CT-2024-010 THE COMPETITION TRIBUNAL IN THE MATTER OF the Competition Act, R.S.C. 1985, c. C-34; AND IN THE MATTER OF certain conduct of Google Canada Corporation and Google LLC relating to the supply of online advertising technology services in Canada;
AND IN THE MATTER OF an application by the Commissioner of Competition for one or more orders pursuant to section 79 of the Competition Act.
B E T W E E N:
COMMISSIONER OF COMPETITION Applicant/Responding Party and
GOOGLE CANADA CORPORATION AND GOOGLE LLC Respondents/Moving Parties
REPLY RECORD
July 23, 2025
-2-
DAVIES WARD PHILLIPS & VINEBERG LLP 155 Wellington Street West Toronto ON M5V 3J7
Kent E. Thomson (LSO# 24264J) Tel: 416.863.5566 Email: kentthomson@dwpv.com Elisa K. Kearney (LSO# 49342T) Tel: 416.367.7450 Email: ekearney@dwpv.com Chantelle T.M. Cseh (LSO# 60620Q) Tel: 416.367.7552 Email: ccseh@dwpv.com Chanakya A. Sethi (LSO# 63492T) Tel: 416.863.5516 Email: csethi@dwpv.com Chenyang Li (LSO# 73249C) Tel: 416.367.7623 Email: cli@dwpv.com
Lawyers for the Respondents/Moving Parties, Google Canada Corporation and Google LLC
TO:
ATTORNEY GENERAL OF CANADA Department of Justice Canada Competition Bureau Legal Services (CBLS) Place du Portage, Phase 1 50 Victoria Street, 22nd Floor Gatineau QC K1A 0C9
Alexander M. Gay Donald Houston John Syme Ian Clarke Katherine Rydel Adam Rossiter
Tel: 613.296.4470 Fax: 819.953.9267
Lawyers for the Applicant, Commissioner of Competition
CT-2024-010 THE COMPETITION TRIBUNAL IN THE MATTER OF the Competition Act, R.S.C. 1985, c. C-34; AND IN THE MATTER OF certain conduct of Google Canada Corporation and Google LLC relating to the supply of online advertising technology services in Canada;
AND IN THE MATTER OF an application by the Commissioner of Competition for one or more orders pursuant to section 79 of the Competition Act.
B E T W E E N:
COMMISSIONER OF COMPETITION Applicant/Responding Party and
GOOGLE CANADA CORPORATION AND GOOGLE LLC Respondents/Moving Parties
INDEX
Tab 1. A.
2. A.
Description Affidavit of Keith N. Hylton dated July 23, 2025 Reply Expert Report of Professor Keith N. Hylton dated July 23, 2025
Affidavit of Rupert Younger dated July 22, 2025 Supplemental Expert Report of Rupert Younger dated July 22, 2025
Page No. 1 – 2 3 – 37
38 – 39 40 – 43
CT-2024-010 THE COMPETITION TRIBUNAL IN THE MATTER OF the Competition Act, R.S.C. 1985, c. C-34; AND IN THE MATTER OF certain conduct of Google Canada Corporation and Google LLC relating to the supply of online advertising technology services in Canada;
AND IN THE MATTER OF an application by the Commissioner of Competition for one or more orders pursuant to section 79 of the Competition Act.
BETWEEN:
COMMISSIONER OF COMPETITION
and
Applicant
GOOGLE CANADA CORPORATION AND GOOGLE LLC
Respondents
AFFIDAVIT OF KEITH N. HYLTON SWORN JULY 23, 2025
I, Keith N. Hylton, of the City of Boston, in the State of Massachusetts, in the United States of America, MAKE OATH AND SAY AS FOLLOWS:
1.
| am a tenured University Chair (the William Fairfield Warren Distinguished
Professor) at Boston University and a Professor of Law at Boston University School of
Law. In 1986,
| obtained my Ph.D. in Economics from MIT, followed by my J.D. from
Harvard Law School in 1989.
2.
| have published more than 150 articles in numerous law and economics journals,
many on antitrust matters. In particular,
| have conducted research and published on the
topics of optimal penalties, including in the antitrust economics context.
2 -2- 3. | have been engaged by Davies Ward Phillips & Vineberg LLP, on behalf of Google Canada Corporation and Google LLC, to act as an independent expert in this proceeding.
| prepared an initial Expert Report dated May 6, 2025, which that same date.
| attached to my Affidavit of
4. | have now prepared a Reply Expert Report dated July 23, 2025, which attached as Exhibit “1” to this Affidavit.
| have
SWORN by Keith Hylton of the City of Boston, in the State of Massachusetts, in the United States of America, before me at the City of Toronto, in the Province of Ontario, on July 23, 2025 in accordance with O. Reg. 431/20, Administering Oath or Declaration Remotely.
DocuSigned by: (cs Commissioner for Taking Affidavits (or as may be)
Signed by: but Hylton KEITHN . HYLTON
CHENYANG LI (LSO# 73249C)
O L V H L
'
D N N O I L V Y V 1 9 O A C ‘ O Z / L E v “ D A Y J O N I A O Y d A H 4 0 1 3 9 ‘ V O I M A N S L L A S N H O V S S V W N Y O M S N O L 1 A H L I S I H X 3 S I S I
H
) n o 0 q ‘ A T I L O I N I N G V I N V L N O 4 0 L V S W 3 H L N I ‘ A I N F N O H L O L « b s
S
N
A
V
N
D
S
1
S
Y
0
g L S ‘ O S H L Z £ c A
0
i
8
b
7
u
-
)
0
e
A N I Z A V
p
v
T
0
q
Y
:
S
S
T
A
r
O
V A L I O 3 L I N N N I S Z O A V G I S
I
8
H
8
N
O
Z
E
0 G L I
O
Z
4
O
7
S A H L 4 0
O
6
O
Y
(
I
L
G
O
S
V
O
V
V
M
L
1
N
N
L
S
V
2
O
O
H
I
L
S
N
S
L
A
L
H
O
M
N
0
0
I
‘
1
4
I
L
1
o
H
N
Y
Reply Expert Report of Professor Keith N. Hylton 1*
Commissioner of Competition v. Google LLC and Google Canada Corporation (CT-2024-010)
July 23, 2025
1*
William Fairfield Warren Distinguished Professor, Boston University and Professor of Law, Boston University School of Law.
4
1
5
TABLE OF CONTENTS PART I INTRODUCTION AND SUMMARY OF OPINIONS ......................................3 A. Qualifications and Assignment ............................................................................................3 B. Duties as an Independent Expert..........................................................................................4 C. Overview of Reply Report and Summary of Opinions ........................................................4 PART II PROFESSOR TADELIS MISAPPREHENDS THE INTERNALIZATION APPROACH ...........................................................................................................9
A.
B.
C.
Prevailing Economic Research and Literature Considers the Internalization Approach to be Superior to the Deterrence Approach ..............................................................................9
(i) (ii)
Internalization Approach is Consistent with Specific Deterrence ...........................9 Quantification and Measure of Economic Categories in Internalization Approach is Possible...............................................................................................................12
Professor Tadelis is Incorrect that I Did Not Address the Deterrence Approach in my First Report ........................................................................................................................13
The Internalization Approach is Not Premised on Monopolies Spurring Innovation and Efficiency ...........................................................................................................................14
(i)
(ii)
PART III
A.
B.
C.
Professor Tadelis Misunderstands the Internalization Approach and the Impact of Efficiencies on the Analysis in my First Report ....................................................14
Evidence of Greater Efficiencies and Innovation by a Monopolist Compared to Competition is Not Necessary ...............................................................................16
PROFESSOR TADELIS'S INCORRECT ASSERTIONS CONCERNING THE PROBABILITY OF DETECTION AND ENFORCEMENT ................19
The Assumption in my First Report Concerning Probability of Detection is Not Based on a Single Study ....................................................................................................................19
The Probability of Detection of Cartel Activity is Not Relevant to Probability of Detection in Monopolization Cases ...................................................................................23
Efforts by Firms to Avoid Transparency Do Not Undermine the Analysis in My First Report .................................................................................................................................23
D. Time is Not Relevant to the Probability of Antitrust Enforcement ...................................25 PART IV
PART V
PROFESSOR TADELIS MISCHARACTERIZES THE ANALYSIS OF HARM FROM MONOPOLIZATION ..............................................................27
PROFESSOR TADELIS'S OPINION THAT THE FINANCIAL PENALTY IS A CAP IS NOT RELEVANT .........................................................................30
2
PART I INTRODUCTION AND SUMMARY OF OPINIONS A. Qualifications and Assignment 1. I have been Davies Google the area of antitrust economics, specializing in optimal penalties in antitrust matters.
6
2. On May 6, 2025, I swore an Affidavit to which I attached my Expert Report of the same First Report of my First Report. My curriculum vitae
3. When I use capitalized terms in this Report that are not otherwise defined, I am relying on the meanings given to those terms in my First Report.
4. Since I submitted my First Report, I am advised by Davies that the Commissioner of Competition amended his Notice of Application in this proceeding on or about July 8, 2025 (the Amended Notice of Application I understand that the Commissioner has, among other things, revised the language used by him in his Application concerning the financial penalty the Commissioner seeks against Google in this proc anti-
5. Additionally, since submitting my First Report, I have received and reviewed an Affidavit sworn by Professor Steven Tadelis on July 4, 2025, which attached his Expert Report of that same Tadelis Report evaluate the opinions that I provided in my First Report. I have been asked by Davies to respond to several key points raised in the Tadelis Report.
6. As with my First Report, I do not purport to address the matters set out in this Report from a legal perspective, and I do not intend to provide an opinion on questions of Canadian law. Moreover, I do not express an opinion concerning the merits of the Com the underlying proceeding, including whether Google has or has not monopolized any product or geographic market alleged by the Commissioner. Instead, I simply assume for the purposes of the
3
7
analysis in this Report that Google has engaged in the alleged monopolizing conduct. Finally, I do not purport to opine upon what the precise quantum of the optimal penalty ultimately determined by the Competition Tribunal would be in this proceeding if liability were to be found, or how such a financial penalty would precisely be calculated.
B. Duties as an Independent Expert 7. As with my First Report, I am aware that as an independent expert, it is my duty to: (i) provide opinion evidence that is fair, objective and non-partisan; (ii) provide opinion evidence that is related only to matters that are within my area of expertise; and (iii) provide such additional assistance as the Competition Tribunal may reasonably require to determine a matter in issue. I understand that this duty prevails over any obligation I might owe to Google, its counsel or to any other party, and that I am Conduct for Expert Witnesses.
8.
9. As I stated in my First Report, I am being compensated for my involvement in this matter at my typical hourly rate. My compensation is in no way contingent on the contents of my First Report, this Report, or the outcome of this proceeding.
C. Overview of Reply Report and Summary of Opinions 10. Notice of Application. Nothing in the Tadelis Report or the Amended Notice of Application causes me to change the substance of the opinions and conclusions I provided in my First Report. It remains my opinion that:
a)
a)
a financial penalty equal to three times the value of the benefit derived from monopolizing conduct would result in inefficient overdeterrence and therefore can fairly be regarded from an antitrust economics perspective as having a punitive purpose; and
a financial penalty equal to three percent of annual worldwide gross revenues, in the case of Google would result in inefficient overdeterrence and can fairly be
4
8
regarded as punitive in nature from an antitrust economics perspective. Moreover, having regard to the extraordinary amounts that I understand are at issue here, I reach the same conclusion even if the Tribunal were ultimately to impose a financial penalty smaller than (or even equal to a fraction of) the statutory maximum.
11. As a preliminary matter, I note that throughout his Report, 2 Professor Tadelis offers opinions regarding the interpretation and purpose of the Competition Act and the broader objectives of competition law. In my view, such opinions concern matters of law. My First Report is an analysis of the financial penalties available under subsection 79(3.1) of the Competition Act from an economics perspective and does not offer opinions on Canadian law. Accordingly, I have s legal opinions and have limited my reply to economic matters raised in the Tadelis Report.
12. I would also note that the study of optimal penalties constitutes a recognized sub-specialty within the broader field of antitrust economics. In this context, I am not aware of any contributions by Professor Tadelis to the literature or practice relating to optimal penalties. To the best of my knowledge, Professor Tadelis has not published any work, whether in peer-reviewed journals or otherwise, addressing the theory or application of optimal penalties in antitrust economics.
13. the substance of my opinions for several reasons:
a)
s Report is flawed and does not impact
Professor Tadelis asserts that the deterrence approach is more appropriate for assessing antitrust penalties. However, as I explain below, this assertion is inconsistent with established antitrust economic theory. As a starting point, this approach comes from the traditional approach to antitrust regulation that I explained in my First Report was outdated. 3 Further, the economic literature
2 See Tadelis Report, at paras. 18, 19, 22, 25, 54, 55 and 58. 3 The traditional approach I refer to is based on the traditional approach to punishment reflected in Bentham, J. (1988). The Principles of Morals and Legislation. Prometheus Books (Original work published 1789). For a discussion of the internalization approach, as contrasted with the deterrence approach, see Becker G.S. (1968), Crime and Punishment: An Economic Approach Journal of Political Economy, 76(2), at pp. 169-217.
5
4 5
b)
c)
9
4 General deterrence refers to efforts aimed at regulating or excluding entire categories of activity, whereas specific deterrence focuses on optimally regulating the conduct of individual companies, organizations or business executives. 5 the realm of general deterrence. This approach seeks to broadly discourage certain categories of activity, even if there are beneficial aspects associated with the conduct in question. In contrast, specific deterrence aims to regulate the harmful conduct of specific actors engaged in commercial activity, without discouraging their beneficial conduct. The internalization approach used in my First Report falls
economics is that the internalization approach is superior because it preserves the beneficial aspects of the conduct in question while effectively addressing its harmful effects. The internalization approach differs from the traditional approach to antitrust regulation in that it seeks optimal deterrence;
Professor Tadelis asserts that my First Report failed to address the deterrence approach to evaluating optimal penalties and focused solely on the internalization approach. This assertion is incorrect. In my First Report, I discussed both the traditional, general deterrence approach advanced by Professor Tadelis and the specific deterrence focused, internalization approach. While I did not conduct an analysis of the financial penalties through the lens of the traditional model of antitrust regulation, I explained why, in the context of antitrust economics, the specific deterrence focused, internalization approach is preferable;
Professor Tadelis further criticizes my use of the internalization approach by arguing that I have not shown that monopolies lead to greater efficiencies or innovation than competition. However, this criticism is misplaced. My First Report does not rely on, or make assumptions about, the relative degree of efficiencies or innovation under monopoly versus competition. In any event, well-established economic literature has long recognized that the prospect of monopoly induces
Calabresi, G. (1970), The Costs of Accidents: A Legal and Economic Analysis (Yale University Press). Calabresi, 1970.
6
d)
e)
10
innovation and efficiencies. Professor Tadelis relies primarily on economic literature from Professor Carl Shapiro in criticizing my alleged assumptions about
contradict mine. Profess
concerned;
6
-
Professor Tadelis seeks to undermine the assumption in my First Report that there is near-certain (i.e., above 80 percent) compound probability of detection and enforcement against monopolization. However, Professor Tadelis ignores that monopoly conduct typically takes place in public and is therefore usually highly detectable. Moreover, even when the actual probability of enforcement may be low, firms particularly those as prominent as Google often behave as if the likelihood of enforcement is high. This is largely because such firms are well aware that their competitors have strong incentives to report any suspected violations of competition law and are easily able to do so either directly or indirectly. 7 Professor Tadelis also erroneously focuses on the probability of successful prosecutions and findings of liability to conclude that the probability of detection and enforcement in monopolization cases is low. This is not the correct approach. The low success rate of unmeritorious complaints has no relationship to whether meritorious complaints against monopolization are enforced;
Professor Tadelis asserts that by focusing on price effects, my First Report presents an incomplete picture of the societal harm caused by monopolization. He suggests that if all potential harms were considered, the internalization approach would likely result in a higher optimal penalty than the deterrence approach. However, his analysis is methodologically unsound because it selectively expands the analysis to
6 Schumpeter, J. (1942), Capitalism, Socialism, and Democracy (Harper & Brothers); Shapiro, C. (2012), Competition and Innovation. Did Arrow Hit the Bull's Eye in J. Lerner & S. Sterns (Eds.), The Rate and Direction of Inventive Activity Revisited (University of Chicago Press), at pp. 361-404. 7 Wu, T. (2012), Taking Innovation Seriously: Antitrust Enforcement If Innovation Mattered Most Antitrust Law Journal, 78(2), at pp. 313-326 ll show you a potential Section 2 case, or at least a lengthy investigation ).
7
14. Parts:
f)
11
include only additional potential harms without accounting for additional potential benefits. A balanced approach that considers both additional harms and benefits would likely result in an optimal penalty under the internalization approach. Moreover, the additional potential harms that Professor Tadelis identified are not so empirically certain and prevalent that they would significantly modify the analysis based on price effects; and
Professor Tadelis suggests that my First Report assumes the final penalty imposed will always be equal to the maximum statutory financial penalty provided for in subsection 79(3.1) of the Competition Act. I made (and make) no such assumption. As is stated above, my opinions in this matter remain unchanged even if the Competition Tribunal were to impose financial penalties upon Google that are smaller than (or even a fraction of) the statutory maximum. This is for two reasons. First, because the statutory maximum penalties are so likely to be excessive in their overdeterrent effect that there is a wide range of potential penalties that will have an overdeterrent effect before the maximum is reached. Second, it is rational for a large firm such as Google to treat the statutory penalties provided for in legislation of this nature as fair indications of what it might be required to pay upon a finding of a violation. The statutory penalties are the only guideposts upon which Google can rationally determine its course of conduct. The mere possibility that the Tribunal might impose a lower penalty in some cases does not eliminate the overdeterrent effect of the financial penalties provided for in the Competition Act.
I have organized the balance of my Report and analysis of the Tadelis Report into four
a)
b)
Part II responds to certain characterizations and statements made by Professor Tadelis concerning the internalization approach to evaluating antitrust penalties;
Part III antitrust detection and enforcement;
s analysis concerning the probability of
8
c)
d)
12
Part IV s analysis concerning the consideration of additional potential harms from monopolization; and
Part V s opinion concerning the characterization of the financial penalties provided for in subsection 79(3.1) of the Competition Act or in this case, fractions of those penalties would result in inefficient overdeterrence.
15. I have included in Report.
PART II
A.
to this Report a list of the documents referenced in this
PROFESSOR TADELIS MISAPPREHENDS THE INTERNALIZATION APPROACH
Prevailing Economic Research and Literature Considers the Internalization Approach to be Superior to the Deterrence Approach
16. In paragraphs 11-12 and 17-24 of his Report, Professor Tadelis sets out his understanding of the internalization approach to evaluating antitrust penalties (i.e., an approach that internalizes consumer harm) and the deterrence approach (i.e., an approach that focuses on completely
9 I disagree with Professor inconsistent with prevailing economic research and analysis.
17.
10
8
He further contends
conclusions, which are
(i) Internalization Approach is Consistent with Specific Deterrence To place Professor
assertions in their appropriate context, it is important to
8 Tadelis Report, at para. 12. 9 Tadelis Report, at paras. 12 and 22. 10 Becker, 1968, at pp. 169-217; Stigler, George J. (1970) The Optimum Enforcement of Laws Journal of Political Economy, 78(3), at pp. 526-536; Landes, W.M. (1983) Optimal Sanctions for Antitrust Violations Chicago Law Review, 50(2), at pp. 652-678; Hylton, K.N. & Lin, H. (2014) Innovation and Optimal Punishment with Antitrust Applications Journal of Competition Law and Economics, 10(1), at pp. 1-26.
University of
9
13
concepts were defined in Calabresi (1970). 11 Calabresi refers to general deterrence as an effort to regulate or exclude entire activities, while specific deterrence is an approach that focuses on
advocated by Professor Tadelis should be understood as a form of general deterrence, as it seeks to completely discourage certain types of activity. In contrast, specific deterrence aims to regulate only the harmful conduct of specific actors, without discouraging their beneficial activities. The internalization approach I used in my First Report aligns with the concept of specific deterrence.
18. (1968), established that the internalization approach is, in many contexts, superior to the general deterrence approach, particularly in the field of modern regulatory enforcement such as antitrust. 12 This is because in areas like antitrust, regulated activities often produce both harmful and beneficial effects. As I explained in my First Report, a policy of complete deterrence, as advocated by the general deterrence approach, would eliminate both the positive and negative aspects of such activities, thereby depriving society of their benefits. 13 In contrast, the purpose of the internalization approach is to preserve the beneficial effects of activities while limiting and regulating the harmful effects. Therefore, when evaluating activities with mixed effects, economists aiming to have an enforcement system that maximizes social welfare will generally favor the internalization approach. 14
19. It is widely accepted among antitrust scholars that monopolization is a mixed activity, comprising conduct that may be harmful to society as well as conduct that is clearly beneficial. 15 This understanding is reflected in the Competition Act itself. The Act does not prohibit mere dominance in a market; rather, it targets only the abuse of such dominance. This distinction
11 Calabresi, 1970. 12 Becker, 1968, at pp. 169-217. 13 See First Report, at paras. 31-32. 14 Viscusi, W.K. (1993) The Value of Risks to Life and Health Journal of Economic Literature, 31(4), at pp. 1912-1946; Polinsky, M., & Shavell, S. (1979) The Optimal Tradeoff Between the Probability and Magnitude of Fines American Economic Review, 69(5), at pp. 880-891. 15 Williamson, O.E. (1968) Economies as an Antitrust Defence: The Welfare Trade-Off American Economic Review, 58(1), at pp. 18-36; Hylton, K.N. (2003), Antitrust Law: Economic Theory and Common Law Evolution (Cambridge University Press); Hylton & Lin, 2014, at pp. 1-26.
10
14
underscores the general consensus that monopoly can, in many cases, generate efficiencies and confer social benefits. 16
20. In light of the mixed nature of monopolistic activity, as a default rule the internalization approach is considered by most antitrust economists, including me, to be unambiguously superior to the deterrence approach, which is largely seen as outdated and unnecessary. This is because the internalization approach preserves the beneficial effects of monopolies while limiting their harmful consequences. 17
21. assumes that there is an optimal level of monopoly. The notion is that if the monopolist can pay a penalty that includes all the social costs of monopoly and still make some profit, then society is
warrants clarification because it suggests that the internalization approach reflects a preference for monopoly over competition. However, no such preference is implied by the approach. As I explained above, the internalization approach approves of monopoly when there are mixed effects (both beneficial effects and harmful effects). The internalization approach aims to deter the specific conduct that results in a substantial lessening or prevention of competition while avoiding deterrence of welfare maximizing conduct.
22. In contrast, the prevailing view among antitrust economists, including myself, is that the deterrence approach is appropriate only in situations where the conduct in question offers no potential benefits to society , 18 like in the case of price fixing which has no redeeming qualities from an economic perspective and is per se illegal. As I explained in my First Report, although such cases exist, they are rare and exceptional in the context of monopolization.
23. Moreover, even in these rare instances, the internalization approach typically provides an effective framework. To illustrate, consider the example from my First Report involving the
16 Hylton, K.N. & Evans, D.S. (2008), The Lawful Acquisition and Exercise of Monopoly Power and its Implications for the Objectives of Antitrust Competition Policy International, 4(2), at pp. 1-47. 17 Becker, 1968, at pp. 169-217; Landes, 1983, at pp. 652-678; Hylton, 2003. 18 Landes, 1983, at pp. 652-678; Hylton & Lin, 2014, at pp. 1-26. For a more general argument touching on areas outside of antitrust, see Hylton, K.N. (1998), Punitive Damages and the Economic Theory of Penalties Georgetown Law Journal, 87(2), at pp. 421-472.
11
15
of conduct is purely anticompetitive, as it produces no offsetting efficiencies or societal benefits. Accordingly, such conduct should be completely deterred, and the penalties should be set at a level that eliminates any prospect of gain from the activity. However, even if the penalty is set at the internalization level, it will generally suffice to deter purely destructive acts. In this scenario, the offend the monopoly profit obtained will be less than the total harm imposed on victims. The gain to the offender consists of the monopoly profit, which is derived from surplus transferred from . In contrast, the total harm to society includes not only this transfer of surplus, but also the deadweight loss and the harm inflicted on rival firms through the destruction of their facilities. The total harm exceeds the
24.
(ii)
Quantification and Measure of Economic Categories in Internalization Approach is Possible
Professor Tadelis also asserts in paragraphs 12 and 23 of his Report that the internalization
r surplus and efficiencies, and
criticism by Professor Tadelis.
25. Consider the difficulties associated with measuring consumer surplus. First, such a measurement is not necessary for the analysis I provided in my First Report and has no bearing on the substance of my opinion. My First Report explains conceptually that is, without requiring any calculation why the penalties contemplated by subsection 79(3.1) of the Competition Act would result in inefficient overdeterrence and should therefore be viewed as punitive. Answering such a question does not require calculation of the actual financial penalty imposed by the Competition Tribunal in a particular case. Second, if necessary, economists are capable of making complex measurements, such as estimating the loss in consumer surplus resulting from monopolization, including the calculation of deadweight loss. As an illustration, I provided a numerical example of such a calculation in paragraphs 49 to 51 of my First Report. This type of calculation would be appropriate when seeking to determine the total consumer harm caused by an antitrust violation.
12
16
26. Similarly, with respect to the difficulties associated with measuring efficiencies, economists can also, with sufficient data, obtain reasonable estimates. I provided an example of such a calculation in my First Report, paragraphs 49-51. The easiest case in which to measure the efficiency gain is where it consists of a reduction in unit costs of an estimable magnitude. This is the example I consider in my First Report. Indeed, efficiencies are commonly evaluated by antitrust economists and enforcement agencies.
27.
B.
For these reasons, Professor
conclusion that my reliance on the internalization
Professor Tadelis is Incorrect that I Did Not Address the Deterrence Approach in my First Report
28. Professor Tadelis asserts that, in my First Report, I did not discuss the deterrence approach to evaluating optimal penalties and instead only considered the internalization approach. 19 Professor Tadelis is mistaken. In my First Report I considered both approaches the general deterrence approach under the traditional approach to antitrust regulation and the internalization approach and explained why, in the context of antitrust economics, the traditional approach is outdated and the internalization approach, which is consistent with specific deterrence, is superior.
29. By way of example only, 20 at paragraph 26 of my First Report, I explained that the traditional general deterrence model of antitrust is viewed by most antitrust economists in the U.S., including me, as outdated and that, as a result, I did not propose to conduct an analysis of the financial penalties sought against Google through the lens of the traditional model of antitrust regulation.
30. However, I also explained in paragraphs 31 and 32 of my First Report that in the absence of market efficiencies generated by the monopolization at issue arising from the conduct in question, the optimal penalties framework would provide that a financial penalty should be set at
19 20
Tadelis Report, at para. 17. See also First Report, at paras. 36-37.
13
31.
17
However, as I discuss in greater detail above, where the conduct at issue generates
i.e., complete deterrence) financial penalties are not optimal and are therefore punitive in nature. A financial penalty that eliminates the benefit(s) received by a firm engaging in market efficiency-enhancing conduct results in inefficient deterrence. A financial penalty that requires not only the disgorgement of the benefit derived from the conduct in question, but multiples of that benefit, is in my view clearly punitive from the perspective of antitrust economics. To promote activity in the marketplace that promotes efficiency, the optimal penalties framework would provide that financial penalties must be moderated to levels that would not result in overdeterrence of those activities (i.e., using an internalization approach).
C.
The Internalization Approach is Not Premised on Monopolies Spurring Innovation and Efficiency
(i)
Professor Tadelis Misunderstands the Internalization Approach and the Impact of Efficiencies on the Analysis in my First Report
32. Professor Tadelis opines in paragraph 25 of his Report that the internalization approach a claim that runs 21 This characterization of the internalization approach is not consistent with my understanding as an expert in the area of optimal penalties. Nor is this characterization consistent with the prevailing view in antitrust economics. 22
33. The internalization approach is designed to account for a range of scenarios. This includes cases where monopoly leads to efficiency gains, situations in which the prospect of monopoly incentivizes innovation, as well as circumstances where neither efficiency nor innovation is realized. As Becker observed almost six decades ago in his 1968 article, when the harm caused by offensive conduct exceeds the gains to the offender, the internalization approach is sufficient to achieve complete deterrence. 23 endorsed by numerous antitrust economists in the lengthy period since his article was first
21 22 23
Tadelis Report, at para. 25. Becker, 1968, at pp. 169-217; Landes, 1983, at pp. 652-678; Hylton, 2003. Becker, 1968, at pp. 169-217.
14
18
published. 24 Antitrust economists have generally accepted that the internalization approach is optimal for efficient specific deterrence, regardless of whether actual efficiencies are realized. In this respect, the internalization approach I employed in my First Report is broader than the deterrence approach advocated by Professor Tadelis. My framework subsumes his as a special case namely, where monopolization does not produce any efficiency or innovation benefits.
34.
In paragraph 26 of his Report, Professor Tadelis cites a paper that I authored in 2015 to
approach and the internalization approach are essentially the same. When monopolization does not
25
While Professor Tadelis accurately
summarizes this aspect of my paper, 26 I disagree with how he applies this concept in his analysis and with the conclusions he ultimately draws from it.
35. As a preliminary matter, dominance and the pursuit of monopoly is not harmful from an economics perspective and there is a general understanding that monopoly not infrequently entails efficiencies and social benefits. 27 This prevailing view is reflected in the statutory frameworks of both the US and Canada. As noted above in the Competition Act, it is the abuse of a position of dominance that violates the Act. Similarly, under the Sherman Act in the U.S., it is wrongful monopolization that violates the Sherman Act, not the existence of a monopoly.
36. proceeding that Google is an innovative firm and there are numerous efficiencies associated with its conduct in respect of its provision of advertising technology tools and services. 28 Indeed, it appears that the product features that the Commissioner complains of in this case were innovations By way of example only, the Response
24 Stigler, 1970, at pp. 526-536; Landes, 1983, at pp. 652-678; Kaplow, L. (2011), An Economic Approach to Price Fixing Antitrust Law Journal, 77(2), at pp. 343 449; Hylton & Lin, 2014, at pp. 1-26. 25 Hylton, K.N. (2015), Antitrust Enforcement Regimes: Fundamental Differences in Blair, R.D. & Sokol, D.D. (Eds.), The Oxford Handbook of International Antitrust Economics (Oxford University Press), at pp. 17-32. 26 First Report, at para. 32. 27 Hylton & Evans, 2008, at pp. 1-47; Schumpeter, 1942; Gilbert, R.J. (2006) Looking for Mr. Schumpeter: Where Are We in the Competition Innovation Debate? in Jaffe, A.B., Lerner, J., & Stern, S. (Eds.), Innovation Policy and the Economy (Vol. 6) (The MIT Press), at pp. 159-215; Hylton & Lin, 2014, at pp. 1-26. 28 , at paras. 250-257.
15
19
benefits of realtime bidding while ensuring that Google can offer publishers and advertisers an acceptably fast speed of service, a better assurance of safety, a higher quality of advertising 29
30
37. Moreover, in paragraph 26 of his Report, Professor Tadelis presents another scenario. He asserts that if the monopoly does not create efficiencies then a penalty of 31 In my view, this conclusion is incorrect and is not supported by my 2015 paper, which Professor Tadelis relies upon in his attempt to advance this argument.
38. Even if we make the unlikely assumption that a particular monopoly does not result in any efficiency gains, it does not necessarily follow that a penalty equal to three times the benefit derived from that monopoly is too low. That is because the threshold at which a penalty becomes a punitive penalty depends on the probability of detection and enforcement. If the likelihood of enforcement is high (for example, close to 100 percent), then an optimal penalty need only be set at a level equal to the total consumer harm. In such cases, a penalty set at three times the benefit is very likely to be greater than the total consumer harm such that the penalty inefficiently deters, is not optimal and, instead, is punitive in nature. 32
39.
(ii)
Evidence of Greater Efficiencies and Innovation by a Monopolist Compared to Competition is Not Necessary
In paragraph 27 of his Report, Professor Tadelis criticizes my First Report for allegedly
alleged monopolization in advertising technology has resulted in efficiencies, citing the U.S. United States v. Google LLC to support his position.
29 30 31 32
, at para. 255(a). , at para. 265. Tadelis Report, at para. 26. See Hylton, 2003.
16
20
40. Professor Tadelis begins his analysis of innovation by stating that monopoly is necessary for innovation or that firms only have incentives to innovate if they have prospects of gaining monopoly power. This statement conflicts with the economic literature. There is a significant economic literature showing that many
Tadelis then goes on to opine in paragraph 3 monopolization is expected to increase innovation, but competition is not expected to result in new technologies and know-
41. Professor critiques and characterizations of the opinions in my First Report are flawed for the following reasons.
42. First, My First Report does not make assumptions about the relative degree of efficiencies or innovation under monopoly versus competition. Rather, the analysis in my First Report is based on the uncontroversial economic principle that the prospect of profit whether through monopoly, the use of intellectual property rights or other means serves as an incentive for innovation. 33 As Joseph Schumpeter, one of the most influential economists of the 20 th century, notably observed, the possibility of temporary monopoly profits can motivate firms to innovate. 34
43. I readily acknowledge that competition can also stimulate innovation and efficiencies and nothing in my First Report suggests otherwise. Competition often leads to the development of new technologies, production methods or business processes, which may in turn provide firms with a temporary competitive advantage or even a short-term monopoly. 35 However, the fact that competition can drive innovation and efficiencies does not mean that a monopoly cannot also foster innovation and efficiencies. 36
33 Schumpeter, 1942. See also Shapiro, 2012, at pp. 361-404; Cass, R.A. (2013), Laws of Creation: Property Rights in the World of Ideas (Harvard University Press); Cohen, W.M., Nelson, R.R., & Walsh, J.P. (2000) Intellectual Assets: Appropriability Conditions and Why U.S. Manufacturing Firms Patent (or Not) National Bureau of Economic Research, Working Paper 7552. 34 Schumpeter, 1942, at ch. 7-8; Mason, E. (1951), Schumpeter on Monopoly and the Large Firm Economics and Statistics, 33(2), at pp. 139-144. 35 Shapiro, 2012, at pp. 361-404. 36 Hylton & Evans, 2008, at pp. 1-47.
Protecting Their
The Review of
17
21
44. Second, in the highly unlikely event that Google has failed to innovate or create efficiency gains, the analysis in my First Report would remain valid. My First Report offers a general analysis of the financial penalties available under subsection 79(3.1) of the Competition Act and the optimal approach to antitrust penalties. This approach and analysis are not dependent on the specific facts of any one case. The conclusions and substance of my analysis remain valid even if, in some cases, the facts and circumstances to which the analysis are applied are altered (for example, some variables in my analysis, such as the measure of the efficiency gain, are set at zero instead of being
cre encompasses instances in which efficiencies are both significant and insignificant. So long as it is possible that a firm such as Google could generate innovation or efficiency gains, the appropriate approach to penalization remains the internalization approach that I explained in my First Report. This is because the internalization approach is optimal, for specific deterrence purposes, whether or not efficiencies actually result from the business activities in question.
45. Third, Professor assertion that Google has generated no plausible efficiencies is highly unlikely to be true. As explained above in paragraph 36, Google is highly innovative, including with respect to the development and implementation of its advertising technology tools and services. Successful companies almost always innovate or create efficiency gains to some degree. If Google were not innovative and did not create any efficiency gains, it is highly unlikely that it would have been a successful company.
46.
In paragraph 29 of his Report, Professor Tadelis comments on the use of the Dynamic
incorporating a subsidy for the monopolist, intended to incentivize innovation. Again, if the monopolist is not engaged in meaningful innovation, or the innovations from competition exceed
essment.
47. Professor innovation, then the Dynamic Model
never engage in meaningful which presupposes that the monopolist is, in fact,
18
22
innovating would have no relevance and would not be used as a method of analysis. The Dynamic Model is specifically designed for situations in which a monopolist undertakes meaningful, substantial or market-creating innovation, sometimes characterized as dynamic efficiencies. The Dynamic Model is widely accepted in antitrust economics as a meaningful methodology to determine optimal penalties that ensure that regulation does not discourage firms from innovating. 37
48. s assertion that monopolists never engage in meaningful innovation is both extreme and empirically unsupported. 38 Even if, for the sake of argument, this assumption was true in a particular case, the deterrence approach Professor Tadelis recommends would still be suboptimal. Specifically, the financial penalties provided for in subsection 79(3.1) of the Competition Act would remain excessively punitive, resulting in over-deterrence, because they would eliminate the prospect of gains created from operational efficiencies. In contrast, the internalization approach would continue to be preferable, even under this extreme assumption, because it accounts for potential operational efficiencies that may arise from monopoly. This alone justifies the superiority of the internalization approach, even where dynamic efficiencies are not present.
PART III
A.
PROFESSOR TADELIS'S INCORRECT ASSERTIONS CONCERNING THE PROBABILITY OF DETECTION AND ENFORCEMENT
The Assumption in my First Report Concerning Probability of Detection is Not Based on a Single Study
49. basis for his assumption of near-
or evaluating monopolization
37 Sidak, J.G., & Teece, D.J. (2009) Dynamic Competition in Antitrust Law Journal of Competition Law & Economics, 5(4), at pp. 581-631; Teece, D.J. (2012) Next Generation Competition: New Concepts for Understanding How Innovation Shapes Competition and Policy in the Digital Economy Journal of Law, Economics and Policy, 9(1), at pp. 328-352; Hylton & Lin, 2014, at pp. 1-26; Spulber, D.F. (2023), Antitrust and Innovation Competition Journal of Antitrust Enforcement, 11(1), pp. 5-50; Petit, N., Schrepel, T., & Heiden, B. (2024), Situating The Dynamic Competition Approach Dynamic Competition Initiative (DCI), Working Paper 1-2024. 38 Even the scholarly articles that appear to be sympathetic to this assertion do not support it. See, for example, Aghion, P., Bloom, N., Blundell, R., Griffith, R., & Howitt, P. (2005) Competition and Innovation: an Inverted-U Relationship Quarterly Journal of Economics, 120(2), pp. 701-728; Shapiro, 2012, at pp. 361-404.
19
23
mergers must be reported to the FTC such that the detection rate for anticompetitive mergers is likely higher than for monopolization cases; and (iii) the study concerns the probability of issuing a complaint without considering the success rate of the complaint. This criticism is misplaced for at least three reasons.
50. First, my opinion that there is a near-certain (i.e., above 80 percent) compound probability of detection and enforcement of monopolizing conduct is not based on a single study. It is based on my examination over a period of many years of the process of public and private litigation in the U.S. in monopolization, merger and cartel cases. In merger and monopolization cases, enforcement authorities are often alerted by rivals about their concerns regarding the potentially monopolizing effects of the actions of dominant firms in the markets in which they operate. 39 In cartel cases, by contrast, the activities of participants in the cartel are often kept secret. It is well accepted in the antitrust economics literature that actions in monopolization cases (unlike in cartel cases) are publicly observable, with the result that there is a high probability of detection and enforcement. 40 I cited the FTC study as only one example to support my opinion. Other examples include studies noting the powerful incentives of private actors to report or to sue under the American antitrust laws. 41
51. Second, merger and monopolization cases share procedural similarities, making the FTC study a highly relevant example. As I explained in my First Report, in both merger and monopolization cases, the actions of the defendant or investigation target are publicly observed. Mergers are often announced and reported in mainstream media and now social media well before they are consummated. Monopolizing acts are often publicly disclosed or announced and reported both in mainstream media and now on social media at an early stage, including in many cases before they are undertaken. Even where the monopolizing acts are less visible, industry insiders are likely to know of the acts long before consummation or realization, and certainly while they are ongoing.
39 Muris, T.J. (2000), The FTC and the Law of Monopolization Antitrust Law Journal, 67(3), pp. 693-723. 40 Kovacic, W.E. (2004) Private Participation in the Enforcement of Public Competition Laws in M. Adenas et al. (Eds.), Current Competition Law (Vol. 2), at pp. 167-179); Brodley, J.F. (1995) Antitrust Standing in Private Merger Cases: Reconciling Private Incentives and Public Enforcement Goals Michigan Law Review, 94(1), at pp. 1-108. 41 See Kovacic, 2004, at pp. 167-179; Brodley, 1995, at pp. 1-108; Hylton, K.N. (2015) Microsoft After Fifteen Years Competition Policy International, 11(1).
20
24
52. Moreover, both merger and monopolization cases may involve reports or complaints to enforcement agencies, with the result that the probability of detection and enforcement is high. In the context of a merger, industry rivals who believe that the proposed merger will be anticompetitive in effect will contact the enforcement agencies to warn them or persuade them to act. 42 The same is true of monopolization, where rival firms may report their concerns or complaints to enforcement agencies. 43 In both contexts, one is more likely to see over-reporting, rather than under-reporting, given that firms have a self-interest in seeing enforcement commenced against their rivals.
53. Third, contrary to Professor assertion, the success rate of a complaint is not a more appropriate metric than the probability of issuing a complaint or challenge (which was the metric used in the FTC Study). The success rate of complainants does not generally provide useful information unless valid complaints or challenges can be accurately distinguished from invalid complaints and challenges, which is a very challenging distinction to make at the complaint stage. Accordingly, the probability of accurate punishment could be 100 percent, and yet we might still observe a low success rate because of false, frivolous or unmeritorious complaints or reports.
54. In paragraph 41 of his Report, Professor Tadelis points to statistics in the Federal Trade -Scott-Rodino Annual Report for 2023 concerning premerger reporting that of 1,805 transactions reported, 37 resulted in second requests, approximately 14 complaints were filed, resulting in 3 litigated wins to support the position that the probability of detection for mergers is less than the 97% indicated by the FTC study. However, these statistics are not instructive or helpful.
55. As Professor Tadelis himself acknowledges, it is possible for the compound detection and accurate enforcement rate to be 100 percent even if the statistics he presents are observed. The data he relies upon do not demonstrate that the compound rate of accurate detection and punishment in monopolization cases is low; in fact, they do not preclude the possibility that the rate is as high as 100 percent. To illustrate this point, consider the following hypothetical: Of the
42 Kovacic, 2004, at pp. 167-179. 43 Kovacic, 2004, at pp. 167-179; Hylton, 2015; Page, W.H. & Lopatka, J.E. (2007), The Microsoft Case: Antitrust, High Technology, and Consumer Welfare (University of Chicago Press) (both discussing lobbying efforts against Microsoft before and during litigation).
21
25
1,805 transactions reported in 2023, suppose that 1,500 posed no risk of violating the law. Of the remaining 305 transactions, assume 200 were withdrawn after regulatory scrutiny or intervention, leaving 105 transactions as the base for potential second requests. If only 37 resulted in second requests and among those, only 3 actually violated the statute with the courts correctly identifying and litigating those three cases then the number of proposed mergers that would have violated the Clayton Act would be 203 (200 resolved by withdrawal and 3 through litigation). In this scenario, the detection and enforcement accuracy would be 100 percent, yet the statistics cited by Professor Tadelis would still be observed. This example demonstrates that the statistics referred by Professor Tadelis do not, in themselves, indicate underenforcement, contrary to his suggestion.
56. Professor Tadelis makes a similar assertion in paragraph 44 of his Report, contending that the opinion in my First Report that the probability of detection and enforcement of monopolization low probability of successful This is a mischaracterization of the opinion on my First Report. My analysis is
-deterrent and punitive features of the statutory penalties provided for in the Competition Act under the assumption that there is a high enforcement probability, including one that is substantially less than 100 percent (which may fall below an 80 percent probability as well).
57. In any event, as with the success rate of complaints, the success rate of antitrust enforcement actions is not, as Professor Tadelis contends, an appropriate metric for determining the probability of detection and enforcement. This is because such metrics fail to account for the fact that some cases are brought in respect of conduct that is legally benign and does not account for cases that are resolved through settlement rather than litigation. In fact, a number of the most compelling cases for punishment are often settled before reaching the stage of a final judgment. 44 Therefore, the observation of a low conviction rate does not imply that the probability of detection and enforcement is low.
44 Baxter, W.F. (1980), The Political Economy of Antitrust in R.D. Tollison (Ed.), The Political Economy of Antitrust (Lexington Books), at p. 23 (noting low plaintiff win-rates in antitrust); Hylton, K.N. (1993) Asymmetric Information and the Selection of Disputes for Litigation The Journal of Legal Studies, 22(1), at pp. 187-210 (discussing antitrust on pp. 188, 190 and 200, noting that low plaintiff win-rates in antitrust are probably due to informed and guilty defendants settling).
22
B.
58.
The Probability of Detection of Cartel Activity is Not Relevant to Probability of Detection in Monopolization Cases
26
In paragraph 42 of his Report, Professor Tadelis opines that economists have studied the
cartel cases is relevant to, or can inform, the probability of detection in monopolization cases such as this one.
59. not a useful or appropriate basis for comparison in the context of assessing detection and enforcement. Cartels, as Professor Tadelis acknowledges in paragraph 43 of his Report, normally
probability of detection of cartels is quite low. In contrast, as explained above, monopolization itself is almost never a secret act. This difference alone is sufficient to render the conclusions of economic studies involving cartels inapplicable and irrelevant to this proceeding.
60. In addition to the factor of detectability, there are important differences in the incentives to report or to proceed with litigation conditional on detection. Rivals in monopolization and merger settings are normally businesses with a range of sophistication and resources. They may have strong incentives to report or to litigate, as well as the resources and will to do so. 45 Consumers in the cartel setting, however, tend not to have strong incentives to report or to litigate, including because the cost and burdens associated with doing so probably exceed the benefit to a single consumer.
C.
Efforts by Firms to Avoid Transparency Do Not Undermine the Analysis in My First Report
61.
suggesting that firms, like Google, face a lower probability of antitrust detection enforcement than contemplated in my First Report. To support this view, Professor Tadelis refers to allegations by
45
Brodley, 1995, at pp. 1-108.
23
27
the Commissioner of Competition in this proceeding that Google took steps to avoid transparency, including the alleged concealment of Project Bernanke. I disagree with Professor assessment for several reasons.
62. First, the internalization approach used in my First Report remains valid even if a firm takes steps to avoid transparency such that the probability of antitrust detection may not be close to 100 percent. As detailed above in paragraph 42, my First Report offers an analysis of the over-deterring and punitive nature of the financial penalties available under subsection 79(3.1) of the Competition Act. This analysis is not dependent on the specific facts or circumstances of any one case. My analysis can be applied to offer an algorithm for computing the optimal antitrust penalty in many different factual scenarios with a wide range of antitrust detection probabilities. Indeed, my First Report expressly considers scenarios in which the probability of enforcement is low (i.e., 1/3 or 60%) and explains that even in those scenarios, my opinion continues to hold.
63. Second, even when the probability of enforcement is low, firms especially those as prominent as Google tend to act as though they face a high probability of enforcement. 46 This is because such firms are acutely aware that they have many rivals, including rivals with ample resources with strong incentives to report any suspected violations of competition law. Additionally, rational enforcement agencies have a vested interest in uncovering violations by high-profile firms like Google. 47 This creates an environment in which the perceived risk of enforcement remains significant, regardless of any efforts to avoid transparency.
64. Third, Professor assertion that Google concealed its allegedly monopolizing conduct is inconsistent with his own opinion in paragraph 35 of the Tadelis Report. There, he
approved by the U.S. antitrust enforcement agencies, including the acquisition of Double Click
46
Wu, 2012, at pp. 313-326
47 Wu, 2012, at pp. 313-326. The United States and a consortium of state attorneys general currently are investigating Google for its possible abuse of market power: Scott Morton, F.M. & Dinielli, D.C. (2020), Roadmap for a Digital Advertising Monopolization Case Against Google (Omidyar Network), at p. 1 (https://omidyar.com/wp-content/uploads/2020/09/Roadmap-for-a-Case-Against-Google.pdf) -based digital platforms are attracting
.
digital advertising in
24
28
and Ad Meld. These agencies would no doubt have scrutinized these proposed acquisitions for any evidence of conduct likely to create or enhance monopoly power. The fact that these acquisitions were approved suggests to me that the agencies found no evidence either of concealed or unconcealed efforts to monopolize.
65. Fourth, even if, for the sake of argument, Google did conceal certain initiatives from market participants or others, this does not demonstrate that Google believed that there was a low probability of antitrust enforcement in relation to any of its business activities, either in relation to the provision of ad tech tools or services or otherwise. Concealment, if it occurred, may be confined to a particular business activity or initiative, or be entirely unrelated to antitrust concerns. In this regard, businesses frequently keep their plans and activities confidential to avoid alerting competitors and to secure a first-mover advantage. Alternatively, if concealment were motivated by antitrust concerns, it could reflect a belief that the firm in question faces an unusually high risk of unfounded antitrust challenges. In such cases, efforts to conceal may reflect a defensive response to the risk of spurious antitrust claims, rather than a belief in a low probability of enforcement in relation to genuine or well-founded antitrust concerns.
D. 66.
Time is Not Relevant to the Probability of Antitrust Enforcement
that his analysis does not account for the considerable time it takes to identify and prosecute a conclusion.
67. The passage of time between the establishment of a monopoly and the initiation of enforcement action does not necessarily indicate a failure of detection or enforcement. Nor will it he probability of antitrust enforcement. In many cases the passage of time is unavoidable. Take for instance the current
appears to take issue with extends as far back as 2008, more than 17 years ago.
48
See Amended Notice of Application, at paras. 150, 155 and 157.
48
25
29
68. Often, enforcement is delayed simply because there is no credible evidence that a firm has engaged in conduct warranting antitrust scrutiny or enforcement activity. In many cases, this absence of evidence reflects the fact that the firm in question has not engaged in anticompetitive behavior or has benefitted those who use its products or services. Additionally, the competitive process itself may result in the displacement of less efficient rivals, who may eventually complain about their losses. However, such complaints are not always immediate or meritorious, and their timing does not necessarily correlate with the presence of anticompetitive conduct. Finally, time is needed for agencies to conduct an investigation before enforcement is initiated, but once an investigation target is aware of the beginnings of an investigation, or even the beginnings of discussions of a potential investigation, the likelihood of an enforcement action becomes palpable. 49
69. Professor Tadelis seems to suggest that immediately upon acquiring a dominant market position, a monopolist must be punished. However, this is not the approach competition enforcement regimes generally take. 50 In practice, enforcement authorities do not presume that dominance alone warrants intervention; rather, they assess whether the firm in question has engaged in conduct that constitutes an abuse of dominance or other anticompetitive behavior. 51
70. detection by collapsing the time period of analysis and asking whether the monopoly is ever successfully prosecuted (but even so the probability is low). This is not the correct way to calculate
despite raising this objection, Professor Tadelis does not propose an alternative method for calculating the probability of detection or enforcement.
71. Professor Tadelis suggests a potential criticism of the framework set out in my First Report based on the difficulty of calculating probabilities. This argument is unpersuasive. First, it is not necessary to determine precise probabilities in order to conclude that the penalties provided for in section 79(3.1) of the Competition Act are likely to be excessive and punitive. For this purpose, it
49 Wu, 2012, at pp. 313-326. 50 Hylton, 2003, at pp. 43-67 and 186-229 (discussing antitrust enforcement in the context of monopolization); Wu, 2012, at p. 327. 51 Hylton, 2003, at pp. 43-67 and 186-229 (discussing antitrust enforcement in the context of monopolization); Wu, 2012, at p. 327.
26
30
is sufficient to establish a plausible range of probabilities. Second, from a technical perspective, if there is concern that my analysis "collapses" the relevant time period, this can be addressed simply by extending the time frame under consideration. Doing so would not alter my general conclusions or my analysis.
PART IV PROFESSOR TADELIS MISCHARACTERIZES THE ANALYSIS OF HARM FROM MONOPOLIZATION 72. In paragraphs 47 to 51 of his Report, Professor Tadelis opines that the analysis in my First Report understates the social harm from monopolization by focusing on price effects and excluding harms such as barriers to entry, lost innovation from competition, umbrella effects (i.e., when a monopolist raises prices, its competitors also raise prices) and impacts on the political process.
monopolization, it is likely that the internalization approach would result in a much larger penalty 52 In my opinion, the concerns expressed by Professor Tadelis are flawed and misplaced for the following reasons.
73. First, as a general matter, it is the typical practice in economic analysis to focus on price effects rather than focusing on other types of alleged harm. Indeed, in paragraph 41 of the Tadelis Report, Professor Tadelis himself focuses on price effects when discussing a 2015 study of consummated mergers completed by John Kwoka. As a result, I do not agree that the analysis in my First Report was unduly narrow.
74. Second, the analysis in my First Report could be expanded to include additional harms, such as those identified by Professor Tadelis (e.g., barriers to entry, lost innovation from competition and umbrella effects). However, given that the internalization approach considers both social harms and efficiencies in arriving at the optimal penalty, I would then also need to expand the internalization framework analysis to include additional benefits (e.g., inducing innovation, 53 quality investments, 54 far-sighted research incentives, etc.). Professor Tadelis would appear to prefer to expand the analysis only to consider additional harms (i.e., additional negative
52 Tadelis Report, at para. 51. 53 Schumpeter, 1942. 54 Spence, A. (1975), Monopoly, Quality, and Regulation Bell Journal of Economics, 6(2), at pp. 417-429 (monopolists sometimes have an enhanced incentive to invest in quality).
27
31
externalities); however, selectively expanding the analysis to include only additional harms, as Professor Tadelis suggests, is methodologically unsound.
75. There are numerous other possible benefits from monopolies besides encouraging innovation through the chance to make profits. Monopolies in developed western economies have often generated substantial external benefits. For instance, the Bell System (which had a monopoly on phone service) helped generate some of the most important scientific innovations of the 20 th century. 55 Quantas, once an Australian air travel monopoly, became known for its excellent safety record. 56 Patents, which give their owners temporary monopolies over products that are subject to patent protection, often lead to more inventions. Alcoa, under government direction, built up the U.S. aluminum industry, which was very important during World War II. 57
76. Although the foregoing examples all involve some form of innovation, I distinguish them from the basic incentive the lure of monopoly provides to invest into the creation of a new marketable product or service (which is what I examine in the Dynamic Model). What distinguishes my examples in this paragraph is that they involve an investment that is typically not rewarded immediately by market participants (dramatic research discovery, safety improvements, national security investments). It is probably not a coincidence that dominant firms historically have made such investments, because some degree of market dominance is likely necessary to provide the security and resources required to make investments that might positively impact an entire industry, above a
77. Third, with respect to price umbrellas in particular, the price response of rivals depends on the circumstances. Sometimes rivals will raise prices. As stated above, this is known in the field of antitrust economics as the umbrella effect. Sometimes, however, rivals will compete more vigorously and lower their prices, 58 with the result that Professor opinion is highly speculative.
55 Gertner, J. (2012), The Idea Factory: Bell Labs and the Great Age of American Innovation (Penguin Press). 56 Earnes, J. (2021), Red Tail Skies: A Big Book of Qantas Stories (Allen & Unwin). 57 Skrabec, Q.R. (2017), Aluminum in America (McFarland). 58 Fudenberg, D. & Tirole. J. (1984), The Fat-Cat Effect, the Puppy-Dog Ploy, and the Lean and Hungry Look The American Economic Review, 74(2), at pp. 361-366 (theoretical); Borenstein, S. (1989) Hubs and High Fares: Dominance and Market Power in the U.S. Airline Industry The RAND Journal of Economics, 20(3), at pp. 344-365 (empirical).
28
32
78. Fourth, with respect to barriers to entry, not every monopolist can successfully create barriers to entry and exclude competitors. Blackberry lost its smartphone monopoly to Apple and other providers. Kodak lost its monopoly power to suppliers of digital technology, including digital cameras, smartphones, etc.
79.
-supported by the economics literature. The most recent study on this question of which I am aware finds no evidence that monopolization leads to enhanced political lobbying power. 59
80. from monopolies, it is important to remember that not every monopoly creates problems like foreclosing new competitors, raising prices across the market (i.e., the umbrella effect), or stopping innovation by others. There is no reason to assume that these additional harms exist in the absence of clear evidence on a case-by-case basis. Accordingly, it was not necessary for me to include these additional harms in the analysis set out in my First Report.
81. Importantly, if I incorporated all the benefits potentially caused by monopolization into my analysis, it is likely the internalization approach would result in a smaller optimal penalty. Conversely, had I incorporated all the potential harms discounted appropriately by their likelihoods of being significant, my conclusions would not have been significantly different even if I had ignored all the potential benefits. Finally, if I had included all the additional potential harms and all the additional potential benefits, my conclusions would have remained the same. In an antitrust setting, such as this, where positive and negative externalities are equally likely, the reasonable working assumption is that they cancel each other out. 60
59 Shahshahani, S. & McCarty, N. (2023), Testing Political Antitrust New York University Law Review, 98(4), at pp. 1169-1264. 60 Hylton, 2003.
29
33
PART V PROFESSOR TADELIS'S OPINION THAT THE FINANCIAL PENALTY IS A CAP IS NOT RELEVANT 82. In paragraph 57 of his Report, Professor Tadelis claims that I assume the penalty the Competition Tribunal will actually impose in this case, and in other cases, will always be equal to the maximum financial penalties provided for in subsection 79(3.1) of the Competition Act. This assertion is incorrect and, in any event, irrelevant. To be clear, I have made no such assumption.
83. The analysis in my First Report is based on the penalties provided for in subsection 79(3.1) of the Competition Act, which are the only clear guideposts available to firms in planning their conduct. Whether the Competition Tribunal imposes the maximum penalties in a given case, it is rational for a firm like Google to treat the statutory financial penalties as the relevant benchmarks for compliance purposes. Professor Tadelis does not identify any alternative, predictable standard firms could use in place of the financial penalties provided for by statute.
84. Furthermore, Professor unsupported assertion that the Competition Tribunal will always weigh all relevant evidence and set a financial penalty no larger than necessary is not a sufficient answer to the concern about overdeterrence. My analysis and opinions do not turn on the Tribunal imposing against Google the maximum financial penalties provided for in section 79(3.1) of the Act. Indeed, having regard to the amounts at issue in this case, some penalties that are smaller than (or a fraction of) the statutory maximum would still in my view as an antitrust economist constitute a punitive penalty.
85. Professor Tadelis does not claim to be able to predict whether a financial penalty determined for a firm such as Google would be less than the amounts provided for in subsection 79(3.1) of the Competition Act, or by how much. Indeed, Professor Tadelis cannot even credibly opine to know whether financial penalties calculated using the factors set out in the Competition Act will be optimal or suboptimal (either from the perspective of the deterrence approach or the internalization approach), with the result that firms that become Respondents in abuse of dominance proceedings commenced by the Commissioner may be subject to an overdeterrent and punitive penalty.
86. Moreover, the financial penalties set out in the Competition Act themselves, by their magnitude and structure, create a credible threat of punitive sanctions, particularly for large firms.
30
34
The mere possibility that the Tribunal might impose a lower penalty in some cases does not eliminate the overdeterrent effect of the financial penalties set out in the Competition Act. And the lower penalties themselves may be punitive in nature.
87. In paragraph 58 of his Report, Professor Tadelis suggests that to use the internalization approach, there must be sufficient information to calculate an optimal penalty, including the s that, as a result, the alternative statutory financial penalty available in the Competition Act of three percent of annual worldwide gross revenues is not applicable. However, this is an oversimplification, and appears to ignore the clear structure, wording and purpose of the provision in question. In practice, the determination of the benefit may be contested or uncertain, and the Tribunal may choose to proceed with the financial penalty of up to three percent of the annual worldwide gross revenues of the Respondent to abuse of dominance proceedings even when some information about the benefit derived by the Respondent from the anti-competitive practices at issue is available. Firms must therefore consider the possibility that some variant of the higher financial penalty will be imposed, regardless of the availability of benefit related information.
88. Finally, even if the benefit derived from the anticompetitive practice is known and the financial penalty is set at three times the benefit derived from the anticompetitive practice, this multiplier is not justified by the internalization standard and risks being punitive, as I have explained in my First Report. Professor Tadelis does not address this point in his Report.
89. of my opinions for the following primary reasons:
a)
assessing antitrust penalties is inconsistent with established antitrust economic theory. The prevailing view in antitrust economics is that the internalization approach is superior because it preserves the beneficial aspects of the conduct in question while effectively addressing its harmful effects. The internalization approach differs from the traditional approach to antitrust regulation in that it seeks optimal deterrence;
31
b)
c)
d)
e)
f)
35
approach to evaluating optimal penalties is incorrect. In my First Report, I discussed both the traditional, general deterrence approach advanced by Professor Tadelis and the specific deterrence focused, internalization approach;
Well-established economic literature has long recognized that the prospect of monopoly induces innovation and efficiencies. Moreover, the sources that Professor Tadelis relies upon to critique my First Report do not contradict my First Report;
there is near-certain probability of detection and enforcement of monopolies is misguided. Professor Tadelis ignores that monopoly conduct typically takes place in public and is therefore usually highly detectable;
the societal harm caused by monopolization is misplaced. The balanced analysis that I considered captures the harms and benefits that would likely result from the imposition of an optimal penalty under the internalization approach; and
imposed will always be equal to the maximum statutory financial penalty provided for in subsection 79(3.1) of the Competition Act is incorrect. The validity of the opinions I expressed in my First Report are not contingent on that assumption.
32
APPENDIX A LIST OF MATERIALS REFERRED TO
36
Baxter, W.F. (1980), The Political Economy of Antitrust , in R.D. Tollison (Ed.), The Political Economy of Antitrust (Lexington Books).
Becker, G.S. (1968), Crime and Punishment: An Economic Approach , Journal of Political Economy, 76(2), 169 219.
Borenstein, S. (1989), Hubs and High Fares: Dominance and Market Power in the U.S. Airline Industry , The RAND Journal of Economics, 20(3), 344 365.
Brodley, J.F. (1995), Antitrust Standing in Private Merger Cases: Reconciling Private Incentives and Public Enforcement Goals , Michigan Law Review, 94(1), 1 108.
Calabresi, G. (1970), The Costs of Accidents: A Legal and Economic Analysis (Yale University Press).
Cass, R.A. (2013), Laws of Creation: Property Rights in the World of Ideas (Harvard University Press).
Cohen, W.M., Nelson, R.R., & Walsh, J.P. (2000), Protecting Their Intellectual Assets: Appropriability Conditions and Why U.S. Manufacturing Firms Patent (or Not) , National Bureau of Economic Research, Working Paper 7552.
Earnes, J. (2021), Red Tail Skies: A Big Book of Qantas Stories (Allen & Unwin). Fudenberg, D. & Tirole. J. (1984), The Fat-Cat Effect, the Puppy-Dog Ploy, and the Lean and Hungry Look , The American Economic Review, 74(2), 361 366.
Gertner, J. (2012), The Idea Factory: Bell Labs and the Great Age of American Innovation (Penguin Press).
Hylton, K.N. & Evans, D.S. (2008), The Lawful Acquisition and Exercise of Monopoly Power and its Implications for the Objectives of Antitrust , Competition Policy International, 4(2), 1 47.
Hylton, K.N., & Lin, H. (2014), Innovation and Optimal Punishment, with Antitrust Applications , Journal of Competition Law and Economics, 10(1), 1 26.
Hylton, K.N. (1993), Asymmetric Information and the Selection of Disputes for Litigation , The Journal of Legal Studies, 22(1), 187 210.
Hylton, K.N. (2015), Microsoft After Fifteen Years , Competition Policy International, 11(1). Hylton, K.N. (2015), Antitrust Enforcement Regimes: Fundamental Differences , in Blair, R.D. & Sokol, D.D. (Eds.), The Oxford Handbook of International Antitrust Economics (Oxford University Press), 17 32.
33
37
Hylton, K.N. (2003), Antitrust Law: Economic Theory and Common Law Evolution (Cambridge University Press).
Kovacic, W.E. (2004), Private Participation in the Enforcement of Public Competition Laws , in M. Adenas et al. (Eds.), Current Competition Law, Vol. 2, 167 179.
Landes, W.M. (1983), Optimal Sanctions for Antitrust Violations , University of Chicago Law Review, 50(2), 652 678.
Mason, E. (1951), Schumpeter on Monopoly and the Large Firm , The Review of Economics and Statistics, 33(2), 139 144.
Muris, T.J. (2000), The FTC and the Law of Monopolization , Antitrust Law Journal, 67(3), 693 723.
Page, W.H. & Lopatka, J.E. (2007), The Microsoft Case: Antitrust, High Technology, and Consumer Welfare (University of Chicago Press).
Polinsky, M., & Shavell, S. (1979), The Optimal Tradeoff Between the Probability and Magnitude of Fines , American Economic Review, 69(5), 880 891.
Schumpeter, J. (1942), Capitalism, Socialism, and Democracy (Harper & Brothers). Scott Morton, F.M. & Dinielli, D.C. (2020), Roadmap for a Digital Advertising Monopolization Case Against Google (Omidyar Network) (https://omidyar.com/wp content/uploads/2020/09/Roadmap-for-a-Case-Against-Google.pdf).
Shahshahani, S. & McCarty, N. (2023) Testing Political Antitrust , New York University Law Review, 98(4), 1169 1264.
Shapiro, C. (2012), Competition and Innovation. Did Arrow Hit the Bull's Eye , in J. Lerner & S. Sterns (Eds.), The Rate and Direction of Inventive Activity Revisited (University of Chicago Press), 361 404.
Skrabec, Q.R. (2017), Aluminum in America (McFarland). Spence, A. (1975), Monopoly, Quality, and Regulation , Bell Journal of Economics, 6(2), 417 429.
Viscusi, W.K. (1993), The Value of Risks to Life and Health , Journal of Economic Literature, 31(4), 1912 1946.
Williamson, O.E. (1968), Economies as an Antitrust Defence: The Welfare Trade-Off , American Economic Review, 58(1), 18 36.
Wu, T. (2012), Taking Innovation Seriously: Antitrust Enforcement If Innovation Mattered Most , Antitrust Law Journal, 78(2), 313 326.
34
CT-2024-010 THE COMPETITION TRIBUNAL IN THE MATTER OF the Competition Act, R.S.C. 1985, c. C-34; AND IN THE MATTER OF certain conduct of Google Canada Corporation and Google LLC relating to the supply of online advertising technology services in Canada;
AND IN THE MATTER OF an application by the Commissioner of Competition for one or more orders pursuant to section 79 of the Competition Act.
B E T W E E N:
COMMISSIONER OF COMPETITION
and
GOOGLE CANADA CORPORATION AND GOOGLE LLC
Applicant
Respondents
AFFIDAVIT OF RUPERT YOUNGER SWORN JULY 22, 2025
I, Rupert Younger, of the City of Oxford, in the County of Oxfordshire, in England, MAKE OATH AND SAY AS FOLLOWS:
1. I am the founder and Director of the Oxford University Centre for Corporate Reputation (the “Centre”). The Centre is the world’s leading research centre studying organisational social evaluations—legitimacy, status, stigma, celebrity, reputation and trust—and their impact on organisations. I have more than three decades of experience in the study, research, and management of corporate reputations.
2. I have been engaged by Davies Ward Phillips & Vineberg LLP, on behalf of Google Canada Corporation and Google LLC, to act as an independent expert in this proceeding.
39 -2-I prepared an initial Expert Report dated May 6, 2025, which I attached to my Affidavit of that same date.
3. I have now prepared a Supplemental Expert Report dated July 22, 2025, which I have attached as Exhibit “1” to this Affidavit.
SWORN by Rupert Young of the C in England, before me at the City of Toronto, in the Province of Ontario, on July 22, 2025 in accordance with O. Reg. 431/20, Administering Oath or Declaration Remotely.
Commissioner for Taking Affidavits (or as may be)
RUPERT YOUNGER
ALISA MCMASTER
This is Exhibit “1” referred to in the Affidavit of Rupert Young sworn by Rupert Young of the in England, before me at the City of Toronto, in the Province of Ontario, on July 22, 2025 in accordance with O. Reg. 431/20, Administering Oath or Declaration Remotely.
Commissioner for Taking Affidavits (or as may be) ALISA MCMASTER
40
SUPPLEMENTAL EXPERT REPORT OF RUPERT YOUNGER
July 22, 2025
41
1.
2.
3.
4.
42
I have been retained by Davies Ward Phillips & Vineberg LLP (“Davies”), lawyers for Google Canada Corporation and Google LLC (collectively, “Google”) in this proceeding, as an independent expert. I previously an Expert Report in this proceeding on May 6, 2025. capitalised terms in my Supplemental Report dated July 22, 2025 have the meanings ascribed to them in my Expert Report dated May 6, 2025.
by counsel for Google that the Commissioner of Competition amended his Notice of Application in this proceeding on or about July 8, 2025 (the “Amended Notice of Application”).
I have reviewed a copy of the Amended Notice of Application. I understand that, among other things, the Commissioner has amended the language in his Notice of Application concerning the Commissioner seeks against Google in this proceeding from an amount “equal to three times the value of the Google’s anti-competitive practice, or if that amount cannot be reasonably determined, 3% of Google’s annual worldwide gross revenues” to “such an amount as the Tribunal decides is appropriate”.
These amendments do not impact or change the substance of the opinions I expressed in my Expert Report. That is so for at least three reasons:
(a)
(b)
(c)
I understand that it remains open to the Commissioner to take the position in this proceeding hree times the Google’s anti-competitive practice, or if that amount cannot be reasonably determined, 3% of Google’s annual worldwide gross revenues” should be imposed against Google, including because the Tribunal could conclude that such an amount is “appropriate”;
it does not appear from the Amended Notice of Application that the Commissioner hree Google’s anti-competitive practice, or if that amount cannot be reasonably determined, 3% of Google’s annual worldwide gross revenues” cannot be imposed by the Tribunal in this proceeding or that the imposition by the Tribunal of a such an amount would not be “appropriate”; and
the risk that against Google in this proceeding continues to exist and does not appear to have materially changed in the period on May 6. Indeed, I note that in his Amended Notice of Application (see paragraphs 142 to 147), it appears that the Commissioner has added additional allegations that Google “abused” a position of dominance in Canada, and that Google did so though “premeditated design and
5.
43
conduct” 1 . These additional allegations of deliberate wrongdoing compound the incremental harm to Google’s various character reputations in Canada, as well as the risk of Google becoming stigmatised in its dealings with a number of its most important stakeholder groups in Canada.
To be clear, my observations and opinions in respect of this matter are not contingent on the imposition by the Tribunal upon Google of the maximum
provisions of the Competition Act. Instead, I understand that the imposition by
imposed in Canada. In my view, for the reasons I have previously explained, the reputational consequences to Google would be the same.
6.
7.
1
The opinions in this Supplemental Expert Report are entirely my own. Although I have been retained on behalf of Google, I acknowledge that I am bound by and will comply with the Competition Tribunal’s code of conduct for expert witnesses. the Competition Tribunal impartially on matters relevant to my areas of expertise; (ii) that duty overrides any duty I might have to a party in this proceeding, including Google; and (iii) I am to be independent and objective, and not an advocate for a party.
I am being compensated for my involvement in this matter at my standard hourly rate. My compensation in this matter is not contingent on the substance of my opinion, or the outcome of the proceeding.
Reference taken from section 1.7 on p. 6 of the Amended Notice of Application.