General Motors Corporation

Text (OCR of the scan at left; may contain errors)

IN THE MATTER OF GENERAL MOTORS CORPORATION MODIFYING ORDER IN REGARD TO ALLEGED VIOLATION OF SEC. 5 OF THE FEDERAL TRADE COMMISSION ACT Docket C-2966. Consent Order, May 18, 1979--Modifying Order, Dec. 21, 1995 This order reopens a 1979 consent order that settled allegations that General Motors ("GM") engaged in unfair and deceptive practices by selling cars with engines and other equipment manufactured by a different GM division without informing purchasers. This order modifies the consent order by allowing GM to display division brand nameplates on engines that are not manufactured by that GM division. In addition, the Commission deleted the provision from the modified order that have expired, concluding that elimination of the expired provisions is warranted. The Commission determined that changed conditions of fact justified reopening the proceeding and modifying the order. ORDER REOPENING THE PROCEEDING AND MODIFYING CEASE AND DESIST ORDER On July 7, 1995, General Motors Corporation ("GM") filed a petition pursuant to Section 5(b)(2) of the Federal Trade Commission Act ("FTC Act"), as amended, 15 U.S.C. 45(b)(2), and Rule 2.51 of the Commission's Rules of Practice, 16 CFR 2.51, to reopen the proceeding and modify the cease and desist order entered against it on May 18, 1979, in Docket No. C-2966 (93 FTC 860).' The cease and desist order settled allegations that GM, a worldwide manufacturer of passenger vehicles: (1) represented that certain standard and optional equipment was manufactured by the particular division that assembled the passenger vehicle when that equipment was, in fact, manufactured by another division; (2) misrepresented the availability of various standard and optional equipment; and (3) substituted equipment other than that represented as being available, and delivered passenger vehicles that contained standard or optional equipment different from that ordered by retail purchasers.

This order will sunset on May 18, 1999, provided that neither the Department of Justice nor the Commission files a complaint in federal court to enforce the order pursuant to Section 5(1) of the FTC Act prior to that date. See Final Rule Regarding the Duration of Existing Competition and Consumer Protection Orders, 60 Fed. Reg. 58514 (Nov. 28, 1995). Modifying Order 120 F.T.C.

Various parts of the order imposing obligations with respect to certain model year passenger vehicles, i.e., Parts V, VI, VII, X.B, X.C, and XI.B, have expired. The remaining provisions of the order remain in full force and effect. Part I of the order defines certain terms used in the order. Part II prohibits GM from misrepresenting the manufacturing source of any engine option and the availability of an option or item of standard equipment. Part III prohibits GM from displaying the name of any GM car division on any engine unless that engine is manufactured by that particular division. Part IV requires GM to notify dealers if passenger vehicles are being equipped with engines other than those specified in sales literature for that passenger vehicle. Part VIII requires GM to make available replacement parts and repair and maintenance information for passenger vehicles equipped with substituted engines to its dealers. Part IX of the order limits application of the order to the United States and its territories. Part X.A prohibits GM from utilizing a wholesale ordering system that prevents dealers from designating specific optional equipment requested by purchasers.

The petition to reopen the proceeding to modify Part III of the order and to add a definition to Part I was placed on the public record for thirty days on August 4, 1995, for the purpose of receiving public comment. See 60 Fed. Reg. 39,958 (1995). No comments were received. GM agreed to an extension of time for Commission action until December 20, 1995, to enable it to develop clarifying information to respond to questions raised by the Commission. STANDARD FOR REOPENING A FINAL ORDER OF THE COMMISSION Section 5(b) of the FTC Act, 15 U.S.C. 45(b), provides that the Commission shall reopen an order to consider whether it should be altered, modified, or set aside if the respondent makes "a satisfactory showing that changed conditions of law or fact" so require.” A satisfactory showing sufficient to require reopening is made when a a a Section 5(b) provides, in part:

[T]he Commission shall reopen any such order to consider whether such order (including any affirmative relief provision contained in such order) should be altered, modified, or set aside, in whole or in part, if the person, partnership, or corporation involved files a request with the Commission which makes a satisfactory showing that changed conditions of law or fact require such order to be altered, modified, or set aside, in whole or in part. The 1980 amendment to Section 5(b) did not change the standard for order reopening and modification, but "codifie[d] existing Commission procedure by requiring the Commission to reopen an order if the specified showing is made," S. Rep. 96-500, 96th Cong., 2d Sess. 9-10 (1979), and added the requirement that the Commission act on petitions to reopen within {20 days of filing. GENERAL MOTORS CORPORATION 1071 1069 Modifying Order request to reopen identifies significant changes in circumstances and show that the changes eliminate the need for the order or make continued application of the order inequitable or harmful to competition. Louisiana Pacific Corp., Docket No. C-2956, Letter to John C. Hart June 5, 1986) at 4.> Reopening may not be required if the changes were reasonably foreseeable at the time of consent negotiations. Phillips Petroleum Co., Docket No. C-1088, 78 FTC 1573, 1575 (1971).

The language of Section 5(b) plainly anticipates that the burden is on the requester to make "a satisfactory showing" of changed conditions to obtain reopening of the order. See Gautreaux v. Pierce, 535 F. Supp. 423, 426 (N.D. Ill. 1982) (requester must show "exceptional circumstances, new, changed or unforeseen at the time the decree was entered"). The legislative history also makes clear that the requester has the burden of showing, by means other than conclusory statements, why an order should be modified.* If the Commission determines that the requester has made the necessary showing, the Commission must reopen the order to determine whether the modification is required and, if so, the nature and extent of the modification. The Commission is not required to reopen the order, however, if the requester fails to meet its burden of making the satisfactory showing of changed conditions required by the statute. The requester's burden is not a light one in view of the public interest in repose and finality of Commission orders.° See S. Rep. No. 96-500, 96th Cong., 2d Sess. 9 (1979) (significant change or changes causing unfair disadvantage); Pay Less Drugstores, Inc., Docket No. C-3039, Letter to H.B. Hummelt (Jan. 22, 1982) (changed conditions must be unforeseeable, create severe competitive hardship and eliminate dangers that the order sought to remedy) (unpublished); See also United States v. Swift & Co., 286 U.S. 106, 119 (1932) ("clear showing" of changes that eliminate reasons for order or such that order causes unanticipated hardship).

4 The legislative history of amended Section 5(b), S. Rep. No. 96-500, 96th Cong., 2d Sess. 9-10 (1979), states:

Unmeritorious, time-consuming and dilatory requests are not to be condoned. A mere facial demonstration of changed facts or circumstances is not sufficient. . . . The Commission, to reemphasize, may properly decline to reopen an order if a request is merely conclusory or otherwise fails to set forth specific facts demonstrating in detail the nature of the changed conditions and the reasons why these changed conditions require the requested modification of the order.

5 See Federated Department Stores, Inc. v. Moitie, 452 U.S. 394 (1981) (strong public interest considerations support repose and finality); Bowman Transportation, Inc. v. Arkansas-Best F reight System, Inc., 419 U.S. 281, 296 (1974) ("sound basis for. . . [not reopening] except in the most extraordinary circumstances"); RSR Corp. v. FTC, 656 F. 2d 718, 721-22 (D.C. Cir. 1981) (applying Bowman Transportation standard to FTC order). Modifying Order 120 F.T.C.

CHANGED CONDITIONS OF FACT WARRANT REOPENING ORDER In 1979, when the Commission issued its order, the Cadillac, Buick, Oldsmobile, Pontiac, and Chevrolet Divisions ("the nameplate divisions") of GM were organized as semi-autonomous motor car companies under GM's North American Car and Truck Operations with individual design, marketing, and manufacturing responsibilities, including engine production. In 1984, GM consolidated its engine manufacturing operations so that the nameplate divisions that existed in 1979 no longer manufactured engines. GM Petition at 6. In 1990, GM further consolidated all of its engine manufacturing operations into a single unit, except that its Saturn division continues to manufacture its own engines. GM Petition at 2. Because GM's nameplate divisions no longer manufacture engines, GM generally is prevented from branding an engine with a division nameplate since Part III of the 1979 order permits a division nameplate to appear on an engine only if the engine was manufactured by that division.* GM maintains that Part III of the order is inequitable and harmful to competition following its reorganization. GM has submitted a market research report prepared by an independent market research organization that suggests that a substantial number of consumers would place a higher value on GM passenger cars if their engines were branded with their division nameplate. Five of GM's competitors, like GM, have separated their engine production organizations from their marketing organizations.’ Unlike GM, however, these competitors are able to brand their engines with their division nameplates rather than their manufacturing source to take advantage of any commercial value associated with the division nameplates. Because Part III of the order has become inequitable and harmful to competition, we conclude that GM has made a sufficient showing to warrant reopening the 1979 order and consideration of its requested modification.

Because the Saturn division manufacturers its own engines, it is permitted under Part III to put the Saturn nameplate on its engines.

? These companies and their marketing organizations are Ford (Ford and Lincoln-Mercury), Chrysler (Chrysler-Plymouth, Dodge, and Jeep-Eagle), Toyota (Toyota and Lexus), Nissan (N issan and Infiniti), and Honda (Honda and Acura).

GENERAL MOTORS CORPORATION 1073 1069 Modifying Order THE REQUESTED MODIFICATION GM requests that the Commission revise Part III of the order and add a definition of "nameplate" (a term that did not appear in the original order) to Part I of the order. The modification GM request would permit it to display division nameplates only on those engines that are materially different from engines in new cars displaying all other division nameplates.* The modification would not permit GM to display different division nameplates on identical engines. In addition, the requested modification would not affect other protections afforded by the order. Part II.A of the order will continue to prohibit GM from displaying a division nameplate on an engine in a manner that misrepresents the manufacturing source of an engine. Part IV of the order also will continue to require GM to notify dealers of any production plans that would result in replacing one engine with a different engine in makes of cars sold by the dealers. Finally, several provision of the order have already expired. Although GM did not request that the Commission delete these expired provisions, we conclude that elimination of these expired provisions is warranted. Accordingly, the expired provisions will be deleted from the modified order.

It is therefore ordered, That the proceeding is hereby reopened and the order modified to read:

ORDER For purposes of this order, the following definitions shall apply: A. The term "GM" shall mean General Motors Corporation, and all of its divisions, its successors, assigns, officers, representatives, agents, and employees, acting directly or through any subsidiary or other device.

B. The term “franchised GM passenger car dealer" shall mean any person, partnership, or corporation which is a party to a franchise Because it is possible that consumers would place a higher value on a GM automobile with a division nameplate on the engine because they assume that its engine is materially different from the engines sold by the other GM nameplate divisions, the modification would permit GM to display a division nameplate on an engine only if it is materially different from engines with other nameplates. Modifying Order 120 F.T.C.

agreement with GM to purchase new GM passenger cars for resale to purchasers.

C. The term "manufacturing source" shall mean the GM division or entity by which the item referred to was produced. D. The term "line" shall mean each make and model of passenger car manufactured by General Motors Corporation and distributed or sold under the Chevrolet, Pontiac, Buick, Oldsmobile or Cadillac name.

E. The term "engine option" shall mean any engine designated by a GM ordering code number (including the standard engine) offered by GM as factory-installed equipment. For purposes of this order, each engine option shall be assigned a single, unique ordering code designation for a given model year which does not vary across division lines.

F. The term "material difference" shall mean any difference which results in a significant difference in engine performance, including but not limited to any difference in Environmental Protection Agency ("EPA") fuel economy ratings, mileage intervals in excess of 1,000 miles for recommended engine maintenance, horsepower and displacement, or which results in a difference or regular maintenance replacement parts.

G. The term "substituted engine" shall mean an engine option installed in any GM line in any area of the country as a replacement for an engine option offered for that line in the same model year, but which is unavailable in such line or area, if the replacement engine option:

(1) Is produced by a division other than that which produced the engine option to be replaced; or (2) Has any "material difference" from the engine option to be replaced.

H. The term "option" shall mean an item of equipment to be installed in a new GM passenger car for which GM provides purchasers a choice of alternatives.

I. The term "purchaser" shall mean a potential buyer, potential lessee, buyer and lessee of any new GM passenger car, but shall not include a franchised GM passenger car dealer. J. The term "nameplate" shall mean the name of the franchise identity through which new passenger cars are sold by GM, such as GENERAL MOTORS CORPORATION 1075 1069 Modifying Order "Cadillac," "Buick," "Oldsmobile," "Pontiac," "Chevrolet," or "Saturn."

II.

It is hereby ordered, That GM is prohibited from misrepresenting as of the time the representation is made by GM: A. The manufacturing source of any engine option; and B. That an option or item of standard equipment offered for a new GM passenger car is available if in fact it is not. II.

It is further ordered, That GM is prohibited from displaying. a passenger car's nameplate on any engine or visible attachment to the engine unless such engine is materially different from engines in new GM cars sold under all other nameplates.

IV.

It is further ordered, That if:

A. GM furnishes or has furnished, during or in preparation for any model year, any information to any franchised GM passenger car dealers regarding any engine offered for any GM line for any model year, and B. The engine described in the information provided to such dealers is to be or has been replaced by a substitute engine for that model year, GM shall notify such dealers in writing, with respect to the affected lines handled by them, forthwith after the decision to substitute has been made. Such written notification shall include the lines in which the substituted engine is offered, its manufacturing source, ordering code number, designation used in the vehicle identification number to identify the type of engine option, and any material differences between the substituted engine and the engine to be replaced. Modifying Order 120 F.T.C.

V.

(Expired) VI.

(Expired) VIL.

(Expired) Vill.

It is further ordered, That GM shall make available, subject to force majeure, labor disruptions, and other causes outside GM's control, replacement parts and repair and maintenance information to franchised GM passenger car dealers adequate to allow such dealers to provide GM warranty service to purchasers of new GM passenger cars equipped with any substituted engine to the same extent as it does in the case of new GM passenger cars equipped with nonsubstituted engines.

IX, It is further ordered, That this order shall be limited in its application to sales of new GM passenger cars in the United States and its territories.

X.

It is further ordered, That:

A. GM is prohibited from utilizing a wholesale ordering system whereby its franchised GM passenger car dealers may not designate the specific options, other than standard equipment, requested by the purchaser. GM shall notify its dealers in writing that purchasers should be given the opportunity to designate the specific options ordered. Provided, that GM shall indicate when an option is required to be paired with another specific option. GENERAL MOTORS CORPORATION 1077 1069 Modifying Order B. (Expired) C. (Expired) XI.

It is further ordered, That:

A. GM shall notify the Commission at least thirty (30) days prior to any proposed change in the corporation such as dissolution, assignment, or sale resulting in the emergence of a successor corporation, the creation or dissolution of subsidiaries, or any other change in the corporation which may affect compliance obligations arising out of this order.

B. (Expired) R.R. DONNELLEY & SONS CO., ET AL. 105 36 Initial Decision NUMBER OF PRINT BUYERS WITH JOBS IN "CORE" IN 1990 NUMBER OF PRINT BUYERS NO LONGER IN "CORE" BECAUSE RUN LENGTH NOW LESS THAN 10 MILLION NUMBER OF PRINT BUYERS NO LONGER IN "CORE" BECAUSE PAGE COUNTS NOW LESS THAN 33 NUMBER OF PRINT BUYERS NO LONGER IN "CORE" BECAUSE WORK NOW "HIGHLY VERSIONED"

NUMBER OF PRINT BUYERS NOT IN "CORE" BECAUSE THEY WILL SWITCH TO OFFSET IN EVENT OF GRAVURE PRICE INCREASE NUMBER OF PRINT BUYERS NO LONGER IN "CORE" BECAUSE THEIR PRINT PROGRAMS NO LONGER NUMBER OF PRINT BUYERS NEVER IN "CORE" BECAUSE THEIR JOB IS NOT FOUR-COLORS NUMBER OF PRINT BUYERS NOT IN "CORE" BECAUSE THEY ARE PROCESS NEUTRAL NUMBER OF PRINT BUYERS NOT IN "CORE" BECAUSE MAGAZINE, IN WHOLE, IS "HIGHLY VERSIONED"

NUMBER OF PRINT BUYERS FOR WHOM PRINT PROCESS PREFERENCE AND PROGRAM DETAILS NOT AVAILABLE NUMBER OF PRINT BUYERS WITH JOBS IN "CORE" AT 2 TIME OF TRIAL 364. Dr. Hausman's analysis of the "core" market led him to conclude that with six gravure printers competing for the work of as few as two print buyers (or, at most, thirty-six), the possibility of price discrimination is unlikely simply because of the number of buyers and sellers (Tr. 6337). It also means that some of the printers with gravure capacity do not have "core" work (Tr. 6337-38). [ ] (RX-665-A-B). Thus, if one of the printer with "core" work attempted price discriminate, one of the printers without "core" work would take the business at a lower price (Tr. 6337-38). Because of the trend toward increased versioning, buyer consolidation, and shorter run lengths, Dr. Hausman believes that the number of participants in the relevant product market will become even smaller (Tr. 6339).

Initial Decision 120 F.T.C.

4. Entry Into High Volume Publication Gravure Printing 365. De novo entry or expansion into gravure printing takes two or more years. For example, over 24 months--from January 1985 to March 1987--were required from project approval to first publication at Donnelley's Reno facility (CX-7-K; CX-66; CX-69). 366. Other examples of entries or expansion over a two, or more, year period include:

367.[ ](CX-8-J).

368.[ ](CX-141-L).

369. Two years--from 1987 to early 1990--were required from project approval to start-up of the seventh gravure press at Donnelley's Spartanburg plant (CX-7-Q; CX-90-V). 370. Two years' lead time was required for Donnelley's most recent gravure press installation, the Warsaw tandem press (CX-9-F- G; CX-63-Z-5; see also CX-117-Z-47).

371. Two years’ lead time was required for Ringier to analyze, plan, purchase and install gravure presses at its Corinth, Mississippi facility (Tr. 1476). Its president acknowledged that the time to acquire new gravure presses generally runs in the two year range (Tr. 1479-80, 1504).

372. Over two years’ lead time (from late 1983 to 1986) was required for Quad Graphics to plan for, purchase, install and print with its first gravure press (Tr. 2351-52). Its gravure operations were not profitable until 1989 (Tr. 2367).

373. Brown Printing required three years, from "early 1989" to "May of 1992," to plan, order, install and begin operation of its new three-meter press in its Franklin, Kentucky plant (Tr. 2525). 374. Wayne Angstrom, a former Donnelley executive, estimated that it would take two to three years for a new firm to enter the gravure market (Tr. 2618-19). Mr. Walter, Donnelley's CEO, estimated that it would take two years to enter and a few more years to reach full efficiency (CX-101-Z-108) (see also CX-106-Z-38-39). Longer time frames (from two and one half to three years) were contemplated for the three-meter presses that Meredith/Burda planned to install at its Lynchburg facility (CX-55-Z-1; CX-251-B; CX-252-B).

375. Factors which contribute to entry or expansion delay include:

R.R. DONNELLEY & SONS CO., ET AL. 107 36 Initial Decision Regulations requiring that appropriate environmental permits be obtained (Tr. 2742-47, 2754-56, 2767-69). Mr. Voss, former president of Meredith/Burda, testified that clean air permits are so difficult to get that obtaining them would require at least four to six months (CX-900-Z-56-66-67). Environmental restrictions will probably be more onerous in the future (Tr. 2533-34, 2755, 2768-69). The reluctance of gravure customers to switch from their current printer to a new one (Tr. 941-42, 1351, 1943-44, 1961, 2156, 2658- 59).

Jerry Ryan, of Service Merchandise, testified: [W]hen you get involved in a large complex program like the one we happen to have, whether you're talking about the overall program or you're talking just about even the fall catalog itself, you are not going to go to the first bozo who happens to have gravure facilities; you're going to go with somebody that you know has a good track record, understands your problems, comes through in a pinch. . . And you may end paying that guy, whether it's three percent or five percent more, than you might be able to pay to somebody else that you wouldn't have the same kind of faith in.

(Tr. 4298).

The reluctance to switch is increased by the long term contracts which Donnelley and other gravure printers have with their customers (in Donnelley's case, at least [ ] of its business involves multi-year contracts) (CX-757-A; CX-1157-Z-7, Z-11,; see also CX- 63-U; CX-102-Z-7-8; CX-159-G; CX-483-Z-236). Due to these contracts, new entrants would find that much of the relevant market would be inaccessible for at least two years. Sunk costs, i.e., costs of entry that are unlikely to be recovered through the redeployment of those assets. The investment made in gravure is "sunk" in the sense that it cannot readily be recouped by sale for other uses (Tr. 4591; see Tr. 1504-05). All existing gravure facilities in the United States have at least three presses except Quebecor's Dallas facility which has two presses (Tr. 301; see CX-501). Those familiar with the industry recommend a minimum of two gravure presses per plant (Tr. 300, 2619, 2798-99, 1503-04).

Donnelley's Reno facility was opened in 1987 at a cost of approximately $94 million (see CX-7-L; CX-501-M). [ ] (CX-501- M; CPF 19).

Initial Decision 120 F.T.C.

Quad Graphics, the most recent gravure entrant (Tr. 1502-03, 3341), needed a minimum of two 96-inch gravure presses to enter high volume publication gravure printing and its total cost of entry approximated $50 million (in 1993 dollars) (see Tr. 2356, 2358, 2361, 2364).

[ ] Individual gravure presses, with associated equipment, cost approximately [ J], depending on web width and other operating parameters (see CX-11-Z-51, Z-55; CX-106-Z-104-A; CX-247-L; CX-248-F; CX-1449-F; CX-1453-A; Tr. 299, 2532-34, 6255). 376. Consolidation, rather than new entry, has occurred in the past several years. In 1981, there were twelve gravure printers in the United States; today there are six (CX-933-G; CX-501; CX-519). The most recent exit of a gravure producer, Standard Gravure, occurred in 1992 (CX-510).

377. The six gravure printers operating in the United States in 1993 are:

Brown Printing R. R. Donnelley (acquired Meredith/Burda) Krueger/Ringier Quebecor Corp. (acquired Arcata Graphics, Maxwell Communications Corp.) Quad Graphics World Color Press PYnNr nM (CX-501; CX-519).

I. Effects Of The Acquisition 1. Market Share and Concentration 378. The market share and concentration figures in the relevant geographic market--the continental United States--in terms of throughput capacity, number of presses, and sales were, for 1990: R.R. DONNELLEY & SONS CO., ET AL. 109 36 Initial Decision TABLE 1 CONTINENTAL UNITED STATES GRAVURE CAPACITY 1990 Company Throughput Percent HHI Capacity Share Contribution Donnelley 2,753 33.1 1,093 Meredith/Burda 1,296 15.6 242 Combined 4,049 48.7 2,368 { ] [ ] Pre-Acquisition HHI 2,041 Post-Acquisition HHI 3,070 Increase in HHI 1,029 (CX-501-B).

TABLE 2 CONTINENTAL UNITED STATES GRAVURE CAPACITY 1990 Number of Percent HHI Company Presses Share Contribution Donnelley 56 35.9 1,289 Meredith/Burda 20 12.8 164 Combined 76 48.7 2,373 [ ] [ ] Pre-Acquisition HHI 2,172 Post Acquisition HHI 3,093 Increase in HHI 920 (CX-501-A) Initial Decision 120 F.T.C.

TABLE 3 CONTINENTAL UNITED STATES GRAVURE SALES 1989 (for prepress and presswork in millions of dollars) Percent HHI Company Sales Share Contribution Donnelley 202.7 27.4 749 Meredith/Burda 117.9 15.5 241 Combined 325.6 42.9 1,840 [ ] [ ] Pre-Acquisition HHI 1,868 Post-Acquisition HHI 2,719 Increase in HHI 850 (CX-501-E).

379. Table 4 shows market shares for certain gravure printing work of more than five million copies produced in 1990: TABLE 4 1990 GRAVURE OUTPUT SHARES.

(PERCENT) (A) (B) 5.0+ 10.0+ million Million copies copies DONNELLEY 36.4 37.4 MEREDITH/BURDA 18.3 16.0 TOTAL 54.7 53.4 R.R. DONNELLEY & SONS CO., ET AL. 111 36 Initial Decision 2. Cancelled or Delayed Expansion Plans 380. Dr. Hilke compiled CX-502-A, a chart which depicts the cancellation or deferment of various Donnelley and Meredith/Burda expansion plans which he believed had a "quite high" probability of going forward absent the Meredith/Burda acquisition (Tr. 3349-50). [ ] 381. In Dr. Hilke's opinion, the cancellation or deferral of these expansion plans had a substantial adverse competitive effect because had they gone forward, significant gravure capacity would have been added to the market at or about the time of the acquisition and this would have resulted in an increase in gravure supply and a reduction in prices (Tr. 3347-54, 3359, 6133-34).

382. Dr. Hilke concluded that, in actuality, Donnelley's acquisition of capacity rather than expansion "represents an alternative which involves higher prices and less competition. . . ." (Tr. 3359).

383. Several documents prepared by or for Donnelley tend to support Dr. Hilke's observations.

384. Donnelley's strategic consultant, BCG, pointed out that if Donnelley continued to bring on new capacity, price erosion would result (CX-701-D-E).

385. Robert A. Revak, of Donnelley's catalog group, stated in a draft of its 1989 strategic plan:

The only way that I can see Donnelley changing the trend of continuing price erosion in the marketplace is to remove our competition through acquisition. (CX-157-B).

386. Donnelley's October 1990 Rating Agency Presentation stated that, "[w]ith the addition of Meredith/Burda's modern, well equipped plants, Donnelley obtains needed additional capacity to better serve and expand share in these markets without adding additional capacity to the industry" (CX-1156-J; see also CX-35-K; CX-40-D; CX-41-V; CX-1061-B).

387. A February 12, 1990 memorandum sent to senior Donnelley management by John S. Oberhill, then president of the magazine group (CX-140-M-N), recognized the price effect of gravure acquisitions:

Initial Decision 120 F.T.C.

Selective acquisitions to reduce supply is the preferred method to stabilize pricing levels and obtain growth in both sales and earnings. Large capacity additions, which significantly exceed market growth rates, must be brought forward cautiously unless there is enabling contract volume.

(CX-154-A).

388. An internal memorandum prepared by Shearson Lehman Hutton, the firm handling Donnelley's debt offering for the acquisition (CX-149-H-I), observed:

Meredith/Burda was one of Donnelley's larger competitors, particularly in the very long-run, very capital intensive gravure segment of U.S. printing. Thus, Donnelley effectively acquires additional, as well as some excess, capacity at less than "new" construction cost without adding to industry capacity, and eliminates some competition in the process. Correspondingly, some $175 million [of] Donnelley's future capital needs were offset through this acquisition. (CX-757-A).

3. Constraints on Meredith/Burda's Independence 389. The initial Donnelley/Meredith Burda acquisition agreement required the latter to obtain Donnelley's approval before any printing contracts exceeding [ ] could be signed by Meredith/Burda, and Meredith/Burda did seek Donnelley's approval for contracts with its customers including Target, a high volume gravure customer, from December 1989 to September 1990 (CX-2-Z-23-24, Z-32-34; CX- 496-D; CX-1052; CX-1062). © 390. The initial agreement required [| ] (CX-496-D). Capital expenditures exceeding $100,000 required Donnelley's approval (CX-2-Z-33). Meredith/Burda sought Donnelley's approval for its contracts with suppliers during the period December 1989 to September 1990 (CX-1106-A) and Donnelley imposed changes on proposed contracts (CX-1106-H-K).

391. Other restrictions on Meredith/Burda's business decisions were also imposed in the initial acquisition agreement of 1989 (CX- 2-Z-32-34; CX-496-E-G).

392. Dr. Hilke acknowledged that agreements curtailing a seller's actions for a short period after an initial sale agreement and before the closing are common and avoid the costs of redetermining and renegotiating the price at the time of the final closing; however, he R.R. DONNELLEY & SONS CO., ET AL. 113 36 Initial Decision believes that extension of such agreements over an extended period of time curtails competition by severely limiting the independence of the acquisition target while it is supposed to remain independent (Tr. 3370-71).

4, Unilateral Market Power 393. The size of Donnelley's post-acquisition gravure market share in the United States (42.9 to 48.7%) suggests that it can unilaterally raise prices to some high volume publication gravure print customers, restrict output or engage in other anticompetitive conduct. 1992 Merger Guidelines of the Department of Justice and the Federal Trade Commission ("1992 Guidelines") Section 2.2. 394. Donnelley officials recognized that the acquisition would strengthen its position in the market. In his handwritten notes to Carl Doty recommending the acquisition of Meredith/Burda, catalog group president Schroder wrote "market control and stabilize" as the first on a list of "strategic issues" associated with the acquisition (CX-40-C; CX-41-I); and BCG's Michael Silverstein noted that the acquisition accorded with his observations that "[p]rice stability is facilitated by very large share of leader" (CX-701-F). 395. Ronald L. Nicol, another BCG consultant, informed Donnelley that the acquisition "creates market power for Donnelley" and “limits customer options" (CX-703-P).

396. Before the acquisition, Donnelley had a leading share in gravure catalogs and inserts (CX-158-Z-13), and some of its customers view it as the major supplier of high volume publication gravure printing (CX-622-M; CX-421-C; CX-632-C; CX-785-B; Tr. 674-75).

397. Dr. Hilke testified that Donnelley could exercise market power with respect to those high volume publication customers who would not switch to gravure in the face of a 5% increase in gravure prices by targeting them for non-cost based price increases, while keeping prices at competitive levels for lower volume gravure printing customers whose demand is more elastic (Tr. 3012-14, 3071, 6163-64).

398. Dr. Hausman agreed that some high volume gravure customers prefer that process over offset (Tr. 5224), but denied that Donnelley could successfully practice price discrimination: Initial Decision 120 F.T.C.

in certain types of situations of price discrimination, you are able to tell whether a customer will switch or not. But in this type of situation . . . you will not be perfectly [able to] identify those customers that you could price discriminate against (Tr. 5196-97), because where fixed costs are high and excess capacity exists, the failure to accurately predict which customers would accept a non-cost based price increase might result in a loss of revenue and, perhaps, require lower prices to obtain replacement business (Tr. 5198-99, 6330-32, 6340-41, 5921).

399. Donnelley has, over the years, asked its salesmen and operations officers to assess customer needs, to report the prices charged them, and to assess the profitability of work done for different accounts. For example, a Donnelley document entitled "Pricing Strategy, Plan, and Policy" stated: At the meeting I outlined the concept we discussed which related toward niche marketing strategies.

Within this plan, I would also suggest identification of targeted accounts and plans for them, as well as a listing of accounts, levels, and categories of accounts. The plans, dates, needs, etc. would be part of this. What do we need to do for these particular salespeople regarding sensitivity, goals, training, special information, etc.? Probably most important to include is the actual strategy and concept of how it would work.

As a side benefit of this, I believe comparative price level information on levels, customers (Graded A, B, C, D, etc.), selected price sensitivity test customers, etc. comparisons should be made.

(CX-560-B).

400. The record contains many other examples of customer analysis (CPF 747-845) which, according to Donnelley, have no sinister implications but is rather what every successful business must do if it is to satisfy customer needs.

401. Dr. Hilke agreed that the "process of targeting" is: not in itself offensive, it just is the underlying set of conditions and practices that could lend themselves to a targeting of customers after an acquisition of anticompetitive concern I'm talking about. (Tr. 3024).

402. Although he denied that Donnelley could successfully engage in an extensive program of selective price increases to high volume publication gravure customers (F 398), Dr. Hausman conceded that some customers might be targeted for such increases: R.R. DONNELLEY & SONS CO., ET AL. 115 36 Initial Decision Judge Parker: One second. Professor, there are still some customers that testified, I believe, the high volume customers, that they simply would never consider switching to offset.

The Witness: Yes. Penney's did, for instance. Judge Parker: What about that? The Witness: Well, I think for those customers -- Judge Parker: Can they be price discriminated against from now on by Donnelley or others? The Witness: If you knew who they were. Theoretically, if you knew who they were and if that same person stays there, because it's usually a personal preference.

(Tr. 5224).

403. Dr. Hausman denies that Donnelley knows who these customers are but, in light of Donnelley's extensive, continuing analysis of customer needs, I agree with Dr. Hilke that targeting might well be successful.

5. Customer Complaints 404. As soon as the Meredith/Burda acquisition was announced, customers expressed concern to the FTC and the parties about the decrease in competition that might result (CX-8-Z-9-25; CX-119-Z- 90; CX-121-Z-69; CX-171; CX-174-A; CX-174-B; CX-176; CX- 177; CX-178-A; CX-179-B; CX-186-B; CX 332-C; CX-467-A; CX- 620-E; CX-624; CX-1006-N; RX-79-T; Tr. 671-72, 705-06, 743, 823, 961-62, 964-65, 1033-34, 1439-40, 1934-36, 2010, 4092-94). 405. Donnelley claims that some of those customers who complained about the acquisition did not express that much concern during their testimony (RPF 256), but it is evident that even those customers who have as yet experienced no adverse effects from the acquisition may still be concerned about its long-term impact. For example, Mr. Sackett of Penney testified:

A. I should answer that our relationship with Donnelley is, as I pointed out earlier, excellent. And there has been no specific adverse impact on our business relationship with them as a result of that purchase. Nevertheless we have a $4 billion dollar [sic] business which is entirely dependent upon the supply of rotogravure capacity and we cannot reasonably applaud an action which results in having one source of supply available to us. (Tr. 617-18).

Initial Decision 120 F.T.C.

6. Coordinated Interaction 406. A concern of Dr. Hilke is that Donnelley's acquisition of one of its primary competitors increases concentration in an already concentrated market with a small number of firms and that it may therefore lead to coordinated interaction, or collusion, among the remaining firms (Tr. 3000, 3391-92, 6155-57, 6190). 407. Where there are, as here, few suppliers, and there is a dominant firm such as Donnelley, coordination of prices is more likely because cheating is easier to detect and punishment is severe (Tr. 3030, 3392, 6149-53, 6157).

408. Coordination of gravure prices is possible because information about competitive activity of industry members is readily available from press manufacturers (Tr. 331-34, 2547), from movement of employees from firm to firm (Tr. 3397; CX-141-Z-95) and from industry meetings (see CX-379-A; CX-391-V; CX-453; CX-454; CX-462; CX-634-Z-3; CX-644; CX-765-B; CX-891-A; CX-892-A; CX-943; CX-944; RX-152-C; RX-153-A). 409. The nature of gravure printing may also facilitate coordination: there are only two major manufacturers of gravure presses (CX-102-Z-48-49; Tr. 6901); all gravure printers use the same process to produce the finished product, and much of the printers’ business is obtained through bidding, which requires an intimate knowledge of industry cost structure and other competitive variables.

410. The probability of Donnelley being able to successfully impose unilateral price increases on its high volume publication gravure customers or of colluding with its competitors with respect to price is limited somewhat by the size of its customers and their ability to switch suppliers. There have been several post-acquisition instances where print buyers have qualified additional gravure printers besides Donnelley or have switched substantial quantities of their printing to other gravure printers (RPF 268, Table F). 411. Nevertheless, many print buyers believe that the Donnelley- Meredith/Burda combination produces higher quality work than other gravure printers; and, there are substantial impediments to switching gravure suppliers with ease (Tr. 3075-77, 5894-95, 5973-75). 412.[ ] (Tr. 2012-13, 2015-26, 2033-34, 5968-74). 413. Mr. Angstrom of St. Ives testified:

R.R. DONNELLEY & SONS CO., ET AL. 117 36 Initial Decision a very large customer who commands -- significant levels of capacity . . . is going to find great difficulty finding another home quickly. (Tr. 2669).

414. Donnelley's David Moeller testified that, when large customers have been with a printer over a long period of time, they are likely to continue the relationship even if they have to pay a higher price to do so (Tr. 4064-65), and the supervisor of the purchasing department at Current, Inc., stated[ ] (Tr. 1943). 415. Indeed, the concerns expressed by their larger customers over the Donnelley-Meredith/Burda acquisition reflects their belief that switching to alternative sources of high volume gravure publication printing would, in some cases, be difficult and time consuming.

7. Efficiencies 416. Dr. Hilke testified that with respect to the existence of merger specific efficiencies or synergies: to the extent I've been able to identify such efficiencies, they don't seem to be ones which would be peculiar to this particular acquisition. (Tr. 3412).

8. Conclusion 417. For the reasons given above, complaint counsel's concern that Donnelley's acquisition of Meredith/Burda may substantially lessen competition in high volume publication gravure printing in the United States is justified; and, since new entry or expansion into this market would require at least two years or more lead time, it would not mitigate the probable anticompetitive effects of the acquisition. Initial Decision 120 F.T.C.

III. CONCLUSIONS OF LAW A. The Relevant Product Market 1. Introduction The purpose of market definition in antitrust cases is to "identify those sections of the economy which may be exposed by the [challenged] transaction to anticompetitive price increases" Owens- Illinois, Inc., D. 9212, at 4 (Feb. 26, 1992). The 1992 Guidelines, Section 1.0, defines a market by application of the so-called "five percent test”:

A market is defined as a product or group of products and a geographic area in which it is produced or sold such that a hypothetical profit-maximizing firm, not subject to price regulation, that was the only present and future producer or seller of those products in that area likely would impose at least a "small but significant and nontransitory" increase in price, assuming the terms of sale of all other products are held constant. A relevant market is a group of products and a geographic area that is no bigger than necessary to satisfy this test. See also Owens-Illinois, Inc., pp. 4-5.

Complaint counsel and Donnelley do not disagree that many print buyers can and do use either offset or gravure printing services for some jobs. Donnelley goes even further: it asserts that practically all print jobs, regardless of run length, page count or number of versions can be done by gravure or offset printers, and are acceptable to their customers. Complaint counsel deny this and claim that there is a product over which Donnelley has market power (i.e., "the ability profitably to maintain prices above competitive levels for a significant period of time” 1992 Guidelines, Section 0.1): the supply of high volume publication gravure printing services, a product which is sufficiently distinct from offset that buyers could not defeat an increase in its price by shifting their purchases to offset. See Hospital Corporation of America, 106 FTC 361, 464 (1985), aff'd. 807 F.2d 1381 (7th Cir. 1986), cert. denied, 481 U.S. 1038 (1987). Whether a product is "sufficiently distinct" so that switching would not occur depends on the "reasonable interchangeability" between the products, which is determined by: R.R. DONNELLEY & SONS CO., ET AL. 119 36 Initial Decision examining such practical indicia as industry or public recognition of the submarket as a separate economic entity, the product's peculiar characteristics and uses, unique production facilities, distinct customers, distinct prices, sensitivity to price changes, and specialized vendors.

Brown Shoe Co. v. United States, 370 U.S. 294, 325 (1962). See also United States v. E.I. dupont de Nemours & Co., 351 U.S. 377, 395 (1956).

2. The Peculiar Characteristics of the Gravure and Offset Processes In addition to the physical differences of gravure and offset presses (F 14-28), there are differences between the two processes which affect customer choice: gravure's greater throughput which makes it more cost-effective than offset for jobs with large numbers of copies and many pages per copy (F 162); the greater durability of gravure cylinders (F 165-169); gravure paper savings resulting from variable cut-off capability (F 172); the ability of gravure presses to stitch and trim on line (F 173-74); less paper waste in the gravure process (F 175); and, gravure's ability to produce better results than offset on cheaper, lighter weight, uncoated paper (F 180). 3. Gravure and Offset Quality Some industry members believe that, for their purposes, gravure and offset offer comparable quality (F 186); some high quality high volume magazines use offset and gravure for different parts of their publication (F 185); and, industry members testifying in this proceeding could not distinguish between the two processes with the naked eye (F 187).

Nevertheless, the firm belief of many print buyers that gravure offers higher quality than offset is a real constraint on their choice of printing processes.

4. The Economics of Gravure and Offset The record supports complaint counsel's claim that, for low version, long run, high page count publications, gravure is less costly than offset. The breakeven point (F 212) at which this usually occurs Initial Decision 120 F.T.C.

is in publications with less than 4 four-color versions, more than 32 pages (F 222) and a run length in excess of 5 million copies (F 216). The gravure advantage for these publications is confirmed by evidence that print buyers have switched to the gravure process as the run length of their publications increased (F 217-19) and in documents written by Donnelley employees (F 227-232). Additional confirmation was supplied by the testimony of buyers who described their print programs and their choice of gravure or offset to meet their demands (F 59-132). Of particular interest is the decision of some buyers, such as National Geographic, to use both processes for their publications because of their unique contribution to specific needs, such as, in the case of offset, its lower cost when a portion of the publication requires a high number of versions (F 121). Donnelley criticizes Dr. Hilke's product market characteristics as vague and contradictory. Some uncertainty is part and parcel of any attempt to define the boundaries of a product market but it is not a fatal flaw if it is, on balance, supported by the record. See United States v. Pabst Brewing Co., 384 U.S. 546, 549 (1966). In this case, there was explicit testimony that for run lengths in excess of 5 or 10 million copies, gravure is more economical than offset (see, e.g., F 234):

Well, 10 million and page counts of 32 or more, typically, in my experience, today, yesterday, tomorrow is going to be done gravure. 5. Gravure and Offset Prices The independence of gravure and offset prices indicates that the cross-elasticity of demand between the two processes is relatively low, and that at this level the processes occupy separate competitive niches:

The outer boundaries of a product market are determined by the reasonable interchangeability of use or the cross-elasticity of demand between the product itself and substitutes for it.

Brown Shoe, 370 U.S. at 325.

Direct evidence of cross-elasticity of demand is often unavailable: Hence, we may apply reasoned judgment in estimating or inferring the relative magnitude of the elasticities in order to assess the degree of market power... . R.R. DONNELLEY & SONS CO., ET AL. 121 36 Initial Decision Grand Union, 102 FTC 812, 1041 (1983). See also B.F. Goodrich Co., 110 FTC 207, 290 (1988) ("persistent price differences" a "surrogate" for direct evidence of cross-elasticity). Evidence of independent gravure and offset prices include Donnelley's price tracking reports (F 205-09) and the testimony of industry members (F 210) confirming complaint counsel's claim that the demand for high volume publication gravure printing is relatively inelastic, i.e., that demand for it does not fall significantly if its price increases by a small but significant and nontransitory amount. F TC v. Bass Brothers Enterprises, 1984-1 CCH Trade Cas. {{ 66,041 at 68,613.

6. Statistical Analyses of Gravure's Advantage Over Offset. Statistical confirmation of gravure's dominance over offset for high volume publication gravure printing is provided by Dr. Hilke's analysis of such publications in CX-1167 (F 236-42) and his analysis of Donnelley's studies of gravure economics (F 244-48). CX-1167 demonstrates that high volume publication gravure printing in 1990 accounted for a substantial amount of commerce and that such printing is predominantly the domain of gravure (F 239, 240).

High volume publication printing is not done exclusively by gravure: in 1990, offset accounted for 11.4% of print runs which exceeded 5 million copies (F 240); however, existence of some competition between the two processes does not negate the conclusion that they occupy separate markets for those customers whose demand for gravure is inelastic. See Columbia Metal Culvert Co. v. Kaiser Aluminum & Chemical Corp., 579 F.2d 20, 30 (3d Cir.), cert. denied, 439 U.S. 876 (1978):

The existence of competition between these product lines does not alone preclude market power within each line, if each product has a cadre of customers in which it enjoys a decisive advantage.

Owens-Illinois, which Donnelley says is controlling in this case, is not inconsistent with Columbia Metal. In Owens, the Commission found that a significant competitor had entered the market with spaghetti sauce packed in metal cans and that its capture of 5% of the market in less than two years showed that metal cans compete with Initial Decision 120 F.T.C.

glass for the packaging of spaghetti sauce. The Commission also predicted "further transfer of market share to cans" if glass container prices rise. Owens-Illinois, at 15-16.

There is no evidence in this case of any recent, dramatic inroads by offset into high volume publication printing and there is no reason to believe that there will be any technical changes in offset which will allow substantial penetration into gravure's domain, even if the M- 3000 press should be successful (F 198-200). 7. Industry Recognition The print industry, both buyers and sellers, recognize the existence of a high volume publication gravure market (F 274-310). See B.F. Goodrich, 110 FTC at 290 ("industry firm perceptions are "surrogates" for direct evidence of elasticity"); Grand Union Co., 102 FTC at 1041. This includes Donnelley, whose employees, in many documents, explicitly or implicitly recognized a gravure market (F 274-79), and who, along with other industry members, made major investments in gravure capacity in the face of excess offset capacity (F 285). This phenomenon, together with Donnelley's purchase of Meredith/Burda's gravure capacity (F 7-8) when it could have bought many more less costly offset presses, is inconsistent with the claim that there is no significant difference between the two processes. 8. Conclusion Donnelley relies too heavily on its analysis of gravure print buyers who have switched to offset 122-32), for it ignores buyers who, like [ ], may switch from offset to gravure after reconsidering its cost (F 85) or who, like Montgomery Ward, may produce new gravure catalogs (F 126) (see also F 315-16). This analysis also ignores the reason for some of the switches-increased versioning--which simply reinforces complaint counsel's claim that the two processes offer unique features (F 106, 108, 123, 125). In addition, versioning information for some buyers is unavailable (F 128).

Thus, although some high volume publication gravure buyers have switched to offset, the totality of the evidence convincingly demonstrates that the demand for high volume publication gravure printing is, for some customers, (i.e., those who would not switch to R.R. DONNELLEY & SONS CO., ET AL. 123 36 Initial Decision offset even if gravure prices were raised 5% (F 273)), inelastic and that these customers and their suppliers operate in the relevant product market alleged in the complaint.

B. The Relevant Geographic Market The geographic area or areas within which the probable effects of this acquisition should be measured is where the "seller operates and to which buyers can practicably turn for supplies." Tampa Electric Co. v. Nashville Coal Co., 365 U.S. 320, 327 (1961); FTC v. Foodtown Stores, 539 F.2d 1339, 1344 (4th Cir. 1976); Midcon Corp., 112 FTC 93, 162 (1989).

The parties agree that the United States is one geographic market, but disagree as to complaint counsel's claim that a significant number of West Coast high volume gravure publication print customers can practicably turn only to West Coast gravure printers for their needs. Since no Elzinga-Hogarty analysis of shipping patterns is possible (F 329), complaint counsel rely on industry perception and industry actions to establish their claims that there is a distinct West Coast market (F 328, 330).

Some West Coast customers, such as Target Stores, which must have timely printing of its inserts (F 331), are cited as examples of print customers who can turn only to West Coast gravure printers for their needs, yet Target's print buyer testified that he might consider switching from West Coast gravure suppliers if they raised their prices by 5% (F 332).

Since a substantial amount of gravure printing is done outside of, and shipped into, the Western United States (F 333) I reject complaint counsel's claim that West Coast print buyers can, in most cases, turn only to West Coast gravure printers to satisfy their needs, and I therefore reject their proposed West Coast geographic market for high volume publication gravure printing. C. The Effects Of The Acquisition 1. Market Share and Concentration The 1990 United States market share/concentration figures for gravure printing were:

Initial Decision UNITED STATES GRAVURE SHARE MEASURES Capacity Capacity (Numbers of Presses) (Press Throughput) Percentage Shares Donnelley 35.9% 33.1% Meredith Burda 12.8% 15.6% Combined Shares 48.7% 48.7% HHI Contributions Donnelley 1,289 1,093 Meredith Burda 164 242 Combined HHI 2,373 2,368 United States HHI Pre-Acquisition HHI 2,172 2,041 Post-Acquisition HHI 3,093 3,070 Increase in HHI 920 1,029 (F 378).

Gravure Sales 27.4% 15.5% 42.9% 7A9 1,840 1,868 2,719 Since these figures encompass all gravure printing, market share data for gravure output in the relevant product market is a more accurate indication of the impact of the acquisition: TABLE 4 1990 GRAVURE OUTPUT SHARES (PERCENT) (A) 5.0+ million copies (B) 10.0+ million copies DONNELLEY 36.4 MEREDITH/BURDA 18.3 TOTAL 54.7 [ ] (F 379).

37.4 16.0 53.4 By any measure--total gravure printing or high volume publication printing--the concentration in the markets exceed the level at which illegality can be inferred. See 1992 Guidelines, Section 1.51(C):

R.R. DONNELLEY & SONS CO., ET AL. 125 36 Initial Decision Post-Merger HHI Above 1800. The Agency regards markets in this region to be highly concentrated. Mergers producing an increase in the HHI of less than 50 points, even in highly concentrated markets post-merger, are unlikely to have adverse competitive consequences and ordinarily require no further analysis. Mergers producing an increase in the HHI of more than 50 points in highly concentrated markets post-merger potentially raise significant competitive concerns, depending on the factors set forth in Sections 2-5 of the Guidelines. Where the post-merger HHI exceeds 1800, it will be presumed that mergers producing an increase in the HHI of more than 100 points are likely to create or enhance market power or facilitate its exercise. The presumption may be overcome by a showing that factors set forth in Sections 2-5 of the Guidelines make it unlikely that the merger will create or enhance market power or facilitate its exercise, in light of market concentration and market shares. See also Hospital Corp. of America v. FTC, 807 F.2d 1381, 1384 (7th Cir. 1986), cert. denied, 481 U.S. 1038 (1987) (combined share of 26 percent, four-firm concentration of 91 percent; held unlawful); FTC v. Warner Communications Inc., 742 F.2d 1156, 1163 (9th Cir. 1984) (combined share of 26 percent; four-firm concentration of 75 percent; presumptively unlawful); RSR Corp. v. FTC, 602 F.2d 1317, 1324 (9th Cir. 1979) (combined market share of 19.2 percent, acquisition held unlawful).

Additional analysis beyond market share statistics demonstrates that the challenged acquisition may pose a "significant threat to competition" United States v. General Dynamics Corp., 415 U.S. 486, 496-99 (1974).

The particular threat to competition which the acquisition may create is its potential for the exercise of market power over high volume publication gravure printing either by one firm (unilateral market power)’ or a group of firms (coordinated interaction). 2 Lessening of Competition Through Unilateral Effects A merger may diminish competition even if it does not lead to increased likelihood of successful coordinated interaction, because merging firms may find it profitable to alter their behavior unilaterally following the acquisition by elevating price and suppressing output. Unilateral competitive effects can arise in a variety of different settings. In each setting, particular other factors describing the relevant market affect the likelihood of unilateral competitive effects. The settings differ by the primary characteristics that distinguish firms and shape the nature of their _ competition. 1992 Guidelines, Section 2.2. 3 Lessening of Competition Through Coordinated Interaction A merger may diminish competition by enabling the firms selling in the relevant market more likely, more successfully, or more completely to engage in coordinated interaction that harms consumers. Coordinated interaction is comprised of actions by a group of firms that are profitable for each of them only as a result of the accommodating reactions of the others. This behavior includes tacit or express collusion, and may or may not be lawful in and of itself. 1992 Guidelines, Section 2.1.

Initial Decision 120 F.T.C.

2. Cancellation of Expansion Plans Donnelley's acquisition of Meredith/Burda resulted in cancellation of its expansion plans, a cancellation which would have added capacity and which might have reduced prices in the relevant product market (F 380-88). It also removed from the competitive arena its major competitor and, to some industry observers, the only other high quality gravure printer (F 338-57). 3. The Unilateral Exercise of Market Power In addition to allowing the reduction of output, the acquisition increased the likelihood that the combined firm, either on its own, or in combination with other gravure firms, would exercise market power. See 1992 Guidelines, Section 2.0.

Donnelley's post-acquisition market share suggests that it can unilaterally raise prices to some high volume publication gravure print customers, restrict output or engage in other anticompetitive conduct.

The probability of this occurrence was evident to Donnelley officials at the time of the acquisition (F 393-94), was of great concern to its print customers (F 404-05), and was conceded, at least as to customers such as Penney, by Dr. Hausman, who agreed that it could be targeted for price increases (F 402). Donnelley disagrees that high volume publication gravure printing customers can be targeted and claims that even if the possibility exists, the number of "core" customers has diminished dramatically since the acquisition (F 364). Donnelley's argument ignores recent test entry into the core market by [ ] (F 262), the probability that other high volume offset customers will do the same, and the possible effect of the recent recession on high volume publications. Reliance on post-acquisition effects which may be the result of an economic downturn ignores the possibility that as the economy improves the "trends" which Donnelley observes will vanish (F 358-62) and skepticism about their long-term effect is warranted, especially those over which Donnelley may have some influence. See Hospital Corp. of America, 807 F.2d 1381, 1384 (7th Cir. 1986), cert. denied, 481 U.S. 1038 (1987). R.R. DONNELLEY & SONS CO., ET AL. 127 36 Initial Decision 4. Coordinated Interaction Proof of actual collusion arising from an acquisition is "not required to show a violation of [Section 7 of the Clayton] Act. Instead [a] predictive judgment . . . is called for." Owens-Illinois (Azcuenaga, concurring). See also B.F. Goodrich, 110 FTC at 303: "As the number of firms in an industry declines, and industry concentration increases . . . it becomes easier for those firms to coordinate their pricing and the likelihood of anticompetitive effects ... increases."

A predictive judgment that Donnelley's acquisition of its major competitor increases the likelihood of collusion can be made with confidence because of the exit of several market participants over the past few years (F 376-77), the small number of remaining firms (F 407), and the ready availability of industry information which can facilitate collusion (F 408). See United States v. Aluminum Company of America, 377 U.S. 271, 280-81 (1964); 1992 Guidelines, Sections 2.0, 2.11, 2.12.

5. Conclusion The acquisition of Meredith/Burda by Donnelley creates an “appreciable danger" of future anticompetitive effects. Owens- Illinois at 29.

That concern could be ignored if there were no barriers to entry into gravure printing, for, in that case, "it is unlikely that market power, whether individually or collectively exercised, will persist for long," B.F. Goodrich, 110 FTC at 207, 296 n.63, but there are substantial barriers to rapid and effective entry into high volume publication gravure printing. Barriers to entry into gravure printing (F 375) would create a more than two year delay between the time entry or expansion is contemplated and ultimately achieved (F 365- 74). Thus, entry into high volume publication gravure printing would not occur swiftly enough to counter the probable anticompetitive consequences of the Meredith/Burda acquisition: In order to deter or counteract the competitive effects of concern, entrants quickly must achieve a significant impact on price in the relevant market. The Agency generally will consider timely only those committed entry alternatives that can be achieved within two years from initial planning to significant market impact. 1992 Guidelines, Section 3.2.

Initial Decision 120 F.T.C.

IV. SUMMARY 1. The Commission has jurisdiction over the subject matter of this proceeding and over defendants R.R. Donnelley & Sons Co. ("Donnelley") and Pan Associates, a limited partnership. 2. Donnelley and Pan Associates were, at all times relevant herein, engaged in commerce, as "commerce" is defined in Section 1 of the Clayton Act, as amended, 15 U.S.C. 12, and their business is in, or affects, commerce as "commerce" is defined in Section 4 of the Federal Trade Commission Act, as amended, 15 U.S.C. 44. 3. The appropriate line of commerce within which to evaluate the competitive effects of the acquisition of Meredith/Burda by Donnelley is the supply of high volume publication gravure printing. 4. The appropriate geographic market within which to evaluate the competitive effects of the acquisition of Meredith/Burda's high volume publication gravure printing business is the continental United States.

5. Barriers to entry into the relevant market are substantial, and substantial harm to competition would occur until new entry could be accomplished.

6. Prior to and at the time of the acquisition, Donnelley and Meredith/Burda were actual, direct and substantial competitors in the supply of high volume publication gravure printing. 7. The effect of this acquisition has been or may be substantially to lessen competition or to tend to create a monopoly in the aforesaid product and geographic market in violation of Section 7 of the Clayton Act, as amended, 15 U.S.C. 18, and Section 5 of the Federal Trade Commission Act, as amended, 15 U.S.C. 45, in the following ways:

(a) It eliminated actual competition between Donnelley and Meredith/Burda and between Meredith/Burda and others in the relevant market;

(b) It significantly increased the already high levels of concentration in the relevant market;

(c) It created a firm whose share of the relevant market is so high that it has achieved the position and market power of a dominant firm;

R.R. DONNELLEY & SONS CO., ET AL. 129 36 Initial Decision (d) It eliminated Meredith/Burda as a substantial independent competitive force in the relevant market; and (e) It increased the likelihood of successful anticompetitive conduct, non-rivalrous behavior and actual or tacit collusion among the firms in the relevant market.

8. All of the above increase the likelihood that firms will increase prices and restrict the output of high volume publication gravure printing.

9. The order entered hereinafter is appropriate to remedy the violation of law found to exist.

V. ORDER I.

It is ordered, That for purposes of this order the following definitions shail apply:

A. "Donnelley" means respondent R.R. Donnelley & Sons Co., its directors, officers, agents, representatives and employees, and its parents, predecessors, subsidiaries, divisions, groups, affiliates, successors and assigns, and their respective directors, officers, agents, representatives, employees, successors and assigns. B. "Meredith/Burda" means the former business entity jointly owned by Meredith Corporation and Pan Associates, L.P. C. "Pan Associates, L.P." means a limited partnership and a holding company for the Burda family with its principal place of business in New York, New York.

D. "Meredith/Burda's Printing Business" means the business of commercial printing acquired by Donnelley from Meredith/Burda, including all of the assets, titles and properties, tangible and intangible, of said business, and its associated interests, rights and privileges, including without limitation all buildings, leaseholds, machinery, equipment, inventory, supply arrangements, funded employee benefit plans, customer lists, copyrights, trade names, trademarks, trade secrets, patents and other property of whatever description, together with all additions and improvements thereto made subsequent to the Acquisition by Donnelley and all other facilities, assets, titles, properties, interests and rights and privileges, Initial Decision 120 F.T.C.

including any business interest in Siegwerk Inc., as may be necessary to reconstitute Meredith/Burda's Printing Business as a viable competitor to the same extent as existed prior to the Acquisition. Meredith/Burda's_ Printing Business shall include: all of Meredith/Burda's printing plants located at Casa Grande, Arizona; Des Moines, Iowa; Newton, North Carolina; and Lynchburg, Virginia.

E. "Acquisition" means the acquisition of Meredith/Burda by Donnelley, pursuant to a Purchase and Sale Agreement entered into on December 21, 1989, and more fully described in J 12 of the Commission's complaint issued in this matter. F. The "Commission" means the Federal Trade Commission. II.

It is further ordered, That Donnelley shall divest, absolutely and in good faith, Meredith/Burda's Printing Business within twelve (12) months from the date this order becomes final. The divestiture shall be only to an acquirer that receives the prior approval of the Commission, and only in a manner that receives the prior approval of the Commission and, if the divestiture of Meredith/Burda's Printing Business is to be accomplished by a public offering of all stock and other share capital of a corporation containing Meredith/Burda's printing business, such public offering shall also only be in a manner that receives the prior approval of the Commission. Provided, however, that for a period of ten (10) years after the date of the public offering, no person who is an officer, director or executive employee of Donnelley or who owns more than one (1) percent of the stock of Donnelley shall be an officer, director or executive employee of the corporation or shall own or control directly or indirectly more than one (1) percent of the stock of the corporation. The purpose of the divestiture is to maintain Meredith/Burda's Printing Business as a viable competitive concern engaged in commercial printing and to remedy the lessening of competition, resulting from the Acquisition, as alleged in the Commission's complaint. In connection with the divestiture required by this paragraph:

A. If any printing plant associated with Meredith/Burda's Printing Business, or any other facilities, assets, titles, properties, interests, rights and privileges associated with such Printing Business, have R.R. DONNELLEY & SONS CO., ET AL. 131 36 Initial Decision been sold, closed shut down, disposed of, or are no longer operational, Meredith/Burda's Printing Business shall include Donnelley's most comparable, as determined by the Commission, printing plant or facilities, assets, titles, properties, interests rights and privileges associated with such printing business that are in the same or better condition than those that were acquired. B. Donnelley shall provide to the acquirer of Meredith/Burda's Printing Business, or to the corporation in the event of a public offering, on a nonexclusive basis, all technology (including patents, licenses and know-how) that was not obtained by Donnelley as part of the Acquisition and is used by Donnelley, or developed by Donnelley for use, in connection with Meredith/Burda's Printing Business; Donnelley shall not interfere with any attempt by such acquirer of Meredith/Burda's Printing Business, or the corporation in the event of a public offering, to employ any personnel previously or presently employed by Meredith/Burda, or previously or presently employed by Donnelley, in connection with the operation of Meredith/Burda's Printing Business nor seek to enforce any employment contract against such personnel. C. Donnelley shall assign to the acquirer of Meredith/Burda's Printing Business, or to the corporation in the event of a public offering, all customer agreements or understandings, formal or informal, and all customer records and files relating to commercial printing supplied by Meredith/Burda's Printing Business. Ill.

It is further ordered, That:

A. If Donnelley has not fully complied, absolutely and in good faith with paragraph II of this order within the times provided in such paragraph, Donnelley shall consent to the appointment of a trustee to divest the assets pursuant to paragraph II of this order. In the event the Commission or the Attorney General brings an action pursuant to Section 5(1) of the Federal Trade Commission Act, 15 U.S.C. 45()), or any other statute enforced by the Commission, for any failure by Donnelley to comply with this order, Donnelley shall consent to the appointment of a trustee in such action. Neither the appointment of a trustee nor a decision not to appoint a trustee under this paragraph shall preclude the Commission or the Attorney General from seeking Initial Decision 120 F.T.C.

civil penalties and other available relief, including a court-appointed trustee, pursuant to Section 5(1), or any other statute enforced by the Commission, for any failure by Donnelley to comply with this order. B. Ifa trustee is appointed by the Commission or a court pursuant to paragraph III. A of this order, Donnelley shall consent to the following terms and conditions regarding the trustee's power, authority, duties and responsibilities:

(1) The Commission shall select the trustee, subject to the consent of Donnelley, which consent shall not be unreasonably withheld. The trustee shall be a person with experience and expertise in acquisitions and divestitures. If Donnelley has not opposed the selection of a proposed trustee, in writing, within fifteen (15) days after notice by the Commission's staff to Donnelley of the identity of the proposed trustee, Donnelley shall be deemed to have consented to the selection of the proposed trustee.

(2) The trustee shall, subject to the prior approval of the Commission, have the exclusive power and authority to divest the Meredith/Burda Printing Business.

(3) The trustee shall have eighteen (18) months from the date of appointment to divest Meredith/Burda's Printing Business. If however, at the end of the 18-month period, the trustee has not submitted a plan for divesting the Meredith/Burda Printing Business or believes that such divestiture cannot be accomplished within a reasonable time, the trustee's period for divesting may be extended by the Commission or, in the case of a court-appointed trustee, by the court.

(4) The trustee shall have full and complete access to the personnel, books, records and facilities of any of the properties of Donnelley, or any other relevant information to divestiture of Meredith/Burda's Printing Business. Donnelley shall develop such financial or other information as the trustee may reasonably request. Donnelley shall cooperate with the trustee, and shall take no action to interfere with or impede the trustee's accomplishment of the divestiture. Any delays caused by Donnelley in meeting the reasonable requests of the trustee shall extend the time for the trustee to divest in an amount equal to the delay, as determined by the Commission or, for a court-appointed trustee, by the court. (5) Subject to Donnelley's absolute and unconditional obligations under paragraph II of this order, the trustee shall use his or her best R.R. DONNELLEY & SONS CO., ET AL. 133 36 Initial Decision efforts to negotiate the most favorable price and terms available in divesting Meredith/Burda's Printing Business. (6) The trustee shall serve, without bond or other security, at the cost and expense of Donnelley on such reasonable and customary terms and conditions as the Commission or a court may set. The trustee shall have the authority to retain, at the cost and expense of Donnelley, such consultants, attorneys, investment bankers, business brokers, accountants, appraisers and other representatives and assistants as are reasonably necessary to assist in the divestiture. The trustee shall account for all monies derived from divesting Meredith/Burda's Printing Business, and for all expenses incurred. After approval by the Commission, or, in the case of a courtappointed trustee, by the court, of the accounts of the trustee, including fees for his or her services, all remaining monies shall be paid to Donnelley and the trustee's power shall be terminated. The trustee's compensation shall be based at least in significant part on a commission arrangement contingent on the trustee divesting Meredith/Burda’'s Printing Business.

(7) Except in the case of reckless disregard of his or her duties, Donnelley shall indemnify the trustee and hold the trustee harmless against any losses, claims, damages or liabilities arising in any manner out of, or in connection with, the trustee's duties under this order.

(8) Within thirty (30) days after the appointment of the trustee, and subject to the prior approval of the Commission, and in the case of a court-appointed trustee, of the court, Donnelley shall execute a trust agreement that transfers to the trustee all rights and powers necessary to divest Meredith/Burda's Printing Business. (9) If the trustee ceases to act or fails to act diligently, a substitute trustee shall be appointed in the same manner as provided in paragraph III of this order.

(10) The Commission, and in the case of a court-appointed trustee, the court, may on its own initiative, or at the request of the trustee, issue such additional orders or directions as may be necessary and appropriate to accomplish the requirements of this order. (11) The trustee shall report in writing to Donnelley and the Commission every sixty days (60) concerning the trustee's efforts to divest Meredith/Burda's Printing Business. Initial Decision 120 F.T.C.

IV.

It is further ordered, That pending any divestiture required by this order, Donnelley shall take all measures necessary to maintain Meredith/Burda's Printing Business in its present or improved condition, and to prevent any deterioration, except for normal wear and tear, and otherwise not cause or permit impairment of the marketability or viability of Meredith/Burda's Printing Business. Donnelley shall not burden Meredith/Burda's Printing Business, or the corporation in the event of a public offering, with any obligations that may impair the viability of the business or frustrate the purposes of the divestiture, and in no event shall any obligations, apart from funded employment benefit pension funds, transferred by Donnelley be any greater than those carried by Meredith/Burda at the time of the Acquisition.

V.

It is further ordered, That for a period of ten (10) years from the date this order becomes final, Donnelley shall not, without the prior approval of the Commission, directly or indirectly, through subsidiaries or otherwise, (A) acquire the whole or any part of the stock, share capital, equity or other interest in, any concern, corporate or noncorporate, engaging in the supply of publication gravure printing within the United States; (B) acquire any assets used for or previously used for (and still suitable for use for) the supply of publication gravure printing within the United States; or (C) enter into any agreement, understanding or arrangement with any concern by which Donnelley would obtain the market share, in whole or in part, of such concern.

VI.

It is further ordered, That:

A. Within sixty (60) days from the date this order becomes final, and every sixty (60) days thereafter until it has fully complied with paragraphs II and III of this order, Donnelley shall file with the Commission a verified written report setting forth in detail the manner and form in which it intends to comply, is complying or has R.R. DONNELLEY & SONS CO., ET AL. 135 36 Initial Decision complied therewith. All such reports shall include, in addition to such other information and documentation as may hereafter be requested: (a) a specification of the steps taken by Donnelley to make public its desire to divest Meredith/Burda's Printing Business; (b) a list of all persons or organizations to whom notice of divestiture has been given; (c) a summary of all discussions and negotiations together with the identity and address of all interested persons or organizations; and (d) copies of all reports, internal memoranda, offers, counter offers, communications and correspondence concerning said divestiture; and B. On the anniversary of the date this order became final, and every anniversary thereafter for the following nine (9) years, Donnelley shall file with the Commission a verified written report of its compliance with paragraph V of this order. VIL.

It is further ordered, That for the purpose of determining or securing compliance with this order and subject to any legally recognized privilege, upon written request and on reasonable notice to Donnelley made to its principal office, Donnelley shall permit any duly authorized representatives of the Commission access, during office hours and in the presence of counsel, to inspect and copy all books, ledgers, accounts, correspondence, memoranda and other records and documents in the possession or under the control of Donnelley and to interview officers or employees of Donnelley relating to any matter contained in this order. Vil.

It is further ordered, That Donnelley shall notify the Commission at least thirty (30) days prior to any proposed changes in the corporation, such as dissolution, assignment or sale resulting in the emergence of a successor corporation, the creation or dissolution of subsidiaries or any other changes that may affect compliance obligations arising out of the order.

Opinion 120 F.T.C.

OPINION OF THE COMMISSION! BY STAREK, Commissioner:

I. INTRODUCTION On September 4, 1990, R.R. Donnelley & Sons Co. ("Donnelley") acquired all interests in the Meredith/Burda printing business ("Meredith/Burda") from Meredith Corporation ("Meredith") and Pan Associates, L.P. ("Pan").’ Prior to the subject acquisition, Donnelley and Meredith/Burda independently provided commercial printing services. Donnelley is the largest supplier of commercial printing services in the United States. IDF {1.° It provides printing services for publications including mail-order catalogs, newspaper inserts, magazines, books, directories, computer documentation, and financial documents. IDF J 1. Prior to the acquisition, Donnelley provided printing services from plants located throughout the United States employing the two primary printing technologies in publication printing: the gravure process and the offset process. IDF {{f 1- 5.* Immediately prior to the acquisition, Meredith/Burda was among the largest commercial printers in the United States, offering both gravure and offset printing services from four plants in the United States for a variety of publications. IDF { 7-9.° The following abbreviations are used in this opinion: ID Initial Decision (page no.) IDF Initial Decision (J no.) OA Tr. Oral Argument Transcript (page no.) RAB Respondent's Appeal Brief (page no.) CAB Complaint Counsel's Appeal Brief (page no.) RRB Respondent's Reply Brief (page no.) CPF Complaint Counsel's Proposed Findings of Fact (¥ no.) RPF Respondent's Proposed Findings of Fact ({ no.) Tr. Administrative Hearing Transcript (page no.) CX Complaint Counsel's Exhibit RX Respondent's Exhibit 2 Prior to the acquisition, Meredith/Burda was a joint venture comprising Meredith/Burda Companies, Inc. (a wholly-owned subsidiary of Meredith) and Pan (a limited partnership owned by members of the Burda. family). Pursuant to a purchase and sale agreement, Donnelley acquired Meredith/Burda. by acquiring (i) all of the issued and outstanding stock of Meredith/Burda Companies, Inc., and (ii) all of the limited partnership interests in Pan. CX-3-B. 3 Donnelley's fiscal 1990, 1991, and 1992 net sales were $3.498, $3.915, and $4.193 billion, respectively. IDF 6.

Donnelley also provides commercial printing services using other processes -- letterpress and flexography -- that are not important to the disposition of this case. See IDF {J 1, 29; RPF J 11. 5 In fiscal year 1989, the year prior to its acquisition by Donnelley, Meredith/Burda reported sales of $456.7 million. IDF { 9.

R.R. DONNELLEY & SONS CO., ET AL. 137 36 Opinion The Commission's complaint, issued October 11, 1990, charges that this acquisition may tend substantially to lessen competition in the supply of high volume publication gravure printing in two geographic markets -- the continental United States and the western United States -- in violation of Section 7 of the Clayton Act, 15 U.S.C. 18, and Section 5 of the Federal Trade Commission Act ("FTC Act"), 15 U.S.C. 45.° The Administrative Law Judge ("ALJ") assigned to this proceeding issued an Initial Decision on December 30, 1993, holding that the effect of the acquisition has been or may be substantially to lessen competition or to create a monopoly in the supply of high volume publication" gravure printing in the continental United States. The ALJ ordered Donnelley to divest the acquired Meredith/Burda business.

Respondents state four bases for appeal from the Initial Decision:’ (1) that the doctrine of issue preclusion bars the ALJ from reexamining factual issues decided in a 1990 hearing by the United States District Court for the District of Columbia on the Commission's motion for a preliminary injunction, FTC v. R.R. Donnelley & Sons Co., 1990-2 Trade Cas. (CCH) { 69,23-9 (D.D.C. 1990); (2) that the ALJ erred in recognizing a relevant product market of "high-volume publication gravure printing”; (3) that the ALJ erred in finding that the acquisition would tend substantially to lessen competition or tend to create a monopoly; and (4) that the liability finding and the order of divestiture are not based on substantial competent evidence.

The Commission reviews this matter de novo. We conclude that the ALJ and, thus, the Commission are not barred from reexamining factual issues decided in the preliminary injunction hearing. We further conclude that "high volume gravure printing" as proposed by Specifically, the complaint charges that Donnelley had entered into agreements with Meredith Corporation ("Meredith") and Pan Associates Limited Partnership ("Pan") that violated Section 5 of the Federal Trade Commission Act, as amended, 15 U.S.C. 45, and that pursuant to these agreements Donnelley had acquired certain business interests of Meredith and Pan in Meredith/Burda, and that such acquisition violated Section 7 of the Clayton Act, as amended, 15 U.S.C. 18, as well as Section 5 of the FTC Act.

7 Either side may appeal the ALJ’s decision to the full Commission, which will then enter its own decision. 16 CFR 3.52. If the Commission finds a violation of law, it may enter an order to divest and for other appropriate relief. 15 U.S.C. 21(b). Under both the Clayton Act and the FTC Act, such an order is subject to review in the court of appeals, 15 U.S.C. 21(c), 45, and, after the record is filed, "the jurisdiction of the court of appeals to affirm, enforce, modify, or set aside orders of the Commission . . . Shall be exclusive.” 15 U.S.C. 21(d). Any order the Commission may issue does not take effect until judicial review is complete. 15 U.S.C. 21(g). Opinion 120 F.T.C.

complaint counsel is not a relevant market for the purposes of assessing the legality of the acquisition under Section 7 of the Clayton Act. Because complaint counsel made no attempt to prove that the acquisition lessened competition in a broader market, this finding could be considered dispositive. It is undisputed that concentration and other characteristics of such a broader market are not conducive to an exercise of market power by the merged firm, unilaterally or in coordination with others. Moreover, if we assume for purposes of further analysis the existence of a relevant market for high volume publication printing, the analysis of potential anticompetitive effects reinforces our conclusion that the acquisition does not violate Section 7. We conclude that neither coordinated nor unilateral anticompetitive effects are a likely result of the acquisition in the assumed market. For the reasons set forth below, the complaint is dismissed.

II. PROCEDURAL HISTORY In July 1990, the Commission moved for a statutory temporary restraining order and preliminary injunction under Section 13(b) of the FTC Act, 15 U.S.C. 53(b), to prevent consummation of the acquisition. The Commission sought such relief "in aid of an FTC administrative proceeding.” E.g., FTC v. Food Town Stores, Inc., 539 F.2d 1339, 1341 (4th Cir. 1976). As provided by Section 13(b) itself, the Commission asked that injunctive relief be granted "pending the issuance of a complaint by the Commission and until such complaint is dismissed by the Commission or set aside by the court on review, or until the order of the Commission made thereon has become final."* Section 13(b) provides in pertinent part:

Whenever the Commission has reason to believe -- (1) _ that any person, partnership or corporation is violating, or is about to violate, any provision of law enforced by the Federal Trade Commission, and (2) That the enjoining thereof pending the issuance of a complaint by the Commission and until such complaint is dismissed by the Commission or set aside by the court on review, or until the order of the Commission made thereon has become final, would be in the interest of the public -the Commission by any of its attorneys designated by it for such purpose may bring suit in a district court of the United States to enjoin any such act or practice. Upon a proper showing that, weighing the equities and considering the Commission's likelihood of ultimate success, such action would be in the public interest, and after notice to the defendant, a temporary restraining order or a preliminary injunction may be granted without bond...

R.R. DONNELLEY & SONS CO., ET AL. 139 36 Opinion The district court judge who heard the case permitted the parties to present a truncated evidentiary hearing under a provision in the applicable local court rules that vests in the district court judge the discretion to determine the scope and nature of any hearing for a preliminary injunction. D.D.C. Local Rule No. 205(d).? The truncated evidentiary hearing lasted five partial days and ended with an additional half day for closing arguments. Neither the Commission nor the respondents objected to the truncated nature of the hearing, and no party moved under Fed. R. Civ. P. 65(a)(2) to have the hearing on the preliminary injunction consolidated with the trial on the merits.'° At all times all parties were engaged in a proceeding that sought only temporary and preliminary relief. On August 27, 1990, the district court denied the Commission's request for a preliminary injunction. The court's findings and conclusions, adopted nearly verbatim from proposed. findings submitted by Donnelley and Meredith, expressly recognized that the proceeding was one for preliminary relief: This matter was heard on the motion of plaintiff Federal Trade Commission (FTC) for a preliminary injunction pursuant to Section 13(b) of the FTC Act, 15 U.S.C. 53(b), seeking to enjoin the acquisition of any stock, assets, or other interest in Meredith Corporation, Meredith/Burda Companies, and Pan Associates, L.P., by R.R. Donnelley & Sons Company, pending the issuance of an administrative complaint by the FTC challenging this acquisition and final action thereon. FTC v. R.R. Donnelley & Sons Co., 1990-2 Trade Cas. (CCH) J 69,239, at 64,854 (D.D.C. 1990). Similarly, an addendum composed by the district court, and added to the proposed findings of Donnelley and Meredith, acknowledged the rushed circumstances in which the court had considered the Commission's motion for preliminary relief: The case became at issue (on August 21) during a week in which the undersigned was serving as the Court's Motions Judge, handling multiple other motions for temporary restraining orders and preliminary injunctions, etc., in cases assigned to This Rule has since been revised slightly, but the present published form is in all material respects the same as it was in August 1990. Any such consolidation would have been improper. While a court may certainly order. consolidation of hearings on requests for preliminary and permanent injunctions where both types of relief have been sought in the same action, it may not consolidate a cause of action not pleaded in the complaint. The Commission's complaint sought only preliminary relief. The granting of such relief was "an end in itself," and "the district court [was] not authorized to determine whether the antitrust laws have been or are about to be violated. That adjudicatory function is vested in FTC in the first instance." FTC v. Food Town Stores, Inc., 539 F.2d at 1342; see also FTC v. Beatrice Foods Co., 587 F.2d 1225, 1229 (D.C. Cir. 1978); FTC v. Lancaster Colony Corp., 434 F. Supp. 1088, 1090-91 (S.D.N.Y. 1977). Opinion 120 F.T.C.

other judges who temporarily are away from the courthouse. Because of that fact, and recognizing that yet another criminal drug trial will begin tomorrow, the Court asked both sides to submit proposed orders respectively granting and denying plaintiff's motion for a preliminary injunction. In the interest of time, the Court obviously has largely adopted defendants' proposed order. Id. at 64,856.

FTC counsel filed a notice of appeal and sought from the district court an injunction against the transaction pending appeal, in order to preserve the Commission's option to pursue appellate review. Following analysis of the decision, however, the Commission determined not to seek further review of the district court's denial of preliminary relief and on August 30, 1990, moved to dismiss its appeal. On September 4, 1990, Donnelley and Meredith consummated their transaction.

On October 11, 1990, the Commission issued an administrative complaint in this proceeding pursuant to Sections 7 and 11 of the Clayton Act, 15 U.S.C. 18 and 21, and Section 5 of the FTC Act, 15 U.S.C. 45, challenging the transaction.'' Section 11(b) of the Clayton Act, 15 U.S.C. 21(b), expressly vests the Commission with jurisdiction to determine the legality of a corporate acquisition under Section 7 and, if warranted, to order divestiture. On January 16, 1991, the ALJ denied Donnelley's cross motion for a summary decision and directed that its res judicata and collateral estoppel defenses to the complaint be stricken. The ALJ explained his rejection as follows:

Respondents argue that a five-day hearing is not a "curtailed" procedure, but it is in comparison with the typical Commission administrative hearing involving antitrust claims. These proceedings usually involve extensive formal discovery into such issues as relevant product and geographic markets, entry barriers and probable competitive injury. All of these issues are present in this case, and it has been my experience that matters of this nature involve, not five-day hearings, but hearings which may last several months and which involve many witnesses and hundreds of documents. Thus, I cannot conclude that the District Court was presented with every fact which bore on the issues before it, and I find that those facts can only be developed by discovery and a trial on the merits in this administrative proceeding. On February 20, 1991, Donnelley filed an "emergency" appeal from the ALJ's decision. Donnelley's motion was not received by the Although named in the original complaint, Meredith was dismissed as a respondent by Stipulation of April 19, 1991, R.R. DONNELLEY & SONS CO., ET AL. 141 36 Opinion Commission, as it was not authorized under Commission Rule 3.23.” Donnelley then filed a petition with the Court of Appeals for the Seventh Circuit seeking review of the ALJ's decision and, effectively, asking that court to direct the Commission to dismiss the complaint under the doctrine of issue preclusion. The Court of Appeals rejected Donnelley's petition for want of jurisdiction under 5 U.S.C. 704, holding that the ALJ's decision was not a reviewable "final order" of the Commission. R.R. Donnelley & Sons Co. v. FTC, 931 F.2d 430 (7th Cir. 1991).

After extensive pretrial discovery, hearings were held in Washington, D.C., and Chicago, Illinois, from January 25, 1993, to June 17, 1993. The parties filed their proposed findings of fact on September 17, 1993, and the ALJ closed the record on October 8, 1993.

On December 30, 1993, the ALJ issued an initial decision ("ID"), pursuant to the Commission's rules of practice. 16 CFR 3.51. The ID holds that the Commission has jurisdiction over the subject matter of this proceeding and over defendants Donnelley and Pan, and that Donnelley and Pan were engaged in "commerce" as defined in Section 1 of the Clayton Act, 15 U.S.C. 12, and Section 4 of the FTC Act, 15 U.S.C. 44. ID at 91.

With respect to the market conditions relevant to the competitive analysis of this acquisition, the ALJ found: (i) that the relevant market within which to evaluate the competitive effects of the acquisition is "the supply of high volume publication gravure printing" in "the continental United States"; (ii) that Donnelley and Meredith/Burda were actual, direct and substantial competitors in this relevant market; and (iii) that "barriers to entry into the relevant market are substantial, and substantial harm to competition would occur until new entry could be accomplished." ID at 91-92. The ALJ held that the effect of this acquisition has been or may be substantially to lessen competition or to tend to create a monopoly in the alleged market in violation of Section 7 of the Clayton Act in the following ways:

Rule 3.23, patterned after 28 U.S.C. 1292, allows Commission review only of specified types of interlocutory rulings by the ALJ (not involved here), or where the ALJ certifies his ruling for interlocutory Commission review. In this case the ALJ did not certify his ruling for Commission review. Instead, the motion was placed on the public record by the Commission's Secretary on February 28, 1990, as an ex parte communication.

Opinion 120 F.T.C.

(a) It eliminated actual competition between Donnelley and Meredith/Burda in the relevant market;

(b) It significantly increased the already high levels of concentration in the relevant market;

(c) It created a firm whose share of the relevant market is so high that it has achieved the position and market power of a dominant firm;

(d) It eliminated Meredith/Burda as a substantial independent competitive force in the relevant market; and (e) It increased the likelihood of successful anticompetitive conduct, non-rivalrous behavior and actual or tacit collusion among the firms in the relevant market.

ID at 92. The ALJ concluded that the ultimate effect of the acquisition is to "increase the likelihood that firms will increase prices and restrict the output of high volume publication gravure printing." ID at 92.

To remedy these anticompetitive effects, the ALJ ordered Donnelley to divest the acquired Meredith/Burda business comprising, inter alia, four gravure printing plants to an acquirer that obtains the prior approval of the Commission. Among other standard provisions, the order prohibits Donnelley, for a period of ten years from the date the order becomes final, from acquiring without prior Commission approval (i) any interest in a firm engaged in the supply of publication gravure printing in the United States, or (ii) any assets used (or suitable for use) for the supply of publication gravure printing within the United States. ID at 93-104. III. ISSUE PRECLUSION Principles of issue preclusion or collateral estoppel do not preclude the ALJ or the Commission from deciding the merits of the complaint.

Respondents argue that the district court's decision denying the Commission's request for preliminary injunctive relief under Section 13(b) of the FTC Act -- on grounds that the Commission had not adequately established the relevant product market -- estops the Commission from adjudicating the question of product market definition in an administrative proceeding under Section 11 of the Clayton Act and Section 5 of the FTC Act. The district court R.R. DONNELLEY & SONS CO., ET AL. 143 36 Opinion determined, after a truncated hearing, that the Commission had not shown the requisite "likelihood of success" to warrant a preliminary injunction pending completion of its administrative proceeding.'? The district court did not, and did not purport to, decide the case, or any of its subsidiary factual issues, on the merits.'* The district court was not called upon to reach a "final resolution” on the antitrust issues, but only to determine whether the FTC made "a showing adequate to justify preliminary relief." FTC v. Alliant Techsystems Inc., 808 F. Supp. 9, 19 (D.D.C. 1992); FTC v. Beatrice Foods, Inc., 587 F.2d 1225, 1229 (D.C. Cir. 1978).'° Under established principles, the court's denial of a preliminary injunction does not estop the Commission from exercising its express statutory authority to adjudicate the legality of the transaction.’ A. General Lack of Collateral Preclusive Effect Ordinarily a court is not estopped from deciding the merits of a disputed issue unless the decision alleged to create an estoppel effect was "final, and on the merits.” 18 Wright, Miller & Cooper, Federal Practice and Procedure Section 4427, at 269 (1981); see also Restatement (Second) of Judgments Section 27 (1982) (an issue is precluded only when it is "determined by a valid and final judgment, 13 See supra pp. 7-9 M4 Indeed, in rejecting the Commission's request for a preliminary injunction, the district court commented at length on perceived gaps in the Commission's preliminary showing and on adjustments the Commission trial staff made in the market definition throughout the five-day hearing. FTC v. R.R. Donnelley & Sons Co., 1990-2 Trade Cas. (CCH) J 69,239, at 64,854-55. Similarly, in subsequently denying the Commission's request for a stay pending appeal, the court again acknowledged the preliminary nature of the case, stating that it had "found little likelihood of success on the merits." The district court thus expressly recognized that it had not heard the case on the merits. 1S wThe question whether the acquisition actually violates the antitrust laws is reserved for the Commission and is not before [the district court in a preliminary injunction hearing]. The Commission meets its burden ‘if it shows preliminarily, by affidavits or other proof, that it has a fair and tenable chance of ultimate success on the merits." FTC v. Southland Corp., 471 F. Supp. 1, 3 (D.D.C. 1979) (citation omitted). The Seventh Circuit has noted: "One of the main reasons for creating the Federal Trade Commission and giving it. . . jurisdiction to enforce the Clayton Act was that Congress. . thought the assistance of an administrative body would be helpful in resolving such [antitrust] questions." Hospital Corp. of America v. FTC, 807 F.2d 1381, 1386 (7th Cir. 1986), cert. denied 431 U.S. 1038 (1987). Therefore, "Section 13(b) does not contemplate a full-blown trial-type hearing in District Court." FTC v. Imo Indus., 1992-2 Trade Cas. (CCH) 69,943 (D.D.C. 1992). Any order issued by the Commission as a result of the administrative proceeding is reviewable in a court of appeals. See 15 U.S.C. 45(c).

'6 See R.R. Donnelley & Sons Co. v. FTC, 931 F.2d 430 (7th Cir, 1991); Southwest Sunsites, Inc., 98 FTC 866, 870-71 (1981!) (interlocutory order) (decisions in a preliminary injunction action brought under Section 13(b) do not collaterally estop the Commission from deciding the merits in a full administrative trial); see also Statement of Federal Trade Commission Policy Regarding Administrative Merger Litigation Following Denial of Preliminary Injunction (June 21, 1995). Opinion 120 F.T.C.

and the determination is essential to the judgment"); accord Ashe v. Swenson, 397 U.S. 436, 443 (1970) ("valid and final judgment"). Because a preliminary injunction hearing is not designed to decide the case either finally or on the merits, decisions made in that context are rarely granted collateral effect. As the Supreme Court has observed:

Since . . . likelihood of success on the merits was one of the factors the District Court and the Court of Appeals considered in granting . . . a preliminary injunction, it might be suggested that their decisions were tantamount to decisions on the underlying merits. .. . This reasoning fails, however, because it improperly equates "likelihood of success" with "success," and what is more important, because it ignores the significant and procedural differences between preliminary and permanent injunctions.

The purpose of a preliminary injunction is merely to preserve the relative positions of the parties until a trial on the merits can be held. Given this limited purpose, and given the haste that is often necessary if those positions are to be preserved, a preliminary injunction is customarily granted on the basis of procedures that are less formal and evidence that is less complete than in a trial on the merits. A party thus is not required to prove his case in full at a preliminary injunction hearing . . . and the findings of fact and conclusions of law made by a court granting a preliminary-injunction are not binding at trial on the merits. .. . University of Texas v. Camenisch, 451 U.S. 390, 394-95 (1981) (citations omitted, emphasis added); see A.J. Canfield Co. v. Vess Beverages, Inc., 859 F.2d 36, 38 (7th Cir. 1988). The law of the D.C. Circuit is particularly pertinent in determining the preclusive effect to be given findings of a district court in that circuit. The D.C. Circuit has repeatedly refused to accord preclusive effect to findings made in a preliminary injunction hearing. In rejecting the contention that categorically-stated findings of a district court in a preliminary injunction proceeding might have preclusive effect, the court observed:"”

Because the ultimate merits are not "necessary" to the outcome of a preliminary injunction under Section 13(b), the collateral estoppel doctrine is not relevant to this case. E.g. Parklane Hosiery Co. v. Shore, 439 U.S. 322, 326 n.5 (1979) ("Under the doctrine of collateral estoppel ... the judgment in the prior suit precludes relitigation of issues actually litigated and necessary to the outcome of the first action."); see Montana v. United States, 440 U.S. 147, 153 (1979) ("Under collateral estoppel, once an issue is actually and necessarily determined by a court of competent jurisdiction, that determination is conclusive in subsequent suits based on a different cause of action involving a party to the prior litigation."); cf Public Service Co. of Indiana v. EPA, 682 F.2d 626, 630-31 (7th Cir. 1982), cert. denied, 459 U.S. 1127 (1983). The district court was authorized to consider only the “likelihood of success.” The actual merits were not litigated, and even if plaintiffs thought they were litigated, the "merits" were not "necessary to the outcome of the first action." Jd. To hold otherwise "improperly equates ‘likelihood of success’ with ‘success,’ and ... ignores the significant procedural differences between preliminary and permanent injunctions." Camenisch, 451 US. at 394. R.R. DONNELLEY & SONS CO., ET AL. 145 36 Opinion To the extent that the findings and conclusions of the District Judge purported to settle finally the questions of law and fact raised by the complaint, those findings and conclusions went beyond the determination the judge was called upon to make, and should not be regarded as binding in further proceedings in the trial court. Industrial Bank of Washington v. Tobriner, 405 F.2d 1321, 1324 (D.C. Cir. 1968) (footnotes omitted); accord Hunter Douglas, Inc. v. Sheet Metal Workers Intl, 714 F.2d 342, 346 (4th Cir. 1983); Berrigan v. Sigler, 499 F.2d 514, 518 (D.C. Cir. 1974); 11 Wright & Miller, Federal Practice and Procedure Section 2962, at 630-31. The Seventh Circuit has stated that "in certain rare instances, decisions granting or denying preliminary relief will be given preclusive, effect." Canfield, 859 F.2d at 38, citing and following Miller Brewing Co. v. Jos. Schlitz Brewing Co., 605 F.2d 990 (7th Cir. 1976), cert. denied, 444 U.S. 1102 (1980). The "rare" situation found in Miller Brewing, however, is not present here. In Miller Brewing, Miller sued Schlitz for trademark infringement based on the latter's use of the term "Lite" for a reduced-calorie beer. The court of appeals summarily affirmed judgment for Schlitz because the court of appeals had, in an earlier preliminary injunction proceeding, decisively held that "Lite" was generic when applied to beer and hence not protected by trademark law. The court of appeals held that its earlier decision was decisive because, among other things, Miller in the earlier injunction proceeding had conceded that the evidentiary record before the court then was as complete as it could ever be on that issue and that the fundamental facts were not in dispute. 605 F.2d at 995. The court, thus, held that its earlier determination that "Lite" is a generic term was "an insuperable obstacle to Miller's claims.""* Miller Brewing is plainly inapposite. The Commission did not present at the expedited preliminary injunction proceeding all the evidence in support of its alleged market definition that it marshalled in the administrative proceeding. Although Donnelley was perfectly 18 Similarly, in Canfield, the court of appeals held that a plaintiff was collaterally estopped from bringing a trademark injunction suit against a competitor's use of the term "chocolate fudge” for diet soda, because, in at least one prior case involving the same plaintiff, another court had made a decisive finding that the term "chocolate fudge” was generic. Notably, however, the court rejected an argument that it had, in the context of a previous affirmance of a preliminary injunction, made a decisive ruling in plaintiff's favor on the issue. The, court explained that in view of the "general presumption against giving preclusive effect to preliminary [injunction] rulings” and in view of the fact that its prior opinion affirming the preliminary injunction "dealt with probabilities only, since we were determining the likelihood of success on the merits," there was no resulting preclusive effect running in favor of the plaintiff. 859 F.2d. at 38.

Opinion 120 F.T.C.

within its rights to insist that the preliminary injunction hearing be conducted expeditiously, it may not elevate that truncated hearing into a decision on the merits.'? Moreover, in this case, the district court lacked authority (even had it purported to do so) to resolve with finality any issue in the case.”” Through Section 11(b) of the Clayton Act and Section 5(b) of the FTC Act, Congress expressly authorized the Commission, in its sole discretion, to determine the legality of corporate acquisitions and other antitrust violations by means of an administrative proceeding, subject to review in the court of appeals. To grant preclusive effect to decisions of the district court in a Section 13(b) injunctive proceeding would usurp the Commission's statutory fact-finding role under these statutes, in plain contravention of the will of Congress. See NLRB vy. Denver Building & Construction Trades Council, 341 U.S. 675 (1951).”"

In creating the Federal Trade Commission and establishing a procedure for administrative determination of the legality of corporate acquisitions and other conduct, Congress recognized the 19 In this regard, the Supreme Court has noted the appropriate procedures for seeking to have a preliminary injunction hearing consolidated with the trial on the merits: Should an expedited decision on the merits be appropriate, Rule 65(a)(2) of the Federal Rules of Civil Procedure provides a means of securing one. That Rule permits a court to "order the trial of the action on the merits to be advanced and consolidated with the hearing of the application." Before such an order may issue, however, the courts have commonly required that "the parties should normally receive clear and unambiguous notice [of the court's intent to consolidate] either before the hearing commences or at a time which will still afford the parties a full opportunity to present their respective cases.” University of Texas v. Camenisch, 451 F.2d at 395 (brackets in original, quoting Pughsley v. 3750 Lake Shore Drive Coop. Bldg., 463 F.2d 1055, 1057 (7th Cir. 1972)). Respondents did not follow this procedure. In any event, such consolidation would have been improper. See supra note 10. 20 In Miller Brewing, the court that decided Miller's motion for preliminary injunction would also decide its request for final relief. It seems clear that Miller Brewing is no more than an application of the view that on an application for preliminary injunction, the trial court (or, as in Miller Brewing, an appellate court on an appeal from the grant or denial of preliminary injunction) may, in an appropriate case, go beyond the issue of preliminary relief and indicate (or direct) that the case should be dismissed in its entirety because it has fully and finally resolved an issue that presents an "insuperable" obstacle to maintaining the case. E.g., Deckert v. Independence Shares Corp., 311 U.S. 282, 287 (1940) ("If insuperable objection to maintaining the bill clearly appears, it may be dismissed and the litigation terminated") (citation omitted).

21 Respondents attempt to distinguish Denver Building on the dual grounds that the NLRB has exclusive jurisdiction over labor matters, while the Commission does not have exclusive jurisdiction over antitrust cases, and that the NLRB statute in that case was designed "to assist in a preliminary investigation," while the Commission in this case had already conducted a pre-merger investigation into the challenged transaction. RAB at 5 n.3. These arguments are not compelling. First, under Section 13(b) of the FTC Act the Commission is the primary (and exclusive) factfinder in all cases in which it seeks preliminary relief under that statute in aid of its adjudicative proceeding. Second, Section 13(b) is also intended "to assist" the Commission's law enforcement efforts. And neither the Hart-Scott- Rodino Act, 15 U.S.C. 18a, nor Section 13(b), nor the federal civil discovery rules, intimate that the Commission's discovery (and investigation) rights terminate at the conclusion of a preliminary injunction action.

R.R. DONNELLEY & SONS CO., ET AL. 147 36 Opinion value of specialized expertise in a complex area of the law. See, e.g., Atlantic Refining Co. v. FTC, 381 U.S. 357, 367 (1965); Stanley Works v. FTC, 469 F.2d 498, 505 (2d Cir. 1972), cert. denied, 412 U.S. 928 (1973).” Section 13(b) of the FTC Act, added to the statute in 1973, was designed to permit the Commission to secure preservation of the status quo pending determination of the case on the merits in an administrative proceeding, in recognition of the difficulty of reconstituting some merged parties as viable entities through divestiture. FTC v. Weyerhaeuser Co., 665 F.2d 1072, 1081 (D.C. Cir. 1981); see also Kenneth Elzinga, The Antimerger Laws: Pyrrhic Victories, 12 J.L. & Econ. 43 (1969). But nothing in the statute or its legislative history purports to divest the Commission of its preexisting statutory authority to adjudicate the legality of acquisitions and other practices on their merits.” To the contrary, courts have expressly recognized that denial of preliminary injunctive relief under Section 13(b) of the FTC Act comes without prejudice to the Commission's authority to reach a contrary conclusion when it adjudicates the merits of the case. See FTC v. Elders Grain Co., 868 F.2d 901 (7th Cir. 1989). Compare FTC vy. Simeon Mgmt. Corp., 532 F.2d 708, 717 (9th Cir. -1976) (preliminary injunction denied with the court, per Judge Kennedy, ‘intimat[ing] no view . . . as to the appropriate disposition" on the merits), with Simeon Mgmt. Corp. v. FTC, 5 7 9 F.2d 1137 (9th Cir. 1978) (subsequent cease and desist order upheld on review of final agency action).” 2 See also Hospital Corp. of America v. FTC, 807 F.2d 1381, 1387 (7th Cir. 1986) (Posner, J.), cert. denied, 481 U.S. 1038 (1987) ("One of the main reasons for creating the Federal Trade Commission and giving it concurrent jurisdiction to enforce the Clayton Act was that Congress distrusted judicial determination of antitrust questions. It thought the assistance of an administrative body would be helpful in resolving such questions and indeed expected the FTC to take the leading role in enforcing the Clayton Act... .. See Henderson, THE FEDERAL TRADE COMMISSION, ch. 1 (1924)"). The risks inherent in a lack of specialized expertise in adjudicating complex antitrust issues are magnified in a preliminary injunction proceeding. See Richard Posner, ECONOMIC ANALYSIS OF LAW 554 (4th ed. 1992) ("The problem for the judge asked to grant a preliminary injunction is that he is being asked to rule in a hurry, on the basis of incomplete information. The risk of error is high."). 23 A proviso to Section 13(b) permits the Commission to request, and the court to grant, a permanent injunction in a "proper case," in lieu of administrative resolution of the matter. See, e.g., United States v. JS&A Group, Inc., 716 F.2d 451 (7th Cir. 1983). However, the choice whether to request such relief rests solely within the Commission's discretion, and the legislative history to the proviso indicates that it is to be invoked where the agency concludes that a case presents no issues warranting detailed administrative consideration. Jd. at 456-57. Neither in this case, nor in any other suit to date under Section 13(b) to preliminarily enjoin a corporate acquisition, has the Commission asked the district court to render a final decision on the merits. In all such cases, as in this one, the Commission has exercised its express statutory authority under the Clayton and FTC Acts to decide the merits. 24 The obverse is equally correct. Compare FTC v. Weyerhaeuser Co., 665 F.2d at 1075 (finding that Commission demonstrated likelihood of success under Section 7 in five day preliminary injunction hearing), with Weyerhaeuser Co., 106 FTC 172, 265 (1985) (Commission finding lack of Section 7 violation and dismissing complaint after administrative hearing). Plainly, in Weyerhaeuser, the Opinion 120 F.T.C.

Thus, under the circumstances here, the general rule -- that decisions made in a preliminary injunction hearing do not have preclusive effect on a subsequent trial on the merits -- applies with particular clarity.

B. No Basis for Exception to the General Rule There is no basis for a departure from the general rule in this case. Donnelley argues that it seeks to create only a narrow exception to the general rule, stating two bases for giving preclusive effect to the decisions of preliminary injunction hearing: (1) where the preliminary injunction proceeding did not afford a "full and fair opportunity to litigate"; and (2) where the preliminary injunction judgment is not vacated.”> In fact, Donnelley's proposed standards for according collateral preclusive effect to a preliminary injunction are no standards at all.

On the first basis, the Supreme Court in Kremer v. Chemical Construction Corp., 456 U.S. 461, 481-82 (1982), equates "full andfair opportunity" with "minimum guarantees of due process. Thus, Donnelley appears to suggest that collateral estoppel attaches each time a court denies a preliminary injunction, unless it does not afford the minimum due process. Under this standard, a preliminary injunction hearing would always bar subsequent litigation on the merits since the proper entry of a preliminary injunction decision must meet standards of minimum due process: Section 52(a) of the Federal Rules of Civil Procedure, which applies to all preliminary injunctions hearings in federal courts, requires that "in granting or refusing interlocutory injunctions the court shall... set forth the findings of fact and conclusions of law which constitute the grounds of its action." See FTC v. Beatrice Foods Co., 589 F.2d at 1235.7 On the second basis for an exception to the general rule, Donnelley cites United States v. Munsingwear, 340 U.S. 36 (1950), for its view that the Commission should have moved the court of appeals to vacate the district court's decision in order to avoid Commission did not even attempt to argue that the respondent was precluded by collateral estoppel principles from litigating the merits in the administrative proceeding that followed the district court's findings that were favorable to the Commission, even though the "administrative complaint charge[d] the same anticompetitive effects earlier alleged in the Commission's complaint for a preliminary injunction." 665 F.2d at 1075 n.8.

25 RAB at 5-10.

26 If the “full and fair opportunity” is something greater than a “minimum guarantees of due process,” Donnelley does not specify what it is. R.R. DONNELLEY & SONS CO., ET AL. 149 36 Opinion preclusion. The remedy of vacating a lower court's opinion, however, is available only in the limited class of cases in which the litigated issues become moot before an appeal can be prosecuted. In essence, Donnelley would require every party losing a preliminary injunction to appeal in order to avoid preclusive effect. This position poses a standard that is contrary to the sound administration of justice. It would force the Commission to pursue an appeal and emergency relief in the court of appeals, even if it believed the district court did not commit any reversible error under the standards applicable to preliminary injunction cases merely to create the circumstances in which the Commission could then ask the court of appeals to vacate the lower court's decision as moot.”’ Munsingwear simply recognizes the principle that a party ought not to be burdened with an adverse decision that it has been denied an opportunity to appeal, through no fault of its own. By contrast, to the extent Donnelley's argument is that every preliminary injunction must be either appealed to its conclusion or vacated without an appeal, Donnelley would place an enormous burden on the appellate courts. And to the extent that Donnelley's argument is that the Commission's case was moot (because Donnelley had consummated the transaction and thereby precluded the Commission from obtaining preliminary relief -- the only relief the Commission sought), Donnelley would require the Commission to ask an appellate court either to reverse or to vacate every preliminary merger injunction the Commission loses. This standard also would impose an inordinate burden on the Any appeal requires more than a mere disagreement with the trial court about the result in the case. The critical evaluation of an appeal entails many considerations, including an awareness that appellate courts distinguish among errors of law, errors of fact, and abuses of discretion in reviewing district court decisions. E.g., Vision Sports, Inc. v. Melville Corp., 888 F.2d 609, 612 (9th Cir. 1989); Baja Contractors, Inc. v. City of Chicago, 831 F.2d 667 (7th Cir. 1987), cert. denied, 485 U.S. 993 (1988) (factual and legal errors constitute an abuse of discretion). The district court's legal analysis fairly correctly articulated the legal standards applicable to an action by the Commission for a preliminary injunction. Assuming the district court erred, its errors concerned questions of fact, not of law. For the Commission to have succeeded in appealing the district court's denial of an injunction, the Commission would have been required to show that the district court's findings of fact were "clearly erroneous." See Fed. R. Civ. P. 52(a) ("Findings of fact, whether based on oral or documentary evidence, shall not be set aside unless clearly erroneous, and due regard shall be given to the opportunity of the trial court to judge the credibility of witnesses"). The standard for judicial review of a district court's factual findings is much closer to the “abuse of discretion" standard than it is to the “error of law" standard. An "abuse of discretion" standard is very difficult for an appellant to meet, and courts will rarely overtum a district court for declining to grant equitable relief, especially when that court finds that the equities do not warrant an injunction. Compare FTC v. Weyerhaeuser Inc., 665 F.2d 1072, with FTC v. PPG Indus. Inc., 798 F.2d 1500 (D.C. Cir. 1986). Opinion 120 F.T.C.

appellate courts. Neither Munsingwear nor any public policy compels such a result.”

The Commission's preliminary injunction action did not seek a resolution of the merits, the court did not decide the case on, the merits, and the Commission was never given "clear and unambiguous notice" -- nor indeed any notice at all? -- that its action for a preliminary injunction was anything more than a preliminary assessment by the district court of the Commission's likelihood of success, given the evidence the Commission presented in that proceeding.” Indeed, the district court's decision clearly contemplates that the merits of this matter will be adjudicated in an FTC administrative proceeding.*' Donnelley's belated attempt to turn that proceeding into something that it was not is rejected. IV. MARKET DEFINITION Section 7 of the Clayton Act prohibits acquisitions, the effect of which "may be substantially to lessen competition or tend to create a monopoly" 15 U.S.C. 18. The language of Section 7 indicates that a plaintiff need not prove that an anticompetitive effect is a certainty. California v. American Stores Co., 495 U.S. 271, 284 (1990) ("plaintiff need only prove that [the acquisition's] effect may be substantially to lessen competition").*? But Section 7 requires a prediction of probable anticompetitive effects, not ephemeral The more appropriate practice is for the Commission (i) to confine its requests that district court decisions be vacated to those moot cases that articulate legal principles the Commission believes are erroneous but which it cannot vindicate on appeal, but (ii) not to seek vacation of the adverse preliminary decisions that turn on their unique facts. The efficacy of this approach by the Commission depends on its continued ability to litigate the full merits of a case in the related administrative proceeding when it is denied preliminary relief in the district court. 29 Pughsley v. 3750 Lake Shore Drive Coop. Bldg., 463 F.2d at 1057; see supra note 19. 30 Although not dispositive, the relative length and scope of the administrative trial, and the volume and quality of evidence presented therein, is informative. The administrative trial lasted five months, as compared with five days for the preliminary injunction hearing. Moreover, complaint counsel presented live testimony from 40 witnesses and presented 1,450 evidentiary exhibits, as compared with 6 witnesses and 100 exhibits at the preliminary injunction hearing. CAB at 74. 31 See supra pp. 8-9.

Although the complaint challenges the acquisition under both Section 7 of the Clayton Act and Section 5 of the FTC Act, the analytical standards for assessing legality in this context are read coextensively. See FTC v. PPG Indus., Inc., 798 F.2d 1500, 1501 n.2 (D.C. Cir. 1986); FTC v. Pepsico, Inc., 477 F.2d 24, 28 n.6 (2d Cir. 1973); Grand Union Co., 102 FTC 812, 1027 (1983). 3 Section 7 does not require "a certainty” or “even a high probability” that an acquisition will substantially lessen competition. United States v. General Dynamics Corp., 415 U.S. 486, 505 (1974); FTC v. Elders Grain, Inc., 868 F.2d 901, 906 (7th Cir. 1989). See also United States v. Western EEC. Co., 767 F. Supp. 308, 330 (D.D.C. 1991), aff'd, 993 F.2d 1572 (D.C. Cir.), cert. denied, 114 S. Ct. 487 (1993) ("probability rather than “certainty” is the applicable standard in antitrust law). R.R. DONNELLEY & SONS CO., ET AL. 151 36 Opinion possibilities. Brown Shoe Co. v. United States, 370 U.S. 294, 323 (1962); see Midcon Corp., 112 FTC 93, 169-170 (1989). The central concern of Section 7 is that acquisitions "should not be permitted to create or enhance market power or to facilitate its exercise." U.S. Department of Justice and Federal Trade Commission Horizontal Merger Guidelines ("Merger Guidelines") Section 0.1. Market power is "the ability profitably to maintain prices above competitive levels for a significant period of time." Jd. Section 0.1. Thus, the ultimate issue under Section 7 is whether the challenged acquisition likely will enable the merging firm, acting unilaterally or collectively with other firms, to increase prices above competitive price levels.*° See, e.g., Hospital Corp. of America v. FTC, 807 F.2d 1381, 1386 (7th Cir. 1986) (Posner, J.), cert. denied, 481 U.S. 1038 (1987) (Section 7 requires "judgment [as to] whether the challenged acquisition is likely to hurt consumers, as by making it easier for firms in the market to collude, tacitly or explicitly, and thereby force prices above, or farther above, the competitive level"). The ultimate question under Section 7 may be stated more broadly to include a prediction of adverse effects in competitive dimensions other than price -- reductions in output, product quality, or innovation. See Merger Guidelines 0.1, n.6; see, e.g., United States v. Philadelphia Natl Bank, 374 U.S. 321, 368-69 (1963); FTC v. PPG Indus., Inc., 628 F. Supp. 881, 885 (D.D.C.), affd in part, 798 F.2d 1500 (D.C. Cir. 1986).

The prediction that an acquisition likely will have adverse competitive effects requires a series of successive determinations outlined in the Merger Guidelines Section 0.2.°%° First, the Commission must find that the acquisition would increase concentration and result in a concentrated relevant market, properly defined and measured. "Determination of a relevant market is a necessary predicate to a finding of a [Section 7] violation." United States v. E.I. dupont de Nemours & Co., 353 U.S. 586, 593 (1957). Section 7 of the Clayton Act prohibits acquisitions "where in any line of commerce . . . in any section of the country, the effect of such acquisition may be substantially to lessen competition." 15 U.S.C. 18 34 4 Trade Reg. Rep. (CCH) ¥ 13,104 (April 2, 1992). 3 See Merger Guidelines Section 1.11 (the prevailing preacquisition price level is generally used as a proxy for the competitive price level). 36 The Commission uses the framework set out in the Merger Guidelines for determining whether to challenge horizontal acquisitions. Although the Merger Guidelines are not binding on courts, courts of appeal have considered them in determining the impact on competition of proposed acquisitions. See, e.g., United States v. Baker Hughes, 908 F.2d 981, 983 n.3 (D.C. Cir, 1990); PPG, 798 F.2d at 1503. Opinion 120 F.T.C.

(emphasis added). The purpose of market definition identify those sections of the economy that may be exposed by the challenged acquisition to a substantial lessening of competition. Owens-Illinois, Inc., 115 FTC 179 (1992) [FTC Dkt. No. 9212 (Feb. 26, 1992), slip op. at 4].*” As suggested by the language of Section 7, the relevant market must be defined along both a product and a geographic dimension. United States v. General Dynamics Corp., 415 U.S. 486, 510 (1974) ("delineation of proper geographic and product markets is a necessary precondition to assessment of the probabilities of a substantial effect on competition within them"). Complaint counsel bear the burden of proving a relevant market within which anticompetitive effects are likely as a result of the acquisition. Kaiser Aluminum & Chem. Corp. v. FTC, 652 F.2d 1324 (7th Cir. 1981). See United States v. Connecticut Natl Bank, 418 U.S. 656, 669 (1974).

The ALJ found that the relevant product market is "high volume publication gravure printing," as alleged by complaint counsel. ID at 81-85. The ALJ further found that the United States constitutes a relevant geographic market within which to assess the competitive effects of the acquisition, and rejected complaint counsel's proposed Western United States market. ID at 85-86. Respondents appeal the ALJ's product market determination, arguing that competition from offset printing services is sufficient to undermine any attempt by respondents, unilaterally or collectively with other gravure printers, to exercise market power with respect to gravure printing services. RAB at 10-44. Neither complaint counsel nor respondents appeal the ALJ's geographic market findings.** We find that the ALJ's conclusions as to geographic market are well supported, but we reject the ALJ's conclusions as to product market. We conclude that "high volume publication gravure printing" as proposed by complaint counsel is not a relevant market for the purposes of assessing the competitive effects of the acquisition. Using the price discrimination methodology proposed by complaint counsel, adjusted to reflect actual substitution possibilities, we find that offset printing is used extensively in high volume publication printing. Complaint counsel estimate that, in 1990, offset accounted 37 See United States v. du Pont, 353 U.S. at 593 ("substantiality [of any lessening of competition] can be determined only in terms of the market affected"). 38 CAB at 51 n.72 (Complaint counsel do not "formally challeng[e]"the ALJ's rejection of the proposed "Western United States" geographic market, but nevertheless contend that the evidence supports such a separate antitrust market). R.R. DONNELLEY & SONS CO., ET AL. 153 36 Opinion for 24.1% of print jobs with more than sixteen pages and print runs of more than five million copies. In the "core" of complaint counsel's proposed market -- print jobs with more than thirty-two pages and print runs of more than ten million copies -- offset accounted for 13.5%.” These conclusions could be considered dispositive: Complaint counsel made no attempt to prove that the acquisition lessened competition in a broader market that includes offset printing, and it is undisputed that concentration and other characteristics of such a broader market are not conducive to an exercise of market power by the merged firm, unilaterally or in coordination with others. Assuming, however, the existence of a relevant market for high volume publication printing -- contrary to our finding that printers cannot identify customers with inelastic demand according to the alleged parameters of the market -- the analysis of potential anticompetitive effects reinforces our conclusion that the acquisition does not violate Section 7. A. Product Market: General Standards A market may be defined as "a product or group of products and a geographic area in which it is produced or sold such that a hypothetical [monopolist] of those products in that area likely would impose at least a ‘small but significant and nontransitory' increase in price. A relevant market is a group of products and a geographic area that is no bigger than necessary to satisfy this test." Merger Guidelines Section 1.0.*' Thus, a relevant product market is the smallest grouping of products whose sellers, if unified by a hypothetical cartel or merger, could profitably increase prices significantly above the competitive level.“ See H.J., Inc. v. International Tel. & Tel. Corp., 867 F.2d 1531, 1537 (8th Cir. 1989) (a market is “any grouping of sales whose sellers, if unified by a hypothetical cartel or merger, could raise prices significantly above the competitive level"); Owens-Illinois, Inc., slip op. at 4-5. Market definition under the Merger Guidelines focuses solely on demand substitution factors -- i.e., possible consumer responses. ° cx. 1167-C-1.

Id. These estimates may understate the proportion of offset printing in the proposed market. See id. (defining the hypothetical monopolist as "a hypothetical profit-maximizing firm, not subject to price regulation, that was the only present and future producer or seller of those products in that area").

2 The prevailing pre-acquisition price level is generally used as a proxy for the competitive price level. Id. Section 1.11.

Opinion 120 F.T.C.

Supply substitution factors -- i.e., possible production responses -- are considered in the identification of firms that participate in the relevant market and the analysis of entry. Merger Guidelines Section 1.0. See Section V., infra. The Commission and the courts use cross-price elasticity of demand as the primary tool for market definition. See E.I. du Pont, 351 U.S. at 394; Merger Guidelines Section 1.11. Under the Merger Guidelines, the Commission evaluates cross-price elasticities of demand through an iterative process that begins with a candidate market of each product of each merging firm and examines the extent to which the price of each such product is constrained by putative substitutes in demand. The Commission asks whether a hypothetical monopolist of that product could profitably impose a "small but significant and nontransitory" price increase in light of successive "next best substitutes." If alternative products are, in the aggregate, sufficiently attractive, an attempt to raise prices would not prove profitable, such that the candidate market would prove too narrow. The candidate market is expanded to encompass those alternative products to which consumers would switch in response to a significant price increase.*° 8 Respondents argue that this approach is "mistaken" and contrary to long-standing Section 7 precedent. RRB at 46 n.44. Although case law often has treated supply substitutability (or supply elasticity) as part of market definition, the precedent does not compel that approach. See, e.g., Kaiser Aluminum & Chem. Corp. v. FTC, 652 F.2d 1324, 1330 (7th Cir. 1981) (explicitly eschewing application of supply substitution to market definition). The Merger Guidelines’ approach was designed to eliminate analytical confusion that has often arisen when demand and supply elasticities are considered concurrently. See, e.g., Gregory Werden, Market Delineation Under the Merger Guidelines: A Tenth Anniversary Retrospective, 38 ANTITRUST BULL. 517, 524-27 (1992). In fact, when both demand and supply substitution are examined with the appropriate focus on market power, consideration of supply substitution in the identification of relevant market participants should yield market shares that are identical to those determined by including supply substitution as part of market definition. Jd. at 525. The Merger Guidelines’ separation of these steps merely clarifies the focus on identifying mergers that create or enhance market power. "This methodology matches the grouping of buyers who are exposed by their demand patterns to supracompetitive pricing with the set of sellers who are both readily able and likely to produce the pertinent output." Owens-Illinois, slip op. at 12. Cross-price elasticity of demand between the product in question and other products is used as the best indicator of own price elasticity of demand for the product in question, which is the ultimate concern of market definition. "The extent to which a monopolist would increase price is largely a function of own-elasticity of demand for the product . . . Cross-elasticity is relevant only because it is closely related to own-elasticity." Gregory Werden, Market Delineation and the Justice Department's Merger Guidelines, 1983 DUKE L.J. 514, 573 (1983). See ABA Section of Antitrust Law, Monograph No. 12, HORIZONTAL MERGERS: LAW & POLICY 107-08 (1986) (hereinafter, ABA Merger Monograph).

5 Direct evidence of cross-price elasticity of demand is often unavailable; “[h]ence, we may apply reasoned judgment in estimating or inferring the relative magnitude of the elasticities in order to assess the degree of market power." Grand Union, 102 FTC at 812. In addition to engaging in the direct crossprice elasticity analysis described in the Merger Guidelines, the Commission may consider "practical indicia” such as "industry or public recognition of the [market for the product at issue] as a separate economic entity, the product's peculiar characteristics and uses, unique production facilities, distinct customers, distinct prices, sensitivity to price changes, and specialized vendors." Brown Shoe Co. v. United States, 370 U.S. 294, 325 (1962). See United States v. E.l. du Pont de Nemours Co., 351 U.S. R.R. DONNELLEY & SONS CO., ET AL. 155 36 Opinion As a general matter, the Commission considers significant "a price increase of five percent lasting for the foreseeable future." Merger Guidelines Section 1.11.

B. Market Defined by Non-Discriminatory Price Increase Donnelley and Meredith/Burda each produce and sell gravure printing services. Thus, absent price discrimination, the initial test is whether a hypothetical monopolist producer of gravure printing services would likely impose a small but significant and nontransitory price increase above prevailing levels. In order to find that gravure printing is a relevant product market under this initial test, the Commission must conclude that substitution by the marginal consumers of gravure printing services -- those who likely would switch to offset printing (or to some other service) in response to a five percent increase from prevailing prices -- would not be sufficient to defeat the price increase.”

The evidence demonstrates that, at the time of the acquisition, a hypothetical monopolist of gravure printing services could not profitably impose a significant nontransitory price increase for all current gravure printing services. Complaint counsel and Donnelley do not disagree -- and the ALJ found -- that buyers can and do use either offset or gravure printing services for a wide range of printing jobs. ID at 84.4” Customers, demand for printing services varies 377, 395 (1956). As with evidence of cross-price elasticity, evidence of such practical indicia is relevant only to the extent that it is probative of the own-price elasticity of demand for the product at issue. See ABA Merger Monograph, supra note 44, at 108-09 ("The hypothetical monopolist paradigm does not expressly incorporate the submarket indicia identified in Brown Shoe, except to the extent that the indicia may assist the factfinder in inferring whether the hypothetical monopolist could profitably impose a price increase." (notes omitted)). Cf Rothery Storage & Van Co. v. Atlas Van Lines, Inc., 792 F.2d 210, 218-19 (D.C. Cir. 1986), cert. denied 479 U.S. 1033 (1987) (treating submarket indicia as “proxies for cross-elasticities . . in predicting a firm's ability to restrict output and hence to harm consumers").

46 Offset printing is clearly the next best substitute for gravure printing for any relevant job specification. Merger Guidelines Section 1.11 n.9 ("Throughout the Guidelines ‘next best substitute’ refers to the alternative which, if available in unlimited quantities at constant prices, would account for the greatest value of diversion of demand in response to a ‘small but significant and nontransitory' price increase"). More specifically, the next best substitute to gravure in high volume publication printing is heatset offset printing. IDF 9 27-28; RPF f 12-36 (respondents refer only to heatset technology); Hausman Tr. 6422-23 (including only offset as competing with gravure in high volume publication printing). The three other processes used in commercial printing -- coldset offset, letterpress, and flexography -- are not economically reasonable alternatives for high volume publication printing and would likely not significantly constrain the competitive conduct of a hypothetical monopolist of gravure and heatset offset printing services. IDF J 29; CPF 4] 85-86, 115-136. In any event, the exclusion of these processes from the relevant market is not critical to the ultimate determination. 7 See generally RAB at 10-44; CAB at 25-26. Opinion 120 F.T.C.

across a number of significant dimensions,* and gravure and offset printing are not perfectly substitutable for all differentiated print jobs demanded. But there is a significant margin of overlapping use.” At this margin, gravure and offset printing are relatively cross-price elastic. Even complaint counsel's expert testified that a price increase to all gravure customers could not be sustained because too many purchasers would substitute to offset printing. This conclusion is not disputed by complaint counsel and is analytically indisputable.*! Thus, absent an ability to price discriminate, the hypothetical gravure monopolist could not increase prices profitably and the relevant market would be expanded to include offset.” Indeed, by proposing a relevant market consisting of a subset of all gravure print jobs, the complaint in this matter contemplates a relevant market defined by the ability to price discriminate between relatively elastic and inelastic customers of gravure printing services. C. Market Defined by Discriminatory Price Increase The complaint alleges a relevant market for "high volume publication gravure printing,” which is approximated by four-color gravure printing jobs with at least five million copies, at least sixteen pages, and fewer than four four-color versions (or the equivalent in one-color versions).°’ Thus, complaint counsel have attempted to These dimensions include, most significantly: number of versions of each publication, number of copies per version, number of pages per copy, print quality, colors, and, of course, price. 49 RX 497 depicts the competition between gravure and offset essentially as follows: See also CPF-Conclusions of Law at 13 (citing this exhibit as showing "some competition between gravure and offset but also . . . significant areas where the processes do not compete” and as standing for the proposition that there is "limited direct competition between the processes"). The testimony of all of complaint counsel's witnesses corroborates this view. See, e.g., Nytko Tr. 1525 (“offset and gravure compete in certain areas, and there are certain areas that your expectation is that offset would be the predominant process [and] other areas where you would expect gravure to be the predominant process").

°° Hilke Tr. 3070-71.

3 See CAB at A-1 (citing Dr. Hilke for the proposition that "a price increase across all gravure printing (broader than high volume gravure printing only) may not be profitable"). See Hilke Tr. 3174-75.

3 CAB at 10; CPF { 657; Hilke Tr. 2997-98, 3419-20, 6149- 50. See CX 1167; CX 1351-B. R.R. DONNELLEY & SONS CO., ET AL. 157 36 Opinion prove that a hypothetical gravure printing monopolist could profitably impose a discriminatory price increase on customers whose printing demand fits these parameters. Complaint counsel further describe the "core" of this proposed market as gravure print jobs with at least ten million copies, more than thirty-two pages, and fewer than four four-color versions (or the equivalent in one-color versions), but does not plead this as an alternative market.™ The Commission and the courts recognize that if a seller or group of sellers can earn substantially different returns from different classes of customers based on their relative demand elasticities for the products of the sellers, the relatively inelastic class of customers may constitute a relevant market. Under the standards set forth in Section 1.12 of the Merger Guidelines, the Commission will define a relevant market for a group of buyers for which a hypothetical monopolist would separately impose a "small but significant and nontransitory" increase in price.© If a hypothetical monopolist can identify and price differently to those buyers (targeted buyers) who would not defeat the targeted price increase by substituting to other products in response to a "small but significant and nontransitory" price increase for the relevant product, and if other buyers would not purchase the relevant product and resell to targeted buyers, then a hypothetical monopolist would profitably impose a discriminatory price increase on sales to targeted buyers. Merger Guidelines Section 1. 12.°”

54 See CAB at 10; CPF § 657; Hilke Tr. 2997-98, 3097; CX 1351-B. Complaint counsel define the "core" of the market "as the area in which a-hypothetical monopolist has the greatest possibility of success." CAB at 10, citing Hilke Tr. 2997-98, 3417. Complaint counsel do not propose the "core" as a relevant market alternative to the relevant market proposed in the complaint. OA Tr. at 41. 38 See, ¢.g., Owens-Illinois, Inc., 115 FTC 179 (1992) [FTC Dkt. No. 9212 (Feb. 26, 1992), slip op.}; MidconCorp., 112 FTC 93 (1989). See also United States v. Grinnell Corp., 384 U.S. 563 (1966) (in monopolization case under Section 2 of the Sherman Act, 15 U.S.C. 2, relevant price discrimination product market for class of inelastic buyers); United States v. Rockford Memorial Corp., 898 F.2d 1278 (7th Cir.), cert. denied, 498 U.S. 920 (1990) (affirming injunction against merger under Section 7 of the Clayton Act). In Rockford Memorial, Judge Posner explained that each category of customers identified with a specific hospital service (i.e., a specific medical indication) could represent a separate relevant product market if a hypothetical monopolist could discriminate in price (or other terms of competition) between such categories based on identified demand elasticities. 5 “This is true regardless of whether a general increase in price would cause such significant substitution that the price increase would not be profitable," Merger Guidelines Section 1.12, as is the case here.

7 Price discrimination consists of "differences in price not based on differences in cost." Midcon, 112 FTC at 168-69. "Price discrimination consists of obtaining different economic profits from different customers for similar products." Hilke Tr. 3498-99; Hausman Tr. 5474-75. See also F.M. Scherer & David Ross, INDUSTRIAL MARKET STRUCTURE & ECONOMIC PERFORMANCE 489 (3d ed. 1990) ("Price discrimination is the sale of different units of a good or service at price differentials not directly corresponding to differences in supply cost"). "Profitability" and "cost" are defined in economic rather than accounting terms. See Merger Guidelines Section 0.1 ("References to profitability of certain actions focus on economic profits rather than accounting profits. Economic profits may be defined as the excess of revenues over costs where costs include the opportunity cost of invested capital."). Opinion 120 F.T.C.

In this matter, the Commission may find that a profitable discriminatory price increase is possible, and therefore sufficient to define a relevant market, if three conditions are satisfied: (1) the hypothetical monopolist can identify gravure customers with sufficiently inelastic demand for gravure printing (i.e., those who will not switch to offset printing in response to a five percent price increase); (2) the hypothetical monopolist can selectively and profitably increase prices to those gravure customers;** and (3) arbitrage of gravure printing (resale by favored elastic customers to targeted inelastic customers) would not be sufficient to undermine the price increase.”

There appears to be no dispute regarding arbitrage: it is generally not feasible in markets for individualized services, and publication printing does not appear to be an exception.®! Assuming that a hypothetical gravure monopolist could impose a discriminatory price increase on a class of buyers whose print jobs fit the proposed market parameters, those buyers would not likely make the price increase unprofitable by purchasing required printing services from elastic buyers who obtained the service at a lower price. Thus, the focus of the remaining product market inquiry is whether a hypothetical gravure monopolist could accurately identify, and profitably increase price to, a class of current printing customers with inelastic demand for gravure printing.

We do not require proof of actual past or present price discrimination to use the possibility of price discrimination to define a market in a Section 7 case. Section 7 addresses likely future effects on competition, so proof of likely future discrimination could support The issue is whether a sufficient number of customers with high volume requirements would not switch to offset in response to a "small but significant and nontransitory" price increase for high volume printing jobs such that the price increase would prove profitable. Complaint counsel need not prove that all current gravure customers have inelastic demand for gravure printing high volume publication printing jobs, only that enough customers are sufficiently inelastic so as to make the price increase profitable.

5 CAB at 12-13; Hilke Tr. 3011-3012.

60 Hilke Tr. 5876-77; Hausman Tr. 5225-26. See Merger Guidelines Section 1.22 n.12 (“arbitrage is inherently impossible for many services"); accord Robert Pitofsky, New Definitions of the Relevant Market and the Assault on Antitrust, 90 COLUM. L. REV. 1805, 1848 (1990); Scherer & Ross, supra note 57, at 489.

©! CPF 4 736-40; Hilke Tr. 3069-70.

62 Hilke Tr. 3070 ("[A]rbitrage is a non-starter in [high-volume publication gravure printing]. I simply don't see any reason why J.C. Penney would find any use for old Best catalogs in trying to sell their product. Once the thing is printed, arbitrage is really not something that's in the realm of interest."). R.R. DONNELLEY & SONS CO., ET AL. 159 36 Opinion the necessary market definition showing.” "Thus the possibility of price discrimination might in appropriate circumstances be enough to justify concern about anticompetitive effects. But, possibilities can be a weak foundation for a prediction of ‘likely,’ ‘substantial’ competitive effects." Midcon, 112 FTC at 169-70. The Commission must be mindful of the analytical hazards of defining markets by reference to possible price discrimination. It is an economic truism that buyers do not have homogeneous preferences or demand elasticities for a given product within a relevant market, and there may often be some conceptual means of identifying classes of customers that appear to have inelastic demand for the product. The potential for this approach to swallow up the market definition principles established by the federal courts and the Commission is substantial. As the Commission warned in Midcon Corp.: "In considering possible markets under [a price discrimination] theory, there is a danger of implicitly assuming the conclusion." 112 FTC at 168. That risk requires a particular rigor in examining the conceptual basis for distinguishing the allegedly inelastic customers and the factual basis for the prediction that price discrimination with respect to those customers is likely. The analytical hazards of defining a relevant market by the possibility of price discrimination are manifest in complaint counsel's theory and the ALJ's decision. Although complaint counsel have described, in theory, a methodology for identifying a category of printing customers whose demand for gravure printing should be relatively inelastic, complaint counsel have not carried the burden of proving that the methodology allows an accurate identification of 3 See Owens-Illinois, slip op. at 39 (Commissioner Azcuenaga, concurring). 4 See Owens-Illinois, slip op. at 36 n.41 (“{T]he absence of price discrimination . . . is not determinative of what is likely to occur in the future. Its presence, however, might have conveyed a warning of appreciable danger . . .”) 6 See Pitofsky, supra note 60, at 1816 ("There will almost always be classes of customers with strong preferences for [differentiated] products, but to reason from the existence of such classes to a conclusion that each is entitled to the ‘protection’ of a separate narrow market definition grossly overstates the market power of the sellers."). 66 The Commission will recognize the possibility of price discrimination as a means of defining a relevant market if there is a conceptually sound methodology, supported by the record, by which a hypothetical monopolist can identify the alleged inelastic customers. See Owens-Illinois, slip op. at 34; Midcon, 112 FTC at 168-69.

Opinion 120 F.T.C.

inelastic end uses and, thus, that a price increase within the identified category likely would be profitable. First, the identification of the parameters of the proposed market reflects a significant analytical error. In particular, the number of versions is not determined exogenously but is instead an endogenous competitive variable; increasing the number of printed versions ("versioning") is a means by which printing customers can and do substitute from gravure to offset. Second, even assuming that the described parameters are a reasonable means of distinguishing elastic from inelastic customers, there is substantial historical and existing use of offset printing within the proposed market, even at existing prices and even within the "core" of the market. Third, given this existing substitution at current (presumptively competitive) prices, a significant and nontransitory increase in gravure prices for high volume printing would likely expand use of offset printing. Eliminating (or discounting) the versions parameter significantly increases the amount of observable substitution between offset and gravure. The dynamic analysis of the Merger Guidelines clarifies that substitution at the margin would likely make unprofitable such a supracompetitive price increase. For these and other reasons described below, we find that "high volume publication gravure printing" does not constitute a relevant market for purposes of evaluating the competitive effects of Donnelley's acquisition of Meredith/Burda.

D. Identification of Inelastic Uses 1. "Breakeven Analysis"

Complaint counsel have attempted to describe a methodology for identifying a category of commercial printing customers whose demand for gravure printing should be relatively cross-inelastic with offset: the "breakeven" analysis of gravure and offset printing costs. complaint counsel argue that above a certain volume level, offset printing becomes an increasingly less viable alternative to gravure as the total number of copies increases. If a hypothetical gravure monopolist can approximate with some degree of confidence the The ability to version is an important variable of competition between gravure and offset for high volume printing. Increasing the number of versions is a means of reaching customers more specifically and is an attribute for which printing customers are willing to pay. Thus, a current gravure customer that chooses to print a greater number of versions and therefore shifts to offset has substituted offset for gravure.

R.R. DONNELLEY & SONS CO., ET AL. 161 36 Opinion volume level at which offset printing becomes an unprofitable substitute for gravure, it likely would increase prices to all customers for jobs above that volume.® The breakeven analysis is based on a fundamental difference between the cost structures of the gravure and offset processes.” The record shows that the cost of a print job can be divided into two basic economic categories: (1) fixed costs, which do not depend upon the number of copies printed, and (2) variable costs, which vary with the number of copies printed.” The cost structure of the gravure process is characterized by relatively high fixed costs and low variable costs for each print run.” Relative to the gravure process, the cost structure of the offset process is characterized by lower fixed costs and higher variable costs for each print run.” The primary difference in fixed costs between the processes is that, for a typical print job, the costs of engraving gravure cylinders and installing them to the gravure press are higher than the costs of producing offset printing plates and installing them into an offset press.” The primary differences in variable costs between the two processes are that gravure presses generally have a higher rate of throughput than offset presses” and that, for a given high volume print job, offset plates are less durable Thus, the relevant "identification" in this case concerns the job, not necessarily the customer. If gravure printers know that offset printing is not a reasonable economic alternative to gravure printing for high volume publication jobs, it need not distinguish among customers. The profit maximizing strategy for the hypothetical monopolist is to approximate perfect price discrimination: setting a separate profit maximizing price for each product and each customer for which it can identify demand elasticity. Hal R. Varian, Price Discrimination, in 1 HANDBOOK OF INDUSTRIAL ORGANIZATION 600 (Richard Schmalensee & Robert D. Willig, eds. 1989). 69 Although there is considerable dispute regarding the volume at which the breakeven point occurs, as well as whether trends in technology have or will significantly alter the cost structures that form the basis of the breakeven analysis, the general validity of the breakeven analysis is supported by internal costs studies of the merging firms and by testimony of both gravure and offset printers. See IDF GF 212-15.

70 The fixed costs of a print job include the preliminary costs of original artwork and color separations, the costs of making the original set of offset plates or gravure cylinders (including proofs), the costs of the initial "makeready" (including ink and paper used in the makeready), and other costs that do not vary for a print run. Hodgson Tr. 186-96; CX 1164-F-G; CX 1165-G-H. The variable costs of a print job include the operating cost of the press and the materials used in printing the product, such as ink and paper. Hodgson Tr. 204-20; CX 1164-F-G; CX 1165-G-H. 1 Nytko Tr. 1484, 1514, 1525; Kelly Tr. 1716; Angstrom Tr. 2614; H. Sullivan Tr. 2789; J. Sullivan Tr. 4871-72. See Hausman Tr. 6308, 6417. n Kelly Tr. 1716; Glazer Tr. 2155; Angstrom Tr. 2614; J. Sullivan Tr. 4871-72. See Hilke Tr. 3090; Hausman Tr. 6308, 6417.

73 IDF Y 165. See, e.g., Hodgson Tr. 244-45; Haight Tr. 1328; Engdahl Tr. 2557-58. See also Hilke Tr. 3514-16.

™ IDF Ff 133-164. See, e.g., Scirocco Tr. 1024-26; Wells Tr. 1916-19; Kaminsky Tr. 2001-02, 2037. Slower input of offset implies, among other things, longer press time per printed page. Opinion 120 F.T.C.

and require more frequent replacement than gravure cylinders.” Thus, average unit costs for a particular print job when printed on gravure decrease with volume more rapidly than when printed on offset.

The breakeven point between gravure and offset is defined as the number of copies at which a print job with given page specifications (number and size) is equally costly to print using either the gravure or the offset process. IDF J 212.”° In other words, the breakeven point is the intersection of the differing average cost curves, as depicted in the following figure.”

\ \ wneen --8a Within the framework of the Merger Guidelines, the relevant point for evaluating the usefulness of the breakeven analysis is the volume that would yield a five percent cost differential between gravure and offset.”* Of course, printing customers do not make purchases based on the relative costs of their suppliers; their purchases are based on relative prices (and other attributes) of the printing services offered. Assuming that both gravure and offset firms have operated competitively (whether or not they are all in one > IDF 165-69. The evidence suggests that, for technology in current use, offset plates for a print run must be replaced at approximately 1-2 million impressions while gravure cylinders for a print run require maintenance ("dechroming and rechroming") at approximately 5-10 million impressions. The ALJ noted other variable costs that are higher for offset than for gravure, including gravure paper savings resulting from variable cut-off capability (IDF ] 172); the ability of gravure presses to stitch and trim on line (IDF ] 173-74); less paper waste in the gravure process (IDF J 175-76); and gravure's ability to produce better results than offset on cheaper, lighter weight, uncoated paper (IDF 180). For purposes of analysis, we can assume that such cost-differences exist. These other factors are relevant to the explanation of any observed cost differences, but specific quantification is not necessary to the evaluation of the breakeven analysis. The ALJ did not attempt such quantification. 76 The breakeven analysis appears most reliably defined in terms of run length. The number of versions are then defined as the number of runs. The cost structure of the gravure printing process suggests that run length determines the relative cost advantage vis-a-vis offset. Hodgson Tr. 254-55, 373-74; Hilke Tr. 3093, 3435-36.

77 See Hilke Tr. 3152 (describing this relationship). 78 Wilke Tr. 3153.

R.R. DONNELLEY & SONS CO., ET AL. 163 36 Opinion antitrust market), relative average prices for gravure and offset printing should maintain a relationship similar to that depicted in Figure 1.”

The record does not indicate clearly the volume at which the breakeven occurs or the volume at which the "breakeven plus five percent" occurs. Although there is some evidence in the record that the breakeven point may occur at a lower point,” the ALJ found that the intersection occurred at the lower bound of complaint counsel's proposed relevant market: approximately five million copies for publications with sixteen pages and fewer than four versions. ID at 82-83. But complaint counsel's economic expert testified -- based on internal studies of relative costs by Donnelley -- that the theoretical gravure-offset breakeven for a sixteen-page catalog may be greater than six million copies.*’ To the extent that these cost-based breakeven estimates do not reflect a five percent differential,” the relevant volume for purposes of market definition is higher. What complaint counsel call the "core" of this market -- more than ten million copies, more than thirty-two pages, and low versioning -may be viewed as an attempt to approximate the relevant "breakeven plus five percent" point.®* The record is not "strongly suggestive" that gravure producers are currently engaged in coordinated interaction or that Donnelley is otherwise behaving non-competitively and, therefore, does not provide a basis for postulating a price increase lower than five percent, or for using a threshold other than prevailing prices from which the price increase is postulated. See Merger Guidelines § 1.11. In this case, complaint counsel have argued that the acquisition has resulted in dominant firm conduct by Donnelley and have presented evidence that capacity utilization of gravure has been high at times when there has been substantial excess capacity on offset. See IDF ¥ 280-284. Other evidence suggests that prices for both gravure and offset continue to fall in this industry and that it is a "buyer's market." See IDF 49 358-59. Moreover, the exact specification of the postulated price increase is not critice| to our conclusions.

80 see IDF 9 227-35.

8 Hilke Tr. 3146-51; CX-1164-B, D; CX-1433-B. The economically relevant breakeven analysis must compare the costs of the most efficient technology of each process. It is not clear whether the isolated cost studies upon which CX-1164 is based reflect a state-of-the-art comparison. Nor is it clear that industry witnesses were making the economically relevant comparison. The evidence taken as a whole, however, provides a reasonable approximation of the costs of the marginal technologies. 82 Dr. Hilke's estimate explicitly excludes the five percent addition. The ALJ's conclusion appears not to incorporate the five percent differential: The record supports complaint counsel's claim that, for low version, long run, high page count publications, gravure is less costly than offset. The breakeven point [IDF { 212} at which this usually occurs is in publications with less than 4 four-color versions, more than 32 pages [IDF ] 222] and a run length in excess of 5 million copies [IDF J 216] .... In this case, there was explicit testimony that for run lengths in excess of 5 or 10 million copies, gravure is more economical than offset (see, e.g., F 234). .

ID at 82-83.

83 An additional complication in the breakeven analysis arises from considerations of print quality. The ALJ's conclusions reflect the conflicting testimony regarding differences in quality between gravure and offset:

Opinion 120 F.T.C.

2. Reservations About the Breakeven Analysis These varying estimates suggest that "breakeven" analysis may be a poor means of distinguishing, with a reasonable degree of confidence, jobs for which gravure is relatively cross-price elastic with offset from those for which it is not. Indeed, one of complaint counsel's witnesses testified that even a "ballpark estimate" would be a "gross generalization."™ In particular, he testified that "it's very dangerous to generalize about . . . the specific crossover point at which gravure is more efficient than offset" in large part because changes in technology shift the crossover point.® The continuing trend of increased productivity and efficiency of offset relative to gravure magnifies our doubts about the value of the breakeven analysis as a means of identifying inelastic gravure uses. Significant trends in process technology between the two putative substitutes should also be considered in the forwardlooking analysis required by Section 7. Over the past decade, offset technology has made significant gains in both quality and productivity relative to gravure technology.® As a result of these relative improvements in offset technology, the margin of competition between gravure and offset has increased in recent years. IDF JJ 194-201.*’ In general, Some industry members believe that, for their purposes, gravure and offset offer comparable quality [IDF J 186}; some high quality high volume magazines use offset and gravure for different parts of their publication [IDF { 185]; and, industry members testifying in this proceeding could not distinguish between the two processes with the naked eye [IDF { 187]. Nevertheless, the firm belief of many print buyers that gravure offers higher quality than offset is a real constraint on their choice of printing processes.

ID at 82. Complaint counsel argue that, whatever similarities may be shown in an isolated comparison, print customers generally perceive a reduction in the quality of printed copies as a print run approaches its conclusion, and that the reduction begins earlier in the print run for offset than for gravure. CAB at 22-24. It appears, however, that the breakeven estimates already account for this consideration: differences in quality are incorporated into the relative cost structures through the inclusion of the costs of changing offset plates and the costs of “dechroming and rechroming" gravure. In any event, although the ALJ appears to have felt constrained to hold "quality considerations" aside when discussing the economics of breakeven, IDF J 214, his conclusion regarding the actual estimate of the relevant breakeven did not do so explicitly. See ID at 83. 84 Niytko Tr. 1485.

85 ig.

86 RX-161-A, RX-163, RX-164 (pre-acquisition internal Donnelley memoranda describing program to improve gravure productivity to compensate for relative improvements in offset); Doty Tr. 4582-86, Weir Tr. 3778-88; RX-150.

87 See Angstrom Tr. 2609 (as offset press technology has improved, "the area of competition between the two processes . . . has definitely gotten broader and . . . there is more work crossing over between the two processes"); CX-1142-Z-43 (industry publication stating that "the latest innovations in web offset . . . [have made] web offset competitive with gravure in long-run printing"). See also RX- 153-A (gravure industry group document stating that "with the ever advancing technology in offset, the competitive advantage of gravure is becoming questionable” and that improvements in gravure technology "will be mandatory to protect our market share within the commercial printing community R.R. DONNELLEY & SONS CO., ET AL. 165 36 Opinion however, these improvements should already be reflected in the measures of substitution described above and in the estimate of the breakeven point for gravure and offset economics. Donnelley claims that the recent introduction of the latest generation offset presses will accelerate this trend and may, in fact, eliminate any gravure cost advantage in high volume printing.** The ALJ found that these newer offset presses "come closer to the performance of gravure presses." IDF J 153. These presses are designed for rated speeds and web widths that suggest substantial unit cost improvements over prior generations of offset presses in high volume publication printing.” In particular, the Heidelberg/Harris M-3000 offset press has a rated speed of 3000 feet per minute, which approaches the highest rated speed for any existing gravure press.” Some evidence suggests that the M-3000 offset press compares favorably with late generation gravure presses in terms of unit cost and productivity for high volume printing jobs.”’ As a result, the M- 3000 and other new generation offset presses may further reduce whatever cost differential exists between gravure and offset, even for print jobs well within the volume and page parameters of the proposed market.”

ys. offset/flexo"; (emphasis added); CX-1171-B (reflecting view that offset's share of an overall printing "market" is increasing while gravure's share is "steady"). These latter two documents, and many others in the record, refer to a "market" that includes at least gravure and offset. They therefore tend to counter complaint counsel's attempt to bolster their alleged market definition by referring to internal Donnelley documents that mention a "gravure market." The ALJ relied on these documents as evidence of industry recognition of a separate "gravure market." IDF 87 JJ 274-79. These documents generally do not refer to a separate high volume publication gravure market as alleged in the complaint.

58 RAB at 36-40.

89 Rockwell Graphics recently introduced an offset press named the G-25W, which has a rated speed of 2500 feet per minute. Nytko Tr. 1547; Weir Tr. 3771-73. Mitsubishi and Man-Roland have also introduced offset presses with rated speeds of greater than 2000 feet per minute. Buchanan Tr. 2055-76; Weir Tr. 3771; RPF { 29.

90 RPF { 27-28; CPF ¥ 490 (CX 120-Z-30-31, 38). io Dr. Hausman performed an econometric analysis comparing the estimated costs of printing the largest catalog described in the record (J.C. Penney's 1000-1500 page catalog) using an M-3000 with the costs of printing the same catalog using a new three-meter gravure press. RX-666. His analysis purports to show that the state-of-the-art offset press maintains a unit cost advantage of approximately 2-4%, depending on varying assumptions. Hausman Tr. 5320-21, 6384-85, 6388-89. Compare CX- 1432; CPF JJ 954-967 (Dr. Hilke's analysis estimating a 11-22% cost advantage for a state-of-the-art gravure press over the M-3000), with Hausman Tr. 6382-89 (Dr. Hausman's analysis of Dr. Hilke's methodology; estimating a 1-3% cost advantage for the M-3000 after replacing some of Dr. Hilke's assumptions with some of his own). See also RX 476; Melton Tr. 2426 (comparison of the M-3000 with a modem gravure press, performed by one of complaint counsel's witnesses, showing M-3000 to be 26% more productive).

2 See, e.g., Angstrom Tr. 2608-10, 2622-23 (M-3000 and Man-Roland's "Lithoman V" offset press considered "cross-over presses" because of their ability to do print jobs traditionally thought of as "gravure" work, including high-volume publications). Opinion 120 F.T.C.

Complaint counsel admit that these and other technological areas of competition between advances may "create a gravure and offset." CPF { 53.9.% In fact, evidence presented by complaint counsel shows that the M-3000 was designed to be cost-competitive with gravure in run lengths of up to ten million copies” and that, when commercially diffused, it would increase substitution between gravure and offset in high volume publication printing.” In essence, the commercial success of the M-3000 (and other new generation offset presses) would shift. the unit cost breakeven point for gravure and offset to a higher run length and larger volume. On the other hand, there is no evidence in the record that the M- 3000 offset presses have achieved commercial success in high volume publication printing services, as defined by complaint counsel.” Although some evidence suggests that other new generation offset presses are now commercially viable,” there is no evidence in the record that these presses have significantly altered the nature of competition between gravure and offset or that they have produced dramatic increases in the use of offset in high volume publication printing.” Thus, there is no assurance that these presses will have a significant impact on competition in high volume publication printing. See IDF J 202; ID at 84. Nevertheless, the continuing trend of increased productivity and efficiency of offset relative to gravure reinforces our skepticism regarding the value of the breakeven analysis as a means of identifying inelastic gravure uses. A small but significant and nontransitory increase in the price of gravure for high volume publication printing certainly would not %3 Citing Angstrom Tr. 2609; CX 1272-G.

94 See IDF { 200; CX-1272 (graphically depicting the "potential market for the M-3000" and showing this offset press as cost-competitive with gravure for volumes as high as ten million copies). Sullivan Tr. at 2806-08.

95 See Hilke Tr. 6069-70 (‘‘[T}]he M-3000 might be effective in enlarging that area of competition between the two processes, but that doesn't eliminate the portion of gravure that is basically beyond the reach of offset under most circumstances” (citing CX 1272-G; emphasis added); CPF J 541. See also Hodgson Tr. 290-91; Coleman Tr. 1802-03; Angstrom Tr. 2609-10; Sullivan Tr. 2807-08 (testimony to the effect that the commercial success of the M-3000 would simply shift the breakeven point to a higher volume).

96 See CAB at 42 (not one page of high volume printing has been contracted to be printed on an M-3000"; emphasis in original). Complaint counsel attribute the lack of success to two phenomena: (i) performance problems in trials, which suggest that it will not be commercially viable in the near future, and (ii) the revolutionary (as opposed to evolutionary) nature of the technology, which suggests that printers will be reluctant to incur the risks of an unproven technology. See generally CPF Jf 500-535. 97 IDF Gg 194-95, 198. See RPF 4] 29-32.

98 1D at 84; IDF J 198-200, 202.

R.R. DONNELLEY & SONS CO., ET AL. 167 36 Opinion reverse this trend, and as discussed above, would shift the breakeven point (wherever it currently lies) to a significantly higher volume. 3. "Versioning" as a Competitive Variable The ALJ's identification of the parameters of the proposed market reflects a significant analytical error that results in a substantial understatement of competition between gravure and offset printing. Specifically, the ALJ's market definition treats the number of versions of a given print job as somehow pre-determined for each customer. In fact, it appears to be a variable of competition between gravure and offset in high volume publication printing. Increasing the number of versions is a means for publishers to target customers more specifically, IDF { 361, and is an attribute for which printing customers are willing to pay.” As the breakeven analysis makes clear, the costs of versioning are greater in gravure-printing than in offset printing. IDF JJ 38, 224, 225. The record clearly shows that many customers, including some of the largest, have shifted their purchases of high volume publication printing services from gravure to offset. IDF 9] 122-32. Recognizing these facts, offset printers attempt to influence customers to increase the number of versions of high volume publications." Complaint counsel argue that the Commission should ignore the competitive relevance of this switching by, in essence, treating these customers as having exited the relevant market. CAB at 10 n.9.'°' The ALJ concurred in this view. ID at 85.'”

Complaint counsel's analysis presumes implicitly that the number of versions of a particular print job is predetermined. In fact, however, the number of versions ordered for a given print job appears “Target marketing” has become increasingly common as retailers and magazine publishers attempt to tailor their publications to the specific preferences of customers or to competitive circumstances. Wyker Tr. 908; Engdahl Tr. 2571; Van Horne Tr. 4637; Doty Tr. 4576-77; Higham Tr. 2321; CX-483-X.

100 RX-19-A; Van Home Tr. 4636-38; Pope Tr. 2843-44, 101 See CPF ¥ 1111 ("In instances where switching between processes occurs, it is usually as a result of a change in the requirements of the printing program. (Bentele Tr. 1441-42; Nytko Tr. 1522- 23)").

102 "Donnelley relies too heavily on its analysis of gravure print buyers who have switched to offset (IDF 9] 122-32), for it ignores . . . the reason for some of the switches -- increased versioning." See IDF ¥ 106 ("During 1993, Wal-Mart began shifting its predominantly gravure printed program to offset. Each issue has a run length ranging from 60 to 70 million . . . each with many localized versions"); IDF J 108 (K-Mart shifting weekly 24-36 page insert from gravure to offsert after increasing versions); IDF ¥ 122-125.

Opinion 120 F.T.C.

to be selected by many buyers after comparing the benefits and costs -- that is, the net profit -- associated with varying the attributes of a particular print job. As the relative prices of gravure and offset printing change, so too will the profits associated with each possible variant of a particular job (e.g., low-version/high-run length versus high version/low-run length). The ability of buyers profitably to adjust these attributes provides a means by which printing customers can and do substitute from gravure to offset.” That substitution of this sort occurs is beyond dispute. Print customers with very large print volumes and page counts have substituted from gravure to offset in response to changes in perceived profit considerations related to versioning. Wal-Mart, for example, recently shifted to offset a large portion of its print program (with print issues of 60 to 70 million copies) in order to version more extensively and to target customers geographically. IDF {J 106, 123- 24.!% K-Mart, one of the world's largest purchasers of high volume publication printing services, also shifted the printing of its weekly national inserts from gravure to offset after deciding to increase versions. IDF] 108, 125.'° The print volumes and page counts in To illustrate, consider a retail merchant that intends to mail a sales catalog nationwide. The characteristics of a catalog -- the number of pages, the page size, the type of paper, the number of versions -- seldom will be fixed in advance, but are instead competitive choices made by the retailer. Focusing on versioning (holding other characteristics constant), the retailer will weigh the potential benefits of greater versioning (i.e., greater sales revenue from a more accurately targeted catalog) against potential costs (i.e., higher unit costs from shorter run lengths). Given existing prices for offset and gravure, the retailer will select the set of characteristics that offers the greatest expected profit. Clearly, this profit calculation will change as the relative prices of offset and gravure change. At one set of relative prices, the retailer may find it most profitable to opt for a single version catalog with a high run length, which might dictate the use of gravure. Were gravure prices to increase, however, a multipleversion publication might yield higher profits, which would dictate the use of offset. Analogously, consider an airline that offers passenger service on a competitive route and is faced with the choice of leasing either one jet that seats 500 passengers or two jets that seat 250 each. Assuming the price of a lease for the larger jet is significantly lower than the price of the lease for the two smaller jets (i.e., a price differential greater than 5%), the airline may choose the larger jet in order to minimize costs and, accordingly, to set a lower passenger ticket price. Alternatively, the airline could choose to lease the two smaller jets at a higher lease price in order to provide additional service to passengers (more departure times) albeit at a higher ticket price. The number of jets selected to provide the service is a competitive variable that is determined by the customer to maximize its profits; it is not an invariable, pre-determined market condition. Thus, it would be incorrect to interpret the airline's choice of the two-jet lease as a departure from the relevant market. The correct interpretation is that the airline selected between two substitutes.

104 ¢ ee Baron Tr. 2261, 2275-76, 2307-08. 105 Specifically, the AL} found: "Today, K-Mart finds the gravure and offset processes to be interchangeable, receives bids on both processes, and uses both processes for its national inserts. .. .One of the reasons K-Mart has moved more to offset is increased versioning." IDF J 125. See Habeck Tr. 4158-59, 4163, 4234-35; RX-652.

R.R. DONNELLEY & SONS CO., ET AL. 169 36 Opinion each of these examples are not only within complaint counsel's proposed relevant market, they are well within the "core." Of course, not all volume of a print customer need be shifted to offset to undermine a gravure price increase.'”’ As the ALJ found, "some buyers, such as National Geographic, use both processes for their publications because of their unique contribution to specific needs, such as, in the case of offset, its lower cost when a portion of the publication requires a high number of versions." ID at 83; see IDF { 121.'°8 Customers who currently use both processes would likely shift additional volume to offset as a result of a relative increase in the price of gravure.”

This evidence shows that the ability to version is an important variable of competition between gravure and offset for high volume printing. In reaching this conclusion, we do not find that all high volume gravure print jobs are currently at this margin of substitution. We need only find that a sufficient number of jobs are at this margin such that a high volume gravure price increase likely would not be profitable. A hypothetical gravure monopolist could profitably raise price only to those customers who could not avail themselves of versioning and other alternatives. As complaint counsel recognize, "the gravure monopolist must take versioning into consideration when targeting price increases. It is inappropriate to ignore versioning when identifying printing likely to be targeted by the hypothetical gravure monopolist." CAB at 9-10 n.8 (emphasis in original). Because a gravure monopolist could not determine ex ante whether current purchasers of gravure services will choose to print a greater number of versions in response to a small but significant and The record reflects other specific examples of customers substituting offset for gravure in order to obtain more versions. See, e.g., RX-355; Gorden Tr. 3954-57 (Levitz Furniture). The record generally reflects that versioning is increasing. See, e.g., Moeller Tr. 4009-10. See, e.g., IDF J 121 (regarding National Geographic: for each 9.8 million copy run length, more than four versions of up to 25 pages are printed offset and fewer than four versions of up to 88 pages are printed gravure); IDF J 111 (regarding Modern Maturity: for each 22. 4 million copy run length, 4-20 versions of up to 12 pages are printed offset and one or two versions of up to 88 pages are printed gravure). Customers who use both processes would likely shift additional volume to offset as a result of a relative increase in the price of gravure. 108 Each National Geographic issue has a volume of nearly ten million copies and the offset portion is 20-25 pages per issue. IDF J 121. Again, this print job is well within complaint counsel's proposed market and approaches the proposed "core" (absent the versioning restriction). See Allen Tr. 1598-99 (AARP plans increased versioning so that certain pages formerly printed gravure will be printed offset, even without a change in relative prices). 109 Likewise, a publication printing customer faced with a relative increase in the price of offset could alter its print program to reduce the number of versions in order to obtain a better quality-adjusted gravure price. Customers with very high volume print requirements, but who are now using offset for highly-versioned programs, cannot be said to be insulated from gravure competition. Opinion 120 F.T.C, nontransitory price increase, it would likely not be able profitably to target such price increases.

By ignoring this method of substituting between offset and gravure, the data and analysis presented by complaint counsel mask the amount of actual substitution that currently occurs between offset and gravure.''° Eliminating the versioning parameter significantly increases the amount of observable substitution between offset and ravure.''’ Given this existing substitution, the dynamic analysis of g g y y the Merger Guidelines clarifies that substitution at the margin would likely make a high volume gravure price increase unprofitable. E. Current and Likely Purchasing Patterns Assuming arguendo that versioning is an appropriate market parameter, the evidence suggests that the boundaries at five million copies and sixteen pages have been drawn too low: at both preacquisition and post-acquisition prices, a substantial portion of print jobs above this line is done using offset printing.''? Complaint counsel estimate that, in 1990, offset accounted for 12.5% of printing in their proposed relevant market: "low versioned jobs over 16 pages" for print runs of more than five million copies.''? Moreover, when the versioning parameter is eliminated to reflect more accurately actual substitution (and substitution possibilities), the proposed market boundaries become more porous still. Complaint counsel To a lesser extent, the analysis in this section with respect to versioning also applies to these other listed variables. For example, customers faced with a supracompetitive price increase for gravure could conceivably adjust the number of pages in order to obtain an economical offset bid. However, the extent to which customers likely would make changes in these other variables in response to a gravure price increase is unclear. By comparison, the demand for a greater number of versions is clearly derived from demand for targeted marketing. Thus, unlike changes in other listed variables, an increase in the number of versions has an identified benefit to the print customer that can be considered along with any increase in unit costs assumed by switching to offset. Because the versioning parameter is selected somewhat arbitrarily, its elimination also eliminates some anomalous results. See RPF 64-73; RX-28 (showing, for example, one job with a total volume of 35 million copies that is excluded from complaint counsel's proposed market because it had several versions, despite the fact that one of the four-color versions had a mun length of 18 million, which would place that version alone in the "core"); Hausman Tr. 5371-81. The finding is also consistent with the testimony of complaint counsel's economic expert that, for a 16-page catalog, the theoretical gravure-offset "breakeven point" may be greater than six million copies. Hilke Tr. 3146-51; CX-1164-D; CX-1433.B. "3 CX.1167-C-1. See CX-1167-C (showing offset accounted for 11.4% of "low versioned jobs over 32 pages" for print runs of more than five million copies). Dr. Hilke prepared CX-1167 based on subpoena responses from 29 printers with gravure and offset printing capacity. He compiled information regarding run length, page counts, and versions for high volume work that the printers perform. Based on this information, he tabulated proportions of gravure and offset for different permutations of these factors. CX 1167-F-G.

R.R. DONNELLEY & SONS CO., ET AL. 171 36 Opinion estimate that, in 1990, offset accounted for 24.1% of print jobs with more than sixteen pages and print runs of more than five million copies.''4 Even in the "core" of complaint counsel's proposed market -- print jobs with more than thirty-two pages and print runs of more than ten million copies -- offset accounted for 13.4%.'!° For purposes of analysis, we will consider these estimates to be representative.''® This substantial existing use of offset printing at current prices would likely increase in response to an attempt by gravure printers to impose a supracompetitive price increase.''’ Historical, or existing, purchasing patterns may indicate demand elasticities at competitive prices, but the focus of market definition under Section 7 is demand elasticities at a future, supracompetitive price as a result of the acquisition.''® The principal object of Section 7 market definition analysis is to determine susceptibility to an exercise of market power.''? Therefore, the Commission employs a hypothetical supracompetitive price increase to define the market. Even where historical purchasing patterns suggest that a product is insulated from a putative substitute under existing, competitive prices, relevant purchasers may readily turn to the substitute if faced with any significant price increase. Current preferences in the competitive equilibrium would be altered by an increase in the relative prices of gravure. Olin Corp., 113 FTC 590, 598 "(1990) ("Evidence of what 4 6x.1167-C-1.

Id.

"6 The estimates may understate slightly the proportion of offset printing in the proposed market. Respondents argue that these proportions underrepresent the amount of high volume work done on offset because the entire universe of gravure printers are represented while only 20 offset printers are represented, even though Dr. Hilke testified that there are "thousands and thousands" of offset printers. Hausman Tr. 5365; Hilke Tr. 3106. Dr. Hilke admitted that this may tend to understate the offset proportion, but that he considered the proportions reliable because the 20 offset printers appear to be those most likely to bid on high volume business. Hilke Tr. 3106-07, 3447. Nevertheless, in response, Dr. Hilke adjusted his estimates and arrived at the estimates stated in CX-1167-C-1. Dr. Hausman made his own estimates based on third-party information regarding the same print job parameters and found, inter alia, that 27% of work in complaint counsel's core was performed on offset (compared with Dr. Hilke's estimate of 4.5%). RX-668. These estimates have their own problems, none of which needs be resolved here.

Similarly, if offset prices increased from prevailing levels, one would expect an increase in the use of gravure, especially at lower volumes: the margin defined by the breakeven analysis would shift to a lower volume and page count.

118 See Robert G. Harris & Thomas J. Jorde, Antitrust Market Definition: An Integrated Approach, 72 Cal. L. Rev. 1 (1984) (examination of historical purchasing patterns is an initial step necessary to permit consideration of evidence bearing on purchasers, likely responses to future price increases); Pitofsky, supra note 60, at 1836 n.141 ("Examination of historical purchasing patterns is a first step in market definition. While it reflects the sensible concern about using actual rather than hypothetical data, it must be corroborated by further analysis in order to be reliable."). ne See ABA Merger Monograph, supra note 44, at 99-10! (and authorities cited therein). Opinion 120 F.T.C.

has happened during a period of equilibrium . . . does not serve as a predictor of what would happen if the price [of the product at issue] ... rose above the competitive level.").

Current substitution provides the strongest evidence that additional marginal substitution is likely to occur in response to a supracompetitive price increase.’ Absent evidence that current offset purchasing within the proposed market is not "competitively relevant," we can conclude that the parameters selected by complaint counsel to define the relevant market have not accurately identified inelastic uses of gravure.'*' There appears to be no basis for concluding that current substitution is not competitively relevant. Several of the largest print buyers in the United States use the offset process for high volume publication printing falling within the specifications of complaint counsel's proposed market.'” Several buyers have switched their high volume printing from gravure to 120 Pitofsky, supra note 60, at 1836 ("The single most reliable line of evidence in relevant market definition is whether, in response to past price changes, buyers promptly shifted to other products, or competitors promptly adjusted sales efforts. when evidence of this type is available, it should outweigh speculation based on theoretical constructs"). 121 oe Adventist Health System/West, Docket No. 9234, slip op. (April 1, 1994). In Adventist, the Commission rejected a proposed relevant market based on existing purchasing patterns similar to those here, as measured by "Elzinga-Hogarty" (““E-H”) statistics. The E-H statistics for a given area are the LOFI (‘little out from inside") -- the percentage of production in the area that is consumed in the area -- and the LIFO (‘little in from outside") -- the percentage of consumption in the area that is produced in the area. See Kenneth Elzinga & Thomas Hogarty, The Problem of Geographic Market Delineation Revisited: The Case of Coal, 23 ANTITRUST BULL. 1, 2 (1978) (proposing a relevant geographic market as an area in which the LIFO and LOFI both exceed 90%); Kenneth Elzinga & Thomas Hogarty, The Problem of Geographic Market Delineation in Antimerger Suits, 18 ANTITRUST BULL. 45, 73-76 (1973) (earlier proposal for 75% thresholds; now called the"weak" market test). In an appropriate case, the approach could be extended to product market definition: (T]he LOFI analogue would ask: what is the smallest group of applications required to account for nearly all shipments of a given product? The LIFO analogue would ask: Of total purchases intended to serve the group of applications identified by the LOFI test, do nearly all consist of the given product? ABA Merger Monograph at 97. In this case, the LIFO value for high volume gravure printing -- i.e., the estimate of current offset use in print jobs above the proposed parameters -- is 75.9%. The LOFI value -- i.e., the percentage of total gravure capacity used in the proposed high volume publication market is less than 31%. See Section V.A.1., supra. The LIFO statistic is more probative of substitution opportunities of consumers in a proposed price discrimination market. In Adventist, the Commission rejected the proposed geographic market based on a LIFO of 75% and lack of proof that existing substitution was not "competitively relevant.” Slip op. at 10-13 (also noting that EH statistics are merely one factor in market definition).

2 Competition is also indicated by bidding between offset and gravure printers. Both complaint counsel and respondent submitted economic and econometric analysis based on pricing data from common sets of print jobs for which both gravure and offset bids were received. For example, compare CX-1177; CX-764; CX-1411; CX-1348; with RX-465-R; RX-465-Z. The very existence of this pricing data suggests competition between the processes. The relevance of this competitive bidding can be discounted somewhat because the cost of submitting a bid may be relatively insignificant. Given our conclusions based on the amount of existing substitution, however, it is not necessary to engage in further analysis of this conflicting evidence. R.R. DONNELLEY & SONS CO., ET AL. 173 36 Opinion offset in recent years, including -- as the ALJ found -- present buyers who testified in the preliminary injunction hearing that they could only use the gravure process for their high volume printing needs and that they would not switch in response to a significant price increase. IDF J 122. As discussed above, some buyers have substituted to offset in order to take advantage of versioning, including Wal-Mart and K-Mart. IDF JJ 123, 125. Many other buyers, some of whose print jobs occupy complaint counsel's proposed core, recently have shifted purchases to offset (without respect to versioning). IDF JJ 124, 126-132 (e.g., Montgomery Ward, Hanover House, J.C. Penney, Damark, and Compuadd).

The experience of one customer in the proposed "core" is also instructive. Victoria's Secret uses offset for seventeen (low versioned) catalogs, each with an average of 100 pages and a run length of 3.5 million to 30 million copies. IDF { 129. Victoria's Secret, like many others, solicits bids from both gravure and offset printers and continues to purchase the bulk of its printing services from an offset printer. IDF J 130, 132. But this customer is "testing" with gravure printing for some catalogs and is comparing the sales results of (consumer responses to) catalogs printed on gravure versus catalogs printed on offset. IDF J 130. Although the first catalog printed on gravure did not meet expectations, Victoria's Secret continues to test, IDF { 131, and may very well shift all or a significant portion of its printing requirements to gravure. The ALJ treats these facts as if they indicate some inevitable trend to gravure for all high volume print jobs. But Victoria's Secret has been purchasing primarily offset for years, despite a print program that is among the largest in the proposed "core." If price alone dictated this customer's printing requirements, it would have been expected to substitute to gravure long ago, particularly in light of the relative decrease in the gravure cost advantage over the last decade. Should this customer substitute gravure for offset in the future, it should not be considered an inevitable and permanent gravure customer any more than Wal-Mart should now be considered an inevitable and permanent offset customer.

Complaint counsel attempt to argue that current substitution is not marginal by arguing that "there is no record evidence that offset has taken a long publication gravure job on the basis of price. .. . there [are] many instances where the program of the retailer changes, and becomes a many version job, and they switch, but none of price Opinion 120 F.T.C.

alone." OA Tr. 48." In fact, there is testimony in the record of gravure-to-offset substitution on the basis of price.’ And, in any event, substitution on the basis of price alone makes little sense in the context of competition between differentiated products. Where putative substitutes are differentiated, "relevant market definition turns on the aggregate decisions of different classes of customers who have different attitudes toward the importance of price and product characteristics in deciding whether to substitute or not."!* A publication printing purchaser's choice of characteristics for a particular print job -- and thus the purchaser's choice of printing technology -- does not turn on any single factor. Each purchaser demands the-package of printing services attributes that will reach its customers most effectively.'** The purchaser will choose the bundle of characteristics that offers the highest profit. This calculation will be determined partly by the prices offered by the two competing printing technologies. Moreover, as the discussion of versioning illustrates, even when substitution is triggered by a change in the relative prices, one would not necessarily expect other characteristics of the job to remain constant -- purchasers would almost surely select a new set of optimum job characteristics (i.e., a new program). It is perhaps plausible that there are few instances where only the print process (but not the print program) changes. The ALJ errs, however, by inferring that this substitution is not economically relevant for the purpose of evaluating the competitive effects of this transaction. '?” In defining relevant markets under Section 7, the Commission and the courts recognize competition -- and the potential elimination of competition -- in variables other than price. In considering the —I 3 See CAB at 35 n.S0.

i 4 The witness from Wal-Mart testified that price was a primary consideration in shifting a substantial portion of its insert printing from"gravure to offset. Baron Tr. 2307-08, 227576; RX-383. See also RX-84-B (Wal-Mart stating that it is seeking bids from both gravure and offset printers to ensure competitive pricing); Moeller Tr. 4003-04 (Donnelley witness referring to substitution of offset for gravure by Hanover Direct, a retailer, based on more favorable price); IDF J 124. 25 pitofsky, supra note 60, at 1835. See Robert D. Willig, Merger Analysis, Industrial Organization Theory, and Merger Guidelines, BROOKINGS PAPERS ON ECONOMIC ACTIVITY (MICROECONOMICS) 281, 299-305 (1991) (economic analysis of merger policy in a differentiated products setting).

126 See, e.g., Hilke Tr. 6028-30 (print customer's choice of paper determined by preferences of purchasers of the publication; choice of paper and number of versions suggest offset more economical than gravure for particular print job).

127 [Redacted]’s experience demonstrates that demand for print services is multi-dimensional; even though the witness testified that the bid prices she receives from gravure printers are generally lower than bid prices for offset, [redacted] continues to purchase primarily offset printing services. IDF TW 262-64.

R.R. DONNELLEY & SONS CO., ET AL. 175 36 Opinion likely reaction of buyers to a price increase for purposes of market definition, the Merger Guidelines state that "the Agency will take into account all relevant evidence, including ... evidence that buyers have shifted or have considered shifting purchases between products in response to relative changes in price or other competitive variables." Section 1.11 (emphasis added). See United States v. Continental Can Co., 378 U.S. 441, 455 (1964) (rejecting argument that different prices for metal and glass containers demonstrate the existence of separate glass and metal container markets because "price is only one factor in a user's choice between one container and another").'78 No evidence in the record appears to suggest that high volume customers using offset are inframarginal, economically irrational, or otherwise irrelevant to market definition. Complaint counsel offer no explanation for the existing use of offset.'”” Thus, there is no reason to believe that the high volume print jobs currently using offset do not represent marginal consumers in high volume publication printing, as defined by the complaint.'*° 8 Cf; Merger Guidelines Section 0.1 n.6 ("Sellers with market power also may lessen competition on dimensions other than price, such as product quality, service, or innovation."). Complaint counsel did not attempt to speculate seriously on existing substitution decisions at oral argument:

COMMISSIONER STAREK: [Why is it that the I.D. found that 11.4 percent of all the low versioned jobs over 32 pages for runs of 5 million or more were done in offset? Why did 11 percent of the purchasers choose offset? MR. DOYLE [for complaint counsel]: I think on the basis of the record evidence on quality and cost, it doesn't make sense. And quite frankly, Commissioner, I don't know. I don't know. CHAIRMAN STEIGER: Does the record indicate whether there could .. . have been a capacity constraint problem or a time problem for these 11 percent, or is that not in the record? MR. DOYLE: Well, . .. given the fact that this is 1990 dated, there are instances when particular suppliers could have been capacity constrained, and that very well could explain that. But beyond the capacity constraint situation, I can’t address that. OA Tr. 48-49.

130 The ALJ also found that evidence presented by complaint counsel of independent gravure and offset prices is inconsistent with the proposition that the two processes compete. ID at 8384; IDF] 211. Complaint counsel attempt to support the proposed relevant market with evidence purportedly showing a lack of correlation between the prices of offset and gravure. There are serious doubts about the quality of the data on which these correlations were based; but even putting these reservations aside, there are serious conceptual problems with using price correlations to delineate antitrust markets. Although some economists have proposed using price correlations to identify relevant markets, others have noted that price correlations often can be very misleading and, in any case, are static measures of competition that are not designed to predict probable demand responses to an exercise of market power. See, e.g., Jonathan Baker, Why Price Correlations Do Not Define Antitrust Markets, FTC Working Paper No. 149 (1987); Luke M. Froeb & Gregory J. Werden, Correlation, Causality, and All That Jazz: The Inherent Shortcomings of Price Tests of Antitrust Market Definition, 8 REV. INDUS. ORG. 329 (1993). Cf. Coca-Cola Bottling Co. of the Southwest, 118 FTC 452 (1994) [FTC Docket No. 9215 (Aug. 31, 1994), slip op. at 46-471. In particular, even if two goods are very close substitutes (as measured by demand cross-elasticity), the correlation between their prices can be low given sufficiently high supply elasticities. Hence, even assuming that the price correlations presented by complaint counsel are valid, this is not dispositive evidence that high volume gravure printing is a relevant product market. In any event, actual evidence of existing competition between gravure and offset confirms that this and other indicia of separate markets should be given little weight. Opinion 120 F.T.C.

The record as a whole shows substantial existing competition between gravure printing services and offset printing services, particularly in publication printing for print jobs with volumes between one million and ten million copies, but the margin (with the versioning parameter appropriately evaluated) appears to extend into even higher volumes. In view of these facts, it appears that gravure printers, if unified by a hypothetical cartel or merger, could not identify inelastic end users effectively and, thus, could not profitably impose a small but significant and nontransitory increase in the price of high volume publication printing.

V. RELEVANT COMPETITORS & CONCENTRATION Based on the foregoing, we conclude that "high volume publication gravure printing” is not a relevant market for purposes of assessing the competitive effects of the acquisition. At a minimum, the "gravure" qualification to that definition is too restrictive: even if we assume the existence of a distinct market for "high volume publication printing,” the relevant competitors in that market would include a significant number of firms that supply both gravure and offset printing services and a significant number of firms that supply only offset printing services.'*' In the remainder of this opinion, we will assume the existence of a "market" for high volume publication printing in order to provide a more comprehensive description of (i) the competitive interaction of firms currently capable of producing high volume publication printing services and (ii) the possibilities of anticompetitive conduct by the merged firm, either unilaterally or in coordination with other firms.

The assumed relevant market must be measured in terms of its participants and concentration. Following the methodology of the Merger Guidelines, the Commission identifies relevant suppliers, assigns to each relevant supplier a market share that reflects its future competitive significance in the relevant market, and then, based on these determinations, calculates market concentration. See id. It is far from clear that this assumption is reasonable. To conclude that high volume publication printing services constitutes a relevant market, the Commission must find that a hypothetical monopolist of printing services could systematically discriminate, on the basis of relative elasticities of demand, between high volume customers -- those satisfying complaint counsel's volume, page count, and color criteria, but not the versioning criterion -- and other printing customers. The evidence suggests that a hypothetical monopolist could not systematically discriminate in this way. Instead, the evidence tends to demonstrate that firms producing gravure and/or offset printing services cannot systematically identify print jobs for which customers have relatively inelastic demand. See Section IV.D.-E. R.R. DONNELLEY & SONS CO., ET AL. 177 36 Opinion Sections 1.3 - 1.5. Market concentration is a function of the number of firms in the relevant market and their respective shares. Jd. Section 1.5. The Commission uses the Herfindahl-Hirschmann Index ("HHI") as the most economically relevant measure of concentration. The HHI is calculated by summing the squares of the shares of relevant market participants, and thereby "gives proportionately greater weight to the market shares of the larger firms, in accord with their relative importance in competitive interactions." Jd.'* A. Relevant Market Participants For purposes of estimating market shares and concentration in the assumed relevant market, firms to be identified as participants are (i) those currently engaged in high volume publication printing at the time of the acquisition and (ii) any "uncommitted entrants" -- firms not currently engaged in high volume publication printing but whose inclusion would more accurately reflect probable supply responses to a supracompetitive price increase. Merger Guidelines Section 1.3. The ALJ identified as relevant suppliers only the nine printers with gravure capacity at the time of the acquisition. IDF {J 377-79.'*? The ALJ erred in failing to identify as relevant participants at least seventeen offset-only printers that, according to complaint counsel, supplied high volume publication services in the United States at that time.'* We find that each offset producer that currently bids on and supplies high volume publication printing is a relevant participant. The ALJ also erred in failing to address whether any printing firms not currently supplying high volume publication printing are properly considered "uncommitted entrants." In the assumed price discrimination product market, such firms would include any offset printers who currently are not supplying high volume publication printing but who are likely to do so "within one year and without the expenditure of significant sunk costs, in response to a ‘small but significant and nontransitory' price increase."'*> Merger Guidelines 132 cee FIC v. PPG Indus., Inc., 7.98 F.2d 1500, 1503 (D.C. Cir, 1986). 133 The ALJ found that, since the acquisition, one firm has exited (Standard Gravure in 1992) and two others have merged (Quebecor acquired Arcata in 1993). IDF § 376. 134 See note 113, supra (describing CX 1167). All existing gravure printers participate in high volume publication printing and have been included as relevant suppliers. Moreover, we conclude that entry into high volume publication printing using gravure capacity would require the expenditure of substantial sunk costs and would entail a lag Opinion 120 F.T.C.

Section 1.32. "Sunk costs are the acquisition costs of tangible and intangible assets that cannot be recovered through the redeployment of these assets outside the relevant market." Jd. Since we find that the relevant market is not limited to gravure printing and have assumed that the relevant market is high volume publication printing, we find that assets used for high volume publication printing represent sunk costs only if the assets cannot be redeployed to lower volume print jobs. Even assets that can be used only in publication printing are not properly considered sunk if they can be used for printing jobs that do not meet the volume and page count criteria that define the assumed relevant market.'*° Because offset printing assets that are used in high volume publication printing can also be used for lower volume publication printing, the acquisition costs of those assets are not properly considered sunk for purposes of this analysis.'*? Complaint counsel did not present, and the ALJ did not find, any other significant sunk costs of entry into or exit from the assumed relevant market.'*® Therefore, a firm not currently bidding on high volume publication printing jobs may be considered an uncommitted entrant and counted among the relevant suppliers if (i) it owns offset printing assets capable of both high volume and lower volume publication printing or (ii) it could acquire such assets and deploy them within one year.'*? The record does not permit a reliable estimate of the likely supply response by such firms to a supracompetitive price of more than one year. See IDF ¥] 365-75 (finding that de novo entry into gravure printing would take more than two years); CPF ¥J 1480-90 (entry into gravure printing production entails substantial sunk investment). Therefore, the only potential uncommitted entrants are printers who could deploy offset capacity to supply high volume publication printing. 16 Complaint counsel argue that asset acquisition costs are sunk if the assets cannot be redeployed for use other than in gravure printing. CPF J] 1477-84. This would be correct in a market defined to include only gravure printing.

This is particularly clear with respect to the more recent generation of offset presses. See Section IV.D.2., supra. To the extent that these presses can be used to produce efficiently in both high volume/high page and low volume/low page jobs, their purchase cannot be considered a sunk cost in either putative "market." See RRB at 56.

138 In the assumed price discrimination market, in which both gravure and offset technologies participate, the only costs identified by complaint counsel that may properly be considered sunk are the investments in "accumulated know-how and familiarity that both printers and customers gain about each other during an on-going relationship." Complaint counsel argue that this sunk investment "means that customers are reluctant to switch suppliers." CPF J 1462. Of course, such costs (assuming they are significant) affect the potential for switching to any new supplier, whether it is an offset printer or a gravure printer. And in any event, the significance of such costs is belied by the substantial switching between gravure and offset at preacquisition and current prices. See Section IV.E, supra; IDF { 122-32. 39 "Uncommitted supply responses may occur .. . by the switching or extension of existing assets to production or sale in the relevant market; or by the construction or acquisition of assets that enable production or sale in the relevant market." Merger Guidelines Section 1.32. R.R. DONNELLEY & SONS CO., ET AL. 179 36 Opinion increase by incumbent high volume publication printers. Nevertheless, given the lack of contrary evidence, we find that some supply response is likely to constrain the conduct of incumbent firms. B. Pre-Acquisition Market Shares & Concentration In general, each firm that is identified as a market participant is assigned a market share that reflects its future competitive significance. Normally, this share is "based on the total sales or capacity currently devoted to the relevant market together with that which likely would be devoted to the relevant market in response to a 'small but significant and nontransitory' price increase." Merger Guidelines Section 1.41. More specifically, when the potential for price discrimination defines the relevant market, each firm should be assigned a share reflecting sales likely to be made in, or capacity likely to be used to supply, the price discrimination market in response to such a price increase. Merger Guidelines Section 1.42. In the assumed relevant market, complaint counsel argue that firms are primarily distinguished on the basis of their relative advantages in serving different groups of customers.'*° Under these conditions, unit sales are generally the best indicator of firms, future competitive significance. Id.

Using complaint counsel's estimates for unit sales at the time of the acquisition, Table 1 depicts market shares and concentration in high volume publication printing defined to include print jobs of at least five million copies and at least sixteen pages per copy, and Table 2 depicts market shares and concentration in the supply of printing services for print jobs of at least ten million copies and at least sixteen pages per copy.'*! 140 coe CPF Y 1737.

M41 Tables 1 and 2 are derived from CX-1167, which summarizes complaint counsel's estimates of relative shares of output of gravure producers (CX-1167-D) and offset printers (CX-1167-E). Opinion 120 F.T.C.

TABLE 1 U.S. UNIT SHARES IN HIGH VOLUME PRINTING (1990) 5 MILLION+ COPIES (16 PAGES/COPY) Billions of Pages Printer Gravure Offset Total Share HHI Donnelley 313.4 10.7 324.1 28.6 818 Quebecor 177.4 2.4 179.8 15.9 253 Meredith/Burda 157.6 11.2 168.8 14.9 222 Ringier 75.8 8.8 84.6 75 56 Valassis 0.0 69.9 69.9 6.2 38 Arcata 58.5 5.2 63.7 5.6 31 Quad 19.8 43.5 63.3 5.6 31 World Color 34.4 19.7 54.1 4.8 23 Sullivan 0.0 20.0 20.0 1.8 3 American Signature 0.0 18.6 18.6 1.6 3 Perry 0.0 17.9 17.9 1.6 3 Standard 10.3 6.5 16.8 1.5 2 Brown 13.8 1.9 15.7 1.4 2 13 Others 0.0 36.6 36.6 3.2 * TOTAL 861.0 273.0 1134.0 100.0 1486* TABLE 2 U.S. UNIT SHARES IN HIGH VOLUME PRINTING (1990) 10 MILLION+ COPIES (16 PAGES/COPY) Billions of Pages Printer Gravure Offset Total Share HHI Donnelley 229,3 2.9 232.2 29.3 858 Quebecor 156.3 0.8 157.1 19.8 392 Meredith/Burda 98.1 0.0 98.1 12.4 154 Valassis 0.0 69.9 69.9 8.8 77 Ringier 55.2 8.4 63.6 8.0 64 Quad 14.1 42.9 57.0 7.2 52 Arcata 49.6 5.2 54.8 6.9 48 Perry 0.0 11.5 11.5 1.4 2 Sullivan 0.0 9.2 9.2 1.2 1 Brown 5.5 0.6 6.1 0.8 1 American Signature 0.0 5.8 5.8 0.7 1 Standard 4.9 0.7 5.6 0.7 1 Century Graphics 0.0 3.4 3.4 0.4 * Graphic Arts 0.0 3.3 3.3 0.4 * Alden 0.0 3.2 3.2 0.4 * World Color 0.0 2.1 2.1 0.2 * 6 Others 0.0 9.1 9.1 1.1 * TOTAL 613.0 179.0 792.0 100.0 1650* R.R. DONNELLEY & SONS CO., ET AL. 181 36 Opinion Two important caveats apply to the interpretation of Tables 1 and 2, each tending to suggest that the tables overstate concentration in the assumed relevant market. First, the tables may understate slightly the proportion of historical high volume publication printing sales made by offset printers. '** Second, neither of these tables reflects any likely supply responses of uncommitted entrants. To the extent that a significant and nontransitory increase in high volume publication printing prices by all or some subset of incumbent sellers would induce a supply response through uncommitted entry, the historical sales share of each incumbent seller would overstate its relative competitive significance. Shares of high volume printing capacity (as an alternative to unit sales) may reflect a more comprehensive measure of the future competitive significance of incumbent firms and uncommitted entrants.'*? However, we lack a reliable basis for making (i) a quantitative estimate of the full measure of offset capacity likely to be devoted to the high volume publication printing in response to a supracompetitive price increase, or (ii) a quantitative adjustment to the historical sales shares reflected in Tables 1 and 2. In sum, although Tables 1 and 2 may not significantly overstate the appropriate measure of relevant market concentration, they likely represent the upper bound of the range of reasonable concentration estimates.

C. Effect of the Acquisition on Concentration Based on this (appropriately qualified) concentration information, the Commission next considers whether the acquisition significantly increases concentration and results in a concentrated market. Other things being equal, market concentration affects the likelihood that 142 See note 116, supra (even complaint counsel's expert admitted that his approach to data collection and estimation may tend to understate the use of offset printing, although not significantly; moreover, respondents’ economic expert found that, even using complaint counsel's market parameters, offset accounted for a much higher percentage of high volume printing jobs than complaint counsel had estimated).

143 In order accurately to portray the relative competitive significance of the identified relevant market participants, all capacity used or likely to be used to supply the price discrimination product market should be included for purposes of calculating market shares and concentration. Merger Guidelines Section 1.42. The evidence suggests that all gravure printing capacity should be included in the relevant market, and that some unspecified, substantial portion of offset printing capacity must also be included.

Opinion 120 F.T.C.

one firm, or a group of firms, could successfully exercise market power. /d. Section 2.0.“ From Table 1, the maximum effect of the acquisition on market share and concentration in print jobs of at least five million copies and at least sixteen pages per copy can be summarized as follows:'* Donnelley-Meredith/Burda Share: 43.5% Increase in HHI: 852 Post-Acquisition Market HHI: 2338 From Table 2, the maximum effect of the acquisition on market share and concentration in print jobs with a volume of at least ten million copies can be summarized as follows:'° Donnelley-Meredith/Burda Share 41.7% Increase in HHI 727 Post-Acquisition Market HHI 2377 In general, the Commission presumes that an acquisition producing an increase in the HHI of more than 100 points and yielding a post-acquisition HHI exceeding 1800 is likely to create or enhance market power or facilitate its exercise. Merger Guidelines 1.51.'*” Based on the concentration information described above, this presumption likely applies.'** This presumption may be overcome by showing that other market conditions make it unlikely that the "4 See B.F. Goodrich Co., 110 FTC 207, 303 (1988) ("As the number of firms in an industry declines, and industry concentration increases, ceteris paribus, it becomes easier for those firms to coordinate their pricing, and the likelihood of anticompetitive effects from an acquisition consequently increases as well."). Conversely, a merger between market participants is unlikely to create or enhance market power or to facilitate its exercise unless it significantly increases concentration in the market. Merger Guidelines Section 1.0.

145 Compare IDF 4 378, Tables | and 2 (based solely on gravure capacity, finding that the acquisition produced an HHI increase in the range of 920-1029 to a post-acquisition HHI in the range of 3070-3093). See CX-501-A-B.

For purposes of the competitive effects analysis that follows, note that both the merged firm's market share is lower in the higher volume segment. 147 In general, the Commission regards as “highly concentrated" markets"with an HHI exceeding 1800 and regards as significant an increase in the HHI exceeding 100 points. /d.; see, e.g., Coca-Cola Co., slip op. at 44 (HHI increase of 443 to post-merger HHI of 3572); Owens-Illinois, slip op. at 27 (using production figures, post-acquisition HHI of 2478 with increase of 852 points); Olin Corp., 113 FTC 400, 610-11 (1990), aff'd, 986 F.2d 1295 (9th Cir. 1993), cert. denied, 114 S.Ct. 1051 (1994) (based on production, post-acquisition HHI of 4122, with increase of 1186); Hospital Corp. of America, 106 FTC 361, 488 (1985), 807 F.2d. 1381 (7th Cir. 1986), cert. denied 481 U.S. 1038 (1987) (postacquisition HHI of 2416 with increase of 395 points). 148 . .

We assume here that an adjustment to compensate for any understatement of actual and likely offset supplies would not be sufficient to reduce market concentration below the highly concentrated threshold or to reduce the HHI increase below 100. R.R. DONNELLEY & SONS CO., ET AL. 183 36 Opinion acquisition will create or enhance market power or facilitate its exercise, in light of market concentration and market shares. VI. LIKELY COMPETITIVE EFFECTS OF THE ACQUISITION An acquisition may give rise to anticompetitive effects in two general forms. First, it may facilitate coordinated interaction: a collective exercise of market power among relevant suppliers. Second, it may allow the merged firm to exercise market power unilaterally. See Generally Merger Guidelines Section 2. The ALJ found that the acquisition is likely to give rise to both unilateral and coordinated anticompetitive effects. ID at 90-91. In addition, the ALJ found that the merged firms already have exercised market power unilaterally by cancelling or deferring pre-acquisition plans to expand printing capacity. ID at 89-90. We reject these findings and conclude that neither coordinated nor unilateral anticompetitive effects are likely in high volume publication printing.'” A. Coordinated Interaction Analysis "Coordinated interaction" is defined broadly as "actions by a group of firms that are profitable for each of them only as a result of the accommodating reactions of the others. This behavior includes tacit or explicit collusion, and may or may not be lawful in and of itself." Merger Guidelines Section 2.1. "Successful coordinated interaction entails reaching terms of coordination that are profitable to the firms involved and an ability to detect and punish deviations that would undermine coordinated interaction." Id.; see Coca-Cola Bottling Co. of the Southwest, 118 FTC 452 (1994) [FTC Docket No. 9215 (Aug. 31, 1994), slip op. at 74]; B.F. Goodrich Co., 100 FTC at 294. The ALJ made few findings related explicitly to the possibility of coordinated interaction, and concluded that the acquisition increases the likelihood of coordinated interaction because the postacquisition market is concentrated and because information about competitive activity of suppliers is readily available. ID at 90-91.'° 149 We reject the ALJ’s finding that Donnelley's “deferral or cancellation" of capacity expansion plans constituted an anticompetitive effect because, inter alia, we find that the merged firm does not have unilateral market power.

150 The ALJ found that "Donnelley's acquisition of one of its primary competitors . . . may lead to coordinated interaction, or collusion, among the remaining firms." IDF 406 He makes five subsidiary findings. First, the acquisition increases concentration in an already concentrated market. IDF ¥ 406. Second, Donnelley is a dominant firm and, as a result, coordination of prices is more likely Opinion 120 F.T.C.

We affirm the ALJ's findings that the significant increase in concentration occasioned by the acquisition increased the probability of successful coordination. Nevertheless, based on the totality of market conditions, we conclude that coordinated interaction to discriminate against high volume publication printing customers would be inherently unstable.

The two principal impediments to coordinated interaction are specific to the theory of discriminatory effects alleged. First, relevant suppliers would have strong incentives to deviate from the coordination by diverting capacity from the elastic lower volume into the more profitable high volume publication printing jobs. Second, since the alleged inelastic printing jobs encompass (by definition) only the largest contracts of the largest customers, the incentives to make such diversions are overwhelming. As discussed below, the significance of each problem is inversely related to the breadth of the relevant discrimination market: if the market is defined narrowly in order to minimize the perceived substitutability of offset for gravure, then each problem is magnified. Other market conditions exacerbate these problems.

1. Potential Diversion of Capacity Coordination directed at some subset of the customers served by a putative collusive group is inherently less likely to succeed than coordination focused on all customers, all other things equal. Owens- Illinois, slip op. at 31. Relative to a nondiscriminatory coordination, a discriminatory coordination creates greater incentives to depart from the terms of the coordination and can be undermined by smaller increases in output into the relevant market by relevant suppliers that either cheat or fail to coordinate altogether. Assuming that relevant because cheating is easier to detect and punishment is severe. IDF { 407. Third, coordination of gravure prices is possible because of the ready availability of "information about competitive activity of industry members." IDF ¥ 408. Fourth, "[t]he nature of gravure printing may also facilitate coordination: there are only two major manufacturers of gravure presses . . . ; all gravure printers use the same process to produce the finished product, and much of the printers, business is obtained through bidding, which requires an intimate knowledge of industry cost structure and other competitive variables." IDF ] 409. Fifth, customers are concerned that switching to new suppliers is difficult. IDF § 411-15. As described below, only the concentration finding appears well-founded. In particular, the record basis for the ALJ's conclusion that competitively-sensitive information is readily available and facilitates coordination is very thin. For example, much of the witness testimony cited in the relevant factual finding, IDF J 408, merely describes speculation about the availability of relevant competitor information. See, e.g., Engdahl Tr. 2547, Hodgson Tr. 333. And the record does not reflect that rival printers can readily obtain competitively sensitive information that would facilitate tacit, or explicit collusion.

R.R. DONNELLEY & SONS CO., ET AL. 185 36 Opinion publication printers can accurately identify printing jobs by relative elasticities of demand, the potential for diversion of printing capacity from elastic lower volume jobs to inelastic high volume publication printing poses a serious impediment to coordination. Under complaint counsel's theory of coordinated anticompetitive effects, relevant suppliers will earn higher economic returns from high volume publication print jobs than from lower volume print jobs.'*' As such, any relevant supplier would have an incentive to divert capacity from the lower volume printing jobs to the higher-return high volume jobs.'* As output is diverted to the high volume jobs (i.e., as supply to the assumed relevant market is increased), prices will fall toward the competitive level. Despite the obvious application of this analysis to the instant case, and the precedent in Owens-Illinois, the ALJ did not address the possibility of diversion. In Owens-Illinois, the Commission found that certain end users of glass containers had inelastic demand for glass (based on the characteristics of the end use), that the inelastic end uses comprised a minor portion of total glass container output, and that capacity used to produce glass containers for elastic end uses could also be used to produce glass containers for inelastic end uses. Under these circumstances, the Commission considered whether glass container manufacturers would collusively price discriminate against customers in the inelastic end uses, and concluded that the incentive to divert capacity would be a powerful force subverting collusion: Any collusive scheme focused on the inelastic end-uses would be threatened not just by the normal incentives to cheat which might in some circumstances undermine even across-the-board collusion; it would face in addition the disruptive force of a pool of readily fungible productive capacity far greater in magnitude than any contemplated output reductions, yet presently devoted to the elastic end-uses and therefore not benefiting from the collusive scheme. This is true regardless of the precise nature of the predicted anticompetitive conduct. As discussed below, whether the coordination takes the form of an agreement (tacit or explicit) on price and other relevant terms or an agreement (tacit or explicit) to allocate customers, the outcome is that suppliers achieve supracompetitive profits with respect to the targeted customers. Therefore, the incentives to obtain supracompetitive profits and the diversion analysis in this section apply to discriminatory coordination generally.

2 Printers have strong incentives to divert capacity to an inelastic end use even if they currently operate at full capacity utilization. The profit maximizing gravure printer would shift production based on changes in the relative profitability: if prices increased in high volume publication printing, every gravure printer would have an increased incentive to obtain the supracompetitive profits available from shifting sales away from elastic uses.

Opinion 120 F.T.C.

Id., slip op. at 31. Following this analysis, discriminatory coordination is less likely, other things equal, the smaller is the percentage of total capacity currently devoted to the inelastic end uses and the more cross-price elastic is supply between the elastic and inelastic end uses.

Even in the case most favorable to complaint counsel -- that is, assuming that only gravure printers are necessary to successful coordination and that only gravure capacity could be diverted to the inelastic high volume print jobs -- it appears that a small diversion would undermine any anticompetitive price increase. The parties’ estimates of the share of total gravure capacity currently devoted to all high volume publication printing range from approximately 9% to 31%, depending on a variety of disputed assumptions about production and capacity.'*? Using the formulation from Owens- Illinois, a gravure output diversion from low volume printing into high volume printing of at most approximately 1.5% -- and perhaps as little as 0.5% -- would defeat a supracompetitive price increase of 5% (the benchmark generally used to define a significant and nontransitory price increase).'** If the hypothetical coordination 153 Respondents’ expert, Dr. Hausman, estimated that "low version gravure jobs’ with more than 32 pages” and more than 5 million copies accounted for 6.8% to 17.6% of total gravure capacity, depending on the measure of capacity. See RX-397; RX-687; RX-688. Using the same capacity denominators, but changing the numerator to include all jobs with more than 16 pages (to conform with complaint counsel's proposed market), low version gravure jobs above 5 million copies account for 8.7% to 22.5% of total gravure capacity. See CX-1351-F; CX-1167 (reducing the page threshold from 32 to 16 increases the numerator almost by 28%. Complaint counsel's expert, Dr. Hilke, calculated "the gravure capacity share of low version gravure jobs," with more than 16 pages. See CX-1351-I. He estimated that such printing jobs greater than 5 million copies accounted for 30.6% of total gravure capacity.

Dr. Hilke also estimated "the gravure capacity share of low version gravure jobs compared to non- Donnelley capacity." See CX-1351-A and G, and "the gravure capacity share of low version gravure jobs compared to capacity outside the [three or four largest firms including Donnelley]." See CX-1351-C and H. Since Donnelley is part of the hypothetical coordination, excluding its capacity unreasonably understates the amount of divertible capacity in the coordinating firms. Excluding the capacity of the largest three or four firms is an even greater distortion. The principal point of the diversion analysis is not to determine whether some “fringe” has sufficient capacity to perform all of the work in the relevant market; it is to determine the amount of diverted output that would defeat coordination at the margin. Nevertheless, assuming a subset of gravure printers could coordinate, "fringe" firms with little or no sales into the assumed market would have stronger incentives to shift output into the market. 154 This diversion percentage is derived as follows (from Owens-Illinois, slip op. at 32 n. 36): Elasticity of demand is defined roughly as a ratio of percentage change in output demanded divided by a given percentage change in price; demand for a product is inelastic if this ratio is less than one. Ina market with inelastic demand, for example, a 5% price increase would result in less than a 5% decrease in output, and a 5% price increase could be defeated by less than a 5% increase in output. Thus, a 5% price increase targeted at allegedly inelastic customers representing 9% of total printing output would be defeated by an increase in output of less than 5% of 9%, or 0.45% of total gravure capacity. That amount of output represents approximately 0.5% of gravure production available for elastic end uses. See also Merger Guidelines Section 1.11 ("In attempting to determine objectively the effect of a ‘small but significant and nontransitory’ increase in price, the Agency, in most contexts, will use a price increase of five percent lasting for the foreseeable future”). R.R. DONNELLEY & SONS CO., ET AL. 187 36 Opinion among gravure printers is limited to imposing supracompetitive prices in the "core" (low-versioned jobs, 32 pages per copy, at least 10 million copies), the output diversion required to upset the coordination is even smaller. The parties’ estimates of the share of total gravure capacity currently devoted to the "core" range from approximately 3% to 15%, depending primarily on disputed assumptions about capacity.’ Thus, a gravure output diversion of less than 1% into the core would defeat a supracompetitive price increase of 5% on core printing jobs.’ This exercise illustrates that a discriminatory coordination is increasingly less stable the more narrowly defined is the alleged market of inelastic customers. Thus, if the market definition properly takes account of actual and potential substitution to offset, the category of inelastic customers (assuming they can be identified by any objective criteria) appears considerably smaller than either complaint counsel's broad market or its alleged "core." Thus, the gravure output diversion necessary to defeat a price increase to any inelastic customers is likely to be considerably less than one percent. In any event, the percentage of total relevant printing capacity currently used in high volume publication printing is not meant to yield a precise prediction of likely results. Instead, it merely illustrates the relatively strong incentives of suppliers, to deviate from a discriminatory coordination and, thus, the inherently greater instability of such coordination, other things equal.'*” 155 See RX-397-A, B, E (using various measures of capacity, Dr. Hausman estimated that core printing accounted for 2.9% to 10.9% of total gravure capacity); RX-687 and RX-688 (using adjusted capacity and production data from CX-S02 and CX-1167, Dr. Hausman estimated that core printing accounted for 3.0% to 10.8% of total gravure capacity). See CX-1351-I (Dr. Hilke adjusted the capacity and production data used to produce RX-397 and estimated that core printing accounted for 14.8% of total gravure capacity).

156 If core printing accounts for 3% of total gravure capacity, diversion of less than 0.2% from lower volumes would defeat the price increase. If core printing accounts for 15% of total gravure capacity. diversion of less than 0.9% from lower volumes would defeat the price increase. Complaint counsel also argue that cheating by colluding printers and diversion by a noncolluding fringe would not "reduce the price obtained by the primary printer." CAB at 69, They state: “where, as in printing, all production is customized and done to order," cheating and fringe firms "would have to increase their production massively to enable customers to switch work from colluding printers to cheaters or to the noncolluders." CAB at 69-70. It is not clear that the form of competition, coordination, or diversion this argument contemplates is economically well-founded. For example, under complaint counsel's customer allocation scenario, cheating (diversion) would take the form of soliciting business from customers who have been allocated to other suppliers. By definition, the cheater would offer terms of sale that are more attractive to the customer than the terms offered by the supplier earning supracompetitive profits; if it cannot offer more attractive terms, then the allocation is a natural consequence of cost conditions and does not require coordination. In addition, the necessary diversion does not require any increase in production by any firm; it merely requires a shift in production in order to obtain higher profits.

Opinion 120 F.T.C.

Having established in theory that a small diversion of output would defeat a discriminatory coordination, even under assumptions most favorable to complaint counsel, we now examine whether such a diversion is practical and economically feasible. Gravure printing Capacity appears to be highly cross-price elastic between low volume and high volume publication printing. Each of the incumbent gravure printers is not only capable of supplying high volume publication printing; complaint counsel assert that "all the gravure printers are already doing high volume publication gravure printing." CAB at 70- 71. Consequently, substantial capacity now being used to produce into the elastic lower volumes likely would be diverted in response to the prospect of earning higher, supracompetitive returns in high volume printing.'** Even the gravure printer with the smallest amount of gravure capacity, Brown Printing, alone has more than sufficient divertible capacity to defeat any coordinated price discrimination.’ Complaint counsel argue that the theoretical diversion calculation understates the actual amount of diversion necessary to defeat a supracompetitive price increase in high volume publication printing. CAB at 69. In particular, complaint counsel assert that diversion of small amounts of high volume publication printing is not economical for customers because printing is characterized by job-specific scale economies and "batch economies" and because customers value color consistency throughout a print job (which, presumably, is best achieved by consolidating one job in a single printer). Jd. Even assuming, however, that it is more efficient to consolidate all of the work on one job in one printer,’ this does not significantly affect the incentives or potential for diversion. Diversion (even in relatively smal] amounts) need not take the form of dividing up a single print job. The high volume printing programs of most publication printing Although the ALJ found that switching costs between printing customers are significant, IDF J 411-15, this finding proves too little and too much at the same time. The finding necessary to the conclusion that relevant diversion would not undermine coordination is that it is less costly for a buyer to switch to a printer who already supplies high volume publication printing than to switch toa printer who does not. The record does not support such a finding. If the finding is simply that it is costly for a buyer to switch to any printer, it suggests that the acquisition at issue would have very little effect on competition.. The finding has greater resonance in the discussion of unilateral effects. See Section VLB., infra.

15 See RPF { 221. Dr. Hausman testified that, following its acquisition of a three meter gravure press (one of the largest and most efficient in the U.S.), Brown alone has more than six times the capacity needed to defeat the price increase. Hausman Tr. 6369-71. ' The record contains numerous references to high volume print jobs that are divided among more than one gravure printer at current prices. See, e.g., Steen Tr. 1110-11, 1132; Doty Tr. 4593. With supracompetitive prices, the incentives to divide jobs would increase, although perhaps not significantly. R.R. DONNELLEY & SONS CO., ET AL. 189 36 Opinion customers entail several jobs, and many customers routinely use more than one gravure printer within a program -- using one printer for one job and another printer for another job.'*' Thus, a buyer would likely attempt to induce diversion to obtain lower prices by switching entire jobs among gravure printers, and could do so without affecting any job-specific scale economies or color consistency. Substantial increases in gravure capacity in recent years have increased both the potential and the incentives of gravure printers to divert output to chase the highest available return. Since 1990, gravure printers other than the merged firm have engaged in substantial capacity expansions.’ In addition, gravure printers could use any excess capacity to undermine the coordinated interaction. The ALJ appears to find that gravure printing capacity has been in shortage, IDF { 282, based largely on evidence that gravure printing capacity at certain locations for certain limited periods was fully utilized.' The totality of the evidence, however, suggests that excess capacity exists. Several industry participants, including witnesses called by complaint counsel, testified "that there is and has been excess gravure capacity." IDF J 358. Moreover, in recent years, "many firms are cutting back or completely cancelling their long-run, high-volume printing programs." IDF J 360. Some buyers have simply eliminated publications -- the most prominent of whom are Sears and Montgomery Ward. Jd. Other buyers have increased versioning, choosing more targeted marketing, and have substituted offset for gravure accordingly. IDF J] 361-62. At the same time, several firms have expanded gravure capacity. IDF JJ 280-84. Consistent with this information, gravure prices have been falling since at least 1985 and continue to fall. This evidence is consistent with two market conditions: (i) gravure capacity exceeds gravure demand or (ii) offset capacity constrains the price of gravure printing. Sackett Tr. 570-72; McCoig Tr. 739-40; Gallo Tr. 828-29; Scirocco Tr. 1029-34; Steen Tr. 1079; Joy Tr. 1166; Haight Tr. 1347-48; Baron Tr. 2279-82. 162 ,,. : ' : :

Using complaint counsel's estimates, gravure printers other than the merged firm have increased total gravure printing capacity by more than 7% between 1990 and 1993. See CX-502-B-C (indicating expansions by Quebecor, Ringier, Quad, and Brown). The post-acquisition expansion by gravure printers other than the merged firm is equivalent to more than 23% of Meredith/Burda's preacquisition capacity. Jd. In fact, post-acquisition capacity expansions by these other gravure printers may exceed 50% of Meredith/Burda's pre-acquisition capacity, measured by the number of presses. Ringier has acquired two gravure presses and is planning to acquire three additional large (3-meter) presses. Nytko Tr. 1474-75; RX-269-I. Brown has added the first 3-meter press installed in the US. IDF 4 373; Engdahl Tr. 2525; RX-269-B. Quad Graphics, which entered in the mid-1980s, has acquired three gravure presses, and has ordered two more (as well as four M-3000 (high volume) offset presses). IDF 4 372; Melton Tr. 2351-52; RX-269-F.

163 coe CPF Yf 1840-46.

Opinion 120 F.T.C.

We have found that offset capacity does constrain the price of gravure printing. See Section IV, supra. In any event, the existence of excess capacity is not necessary to a finding that diversion is likely.'“ The potential for diversion is not dispositive.'® Relative to nondiscriminatory coordination, however, the existence of readily available supracompetitive returns from shifting sales from one group of customers to another creates relatively stronger incentives to divert production away from customers for whom the return is lower. 2. Characteristics of Buyers and Transactions Since the allegedly inelastic printing jobs, by definition encompass very large contracts, publication printers would have substantial incentives to divert production from the elastic print jobs to obtain supracompetitive returns. The ALJ found that the probability of coordinated interaction is limited by the size of buyers with an ability, and demonstrated willingness, to switch suppliers. IDF J 410.'® We affirm this finding and expand on it. Other things equal, the fewer the number of major buyers and the larger the size of single transactions in a relevant market, the less likely it is that coordination among the sellers will succeed. The larger the payoff from each discrete instance of cheating, the more likely that firms will cheat.

The bigger a buyer is, the more easily and lucratively a member of the cartel can cheat on his fellows; for with a single transaction, he may be able to increase his 164 "[P]redictions about the size of a firm's probable supply response require an evaluation of both the technical feasibility of making supply adjustments and the economic opportunity cost of doing so." ABA Merger Monograph, supra note 44, at 160-61 (citations omitted). For a firm considering increasing its output to inelastic end uses, the opportunity coast of using excess capacity will be lower than the opportunity cost of using capacity currently used to supply the elastic uses. Nevertheless, assuming it is practically feasible, a profit-maximizing firm will shift output to the inelastic end-uses in response to a supra-competitive price increase even if its capacity is fully utilized prior to the price increase.

165 The import of the percentages can easily be overstated. After all, the Commission's simple diversion calculation would show that a nondiscriminatory collusive price increase would be undermined by a "small" percentage increase in output. For example, a 5% price increase would be generated by something less than a 5% output reduction and defeated by an equal output increase. In fact, any nontrivial output increase by one of the coordinating sellers would tend to destabilize the coordination and lead to output increases by other firms seeking to avoid loss of sales to sellers deviating from the terms of the coordination.

166 Id. ("There have been several post-acquisition instances where print buyers have qualified additional gravure printers besides Donnelley or have switched substantial quantities of their printing to other gravure printers (RPF 268, Table F)"). R.R. DONNELLEY & SONS CO., ET AL. 191 36 Opinion sales and hence profits dramatically. But with all the members thus vying for the large orders of big buyers, the cartel will erode. Hospital Corp. of America v. FTC, 807 F.2d 1381, 1391 (7th Cir. 1986), cert. denied, 481 U.S. 1038 (1987), citing George Stigler, A Theory of Oligopoly, in THE ORGANIZATION OF INDUSTRY 38 (1968); FTC v. Elders Grain, Inc., 868 F.2d 901, 905 (7th Cir. 198- 9); Owens-Illinois, slip op. at 32 ("As buyer concentration . . increases, the benefits from cheating to capture a customer's business increase relative to the magnitude of gains from collusion").'®’ By its terms, complaint counsel's proposed relevant market includes only the largest print jobs in publication printing. Indeed, according to the theory of anticompetitive effects, the customers with the largest print programs are the "most vulnerable to a price increase." CAB at 67. Many of these printing programs generate revenues for printers in excess of $10 million annually, and some as high as $100 million annually.’ Each high volume printing program represents substantial incremental revenue for any one printer. Thus, large contracts are likely to attract gravure printers to cheat on the terms of any coordination. It follows that coordination in the "core" the category of the largest and ostensibly most vulnerable printing jobs -- appears to be particularly vulnerable to cheating.'® Large high volume print buyers often use procurement techniques designed to ensure negotiating leverage vis-a-vis printers and to undermine coordination. For example, most large buyers solicit and obtain multiple bids and privately negotiate line items based on the best of the bids.’ Such techniques increase uncertainty in See also Richard Posner, ANTITRUST LAW: AN ECONOMIC PERSPECTIVE 53-54 (1976) (noting that large buyers and large transactions also make it more difficult to detect deviations from coordination, since a substantial loss of sales may be attributable to random defections rather than seller conduct that should be punished).

168 tien Tr. 1573. See generally IDF Y 59-121. 169 At the time of the acquisition, four print buyers accounted for 60 to 80 percent of purchases in the "core." Complaint counsel estimate that this amount is approximately 60%, although the derivation of this figure is not clear. CAB at A-8, citing CX-1374-B. Respondents estimate that the same four buyers account for nearly 80 percent of the "core." RAB at 45, fn. 38 (comparing RX 665 with CX 1167). Buyer concentration in the core appears to have increased since that time, as firms have “exited” the bore (through substitution to offset or actual exit). IDF J 363-64. At the time of the acquisition, complaint counsel's broad putative market comprised approximately 100 purchasers with approximately 1,000 print jobs, and the putative "core" comprised 36 purchasers with approximately 200 jobs. CPF J 2011; RX-665. Dr. Hausman found that increased versioning, reduced print programs, and elimination of print programs, among other things, substantially reduced the number of firms in the “core” to as few as nine (including the buyers about whom he lacked information). IDF {ff 364-65. These are precisely the buyers who are best positioned to induce deviations from coordination. 170 TOF q 43; RPF J 261.

Opinion 120 F.T.C.

coordinating firms and make detection and punishment of deviations from coordination more difficult. See B.P. Goodrich Co., 100 FTC at 325. More importantly, large purchasers commonly offer longterm contracts to printers. IDF J{ 42, 375.'" Common use of longterm contracts on large transactions increases the potential gains from deviating from coordinated interaction. The longer the term of a printing supply agreement, the greater are the revenues available from a single deviation, and thus the stronger the incentives to deviate, other things equal. See Merger Guidelines Section 2.12.'” The use of long-term contracting and other sophisticated procurement techniques may be expected to increase in response to any anticompetitive conduct by high volume publication printers.!” 3. Other Impediments to Successful Coordination Compounding these fundamental weaknesses in complaint counsel's theory of coordinated effects are a variety of other significant obstacles to coordination in high volume publication printing. First, coordinating printers could not likely agree on the identity of inelastic users and, thus, could not tacitly identify the margin. RRB at 73. Although we have assumed to this point in the competitive effects analysis that this problem does not exist, it is, in fact, an imposing problem for the hypothetical coordination among publication printers. See Section IV, supra. Second, printers have varying cost structures and vary in other important ways that undermine the probability of achieving a consensus on the terms of coordination. See B.F. Goodrich Co., 110 FTC 207, 321 (1988). As complaint counsel concede, even if the roster of relevant participants is limited to gravure printers, such firms differ in a variety of ways.'“ See generally IDF J] 340-57. al . . . oe CPF ¥ 2014-15. Many of these contracts provide the buyer with the right to audit price increases. CPF J 2016; RPF { 261-62.

172 Section 2.12 of the Merger Guidelines states that long-term contracts will induce deviations only “where the duration, volume and profitability of the business covered by such contracts are sufficiently large as to make deviation more profitable in the long term than honoring the terms of coordination, and buyers likely would switch suppliers." This contemplates incentives to cheat in the aggregate. More precisely, a particular seller is more likely to deviate to obtain a particular contract the greater are the revenues from the contract relative to the seller's total revenues. 173 ws . .

In addition, any firms that are under long-term contracts with price and other terms fixed are at least partially protected from post-acquisition anticompetitive effects for the term of the contract. Complaint counsel state that "gravure printers differ in terms of capacity, utilization, efficiency of operation, press capabilities, service, distribution and sales, reputation and track record, and reputation for quality.” CAB at 71. Complaint counsel's industry expert testified to the differences in cost structure. Hodgson Tr. 367-69.

R.R. DONNELLEY & SONS CO., ET AL. 193 36 Opinion When the hypothetical coordination in the assumed market is expanded to account for the conduct of many offset printers who currently supply the market, the asymmetries among firms are even more pronounced. The different cost structures of gravure and offset printing would further confound any attempt to reach consensus.!”° Third, because high volume printing jobs are performed on a custom basis, relevant suppliers would be required to achieve consensus on a great number of variables and would have multiple Opportunities to cheat by shading on hidden variables.'”° Complaint counsel respond by positing a simple customer allocation scheme that ostensibly obviates consensus on multiple variables. CAB at 67-68. It is not clear, however, that all firms would maximize profits by settling on the current allocation -- for example, firms that have recently expanded capacity likely would not.'” In any event, assuming customer allocation is the most effective form of coordination among printers, it entails the essential incentives to deviate common to any form of coordinated discrimination against high volume publication printing customers, as discussed in Section VI.A.1.2., supra.

These market conditions, taken together, indicate that the merger is very unlikely to give rise to coordinated, discriminatory anticompetitive effects in high volume publication printing. At the same time, the differences among relevant suppliers, in conjunction with Donnelley's post-acquisition share, suggest the possibility of unilateral anticompetitive effects.

B. Unilateral Effects Analysis "A merger may diminish competition even if it does not lead to increased likelihood of coordinated interaction, because merging firms may find it profitable to alter their behavior unilaterally following the acquisition by elevating price and suppressing output." Merger Guidelines Section 2.2. An individual firm has "unilateral" market power if it can raise price above the competitive level without inducing customers to reduce their purchases to a degree that makes TS See Section IV.D.1., supra.

176 see CAB at 69; RRB at 73, 78-79.

177 . se . .

See note 162, supra, and accompanying text (regarding increases in capacity). Opinion 120 F.T.C.

the price increase unprofitable.’ A merger yields unilateral anticompetitive effects if it permits the merged firm to impose a significant and nontransitory price increase without inducing so much substitution to other relevant suppliers that the price increase becomes unprofitable.'” A merger may facilitate a unilateral exercise of market power in two general market "settings": where firms are "distinguished primarily by differentiated products," Merger Guidelines Section 2.21, and where firms are "distinguished primarily by their capacities." Jd. Section 2.22.'*° Complaint counsel allege that both settings are applicable to this case and that the acquisition ‘likely will permit the merged firm unilaterally to increase prices. Without specifying the prerequisite market setting, the ALJ held that "Donnelley's post-acquisition market share suggests that it can unilaterally raise prices to some high volume publication gravure print customers, restrict output or engage in other anticompetitive conduct." ID. at 90.'8! We reject the ALJ's conclusions and find that the acquisition is unlikely to give rise to unilateral anticompetitive effects under any theory.

Merger Guidelines Section 0.1. ("Circumstances also may permit a single firm, not a monopolist, to exercise market power through unilateral or non-coordinated conduct -- conduct the success of which does not rely on the concurrence of other firms in the market or on coordinated responses by those firms."); Robert D. Willig, Merger Analysis, Industrial Organization Theory, and Merger Guidelines, BROOKINGS PAPERS ON ECONOMIC ACTIVITY (MICROECONOMICS) 281, 293 (1991) ("A unilateral effect would arise, in contrast, when a merger between sellers of close substitutes impels them to raise prices profitably whether or not rivals in fact follow"). See also William Landes and Richard Posner, Market Power in Antitrust Cases, 94 HARV. L. REV. 937 (1981). 179 See generally Roscoe B. Starek WI and Stephen Stockum, What Makes Mergers Anticompetitive? "Unilateral Effects" Analysis Under the 1992 Merger Guidelines, 63 ANTITRUST L. J. 801 (1995).

180 Although Section 2.2 of the Merger Guidelines does not limit itself to a particular theoretical economic model of competitive behavior, the general distinction between market settings appears to be based on the distinction between two general models: the differentiated product Bertrand model and the homogeneous product Cournot model. See generally Willig, supra note 178, at 292-93. In the differentiated products Bertrand model, a merger between two firms that produce close substitutes in a market of differentiated products will generate an increase in market power. In the Cournot model, a merger will generate market power if nonmerging rivals would not find it profitable to expand output by an amount sufficient to offset the output reductions of the merged firm. See, e.g., Martin K. Perry and Robert H. Porter, Oligopoly and the Incentive for Horizontal Merger, 75 AM. ECON. REV. 219 (1985); Joseph Farrell and Carl Shapiro, Horizontal Mergers: An Equilibrium Analysis, 830 AM. ECON. REV. 107 (1990); see also Willig, supra note 178, at 295-98. 181 The ALJ based this holding primarily on the basis of Donnelley's post-acquisition market share, "admissions" that Donnelley could increase prices following the acquisition, and testimony from customers stating "concerns" about the effect of the acquisition. ID. at 90; IDF [J 393-405. R.R. DONNELLEY & SONS CO., ET AL. 195 36 Opinion 1. Potential Reduction in Localized Competition Where different products within a market differ in the degree of their substitutability for one another, competition can be "localized" so that a seller competes more directly with those firms selling relatively close substitutes. Similarly, competition may be localized in markets where sellers are differentiated by their relative advantages in serving different groups of buyers, and buyers negotiate individually with sellers.' In the context of localized competition, unilateral anticompetitive effects may result from (i) a merger between rival sellers that produce relatively close substitutes in a market for differentiated products and (ii) a merger between rival sellers that have similar advantages in serving a particular group of buyers. See Merger Guidelines Section 2.21. Such a merger may enable the merged firm unilaterally to increase prices above premerger levels because some of the sales lost by one firm due to the price increase will be diverted to the other firm. "[C]apturing such sales loss through merger may make the price increase profitable even though it would not have been profitable premerger." Jd. Whether the merger facilitates a unilateral exercise of market power in this setting depends on the “closeness of the products of the merging firms" and the "ability of rival sellers to replace lost competition." Jd.

Substantial unilateral price elevation in a market for differentiated products requires that there be a significant share of sales in the market accounted for by consumers who regard the products of the merging firms as their first and second choices, and that repositioning of the non-parties' product lines to replace the localized competition lost through the merger be unlikely. Id. See Coca-Cola Bottling Co. of the Southwest, 118 FTC 452 [FTC Dkt. No. 9215 (Aug. 31, 1994), slip op.]; State of New York v. Kraft General Foods, Inc., 1995-1 Trade Cas. (CCH) § 70,911 (S.D.N.Y. 1995). In other words, to show that the acquisition facilitated a unilateral exercise of market power, complaint counsel must demonstrate that (i) customers of the two merging firms have relatively inelastic demand for a particular type or quality of printing services that is currently provided only by the merging firms, and (ii) the supply elasticity of other relevant printers (gravure and offset) is See Merger Guidelines Section 2.21 n. 21.

Opinion 120 F.T.C.

insufficient to meet the demand of such buyers that would otherwise switch to a repositioned printer.'® Economic analysis indicates that the perceived substitutability between two firms’ products (i.e., their "closeness") is the primary factor determining the market power that will be created by a merger in a differentiated product setting, and that market concentration plays a lesser role.'* The closeness of the merging firms' products has a critical effect on the profitability of a post-merger price increase because the more closely substitutable are two products (relative to their substitutability with other products), the greater is the degree to which substitution away from each of the products of the merging firms due to a price increase will be "internalized" by the merged firm. Therefore, information that directly reflects customers’ actual preferences is more probative than market concentration data in assessing the relative substitutability of differentiated products.'® Nevertheless, market concentration can sometimes act to reinforce other evidence bearing on the closeness of the products of the merging firms:

Where market concentration data fall outside of the safeharbor regions of Section 1.5, the merging firms have a combined market share of at least thirty-five percent, and where data on product attributes and relative product appeal show that a significant share of purchasers of one merging firm's product regard the other as their second choice, then market share data may be relied upon to demonstrate that there is a significant share of sales in the market accounted for by consumers who would be adversely affected by the merger. Merger Guidelines Section 2.211.

Although respondents argue that printing services of the different gravure printers are not differentiated,'®° it appears that printers may be differentiated in their relative advantages in dealing with particular customers.'®’ Printers appear to be distinguished in a variety of See Landes and Posner, supra note 178.

184 see Willig, supra note 178, at 300-01. 185 Id. at 301 (The concentration presumption is "unlikely to be valid in many areas of application where specific information can be developed about product characteristics and about consumer preferences for them. For such applications, merger analysis that focuses exclusively on market shares is likely to go awry.").

86 RAB at 51-52.

187 ce CPF] 1737 ("Each print job is unique and firms are perceived to have differing capabilities to produce a job to a customer's satisfaction"); CPF J 1738 ("Suppliers of printing services have differing equipment capabilities and differing reputations for quality, service and reliability"); RAB at 56. Printing firms, differing abilities to serve given customers or categories of customers may be based not on any particular aspect of the service, but on the totality of their printing operations -- product R.R. DONNELLEY & SONS CO., ET AL. 197 36 Opinion dimensions, and the merging firms had many common characteristics prior to the merger. IDP {J 338-357. Donnelley certainly perceived Meredith/Burda as one of its most significant competitors.'® Moreover, many customers testified that they "ranked Donnelley and Meredith/Burda as the highest quality gravure printers, and viewed them as vigorous competitors prior to the acquisition." IDF { 405; see IDF ¥ 411.'° The concerns expressed by large customers of the merging firms may reflect their belief that. switching to alternative printers would be difficult. IDF [ 415. Where such evidence suggests that market shares may reflect each firm's relative appeal as the first and second choices of their current customers, the merged firm's substantial share of the assumed relevant market may give rise to a presumption that the firms are particularly close substitutes for a significant share of customers. Merger Guidelines Section 2.211. Here, the structural presumption is very weak. Many customers view the printing services of other printers to be good substitutes for the services of the merging firms and did not consider the merging firms to be the first and second choices for quality and service.’ And there appear to be no general, objective criteria by which buyers with relatively inelastic demand for the services of the merged firm can be identified.'®' The best objective evidence that the level of differentiation between printers is not significant -- and that other printers are reasonable substitutes that constrain the merging firms -is that each of the other gravure printers, and many offset printers, are quality, technical expertise, distribution capabilities, bindery capacity, geographic proximity, and other aspects of the relevant service.

188 see generally CPF ff 1770-1815.

189 There is some danger in relying on these customer complaints to draw any general conclusions about the likely effects of the acquisition or about the-analytical premises for those conclusions. The complaints are consistent with a variety of effects, and many -- including those the ALJ relied upon -directly contradict complaint counsel's prediction of unilateral price elevation. In fact, some suggested that the effect would be the opposite. See, e.g., Deutsch Tr. 961-62 (predicting that the merged firm would reduce prices following the acquisition "and keep them low enough to drive some of their competition out of business"); Bentele Tr. 1439-40 (concerned that Donnelley would set post-acquisition prices at a level that would not allow other printers "to exist in the marketplace"). 190 Henry Tr. 718; Gallo Tr. 829, 838; Owens Tr. 1284-85; Allen Tr. 1577-78; see RX-396. Many customers did not consider Donnelley to provide the kind of quality and service that would make it a closer substitute for Meredith/Burda than were other firms. See, e.g., CPF ff 1873, 1877-78, 1880. The concerns expressed by these customers may be interpreted as the result of a change in management rather than as a change in the nature or level of competition. 191 As discussed previously, the volume, versions, and page parameters of the assumed relevant market do not yield accurate predictions about demand elasticities for gravure generally. Moreover, they do not provide a basis for estimating demand elasticity for the services of the merging firms -- all gravure printers supply printing services that meet the parameters. Therefore, the criteria by which inelastic customers are identified must be different from the alleged market parameters. Opinion 120 F.T.C.

currently selling into the alleged inelastic uses.” Among the "core" customers complaint counsel identifies as most susceptible to anticompetitive effects, Quebecor's share was substantially higher than Meredith/Burda's prior to the acquisition, and seven firms had a premerger market share of approximately seven percent or higher.'* Each of the gravure printers other than the merged firm is currently supplying printing services to customers who have expressed "concerns" about the acquisitions.'™ Even assuming that printers are sufficiently differentiated in their abilities to perform work for groups of customers, and that customers who regarded the merging firms as their first and second choices prior to the merger comprise a significant share of sales in the market, unilateral effects are not likely. In high volume publication printing, incumbent printers "likely would replace any localized competition lost through the merger by repositioning their product lines." Merger Guidelines Section 2.212.'% Each of the incumbent gravure printers is currently supplying high volume publication gravure printing,'” implying that the customers for whom the merging firms were first and second choices represent a minor portion of a proposed relevant market that already represented a minor portion of total gravure capacity. The capacity available for diversion from elastic customers to alleged inelastic customers is similar to that in the coordinated effects scenario.'”’ The other printers would have strong incentives to increase sales to the alleged inelastic customers because, by definition, they obtain no benefit from the merged firm's unilateral anticompetitive conduct. Thus, as long as their supply is relatively elastic, these firms should be expected to reposition with alacrity. "In markets where it is costly to evaluate product quality, buyers who consider purchasing from both merging parties may limit the total number of sellers they consider." Merger Guidelines Section See Section V.B. (Tables 1 and 2), supra.

193 see Id. (Table 2).

194 See, e.g., RAB at 53.

195 Willig, supra note 178, at 304 (Absent this important consideration, as Willig notes, the analysis would include an "implicit assumption that the pattern of demand relationships and products’ characteristics are not subject to endogenous change. Although this may be an accurate description in many contexts, others firms may be readily and quickly able to reposition their products in response to market incentives.").

ior CAB at 70-71.

7 See Section VI.A.L, supra. In the unilateral effects calculus, the merged firm's capacity is removed from the denominator, and high volume firms that did not view the merging firms as first and second choices are removed from the numerator. See Hausman Tr. 5450-58; RX-397. R.R. DONNELLEY & SONS CO., ET AL. 199 36 Opinion 2.212.'% High volume publication printing meets this description. Contracts are complex, purchase decisions are multidimensional, and buyers often limit the number of potential suppliers through a qualifying process. IDF J 43. Under these circumstances, "[i]f either of the merging firms would be replaced in such buyers' consideration by an equally competitive seller not formerly considered, then the merger is not likely to lead to a unilateral price elevation." Merger Guidelines Section 2.212. Substantial evidence has been presented that many buyers would incur non-trivial costs of qualifying an additional printer to replace one of the merged firms in their bidding consideration. IDF { 411-15. The actions of relevant buyers, however, belie the competitive significance of these "switching costs." Following Donnelley's acquisition of Meredith/Burda, several customers have shifted substantial quantities of business from the merged firm to other gravure printers (as well as to offset printers); and many buyers have replaced one of the merging firms by qualifying other gravure printers to bid on and supply their high volume publication printing services. IDF { 410.'” The merged firm's competitors have actively sought to expand their sales to high volume publication customers, with observable success.” Under these circumstances, unilateral anticompetitive effects through the loss of localized competition are unlikely. 2. Unilateral Effects Under Capacity Constraints Unilateral anticompetitive effects may also result from horizontal mergers in markets in which products are "relatively undifferentiated."""' In these markets, where "capacity primarily distinguishes firms and shapes the nature of their competition," merging firms with a high combined market share may find it profitable unilaterally to raise price and suppress output after merger: In such markets, rivals of the merging firms may need to induce customers to incur whatever costs are necessary to evaluate a repositioned product. See Section IV.E., supra.

See, e.g., RX-99-A (Brown's efforts to solicit Meredith/Burda accounts immediately following the merger).

201 Because the model of competitive effects in this Section assumes that firms are not differentiated by their ability to serve particular groups of customers, it is mutually exclusive with the model described in Section V.B.I., supra. Nevertheless, complaint counsel have essentially argued these models in the alternative.

Opinion 120 F.T.C.

The merger provides the merged firm a larger base of sales on which to enjoy the resulting price rise and also eliminates a competitor to which customers otherwise would have diverted their sales. ... [MJerged firms [with a high market share] may find it profitable to raise price and reduce joint output below the sum of their premerger outputs because the lost markups on the foregone sales may be outweighed by the resulting price increase on the merged base of sales. Merger Guidelines Section 2.22. Under the Merger Guidelines, the Commission recognizes that a combined share of greater than thirtyfive percent raises the possibility of such effects. Id.?” Although the merger may have given Donnelley the requisite market share for further analysis, the record establishes that nonmerging firms would be able economically to respond to the merged firm's price increase "with increases in their own outputs sufficient in the aggregate to make the unilateral action of the merged firm unprofitable." Jd. Assuming that the services of other gravure printers are relatively undifferentiated, anticompetitive effects are unlikely because those other gravure printers are able quickly and easily to expand their output to the category of customers allegedly subject to discrimination by the merged firm.2% As discussed previously, the merged firm's gravure and offset competitors are under no binding capacity constraint that would keep them from expanding output to high volume publication printing.” Indeed, each relevant printer has substantial existing capacity that currently is not used to supply the assumed relevant market but could be used in response to unilateral anticompetitive conduct by the merged firm.”” To undermine a discriminatory unilateral price increase by the merged firm, each other supplier needs only to shift output from the elastic to the allegedly inelastic print jobs; it need not have excess 202 The logic of this analysis does not depend on a particular market share. Only if a consumer has limited opportunities to substitute will he be willing to pay an anticompetitive price. If customers have access to suppliers that are able to supply them with a relatively undifferentiated product, then the market share of these firms is of limited significance to the effect of potential substitution on a firm's market power. The importance of market share in this type of industry is only its potential reflection of constraints on firms’ productive capacities. See Willig, supra note 181, at 295. Ifa firm has a two percent share of a homogeneous good market largely because it can only produce additional output at substantially higher marginal cost, its economic significance is very different than if it can expand its production at relatively constant marginal cost. See Landes & Posner, supra note 178, at 945. See also United States v. General Dynamics Corp., 415 U.S. 486 (1974). 203 Under the standards of the Merger Guidelines, if the firm can expand output within one year without significantly increasing costs, then its expansion can be presumed to constrain a unilateral price increase by the merged firm. /d; see id. Section 1.3 (uncommitted entry). 204 oe Sections V.B., VI.A.L, and VI.B.1., supra. 205 Id.

R.R. DONNELLEY & SONS CO., ET AL. 201 36 Opinion printing capacity overall, and it need not expand its capacity. Moreover, the merged firm's gravure competitors have actively engaged in capacity expansions under competitive conditions, adding more efficient capacity that could easily be used to supply customers against whom the merged firm attempts to impose a discriminatory price.”

C. Deferral of Capacity Expansion Plans The ALJ's finding that Donnelley's post-acquisition deferral or cancellation of capacity expansion plans was anticompetitive describes a theory of competitive effects that appears to be, at best, only casually related to the theory of discriminatory anticompetitive effects in high volume publication printing. Prior to the acquisition, Donnelley and Meredith/Burda independently had given various levels of consideration to purchasing and installing gravure presses.”” Following the acquisition, Donnelley deferred these considerations "so that the firms could be integrated and a complete assessment of any operating efficiencies could be made." RAB at 62. For much of the capacity that was considered, the deferrals appear to have become cancellations.’ Respondents state that these cancellations or deferrals were the result of the onset of economic recession immediately following the acquisition and the significant decrease in the demand for gravure printing services. RAB at 62. Complaint counsel allege that the cancellation or deferral of "planned" capacity expansions is, by itself, an anticompetitive effect of the acquisition. CAB at 50-51. The ALJ found that "the cancellation or deferral of these expansion plans had a substantial adverse competitive effect because had they gone forward, significant gravure capacity would have been added to the market at or about the time of the acquisition and this would have resulted in an increase in gravure supply and a reduction in prices." IDF J 381; see ID at 89.90.7° The ALJ's finding is based in large part on evidence that Donnelley believed that, by acquiring Meredith/Burda, it could increase its own capacity and could thereby avoid investing in See note 162, supra, and accompanying text. 207 cee IDF 1380; RAB 61-62.

208 See, e.g., Moeller Tr. 4124 (Donnelley now appears to have no plans to order a press that it had planned, prior to the acquisition of Meredith/Burda, to install at its Reno facility). 209 See Hilke Tr. 3347-54, 3359 (Donnelley's acquisition of capacity rather than expansion “represents an alternative which involves higher prices and less competition"). Opinion 120 F.T.C.

internal expansion. IDF {J 381-88. Such evidence likely would exist regarding a vast number of horizontal acquisitions for which no other theory of anticompetitive effects could be articulated.”"° Capacity expansion deferrals (or cancellations) likely occur in connection with many procompetitive acquisition. It will frequently be the case that a firm that buys existing assets from another firm also considered, as a possible alternative, the creation of new assets, but opted for the former course of action as the least costly alternative. In such cases, the decision to acquire existing assets will frequently (if not usually) cause the acquirer to decide against the contemplated new capital expenditures (e.g., the firm only needs one new factory, not two). Without more, we cannot infer that the cancellation was anticompetitive. A firm's decision to forego internal expansion in favor of acquisition may be anticompetitive only if the merger creates or enhances unilateral market power.”!! Respondents concede that a cognizable anticompetitive effect based on allegations of this type could be found under an analysis similar to an actual potential competition analysis: (i) Would the expansions likely have taken place absent the acquisition? If so, (ii) Would the expansions have increased competition (reduced qualityadjusted prices) in the alleged relevant market relative to what exists today??? First, the Commission would need to find that the capacity expansions were reasonably probable in a timely manner. Here, however, some of the plans appear to have been inchoate.”!? It also appears that Donnelley completed some capacity expansions after a deferral following the acquisition.’'* See United States v. Amax, Inc., 402 F. Supp. 956, 959 (D. Conn. 1975) (the argument that defendant's acquisition is anticompetitive because it could have expanded internally rather than through acquisition is an argument that "can be made, at least in theory against any horizontal merger, and Section 7 of the Clayton Act has not been interpreted to outlaw all such mergers."). 211 cy Robert H. Bork, THE ANTITRUST PARADOX 206-08 (1978) (internal expansion is preferable to merger only where the merger would create unilateral market power in the merged firm; otherwise, "we must assume that the firm makes the choice between internal expansion and merger on the basis of the relative costs of the two routes to larger size”). See RAB at 62; CAB at 50 n. 70 (conceding standard and arguing that Donnelley gained unilateral markei power through the acquisition). 712 eee, e.g. B.A.T. Indus., Lid., 104 FTC 852, 930 (1984). 213 For example, the record describes Meredith/Burda as "considering" and "contemplating” an expansion at its Lynchburg facility, and taking only preliminary steps toward such an expansion. See CPF J 1597-1627. There appears to be no evidence that Meredith/Burda's board or general management ever approved these expansion plans. See, e.g., RX-280 (Meredith/Burda's capital budgets prior to the acquisition did not reflect plans for expansion at Lynchburg). 214 For example, Donnelley appears to have proceeded with expansion at its Warsaw facility after deferring the projected date of completion from the fall of 1991 to the fall of 1994. CPF ] 1565. R.R. DONNELLEY & SONS CO., ET AL. 203 36 Opinion Second, assuming Donnelley cancelled certain planned expansions, it appears unlikely that the cancellations were the result of a unilateral exercise of market power.?'° To prove that the actions constituted a unilateral exercise of market power resulting from the acquisition, we would be required to conclude that the merging firms are uniquely able to expand capacity to the extent allegedly cancelled. If other firms are equally well-positioned to expand capacity, or could quickly and economically expand capacity in response to a supracompetitive price increase, Donnelley's actions would not result in a reduction in capacity relative to the hypothetical world that would have existed "but for" the acquisition. Clearly, other firms have expanded gravure capacity since the acquisition.”!® And the record does not indicate the extent to which these other firms would have expanded capacity if Donnelley had completed the expansions allegedly considered.”””

More importantly, the relevant capacity expansion involves not capacity to produce gravure printing but rather capacity to produce high volume publication printing. As discussed throughout this opinion, the only relevant market alleged by complaint counsel comprises a small share of total printing capacity. Thus, the relevant measure of available capacity under this theory is the capacity to produce for the high volume customers whose demand is allegedly inelastic. Most of the "cancelled" capacity would have been used in printing jobs for which there was no threat of post-acquisition anticompetitive conduct. The assumed relevant market consumes substantially less than one-third of total gravure capacity, and an even lesser share of gravure-plus-offset capacity. Hence, complaint counsel's theory requires that Donnelley would forego the returns it would have earned on the capacity's competitive uses (e.g., in lower volume publications) -- returns which constitute the major share of the total returns on the assets -- in order to earn supracompetitive returns on activities which constitute a minor portion of the assets’ total uses. Even if the supracompetitive returns in the relevant market were assured to the merged firm, the conditions necessary for The cancellation would also be anticompetitive if it had been coordinated with the other gravure printers. Complaint counsel do not allege such a coordinated action. See note 162 supra, and accompanying text. 7 Other firms may not have proceeded with their expansion plans if Donnelley had completed all of the expansions allegedly planned. There is no evidence in the record allowing the Commission to predict the extent to which the acquisition affected other firms, plans for expansion. Therefore, we cannot assess the likely effect of the acquisition on total industry capacity or output. Concurring Opinion 120 F.T.C.

this tradeoff to make economic sense seem implausible. The monopoly profits in the putative relevant market would have to be substantial.

In any event, we have concluded that the merger did not create or enhance unilateral market power in Donnelley even in the narrow price discrimination market. Given the conclusions of the relevant market analysis and the competitive effects analysis above, the record does not reveal that the cancellation of the premerger plans for expansion was anticompetitive.”

VII. CONCLUSION The complaint's proposed market definition does not sufficiently account for actual and potential substitution and is therefore too narrowly drawn. Even assuming the relevance of a high volume publication printing market, there is no theory of anticompetitive effects that withstands scrutiny. The complaint is dismissed for failure to prove that the acquisition is likely substantially to reduce competition in a relevant market.

CONCURRING OPINION OF COMMISSIONER MARY L. AZCUENAGA I concur in the decision of the Commission to dismiss the complaint. To support the product market alleged in the complaint, "high volume publication gravure printing," complaint counsel attempted to show that a discriminatory price increase could be imposed on customers who purchase high volume gravure printing services. Many such customers do not regard offset printing as a substitute for gravure printing. In defining the product market using the analytical approach set forth in the 1992 Horizontal Merger Guidelines, the question is whether enough customers would switch from high volume gravure printing to offset printing to defeat a price increase. In my view, the opinion of the Commission understates the strength of complaint counsel's case in support of the product market. Nevertheless, the evidence of actual switching from high volume gravure to offset printing is sufficient that I am unable to conclude that the weight of the evidence supports the proposed product market. Not having found a relevant market in which to assess competitive A similar theory was proposed and rejected in Owens-Illinois [IDF ¥ 10], slip op. R.R. DONNELLEY & SONS CO., ET AL. 205 36 Final Order effects, I would dismiss the complaint, and I do not reach the other issues discussed in the opinion of the Commission. - FINAL ORDER The Commission has heard this matter on the appeal of respondents R.R. Donnelley & Sons Company and Pan Associates, L.P. from the Initial Decision and on briefs and oral argument in support of and in opposition to the appeal. For the reasons stated in the accompanying Opinion, the Commission has determined to grant the appeal.

Accordingly, It is ordered, That the complaint is dismissed. Complaint 120 F.T.C.