Food Lion, LLC v. Dean Foods Co.

312 F.R.D. 472, 2016 U.S. Dist. LEXIS 8152, 2016 WL 310432

• Court: District Court, E.D. Tennessee
• Decided: January 25, 2016
• Docket: No. 2:07-CV-188
• Status: Published

Opinion

MEMORANDUM OPINION AND ORDER

J. RONNIE GREER, UNITED STATES DISTRICT JUDGE

This matter is before the Court on the motion of plaintiffs for class certification, [Docs. 201, 202]. The defendants have responded in opposition, [Doc. 228], and plaintiffs have replied, [Doc. 295]. Plaintiffs filed a supplemental memorandum in support of the motion for class certification, [Doc. 669], and defendants responded to the supplemental filing, [Doc. 719]. The Court heard expert testimony on June 23-24, 2015, [see Docs. 739, 740], and heard oral argument on September 17, 2015, [see Doc. 771], The motion is ripe for disposition and, for the reasons set forth below, will be DENIED.

I. Background

Plaintiffs are retail sellers of processed milk who purchase directly from Dean Foods Company (“Dean”) and/or Dairy Farmers of America, Inc. (“DFA”), a dairy cooperative which owns, or owns an interest in, milk processing plants. Plaintiffs bring this putative class action complaint under §§ 1 and 2 of the Sherman Act, 15 U.S.C. §§ 1-2, and § 3 of the Clayton Act, 15 U.S.C. § 14. Plaintiffs’ complaint originally included counts alleging an agreement not to compete (Count 1), conspiracy to restrain trade (Count 2), monopolization and attempted monopolization (Counts 3 and 4), and conspiracy to monopolize (Count 5). Counts 2 through 5 have been dismissed, [Docs. 537, 667]¹, and those claims have now been abandoned by plaintiffs. Count 1 alleges that Dean, DFA and National Dairy Holdings, L.P. (“NDH”), agreed, in violation of the Sherman Act, “to lessen competition for sales of processed milk in the Southeast” (defined by plaintiffs as Federal Milk Marketing Orders (FMMO) 5 and 7). Here is how plaintiffs describe the alleged conspiracy:

A. Dean and Suiza’s Rapid Consolidation of the Milk Bottling Business

In the years before the 2001 Dean/Suiza merger, the milk bottling business was highly fragmented. In the late 1990’s, two competitors — Dean and Suiza — employed “arms race” acquisition strategies, substantially increasing the number of plants they owned and significantly consolidating milk bottling. Simultaneously, the suppliers to milk bottlers — dairy cooperatives (“co-ops”) — also were consolidating, and in 1998, four dairy cooperatives merged to form Defendant DFA. DFA supplied Sui-za’s bottling plants pursuant to full-supply agreements, while Dean’s bottling plants were supplied largely by independent dairy farmers.

Suiza, Dean, and DFA were each controlled by a small group of executives. At that time, Suiza was run by CEO Gregg Engles, Vice Chairman Tex Beshears, and CFO and Vice-President of Corporate Development, Tracy Noll, while Dean was run by CEO Howard Dean. DFA’s co-op operations were run by CEO Gary Hanman and CFO Gerry Bos. DFA’s business interests also included ownership in milk bottlers, the most significant of which were managed and partially owned by Pete Schenkel and Alen Meyer.

The enormous sums of money these executives received over a short period of time make the professional and financial interrelationships between the DFA, NDH and Suiza executives highly relevant the Retailer Plaintiffs’ claims. For example, in 1999, DFA acquired one-third of Suiza’s Dairy Group in exchange for selling certain of DFA’s joint-venture interests to the Suiza Dairy Group. As part of that transaction, Pete Schenkel netted $100 million personally and became the President of Suiza Dairy. DFA also acquired the milk supply rights to the remaining Suiza plants in that transaction. Earlier in 1998, DFA had gained supply rights to Suiza plants in three states within the Southeast with the sale of DFA’s Land-O-Sun joint venture plants to Suiza. The DFA-Suiza Land-O-Sun transaction netted Allen Meyer $70 million. From 1998 to 2004, Tracy Noll was paid over $34 million related to his milk industry positions and relationships. During that same time, Tex Beshears was paid well over $100 million. Finally, Dean’s CEO, Gregg Engles, one of the highest paid food industry executives, has received $103 million since 2003. Thus, collectively these five executives alone made in excess of $400 million in the past 10 years or so in milk related transactions. These handpicked loyalists were well-placed to effectuate the market-wide plan described below.

B. Suiza, DFA, and Dean Engineer the Dean/Suiza Merger

By 2000-2001, Dean and Suiza operated milk bottling plants throughout the Southeast, in overlapping geographic areas. Sui-za was the largest milk bottler in the United States, operating 67 dairy processing plants in 29 states, and by 2001, Dean was the second largest milk bottler, operating 43 dairy processing plants in 19 states. By a large margin, Dean and Suiza were the largest two competitors in the highly fragmented industry. Meanwhile, by 2001 Sui-za’s supplier DFA was the largest dairy co-op in the United States.

On December 1, 2000, as the competitive race proceeded, the presidents of Suiza, DFA, and Dean — Gregg Engles, Gary Hanman, and Howard Dean — went on a hunting trip. This meeting ultimately led to the December 2001 merger of Suiza and Dean. Before the hunting trip, Engles had been actively pursuing the acquisition of Crowley-Marigold, two key groups of bottling plants in the Northeast and the Midwest. Suiza’s Vice-President of Corporate Development, Tracy Noll, was in charge of conducting the due diligence review of Crowley-Marigold. Upon returning from the hunting trip in early December 2000 with the CEOs of Dean and DFA, Engles told his Suiza management team that Suiza would seek to buy Dean or Crowley-Marigold — but not both. Suiza could not acquire both Dean and the Crowley-Marigold plants because of antitrust concerns — i.e., Suiza would control and own too many milk bottling plants. Engles instructed Noll to continue exploring the Crowley-Marigold acquisition. Noll retained Suiza’s long-time acquisition consultants, Deloitte and Touche, to assist with due diligence. Suiza simultaneously explored its ability to acquire Dean.

By the end of January 2001, Suiza chose to acquire Dean instead of Crowley-Marigold; the newly merged Dean/Suiza company would be called “Dean.” Furthermore, in mid-January 2001, Noll told Engles that Noll would be leaving Suiza to rejoin Cletes “Tex” Beshears at the predecessor of NDH. As the Dean/Suiza deal began to take shape, it was obvious that consolidation of the two bottling giants would create substantial antitrust concerns on the part of the DOJ. Because of the unprecedented level of concentration in the milk bottling business that would result from the merger, particularly in the Southeast, Suiza and Dean expected that the DOJ would likely (1) require divestitures of several Dean/Suiza plants and (2) mandate that Dean/Suiza not purchase any other plants.

A solution, however, was orchestrated through a four-way deal. In mid-February 2001, Suiza’s Greg Engles and Pete Schenkel met with DFA’s Hanman and Bos and others about acquiring the eventual spin-off plants from the Dean/Suiza merger. DFA established an entity — Defendant NDH — to first purchase the Crowley-Marigold plants that Suiza previously had been pursuing and then to capture the inevitable plant divestitures from the Dean/Suiza merger. That new company, was immediately stocked with senior executives who had previously worked for Sui-za or DFA. DFA, along with Allen Meyer (a long-time business partner in the bottling business with DFA), Tracy Noll (a former Suiza executive) and Tex Beshears (another former Suiza executive) became NDH’s owners. Curiously, NDH’s Allen Meyer was paid a $1.6 million bonus for “originating, negotiating and facilitating” the transaction with Suiza. When asked in a recent deposition about the specifics of Meyer’s involvement and the bonus Meyer was paid, however, NDH’s Tracy Noll “took the Fifth.”

Ultimately, in a successful attempt to placate DOJ and win approval for the merger, Dean agreed to divest 11 bottling plants to NDH. In numerous presentations during the merger review process, Dean convinced the DOJ that the divested plants would provide vigorous competition and that following the merger, there would remain “many” local and regional milk bottlers to compete for sales to retail customers.

For example, Dean provided the DOJ with maps of the Bristol, Virginia area and the Huntsville, Alabama area and labeled the maps, “Fluid Milk — Vigorous Local and Regional support acquisitions increasing its investment at times, primarily through preferred equity investments _”. At the end of 2007, NDH expected that DFA would have $390.3 million in preferred equity invested in NDH, with DFA having an 87.5% common ownership stake in Competition,” Dean and Suiza also assured the DOJ that NDH planned to make further acquisitions of milk bottling facilities. Those efforts ultimately led the DOJ to announce on December 18, 2001, that it had approved the merger. The DOT’S announcement extolled the fact that “these divestitures ensure the fact that consumers of milk ... continue to get the benefits of competition” and that there would be increased choices and lower prices. The merger created unprecedented market concentration in the fluid milk business in the Southeast United States. The resulting Dean is a publicly-traded, for-profit corporation based in Dallas, Texas, with 2007 net sales of $11.8 billion; Dean has more than 100 milk bottling plants, located in thirty-six states.

In summary, when the dust settled from the Dean/Suiza merger: (1) Suiza had accomplished its goal of eliminating its biggest competitor, Dean; (2) DFA had accomplished its goal of contractually controlling access to supply all-interested bottling plants through full supply agreements, including the combined Dean/Suiza bottling plants, the spun-off NDH bottling plants, and the Crowley-Marigold bottling plants acquired by NDH; (3) Dean/Suiza’s new biggest “competitor,” NDH, was run by insiders from both Suiza (Noll and Beshears) and DFA (Meyer); and (4) the newly-formed co-conspirator, NDH, captured the Dean/Suiza divested plants and the Crowley-Marigold plants — with NDH executive Noll even using the due diligence commissioned by Suiza when Noll previously analyzed the Crowley-Marigold acquisition on Suiza’s behalf.

Defendants’ ultimate mission — which forms the basis for Retailer Plaintiffs’ claims — was to enlist DFA’s allegiance in ensuring that NDH remained a compliant and cooperative partner rather than a “vigorous” competitor. The conspiracy manifested itself in four key ways. First, NDH was complicit in undermining the competitive significance of the spun-off plants. Second, Dean and NDH cooperated with each other rather than competing vigorously as Dean told the DOJ they would. Third, Dean, NDH and DFA minimized their opportunities for robust competition by retreating to separate geographic portions of the Southeast. Fourth, Dean, NDH and DFA acted together to block new entrants of other milk bottlers in the Southeast.

1. Undermining the Divestitures — Dean and NDH Move Customers to Dean to Facilitate Closing Competing Plants

Despite the fact that Dean and Suiza told the DOJ that NDH would be “an aggressive expanding competitor,” Suiza acted even before receiving DOJ’s merger approval to undermine the competitive position of at least two of the plants that NDH acquired pursuant to the merger.

For example, before the former Suiza Meadow Gold plant in Huntsville, Alabama was spun-off to NDH, Dean/Suiza knowingly and intentionally hobbled the plant by moving the Wal-Mart business supplied by Huntsville to a Dean plant in Birmingham. Because Wal-Mart was Huntsville’s largest customer by far, representing almost 40 percent of the plant’s total sales, losing that business assured that the Huntsville plant would become immediately unprofitable.

Moreover, NDH co-owner Meyer testified that NDH knew even before the Dean/Suiza merger closed that NDH would lose money on the hobbled Huntsville plant, and admitted that “[i]t should have been closed day one.” It was not until 2003, however, that NDH closed the Huntsville plant after posting substantial financial losses for many months. Likewise, Dean also switched Wal-Mart business away from the former Suiza Flav-O-Rich plant in Bristol, Virginia that was also spun-off to NDH in the merger. As a result, two of the five spin-off plants located in the Southeast market were instantly rendered financially non-viable, and NDH ceased milk bottling at both plants.

2. Customer Allocation and “Courtesy Bidding” — NDH, at Dean’s “Instruction” Declines to Compete

Although discovery is still on-going, Retailer Plaintiffs have already uncovered evidence that NDH agreed, at Dean’s instruction, not to bid aggressively for certain customer business. In a December 2002 e-mail exchange between NDH President — and former Suiza executive — Tracy Noll and NDH’s Rob Cottet (also a former Suiza employee), the latter a manager for several Southeast plants, Noll and Cottet discussed NDH’s unexpected grab of Associated Grocers (“AG”), Dean’s largest customer at least at one of its bottling plants in Florida. But NDH was unhappy with this “win,” however, because it violated the agreement between Defendants to allocate customers. As Cottet put it:

We were actually approached last spring/summer for the [AG] business and we did not bid aggressively at the instructions of our former owners.

(emphasis added). The “former owner[]” who instructed NDH not to “bid aggressively” was Suiza, which became the new Dean — NDH’s ostensible competitor.

The December 2002 e-mail exchange was not the only time that Noll and Cottet discussed NDH’s agreement not to compete with Dean. Following the loss of the Wal-Mart business at its Huntsville and Bristol plants, Cottet, perhaps unaware of the anticompetitive scheme that was in place, attempted to find new customers to replace that lost business. Cottet’s initial efforts met with some modest success, including winning the AG business discussed above. But before Cottet could fully implement his new customer strategy, he received a phone call from NDH’s President, Tracy Noll. Noll told Cottet that “Dallas” (which Cottet understood as shorthand for Dallas-based Dean) had complained about NDH’s going after Dean-allocated customers. Noll told Cottet that Cottet could not solicit any more of Dean’s customers. When Cottet complained that without new customers, the Huntsville and Bristol plants would continue to lose money and that they should then be immediately shut down, he was told that NDH could not close the plants just yet — because it was too soon after the merger to do so. Plainly, NDH hoped to keep up appearances of competition where there was none, in case the DOJ was still watching. Absent the conspiracy, there was no rational reason for NDH to care how things looked to the DOJ.

The evidence also indicates that NDH engaged in collusive “courtesy bidding” to avoid taking business from Dean while giving purchasers the appearance of competition. For example, a 2005 NDH email details how a Dean customer called Cowboy, a chain of approximately 40 retail stores in several states, solicited NDH to bid on its ice cream business, presumably to see if the retail chain could obtain better pricing. An NDH employee recommended to CEO Brian Haugh that NDH make a “courtesy bid” that would allow Dean to keep the Cowboy’s business:

I believe most of the stores are worked by Mayfield [Dean plants in Braselton, Georgia and Athens, Tennessee]. He is also looking for us to bid on this, I would recommend a courtesy bid letting May-field keep it.

This document — which also notes that NDH “gave” bottled milk customers to Dean, shows the common interests and coordination between Dean and NDH to share the market and to not compete with each other.

3. Territorial Division and Capacity Restriction — Dean, NDH, and DFA Allocate Territories in the Southeast Through Strategic Acquisitions and Plant Closings

In addition to Defendants’ efforts to avoid competing aggressively for customers, the Retailer Plaintiffs’ will prove that Dean, DFA, and NDH engaged in coordinated plant closings and acquisitions designed to physically separate their plant locations throughout the Southeast. For example, as NDH was preparing to withdraw from Huntsville, Alabama and cede the area to Dean, Dairy Fresh — at the time one of the few remaining independent bottlers in the Southeast — saw an opportunity to expand by purchasing the plant. Instead of obtaining full value from Dairy Fresh by selling the plant as a going concern, however, NDH sold the plant to the city of Huntsville at land value, after DFA “killed the deal” with Dairy Fresh. NDH was therefore willing to take less money to ensure that the plant would not be available to provide competition to Dean.

Less than a year after DFA killed Dairy Fresh’s effort to purchase the Meadow Gold plant as a going concern, NDH acquired Dairy Fresh, using DFA’s financial backing — thereby extinguishing the threat the Dairy Fresh posed as an independent bottling company to Defendants’ geographic market allocations.61 NDH also acquired a local bottling plant in Chattanooga, Tennessee in October 2003 and then closed it in May 2006. In total, since 2001, Dean, NDH, and DFA have acquired seven Southeast bottling plants and have shut down seven plants.

Following these actions, including NDH’s closing its milk bottling plants in Huntsville, Alabama, and Roanoke, Virginia, and ceasing its milk bottling operations at its Bristol, Virginia plant, the competitive milk bottling landscape in the Southeast looks very different today from the scene of vigorous competition promised to the DOJ in 2001. Immediately following the Dean/Suiza merger, there was an NDH plant in relatively close proximity to almost every Dean plant located in Order 5. Likewise, in Order 7, as of early 2002, most wholesale milk buyers were within reasonable distance of more than one milk bottler. Accordingly, retail customers had a choice of bottlers from whom they could purchase processed milk. By 2008, Dean and NDH had, in essence, “moved” their facilities to separate portions of the Southeast. Dean’s milk bottling plants are now heavily concentrated throughout Order 6 and in the eastern half of Order 7. NDH’s bottling plants are located mostly along the Gulf Coast, while the DFA joint venture bottling plants are located chiefly in the western portions of Order 7. New NDH or DFA joint venture plants are now located near Dean plants, and few Dean plants are now located near the majority of the remaining NDH plants. As a result, each Defendant bottler is now located primarily in a separate area of influence and control within the Southeast, with only token competition (if that) throughout much of the rest of the Southeast.

This separation of plants was important to accomplish the goals of the conspiracy. Indeed, Ernie Yates, head of Dean milk procurement, and John Wilson, a senior DFA executive, met in 2007 to discuss “plant rationalization” and collusively reducing bottling output to increase retail margins. Despite the fact that DFA/NDH and Dean are ostensibly competitors in the bottling market, Wilson reported to DFA’s CEO Rick Smith that Wilson and Yates:

talked some about plant rationalization. He, you know, between he and I, we were kind of concluding, you know, that’s really our fundamental problem right now on profitability of bottling. Is that the plants there’s just a few too many plants out there with too much capacity and if we could take, the industry could take some strategic plants out of the system it would certain allow folks to be more aggressive on pricing to retailers and help things. Obviously, we did not get specific on locations but it was generally the spirit there.

(emphasis added). “[PJlant rationalization” — i.e., the reduction of output in order to increase prices — by competitors is, of course, per se illegal under the antitrust laws.

4. Payments to Potential Competitors to Exit the Market and Retaliation Against Entrants — Dean, NDH and DFA Use Economic Leverage to Block Competitor Entry into the Southeast

Defendants were not satisfied to divide customers and territories. Indeed, if other milk bottlers entered the Southeast market, Defendants’ plan to charge and maintain illegally high prices for processed milk would be much less profitable. Therefore, Defendants also constructed a market-wide blockade, working together to keep would-be competitors out of the milk bottling business.

a. Defendants coordinate to block competition from “Red Oak” bottling plant

For example, an investor group (which included a DFA-rival co-op, Maryland and Virginia Milk Producers (“MD/VA”)), began building a new milk bottling facility, known as the “Red Oak” facility, in Baxley, Georgia. Dean was immediately alarmed. Dean concluded that the plant would be its “biggest threat” in the area. Indeed, the Red Oak owners expected the plant to be profitable in its first or second year of operation.

Dean made an initial offer to purchase Red Oak, but it was rejected — partly because Dean would not commit to actually completing or operating the plant — and the Red Oak investors preferred an investment from northeast bottler, HP Hood, Inc. (“Hood”). In fact, Jay Bryant at MD/VA testified that instead of actually intending to utilize the bottling plant, Dean wanted only to eliminate a potential competitor. Dean, working together with DFA and NDH, used economic leverage in other deals to dissuade the Red Oak investors from completing the plant. After DFA and NDH intervened, Hood backed away from the Red Oak deal. Hood, which had been in discussions with DFA/NDH regarding selling plants to NDH, fortuitously found itself purchasing a number of NDH bottling plants in the northeast, and MD/VA received supply rights to at least two bottling plants. DFA had the exclusive right to supply them — one owned by Hood and one owned by Dean. With Hood and MD/VA appeased with other plants and supply rights, Dean purchased the Red Oak facility and, as part of the deal, Dean sought and secured a wide-ranging, five-year, non-compete agreement from MD/VA in the process. The MD/VA board was told to keep the deal terns confidential and that neither MD/VA nor Dean would be issuing a press release announcing the transaction,

b. Coordinated retaliation against coop/bottler Southeast Milk

Defendants also choreographed a retaliation scheme against Southeast Milk, Inc. (“Southeast Milk”), a dairy farmer cooperative with operations in Florida, Georgia and Tennessee. Defendants were concerned that Southeast Milk would infringe on their bottling plant footprint in the Southeast.

Beginning in 2003, DFA began using its control of supply rights to certain NDH plants to extort payments from Southeast Milk. Southeast Milk was supplying the NDH plants with DFA’s acquiescence, but DFA required payments from Southeast Milk designed to subsidize DFA pay prices for raw milk in south Georgia and to suppress raw milk pay prices paid by Southeast Milk in the Southeast. Southeast Milk — much like MD/VA before it — began exploring plant acquisitions to secure a market for its raw milk after DFA effectively cornered access to bottlers through use of its full supply agreements.

In April 2004, Southeast Milk purchased the Gustafson’s bottling plant in Northeast Florida. Almost immediately, NDH noted internally its desire to “[s]hut the plant down if we can. We could make a little money.” In the short-term, DFA continued to allow Southeast Milk to supply the two NDH plants in Florida so long as Southeast Milk continued to make the tribute payments demanded by DFA

However, in late 2004 and early 2005, a team of investors including Southeast Milk sought to build support for a major push to expand into Southeast milk bottling— “Flagship.” In December 2004, Flagship approached MD/VA and asked if MD/VA would join forces with it. Having just accepted payment from Dean and DFA (ie., supply rights to Dean’s Red Oak facility and Hood’s Winchester facility) to abandon Red Oak and for the accompanying non-competition agreement, MD/VA decided not to join Flagship. Instead, it divulged Flagship’s plans to both DFA and Dean. Those plans included Flagship purchasing the New Atlanta Dairies from Parmalat in Atlanta, Georgia and the Giant plant in Landover, Maryland. MD/VA expressed its fear to both Dean and DFA that by not joining Flagship, MD/VA stood to lose its current supply rights to the Giant Land-over plant.

Dean was very concerned about the Flagship initiative and approached DFA with a plan to jointly placate MD/VA and to keep individual farmers from joining coops (namely Southeast Milk and USA Milk) that would potentially supply rival bottling ventures such as Flagship. In a transcribed voicemail, Dean Dairy Group President, Pete Schenkel and Dean’s Ernie Yates note that Dean has spent millions on “antitrust” and we “are concerned about losing millions more in some sort of milk situation where effectively due to people looking for alternatives and making dumb moves by buying underutilized plants they are going to lose more margins.”

Defendant’s overall plan to prevent new competitors’ entry, articulated at length in the transcribed voicemail, was to: (1) have DFA relax the full-supply agreement between Dean and DFA; (2) give MD/VA supply rights to Dean’s Shenandoah’s Pride plant to “appease” MD/VA’s members, and (3) allow Dean’s Barber and Purity plants to purchase raw milk supply directly from independents and non-DFA cooperatives in order to keep farmers— frustrated with DFA and DMS — within the Dean system rather than supplying Flagship and other rival milk bottlers.

As planned, shortly after this conversation, MD/VA received supply rights to Dean’s Shenandoah’s Pride plant outside of the DFA full-supply agreement, but with DFA’s acquiescence. Contemporaneous notes from MD/VA’s COO Mike John explicitly refer to MD/VA’s new Shenandoah’s Pride supply rights as a payment to MD/VA compensating it for the Giant, Landover sales MD/VA stood to lose by not joining the Flagship venture. Also, as planned and again outside of the DFA full-supply agreement, Dean began purchasing raw milk directly from independent producers to keep that raw milk out of competitor bottlers’ hands.

In addition to paying-off MD/VA for not joining Flagship and coordinating their approach to farmers contemplating joining rival co-ops pursuing bottling alternatives, DFA and Dean along with NDH finally were prepared to collectively punish Southeast Milk. They attacked Southeast Milk on all fronts: (1) NDH no longer accepted Southeast Milk’s supply at its Florida plants, and NDH noted that it was “forced to seek DFA support” from out-of-state to “subsidize competitive activity;” (2) Dean — nearly simultaneously — notified Southeast Milk that Dean intended not to automatically renew Southeast Milk’s raw milk supply agreement at Dean’s T.G. Lee and McArthur dairies; and (3) DFA reiterated its demand that Southeast Milk lower the price it was paying its cooperative milk farmer members in Georgia, or DFA would destroy raw milk prices in Southeast Milk’s home Florida market by flooding the area with additional raw milk.

[Doc. 731 (originally filed as Doe. 202) at 10 - 30].

II. The Proposed Class

Plaintiffs seek certification of a class defined as follows:

All persons, other than schools and school districts, within the Southeast United States who have purchased, at any time from January 1, 2002 until December 31, 2009, from any defendant, fresh white fluid milk which has been pasteurized and processed for human consumption and then packaged into containers which are sold to retail outlets and other customers.

[Doc. 669 at 12]. The proposed class has been modified from the class originally proposed by plaintiffs in two ways in response to criticism raised by defendants’ initial opposition. First, plaintiffs have modified the time period for qualifying purchases from “January 1, 2002 until the present” to “January 1, 2002 until December 31, 2009.” Second, they substitute the term “fresh, white, fluid milk” for the term “grade A milk.” [Id. at 11-12].

III. Class Certification Standard

“Class certification is appropriate if the [district] court finds, after conducting a ’rigorous analysis’, that the requirements of Rule 23 have been met.” Rikos v. Procter & Gamble Co., 799 F.3d 497, 504 (6th Cir.2015) (quoting In re Whirlpool Corp. Front-Loading Washer Prods. Liab. Litig., 722 F.3d 838, 851 (6th Cir.2013) (quoting Wal-Mart Stores, Inc. v. Dukes, 564 U.S.338, 131 S.Ct. 2541, 2551, 180 L.Ed.2d 374 (2011)). “The class action is ’an exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only.’” Dukes, 131 S.Ct. at 2550 (quoting Califano v. Yamasaki, 442 U.S. 682, 700-701, 99 S.Ct. 2545, 61 L.Ed.2d 176 (1979)). The burden of establishing the propriety of class certification rests with the proponent of the class. Beattie v. CenturyTel, Inc., 511 F.3d 554, 560 (6th Cir.2007). A class may be certified only if (1) the class is so numerous that joinder of all members would be impracticable (“numerosity”); (2) there are questions of fact and law common to the class (“commonality”); (3) the claims of the representatives are typical of the claims of the unnamed members (“typicality”); and (4) the named representatives will be able to represent the interests of the class adequately (“adequacy”). Fed. R. Civ. P. 23(a).

In addition to the requirements of Rule 23(a), at least one of the three alternative subsections of Rule 23(b) must be satisfied before the class can be certified. Fed. R. Civ. P. 23(b); Sprague v. General Motors Corp., 133 F.3d 388, 397 (6th Cir.1998) (“No class that fails to satisfy all four of the prerequisites of Rule 23(a) may be certified, and each class meeting those prerequisites must also pass at least one of the tests set forth in Rule 23(b).”). Here, plaintiffs seek certification under Rule 23(b)(3), which requires them to demonstrate (1) that questions of law or fact common to class members predominate over any questions affecting only individual members, and (2) that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy (“predominance” and “superiority”). Fed. R. Civ. P. 23(b)(3); In re Whirlpool, 722 F.3d at 850-51.

Since Rule 23 does not impose a mere pleadings standard, it may be necessary for the court to probe behind the pleadings in conducting its certification inquiry and to conduct “a rigorous analysis” of the evidence. Dukes, 131 S.Ct. at 2551; Gen. Tele. Co. of Southwest v. Falcon, 457 U.S. 147, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982). Evidentiary proof is required to show compliance with Rule 23, Comcast Corp. v. Behrend, — U.S. -, 133 S.Ct. 1426, 1432, 185 L.Ed.2d 515 (2013), and ’“rigorous analysis’ may involve some overlap between the proof necessary for class certification and the proof required to establish the merits of the plaintiffs’ underlying claims.” In re Whirlpool, 722 F.3d at 851 (citing Dukes, 131 S.Ct. at 2551). But inquiry into the merits is limited at the class certification stage and merits questions may be considered only to the extent relevant to the Rule 23 prerequisites. Id. at 851 (citing Amgen, Inc. v. Conn. Retirement Plans & Trust Funds, — U.S. -, 133 S.Ct. 1184, 1194-95, 185 L.Ed.2d 308 (2013)). “[A] court may not refuse to certify a class on the ground that it thinks the class will eventually lose on the merits.” Loeb Ind., Inc. v. Sumitomo Corp., 306 F.3d 469, 480 (7th Cir.2002) (citing Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974)).

A. The Rule 23(a) Requirements

1. Numerosity

The class must be so numerous that joinder of all members is impractical. Fed. R. Civ. P. 23(a)(1). According to plaintiffs, this requirement is met because the proposed class contains thousands of members. Generally, the number of members of the proposed class, if more than several hundred, easily satisfies the requirements of Rule 23(a)(1). Bacon v. Honda of Am. Mfg., Inc., 370 F.3d 565, 570 (6th Cir.2004); see also Bittinger v. Tecumseh Prods. Co., 123 F.3d 877, 884 n. 1 (6th Cir.1997) (joinder of parties impracticable for class with over 1100 members and “[t]o reach this conclusion is to state the obvious.”). Defendants do not appear to contest plaintiffs’ position on numerosity. The requirements of Rule 23(a)(1) are met.

2. Commonality

Rule 23(a)(2) provides that “[o]ne or more members of a class may sue or be sued as representative parties on behalf of all members only if ... there are questions of law or fact common to the class,” Fed. R. Civ. P. 23(a)(2), and commonality requires the plaintiffs to demonstrate that the class members ’have suffered the same injury.’” Dukes, 131 S.Ct. at 2551 (quoting Falcon, 457 U.S. at 157, 102 S.Ct. 2364). “[Plaintiffs must show that their claims ’depend upon a common contention’ that is ’of such a nature that it is capable of classwide resolution — which means that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke.’” Rikos, 799 F.3d at 505 (quoting Dukes, 131 S.Ct. at 2551) (emphasis in original). One common question is sufficient. Powers v. Hamilton County Pub. Defender Com’n, 501 F.3d 592, 619 (6th Cir.2007).

In antitrust cases, the commonality requirement is often easily met. “Price-fixing conspiracy cases by their very nature deal with common legal and factual questions about the existence, scope, and extent of the alleged conspiracy.” In re Foundry Resins Antitrust Litig., 242 F.R.D. 393, 404-05 (S.D.Ohio, 2007) (citing In re Workers’ Comp., 130 F.R.D. 99, 105 (D.Minn.1990)). In this case, plaintiffs argue that “their claims, along with those of all other absent class members, depend upon a common contention: that defendants conspired to, and did, prevent NDH from becoming a strong competitor, and that defendants agreed not to compete on price as vigorously during the conspiracy.” [Doc. 699 at 13]. Defendants do not appear to challenge plaintiffs’ assertion that the commonality requirement has been met and the Court finds that the requirements of Rule 23(a)(2) is met.

3. Typicality

Rule 23(a)(3) requires plaintiffs to show that “the claims or defenses of the representative parties are typical of the claims or defenses of the class.” Fed. R. Civ. P. 23(a)(3). Often, “[t]he commonality and typicality requirements of Rule 23(a) tend to merge.” Rikos, 799 F.3d at 508 (quoting Dukes, 131 S.Ct. at 2551 n. 5). There are differences, however. Commonality traditionally refers to characteristics of the class as a whole, while typically “refers to the individual characteristics of the named plaintiff in relation to the class.” Prado-Steiman, ex rel. Prado v. Bush, 221 F.3d 1266, 1279 (11th Cir.2000). Typicality is ordinarily established in the antitrust context when the named plaintiffs and all class members allege the same antitrust violation by defendants. Thomas & Thomas Rodmakers, Inc. v. Newport Adhesives & Composites, Inc., 209 F.R.D. 159, 164 (C.D.Cal.2002); (citing In re Playmobil Antitrust Litig., 35 F.Supp.2d 231, 244 (E.D.N.Y.1998)).

Defendants argue that plaintiffs do not meet the typicality requirement of Rule 23 “for the same reasons that common issues do not predominate.” [Doc. 728 at 47]. Defendants point to “the tremendous diversity among purchasers and the competitive conditions they faced,” and argue that “there is no putative class member whose claims could be considered “typical” of all others. [Id.]. With respect to Breto, defendants assert that there are three reasons why he fails to meet the typicality (and adequacy) requirements: (1) Breto operated one small convenient store for only a short period of time between June, 2006 and December, 2008; (2) Breto purchased milk at list prices; and (3) Breto sold his business. As for Food Lion, defendants argue it fails the test for Rule 23(a)(3) too because: (1) Food Lion operated only in the Eastern portion of the relevant geographical market; (2) the competitive conditions it faced were dissimilar to those faced by smaller, local retailers, large national retailers, or regional retailers; and (3) Food Lion was the only customer of two Dean processing plants in the Southeast, which gave it significant economic leverage and its cost-plus purchasing arrangement makes it subject to a unique defense. [Id. at 48]. Plaintiffs, of course, disagree, relying on Professor Cotterill’s expert conclusion that the alleged “conspiracy would have permeated all [] transactions, causing market-wide impact.” [Doc. 699 at 15], Because these are the same issues raised by defendants in support of their argument that common issues do not predominate under Rule 23(b)(3), the Court will discuss them below and finds, for the same reason that it finds plaintiffs’ claims to lack predominance, that the typicality requirement is not met in this case.

4. Adequacy

The final requirement of Rule 23(a) is that the representative parties must “fairly and adequately represent the interests of the class.” Fed. R. Civ. P. 24(a)(4). “It is axiomatic that a putative representative cannot adequately protect the class if his interests are antagonistic to or in conflict with the objectives of those he purports to represent.” 7A Wright, Miller & Kane, Federal Practice and Procedure Civil, 3d § 1768. The adequacy inquiry under Rule 23(a)(4) serves to uncover conflicts of interests between named parties and the class they seek to represent. Amchem Products, Inc. v. Windsor, 521 U.S. 591, 625, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997). The burden of proof is on the party advocating class certification to demonstrate that class representation will be adequate. Valley Drug Co. v. Geneva Pharms., Inc., 350 F.3d 1181, 1187 (11th Cir.2003). Whenever named plaintiffs have interests that are actually or potentially antagonistic to the interests and objectives of other class members, the concern is that the named plaintiffs cannot “vigorously prosecute” the interests and objectives of the class. Valley Drug, 350 F.3d at 1189. The adequacy of representation requirement “overlaps with the typicality requirement” because a class representative has no incentive to pursue the claims of the other members absent typical claims. In re Amer. Med. Sys., Inc., 75 F.3d 1069, 1083 (6th Cir.1996).

“[Disparate groups cannot be mixed together under Rule 23(a) where the economic reality of the situation leads some class members to have economic interests that are significantly different from — and potentially antagonistic to — the named representatives purporting to represent them.” Valley Drug, 350 F.3d at 1195. And, a class cannot be certified when some members of the proposed class benefitted from the allegedly wrongful conduct. Pickett v. Iowa Beef Processors, 209 F.3d 1276, 1280 (11th Cir.2000) (“Thus, a class cannot be certified when ... it consists of members who benefit from the same acts alleged to be harmful to other members of the class.”).

As noted in the preceding section, defendants argue that plaintiffs do not meet the adequacy requirements of Rule 24(a)(4) “for the same reasons that common issues do not predominate — ” As with the typicality requirement, the Court will largely discuss defendants’ argument in the section below discussing the predominance requirement. Even absent these concerns, however, it would be difficult for the Court to find that Breto and/or Food Lion could adequately represent the proposed class here. Plaintiffs’ proposed class definition includes a wide variety of individuals and entities, from huge corporate entities like Wal-Mart to single location businesses that simply use milk as an ingredient in the products they sell. Some class members are national in scope, while some are regional and other operate on a local basis. Large volume purchasers appear to have had significant bargaining power, while the smaller operators had virtually none. Some purchasers simply purchase from a price list, while others negotiate prices or obtained “pay-to-stay” payments from Dean in exchange for business.

Breto’s adequacy is of special concern for the Court. As noted above, Breto operated a convenience store at a single location in Jonesborough, Tennessee and purchased only about $5,000 worth of processed milk between June 2006 and December 2008. Bre-to bought from Dean price lists, even though he had been told he could obtain product from a cheaper supplier. Breto has sold his business and, in the Court’s view, has little incentive to pursue the perhaps much more significant interests of other absent class members despite the fact that he is represented by very competent and qualified class counsel. So far as the Court is aware, Breto has personally had very little involvement in the litigation except for a deposition. He has not attended court hearings in the matter, including the evidentiary hearing and oral argument on the class certification motion. Indeed, the attorney representing Breto and Food Lion acknowledged at oral argument that he did not know Breto and could not represent to the Court that he had attended any hearing in the case. [Doc. 776 at 29],

Food Lion as a class representative presents other, and potentially far more serious, concerns. Food Lion purchases hundreds of millions of dollars worth of processed milk annually and has approximately 1,300 supermarkets in 11 different states. It has distribution centers in North Carolina, Pennsylvania, Virginia, Tennessee, South Carolina, and Florida. Food Lion operations are located in only the eastern part of FMMOs 5 and 7, plaintiffs’ proposed geographic market, with the rest outside orders 5 and 7. Food Lion, a large-volume customer, has been able to extract “paid to stay” concessions from Dean. In 2005, Dean agreed to pay Food Lion one-percent of 2005 net milk sales for the right to continue to supply milk products to Food Lion’s distribution centers 1 through 4, an amount that exceeded $1.6 million. Other plants have likewise made similar payments to Food Lion. Maybe most importantly, Food Lion has operated under a “cost-plus” arrangement with Dean under which Food Lion’s prices have been determined pursuant to a formula which includes the cost of raw milk, determined on published government figures, processing costs, based on figures published by the Virginia Milk Commission, plus an agreed upon profit margin. These facts raise the possibility that Food Lion may have suffered no injury from the alleged conspiracy based on its purchases pursuant to the negotiated price, apparently a majority of its sales. Other large volume purchasers apparently have similar arrangements. This situation may leave Food Lion with an incentive to settle its claims for less than the full amount of potential damages at the expense of all other absent class members. It likewise “generates unwarranted pressure” on [defendants] to settle [potentially] nonmeritorious or marginal claims.” In re Rail Freight Fuel Surcharge Antitrust Litig., 725 F.3d 244, 252 (D.C.Cir.2013). The Court will not specifically decide whether all this means plaintiffs have not met their burden to show adequacy in light of the Court’s discussion below and decision that the requirements of Rule 23(b)(3) are not met in this ease.

B. Rule 23(b)(3) Requirements: Predominance

Meeting the predominance requirement of Rule 23(b)(3) demands more than common evidence that defendants colluded to raise prices for processed milk. The plaintiffs must also show that they can prove, through common evidence, that all class members were in fact injured by the alleged conspiracy. Amchem, 521 U.S. at 623-24, 117 S.Ct. 2231. That is not to say, however, that plaintiffs must be prepared at the certification stage to demonstrate through common evidence the precise amount of damages incurred by each class member. See Dukes, 131 S.Ct. at 2558.

Rule 23(b)(3) tests “whether proposed classes are sufficiently cohesive to warrant adjudication by representation,” Amchem, 521 U.S. at 623, 117 S.Ct. 2231, but it is far more demanding than the commonality, typicality, and adequacy inquiries of Rule 23(a). Comcast, 133 S.Ct. at 1432; Amchem, 521 U.S. at 623-24, 117 S.Ct. 2231. To satisfy Rule 23(b)(3), the questions in a class action that are subject to generalized proof, and thus applicable to the class as a whole, must predominate over questions that are subject only to individualized proof. Beattie, 511 F.3d at 560.

In conducting the predominance inquiry, courts must “take into account ’the claims, defenses, relevant facts, and applicable substantive law,’ ... to assess the degree to which resolution of the elasswide issues will further each individual class member’s claim against the defendant.” Klay v. Humana, Inc., 382 F.3d 1241, 1254 (11th Cir.2004), abrogated on other grounds by Bridge v. Phoenix Bond & Indem. Co., 553 U.S. 639, 128 S.Ct. 2131, 170 L.Ed.2d 1012 (2008) (quoting Castano v. Am. Tobacco Co., 84 F.3d 734, 744 (5th Cir.1996)). “If proof of the essential elements of the cause of action requires individual treatment, then class certification is unsuitable.” In re Hydrogen Peroxide Antitrust Litig., 552 F.3d 305, 311 (3d Cir.2008) (citing Newton v. Merrill Lynch, Pierce, Fenner & Smith, Inc., 259 F.3d 154, 172 (3d Cir.2001)). Although individual treatment of the essential elements of a case precludes certification, it is not necessary that all questions of fact be common, but only that some questions are common and that they predominate over individual questions. Id.; see Amgen, 133 S.Ct. at 1196 (“Rule 23(b)(3) ... does not require a plaintiff seeking class certification to prove that each element of her claim is susceptible to classwide proof,”) (internal quotations omitted).

A “close look” must be taken at whether common questions predominate over individual ones and a “rigorous analysis” must be conducted that may “entail overlap with the merits of the plaintiffs underlying claim.” Comcast, 133 S.Ct. at 1432 (internal quotations omitted). Free-ranging merits inquires are not permitted at the certification stage, however. Amgen, 133 S.Ct. at 1194-95. “Merits questions may be considered to the extent — but only to the extent — that they are relevant to determining whether the Rule 23 prerequisites for class certification are satisfied.” Id. at 1195.

The predominance inquiry begins with the elements of the underlying cause of action. Erica P. John Fund, Inc. v. Haliburton Co., 563 U.S. 804, -, 131 S.Ct. 2179, 2184, 180 L.Ed.2d 24 (2011). As set forth above, plaintiffs have a single claim remaining in this case, that is, a claim of conspiracy to violate the antitrust laws. To prevail on their antitrust claims based on allegations of a conspiracy, plaintiffs must demonstrate (1) a violation of antitrust laws (i.e. the conspiracy), (2) direct injury (or impact) from the violation, and (3) measurable damages. See Hydrogen Peroxide, 522 F.3d at 311; In re Polyurethane Foam Antitrust Litig., 2014 WL 6461355 at *8 (N.D.Ohio Nov. 17, 2014).

1. The Conspiracy

“Predominance is a test readily met in certain cases alleging ... violations of the antitrust laws” and generally “proof of the conspiracy is a common question that is thought to predominate over the other issues of the case ...” In re Scrap Metal Antitrust Litig., 527 F.3d 517, 532, 535 (6th Cir.2008) (quoting Amchem, 521 U.S. at 625, 117 S.Ct. 2231 and citing 7AA Wright & Miller § 1781). Plaintiffs allege here that defendants engaged in a single, market wide conspiracy to reduce competition in the sale of fluid white milk, which they will prove with evidence common to the class. “Courts have fairly consistently found ... that common issues regarding the existence and scope of the conspiracy predominate over other questions affecting only individual members in antitrust price fixing cases.” In re Southeastern Milk Antitrust Litig., 2010 WL 3521747 at *9 (E.D.Tenn. Sept. 7, 2010). That makes sense because determination of the conspiracy issue will focus on the conduct of the defendants, not the individual class members. See Merenda v. VHS of Mich., Inc., 296 F.R.D. 528, 548 (E.D.Mich.2013). The existence of a conspiracy is central to the claims of all class members and thus is appropriate for resolution generally on a classwide basis. Since the parties have focused on the second element, impact, and because that element poses the more serious impediment to certification, the Court will likewise focus its attention there.

2. Impact

In addition to proving an antitrust violation, plaintiffs must also establish “actual” injury “attributable to an antitrust violation,” J. Truett Payne Co., Inc. v. Chrysler Motors Corp., 451 U.S. 557, 561-62, 101 S.Ct. 1923, 68 L.Ed.2d 442 (1981) (citing Perkins v. Standard Oil Co., 395 U.S. 642, 648, 89 S.Ct. 1871, 23 L.Ed.2d 599 (1969)), commonly referred to as antitrust impact. The Third Circuit has described the element of antitrust impact as follows:

... [individual injury (also known as antitrust impact) is an element of the cause of action; to prevail on the merits, every class member must prove at least some antitrust impact resulting from the alleged violation. In antitrust cases, impact often is critically important for the purpose of evaluating Rule 23(b)(3)’s predominance requirement because it is an element of the claim that may call for individual, as opposed to common, proof. Plaintiffs’ burden at the class certification stage is not to prove the element of antitrust impact, although in order to prevail on the merits each class member must do so. Instead, the task for plaintiffs at class certification is to demonstrate that the element of antitrust impact is capable of proof at trial through evidence that is common to the class rather than individual to its members. Deciding this issue calls for the district court’s rigorous assessment of the available evidence and the method or methods by which plaintiffs propose to use the evidence to prove impact at trial.

Hydrogen Peroxide, 552 F.3d at 311-12 (citations omitted).

“Establishing causation” or “fact of damage” requires plaintiffs to demonstrate a causal connection between the antitrust violation and the antitrust plaintiff, and the class members, through proof common to the class. Bell Atlantic Corp. v. AT & T Corp., 339 F.3d 294, 302 (5th Cir.2003). See also Bail Freight, 725 F.3d at 252 (“The plaintiffs must also show that they can prove, through common evidence, that all class members were in fact injured by the alleged conspiracy.”); Rikos, 799 F.3d at 507 (affirming the “normal rule” that named plaintiffs must show “’that they can prove, through common evidence, that all class members were in fact injured by the alleged conspiracy.’”) (quoting Rail Freight, 725 F.3d at 252) (emphasis added). While impact must be determined by a common methodology or evidence, Rule 23(b)(3) does not require identical damages for each class member. Messner v. Northshore Univ. HealthSystem, 669 F.3d 802, 815 (7th Cir.2012). In other words, plaintiffs must show that every class member was impacted to some degree by the antitrust violation.

Whether the plaintiffs have offered a reliable method of proof of classwide impact is at the heart of the parties’ dispute on this motion. Plaintiffs first argue that “impact is often presumed to be common to a class in conspiracy cases,” but even if not presumed, plaintiffs “have gone well beyond the presumption and have provided an expert methodology to prove impact on a class-wide basis.” [Doc. 731 at 45-46]. To prove common impact and the accompanying damages, plaintiffs offer the expert testimony of Professor Ronald W. Cotterill; in rebuttal defendants offer the testimony of Professor Joseph P. Kalt. These experts, both of whom testified at the evidentiary hearing on this motion, very predictably disagree. Professor Cotterill has conducted a regression analysis using data from defendants’ records which plaintiffs assert shows antitrust impact across the whole class. Defendants attack the Cotterill regression analysis in numerous ways and with various iterations of the same argument. The Court, however, will address those which are of most concern with respect to the class certification motion.

a. Plaintiffs’ Damages Model

Professor Cotterill employs a “multiple regression reduced form spatial model” to estimate damages. According to Professor Cotte-rill, his estimate require a two-step process. He first “estimate[s] the regression model;” then he “calculate^] but-for prices predicted by this model once competition is restored to levels existing prior to the anticompetitive conduct” alleged by the plaintiffs in this case. The but-for price results in an “overcharge” which is multiplied by purchases to obtain “estimated damages.” The Sixth Circuit briefly discussed Professor Cotterill’s regression analysis in its opinion and noted that “[a] multiple regression analysis is useful in quantifying the relationship between a dependent variable (i.e. the price of milk) and independent variables (e.g. energy costs and/or demand factors).” 739 F.3d at 285 (citing Wiesfeld v. Sun Chemical Corp., 84 Fed.Appx. 257, 261 n. 3 (3d Cir.2003)). The model used by Professor Cotterill is similar to the regression model used by the Department of Justice in its analysis of the 2001 Dean-Suiza merger.

The model estimated its “price observations” from sales data from the defendants. Professor Cotterill calculates the price of milk by customer zip code, by product and facility from which it was produced using approximately seven million price observations/data points. Defendants produced sales data from 2000 through 2007, two years before the beginning of the conspiracy alleged by plaintiffs (i.e. two years before the Dean-Suiza merger which was completed in December, 2001), and five years during which competition was allegedly lessened. The model relies on several supply factors, such as the cost of raw milk, electricity and diesel fuel costs, cost of resins, a dairy manufacturing wage rate, the prime interest rate, and limited demand factors measured by per cap-ita income. The modeling also uses certain “fixed effect variables” to account for such things as “fat content, package size, brand name, customer location, or production facility” and other factors which could potentially affect milk prices.² The model then measures the before and after effect of all these variables on the price of processed milk to predict what the price of processed milk should be for the “after” period. Professor Cotterill then concludes based on his modeling that the price of processed milk in the after period is 7.2 percent higher than what is explained by the supply/demand variables and thus attributable to the lessening of competition. Professor Cotterill concludes that “in virtually all areas of orders 5 and 7 we did find an impact. In some areas, its more than in others.”

Professor Cotterill adds certain competition variables to his model to “measure the validity of plaintiffs’ allegations,” where, in a competitive market, the presence of a plant operating nearby would indicate lower prices to a customer. Professor Cotterill adds these “explanatory variables” to measure the impact of a Defendant plant within 200 miles of a customer, “because defendants’ data show that over 90% of processed milk sales occur to zip codes within 200 miles of a customer.” Professor Cotterill includes the following competition variables in his regression model: (1) DFA JV within 200 miles, pre-eon-duct; (2) DFA JV plant within 200 miles, post-conduct; (3) Legacy Dean plant within 200 miles, pre-conduct; (4) Suiza plant within 200 miles, pre-conduct; (5) Dean-Suiza plant within 200 miles, post-conduct; (6) NDH plant within 200 miles, post-conduct; and (7) Independent plant within 200 miles.

b. Assumed Impact/Averaging

According to Professor Cotterill, his analysis of nearly seven million pieces of sales data received from defendants, using his regression model, establishes that “99.9 percent of the [proposed] class members were affected by this overcharge,”³ thus meeting the requirement that plaintiffs offer a reliable method of proof of classwide impact. The late Professor Catherine Morrison Paul, defendants’ original expert, and Professor Kalt respond that Professor Cotterill simply assumes common impact and his economic model only calculates the effect of Defendants’ alleged unlawful activity as a forced average across all types of products within each zip code. As Professor Kalt explains, Professor Cotterill’s model “analyzes only average zip code-level prices for products and plants, and calculates only an average zip code-level overcharge on average milk purchases. Professor Cotterill then assumes that the average overcharge apples to each individual class member and each product in each zip code.” Professor Cotterill predictably responds that he neither assumes impact nor does he average.

What Professor Cotterill does in fact do is quite clear. He takes nearly seven million pieces of sales data produced by defendants and uses his regression model to find a common impact for each zip code in orders 5 and 7, even though he has no before and after data for all zip codes. He comes to his conclusion based on his further conclusion that both the supply and demand and competition variables “ripple” across all zip codes in orders 5 and 7, what defendants refer to as “assumed” impact.

Professor Kalt, using Professor Cotterill’s regression model, ran the same data used by Professor Cotterill, but on a zip code by zip code basis. Of the 5,489 zip codes in orders 5 and 7, Professor Cotterill’s model shows no evidence of injury for 1,446 (26.3%) zip codes. This, Professor Kalt says, means that Professor Cotterill “is only able to assert that these zip codes [and the individual class members within them] suffered injury because his modeling assumes that the facts of zip codes where his model can calculate effects of defendants’ allegedly unlawful conduct stand for the facts of the zip codes where his model cannot calculate such effects.” Furthermore, only a little more than half (56.8%) of the 5,489 zip codes in orders 5 and 7 have positive and statistically average overcharges when the regression model is run on a zip code by zip code bases. Of the remaining zip codes where the model can estimate overcharges, 571 zip codes have either zero or not statistically different from zero zip code average overcharges, while 356 zip codes have statistically significant negative overcharges.

Professor Kalt also tested, again using Professor Cotterill’s regression model, the conclusion of Professor Cotterill that defendants’ alleged non-competitive conduct resulted in the overcharges. Professor Cotte-rill’s analysis looked specifically at the impact of eight plant closures alleged by plaintiffs to be at the core of defendant’s allegedly unlawful conspiracy to lessen competition. For that analysis, Professor Cotterill’s model is constructed to apply the same effect of a plant closure within a 200 mile radius of the plant, a radius chosen by Professor Cotterill, as noted above, “because Defendants’ data show that over 90% of processed milk sales occur to zip codes located within 200 miles of the facility.” To test Professor Cotterill’s conclusion, Professor Kalt applied Professor Cotte-rill’s model to his data, once again by zip code, for the ten largest zip codes by transaction volume. When run this way, Professor Cotterill’s model is not capable of providing any information on overcharges at the level of individual class member purchases 60 to 100 percent of the time. The model finds positive and statistically significant overcharges for only about one-quarter of class member purchases, negative and significant changes up to 5% of the time, and damages which are zero 15 percent of the time. Thus, Professor Kalt concludes, the Cotterill model does not provide a reliable basis for common classwide impact.

Despite Professor Cotterill’s conclusory claims that his model does not assume common impact or rely on averaging, the Court finds, regardless of the nomenclature used, that the model does in fact assume, largely by the coefficients assigned to competitive variables, i.e. changes in plant ownership or plant closures, common impact and does employ averaging,⁴ in the ordinary sense of the word, to find impact within zip codes where data is otherwise insufficient to draw the conclusion of impact or where the available data for that particular zip code shows positive benefits or neutral impact. Professor Cotterill protests that he does not do averaging but his own report refers to average price effect by zip code for all changes in “spatial” competition (which the Court understands to refer to changes in plant ownership or plant closures within the model’s 200 mile radius), referred to by Professor Cotterill as “proxies” for changes in competition and states that the model “aggregates” the individual customer’s purchases from a particular plant by “taking the average price.” Professor Cotterill also candidly acknowledges that the model cannot distinguish between market power exercised unilaterally from that exercised by coordination. The model simply finds an effect that correlates with the change in ownership. Professor Cotterill also notes that his model is constructed to allow damage estimates to be recalculated if the jury ultimately decides that certain plant closures were for normal business reasons while others closed as the result of an illegal price fixing conspiracy.

All of this is particularly problematic in light of the geographic terms of the proposed class, i.e. orders 5 and 7. The Court, of course, looks to the merits of plaintiffs’ claim only to the extent necessary to resolve the class certification motion and leaves most merit questions for resolution by a jury. Resolution of the major question, that of common impact, however, does not actually require the Court to resolve disputed facts. Professor Kalt analyzed the very same data as Professor Cotterill by the same methods used by Professor Cotterill but relying on actual customer-level data rather than the aggregates based on average prices used by Professor Cotterill. When the data is analyzed based on actual prices from the data set, the average price paid by a customer in orders 5 and 7 for 19% of the purchases was lower during the class period than before it, and there was no statistically significant difference during the class period than before it. Thus, only 52.4% of the purchases showed any statistically significant increase in prices and, when weighted by volume, only 39.9% of the transactions were statistically higher than before the class period began. Professor Cotterill’s model thus cannot prove impact for all, or even substantially all of the putative class members, largely because the model assumes the same impact from being located in a zip code within 200 miles of a post-merger Dean plant. [“Average prices] say [ ] nothing about the actual price paid by each impacted class member. Average prices falter as a method for proving class-wide injury, because averaging by definition glides over what may be important differences.” See Sheet Metal Workers Local 441 Health & Welfare Plan v. GlaxoSmithKline, PLC, 2010 WL 3855552 at *30 (E.D.Penn. Sept. 30, 2010) (citing Reed v. Advocate Health Care, 268 F.R.D. 573, 591-92 (N.D.Ill.2009) (quoting In re Graphics Processing Units Antitrust Litig., 253 F.R.D. 478, 494 (N.D.Cal.2005)).

C. False Positives/Formula Pricing

Beginning before the onset of the alleged conspiracy⁵ and continuing throughout the entire conspiracy period, Food Lion purchased a significant portion, approximately 80%, of its processed milk pursuant to oral “cost-plus” pricing arrangements.⁶ These “private-label” purchases were apparently made from two Dean plants in Winston-Salem and High Point, North Carolina. The prices paid by Food Lion were determined by three factors: (1) raw milk prices announced by FMMO and/or analogous state agencies,⁷ (2) a processing cost element based on a monthly processing cost survey published by the Virginia Milk Commission, and (3) an agreed-upon profit margin.⁸ Two of these components are indisputably beyond the effects of the alleged conspiracy. Professor Cotterill acknowledged these “formula pricing” arrangements in his March 5, 2010 expert report:

11.4. The impact of formula pricing

Many purchasers of milk have “contracts” that link the price of processed milk to known cost factors. This type of pricing is referred to as formula pricing. Although the phrase “formula pricing” refers to a variety of arrangements between bottlers and their customers, a primary component of formula pricing for retailer customers is the cost of raw milk. Review of written formula pricing contracts entered into by Dean and NDH shows that typically pricing adjustments for retail customers are made on a monthly basis. Although the formulas in these written contracts vary regarding the factors taken into account (e.g. some formulas include the price of resin, others do not), the formulas invariably contain a provision for an adjustment in price tied to monthly changes in the cost of raw milk. Further, the cost of raw milk in the formula normally includes the announced Federal Milk Marketing Order price (or state order, as the ease may be), plus any applicable premiums (e.g. over-order premiums) or surcharges charged by cooperatives. Accordingly, as the cost of raw milk changes, so does the price of processed milk for retailer customers whose prices are formula-based. In some cases, changes in costs may have a delayed effect on a customer’s formula price.

Formula pricing is common in the processed milk industry, even though the majority of pricing arrangements are not based on a formal written contract. Major customers such as Albertsons, Costco, Bi-Lo, Rite-Aid, Sodexho, and Food Lion are all formula pricing customers who were automatically billed for changes in the over-order premium. Rick Beamon, a former executive at NDH as well as former Chief Operating Officer of Dean’s Southwest Division, testified that 73 to 78 percent of NDH’s total customer base had prices determined by a formula. Similarly, he also testified that close to 80% of Dean’s customers under his management had a formula pricing arrangement. Although such arrangements may, in some eases, take the form of a written contract, Dean has stated that it does not maintain formal written contracts with the majority of its customers. For example, Dean’s Dairy Fresh-Winston-Salem plant manager Byron Meredith testified he was unaware of a formal contract document ing Dairy Fresh’s formula pricing agreement with Food Lion, a longstanding formula-pricing customer.

Although the “formula” or calculation used may vary across customers or plants, inputs typically include raw milk costs. Food Lion’s current formula price, for example, relies on a blend of prices from the Virginia Milk Commission (“VMC”) and FMMO prices charged to Dairy Fresh by DFA as inputs based on the percentage of Dean’s sales to Food Lion in Federal Order 5 and the unregulated portion of Virginia. The contract between NDH and The Pantry states that The Pantry’s unit price will increase or decrease by an amount equal to the sum of the change in NDH’s unit cost of raw milk, the change in cost of ingredients and packaging, and the change in the cost of resin. The cost of raw milk is “determined based on the applicable Federal Milk Market Order price for raw milk plus or minus cooperative adjustments.”

As described in Section 11.3, raw milk is the principal cost component of processed milk, and given the common use of formula pricing, one would expect changes in processed milk prices to follow changes in raw milk prices. Thus, Figure 7 compares the price paid by Food Lion for gallon-sized private label whole milk from Dean’s Dairy Fresh-Winston-Salem facility to Dean’s estimated raw milk cost. Estimated cost to Dean is calculated as the sum of the Class I Federal Minimum, applicable statutory transportation, administration, and hurricane assessments, and Charlotte announced Coop Class I over-order premium. Not surprisingly, the two lines generally exhibit similar movements. While processed milk prices do respond to changes in raw milk costs, the gap between processed and raw milk prices on Figure 7 increases over time, indicating that processed prices have increased more than raw milk costs would otherwise indicate. For example, in July 2001, the graph shows that Food Lion’s sales price of processed milk was $2.28, while raw milk cost was $1.68, resulting in a difference of $0.60. However, in July 2006, the graph shows that the sales price of processed milk was $2.32, while raw milk cost was $1.50, resulting in a difference of $0.82. Thus, prices have been increasing relative to raw milk costs. I explore this further in Section 111.5.2 and Section V.3 below.

[[Image here]]

111.5.2. Formula pricing does not immunize retailer Plaintiffs from harm

Defendants claim that impact to class members arising from the anticompetitive conduct at issue cannot be isolated and quantified because “to the extent Plaintiffs suffered any injury or economic loss as alleged in their Amended Complaint, those injuries would arise from a variety of factors unrelated to any allegedly illegal conduct by Dean.” Specifically, Defendants incorrectly claim that retailers whose processed milk prices were set according to a pricing formula were immune from Defendants’ conduct, stating:

“Some retailers (including Food Lion) can and do negotiate pricing formulas or mechanisms pegged to various objective factors. Fluctuations in price to these customers are a function of changes in the values of the underlying objective factors, not any alleged antitrust violation.”

I examine Defendants ’ claim empirically and find that readily-available data contradict their claim and instead support the premise that retailers with formula pricing arrangements were impacted.

If Defendants’ claims were true, one would expect to observe Food Lion’s price moving similarly to raw milk cost, the “the underlying objective factor” in Food Lion’s pricing formula calculation described in Section II.4. In other words, the spread between price and cost would remain constant over time. In order to test Defendants ’ claims empirically, Figure 11 examines the price paid by Food Lion for private label one gallon whole milk from Dean’s Winston-Salem facility to customers located in Orders 5 and 7 (upper line) calculated from Defendants’ sales data. It also shows the raw milk cost over time (lower line), calculated as the Class I Federal Minimum milk price, plus the zone diffex-ential applicable at Winston-Salem, the Charlotte, NC announced Coop Class I over-order premium, and applicable Federal Order 5 assessments. The difference between these two lines is an estimate of the amount above the formulaic cost that Food Lion pays.

[[Image here]]

Figure 12 shows the difference over time between the price paid by Food Lion and the raw milk cost to Dean illustrated on Figure 11. The difference is the plant’s gross margin and as such represents the amount above raw milk cost that Food Lion paid for private label one gallon whole milk from the Winston-Salem facility. The yellow trend line indicates that there is a positive linear trend calculated from 2002 onward. Thus the plant’s gross margin on Food Lion sales increased over time, notably during the period 2004 through 2006. The drop m 2007 corresponds to Food Lion’s negotiations with. Dean to reduce the price paid for milk; however, this graph indicates the price/cost difference in 2007 was still higher than it was in 2003 and earlier. This indicates that Food Lion’s price increased relative to raw milk costs.

Figure 12: Difference between raw milk cost and Food Lion private label one gallon whole milk price from the Winston-Salem plant

[[Image here]]

V.3. Comparison of processed milk prices and raw milk costs

As described earlier, raw milk is the largest component of production costs for processing milk. Thus, it is instructive to examine how processed milk prices move in relation to raw milk costs. Admittedly, other production inputs can (and do) affect milk processing, hence why they are included as control variables in my statistical model. However, this section focuses only on raw milk costs to serve as a corroborating analysis examining whether the results of my regression analysis are consistent with a less sophisticated approach I examine the difference between processed milk prices and raw milk costs. To the extent competition remained largely the same, and other cost factors changed little (or had a small impact), then one would expect a similar difference over time between processed milk prices and raw milk costs. However, if competition in the Southeast decreased, as Plaintiffs claim, then Defendant bottlers’ increase in market power may have enabled them to raise processed milk prices without fear of losing sales to a rival. If Plaintiffs’ allegations are true, and Defendants conduct increased their ability to raise prices, then one would expect to observe a widening gap between processed milk prices and raw milk costs over time.

Figure 32, repeated from Figure 11 in Section II1.5.2 above, shows the price paid by Food Lion for private label one gallon whole milk from Dean’s Winston-Salem facility (upper line) calculated from Defendants’ sales data. It also shows the raw milk cost over time (lower line), calculated as the Class I whole milk price, plus the base zone (Meck-lenburg County, NC) differential, the Charlotte, NC announced over-order premium, and applicable Federal Order 5 assessments.288 In July 2001, the graph shows that the sales price of processed milk was $2.28, while raw milk cost was $1.68, resulting in a difference of $0.60. Over the 2000 to 2001 pre-conduct period, this difference was $0.56 on average. However, in post-conduct July 2006, the graph shows that the sales price of processed milk was $2.32, while raw milk cost was $1.50, resulting in a difference of $0.82. Thus, Dean’s markup of processed milk price over raw milk cost in July 2006 was $0.26 higher than the average 2000-2001 markup.

Figure 33 shows prices generated when assuming that the 2000-2001 processed-to-raw milk difference persisted from 2002 to 2007, when Defendants’ sales data ends. The yellow line shows the same processed milk price shown on Figure 32, while the red line shows the price that would have prevailed if the processed-to-raw milk difference remained at 2000-2001 levels ($0.56 on average).

[[Image here]]

The two lines in Figure 33 show that actual milk prices on average were $0.15 higher, or 6.0%, than movements in raw milk costs would suggest. Replicating the analysis in Figure 33 using only the Federal minimum for raw milk costs, i.e., excluding the over-order premium and other assessments, results in an average difference of 8.8%.

As illustrated by these excerpts from Professor Cotterill’s expert report, Professor Cotterill ties his estimate of impact from these purchases to the alleged conspiratorial conduct of defendants related to raw milk prices established by government agencies and unrelated to the alleged conspiracy. Defendants therefore argue that the “overcharge” found by Professor Cotterill cannot have anything to do with the alleged conspiracy, resulting in a false positive as a result of the analysis of the data by Professor Cotte-rill’s model, and the fact that he does not find zero overcharges for these purchases means the model does not reliably measure the effects of the alleged conspiracy. Defendants rely largely on Rail Freight In Rail Freight, the district court certified a class of railroad shippers who accused the nation’s largest freight railroads of conspiring to fix fuel surcharges. As in this case, the plaintiffs offered a regression model designed by their damages expert that purported to estimate the portion of fuel surcharges attributable to the conspiracy. The debate over certification “centered on the predominance requirement and whether the plaintiffs could show, through common evidence, injury in fact to all class members from the alleged priee-fixing scheme, ...” 725 F.3d at 249 (emphasis added). That debate, in turn, focused on whether the damages model yielded false positives because it found injury to certain “legacy” shippers who were bound by rates negotiated before any conspiratorial behavior was alleged to have occurred.

Citing Amchem, the D.C. Circuit noted that “[mjeeting the predominance requirement demands more than common evidence the defendants colluded to raise [prices]. The plaintiffs must also show that they can prove, through common evidence, that all class members were in fact injured by the alleged conspiracy.” Id. at 252 (citing Amchem, 521 U.S. at 623-24, 117 S.Ct. 2231). The D.C. Circuit found that while plaintiff need not be prepared at the certification stage to prove the amount of damages incurred by each class member, they must, through common evidence, show that “all class members suffered some injury.” Id. (emphasis in original). See also Rikos, 799 F.3d at 522. The D.C. Circuit vacated the district court’s grant of class certification because the model found damages as to the legacy shippers, producing false estimates and “detect[ing] injury where none could exist.” Id. at 252. Defendants argue that Professor Cotterill’s model suffers from the same infirmity and the Court agrees. Neither Professor Cotterill nor plaintiffs’ counsel have provided the Court with a plausible explanation for how Food Lion, or any other purported class member with a negotiated formula, has suffered injury from the alleged conspiracy, although they have attempted to do so. At oral argument, plaintiffs’ counsel suggested that the formula pricing agreements in this case were different from the Legacy contracts in Rail Freight because the contracts here were oral and terminable at will. Plaintiffs argue that Food Lion’s injury therefore was loss of the opportunity to negotiate a better contract with another supplier because of the conspiracy. There are two problems with the explanation, however. First, Food Lion never sought to terminate its agreements with Dean, but rather was able to negotiate and receive large rebate payments and other concessions from Dean under threat of termination of the formula-pricing agreements, belying Food Lion’s argument that it had no other supply options. Second, no evidence in the record supports counsel’s claim.

Other courts have come to similar conclusions. For instance, in Piggly Wiggly Clarksville, Inc. v. Interstate Brands Corp., the Fifth Circuit affirmed the district court’s denial of a class certification motion where “many of the class members negotiated a price rather than being charged strictly on price lists.” 100 Fed.Appx. 296, 297 (5th Cir.2004). Even though plaintiffs had offered an expert’s view, as here, that damages could easily be calculated, the district court was not required to accept the expert’s opinion, and was within its discretion to deny class certification given the degree of inquiry needed into the individual facts of thousands of class members and potentially thousands of transactions. Id. at 299.

Here, inquiry will be needed as to which members of the class had negotiated prices, the terms of their negotiated agreements, when the formula was established, and the extent of their damages, if any, as to both formula-pricing and non-formula-pricing purchases.⁹ Counsel for plaintiffs reluctantly conceded at oral argument that the issues related to Food Lion’s formula-pricing agreements would likely need to be submitted to the jury separately, likely through interrogatories, and that the jury could find that Food Lion suffered no injury as to these negotiated transactions. This, of course, raises the possibility that the jury could find in favor of Food Lion with respect to its non-negotiated transactions but against Food Lion as to its formula-pricing transactions. The same would ultimately be true of all other purchasers of processed milk, a vast majority of the class apparently, who bought pursuant to negotiated prices and/or received pricing concessions, rebates, and the like. These damages issues would ultimately overwhelm the Court’s inquiry into questions common to the class. See Comcast, 133 S.Ct. at 1433. Predominance and manageability (relevant on the superiority question) may be destroyed solely by the complexity of determining damages. See Windham v. Am. Brands, Inc., 565 F.2d 59, 67-68 (4th Cir.1977). See also O’Sullivan v. Countrywide Home Loans, Inc., 319 F.3d 732, 745 (5th Cir.2003) (district court abused discretion in certifying class “[i]n light of the individual calculation of damages that is required.”). Under these circumstances, it is hai'd to say that a class action here will lead to economies of time and expense.

In Deiter v. Microsoft Corp., the Fourth Circuit affirmed a district court order excluding from a certified class purchasers of software licenses through Microsoft’s enterprise program who had negotiated individual three-year purchase agreements. 436 F.3d 461, 465 (4th Cir.2006). The Fourth Circuit stated:

In proving their case, however, the plaintiffs would hardly prove a case on behalf of Microsoft’s Enterprise customers. These customers, who purchased at least 250 licenses, did not purchase online or by telephone, nor did they pay prices established in advance by Microsoft. The prices that Enterprise customers paid were negotiated and, as a consequence, were both discounted and

unique to each transaction.

“Moreover, to prove that Microsoft over charged the Enterprise customers would require new and different proof because the Enterprise customers were able to negotiate their deals in a different competition context from that involving the plaintiffs. ...

Id. at 468. Although the Fourth Circuit appears to have decided the Deiter case on the basis that plaintiffs claims were not typical of the claims of Microsoft’s Enterprise customers, its logic applies fully to the predominance question as well.

IY. Conclusion

For the reasons stated herein, plaintiffs have not “affirmatively demonstrate[d] their compliance with the Rule,” Dukes, 131 S.Ct. at 2551, and the motion to certify class is DENIED.

So ordered.

1 . Count 1 of the complaint was also dismissed by the Court. That decision, however, was reversed by the Sixth Circuit Court of Appeals. See Food Lion, et al. v. Dean Foods Company, et al., 739 F.3d 262 (6th Cir.2014) and "remanded for further proceedings consistent with the opinion.” Id. at 286.

2 . The model has approximately 7,000 fixed effect variables.

3 . The Court admits some confusion about Professor Cotterill’s actual conclusion on this point in light of what appears to be conflicting and contradictory testimony at the evidentiary hearing. He testified that tire overcharges "potentially" affect every potential class member, [Doc. 739 at 47], that the overcharge affects "virtually all of orders 5 and 7,” [id. at 39-40], and that "a great number of people in the Southeast, 80, 90 percent or more, suffered impact.” [Id. at 48]. Finally, he gave the testimony quoted above, leaving the Court with some doubt based on Professor Cotterill's testimony whether the claimed impact is indeed classwide.

4 . In some ways, Professor Cotterill's insistence that his model does not rely on averaging is baffling. Multiple regression analysis, by definition, relies on averaging the underlying trends in a particular data series. Fed. Jud. Center, Reference Guide On Multiple Regression (3d Ed. 2012) at 334.

5 . The formula was first established in the mid-1990s.

6 . The record does not establish which other purported class members also made purchases pursuant to a negotiated formula; however, the plaintiffs do not dispute that many, especially the larger purchases, did so. Professor Cotterill identifies some of them in his report, i.e. Albertsons, Costco, Bi-Lo, Rite-Aid, and Sodexho.

7 . Plaintiffs originally claimed in Count II of their complaint that defendants inflated raw milk prices as part of the illegal conspiracy. The Court dismissed Court II and plaintiffs have now abandoned that claim.

8 .The profit margin started out in the 5-7% range and fluctuated monthly based on audited indices published by the Virginia Milk Commission. In 2007, Food Lion demanded and received large incentive payments from Dean in order to continue the purchases and the parties switched to a fixed six percent profit margin.

9 . Plaintiffs appear to focus on the agreed upon profit margin as the part of the formula where defendants’ alleged anticompetitive conduct is reflected; however, the profit margins were generally between 6 and 7%. Professor Cotterill found average overcharges of greater than 7%, an amount equal to or greater than the actual profit margins. If this is plaintiffs’ focus, the appropriate measure of damages would likely be the difference between the "anti-competitive” margin and a "competitive” margin, not Professor Cotterill’s overcharges.