"The Apple iPod iTunes Anti-Trust Litigation"

Filing 269

Memorandum in Opposition re 240 MOTION for Decertification of Rule 23(B)(3) Class filed byMelanie Tucker. (Sweeney, Bonny) (Filed on 10/19/2009)

"The Apple iPod iTunes Anti-Trust Litigation"

Doc. 269

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page1 of 34

1 COUGHLIN STOIA GELLER RUDMAN & ROBBINS LLP 2 JOHN J. STOIA, JR. (141757) BONNY E. SWEENEY (176174) 3 THOMAS R. MERRICK (177987) 655 West Broadway, Suite 1900 4 San Diego, CA 92101 Telephone: [phone] (fax) [e-mail] 6 [e-mail] [e-mail] 7 THE KATRIEL LAW FIRM 8 ROY A. KATRIEL (pro hac vice) 1101 30th Street, N.W., Suite 500 9 Washington, DC 20007 Telephone: [phone] (fax) [e-mail] 11 Co-Lead Counsel for Plaintiffs 12 [Additional counsel appear on signature page.] 13 UNITED STATES DISTRICT COURT 14 NORTHERN DISTRICT OF CALIFORNIA 15 SAN JOSE DIVISION 16 THE APPLE IPOD ITUNES ANTI-TRUST ) Lead Case No. C-05-00037-JW(RS) ) 17 LITIGATION ) CLASS ACTION ) 18 ) PLAINTIFFS' MEMORANDUM IN This Document Relates To: ) OPPOSITION TO DEFENDANT'S MOTION 19 ) FOR DECERTIFICATION OF RULE ALL ACTIONS. 20 ) 23(b)(3) CLASS 21 22 23 24 25 26 27 28 Judge: Date: Time: CTRM: Hon. James Ware November 23, 2009 9:00 a.m. 8-4th Floor

Dockets.Justia.com

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page2 of 34

1 2 3 I. 4 II. 5 6 7 III. 8 9 10 11 V. IV.

TABLE OF CONTENTS Page INTRODUCTION ...............................................................................................................1 PROCEDURAL BACKGROUND......................................................................................2 A. B. The Court Certified the Class Based on Professor Noll's Report............................2 Apple Thwarted and Continues to Thwart Discovery .............................................3

MOTIONS TO DECERTIFY ARE IMPROPER WHERE NO NEW FACTS OR LAW EXIST ........................................................................................................................5 PROFESSOR NOLL REVIEWED INFORMATION SUFFICIENT TO CONCLUDE THAT AT LEAST ONE OF HIS PROPOSED DAMAGES METHODOLOGIES WOULD WORK ..............................................................................6 DR. BURTIS'S REPORT IS SO INSUBSTANTIAL, IT SHOULD BE GIVEN NO WEIGHT .......................................................................................................................7 APPLE FAILS TO IDENTIFY A SINGLE DEFECT IN PROFESSOR NOLL'S METHODOLOGIES THAT DEFEATS CLASS CERTIFICATION ................................8 A. B. Any Comparison of Professor Noll and Dr. French's Analyses Is Meaningless ...........................................................................................................11 Professor Noll's Methodologies ............................................................................12 1. Before-After Method .................................................................................13 a. Professor Noll Sufficiently Explained How He Will Account for Product Differences and Other Necessary Variables ........................................................................................14 "Coolness" Does Not Prevent Construction of an Effective Damages Model .............................................................................18 Apple's "Pricing Strategy" Does Not Preclude an Effective Damages Model .............................................................................19 Dr. Burtis Does Not Explain Why the Before Period Is Too Short...............................................................................................20

12 VI. 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 VII.

b. c. d. 2. 3.

Professor Noll Has Demonstrated that the Mark-Up Method Can Be Used in This Case.................................................................................20 Professor Noll Has Demonstrated that the Yardstick Method Can Be Used in This Case.................................................................................22

THE COURT ALREADY CONSIDERED AND REJECTED APPLE'S ARGUMENT THAT A POSSIBLE UNDERCHARGE OF ITS FILES MUST BE CONSIDERED IN THE DAMAGES MODEL ................................................................23 -i-

PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page3 of 34

1 2 3 VIII. 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS- C-05-00037-JW(RS)

Page CONCLUSION..................................................................................................................25

- ii -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page4 of 34

1 2 3 CASES

TABLE OF AUTHORITIES Page

4 Freeland v. AT&T Corp., 238 F.R.D. 130 (S.D.N.Y. 2006) .......................................................................................15 5 Gonzales v. Arrow Fin. Servs. LLC, 6 489 F. Supp. 2d 1140 (S.D. Cal. 2007)................................................................................5 7 Heffelfinger v. Elec. Data Sys. Corp., 580 F. Supp. 2d 933 (C.D. Cal. 2008) ...............................................................................23 8 In re Carbon Black Antitrust Litig., 9 No. Civ. A. 03-10191-DPW, MDL No. 1543 2005 WL 102966 (D. Mass. Jan. 18, 2005) ........................................................................................10, 14, 23 10 In re Citric Acid Antitrust Litig., Nos. 95-1092, C-95-2963 FMS, 1996 WL 655791 11 (N.D. Cal. Oct. 2, 1996)...............................................................................................13, 14 12 In re Dynamic Random Access Memory (DRAM) Antitrust Litig., No. M 02-1486 PJH, 2006 WL 1530166 13 (N.D. Cal. June 5, 2006) ............................................................................................ passim 14 In re Foundry Resins Antitrust Litig., 242 F.R.D. 393 (S.D. Ohio 2007) ......................................................................................10 15 In re Linerboard Antitrust Litig, 16 203 F.R.D. 197 (E.D. Pa. 2001).........................................................................................10 17 In re Live Concert Antitrust Litig., 247 F.R.D. 98 (C.D. Cal. 2007) .............................................................................13, 15, 22 18 In re Polyester Staple Antitrust Litig., MDL No. 3:03 CV 1516, 2007 WL 2111380 19 (W.D.N.C. July 19, 2007) .................................................................................................15 20 In re Rubber Chems. Antitrust Litig, 232 F.R.D. 346 (N.D. Cal. 2005).................................................................................10, 13 21 In re Static Random Access (SRAM) Antitrust Litig., 22 No. C 07-01819 CW, 2008 WL 4447592 (N.D. Cal. Sept. 29, 2008) .................................................................................................10 23 In re Vitamins Antitrust Litig., 24 209 F.R.D. 251 (D.D.C. 2002)...........................................................................................14 25 In re Wellbutrin SR Direct Purchaser Antitrust Litig., No. 04-5525, 2008 WL 1946848 (E.D. Pa. May 2, 2008)...........................................23, 24 26 J. Truett Payne Co., Inc. v. Chrysler Motors Corp., 27 451 U.S. 557, 101 S. Ct. 1923 (1981)................................................................................11 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

- iii -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page5 of 34

1 2 Page

3 Los Angeles Mem'l Coliseum Comm'n v. NFL, 791 F.2d 1356 (9th Cir. 1986) ...........................................................................................10 4 Meijer, Inc. v. Abbott Labs., 5 No. C 07-5985 CW, 2008 WL 4065839 (N.D. Cal. Aug. 27, 2008) .................................10 6 Royal Printing Co. v. Kimberly-Clark Corp., 621 F.2d 323 (9th Cir. 1980) .............................................................................................24 7 Seigel v. Chicken Delight, Inc., 8 448 F.2d 43 (9th Cir. 1971) ...............................................................................................24 9 Slaven v. BP Am., Inc., 190 F.R.D. 649 (C.D. Cal. 2000) .........................................................................................5 10 Somers v. Apple, Inc., 11 No. 07-CV-06507 JW (N.D. Cal.) ............................................................................. passim 12 Sun Microsystems Inc. v. Hynix Semiconductor, Inc., 608 F. Supp. 2d 1166 (N.D. Cal. 2009) .......................................................................15, 18 13 STATUTES, RULES AND REGULATIONS 14 Federal Rule of Civil Procedure 15 Rule 23(b)(3)..............................................................................................................1, 3, 25 16 SECONDARY AUTHORITIES 17 Daniel L. Rubinfeld, Quantitative Methods in Antitrust, 1 Issues in Competition Law And Policy 723 (ABA Section of Antitrust Law) (2008) .................................................15 18 19 20 21 22 23 24 25 26 27 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS- C-05-00037-JW(RS)

- iv -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page6 of 34

1

Plaintiffs Somtai Troy Charoensak, Mariana Rosen, and Melanie Tucker (collectively,

2 "plaintiffs") respectfully submit this memorandum in opposition to defendant's ("Apple") Motion 3 for Decertification of Rule 23(b)(3) Class ("Def's Motion"). 4 I. 5 INTRODUCTION After full briefing and oral argument, this Court certified a Rule 23(b)(3) Class on plaintiffs'

6 monopolization, attempted monopolization and related state law claims in December 2008. 7 Dkt. No. 196 ("Class Certification Order"). Apple now wants another bite at the apple. Citing no 8 new law or facts, and relying largely on a retread report by an expert it had retained prior to the 9 original class certification briefing, but chose not to use, Apple moves to decertify the Class. 10 In its Class Certification Order, the Court identified numerous common issues of law and fact

11 that supported class certification, including market definition, market power, anticompetitive 12 conduct, harm to competition, Apple's willful acquisition and maintenance of monopoly power, and 13 anticompetitive injury. Class Certification Order at 6-8. Apple's motion challenges none of these 14 findings. 15 Apple moves to decertify the class based solely on the report of its expert, Dr. Michele M.

16 Burtis. Dkt. No. 241. Dr. Burtis belatedly challenges the sufficiency of the 60-page declaration 17 submitted by plaintiffs' expert, Stanford Professor Emeritus Roger G. Noll, in support of class 18 certification, and only with regard to his proposed methodologies to show classwide damages. In 19 addition to his detailed initial report, Professor Noll has also submitted a 56-page reply declaration,1 20 filed concurrently herewith, that provides even more detail as to why he believes that his proposed 21 methodologies can be utilized to show classwide damages. 22 As set forth in more detail below and in Professor Noll's reply declaration, Dr. Burtis

23 attempts to re-write Professor Noll's thorough, detailed analysis in order to compare it to the analysis 24 of a completely different expert, Dr. Gary L. French, in the indirect purchaser case, Somers v. Apple, 25 26 27 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)
1

See Reply Declaration of Roger G. Noll, filed concurrently ("Noll Reply Decl.").

-1-

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page7 of 34

1 Inc., No. 07-CV-06507 JW (N.D. Cal.). The two reports bear only minimal resemblance, and the 2 Court's concerns with Dr. French simply do not apply to Professor Noll's analysis. 3 Further, Apple seeks to use its own refusal to produce anything other than exemplar data

4 prior to the original class certification briefing, and delay in producing merits-based discovery now, 5 to support decertification. Its "gotcha" litigation tactics are inappropriate and certainly do not meet 6 its burden to overturn the Court's original class certification decision. 7 Finally, when the reports of Dr. Burtis and Professor Noll are compared substantively,

8 Professor Noll's is far more sophisticated and reasoned. In forming his opinions, Professor Noll has 9 relied on data both from Apple and from publicly-available sources ­ scholarly articles, economic 10 treatises, analyst reports, etc. ­ in addition to his years of experience and intimate knowledge of the 11 industry and expertise in formulating damages models in antitrust cases. In sharp contrast, Dr. 12 Burtis testified that she examined none of the sources specifically cited by Professor Noll. She just 13 looked at the same things she looked at to prepare her report in rebutting Dr. French in Somers. Her 14 opinions bear out this paucity of preparation. Dr. Burtis, for example, is of the opinion that even if 15 plaintiffs make a clear showing that Apple has monopolized or attempted to monopolize the market, 16 she does not believe anyone ­ either on a class or individual basis ­ could prepare a damages model 17 to hold Apple accountable. 18 Apple's limited argument attacking Professor Noll's proposed damages methodologies

19 cannot satisfy its burden to overcome the many common issues of law and fact that the Court 20 identified in its original Class Certification Order. Apple's motion should be denied. 21 II. 22 23 PROCEDURAL BACKGROUND A. The Court Certified the Class Based on Professor Noll's Report

On July 21, 2008, plaintiffs filed their motion seeking certification of a class of direct

24 purchasers of iPods, supported by the detailed 60-page expert declaration of Professor Noll. Dkt. 25 Nos. 165, 166-1, (Declaration of Roger G. Noll, filed July 21, 2008). Relying on exemplar cost and 26 revenue data produced by Apple, numerous analyst reports, public information on iPod 27 characteristics and pricing, his extensive experience in technology markets, and the pleadings and 28 depositions in this litigation, Professor Noll opined that three widely-accepted methods were
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

-2-

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page8 of 34

1 available to calculate damages to the class, and that each of those methods relied on evidence 2 common to the class: (1) before-after method; (2) yardstick method; and (3) mark-up method. Dkt. 3 No. 166-1 at 52-59. 4 On September 8, 2008, Apple retained its economic expert, Dr. Michele M. Burtis. On

5 September 19, 2008, Apple conducted a day long deposition of Professor Noll. See Declaration of 6 Thomas R. Merrick in Support of Plaintiffs' Memorandum In Support of Opposition to Defendant's 7 Motion for Decertification of Rule 23(b)(3) Class, filed concurrently ("Merrick Decl."), Ex. 1, 8 Deposition of Roger G. Noll, taken September 19, 2008. During the deposition, Professor Noll 9 responded to extensive questioning by Apple concerning his report and specifically about each 10 methodology he proposed, including the possible uses of regression analysis. See, e.g., id., Ex. 1 at 11 68:13-75:24. 12 Apple filed its opposition to plaintiffs' motion for class certification three months after

13 plaintiffs served their motion, on October 17, 2008. Although it retained Dr. Burtis over a month 14 earlier, Apple submitted no expert report to rebut Professor Noll. Dkt. No. 181. After briefing and 15 oral argument, on December 22, 2008, the Court certified a class of direct purchasers of iPods for 16 monopolization and attempted monopolization of the portable digital audio player market.2 Class 17 Certification Order. 18 19 B. Apple Thwarted and Continues to Thwart Discovery

Prior to class certification, at Apple's request and over plaintiffs' objections, the Court

20 sharply limited the scope of discovery. Even as to this limited discovery, Apple objected to each of 21 plaintiffs' discovery requests on the grounds that it was not relevant to class certification. See Dkt. 22 No. 261, Declaration of Bonny E. Sweeney in Support of Plaintiffs' Opposition to Apple's 23 Administrative Motion to Set Briefing Schedule for Decertification Motion, ¶¶2-3. Plaintiffs were 24 forced to file a motion to compel seeking financial information, including profit and loss statements, 25 revenue, cost and sales data. Dkt. No. 137. 26
2 Court has requested the scope of 27 hearingThethat issue will be heldfurther briefing on this motion. the injunctive relief class and a on concurrently with 28

PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

-3-

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page9 of 34

1

Apple took the position that plaintiffs were not entitled to production of this financial data

2 because, according to Apple, such information was not relevant to class certification. Dkt. No. 261, 3 ¶¶5, 7. Just prior to the hearing on the motion to compel, Apple proposed a compromise. 4 Ultimately, plaintiffs agreed that Apple could produce exemplars of the requested data, as opposed 5 to all of the data, so that Professor Noll could make a determination whether the data, when it was 6 eventually produced, would suffice. Plaintiffs agreed because Apple's counsel, orally and in writing, 7 took the position that, at the class certification stage, plaintiffs are not required to have completed 8 their damages study, but rather are required to show that they can rely on common evidence to show 9 classwide impact. See In re Dynamic Random Access Memory (DRAM) Antitrust Litig., No. M 0210 1486 PJH, 2006 WL 1530166, at *8-*9 (N.D. Cal. June 5, 2006).3 As Apple's counsel said: 11 12 13 To be clear, I did not agree to, or anticipate, that we would produce the actual data beyond an exemplar of the type of data that are available. I understood that your expert wanted to know what type of data is available rather than acquiring all the data now because he does not intend to actually produce a damage study at this point. That's the compromise we reached, and I thought that met your pre-cert needs.

14 Dkt. No. 261-3. 15 After the Court lifted the partial discovery stay following class certification, on May 22,

16 2009, plaintiffs served amended discovery requests seeking, inter alia, the complete data as to which 17 Apple had produced only exemplars, profit and loss information, information regarding the markets 18 for iPods and digital audio and video files, and various aspects of Apple's DRM and its impact on 19 interoperability of iPods and iTS with competitors' products. Apple, in contrast, has conducted no 20 new discovery following certification. The parties have met and conferred several times over the 21 past four and a half months regarding Apple's objections and the scope of plaintiffs' requests. To 22 date, plaintiffs have received only quarterly iPod and iTS sales figures, some limited data on cost 23 factors for iPods and iTS and some documents related to Apple's dispute with Real Networks over 24 25 26 27 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)
3

Citations are omitted and emphasis is added, here and throughout, unless otherwise noted.

-4-

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page10 of 34

1 its Harmony product. Plaintiffs continue to meet and confer in good faith, but believe another 2 motion to compel will be required.4 3 Despite the lack of any new evidence or law, and only eight months after the Court's Class

4 Certification Order, Apple now moves to decertify the Class. Apparently seeking to capitalize on its 5 relative success using Dr. Burtis in the indirect purchaser Somers action, Apple belatedly submits the 6 report of Dr. Burtis here. Dkt. No. 241. As set forth more fully below, Dr. Burtis's declaration is 7 based on rebutting a different expert (Dr. French) in a different case (Somers) based on a different 8 damages standard (indirect purchasers' pass-through damages). 9 III. 10 As a threshold matter, Apple's motion does not meet the Ninth Circuit standards for 11 decertification. In motions to decertify a class, the defendant must meet a "heavy burden" in 12 demonstrating that decertification is appropriate. Gonzales v. Arrow Fin. Servs. LLC, 489 F. Supp. 13 2d 1140, 1154 (S.D. Cal. 2007); see also Slaven v. BP Am., Inc., 190 F.R.D. 649, 651 (C.D. Cal. 14 2000) ("Defendants' burden in urging decertification is relatively heavy."). 15 decertification, Apple must raise some new controlling law or facts to support its argument that the 16 17 Apple implicitly recognized the sufficiency of Professor Noll's proposed damages 18 methodology when it opposed plaintiffs' motion for class certification twelve months ago. Apple 19 retained Dr. Burtis approximately two weeks prior to Professor Noll's deposition in 2008 (Merrick 20 Decl., Ex. 2 at 147:24-148:7, Deposition of Michelle M. Burtis, Ph.D., taken September 30, 2009) to 21 assist Apple's counsel in preparing for the deposition. She read Professor Noll's declaration and 22 discussed his proposed methodologies with Apple's counsel. Id , Ex. 2 at 77:22-78:4, 78:7-11, 23 24 Of course, plaintiffs' discovery does not focus solely on the information plaintiffs will ultimately need to generate their damages report. But Apple's repeated delays and refusals to 25 produce are representative of its entire approach to discovery. 26 In their Opposition to Apple's Administrative Motion to Set Briefing Schedule for Decertification Motion, plaintiffs discussed why Apple's motion was improper. Dkt. No. 260 at 227 3. Rather than repeat them here, plaintiffs incorporate by reference the arguments made therein. 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)
5 4

MOTIONS TO DECERTIFY ARE IMPROPER WHERE NO NEW FACTS OR LAW EXIST

To justify

initial class determination was in error.5 Apple cannot do so. Apple simply wants a "do-over."

-5-

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page11 of 34

1 79:10-15, 79:24-80:12. After the deposition, she read the transcript and again spoke with Apple's 2 counsel. Id., Ex. 2 at 151:5-9. Still, Apple made the tactical decision not to submit a report to rebut 3 Professor Noll. Id., Ex. 2 at 151:10-21. 4 Apple's sole basis for this motion is the Court's ruling on a motion for class certification

5 brought by indirect purchasers in Somers using an entirely different expert, Dr. Gary French. See 6 Def's Motion at 7. Apple points to no new facts or law to support its argument that the Court did not 7 properly consider and evaluate these issues in its original class certification decision. The Court's 8 order denying certification in Somers did not raise any new facts or law relevant to this action, but 9 instead denied certification of the indirect purchaser class on the basis that Dr. French did not 10 proffer a reliable method of proving impact and damages on a classwide basis. Somers Dkt. No. 80 11 at 12. 12 In the earlier class certification briefing, argument and the Court's Order, neither the Court

13 nor Apple ever raised concerns regarding Professor Noll's methodologies. Professor Noll has 14 reviewed all the data available to him, has significant knowledge and experience in technology 15 markets, and has provided substantial detail on his proposed methodologies. See Dkt. No. 166-1 at 16 52-59. By contrast, Dr. Burtis's expert report contains little more than excerpts from her report 17 submitted in Somers to rebut Dr. French, and is devoid of any analysis or data. Apple makes no 18 independent assessment of Professor Noll's methodologies and raises no new facts or law. 19 IV. 20 21 PROFESSOR NOLL REVIEWED INFORMATION SUFFICIENT TO CONCLUDE THAT AT LEAST ONE OF HIS PROPOSED DAMAGES METHODOLOGIES WOULD WORK Apple's criticism that Professor Noll has "collected little if any data" (Def's Motion at 7), is

22 disingenuous given Apple's insistence that pre-certification discovery be limited, and its insistence 23 that revenue, cost, profit and loss and sales data are not relevant to class certification, eventually 24 producing only "exemplar" data for Professor Noll's use in determining whether the mark-up 25 method would work. Besides being disingenuous, Apple's argument is incorrect, because Professor 26 Noll reviewed extensive pricing and product characteristics data that was publicly available. Dkt. 27 No. 166-1 at 4-6 (citing scholarly papers, analyst reports, trade press reports, public information on 28 features and pricing of iPods, court submissions in this litigation, and four decades of experience in
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

-6-

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page12 of 34

1 studying the entertainment and information technology industries).6 All of this information, like the 2 revenue, cost and sales data, is common to the Class. Combining this information and his 3 experience, Professor Noll concluded that three methods were potentially available to calculate 4 damages. Id. at 5-6. 5 Moreover, Dr. Burtis has no basis for her assertion that Professor Noll "has not collected any

6 data or shown that data exist to implement these methods." She never reviewed any of the exemplar 7 data, product feature or pricing data relied on by Professor Noll. During her deposition, Dr. Burtis 8 stated that she only received and reviewed the data listed in Exhibit 2 to her report. Merrick Decl., 9 Ex. 2 at 13:23-14:2, 100:1-13. This was the same list of documents she reviewed in Somers. 10 Notably missing from this list is the exemplar revenue, cost and sales data produced by Apple, 11 hundreds of analyst reports, Apple quarterly earnings calls from 2002 through 2007, scholarly 12 articles, and several other trade press reports that were all reviewed and considered by Professor 13 Noll. Dkt. No. 166-1 at 5-6. Accordingly, Dr. Burtis cannot make a genuine assessment of the 14 sufficiency of the bases of Professor Noll's opinions. 15 V. 16 Dr. Burtis's critique of Professor Noll's Report should be given no weight. Besides the vast 17 disparity in the experience and accomplishments of Dr. Burtis and Professor Noll (see Dkt. No. 24118 2, Ex. 1 (Dr. Burtis CV); Cf. to Noll Reply Decl., Ex. 1 (Professor Noll CV), Dr. Burtis conducted 19 almost no independent analysis, as demonstrated by the following examples: 20 First, Dr. Burtis merely cribbed her critique of Dr. French, ignoring several critical 21 differences between the two experts' approaches. She did not recognize, for example, that Professor 22 Noll's methodologies do not rely on averages (unlike Dr. French), that Dr. French did not propose 23 relying upon a yardstick or mark-up method (unlike Professor Noll), or that Professor Noll did 24 25 DRAM, 2006 WL 1530166, at *8 (accepting Noll's antitrust impact and proposed damages analysis for class certification based on "his analysis and conclusions on actual market share 26 estimates, review of contracts entered into between defendants and various DRAM purchasers, and trade publications reflecting DRAM pricing information, and actual sales and price data 27 industryproduced in discovery"). thus far 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)
6

DR. BURTIS'S REPORT IS SO INSUBSTANTIAL, IT SHOULD BE GIVEN NO WEIGHT

-7-

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page13 of 34

1 identify a beginning date for the period of anticompetitive conduct (unlike Dr. French). At her 2 deposition, Dr. Burtis could not identify even a single difference between Dr. French and Professor 3 Noll, even though Dr. French was retained by indirect purchasers. Merrick Decl., Ex. 2 at 84:64 88: 21. 5 Second, Burtis revealed at her deposition how little effort went into her report. She did not

6 write it herself, could not estimate how much time she spent on it, did no data analysis herself, and 7 did not even inquire into the kinds of data Apple maintains about iPods. Id., Ex. 2 at 12:21-16:1, 8 24:2-13, 97:9-98:7. 9 Third, although she claimed to be aware that the certified direct purchaser class includes

10 resellers (unlike the indirect purchaser class), her report and testimony reflected a complete absence 11 of any thought about how the damages methodologies applied to them. When asked whether she 12 thought the "net overcharge" adjustment had to be applied in the case of resellers that do not buy 13 iTunes files, she testified: "It's an interesting question. Clearly, Best Buy doesn't purchase music, 14 at least as far as I know. But I don't know. Maybe they do. I don't know the answer to that. I'm 15 not sure. I would have to think about that." Id., Ex. 2 at 44:6-45:12. 16 Fourth, Dr. Burtis displayed a palpable discomfort when asked when she first concluded that

17 Professor Noll's methodologies could not work to calculate damages. Id., Ex. 2 at 76:22-81:17. 18 Apple and its expert obviously initially viewed Professor Noll's proposed damages methodologies as 19 unassailable, or they would have attacked them in their opposition last year. Following this Court's 20 ruling in Somers, however, Apple and Dr. Burtis had a sudden (and unjustified) change of heart. 21 VI. 22 As Professor Noll's 56-page reply declaration demonstrates, Apple's and Dr. Burtis's 23 challenge is without support. Professor Noll addresses in detail each of Dr. Burtis's criticisms of his 24 report. He makes clear that Dr. Burtis fails in her attempt to simplify his analysis so that it fits 25 within Dr. French's report in the indirect purchaser case. The two reports bear only the most 26 superficial of resemblance to each other. Noll Reply Decl. at 11-17. He also provides further detail 27 as to why the three methodologies he originally proposed ­ the before-after, mark-up and yardstick 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

APPLE FAILS TO IDENTIFY A SINGLE DEFECT IN PROFESSOR NOLL'S METHODOLOGIES THAT DEFEATS CLASS CERTIFICATION

-8-

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page14 of 34

1 methods ­ can be utilized, based on evidence common to the Class. In short, Professor Noll rebuts 2 Dr. Burtis's criticisms and shows why he stands behind the conclusions he reached in his report, 3 which the Court found to be sufficient to support class certification. Id. at 2 ("My overall conclusion 4 is that neither the Burtis Report nor any other material I have read causes me to change any of the 5 opinions that I expressed in the Noll Report."). 6 Apple and Dr. Burtis's critique of Professor Noll's declaration is narrowly limited to whether

7 he has proffered a reliable methodology for calculating damages. Def's Motion at 1-2; Dkt. No. 241 8 at 3. Neither Apple nor Dr. Burtis challenge Professor Noll's conclusions that economic analysis 9 relying on common evidence could be used to establish market definition, market power, 10 anticompetitive conduct and harm to competition. Apple's decision not to challenge Professor Noll 11 on liability issues is especially important because damages and liability are closely related. Noll 12 Reply Decl. at 9-10. One can use the same methods to show both that Apple's exclusionary conduct 13 increased market power and thereby caused harm to competition, and that the degree of price 14 discrimination increased when Apple began acting anticompetitively. Id. at 10. 15 In its Class Certification Order, the Court identified numerous common issues of law and fact

16 that supported class certification, including market definition, market power, anticompetitive 17 conduct, harm to competition, Apple's willful acquisition and maintenance of monopoly power, and 18 anticompetitive injury. See (Class Certification Order); Noll Reply Decl. at 6-8. Plaintiffs 19 incorporate by reference their argument and authority contained in their motion for class certification 20 (Dkt. No. 165), and reply memorandum in support of class certification (Dkt. No. 188). Apple's 21 motion challenges none of these findings. Nor do they challenge Professor Noll's opinion that all 22 damages models will use evidence common to the Class. Rather, Apple and Dr. Burtis merely assert 23 that damages cannot be calculated, whether for an individual or the class, because there is no 24 appropriate benchmark to which iPod prices can be compared. Def's Motion at 7-9; Dkt. No. 241 at 25 3-4; Merrick Decl., Ex. 2 at 30:21-31:16, 35:6-36:1. 26 Professor Noll proposes three potential methodologies for calculating damages: (1) before-

27 after; (2) yardstick; and (3) mark-up. Dkt. No. 166-1 at 54-59; Noll Reply Decl. at 27-56. All three 28 are valid, widely-recognized economic methods accepted by courts in antitrust cases. Noll Reply
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

-9-

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page15 of 34

1 Decl. at 12-14. See DRAM, 2006 WL 1530166, at *10 (finding Professor Noll's before-after, 2 yardstick, and mark-up methods were valid means of proving damages on class-wide basis); In re 3 Static Random Access (SRAM) Antitrust Litig., No. C 07-01819 CW, 2008 WL 4447592, at *6-*7 4 (N.D. Cal. Sept. 29, 2008) (same). At this stage, plaintiffs are not required to select which method 5 they will ultimately use "as long as they offer a methodology that is generally accepted." In re 6 Carbon Black Antitrust Litig., No. Civ. A. 03-10191-DPW, MDL No. 1543 2005 WL 102966, at *20 7 (D. Mass. Jan. 18, 2005) (quoting In re Linerboard Antitrust Litig, 203 F.R.D. 197, 217-18 (E.D. Pa. 8 2001)). Nor would it be appropriate without the necessary data. See Merrick Decl., Ex. 2 at 69:109 70:9. As Professor Noll explains in his reply declaration, determining precisely what methodology 10 and equations will be used before the data have been produced would be irresponsible. Noll Reply 11 Decl. at 29. Plaintiffs have more than satisfied their burden of coming forward with "seemingly 12 realistic" methodologies for calculating damages on a classwide basis. DRAM, 2006 WL 1530166, 13 at *8; see also In re Rubber Chems. Antitrust Litig, 232 F.R.D. 346, 354 (N.D. Cal. 2005) (at class 14 certification stage "the court's inquiry is limited to whether or not the proposed methods are so 15 insubstantial as to amount to no method at all").7 Apple, which bears the burden of proof on this 16 motion, has made no showing otherwise. 17 In addition, plaintiffs in antitrust cases are not required to calculate damages with the level of

18 certainty required in some other areas. Los Angeles Mem'l Coliseum Comm'n v. NFL, 791 F.2d 19 1356, 1360 (9th Cir. 1986). This is because "[t]he vagaries of the marketplace usually deny us sure 20 knowledge of what plaintiff's situation would have been in the absence of the defendant's antitrust 21
7 At best, the attack on Professor Noll's merits-based and not 22 appropriate at the class certification stage. See proposed methodologies areat *6 ("The validity of SRAM, 2008 WL 4447592, be adjudicated at trial based upon economic theory, data sources, and 23 [the proposed] methods `will entirely common to the class.'"); see also Meijer, Inc. v. Abbott Labs., statistical techniques that are

*10 (N.D. Cal. Aug. 27, 2008) (rejecting as a 24 No. C 07-5985 CW, 2008 WL 4065839, atplaintiff's proposed damages methodologies merits argument, defendants' contention that were inappropriate and required individualized evidence); DRAM, 2006 WL 1530166, at *10 (defendants' 25 argument that the "`variability'" of the DRAM market invalidated Dr. Noll's proposed damages decided at this 26 models was a merits argument that "need not be 400 (S.D. Ohio stage of the litigation."); In re Foundry Resins Antitrust Litig., 242 F.R.D. 393, 2007) (weaknesses in expert's proposed methodologies irrelevant at the class certification stage where the proposed methodologies 27 are reliable). 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

- 10 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page16 of 34

1 violation." J. Truett Payne Co., Inc. v. Chrysler Motors Corp., 451 U.S. 557, 566-67, 101 S. Ct. 2 1923 (1981). The wrongdoer whose anticompetitive conduct has distorted the marketplace thus 3 cannot "insist upon specific and certain proof of the injury which it has itself inflicted." Id. This 4 standard is especially relevant here, because Apple makes the unsupported ­ and unsupportable ­ 5 argument that no plaintiff, not even an individual plaintiff, can calculate damages in this case. See 6 Merrick Decl., Ex. 2 at 30:21-31:16. 7 8 As Professor Noll explains, given the detailed analysis he submitted and the differences 9 between a damages analysis for indirect and direct purchase classes, Dr. Burtis's comparisons of 10 Professor Noll and Dr. French are specious at best. Professor Noll and Dr. French's reports are the 11 same in only two respects: the two discuss the same three methods for calculating damages and both 12 propose the before-after method as a plausible method. This is not surprising given courts' routine 13 approval of these methods in antitrust cases to calculate a but-for price of a product. Noll Reply 14 Decl. at 11-14. However, this is the extent of their similarity. 15 Professor Noll and Dr. French's analyses differ in critical ways. First, because this is a direct 16 purchaser action, Professor Noll's analysis is focused on the overcharge direct purchasers paid for 17 iPods. Dkt. No. 166-1 at 52; Noll Reply Decl. at 14. By contrast, Dr. French's proposed methods 18 for calculating damages necessarily included both the overcharge to direct purchasers and pass19 through to indirect purchasers. See Somers Dkt. No. 42-1, Affidavit of Gary L. French, ¶64. 20 Second, Professor Noll provides substantially more analysis on estimating overcharges to 21 direct purchasers. His discussion is twice as long as Dr. French's discussion of both overcharges and 22 pass-through put together. Dkt. No. 166-1 at 52-59; Cf. Somers Dkt. No. 42-1, ¶¶64-72. His 5623 page reply declaration provides even more thorough analysis. The Court's concerns about the 24 vagueness of Dr. French's analysis simply do not apply to Professor Noll's detailed reports. See 25 Somers Dkt. No. 80 at 12 (Dr. French "has done nothing more than make a vague five-paragraph 26 long collection of proposals for accomplishing what the Court sees as a daunting task."). 27 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

A.

Any Comparison of Professor Noll and Dr. French's Analyses Is Meaningless

- 11 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page17 of 34

1

Third, unlike Dr. French, Professor Noll actually proposes and discusses yardstick and mark-

2 up as possible methods of calculating damages and collected exemplar data to confirm that these 3 methods would work. Dkt. No. 161-1 at 56-59; Noll Reply Decl. at 46-56. As Apple points out, Dr. 4 French provides one-line explanations of the general approach of the yardstick and mark-up methods 5 (see Somers Dkt. No. 42-1, ¶65), but only proposes using the before-after method. Def's Motion at 6 9; Somers Dkt. No. 42-1, ¶¶66-68. This is why Dr. Burtis's critique of Dr. French in the indirect 7 case did not address these methods. Somers Dkt. No. 74 at 8-15. Thus, Dr. Burtis's claim that 8 "Professor Noll's proposed methods suffer from the same basic flaws this Court found in Dr. 9 French's three methods." (Dkt. No. 241 at 3; see also Def's Motion at 9) is patently false. Neither 10 Dr. Burtis nor the Court found any "flaws" in Dr. French's yardstick and mark-up methods because 11 Dr. French did not propose them. 12 Fourth, Professor Noll's proposed methodologies are based on actual transaction data and list

13 prices for each iPod model, not averages. Noll Reply Decl. at 17; Dkt. No. 161-1 at 53. By contrast, 14 Dr. French proposed aggregating individual transaction prices of retailers and distributors for months 15 and models of iPods to calculate the amount of pass-through to indirect purchasers. Somers Dkt. No. 16 42-1, ¶67. None of Professor Noll's methods are based on average prices. 17 Finally, and as discussed more fully below, unlike Dr. French, Professor Noll identifies cost

18 and demand variables, such as technological changes in iPods, and identifies the date when the 19 period of anticompetitive harm and damages began. Dkt. No. 161-1 at 49, 55. 20 Accordingly, Dr. Burtis's attempt to lump Professor Noll into the same category as Dr.

21 French in an effort to discredit Professor Noll's opinions is entirely without merit. 22 23 B. Professor Noll's Methodologies

In contending that his proposed methods will not work, Dr. Burtis and Apple state that

24 Professor Noll did not: (1) develop any actual models, propose any equations, or identify variables to 25 be used in calculating damages; (2) collect the necessary data; or (3) propose ways to overcome 26 obstacles to implementing his methods. Def's Motion at 7-9; Dkt. No. 241 at 3-4. Besides ignoring 27 the impact of Apple's refusal to produce data, and the fact that plaintiff is not required to produce a 28 completed damages study at the class certification stage, Apple's arguments are based on a flawed
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

- 12 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page18 of 34

1 analysis and a distortion of Noll's Report. As demonstrated below, Apple's arguments fail as to all 2 three of Professor Noll's proposed damages methodologies. 3 4 1. Before-After Method

The before-after method of calculating damages compares the price of products in the

5 "during" period, the period when the anticompetitive conduct affected the price of the products, with 6 prices of products in the periods when the anticompetitive conduct had no effect. Dkt. No. 161-1 at 7 55-56; Noll Reply Decl. at 27-28. This is a recognized method of proving impact and damages in 8 antitrust cases. See, e.g., DRAM, 2006 WL 1530166, at *10; Rubber Chems., 232 F.R.D. at 353; In 9 re Citric Acid Antitrust Litig., Nos. 95-1092, C-95-2963 FMS, 1996 WL 655791, at *7 (N.D. Cal. 10 Oct. 2, 1996).8 Where the product at issue is available in different models or with different 11 characteristics, a regression analysis is commonly implemented. Dkt. No. 161-1at 55; Noll Reply 12 Decl. at 27-28; In re Live Concert Antitrust Litig., 247 F.R.D. 98, 145 (C.D. Cal. 2007). 13 Here, Professor Noll proposes comparing the price of iPods in the during period with prices

14 of iPods when Apple's anticompetitive conduct had no effect. Dkt. No. 161-1 at 55. This will 15 include a regression analysis to account for technological advances in product features of iPods over 16 time as well as other cost and demand factors. Id.; Noll Reply Decl. at 29-30. 17 Dr. Burtis contends Professor Noll's before-after method of proving damages is flawed for

18 three reasons: (1) Professor Noll fails to take into account that there were some models of iPods sold 19 in the before period that were not sold in the during period and vice versa; (2) Professor Noll omits 20 certain supply and demand variables that are explanatory of the price of iPods;9 and (3) there is 21 insufficient pricing data in the before period to conduct the analysis. Def's Motion at 7-8; Dkt. No. 22 241 at 5-10; Merrick Decl., Ex. 2 at 30:21-31:16. 23 24 25 Dr. Burtis herself acknowledges that such analyses are routinely conducted. See Merrick Decl., Ex. 2 at 62:20-65:11, 153:14-154:8, 187:2-24. 26 9 27 Decl. atProfessor Noll assumes that by "supply" Dr. Burtis is referring to cost variables. Noll Reply 37. 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)
8

- 13 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page19 of 34

1 2 3

a.

Professor Noll Sufficiently Explained How He Will Account for Product Differences and Other Necessary Variables

The fact that multiple models of iPods exist does not make the before-after model

4 unworkable. Indeed, courts routinely find the before-after method appropriate even where there are 5 hundreds of different product variations within a particular market. See, e.g., DRAM, 2006 WL 6 1530166, at *10 (in a case with hundreds of product variations, the court rejected defendants' 7 argument at the class certification stage that Professor Noll's application of the before-after method 8 was unworkable because of the "`variability' present in the DRAM market"); In re Vitamins 9 Antitrust Litig., 209 F.R.D. 251, 268 (D.D.C. 2002) (upholding method of establishing damage in the 10 "but-for" world despite fact that varied products existed); Citric Acid, 1996 WL 655791, at *6-*7 11 ("Diversity of products and pricing does not necessarily mean that plaintiffs cannot show class-wide 12 impact . . . ."). 13 The DRAM case is particularly instructive. There, the market at issue was much more

14 complex than here. It consisted of many more varieties of DRAM than there are Apple models. 15 DRAM, 2006 WL 1530166, at *1. There, as here, defendants challenged Professor Noll's testimony 16 in support of class certification. Id. at *8. "They contend that the complexity of the DRAM market, 17 and the diversity of DRAM products and prices present therein, makes common proof of impact 18 impossible." Id. Despite this very complex market, the court in DRAM held that Professor Noll's 19 three methodologies for showing impact and damages ­ the same three methodologies he proposes 20 employing here ­ were sufficient to support class certification. Id. at *9; see also Carbon Black, 21 2005 WL 102966, at *19-*20 (class certified where before-after method used to calculate damages 22 for varying products with varying prices). 23 Here, a before-after model will be used to determine the price of a single product ­ an iPod ­

24 in a significantly less complicated market. See Merrick Decl., Ex. 1 at 70:25-71:12. There are only 25 43 models of the iPod, all produced by Apple. Although "different models of iPod have different 26 functionality and have different prices. . . , [t]he differences among them and the changes in prices 27 through time . . . should in part be determined by their attributes." Id., Ex. 1 at 246:21-247:4; Dkt. 28 No. 161-1 at 55-56; Noll Reply Decl. at 19-20. This can be determined with the use of objective
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

- 14 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page20 of 34

1 criteria obtained either from Apple's internal records or public sources, all of which are common to 2 the Class. Dkt. No. 161-1 at 49; Noll Reply Decl. at 19-20, 37; Merrick Decl., Ex. 1 at 70:25-71:12. 3 Professor Noll proposes using a hedonic regression "[b]ased on the qualitative attributes of

4 the product, and the prices of other products, and other factors that would increase demand." 5 Merrick Decl., Ex. 1 at 81:23-82:4; Noll Reply Decl. at 39-40. "Hedonic regressions are widely 6 used in economics to analyze differentiated products, especially with frequent product innovation, 7 including markets for consumer electronics." Noll Reply Decl. at 39 and n.86 (citing examples).10 8 Professor Noll would start by analyzing all plausible factors that might be significant in

9 explaining price. Dkt. No. 161-1 at 55; Noll Reply Decl. at 28-30; Merrick Decl., Ex. 1 at 79:1910 81:16, 246:15-247:21. Professor Noll has identified product features such as video and photo 11 capabilities, internet access, memory, size, weight, battery life, color, as well as features available on 12 other products and their prices, the availability of digital downloads, and the life cycle of the iPod, as 13 possible cost and demand variables that may be determinative of price. Dkt. No. 161-1 at 55; Noll 14 Reply Decl. at 29; Merrick Decl., Ex. 1 at 82:13-83:4, 83:12-85:13, 97:10-102:2, 101:15-102:2, 15 112:23-113:8, 128:20-130:1; Merrick Decl., Ex. 2 at 135:8-12. 16 Dr. Burtis contends that because the iPod models sold in the before period have different

17 features than the models in the during period, it is impossible to conduct a before-after analysis. 18 Dkt. No. 241 at 7-8; Merrick Decl., Ex. 2 at 30:21-31:9, 31:17-32:15. This simplistic complaint 19 completely ignores the purpose of a regression analysis. See Noll Reply Decl. at 35, 36. A 20 See also Sun Microsystems Inc. v. Hynix Semiconductor, Inc., 608 F. Supp. 2d 1166, 1207-08 21 (N.D. Cal. 2009) (using multiple regression analyses in a before-after model to control for supply determinates of the DRAM); Merrick Ex. 3, Daniel L. Rubinfeld, 22 and demand Methods in Antitrust, price of in Competition LawDecl.,Policy 723, 724 (ABA Section Quantitative 1 Issues And reduced-form model might explain the variation in the price of 23 of Antitrust Law) (2008) ("A typicalvariables relating to cost, demand, and market structure."); Live a product as a function of a series of Concert, 247 F.R.D. 98 (using regression to estimate the concert tickets 24 while controlling for artist quality and otheranalysis affecting price); but-for price ofStaple Antitrust factors In re Polyester Litig., MDL 2007 WL 2111380, at *29 July 19, 2007) ("`In 25 estimating theNo. 3:03 CV 1516,regression analysis attempts to (W.D.N.C.ordinary market factors "but-for" price, the control for of the conspiracy, if any, on price . .'"); Freeland v. AT&T Corp., 26 so that the effect 149 n.15 (S.D.N.Y. 2006) (citing can be isolated . .Statistics, U.S. Dep't of Labor, 238 F.R.D. 130, Bureau of Labor Handbook of Methods, Ch. 14 at 5 (available at 27 BLS http://stats.bls.gov/opub/hom/pdf/homch14.pdf)). 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)
10

- 15 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page21 of 34

1 regression analysis, like the one proposed by Professor Noll, controls for cost and demand variables 2 that are determinative of price and change over time, including product features. Dkt. No. 161-1 at 3 55; Noll Reply Decl. at 13-14, 18, 28-30; Merrick Decl., Ex. 1 at 70:13-24, 80:12-83:5; Cf. Merrick 4 Decl., Ex. 2 at 10:13-11:2, 33:5-9, 60:15-23 (agreeing that it is possible to take into account changes 5 in products features). 6 Importantly, as Professor Noll explained, omitted variables are only significant if they are

7 positively correlated to the period of exclusivity. Noll Reply Decl. at 33; Merrick Decl., Ex. 1 at 8 81:6-16, 111:15-112:11; Merrick Decl., Ex. 3 at 726. It is incorrect to say, as Dr. Burtis does, all 9 factors that are correlated to iPod prices must be included. Dkt. No. 241 at 6. As Professor Noll 10 explained, "[i]t's not that [the variables] have a correlation to price. It's that they ­ in order for the 11 specification error of leaving those variables out to produce an inconsistent or biased estimate of the 12 effect of exclusivity, it has to be the case that the exclusivity period is correlated with this alternative 13 source." Merrick Decl., Ex. 1 at 112:5-11; Noll Reply Decl. at 33. 14 In fact, if the omitted factors are negatively correlated to the conduct, then the impact of the

15 anticompetitive conduct is underestimated. Merrick Decl., Ex. 1 at 81:6-16; Noll Reply Decl. at 34. 16 For example, omission of the "coolness" factor and/or Apple's "pricing strategy," are only possibly 17 important if they are correlated to the period of anticompetitive conduct. Dr. Burtis provides no 18 evidence that this is the case or that Professor Noll would leave these factors out of a regression 19 analysis if they were determinative of price.11 20 Dr. Burtis also claims that Professor Noll has failed to prove that data exists to measure the

21 variables needed to run a regression analysis. Dkt. No. 241 at 7. The appropriate variables can be 22 identified, and the effect therefrom measured, by analysis of Apple's own documents, including 23 Apple's product specifications. As Professor Noll notes, it is not plausible that "Apple does not 24 know or cannot produce in discovery the specifications for its own products." Noll Reply Decl. at 25 37. And "[e]ven if a tsunami has destroyed Apple's records, all is not lost," because plaintiffs can 26
11 As Professor Noll explains, even variables that are 27 regression using the "instrumental approach." Noll Reply not "observable" can be accounted for in a Decl. at 33-34 (explaining this approach). 28

PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

- 16 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page22 of 34

1 examine the 43 iPod models, consult industry trade sources that provide "teardown reports" of 2 electronics products components or consult product reviews with extensive information about 3 technical features. Id. Professor Noll has already reviewed public information on pricing and iPod 4 product attributes to confirm that such a regression could work. Dkt. No. 161-1 at 53; Noll Reply 5 Decl. at 37. 6 By contrast, Dr. Burtis has conducted no analysis of her own to support her contention that

7 such a regression is impossible. Merrick Decl., Ex. 2 at 12:21-13:4, 15:9-16:1, 26:2-5. Dr. Burtis's 8 theory that the before-after analysis will not work here because "you have characteristics of products 9 that exist in only one of the two periods" (id., Ex. 2 at 61-62) is, as Professor Noll explains, "simply 10 incorrect as a matter of economic methodology." Noll Reply Decl. at 38. In fact, Dr. Burtis's 11 argument makes sense only if the measure of anticompetitive conduct is nearly perfectly correlated 12 with the measure of product features. Id. This would only happen if an important product 13 characteristic was present when Apple's exclusionary conduct began and remained unchanged over 14 time. Id. at 39. As Professor Noll explained, he has examined iPod product features and is not 15 aware of any such characteristic. Id. ("In examining the features of the iPod that have been 16 introduced since April 2003, I am not aware of any that were introduced around that date and have 17 remained unchanged since that time. If there is not near-perfect overlap between the measure of an 18 important feature and the measure of the exclusionary conduct, Dr. Burtis's claim is incorrect.").12 19 Professor Noll has not yet written an equation specifying all the variables to be used in a

20 regression analysis because not all necessary data has been produced. Noll Reply Decl. at 18; 21 Merrick Decl., Ex. 1 at 96:24-97:9. Once full data is collected, those variables that are "statistically 22
12 Dr. effects Katrina on an 23 example of Burtis references her article on theprices inof Hurricaneworld. Dkt. oil refineries asDr. a before-after test that calculated the but-for No. 241 at 2. able to perform the test because she had sufficient data 24 Burtis states that she wasconverted into a hypothetical antitrust damages example,to forecast prices after Katrina. Id. When this before-after

by 25 regression would not fit the standard for a valid damages model as describedhas Dr. Burtis. Noll Reply Decl. at 40-43. Rather, suffering from the alleged defects Dr. Burtis found in the Dr. French and Professor Noll reports, the Katrina regression uses average prices, assumes all refineries 26 have the same output mix, assumes that relative quantities of refinery outputs are fixed over time and the same as before hurricane, and 27 among refineries, assumes that demand after the hurricane isrelated to weather. the at 43. uses seasons, as "a crude instrument," for omitted variables Id. 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

- 17 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page23 of 34

1 significant and quantitatively important in explaining price" will be identified. Noll Reply Decl. at 2 28-32. Even Dr. Burtis agrees with this basic approach. Merrick Decl., Ex. 2 at 69:10-70:9. 3 Finally, to the extent Dr. Burtis and Apple attack the variables Professor Noll will account for

4 in his regression analysis, this is a merits argument not appropriate for the class certification stage. 5 See Sun Microsystems, 608 F. Supp. 2d at 1208-09 ("[T]o the extent that defendants challenge the 6 accuracy or propriety of these variables, it is an issue that goes to the weight, rather than the 7 admissibility, of [the expert]'s testimony."). 8 9 Dr. Burtis claims that identification and quantification of the determinative factors in this 10 case is "unusually difficult" (Merrick Decl., Ex. 2 at 102:21-103:12, 115:13-116:2; Cf. id., Ex. 2 at 11 23:2-25:9), but she has not collected any data or conducted any analysis to support this conclusion. 12 Id., Ex. 2 at 12:21-13:12, 170:18-171:11, 183:20-184:6. Dr. Burtis's failure to do so is telling, 13 because as Apple's expert, she had access to any Apple data she requested. Id., Ex. 2 at 13:17-15:8. 14 Nor is she able to identify any plausible factors beyond those discussed by Professor Noll except for 15 16 Dr. Burtis opines that "coolness" is "an important determinant of prices" and omission of this 17 variable "would bias an estimate of the alleged overcharge." Dkt. No. 241 at 7; Merrick Decl., Ex. 2 18 at 183:20-184:2. Yet, Dr. Burtis has not done even the most cursory analysis to support this 19 conclusion. Id., Ex. 2 at 104:1-9; 184:3-6. As Professor Noll explains, "[g]iven that Apple products 20 are commonly referenced as cool and that Apple's CEO has been interviewed about the source of 21 coolness, the obvious next step is to learn through discovery the role of coolness in Apple's 22 marketing and product planning." Noll Reply Decl. at 24. Dr. Burtis did none of this. 23 Indeed, Professor Noll concluded from his own research that "there is no reason to treat 24 coolness seriously in a damage analysis," because "there is absolutely no reason to believe that the 25 26 Notably, Dr. Burtis admitted she had never used "coolness" in any other economic context. Merrick Decl., Ex. 2 at 103:20-25. This is a word commonly used by Apple throughout this 27 litigation in an effort to raise individual issues where none exist. See, e.g., Dkt. No. 200 at 15. 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)
13

b.

"Coolness" Does Not Prevent Construction of an Effective Damages Model

the iPod's "coolness."13 See Dkt. No. 241 at 7; Merrick Decl., Ex. 2 at 103:13-19; 106:21-107-24.

- 18 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page24 of 34

1 coolness of iPods has increased" since their introduction in 2001. Noll Reply Decl. at 24. For this 2 reason, there is "no danger that the effects of the alleged anticompetitive conduct are likely to be 3 confounded with an increase in the attachment of consumers to the product." Id. Moreover, unlike 4 Dr. Burtis, who merely asserted that she does not "know of any way to measure [coolness]" (Merrick 5 Decl., Ex. 2 at 198:7-15), Professor Noll articulated a method of measuring coolness, if needed, by 6 identifying market and academic research on coolness and what it means in the iPod market. Noll 7 Reply Decl. at 20-25. During his deposition, he explained that while "coolness" could not be 8 measured directly, it could be inferred "indirectly from the results." Merrick Decl., Ex. 1 at 87:6-15. 9 This would be done by measuring objective attributes that people find "cool" and thus, encourage 10 them to purchase an iPod. Id. His theory was confirmed by his market and academic research which 11 found that coolness is derived from the attributes of a product and not from unmeasurable success in 12 the marketplace. Noll Reply Decl. at 21-23. Such objective attributes can be measured and included 13 in a regression analysis.14 Merrick Decl., Ex. 1 at 86:6-89:10; see also Merrick Decl., Ex. 3 at 730 14 (describing the importance of accounting for "marketing variables" with branded products that may 15 affect demand). 16 In short, Dr. Burtis's contention that "coolness" renders a before-after methodology

17 unworkable has no basis. 18 19 With respect to Apple's "pricing strategy," Dr. Burtis contends Apple's pricing of iPods 20 "may be more complex" because there are relatively few price changes of iPods over time. Merrick 21 Decl., Ex. 2 at 106:21-107:9. Still, Dr. Burtis provides no data to support this contention and indeed 22 23 Merrick Decl., Ex. 2 at 14:21-24, 109:23-110:2, 112:10-113:10, 175:5-22. As Dr. Burtis admits, if 24 25 Indeed, the fact that people buy iPods for various reasons is not unique to iPods and thus does not pose a unique problem to this case. See Merrick Decl., Ex. 2 at 106:16-20. 26 15 If "pricing strategy," it is not 27 consumerApple uses this type ofDecl., Ex. 1 at 130:2-10. unique to Apple but is present with most electronics. Merrick 28
PLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)
14

c.

Apple's "Pricing Strategy" Does Not Preclude an Effective Damages Model

has not collected any data from Apple or done any analysis of data herself.15 Dkt. No. 241 at 7;

- 19 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page25 of 34

1 Apple's pricing strategy is a determinative factor of price, it would be identifiable by objective 2 information common to the class (Merrick Decl., Ex. 2 at 112:10-113:10, 172:14-19) and could be 3 accounted for in a regression analysis. See Dkt. No. 241 at 7 ("A model that includes demand and 4 supply variables will show that those variables do not influence Apple's prices, but without some 5 explanatory variable to capture Apple's strategy, the model will fail to explain prices at all."). Thus, 6 without full discovery a regression model could not be competently specified. Noll Reply Decl. at 7 29-31. 8 9 Again without support, Dr. Burtis makes the blanket conclusion that "[t]he period before the 10 alleged violation was too short . . . ." Dkt. No. 241 at 7-8. Although not entirely clear, Dr. Burtis 11 seems to contend that the period is too short because there is "limited price data available." Id. at 8; 12 Merrick Decl., Ex. 2 at 89:21-90:11, 152:3-19. However, Dr. Burtis has not collected any pricing 13 data and has done no analysis to determine if the available data is sufficient. Id., Ex. 2 at 14:21-15:8, 14 109:23-110:2. Nor was she able, at her deposition, to identify any antitrust case in which a before15 after methodology was used and the "before" period was longer than 18 months. 16 Professor Noll has demonstrated how changes in product features over time do not preclude 17 using the before-after method to calculate damages. To the extent Dr. Burtis is arguing that 18 Professor Noll does not identify when Apple's anticompetitive conduct began, he reiterates that the 19 class period began when Apple launched iTS and why this is a reasonable starting period for the 20 class. Noll Reply Decl. at 45. Further, in his reply declaration, Professor Noll addresses how 21 Apple's change in DRM policy in 2009 may impact the length of the after period. Id. at 44-46. 22 None of Apple's and Dr. Burtis's criticisms of the before-after method support decertification 23 of the Class. In fact, Professor Noll is confident that a before-after model can be implemented in this 24 case. Merrick Decl., Ex. 1 at 70:25-72:1. 25 2. 26 27 Professor Noll Has Demonstrated that the Mark-Up Method Can Be Used in This Case d. Dr. Burtis Does Not Explain Why the Before Period Is Too Short

Professor Noll is also confident that the mark-up method can be successfully utilized. Id.,

28 Ex. 1 at 73:10-19, 75:2-13. Dr. Burtis makes the conclusory and incorrect statement that the markPLAINTIFFS' MEMORANDUM IN OPPOSITION TO DEFENDANT'S MOTION FOR DECERTIFICATION OF RULE 23(b)(3) CLASS - C-05-00037-JW(RS)

- 20 -

Case5:05-cv-00037-JW Document269

Filed10/19/09 Page26 of 34

1 up and yardstick methods are "effectively the same." Dkt. No. 241 at 4, 12-13. In his reply 2 declaration, Professor Noll discusses in detail how and why the two approaches differ (Noll Reply 3 Decl. at 46-49) and how his two proposed versions of the mark-up method can be used to calculate 4 damages on a classwide basis. Id. at 46-51; see also Dkt. No. 161-1 at 58-59. 5 Unlike the yardstick method, Professor Noll's proposed mark-up method does not require

6 identific