TY - JOUR
AU - Kwok, Kelvin Hiu, Fai
AB - I. INTRODUCTION Nutanix1 is the very first case brought and decided under the Hong Kong Competition Ordinance (HKCO) (Cap 619), which came into full effect back in 2015 as the first piece of cross-sector competition legislation in Hong Kong. Under the HKCO’s ‘judicial enforcement’ model, the case was brought by the Hong Kong Competition Commission (HKCC) and decided by the Hong Kong Competition Tribunal (HKCT).2 The decision addressed a number of substantive issues concerning the elements of the First Conduct Rule (FCR) in the HKCO3 (which is modelled upon Article 101(1) of the Treaty on the Functioning of the European Union (TFEU)), as well as procedural issues such as the standard of proof. For the purpose of this article, I will focus on two particular issues about the ‘agreement’ element under the FCR. The other substantive issues, including the finding of an anti-competitive object, are relatively less controversial as the case is about bid-rigging. The facts of the case may be briefly stated as follows. Nutanix is a supplier of server systems. It has a distribution network consisting of both distributors (including Innovix and SiS) and resellers (including BT and Tech-21). The Hong Kong Young Women’s Christian Association (YWCA) sought to procure the Nutanix-NX server solution in accordance with its tender procedures. The first tender—in which only BT took part—failed due to an inadequate number of tenders. BT proceeded to collude with other members of Nutanix’s distribution network through the coordination of Nutanix to win the second tender by having the other members submit ‘dummy bids’, ie bids ‘not intended to win but just to make up the numbers’.4 The HKCC argued that there was an ‘object’ infringement of the FCR based on the following agreements5: an agreement between Nutanix (through Hung as Nutanix’s Territory Account Manager) and BT (through Chan as BT’s Technical Pre-Sales Manager), whereby Nutanix would procure dummy bids from members of Nutanix’s distribution network (through employees who are Hung’s friends); an agreement between Nutanix (through Hung), BT (through Chan) and Innovix (through Siu as Innovix’s Product Manager), whereby Innovix would submit a dummy bid; an agreement between Nutanix (through Hung) and Tech-21 (through Ng as Tech-21’s Sales Manager), whereby Tech-21 would submit a dummy bid; and an agreement between Nutanix (through Hung) and SiS (through Shek as SiS’s Product Manager), whereby SiS would submit a dummy bid. The HKCT held that agreements (1)–(3) were established on the facts, but not agreement (4)—there was no agreement with SiS because, unlike the other employees, Shek’s behaviour should not be attributed to SiS.6 The HKCT proceeded to find that agreements (1)–(3) had the object of restricting competition, contrary to the FCR.7 This article, in examining the ‘agreement’ aspect of the decision, questions the approach taken to the analysis of a ‘hub-and-spoke’ arrangement and that of corporate attribution in Nutanix. Both of these issues are relevant beyond Hong Kong and chosen precisely for this reason. II. FIRST ISSUE: ‘HUB-AND-SPOKE’ ARRANGEMENT Nutanix involves a ‘hub-and-spoke’ arrangement, namely a collusive arrangement between actual or potential competitors (as ‘spokes’) facilitated by a player at a different level of the supply chain—say a supplier or customer (as the ‘hub’).8 The US case of Interstate Circuit provides a classic example of a ‘hub-and-spoke’ arrangement.9 Eight film distributors received the same letter from Interstate Circuit, a major first-run theatre operator, requesting the distributors to place certain restrictions on subsequent-run theatres. Such restrictions would undermine the competitiveness of subsequent-run theatres by forcing them to elevate their prices. The letter listed the name of each of the distributors, who responded by imposing the restrictions as requested. The US Supreme Court, in holding that the distributors violated section 1 of the Sherman Act, did not consider it necessary to find an express agreement (based on express communication between the distributors), given the evidence supported the finding of a ‘tacit agreement’10 in the following manner: It was enough that, knowing that concerted action was contemplated and invited, the distributors gave their adherence to the scheme and participated in it. Each distributor was advised that the others were asked to participate; each knew that cooperation was essential to successful operation of the plan. They knew that the plan, if carried out, would result in a restraint of commerce, which … was unreasonable within the meaning of the Sherman Act, and knowing it, all participated in the plan. … Acceptance by competitors, without previous agreement, of an invitation to participate in a plan, the necessary consequence of which, if carried out, is restraint of interstate commerce, is sufficient to establish an unlawful conspiracy under the Sherman Act.11 By ‘previous agreement’, the Court was referring to a prior express agreement amongst the film distributors. The Court’s point was that in the absence of an express agreement, section 1 of the Sherman Act would still be triggered by a tacit agreement (ie ‘concerted action’ contemplated, invited, and accepted), here being the distributors’ tacit agreement as facilitated by their customer, namely Interstate Circuit,12 who was supposedly also a party to the agreement.13 The agreement embodied the parties’ joint intention to impose restrictions on subsequent-run theatres. The agreement took the form of a ‘hub-and-spoke’ arrangement, whereby Interstate Circuit, as the ‘hub’, coordinated and facilitated the collusion between the competing film distributors as the ‘spokes’.14 Similar to other antitrust agreements,15 the ‘hub-and-spoke’ agreement may be conceptualized as the product of offer and acceptance16: the ‘hub’ issued to the ‘spokes’ an offer to participate in an anti-competitive scheme, which the ‘spokes’ accepted by ‘g[iving] their adherence to the scheme and participat[ing] in it’.17 The ‘hub-and-spoke’ arrangement may instead be analysed as a concerted practice rather than an agreement under a EU-based legal framework such as the FCR.18 The ‘hub-and-spoke’ arrangement in Nutanix was set up to coordinate the submission of bids by several IT equipment vendors (as ‘spokes’) in YWCA’s second tender. Nutanix, the server system supplier, acted as the ‘hub’ in coordinating and facilitating the collusion between the bidders. The ‘hub’ and the ‘spokes’ all shared the intention to rig the tender process by using ‘dummy bids’ to increase the chance of success of one of the bidders (which happened to be BT).19 Unlike the arrangement in Interstate Circuit, the role of the ‘hub’ was played by a supplier, rather than a customer, of the colluding competitors. Another difference is the incentive on the part of the ‘hub’. In Interstate Circuit, the collusion amongst the distributors (the ‘spokes’) to impose restrictions on subsequent-run theatres directly benefited the ‘hub’, namely Interstate Circuit, as a competitor of those theatres. In these circumstances, the ‘hub’ would have a natural incentive to ensure the success of the spokes’ collusion. In Nutanix, however, it may be questioned why the ‘hub’, namely Nutanix, had the incentive to orchestrate the bid-rigging scheme to favour one specific member of its distribution network, namely BT.20 After all, the tender was specifically for a Nutanix server system: so long as the tender was successful, Nutanix would presumably earn a profit as the server system supplier irrespective of who the successful bidder was.21 If the concern were a lack of interest in bidding (given that YWCA needed a minimum of five bids for a successful tender22), could Nutanix not have overcome the problem by encouraging members of its distribution network to participate in the tender process, rather than having BT as the ‘chosen winner’?23 Nutanix’s incentive to orchestrate the collusion in favour of BT is somewhat a mystery, although the HKCT did note from the evidence that ‘Nutanix was very keen to make sure that BT would win the deal’.24 Notice that BT was party to agreement (2) (with Innovix), but not agreement (3) (with Tech-21) or (the alleged) agreement (4) (with SiS). This difference was apparently due to awareness (or lack of awareness) of BT’s involvement. Agreement (2) was constituted in circumstances where Innovix was aware that its cover bid was meant specifically to ‘help BT’ (as stated in Hung’s email to Siu).25 In contrast, there was no evidence suggesting that Tech-21 (through Ng) or SiS (through Shek) was aware of BT’s participation in the tender process, or the identity of other bidders.26 However, are we right in assuming that knowledge of another party’s identity or existence ought to be a prerequisite for having an agreement with that party?27 In The Satanita, an English contract law decision, the Court of Appeal was satisfied that participants of a yacht race had a contractual agreement with one another (and with the managing committee) based on the sailing rules adopted by the committee.28 The participants might not know each other, or indeed how many people were taking part in the race, but this did not prevent the Court from finding an agreement between all of them.29 This finding is justified by the fact that when a participant joined the race, the participant shared the same objective intention with other participants (whoever they were and however many of them) and the committee to conduct the race on the basis of the rules. Applying the same logic to antitrust agreements—which, as I have argued elsewhere,30 should be subject to an even lower standard of precision than contractual agreements31—the Tribunal could have found that BT was party to not only agreement (2), but also agreements (3) and (4) (subject to the corporate attribution issue) notwithstanding BT’s involvement was unknown to Tech-21 or SiS, on the basis that Nutanix, BT and the other distribution partner (Innovix, Tec-21 or SiS) shared the common objective intention to have the other partner submit a dummy bid.32 Besides, was it actually necessary for the HKCT to treat the ‘hub-and-spoke’ arrangement as four separate agreements? Although the multiple agreement approach is consistent with how ‘hub-and-spoke’ arrangements are analysed in the UK,33 could the HKCT not have found one overarching agreement between Nutanix, BT and the other distribution partners (ie Innovix, Tech-21 and SiS34)? Notice that the joint intentions under agreements (1)–(4) regarding the procurement and submission of dummy bids can be generalized into a joint intention amongst all parties to rig YWCA’s tender process by way of dummy bids in order to increase the chance of success of the ‘chosen winner’, whoever it might be. The ‘chosen winner’ happened to be BT, of which Nutanix, Innovix and BT itself were well aware. But the generalised joint intention of bid-rigging stated above—which already amounts to an antitrust agreement (albeit a tacit one)—need not actually identify who the chosen winner is. It follows that the finding of a single bid-rigging agreement amongst Nutanix, BT and the cover bidders based on the generalized collusive intention was not prevented by the fact that two of the cover bidders (ie Tech-21 and SiS35) were unaware that BT was involved as the ‘chosen winner’. Nor was such a finding prevented by the fact that Nutanix, as the ‘hub’, played only a facilitative role at a different level of the supply chain than the collusive ‘spokes’. As between this single agreement approach and the HKCT’s multiple agreement approach to the analysis of ‘hub-and-spoke’ arrangements, the former approach should be preferred for its elegance. III. SECOND ISSUE: CORPORATE ATTRIBUTION As mentioned earlier, agreement (4) (ie the alleged agreement between Nutanix and SiS) was not established as the HKCT declined to attribute the behaviour of Shek (ie SiS’s employee) to SiS. It might be thought that corporate attribution is generally a straightforward matter as between an undertaking and its employees.36 Indeed, arguments against attribution are rarely, if ever, successful in the EU or UK context.37 Specifically, arguments that an employee has not been instructed to act anti-competitively, or that an employee has deviated from instructions or prohibitions, have been rejected as reasons against attribution.38 In response to SiS’s arguments against attribution, the HKCC argued (as its primary position) that ‘an undertaking is responsible for acts of its employees carried out during their employment which give rise to [the] contraventions’.39 Support for this broad, simple approach to attribution40 can be found in the recent European Court of Justice decision in VM Remonts, where it was held that ‘any anti-competitive conduct on the part of an employee is … attributable to the undertaking to which he belongs and that undertaking is, as a matter of principle, held liable for that conduct’.41 The HKCT nonetheless rejected this broad approach, premised on what appears to be a ‘purely temporal test’ requiring only the act to be ‘done during the employee’s working time’,42 and preferred, instead, to rely on previous authorities that limit attribution to the acts of employees ‘authorised to act on behalf of the undertaking’ (as in the EU) or ‘those acts [that] were carried out in the course of the employment’ (as in the UK).43 In the HKCT’s view, it is important for the rule of attribution44 to strike ‘a proper balance between effectiveness in enforcement of competition law on the one hand, and avoiding hardship and injustice on the other hand’45; and this balance would be best achieved by a middle-ground test that looks for ‘a sufficient connection between the acts of the employee in question and the undertaking’ in order for the employee to be considered as ‘acting in the course of his employment’ and, accordingly, for the undertaking to be liable for the employee’s acts.46 The HKCT noted that: Where the employer has put the employee in his position to do the kind of acts in question, it will often be fair to conclude that the employer has engaged in economic activity through the employee and is answerable for the manner in which the employee has conducted himself in that appointment.47 The HKCT eventually decided against the attribution of Shek’s misbehaviour, highlighting the following factors that indicated the lack of ‘sufficient connection’: Shek was a junior employee whose general duties did not include submission of tender or even provision of any binding quotation, and had no authority to bind SiS in relation to any commercial commitment. SiS was a distributor whose business did not include sale to end-users. The Commission also failed to show that Shek’s seniors were cognisant of his arrangements with Hung, as Shek alleged.48 The concepts of ‘sufficient connection’ and ‘acting in the course of employment’ should sound familiar to those with basic familiarity with the common law of torts.49 In fact, ‘close connection’ (as opposed to ‘sufficient connection’) is precisely the test for ascertaining whether an employee’s tort arises in the course of employment, and accordingly whether an employer is vicariously liable for that tort as a matter of Hong Kong (as well as UK) law.50 Tests such as ‘sufficient connection’ and ‘close connection’, while providing great flexibility and discretion, suffer from the inherent problem of indeterminacy.51 Connection is a matter of degree, and words such as ‘sufficient’ and ‘close’ do little to clarify the boundaries of an employer’s liability under competition law or tort law.52 In the Hong Kong Court of Final Appeal decision in Ming An Insurance Co (HK) Ltd v Ritz-Carlton Ltd, the leading Hong Kong authority on tortious vicarious liability, Bokhary PJ explained that: By ‘close connection’ is meant a connection between the employee’s unauthorised tortious act and his employment which is so close as to make it fair and just to hold his employer vicariously liable. I consider close connection to be an intellectually satisfying and practical criterion for vicarious liability. It imposes vicarious liability when, but only when, it would be fair and just to do so. And it provides a workable concept, namely a sufficiently close connection, for determining in each case whether doing so would be fair and just.53 To define ‘close connection’ in terms of what is ‘fair and just’ is, with due respect, unhelpful. As Rick Glofcheski remarked: Under [Bokhary PJ’s] approach, judges are invited to ‘openly confront’ the question of vicarious liability on the basis of what justice and fairness requires, but with little guidance as to the criteria on which the assessment of close connection and justice and fairness should be based. … Certainly, something more precise is required ….54 It might be argued that considerations such as ‘whether the employer has put the employee in his position to do the kind of acts in question’55 and whether the wrong ‘involve[s] a risk sufficiently inherent in or characteristic of the employer’s business’56 would serve as reliable, concrete indicators of sufficient or close connection. The former consideration would presumably depend on the employee’s general scope of authority,57 but this is not always clear. The latter consideration of risk may likewise be criticized as ‘too vague’, bearing in mind that ‘[a]ll enterprises involve the creation of risks’ of some sort.58 As the HKCT noted, the Nutanix situation ‘boils down to a case where Shek, a relatively junior employee whose responsibility did not include submitting tenders, had gone rogue’.59 On the one hand, the fact that Shek was not empowered to submit tenders did not rule out the risk that Shek would become involved in a tender collusion through business contacts. On the other hand, the presence of the risk of rogue employees cannot be determinative per se, as the HKCT’s decision against attribution would suggest. It would be highly desirable, if not necessary, for the test of ‘sufficient connection’ to be clarified through future decisions of the Hong Kong courts. There is a major distinction between tortious vicarious liability and competition law liability insofar as corporate attribution is concerned. Not holding an employer vicariously liable does not affect the primary liability of the employee for the torts he has committed, hence the victim may still go after the employee (although the employer is often a more worthwhile defendant with deeper pockets).60 In contrast, a decision of non-attribution, as in the case of Nutanix (as between SiS and its employee Shek), would likely lead to the denial of both the primary liability of the employer undertaking and the secondary liability of the employee (as a ‘person being involved in a contravention’61) under the HKCO, since the latter form of liability is generally parasitic on the former.62 This suggests that, all things equal, non-attribution would be more obstructive to the attainment of deterrence under competition law than tort law,63 and that a ‘stricter’ approach is warranted under competition law.64 Consistent with this observation, the test of attribution preferred by the HKCT in Nutanix—that of ‘sufficient connection’—appears to reflect a stricter standard than the test of ‘close connection’ under tort law. It remains to be seen whether the practical application of the ‘sufficient connection’ test is more likely to favour attribution than the ‘close connection’ test. IV. CONCLUSION This article examines two important aspects of the Nutanix decision in Hong Kong, both of which pertain to the ‘agreement’ element under the FCR in the HKCO. The first aspect is the analysis of a ‘hub-and-spoke’ arrangement; the second aspect is corporate attribution. I have argued, in respect of the first aspect, that the ‘hub-and-spoke’ bid-rigging arrangement in Nutanix need not be analysed in terms of multiple agreements between the ‘hub’ and the ‘spokes’. The preferred approach would be to analyse the arrangement as a single antitrust agreement based on a general, joint intention of bid-rigging shared amongst the ‘hub’ and the ‘spokes’. Significantly, it is not necessary for the ‘spokes’ to be aware of one another’s identity or existence for them to be considered as parties to the agreement, as an English contract law precedent would suggest. I have also argued, in respect of the second aspect, that, while the HKCT justifiably preferred a stricter standard of attribution for competition law liability—namely, the test of ‘sufficient connection’ between the employee’s behaviour and the employer undertaking—than tortious vicarious liability (ie the test of ‘close connection’), the ‘sufficient connection’ test suffers from the same problems of vagueness and indeterminacy as the tort law standard. The test would thus benefit from further judicial clarification and particularization. Footnotes 3 HKCO, s 6(1) (‘An undertaking must not– a. make or give effect to an agreement; b. engage in a concerted practice; or c. as a member of an association of undertakings, make or give effect to a decision of the association, if the object or effect of the agreement, concerted practice or decision is to prevent, restrict or distort competition in Hong Kong.’). 1 Competition Commission v Nutanix Hong Kong Ltd & Ors [2019] HKCT 2. For a fuller discussion of the decision, see Thomas Cheng and Kelvin Kwok, Hong Kong Competition Law: Comparative and Theoretical Perspectives (CUP) (forthcoming). 2 Kelvin Kwok, ‘The New Hong Kong Competition Law: Anomalies and Challenges’ (2014) 37 World Competition 541, 563–65; Kelvin Kwok and Thomas Cheng, ‘Procedural Fairness in Hong Kong Competition Law’ in Daniel Sokol and Andrew Guzman (eds), Antitrust Procedural Fairness (OUP 2019). 4 Nutanix (n 1) [265]. 5 ibid [2]–[3], [117], [181]–[287], [423], [554]–[558]. 6 ibid [381], [556]. 7 ibid [559]. No argument was raised on the applicability of the ‘efficiency exclusion’ under s 1 of sch 1 to the HKCO (which is modelled upon art 101(3) TFEU), but only that of the exclusion for ‘agreements of lesser significance’ under s 5 of sch 1, which was dismissed on the basis that it ‘had not been properly raised’: Nutanix (n 1) [560]. 8 Oliver Black, Agreements: A Philosophical and Legal Study (CUP 2012) 78–79; William Page, ‘Tacit Agreement Under Section 1 of the Sherman Act’ (2017) 81 Antitrust Law Journal 593, 621–24. 9 Interstate Circuit v United States, 306 US 208 (1939); Page (n 8) 621–24. The following summary is also based, in part, on: First National Bank of Arizona v Cities Service, 391 US 253, 286–87 (1968). 10 First National Bank (n 9) 287. 11 Interstate Circuit (n 9) 226–27. 12 Page (n 8) 622. 13 The point that Interstate Circuit was an agreement party (alongside the film distributors) was, however, not stressed by the Court. Strictly speaking, the agreement was tacit only as between the distributors; it was explicit as between Interstate Circuit and each of the distributors given their express communication: Page (n 8) 624. 14 Page (n 8) 624. 15 See In re High Fructose Corn Syrup Antitrust Litigation, 295 F 3d 651, 654 (7th Cir 2002); Kelvin Kwok, ‘The Concept of “Agreement” under Article 101 TFEU: A Question of EU Treaty Interpretation’ (2019) 44 European Law Review 196, 204–05 (and the sources cited therein); Black (n 8); Oliver Black, ‘Agreement: Concurrence of Wills, or Offer and Acceptance?’ (2008) 4 European Competition Journal 103; Oliver Black, ‘Two Theories of Agreement’ (2007) 13 Legal Theory 1. 16 cf Black (n 8) 77–79; Edwin Peel, The Law of Contract (14th edn, Sweet & Maxwell 2015) 53–54. 17 Interstate Circuit (n 9) 226. 18 See Nutanix (n 1) 266; Case C-194/14 P AC-Treuhand AG v Commission EU:C:2015:717; Argos Ltd and Littlewoods Ltd v Office of Fair Trading [2006] EWCA Civ 1318; Black (n 8) 288; Richard Whish and David Bailey, Competition Law (9th edn, OUP 2018) 353–55. 19 Nutanix (n 1) [2], [182], [205]. 20 Nutanix (n 1) [221]. 21 ibid. 22 ibid [2]. 23 See Nutanix (n 1) [400], [405]. 24 Nutanix (n 1) [222]. 25 ibid [145], [252]–[253], [557]. 26 ibid [233], [237], [286]. 27 See Argos (n 18) [31]; Black (n 8) 78; Jeremy Phillips, ‘Contracts between Users of a Service Supplied by a Third Party’ (1976) 92 Law Quarterly Review 499, 500. 28 The Satanita [1895] P 248, in particular 260–61; Phillips (n 27) 500; Black (n 8) 77–79; John Cartwright, Formation and Variation of Contracts (Sweet & Maxwell 2014) para 3–20 fn 102; Peel (n 16) 53–54. 29 ibid. 30 Kwok (n 15). 31 ibid. 32 Cf Nutanix (n 1) [265]–[266]. 33 See Argos (n 18) [31]. 34 Subject to the corporate attribution issue discussed below. 35 ibid. 36 See Kelvin Kwok and Ernest Lim, ‘Optimal Deterrence, the Illegality Defence, and Corporate Attribution’, European Business Organization Law Review (forthcoming); Kelvin Kwok, ‘Object and intention under Article 101 TFEU: Lessons from Australia, New Zealand and Analytical Jurisprudence’ (2019) 48 Common Law World Review 114, 139 fn 132; Frank Wijckmans and Filip Tuytschaever, Horizontal Agreements and Cartels in EU Competition Law (OUP 2015) 238. 37 See Nutanix (n 1) [323]–[340] for a survey of relevant EU and UK cases. 38 ibid. 39 Nutanix (n 1) [320]. 40 Referred to as ‘the relatively strict approach to attribution adopted in the EU and UK’: Nutanix (n 1) [346]. 41 Case C-542/14 VM Remonts v Konkurences padome EU:C:2016:578 [24]; Nutanix (n 1) [330]–[331]. 42 Nutanix (n 1) [331], [337]. 43 See n 37. 44 As a special rule of attribution, in contrast to primary and general rules of attribution: Nutanix (n 1) [315]–[319]. 45 Nutanix (n 1) [362]. 46 ibid [370], [372], [377]. 47 ibid [372]. 48 ibid [556]. For further details, see Nutanix (n 1) [375]–[377]. 49 Nutanix (n 1) [337]. 50 Ming An Insurance Co (HK) Ltd v Ritz-Carlton Ltd (2002) 5 HKCFAR 569 [13]–[25]; Rick Glofcheski, Tort Law in Hong Kong (4th edn, Sweet & Maxwell 2017) 488–501; Edwin Peel and James Goudkamp, Winfield and Jolowicz on Tort (19th edn, Sweet & Maxwell 2014) para 21–021; Nick McBride and Roderick Bagshaw, Tort Law (6th edn, Pearson 2018) 847–48. 51 Rick Glofcheski, ‘A Frolic in the Law of Tort: Expanding the Scope of Employers’ Vicarious Liability’ (2004) 12 Tort Law Review 18, 31–32. 52 ibid. 53 Ming An Insurance (n 50) [19]. 54 Glofcheski (n 51) 32. 55 Nutanix (n 1) [372]. 56 Ming An Insurance (n 50) [35] (Litton NPJ). 57 Nutanix (n 1) [331], [342], [376], [556]. 58 Glofcheski (n 51) 32. 59 Nutanix (n 1) [377]. 60 Glofcheski (n 50) 478. 61 HKCO, s 91. 62 Nutanix (n 1) [346]–[359]. 63 See Nutanix (n 1) [360]; Glofcheski (n 51) 36. 64 cf Nutanix (n 1) [346], [359]. © The Author(s) 2020. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - ‘Hub-and-spoke’ bid-rigging and corporate attribution under Hong Kong Competition Law
JO - Journal of Antitrust Enforcement
DO - 10.1093/jaenfo/jnaa005
DA - 2020-03-01
UR - https://www.deepdyve.com/lp/oxford-university-press/hub-and-spoke-bid-rigging-and-corporate-attribution-under-hong-kong-wrW6tMHP37
SP - 223
VL - 8
IS - 1
DP - DeepDyve
ER -