Opinion
Civil Action 7:19-cv-00269
04-04-2023
VTX COMMUNICATIONS, LLC, et al., Plaintiffs, v. AT&T INC., et al., Defendants.
ORDER AND OPINION
Micaela Alvarez, United States District Judge.
The Court now considers Plaintiffs' motion to strike Edward Rock's expert report and exclude his testimony, Defendants' motion to exclude the testimony of Thomas Z. Lys and Coleman Bazelon, and alternative motion, responses, replies, and exhibits.
Dkt. No. 79.
Dkt. No. 87.
Dkt. Nos. 88, 91.
Dkt. Nos. 104, 107.
I. Background and Remaining Pre-Trial Issues
In ruling on the parties' motions for summary judgment, the Court found liability for- inter alia-Defendants' operation of Cricket in competition with Plaintiff Partnerships and Defendants' decision to retain spectrum licenses to extract value from the Partnerships in abrogation of their contractual and fiduciary duties. Now, the parties have moved to strike certain of each other's expert reports and exclude those experts from trial.
Dkt. No. 119.
II. Legal Standard for Expert Opinion Admissibility
“[T]he Federal Rules of Evidence control the admission of expert testimony.” Rule 702 provides:
Mathis v. Exxon Corp., 302 F.3d 448, 459 (5th Cir. 2002).
A witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or otherwise if:
(a) the expert's scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue;
(b) the testimony is based on sufficient facts or data;
(c) the testimony is the product of reliable principles and methods; and
(d) the expert has reliably applied the principles and methods to the facts of the case.
“Expert testimony which does not relate to any issue in the case is not relevant and, ergo, nonhelpful. Similarly, low probative value, or a total lack of it, will render proposed expert testimony unhelpful and, therefore, inadmissible under Federal Rule of Evidence 702.” The Court scrutinizes proposed expert testimony more searchingly than lay witness testimony for its pertinency and potential prejudice.
4 Jack B. Weinstein & Margaret A. Berger, Weinstein'S Federal Evidence, § 702.02[5] (Mark S. Brodin, ed., 2d ed. 1997) (cleaned up), quoted in Daubert v. Merrell Dow Pharm., 509 U.S. 579, 591 (1993).
Rule 702 of the Federal Rules of Evidence Is Sound; It Should not be Amended, 138 F.R.D. 631, 632 (1991) (Weinstein, J.), quoted in Daubert, 509 U.S. at 595.
There is a relevancy problem where an expert opinion is primarily legal in nature because “[e]xperts cannot render conclusions of law or provide opinions on legal issues.” In a jury trial, the judge determines the applicable law, so legal opinions are not helpful to the jury as trier of fact. Furthermore, “the trial judge ought to insist that a proffered expert bring to the jury more than the lawyers can offer in argument.”
Renfroe v. Parker, 974 F.3d 594, 598 (5th Cir. 2020) (internal quotations omitted).
Salas v. Carpenter, 980 F.2d 299, 305 (5th Cir. 1992).
Once testimony is deemed relevant, the reliability of the expert's methodology becomes the touchstone. The Daubert test is a flexible one, and “under Daubert, any step that renders the analysis unreliable renders the expert's testimony inadmissible. This is true whether the step completely changes a reliable methodology or merely misapplies that methodology.”
Daubert, 509 U.S. at 594.
Moore v. Ashland Chem. Inc., 151 F.3d 269, 278 n.11 (5th Cir. 1998) (quoting In re Paoli R.R. Yard PCB Litig., 35 F.3d 717, 745 (3d Cir. 1994) (cleaned up).
To test reliability, the Court assesses the intellectual rigor of the proposed expert testimony, which must be validated by an independent and objective source beyond the expert's assurances, and the Court “should ensure that the [expert] opinion comports with applicable professional standards outside the courtroom and that it will have a reliable basis in the knowledge and experience of [the] discipline.” However, an expert report or opinion need not be in lockstep with the relevant discipline's prevailing view in order to be admissible. “Certain more specific factors, such as testing, peer review, error rates, and ‘acceptability' in the relevant scientific community . . . might prove helpful in determining the reliability of a particular scientific ‘theory or technique.'” Reliance on studies that do not support a contention, cherry-picked data, or a dubious methodology may be grounds to reject expert testimony.
Pipitone v. Biomatrix, Inc., 288 F.3d 239, 244 (5th Cir. 2002) (quoting Kumho Tire Co. v. Carmichael, 526 U.S. 137, 152 (1999)).
Brown v. Ill. Cent. R.R., 705 F.3d 531, 536 (5th Cir. 2013) (quoting Moore v. Ashland Chem. Inc., 151 F.3d 269, 276 (5th Cir. 1998) (en banc)); see Hathaway v. Bazany, 507 F.3d 312, 318 (5th Cir. 2007) (alteration and quotation omitted) (“But the existence of sufficient facts and a reliable methodology is in all instances mandatory. Without more than credentials and a subjective opinion, an expert's testimony that “it is so” is not admissible.”).
Watkins v. Telsmith, Inc., 121 F.3d 984, 991 (5th Cir. 1997) (second alteration in original) (quotation omitted).
Whitehouse Hotel LP v. Comm'r, 615 F.3d 321, 332 (5th Cir. 2010) (rejecting the argument that compliance with uniform published professional standards goes to admissibility rather than credibility); see Daubert, 509 U.S. at 588 (“Nothing in the text of this Rule establishes ‘general acceptance' as an absolute prerequisite to admissibility.”).
Kumho Tire Co. v. Carmichael, 526 U.S. 137, 141 (1999) (quoting Daubert, 509 U.S. at 593-94).
Burst v. Shell Oil Co., 650 Fed.Appx. 170, 174 (5th Cir. 2016) (per curiam).
The Court's task at this stage is gatekeeping, not premature factfinding. “As a general rule, questions relating to the bases and sources of an expert's opinion affect the weight to be assigned that opinion rather than its admissibility and should be left for the jury's consideration.” Indeed, the Fifth Circuit has cautioned against transforming a motion to exclude an expert into a trial on the merits, because the factfinder may be entitled to accept or reject an expert's testimony including by judging whether the predicate facts on which an expert relied are accurate. In short, experts may rely on disputed facts, but not unsubstantiated assertions. Cross-examination and presentation of competing evidence, rather than exclusion for inadmissibility, are the traditionally favored ways to challenge an expert opinion. “It is the role of the adversarial system, not the court, to highlight weak evidence.”
United States v. 14.38 Acres of Land, more or less Situated in Leflore Cnty., 80 F.3d 1074, 1077 (5th Cir. 1996) (quoting Viterbo v. Dow Chem. Co., 826 F.2d 420, 422 (5th Cir. 1987)); see FED. R. EVID. 702 ADVISORY COMMITTEE'S NOTE TO 2000 AMENDMENT (“WHEN FACTS ARE IN DISPUTE, EXPERTS SOMETIMES REACH DIFFERENT CONCLUSIONS BASED ON COMPETING VERSIONS OF THE FACTS. THE EMPHASIS IN THE AMENDMENT ON ‘SUFFICIENT FACTS OR DATA' IS NOT INTENDED TO AUTHORIZE A TRIAL COURT TO EXCLUDE AN EXPERT'S TESTIMONY ON THE GROUND THAT THE COURT BELIEVES ONE VERSION OF THE FACTS AND NOT THE OTHER.”).
See Pipitone, 288 F.3d at 250.
See MM Steel, L.P. v. JSW Steel (USA) Inc., 806 F.3d 835, 852 (5th Cir. 2015) (quoting Daubert, 509 U.S. at 596; see 14.38 Acres of Land, 80 F.3d at 1078 (“[T]he trial court's role as gatekeeper is not intended to serve as a replacement for the adversary system.”).
Primrose Operating Co. v. Nat'l Am. Ins. Co., 382 F.3d 546, 562 (5th Cir. 2004).
III. Application
Defendants move to exclude testimony of Thomas Lys and parts of the testimony of Coleman Bazelon. Defendants allege that Dr. Bazelon's reply report raises new opinions and should be stricken, or alternatively, they should be granted leave to designate a reply expert. Plaintiffs move to strike the rebuttal report and exclude trial testimony of Edward Rock as a legal opinion.
A. Coleman Bazelon
Plaintiffs' expert Coleman Bazelon is an economic consultant specializing in telecommunications and intellectual property. He consults on regulation, wireless license auctions, competition policy, patent infringement, and asset valuation. Here, he includes his opinions on cannibalization of Partnership customers. Cannibalization occurs when there is competition among affiliates; when an existing operator (in this case, AT&T) launches a new brand (in this case, Cricket), it predicts how many customers will leave the legacy brand for the new one. Those customers are said to be “cannibalized,” while new customers that the legacy brand would not have picked up otherwise are “additive.”
Dkt. No. 87-3 at 5, ¶¶ 1-3.
Dkt. No. 92 at 593 (Deposition of Jennifer Van Buskirk).
Dkt. No. 87-3 at 62, ¶ 94.
Defendants argue that Dr. Bazelon's cannibalization figures rely on invalid assumptions and that he changes his cannibalization methodology between his initial report and his reply report. They also argue that he uses very late ex post evidence and unreliably extrapolates from a single data point to value the relevant spectrum licenses.
1. Cannibalization
Dr. Bazelon's cannibalization methodology is produced in his initial report as a formula.In words, it calculates cannibalized subscriber additions in a given market in a given quarter as a function of Cricket's gross additions in south Texas times the proportion of the south Texas market in the given partnership, minus the estimated gross additions for a hypothetical, independent Cricket (the “baseline”). The resulting figure-how many subscriber additions are a result of AT&T-is adjusted for diversion and churn.
Id. at 114, ¶ 3.
Defendants challenge Dr. Bazelon's use of a regression analysis to construct data for certain periods, including January and February of 2014 (before the merger) and periods in 2014 where the data is unusually low. But the rationale for the regression analysis is clear: the de minimus data from January and February 2014 come from AT&T's legacy separate pre-paid service, Aio. Pre-merger (i.e., non-AT&T) Cricket was larger than that.
Dkt. No. 87 at 35, ¶ 49.
Dkt. No. 87-7 at 68-69, ¶ 11, n.177.
See Dkt. No. 88 at 45, ¶ 85, n.157.
Dr. Bazelon uses “2015 post-acquisition data to estimate the 2014 Cricket but-for baseline[, which] incorporates any increases in gross additions that Cricket would have experienced from the AT&T acquisition.” On this issue, Plaintiffs' calculation of damages is the difference between the baseline and the actual. But for January and February 2014, the baseline and actual have yet to diverge because the acquisition has not happened yet. Use of regression analysis to project unusually low data in other periods (after the merger) will be addressed below.
Dkt. No. 87-7 at 40, ¶ 69.
In his reply report, Dr. Bazelon changes several relevant items, which Defendants now challenge as unfounded and deserving of another reply report. He (1) opts for a monthly analysis instead of a quarterly analysis, (2) zeros out negative cannibalization, (3) accepts unusually low post-merger data that he previously rejected in favor of his own regression results, and (4) creates a Covid-19 adjustment for the 2020 and 2021 baselines.
A reply report is subject to the same reliability standards. Furthermore, the Court looks to four factors in deciding whether to exclude expert testimony as untimely: “(1) the explanation for the failure to produce the report earlier; (2) the importance of the testimony; (3) potential prejudice in allowing the testimony; and (4) the availability of a continuance to cure such prejudice.” For reasons that will become clear, the Court will not exclude Dr. Bazelon's report for timeliness.
AIG Eur., Ltd. v. Caterpillar, Inc., 831 Fed.Appx. 111, 115 (5th Cir. 2020).
Why did the reply report switch to a monthly analysis? Defendants assert that Dr. Bazelon “changed his methodology to a monthly view and-without explanation or any basis for doing so-simply excluded any months where negative cannibalization resulted from his calculations.”
Dkt. No. 87 at 34, ¶ 46.
Defendants know full well why the initial report looks at quarters and the reply report looks at months; they did not produce the more granular, responsive documents until after their own experts used them. While Plaintiffs could dial it down a bit on the ad hominem attacks on counsel, the Court is satisfied that the changes in the reply report were necessitated by Defendants' lack of production.
Dkt. No. 88 at 41-16.
We have all seen Vice Chancellor Laster's opinion by now.
Defendants did not produce the monthly data that covered post-2016 periods until June 23, 2022-after Plaintiff's expert designation deadline, but just one day after Defendant's expert designation deadline. Defendants have a duty to timely supplement production when it finds responsive documents even if Plaintiffs do not complain of incompleteness or file a motion to compel. Plaintiffs do not know what they do not have. In a case this old, the Court is highly skeptical that the responsive documents were serendipitously found between May 16th and June 23rd. The Court will not exclude the reply report for timeliness issues, nor will it grant Defendants' motion to designate a new expert since its expert had a first look at the new data that inspired the reply report. Plaintiffs have sufficiently explained the need for Dr. Bazelon's reply report, and there is no prejudice in Plaintiffs' experts using the same data used by Defendants' experts. Defendants' motion for leave is therefore DENIED.
See Dkt. No. 88 at 49-50, ¶ 93 (and accompanying citations).
That is, between the parties' expert designation deadlines.
In the reply report, Dr. Bazelon collapses all “negative cannibalization” to zero. Negative cannibalization is not the Partnerships taking back Cricket customers. It is when AT&T's Cricket underperformed compared to the gross additions Dr. Bazelon projected for hypothetical Cricket not owned by AT&T. Dr. Bazelon's choice to shift from a “conservative” approach in the initial report (given the incomplete data) to what he considers a more accurate approach in the reply report may be fair game on cross examination. But that choice does not make his whole cannibalization methodology unreliable.
Dkt. No. 87-7 at 70-71, ¶ 17.
That is accounted for by diversion and churn. See Dkt. No. 87-7 at 34, ¶ 63.
Dkt. No. 87-3 at 114, ¶ 4.
The apparent inconsistency in Dr. Bazelon's acquiescence on the unusually low postmerger data but continued use of regression data for pre-merger gross adds is explained. In analyzing the initial data, he “could not distinguish between a transition where there may have been a ramp up in gross additions and data errors. However, the data that Ms. Taylor [Defendants' expert] uses in her analysis . . . includes these apparently anomalous data points [so Dr. Bazelon] incorporates these early months into [his] updated analysis.” In other words, since Defendants doubled down on those figures, he decided to take them at their word. In contrast, the de minimus data from January and February 2014 continues to need replacing, in his opinion.
Dkt. No. 87-7 at 35-36, ¶ 65.
Dr. Bazelon applies a Covid-19 adjustment to gross additions “to account for the depressed level of gross additions” during 2020 and 2021. Defendants put forth a theory that this was an outcome-oriented hack that Dr. Bazelon pulled off when he “found he had grossly overestimated [Cricket] performance [and] needed a new and lower baseline to maintain or increase damages.”But if the baseline is meant to project hypothetical Cricket's performance through 2021, then the same market forces created by Covid-19 which effected real Cricket would effect the baseline. Dr. Bazelon's model takes this into account.
Dkt. No. 87-7 at 34, ¶ 63.
Dkt. No. 107 at 32, ¶ 50.
Defendants point out that the report's tables indicate several high individual p-values and ask the Court to find the Covid-19 adjustment statistically insignificant. Dr. Bazelon explains that to construct the Covid-19 adjustment, he “use[s] two dummies for the year 2020 and 2021 and their interactions with the time trend and the quarter dummies. The F-stat[istic] shows that the covariates are jointly significant.”
Dkt. No. 87 at 38, ¶ 55.
Dkt. No. 87-7 at 34, ¶ 63, n.100. Plaintiffs provide an example of this type of statistical dispute being resolved in favor of their position by another District Court. In re Elec. Books Antitrust Litig., 2014 U.S. Dist. LEXIS 42537 at *79-80 n.34 (S.D.N.Y.) (J. Cote) (“Multicollinearity occurs when several variables are correlated with each other -that is, they overlap in the sense that they each capture some part of the same effect -- such that it is impossible from the data set to objectively determine the influence of one variable as opposed to the influence of these others. In that case, . . . while the overall effect of these variables may be strongly statistically significant, each coefficient may not appear statistically significant on its own.”) (citing JEFFREY M. WOOLDRIDGE, INTRODUCTORY ECONOMETRICS: A Modern Approach, 148-49 (5th ed. 2013).
An insignificant p-value would be problematic if Dr. Bazelon were running a t-test to determine if Covid-19 caused the depressed year-over-year gross addition numbers. But that is not what is going on here. Instead, the reason for this data is to project baseline gross addition numbers for hypothetical Cricket that reflect the same market forces impacting real Cricket. For 2020 and 2021, real Cricket's numbers are conspicuously lower. Conspicuousness does not equate to statistical significance, but it does not need to. Even if Covid-19 is not the sole cause, there is no reason to think that the global causes impacting real Cricket would not also impact hypothetical Cricket. The report calls this the “COVID adjustment” since it correlates with 2020-2021 and there is statistical significance on a quarter-over-quarter basis. But the adjustment's validity does not actually depend on isolating Covid-19 as the statistically significant reason for the lower numbers.
Defendants have not put forth an alternative explanation for Cricket's low numbers that would not apply to Dr. Bazelon's hypothetical Cricket, such as mismanagement by AT&T.
2. Spectrum
Defendants argue that Dr. Bazelon's valuation of the portion of the spectrum licenses in the Partnerships' service areas is unreliable because it relies on ex post FCC auction data from up to nine years after AT&T's license acquisitions and that it improperly extrapolates from a single data point. Contrary to Defendants' vigorous arguments, courts rarely establish hard rules that certain methodologies are always unreliable and excludable. It matters what is being measured, and why.
Dkt. No. 87 at 44-49.
What does Dr. Bazelon measure with respect to the spectrum licenses? He does not give an absolute valuation (i.e. the fair market price) for the licenses; he calculates the portion of the overlapping license's book value (i.e. the price AT&T paid) associated with the Partnerships' service areas. So long as it is applied consistently-by using the same valuation metrics on both sides of the Partnership boundary and apportioning the relative book value to each side-this methodology does not require estimation of the license's absolute price.
Dkt. No. 87-3 at 101-103, ¶¶ 135-137.
This consistency appears to have been implemented in Dr. Bazelon's report. Dkt. No. 87-3 at 105, ¶ 141.
While very late ex post auction data would likely be unreliable as to absolute value, its reliability as to relative value is properly explained. A subsequent auction may be a bad measure for absolute value because the market's projection for that spectrum band's use (and thereby, value to an operator) might change dramatically over time. But the proportional share of a license's value as between two geographic areas is unlikely to change dramatically over time barring rapid population swings.
Thus, Dr. Bazelon uses auction data to construct relative value indexes, which allow him “to value one particular geography relative to another geography, without having to know the absolute value of either.” The book value paid by AT&T will be plugged in as the absolute value of the overlapping license, so the age of the auctions from which the relative value indexes are developed does not render the methodology unreliable.
Dkt. No. 87-3 at 105, ¶ 141.
While Defendants assert several times that Dr. Bazelon relies on “a single transaction” or “a single data point,” they do not cite to his report nor explain their claim. From the Court's review of the report, it appears that constructing the relative value indexes involves comparing several licenses that envelop the same service area to triangulate the effected Partnerships.
Dkt. No. 87 at 44, 48, ¶¶ 70, 80.
E.g. Dkt. No. 87-3 at 108, ¶ 145.
Auction data is used here to empirically capture data that effects spectrum pricing beyond population alone. This methodology is necessary because-as Dr. Bazelon points out- Defendants have attempted to have their cake and eat it too by (1) their appraiser, Duff & Phelps, using extra-populational metrics such as income, terrain, and population density to arrive at a fair value but (2) AT&T charging the Partnerships their allocated portion based on population alone.Dr. Bazelon's use of empirical data here also highlights a difference between his testimony in this case and his testimony in the Delaware case that Defendants would like to compare it to. There, contrary to Defendants' incorrect briefing, Dr. Bazelon was permitted to testify at trial, but the weight of his testimony was discounted for relying on projections of absolute spectrum valuation when examples of real market transactions were available. Here, Dr. Bazelon's testimony makes use of the real world market evidence that he was criticized for ignoring. For all of the aforementioned reasons, Defendants' motion is DENIED as to the Bazelon report.
Id. at 103-104, ¶¶ 138-139.
ACP Master, Ltd. v. Sprint Corp., Nos. 8508-VCL, 9042-VCL, 2017 Del. Ch. LEXIS 125, at *94 (Ch. 2017).
Dkt. No. 107 at 38, ¶ 63, n.126 (describing the Delaware court as “excluding Bazelon” for his methodology).
ACP Master, 2017 Del. Ch. LEXIS 125, at *94.
B. Thomas Lys
Defendants argue that Dr. Lys' report constitutes an improper legal conclusion, that his analysis of the processes used by AT&T contains no particular skill, and that he improperly relies on the Bazelon report.
First, “[a]n opinion is not automatically objectionable just because it embraces an ultimate issue.” Defendants' cited authorities stand for the proposition that expert opinions on what law applies or how it should apply are improper. That is the Court's domain. But there is nothing improper about an expert arriving at a conclusion based on the facts analyzed and the law provided by the Court. Dr. Lys' report shows a meticulous effort to use only the interpretations of Delaware law and the parties' Partnership Agreements which have already been endorsed by the Court.
Since then, the Court has found in Plaintiff's favor on many of these ultimate issues of liability. Since the jury will be instructed consistent with that opinion anyway, there is nothing improper about Dr. Lys' dependance on those findings. Furthermore, finding liability is a threshold issue for assessing damages-the job Dr. Lys was hired to do. If he did not think AT&T was liable for underpaying the Partnerships, he would not quote an underpayment figure.
Dkt. No. 119.
Second, the Court finds that Dr. Lys' opinion is based on specialized knowledge and skills that will be helpful to the jury. Two examples will suffice. The report provides an opinion on how typical reseller mobile virtual network operator (MNVO) contracts function-including volume minimums and maximums-and how AT&T's use of Partnership networks at a reseller rate differed from those contracts. Another example is his application of accounting standards to AT&T's expense offsetting as a form of compensation.
Dkt. No. 87-2 at 32-33, 49, 54.
Id. at 69-70.
Third, because Dr. Bazelon's testimony is admissible, Dr. Lys' reliance on it is not improper. Defendants' motion is DENIED as to the Lys report
C. Edward Rock
Finally, Professor Rock's rebuttal report to Dr. Lys differs from the other reports in type, not degree of reliability. It is a legal brief by a law professor and legal expert. It deals with interpretation of Delaware law and interpretation of the parties' contracts. These opinions attempt to usurp the role of the Court and are easily excludable under Rule 702(a). The Court-not a rebuttal witness-will ensure that Dr. Lys' opinions comport with the correct legal standards.
Defendants cast Professor Rock's opinion as one about the Partnerships' organizational design as opposed to an interpretation of the Partnership Agreements themselves. But it is clear from reading Professor Rock's report that his arguments against Dr. Lys' organizational design opinions are legal arguments based on the Agreements and Delaware law. These points can be made by the lawyers.
Dkt. No. 91 at 18.
Dkt. No. 79-1 at 158-159, ¶¶ 84-85.
Still, the Court identifies two topics in Professor Rock's report with potential probative value at trial. First, Professor Rock provides an opinion on the practicality of negotiating rates with limited partners, as opposed to unilaterally setting nationwide rates as AT&T did. But he provides no basis for the alleged impracticability, and instead resorts to a string of rhetorical questions. His assertion that negotiation would be impractical is not supported by any facts or data. Similarly, he opines that AT&T consulting Ernst & Young and Duff & Phelps was “a reasonable way of handling the complexities created by the formation of these partnerships during the 1980s” but provides no basis for this assertion.
Dkt. No. 79-1 at 160.
Id. ¶¶ 92-93; see also id. at 157, ¶ 75.b.ii. (describing individualized negotiation as “time-consuming, disruptive, inconclusive, and damaging to running a regional or national network efficiently”).
Id. at 161, ¶ 102.
Accordingly, the Court finds that admitting Professor Rock's testimony would provide no reliable, probative value and would serve only to confuse the jury by providing opinions on matters that are not for the jury to decide. His rebuttal report is hereby STRICKEN, and his testimony EXCLUDED.
IV. Conclusion
Accordingly, the Court hereby GRANTS Plaintiffs' motion and DENIES Defendants' motions.
IT IS SO ORDERED.