenario II; Hip Yick (R6) stated it took 6 man‑hours to paint a flat in Scenario I, but 10 in Scenario II, or 4 man‑hours to install a door frame in Scenario I, but 6 in Scenario II; Wai Sun (R8) stated the cost of painting a flat is $2,000 in Scenario I, but $3,000 in Scenario II. These figures are not supported by any underlying documents nor in many cases credible calculations. In cross‑examination Dr Pang said she had herself recognised that there might have been an element of exaggeration in the figures supplied by the respondents, but she did not investigate it as she did not consider the magnitude of the cost savings to be an important part of her analysis. G4b(iii). Out‑of‑pattern flats 258.If the level of efficiencies claimed did result from a respondent being able to work on multiple flats on the same floor, this would mean that the respondent’s cost or expected cost of the work for an out‑of‑pattern flat would be higher. Dr Jenkins examined all the 35 out‑of‑pattern flats and identified 8 of them for which she could find suitable comparators that were decorated by the same respondent, are in the same building, have the same flat layout, and involved similar work scopes. Her analysis shows no systematic pattern of the works for out‑of‑pattern flats being priced higher than those for in‑pattern flats. She found that 4 of the out‑of‑pattern flats were priced lower than their comparators, 2 were materially the same, and 2 were higher. Admittedly the sample size was small and one could quibble about minor differences in the customers’ precise requirements between the out‑of‑pattern flats and the comparators (even though it was the Package items that were being compared), but these criticisms do not in my view remove all weight from the results. The respondents have not themselves sought to show empirically that in‑pattern flats were priced lower than out‑of‑pattern flats because of the expected substantial efficiencies claimed to flow from the Floor Allocation Arrangement. Moreover, the respondents’ own evidence shows that in some cases (eg Mau Hang (R3)), higher discounts were given for out‑of‑pattern flats than for in‑pattern flats. 259.The respondents themselves gave no explanation why there were no detectable signs at all that out‑of‑pattern flats were priced higher as they could not benefit from the efficiencies claimed. In the joint report Dr Pang surmised it could be because most of the relevant out‑of‑pattern flats were on lower floors and the respondents could have priced lower because of their preference to work on lower floors (though the respondents’ own evidence on their preferences was conflicting – see §123 above). In cross‑examination, however, Dr Pang came up with a theory that the respondents priced on the basis of expected efficiencies across whole buildings and those prices would be applied to all flats decorated by them without distinction. In other words, the out‑of‑pattern flats were “cross‑subsidised” by the in‑pattern flats. I am unable to accept this theory which is not only unsupported by any evidence from the respondents but seems to me to be contradicted by their evidence (eg Chan Kam Shui (R3)) that suggests the floor a flat was on was material to the discount they were prepared to give for that flat. G4b(iv). Magnitude of the efficiency gains 260.The concept of “improvement” in s 1(a)(i) of Schedule 1 imports a balancing of an agreement’s deleterious effect on competition against its benefits. As stated in Bellamy & Child, European Union Law of Competition (8th ed), §3.030, “[t]here will be no ‘improvement’ within the meaning of Article 101(3) if the disadvantages of an agreement in terms of competition are greater than its objective benefits”. Logically, as stated in §51 and §96 of the Article 101(3) Guidelines, the magnitude of the efficiency gains is a matter to be taken into account. 261.For the reasons above, there is no reliable basis in the present case to assess the magnitude of the claimed efficiencies. The respondents’ expert has herself disavowed having seriously assessed the magnitude of the benefits. Without any sense of magnitude, one lacks even a rudimentary basis on which to carry out any balancing. G4b(v). Disadvantages from the restrictions 262.On the other scale of the balance must be placed the disadvantages that the agreement creates in the field of competition. There is however very little evidence from the respondents in this respect, besides the constant refrain that they only won 33% and that there were “hundreds” of outside contractors competing. As I have explained above, I take the view that there was serious harm to competition resulting from the respondents’ agreements. The respondents have failed to grapple with the effect of the restriction of competition brought about by their agreements. G5. Condition 2 — benefit passed on to consumers 263.The second condition in s 1 of Schedule 1 is that the agreement must allow consumers a fair share of the resulting benefit. 264.In the case of cost efficiencies, it is relevant to see whether there is sufficient pressure of competition so as to induce the undertakings concerned to pass on a fair share of the cost savings to consumers. As pointed out in Bellamy & Child, European Union Law of Competition (8th ed), §3.038:
265.There is no rigorous assessment in the respondents’ evidence of the extent to which the cost savings claimed to result from the Floor Allocation Arrangement were likely to be passed on to the tenants, by reference to the characteristics and structure of the market, the nature and magnitude of the efficiency gains, the elasticity of demand, the magnitude of the restriction of competition,[34] the tenants’ price‑sensitivities, and the respondents’ own capacity constraints. For example, the incentive to pass on cost efficiencies may depend on whether an increase in demand will be generated by lowering prices, and on whether such increase in demand can be met by the respondents’ capacities. Instead, Dr Pang observed that:
266.Based on microeconomic theory, however, Dr Jenkins has demonstrated that with the Floor Allocation Arrangement in place, a respondent faced no competitive pressure from the other respondents and would be likely to price up to just below the outside contractors. There would be lower prices reflecting a transfer of the efficiency gains only if outside contractors have lower costs than all but the cheapest Appointed DC, for the equivalent service and quality offered. In her own report Dr Pang did not suggest this condition was met in practice. Instead, she stated that compared to outside contractors, the respondents were the lower‑cost providers because of the efficiency gains and scale economies, that the prices offered by the outside contractors were generally higher than those of the respondents (in the case of two respondents, 15% higher), and that even if some of the outside contractors’ prices might appear lower than the respondents’, the overall quality of the work offered by the respondents was likely to be superior (other things being equal). In this connection, it can be seen from Appendix 2 to this judgment that even the least efficient respondent according to the table at §230 above (ie Tai Wah (R7)) was able to secure contracts for 73 flats despite the presence of “hundreds” of outside contractors. 267.Given the attributes that differentiated the Appointed DCs from outside contractors for tenants to whom these attributes mattered, it would not be difficult for the respondents to price‑discriminate against them, ie charge a higher price to a tenant who wished to engage an Appointed DC (as opposed to an outside contractor) and who had little choice but to engage a particular respondent because of the Floor Allocation Arrangement. 268.The respondents further rely on the prices published on 13 promotional flyers obtained from a neighbouring public housing estate, namely, On Tai Estate, which had been completed more recently. It is said that the prices offered by the Appointed DCs in that estate were higher than those offered by the respondents in On Tat Estate, even adjusting for inflation, and that this is proof that the respondents were able to and did lower their prices by virtue of the Floor Allocation Arrangement. With respect, I am unable to place weight on this evidence. The expert report stated the flyers were collected by the expert at the site, but Dr Pang said in oral evidence that they were in fact given to her by the representative of one of the respondents, namely, Chan Yiu Kwai (R6). These were put forward as flyers for On Tai Estate (Phase 4), but many were in fact flyers for Phase 3. Further, there is no basis to consider On Tai Estate (Phase 4) an appropriate proxy for the “but for” world in relation to On Tat Estate (Phase 1): there is no information who precisely the Appointed DCs were and what their cost structures were, what pattern the distribution of flats decorated by the Appointed DCs took, and what mutual agreements or arrangements (if any) they had entered into. It also appears that the Appointed DCs did not share a site office in On Tat Estate (presumably with each setting up its own) or adopt any other cost saving joint endeavours that the Commission has not claimed to be infringements of the Ordinance. 269.On the basis of the data provided by the respondents themselves to Dr Pang, they had significantly different costs. Kam Kee (R5), for example, was substantially more efficient than the others judging from the data set out in the table at §230 above. With the Floor Allocation Arrangement in place, the less efficient respondents were insulated from any competition from the more efficient ones. The more efficient respondents had no incentive to price below the levels of the jointly fixed Package Prices which were profitable even for the less efficient respondents (eg the Flyer price of $21,880 for Package B for a 2‑3 person flat was still 25% higher than the cost of Tai Wah (R7) (ie $17,535), who had the highest cost among the respondents according to the table at §230 above). With competition among the Appointed DCs, the tenants would benefit from the incentive for the respondents to price closer to their underlying expected costs to increase the chance of winning contracts. 270.The elimination of rivalry from their closest competitors conferred upon the respondents the incentive and ability to raise prices. The tenants were effectively deprived of any choice between Appointed DCs because only one of them would be prepared to take on work on any particular floor. It is of particular importance in a case such as the present to balance the effect of the restriction on competition and the claimed benefits passed on to consumers. But the same deficiency in the respondents’ analysis of the negative effects of the restrictions flowing from the impugned agreements (see §262 above) is reflected in their failure to assess whether the cost savings said to have been passed on to the tenants are sufficient to compensate for the anti-competitive disadvantages. 271.On the evidence I do not think there is any basis to conclude that the tenants who engaged the respondents had in fact received a fair share, indeed any share, of the cost efficiencies claimed. G6. Condition 3 — indispensability 272.As the third condition for the defence, the statute requires that the agreement does not impose on the undertakings concerned restrictions that are not indispensable to the attainment of the efficiencies relied upon. Implicitly, this means that both the agreement and also the individual restrictions flowing from it must be reasonably necessary for the attainment of the efficiencies: see Article 101(3) Guidelines, §73. It is therefore relevant to inquire whether the efficiencies could be achieved by the undertakings on their own and, even if an agreement is necessary, whether another less restrictive type of agreement would suffice and whether the individual restrictions resulting from the agreement are each reasonably necessary in order to produce the efficiencies. 273.In my judgment the respondents fail on both counts. In her report, Dr Pang stated that she could think of no other less restrictive mechanism that would ensure that each respondent could concentrate their decorative work on a small number of specific floors so that each could take advantage of the scale economies and efficiencies from being able to conduct work in a larger number of flats on the same floor. 274.Dr Pang’s opinion that a market allocation agreement among the respondents was necessary to generate floor‑based concentration of work and hence the efficiencies relied on is clearly based on her conception of the but‑for world. For the reasons stated above in §§232-247, that hypothesis cannot be accepted. It seems to me that if the alleged efficiencies existed to the extent claimed, then, as many of the respondents’ witnesses testified, substantial concentration by floor would also be achieved under normal conditions of competition. 275.Further, even if some form of agreement was necessary for achieving floor‑based concentration of work, the respondents have failed to substantiate that the extent and intensity of restrictions imposed by the Floor Allocation Arrangement were reasonably necessary. One can think of other arrangements which, though anti‑competitive, would be less restrictive of competition. One example posited by the Commission is an arrangement between Appointed DCs that whichever of them wins the first flat on a floor wins the floor. This arrangement is feasible provided there is a mechanism among the respondents to determine and publish who won the first flat on a particular floor. Such a scheme, while still a contravention of the first conduct rule, would at least enable competition for entire floors — the respondents would compete with each other for the first flat on each floor, with benefit enuring to the tenants from this competition. The respondents have no valid answer to show why this will not work. 276.In sum, the respondents have clearly failed to show that the Floor Allocation Arrangement, both as an agreement among themselves and in terms of the actual restrictions it imposed, was indispensable for attaining the efficiencies alleged. G7. Condition 4 — possibility of eliminating competition 277.According to the fourth condition, the agreement must not afford the undertakings concerned the possibility of eliminating competition in respect of a substantial part of the goods or services in question. This implies that efficiencies attained by agreements must never be allowed entirely to outweigh the protection of competition. 278.This condition was not the direct subject of expert evidence in the present case. While by virtue of the Floor Allocation Arrangement competition among the respondents themselves was excluded, it does not necessarily follow that all competition was eliminated. On the other hand, while it is true that the respondents together won only 867 (ie 33.58%) out of the 2,582 flats, or (on my assumption, see §128 above) 38% of those that engaged decoration contractors, and while outside contractors decorated a substantial proportion of the other flats, it does not follow from this alone that the fourth condition is satisfied. The impugned arrangement virtually removed all internal competition among all contractors that were subject to HKHA’s supervision and regulation. There is some evidence that each Appointed DC’s capacity was only about 50‑70 flats at most[35] which would suggest the respondents had already virtually worked to their limit in On Tat Estate. Further analysis of the market and the services would be required before one can conclude that competition in respect of a substantial part of the services was not eliminated, taking into account factors such as to what extent outside contractors’ services were considered to be substitutable in the eyes of the tenants, the actual costs of production and capacity constraints. That analysis was not undertaken in this case. 279.It must be concluded therefore that the respondents have also failed to establish that the Floor Allocation Agreement satisfied the fourth condition in s 1 of Schedule 1. G8. Conclusion 280.For the above reasons, the respondents have in my view fallen very far short of satisfying the conditions of the efficiency defence. 281.The respondents have not sought to justify the Package Price Arrangement separately by reference to the efficiency defence. Instead, it is simply said that had the Package Price Arrangement not existed, the purpose of the Floor Allocation Arrangement would have been undermined. Since the defence fails in relation to the latter, it is equally to be rejected in relation to the former. H. THE SUB‑CONTRACTOR DEFENCE H1. The parties’ contentions 282.As mentioned above (§5(3)), both W Hing (R1) and Wide Project (R9) contend that they are not liable for any contravention of the first conduct rule on the ground that the renovation business in relation to On Tat Estate (Phase 1) was carried out by a separate undertaking. They only let their names be used by their respective “sub‑contractors”, and if any agreement had been made infringing the Ordinance, it was their sub‑contractors, as separate undertakings of which these two respondents did not form part, who had entered into such agreement without their knowledge. As their arguments are similar, I deal with them together after setting out the relevant facts respectively. 283.W Hing (R1) has submitted that the “sub‑contractors” are more accurately described as “assignees” of the Licence. Nothing however turns on the label used provided the actual relationship is clear, and for convenience I shall continue to use the term “sub‑contractors”. 284.For its part, the Commission does not dispute the facts relied on for the sub-contractor defence, but contends that on those facts the proper conclusion is that W Hing (R1) and Wide Project (R9) each formed a single economic unit with its sub‑contractor and, therefore, an undertaking, within the meaning of the Ordinance. As its secondary positions the Commission submits that the sub‑contracting arrangements should as a matter of public policy be refused recognition so that the respondent should be held responsible for all the acts of its sub‑contractor in On Tat Estate (Phase 1), and, further or alternatively, that on the basis of the approach adopted in Case C‑542/14 SIA ‘VM Remonts’ v Konkurences padome, the respondent should be held liable for the acts of its sub‑contractor on the ground that the respondent could reasonably have foreseen the anti‑competitive acts of the sub‑contractor and was prepared to accept the risk that they entailed. H2. Facts concerning W Hing (R1) 285.W Hing had been a decoration contractor on HKHA’s Reference List since 1994. However, because the scale of decoration works for public housing units was small with a relatively low profit margin, it was not W Hing’s business focus. Every time W Hing came top in the priority list and became an Appointed DC, it “sublet” the business to a “sub‑contractor” in return for a lump sum fee and did not become involved in the actual day‑to‑day operation of that business. 286.In the case of On Tat Estate (Phase 1), two days after receiving HKHA’s offer to appoint W Hing as an Appointed DC on 3 August 2016, one Ng Tat Leung George (a director of W Hing) signed the reply slips stamped with W Hing’s circular chop and sent them back to HKHA, accepting the appointment. 287.Shortly after W Hing indicated acceptance of the appointment, Philip Wong, another director of W Hing, was introduced to WC Wong, through an intermediary, as someone interested in undertaking decoration works for tenants in the Estate. Philip Wong did not know WC Wong before but he trusted the friend who introduced him, and agreed with WC Wong orally that W Hing would “sublet” the works to him on the following terms:
288.Pursuant to the agreement, WC Wong paid $200,000 to W Hing by a cheque dated 14 August 2015. 289.There was some confusion in Philip Wong’s oral evidence as to how the surety bond was obtained, but I am prepared to accept the consistent stance of Philip Wong and WC Wong that it was arranged by WC Wong, at his own cost, with Wing Lung Insurance Co Ltd. That evidence is inherently credible and the money trail for the payment could be readily verified with the insurer if need be. The bond was, however, sent in draft to W Hing and stamped with W Hing’s seal and signed by Philip Wong and another representative of W Hing. The bond was issued as a guarantee to HKHA for the performance of W Hing’s obligations as the Licensee. 290.On 10 June 2016, WC Wong also procured a contractors all risks policy issued by Wing Lung Insurance Co Ltd, pursuant to the requirement of the Licence. WC Wong paid the premium of $5,000 but the policy was issued to W Hing. The insured was stated to be “W Hing Construction Co Ltd as contractor &/or all its sub‑contractors and Hong Kong Housing Authority and Nice Property Management Ltd and Tenants as principal”; the contract was stated to be “Interior decoration and renovation works at On Tat Estate (Phase 1)”. 291.On 20 August 2015, WC Wong (and an assistant of his) attended the briefing session organised by HKHA together with Chu Lok Yin (W Hing’s Project Director) and Lui Kwok Leung (W Hing’s Quantity Surveying Manager). Mr Chu and Mr Lui signed the Licence on behalf of W Hing while WC Wong signed as a witness. Mr Chu and Mr Lui also each signed an authorisation letter to allow the Police to provide their criminal record searches to HKHA. 292.WC Wong arranged for one Yu Sung Ching (“Yu”) to attend the further briefing session on 30 December 2015. Yu was named as site supervisory staff for W Hing in the form submitted to HKHA. In around end of May 2016, Yu signed an authorisation letter for the police to disclose to HKHA information about any criminal record of himself. Philip Wong signed that letter as a witness, in his capacity as an authorised signatory of W Hing. The letter was also stamped with W Hing’s usual circular chop. 293.The respondents’ joint letter dated 30 May 2016 requesting for approval for the setting up of the Site Office was signed by Yu on behalf of W Hing with an oval chop bearing the Chinese name of W Hing. The agreement dated 9 August 2016 for use of the Site Office was signed by WC Wong and stamped again with the oval chop. That chop was apparently prepared by WC Wong or Yu for use in relation to On Tat Estate (Phase 1). 294.As a matter of fact, WC Wong and his associates were responsible for setting up the Site Office, printing the Flyers, approaching and signing contracts with tenants, and carrying out the decoration works in the name of W Hing. W Hing’s directors and “regular staff” were not involved. Name cards of Yu and others bearing the name of W Hing were printed though they were not in the usual design of the name cards of W Hing’s “regular staff”. The receipts issued to the tenants in the name of W Hing were signed by Yu and stamped with the oval chop. H3. Facts concerning Wide Project (R9) 295.Wide Project (R9) is named in these proceedings as an individual, Mr Yeung Kwok Yee (“Yeung”), trading in sole proprietorship in the name of Wide Project Engineering & Construction Co. Since around 1984, Wide Project had been approved as a decoration contractor on HKHA’s Reference List, but works for individual HKHA flats had never been Wide Project’s primary business. Instead, whenever Wide Project was offered the opportunity to become an Appointed DC, Yeung would seek a referral from one Mr Yue Tak Shing, a trusted senior person in the trade, for a reliable contractor to be appointed in Wide Project’s name. Since around 2005, Mr Yue had always referred one Lee Kwok Keung (“Lee”) to Yeung. Lee at first traded in the name of Park Lee Interior Design & Decoration but had since become a director of a company incorporated in 2011 called Park Lee Interior Design & Decoration Ltd. 296.When Yeung received HKHA’s notice dated 3 August 2015 offering to appoint Wide Project as an Appointed DC in relation to On Tat Estate (Phase 1), Yeung accepted the appointment the next day, signing the reply slip and stamping it with Wide Project’s chop. At around the same time, he sought a referral from Mr Yue who again suggested Lee for the purpose of taking up the appointment in Wide Project’s name. Yeung and Lee had a discussion and came to an oral agreement as follows:
297.On 11 August 2015, a surety bond was issued by Wing Lung Insurance Co Ltd to guarantee the performance of Yeung trading as Wide Project. It was signed and stamped by Yeung but presumably paid for by Lee. 298.Yeung did not attend the briefing session on 20 August 2015 as he was travelling on that date. Lee attended instead in the name of Wide Project. Yeung subsequently attended the management office of On Tat Estate on 24 August 2015 to sign the Licence. He also signed an authorisation letter for HKHA to make a criminal record check on himself with the Police. 299.Thereafter Yeung let Lee handle every aspect of the decoration business of On Tat Estate in Wide Project’s name. Yeung signed and stamped Wide Project’s chop on certain forms in June 2016 to report the names of workers to HKHA so that the necessary criminal record check could be conducted, though the information was filled in by Lee or his team. A contractors all risks policy was issued in favour of Wide Project on 8 June 2016 in terms similar to the policy obtained by W Hing (R1) described above, which must have been based on an application signed by Yeung. But, as with the bond, I proceed on the basis that it was paid for by Lee. 300.Lee was not an employee of Wide Project. In fact, at the time Wide Project had only one employee, who was Yeung’s secretary. Lee enlisted workers or partners (including Mak Sau Kwan, who shared the profits with Lee) to work in On Tat Estate though they were reported by Wide Project to HKHA as Wide Project’s workers. The work orders did not specify the decorator as Wide Project but there can be no doubt that Lee and his team were working in the name of Wide Project and the tenants engaged them on that basis. In his evidence Yeung also accepted Wide Project was party to the decoration contracts. It is accepted that Lee did not have to apportion his profits with or account for them to Yeung, and did not report to Yeung in relation to the decoration works at On Tat Estate. H4. The legal principles 301.As defined in s 2 of the Ordinance, an undertaking
302.Following EU law, the competition rules in the Ordinance apply to the actions of undertakings. The key organising concept is that of the single economic unit. Thus the first conduct rule, as formulated, prohibits relevant agreements between undertakings rather than between persons. As the CJEU stated in Akzo Nobel and Others v Commission (Case C‑97/08P) (at §55), the concept of an undertaking “must be understood as designating an economic unit even if in law that economic unit consists of several persons, natural or legal …” and (at §56): “When such an economic entity infringes the competition rules, it falls, according to the principle of personal responsibility, to that entity to answer for that infringement”. 303.While the competition rules target undertakings, in accordance with general legal principles, applications brought in the Tribunal and orders made by the Tribunal are necessarily addressed to legal entities. Thus ss 92‑94 of the Ordinance allow the Commission to apply for and the Tribunal to make orders against persons who have contravened or been involved in a contravention of a competition rule. Accordingly, when liability is to be attached for infringement by an undertaking, the infringement must be imputed to one or more person against whom legal proceedings may be instituted and on whom orders including financial penalties may be imposed. 304.There have been many cases in the EU raising the question whether or not separate entities form a single economic unit and therefore one undertaking. The question has arisen in different contexts, such as whether two or more entities form a single undertaking so that any agreement among them does not fall within Art 101(1) of the TFEU as it is not an agreement “between undertakings”,[36] as in C‑73/95P Viho v Commission. Another context, more pertinent for present purposes, is whether certain entities together form an undertaking so that the infringing conduct of one entity is attributable to the others on the basis they are comprised within the same undertaking, as was the question in Akzo Nobel. 305.The question has arisen in the cases with regard to a parent company and its subsidiary, a principal and his agent, and a firm and its commercial representative. Various factors may be relevant to the question and different factors may attract different weight in different types of situation. Thus, in the context of a parent and subsidiary, an important factor is whether the parent exercises “decisive influence”[37] over the relevant activities or policies of the subsidiary. In relation to principal and agent, the case law indicates that where the agent bears no financial risk, even if it might look like a separate entity, it is to be treated as an “auxiliary organ” of the principal, and on that basis arrangements such as the principal dictating pricing to the agent are not caught as price fixing agreements because they are between two parts of the same undertaking.[38] 306.Another approach that has been used, which seems to me to be more apt for present purposes, is to ask whether or not there is unity in the conduct of the relevant entities on the market: Case T‑66/99 Minoan Lines SA v Commission, §123; Case 48/69 ICI v Commission [1972] ECR 619, §140; Case T‑541/08 Sasol v Commission, §31. 307.Whether third parties perceive the principal and agent as forming a single economic entity is a relevant factor, as shown by Minoan Lines. In that case the EU Commission had found that several undertakings (including Minoan Lines) which provided ferry services between Greece and Italy had infringed Art 85(1) of the EC Treaty by agreeing on the levels of fares on certain routes. Minoan Lines attacked the decision on the ground that the Commission had wrongly imputed to it the activities of European Trust Agency (ETA). The Court of First Instance of the EU rejected the challenge, laying emphasis on the fact that the relevant actions and initiatives of ETA were not undertaken by ETA in its own name and on its own behalf, but in the name of and on behalf of Minoan Lines (see §§109 & 129). It will be noted that the facts of that case are different from those here, in particular because ETA took on no financial risk in connection with the relevant business. 308.W Hing (R1) placed reliance on EU Commission Regulation No. 1407/2013 which laid down an exhaustive list of criteria for determining when two or more enterprises within the same Member State are to be considered as a single undertaking. This regulation, however, concerns the application of Arts 107 and 108 of the TFEU in connection with the question of the de minimis state aid that a single undertaking may receive per Member State. It seems to me the list of criteria which was introduced in order to provide certainty and reduce the administrative burden in that specific area provides no assistance in the present case. H5. Discussion 309.The issue arising in the present case is not similar to that in parent‑subsidiary cases or principal‑agent cases, and neither control and influence nor financial risks seem to me to be the focal point of enquiry. Rather, in a case such as this whose facts are not comparable to any decided cases counsel can locate, one needs to begin with the statute. In defining “undertaking”, the statute directs attention to the questions of (i) what is the relevant economic activity that was engaged in; (ii) what is the entity engaged in that activity; and (iii) what persons are comprised in that entity. 310.The Licence was offered to and signed by W Hing (R1) and Wide Project (R9). Only they and not their sub‑contractors were on the Reference List and as such eligible for the Licence. It was the respondent who had to have fulfilled the relevant financial and other eligibility requirements. It was the respondent who was the Appointed DC. It is not in dispute that the Licence was a contract between HKHA and the respondent, not its sub‑contractor. While the respondent in effect purported to give the sub‑contractor the benefits of this contract, its obligations thereunder could not be assigned. It is the respondent who was and remained throughout responsible to HKHA for compliance with the conditions of the Licence. 311.The terms of the Licence made clear it was the respondent who would be carrying out the decoration works in the Estate. No rights or obligations under the Licence could be either sub‑contracted or assigned by the respondent without written approval of HKHA, which was never sought by W Hing (R1) or Wide Project (R9). Indeed, HKHA was not aware of the existence of the sub‑contractors. 312.As was accepted by W Hing (R1) and Wide Project (R9), if the performance of the decoration works was poor, HKHA would hold them responsible. The monitoring scheme applied over a rolling three‑year period and thus covered the works by the respondent on any projects during that time irrespective of the personnel undertaking the actual work. Further, under the Licence it was the respondent who was required to indemnify HKHA in respect of any loss or damage. 313.Under the Licence it was the respondent who was required to maintain a surety bond and proper insurance cover. The bond and the policy were both issued in the name of the respondent, not the sub‑contractor. It was the respondent, not the sub‑contractor, who was party to the relevant contracts with the insurance company. 314.The Site Office was applied for in the name of the respondent, not the sub‑contractor. The licence to occupy the Site Office was issued to the respondent. The desk in the Site Office was labelled with the name of the respondent. It was the respondent who was contractually liable to HKHA for the rent. 315.The decoration works were promoted to the tenants in the name of the respondent. The sales staff distributed business cards bearing the name of the respondent. The work orders and receipts for payment were issued to the tenants in the name of the respondent. The contracts with the tenants were plainly made in the name of the respondent. If a tenant defaulted in payment, a claim could only be made against him by the respondent. A claim made in the name of WC Wong or Lee would be liable to be struck out. Likewise, if a tenant made a contractual claim in relation to the works, whether for defective workmanship or otherwise, it would have to be made against the respondent with whom there was a contract. If a tenant wished to enforce the six‑month warranty contained in the standard terms and conditions printed on the work order, he would have to pursue the respondent (although the respondent might then ask the sub‑contractor to rectify the defects). 316.The picture emerged clearly in the cross‑examination of the witnesses of W Hing (R1) and Wide Project (R9) respectively:
317.On these facts, the answers to the questions posed in §309 above are in my view clear. The relevant economic activity was the commercial provision of decoration services to individual tenants in On Tat Estate (Phase 1) in the period from June to November 2016. The entities engaged in that activity were, so far as relevant, the entities operating there in the name of W Hing and Wide Project respectively. In each case the entity comprised both the respondent and the sub‑contractor. In order to undertake and deliver the economic activity which is the subject matter of the infringements, both the respondent and the sub‑contractor were necessary and had to act together. Decoration services were offered in the name of the respondent. It was the respondent that entered into contractual relationship with and undertook obligations to HKHA and the tenants. The respondent, for a fixed sum, permitted the sub‑contractor to provide decoration services in the name of the respondent and in discharge of the respondent’s obligations to the tenants, in return for fees paid by the tenants which the sub‑contractor was allowed to keep. The respondent completely delegated to the sub‑contractor the physical carrying on of the business in its name. There is, as it seems to me, unity in their conduct on the market. It was a joint exercise in which the respondent brought in its status as an Appointed DC and the associated rights and took on responsibilities and obligations towards HKHA and the tenants, whilst the sub‑contractor brought in workers, raw materials, site supervision and management. 318.W Hing (R1) and Wide Project (R9) emphasise that their sub‑contractors bore the risks of profit and loss on the decoration business. It is argued that this makes them independent traders on their own account and separate undertakings from the respondents. Reliance is placed on parts of the European Commission’s Guidelines on Vertical Restraints (2010/C 130/01) (§§13‑21) and an earlier Notice on exclusive dealing contracts with commercial agents (OJ 139, 24 December 1962, pp 2921‑62). These materials, which concern the question whether an agreement is a true “agency agreement” so that it falls outside the scope of Art 101(1), do not in my view go to the crux of the problem in this case. Nor do the decisions in Case C-217/05 Confederación Española de Empresarios de Estaciones de Servicio v Compañía Española de Petróleos SA andCase T‑418/10 Voestalpine AG v Commission,which are principal and agent cases relied upon by Wide Project (R9), meet the focus of this case. 319.Although as between the respondent and its sub‑contractor, it was the latter who took the profit or bore the loss, it would be incorrect to suggest that the respondent bore no risks. The respondent was the contracting party with the outside world and clearly assumed the risks externally. If the sub‑contractor disappeared or became insolvent or otherwise incapable of further performance, it was the respondent who had to pick up the pieces vis-à-vis HKHA and the tenants. If the sub‑contractor performed poorly, it would be the respondent who had to face disciplinary action by HKHA and claims by the tenants. The agreement between the respondent and the sub‑contractor could hardly be raised as a defence to any action by HKHA or the tenants against the respondent. 320.W Hing (R1) and Wide Project (R9) rely on the fact, inter alia, that: (i) each of them received a fixed fee of $200,000 from the sub‑contractor but did not thereafter derive any profit or income from the decoration works; (ii) they did not incur any expenditure in relation to the performance of the decoration works; (iii) the workers who actually carried out the decoration works were not employees on their payroll, but were engaged by the sub‑contractors; and (iv) the surety bond and insurance policy were procured at the cost of the sub‑contractors. These are not in dispute but they concern without exception the internal, private relationship between the respondent and the sub‑contractor. 321.From the point of view of HKHA, the other respondents, the tenants who engaged them for decoration works and, indeed, the rest of the world, a very different picture emerged, as explained above. The part played by the respondent is not just a matter of appearance, but is substantiated by a wide range of legal and economic obligations actually retained by W Hing (R1) and Wide Project (R9) in relation to the works undertaken by their “sub‑contractors”. 322.It would simply be wishful thinking for the respondent to assume, as submitted by counsel, that it dropped “completely out of the picture” upon admitting a subcontractor to “take over” the works. As far as the outside world was concerned, the sub‑contractor never came into the picture and the respondent never went out of it. 323.It has been submitted on behalf of W Hing (R1) and Wide Project (R9) that it would be open to the Commission to bring proceedings against the sub‑contractors instead of against them. In my judgment, while the sub‑contractor may well be liable for the infringing conduct as another person comprised in the same undertaking, this does not preclude the Commission from proceeding against the respondent who, on the above analysis, must be regarded as a person comprised in the undertaking in question and therefore liable for the contraventions of the first conduct rule. H6. Tai Dou (R4) 324.Tai Dou (R4) is a partnership consisting of 4 partners, being the two persons named in the Originating Notice of Application (namely, Cheung Yiu Fai Danny and Wong Tung Hoi) as well as Mr Pacquet Wong and Madam To Suet Chun.[39] It attempted at a very late stage to raise a similar argument but, for the reasons set out in this Tribunal’s Reasons for Decision dated 31 October 2018, was not allowed to do so. In so far as counsel sought in closing submissions to run the sub‑contractor defence on behalf of Tai Dou (R4), I consider that the argument is inadmissible. 325.On the basis of its existing Response (filed jointly with the 2nd, 3rd, 6th, 7th, 8th and 10th respondents), Tai Dou (R4) admits that it attended the worship ceremony at the Site Office on 16 June 2016 and that lots were drawn there (though it is averred it was for the purpose of allocating desks in the Site Office). From the evidence, it can be seen that Pacquet Wong, an undisputed partner of Tai Dou (R4), signed the reply accepting the appointment as an Appointed DC, signed the surety bond, signed the Licence, signed and submitted to HKHA a list of staff and workers, attended the briefing session on 20 August 2015 and named KC Ho as the contact person. The evidence shows that those working in On Tat Estate (Phase 1) in the name of Tai Dou (R4) included KC Ho and Ma Yick Yin. 326.A witness statement of KC Ho was originally filed on behalf of Tai Dou (R4), in which he described himself as the General Manager of Tai Dou (R4). In early October 2018, Tai Dou (R4), having replaced its previous lawyers, indicated that it would not be calling KC Ho. The 2nd, 3rd, 5th, 6th, 7th, 8th and 10th respondents then indicated their intention to call him and also filed a witness statement of Ma Yick Yin. Eventually, KC Ho was not called to testify, and his statement is therefore to be disregarded. Ma Yick Yin was called, and said that he considered KC Ho the general manager of Tai Dou (R4), that he worked in On Tat Estate (Phase 1) in accordance with KC Ho’s instructions, and that he and the workers used business cards with Tai Dou (R4)’s name. There are also receipts issued to tenants bearing an oval chop stating “Tai Dou Building Co” in both Chinese and English. 327.Despite KC Ho’s absence, based on all the other evidence there is no doubt that KC Ho and the team working there in the name of Tai Dou (R4) had engaged in the Floor Allocation Arrangement and the Package Price Arrangement. Counsel’s argument that there is a possibility that Tai Dou (R4) did not enter into any arrangement with others but simply obtained decoration work from the remaining floors because they were not allocated to any of the other respondents, seems to me, with respect, to be wholly unrealistic. 328.It is clear that the undertaking operating in On Tat Estate (Phase 1) was the partnership Tai Dou. The two named persons each being a partner must be held responsible as persons comprised in the undertaking. Counsel’s argument that there is nothing to show that they personally took part in or were aware of the conduct impugned is beside the point. H7. Conclusion 329.For the above reasons the sub‑contractor defence fails. It is accordingly unnecessary to deal with the Commission’s arguments based on illegality or public policy and on VM Remonts. I. SUMMARY AND CONCLUSION 330.Between June and November 2016, the respondents, each an Appointed DC in On Tat Estate (Phase 1), made and gave effect to an agreement — the Floor Allocation Arrangement — with one another, whereby they were each allocated 4 floors in each of the 3 buildings in the pattern set out in §29 above for the purposes of taking up decoration business. They agreed not to actively seek business from tenants on floors allocated to the other respondents. If approached by those tenants, they would direct them to the respondents who had been allocated the relevant floors and, unless the tenants insisted, they would decline business from them. This agreement was implemented, resulting in the pattern of distribution of the flats decorated by the respondents as shown in Appendix 2. Such conduct consisted of the allocation of market for the supply of services and hence constituted “serious anti-competitive conduct”. 331.During the same period of time, the respondents made and gave effect to an agreement — the Package Price Arrangement — with one another, whereby they agreed on the prices of the Packages to be put on a joint Flyer for On Tat Estate (Phase 1). Each of them agreed to and contributed financially to the printing of the Flyers. The Flyers were used by the respondents for promotional purposes and distributed to tenants. The prices in the Flyer served as a starting point or anchoring reference point for negotiations with many if not all customers where their requirements included the Package items. In a number of contracts, the prices in the Flyer were or were near the final prices. Such conduct consisted of fixing the price for the supply of services and therefore also constituted “serious anti-competitive conduct”. 332.Both the Floor Allocation Arrangement and the Package Price Arrangement had the object of preventing, restricting or distorting competition in Hong Kong. 333.As a matter of construction, the burden of establishing the efficiency defence on the balance of probabilities lies on the respondents. While this engages the presumption of innocence, the incidence of the burden satisfies the 4-step proportionality analysis. 334.The efficiency defence fails as this Tribunal is not satisfied that the agreements in question generated efficiencies that compensated for the harm resulting from the restriction of competition imposed by the agreements, or that they allowed the respondents’ customers a fair share of the resulting benefit, or that they imposed restrictions that were indispensable to the attainment of the objectives, or that they did not afford the respondents the possibility of eliminating competition in respect of a substantial part of the services in question. 335.Notwithstanding that W Hing (R1) and Wide Project (R9) had, in return for a fixed lump sum fee received in advance, let other persons carry out the operations of the decoration work at On Tat Estate (Phase 1) in their names, they were comprised in the undertakings that were engaged in the provision of decoration services there and which made and gave effect to the agreements impugned. 336.Accordingly, each of the 10 respondents is a person or (in the case of partnerships) are persons who have contravened the first conduct rule. 337.The orders sought by the Commission include (i) a pecuniary penalty under s 93 in respect of each contravention; (ii) declarations of contravention; (iii) an order under s 96 for the respondents to pay to the Government sums in respect of the costs of and incidental to the investigation into their conduct; and (iv) costs of the proceedings in the Tribunal. These matters are to be determined after a further hearing.
Mr Daniel Beard QC, Mr Abraham Chan SC and Mr Byron Chiu, instructed by King & Wood Mallesons, for the Applicant Mr Harrison Cheung, instructed by Wong & Lawyers, for the 1st Respondent Mr Carter Chim, Mr Thomas WK Wong and Ms Joyce YY Kwok, instructed by JCC Cheung & Co, for the 2nd and 3rd Respondents Ms Sharon Ng, instructed by Henry Wan & Yeung, for the 4th Respondent Mr Hectar Pun SC (on 18-20 December 2018 only), Mr Anson Wong Yu Yat and Ms Allison Wong, instructed by JCC Cheung & Co, for the 5th Respondent Mr Richard Yip and Ms Tara Liao, instructed by JCC Cheung & Co, for the 6th, 7th, 8th and 10th Respondents Ms Connie Lee and Mr Tommy Cheung, instructed by Benjamin Au & Billy Chan, Solicitors, for the 9th Respondent APPENDIX 1
Oi Tat House
Shing Tat House
[1] [2018] HKCT 6 (Reasons for Decision dated 31 October 2018). [2] HKHA also operates a “Home Ownership Scheme” (“HOS”) whereby flats in HOS buildings erected by HKHA are sold to successful eligible applicants at subsidized prices. The Decoration Contractor System is also operated for new HOS buildings. For the purposes of this case, however, references to HOS buildings and flats will be omitted. [3] R2‑3’s Closing Submissions, §§43‑47; R4’s Closing Submissions, §§38‑44; R5’s Closing Submissions, §52; R6‑8 & 10’s Opening Submissions, §4. [4] The quoted version is taken from the supplemental statement of Chim Lap Fan (R2), para 10(2). There were very minor but immaterial variations for the other witnesses. [5] §299. [6] §394. [7] which R4 subsequently unsuccessfully sought to disown: see the Tribunal’s Reasons for Decision dated 31 October 2018 [2018] HKCT 6. [8] Annex to the Commission’s Closing Submissions. [9] See United States v Socony-Vacuum Oil Co, 310 US 150, 221 (1940). [10] Verizon Communications v Law Offices of Curtis V. Trinko, 540 US 398, 408 (2004). [11] Now the Competition and Consumer Act 2010, as amended by the Competition and Consumer Amendment (Competition Policy Reform) Act 2017. [12] Treaty on the Functioning of the European Union. [13] Bills Committee on Competition Bill (Responses to follow-up questions arising from previous meetings) — paper for discussion on 21 June 2011, CB(1)2420/10‑11(02), §2. [14] See clause 2 of the Notice. [15] Eg Ibáñez Colomo and Lamadrid, “On the Notion of Restriction of Competition: What We Know and What We Don’t Know We Know”, in Gerard, Merola and Meyring (eds), The Notion of Restriction of Competition: Revisiting the Foundations of Antitrust Enforcement in Europe (Brussels: Bruylant, 2017), pp 333‑373. [16] See para 107 of Van den Bergh Foods which in turn cites a long line of decisions of the General Court. [17] Art 101(3), which is the equivalent of s 1 of Schedule 1 to the Ordinance, is quoted in §148 below. [18] 867 ÷ 2,582 × 100% [19] In the case of the 2nd, 3rd, 4th, 6th, 7th, 8th and 10th respondents — see para 80 above. [21] Section 8 of Hong Kong Bill of Rights Ordinance (Cap 383). [22] It may also be noted that, in relation to abuse of dominance, it has also been held by the Grand Chamber of the CJEU that where a dominant undertaking contends that the exclusionary effect its conduct has produced may be counterbalanced by efficiency gains, it is for the undertaking to show that the efficiency gains counteract any likely negative effects on competition, that such conduct is necessary for the achievement of those efficiency gains and that it does not eliminate effective competition: Case C-209/10 Post Danmark A/S v Konkurrencerådet, §42. [23] Pursuant to Art 51(1) of the Charter: “The provisions of this Charter are addressed to the institutions, bodies, offices and agencies of the Union with due regard for the principle of subsidiarity and to the Member States only when they are implementing Union law. They shall therefore respect the rights, observe the principles and promote the application thereof in accordance with their respective powers and respecting the limits of the powers of the Union as conferred on it in the Treaties.” [24] Art 48(1) reads: “Everyone who has been charged shall be presumed innocent until proved guilty according to law.” [25] It provides: “The section 54 prohibition [ie prohibition of mergers resulting in a substantial lessening of competition] shall not apply to any merger if the economic efficiencies arising or that may arise from the merger outweigh the adverse effects due to the substantial lessening of competition in the relevant market in Singapore.” [26] It provides: “The Tribunal shall not make an order under section 92 if it finds that the merger or proposed merger in respect of which the application is made has brought about or is likely to bring about gains in efficiency that will be greater than, and will offset, the effects of any prevention or lessening of competition that will result or is likely to result from the merger or proposed merger and that the gains in efficiency would not likely be attained if the order were made.” [27] See paragraphs 1.14‑1.15 & 4.83 of the Guidelines on Mergers and Acquisitions in Hong Kong Telecommunications Markets issued in 2004 by the Telecommunications Authority pursuant to s 6D(2)(aa) of the Telecommunications Ordinance. [28] Vivien Rose, “Reverse burdens of proof: the golden thread in Article 101(3) TFEU”, BEG/ALBA Annual Conference (May 2011). [29] Paras 38 and 39. [30] Para 45. [31] Para 52. [32] Kist & Centella, Coherence and Efficiency in a Decentralised Enforcement of EC Competition Rules: Some Reflections on the White Paper on Modernisation, in Ehlermann and Atanasiu (eds), European Competition Law Annual 2000: Volume 5, The Modernisation of EC Antitrust Policy (Hart Publishing 2001), at p 375. [34] These four factors being factors suggested in Article 101(3) Guidelines, §96. [35] According to Chan Yiu Kwai (R6). [36] Although the first conduct rule in the Ordinance does not contain the phrase “between undertakings”, the Commission has recognised that the same approach applies: see the Commission’s Guideline on the First Conduct Rule, §§2.6-2.10. [37] Akzo Nobel, para 60. [38] See the Commission’s Guideline on the First Conduct Rule, §2.14. [39] It is not clear why only two of the four partners have been named as respondents, but the Commission accepts that as such, only the two named persons can be subjected to any sanctions and orders imposed by the Tribunal in these proceedings. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
) or 33.6% of finding the “correct” respondent. Such an approach does not seem to me realistic. Competitive markets do not work in that manner. If the alleged efficiencies existed, the respondents would compete in price, and the tenants would be incentivised to search for the contractor that “matched” his floor. Dr Pang criticised that this assumed each respondent’s cost of approaching and marketing to tenants is negligible. It is to be recalled, however, that savings in marketing cost resulting from a market allocation agreement cannot be treated as an objective benefit (see §213 above).
Cases cited in this judgment
Other judgments that cite this case
Further hearings and rulings under CTEA 2/2017