AustLII Home | Databases | WorldLII | Search | Feedback

University of Melbourne Law School Research Series

Melbourne Law School
You are here:  AustLII >> Databases >> University of Melbourne Law School Research Series >> 2005 >> [2005] UMelbLRS 2

Database Search | Name Search | Recent Articles | Noteup | LawCite | Author Info | Download | Help

Beaton-Wells, C --- "Customer Testimony and Other Evidence in Australian Antitrust Assessments - Searching for the Oracle" [2005] UMelbLRS 2

Last Updated: 24 September 2009

Customer testimony and other evidence in
Australian antitrust assessments – searching
for the oracle

Caron Beaton-Wells*

In United States of America v Oracle Corporation 331 F Supp 2d 1098 (ND

Cal 2004) the United States Government failed in its bid to enjoin the

acquisition by Oracle Corporation of one of its largest rivals, PeopleSoft Inc,

under the antitrust laws of that country. To a significant extent, the

government’s failure was attributable to weaknesses in its evidentiary case.

In particular, the court was not persuaded by the testimony of software

customers, called as witnesses to express a view on the scope of the

relevant market and the likely anticompetitive effects of the proposed

acquisition in that market. The approach taken by the court to the customer

testimony, as well as to expert evidence adduced in the proceeding,

provides useful insights for those dealing with Australian merger cases

under the Trade Practices Act 1974 (Cth). These insights will be relevant to

litigation in the Federal Court, as well as to decisions made by the

Australian Competition and Consumer Commission and the Australian

Competition Tribunal in relation to the clearance and authorisation of

merger proposals.

1. INTRODUCTION

On 6 June 2003 Oracle Corporation, the world’s third largest software company, initiated a takeover
bid for the shares in one of its largest competitors, PeopleSoft Inc. The United States Government
(the government), acting through the Department of Justice and joined by seven States, brought
proceedings to prevent the acquisition, alleging that it would violate s 7 of the Clayton Act.1 Section 7
prohibits acquisitions “where in any line of commerce in any section of the country, the effect of such
acquisition may be substantially to lessen competition, or to tend to create a monopoly”. The District
Court for the Northern District of California (constituted by Walker J) delivered its judgment, against
the government, on 9 September 2004.2 No appeal was lodged and on 1 June 2005, Oracle’s
acquisition of PeopleSoft became final.3

In Australia, despite the fact that the Trade Practices Act 1974 (Cth) (TPA) is entering its fourth
decade of operation, there have been only a handful of merger cases that have been prosecuted to a
fully contested trial and judgment. There have also been, relatively speaking, few merger
authorisation cases determined by the Australian Competition Tribunal (the Tribunal) on review from
a decision of the Australian Competition and Consumer Commission (the Commission). It is
instructive therefore to be aware of and learn from merger litigation in other jurisdictions in which
similar legal principles and an adversarial system of justice apply. United States of America v Oracle
Corporation 331 F Supp 2d 1098 (ND Cal 2004) (Oracle) is a case in point.

* BA/LLB (Hons), LLM, PhD (Melb), Senior Lecturer, Melbourne Law School, University of Melbourne; Victorian Bar. This
article is a slightly modified version of a paper given by the author at the Trade Practices Conference hosted by the Law
Council of Australia in Sydney on 26-28 August 2005.
1 15 USC § 18.
2 United States of America v Oracle Corporation 331 F Supp 2d 1098 (ND Cal 2004).
3 See the announcement made at http://www.oracle.com/peoplesoft/integration.html (viewed 16 August 2005).

There are significant insights to be derived from Oracle for litigants in Federal Court proceedings
under s 50 of the TPA regarding the evidence presented on this issue.4 Those insights are likely to be
applicable equally to proceedings before the Tribunal, the case load of which is set to increase should
the 2005 amendments to the Act, as proposed by the government, be passed.5 Tribunal proceedings
share many of the characteristics of Federal Court proceedings,6 despite the fact that the Tribunal is
not bound by the rules of evidence.7

Oracle should also be of substantial interest to the Commission and parties with which it deals in
connection with merger clearances8 given that, like legal proceedings, such matters involve the
collection and evaluation of market “evidence” (including information from customers). Indeed, the
vast majority of merger matters are resolved at this stage (most unopposed)9 and hence it is important
to note that the practical insights identified in this article are of equal relevance to participants in these
non-litigious processes as they are to those involved at the stage (should it ever be reached) of legal
proceedings. Notwithstanding this, the language of “litigation” and related concepts is used generally
in the article for the sake of brevity.

Adopting classifications that have been employed elsewhere,10 Oracle is concerned particularly
with the categories of industry evidence and expert evidence. The term “industry evidence” is used
here to refer to evidence that is derived from the industry relevant to the case at hand and concerns,
broadly speaking, competitive dynamics in that industry. Of particular interest for present purposes is
the approach taken in Oracle to evidence derived from customers, on the questions of market
definition and anticompetitive effects. The term “customers” is intended in this context to be confined
to industrial buyers, and is to be distinguished in that regard from consumers (or end-users). The term
“expert evidence” needs no explanation in this context.

Following a brief outline of the relevant legal principles and the parties’ contentions, the
evidence that was led in each of these categories and the court’s treatment of it are summarised
below. An analysis of the key insights and possible implications for Australian competition law and
practice follows. In particular, issues highlighted by Oracle with respect to customer evidence are
examined. In short, such issues relate to:

from which such witnesses might be called;

the future effects of the impugned transaction; and

that involve highly differentiated products.

4 Section 50 prohibits direct or indirect acquisitions of shares or assets if the acquisition “would have the effect or be likely to
have the effect of substantially lessening competition in a market”.
5 Pursuant to those amendments, to be made by the Trade Practices Legislation Amendment Bill 2005, applications for merger
authorisations will be determined by the Tribunal rather than the Commission. As at the date of writing, the Bill had been
amended by the Senate excising the schedule dealing with mergers and it remained unclear as to whether the amended Bill
would be passed by the House of Representatives.
6 As exemplified recently by the approach taken by the Tribunal to expert evidence in Re Qantas Airways Ltd [2004] ACompT
9
, discussed below.
7 See s 103 of the Trade Practices Act 1974 (Cth).
8 While the Commission’s practice of entertaining applications for informal clearances is longstanding and non-statutory in
nature, a parallel procedure involving the assessment of applications for formal clearances is proposed in the 2005 amendments
to the Act (see Trade Practices Legislation Amendment Bill 2005), in accordance with recommendations made by the Dawson
Committee of Inquiry in its Review of the Competition Provisions of the Trade Practices Act 2003. As to the fate of this Bill, see n 5, above.
9 In 2003-2004, for example, the Commission examined 189 mergers and acquisitions for clearance purposes, 183 of which
were not opposed and of the remaining 6, 2 were allowed to proceed upon the acceptance of undertakings: Australian
Competition and Consumer Commission, Annual Report, 2003-04, p 76, available at http://www.accc.gov.au/publications
(viewed 16 August 2005).
10 See Beaton-Wells C, Proof of Antitrust Markets in Australia (Federation Press, 2003).
2. OUTLINE OF RELEVANT LEGAL PRINCIPLES

It is necessary to outline the relevant legal principles governing a case such as this given that such
principles have a bearing on the nature of the evidence that is led, as well as the court’s assessment of
the evidence. In order to appreciate the insights for the Australian context it is also useful to have an
appreciation of the key similarities and differences between the principles governing merger cases in
the United States and those governing such cases here. It is to be borne in mind, however, that similar
principles apply to cases involving several of the other prohibitions in Pt IV of the TPA11 and hence
much of the discussion in the article will have application beyond the merger context.
Consistent with the approach taken in applying s 50 of the TPA, an inquiry under s 7 of the
Clayton Act is commenced by the definition of the relevant market(s), in terms of both product and
geographic area. This is seen as a “necessary predicate” to making a determination about
anticompetitive effects.12 Thus, in broad terms and applying the language of s 7, the approach taken in
cases of this kind is for the court to determine: (1) the “line of commerce” or product market in which
to assess the transaction; (2) the “section of the country” or geographic market in which to assess the
transaction; and (3) the transaction’s probable effect on competition in those product and geographic
markets.13

In defining the market, the courts generally adopt the approach favoured in the Merger
Guidelines published by the Department of Justice and Federal Trade Commission (US Merger
Guidelines).14 It is an approach that relies on demand-side responses. Starting with the smallest
possible group of competing products, the Guidelines then ask whether a hypothetical monopolist
over that group of products would profitably impose at least a small but significant and nontransitory
price increase (SSNIP), generally deemed to be about 5% lasting for the foreseeable future.15 If a
significant number of customers respond to an SSNIP by purchasing substitute products having a
considerable degree of functional interchangeability for the monopolist’s products, then the SSNIP
would not be profitable.16 Accordingly, the product market must be expanded to encompass those
substitute products that constrain the monopolist’s pricing. The product market is expanded until the
hypothetical monopolist could profitably impose an SSNIP.17 Similarly, in defining the geographical
market, the Guidelines hypothesise a monopolist’s ability profitably to impose an SSNIP, again
deemed to be about 5%, in the smallest possible geographic area of competition.18 If customers
respond by buying the product from suppliers outside the smallest area, the geographic market
boundary must be expanded.19

The hypothetical monopolist test is endorsed in the Merger Guidelines published by the
Commission (Australian Merger Guidelines),20 albeit in conjunction with the so-called QCMA or
price elevation test.21 Three key differences between the two tests have been identified.22 First, the
hypothetical monopolist test quantifies the degree of constraint necessary to include competing
products or geographic areas in the relevant market whereas the QCMA test is non-prescriptive in this

11 See ss 45, 46, 47 of the Trade Practices Act 1974 (Cth).
12 Brown Shoe Co v United States 370 US 294 at 335 (1962).
13 United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1110-1111 (ND Cal 2004).
14 Department of Justice and Federal Trade Commission, Horizontal Merger Guidelines (2 April 1992, as revised 8 April 1997)
(US Merger Guidelines).
15 US Merger Guidelines, n 14, § 1.11.
16 US Merger Guidelines, n 14, § 1.11.
17 US Merger Guidelines, n 14, § 1.11.
18 US Merger Guidelines, n 14, § 1.21.
19 US Merger Guidelines, n 14, § 1.21.
20 Australian Competition and Consumer Commission, Merger Guidelines, June 1999 (Australian Merger Guidelines), [5.44].
It is also the test used in Europe: see European Commission, Notice on the Definition of the Relevant Market for the Purposes
of Community Competition Law, Official Journal OJ C372, 9 December 1997.
21 See Australian Merger Guidelines, n 20, [5.41]-[5.42], referring to the landmark statement on the test for market definition
articulated by the then Trade Practices Tribunal in Re Queensland Co-Operative Milling Association Ltd and Defiance
Holdings Ltd (1976) 25 FLR 169 at 189; 8 ALR 481.
22 Brewster D and O’Bryan M, “Market Definition – Drawing Imaginary Lines?” (Paper presented at Law Council of Australia
Trade Practices Conference, 2004) pp19-21.
regard.23 Second, the QCMA test takes account of supply-side substitution, as well as demand-side
substitution, whereas in the United States the potential for new entry is generally only considered
once the market has been defined. Thirdly, the United States test is, as its name suggests, hypothetical
whereas the QCMA test is concerned with the behaviour of the actual firm(s) that are the subject of
the proceeding. It is this third difference that has been cited as the reason why the hypothetical
monopolist test has not been and is not likely in the future to be employed by the Federal Court,
notwithstanding its endorsement in the Australian Merger Guidelines.24

In determining anticompetitive effects, the first step in the United States traditionally has been to
calculate market shares of the firms involved in the transaction, as well as the overall concentration
levels and trends in the industry. Where the post-merger share is shown to be a certain percentage
(identified as 30% in United States v Philadelphia Nat Bank 374 US 363 at 364 (1963); 35% in the
US Merger Guidelines),25 the transaction is presumed to violate s 7. The burden then shifts to the
defendant to rebut this presumption of illegality by proving that market shares do not reflect
accurately the probable state of competition in the post-merger market. Presumptions also operate in
relation to concentration measurements. Under the US Merger Guidelines, these measurements are
made quantitatively, using what is known as the Herfindahl-Hirschman Index (HHI).26 A significant
trend toward concentration creates a presumption that the transaction violates s 7.27 In other words, in
the United States, plaintiffs establish a prima facie case of a s 7 violation by “show[ing] that the
merger would produce ‘a firm controlling an undue percentage share of the relevant market, and
[would] result in a significant increase in the concentration of firms in that market’ ”.28

Since the 1960s the market share and concentration inquiry has been recognised as merely the
starting point in a merger analysis,29 as it is in the Australian Merger Guidelines (in which similar
thresholds are applied).30 In similar fashion to the approach required by s 50(3) of the TPA, the
effects of the proposed acquisition thus are examined having regard to a range of factors, both
structural and behavioural in nature.31 In both jurisdictions the exercise is essentially predictive in
nature in that, in the language adopted in Australian context, it involves a comparison of future

23 See also the comments in Smith R and Walker J, “Australian Trade Practices and the Emerging Role of ‘Commercial
Reality’ versus Substitution in Market Definition” (1997) 5 CCLJ 1 at 6; Sweeney C and Hay D, “Quantitative Economic
Evidence in Australian and New Zealand Courtrooms” (2003) 10 CCLJ 284 at 291.
24 Brewster and O’Bryan, n 22, p 21.
25 US Merger Guidelines, n 14, §2.211.
26 The HHI is calculated by squaring the market share of each participant, and summing the resulting figures. The concentration
standards in the Guidelines concern (1) the pre-merger HHI (HHI1), (2) the post-merger HHI (HHI2) and (3) the increase in the
HHI resulting from the merger, termed delta HHI (DELTA SYMBOL HHI). See Gavil AI, Kovacic WE and Baker JB,
Antitrust Law in Perspective: Cases, Concepts and Problems in Competition Policy (Thomson West, 2002) pp 480-484. The
Guidelines specify safe harbours for mergers in already concentrated markets that do not increase concentration very much. For
example if the post-merger HHI is between 1000 and 1800 (a moderately concentrated market) and the DELTA SYMBOL HHI
is no more than 100 points, the merger is unlikely to be presumed illegal. See US Merger Guidelines, § 1.51. Likewise, if the
post-merger HHI is above 1800 (a highly concentrated market) and the DELTA SYMBOL HHI is no more than 50 points, the
merger will not be presumed illegal.
27 United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1110 (ND Cal 2004), citing United States v Baker
Hughes Inc 908 F 2d 981 at 982-983 (DC Cir 1990) (Thomas J).
28 United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1110 (ND Cal 2004), citing Federal Trade
Commission v H J Heinz [2001] USCADC 59; 246 F 3d 708 at 715 (DC Cir 2001).
29 See, eg, Hospital Corp of Am v Federal Trade Commission 807 F 2d 1381 at 1386, 1388-1389 (7th Cir 1986); United States
v Waste Management [1984] USCA2 813; 743 F 2d 976 (2d Cir 1984).
30 See Australian Merger Guidelines, n 20, [5.95].
31 The US Merger Guidelines identify five relevant factors: (1) whether the merger would significantly increase concentration
and would result in a concentrated market, properly defined; (2) whether the merger raises concerns about potential adverse
competitive effects; (3) whether timely and likely entry would deter or counteract anticompetitive effects; (4) whether the
merger would realise efficiency gains that cannot otherwise be achieved; and (5) whether either party would be likely to fail in
the absence of the merger. See US Merger Guidelines, n 14, § 0.2
competition in the relevant market(s) with (the factual) and without (the counterfactual) the
transaction in question.32

In particular, consideration is given to predicting a substantial lessening of competition as a result
of: (1) coordinated interaction between the merged firm and remaining rivals (coordinated effects);33
and/or (2) unilateral action on the part of the merged firm (unilateral effects).34 Coordinated effects
were not alleged by the government in Oracle 331 F Supp 2d 1098 (ND Cal 2004) and hence are not
referred to further here.35

Unilateral effects are said to result from “the tendency of a horizontal merger to lead to higher
prices simply by virtue of the fact that the merger will eliminate direct competition between the two
merging firms, even if all other firms in the market continue to compete independently”.36 Such
effects are thought to arise in primarily two situations,37 only the second of which was alleged in
Oracle. The first situation involves a “dominant firm and a ‘fringe’ of competitors producing a
homogeneous product”.38 In this situation, the dominant firm has a substantial cost advantage over the
fringe competitors and, therefore, can restrict output to obtain an above-marginal cost price. The
second situation, and the one that was applicable in Oracle, concerns differentiated products.39

Four factors were identified by the court as necessary to establish a differentiated products
unilateral effects claim:40

differentiated if no “perfect” substitutes exist for the products controlled by the merging firms.41

close substitutes if a substantial number of the customers of one firm would turn to the other in
response to a price increase.

32 This test has its origin in the reasons of the Full Court of the Federal Court in Outboard Marine Australia Pty Ltd v Hecar
Investments (No 6) Pty Ltd [1982] FCA 265; (1982) 66 FLR 120. See, more recently, Stirling Harbour Services Pty Ltd v Bunbury Port Authority
[2000] FCA 1381; [2000] ATPR 41-783 at 41,267 [12].
33 See US Merger Guidelines, n 14, § 2.1; Australian Merger Guidelines, n 20, [5.167]-[5.170].
34 It was explained in United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1113 (ND Cal 2004) that
“unilateral effects” is primarily a new term employed in United States jurisprudence and literature to address antitrust issues
that courts have considered in other contexts for quite some time. In Australia, it is generally understood that the merger test
was amended in 1993, lowering the standard from dominance to substantial lessening of competition, so as to catch not only
acquisitions that would enable the merged entity to exercise unilateral market power, but also those that might facilitate
coordinated market conduct. This amendment was effected from 21 January 1993 by the Trade Practices Legislation
Amendment Act 1992
(Cth), consequent upon the recommendations of the Senate Committee on Legal and Constitutional
Affairs, Mergers, Monopolies and Acquisitions – The Adequacy of Existing Legislation Controls (Canberra, 1991) (the Cooney
Committee).
35 There was a belated attempt to raise them in the government’s post-trial submissions suggesting the possibility of tacit
collusion between the post-merger Oracle and SAP. However, as no evidence had been presented in support of such a
possibility, it was dismissed out-of-hand by the judge: see United States of America v Oracle Corporation 331 F Supp 2d 1098
at 1165-1166 (ND Cal 2004).
36 United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1113 (ND Cal 2004), citing Shapiro C, “Mergers with
Differentiated Products” (1996) (Spring) 10 Antitrust 23 at 23.
37 United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1113 (ND Cal 2004), citing Starek III RB and
Stockum S, “What Makes Mergers Anticompetitive?: ‘Unilateral Effects’ Analysis Under the 1992 Merger Guidelines” (1995)
63 Antitrust Law Journal 801 at 803; US Merger Guidelines, n 14, §§ 2.21, 2.22; Areeda PE, Hovenkamp H and Solow JL,
Antitrust Law (rev ed, Aspen, 1998) at [910] (subdividing unilateral effects theories into four categories).
38 United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1113 (ND Cal 2004), citing Starek and Stockum, n 37
at 803.
39 United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1113 (ND Cal 2004), citing Starek and Stockum, n 37
at 803.
40 United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1117-1118 (ND Cal 2004). This is the court’s
summary based on a brief review of the relevant literature and case law, albeit acknowledged as substantially tracking the
analysis in Areeda, Hovenkamp and Solow, n 37 at [914f] at pp 68-69. Notably, the court’s analysis is somewhat different from
the analysis outlined in the US Merger Guidelines.
41 United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1117-1118 (ND Cal 2004).

firms so that a merger would make an SSNIP profitable for the merging firms. In short, in a
unilateral effects case, a plaintiff is attempting to prove that the merging parties could increase
prices unilaterally. Accordingly, a plaintiff must demonstrate, in effect, that the merging parties
would enjoy a post-merger monopoly or dominant position, at least in what is referred to as a
“localized competition” space.42

demonstrate that the non-merging firms are unlikely to introduce products sufficiently similar to
the products controlled by the merging firms to eliminate any significant market power created
by the merger.43

In this case there was no issue that the relevant products were highly differentiated and that the
products of Oracle and PeopleSoft were close substitutes. The issues were centred on the third and
fourth factors of a unilateral effects claim posited above. The government sought to prove that Oracle
and PeopleSoft were in a localised competition sphere within the proposed market, a sphere that
excluded any other vendors, and that the proposed merger would affect adversely competition in this
sphere. It also set out to prove that their nearest competitor, SAP, could not reposition itself to replace
the localised competition that would be lost if the merger went ahead (at 1166).

Finally, it should be noted that, unlike in Australia, defendants to a merger suit in the United
States are entitled to raise an efficiencies defence. Given that this issue does not arise (at least not
directly) in litigation brought under s 50 of the TPA and given that, in any event, it played a relatively
minor role in the outcome of Oracle and involved evidence that was separate and distinct from the
evidence adduced on market definition and anticompetitive effects (such evidence being the principal
focus of this article), no more will be said here about it.

3. OUTLINE OF PARTIES’ CONTENTIONS

The government’s case against Oracle was based on the establishment of a market that was extremely
limited in terms of the products, customers, vendors and geographic area involved.

In terms of products, the relevant market was said to be for enterprise resource planning (ERP)
system software. ERP is packaged software that integrates most of an entity’s data across all or most
of its activities. These copyrighted software programs are licensed to end-users along with a
continued right to use the licence which usually includes maintenance or upgrades of the software. To
the customer, the fees to license and maintain ERP software are generally a small part, 10-15%, of the
total cost of the installation and maintenance of an ERP system. An ERP installation, because of its
complexity, usually requires substantial and expensive personnel training, consulting and other
services to integrate the program into the customer’s pre-existing or “legacy” software (at 1101).

Many ERP programs are developed to address the needs of particular industries, referred to in
industry lingo as “verticals”. Vertical-specific ERP programs often are not well suited to the needs of
firms in other verticals. Thus, an enterprise the operations of which embrace more than one vertical
faces the task of integrating the programs. The largest and most complex organisations face particular
difficulty (at 1101-1102).

ERP programs have been developed to handle the full range of an enterprise’s activities.
Although ERP encompasses many so-called “pillars”, the government sought to confine the market in
this case to two – Human Resources Management (HRM) and Financial Management Systems

42 United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1118 (ND Cal 2004). Whether Walker J was correct
in his approach to this particular requirement has been debated in the United States: see, eg, the commentary on the case in
“Unilateral Effects Analysis After Oracle” (2005) (Spring) 19(2) Antitrust 8.
43 As was observed by the judge in the Oracle case, these four factors reflect traditional antitrust analysis in that they involve
consideration of demand-side substitutability (factors 1-3) and supply-side substitutability (factor 4): United States of America
v Oracle Corporation 331 F Supp 2d 1098 at 1118 (ND Cal 2004).
(FMS). Each ERP pillar consists of “modules” that automate particular processes or functions.44
Some of these modules might be regarded as “core” and others as more peripheral.

In terms of customers, the government focused on “large complex enterprises” (LCEs) that have
“high function software” needs. The satisfaction of these needs was said to require software with
particular performance characteristics. Software with the relevant characteristics was described as
scalable; highly configurable; seamlessly integratable; able to accommodate rapid growth,
acquisitions and reorganisations; able to reflect actual units of business; and able to adapt to industry
specific requirements (at 1125). LCEs rarely, if ever, buy core HRM or FMS modules in isolation.
Customarily, FMS and HRM software are purchased in bundles with other products.45

In terms of vendors, the government argued that only Oracle, PeopleSoft and the German
company, SAP AG, should be considered. Although by no means alone in the ERP business (many
firms develop, produce, market and maintain ERP software), these three firms have the most
comprehensive of ERP offerings (at 1104). The proposed market excluded:

including Lawson, AMS and Microsoft) (at 1105);

as Siebel that sells individual CRM pillars);

another firm, such as Accenture, Fidelity, Hewitt or Aon; some of these service providers may
purchase software from an ERP vendor such as Oracle, while others use internally created
software) (at 1106);

sometimes referred to as “legacy” software, that it used prior to the development of ERP HRM
and FMS software) (at 1125).46

In terms of geographic area, the relevant area of competition between these three firms was
alleged by the government to be confined to the United States.

It was on these contentions about the scope of the market that the government’s substantive case
effectively rose and fell, as the essence of its complaint about anticompetitive effects was that the
proposed acquisition would constrict what was an already highly concentrated oligopoly to a duopoly
involving SAP and the merged Oracle/PeopleSoft (at 1107).

Oracle contended that the government’s market definition was both legally and practicably too
narrow. Paraphrasing from the judgment (at 1107-1108), its defence was, in summary, that:
“high function” HRM and FMS software does not exist; “high function” is simply a label created
by the government for the purposes of the litigation;

developing, producing, marketing and maintaining HRM and FMS ERP software;

44 HRM and FMS software each consists of numerous modules. HRM modules include such functions as payroll, benefits, sales
incentives, time management and many others. FMS modules include such functions as general ledger, accounts receivable,
accounts payable, asset management and many others: United States of America v Oracle Corporation 331 F Supp 2d 1098 at
1102 (ND Cal 2004).
45 For example, customers may purchase a cluster of products such as Oracle’s E-Business Suite that provide the customer with
a “stack” of software and technology, which may include core HRM or FMS applications, add-on modules, “customer-facing”
business applications such as customer relationship management software, and the infrastructure components (application
servers and database) on which the applications run: United States of America v Oracle Corporation 331 F Supp 2d 1098 at
1103, 1107 (ND Cal 2004).
46 Explained in the government’s post-trial brief, available at http://www.usdoj.gov/atr/cases/oracle.htm (viewed 16 August
2005).
price competition comes from sources in addition to ERP software vendors and includes
competition from firms that provide outsourcing of data processing and from the durability and
adaptability of enterprises’ incumbent or legacy systems;

Europe;

any market power by a merged Oracle/PeopleSoft; and

merger will not have an anticompetitive effect.

4. INDUSTRY EVIDENCE

Most of the evidence presented in Oracle was evidence that may be characterised as industry
evidence in that it involved testimony and documents from people and firms participating in the ERP
software industry. The government presented evidence from 10 customer witnesses, five vendor
witnesses and two systems integration/consultant witnesses. Oracle presented witnesses in each of
these categories also.

4.1 Customer witnesses

Described by counsel as their “strongest witnesses” (at 1125), the government led evidence from 10
senior managers of large enterprises (both public and private), including DaimlerChrysler, the State of
North Dakota, Pepsi Americas, and Greyhound Lines. Their evidence was remarkably consistent,
with several recurring themes (at 1125-1130):

software;

PeopleSoft or SAP (some did not even include the latter in their consideration set);

vendors had been or was in the future to be 10% higher (and in some cases even if greater than
10%), they still would not have considered and would not in the future consider a possible
alternative.

The court found the testimony of the customer witnesses “largely unhelpful to the government’s
effort to define a narrow market of high function FMS and HRM” (at 1130). The credentials of the
witnesses was not doubted and nor was the sincerity of their beliefs in the testimony that they gave.
However, Walker J questioned the grounds upon which the witnesses offered their opinions on the
definition of the product market and competition within that market (at 1130-1131). Their opinions
amounted, in the judge’s view, to a statement of preferences rather than a statement about
substitutability, actual or potential. As the court saw it (at 1131):

The preferences of these customer witnesses for the functional features of PeopleSoft or Oracle

products was evident. But the issue is not what solutions the customers would like or prefer for their

data processing needs; the issue is what they could do in the event of an anticompetitive price increase

by a post-merger Oracle.

Furthermore, on the issue of anticompetitive effects, Walker J regarded the evidence of these
witnesses as failing to rise above speculation, “speculation [that] was not backed up by serious
analysis that they had themselves performed or evidence they presented” (at 1131). There was, as the
judge pointed out, scant testimony by these witnesses about what they would or could do or not do to
avoid a price increase from a post-merger Oracle. While each testified, “with a kind of rote” (at
1131), that they would have no choice but to accept a 10% increase by a merged Oracle/PeopleSoft,
none gave testimony about the cost of alternatives to the hypothetical price increase: for example,
how much outsourcing would actually cost, or how much it would cost to adapt other vendors’
products to the same functionality that the Oracle and PeopleSoft products offer (at 1131). Walker J
went on to say (at 1131):
If backed by credible and convincing testimony of this kind or testimony presented by economic
experts, customer testimony of the kind the government had offered was recognised by the court as
capable of putting a human perspective or face on the injury to competition that was alleged. But
unsubstantiated customer apprehensions do not substitute for hard evidence.

Oracle, too, presented customer witnesses, albeit a far lesser number (from the judgment it
appears to have been only two). One such witness spoke of his experience at Fleet Boston and Bank
of America in which both firms had turned to outsourcing to meet their HRM needs (at 1132-1133).
The other, a representative of the Emerson Electric Company, testified as to the range of options
available to his company for handling its HRM and FMS needs and gave specific instances in which
in-house software and outsourcing were used instead of purchasing from an ERP vendor (at 1132-
1133).

The testimony of the Oracle witnesses, like that of the government’s customer witnesses, was
recognised by the court as entailing some speculation about their future options. But the
distinguishing factor as far as Walker J was concerned was that the Oracle witnesses testified about
concrete and specific actions that they had taken in order to meet their firms’ information processing
needs. The judge found on this basis, as well as an assessment of the witnesses’ credibility, that the
testimony of the Oracle customer witnesses was “more believable” than that of the government’s
witnesses (at 1133).

4.2 Vendor and consultant witnesses

In addition to its customer witnesses, the government in Oracle led evidence from two executives
from PeopleSoft, one from Microsoft and one formerly of JD Edwards, an ERP software company
that had been acquired by PeopleSoft. It also adduced testimony from senior representatives of two
consulting companies, BearingPoint and IBM, that specialise in advising customers on the purchase
and implementation of ERP software. These consulting firms generally have an alliance with a
particular ERP vendor – in these instances, BearingPoint with Microsoft and IBM with PeopleSoft.
On behalf of Oracle, evidence was given by executives from ERP vendors, Lawson and SAP, a
consultant witness from Accenture and two witnesses from outsourcing firms, Fidelity and ADP.

The government’s PeopleSoft witnesses gave evidence regarding the bases upon which they
distinguished so-called high function customers from mid-market customers; evidence as to why, in
their view, firms such as Lawson were able to service only mid-market customers; and evidence as to
why alternative solutions such as outsourcing, “do-nothing” and best-of-breed solutions were not
realistic options for high function customers. They also gave evidence explaining their view that there
was localised competition between Oracle and PeopleSoft that excluded SAP, for the purposes of
establishing the government case regarding unilateral anticompetitive effects (at 1136-1142).
The Microsoft witness gave evidence intended to support the government’s characterisation of
his company as restricted to mid-market business. He also gave evidence aimed at explaining why
neither Microsoft’s previous unsuccessful attempt at acquiring SAP nor the successful alliance that it
has established with BearingPoint indicated its intentions to enter the high function market (at 1143-
1144).

The JD Edwards witness was called to establish that there were high barriers to entry to the high
function market through evidence of his company’s failed attempt to reposition itself as a player in
that market, prior to its acquisition by PeopleSoft (at 1144).
The two consultant witnesses, from BearingPoint and IBM, were called to provide confirmatory
testimony as to the ERP software needs of LCEs and the capacity of only Oracle, PeopleSoft and SAP
to service those needs (at 1134-1136).

Notwithstanding this evidence, the government’s vendor and consultant witnesses failed to
satisfy the court that there was a clear distinction to be drawn between high function and mid-market
customers for the purposes of defining the relevant market in the case. If anything, the evidence of
these witnesses only undermined the government’s proposed market definition in that they each
employed different criteria relevant to distinguishing between these two categories of customer.
Moreover, one of the PeopleSoft executives conceded (at 1138-1139) that there was no “clear-cut”
dividing line between them and the other conceded (at 1141) that the day prior to Oracle’s tender
offer, PeopleSoft’s demarcation line had been moved from $500 million and/or 2,000 employees to
$1 billion in revenue only – under the former standard the high function market clearly would have
included competitors other than Oracle, PeopleSoft and SAP. The deficiencies in this aspect of the
government’s evidence were exacerbated by the evidence given by Oracle’s vendor witnesses, the
SAP witness testifying that the term “high function” has no recognised meaning in the industry and
that characterising a customer is by no means an “exact science” (at 1152).

As to the balance of the vendor/consultant evidence adduced by the government, it is fair to say
that the court rejected almost all of it on credibility grounds. The attempt by the PeopleSoft witnesses
to persuade the court that so-called mid-market firms posed no competitive threat to the big three was
described as “self-serving” (at 1139). This adverse impression was reinforced when, on crossexamination, their evidence was contradicted by PeopleSoft’s own internal records. Contrary to the testimony that Lawson was not a serious competitor for PeopleSoft, for example, a PeopleSoft record documented Lawson as having been an enterprise competitor 27 times, SAP 33 times and Oracle 38 times over the same period (at 1138).

Such was Walker J’s cynicism about the evidence of these witnesses that he coined the term
“Lawson amnesia” to describe the weaknesses exposed in their testimony (at 1139). The judge
included in this description (at 1136), the witness from IBM whose evidence was discounted on the
additional ground of possible bias, having regard to IBM’s potential loss of PeopleSoft
implementation business should the merger proceed. The final nail in the proverbial coffin on this
issue was hammered in by the Lawson witness, called by Oracle, who gave uncontradicted evidence
(at 1150) regarding Lawson relationships with some of the largest enterprises in the United States,
customers that exceed $1 billion in revenues, employ more than 10,000 people and are listed among
the Fortune 1000.

Walker J was also not persuaded by the PeopleSoft evidence that solutions such as outsourcing
were not a feasible alternative for LCEs. In the case of one of these witnesses, his testimony to this
effect was found to have been “impeached” by a PeopleSoft document showing the company to have
competed against ADP, an outsourcing firm, 15 times over a relevant period (at 1142). By
comparison, the judge found the testimony of an Oracle consultant witness from Accenture to be
“reliable and informative”, such evidence having included a number of specific examples of high
function clients that had chosen the outsourcing route (at 1149). This account was supported by the
outsourcing witnesses themselves, whose evidence was also characterised as “reliable” and “amply
supported by specific examples of high function customers that had chosen to outsource with Fidelity
or ADP as an ERP alternative” (at 1153).

The evidence of the PeopleSoft witnesses as to localised competition between PeopleSoft and
Oracle was undermined by concessions given in cross-examination of the one witness and the
deposition statement of the other that there was no vertical segment of the market in which SAP was
not a competitor for high function customers (at 1168). This was followed by evidence of the
Accenture witness of a SAP/Accenture alliance that was aimed at developing a product to see SAP
become competitive with Oracle and PeopleSoft in the banking industry (the only vertical from which
it was then missing) (at 1149). There was also the evidence from the SAP witness, found to have been
“reliable and uncontradicted”, regarding its various high function clients and specific instances in
which SAP had competed head to head with Oracle and other ERP vendors (at 1151-1153).

The evidence of the Microsoft witness was discounted entirely by Walker J without any
assistance, so it seems, from cross-examination. The judge described (at 1144) the witness’s “Uriah
Heep like humility about Microsoft’s intentions regarding the failed SAP alliance and the successful
BearingPoint alliance unconvincing”, observing that “it strains credulity to believe Microsoft would
offer billions of dollars to acquire SAP merely to make data processing easier for customers who use
both Microsoft Office and SAP ERP” (as had been suggested by the witness). The witness’s
testimony regarding the supposed limited functionality of some of its products was also found to be
contradicted by statements made on the BearingPoint website, extolling the multi-dimensional
capabilities of the same products (at 1144). Furthermore, the BearingPoint witness himself gave
evidence that Microsoft had the capacity and the intentions to enter the market space occupied by
Oracle, PeopleSoft and SAP (at 1134). Not surprisingly in these circumstances, Walker J found the
evidence of the MicrosSoft witness to be “incredible” (at 1160) and the evidence of the BearingPoint
entry to the alleged market (at 1134).

5. EXPERT EVIDENCE

The government called three witnesses to give expert evidence:
1. a professor of business administration at Harvard Business School with particular expertise in
operations management and information technology and experience also as a consultant to
software companies (Iansiti);
2. a professor of the University of Virginia and, as described by the court (at 1143), “a well-known
and highly regarded economist” (Elzinga); and
3. a professor the provenance and area of expertise of whom was not identified in the judgment but
from whose testimony (as described by the judge) it was apparent that he was an economist
(McAfee).

Oracle called two expert witnesses:
1. an industrial organisation economist at MIT (Hausman); and
2. dean of the Haas Graduate School of Business at the University of California and former director
of the Federal Trade Commission (Campbell).47

5.1 Iansiti

Iansiti was essentially an industry expert, as distinct from an economic expert. He was described by
the court as bringing “an academic perspective that basically echoed the testimony of the customer
witnesses” (at 1133). Based on his review of the product documents and analyst’s reports of 148 ERP
vendors, the witness opined that only the products of Oracle, PeopleSoft and SAP possess the
functionality adequate to meet the needs of LCEs. He elaborated on reasons as to why neither Lawson
nor Microsoft specifically are and would not readily in the future be in a position to compete with the
products of the big three for this business.

Walker J was unmoved by this testimony on the basis primarily, so it seems, that Iansiti had not
(and did not claim to have) performed an “economic study of the ERP industry” (at 1134). It is not
entirely clear what the judge had in mind in relation to such a study, but from his subsequent
treatment of the other expert evidence, it would appear that he meant an econometric or some other
quantitative exercise. Notably, Iansiti’s evidence was not seen as bolstering the industry evidence
simply by providing an independent and consistent view of customer requirements and the capacity of
competing products to satisfy them. Such, perhaps the judge perceived, were matters that fall within
the domain of the industry witnesses and hence, an expert would have to bring a somewhat different
approach to the analysis in order to add substantive value to the testimony of the lay witnesses.

5.2 Elzinga

Elzinga was referred to by Walker J as “[b]y far the most important of [the government’s] witnesses”
(at 1145). His opinion was that the relevant product market should be limited to high function FMS
and HRM software on the basis that a hypothetical monopolist could profitably impose an SSNIP in
that market and that it should exclude mid-market vendors, best-of-breed solutions, incumbent or
legacy solutions and the services of outsourcing firms (at 1145). He reached this opinion having
analysed four “strains” of evidence:

47 Oracle also adduced evidence from two experts on the integration layer of software technology in an attempt to argue that
developments in this layer suggest that it should be included in the product market. That argument was only a minor aspect of
Oracle’s rebuttal of the government’s case and was rejected. It does not warrant discussion in any detail in this article and
hence the evidence of these two experts is not referred to further.

Having reached the conclusion that the market should thus be limited to Oracle, PeopleSoft and
SAP, Elzinga used sales data (applying a minimum threshold purchase of $500,000 per customer) to
calculate market shares and perform the HHI concentration calculations. In summary, his calculations
produced results that, if accepted, would have attracted the presumptions of illegality (at 1148).

Oracle’s experts criticised Elzinga’s testimony on product market definition, as well as the
evidence given by the government’s industry witnesses, describing it as “vague, unrealistic and
underinclusive” (at 1153). As to the first of these criticisms, Hausman and Campbell argued that the
term “high function” software is too imprecise and much was made of the fact that there were no
“quantitative metrics” that could be used to distinguish high function vendors from other vendors
(something Elzinga had conceded) (at 1154). Oracle claimed that “there must be a clear break in the
chain of substitutes in order for separate markets to be found” (at 1154) and Walker J agreed.

The proposed definition was said to be unrealistic or disconnected in that it failed to acknowledge
that ERP software was bought as a bundle; rarely are FMS or HRM pillars bought on their own.
Hence, any discounts offered are blended discounts and, if not acceptable to the customer, then (as
Hausman argued), the customer can always threaten to turn to a best-of-breed solution as an
alternative for a particular pillar (at 1154-1155).

Finally, Oracle’s experts argued and Walker J agreed that there were viable substitutes to high
function ERP that had to be included in the market, pointing to evidence that demonstrated numbers
of LCEs opting for mid-market, outsourcing and other solutions in preference to the offerings by
Oracle et al (amongst them, ironically, the Department of Justice that, two weeks after bringing the
case, had chosen a mid-market firm, AMS, from which to buy its FMS software for $24m) (at 1155-
1156).

Given that the industry evidence had failed to establish the government’s case on market
definition, “the full weight of the plaintiffs’ product market burden fell at trial on Elzinga” (at 1158).
Elzinga, the court found, was unable to discharge that burden. Instead, “in resolving the battle of the
expert witnesses”, the court was of the view that “Oracle’s witnesses presented the better and more
convincing case”. Elzinga, the judge considered, “for all of his indubitable credentials as an
economist seemed mostly to apply the techniques of his avocational interest in mystery writing” (at
1158). The evidence that he had marshalled was criticised by the judge as “circumstantial and highly
qualitative”, his concentration statistics as flawed having been based on sales that did not separate
FMS and HRM pillars from other pillars included in the sale, and his other statistical tabulations as
“sketchy” (for example, the customer surveys were based on a sample of only 28 sales opportunities)
(at 1158-1159).

Elzinga also carried the evidentiary burden with respect to the government’s proposed
geographic market (confined to the United States). In relation to this aspect of the case, he attempted
to argue that it would be inappropriate to employ the Elzinga-Hogarty (E-H) test, a test frequently
applied in defining geographic markets in United States antitrust cases and, as the name suggests, a
test of which he himself is the co-author.48 If that test was employed, he conceded, it would support

48 See Elzinga K and Hogarty T, “The Problem of Geographic Market Delineation in Antimerger Suits” (1973) 18 (Spring)
Antitrust Bulletin 45. In general terms, the E-H test “measures the accuracy of a market delineation by determining the amount
of either imports into or exports from a tentative market. The test is based on the assumption that if an area has significant
exports or imports, then that area is not a relevant geographic market. Under the [test], exports or imports greater than 10%
Oracle’s argument that the market should be seen as a global one (at 1163). Elzinga also argued that
the geographic market should be defined without regard to the site of manufacture of the software (a
reference to the manufacture of SAP software in Germany) but rather, focusing on the vendorcustomer
relationship which entails installation, implementation, maintenance and upgrade and hence
is necessarily local in character (at 1161-1162). Elzinga also pointed to the absence of arbitrage in this
market and to the fact that prices in the United States are not affected by prices in Europe and vice
versa (at 1162).

Again, on the question of geographic market, Walker J preferred the analysis of Oracle’s experts
over Elzinga’s analysis. The explanation proffered by the latter for failing to apply the E-H test in this
case was considered “unpersuasive” (at 1164); it has been used in other cases to define the relevant
geographic market for the purposes of merger analysis and is seen as especially important when
vendor-customer relationships are involved. There were many other markets involving such
relationships that could not on any tenable basis be limited to the United States (eg computer sales).
Furthermore, the “relationship” factor could not be seen as all-important given the evidence
indicating that “non-relationship” solutions were also favoured by customers (eg outsourcing) (at
1164-1165). Finally, there was, as noted by Hausman, empirical evidence showing that prices in
Europe constrain prices in the United States and vice versa (at 1163-1164).

Having rejected the government’s proposed market in the case, the court had to conclude (at
1165) that the market share and concentration statistics presented by Elzinga were of no assistance.49
Thus, without the benefit of the presumptions, the court turned next to the evidence of anticompetitive
effects. The government rested its theory of anticompetitive effects on an attempt to prove that Oracle
and PeopleSoft are in a “localized” competition sphere (a “node”) within the high function FMS and
HRM market, a sphere excluding SAP and one into which SAP could not be repositioned to replace
the competition that would be lost by an Oracle/PeopleSoft merger (at 1166). In attempting to prove
localised competition between Oracle and PeopleSoft, the government relied on virtually the same
evidence that was used to prove the product market (at 1166). There was, however, additional expert
testimony on this aspect of the case, derived from McAfee.

5.3 McAfee

This expert supported the government’s unilateral anticompetitive effects claim based on three
analyses. First, like Elzinga, he analysed Oracle discount approval forms with a view to showing the
vigorous nature of the competition between Oracle and PeopleSoft (at 1168). Second, he ran
regression analyses (using the variables of competitor, next revenue and discount percentage based on
sales representative surveys), aimed at demonstrating that, when competing against PeopleSoft,
Oracle offers greater discounts than in any other competitive situation (at 1168-1169). Third, he
conducted a merger simulation analysis designed to show how competitive the market was premerger
and compared it with the likely scenario post-merger.50 Based on this analysis, he concluded
that the merger would lead to a unilateral price increase – from 5% to 11% in the FMS market and
from 17% to 30% in the HRM market (at 1169-1170).

Oracle’s experts attacked the government’s case on unilateral effects on several fronts. Campbell
argued that the unilateral effects theory was not applicable in this type of case because it is predicated
on a market in which buyers have little or no power whereas ERP software buyers are plainly

suggest that the market examined is not a relevant market”: United States of America v Oracle Corporation 331 F Supp 2d
1098 at 1161 (ND Cal 2004), citing United States v Country Lake Foods, Inc 754 F Supp 669 at 672 n 2 (D Minn 1990).
49 The words used were in fact “wholly inapplicable”.
50 The simulation works by putting in necessary variables and assumptions, such as market shares and percentage of wins in
head-to-head competition. Once these variables were accounted for, McAfee still had to set a variable for “how competitive the
market [was] pre-merger”. One way of creating such a measurement is by estimating the “total value of the product that
accrues to the buyer” (ie, how “much of the value of the software to the buyer actually accrues to the buyer and how much
accrues to the vendors in the form of price”). McAfee ran the simulation based upon five different “buyer accrual” estimates:
0.5 (only 50% accrual) to 0.9 (90% accrual). McAfee used the market shares calculated by Elzinga as his market shares
variable. Once all the data are compiled and the variables accounted for, the merger is simulated by merging the shares of the
two merging firms. Once this is done, data can be calculated showing how much the price of the relevant product is expected to
increase. See United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1170 (ND Cal 2004)
knowledgeable and sophisticated (at 1171). McAfee’s analysis of the discount approval forms showed
no more than that Oracle and PeopleSoft are frequent and close rivals; it says nothing about
competition between these two firms and SAP (at 1171). McAfee’s regression analysis was flawed
from the outset as it was based only on data involving broader suites of enterprise application
software, rather than on data confined to FMS and HRM software (at 1171). His merger simulation
analysis was “simplistic” and “spurious” in that it employed an inappropriate simulation model and it
was based on Elzinga’s market share statistics which had already been showed to be artificial – being
based on the government’s “gerrymandered” market definition (at 1172).

Finally, and said to be most importantly, Oracle complained that the government had failed to
provide “econometric calculations in trying to prove localisation”. Proving localisation, it was argued,
requires “extensive econometric analysis”, such as diversion ratios, price-cost margins and the like
and the government had offered none (at 1172). Walker J accepted most of the criticisms made of the
government’s evidence by Oracle’s experts. He appears to have been influenced, in particular, by the
argument concerning the absence of econometric evidence, pointing to several other cases in which
courts have based their rulings on unilateral effects, at least in part, of such evidence (at 1172).

6. INSIGHTS AND IMPLICATIONS

In analysing the insights to be derived from and possible implications of Oracle for Australian
antitrust assessments, it is useful to consider the various categories of evidence dealt with in the case
separately.

6.1 Industry evidence generally

Industry evidence, of which customer evidence is a part, has been the most crucial category of
evidence on issues such as market definition and competitive dynamics in Pt IV proceedings to date.51
It has been valued highly and consistently by courts in ensuring that the findings made on such issues
reflect “commercial realities”52 and, for this reason, has often been preferred over expert evidence.53
As in Oracle, industry evidence in Australian litigation has taken the form of testimony from
members of the industry (including customers) as well as business records that emerge from the
discovery process or are subpoenaed by the parties. Business records, in particular, have been valued
for their contemporaneity and have been seen as a rich and reliable source of evidence about what
firms actually think and what they actually do in the normal course of their business.54

Similarly, industry submissions play a significant role in the deliberations of the Commission in
dealing with clearance or authorisation proposals. As is stated in the Australian Merger Guidelines,
the Commission has a practice of seeking the views of participants in the relevant industry, including
competitors, suppliers and customers, amongst others, with respect to such proposals.55 In the case of
customer views specifically, the Guidelines indicate that they will be taken into account in defining

51 In News Ltd v Australian Rugby Football League Ltd [1996] FCA 1256; [1996] ATPR 41-466 at 41,679, for example, Burchett J pointed out
that: “[i]t is accepted that an industry’s own perception of the breadth of the market in which a product is involved, as
evidenced by the conduct of those engaged in the industry, may be persuasive to show the true extent of the market.” More
recently, in Boral Besser Masonry Pty Ltd v Australian Competition and Consumer Commission [2003] HCA 10; (2003) 215 CLR 374; 195
ALR 609 at [257], McHugh J commented that “the views and practices of those within the industry are often most instructive
on the question of achieving a realistic definition of the market”.
52 This expression has been coined in several cases. See, for example, the comments made in Australia Meat Holdings Pty Ltd v
Trade Practices Commission [1989] FCA 25; [1989] ATPR 40-932 at 50,105 (Pincus J); Singapore Airlines Ltd v Taprobane Tours WA Pty Ltd
[1992] ATPR 41-159 at 40,170 (French J); News Ltd v Australian Rugby Football League Ltd [1996] FCA 1256; [1996] ATPR 41-466 at 41,668;
Australian Competition and Consumer Commission v Universal Music Australia Ltd [2002] ATPR 41-855 at 44,676 [351]. For
a discussion of whether an approach based on “commercial reality” is any different from an approach based on the principles
enunciated in QCMA (1976) 25 FLR 169, see Smith and Walker, n 23.
53 See, for example, the approach taken at first instance in Trade Practices Commission v Australia Meat Holdings Pty Ltd
[1988] FCA 244; [1988] ATPR 40-876 and endorsed on appeal: Australia Meat Holdings Pty Ltd v Trade Practices Commission [1989] ATPR
40-932.
54 The value of this evidence was noted early on by Wilcox J in Mark Lyons Pty Ltd v Bursill Sportsgear Pty Ltd [1987] ATPR
40-809 at 48,798.
55 Australian Merger Guidelines, n 20, [4.11], [6.18].
relevant markets.56 The Commission also generally requires the firms involved in the proposed
transaction to produce internal documents and is prepared to invoke its powers under s 155 of the
TPA if they are not provided voluntarily.57

Given the priority assigned to it by both the courts and the Commission, the selection and use of
industry evidence (to the extent that such matters lie within a party’s control) require substantial care.
As was seen in Oracle, this evidence is as capable of losing a case for a party as it is capable of
winning it. In my view, the key insight to be derived from the Oracle experience in this regard is that
the greatest prospects for success lie in proposing a market that is widely recognised by the industry
itself and is reflected in industry attitudes and practices that may be established through the evidence.
Any other market almost certainly will be found to be divorced from “commercial realities”, a finding
that in the Australian context at least, may well be fatal. In the process, the testimony of industry
witnesses is likely to be characterised, at best, as unsubstantiated or inconsistent with other evidence
(such as that found in business records); at worst, as self-serving or simply lacking in credibility.

The exposure of such weaknesses in the industry evidence may have implications that resonate
beyond the issue of market definition given that the same evidence is often treated as relevant also to
the substantive competition issues. In unilateral effects cases, the relevance of the industry evidence
in this respect flows from the fact that, applying the United States’ approach, the assessment of such
effects has become an exercise involving the identification of what is essentially a narrow version of a
market (a submarket, perhaps).

Notably, commentators in the United States have questioned whether Oracle has elevated
“industry recognition” from the status of a relevant factor58 to that of a necessary condition for
judicial acceptance of a proposed market definition.59 This may be over-reaching but it does
underscore the importance of this category of evidence in both jurisdictions. Whether or not
Australian courts are correct in the emphasis that they give to industry evidence is a matter that has
attracted substantial debate. It is not intended to rehearse the arguments and counter-arguments on
that score here, suffice it to say that, in general terms, commentators have sought to emphasise the
distinction between the concept of a market as employed in the business context, and the concept as it
applies in the statutory antitrust context.60

The “commercial realities” approach is defensible, in my view, to the extent that it reflects the
fact that effective constraints are a function of dynamics in the business world and it is such
constraints, after all, that lie at the heart of competition law analysis. However, as I have argued in
detail elsewhere, it should not be used as a substitute for rigorous testing of the industry evidence.61
For the reasons discussed below, this is of particular relevance for present purposes to evidence
derived from customers.

6.2 Customer evidence

Oracle highlights the particular difficulties that may be involved in establishing a market and
assessing future market dynamics based primarily on customer evidence. There are four such
difficulties that are exposed by the case.

(a) Selection of customer evidence – how many customers

The judge in Oracle was concerned that the views expressed by the government’s customer witnesses
were not sufficiently representative of customer views generally in the ERP software industry. This

56 Australian Merger Guidelines, n 20, [5.59], [5.62].
57 Australian Merger Guidelines, n 20, [4.10].
58 “Industry recognition” traditionally has been seen as just one of seven factors relevant to establishment of a submarket
(which, in the United States, is equivalent to a market for liability purposes), identified in Brown Shoe Co v United States 370
US 296 (1962).
59 “The Oracle/PeopleSoft Decision: The Implications for Merger Analysis in High-Tech Industries” (Paper presented as part of
Antitrust TeleSeminar Series organised by the Antitrust Law Section of the American Bar Association, 4 November 2004
(slides on file with author)).
60 For a summary of these references, see Beaton-Wells, n 10, pp 118-119.
61 Beaton-Wells, n 10, pp 330-331, and the discussion in Ch 4 generally.
concern was reflected in his repeated observation that, while the government’s customer witnesses
clearly had particular views about the feasibility of vendors or solutions other than Oracle and
PeopleSoft, there were plainly other customers (including so-called LCEs) that, as evidenced by the
Oracle witnesses, held very different views (at 1131).

As this experience demonstrates, parties selecting customer witnesses face the perennial
challenge of persuading the court that their evidence is representative of the broader class of
customers to which they belong. Lack of representativeness is a criticism that is readily made and can
be substantiated fairly easily given that, in most instances, for every witness with one view there is a
second with another.

Undertaking a survey may avoid this issue but inevitably will raise issues of a different kind.
While hurdles to the reception of survey evidence based on the hearsay rule were dismantled in
Arnotts Ltd v Trade Practices Commission [1990] FCA 473; (1990) 24 FCR 313 it remains the case that evidence in
this form is vulnerable to attack on methodological grounds and, as a general rule, is given little
weight.62

One way of overcoming the representativeness criticism may be to adduce evidence from a
consulting firm that has had experience in advising numerous customers on their purchases. Of
course, it is preferable that such a firm not be allied with any particular vendor given that the
neutrality of their evidence is likely to be undermined by any such alliance (as it was in the case of
the consultant witnesses called by the government in Oracle).

Another option may be to adduce evidence from an industry expert, perhaps an academic who
has studied the industry and may even have some consulting experience. Such a strategy may mean
that credibility concerns are avoided. However, as illustrated by the judicial response to Iansiti’s
evidence in Oracle, there is always the risk with such a witness that his or her perspective is seen as
simply too theoretical and as adding little value to the evidence derived from the lay witnesses.

The representativeness dilemma illustrated by the government’s experience in Oracle, and the
most effective way of overcoming it, were addressed recently by a former Deputy Director of the
Federal Trade Commission’s Bureau of Competition, George Cary, where he said:

back in the old days we used to wonder whether you could bring a case without customer testimony,

and now we wonder whether you could bring a case without econometric testimony, no matter how

many customer witnesses you have. The key is that you need systematic proof that a particular

customer’s testimony can be generalized to the market as a whole. Obviously, you cannot bring in

every customer, and you cannot, at least not often very credibly, do a poll of customers. You need

some linkage of evidence, whether documentary evidence plus econometrics or documentary evidence

plus customers plus econometrics; you need some combination that allows you to say that what these

customers are testifying to does not apply uniquely to them; it applies across the board, and the proof

of that is: (fill in the blank). The correct standard has never been “how many customers can you get to

say that a merger is good or bad”; it has always been, “does the customer testimony exemplify the

reality of the marketplace.” But, as a result of the Oracle opinion, there will be more attention paid to

affirmatively showing how it is that the conclusions from a group of customers can be systematized

and made consistent with other evidence.63

In Australia the issue of representativeness has arisen predominantly in the context of evidence
from consumers and then more so in proceedings under s 52 of the TPA (alleging misleading and
deceptive conduct and related intellectual property infringements) than under Pt IV. It does not appear
to have been an issue in relation to the evidence adduced from industry (including customer)
witnesses in competition cases. There have been several such cases in which a selected handful of
customers have been presented as witnesses (just as in Oracle) and yet it is not apparent that the
representativeness of their testimony was a matter for concern (at least insofar as it failed to rate a

62 See the discussion of the weaknesses of survey evidence in Beaton-Wells, n 10, pp189-210.
63 “Unilateral Effects Analysis After Oracle” (2005) 19(2) (Spring) Antitrust 8 at 16.
mention in the judgment).64 The reasons for this are unclear but it should not be taken to mean that
parties can afford simply to ignore the issue in future proceedings.

(b) Selection of customer evidence – which customers

In Oracle the testimony of the government’s customer witnesses about their insensitivity to a future
price increase may have been seen as pre-determined, in effect, by the past decisions of their
particular firms. By those decisions the firms in question had locked themselves into long-term
relationships with either Oracle or PeopleSoft. As shown by the evidence in the case, such
relationships involve a significant investment of time and money in their establishment and
maintenance and, by their nature, are very difficult to undo. Albeit not framed explicitly in these
terms, concern about the implications of this for the probative value of the testimony may have
underpinned the comment by Walker J (at 1167) that:

the court cannot take the self-interested testimony of five companies which chose to eliminate SAP

from consideration, and from that sample draw the general conclusion that SAP does not present a

competitive alternative to Oracle and PeopleSoft. Drawing generalized conclusions about an extremely

heterogeneous customer market based upon testimony from a small sample is not only unreliable, it is

nearly impossible.

In economic terms the criticism that could be made about the government’s witnesses is that they
represented so-called “inframarginal” customers, that is, customers who, for whatever reason, would
not be responsive to an SSNIP. Identifying a group of customers who would be unwilling to switch in
response to a price increase does not mean necessarily that they or the products to which they are
loyal constitute a relevant antitrust product market. Rather, economists argue, it is marginal customers
(those who would be responsive to an SSNIP) whose views should be considered in defining a market
for this purpose.

Firms consider the likely reaction of marginal customers in determining whether or not to
increase prices; hence, it is these customers that act as a constraint on pricing. If marginal customers
represent a sufficiently large percentage of a firm’s sales, the firm will lose more through the
imposition of an SSNIP than it will gain from sales under the new price from its inframarginal
customers. In differentiated product markets, it is usually the case that only a relatively small
percentage of a firm’s sales need to be made to marginal customers in order for a price increase to be
unprofitable (and thus for the firm to be constrained in imposing such an increase).65

Applying this approach, it is arguably evidence of the attitudes and behaviour of marginal rather
than inframarginal customers that will assist a court most in making determinations about market
definition and anticompetitive effects. This may have been what Walker J was referring to, amongst
other things, when he described the testimony of the government’s customer witnesses as “largely
unhelpful” (at 1130).

(c) Scope of customer evidence

Generally speaking, there can be little doubt that the most effective evidence from customer witnesses
is evidence concerning the past and present behaviour of their firms and others in the marketplace (for
example, evidence as to their purchasing decisions and the criteria and processes applied in making
them). As was remarked by the judge in Oracle, reflecting specifically on the customer evidence (at
1167): “the most persuasive testimony from customers is not what they say in court but what they do

64 See, eg, Eastern Express Pty Ltd v General Newspapers Pty Ltd [1991] ATPR 41-128 at 52,891 in which Wilcox J accepted
the “unchallenged evidence of numerous [real estate] agents as to the importance of locally advertising properties” in preferring
a market definition that excluded national newspapers and, more recently, Australian Competition and Consumer Commission
v Rural Press Ltd [2001] ATPR 41-804 at 42,737 [107] in which Mansfield J relied on the evidence of advertisers to support
confining the relevant market to regional newspapers, and excluding other forms of media such as radio.
65 See Hausman JA and Leonard GK, “Economic Analysis in Differentiated Products Mergers Using Real World Data” (1997)
5 George Mason Law Review 321 at 323-324.
relation to industry evidence generally.66

However, there may be occasions on which customer witnesses may be invited to opine on a
likely state of affairs in the future. Oracle raises questions about the extent to which such opinions
might properly or usefully be given by such witnesses.

Undoubtedly it is legitimate for a customer witness to express a view as to the likely conduct or
decisions of his or her firm in the future, based on its experiences, needs, resources, information about
options and the like. Such a view would have to be expressed on the assumption that such things will
remain unchanged and, for this reason, its utility necessarily may be limited.67 Furthermore, there may
be some debate as to whether the proper question in this context is what the firm “could” do (say, in
response to an SSNIP) or whether the question that ought be asked is what the firm likely “would” do.
In Oracle the judge appeared to favour the former (at 1131).

In my view, the probative value of evidence given in response to the “could” question will be
limited as it is a question arguably aimed at determining the range of decisions that are theoretically
open to a customer rather than the range of decisions that a customer realistically might make. It is
artificial and quite unhelpful, in my opinion, to inquire of such witnesses whether they “could”
possibly act in a certain way if it is their evidence that, for particular reasons, they “would” not so act.
As was pointed out in Australia Meat Holdings v Trade Practices Commission [1989] FCA 25; [1989] ATPR 40-932
at 50,092 (Davies J), one of the few Australian cases in which customer evidence has played a major
role in securing a particular market definition, “[t]he existence of a market, a concept of economics
and commerce, ought not to be determined by reference to theoretical possibilities”.68

Of course, this distinction depends on what is meant by “could” in this context. If it is taken to
refer only to decisions that make economic sense to a rational customer, then it is probably
indistinguishable in practice from “would”. Notably, the US Merger Guidelines make the point that
the analysis carried out under those guidelines is focused on whether “consumers or producers ‘likely
would’ take certain actions, that is whether, the action is in the actor’s economic interest”.69

This issue aside, it is difficult to quibble with the insistence by Walker J that industry witnesses
give specific evidence to support their assertions. In this instance, as suggested by the judge, such
evidence might have included an explanation of the detailed cost/benefit analyses that would be
undertaken by a customer in selecting an ERP vendor (at 1131). As one commentator has remarked,
there is some difficulty in understanding exactly the point that the judge was attempting to make in
Oracle, but it is possible that he:

could have been thinking that the customers did not know what they would do in a world different

from the world they were living in. The customers knew exactly what to do in the current world, and

they were doing it, but they had not investigated all of the possible options, because it wasn’t sensible

to investigate all of the possible options. But if the world changes, the customers may have to get more

information and reevaluate their decisions, and they may come to different conclusions. I think what

Judge Walker may be saying is that the customers had not gone out and gotten that information yet, so

they could not tell us what they would do after this merger. You don’t have to read the testimony that

66 While it is true that there have been occasions on which Australian courts have valued the opinions of industry witnesses,
including on the issues that fall ultimately for judicial determination in the case at hand, such evidence generally has not been
relied on in isolation but rather has been noted as confirmatory of other categories of evidence. See the discussion in Beaton-
Wells, n 10, pp 168-172.
67 Indeed, for this reason it seems, evidence from customers as to their likely reaction to a future hypothetical price increase has
been described by the leading United States antitrust authors, Areeda, Hovenkamp and Solow, as the “least reliable” form of
evidence; the “most reliable” evidence on such matters being identified as historical shifts in prices and trading patterns:
Areeda et al, n 37 at [538b], p 230.
68 In this case, the evidence of some 32 cattlemen (the relevant customers) that transporting their cattle to distant abattoirs was
neither feasible nor likely was highly influential in Wilcox J’s decision to confine the market to a region of Queensland: see
Trade Practices Commission v Australia Meat Holdings [1989] ATPR 40-876.
69 US Merger Guidelines, n 14, § 0.1.

way, but if you do, ... then there isn’t much of a lesson for the future except that complicated facts

make it difficult to win a merger case.70

In fairness to Walker J, his criticisms might also have been prompted as much by the style of the
customer testimony as by its substance (or lack thereof). The government’s customer witnesses
apparently each recited in fairly formulaic fashion (as if “by rote” (at 1131)), their view as to the
likely effect of a post-merger price increase by Oracle/PeopleSoft. In light of this, as well as the
supposed paucity of specific evidence given to support such views, it was hardly surprising that the
court found these witnesses less than “believable” (at 1133). The reaction of an Australian judge no
doubt would be similar.71

Of more fundamental concern in relation to the scope of customer evidence is the extent to which
customer witnesses may be asked to opine on the likely effects on competition of the transaction in
question, that is evidence purporting to predict market-wide effects as distinct from evidence confined
to the effects on the decision-making of the particular customer. There is a good argument that
customers should not be asked to speculate on the probable state of competition in the market in the
future with or without the relevant transaction. This is simply because customer opinions on such
matters necessarily will be subjective and based on incomplete information. Instead, predictions about
anticompetitive effects are matters on which additional expert evidence should be led. Customer
testimony about past purchases and decision-making will be one of the inputs to such evidence. Other
inputs might include information about barriers to entry, levels of capacity, any vertical integration
and so on.

This limitation on the proper use of customer evidence was argued forcibly by another United
States judge in explaining his reasons for rejecting such evidence in a merger decision handed down a
month prior to the Oracle decision – Federal Trade Commission v Arch-Coal Inc 329 F Supp 2d 109
(DDC 2004). In Arch-Coal the government presented testimony from numerous customers, which in
that case were the utilities that converted coal into electricity, in which the view was expressed that
the challenged merger would lead to higher prices. In a subsequent extra-curial speech, Bates J
explained that, in his view, customer testimony is useful on issues about which such witnesses have
actual knowledge, such as evidence that explains certain market behaviour and casts light therefore on
market definition or substitutability. It is far less useful, if not incompetent, on issues that involve the
court in a forward-looking inquiry, in predicting the likely effects on competition of a proposed
merger, for example.72 In comments that reflected on the reasoning in Oracle, as well as in Arch-
Coal, he went on:

customers do not have any particular expertise that allows them to speculate on the likely effect of a

merger ...

Absent such expertise or meaningful in-house empirical studies or experiences with similar mergers in

the past, any speculation by customers on the likely anti-competitive effects of a proposed merger is

likely to be no more than that – speculation. Witnesses can do no more than repeat broad economic

principles that they think will result in certain effects in the market. I described this in my opinion [in

Arch-Coal], and it appears that Judge Walker had the same reaction in Oracle....

... Ultimately customer testimony was unhelpful in either case because those witnesses, unlike experts

who have studied the market and have a basis for offering opinions, are not competent to provide what

amounts to largely ungrounded opinions as to likely future collusion among their suppliers. Their

testimony is simply not rationally based in their experience or expertise.

70 Werden G, Comments made in an American Bar Association Section of Antitrust Law Brown Bag Program on “Coordinated
Effects Analysis: the Arch-Coal Decision”, 27 October 2004, available at www.antitrustsource.com, March 2005 (viewed
16 August 2005).
71 As was observed recently by Gyles J (himself an experienced advocate, including in trade practices litigation, prior to
judicial appointment) in Australian Rugby Union Ltd v Hospitality Group Pty Ltd [2000] FCA 823; [2000] ATPR 41-768 (ARU), the industry
witnesses in that case may not have been “guilty of any known untruths”; however, he could not be expected “[s]imply to
accept at face value each statement made by a witness in a carefully prepared affidavit”: at 41,047 [44].
72 Justice Bates, “Customer Testimony of Anti-Competitive Effects in Merger Litigation” [2005] Columbia Business Law
Review 279.

Part of the problem lies in the essential nature of an action to block a proposed merger. Unlike in most

antitrust cases, the question of anti-competitive effects in the merger context involves predictions about

the future, and the game of predicting the future poses a unique set of dilemmas for counsel and the

fact-finder alike. The predictive power of economics far outreaches the similar power of any individual

in the industry, no matter how intelligent or experienced that individual may be. Working at Oracle

does not make one an Oracle, and unless the executive moonlights as a professor or can travel through

time, she will not be able to testify directly as to the future events that are at the center of the court’s

inquiry. Of course, the executive will be able to provide essential information on the nature of the

industry, the likelihood of entry, the history of pricing activity, and the predatory activities of firms

with market power in a pre-merger world. All of these variables are certainly critical inputs in the

economic analysis and will therefore shine light indirectly on the likely results of the merger. However,

emotional appeals or predictions of dire consequences by customers will rarely carry the day in the

antitrust context.73

As is evident from these comments (as well as those made by other commentators),74 there is a
persuasive argument that the best evidence regarding anticompetitive effects is economic evidence,
presented through an expert witness, preferably based on some form of quantitative analysis. This
category of evidence and its treatment in Oracle, as well as the experience with its use in Australian
proceedings, are examined below.

(d) Conclusiveness of customer evidence

Difficulties in relying on customer evidence will be especially acute in an industry in which the
relevant products are highly differentiated and products are marketed across a broad range of prices
and quality.75 Differences between products in such industries will be very much a matter of
subjective customer perception and will be assessable on multiple dimensions. Such difficulties will
be accentuated further where the industry is undergoing rapid development, as a consequence of
which customer needs and perceptions can be expected to be fluid.

In such circumstances it will be near impossible to demonstrate a clear break in the chain of
substitutes, distinguishing one set of products or customers from another and enabling bright-line
market boundaries to be drawn. Inevitably as a result, any attempt to define a market based on
customer attitudes and behaviour will be susceptible to the criticism that it is so vague as to be
manipulable for the purposes of the substantive competition analysis. This argument was made by
Oracle’s experts and was accepted by the judge in relation to the customer-based market proposed by
the government in that case.76

Putting concerns about gerrymandering to one side, it is clear that a proposed market should not
be rejected on the grounds simply that its boundaries are blurred. As has been acknowledged
repeatedly in the Australian context, a certain amount of boundary blurring is to be expected with
respect to any market.77 It is, however, as with almost every aspect of competition analysis, a matter
of degree. At some point a line has to be drawn. At least, this would appear to be the case in Australia
in which market definition is necessitated by the terms of the relevant prohibitions in Pt IV of the Act.
In the United States it is argued by some (most actively in the recent past, by representatives of the

73 Bates, n 72 at 286.
74 See, for example, Harkrider J, “Proving Anti-Competitive Impact: Moving Past Merger Guidelines Presumptions” [2005]
Columbia Business Law Review 317.
75 See the observations made about this in United States of America v Oracle Corporation 331 F Supp 2d 1098 at 1120-1121
(ND Cal 2004).
76 Hausman, for example, argued that in formulating this market, the government had “clearly worked backwards from their
desired result”: identifying a group of customers all of which had bought Oracle, PeopleSoft or SAP products, and then
claiming that those customers were “similarly situated” as a means of defining the market. Hausman also noted that, at trial, the
government changed its position and attempted to argue that the high function market was based on the performance attributes
of the software (such as its functionality and scalability), and not the customers who buy it. See United States of America v
Oracle Corporation 331 F Supp 2d 1098 at 1154 (ND Cal 2004).
77 In its first major decision in relation to Pt IV, the High Court pointed out that markets cannot be defined with precision; that
market definition necessarily involves questions of degree and the making of value judgments See Queensland Wire Industries
Pty Ltd v Broken Hill Proprietary Co Ltd [1989] HCA 6; (1989) 167 CLR 177 at 196 (Deane J).
regulatory authorities) that market delineation, as traditionally applied in the antitrust context, in fact
is not required in unilateral effects cases in differentiated product industries.78

This argument aside, it is trite to observe that if the boundaries are drawn so as to widen the
market unduly, it will frustrate the competition policy objective of preventing or restraining excess
market power. Equally, if drawn too narrowly, it will impede firms in the legitimate conduct of their
business. From a practical perspective, given the purposive nature of the market definition task, what
matters is that a view about the relevant boundaries is settled upon with sufficient confidence by the
decision-maker and those affected by the decision to be made. In many instances, for the reasons
previously stated, it may be difficult to attain the requisite degree of confidence with respect to
markets based solely on customer evidence.

Campbell, one of the Oracle experts, described the government’s customer-based product market
as “unprecedented” and “unusual” (at 1154). In Australia, attempts to establish a market in Pt IV
proceedings based entirely or even substantially on customer evidence are not unprecedented but they
could probably be said to be unusual.79 As previously indicated, by comparison, material provided by
customers regularly plays an important role in Commission decisions concerning clearances and
authorisations. In the United States, the Federal Trade Commission too has a history of heavy reliance
on customer opinion.80 While obviously the same evidentiary standards applicable to legal
proceedings do not apply to Commission intelligence gathering, there is no reason why regulators
should not approach customer “evidence” with the same degree of caution as it was approached by
the court in Oracle and for the same reasons.81

6.2 Expert evidence

The evidence given by the economists in Oracle is illustrative of the expansive role that, despite some
initial judicial resistance, Australian economists also now routinely play as expert witnesses in Pt IV
proceedings. That role involves the identification, selection and compilation of evidence, from which
inferences and conclusions are then drawn by the expert, applying economic theory and modes of
analysis, relevant to issues that include the so-called “ultimate” issues in the proceeding. Recognition
of such a role for economists has meant understanding that their value lies in contributing a whole
process of reasoning rather than just a set of abstract theoretical principles.82 However, it also raises
78 See discussion of this argument in Werden G and Rozanski G, “The Application of Section 7 to Differentiated Product
Industries: The Market Delineation Dilemma” (1994) 8 Antitrust 40; Stoll N and Goldfein S, “Markets! We Don’t Need
Markets!” (2003) 29 New York Law Journal 3; Keyte J and Stoll N, “ ‘Markets, We Don’t Need No Stinking Markets!’ The
FTC and market definition” (2004) 49 (Fall) Antitrust Bulletin 593.

79 Australia Meat Holdings [1989] FCA 25; [1989] ATPR 40-932 (referred to at n 68) stands out as a success story in this regard, the customer
evidence in that case having plainly been more robust than that presented by the government in the Oracle proceeding. There
also have been attempts to establish markets based primarily on consumer needs or perceptions but they have generally been
unsuccessful. In Aut 6 Pty Ltd v Wellington Place Pty Ltd [1993] ATPR 41-202, for example, an attempt to establish a luxury
car market failed at first instance, as did the attempt to restrict the market in Taprobane Tours WA Pty Ltd v Singapore Airlines
Ltd [1990] FCA 325; [1990] ATPR 41-054 to holiday packages to the Maldives (while the argument succeeded at first instance, on appeal the
airline was successful in having the trial judge’s finding that the market should be so restricted overturned: Singapore Airlines
Ltd v Taprobane Tours WA Pty Ltd [1992] ATPR 41-159). Both attempts were based on the argument that consumers
associated particular, mostly intangible, benefits with the products in question such that, in the minds of consumers at least,
they were not closely substitutable with what, on any objective view, would be obvious alternatives.
80 Harkrider, n 74 at 338, noting that between 1996 and 2003, the FTC took enforcement action in 98% of cases where there
were strong customer complaints. See also Whitener M, “Editor’s Note: The Regulators” (2005) 19(2) (Spring) Antitrust 5 at 6,
noting that in the United States: “traditional view is that customer testimony should be given significant weight in merger cases
... much more weight than the views of competitors – because customers’ interests are generally more likely to be aligned with
the goals of the merger laws.”
81 Indeed, it was indicated recently by the Assistant Director for Antitrust in the Federal Trade Commission’s Bureau of
Economics, Michael Vita, that, as a result of Oracle, the Commission in future would “press customers hard to identify the
basis for their opinions on, say, their likely reaction to an attempted post-merger price increase. Given the way the customer
testimony was dismissed in the Oracle case, if that decision does become a model for other judges, we will have to make sure
that the testimony we elicit from complaining customers can stand up to that heightened standard.” See “Unilateral Effects
Analysis After Oracle”, n 63 at 16.
82 See the discussion in Beaton-Wells, n 10, pp 303-314.
qualitative debate) as well as the nature of such opinions (the advocacy versus evidence debate).

(a) Quantitative versus qualitative evidence

What is perhaps most striking for Australian observers in relation to the treatment of the expert
evidence in Oracle was the proposition, submitted by the Oracle experts and accepted by the judge,
that such evidence should be supported by some quantitative analysis in order to be effective in
establishing a position on the issues of market definition, market shares and unilateral anticompetitive
effects. As indicated above, much emphasis was placed by Campbell on the absence of any
“quantitative metrics” that could be used to distinguish high function vendors from other vendors (as
well as on the fact that Elzinga had conceded as much) (at 1157). Walker J evidently was disturbed by
this omission, critical of Elzinga’s testimony that there was “something different” about high function
software (at 1157):

the court cannot delineate product boundaries in multi-million dollar merger suits based upon the mere

notion that there is “something different” about the merging products and all others.83

The employment of empirical analysis to support expert opinions has not been a common feature
of Pt IV proceedings. There have been very few Australian cases in which evidence of this nature has
been adduced and in those cases in which it has been presented it has been ineffective.84 Reservations
by Australian judges towards this category of evidence in Pt IV cases were illustrated most recently in
Australian Gas Light Co v Australian Competition and Consumer Commission (No 3) [2003] ATPR
41-966, the first substantive merger to come before an Australian court in over a decade.85 The
quantitative evidence presented by the Commission was considered in some depth by French J but
ultimately failed to persuade him that the proposed acquisition met the test in s 50. For some, the
treatment of the evidence by French J has only reinforced the impression that there may be limited
utility in a party investing the time and expense in compiling quantitative evidence in Pt IV
proceedings.86

83 The judge rejected Elzinga’s claim that it was not possible to present reliable quantitative data because the proposed high
function market was “shot through with price discrimination”. There was said to be price discrimination given that, based on
the discount approval forms, it appeared that Oracle offered different discounts to different customers. Of this argument,
Walker J retorted: “Elzinga’s assertion that this market is ‘shot through’ with price discrimination because ‘somehow’ Oracle
was able to determine what level of discount it could offer to different customers uncannily resembles his argument that there is
‘something different’ about Oracle, PeopleSoft and SAP. Again, the court refuses to sustain plaintiffs’ inarticulable
contentions” (at 1173). In defence of Elzinga it should be noted that the evidence did suggest that LCEs and high function ERP
vendors enter into relationships as distinct from transactions and that the relationships are complex, long-term, to a large extent
services-oriented and hence involve many intangible elements. It also suggested that each of these relationships was unique,
involving a bundle of products and services tailored to meet the particular needs of the customer and priced accordingly. In
such circumstances, Elzinga’s failure to undertake an econometric study (involving cross price elasticities, for example) might
have been viewed more sympathetically. It should also be pointed out that neither of the Oracle experts conducted such a study
(albeit, admittedly, they did not bear the same burden of proof borne by Elzinga) and nor did they appear to suggest that data of
the kind necessary to carry out an econometric study was in fact available.
84 In Trade Practices Commission v Australia Meat Holdings Pty Ltd [1988] FCA 244; [1988] ATPR 40-876 the court declined to place any
weight on evidence of price correlations and gave specific reasons for doing so. In News Ltd v Australian Rugby Football
League Ltd [1996] FCA 1256; [1996] ATPR 41-466 the court either overlooked or chose to ignore the quantitative evidence (aimed at showing
limited substitutability for rugby league). In Australian Rugby Union Ltd v Hospitality Group Pty Ltd [2000] FCA 823; [2000] ATPR 41-768
evidence of cross-elasticities was found not to be statistically significant and in Australian Competition and Consumer
Commission v Universal Music [2002] ATPR 41-768, qualitative evidence was preferred to the quantitative evidence presented
on the issue of market power.
85 The Commission presented detailed quantitative evidence to support a decision that it had made to refuse informal clearance
of AGL’s proposed acquisition of a 35% stake in the Loy Yang Power Station and Coal Mine. The analysis was made possible
by the availability of data concerning electricity trading in the national electricity market.
86 In a recent gloomy comment on the case, a solicitor who acted for the Commission has said: “In stark contrast to judgments
such as News Ltd, ... French J engaged with the quantitative evidence in a manner never before seen in an Australian
competition case. For example, he closely examined the residual demand curves presented to the court by an ACCC expert, and
carefully considered the different methodologies put forward for the derivation of residual demand curves and the calculation
of negative inverse elasticities of demand. One would be hard-pressed to find another Australian judgment concerning Part IV
of the TPA which demonstrates a more concerted effort to engage with, understand and critique complex economic evidence.
There are good reasons, related primarily to problems with data availability and methodology,
that explain why quantitative evidence has not been tendered to a greater extent in Pt IV proceedings
than has been the case to date. What is of particular note in the present context, however, is that
Australian courts generally have not been critical of the lack of such evidence presented to them on
Pt IV issues and when it has been presented have demonstrated a distinct preference for qualitative,
principally industry-sourced, evidence – evidence of just the kind regarded as vague, unreliable, and
highly circumstantial in Oracle.

In the United States an insistence on quantitative evidence may be explained, to some extent, by
reference to the relevant legal tests employed in antitrust cases. The hypothetical monopolist test, for
example, requires that it be shown that an SSNIP of approximately 5% be profitable before market
boundaries are closed. Market concentration involves the use of the Herfindahl-Hirschmann index.
The unilateral effects theory also invites quantitative analysis so as to establish the likelihood of a
unilateral price increase by the merged entity. Indeed, in his discussion of the relevant principles
governing the Oracle proceeding, Walker J cautioned against the use of “qualitative factors” in the
task of identifying a localised competition space for the purposes of a unilateral effects claim (at
1118-1119). Such factors, he pointed out, have been associated traditionally with the now largely
discredited submarkets doctrine and have been recognised as misleading courts into “identify[ing]
artificially narrow groupings of sales on the basis of non-economic criteria having little to do with the
ability to raise price above cost” (at 1119).87 By contrast, the judge observed, “modern econometric
methods hold promise in analyzing differentiated products unilateral effects cases” (at 1122). In
particular, Walker J referred to merger simulation models that “may allow more precise estimations
of likely competitive effects and eliminate the need to, or lessen the impact of, the arbitrariness
inherent in defining the relevant market” (at 1122).88

In relation to market definition, in Australia, the degree of constraint required before a competing
firm or product or geographic area is included in the proposed market has not been quantified in the
same way as under the hypothetical monopolist test. The QCMA test is far less prescriptive,
employing broad concepts such as “close competition”, “strong substitution”, a “sufficient price
incentive” and “relatively high” cross-elasticities.89 Similarly, broad descriptive terms appear in s 50
(for example, “substantial” and “likely”) and in s 50(3)(e), in particular, referring to “the likelihood
that the acquisition would result in the acquirer being able to significantly and substantially increase
prices or profit margins”. Furthermore, such increases are only one factor that the court must take into
account in determining the likely effects of an acquisition. The others, by comparison, are largely
structural in nature (for example, level of import competition; height of barriers to entry; degree of
countervailing power; and the nature and extent of vertical integration).90

In light of this direction provided by the statute and by the Tribunal, and taking account of the
other difficulties associated with the presentation of quantitative evidence referred to above, it can be
expected, in my view, that qualitative industry evidence will continue to predominate in Pt IV
proceedings for the foreseeable future.

Given French J’s conclusions, however, one wonders whether this has in fact been more detrimental to the cause of quantitative
analysis than if he had completely ignored it. For example, after providing a 30 page exposition of the economic evidence put
before the court, he turned to the evidence of an industry expert, which he clearly preferred, stating: this evidence ‘had a ring of
commercial reality about it.’” See Merrett A, “Quantitative Analysis Again Up in Lights” (2005) 13 TPLJ 90 at 96. See also the
observations made by Allsop J about the limitations of economic modelling, specifically the simplicity of the assumptions on
which they are invariably based and hence their inaccuracy in representing reality, in Australian Competition and Consumer
Commission v Baxter Healthcare Pty Ltd [2005] ATPR 42-066; [2005] FCA 581 at [512]- [513].
87 Referring to Areeda, Hovenkamp and Solow, n 37, [914a] at p 60. These commentators have been concerned about the
potential for localised competition analysis to devolve into an unstructured submarket-type analysis.
88 See generally in relation to these models, Werden G and Froeb L, “The Effects of Mergers in Differentiated Products
Industries: Logit Demand and Merger Policy” (1994) 10 J L Econ & Org 407; Werden G, “Simulating the Effects of
Differentiated Products Mergers: A Practical Alternative to Structural Mergers Policy” (1997) 5 George Mason Law Review
363; Baker J, “Contemporary Merger Analysis” (1997) 5 George Mason Law Review 347.
89 Re Queensland Co-Operative Milling Association Ltd (1976] ATPR 40-012 at 17,247.
90 See s 50(3)(a), (b), (d), (i) of the Trade Practices Act 1974 (Cth).
This should not be taken to suggest that quantitative evidence has been a major feature or that it
has been embraced unconditionally by the judiciary in the United States (as compared, for example,
with the attitudes of the regulators by whom quantitative studies have been employed regularly, not
only in the United States but increasingly also in other jurisdictions such as New Zealand, Canada and
Australia).91 Issues with the data availability (merger simulation analyses, for example, generally
require detailed high frequency data in order to estimate demand) and with methodology (including
the nature of the underlying assumptions made in such analyses) plague litigants in the United States,
just as they do in Australia.92 As a consequence, litigants in neither jurisdiction are in a position to
adduce such evidence with confidence that the investment will prove to be worthwhile.

In the United States, in Federal Trade Commission v Staples 970 F Supp 1066 (DDC 1997), for
example – the first major merger case in which the regulatory authorities presented extensive
quantitative analysis in a contested proceeding – the evidence had a most equivocal reception.93
While the court granted the injunction sought by the FTC, it did so without explicit reliance on the
quantitative analysis. This has not stopped the Commission Director from claiming that the court’s
conclusions were based “demonstrably” on the econometric evidence.94 Others have been less
generous, arguing that the studies presented by opposing experts in the case were so vigorously
contested that they effectively cancelled each other out, forcing the court to revert to the documentary
industry evidence (an arguably inevitable outcome whenever well-prepared, reasoned, supported and
presented expert evidence is given on each side).95 The result was said to provide “further
confirmation that econometric analyses will be more persuasive when key modelling choices are
consistent not only with economic theory but also with the documentary and other evidence available
about the market and tested against plausible alternatives”.96

Thus, so it seems, even in the United States the prospects for greater use of econometric analysis
outside the regulatory framework are not as assured as some might imagine.97 The judge in Oracle
may have been critical of the supposed lack of econometric analysis presented by the government in
support of its case. However, it should not be overlooked that the government did present evidence,
through McAfee, by way of a merger simulation analysis. The effectiveness of the criticisms levelled
by the Oracle experts at this evidence and the short shrift that it was given by the judge only
reinforces, in my view, the need for caution in judging the extent to which such evidence might
profitably be employed in legal proceedings in the future.

91 See the developments in this regard canvassed in Evans L, “Economic Measurement and the Authorisation Process: The
Expanding Place of Quantitative Analysis” (1999) 7 CCLJ 99 and also Merrett, n 86.
92 For a discussion of these issues in the United States context and an argument that empirical analysis cannot and should not
replace structural analysis, see Note, “Analyzing Differentiated-Product Mergers: The Relevance of Structural Analysis”
(1998) 111 Harvard Law Review 2420.
93 The Federal Trade Commission had performed a systematic empirical study of Staples’ pricing that was presented in court by
an econometric expert. The Commission also relied on an econometric study of the rate that Staples historically passed through
cost savings to consumers by way of lower prices. The merging firms presented alternative statistical analyses of pricing, as
well as econometric studies of the determinants of Staples’ price-cost margins and the effect on revenues at Staples’ stores of
nearby store openings by possible rivals. For a more detailed explanation of the evidence, see Baker JB, “Econometric Analysis
in FTC v Staples” (Paper presented before the American Bar Association’s Antitrust Section Economics Committee, 18 July
1997, revised 31 March 1998 (at http://www.ftc.gov/speeches/other/stspch.htm, viewed 16 August 2005).
94 See Baker, n 93.
95 As was observed by Shapiro in “Unilateral Effects Analysis After Oracle”, n 63 at 12-13: “By downplaying the role of
customer testimony and arguably emphasizing the role of this type of simulation method or expert testimony, Judge Walker has
encouraged some sort of battle of the experts, and everybody who does simulation models knows there is a lot to argue about:
you have to make a bunch of assumptions in those models. I do not know what judges are going to do when, inevitably they see
one expert get up with this complex model with all these assumptions and another expert gets up and testifies that those are
crazy assumptions and this whole thing is not reliable. How is a judge supposed to sort that out?”
96 See Evans, n 91 at 111. Also, Shapiro, “Unilateral Effects Analysis After Oracle”, n 63 at 12 for a similar point made about
the use of merger simulation analyses by the Federal Trade Commission.
97 For a review of recent United States cases indicating that the use of econometrics in lieu of traditional structural analysis
lacks judicial support, see Hill R, “Practicing What They Preach: One Lawyer’s View of Econometric Models in Differentiated
Products Mergers” (1997) 5 George Mason Law Review 393.
(b) Evidence versus advocacy

In Oracle it appears fairly plain that the experts were engaged in an exercise of advocacy on behalf of
the parties that they represented – Elzinga’s attempt to distance himself from his own geographic
market test on the grounds that it did not suit the government’s case was a prime example of this.
Their evidence did not attract any criticism from the judge on this basis, even if some of his more
cryptic remarks in relation to Elzinga’s testimony suggested a degree of scepticism concerning the
impartiality of this witness.

The question as to the extent to which an expert can act as an advocate for the party that he or she
represents, and related issues of expert impartiality, have been the subject of considerable debate in
Australia. This debate has been accentuated in connection with economists giving evidence in
competition cases by the recognition that their evidence may take the form of submission or argument
(as distinct from evidence)98 and by the practice of “hot-tubbing”99 that in itself could be said at least
to empower if not encourage the economists involved to act as advocates. In light of this, it is notable
that there have been hardly any Pt IV cases in which the impartiality of the economic experts has
attracted adverse judicial comment (let alone adverse findings).100

In 2003 the Federal Court revised its 1998 Practice Direction on expert witnesses, incorporating
suggestions as to ways in which the criticism of partiality might be avoided.101 The revised version
retained an emphasis on the general duties of an expert witness, being, first and foremost, to assist the
court and not to act as an advocate for a party.

Interestingly, this statement of duties was drawn on recently by the Tribunal in support of
criticisms made of experts who gave evidence in Re Qantas Airways Ltd [2004] ACompT 9 at
[217].102 The Tribunal drew a distinction between an expert advocating an opinion, which is
permissible, and an expert advocating for a party, which is impermissible. Additionally, the Tribunal
commented (at [216]) that:

The role of expert witnesses appearing before the Tribunal is to instruct on areas of specialist

knowledge in a manner that is ultimately designed to inform rather than to advocate a particular view.

Obviously, parties will call upon experts whose opinions support their view of the case. However, it is

not appropriate for an expert witness to act as an advocate for the instructing party at all costs, and

professional witnesses should be willing to concede points which, whilst not advancing the case of the

party engaging them, they believe to be open as a fair and reasonable assessment on the material before

them. The Tribunal will be assisted by expert witnesses who can clearly explain the relevant issues and

concepts and can pinpoint the differences between opinions in the profession and the reasons for such

differences so that an informed decision can be made as to which opinion should be accepted on the

available evidence. The Tribunal will not be assisted by experts who uncritically push a party line,

avoid challenging questions, and seek to obscure the real issues in contention.

98 See Federal Court Rules, O 10, r 1(2)(j); Robertson D, “Expert Economic Evidence: Challenging the Paradigm” (2003) 11
TPLJ 70
99 This practice is explained in CCH, Australian Trade Practices Reporter, vol 1 at [16-015.91]. See also Federal Court Rules,
O 34A, r 3(2)(e)-(i).
100 There is only one Pt IV case of which I am aware in which the judge considered it necessary to make some comment on the
subject of expert impartiality. In Australian Competition and Consumer Commission v Universal Music Australia Ltd [2002]
ATPR 41-855 at 44,675 [345], Hill J observed of the two economists who gave evidence in the case that “[e]ach can fairly be
said to have been an advocate for the party for whom he was called”. Having made this observation, it did not appear thereafter
to have any significant influence on the use that Hill J made of the expert opinion evidence, except insofar as it might be seen
as connected with his criticisms of the extent to which each expert expressed views on the ultimate issues in the case. Hill J
declined to accede to the submission that he should disregard the evidence of the Commission’s expert (Ergas) on the grounds
that he was too biased in favour of the Commission, noting in that regard that the same might well have been said of the
respondent’s expert (Hausman), given that he had been engaged as a consultant for the past three years to the major record
companies in the United States.
101 Federal Court Practice Direction, Guidelines for Expert Witnesses in Proceedings in the Federal Court of Australia
(4 September 2003).
102 The Guidelines, it was said, should apply to Tribunal proceedings notwithstanding the Tribunal is not bound by the rules of
evidence (see s 103 of the Act).
These comments may reflect a degree of conservatism on the part of Australian adjudicators
towards the conduct of expert witnesses that may not be shared by their United States counterparts. It
might also suggest that expert witnesses can expect harsher criticism from their professional
colleagues (one of the members of the Tribunal panel in the Qantas case having been Professor
Round, an economist who himself has experience in acting as an expert witness in Pt IV cases) than
from members of the judiciary sitting alone. Such suggestions are speculative, at best.
However, what is clear is that for Australian expert witnesses at least, there are limits on the type
of advocacy that will be persuasive and perceived independence will be as important as actual
independence in ensuring that such witnesses perform their function effectively.

6.3 Relationship between industry evidence and expert evidence

Oracle raises interesting questions regarding the relationship between industry evidence and expert
evidence. On the one hand, the treatment of Iansiti’s and Elzinga’s evidence suggests that where the
court has taken a dim view of the industry evidence on an issue such as market definition, it may be
difficult for an expert to persuade the court to adopt a different, more favourable, view of that
evidence. This is particularly so where the expert acts solely or primarily as an analyst or interpreter
of the same evidence derived from the industry that the court itself has been in a position to analyse
and interpret. In such instances, the expert may even be superfluous.

On the other hand, the analytical/interpretive role of an expert may be employed more effectively
in a negative fashion, as was the case with the Oracle experts, to criticise and cast doubt on the
industry evidence. Greater value may be gained from an expert used in this capacity as it is not
uncommon for a court simply to adopt the expert’s criticisms of the industry evidence in much the
same way as legal submissions are adopted. It is notable that in Oracle those acting for the defendant
made an obvious and deliberate decision not to attempt to present a positive case establishing an
alternative market to that proposed by the government. Rather, Oracle employed a defensive strategy,
described by the judge as “pick[ing] apart [the] plaintiff’s market definition piece by piece” (at 1165).
It was a strategy that worked. As Walker J pointed out, the burden of proof lay on the government and
the government failed to discharge it (at 1165).

While the Oracle strategy and the judge’s reaction to it are neither surprising nor controversial,
they do not reflect fully the complexity involved in the forensic choice that must be made by a
respondent between an approach of positively disproving the applicant’s case and an approach of
simply criticising it with a view to establishing that the applicant has failed to discharge its burden of
proof. This choice is of particular relevance in relation to the expert evidence. It is now accepted that
experts may provide an opinion on the ultimate issues in a proceeding.103 Indeed, in Pt IV
proceedings, in light of the special status of the expert evidence by economists, it might even be said
that such witnesses will be expected to do so. Such is the strength of expectation in this regard that an
expert witness may be criticised for failing to offer a positive opinion on such matters.

In Australian Rugby Union Ltd v Hospitality Group Pty Ltd [2000] FCA 823; [2000] ATPR 41-768, for example,
the economist, Officer, was engaged by ARU to give evidence critical of the expert evidence adduced
by the other side. As was observed by Gyles J, Officer “did not give evidence as to his opinion on the
relevant market or markets and made [it] clear that he would not be in a position to express such an
opinion without further work”. Of this omission and of his rebuttal of the evidence by the opposing
experts, Gyles J was unreserved in his criticism (at 41,061 [79]):

As a tribunal of fact I did not find this evidence of any great value. Indeed much of it was of doubtful

admissibility, as it was argumentative as to mixed questions of fact and law. Furthermore, the general

economic principles which underlie Pt IV and its overseas progenitors are well understood, were no

doubt taken into account by the legislature and do not need to be restated. By and large, his evidence

added little to the cross-examination of the applicant’s experts by counsel for the respondent which

was, no doubt, based in part upon assistance from Professor Officer. Even that cross-examination

would have been more effective if deployed to support a positive view as to market based upon

103 See s 80 of the Evidence Act 1995 (Cth) which abrogated the common law rule that evidence on the ultimate issue in a

proceeding is inadmissible.

admissible expert evidence. It would be correctly seen as misconceived in a personal injury case for the

defendant to call a specialist physician, who had not seen the patient and had expressed no opinion as

to the plaintiff’s medical condition, who proceeded to take pot shots at a well-qualified specialist called

by the plaintiff who had examined the plaintiff and had expressed an opinion as to the plaintiff’s

condition.

As this reaction by Gyles J suggests, ultimately, the decision whether to adopt an offensive or a
defensive strategy in formulating an approach to the evidence (that of the experts, particularly) should
rest, at least in part, on an assessment of the audience in the particular case.

6.4 Last words (if there can ever be) about market definition

It would seem remiss to conclude without noting what is perhaps the most obvious insight extractable
from Oracle, concerning the significance of the issue of market definition. In Australia, the question
of the relevant market has been described as “the most vigorously (and expensively) litigated and
discussed question in our courts and tribunals”.104 The reasons for this are more than borne out by the
experience in the Oracle litigation. As that experience demonstrates, while the principles might be
settled, the task of establishing a relevant market as a matter of evidence remains complex and often
controversial.

Market definition is also an exercise that, albeit intended (at least theoretically) to be only a
precursor to the “real” analysis, can prove to be determinative of the outcome of the proceeding.105 It
is therefore an exercise that parties must get right. In effect, as shown by the demise of the
government’s case in Oracle, this means getting the industry evidence right. It may well be that
greater employment of unilateral effects analysis and merger simulation studies,106 has the potential to
diminish the significance of market definition.107 However, as things presently stand (at least in
Australia, if not also the United States), the prospects of such studies being employed regularly and
effectively in contested court proceedings should not be overestimated.

104 Baxt R, “The Australian Concept of Market – How it Came to Be” in Richardson M and Williams P (eds), The Law and the
Market (1995) p 28.
105 See the comment by the United States Supreme Court in Eastman Kodak Co v Image Technical Services [1992] USSC 73; 504 US 451 at 469
(1992): “market definition generally determines the result of the case.”
106 Economic analysis regarding unilateral effects is more amenable to quantification than is economic analysis of coordinated
effects (collusion): see Shapiro, n 36 at 23.
107 The likelihood of this occurring is discussed in Beaton-Wells C, “Mergers without Markets? Unilateral Effects Analysis in
the United States and its Prospects in Australia” (2006) (forthcoming).


AustLII: Copyright Policy | Disclaimers | Privacy Policy | Feedback
URL: http://www.austlii.edu.au/au/journals/UMelbLRS/2005/2.html