General Foods Corporation

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

IN THE MATTER OF GENERAL FOODS CORPORATION DISMISSAL ORDER, OPINION, ETC., IN REGARD TO ALLEGED VIOLATION OF SEC.5 OF THE FEDERAL TRADE COMMISSION ACT AND SEC. 2(a) OF THE CLAYTON ACT Docket 9085. Complaint, July 197 Final Order, April, 1984 In this Final Order, the Commission denied complaint counsel's appeal against an ALJ' s Initial Decision dated Jan. 25, 1982, sustained the ID for reasons set forth in its accompanying Opinion, and dismissed the Complaint charging a leading processed food manufacturer with violating Section 5 of the Federal Trade Commission Act and Section 2(a) of the Clayton Act, in connection with the marketing of Regular Maxwell House Coffee.

Appearances For the Commission: Edward J Carnot, Michael Goldenberg, Jane Seymour, Wayne Kaplan and Richard F. Silvestri. For the respondent: Paul C. Warnke, John F. Kovin, John G. Calender, Robert P. Reznick and Don Scott DeAmicis, Clifford Warnke Washington, D.C. and Robert Y. Fox in-house counsel, White Plains COMPLAINT The Federal Trade Commission, having reason to believe that General Foods Corporation is in violation of Section 5 ofthe Federal Trade Commission Act (15 C. 45) and Section 2(a) of the Clayton Act, as amended by the Robinson-Patman Act (15 C. 13(a)), and believing that a proceeding by it in respect thereof is in the public interest hereby issues its Complaint charging as follows: DEFINITIONS 1. For the purpose of this Complaint, the following definitions shall apply:

(a) Regular coffee is coffee processed from unroasted coffee beans by means of blending, roasting, and grinding into varying granular sizes and which must be heated and steeped in water before being consumed. It is generally packed and sold in vacuum tin containers and paper bags. It is to be distinguished from soluble (instant) coffee. 204 Complaint (b) Consumer promotions are discounts or other offerings to the consumer to encourage him to purchase the brand promoted. (c) Trade promotions are discounts or other financial incentives to the grocer or wholesaler. (2) RESPONDENT 2. General Foods Corporation (hereinafter "General Foods ) is a Delaware corporation with its executive offce located at 250 North Street, White Plains, New York.

3. Coffee is one of General Foods' principal product lines. Coffee sales make up approximately 29% of the total revenues of General Foods and regular coffee constitutes a substantial portion of such coffee sales.

4. In fiscal year 1975 General Foods had total sales of approximately $3.7 bilion, in fiscal year 1974, $3.0 bilion, in fiscal year 1973, $2. bilion, in fiscal year 1972, $2.4 bilion and in fiscal year 1971 , $2. bilion.

5. The Maxwell House Division of General Foods is principally responsible for the manufacture, distribution and sale of coffee. Sales volume for the Maxwell House Division approximated $900 milion in fiscal year 1974, $763 milion in fiscal year 1973, $703 milion in fiscal year 1972 and $662 million in fiscal year 1971. 6. The relevant market is the production, distribution and sale of regular coffee packaged for sale at retail in Maxwell House Division eastern region. Relevant geographic submarkets include the Cleveland and Pittsburgh market areas of Maxwell House Division Youngstown, Ohio District, the Wilkes Barre and Philadelphia market areas of Maxwell House Division s Philadelphia, Pennsylvania District and each market area of Maxwell House Division s Syracuse, New York District.

7. Concentration of sales for regular coffee packaged for sale at retail in the United States has increased from approximately 66% for the four largest producers in 1971 to approximately 72% in 1974. 8. General Foods, through its Maxwell House Division, is the largest seller of regular coffee packaged for sale at retail in the United States. It accounted for approximately 31 % of such sales in the United States in 1972, 34% in 1973, and 35% in 1974. (3) 9. The Maxwell House Division of General Foods was the dominant seller of regular coffee packaged for sale at retail in the relevant market and each relevant submarket during the years 1971 through 1975, accounting, in each year, for approximately 45% or more of sales of regular coffee packaged for sale at retail in the relevant market and approximately 50% or more of such sales in the relevant submarkets.

Complaint 103 F.

OFFENSES ALLEGED 10. General Foods, through its Maxwell House Division, has used and is using its dominant position, size and economic power to frustrate the growth of smaller regular coffee producers and to frustrate entry of other regular coffee producers into the relevant market and submarkets and to prevent, hinder or lessen competition in the production, distribution or sale of regular coffee packaged for sale at retail in the relevant market and submarkets. More particularly, General Foods, through its Maxwell House Division, since at least October, 1971, has adopted and placed into effect and carried out various policies, acts, practices, or methods of competition to foreclose entry and to lessen, restrain, eliminate or prevent the production, distribution or sale of regular coffee packaged for sale at retail by others engaged in the production, distribution or sale of such product in each market and submarket identified and alleged herein. Among such policies, acts, practices or methods of competition, General Foods, through' its Maxwell House Division, has engaged in one or more of the following:

(a) Sellng regular coffee packaged for sale at retail below cost or at unreasonably low prices;

(b) Using extensive consumer and trade promotions and advertising to forestall or foreclose market entry or to lessen competition; (4) (c) Engaging in geographically discriminatory pricing and promotional and advertising practices to forestall or foreclose entry or lessen competition;

(d) Foreclosing and deterring new entry by increasing advertising and promotional expenditures in previously established marketing areas of the new entrant in retaliation for entry into the relevant market and submarkets.

(e) Using a fighting brand of regular coffee to forestall or foreclose entry or lessen competition.

EFFECTS 11. The aforesaid policies, acts, or practices have or may have the following effects in one or more ofthe relevant market and submarkets alleged herein, among other things:

(a) Monopolizing the production, distribution or sale of regular coffee packaged for sale at retail;

(b) Increasing entry barriers in the production, distribution or sale of regular coffee packaged for sale at retail; (c) Preserving, maintaining, or furthering highly concentrated market structures;

204 Complaint (d) Hindering, restraining, foreclosing or frustrating competition in the production, distribution or sale of regular coffee packaged for sale at retail;

(e) Depriving consumers of the benefits of free and open competition.

JURISDICTION 12. The policies, acts and practices of General Foods' Maxwell House Division as alleged herein at all times relevant hereto have been in or have affected commerce within the meaning of the Federal Trade Commission Act. (5) 13. In the course and conduct of its business, General Foods' Maxwell House Division is now, and for many years past has been, engaged in commerce, as "commerce" is defined in the Clayton Act, in that it has sold and distributed, and is now sellng and distributing, regular coffee packaged for sale at retail to purchasers thereoflocated in States other than the State of origin of shipments and has, either directly or indirectly, caused such products, when sold, to be shipped and transported from the State of origin to purchasers located in other States, as part of a constant course and flow of trade and commerce in such products between General Foods' Maxwell House Division in the State of origin and purchasers thereof located in other States and the District of Columbia.

General Foods, through its Maxwell House Division, has shipped and sold regular coffee packaged for sale at retail to purchasers with places of business located throughout the several States ofthe United States and the District of Columbia for resale to customers within the United States.

VIOLATIONS 14. In the relevant market and in one or more of the relevant submarkets, General Foods, through its Maxwell House Division, has maintained monopoly power over the production, distribution or sale of regular coffee packaged for sale at retail through all or some of the policies, acts and practices set out and alleged in Paragraph 10 in violation of Section 5 of the Federal Trade Commission Act. 15. In the relevant market and in one or more of the relevant submarkets, General Foods, through its Maxwell House Division, has attempted to monopolize the production, distribution or sale of regular coffee packaged for sale at retail through all or some of the policies, acts and practices set out and alleged in Paragraph 10 in violation of Section 5 of the Federal Trade Commission Act. 16. General Foods, through its Maxwell House Division, has engaged in a course and pattern of conduct which constitutes unfair Initial Decision 103 F. methods of competition or unfair acts or practices in violation of Section 5 of the Federal Trade Commission Act. (6) 17. In the course and conduct of its business in commerce, General Foods, through its Maxwell House Division, has since October 1971 discriminated and is now discriminating in price in the sale of regular coffee packaged for sale at retail by sellng such products oflike grade and quality at different prices to different purchasers. Included in, but not limited to, the discriminations in price as above alleged, respondent has discriminated in price in the sale of said products to retailers and other purchasers in the Cleveland and Pittsburgh market areas of its Youngstown, Ohio District and in the Wilkes Barre and Philadelphia market areas of its Philadelphia Pennsylvania District, and in each market area of its Syracuse, New York District by charging said retailers and other purchasers substantially lower prices than charged by said respondent for the sale of said products of like grade and quality to retailers and other purchasers located in other of respondent' s trading areas throughout the nation, including at times other ofthe market areas described herein. The effect of such discriminations in price by respondent in the sale of regular coffee packaged for sale at retail has been or may be substantially to lessen competition or to tend to create a monopoly in the line of commerce in which said respondent is engaged, or to injure destroy or prevent competition between respondent and its competitors in the production, distribution or sale of such products. The discrimination in price as herein alleged violate subsection 2(a) of the Clayton Act, as amended.

INITIAL DECISION BY LEWIS F. PARKER, ADMINISTRATIVE LAW JUDGE JANUARY 25, 1982 1. HISTORY OF THE PROCEEDING The Commission issued its complaint in this case on July 14, 1976. It charges General Foods Corporation with violations of Section 5 of the Federal Trade Commission Act, 15 U.s.c. 45 and Section 2(a) of the Robinson-Patman Act, 15 U.s. C. 13(a).

The complaint alleges that General Foods, through its Maxwell House Division, has used and is using its dominant position in regular coffee and its size and economic power to frustrate the growth of smaller regular coffee producers, to frustrate the entry of other regular coffee producers into certain relevant markets, and to prevent 201 Initial Decision hinder or lessen competition in the production, distribution or sale of regular coffee packaged for sale at retail in the relevant markets. The complaint claims that General Foods achieved these results by sellng regular coffee below cost or at unreasonably low prices, (2) by using extensive consumer and trade promotions and advertising to forestall or foreclose market entry or to lessen competition, by engaging in geographically discriminatory pricing, promotional and advertising practices, by increasing advertising and promotional expenditures in previously established marketing areas of a new entrant to deter its entry into the relevant markets, and by using a fighting brand to forestall or foreclose entry or lessen competition. According to the complaint, the effects of these practices have been the monopolization of the production, distribution or sale of ground coffee packaged for sale at retail, the increase of entry barriers, the preservation of highly concentrated market structures, the hindering of competition and the deprivation to consumers ofthe benefits offree and open competition.

After extensive discovery, hearings began on August 20, 1979. They ended on June 19, 1981. Over twelve thousand pages of testimony were recorded and several thousand pages of documents were received in evidence. The record was closed on June 26 1981. Complaint counsel fied their proposed findings offact and conclusions oflaw on August 27, 1981. General Foods fied its answer to the findings on October 26, 1981 and complaint counsel fied their reply on December , 1981. At my request, the Commission granted me an extension of time to January 26 , 1982 to fie this initial decision. This decision is based on the transcript of testimony, the exhibits which I received in evidence, and the proposed findings of fact and answers thereto fied by the parties. I have adopted verbatim several findings proposed by complaint counsel and counsel for General Foods. Others have been adopted in substance. All other findings are rejected either because they are not supported by the record or because they are irrelevant.

II. FINDINGS OF FACT A. General Foods Corporation 1. General Foods Corporation is a Delaware corporation with its principal executive offces located at 250 North Street, White Plains New York (Complaint n2, Answer n3). It is one of the world' s leading processors of packaged grocery products and (3) markets a variety of these products in the United States under more than thirty major Initial Decision 103 F. brand names (CX 438B).

2. In fiscal year 1972 (year ending April 1, 1972),2 General Foods had net world-wide sales of over $2.5 bilion and net earnings (from continuing operations) of approximately $113 millon. In fiscal year 1976, General Foods had net world-wide sales of approximately $4 bilion and net earnings of over $150 million (CX 438F). 3. Coffee is one of General Foods' principal product lines (Complaint TI3; Answer TI4). From April 1971 through April 3, 1976, General Foods' domestic coffee sales made up from 26% to 30% of its consolidated net sales (CX 438B). Regular coffee (coffee processed from unroasted coffee beans, as distinguished from soluble (instant) coffee) (Complaint Ul(a); Answer TI2) constituted a substantial portion of such coffee sales (Complaint 113; Answer TI4). The sales volume of the regular coffee products sold by General Foods' Maxwell House Division approximated $380 milion in fiscal year 1971, $386 milion in fiscal year 1972, $433 milion in fiscal year 1973, $504 milion in fiscal year 1974, $529 milion in fiscal year 1975, $567 millon in fiscal year 1976, and $900 millon in fiscal year 1977. In these years, earnings before taxes for all ground coffee products approximated: $26 milion in 1971, $24 million in 1972, $13 milion in 1973, $17 millon in 1974 $5 milion in 1975, $25 millon in 1976, and $51 milion in 1977 (CX 991A-N).

4. General Foods' largest sellng regular coffee brand during the relevant time period was Regular Maxwell House. For the following fiscal years, the sales volume and earnings before taxes (in parenttheses) for that brand approximated: $303 (4) million ($21 milion) in 1971, $293 milion ($19 milion) in 1972, $300 millon ($9 milion) in 1973, $351 milion ($4 milion) in 1974, $369 million (loss of over $1 milion) in 1975, $395 milion ($22 milion) in 1976, and $654 million ($35 million) in 1977 (CX 991A-N).

B. Interstate Commerce 5. In the conduct of its business, General Foods' Maxwell House Division distributes regular coffee packaged for sale at retail to purchasers located in states other than the state in which the shipments originated as part of a constant course of trade and commerce in 1 The following abbrevi"tions are used in I.his decision: Commission Exhibit ex - RX - Respondent's Exhibit Transcript of testimony Tr - CPF - Complaint counsel's proposed findings CLA - Complaint counsel's legal argument RPF - Respondent s proposed findings RLA - Respondent's legal argument CRF - Complaint counsel's reply findings CRLA - Complaint counsel's reply legal argument 2 General Foods' fiscal years run from April through March of the following year (CX 438B) 204 Initial Decision regular coffee between the state of origin and purchasers located in other states and the District of Columbia (Complaint TILE; Answer TILA).

6. General Foods produces its regular coffee products in four plants in the United States: Hoboken, New Jersey; Jacksonville, Florida; Houston, Texas; and San Leandro, California. Its products are shipped from these plants to locations throughout the United States (CX' 387-94; Tr. 5123-25).

C. Regular Coffee Producers 7. Green coffee beans constitute the principal cost component in the production of regular coffee packaged for sale at retail (CX 438B). These beans are all imported into the United States (Tr. 4772). 8. Regular coffee is manufactured by roasting the green beans, grinding and blending different types of roasted beans to achieve certain taste characteristics, and then packing the product in vacuum packed cans or in bags for distribution (Tr. 4828). The manufacture of ground coffee is not a capital intensive industry (Tr. 129). 9. Regular collee is marketed in three pack sizes-one pound, two pound and three pound (Tr. 1627) and comes in different grindsregular, drip, fine, and all purpose, as well as grind or blend variations suited for use in electric percolators and in automatic drip coffee makers. There are decaffeinated regular coffee products, coffee products that contain chicory or other extenders, coffees that are premium blends " and expresso coffees. Coffee is also marketed in premeasured fiter rings or pouches. Recently, manufacturers (5) have introduced increased cup yield collees which deliver increased cup extraction as a result of different roasting and grinding processes (Tr. 4837-49).

10. There are several large firms currently operating in the regular coffee industry in the United States: General Foods Corporation; The Procter & Gamble Co., which markets the Folger s and High Point brands; The Coca-Cola Co., which markets the Maryland Club and Butternut brands; Nabisco Brands, formerly Standard Brands, which markets the Chase & Sanborn brand; The Hills Bros. Coffee Co., Inc. a wholly-owned subsidiary of Copersucar, a Brazilian conglomerate; and The MJB Coffee Co. (Tr. 2122, 3224, 4496, 10 541). 11. Many major retail chain grocery stores produce or have produced for them their private brands of coffee which are sold in their stores in competition with brands sold by independent coffee manufacturers (CX 437F; Tr. 12 335). Some ofthe larger chains selling their own brands of coffee are The Kroger Co., the Great Atlantic and Pacific Tea Company ("A&P"), Safeway, and First National/Pick ' Pay (Tr. 1641, 6276, 9538).

Initial Decision 103 F. 12. There are also a substantial number of smaller roasters, many of whom are significant factors in their areas of distribution: Chock Full O'Nuts Co.; American Maize Co., which markets the Savarin brand; J. Lyons & Co., which markets the Martinson s brand; and the JFG Coffee Co. (Tr. 4278-2, 8338, 10 527).

D. The Consumption Of Regular Coffee 1. The Use Of Regular Coffee As A Traffc Builder 13. During an average three-month period, 80% of households in the United States purchase regular coffee at least once. During a 13-week period, these households purchase, on average, 7 1/2 pounds of coffee (CX 190B).

14. Retailers who sell ground coffee recognize that their customers are very sensitive to its retail price. Mr. Theodore Engel of the Kroger Co. testified that it regards regular coffee as a "board item " a term used to identify "the basic items we feel from experience are the most price-sensitive items that we carry. . . " (Tr. 1707). 15. Maxwell House Division documents also disclose the importance of price to regular coffee consumers: (6) Consumers have been educated to buy ground com on price. Ground coHee shares are dramatically responsive to changes in price levels vs. competition. Many people buy every week or two (CX 20Z-38).

Retail price is an extremely important factor in the ground coffee industry affecting both coffee market trend and shares of individual brands within it (CX's 38Z-30 39Z-17, 40S).

16. This sentiment was echoed in a May 1973, Procter & Gamble memorandum:

(a)11 our experience in the coffee business tells us that it is extremely responsive to pricing; and pricing equity is regarded as a necessity even in areas where we have a strong leadership share position (CX 524).

17. Because of this consumer attitude toward regular coffee, it has historically been an actively merchandised category by the Grocery Trade-promoted as a ' traffc builder' or a 'loss leader' " (CX 190B), and this attitude has "led to the use by retail grocers, of coffee as a trade building item" (CX 205A).

18. Louis Epstein, president of Golden Dawn Foods, Inc., a company that wholesales groceries to a voluntary group of supermarkets in western Pennsylvania and eastern Ohio, has purchased ground coffee from manufacturers since 1970 (Tr. 12 157-60). According to Mr. Ep- 204 Initial Decision stein, it is a common practice" in the grocery business to sell a highly desirable item such as coffee at a reduced price to attract consumers: (T)he assumption being that if a consumer walks into the store for the reason of buying that product at a very low price, she will buy the rest of the order there and then you wil balance Qut and make money on the balance ofthe order. That is loss leader selling and it's a common practice in the industry (Tr. 12,187; see also Tr. 2881, 5753, 5818 5938, 6611). (7) 19. David Graham, who had extensive experience first in the marketing of ground coffee brands produced by General Foods, and was thereafter vice president and general manager of the Martinson Division of the Squibb Corporation, responsible for the marketing of Martinson s and Beechnut coffee brands, and who later participated at Standard Brands in major management decisions affecting the Chase & Sanborn brand of coffee, testified that coffee is an "extremely competitive" industry. Certain characteristics make it particularly competitive. It is a large category in terms of physical volume and dollar sales, and represents a "very meaningful part" of the total revenue of grocery store sales. It is a product which is widely consumed and purchased frequently, and it is used by grocers to draw people into their stores (Tr. 10 536-37).

2. The Trend In Regular Coffee Sales And The Increase In Competitive Activity 20. In fiscal 1968, the consumption of regular coffee in the United States peaked at 122.4 million units, or 1.468 bilion pounds (RX 1l05B). In the next ten years, regular coffee consumption declined by almost one-third (Tr. 4335-36).

21. This decline in consumption has caused manufacturers of ground cofiee to intensify their promotional activity in order to maintain their business (Tr. 5722). Any coffee company seeking business in new areas of distribution would have only one source for this business, from the other brands existing in that area (Tr. 4336). Hills Bros. recognized that Procter & Gamble s expansion of Folger coffee did not expand the ground coffee market, but that Folger s obtained its share by taking it from other brands (Tr. 2196). 22. Several Maxwell House Division documents indicate that as a result of this decline in consumption, producers have increased marketing expenditures in an attempt "to maintain volumes in a declining market" (CX 17T) and that trade promotion expenditures have increased because producers realized that it was necessary and more effcient to trade promote in order to utilize more capacity (CX 20Z- 38). Local and regional roasters' reliance on trade dealing as their Initial Decision 103 F. single marketing tool is Itreflective of incremental economies in an industry with idle capacity" (CX 190B).

23. The importance of this decline in consumption and attendant increased marketing expenditures was recognized within (8) the Division as a potential cause of increased competitive activity by roasters: Folger, in quest of "at least" static volumes in a declining ground market, wil continue to grow in their established markets with accelerated trade spending. Hills, MJB and other competition will continue to lose share intensifying trade spending along the way.

In the HFNI, Savarin, CFON, C&S and Martinson s will continue to spend to solidify share (eX 16H; See also ex' s 13M, lSD, 16Z-36). 24. This increase in competitive activity has not been limited to just a few ofthe Maxwell House Division s sales districts but is a nationwide phenomenon (CX's 659B, 670A, 672C, F, 673D, 677 A, C; RX 1207 A-C, F).

E. The Marketing Of Regular Coffee By The Maxwell House Division 1. Introduction 25. During the 1960's, its Maxwell House Division accounted for approximately 35% of General Foods' profits (Tr. 4495). The Division s major regular collee brand is Regular Maxwell House; other brands sold by the Division are Maxim, a freeze-dried coffee, Sanka which is produced in freeze-dried, regular and soluble form, Yuban a premium colIee, Brim, a decaffeinated ground and freeze-dried coffee, and Master Blend, Electra-Perk, ADC, Mellow-Roast and Max- Pax, which are regular coffees (Tr. 4492- , 4836-37, 4847--8, 4858). 2. The Marketing Mix 26. Prior to fiscal 1972, the Division s coffees were sold in 27 sales districts. In fiscal 1972, the Division s national sales manager reduced the districts from 27 to 20 and extended (9) the boundaries of the remaining sales districts. The districts contain 52 marketing areas. The Division s coffees are sold direct to retail grocers and to wholesalers. Ofthe 200 000 retailers in the United States, some 30 000 are sold direct; the remaining 170 000 buy their Maxwell House Division coffees from wholesalers (Tr. 5723, 7130-31, 7159). The Maxwell House Division (and other roasters) promotes its colIee products through a so-called "marketing mix" which consists of advertising, trade dealing and consumer promotions (Tr. 7115).

27. Advertising, or ttmedia " consists of television programming, 204 Initial Decision magazines, newspapers, Sunday supplements, radio, and outdoor bilboards. Television advertising is either network or spot. Network advertising is bought from the major networks, CBS, NBC and ABC in New York City. Spot advertising is purchased from local television stations (Tr. 7820).

28. Trade deals consist of non-performance and performance allowances which are offered to the Division s retail customers (when the Division sells to wholesalers, these allowances are passed on to the wholesalers' customers) (Tr. 7130-31).

29. Non-performance allowances such as buying allowances are discounts from the list price for each case of coffee. The retailer is not required to perform in order to receive the allowance (Tr. 1374-75 5754).

30. A performance allowance is a discount from the list price for each case of coffee, with the requirement that the retailer engage in specified performance. Performance allowances include display allowances, advertising allowances, reduced price features and count and recount. A display allowance requires that the retailer display the coffee in a particular fashion. An advertising allowance requires that the retailer advertise the product in order to get credit for the allowance on all volume purchased. A reduced price feature requires the retailer to reduce the price of coffee to the consumer. A count and recount is an allowance based on cases moved from the trade s warehouse to the retail store serviced by that warehouse (Tr. 5846-7 6655-56).

31. Consumer promotions take the form of coupons, of which there are two basic types: consumer coupons and retailer coupons (RECUs). Retailer coupons (RECUs) are coupons that a retailer runs in its newspaper advertisement which offer the consumer a discount on the purchase price of a particular brand. When the consumer buys the product and presents the coupon to the retailer, the consumer receives a price reduction equal to the face amount of the coupon. Upon submittal of the coupon to a manufacturer, the retailer receives an amount equal (10j to the price reduction from the manufacturer (Tr. 2009- , 2505, 2747 , 12 100-3, 12 183). Consumer coupons are redeemed like RECU' s by the retailer, but they are either mailed direct to the consumer by the producer ("DMCU"), appear in newspapers or magazines C'NECU"), or are packed in the product ("PIC' ) (Tr. 1601 2013- , 2505-06, 2725). RECU' s are effective for a four-or-six-week period. During this period, a retailer can only participate once in the RECU offer and the RECU must expire within a week from the date of publication (Tr. 5846, 5998, 7123). Consumer coupons can be redeemed for an indefinite period (Tr. 4114). Initial Decision 103 F. F. Folger s Entry Into The HFNI 32. Although Regular Maxwell House is sold nationwide, the bulk of its volume has traditionally been sold in the "HFNI " an acroynm for "High Franchise Non-Infringement." The HFNI is the area of the United States east of the Mississippi River where Regular Maxwell House has been the leader in regular coffee sales. In the western part of the United States, the so-called "Low Franchise" area (CX 448A-C; Tr. 1398-99), Regular Maxwell House has been faced with competition from the leader in that area, the Folger Coffee Company (Tr. 2709- , 4502).

33. The Maxwell House Division s major competitor has been and is the Folger Coffee Company. In 1963, Folger was acquired by the Procter & Gamble Company (Tr. 4506) and, until the fall of 1971 restricted its competitive activities, with minor exceptions (some sales in Indiana, Cincinnati, Ohio, and Florida), to the western United States (Tr. 2709- , 4502).

34. In October 1971, Folger began sellng its coffee in Cleveland Ohio as a test market, with the intention, ifits introduction into this market were successful, to expand into the balance of the eastern United States on an orderly schedule (Tr. 2710-11). After waiting for three months, the Maxwell House Division decided to respond to Folger s eastern thrust. It is this response to Folger s entry into Cleveland and other eastern areas and the consequent results which prompted the Commission to issue the present complaint. G. The Relevant Product Market 35. The parties have agreed that the relevant product market is all regular coffee products packaged for sale at retail, including caffeinated, decaffeinated, and extended (11) ground coffee products (Prehearing Conference, April 12, 1978, Tr. 79-84). H. The Relevant Geographic Market 1. Introduction 36. While Dr. Kenneth Elzinga, General Foods' major economic expert, agreed with few of the conclusions drawn from this record by Dr. H. Michael Mann, complaint counsel's main expert 3 he and Dr. Mann agreed that a relevant market is an "area that encompasses the primary demand and supply forces which determine price" (Tr. 3528 9415, 9458-57). However, as is not uncommon when economists (or lawyers) apply a simple principle to a concrete factual situation, the experts' conclusions differed: Dr. Elzinga testified that the relevant J Dr. Elzinga is a professor of e onomics at the University of Virginia (Tr- 9401)- Dr- Mann is a professor of &onorncs at Boston College (Tr. 3515).

204 Initial Decision geographic market for coffee is probably nationwide (Tr. 9430), while Dr. Mann stated that the market is no larger than a General Foods sales district (Tr. 3527).

37. The following discussion analyzes the demand forces which are emphasized by Dr. Mann and complaint counsel and the supply forces which Dr. Elzinga and counsel for General Foods claim are the primary forces which define the relevant geographic market. 2. The Demand Forces (a) Industry Recognition (1 General Foods 38. General Foods recognizes in dealing with its customers that the demand for regular coffee is not uniform throughout the nation but that it differs from area to area. (12) 39. General Foods' fiscal 1971 marketing plan4 for Maxwell House is an example of the way it views the demand for coffee: Ground Coffee Market Structure The National Markets us. The Market Places. . National shares are only an average of widely divergent market places, because RMH5 and Folger are strongly regionalized and smaller brands have only local zones of influences. National share relationships have little relevance to the structure within marketplace.

. No region, district, market area or dealing area even vaguely resembles the national profile.! (Emphasis in the original.

It is in each market place that consumers purchase coffee and the structure of the individual marketplace is the arena that defines the choices the consumer has. To influence her choice marketing effort.'::; must be tailored to her marketplace (CX 20Z- 37).

40. Other statements in General Foods documents reveal the lessthan-national demand forces which it faces when it markets its coffee products:

BACKGROUND THE GROUND COFFEE MARKET STRUCTURE IMPLICA TIONS Marketing Strategy . In the ground coffee market, a flat national dealing/spending level is not a good fit of marketing strategy to market structure. (13) . Due to the responsiveness of ground coHee shares to changes in price, RMH must 4 Marketing plan are tatements of brand objectives and brand strategie and outline the strategie to accomplish those goals (Tr. 1521).

(; Regular Maxwell House.

, , Initial Decision 103 F. respond to market-by-market conditions to remain competitive to locally powerful brands.

The effectiveness of broad marketing programs like advertising and consumer promotion, depends on RMH' s ability to neutralize the impact of on-shelf/feature differentials vs. competition as and where they occur (eX 20Z-39). 41. The fiscal 1976 marketing plan, under the heading "dealing principles" states: "Trade deal on (aJ market by market basis to achieve Brand' s overall objectives" (CX 19A, Z-26; see also CX 18Z- 16). The fiscal 1976 media plan for Maxwell House describes planned media expenditures for that year and states: "the Brand's base budget was allocated by dollars by market according to the Brand' s Development 42. Index"The fiscal 1975(CXmarketing plan 457S).contains a chart RMH Financial Summary Fiscal 1975 " that shows differences in trade and consumer promotions between the sales districts of Youngstown Philadelphia, Kansas City, and Dallas (CX 63S). Another chart contains a series of statements reflecting differences in trade dealing strategy for Maxwell House in each sales district (CX 63Z-15). 43. The fiscal 1974 marketing plan under the heading "Deal Strate- " states: "RMH uses case rates tailored to each market to achieve parity shelf pricing" (CX 15Z-).

44. The fiscal 1972 marketing plan contains a section entitled "Consumer Promotion" (CX 13Z6). In this section, individual markets are broken down into four categories, with examples given of each: (1) Shelf Market-Strong Franchise" (examples: Washington, Memphis); (2) "Feature Market-Strong Franchise" (examples: New York, Cincinnati); (3) "Shelf Market-Weak Franchise" (example: Los Angeles); and (4) "Feature Market-Weak Franchise" (examples: San Francisco, Dallas). Separate regular Maxwell House consumer promotion objectives and strategies are listed for each category of markets (CX 13Z&-9). 45. The fiscal 1971 marketing plan indicates that Regular Maxwell House media funding was allocated on a district-by-district basis according to several factors, including Folger s media spending (CX 12P; see also Tr. 1177). (14) 46. Still more General Foods documents contain statements and tables indicating that demand conditions differ between its sales districts:

a) The "State of the Business " contains a chart entitled "Trade Spending Analysis FY '75 AFP" that shows differences among Maxwell House Division sales districts in terms of Maxwell House trade dealing strategies (CX 1721).

b) The March 1, 1976 Trade Promotion Review " states: "Overall trade dealing provides a mechanism for implementing market-by- , 204 Initial Decision market pricing action, consistent with the nature ofthe industry and competitive circumstances" (CX 190C, CX 191C). c) The August, 1976 RMH Trade Dealing Principles Review tracks Maxwell House share versus competition for each district market area, and dealing area (CX 195A-Z8). Depending on the Maxwell House to competitor share ratio, Maxwell House dealt at a different level, keying only to certain competition (CX 1950). Under the heading "How Deal Levels Are Calculated " the document explains how General Foods calculates Maxwell House trade dealing levels separately for each market "depending on the RMH to competitive share ratio" with lower deal levels being designated for areas in which Maxwell House s market position is stronger (CX 195T; see also 194Q).

d) A 1975 "RMH Trade Dealing Principles " document states: The RMH trade principles are based on pack size franchise relationships to competition on an individual market basis (CX 188B, emphasis in original). This document also states that Maxwell House s "specific dealing programs are on an individual market basis" (CX 188B). e) A "Regular Maxwell House Trade Dealing Review" analyzes on a sales district by sales district basis the different competitive factors in each sales district (CX 187R- , CX 187Z3-14). (15) 1) In another "RMH Trade Dealing Principles Review " General Foods again recognized that it must deal with different demand forces from area to area:

Offers by RMH must be structured to appeal to each area s trade factors with an understanding of their merchandising policies, RMH' s relative pack size strengths and competitive offer levels (CX 194D).

47. The marketing plans of other General Foods brands also reveal General Foods' perception that customer demands differ from area to area. The following are examples:

a) The Brim fiscal 1977 marketing plan states that Brim s trade promotion spending would be initiated to provide "(m)erchandising support for drive period promotions on a responsive district basis only." (CX 61N).

b) The fiscal 1976 marketing plan for Max-Pax contains an outline that shows separately for individual Maxwell House Division sales districts and market areas the estimated trade rates per unit for Max-Pax and the trade deal strategies to be followed by Max-Pax in fiscal 1976. The outline reveals differences among the Maxwell House Division sales districts and market areas in relation to these factors (CX 44Z17-20).

c) The fiscal 1974 Yuban marketing plan contains a chart entitled Initial Decision 103 F. Trade Deals/Pricing Objectives" that shows differences among Maxwell House Division sales districts in terms ofYuban s trade deals and pricing objectives for fiscal 1974 (CX 35Z3). d) The fiscal 1975 Yuban Marketing Plan under the heading Spending Principles" discusses the concentration of advertising effort in particular sales districts and states: Beginning in F' 74 , Yuban s basic spending strategy wil be to concentrate the Brand' advertising and (16) promotion expenditures in key markets only. This "concentration strategy recognizes that the greatest areas for growth for Yuban lie in current areas of strength where the brand already possesses consumer/trade leverage. Seven districts (N. , Boston, Dallas, St. Louis, Portland, San Francisco, Los Angeles) which account for approximately 80% of Brand volume wil receive all the Brand' s media support and over 90% of its promotion expenditures (CX 35U-V). 48. As would be expected from the above statements, General Foods' marketing strategies are tailored, in many cases, to individual Maxwell House sales districts or market areas. Maxwell House s assistant product manager, who was responsible for selling volumes and deals, allocated deals to individual sales districts with the objective of delivering certain volume or share levels in a sales district (Tr. 1373 1376-78).

49. The Maxwell House Division regional sales managers set volume targets for each of the districts (Tr. 6457, 6501, 6603). They also reviewed the regular coffee pricing situation in their regions on a district-by-district basis (Tr. 6658, 6663). Each of the district managers was assigned certain volume objectives, and based on these objectives, certain deals were authorized for those areas in order for the district managers to accomplish the objectives (Tr. 1378). 50. Trade deals were cut for each sales district separately. The same trade deal was never cut for the entire country (Tr. 1377-78), and trade deal rates generally differed among sales districts (Tr. 1660 1963, 5815-16, 5828).

51. Advertising levels for the Regular Maxwell House brand also varied among sales districts (CX's 12P, 4578, 794A-D). 52. The reason that the Maxwell House Division offered varying deals in different districts is that levels of competition differed between the areas (Tr. 7159). By way of ilustration, Mr. Keller, Division s southern region sales manager, testified: . . . I would love to have one deal that I could run the entire region. But there are too many local competitors and too many (17) different competitors and difierent competitive situations in each or the markets that doesn t allow me to do that. It doesn t allow me to be-to run one deal-and as a result I need to have different deals in different districts (Tr. 6675).

204 Initial Decision 53. In some cases, the Maxwell House Division even offered trade deals in an area smaller than a sales district (Tr. 5815). For example before Folger entered Cleveland, the Maxwell House Division dealt the Youngstown sales district as an entity (Tr. 7161). After Folger entered Cleveland, the Division responded to Folger s entry only in Cleveland instead of the entire Youngstown sales district (Tr. 4554 7161--2). The Division s increases in its trade rates, consumer promotions, and advertising in response to Folger s entry in Cleveland were limited to the Cleveland area because the need for them only arose in that area (Tr. 5833-34).

54. When the Maxwell House Division responded to Folger s entry into Pittsburgh, it limited its response to the Pittsburgh market area (Tr. 4554-55, 7161--2), and its increases in trade promotions, consumer promotions, and advertising in response to Folger s entry into Syracuse did not extend beyond the Syracuse sales district (Tr. 5836 5997- , 6022- , 6103, 7162).

55. When the Maxwell House Division responded to Folger s entries into the Philadelphia and Syracuse sales districts, part of its response was the introduction of a new regular coffee brand, Horizon. Horizon s introduction was limited to the Wilkes-Barre/Scranton market area of the Philadelphia sales district and to the Albany market area of the Syracl'se sales district. Its entry was timed to coincide with Folger s entry into these areas (Findings 424-27). 56. When Maxwell House Division responded to Folger s entries into Cleveland and Pittsburgh, part of its response involved increased advertising and deal rates in two of Folger s largest and most profitable areas, Kansas City and Dallas (Findings 428-30). 57. The same area-by-area response by the Maxwell House Division occurred when a competitor, Hills Bros., expanded eastward in the late 1960's (Tr. 4555, 5837, 6097-99).

58. The Maxwell House Division also maintains records of certain business activities on a district, or smaller, basis: "district profitabili- " statements show the following categories of information with respect to Regular Maxwell House (18) for each Maxwell House Division sales district: (1) "Volume Units"; (2) "P/S/S" (pack size split); (3) " (contribution margin); (4) " " (gross profit); (5) "Trade Deals; (6) "Consumer Deals; (7) "Total Deals; (8) "Advertising; (9) All Other; and (10) "PBIT" (profit) (CX 424A-B). The Division also maintains records of sales volumes for its regular coffee products on a market area and sales district basis (Tr. 5830), and generally calculates volume objectives for Regular Maxwell House sales promotion plans by sales district or marketing area (Tr. 10 610). 59. In addition, the Maxwell House Division maintains history data on the amount of trade rates in various eastern markets (Tr. 1587), Initial Decision 103 F. develops "deal close" documents which show deal rates in each sales district (CX 136A-G; Tr. 1404-5), and in its weekly deal control documents, records deals on a market area by market area basis (CX' 1061 , 1064).

60. The Division also gathered Nielsen data showing the market share for each of its regular coffee brands and for its competitors on a sales district basis and sometimes on a market area basis (RX' 1106-07; Tr. 4328-30), and one of its advertising agencies, Ogilvy & Mather, prepared summaries of Regular Maxwell House s advertising expenditures by sales district (Tr. 1177). In addition, Ogilvy & Mather prepared summaries of advertising expenditures by Regular Maxwell House and Folger in various local markets (Tr. 1543). (2) The Maxwell House Division s Competitors 61. The Maxwell House Division s competitors also recognize that demand forces for regular coffee differ from region to region and take these differences into account when they market their coffee. 62. Mr. Allen Toy, vice-president of corporate development of Hils Bros. (Tr. 1991), testified that his company s sales promotion areas were established for the purpose of setting deal promotions with customers in those areas (Tr. 2006). While Mr. Toy is only generally aware of the dealing areas of his competitors, he believes that all competitors would consider major metropolitan areas as part of one promotion area. Differences between competitors' trade areas would occur between cities. For example, the area between Cleveland and Pittsburgh could be included in one sales district for General Foods and another one for Hils (Tr. 2008). (19) 63. Hills Bros. offers different deals, including different consumer deals in different promotional areas, as does its competitors (Tr. 1660 2016-18). Its advertising levels also vary among its promotion areas (Tr. 2021) 64. Folger divides the country into smaller geographic areas called sales promotion areas" (Tr. 2907). The boundaries of Folger s sales promotion areas were designed to comport with the areas in which Folger s competitors were making their offers (Tr. 2907-08), and these areas compare very closely with those of its competitors (Tr. 2908). 65. Folger sets its trade deals by promotional areas, and there are differences in Folger s dead-net prices among its sales promotion areas (Tr. 2908-9).

66. According to Mr. David W. Graham, who worked for both Standard Brands, which markets Chase & Sanborn coffee and Martinson/ Beechnut (Tr. 10 526-28), Standard Brands' sales districts were generally similar to the Maxwell House Division, with one differencethe Division, at the time he was talking about, had 20 sales districts 204 Initial Decision while Chase & Sanborn had 23. In the case of Martinson/Beechnut he estimated that there was probably 95% agreement between its sales districts and Maxwell House s (Tr. 10,643). 67. Hils Bros., Folger and Standard Brands kept records of or studied certain business activities (trade and consumer deals, advertising, and profitability) by district (CX 866I; Tr. 2019, 3152- , 3169 658).

(b) Wholesale And Retail Distribution (1) Grocery Distribution Patterns 68. Mr. John J. MacDonald, a long-time employee of General Foods Maxwell House Division testified that: "There are no national retailers. None. There are no national wholesalers. Each of them fills a particular piece of geography" (Tr. 7159). Within their particular piece of geography, wholesalers and retailers may use multiple buying offces and several distribution centers. For example, Kroger purchases its regular branded coffee in fourteen locations for delivery to its nineteen distribution centers (Tr. 1660), and Flickinger, a wholesale grocery company located in Buffalo, New York, has six branches in New York, Ohio and Pennsylvania (Tr. 11,845). (20) 69. The Maxwell House Division and other coffee producers recognize that the demand for their products varies between sections ofthe country and structure their sales and promotion practices accordingly. The Division s sales districts and the distribution system that accompanied them were set up to match the particular needs of the retail coHee business (Tr. 4557). The actual pattern of retailer locations and the distribution to those retailers of coffee from General Foods' plants were among the many factors considered in determining what areas should constitute separate sales districts (Tr. 4557), and these factors continued to be considered in delineating sales districts during the 1970' s (Tr. 4558).

70. According to Mr. Toy of Hills Bros., all industry members consider major metropolitan areas as one promotion area. A city like Cleveland could not be separated into two promotion areas because retailers in Cleveland cover the city wide area. Differences among the dealing area boundaries of industry members only exist in outlying areas (Tr. 2008).

71. The most important factor that Hils Bros. considers in establishing its sales promotion areas is the retailers' areas of distribution. This is most important because it is "a very diffcult job to sell if you G On the other hand, the Maxwell House Division s decision in !972 to reduce the number of sales distrjc (Finding 26) indicates that fa.dors other than customer demand were important considerations in setting up district boundaries Initial Decision 103 F. "7 Hils Bros. cut across the retailer s warehouse distribution area. includes all the stores receiving products from a warehouse in a single sales promotion area (Tr. 2007). Hils Bros. also aligns sales promotion areas to retailers' needs to minimize transshipment problems (Tr. 2022).

72. Folger s sales promotion areas were also designed to reflect trade distribution boundaries. These boundaries are the areas which the wholesale and retail trade ship their products (Tr. 2907-08). (21) (2) Overlap Of Sales Districts And Grocery Distribution Patterns 73. Complaint counsel commissioned a study by Audits and Surveys in an attempt to demonstrate that there is a correlation between the buying and warehousing patterns of grocers and the boundaries of Maxwell House sales districts. Mr. Richard Lysacker, president of this firm, testified about the study which it conducted. Mr. Lysacker has had extensive academic and business training in survey techniques and statistics (Tr. 1743-52).

74. Mr. Lysacker s study analyzed for 1971 and 1973: (1) the extent to which merchandise flows from wholesaler warehouses located in a sales district to stores located in that same district; and (2) the extent to which General Foods' sales districts align with retail market areas defined by an independent source that considered grocery distribution patterns (Tr. 1811-12).

75. The Grocery Distribution Guide is an annual report containing data that identifies the locations of grocery warehouses and the retail stores that the warehouses supply (Tr. 1791-92), and it was used by Mr. Lysacker to analyze to what extent food sold in grocery chain stores within Maxwell House sales districts were also warehoused in those districts (or outside of them) (Tr. 1795). 76. Mr. Lysacker satisfied himself that the Guide was accurate by having his staff conduct 200 randomly selected interviews with chain warehouses and buying offces and he concluded (a) that the interviews demonstrate that coffee is warehoused in the same location as other food products included in the Guide and (b) that his study applies, therefore, to coffee as well as to grocery products in general (Tr. 1812- , 1818-19, 185&-57).

77. Mr. Lysacker s analysis of warehouse distribution patterns in 152 metropolitan areas for 1971 and 1973 led him to conclude that merchandise sold in a sales district was by and large also warehoused in that same sales district (CX 1029A-C; Tr. 1798). 78. This analysis reveals, according to Mr. Lysacker, that for 1971 1 And, as will be disCllS!cd below, this phenomenon often leads to a breakdown in the cont.ainmeat ofumnand forces within sales distric1.

204 Initial Decision and 1973 96% of the grocery products sold within a Maxwell House Division sales district were also warehoused in the same sales district and that trade flow across district boundaries was minimal. The 96% figure is an average among sales districts weighted on the basis of area volume (CX 1029A-C; Tr. 1914-15). (22) 79. Mr. Lysacker conceded that in the northeast, some shipments cross sales district boundaries, but, according to him, the amount is only 3% in Boston, 4% in New York, 6% in Philadelphia and 6% in Syracuse (Tr. 1917-18).

80. Progressive Grocer is a widely used marketing guide book that supplies numerous details relating to the grocery business in 79 different marketing areas. These areas were developed by taking into account merchandise flow between retail stores and warehouses that service them and media coverage patterns (Tr. 1761-63). 81. After comparing maps of General Foods sales districts with the Progressive Grocer marketing areas8 (CX's 1017A-Y and 1018A-L), Mr. Lysacker concluded that the sales districts had a "strong degree of correspondence " with the Progressive Grocer marketing areas (Tr. 177G-71; CX 1027A-E).

82. Although Mr. Lysacker s study of Maxwell House Division sales districts and Progressive Grocermarket areas reflected, in his opinion the alignment of sales districts with television markets, he also undertook an analysis focusing specifically on how television reception areas correspond to the boundaries of Maxwell House Division sales districts (CX 1015A-F; Tr. 1778-91).

83. Arbitron is a company that measures television on a market-bymarket basis (Tr. 1778-79). It constructs advertising coverage areas that encompass the geographic area predominantly served by the stations located within each area (Tr. 1779--0). These 210 areas are referred to as ADIs (areas of dominant influence) (Tr. 1778-0). Interviews with television viewers are used to establish the ADI to which an area wil be assigned (Tr. 1779--0).

84. A comparison ofthe Maxwell House Division sales districts and Arbitron s ADI areas was made using a random sample of six sales districts for 1971 (CX 1028A-C; Tr. 1781- (23) 85), and Mr. Lysacker believes that there is a high degree of correspondence between Maxwell House Division sales districts and ADI areas, and that the spilover of advertising messages across sales districts is relatively small (Tr. 1788-7, 1790). Mr. Lysacker s study indicates that for the sales dis- "The study compared counties whith are induded within the confines of bath a Gencml Food!! sales district and Progressive Gmcermarketing area with counties which are included only within a General Foods sales district and ounties which are induded omy in a Progressive Grocermarketing area (Tr. 1764). The atudy revealed that 91% of food sales tell in counties which are common to General Foods sales districts andProgressive Grocer marketing areas (Tr. 1767-68).

226 FF;DERAL TRADE COMMISSION DECISIONS Initial Dccision 103 F. tricts analyzed, the overlap contained over 90% of the food sales in the combined areas (CX 1028A-C).

(3) Trade Flow And Transshipment 85. Transshipment is the purchase of a product by a retailer for one warehouse and its shipment to a different warehouse in another area (Tr. 1664 5729-30). Trade flow occurs when a single warehouse serves retail outlets in more than one sales district (Tr. 6558). 86. On the basis of his study, Mr. Lysacker concluded that trade flow was insignificant in 1971 and 1973 and that it did not prevent manufacturers from marketing in a different fashion from one sales district to another (Tr. 1774).

87. Retailers who sell the Maxwell House Division s regular coffee products are free to ship coffee from one district to another, but the Division has attempted to discourage these shipments by announcing that it wil not honor offers available in one area when retailers seek to take advantage of that offer in other areas. (4) Attempts To Limit Trade Flow And Transshipment 88. Retailer coupons may be employed as a method of lessening transshipment (Tr. 5813-14), and one Maxwell House Division retailer coupon assessment states that their use "avoids trade flow problems" (CX 186C). Another trade deal document states that: RECD's are financially effcient for RMII and offer the brand a tactical tool to deal with distinct competitive situations where trade flow prevents an economical case rate response (CX 190L).

89. The Maxwell House Division insists that retailer coupons be redeemed only by stores in the sales district or area in which the Division runs the offer (Tr. 5814, 6528). The deal (24) bulletin for a retailer coupon offer defines the district or trade dealing area in which the retailer coupon is to be run (Tr. 6528). Retail chains outside the area in which a retailer coupon offer is made cannot take advantage of the offer unless they have stores within the geographical boundaries ofthe offer. In such a case, only the retailers' stores within the geographical boundaries can take advantage of the retailer coupon oller (Tr. 5814, 5997-98).

90. For example, if a retailer in the Syracuse sales district ran a coupon that was offered only in the New York sales district, the redemption center would void those coupons redeemed in Syracuse. This would be unauthorized performance performance outside the authorized area (Tr. 6095-98).

91. The offer number associated with a particular retailer coupon 204 Initial Decision offer must appear in the coupon that the retailer prints in the newspaper, and it can be determined quickly whether a retailer coupon offered in one sales district has been redeemed in another sales district (Tr. 5854-55).

92. Under the terms of a Maxwell House Division retailer coupon offer, a consumer could not cut a retailer coupon out of a newspaper ad for Kroger in Cincinnati and redeem the coupon at a Kroger store in Cleveland (Tr. 5853). Controls that exist in General Foods' coupon redemption centers are designed to monitor and control that kind of redemption (Tr. 5854).

93. The Maxwell House Division s promotional announcements limit the geographic area of the offers (Tr. 3555-56). A typical promotion announcement identifies the geographic area in which the offer is available and states that payment on the allowance wil be made only on volume shipped to retail outlets located in the offer area (CX 1125).

94. According to Mr. Engle, a Kroger offcial, the Maxwell House Division would not permit his company to issue a purchase order for regular coffee for delivery at one location at a price being offered by a promotional announcement in another part of the country (Tr. 1662). More specifically, since 1971, the Division would not permit Kroger to issue a purchase order for regular coffee to be delivered in Cleveland at a price being offered by the Division to retailers in Texas (Tr. 1661) 95. Transshipment can also theoretically be controlled by persuasion, restructuring of allowances, and allocations (Tr. 1664). "Allocation" means that a manufacturer wil limit retailers' acceptance of offers to a specific number of cases (Tr. 1666-67). (25) 96. Allocation can control transshipment because only the amount of product actually needed for the market in question would be sold to the retailer. The retailer would not be sold enough volume to ship to other areas (Tr. 1666-67).

97. In at least one instance, the Maxwell House Division monitored transshipment. The Syracuse District Action Plan, under the heading Transshipment Monitor/Control " states:

The higher case rates on RMII could lead to higher than usual transshipment into neighboring districts. As a means of monitoring this flow of product the Brand will continue to ship the Fresh-Lock Lid into the Syracuse District. This will provide a ready and visible means of tracing the shipment of Syracuse product. Secondly, key account purchases will be monitored in White Plain.,, and compared against historical purchases. Consistent and unusually heavy purchases wil indicate transshipment.

. . .

Initial Decision 103 F. These two monitoring devices will provide data to use in support of discouraging transshipment at our customer headquarters (CX 710K) (Emphasis in original). In a covering memorandum, Mr. Einloth of General Foods concluded that "we believe we have the means to monitor and, therefore, exercise control with our key accounts" (CX ?loa). (5) The Increase In Trade Flow And Transshipment 98. While Mr. Lysacker does not believe that trade flow between sales districts is significant, his conclusion is based on 1971 and 1973 data, and General Foods has presented evidence which indicates that it is often unable to control trade flow as well as transshipment. In fact, trade flow and transshipment are increasing as wholesalers and retailers expand their trading areas. (26) 99. One example of the Maxwell House Division s inabilty to discourage transshipment is the result of its monitoring effort described just above. According to Mr. Salesman, the Maxwell House Division Syracuse district manager at the time Regular Maxwell House converted to straight case rates in Syracuse, he received calls from district sales managers across the country reporting the appearance of fresh-lock lid product in their districts (Tr. 5933). Specifically, that peculiarly identifiable product was transshipped from Syracuse to Portland, Oregon, to Los Angeles, California, to Boston, Massachusetts, to Philadelphia, Pennsylvania, to Jacksonville, Florida and to Youngstown, Ohio (Tr. 5934), and the Division s southern region manager testified that Daylight Grocery, a small grocery chain in the Jacksonvile area, had fresh-lock lid products on its shelf during the time of the Syracuse test (Tr. 6586).

As explained by Mr. Salesman:

The rate differential that existed between what they could buy it for in Syracuse and what the trade rate that existed in another area was-the differential was so great that it more than covered the freight cost plus any profit that they might make on that. Transshipping is not only done by the grocery trade, but there are people that we have referred to as bandit brokers, and Wiley and Trepel out of New York have made a business out of this. They know what the going trade Irate) is, and they can buy that coffee in that area and stil pay the freight and make a profi on it (Tr. 5934-35). 100. Other coffee producers have also experienced transshipment of their product (CX 540B; Tr. 2909, 10 630).

101. John Mann, the Maxwell House Division s eastern region manager, gave specific examples of transshipment into and out of sales districts in the eastern region. In summary: Into the Youngstown District, there is transshipment from the New York, Boston Chicago and Dallas districts by Seaway, Fisher Fazio, Pick- Pay and First National. 204 Initial Decision Into the Syracuse District, there is transshipment from the Cincinnati, Philadelphia New York and Bosion districts (27) by S.M. FJickinger, P&C, Grand Union, Price Chopper, Springfeld Sugar Co. and C&S Wholesale Grocers. Into the Cincinnati District, there is transshipment from the Syracuse district by M. Flickinger.

Into the Philadelphia District, there is transshipment from the Syracuse and Boston districts by S.M. Flickinger, New England Grocers and C&S Wholesale Grocers. Into the Boston District, there is transshipment from the Youngstown district by Springfield Sugar Co. ('fr. 6463-72), 102. Because trade flow also occurs, the Maxwell House Division has sometimes offered similar deals to stores in different districts. Mr. Mann testified that trade flow is common in the eastern part of the country (Tr. 6446). 01'86 major grocery trade customers in the eastern region, at least 40 trade flow from one district into another (Tr. 6449). According to other witnesses, trade flow is so widespread that it would be impossible to develop district boundaries to contain it (Tr. 5741 6288). Because of the prevalence of trade flow, trade deals must be structured to take it into account (Tr. 6575). 103. Over the years, the Maxwell House Division has referred to the Boston, New York, and Philadelphia sales districts and part of the Syracuse sales district as "the complex" because it is a contiguous area in which trade flow necessitates that all the areas be trade dealt together (Tr. 5730, 6454-56). Close to 70% ofthe eastern region s total volume of coffee sales is so significantly influenced by trade flow in the complex that it is necessary to offer trade deals of an equal value throughout that area (Tr. 6455-56).

104. Mr. James Keller, southern regional sales manager for the Maxwell House Division, and Mr. Mann detailed the regularity of trade flow into and out of districts in their regions. In summary: From the Charlotte sales district, there is trade flow into the Philadelphia, Youngstown, Cincinnati, Atlanta and Jacksonvile districts by the following trade factors: Rich Foods, Virginia Foods, Acme Stores, Kroger, Merchants (28J Distributors, Thomas & Howard, Engles, Winn-Dixie, Bi- , Piggly-Wiggly and Wetterau. From the Philadelphia sales district, there is trade flow into the New York and Charlotte districts by the following trade factors: A&P, Safeway, Giant Foods, Food Fair and B. Greene.

From the Youngstown sales district, there is trade flow into the Syracuse, Cincinnati Detroit, Cleveland and Philadelphia districts by the following trade factors: Betsy Ross, Golden Dawn, Kroger, Seaway, Giant Eagle, Riverside, Fisher Foods, McClain Grocery and Thorofare Supermarkets.

From the New York sales district, there is trade flow into the Syracuse, Philadelphia Boston and Miami (Jacksonville) districts by the following trade factors: Shop-Rite Stores, Pathmark, Krasdale, White Rose and Waldbaum From the Syracuse sales district, there is trade flow into the Doston, New York and Youngstown districts by the following trade factors: P&C, Price Chopper and S. Flickinger.

Initial Decision 103 FTC. From the Boston sales district, there is trade flow into the New York, Syracuse and Philadelphia sales districts by the following trade factors: Stop.N-Shop, First National Springfield Sugar, Bozzuto s and Waldbaum s (Tr. 6446-8 6560-72). 3. The Supply Forces (a) The Shipment Of Regular Coffee From Plant To Customer 105. The Maxwell House Division ships ground coftee to all areas of the country from its plants located in Hoboken, New Jersey; Jacksonvile, Florida; Houston, Texas; and (29) San Leandro, California. All of the Division s brands of regular coffee are in nationwide distribution (CX 1072D-G). Because these coffee plants ship the Division products to almost anywhere in the country, its quality assurance department was formed to ensure that the products produced at each plant conform to uniform quality specifications (Tr. 5082-83). 106. No single plant has the exclusive responsibility for shipping the Division s regular coffee products to a specific area (Tr. 5132). There are a number of factors which determine which plant ships coffee to which area, one of the most important being which plant is the most effcient or least-cost producer (Tr. 5125). For example, the Jacksonvile plant, because of effciencies realized there, produces 50% of the Regular Maxwell House colIee distributed in the United States and 60% of all the regular coffee distributed by the Division (Tr. 5125).

107. Regular Maxwell House is produced at each of the four Maxwell House Division plants, none of which is the exclusive source of supply for any area of the country (Tr. 5124). All the Division s plants produce Yuban regular coffee, with the single exception of Yuban Electro-matic brand which is produced only at the San Leandro plant and distributed nationally from there (Tr. 5133). 108. Max-Pax is produced only at the Jacksonvile plant and distributed nationally from there whereas Mellow-Roast is produced only at the Hoboken plant and distributed nationally from there (Tr. 5134-35).

109. The beans used in the production of Sank a and Brim are decaf: feinated only at the Hoboken and Houston plants, but are processed into regular coffee at all four plants. There are no limitations on the areas to which a particular plant can ship Sanka or Brim (Tr. 5133). 110. There is a strong market in the United States for Kosher coffee products. The Division s Hoboken plant is not authorized to produce Kosher for Passover products because Mellow Roast, an extended nocoffee product manufactured at Hoboken, uses grain. Thus, products from the Hoboken plant conform to the Kosher for Passover dietary requirements, and, during certain periods, the Hoboken plant does not ship Maxwell House Division brands across the river to the 204 Initial Decision New York City area, where there is a great demand for Kosher for Passover products. All of the Division s Kosher for Passover products are supplied to the New York City area from its other plants (Tr. 512&-30). (30) 111. Folger has three regular coffee plants, one in San Francisco, California, one in New Orleans, Louisiana, and one in Kansas City, Missouri (Tr. 3224). When Folger expanded into the east, it did not build any new production capacity. In fact, it had recently divested itself of its Houston plant in compliance with an FTC consent order. The New Orleans and Kansas City plants supplied the coffee for Folger s eastern expansion (Tr. 3223).

112. Folger s flaked coffee is manufactured only in the San Francisco plant and distributed nationally from there (Tr. 3227), and its High Point brand is manufactured only in the San Francisco plant and shipped to Portland, Oregon and St. Louis, Missouri (Tr. 3226). 113. During much ofthe 1970' , Standard Brands' Chase & Sanborn brand of coffee was in national distribution (CX 1072A-G). Since the early 1970' , Standard Brands has distributed Chase & Sanborn nationally from only one manufacturing plant in New Orleans, Louisiana (Tr. 10 585). Its other coffee roasting plants in San Francisco California, Chicago, Ilinois, and Hoboken, New Jersey were closed during the sixties and early 1970's (Tr. 10 585). 114. Hils Bros. has two coffee plants, one in San Francisco, California and the other in Edgewater, New Jersey (Tr. 2664-5). During the period 1970-1976, the San Francisco plant generally serviced an area west of a line extending diagonally from eastern Montana to the eastern-most point in Texas, while the Edgewater plant generally serviced the area east of that line. Specifically, the Edgewater plant serviced areas as far west as Oklahoma City, Chicago, most of North and South Dakota, and Des Moines, Iowa (RX 1086; Tr. 2666-68). 115. During an eight-month period from mid-1977 to early 1978, all of Hils Bros.' regular coffee distributed in the United States was being produced at the Hils Bros. Edgewater, New Jersey plant (RX 1276).

116. Regional producers also ship their coffee over great distances. Chock Full O'Nuts, which is produced and principally marketed in the New York City area, ships its products as far west as California (Tr. 10 299, 10 542). During the period 1971-1977, Savarin coffee which was produced and principally marketed in the New York City area, was being distributed in the following Maxwell House Division sales districts: Jacksonvile, Florida; Youngstown, Ohio; Portland Oregon; and Denver, Colorado (CX lO72A-G; Tr. 10 542). The MJB brand, which was produced and principally marketed on the west coast, was (31) distributed during fiscal 1971-1977 as far east as Dal- Initial Decision 103 F. las, Cincinnati, Minneapolis and Jacksonvile, Florida (CX's 437G 1072A-G).

(b) Coffee Producers ' Areas Of Distribution 117. Some regular coffee producers in business during the 1970' sold their product only in one Regular Maxwell House sales district. For example, the Victor, Autocrat and Maplewood hrands, were sold only in Boston/Providence; Breakfast Cheer and Columhia were sold only in Pittshurgh; Wilkins appeared only in the Baltimore/Washington area; the Seven Thirty brand was sold only in Charlotte; and the Red Diamond, Royal Cup and Bailey s Supreme hrands were sold in Atlanta, Birmingham and Montgomery (CX 132G-P; Tr. 4559). 118. As to other hrands which were sold in more than one sales district, there were variations in the market shares of these hrands from district to district (CX's 20Z-37; 132G-P). 4. Expert Analysis Of The Demand And Supply Forces (a) Dr. Elzinga 119. Approximately 25% of regular coffee is shipped between 500 and 1000 miles from the plants where it is produced CRX 1227). During fiscal 1971-1977, about 40% of the Maxwell House Division s regular coffee hrands were shipped more than 500 miles (RX 1225). The fact that one can ship coffee long distances at freight rates that are rather inconsequential in terms of overall cost of the product (Tr. 9432), the ahsence of legal harriers to the nationwide shipment of coffee, and evidence that the four Maxwell House Division plants, and plants of other producers, ship regular coffee nationwide, led Dr. Elzinga to conclude that the geographic market for regular coffee is prohahly nationwide (Tr. 9430-32, 9445-6).

120. Dr. Elzinga used the word "prohahly" hecause the purpose of his testimony was to challenge Dr. Mann s conclusion as to the relevant geographic market, not to determine how extensive it actually is. Dr. Elzinga s challenge was based on his application of the Elzinga- Hogarty test to determine whether (32) the Youngstown, Syracuse and Philadelphia sales districts or the comhination of those sales districts qualified as relevant geographic markets for the sale ofregular coffee (RX 1228; Tr. 9446, 11 987).

121. The Elzinga-Hogarty test for determining geographic markets was published in the Antitrust Bulletin in 1973. It was developed in conformity with the economic theory of geographic market delineation and was designed to he applicahle in a scientific manner practically workahle and capahle of replication. Under this test, a particular area qualifies as a relevant geographic market iftwo condi- 204 Initial Decision tions are met: (1) very little of the total production in that area is exported; and (2) consumers in that area are consuming goods primarily produced there. If there are substantial exports by producers located in that area or significant imports into that locale, or both then the area is drawn too narrowly to constitute a real world geographic market (Tr. 9420-21).

122. Dr. Elzinga s test has been used to determine the relevant geographic market for cement, bulk electric power, coal, beer, and 9 in his testimony and in acrude oil (Tr. 9429). Dr. Douglas Greer textbook authored by him, describes the Elzinga-Hogarty test as one ofthe few constructive analytical procedures that has been advanced 595; Industrialfor determining relevant geographic markets (Tr. 11 Organization and Public Policy, at 162). 123. The Elzinga-Hogarty test embraces the two economic factors pertinent to the determination of a relevant geographic market. If transportation costs are so high relative to the value of the product that an area of the country is essentially sealed off from outside competitive pressures, that fact would be revealed by the absence of significant shipments to or from the area. In like fashion, the shipment data would reveal whether a legal barrier prevented commercial transactions from occurring between particular areas (Tr. 9424). 124. Each of the sales districts examined by Dr. Elzinga failed the Elzinga-Hogarty test by a considerable margin (Tr. 9455). In the Youngstown sales district, more than 80% of the ground coffee consumed there was imported from other areas (Tr. 9449). At least 85% of the coffee consumed in the Syracuse sales district was importeu (Tr. 9451-52); and, in the combined (33) Syracuse, Youngstown and Philadelphia sales districts, about 75% ofthe coffee consumed in that area was imported (Tr. 9454-55) Thus, none of these areas qualify as a relevant geographic market under the Elzinga-Hogarty test (Tr. 9453).

125. Dr. Hilke, an FTC economist called by complaint counsel in rebuttal, presented a hypothetical challenging Dr. Elzinga s contention that the Elzinga-Hogarty test is conservative.lO Although Dr. Elzinga conceded that Dr. Hilke s hypothetical is very provocative and clever (Tr. 11 999), he concluded, and I agree with him, that it has no relationship to shipment data in the coffee industry (Tr. 11 998). 126. The Elzinga-Hogarty test is an accepted method of presenting s decision in Tampaa numerical description of the Supreme Court' Electric, infra which placed great emphasis on shipping patterns in the determination of relevant geographic markets, and absent con- A professor of economics at San Jose State Univ rsity who was called as a rebutta expert by complaint CQunscl. 10 That is, on the basis of shipments data, you might designate two areas as separate geographic market areas when in point of fact, if there was a slight pri e increase in one, it would elicit shipments from the other, which would show that the two areas in fact should have been con!1trued aB one geographic area (Tr. 11,998-99). Initial Dccision 103 F. vincing countervailng arguments, I believe that it should be given primary consideration in deciding this issue. I now turn to the countervailing arguments presented by complaint counsel's experts. (b) Dr. Mann 127. Dr. Mann relied on the following facts in arriving at his conclusion that the relevant geographic markets for regular coffee were larger than a Maxwell House Division sales district: (a) Statements in Maxwell House Division documents which recognize different levels of trade dealing between the sales districts (Tr. 3530-33). (34) (b) The fact that the "economic" price and the "dead net" price of Regular Maxwell House, differ between sales districts and that they do not converge as one would expect them to do if the sales districts were not separate geographic markets (Tr. 3536-2). (c) The appearance of different competing sellers in various sales districts (Tr. 3553).

(d) Restrictions on transshipment by the Maxwell House Division (Tr. 3554).

(e) Retail distribution patterns as revealed by the Lysacker study which indicate that 90% of regular coffee stays within the confines of a sales district (Tr. 3564-5).

(c) Dr. Greer 128. Another fact, product differentiation, was considered by Dr. Greer, for he believes that the scope of the geographic market can be limited by this phenomenon (Tr. 11 512). Thus, according to him, if one finds that a particular brand of coffee is the leading brand in a limited geographic area, one could conclude that this area constituted a relevant geographic market (Tr. 11 602).

5. Conclusion 129. After considering all of the expert testimony on this point, I have come to the conclusion that the Elzinga-Hogarty test describes more accurately than any other fact or set of facts relied upon by complaint counsel's expert witnesses the primary forces which determine the relevant geographic market for regular coffee. 130. The only fact relied upon by Drs. Mann and Hilke which even suggests that the market for regular coffee is regional is the claimed difference in the economic and dead net prices of Regular Maxwell House between sales districts, for one can expect a convergence of prices within a relevant geographic market. However, after analyzing the testimony ofthese (35) gentlemen, I have decided that Dr. Elzinga 204 Initial Decision is correct: price data is so complex and ambiguousll that it is oflittle practical utility in resolving the issue of relevant geographic market (Tr. 11 992).

131. The other facts relied upon by Dr. Mann are of no significance. The statements by Maxwell House Division employees and other industry members relied on by him reveal only that demand for coffee differs among sales districts; they indicate nothing about the supply side of the equation, the issue of most importance. The same is true with respect to retail distribution patterns. Even ifthe Lysacker study is valid 12 it reveals only the pattern of shipments from warehouse to retail store. It does not tell me which roasters retailers can turn to for their supply of regular coffee. Furthermore, the pattern of distribution disclosed in the Lysacker study is changing. General Foods is unable, as a practical matter, to control trade flow and transshipment and more grocers are expanding their area of distribution across sales district boundaries. Finally, the fact that different sellers appear in different sales districts means only that they choose to restrict their area of distribution, not that they are limited to those areas by market forces.

132. Dr. Greer s injection of product differentiation as an additional factor leads to a rather bizarre result--ach seller of a branded product would operate within a geographic (36) market defined by his area of distribution (Tr. 11 520). Whatever the merits ofthis approach may be when other issues are considered, antitrust analysis would be impossible ifone accepted this theory. Furthermore, Dr. Greer agreed that the implication ofthis theory is that the geographic market for Regular Maxwell House is nationwide since it is sold nationwide (Tr. 11,521), a point which complaint counsel obviously do not want made.!3 133. In conclusion, I accept the Elzinga-Hogarty test as a valid means oftesting whether or not a given area is a relevant geographic market. The Youngstown, Syracuse and Philadelphia sales districts individually and in combination fail the test by considerable margins. For this reason, and for the reasons given in my conclusions of law infra I find that neither these sales districts, nor the other Maxwell 11 For example, price differences may be the result of different marketing mixes in use in different areas of the country or distribution cost. (Tr. 9466-7). Furhermore, analyzing the price of Reguar Maxwell House tells one nothing about the prevailing market price ofthe coffee sold by all producers and relying OIl the price of onc brand of coffee might result in different geographic markets for each brand of coffee (Tr. 11 994). Finaly, price data is subject to varying interpretations. ex 1380, which compares Regular Maxwell House s average annual prices in the Youngstown and Syracuse sales districts from fiscal 1971-1977 disproves, according to Dr. I-Iike, a national market (Tr.l0 929-30). However, Dr. Elzinga pointed out that ifone looks at the prices over the seven year period they differed only by 1% (Th. 11 990).

12 General Foods argues that it is seriously flawed (RPF 4-\07 to 4-110) but I need not decide this question because its conclusions are of no relevance in the circumstances of this case. Complaint counsel argue that. Dr. Greer s swtement can be "misconst.rued" (CRF 3-1) but I believe that the implication of his theory is as he stated Initial Decision 103 F. House Division sales districts (or any smaller geographic areas such as the Cleveland or Pittsburgh market areas) are relevant geographic markets for the sale of regular coffee.

134. Since complaint counsel allege that General Foods attempted to monopolize the sale of regular coffee only in certain sales districts further analysis of their Section 5 claim is unnecessary, for General Food' s market power or its sales below cost in individual sales districts are irrelevant in terms of their effect on competition in the actual relevant geographic market which is much more extensive than the sales districts. Nevertheless, the parties are entitled to findings on the other issues presented by complaint counsel for purposes of possible appeal, and I will discuss them at some length. 1. The Alleged Attempt To Monopolize 1. The Definition Of Price (a) Introduction 135. Before discussing complaint counsel's claim that General Foods attempted to monopolize "the Eastern ground coffee markets (CLA, p. 1) by engaging in price and non-price predation, one must decide what price one wil use in analyzing that claim. (37) 136. The Maxwell House Division sells its regular coffee to wholesalers and retailers, but the price which these customers pay (and the revenues which General Foods receives) is not solely a function of its list prices, for price is also affected, if one does not view them as costs by the allowances which General Foods offers to its customers in trade deals (non-performance and performance allowances) and to its customers' customers, the ultimate consumer, in the form of consumer promotions (RECUs and consumer coupons).

137. The definition of price may also be affected by the purpose for which it is used-that is, whether one is analyzing complaint counsel's attempt to monopolize or Robinson-Patman claims. The present discussion deals with the former claim. The Robinson-Patman price is discussed more fully in the section dealing with that statute. (b) Economic " Price 138. Dr. Mann testified that "economic price" equals average revenue received by the seller per unit, and that this price is calculated by deducting from net sales revenue "other forms of price reductions (Tr. 3674).

139. In calculating price, economists do not treat costs, such as advertising, as reductions in revenue, and costs do not, therefore affect that evaluation (Tr. 3687).

140. The parties agree that non-performance allowances are price 204 Initial Decision reductions and that they should be deducted from list price. The parties disagree on the proper treatment of the other promotional allowances which the Maxwell House Division offers to its customers and to the ultimate consumer. Complaint counsel argue that these allowances are price reductions whereas General Foods urges that they are costs and should not be deducted from list price. 141. Dr. Mann testified that the economic price for Regular Maxwell House equals net sales less promotions divided by volume. He included all trade and consumer promotions in his definition of promotions (Tr. 3676). Professor John Dearden, who computed the profitability of the Maxwell House Division s regular coffees agreed that promotions are price reductions, whether they are offered to the retailer or to the consumer (Tr. 66-7). (38) 142. Dr. Elzinga offered a definition of economic price which differs markedly from that proposed by complaint counsel's experts, for he would not treat as a price reduction any offer which is not unambiguously welcomed by a customer (Tr. 9462--3). Since some customers turn them down, Dr. Elzinga does not, therefore, view consumer coupons or performance-related offers as price reductions (Tr. 9464). 143. Dr. Greer was called in rebuttal by complaint counsel. He testified that Dr. Elzinga s treatment of consumer coupons and performance-related offers is inconsistent with the law of symmetry, that , that a price reduction to the buyer must result in a reduction in the price received by the seller (Tr. 11,338, 11 344-6). 144. I see no reason why Dr. Mann s definition of economic price as revenue received should not be accepted for Section 5 purposes. Given this, I cannot agree with Dr. Elzinga s analysis, for his test proposes that the definition of price depends on the type of promotional allowance which is offered, and to whom it is offered, even though, from General Foods' point of view, its revenue is reduced by the same amount whether a 10% promotional or buying allowance is given to a retailer, or a ten cent coupon is redeemed by a consumer. 145. I therefore agree with Dr. Greer that coupons given to consumers by the Maxwell House Division, either directly or through retailers, reduce the consumer s (but not the retailer s)14 purchase price (Tr. 399) and that, to the extent they are redeemed, they reduce General Foods' revenue (i. Dr. Mann s economic price) by a corresponding amount. The same logic also applies to performance offers which are accepted by the trade, for to the extent they are accepted, they reduce the Division s revenue.

146. In conclusion, I believe, with some reservations, that the Maxwell House Division s revenue, as reduced by its non-performance and 1\ See my disr.ssion infra of the definition of RobinSOD-Patman price 238 EDERAL TRADE COMMISSION DECISIONS Initial Decision 103 F. performance offers and its consumer coupons, is a price which may be used to analyze the attempt to monopolize charge. 147. The reservations I have with Dr. Mann s and Dr. Greer s definition of price is caused by the fact that the (39) definition assumes that retailers always accept promotional offers and that consumers redeem every coupon. This is demonstrably untrue. While there is a high rate of acceptance ofthe Division s performance offers, some are not accepted (Tr. 2391). The rate of coupon redemption is much lower. Complaint counsel's accounting expert, Mr. Rowe, estimated that it is 33% (Tr. 2392). Thus, one cannot assume that all coffee sold under a particular mix of promotional offers is sold at the same economic price. For example, if the Maxwell House Division sells one-pound cans of coffee for 90 cents and offers a 10 cent coupon to consumers General Foods' revenue, and the "economic price" ofthat coffee is 80 cents only to the extent that the 10 cent coupon is redeemed. If only 10% of coupons are redeemed, then only one out often cans of coffee are sold at the 80 cent economic price. The other nine are sold at a higher price.15 (c) Dead Net" Price 148. While complaint counsel state that: "Economic price provides a more accurate measurement than dead net price of the prices that respondent actually received " some oftheir price analyses, including their Robinson-Patman analysis, use the latter price, defined by them as "total Maxwell House sales revenue, net of trade deals (buying allowances, performance allowances, and retailer coupons), divided by sales volume" (CPF 12-1; CLA, p. 87, n. 312). 2. The Maxwell House Division s Response To Folger s Eastern Expansion (a) Cleveland 149. At the time of Folger s expansion into the HFNI, Mr. Laster was president of the Maxwell House Division. (40) Although he was charged with developing a plan of defense, it had not been developed when, in October 1971, Folger entered Cleveland, and the Division response was modest (Tr. 2743 6952-53).

150. Regular Maxwell House s share in Cleveland at the time of Folger s entry was 43.5%; Hills had a 22.4% share; Chase & Sanborn 10 The omewhat theoretical nature of"ecanomic price " was recognized by Dr- Mann when he WID asked ifit was "the actual price paid by any customer of General Foods," He replied: "It is what the average customer pays. If you pick one single customer out in the real world this may not he the price paid, but it is what the average customer pays" (Tr. 3676). This colloquy also reveals Dr. Mann s diffculty in keeping his definition of price cOIlistent, for his definition of price as "revenue received" is not nece.srily equal to the price which the Division customers pay (See Finding 448, n. 41).

204 Initial Decision had a 9.8%; and COB share was 17.1%. Because of the trade s emphasis on "hot features," share levels had fluctuated significantly: Regular Maxwell House from 33% to 52%, Hils from 14% to 37%, and Chase from 6% to 15% (CX 85B).

151. Mr. Laster recalled that Folger s came in "with a very good product" they believe in quality as we do," a very strong market plan using proven advertising at high levels, consumer promotions at significant levels" and "trade deals that were significantly higher than the deals that were in the Cleveland area prior to their coming (Tr. 6958). Folger s consumer promotions also included delivery offree coffee to homes as a sampling device (Tr. 6958). General Foods perceived the Procter & Gamble introduction as "higher than a normal introduction would warrant" (Tr. 7040).

152. On the other hand, Mr. Hunter, Folger s general manager at the time of the Cleveland test, believed that its planned spending was similar to that of recent Hils Bros. and Taster s Choice introductions (Tr. 2734-38). Folger s initial share objective in Cleveland-20%and its 22-month payout!7 objective were considered to be "about on the line" and "modest" by him (Tr. 2711, 2713, 2738, 3022). 153. On October 8, 1971, the Maxwell House Division computed Folger s net cost to the trade (not counting RECU's) as lower than Regular Maxwell House s (CX 85E), and another Division memorandum reported that Folger s introductory rates on two pound and three pound were "considerably higher than the traditional rates for Cleveland" and, in fact, were 100% greater than the traditional Cleveland trade rates (CX 694A).

154. In November 1971, Mr. Laster and Mr. Nelson, the Division national sales manager, made a personal inspection of the Cleveland area. Mr. Laster concluded from the inspection that "there was no question in my mind that Folger s had established itself in the marketplace at an extremely high level and that their brand in a very short period of time had become (41) extremely successful" (Tr. 6961). Mr. Nelson reported in a November 16, 1971 memorandum that Folger s introduction into Cleveland had been "disturbingly successful" and that Folger s had obtained an "outstanding coffee section" in many stores (CX 122). The same point was reported by Mr. Tower, the Division s marketing manager in a January 3, 1972 memorandum to Mr. Tanck, then ground category manager of the Division. He reported that Folger had attained excellent distribution of its cofl"ee, and had an excellent retail shelf position, in some cases better than Regular Maxwell House. Folger s shelf prices and feature prices were equal to Regular Maxwell House, it was experiencing strong trial by con- 16 Chai!J own brand, " or private label cofree J7 "Payout" is the time needed to recoup introductory investment (Tr. 2713) Initial Decision 103 FTC. sumers, was employing an effective advertising copy and had an excellent package and product (CX 692A).

155. Folger s perceived initial success in Cleveland convinced Mr. Laster that it would soon be a national brand and that if Regular Maxwell House wanted to maintain its share position it had to adopt a different posture than that adopted during Folger s first three months in Cleveland (Tr. 6967). Mr. Laster felt it "was important not to have (Folger) expand rapidly to the rest of the East before we have thoroughly learned from the Cleveland test market. . . . " (Tr. 6979), and he felt that it was important to defend so that Procter & Gamble would realize that General Foods "would not be the Scott Paper ofthe coffee industry" (Tr. 6978).8 156. Consequently, on March 23, 1972, Mr. Laster presented a recommendation to General Foods' management on its response to Folger s national expansion (CX 130A-I; Tr. 6967). 157. CX 130A is a cover memorandum dated March 23, 1972, from Mr. Laster to Mr. R. Bohm. At the time of this memorandum, Mr. Bohm was executive vice-president of General Foods with line responsibilities for the Maxwell House Division (Tr. 6950-51, 6966-7, 7029). Mr. Laster s memorandum states:

Attached you will find a deck showing assumptions, strategies and calculations for the various Folger defense options. (42J Basically, this covers some of the material discussed at the recent Operating Committee meeting. Plea."e let lis know if we can elaborate on the attached material (CX ) 30A). The attached "deck " CX 130B- , is entitled "Folger s Defense Options" and is dated March 16, 1972 (CX 130B). 158. The "Operating Committee" was a senior management group in General Foods that was used as an "advisory group" to Mr. Bohm when he had to make significant decisions (Tr. 6967, 7029). Mr. Bohm initiated the formation of the Operating Committee, which was made up of many ofthe high-ranking executives of General Foods, including some division presidents and corporate staff personnel. Mr. Bohm was chairman ofthe Operating Committee at the time of CX 130A-I (Tr. 4545-46).

159. CX 130B-I is an analysis presented to General Foods' corporate management which shows the financial impact of following the options of ('not defending, defending now " or "defending later" (Tr. 6963-64).

160. On the basis of CX 130B- , Mr. Laster recommended to the IS A reference to the experience of the Scott Paper Company when Procter & Gamhle c)(panded into ilg "uHkct" You look at Scol! Paper or a couple of other companies that were fairly passive in their re3ctiolJ, they litemlly almost went down the tubes (Trone, Tr- 14118) 204 Initial Decision Operating Committee that General Foods "defend. . . from the very first day that P&G moves into new markets. . . ." (Tr. 6964). The defend now" strategy described in CX 130A-I was adopted at the Operating Committee meeting (Tr. 6994, 7031) 161. Under the "defend now" strategy, the study did not project losses on a national basis but predicted a loss of $4 milion in the HFNI sales districts in year one. A positive profit before taxes of $6 milion was projected for year two and $7.5 milion was expected in year three (CX 130F).

162. In Mr. Bohm sjudgment, the best strategy for Maxwell House to follow in response to Folger s expansion was the "defend now strategy, and the figures set forth on CX 130F were one ofthe reasons for his judgment (Tr. 4547). The maintenance of Maxwell House level of profiability in the high franchise non-infringement sales districts was a factor that influenced Mr. Bohm in his conclusion (Tr. 4547-48). Mr. Laster also testified that an objective of the Folger defense was to maintain the long term profitabilty of regular Maxwell House (Tr. 7011). He explained:

lMJaxwell House coffee is a very important division to General Foods and as such, the profitability of our brands is highly significant to the corporation (Tr. 7012). (43) 163. The Maxwell House Division expected its "defend now" strategy to increase future division-wide profits ofthe Maxwell House Division by several milions of dollars as compared to the profits that would be made if the other proposed plans were followed. For fiscal years 1973 through 1977, the "defend now" strategy would result in $16.2 million dollars more profit than the "no defense" strategy, and $23.1 million more than the "defend later" strategy (CX 130I). By fiscal 1977, the Division would earn $77.2 milion under the "defend now" strategy, as contrasted with $62.6 milion and $66.5 milion respectively under the "no defense" and "defend later" strategies (CX 130G).

164. While the "defend now" strategy outlined in CX 130 shows a higher national profitability than either the "no defense" or "defend later" strategies, the projected profits resulting from the "defend now" strategy were lower than Regular Maxwell House s "base" profits, or profits prior to the Procter & Gamble expansion. Prior to Folger s expansion, Regular Maxwell House s national !!base" profit was $19.8 milion. This base profit would be expected to decline to $8 milion in the third year of a "no defense" strategy, to rise to $11 million in the third year of a "defend later" strategy and to equal $15.5 milion in the third year of "defend now" strategy (CX 130F). 165. The Maxwell House Division also expected its response to Initial Decision 103 F. Folger s entry to increase Regular Maxwell House s market shares and limit Folger s market shares in the HFNI sales districts: Regular Maxwell House and Folger s wil be able to achieve the following on-going HFNI shares under the alternative defense options: Share Base Plan No Defense Defend Late Defend Now RMH 39% 30% 39% 42% Folger 20% 20% 10% (CX 130C).

166. Although the plan described in CX 130 was accepted by management, plans such as these were revised based on new (44) information and experience. For example, as a result of an ongoing debate within General Foods on whether it was necessary to match Folger marketing levels, when Folger s expanded into Syracuse General Foods abandoned the strategy of attempting to meet Folger s trade deal rates in favor of an approach that would result in less trade promotion expenditures for Regular Maxwell House (Tr. 6994). 167. A later document prepared between mid-1974 and mid-1975 attempted to quantify the financial impact of the "defense" versus no defense" options (Tr. 8231-32).

168. Under the defense option, Maxwell House anticipated losing $100 000 on a national basis in year one (RX 518E; Tr. 8245, 8866), but its profit would be $43 milion by year ten (RX 518E; Tr. 8867). 169. Under the defense option between year one and year ten Maxwell House s trade deal rate per unit would decrease by about 250/0, its consumer promotion rate per unit would decrease by over 50%, and its advertising rate per unit would decrease by over 50%. Its total marketing expenditure rate per unit would decrease by almost one third (RX 518E; Tr. 8867-69).

170. RX 518A also describes the financial and market share impact of defending versus not defending against the Folger introduction. Under the defense option, Maxwell House s market share would increase by 5% (RX 518A-B; Tr. 8238, 8865). Under the "no defense option, Folger would reach a 25% share and Maxwell House would lose market share in proportion to its share at the time of entry (RX 518A-B; Tr. 8238, 8859-60). Under the defense option, Maxwell House s volume would increase from its pre-entry levels by 530 000 units in year one, by 505 000 units in year two, and by 480 000 units on an on-going basis (RX 518A; Tr. 8239, 8865). Each unit represents twelve pounds of coffee (Tr. 8239).

171. RX 518A indicates that a decision to defend against Folger entry, as opposed to not defending, would generate a profit for General Foods of 34% in terms of incremental ROFE before taxes, which was "favorable when compared to RMH's (then) present ROFE of 204 Initial Decision 22%" (RX 518A; Tr. 8235). "ROFE" refers to return on funds employed (Tr. 8235).

172. At the beginning of the March quarter of fiscal 1972, the Division concluded that it had to match Folger s marketing levels in Cleveland (Tr. 696). This matching strategy was adopted because a disparity in trade deals would put Regular Maxwell House "in a very significant disadvantage" (Tr. 6977). Disparity in trade deal rates could result in a disparity in (45) shelf price so that Regular Maxwell House would be sellng at a premium versus Folger s. This would result in Folger s being featured more frequently and aggressively by the trade, and would increase the trade s purchases of the Folger brand because it would be a more profitable item (Tr. 6978). 173. General Foods imposed two restrictions on the Division s ability to respond to Procter & Gamble. First, all sales of Regular Maxwell House during a four-to-six week trade promotion period had to exceed its reasonably anticipated variable costs (as defined by General Foods) (Tr. 7051- , 9073-74). Second, the Division was not to "aggress, which meant that Regular Maxwell House could not lead any element of the marketing mix and could not have higher levels of consumer promotions than Procter & Gamble (Tr. 6973). 174. Documents written by Division personnel during the Cleveland test reveal that the Division was attempting to meet Folger s trade dealing, and, hopefully, to de-escalate it (CX's 14P, 1l0A, 633J, 643H 644W, 649D, 651E).

(b) Defense Expenditures In Cleveland 175. In January 1972, the Maxwell House Division launched what Mr. Hunter termed a "heavy counter-attack" (CX 494). The counterattack included the use ofa $.50 mailed coupon, as opposed to Folger introductory $.35 coupon and a planned follow-up coupon of $.10 (CX 494; Tr. 2086, 2745). The Division also used in-pack coupons of 10 cents, 29 cents and 39 cents on one, two and three pound coffees, as opposed to Folger s 25 cent and 35 cent coupons on two and three pound coffee (CX 494; Tr. 2745-46). Third, the Division offered a trade allowance on the three pound size of $2.50 a case, a $.50 retailer coupon and a $.50 display allowance. According to Mr. Hunter, the $2.50 trade allowance was substantially higher than the allowance previously ofiered in the market (CX 494; Tr. 2745-47). 176. The Maxwell House Division s response must, however, be considered in light of Folger s distribution of free coffee to Cleveland consumers (Tr. 6291). Furthermore, Folger s advertising expenditures in its first year in Cleveland were $596 000, while Regular Maxwell House s advertising expenditures during the same period were actual- . . . . , Initial Decision 103 F. ly less ($540 000) (CX's 553A, 1389). Folger also exceeded Regular Maxwell House s per case advertising rate (CX's 553A, 954). (46) 177. Complaint counsel argue that these and other comparisons are meaningless and they offer alternative methods of assessing the Division s response which take into account, they claim advertising weight" and the number of times each firm used coupons (CRF 1to 1-80).

178. While it may be true that "For the entire year (in Cleveland) Maxwell House had 53% more GRP's than Folger" (CX 450K) (and even this is disputed by General Foods (RPF 6-86)), complaint counsel do not explain the significance ofthis fact. With respect to the number of times a coupon is used, one would expect the company with the larger market share to spend more money because it has greater volume than the new entrant. In fact, complaint counsel argue that the reasoning behind the Division s western retaliation (infra): (W)as clear. General Foods' response to Folger s introduction in Cleveland had cost it a great deal of money. The Western Retaliation was devised so that General Foods could go into highly profitable Folger markets and equalize those financial expenditures (CPF 5-11).

179. In another proposed finding (CPF 5-5(a)), complaint counsel argue that the purpose of the retaliation was "to force Folger to increase its spending." Mr. Trone explained the reasoning behind the retaliation:

A. Well, as Folger s came into the east and came into our high volume, high profit centers, the idea was to go into the west. into one of their markets that they had very high share levels in and were very dependent on profits. Q. How would this. strategy help to achieve the objective of increasing Folger payback? A. Because our volume base was so small and their volume base was so high in those areas; in order for them to meet our deal rate structures, they would have to spend an awful lot of money. ... (Tr. 1399-1400). (47) 180. It seems inconsistent to me to argue that the Division s defense expenditures in response to Folger s blitz in the HFNI, however measured, were inordinate while arguing that Folger s expenditures in the West were "forced" by the Division s retaliation. Thus, while there may be several valid ways to estimate the intensity of the Division s response, I believe that the one proposed by General Foods offers as much insight as do complaint counsel's compa isons. Using these, even as expanded by complaint counsel (CRF 1-81), one sees that while the Division spent more than Folger for consumer promotion and advertising after the introductory period, it was not, with one exception (1-3/72 in Cleveland) by very much. , 204 Initial Decision (c) Pittsburgh 181. At the time of Folger s expansion into Pittsburgh, Regular Maxwell House s share was 45%, the Breakfast Cheer brand had an 18% share and Hils had a 9% share. Breakfast Cheer, according to a Maxwell House Division memorandum, enjoyed Ustrong consumer/ trade acceptance" in Pittsburgh and was expected to "vigorously detend" (CX 94B).

182. Folger s formal introduction into Pittsburgh in March 1973 consisted of case rates 4011 higher than Regular Maxwell House 40-day credit terms, and distribution of one pound free samples of Folger s coffee in the Johnstown/ Altoona area and six ounce free samples in the metropolitan Pittsburgh area (CX 57B). Under 40-day credit terms, the grocery trade received a 2% prompt payment discount if they paid Folger s within 40 days. The prevailing prompt payment terms in Pittsburgh were 2% - 10 days, which offered a 2% discount only up until the 10th day (Tr. 8580). Following its initial offerings, Folger then escalated its trade deal rates to 40 to 50% higher than prevailing rates (Tr. 6981).

183. Due to the heavy cost of distributing free coffee to consumers General Foods did not match Folger s free coffee program. Instead General Foods responded to the Folger s six ounce free sample with a 2511 DMCU and to the one pound free sample with a 3911 DMCU (CX 169A). General Foods did not respond to Folger s more attractive payment terms, nor did it respond to Folger s increased introductory trade rates (CX 169A). As in Cleveland, the objectives of General Foods in Pittsburgh as transmitted to the Maxwell House Division national sales manager were to maintain Regular Maxwell House share there (Tr. 5772). Regular Maxwell House was allowed to meet but not (48) beat Folger s trade rates, with the overriding restriction that it was not to be sold below cost (as defined by General Foods) (Tr. 5774 , 5785; see also CX' s 57C, 150C, 735A). (d) Philadelphia 184. Procter & Gamble introduced Folger s into Philadelphia in February 1973. Procter & Gamble recognized that expansion into Philadelphia was a "risky venture" because the grocery trade was selling coffee at five to twelve cents a pound below its "dead net cost" (CX 522; Tr. 3071-73). " Dead net cost as used in this exhibit, refers to list price less trade deals only and does not take RECU's into account (Tr. 3073). It was anticipated that the grocery trade would therefore, refuse to buy Folger simply saying we don t want another brand of coffee to help us lose money" (CX 522). 185. Procter & Gamble s introduction into Philadelphia used strong Initial Decision 103 F. trade promotions and strong advertising and consumer promotions including distribution of free coffee to 80% of the households in the area (Tr. 6980). As in Pittsburgh, Folger s trade rates were 40 to 50% higher than the previously prevailng rates (Tr. 6981). (e) Defense Expenditures In Pittsburgh And Philadelphia 186. According to Mr. Hunter, after Folger s entry into Pittsburgh and Philadelphia, the Maxwell House Division employed a series of repetitive coupons, including several direct mail coupons, that far exceeded the amount of couponing used by Folger (Tr. 2813). 187. During the first year of Folger s entry into Pittsburgh and Philadelphia, General Foods, according to complaint counsel, continued to have a strategy of advertising more heavily than Folger (CX' s 462P, Q, W, 480D).

188. General Foods responds that in Pittsburgh from Aprij-September 1973, Procter & Gamble outspent Regular Maxwell House on consumer promotions ($1 027 000 vs. $532 800) and that during the first year of its expansion, Procter & Gamble s expenditures on consumer promotions were $1 131 000 as compared to Regular Maxwell House s $665 000 (CX's 553E, 956, 1389). (49) 189. General Foods also points out that Folger exceeded Regular Maxwell House s consumer promotion expenditures on a per-case basis in the April-June and July-September 1973 periods (CX 57B) and emphasize that Folger, unlike Regular Maxwell House, also had an extensive free coffee promotion in Pittsburgh. 190. As to advertising, Folger outspent Regular Maxwell House (CX' s 449Z-5 450Z-). Furthermore, Folger s trade rates in Pittsburgh and Philadelphia were in many instances higher than Regular Maxwell House s (Tr. 2847, 2878).1 (I) Syracuse 191. Prior to Folger s expansion into the Syracuse/Buffalo area Regular Maxwell House s share there was 41.5%. Other competitors in that area included Hills, with a share of 11.3% and Chase & Sanborn with a share of 13.7% (RX 1114A).

192. Procter & Gamble introduced Folger in Syracuse in October 1974 with trade deaJ case rates about four to five times higher than the historic rates and substantially greater than Regular Maxwell House (Tr. 5946, 6199). Folger did not offer RECU' s in Syracuse but instead focused on higher trade deaJ allowances that the grocery trade viewed as more attractive than RECU' s (Tr. 11 872). 193. General Foods did not match Procter & Gamble s trade deals I' This is because Folger began to meet the trade rates or the so-called "secondary brands" in these cities, and these brands generally offered Imlt"f trade deals than the leader (Tr- 2844 , 2846, 3051) 204 Initial Decision but went significantly "below them " meaning that the value ofRegular Maxwell House s trade deals were lower. According to Mr. Laster there was a "great debate" in the Division as to whether it was important to match Folger s trade deals during the introductory period, or whether the limited marketing funds available could be better spent on advertising and consumer promotions (Tr. 6991). 194. The Division responded to Folger s trade deals with trade deals of lesser value plus RECU's. This strategy was not regarded by the Maxwell House Division as a true match of Folger s trade deals and caused Regular Maxwell House to lose significant share in the Syracuse area (Tr. 6992). (50) 195. Folger s higher trade deals resulted in lower everyday retail shelf prices for Folger (Tr. 6105). Since Regular Maxwell House retail shelf price was higher than Folger, and because ofthe frequency of coffee purchases, this everyday shelf price disadvantage had a negative effect on Regular Maxwell House s business (Tr. 6105). Under the "rate-RECU" program employed by General Foods, Regular Maxwell House s consumer dead net was competitive with Folger s everyday shelf price only one week of a four- to six-week trade promotion period, provided that the grocery participated in the RECU offer and the consumers clipped the RECU from the newspaper and presented it for redemption (Tr. 5933-35).

196. Share data for Folger s third month in the Syracuse/Buffalo area show the effect on Regular Maxwell House of its strategy. In its third month, Folger s attained a share of 33.7%, while Regular Maxwell House s share declined from 41.5% to 36.7% (RX 1114B). 197. The refusal to match Folger s levels of dealing activity continued to cause erosion of Regular Maxwell House s share through the first half of 1975. A May 9, 1975 Maxwell House Division memorandum reported that Regular Maxwell House s share had reached all time lows due to poor feature activity and shelf price disparities: RMH' s share dropped 17.3 points to 30.3%, our lowest share to date. Folger s gained 1 points to 20.5% and C&S gained 9.0 points to 17.8%. Regular Maxwell House decline can be attributed primarily to poor feature activity and widespread shelf disparities in Buffalo (CX 751A).

198. Regular Maxwell House, according to an August 7 1975 memorandum, reached another record low share in June 1975 while Folger s attained an all time high share of almost 27%: RMH dropped 7. 2 points to 30.1 %, its lowest share to date and significantly below share objective. Conversely, Folger s gained 7.2 points for a new high of26.9%. RMH' s decline can be attributed, in part, to feature activity which while greater than Folger . . . (g) , Initial Decision 103 F. is (51) consistent with the share obtained. In addition, shelf pricing disparity continues to plague the Brand (CX 752A).

199. Due to this erosion of Regular Maxwell House s business, Mr. Einloth, in a July 10, 1975 memorandum, recommended that Regular Maxwell House convert to trade deal rates equal to Folger s levels and eliminate RECU offers (CX 710A). In July, the decision was made to terminate the rate and RECU strategy and to meet Folger s trade deal rates head on dollar for dollar" (CX 715A). Defense Expenditures In Syracuse 200. Complaint counsel claim that General Foods planned to employ a consumer promotion and advertising plan stronger than Folgs in Syracuse (CX 710m, that Maxwell House increased some trade deals after the announcement of Folger s introduction into Syracuse (CX 547A, C), and that, according to its advertising agency, General Foods out-advertised Folger in Syracuse (CX 462Z-12). 201. However, during the first three quarters of its expansion 198 000 was expended on Folger s consumer promotions by Procter & Gamble as compared to $769 000 on Regular Maxwell House for the same period (CX's 553G, H, 957). During the first year, Procter & Gamble spent $1 257 000 on consumer promotions as compared to $969 000 for Regular Maxwell House (CX's 553G, H, 1389). Procter & Gamble spent $1 380 000 to advertise Folger s during the first three quarters, while Regular Maxwell House s advertising expenditures during the same period were $720 000 (CX's 553G, H, 957). On a rate per case basis, Folger s also outspent Regular Maxwell House on consumer promotion and advertising in Syracuse (CX's 553G, H, 957). (h) Documents Discussing The Maxwell House Division Defense Efforts (1) Introduction 202. Complaint counsel rely extensively on contemporaneous statements by Regular Maxwell Division employees or agents which they say evidence General Foods' predatory intent (CPF's 3-1 to 3-184). There is no doubt that documents such as CX 130A-I and RX 518Aestablish what Division policy was at the time they were written. However, General Foods, citing statements (52) which I made when some documents were offered in evidence, warns that unless the author is known or unless complaint counsel establish that a proposal was relied upon, little weight should be given to statements made in the documents. General Foods also argues that inferences should not be drawn unless they are clearly justifiable (RPF 2-11 to 2-24). 203. All of this is sensible, but the fact remains that General Foods -.v 204 Initial Decision also relies on statements in many of the documents offered in evidence by complaint counsel, a recognition, I believe, that the documents, if analyzed carefully in light ofthe circumstances surrounding their production, offer some insight into its intent when it responded to Folger s eastern market thrust.

(2) Intent To Maintain Profitability And Market Share 204. CX 130A-I and RX 518A-E, discussed above, reveal that one goal of the Maxwell House Division s response to Folger s eastern thrust, using the "defend now" strategy, was to maintain the Division s profitability and market share, and perhaps even increase it even though that strategy might result in an initial loss. 205. Other documents indicate that this strategy was considered and discussed throughout Folger s eastern expansion. A contingency plan dated June 9, 1976, states that:

RMH is clearly in a better franchis, volume and fmancial position by defendingwhen Folger enters the marketplace (CX 640M) (emphasis in original). 206. A September 13, 1972, memorandum from Mr. Tower to Mr. Laster (CX 633A) attached a contingency plan which listed as objectives:

Provide a sound platform for restoring long-term profitability to RMH. Aggressively defend the HFNI against Folger s introduction and build share by 10+ %. (53) Prevent Folger from exceeding a Year I 13% SOM in HFNI (CX 633F). Mr. Einloth also wrote and confirmed that the objective of the defenses in Cleveland, Pittsburgh, and Philadelphia were to: Prevent Folger from exceeding a 13% share of market in non-infringement areas while building the RMH business by at least 10%.

Provide a strong platform for restoring long-term profitabilty to RMH (CX 725A; Tr. 1581).

207. A plan was developed by the Boston Consulting Group (CX 724A-G) to determine "What are the appropriate competitive and corporate goals for the MHD relative to the Folger Company" and How can Folger s geographical expansion be halted or extremely constrained and at what cost to GF?" (CX's 724A, 725A). Complaint counsel argue that this document unmistakably indicates predatory Initial Decision 103 F. intent, yet the plan, although endorsed by Mr. Einloth, was never implemented (Tr. 1622).

208. Contingency and defense plans written between 1975 and 1977 reveal the authors' hopes that by defending, the Division could maintain profitability and build Regular Maxwell House s share from 5-10% (CX's 640J, 709G, H, ?loa, D), and other documents written between 1971 and 1977 contain similar hopes (CX's 87 A, C, 659D). (3) Intent To Delay Further Expansion By Folger 209. Both Mr. Laster and Mr. Bohm testified that one of General Foods' defense objectives was to delay Folger s expansion into the eastern markets (Tr. 4537, 6984), and Maxwell House Division documents confirm that this was one of its goals (CX's 14C, F, 15E, 73H 74E, 76C, 80F, 150B, 168A). Statements in other documents suggest that delaying Folger s expansion would maintain profits in the eastern sales districts:

(tJhe value of delaying (Folger sJ expansion into the HFNI by two years (54) almost pays for our efforts (CX 724A; see also CX15W, 74Z-1, 80F, 82C, B8D, 19Z-4). 210. Some documents even reveal hopes by the authors that a vigorous defense would discourage further Folger s eastern expansion. While recognizing that "finite predictions are obviously impossible, Mr. Einloth, in a confidential memorandum stated that: If ground Folger were to match our spending levels, they would be placed in an LBT position of$12-15 (milion J annually without any additional expansion. Expanding into the HFNI would increase that to $25-35 (millon) (CX 724E; see also ex' s 14C, 86M, 163A) (4) Intent To Increase Payback 211. The fiscal 1973 marketing plan stated that one objective for Maxwell House was to "(iJncrease Folger s payback" (CX 14C) and Mr. Einloth confirmed that this was a Maxwell House objective (Tr. 1549- 50). The phrase "increase Folger s payback" meant that General Foods wanted to lengthen the time it would take Folger to break even after having entered a given market (Tr. 1608). 212. Another document indicates that Mr. Tower wanted to: develop plans that wil prove to Procter that entry into our high-franchise markets will be expensive for them (CX 121B; see also ex' s 74E, 77B, 87A, 133). 204 Initial Decision (5) The Statements In Perspective 213. The statements discussed just above must be considered in light of actual developments. Thus, the ambitious projection (55) that Regular Maxwell House could defend against Folger and increase its share in the HFNI was rejected by General Foods' management. Mr. Einloth testified that:

We chose (the goal of preventing Folger s from exceeding) 13 per cent. Because if everything else held constant, they got 13 in the HFNI and we built by 10, we would retain national share leadership by about one-tenth of one percentage point. That was my objective as far as I was concerned-I had to convince a lot of people that that was important. I couldn t convince them to defend, let alone maintain, national share leadership. That is how I arrived at the 13. MR. SPIEGEL: Did I understand you to say that you had to convince certain people at General Foods about these objectives? A. Damn right I did.

Q. Who were the people that you had to convince? A. Tower, Laster, primarily those two.

Q. Did you convince them'? A. No.

Q. You didn i convince them? A. Absolutely not. You know what General Foods' share of market is today'? Q. I am talking about at this point in time. A. If I convinced them of national share leadership'? Absolutely nol I did not (Tr. 1573-75). (56) 214. Mr. Einloth felt that maintaining national leadership was "absolutely vital " but this was an objective which he was not able to persuade management to adopt (Tr. 1613-15). The "real question" for management at the time of Procter & Gamble s expansion was whether funds should be spent for defending at all (Tr. 1615). According to him, it was a "constant battle" to sell management on the recommendation of the product group that leadership should be maintained. Eventually, the leadership issue "dropped out" (Tr. 1616). 215. Furthermore, despite the widespread belief among Maxwell House Division personnel that an immediate response to Folger entry into the test markets was essential, an effective response in Syracuse was not attempted until Folger s success was almost assured (Findings 191-99), an indication that management was still reluctant to commit itself to an all-out defense.

(i) The Reasonableness Of The Maxwell House Division s Defense Efforts 216. After considering the defense mounted by the Maxwell House Division in Cleveland, Pittsburgh, Philadelphia, and Syracuse, the Initial Decision 103 F. documents describing those efforts, and Procter & Gamble s prior success in marketing new products, I find that the Division s response to Folger s entry was not "inordinate" as claimed by complaint counsel.20 217. Complaint counsel view the Division s response as "inordinate" because new entrants must use high promotion and advertising levels. They claim that Folger s entry spending plans were reasonable in view ofthat requirement. It is true that when introducing a product in a new geographic area, an entrant must employ higher levels of advertising and consumer promotion than the already-established brands to convince consumers to try the new product; and higher levels of trade promotion are also needed to convince retailers to stock and market the product (Tr. 2070-72, 2091- , 2722- , 2726, 2731- 32). It does not follow, however, that the Division s decision (57) to meet Folger s expenditures in some areas was !!inordinate," for Folger was not the typical new entrant.

218. From July 1, 1970 through June 30, 1971, the period immediately prior to the start of Folger s expansion into the HFNI, Procter & Gamble s worldwide sales were $3.2 bilion, with net earnings of $238 milion (CX 642F). In fiscal 1972, General Foods' worldwide sales were $2.5 bilion, with net earnings of $113 milion (CX 438F). At the time of its expansion of Folger, Procter & Gamble marketed detergent products, food products, paper products and toilet goods under the following brand names: Tide, Cheer, Ivory, Joy, Safeguard, Comet Cascade, Mr. Clean, Spic & Span, Crisco, Duncan Hines, Jiff, Pringle, Charmin, Puffs, Bounty, Pampers, Crest, Gleem, Head & Shouldders and Prell (CX 642G).

219. Gordon Wade, who was employed by Procter & Gamble for nine years in various product management positions and was qualified as an expert in the sale of food and grocery products, testified that Procter & Gamble is the "best marketer of packaged goods in the world" (Tr. 4607). Although there have been some failures, Procter & Gamble has enjoyed great success in entering various consumer goods categories and attaining a leadership position while relegating prior market leaders to number two positions (Tr. 4607-11). 220. Mr. Wade s opinion is shared by the grocery trade. Mr. Engel regards Procter & Gamble as the "best packaged goods marketer in the United States" and testified that there was no way General Foods could keep Procter & Gamble s Folger s brand out of Cleveland or out of any other area Procter & Gamble desired to enter (Tr. 1702--5). 221. The management of General Foods also had great respect for Procter & Gamble (Tr. 4509). According to Mr. Laster, the president 20 I discuss infra complaint counsel's aJJegalions that the Division s below cost saes iw use of Horizon, its western retaiation and it. "Cora" campaign were illegal. , 204 Initial Decision of the Maxwell House Division everything we knew about Procter & Gamble, everything we knew about Folger s, would lead us to believe that we were up against a very strong and very formidable competitor" (Tr. 6955). In fact, Procter & Gamble "was the most formidable competitor we could have encountered" (Tr. 7045). 222. According to Mr. Nelson, the Division s national sales manager:

Procter & Gamble is a huge company. The Folger Division of Procter & Gamble is probably-probably represents 6 to 8 (58) percent of their total business. If they have a desire to move in to any piece of geography, they wil do it; and in all candor if they wanted to, they. could eat us up like a Hershey bar, and the street is littered in the grocery industry with competitors of Procter & Gamble who thought that they were not going to eat them up (Tr. 5773).

223. General Foods had already experienced one major encounter with Procter & Gamble in the cake mix business. Procter & Gamble bought a regional brand, Duncan Hines Cake Mixes, and expanded it nationally, gaining significant share and volume at the expense of General Food's Swansdown brand and other brands of cake mixes (Tr. 1971). General Foods shortly thereafter discontinued the marketing of Swansdown cake mixes (Tr. 6954).

224. General Foods' management was also aware that prior to its eastern expansion Folger was the leading brand in areas where it competed with Regular Maxwell House. In this area, Regular Maxwell House s share was about 9%, while Folger s share had increased from 31 % in fiscal 1967 to 40% in fiscal 1973 (eX 648Y). Furthermore management believed that when Folger s expanded into areas where Regular Maxwell House had a major or leading share, Folger volume grew at its expense. For example, Folger expanded into the Jacksonville district in fiscal 1959. Regular Maxwell House s share at the time was 57%, but by fiscal 1968, its share had declined to 42%, while Folger s share had increased from 5% to 22% (CX 109-E). 225. In 1959, Folger s expanded into the Chicago area with a "vigorous entry campaign" (Tr. 4503). At that time, Hills Bros. had the leading share of 35%, and Regular Maxwell House had a share of 25%. By fiscal 1972, Folger s had overtaken Hils as the leading brand, and Regular Maxwell House s share had fallen to 10. 1 % (CX 1072-B).

226. The strength of the Folger brand was evidenced by "frequent and repetitive" efforts to improve Regular Maxwell House s business in the west (Tr. 4499). Maxwell House Division efforts in the west were of such frequency and intensity that the A.C. Nielson Co. concluded in a report presented to the Division in mid-1971 that one reason for Folger s expansion eastward would be to force the Division Initial Decision 103 F. to focus on the defense of its high franchise areas and alleviate its competitive pressures on Folger s high franchise areas (CX 642Z4). (59) 227. Finally, because coffee contributed so much to General Foods profits (Tr. 4641-43) and was so much more important to it than to Procter & Gamble, as Mr. Hunter recognized (Tr. 3045), management was aware of the profound effect a successful eastern penetration by Folger would have on those profits.

228. After considering all of the facts- , Procter & Gamble prior marketing successes, the strength of its Folger brand in the west, the importance of coffee to General Foods, and Folger s perceived initial success in Cleveland Z! I find that while Folger as a new entrant into the HFNI had to offer consumer and trade promotions below the going rates in the test markets, General Foods was justified in defending its market share and profits vigorously after realizing that Folger s entry was successful. If it had not done so, management might have threatened General Foods with a substantial decline in profits, for Mr. Klein stated in discussing the financial justification for defending (RX 518):

I thought that the financial analysis was reasonable. It dealt with only the incremental impact, which is the proper way to deal with this kind of a decision, and in fact, it was a fairly simple decision to make. The magnitude of the numbers, the negative impact that we expected if we did nothing or did not defend, was so large that it made the decision to defend easy. It would be irrational not to defend (Tr. 8249). 229. Mr. Hunter agreed that the effect on General Foods would have been adverse if it had chosen not to defend: Q. . . . what would be the particular effects if General Foods had elected not to defend? A. There would be no question about it. Their Maxwell House business would have been adversely affected. It would have declined. (60) Q. Would that adverse efiect have been a significant adverse affect? A. Yes, I assume it would be. Because a proportionate loss of share would have been very substantial (Tr. 3043--4).

230. Under the circumstances described above, the Maxwell House Division s belated decision to defend its market share and profits by delaying Folger s expansion into the HFNI did not result in the expenditure of an c inordinate" amount of advertising and promotion funds for, given its substantial market share in the entry cities, it is inevitable that the Division would have to spend large sums to defend (just as its western retaliation would force Folger into large defense expenditures). And, I reject complaint counsel's argument that the 'I Whether, in hindsight, Folger was as successful in jL Cleveland introduction as General Foods claims (RPF 6-16 6-17) or only minimally successful, according to the Nielsen figures (CRLA, p- 10), management believed at the time that Folger was a significant threat (Findings 15.1-55) (j) 204 Initial Decision Division should have chosen a less expensive defense option, for the record reveals that any other decision than "defend now" would have resulted in excessive share and profit losses. The Probability Of Successfully Excluding Folger From The HFNI 231. The immediately preceding discussion leads to a further conclusion: while General Foods hoped to delay Folger s expansion into the HFNI by making that expansion costly, no reasonably knowledgeable member of the industry would have believed that it could be excluded from the HFNI. For example, Mr. Graham of Martinson said when asked if his company or any other coffee company could stop Folger s eastern expansion:

No. We didn t feel that there was any way to stop Folger s from coming into that part of the country when they decided to expand into the area (Tr. 10 556; see abw Findings 21&-25). (61) 232. This conclusion is borne out by the financial analyses justifying the Maxwell House Division s defensive efforts. None of these analyses assumed that Folger could be stopped from expanding into or be excluded from the HFNI. Rather, all of the financial projections anticipated expansion throughout the HFNI by Folger (Tr. 8248-51). 3. The Effects of Folger s Entry Into The HFNI And The Maxwell House Division s Response (a) Introduction 233. Folger entered the eastern markets with established profit and payout period objectives. After some initial hesitation in the first three test markets, the Maxwell House Division decided to compete vigorously with Folger for market share in these markets. As a result of the intense competition from the Division, Folger was not only unable to achieve its objectives on schedule but experienced losses in the eastern United States.

234. The intense competition between Folger and Maxwell House also resulted in losses for some roasters in these markets. Smaller companies such as Paul de Lima, Euclid and Hils were unable to compete against their larger and stronger rivals. Folger s entry and the Maxwell House Division s response were the major reasons for the decline ofthese companies, although there were other factors which caused the losses.

1 Whether General Food's profits were excessive and its defense caused it to sell its Maxwell House brand below average total or average variable cost in some of iL sales districts or market areas is a separate quc tion whjch is djscus ed infra Initial Decision 103 F. (b) The Maxwell House Division 235. In Cleveland, where Regular Maxwell House s pre-entry share was 44.6% (CX 1119C), its market share grew to 49.5% in the first year after entry, and then went to 46.9%, 48.5%, and 55.7%, respectively, for years two, three and four (CX 1119C). 236. In Pittsburgh, Regular Maxwell House s share grew from the pre-entry base of 44.2% to 51.5% and 53.2%, respectively, for years one and two (CX 1119E).

237. For the entire Youngstown sales district, Regular Maxwell House s market share grew from a pre-entry base of 42.0% to 45.3%, 45.7%, 49.2%, 53.3%, and 51.3%, respectively, for years one, two, ' three, four, and five (CX 1387).

238. In Syracuse, without the inclusion of Horizon s market share, Regular Maxwell House s market share moved from the (62) fiscal 1974 base of39.2%, the year prior to Folger s entry, to 41.7%, 38.4% and then 45.4%, respectively, for years one through three (CX 1387). 239. In these districts, Maxwell House was able to maintain its market share position relative to the number two brand (CX 1387). (c) Folger 240. Folger s eastern expansion and Maxwell House s response resulted in losses for Folger. Its comptroller, Mr. Clark, estimated that through June 1976, in Cleveland, Pittsburgh and Syracuse, Folger had cumulative losses of $10 369 000. In Cleveland, 57 months after its entry, Folger had a cumulative loss of $3 050 000 (CX 980). In Pittsburgh, 44 months after its entry, Folger had a cumulative loss of $3,174 000 (CX 981). In Syracuse, 21 months after its entry, Folger had a cumulative loss of $4 145 000 (CX 982). 241. The Maxwell House Division s response to Folger s expansion into the east was undoubtedly the main reason for the substantial losses incurred by Folger, but Mr. Hunter recognized that increased competitive expenditures by Chase & Sanborn and Hils Bros. in response to its entry also affected Procter & Gamble s ability to attain its objectives (CX 570A; Tr. 3050-51).

242. Although Folger suffered losses in the sales districts that entered, Mr. Clark testified that an analysis of the profitability of Folger s coffee is meaningful only on a national basis (Tr. 3143), and Folger s national sales and profitability reveals that it suffered no overall competitive injury (RX 1080C).

243. For the year ending June 30, 1970, Procter & Gamble s sales of Folger s coffee were 284.4 milion pounds. For the year ending June 30, 1976, sales of Folger s coffee had increased to 328.6 milion, an increase of 15% (RX 1069B). During roughly the same period, total 204 Initial Decision consumption of ground coffee declined nationally by 10% (RX H05B). During that same period, sales of Regular Maxwell House coffee declined by approximately 12% (RX H05B, H07A). 244. Procter & Gamble financed the expansion of Folger s coffee " the area of thefrom Folger Coffee Co. profits in its "Folger area country where it had the leading share (Tr. 2891). An exhibit specially prepared by Procter & Gamble for complaint counsel reported the " (63)following profits on Folger s in the "Folger area: Calendar 1972 $17 698.000 1973 $ 7 961 000 1974 $18 884,000 1975 $33 689.000 (CX 552A-B).

245. Maxwell House s response resulted in a longer payout period for Folger. Payout refers to the length of time it will take to reach a break even position on a cumulative basis. It is the time it takes for cumulative profits to offset initi ' introductory spending (Tr. 3214). 246. Folger s estimated payout for the plan implemented in Cleveland was 22 months (CX 491B; Tr. 2713). This estimate anticipated a response by the established brands (CX 491B), but the response by General Foods was not what was expected (CX 495A- , 520A). 247. Folger estimated that increasing its marketing effort to respond to a second round of increased Maxwell House activity would raise the Cleveland payout to 33 months (CX 497F, 518A). 248. By the end of 1974, approximately 38 months after entry, Folger had not achieved payout in Cleveland (CX 553I; Tr. 3246-7). Folger did not continue to calculate its payout for Cleveland. Howev- , given the volume performance and the amount of money Folger was spending in Cleveland, the payout would probably have been (Tr. 3103).according to Mr. Hunter "infinity" 249. The principal reason for Folger s inability to achieve its payout goal in Cleveland was the Maxwell House response (Tr. 278&-87). 250. After its experience in Cleveland, Folger increased the projected payout period for its expansion plan for Pittsburgh to 35 months (CX 498A, 538B).

251. Forty-two months after its entry into Pittsburgh, Folger had not achieved payout (CX 553, 981) The unusually strong response of Maxwell House was the key reason for Folger s lack of success in Pittsburgh (CX 539A-B).

252. General Foods concluded that Maxwell House s response had resulted in Folger s payout being extended to 6&-5 months (CX 649E).

Initial Decision 103 F. 253. While there is some testimony that Folger s projected payout was unrealistic (CX 493A; Tr. 4625-26), the Maxwell House Division response was certainly one cause of Folger s failure to meet it. (64) (d) Other Roasters 254. Folger s eastern expansion and the Maxwell House Division response severely restricted the ability of some other industry participants to compete and caused them substantial monetary and volume losses.

(1) Paul de Lima 255. The Paul de Lima Coffee Co. is a small coffee roaster in the Syracuse, New York area that reported a net worth of $512 900 in 1975 and a net worth of$547 636 in 1977 (RX 1072A; Tr. 2577-79). The de Lima Coffee Company marketed coffee to institutional customers to offce coffee service customers, and to grocery customers (Tr. 2503). 256. Prior to Folger s entry in 1975, de Lima sold regular coffee to retailers in a 60 mile radius around Syracuse (Tr. 2504). de Lima retail business was profitable prior to October 1974 (Tr. 2513). 257. de Lima lowered its price during the first nine months of1975 in an attempt to remain competitive with Folger and Regular Maxwell House (Tr. 2521). During the summer of 1975, de Lima realized that it was losing too much money in grocery operations. de Lima had to make a choice as to whether to continue sellng below cost, thereby endangering the whole company, or to attempt to price at a more realistic level and take the chance oflosing volume. de Lima made the decision to raise its wholesale prices to a level it considered more realistic (Tr. 2522).

258. After de Lima chose to stop selling its cofiee at a loss, the firm suffered an 80% drop in retail sales (CX 1068C; Tr. 2522). The low prices of Regular Maxwell House and Folger also caused de Lima to reduce its distribution area from a 60 mile radius around Syracuse to a 10 mile radius (CX's 607, 1068D; Tr. 2526-27). Because ofthe severe drop in sales, de Lima was forced to dismiss two of its three salesmen (Tr. 2527).

259. Because of the battle between Folger and the Maxwell House Division, de Lima s retail coffee business incurred substantial losses (Tr. 2536); however, the Division s response to Folger s expansion into the east was not the only cause of the losses suflered by de Lima. (65) 260. In the eight years prior to the Procter & Gamble expansion into the Syracuse sales district, de Lima s grocery sales of ground coffee had declined from 976 000 pounds in 1966 to 613 857 pounds by 1973 a decline of approximately 37% (RX 887).

261. Prior to Folger s expansion into Syracuse, de Lima had not , 204 Initial Decision advertised its coffee, had not used any consumer promotion devices or retailer coupons and only used trade promotions from time to time (Tr. 2506-8, 2566). According to Mr. de Lima (w)e have never been convinced of the merit of (retailer coupons and Consumer coupons) and we do not have the administrative facilities to control them" (Tr. 2506).

262. Mr. de Lima testified that one factor in his loss of business was the increase in the price of coffee caused by the Brazilian frost, increases which, because of his small inventory of coffee and inventory accounting system, resulted in immediate increasing costs for his coffee (Tr. 2586-7). Due to this increase in costs, Mr. de Lima s wholesale price was above the price of Regular Maxwell House and Folger (Tr. 2593); nevertheless, most of de Lima s sales losses were not due to the frost (Tr. 2597).

263. de Lima is still in the coffee business today and is the leading seller of institutional coffee in the Syracuse area (Tr. 2572, 2590). (2) Euclid 264. Since 1928, Euclid Coffee, located in Cleveland, Ohio, has sold its regular coffee products to retailers in all of Ohio and surrounding areas (Tr. 2601-02).

265. In 1970, Euclid's conee sales were about $2.25 millon (Tr. 2608), The Euclid Coffee Company was profitable in 1970, making about four or five cents per pound (Tr. 2608). 266. The increased levels of promotions and couponing by Maxwell House and Folger after its entry into Cleveland drastically reduced Euclid' s sales (CX 611; Tr. 2615). For example, Euclid's yearly sales to Fisher Foods, its largest retail customer, dropped from 14 580 cases (10 380 tins + 4200 bags) in 1971, to 2 045 cases in 1974 (CX 612A; Tr. 2621).

267. Euclid's sales drastically declined because its products were selling for a higher retail price than Maxwell House s and Folger and the product did not, therefore, move on' the shelf (Tr. 2616-17). (66) (3) Hils 268. During the period of Folger s entry, recommendations were made that Hils Bros. should meet the level of promotion existing in the Youngstown and Syracuse markets. Hills Bros. could not aflord to meet these promotions, because to do so would, in the opinion ofMr. Toy, result in the sale of its coffee below its cost of goods (Tr. 2213). 269. The efleet of the decision by Hils Bros. not to meet the Maxwell House and Folger promotions was that Hils Bros. suffered severe market share losses. For example, during 1971 through 1977, Hills Initial Decision 103 F. Bros.' market share in the Youngstown sales district dropped from a pre-entry level of15.8% (CX lO72A) to 8.5% (CX 1072G). In the Syracuse sales district, Hills Bros.' market share went from a pre-entry level of 11.0% (CX 1072D) to a 1977 level of 6.9% (CX 1072G). 270. Hills Bros.' market share losses were nearly 50% of its preentry level shares. Had Hills Bros.' market share losses been proportional to its previous share position, it would have lost less than 20% (CX 130C; RX 518A; Tr. 3010).

271. The pricing of Maxwell House and Folger was not the only cause of Hills Bros. decline, however, for prior to the time of Procter & Gamble s expansion into the Cleveland and Pittsburgh area, the grocery trade there did not regard Hils Bros. as a successful marketer of ground coffee because of its failure to maintain a consistent product quality and its lack of consumer promotions (Tr. 1698-1700). In fact Mr. Engel of Kroger regarded Hils Bros.' merchandising activities in Cleveland as "dumping" the product there when Hils Bros. had excess capacity (Tr. 1698).

272. According to Maxwell House Division salesmen, grocery trade factors in Syracuse also reported a decline in the quality of Hils Bros. coffee (Tr. 5992-93).

273. Hils Bros.' share has declined not only in the areas of alleged predation, but throughout all areas of the country. In Hils Bros.' three areas of greatest share strength, its share dropped by 24%, 39% and 48% during the period fiscal year 1972 through fiscal year 1980 year to date (RX 1117). Specifically, Hils Bros.' share declined from 32.7% to 24.8% in Chicago, from 23.7% to 12.4% in Detroit, and from 11.8% to 7.2% in Los Angeles (RX 1117). These three areas in fiscal year 1972 accounted for over 53% of Hils Bros.' total national volume (RX 1117). (67J 4. Regular Maxwell House s Market Power (a) Introduction 274. "Market power " which Drs. Mann and Elzinga agree is synonymous with the term "monopoly power" (Tr. 3630, 9501-D2) is, according to Dr. Mann, the ability of a firm to maintain price persistently above the levels that would occur under full competitive pressure (Tr. 3577 3630). Market power is identified by evaluating the structure of an industry within a relevant geographic market. This analysis looks at market share relationships, pricing conduct, and performance (profiability) (Tr. 3577- , 9498-9500).

204 Initial Decision (b) National Market Shares Of Regular Coffee Producers 275. In fiscal 1971, Regular Maxwell House s share of the national regular coffee market was 24.1 %; Folger s national share was 19. 1 % (RX 1107 A). 23 At the time of Procter & Gamble s expansion into Cleveland, its Folger s brand was sold in an area of the country that included 55% of the national population and that accounted for almost 60% oftotal ground coffee sales (Tr. 3016). Folger s had the leading share throughout that area, running about 33% to 34% (CX 491A; Tr. 3017). ID certain parts of its area of distribution, Folger s had over a 50% share (CX lO72A).

276. As a result of share gains from its eastern expansion, Folger became the leader in the national regular coffee market in 1979. This dominance continued in 1980 (RX's 1107 A, 1279). (c) Regular Coffee Shares In The East 277. In fiscal 1972, which marked the beginning of Folger s expansion into the eastern United States, Regular Maxwell (68) House share in that area was 38.5%; Folger s share was 5.8%. The latest Nielsen data shows that Regular Maxwell House s share has declined in that area while Folger s has grown substantially (RX 1279). (d) Regular Coffee Shares In The Eastern Sales Districts 278. Absolute market shares of General Foods' regular coffee brands and its Maxwell House brand in the eastern sales districts from 1971-1977 were:

Market Shares Of All General Foods Regular Coffee Brands Sales District 1971 1972 1973 1974 1975 1976 1977 Atlanta 37. 39. 44. 47. 48.4 42. 45. Boston 44. 45. 49. 50. 49. 47. 43. Charlotte 47. 47. 50. 49. 50. 48. 38. Jacksonville 44. 46. 48. 50. 47. 48. 45. New York 35. 36. 40.4 40. 40. 38. 34. Philadelphia 49. 47. 55. 55. 51.9 49. 41.4 Syracuse 41. 44. 48. 48. 51. 46. 52. Youngstown 46.4 49. 53. 56. 61.4 60. 55. 23 Although General Foods objects to any market alJalysis which focuses ouJy on Regular MaxweU House (RPF 3-1), Dr.Ebinga testjfied that anaJysia ofthjs brand's market power is apprupriate because it is General Food' flagship" coffee brand (Tr. 9787--8) Initial Decision 103 F.

Market Shares Of Regular Maxwell House Sales District 1971 1972 1973 1974 1975 1976 1977 Atlanta 32. 34. 37. 40. 39. 34. 38. Boston 40. 40. 40. 40. 39. 38. 35. Charlotte 41. 40. 41. 41. 41. 40. 32. Jacksonvile 41. 42. 42. 43. 39. 41. 40. New York 29. 29. 29. 28. 29. 28. 26. Philadelphia 45. 42. 45. 45. 41. 39. 33.4 Syracuse 36.4 37. 40. 39. 41.7 38.4 45.4 Youngstown 42. 45. 45. 49. 53. 51. 48. 279. The latest Nielsen data reveal that Regular Maxwell House share of regular coffee sales in the Youngstown, Philadelphia and Syracuse sales districts has declined from its market share in those sales districts at the time of Folger s entry into each ofthem, whereas Folger has achieved a substantial share in each of these districts (Compare Finding 278 with RX 1279). (69) 280. The shares of Regular Maxwell House and its next largest competitors in the eastern sales districts from 1971-1977 were: Market Shares Of Regular Maxwell House And The Next Largest Competitor Sales District 1971 1972 1973 1974 1975 1976 1977 Atlanta RMH 32. 34. 37. 40. 39. 34. 38. #2 Brand 15. 12. 12. 11. 11. (CS) (CS) (CS) (CS) (CS) (CS) (CS) Boston RMH 40. 40. 40. 40. 39. 38. 35. #2 Brand 10.

(H)' (CS) (CS) (CF) (H) (CF) (CF) Charlotte RMH 41. 40. 41.8 41. 41. 40. 32.2 #2 Brand 14. 13. 12. 12. 11. 12. (CS) (CS) (CS) (CS) (CS) (CS) (CS) Jacksonville RMH 41. 42. 42. 43. 39. 41. 40. #2 Brand 20. 21. 20.2 17. 19. 14. 14. #2 Brand (F) (F) (F) (F) (F) (F) (F) RMH 29. 29. 29. 28. 29. 28.4 26. #2 Brand 13. 14. 12. 14. 18. 18. 12. (S) (S) (S) (S) (CF) (CF) (CF) Philadelphia RMH 45. 42. 45. 45. 41. 39. 33.4 #2 Brand 10. 10.4 10. 10. (H) (CS) (CS) (CS) (F) (H) (F) , p.

204 Initial Decision Sales District 1971 1972 1973 1974 1975 1976 1977 Syracuse RMH 36.4 37. 40. 39. 41. 38.4 45.4 #2 Brand 17. 17. 14. 14. 11. 17. 21.7 (H) (CS) (CS) (CS) (CS) (F) (F) Youngstown RMH 42. 45.3 45.7 49.2 53.3 51.3 48. #2 Brand 15.8 12.5 11.3 13.8 14.7 13.1 16. CS = Chase & Sanborn; CF = Chock Full D' Nuts; F = Folger; H = Hils; S = Savarino . (CS) also had a share of 10. (70) 281. The relative market share ratios (Regular Maxwell House share divided by share of second largest brand) were: Relative Market Shares Sales District 1971 1972 1973 1974 1975 1976 1977 Atlanta- Ratio 3.46 #2 Brand (CS) (CS) (CS) (CS) (CS) (CS) (CS) Boston Ratio #2 Brand (H)" (CS) (CS) (CF) (H) (CF) (CF) Charlotte Ratio #2 Brand (CS) (CS) (CS) (CS) (CS) (CS) (CS) Jacksonville Ratio 2.48 #2 Brand (F) (F) (F) (F) (F) (F) (F) Ratio #2 Brand (S) (S) (S) (S) (CF) (CF) (CF) Philadelphia Ratio #2 Brand (H) (CS) (CS) (CS) (F) (H) (F) Syracuse Ratio #2 Brand (H) (CS) (CS) (CS) (CS) (F) (F) Youngstown Ratio #2 Brand (H) (H) (H) (F) (F) (F) (F) (CX 1072) (71) 282. Complaint counsel argue that Regular Maxwell House s absolute and relative share dominance of regular coffee sales in the western districts, its "product differentiation" advantage, its ability to command a price premium, and its supracompetitive profits (CLA 111) evidence its market or monopoly power-that is, the power to control prices or exclude competition (CLA, p. 109). They claim support for their conclusion in statements made by industry members. Initial Decision 103 F. (e) Regular Maxwell House's Market Power As Viewed By Its Competitors, Employees And Agents (1) Trade Dealing 283. Trade dealing, which accounts for 80-90% of Maxwell House marketing funds, is the most important element ofthe marketing mix for regular coffee (CX's 14Z, 203D; Tr. 1555). Mr. Einloth testified that:

The concept that was commonly accepted wa.c; that if Maxwell House has a larger share ufihe business than competition and the strengths ufthe franchises is (sic) reflecte in that share from the standpoint of trade dealing, we should be able to spend less than competition ('r. 1589).24 (72) 284. Other General Foods documents recognize that share dominance makes it possible to keep trade dealing down and increase regular coffee profitability (CX's 748, 90E, 96E, 190R, 205Z-5, 633F 637L, 646L, 691C).

285. Ogilvy & Mather issued studies in 1972 and 1975 which argued that the larger the market share of Regular Maxwell House relative to the share of the second largest brand in the market, the less the Maxwell House Division had to trade deal and the greater was Regular Maxwell House s profitability. These studies found that there was a high statistical correlation between the level of Regular Maxwell House s share dominance over the second leading brand and its ability to have lower trade deals and make greater profits. The statistical correlations were.7 and.9 (the highest possible correlation is 1) (CX' , 679A).

286. The first analysis indicated that in Maxwell House s high franchise sales districts "there is a high level correlation between the extent of share dominance and profitability. The coeffcient is .7. This is very significant when you consider that a coeffcient of 1.0 would indicate that share dominance explains all the variation in profit" (CX' s 1L, 468K).

287. Mr. William Philips, Ogilvy & Mather s chairman, testified that the analysis proves "something everybody knows " that is, when your franchise is dominant in a trading area: You have less need to promote with the trade to get trade cooperation in the consumer offers and, if there are two or three brands about the same in share and position, its "" II) a donunent which he authored, Mr. Einloth stated. Evidence that 2X share leverage results in ability to underspend competition can he found in RMH history in the East. In nearly all of our Eastern or noninfringement markets (areas where Folger is not in distribution) RMH enjoys a minimum share advantage by pack-size of twice t.hat. of its nearest competit.r. Competitive trade dealing h"s always been considerably higher than RMH, sometimes by as much as $3.oo/case, but the Brand has been able to withstand it., build its share and improve its margin less deal position (CX 725E) 204 Initial Decision a much more competitive environment and your trade deals are high and your contribution25 is lower (Tr. 1947). (73) 288. A second study by Ogilvy & Mather came to the same conclusion as the first, except that the correlation between share leadership and Maxwell House profitability had increased from. 7 to .9 (CX 679A).

289. An additional conclusion of the first study was that when another brand reached the 20% market share range, profitability for Maxwell House (CMLTD) began to decrease (CXID, 468C). When there is a second brand with a market share above 20%, it meant that Maxwell House s profit decreased by about $.50 per unit (CX's ID 468C). The analysis also concluded that if Maxwell House could keep Folger below the 20% share level, Maxwell House could earn an additional $8 milion in long-term profit (CXID, 468C; Tr. 1439-40). 290. The Maxwell House Division adopted trade dealing practices which recognized the significance of share dominance: For each district, market area, and dealing area, RMH share VB. competition is tracked. Depending on the RMH to competitive share ratio, RMH deals at a different level keying only to significant competition (CX' s 194Q, 195Q). 291. The Maxwell House Division s fiscal 1973 marketing plan describes the HFNI as the area where:

RMH is the dominant brand with no significant competition.26 RMH' s leverage27 is great with spending generally independent of competition (eX 14Z-1). (74) 292. The fiscal 1972 throu5'h 1976 marketing plans for Maxwell House state that in the HFNI, the Division could "spend below competition" on trade deals (CX's 13D, M, Z-10, 14P, Z-1, 15G, 16K, W, 18K Z-18, 19R, Z-28).

293. General Foods recognized that in the high franchise sales districts, Regular Maxwell House s trade rates were less than in the low franchise sales districts. In the fiscal 1975 Marketing Plan, the trade rates per unit for 1973 and 1974 were:

Z5 Contribution margin is the amount remainil1g from net revenue after deducting the cost of goods Bold. It is the amount available for marketing and profit (Th. 1248). 26 Mr. Einoth testified that the phrase "no signficant competition" referred to the fact that "ltJhere was no single competitor whose presence war found throughout the entire non-infringement area" but that "(tJhere was signifir cant competition by area within the non-infringement markets" ('r. 1611). rI Meaning that the Division War able to use lower trade deals than its competitorsin the area ('. 1417 1552-3) Initial Decision 103 F. HFNI HFI LF 1973 1.3 1.42 2.40 1974 1.42 1.54 2.

(CX 16Z-36).

294. Its competitors confirmed that the Division trade dealt less heavily in areas where it had a dominant market share (Tr. 2026 2847, 2878), and, in some areas, had higher dead net prices than its competitors.

295. Mr. Trone testified that Regular Maxwell House had higher dead net prices than Hils Bros. and Chase & Sanborn (Tr. 1384, 1505). 296. In Cleveland and Pittsburgh, prior to Folger s entry, Regular Maxwell House s case rate (non-performance and performance allowances) generally was less than other competitors, up to $1.50 per case less. Thus, the list price less case rate for Regular Maxwell House was up to $1.50 higher than the other competitors (Tr. 1647-48, 1681 82). 297. Prior to Folger s entry into the Charlotte and Atlanta sales districts, Regular Maxwell House did not meet competitors' trade rates on a continuing basis; when it was not meeting the competitive trade rates, its dead net price was above the dead net prices of its competitors (Tr. 6636-37).

298. Prior to Folger s entry into Syracuse, Regular Maxwell House trade deals were lower than those of Hils Bros. and Chase & Sanborn. Under both "trade dead net" (list price less (75) non-performance and performance allowances) and "consumer dead net" (list price less non-performance and performance allowances and retailer coupons) it had a higher price than Hils Bros. and Chase & Sanborn. In terms of "consumer dead net " there could be a $.10 per pound differential (Tr. 6026- , 6035-39).

299. An analysis by Mr. Toy of the dead net price differences between Hills, Regular Maxwell House, and Chase & Sanborn in the east prior to the entry of Folger indicates that Regular Maxwell House commanded higher prices than Hills Bros. and Chase & Sanborn. In this analysis, Mr. Toy cumulated the net price differences between Regular Maxwell House, Hils Bros., and Chase & Sanborn on all three pack sizes for a period of seven months (CX 866H-I; Tr. 2046, 2152-54).

300. The analysis for the areas dominated (as Mr. Toy defined the word) by Maxwell House indicates that:

In Philadelphia, Regular Maxwell House s cumulative dead net price was 15 cents 2S Defined by Mr- Toy ufHills Bros. as the situation in which "Maxwell House has far and away the largest share ofmarkel and the next nearest competitors have roughly less than half of the share that Maxwell HOllse wuuld have" (Tr. 2026).

204 Initial Decision higher than Hils' and 23 cents higher than Chase In Pittsburgh, Regular Maxwell House s cumulative dead net price was 7 cents higher than Hils' and 12 cents higher than Chase In Cleveland, Regular Maxwell House s cumulative dead net price was 26 cents higher than Hils' and 17 . cents higher than Chase In Detroit, Regular Maxwell House s cumulative dead net price was 16 cents higher than Hills' and 24 cents higher than Chase In Boston, Regular Maxwell House s cumulative dead net price was 4 cents higher than Hills' and 13 cents higher than Chase In Albany, Regular Maxwell House s dead (76) net price was 5 cents higher than Hils and 17 cents higher than Chase In Buffalo, Regular Maxwell House s cumulative dead net price was 6 cents higher than Hils' and 15 cents higher than Chase s (CX 8661; Tr. 2045-6, 2050). 301. In a study of Philadelphia, Mr. Toy found that Maxwell House sold at higher net prices to the trade than Hils Bros. and Chase & Sanborn for 21 consecutive months between February 1968 and October 1969. The average difference in price between Maxwell House and Hils Bros. over the two years was over 2.6 cents per pound and the average difference between Maxwell House and Chase & Sanborn was over 2.9 cents per pound (CX 872C; Tr. 2059). (2) Price Leadership 302. The Maxwell House Division s competitors sometimes priced their products with reference to Regular Maxwell House. From 1971 to 1977, Kroger "pegged" the price of its private label coffee to the price of Regular Maxwell House. Its prices were set about 5 cents below Regular Maxwell House s prices (Tr. 1651). de Lima Coffee Company primarily looked at the price of Regular Maxwell House in setting the price for its coffee because Regular Maxwell House was the number one brand in the market. It set its price so that it would have a lower shelf price than Regular Maxwell House (Tr. 2507418). Hils Bros. also priced its coffee in relationship to Regular Maxwell House prices so that it could obtain a lower shelf price, for it believed that it required a lower shelf price than Regular Maxwell House in order to maintain its sales volume (Tr. 2036-37). 29 Prirc per pound over a seven month period. Thus, tht! 15 rent difference averaged, per month, 2 cen\. per pound (Tr- 2045-6).

, Initial Decision 103 F. (3) Competition Within The HFNI 303. All of the above notwithstanding, Regular Maxwell House share dominance in the HFNI did not mean that it could ignore the pricing and trade and consumer dealing of its competitors, nor was its dominance so great that it could exclude competitors from the HFNL 304. Thus, despite the statement in ex 14Z-1 that Regular Maxwell House s spending is independent of competition (Finding 291), the same document points out that trade promotions "maintain competitive pricing and feature activity (77) necessary for the short-term health of the franchise" (CX 14Z), and that 45% of the $40 milion trade deal expenditures for the brand would be spent in the HFNL This is convincing evidence that the Division could not simply ignore its competition in that area.

305. Mr. Graham, who has had extensive experience with the Maxwell House Division, Martinson s and Standard Brands (Chase & Sanborn) testified that Regular Maxwell House was, and had to be competitive in the east prior to Procter & Gamble s eastern expansion. Had it not been the grocery trade would have tended to lean much heavier toward the brands where they received larger offers and Maxwell House would have lost ground, would have lost sales and share of the market" (Tr. 10 544-5). In the east, prior to Procter & Gamble s expansion, according to Mr. Graham, Regular Maxwell House could not charge higher than competitive prices to the grocery trade, could not control the price of coffee, and could not command a premium price over other coffees (Tr. 10 545). In fact, Martinson s and brands other than Regular Maxwell House were sold at a premium price in the east (Tr. 10 563-4).

306. Mr. Graham considered the eastern United States even prior to the expansion of Procter & Gamble to be the most competitive portion of the country (Tr. 10 541). This is because the east had a number of strong regional brands that were marketed vigorously in addition to a number of national brands. The national brands, in addition to Regular Maxwell House, were Hils Bros. and Chase Sanborn, and the strong regional brands were Savarin, Chock Full Nuts and Martinson, all of which were marketed "on almost the entire eastern seaboard" (Tr. 10 542). Mr. Graham s opinion as to the competitiveness ofthe east is shared by the Maxwell House Division national sales manager (Tr. 5747).

307. Again, despite the sweeping statements in the documents discussed above, knowledgeable witnesses testified that Regular Maxwell House could not ignore its competitors in the eastern sales districts. Mr. Trone testified that New York "used to give us a fit" in terms of competitive marketing expenditures (Tr. 1473), and that 204 Initial Decision Regular Maxwell House competed with Standard Brands, Hils and was faced with "formidable COB (Chains' own brand) competition. Because coffee was so price sensitive, the consumer would switch to a coffee whose trade deals were reflected on the shelf, and General Foods would have to react to that or lose volume (Tr. 1474-76). 308. Mr. Mann, who is the Maxwell House Division s eastern region sales manager, testified that in New York, Regular Maxwell House responded to major competitors through a diversified marketing approach using consumer promotions, advertising and trade deals (Tr. 6481). It did not always match (78) its competitors penny for penny on trade deals, but also employed advertising and consumer promotions. By contrast, its competitors in New York concentrated on trade deals "with very little attention to the advertising and consumer promotion components of the mix." In some instances, Regular Maxwell House matched competitive deals penny for penny or even exceeded the level of competitive deals (Tr. 6482-83). Despite these efforts, during the three years immediately prior to Procter & Gamble s entry into New York, Regular Maxwell House had been losing share (Tr. 6487).

309. In Boston, Regular Maxwell House s major competitors prior to Procter & Gamble s expansion were Chock, Hills Bros., Chase & Sanborn and COB (Tr. 6479). Regular Maxwell House responded to the trade deals of competitors in Boston in much the same way as in New York, sometimes matching them penny for penny, sometimes exceeding them, and sometimes not matching them exactly but relying instead on advertising and consumer promotions (Tr. 6482). Despite these efforts, Regular Maxwell House in Boston, as in New York had been losing share during the three years prior to the Procter & Gamble expansion (Tr. 6487).

310. A June 1, 1976 Maxwell House Division memorandum entitled Eastern Complex Revised Dealing Strategy reported that Regular Maxwell House s share was declining in all areas of the complex which is comprised of the Boston, New York and Philadelphia sales districts. The cause was stated to be chains-own-brands: "the magnitude of RMH's losses in Boston and Philadelphia reflect significant on-going shelf/feature price disadvantages to RMH's largest competitor in both markets, COB" (CX 192B). The share decline in New York was attributed to significant shelf and feature price disadvantages versus COB, but "major shelf price disadvantages to CFON (Chock Full O'Nuts) have impacted RMH more significantly than COB in New York" (CX 192C). The memorandum concluded that "RMH will continue to risk suffering share declines throughout the complex as long as feature activity remains suppressed and RMH experiences differentials VB. COB and CFON ofthe current magnitude" (CX 192E). , Initial Decision 103 F. 311. A December 2, 1976 Maxwell House Division memorandum reported on the inability of Regular Maxwell House to ignore competition in New York and Boston:

In New York and Boston, CFON and Savarin are taking their toll on RMH. We are attempting to stay competitive but are finding more and more trade resistance to featuring coffee (CX 689B). (79) 312. Mr. Keller, the Maxwell House Division s southern regional sales manager, testified that the Division had to be responsive to competitors in that area (Hils Bros., Chase & Sanborn, Chock Full Nuts, COB and smaller local competitors) (Tr. 6590-94). 313. In Detroit, Atlanta and Charlotte, Regular Maxwell House used a balanced marketing mix which included advertising and consumer promotion as well as trade promotion. Most of Regular Maxwell House s competitors used trade offers as their primary means of marketing. They did very little advertising or consumer promotion in Mr. Keller s opinion (Tr. 6595).

314. To assure that the retail grocery trade would price it competitively with other brands, Regular Maxwell House had to offer trade deals competitive with other manufacturers in Mr. Keller s region. Otherwise the consumer wil walk right by us and pick up a competing product " because if competitors had better trade deal programs the trade would feature them and not Regular Maxwell House (Tr. 6596).

315. Procter & Gamble was an established competitor in the Jacksonville sales district in fiscal 1971 with a share of20.8% (CX 1072A). The Jacksonvile sales district was part ofthe high franchise infringement area, an area where the Division regarded Folger s and Regular Maxwell House as equal in size and strength and where the Maxwell House Division regarded the competition as being "fierce" (CX 14Z-1). 316. According to the head buyer for S.M. Flickinger, a major wholesaler in the Buffalo area, Regular Maxwell House s list price and trade promotional allowances equalled those of Hills Bros. and Chase & Sanborn prior to the time of Folger s expansion there (Buffalo is in the Syracuse sales district) (Tr. 11 859-60), and he believed that if it was not competitively equal with Hils Bros. and Chase Sanborn, it would cease to be a popular brand: It' s that competitive. We wouldn t feature them. They wouldn t have the same retail (prices and the consumer wouldn t pick them up (Tr. 11 860). 317. The Division s Syracuse sales district manager also confirmed that during the period 1964 up to the time of Folger s entry, Regular Maxwell House would be competitive with the offerings of Hils Bros. 204 Initial Decision Chase & Sanborn and Beechnut by making (80) its dead net to the consumer30 the same as these brands (Tr. 5943). Specifically, Regular Maxwell House had to respond to the offerings of these brands: Anyone of those two brands could have had Maxwell House for lunch if we did not respond to their trade offer. Coffee is a very volatile, volatile product. The customer is well aware of what the price is. It's featured quite frequently. If the price is right and the quality is there, brand loyalty goes out the window (Tr. 6156). At times from 1970 to the date of Folger s expansion, there were periods when Regular Maxwell House s consumer dead net and trade dead net were lower than that of Hills Bros., particularly during holidays or other special merchandising events (Tr. 6196-97). 318. Mr. Engel of Kroger testified that prior to the expansion of Procter & Gamble, Regular Maxwell House competed with Hils Bros., Chase & Sanborn, COB, Chock Full O'Nuts and Breakfast Cheer (Tr. 12 063), and he stated that Regular Maxwell House s prices were comparable with other brands, as were its promotional activities (Tr. 12 064).

319. He also stated that Regular Maxwell House had the leading share in both Cleveland and Pittsburgh at the time of Folger s entry because it had the product quality desired by the consumer, it utilized consumer promotion and it provided "incentive to the trade to support the product at retail" (Tr. 1696).

320. Mr. Engel further testified that Regular Maxwell House did not have the ability to control either the retail price of coffee or the price charged by others to the grocery trade in the Cleveland and Pittsburgh areas, and Mr. Metzger stated that the grocery trade in the Buffalo area would not feature Regular Maxwell House if its trade promotions did not equal Hills Bros.' and Chase & Sanborn s (Tr. 860, 12 070).

321. Convincing evidence that the Maxwell House Division could not exclude competitors from the HFNI is the successful (81) expansion of both Hils and Folger into that area, and the expansion of regional roasters within the HFNI.

322. Hils Bros. increased its national share from 9.6% in fiscal 1963 to 11.2% by fiscal 1968, with most of the increase attributable to expanded distribution in the "new Eastern areas" (CX 11Z-7). Hils Bros. formally introduced its coffee into the Cleveland area in 1962 the Pittsburgh area in 1963, the Syracuse sales district in 1964, the Boston sales district in 1965, the Philadelphia sales district in 1966 and the New York sales district in 1967 (RX 1116). 323. As of fiscal 1971, Hils had established a share of17.1 % in the Ji List price minus buying allowances, perfommDce offers and RECC' s. "Trade dead net" excludes RECD's (Tr. 5941).

Initial Decision 103 F. Syracuse sales district and 15.8% in the Youngstown sales district (CX 1072A). Hils Bros.' eastern expansion was regarded as a successful expansion by both Hills Bros. and by the Maxwell House Division (Tr. 2130, 5793).

324. Other Maxwell House Division competitors have increased their business within the HFNI. Through the use of a new advertising campaign and !!aggressive" use of trade promotions, Martinson s increased its share in New York from 9% in 1971 to 13.5% in 1973 (Tr. 562--3). In the Syracuse, Boston and New York sales districts Chock Full O'Nuts has increased its share above its fiscal 1978 levels (RX 1278). Savarin coffee increased its share in the New York sales district over its fiscal 1978 share level (RX 1278). It also increased its share considerably in the Jacksonville sales district during the period fiscal 1971 through 1977 (CX lO72A, G). In early 1976, Savarin commenced a !!major introduction" in the Baltimore-Washington area (CX 681E).

(4) Conclusion 325. Although it is apparent that the Maxwell House Division did not believe it was necessary to meet the trade deals of all competitors at all times in the HFNI, the Division did not control the amount of trade dealing in the sales districts, for it ignored its competitors' offers at its peril, even before Folger entered the HFNI. Stated simply, ifthe Division had not trade dealt along with some of its competitors in the HFNI, it would have lost market share.

326. Furthermore, complaint counsel emphasize trade dealing and ignore the other aspects of the marketing mix advertising and consumer promotion. Although expenditures in these areas were less than for trade promotions (10-20% of the marketing mix), the Division used advertising and consumer promotions more than its competitors, and concentrating solely on its trade promotions gives a false picture of its competitive (82) activity in the HFNI. For example according to Professor Dearden (infra), Regular Maxwell House spent more in HFNI sales districts than in low franchise districts on advertising and consumer and trade promotions in 1973 and 1977 (CX' 987Z-7 to Z-85, Z-147 to Z-166). While these expenditures were undoubtedly influenced by the response to Folger s and Regular Maxwell House s greater sales volume in the HFNI, they certainly cast doubt on the claim that Regular Maxwell House had no significant competition in the east.

327. Furthermore, ignoring consumer promotions is inconsistent with Dr. Mann s definition of "economic price" which deducts these promotions from list price. Because the Division emphasized consumer promotions, ignoring them tends to raise the price the Division 204 Initial Decision charged vis-a-vis its competitors. The same problem exists with Mr. Toy s charts, for they do not include RECU's in his definition of price (Tr. 2040).

328. The problems with complaint counsel's inconsistent definition of price and the testimony of industry members indicate to me that despite the inflated rhetoric of some Maxwell House Division personnel, the Division could not ignore competition in the HFNI, did not control" price in the HFNI, and, in view of Hils' and Folger s successful expansions into that area, could not exclude competition from the HFNL 329. Finally, aside from making the claim, complaint counsel offer no evidence that Regular Maxwell House had a product differentiation advantage over other coffees. While it was and is the dominant brand in the HFNI, according to Mr. Graham, aside from some private label coffees, it did not sell at higher than competitive prices to the grocery trade (Tr. 10 57G-71).31 Ifby "product differentiation advantage" complaint counsel mean simply that it is a product which commands consumer loyalty, then Regular Maxwell House had no advantage over other branded coffees which, one must assume, also had a loyal following. If they did not, they could not remain viable competitors. (83) (D Share-Price-Profit Analysis 330. Dr. Mann conducted statistical tests that were designed to determine whether the share-price-profit relationship which he believes indicates market power is reflected in the relationship between the market shares, prices, and profits of Regular Maxwell House (Tr. 3595, 3608, 3698-99).

331. His study analyzed five different statistical relationships: the relationship between absolute share and profit (CX's 1104, 1110); the relationship between relative share and profit (CX's 1106, 1112); the relationship between absolute share and price (CX's 1105, 1111); the relationship between relative share and price (CX's 1108, 1114); and the relationship between price and profi (CX's 1107, 1113). 332. Dr. Mann was satisfied, based upon correlation analyses (which measure the degree of association between two pieces of data) (Tr. 3545) that each of his five studies established a statistically significant positive relationship between share, price and profit (Tr. 3596-99, 3603-07, 3609- , 3614-18, 4021- , 4035-36). 333. Since these correlations showed that Regular Maxwell House 3\ A General Foods docwnent de cribes Regular Maxwell House as a brand "that exists to satisfy the coffee drinking needs of ground coHee users who want a consistently high qualty and satisfying coffee beverage at a popular price" (CX llZ-9; see o/so ex' s 23H 44J, 45Z-4). " Evaluation of statistical signficance is a process that reveals how likely it is that two variables are related to each other (Tr- 3548) (g) Initial Decision 103 F. pricing behavior and profiability were influenced by absolute share and relative share, he concluded that "some of the profitability levels enjoyed by the Maxwell House brand were consistent with the presence of market power" (Tr. 3688). Dr. Greer also agreed that this analysis established the presence of market power (Tr. 11 582-83). 334. On the other hand, Dr. Elzinga warned that high economic profits are not necessarily explained by the presence of monopoly power. Other factors such as superior management or simple good luck may well account for such profits (Tr. 9821). Correlations such as those prepared by Dr. Mann thus do not necessarily explain Regular Maxwell House s profiabilty33 or (84) lead to the conclusion that it possessed market or monopoly power in the HFNI. In fact, the findings immediately above, which are based on testimony by knowledgeable industry members, reveal that the Maxwell House Division had neither control over price in the HFNI nor the ability to exclude competition in that area, and I reject Dr. Mann s and Dr. Greer conclusions for they are based on statistical analyses which do not reflect the actual state of the regular coffee market in the HFNI or in its sales districts.

Regular Maxwell House s Profitability (1) Professor Dearden s Calculations 335. Professor John Dearden of Harvard University, who is an accounting expert, calculated the profitability of General Foods' regular coffee brands by sales district (Tr. 29). 336. Professor Dearden followed two principles in his profitability calculations: A business segment must be charged with all costs that benefit that segment (Tr. 36-38, 106-7); and, a common denominator must be found that wil allow a comparison ofthe calculation with the profitability of other businesses. The common denominator selected by Professor Dearden was ROFE, or "return on funds employed" (Tr. 36-38). (85) 337. ROFE's for General Foods' regular coffee products by sales district were calculated by assigning as much of the cost and revenue as possible direcUy to the sales district and allocating the remainder in the most equitable way possible (Tr. 214). 1.1 Historically, analysts have often jumped to unjustified conclusions by mistaking an olmerved correlation for a cause-effect relationship. A high sample correlation coeffcient docs not. necessarily signify any causal relation between two variables (Bhattacharyya & .Johnson Wiley Series in Profitability Mu.themu.tico/ Statistics at 406-7 1977).

It is not diffcult by associating different pairs of variables to find a high degree of correlation among some oftbem- In such a situation, itis tempting to conclude that one variable causes t.he other tu change-to a& ume OIutomaticaJly a cause.and-effect relationship (Mandel Statistics (OT Management at 37, University of Ba.ltimore).

The fact that we find a relationship between two variables docs not. imply a cause-and-effect relationship (Salvature & Berliner Statistics at 173).

204 Initial Decision 338. General Foods routinely allocates and assigns costs to determine product line profitability (CX 1006; Tr. 144), and Professor Dearden testified that what he did-assigning revenues, costs, and funds employed to a geographic area-is no different than assigning them to a product (Tr. 21-23). Thus, his district profitability calculations, according to him, give "a fair picture ofthe profitability among sales districts" (Tr. 310).

339. On at least two occasions, General Foods prepared analyses similar to those of Professor Dearden. From 1971 to 1973, the Maxwell House Division prepared district profitabilty statements that reported a quarterly profit figure by brand for each sales district (CX 993A- W) and, in 1977, General Foods financial planning and analysis department prepared a "Max-Pax regionalization study" which calculated profitability quarterly on a sales district basis (CX 815K). 340. The Regular Maxwell House ROFE's calculated by Professor Dearden for the eastern sales districts ranged from about 30 to 70% pre-tax in fiscal 1971 and 1972. The ROFE's for the western sales districts were lower, and in some cases were negative (CX 1115A). The national profits for Regular Maxwell House, according to Professor Dearden s calculations, were 42.40% pre-tax in fiscal 1971 and 37.20% pre-tax in fiscal 1972 (CX 1386).

(2) Mr. Klein s Calculations 341. Mr. Donald Klein, director of financial planning and control for General Foods (Tr. 8217) testified that Professor Dearden improperly treated certain Maxwell House Division cost and investment data and misinterpreted other data, and he revised Professor Dearden figures to present what he believes is a more accurate calculation of profitability (RX 1170; Tr. 8457-68).

342. In addition to selling coffee packaged for sale at retail, General Foods also has a separate military business which sells to military commissaries and other military outlets. The military business is handled by a separate (86) organization which has its own separate overhead expense, general and administrative expense and sales force (Tr. 8461 , 8467).

343. Professor Dearden s calculations attributed certain grocery coHee expenses to the military sales business. Mr. Klein returned these grocery coHee expenses to the grocery coffee business, which has the effect of reducing the Dearden profit before tax figure for the grocery coffee business. This in turn reduces the ROFE profitability estimate (RX 1170A-H; Tr. 8468).

344. When General Foods decaffeinates coffee during the production of its decaffeinated coffee products, caffeine is produced as a by-product, and the processing and sale of this caffeine is treated by Initial Decision 103 F. General Foods as a separate business with its own separate profit and loss statement (Tr. 8411'19). In the ordinary course of its business General Foods does not include profits from the sale of caffeine in its ground coffee profits (CX 991A-N; Tr. 8473). 345. Mr. Klein testified that the demand factors affecting the profitability of General Foods' caffeine business are unrelated to the ground coffee business. Caffeine is basically an industrial product and is not sold through the same channels as ground coffee (Tr. 8411'20). 346. General Foods claims that inclusion of caffeine profits and exclusion of the investment needed to produce these profits inflates sales district profitability and I agree with this argument (RPF 1'39); however, Professor Dearden assigned no caffeine profits to Regular Maxwell House, and Mr. Klein s adjustments do not affect profitability figures for this brand (Tr. 9030-31).

347. The Maxwell House Division self-manufactures about one half of the coffee cans used to package its regular coffee products. This results in a saving to the Division over the purchase price of the can and represents a unique saving as a result of packaging effciency, and Mr. Klein does not believe that savings resulting from self-manufacture of cans belong in any co,.putation of profitability for comparison with a profitability benchmark which does not reflect such packaging effciencies (Tr. 8407).

348. The Division transfers these self:manufactured cans to itself at the prevailing market price and accounts for savings due to selfmanufacture in a separate profit and loss statement (Tr. 8408). In the ordinary course of its business, the Division does not include can plant savings in its calculation of ground coffee profits (CX 991A-k; Tr. 8473). (87) 349. Professor Dearden added these savings to profit before taxes and included the investment in the can plant in his calculations (Tr. 8411). This, according to Mr. Klein, inflated General Foods' profis for purposes of comparabilty with other businesses (Tr. 8409), and his revision excludes can plant savings from profits and can plant investment from funds employed (RX 1170A H; Tr. 8472). 350. Other revisions by Mr. Klein to Professor Dearden s calculations were use of an actual tax rate of 51 % (rather than assumed rates of 48% or 52%) which, in some cases, decreased profitability (Tr. 8500-1); inclusion of interest expense, which increases profitabiliy (Tr. 8413- , 8500-1); adjustment of accounts receivable data to reflect actual accounting experience which in most cases reduces the ROFE estimate (Tr. 8503-07); revision of accounts payable data which increases the funds employed (Tr. 8527-28); inclusion of certain cash investment items overlooked by Professor Dearden which increases 204 Initial Decision investment (Tr. 8532-37); and, revision of fixed asset allocations (Tr. 8541).

351. The following chart compares Professor Dearden s calculations (CX 1115C) with Mr. Klein s (note, however, that Mr. Klein s calculation is for all of the Division s regular coffee) (RX 1171A): CX 1115C RX 1171A Regular Maxwell GF Total Regular House Estimated Coffee Estimated Return Return Boston 25. 15. New York 18. 10.4 Philadelphia 22. 12. Syracuse 10. Youngstown Charlotte 38. 27. Atlanta 35. 27. Detroit 13. 12. Total Eastern Area 20. 12. 352. Mr. Klein also computed an "economic" as opposed to an "accounting" rate of return. Dr. Elzinga explained that accountants expense items which economists recognize as having a useful life beyond the period of expenditure (Tr. 9819). Items which an economist would amortize or depreciate to reflect their useful life are advertising, promotion and research and (88) development (Tr. 9580, 9819). Dr. Elzinga concluded that an amortization rate of 36.8% was appropriate for costing the advertising and promotion of the Maxwell House Division, and that this rate was supported in the economic literature (Tr. 9770-75, 9817-32). The following chart compares Professor Dearden s estimated Regular Maxwell House accounting return with the economic return for all of the Division s regular coffee as estimated by Mr. Klein:

Initial Decision 103 F. CX 1115C RX 1171B Regular Maxwell General Foods' Regular House Estimated Coffee Estimated Accounting Return Economic Return Boston 25. New York 18. Philadelphia 22. Syracuse 10. Youngstown Charlotte 38. 17. Atlanta 35. 16. Detroit 13. Total Eastern Area 20. (3) Criticisms Of ROFE Calculations By Sales Districts 353. Dr. Roman Weil, a professor of accounting at the University of Chicago (Tr. 8018) may have been exaggerating somewhat, but I believe that he reflects the general attitude of the accounting profession about allocations:

(T)he Dearden procedures carry out allocations and those allocations are meaningless. The reason that I use the word meaningless is that those are the very words that Professor Dearden uses in his own writings. The views that I hold about the meaningless-the general feeling I have about these allocations of costs that are not directly (89) attributable are widely held among accounting theorists, accounting writers, and one ofthe most noted writers on the subject of the meaningless of these allocations is John Dearden of Harvard University (Tr. S045-6).

354. While General Foods has, on occasion, calculated district profitability, Mr. Klein stated that national financial data cannot be retroactively assigned to districts to arrive at an accurate determination of sales district profitability. For example, to determine profitability for a sales district, one would need to know which of General Foods' four coffee plants, each with differing costs of producing coHee, supplied coffee to that sales district (Tr. 8260-62). 355. In response to a question by complaint counsel, Mr. Graham of Standard Brands stated:

Q. And you considered the Standard Brands' own calculations of Chase & Sanborn profitability by sales district to be a good measurement tool and very useful to you, did you not? A. The concept was useful. The results weren t very good. . . . J, Prufe5Sor Dearden disagrees (In this p(J;nt, ilrguing that sillce green bean and packaging material cost rel'resentneilrly98% ufalJvarjablcmanufacturingc(Jst incurredatthecofTeel'!antsaDdare purchaseu centrally, the use of national costs is appropriate (Tr. 58- 59). 204 Initial Decision And somehow I just felt the way the costs had been accumulated, there was something very imprecise in the system. So, we quit using the system (Tr. 10,658- 59).

356. Charles Clark, Comptroller of The Folger Coffee Co., testified that Folger, in the ordinary course of its business, (90) does not compute profit data by geographic area because it is not meaningful: Q. Does the Folger Coffee Company usually compile information regarding its revenues, costs and profits on a less than total marketing area basis? A. No, we do not compile that information normally on a less than total U.S. basis. Q. Why is that, sir? A. We have not felt that it was meaningful in the operation of OUT business to do that (Tr. 3171).

(4) Comparison Of Regular Maxwell House Profitability With The "Benchmark"

357. Dr. Mann testified that firms with market power have the ability to raise prices above cost and hold them there, and that high profits over a sustained period oftime indicate that a firm has market power (Tr. 3577, 3630-31, 3991-93). To determine whether Regular Maxwell House enjoyed market power, Dr. Mann compared its sales district ROFE's with a "benchmark." An appropriate benchmark is one which approximates the competitive rate of return (Tr. 3621). Dr. Mann used two benchmarks, one based on the average rate of return for food and kindred products and the other on the average rate of return for all manufacturing, as reported by the ITC' s Quarterly Financial Report (QFR). Each average was calculated for the period 1971-1977 (CX 1085B, D; Tr. 3619, 3622- , 3627-29). 358. Dr. Mann used a seven year average because an average over that long a period of time establishes the ability of a firm to hold prices persistently above the competitive level, evidencing a firm ability to prevent competitors from eroding those profits by expanding and undermining the price structure (Tr. 3630-31). 359. Dr. Mann multiplied the average by 1.5 to obtain the rate of return that he used in his market power analysis (Tr. 3629-31). He used the 1.5 multiplier because he believes (91) that studies of the relationship between profit and share suggest that firms that are expected to have monopoly power have profit rates about 1.5 times the benchmark, and because 1.5 is a conservative number that adjusts for any imprecision in the calculations (Tr. 3631, 3692, 4076). A rate of return greater than 1.5 times the QFR benchmark is, in his opinion higher than a competitive rate of return (Tr. 3645). 360. The all manufacturing seven year average was 11.04% and the food and kindred products seven year average was 11.49%. One point Initial Decision 103 F. five times the average equals 16.56% for all manufacturing and 17.24% for food and kindred products (CX 1085; Tr. 3626-31). 361. Using the Dearden estimates, Regular Maxwell House ROFE' in the following eastern sales districts exceeded the food and kindred products 17.24% standard:

Regular Maxwell Sales District House ROFE % Atlanta 38. Boston 25. Charlotte 35. Jacksonville 20. Memphis 20. Philadelphia 22. (CX' s 1085, 1115C).

362. Dr. Mann testified that the profitability figures in the Y oungstown and Syracuse districts did not meet the average profitability standard of 1.5 times the QFR benchmark because of predatory pricing which he believes occurred in the middle of this period (Tr. 3700 3705).

(5) Criticisms Of The QFR 363. Dr. Mann believes that the QFR average is conservative because it is biased upward by inclusion of monopolists which earn above average rates of return (Tr. 3639, 3692). On the other hand, General Foods argues that, conservative or not, the QFR is so flawed that it cannot be used as a benchmark.

364. Mr. R.T. McNamar was executive director of the FTC from 1973-1977. He was given responsibility for managing the development of the FTC's line of business reporting program and (92) in that role undertook an analysis of the QFR program (Tr. 9595-96). 365. Under QFR procedures, a reporting company is initially classified in the mining, manufacturing, wholesale trade or retail trade industry groupings based upon which of the above activities accounted for most of that firm s gross receipts. A corporation assigned to the manufacturing division is then further classified into two-digit Enterprise Standard Industrial Classification ("ESIC") groups, such as food and kindred products, based upon that two-digit ESIC group "which accounts for more gross receipts than any other two-digit Manufacturing Group" (RX 1182I). The QFR publication warns that this procedure results in entire large conglomerate companies being assigned to an industry group which accounts for only a small portion of that company s gross receipts:

. . . . . . . . 204 Initial Decision It should be noted that these procedures may lead to a conglomerate corporation being assigned to an industry group from which only a small proportion of its receipts are obtained (RX 11821).

366. Mr. McNamar believes the primary industry classification scheme not only distorts data by including operations of firms not engaged in that business, but also distorts data through introduction of items not at all related to the operations of a QFR industry classification. According to Mr. McNamar in a memorandum on the QFR which he sent to the director of the Bureau of Economics: To ilustrate, a company s "primary industry" based on its gross receipts may be a relatively low capita intensive one. Yet, within the same company in filling out its QFR form other aset intensive industries cost structures are included in the "primary industry." These aggregated results are then combined with other similarly treated firms to beome QFR data. Hence, the high aset structure of a " secondary. industry (line of business) is combined with the sales from a less capita intensive one. Obviously, the operating expenses, depreciation charges, etc., are inappropriately subtracted from the (93) primary industry s sales to derive net income before taes. Clearly, the QFR's primary clasc;ification scheme violates sound accounting principles with no apparent justification precedent. More importatly, the reporting approach by design masks the operation of any one industry, and I would submit makes inter-industry comparisons meaningless (RX 1181H; Tr. 9616-19). 367. Mr. James Folsom, when deputy director of the FTC's Bureau of Economics, also cited several concerns about QFR data in a February 7 1975 affdavit fied in FTCv. American Standard. On page 3 of his affdavit, Mr. Folsom stated:

The QFR data are deficient in three major respects. First, the industry categories are to broad to be of maximum utility in analyzing individual industries. Second, the QFR data are assembled by assigning all activity of a firm to that single industry which accounts for the plurality of its sales. This leads to a problem known as contamination. (RX 1155C).

368. A report by the FTC Bureau of Economics staff on the justifications for the FTC LB Program also recognizes deficiencies in QFR: There are two serious problems with the QFR data which severely limit their usefulness. The industry groupings are excessively aggregated, and all of the operations of a company are assigned to a single industry (RX 11798). The report also stated:

(Some) activities are included in each of the categories which should not be, and some activities which should be included are not. (94) Initial Decision 103 F. The magnitude of this problem can be illustrated by some results of a recent internal analysis. This analysis covered slightly more than 100 companies, with a total value of gross receipts of more than2S0 bilion dollars. For these companies approximately one-third of total gross receipts was from QFR categories other than the one to which they were assigned. In four ufthe 31 categories, from 30 to 39 percent of gross receipts belong in other categories; in two categories, the percent is between 40 and 49; in one category, the percent is between 50 and 59; and in other categories is between 60 and 69. As these data show, the distortions introduced in the QFR data by diversification are demonstrably both severe and pervasive. And, given the magnitude of the distortion which (diversification) causes, extreme skepticism is called for in any of the various uses of industry data on sales, costs and profits (RX I179S-W).

369. While complaint counsel argue that criticism of the QFR applies to its use to determine the profitability of specific product markets, and not to its use as a benchmark (CRF 4-13), Dr. Mann, when he was director of the FTC's Bureau of Economics, testified before a Congressional subcommittee that:

Most major firms are so highly diversified that any particular 4-digit sic category accounts for so little of the company s business that it is very misleading to claim we are measuring industry profitability (RX 1275E). (95) (6) General Foods' Benchmark 370. Given the above criticisms and other problems which General Foods sees in the QFR 35 it offered a benchmark consisting of a sample of 25 firms with at least 50% of their sales in branded consumer grocery products (RX 8-123). The problem with this benchmark is that it is most likely more biased and contains more sampling error than QFR. First, the sample is very small, whereas QFR samples over 000 firms (RX 1182U; Tr. 3619).

371. Furthermore, the benchmark's profit rate is probably biased upward because at least 18 of the 25 firms operate in one or more product markets with four firm concentration ratios ranging from 52% to 90% (Tr. 9206-14), where, presumably, profits are higher than in markets with less concentration. Thus, I do not believe that the alternative benchmark offers any more insight into the monopoly profits issue than does QFR.

:I QFRdata is estimated and unaudited (RPF 8-96 to 8-99); it suiers from "significant discontinuities" (RPF 8-100 to 8-102); it includes bankrupt companies (RPF 8-108 to 8-109; and, it report accounting rather than economic rates of retur (RPF 8-110).

JO Because ofthis Dr. Elzinga stated that sampling is not a problem with the QFR (Tr- 9840-1) 204 Initial Decision (7) Conclusion 372. I do not agree with Dr. Mann that the comparison of Regular Maxwell House s estimated profits in certain eastern sales districts with his QFR-based benchmark establishes that the brand enjoyed market or monopoly power in those districts. 373. I use the term "estimated profits" deliberately, for I do not believe that even Professor Dearden would argue that the figures which he calculated are anything other than estimates. They are, to be sure, estimates of a highly-respected expert and are entitled to serious consideration; however, they are based on an assumption which I cannot accept.

374. The assumption is that aggregated accounting data can be broken down so accurately that precise economic conclusions can be made about the profitability of one brand of regular coffee in one Maxwell House Division sales district. Furthermore, even if Professor Dearden s estimates are accepted (96) at face value, the possibility that the "excess" profits identified by Dr. Mann are due to something other than monopoly or market power has not been excluded. And finally, the fact that QFR may have been used by economists as a benchmark in other cases or in articles analyzing industry power (CPF 11-258 to 11 266) does not mean that it must be accepted in this case. Enough problems with the QFR have been identified in this case to make me very wary of accepting it whole-heartedly as a benchmark here. In conclusion, I agree with Dr. Elzinga s summary ofthe proper use of QFR- , it may be used for insight, but using it to draw firm conclusions in this case is very risky:

Where I would disagree with my friend Professor Mann, is that instead oftaking the data for insight, he uses the data to drawn conclusions. And I have two reactions to that. One is that high profits relative to a QFR contaminated benchmark do not necessarily mean monopoly. Before I ascribed an antitrust violation to that situation I would want to be very certain that those high profits were not simply the result of a better run company.

The second thing is, is that I've made an examination of the profitability of the Maxwell House Division 37 and even on the basis of accounting data, this is no money machine. Its average rate ofrcturn approximates that of the average of all manufacturing (Tr. 10 493).

Now, economists are always making judgment calls. It' s one thing to make a judgment call when your numbers represent the phenomenon you are actually measuring. But here we are using a rubber yardstick with accounting data and consequently I think it is imprudent to take numbers that are as fuzzy as accounting numbers and from it draw the conclusion that (97J profits are too big and they have lasted too long, especially when your benchmark has some of the problems that I have already a.--cribed to it (Tr. 9846).

17 An apparent reference to Mr. Klein s two calculations of the profitability of all oftbe Division s regular coffee (Finding 351 and 352) Initial Decision 103 F. 375. Mr. Keith B. Anderson, a staff economist called by complaint counsel, agreed with Dr. Elzinga:

A. . . . One would not look at profit data alone. One would look at it in the context of an overall examination including conduct, structure and performance variables. Q. You would not conclude that monopoly power existed solely on the basis of comparisons with QFR data, is that correct, Mr. Anderson? A. I would not conclude that monopoly power existed solely on a comparison of a firm s performance, a firm s profitability to a benchmark from the QFR. I would look at the whole range of information available (Tr. 10 826). 376. The danger of relying solely on a profi-benchmark analysis is evident, for when one compares Professor Dearden s profitability estimates of General Foods' Yuban brand with Dr. Mann s benchmark one comes to the conclusion that a brand with no more than a 3. market share in seven sales districts would be deemed to have market or monopoly power:

Groun an FV 1971- 1??7 After Tax ROFE Share Of Market Charlotte 43. Atlanta 50. Jacksonvile 41.

Cincinnati 19. 0.4 Chicago 19. Denver 33.7 Portland 18. (RX 1268). (98) Dr. Elzinga properly concluded that these figures show that the uncritical acceptance of Dr. Mann s theory can lead to "some very bizarre, if not absurd, results" (Tr. 10 502-D3). 377. Looking at all of the available evidence, I find that Regular Maxwell House did not possess, or come close to possessing, market or monopoly power either in the nation, the HFNl, or in the HFNI sales districts. I do not accept Dr. Mann s analysis of the brand' performance because his profi-benchmark theory can lead to the conclusion that a brand which clearly does not possess monopoly power could be found to have that power (Finding 376). Furthermore, even if Dr. Mann s theory provided some insight into this issue, the structure ofthe market, and the brand' s performance in that market show its lack of market power, for in none of the areas mentioned above did Regular Maxwell House have enough share to cause concern about, or to lead to the conclusion that it had, monopoly power (Tr. 9503-D4, 11 963). Moreover, its market conduct indicates that 204 Initial Decision Regular Maxwell House did not control the price of regular coffee and could not exclude competition from any area-the nation, the HFNI or the HFNI sales districts.

5. The Alleged Predatory Conduct (a) Introduction 378. According to complaint counsel, General Foods' response to Folger s eastern expansion amounted to an attempt to monopolize the eastern regular coffee markets by (1) engaging in predatory pricing, that is, sustained pricing below average variable and average total cost; (2) using a "fighting brand"; and (3) attempting to impose an allocation of markets on a competing firm. They also claim that General Foods copied a Folger advertising campaign to impede its entry into the HFNL (b) Proof Of Sales Below Cost (1) Complaint Counsel's Exhibits 379. Complaint counsel contend that Regular Maxwell House coffee was sold below average variable cost and average total cost as revealed on exhibits introduced by Mr. Ronald Rowe (CX's 954-- 1389; Tr. 2249). At the time of his testimony, Mr. Rowe had been a staff accountant with the Federal Trade Commission for 19 years ('!r. 2237 38), and he has previously testified as an expert accountant. He has also been assigned to (99) cases involving issues of sales below cost and area price discrimination (Tr. 2239-43).

380. Mr. Rowe prepared exhibits showing revenues and costs per unit for Regular Maxwell House coffee on a quarterly average basis for the Cleveland market area for the period April 1971 through December 1974 (CX 954--55); for the Pittsburgh market area for the period from April 1972 through March 1975 (CX 956); and for the Syracuse sales district for the period from October 1973 through December 1976 (CX 957; Tr. 2250-52).

381. Mr. Rowe determined the method of calculating the figures appearing on CX 954 through CX 959, and he performed the calculations (Tr. 2252-53). He explained in detail his methods of calculating each ofthe revenue and cost figures appearing on CX 954 through CX 959, and described the specific sources of the data he used for each calculation (Tr. 2254-313).

382. The data Mr. Rowe used to make these calculations came from information that General Foods submitted in summary format or from underlying General Foods business records, with the exception of one line item on the Syracuse chart (CX 957). Mr. Rowe developed the figures on that line from data obtained from General Foods' adver- Initial Decision 103 F. tising agency (Tr. 2253). Mr. Rowe testified that, in his opinion, the accounting data shown on CX 954 through CX 959 are very reliable data, and that they accurately show the revenues and costs per unit for the respective geographic areas and time periods (Tr. 2313). 383. Regarding most ofthe information used by Mr. Rowe, a letter to complaint counsel from Mr. Robert Y. Fox, then assistant general counsel of General Foods, stated that the information had been organized "to show as closely as possible the actual, local, geographic cost data" (CX 436B; Tr. 2313-15). Mr. Rowe also considered reliable the information he received from General Foods after the date of Mr. Fox s letter. He considered this additional information reliable based on the deposition statements of Mr. Donald Klein, a General Foods employee and financial consultant, that Mr. Klein had supervised the preparation ofthe information and that it was generated in a manner consistent with the information referenced in Mr. Fox s letter (Tr. 2315).

384. CX 954 and CX 955 show that General Foods' revenues for Regular Maxwell House in Cleveland were below total variable cost for four ofthe first 12 quarters following Folger s entry into Cleveland in October 1971 (Tr. 2322). Those quarters were the first quarter of 1972, and three ofthe four quarters of1974 (Tr. 3712-13). On a yearly average basis, General Foods' revenues for Regular Maxwell House in Cleveland were $.168 below total variable cost per unit from October 1973 through September 1974 (CX 1389). (100) 385. CX 956 shows that General Foods' revenues for Regular Maxwell House in Pittsburgh were below total variable cost for seven of the eight quarters following Folger s entry into Pittsburgh (Tr. 2322). Starting with the June quarter of 1973, General Foods' revenues for Regular Maxwell House in Pittsburgh were below total variable cost for seven consecutive quarters (Tr. 3720). On a yearly average basis revenues for Regular Maxwell House in Pittsburgh were below total variable cost by $.855 per unit from April 1973 through March 1974 and by $2.90 per unit from April 1974 through March 1975 (CX 1389). On a yearly average basis for the Youngstown sales district as a whole, General Foods' revenues for Regular Maxwell House were 329 below total variable cost per unit from April 1973 through March 1974 (CX 1389).

386. CX 957 shows that General Foods' revenues for Regular Maxwell House coffee in the Syracuse sales district were below total variable cost in seven of nine consecutive quarters starting with the quarter of Folger s entry in October 1974 (Tr. 2323, 3724). On a yearly average basis, General Foods' revenues for Regular Maxwell House in the Syracuse sales district were below total variable cost by $.516 per 204 Initial Decision unit from October 1974 through September 1975, and by $.375 per unit from October 1975 through September 1976 (CX 1389). 387. If Mr. Rowe s calculations are accurate, Regular Maxwell House was sold below average variable cost for a total of 18 quarters or 54 months in the three areas. General Foods agrees that it sold below average total cost in these areas (RLA , p. 92). 388. General Foods parries Mr. Rowe s charts with charts prepared by Mr. Klein (RX 1137-42). However, these charts were not based on quarterly data and did not have separate information for Cleveland and Pittsburgh in the Youngstown sales district. General Foods also makes several objections to Mr. Rowe s data. (2) General Foods' Objections To Complaint Counsel' s Exhibits 389. There are three principal reasons why General Foods challenges Mr. Rowe s exhibits. First, in attempting to calculate the total costs for manufacturing the Regular Maxwell House product sold during a particular period, Mr. Rowe allegedly substituted an inventory accounting system that greatly inflated General Foods' actual costs. Second, in (101) classifying costs as fixed and variable 38 several of Regular Maxwell House s fixed costs were treated as variable costs. And third, the data was presented in such a constricted and fragmented manner as to be an inaccurate representation ofthe business (RPF 16). These three objections wil be discussed seriatim. (aa) The Inventory Accounting System Used By Mr. Rowe 390. Cost of production is the cost required to produce a certain can of coffee during a given period and includes raw material costs, packaging material costs and labor costs (Tr. 8666-71). Cost of production is calculated for each Maxwell House Division plant (Tr. 8670). Cost of sales, on the other hand, reflects the cost of goods sold during a particular quarter and is a national average ofthe cost of production ofthe inventory on hand and current cost of production (Tr. 8666-71). General Foods uses cost of sales as a measurement of its costs (Tr. 8666).

391. Mr. Rowe employed cost of production in computing Regular Maxwell House s costs. When compared with cost of sales data, this tends to increase, or, as General Foods puts it, to "inflate the cost of goods sold during a particular period" (Tr. 8671; RPF 9-17). Since it uses cost of sales as a measurement of its cost (Tr. 8666), General Foods argues that Mr. Rowe should have used this data instead of cost of production figures.

38 Variable costs are those costs that increase or decrease with some relationship to output ('r. 2316). Fixed cost.. are those costs that would be incurred by a firm ifit were producing at the zero output level those costs that the firm incurs DO matter what level of output it produces ('r. 3707). Initial Decision 103 F. 392. Complaint counsel reply that cost of production is the correct methodology for computing below cost sales because pricing decisions are made on the basis of current costs, rather than on the value of inventory, and they point to Mr. Klein s testimony that General Foods does making pricing decisions on this basis: The market prices for green coffee, which we have to pay cash for, on a daily basis, (102) increased by almost the same amount: So our net sales rate had to increase by the same amount to cover our higher cash cost in the business (Tr. 8806). Mr. Hunter also testified that the normal practice in the industry is to price to recover the replacement cost of green beans regardless of the firm s accounting system (Tr. 3084).

393. The choice by the Maxwell House Division of an average inventory system should not affect an analysis whose purpose is to determine whether it was charging prices below contemporaneous costs. Thus, I agree with complaint counsel that cost of production is the fairest method of computing sales below cost because it is the only method that permits comparison of current costs and prices. (bb) The Determination Of Whether Certain Costs Are Fixed Or Variable 394. There are differences between General Foods and complaint counsel over the treatment of advertising, promotional costs and warehousing.

(i) Consumer Coupons 395. General Foods claims that some consumer coupons are not variable costs. In support of this assertion, General Foods demonstrates that over a seven-year period, while the volume sold ofGeneral Foods total regular coffee and Maxwell House has been relatively stable, consumer promotion costs have fluctuated greatly (RX 1153A- B; Tr. 8626-35).

396. Many of Regular Maxwell House s consumer promotion offers are not run on a sales district basis, but instead on either a national or regional basis (Tr. 8652). Even though the volume of coffee sold in a particular sales district may increase or decrease, the costs for redeeming coupons remain relatively constant for these national and regional consumer promotions (Tr. 8651). A single commitment is made to run a national or multi-district consumer promotion, and it cannot be altered to exclude a particular sales district (Tr. 8648-9). Nor is it possible to cancel a consumer promotion once it is issued. 397. Complaint counsel, on the other hand, argue that consumer coupons are variable costs. The requirement for (103) payment on a 204 Initial Decision consumer coupon is the purchase of a can of coffee, and they argue that nothing more directly varies with volume than the one-to-one relationship between a coupon redemption and a sale (CRF 1-18). 398. General Foods concedes in its charts that some consumer coupons are variable costs. For example, all consumer coupons issued under separate sales district promotion sheets are treated as 100% variable. On the other hand, consumer coupons which are issued on a multi-sales district basis, are treated only as 50% variable (Tr. 148).

399. I believe such treatment is ilogical, since for each multi-district coupon redeemed in a sales district there is still a one-to-one a full sale isrelationship with volume: for each coupon payment, made, not half a sale. Thus, these coupons vary directly with volume. General Foods' position ignores the fact that when a coupon is released, whether for an individual sales district or on a multi-district basis, the full impact of that coupon is felt by competition in each area. For example, Paul de Lima, who only sells in Syracuse, would have to compete with a 50 cent Ma):well House coupon in Syracuse regardless of whether or not th same value coupon were also distributed somewhere else.

400. Coupons are directly related to volume. For each coupon redeemed, a sale must be made. General Foods' proposal to exclude half ofthe value of multi-district coupons is thus without merit. Accepting its theory would exclude from consideration substantial variable costs that are incurred in a specific sales district. (ii) Advertising 401. General Foods claims that some advertising is not a variable cost. Complaint counsel, on the other hand, argue that all advertising is a variable cost (CRF 1-23).

402. General Foods treats spot advertising as variable in its own computation of costs (RPF 9-72). The only real dispute is over the inclusion of network advertising as a variable cost in estimating costs at the sales district level.

403. General Foods' assertion that network advertising is a fixed cost is based upon the fact that national network advertising is not controllable on a sales district basis, and the fact that it does not run national advertising campaigns to stimulate volume in particular sales districts (Tr. 8654). (104) 404. The allocation of national advertising costs to prove predation in local markets troubles me somewhat, for the costs are controllable only at the national level. On the other hand, the effects of such advertising are apparent at the local level, for it certainly affects the sales oflocal competitors, just as does spot advertising, which General Initial Decision 103 F. Foods agrees is a variable cost (RPF 9-67). The ability of advertising to affect rival firms is the primary reason that Areeda and Turner maintain that advertising should be included as a variable cost in a predatory pricing analysis (III. P. Areeda & D. Turner Antitrust Law 11 721a, at 191 (1978) (Hereafter Antitrust Law). Since network advertising has the same effect on rival firms as spot advertising, it should not be excluded from such an analysis.

(iii) Warehousing 405. Finally, General Foods argues that only the shipping and labor components of the warehousing cost are related to volume on an annual basis. Warehouse management costs, warehouse overhead costs and taxes and insurance charges related to the operation of warehouses do not vary on the basis of volume. Therefore, Mr. Klein and Dr. Elzinga classified only 50% of warehousing costs as variable (Tr. 8619-22). Complaint counsel argue that General Foods, in the ordinary course of business, assigns warehousing costs to individual products on the basis of volume (Tr. 8720), but it does appear that General Foods is correct in recognizing that certain components of warehousing costs are fixed and it seems fair to consider 50% of that cost as fixed. The effect of such changes on Mr. Rowe s calculation is nevertheless, minimal because warehousing cost is a small percentage of total variable costs (See CX' s 954-57). (cc) The Use Of Quarterly Data For Geographic Areas As Small As A Sales District Or Market Area 406. General Foods maintains that to prove predatory pricing one must show at least one full year of sales that are below average variable cost. Mr. Klein testified that annual or long-term data is preferable to quarterly data because it captures the full business cycle and presents a more accurate accounting picture (Tr. 8672-74), and he claimed that this is especially true in the coffee industry because the volatile nature of green coffee prices can heavily influence the relative price and cost levels (Tr. 8673). Dr. Elzinga argued that, given the entry and exit characteristics ofthe regular coffee industry, annual data, at a minimum, is preferable to quarterly (105) data. Predatory conduct, according to him, is not likely to be successful in only a 90-day period (Tr. 9792-94).

407. Dr. Roman Weil, who was qualified as an expert in accounting, also testified that the most appropriate time period for measuring prices and costs in the coffee industry is "very likely longer than a year" (Tr. 8084). A longer period than a year is necessary in order to capture two important cycles in the coffee business. The first cycle relates to the fact that consumers consume less coffee in hot weather 204 Initial Decision than in cold. The second cycle relates to the growing cycle for coffee beans, which is reflected in the volatilty of green coffee bean prices (Tr. 8084-85). For purposes of properly capturing the seasonal nature of consumer purchases, as well as for purposes of properly encompassing green coffee growing cycles, periods of at least one year and possibly up to seven years may be necessary (Tr. 8085). 408. Complaint counsel point out that this argument is inconsistent with General Foods' business practice, for its own contribution margin test (infra) was for a four to six-week period (RPF 9-115). General Foods also maintains quarterly business records and thus evaluates its business on a quarterly basis (Tr. 9044-8). 409. The reason that Dr. Elzinga chooses a long-run test for predation is that he believes that if General Foods were to sell below average variable cost for a quarter, the smaller firms could simply sit back and let General Foods "bathe in their red ink" (Tr. 9964). This appears to be unrealistic for the record reflects that once share is lost it is hard to regain (RPF 5-32; Tr. 1680, 2528-29).

410. I agree with the tradition)al view that predation consists ofthe sacrifice of short-run revenues to make larger rates of return in the long-run, and I do not accept General Foods' argument that predation can be measured only over a one-year period. For example, Areeda & Turner would not permit a monopolist to engage even in promotional pricing which, by definition, is short run, if it results in sales below marginal cost. II Antitrust Law n 716 , at 177. (dd) Using Sales Districts And Market Areas To Measure Predation 411. Dr. Elzinga testified that because ofthe superior staying power of Procter & Gamble and the ease with which competitors can ship their coffee throughout the United States, sales below cost in a particular sales district or marketing (106) area could not drive out Procter & Gamble, nor could Procter & Gamble and other competitors be kept out during the recoupment period. He believes, therefore, that the presentation of sales district price-cost data is not relevant to a predation analysis when the absence of shipment barriers and actual shipment patterns evidence a much larger geographic market (Tr. 9793-95). While I agree with Dr. Elzinga, I am assuming that the geographic market is no larger than a sales district. (ee) Other Objections By General Foods To Mr. Rowe's Data 412. General Foods claims that Mr. Rowe used inaccurate data in assigning plant costs and in calculating coupon and advertising costs for he relied on planning documents which do not record which plants actually shipped to a certain district (CX 387; Tr. 2270-72, 5119-23) and that, in calculating coupon costs, he relied on projected cost data .. .. ..

Initial Decision 103 F. submitted by General Foods rather than campaign analysis documents which give the final actual cost of a particular coupon promotion. Finally, in calculating spot advertising expenditures, it is claimed that Mr. Rowe relied on Broadcast Advertiser s Report data which is a poor estimate of actual costs.

413. Complaint counsel' s answer to these objections is that Mr. Rowe used the best available data. One of the reasons Mr. Rowe considered his data accurate was a letter from Mr. Robert Y. Fox, then Assistant General Counsel of General Foods, which states that the information had been organized "to show as closely as possible the actual, local, geographic cost data" (CX 436B; Tr. 2313-15). General Foods claims that this letter actually informed Mr. Rowe that the data in volumes 15 and 16 were estimates, because they were recent and based on projections (CX 436C). However, General Foods ignores the next sentence in the letter, and a later sentence, that states that it was providing the accurate data in volume 19. The deal rate information in volume 19 applicable to each quarter is therefore the data that should now be used.

The advertising rates in Volume 19 are therefore the ones that should now be used (CX 436C).

Mr. Rowe did use the data from Volume 19 (Compare CX 954 and CX 955 with CX 263 and CX 264 (the 19 in the upper corner stands for the volume number)). (107) (3) General Foods' Exhibits 414. General Foods introduced its own charts, prepared by Mr. Klein, which analyzed whether it sold below cost (RX 1137-42). However, the time period analyzed was one year rather than the quarterly data of Mr. Rowe, and Mr. Klein analyzed costs in the Youngstown sales district rather than in the separate market areas of Cleveland and Pittsburgh. As did Mr. Rowe, he analyzed costs in the entire Syracuse sales district. As previously discussed, I believe that quarterly data is to be preferred over annual data for an average variable cost test (Finding 412), and since I am assuming for purposes of this analysis that complaint counsel have correctly delineated the relevant geographic markets, analysis of costs in the entire Y oungstowD sales district is not appropriate.

(4) Conclusion 415. I am generally satisfied with the broad implications of Mr. Rowe s cost analysis although General Foods does make some valid objections. For example, it is certainly to be expected that there would 204 Initial Decision be some problems with calculating sales district and market area profits and costs for a company which is national, just as there are (supra). Furthermore, Mr. Rowe s esti-when ROFE' s are calculated mates would have been better ifmore accurate data for some costs had been available. Nevertheless, I do not believe that these problems are serious enough to totally invalidate Mr. Rowe s analysis of average variable and average total cost in the Cleveland and Pittsburgh market areas and the Syracuse sales district. My conclusion finds support in General Foods' own documents which reveal that it did sell Regular Maxwell House below average variable cost in Pittsburgh and Cleveland.

416. Sales below cost occurred in these areas when trade deals on Regular Maxwell House equalled or exceeded Regular Maxwell House contribution margin. Contribution margin" as it is used in General Foods' documents is defined in CX 1024 (CX 1024; Tr. 2335-36). The document states:

Contribution margin is a measure used by General Foos of the variable profitability of a product or total business. It is calculated by subtracting variable sales deduction distribution, and manufacturing costs from gross revenues a', follows: (108) Gross Revenue Less: Cash Discount Expense Returns and Allowances Transportation Expense Warehousing Expense Raw Material Cost Packaging Material Cost Variable Manufacturing Labor Variable Manufacturing Expense Contribution Margin General Foods reporting formats identified contribution margin as the key measure of variable profitability from F' 1971 through F' 1974. Beginning with F'1975, General Foods reporting formats were revised with the term "volume contribution" replacing contribution margin.

The two terms are synonymous, however, and are calculated in the same manner (CX 1024).

417. Simply stated, "contribution margin" is net revenue minus variable costs of manufacturing and distribution (Tr. 7051). Mr. Laster explained further:

Contribution margin is the amount of money that is left in a business after the variable (or) the direct costs have been absorbed, and that money can then either all go into profits, some of it can go into advertising, some of it can go into salaries. But it' basically what is left in the P&L after all direct costs have been absorbed Crr. 7051-52). Initial Decision 103 F. 418. General Foods' definition of contribution margin does not include reductions for trade deals, consumer promotions, and advertising (CX 1024; Tr. 7052, 8862). "Trade deals" includes performance offers, buying allowances and retailer coupons (Tr. 7054-55). If, in a given market area or sales district, General Foods' trade deal rate for Regular Maxwell House equalled its contribution margin for Regular Maxwell House, it would be sellng that coffee below variable cost (Tr. 2336). It would be selling below variable cost by the amount of Regular Maxwell House s other variable costs- consumer (109) promotions and media advertising combined (Tr. 2337). In fact, for the purpose of analyzing sales below cost in Syracuse and Youngstown, Mr. Klein classified all spot television expenditures, all trade deals, all direct consumer promotions, and half of the regional consumer promotions as variable costs (Tr. 9238). Thus, even by its own classification of costs, when General Foods' trade deals were at the contribution margin level, it was sellng Regular Maxwell House below variable cost due to consumer promotions and expenditures on spot advertising.

419. General Foods' own documents show that in Pittsburgh and Cleveland, Regular Maxwell House s trade deals were greater than the amount of Regular Maxwell House s contribution margin. In that instance, Regular Maxwell House s revenues did not even cover the amount of its trade deals (CX 103A-I).

420. The "RMH Contingency Plan Re-Visited " dated November 6 1973 (CX 103A-I), indicated that Regular Maxwell House s trade rate was above contribution margin at that time (CX 103E; Tr. 1576-79). The conclusions in the Contingency Plan were based on CX 103Athat showed figures for Regular Maxwell House s CMLTD (contribution margin less trade deals) for the Pittsburgh market area and the Cleveland market area, which together make up the Youngstown sales district. The exhibit also shows Regular Maxwell House CMLTD figures for individual pack sizes in the Pittsburgh and Youngstown market areas (CX 103I; Tr. 1578). Contribution margin less trade deals does not include reductions for consumer promotions and advertising expenditures (Tr. 8862).

421. The exhibit shows that in Pittsburgh, for the December quarter of fiscal 1974, contribution margin less trade deals for three pound Regular Maxwell House was estimated to be a negative 63.5 cents per unit and a negative $152 100 in total dollars (CX 103I). It shows that contribution margin less trade deals for all pack sizes of Regular Maxwell House in total in Pittsburgh for the December quarter fiscal 1974 was estimated to be a negative 2.7 cents per unit and a negative $9 000 in total dollars (CX 103I; Tr. 1578-79). 422. CX 103 also states that in Cleveland "RMH trade spending is 204 Initial Decision currently above contribution margin on 3# (the 31b. size)" (CX 103E). The exhibit shows that in Cleveland, for the December quarter of fiscal 1974, contribution margin less trade deals for three pound Regular Maxwell House was estimated to be a negative 62.4 cents per unit and a negative $90 700 in total dollars (CX 103I). The estimated losses shown on CX 103I were based upon contribution margins supplied to the Maxwell House Division from the comptroller of the Maxwell House Division. The data was internally generated (CX 103I; Tr. 1579).

423. Thus, while there are some problems with Mr. Rowe s calculations, it is quite clear from its own documents and (110) Mr. Rowe analyses that General Foods did price below average variable cost the Pittsburgh and Cleveland market areas and the Syracuse sales district during a substantial period of time in the early 1970' (c) General Foods' Use Of The Horizon Brand 424. General Foods' Horizon Brand regular coffee was first test marketed in Washington, D.C. in the 1960's; it was later withdrawn (Tr. 1592-2099, 5791-92).

425. The next use of Horizon coffee was in General Foods Red Can offensive" whose purpose was to introduce that brand at the same time Folger entered into areas of the HFNI (CX's 112B, 115B; Tr. 1420). The areas chosen were Wilkes-Barre and Albany (CX's 521D 534A; Tr. 1591- , 2763, 5792).

426. General Foods never intended to use the Horizon brand in its second incarnation as anything other than a "fighting brand" , a brand that is aimed at a specific competitor to disrupt its marketing efforts (Tr. 2849-51) 427. This conclusion is inevitable given statements in General Foods documents outlining the reasons for the use of Horizon39 and the similarity between the Folger can and the Horizon can which was introduced into Wilkes-Barre and Albany (Compare CX 818, which is the Horizon can used in the Washington, D. C. test with CX 819, used during Folger s entry, and the Folger can, CX 820). Mr. Oliver Trone a former General Foods employee, testified that General Foods decided to (111) introduce Horizon into the marketplace at the same time as Folger s with a can that "was about as close to them as our lawyers would allow us to go" (Tr. 1420).

19 A char in "RC Bllinefl Proposition " FehrlJaTY 8 1973, calculates the firml1cial impact of using Horizon in Fong-er s test markets. The chart shows the impact 011 the profit before !.x for the ground category of the Maxwell House Division. By using Horizon, the documenl concludes that the ongoing profit for the Ground Category with the Horizon effort would be $21. 1 millon, whereas, without Horizon, the ongoing profit would be only $159 milloll (CX 1148). Other statements in this domment, although arguably discussing a long-term development program for Horizon, reveal that the primary purpose of the Horizon marketinlJ effortwas to impede Folger s entry into the HFN'I (CX 114C , 114. , 114U). This purpose is explicitly stated in CX 107B: "The introduction of Horizon would severely hamper Folger s iJ:troduction in our HFNI area-It can extend Folger s payback from three to ten years. Initial Decision 103 F. (d) General Foods' Western Retaliation 428. At the time of Folger s expansion into the HFNI, Kansas City and Dallas were among its key western areas because they were high in volume, share, and profitability (Tr. 1399). Kansas City was Folger s hometown; it had a high market share there (55% 0%), and it contributed approximately 20% of Folger s national profits (CX l70C).

429. General Foods decided to step up its marketing efforts in these cities "to retaliate in the West on a one-for-one basis" as Folger expanded into the HFNI (CX 14C; Tr. 1399-1400). The purpose of this retaliation was disclosed by Mr. Bert Einloth, at the time the group product manager for Regular Maxwell House, in a 1973 memorandum:

In addition RMJ- has initiated high-level marketing efIbrts in two major Folger markets, Kansas City/Dallas, intended to drain considerable funds from Folger which could be used in their new markets (CX 725A).

New markets" as used in this sentence referred to current and future introductory markets in the east where Folger did not yet have distribution (Tr. 1584). Mr. William Philips, currently chairman of Ogilvy & Mather, colorfully characterized the retaliation in Dallas as the same as bombing Hanoi" and viewed the increased marketing expenditures in Kansas City as "generally the same" (Tr. 1964-65). 430. The western retaliation was designed to cost Folger moneyhopefully as much as $1 milion per year in each western area where General Foods stepped up its marketing efforts (CX' s 76C, 77B, 97K 170C). While increasing its own market share in the west would have been welcome, to General Foods this was only a secondary objective of the western retaliation (CX's 17P, 19G, 76G). (e) General Foods' "Cora " Campaign 431. Mr. Morgan Hunter, for four years Folger s general manager (Tr. 2707), testified that it had used a spokesperson named "Mrs. Olson" for a long time in the west and that Folger (112) had enjoyed consistent significant share growth for a number of years dating from the time that the "Mrs. Olson" advertising was put on the air (Tr. 3017).

432. Folger had proven and effective advertising copy in the "Mrs. Olson" campaign and Folger continued to use it when it entered the Cleveland market area (Tr. 3059 0).

433. The Maxwell House Division s "Cora" advertising campaign, instituted in July 1972 (CX 463Z125), was considered by Mr. Hunter with some justification, to be a direct copy of Folger s "Mrs. Olson 204 Initial Decision campaign. The campaigns were similar in the development of a character and the dialogue that took place between the young couple and the older woman, Mrs. Olson or Cora. Both involved the idea of the young housewife who cannot make a good cup of coffee, the husband complaining about it, and Mrs. Olson or Cora tellng her what to do about it, thus solving the problem (Tr. 2760). 434. The "Cora" campaign was designed to appropriate for Regular Maxwell House the "slice-of-life" problem-solving approach of the Mrs. Olson" campaign prior to Folger s entry and to make HMrs. Olson" look like a "second rate" imitation when she arrived (CX's 1G; 468F). This would, if successful undermine the effectiveness of the Mrs. Olson" campaign (CX's 88L, 171C).

435. The "Cora" commercial began running in 1972, shortly after Folger entered Cleveland (Tr. 5533, 5535, 5557, 5560-1) CD Conclusion 436. Although General Foods' western retaliation coincided with Folger s eastern expansion, I find nothing inherently anticompetitive in it. It was, to be sure, a signal to Folger that it might have to fight a battle for market share in both the east and west, but it seems to me perfectly legitimate for General Foods to defend its national market share by trying to increase its market share in the west while discouraging Folger s expansion plans. Furthermore, Folger s entry into the HFNI can be viewed as having the same purpose as the Division s western retaliation-to put pressure on the Division in the HFNI so it would concentrate less on its western efforts (Finding 226). Thus, I believe the western retaliation was a legitimate business endeavor, just as was Folger s entry into the HFNI. 437. The "Cora" campaign was also a perfectly reasonable reaction to Folger s eastern expansion. The fact that it copied some of the features of Folger s "Mrs. Olson" advertisement does (113) not mean that it was an anticompetitive act. The antitrust laws do not require a company faced with increased competition to accept only non-imitative advertising ideas from its ad agencies. 438. The Horizon campaign in Syracuse was a different matter: Horizon was not adopted because General Foods believed that it could be made into a successful brand. Its sole function was to blunt Folger Syracuse entry by imitating its packaging. However, the introduction of Horizon and the Maxwell House Division s sales below total and average variable cost in the Cleveland and Pittsburgh market areas and the Syracuse sales district occurred in areas which are not relevant geographic markets; complaint counsel have, therefore failed to prove that these acts occurred in an economically meaningful area. 439. The relevant geographic market in which these practices oc- Initial Decision 103 F. curred is nationwide. Considering Regular Maxwell House s market share in that area at the time of Folger s eastern expansion, even complaint counsel would agree that it did not possess or come close to possessing market monopoly power.

440. None of the prerequisites for predation existed in the relevant geographic market. Regular Maxwell House was not sold below total or average variable cost in this area, and it did not possess market or monopoly power in this area; thus, even ifit had been sold below cost in this area, its prices could not have been raised later to recoup the Division s losses. This is true even if the sales districts are considered to be relevant geographic markets, for new entrants would have been attracted into these areas (Tr. 9540, 9544-50 , 11 96iW4). In short General Foods did not engage in predatory conduct through its Maxwell House Division in the relevant geographic market. J. The Robinson-Patman Charge 1. Introduction 441. Complaint counsel claim that the lowest prices "charged to retail customers" (CPF 12-4) by the Maxwell House Division differed between so-called "predatory" areas (Cleveland, Pittsburgh, Syracuse, Kansas City, and Dallas) and "comparison " areas. 2. The Robinson-Patman Price 442. The lowest price charged to retail customers is defined by complaint counsel as "list price less trade deals. Trade deals are made up ofthree basic elements; buying allowances, performance deals, and retailer coupons" (CPF 12-1). (114J 443. Complaint counsel claim support for their definition of price in a statement by Mr. Rowe that it is "acceptable to an accountant" (Tr. 2372), in statements by industry members that buying allowances, performance deals, and retailer coupons are !!trade deals" (Tr. 8863-64), and in the fact that "dead net" price is defined by some industry members as list price less buying allowances, performance allowances, and retailer coupons (Tr. 2011 , 2508-- , 2749- , 2766 6228-29).

444. General Foods does not disagree with complaint counsel's definition of dead net price, but it argues--orrectiy I believe-that this price does not equal the price which it charges its customers, or which its customers pay General Foods (RPF 10-10). 445. For example, Mr. Hunter defined dead net as "the lowest price that the trade could sell that item and not lose any money on it" (Tr. 2749), not the price that Folger charged its customers. Mr. Zurcher referred to dead net as a term used in sales presentations to the trade 204 Initial Decision to show his customers what they could feature to the consumer (Tr. 6228-29)-again, not the price which the Maxwell House Division charges its customers. Mr. Salesman referred to "dead net to the consumer -not to General Foods' customers-as reflecting buying and performance allowances and RECD's (Tr. 5941), and Mr. Mac- Donald denied that dead net "aligns itself with the price we are charging" (Tr. 7133-34; see also Tr. 11, 887). 446. Despite the statements described above, there may be some industry members who believe that performance allowances or retailer coupons reduce the price charged by the Maxwell House Division to its customers, but their opinion does not comport with the way price has been defined by many of those who have been forced to grapple with the complexities of the Robinson-Patman Act,4o (a) RECU's 447. Although complaint counsel argue in the section of their findings devoted to the definition of price that "coupons both RECD' and consumer coupons-are price reductions Hgiven by General Foods" (CPF 3-190 to 3-192), they claim that only RECD' s (along with buying allowances and performance (115) offers) should be deducted from list price to arrive at the Robinson-Patman price. 448. The reason for the difference in treatment is obvious: it is literally impossible for Robinson-Patman purposes to conceive of a coupon which is packed in every can of coffee or featured in every edition of a newspaper, which involves no retailer in its preparation or distribution, and which identifies no particular retailer as a price reduction " " the retailer. It clearly is not: it is simply a way in which a consumer may reduce his purchase price if he redeems the coupon at a retail store. In this case, the retailer serves merely as a convenient way in which the discount to the consumer is passed on. The redemption of the coupon by the retailer quite simply does not affect the price he pays to the Maxwell House Division or the price which the Division charges him 41 for he receives from the Division only the amount which he previously gave to his customer (plus a handling charge). From the point of view of the retailer, the transaction is a wash.

449. The only difference between a RECD and a consumer coupon is that particular retailers arrange the publication ofthe former, and identify themselves as the place where the coupon may be redeemed. The effect from the point of view of the consumer and the retailer is '0 Complaint counsel recognize that the definition of the Robinson-Patman price "ill ultimately a legal conclu jon. TI,C perceptions of industry witnesses as to what COIltitutel! price cannot be detcrmiuativc" (CPF 12-). ;1 OfcourEu, as I fmd above, a redeemed coupon doc reduce the Division s revenue, but "revenue received" or mill net" is IJot a Robin on-Patman price Initial Decision 103 F. identical, and neither coupon effects a reduction in the price that retailers or wholesalers pay to the Division for their coffee. 450. Theodore Engel, a vice president of The Kroger Co., a large retail chain, stated that neither kind of coupon reduces his price: Q. Mr. Engel, why haven t you included the retailer coupon in your definition of price? A. Because that's all passed on to the consumer. Q. So the store is really just a conduit for the coupon? (116) A. Right (Tr. 1656-7).

During surrebuttal testimony, the point was reiterated: Q. Mr. Engel in your explanation of how you go about determining the price that you pay to Regular MaxwelllIouse and other coffee roasters, you did not mention retailer coupons or manufacturer s in-ad coupons as being a deduction in determining that price. Why is that, sir? A. Well, that is b8.ically a transaction between the vendor and the customer. I am a conduit for that. I transfer the funds, but that is the sum and substance of what happens (Tr. 12 083-4).

Q. SO then you don t consider any COUpon as a part of the price that you pay to General Foods? A. That I pay, no, sir (Tr. 12 085).

451. Roger Metzger, head buyer for S.M. Flickinger & Co., a major wholesaler in the Buffalo area, testified that RECD's are not reductions in the price that a wholesaler pays for coffee: Q. So then there is no reflection of a RECU or other coupon in the price that you pay to Maxwell House Division? A. No. As a wholesaler we re not really-in our pricing-involved in that. Q. So (doj you consider any type of coupon whatsoever as effecting a reduction in the price that you pay General Foods? A. No, we don t (Tr. 1l 5). (117J 452. Louis Epstein, president of Golden Dawn Foods, Inc., a major wholesaler in the Cleveland-Pittsburgh area, confirmed that coupons are not considered by him, as a customer of General Foods, as effecting a reduction in the price paid to General Foods. He considered RECD' s as:

. . . a device for lowering the effective price that the consumer pays, but after a sale is over in which a coupon is used, our inventory is valued at the list less the performance and the advertising allowances.

204 Initial Decision The coupon is all over. So we don for that reason--consider it a reduction in the value of our inventory or in the price we paid for it (Tr. 12 166). 453. Allen Toy of Hills Bros. explained the manner in which retailer coupons are used:

A. She would clip out the coupon, take it to her store, and purchase coffee and at the checkout counter she would be reimbursed for either the amount of the coupon or she would receive the coffee at whatever the reduced price is in the coupon. A. The retailer would then give it back to the manufacturer who offered the allowance and he reimbursed for the coupon (Tr. 2010). 454. Mr. Toy testified further that it is the industry practice for manufacturers to reimburse retailers only for coupons properly redeemed and that both RECU's and consumer coupons are redeemed in the same manner:

JUDGE PARKER: Does General Foods pay the dealer money for every (118) dealer coupon he puts in his ad or is it just on the ones redeemed-returned to you? THE WITNESS: lThatJ is the industry practice, yes, sir. JUDGE PARKER: So, in that sense, it is the same as a consumer coupon, or is it? In other words, you only pay for the ones actually turned in by some consumer? THE WITNESS: Yes, that is the practice (Tr. 2014-15). 455. When asked what effect a consumer coupon had on the price paid for coffee by the consumer, Mr. Toy answered: The net result would be that it would enable the consumer to buy coffee at a discount from the regular price (Tr. 2015).

And, in a follow-up exchange with, Mr. Toy responded to this question:

JUDGE PARKER: What about dealer Coupons, same thing, same result? THE WITNESS: Yes. Same result for the consumer (Tr. 2015). 456. While the fact that retail grocers may not participate in RECU' s does not, in my opinion, affect the definition of "economic price, 42 it does indicate that RECU' s are not a reduction in price to them, and there is evidence that some (119) wholesalers and retailers do not always use them. Mr. Epstein of Golden Dawn Foods explained that:

.2 For, as I tind above, to tbe extent that HEeD'g are used hy retailers and redeemed by them, they afect the Maxwell House DivisiotJ s net revenue Initial Decision 103 F. . . . we try to balance our advertising program to include items that the consumer would want on a periodic basis and if, for example, there is a flood of coupons on a particular commodity category, we would decline to use them all because of these not being a balanced program (Tr. 12 168).

457. Mr. Engel recalled that Kroger did not participate in all RECU promotions offered by the Maxwell House Division in Cleveland and Pittsburgh following Procter & Gamble s introduction. In fact, Kroger rarely used the one pound RECU's offered on Regular Maxwell House in Pittsburgh (Tr. 12 084).

458. Mr. Metzger also testified that the Buffalo area retailers serviced by Flickingers do not use all of the retailer coupons offered: ... there are many om rings, some ofthem they use, some of them they don t have room for. There is really more than they can use. And it's a process of looking for the best that they can offer to get the-to attract that customer into our store each week (Tr. S70).

459. Mr. Keller pointed out that all customers in the Division southern region, such as Winn Dixie and Colonial stores, do not use every RECU offered them and generally indicated that: ... quite a few customers don t take advantage ofRECU's at all They don t participate in them, don t want them, don t like them for any number of rea.,-ons: too messy, too much work, too much trouble (Tr. 6616).

460. John Mann described an "extensive list" of customers who did not participate in the Division s RECU offers (Tr. 6516), and Mr. Salesman agreed that some customers do not participate in RECU' (Tr. 5941). (120J 461. The refusal of retailers to eagerly embrace RECU's at all times reveals that from the point of view of the retailer, the RECU' s are not price reductions to them. This, and the indisputable fact that redemption of the coupon results in no reduction in the price retailers pay requires the inevitable conclusion that RECU's should not be deducted from the Maxwell House Division s net price to arrive at the Robinson-Patman price.

(b) Performance Allowances 462. General Foods agrees that while RECU' s do not put any money in the retailer s pocket, both non-performance and performance allowances do (RPF 10-54),43 and complaint counsel argue that since the amount of money received by the retailer in performance allow- 1 General Foods agrees that buying allowances, for which no performance is required, are price reductions (RPF 10-3) 204 Initial Decision ances from the Maxwell House Division far exceeds the cost of performance, performance allowances are really price reductions. 463. The Maxwell House Division performance allowances permit unlimited purchases regardless of the level of performance, for the offers provide that if the retailer performs, he wil receive an allowance of a certain amount per unit on all the coffee he buys during the offer period. There is no maximum limit on the volume the retailer can buy and the amount of the discount he can receive. (CX 1148; Tr. 5846 6003-08 6016-18 5635-36, 6542-45 11,888). On most performance offers, the retailer only has to perform once to get credit for the allowance on all the volume purchased during the offer period (CX 1148; Tr. 6003-04, 6542-45, 11,899, 12 182). 464. Although there is no direct evidence as to the difference between the amount of performance allowance received by retailers and the cost of performance, it is unquestionably large in most cases, for the Division s offers do not specify the size of the advertisement the grocer has to run; for example, a retailer can run one very small advertisement, similar to the small print used in obituary columns and stil collect the allowance on all the volume purchased (CX 1148; Tr. 6005).

465. The value of performance allowances to retailers is obvious, and the rate of acceptance of such deals is very high-(121) around 90% (Tr. 2391 , 2393, 7494). Dr. Greer testified that the rate of acceptance of performance allowances indicates that they clearly are price reductions (Tr. 11 419-21), but other matters-both factual and legal -must be considered before a decision on the status of performance offers can be given.

466. The most important matter is the intent of the offeror and the recipient. If neither the Maxwell House Division nor the retailer believe that performance is a real condition of the offer, then it can be said with confidence that the performance offer is simply a disguised price reduction.

467. That is not the case here, for the Division insists upon performance, minimal though it may be." For example, Mr. Engel testified: Q. Why do you always anticipate that you re going to perform on a performance offer? A. One reason is to assure payment. Some we default and pay the vendor back (Tr. 1653).

Mr. Engel also stated during later testimony: Q. What happens, Mr. Engel, if you don t perform? A. We refund the money to the vendor.

;4 The word "minimal" refers to the cost incurred by the retailor, but I also recognize that the benefit received by the Division may well be much greater because of the exposure given to its coffee by retailer advertisemenl. . . .

Initial Decision 103 F. Q. And does General Foods check up on your performance? A. Yes sir.

Q. Could you say (that) General Foods checks up more or less than other grocery product manufacturers? (122) A. General Foods and Procter & Gamble are the most diligent in making sure that we perform per their contracts (Tr. 12 080). 468. Mr. Metzger and Mr. Epstein, both wholesalers, explained the manner in which they pass through performance offers to their retail customers and police performance by such retailers jointly with General Foods:

Q. Then what happens, Mr. Metzger if a retailer does not perform in accordance with the requirements of the offer? Does he get the allowance anyway? A. He would get the allowance initially, but at the point where the performance was not produced, we would bill him back for the cases that we sold him (Tr. 11 867). 469. After Mr. Metzger confirmed that Flickinger s had indeed biled back its retail customers for failure to perform, he described the procedure for determining performance:

Q. Is Flickinger s responsible for determining whether or not the performance has been produced? A. We jointly with the manufacturer determine the performance (Tr. 11 867). When asked whether the allowances under performance offers were passd on to the retail stores serviced by Golden Dawn, Mr. Epstein answered:

We do. It' s the nature of a voluntary group that we act as an agent for all of our stores. We see to it that they perform through advertising groups and in doing that we pass along to them, at our billng price, all of the performance allowances up front (Tr. 12,164). (123) 470. In testimony which followed, Mr. Epstein described the policying or checking up done by General Foods to determine whether the stores served by his firm were performing in a manner consistent with the terms of performance offers and, after explaining that all those retail stores did not participate in all Regular Maxwell House performance oilers, he concluded:

Q. And when you do receive the performance allowance, it' s because you have decided that performance will take place? A. That's correct (Tr. 12 167).

471. Mr. Keller ofthe Maxwell House Division explained the procedures followed and emphasized that if the customer fails to match the .. . .. . , 204 Initial Decision specifications of a performance deal then we don t pay. . . " (Tr. 6605).

we do police all our offers to satisfy ourselves that they have occurred. The policing is done in several ways. One that I mentioned earlier was we have a retail sales organiztion, 120 people out there that watch and monitor every ad that occurs. . (Tr. 6604) 472. Mr. Keller continued . . . every time a customer performs, he has a responsibility of fillng out a 5812 and signing a certification that he has performed. To that 5812 certification, he attaches his proof of performance that identifies this as proof of performance. He verifies that. . . . So we have that piece of document, as well as our own retail follow-up to make sure that our customers are performing " (Tr. 6605).

473. Mr. Mann also described the manner in which the Division insured compliance with the performance requirements specified in its offers (Tr. 6504-505). Mr. Mann explained that a customer who deducted performance offers before performing was biled back ifthe required performance was not subsequently provided: Q. And does the performance ufthe particular customer also follow the (124) requirements that are specified in the deal bulletin? A. If they don t they don t get paid. So yes, they must perform according to the specifications as outlined on the deal bulletin and comply with them in all of their stores, either by virtue of advertising or displaying, in order to receive the trade deal in effect by virtue of the deal.

Q. Now in some instances you indicated that the allowance is taken off the invoice. A. That' s correct.

Q. In that case, supposing that the customer does not complete a 5812 and furnished the required proof of performance; what happens then? A. We bill the customer back. . (Tr. 6514) 474. Complaint counsel concede that performance allowances require "some performance" (CPF 3-254) but argue, as does Dr. Greer that: "A price reduction when associated with some performance on the part of the buyer is still a price reduction" (Tr. 11 412). While I respect Dr. Greer s economic credentials, he is not an expert on Robinson-Patman law. When its case law is analyzed (infra), I come to the conclusion that the Division s performance allowances are not discounts from price, but are advertising allowances which, along with RECU' , are cognizable only under Section 2(d), not Section 2(a). 475. Since complaint counsel's price discrimination claim is based upon treatment ofRECU' s and performance allowances as deductions from list price, their "price discrimination" charts (CX's 924-953) do not establish that General Foods' Maxwell House Division charged Initial Decision 103 F. different prices for its coffee in different areas of the country. Therefore, the Robinson-Patman Act count of the complaint wil be dismissed.

3. General Foods' Meeting Competition Defense (a) General Foods' Intent 476. Although I wil dismiss the Robinson-Patman count because complaint counsel have failed to prove that the Maxwell (125) House Division sold regular coffee to its customers at different prices, the parties are entitled to my views on General Foods' meeting competition defense.

477. Several General Foods employees and customers testified about its general response to Folger s entry into those sales districts where the Maxwell House Division allegedly sold at lower prices than in other sales districts.

478. Mr. Bohn and Mr. Laster testified that General Foods' policy was to meet Folger s offers but not to sell below General Foods' definition of cost in Cleveland, Pittsburgh and Syracuse (Tr. 4516, 4535 6970, 6975). The basic principle was not to aggress: Q. Now, you have also said, Mr. Laster, that another principle that was laid down was the principle not to aggress. Would you state for us what the factual content was of the principle not to aggress? A. The factual content was that our policy prohibited the organization from leading in any element of the marketing mix in the Cleveland market (Tr. 6972). 479. Mr. Epstein, president of Golden Dawn Foods of Sharon, Pennsylvania, recalled the entry of Folger s into Cleveland: . . . when Folger entered the market, they attempted to meet the lowest tier of competition which at the time was Hils Bros. Hils. Bros. deal rates were the best in the industry at the time and (at the time of) the Cleveland introduction, Folger being under a consent (126) decree, was careful only to meet competition and the lower competition at that time was rrils firos., as I remember. Q. How do you know Procter & Gamble was operating in a fashion which enabled them only to meet the highest deal rates ofiered by any competitors? A. I discussed that with their sales representative at the time (Tr. 12 168-9). 480. Mr. Epstein testified further that Regular Maxwell House responded to the "heating up of competition " and Folger s increased rates, by meeting the competition (Tr. 12, 169, 12 178). 481. Mr. Lloyd Nelson, the Maxwell House Division s national sales Thatis Not to have bigherirade deals than Procter & Gamble And the same for consumer promotions and advertising ('Ir. 6973).

204 Initial Decision manager, testified that exceeding Folger s trade rates "was an abso- I ute no-no" (Tr. 5765) and that:

As I mentioned earlier, we would never exceed the competitive rate in the marketplace. Number two, we had to have what we call hard data which was an absolute clear-cut confirmation of the competitive rate. Three, we couldn t sell below cost and I guess those are kind ufthe three guidelines that the sales organization was operating under etr. 5769).

482. Mr. Engel, an executive of the Kroger Company, and a customer of both General Foods and Procter & Gamble at the time of Folger entries into the Cleveland and Pittsburgh areas, testified that the introduction of Folger s into both areas raised the value of buying allowances, performance offers and RECUs offered by coffee manufacturers and that Folgers' initiated the escalation of these promotional offerings. He further recalled that, in Cleveland, Regular Maxwell House made no initial response to Folger s but later on in year one, met Folger s promotional activity (Tr. 12 076-77). 483. Mr. Epstein recalled that, when Folger entered Pittsburgh Regular Maxwell House matched Folger s escalation of trade deals (Tr. 12 170) and Mr. Nelson testified that in Philadelphia and Pittsburgh the Division attempted to meet Folger s promotional programs: (127) . . . in the Pittsburgh area we met them in what I would call across the board on trade deal, consumer promotions, advertising. In Philadelphia, if my memory is correct, we did not meet them on the trade deal line, again for some ofthe reasons that we talked about earlier, relative to trade flow (Tr. 5771) 484. Mr. Zurcher testified that Regular Maxwell House employed a "following" posture in arriving at a matching of Folger s dead nets in Pittsburgh:

So we adopted a following posture, again very similar to which we eventually got into in Cleveland where through those sources I described, I could confirm and reported neither and, of course, I had to wait until 1 get approval back, and then we would match their dead nets (Tr. 6289).

485. At first, the Maxwell House Division decided not to match Folger s trade deals in Syracuse (Tr. 5785--6). According to Mr. Salesman, because of the Division s response (a trade rate plus RECU as compared to Folger s straight trade rate promotion (Tr. 5952)): We were at a significant disadvantage because we were dead netting to the consumer. We were not dead netting to the trade. In other words, our buying allowances and performance over was not as heavy, not as high as Folger s (Tr. 5954). Initial Decision 103 F. 486. Later, Regular Maxwell House matched Folger s trade rates (Tr. 5956); however, Mr. Salesman stated that he stil had to ascertain Folger s rate before he responded:

As I said earlier, we ran a lag strategy. We followed. I could not initiate an offer, nor was an offer (128) initiated until we verified what the Folger offering was in the marketplace with hard copy of what that ofiering was (Tr. 5957). (b) Mr. MacDonald's Charts 487. In addition to the testimony discussed above, General Foods presented two charts prepared by Mr. MacDonald (RX's 1207 and 1292) which attempt to reconstruct his employer s response to Folger in Cleveland, Pittsburgh and Syracuse, on one, two and three pound coffee for approximately a year after Folger s entry into those cities. There are, however, some problems with these charts which are so significant that they cannot be accepted as support for General Foods claim that it was simply meeting Folger s offers in those areas. 488. RX 1207 represents Mr. MacDonald's first attempt to prove that General Foods was meeting Folger s competition, but his reconstruction suffers from the fact that the flow charts upon which he relied and which were prepared by General Foods' employees are incomplete. For example, in Syracuse, there were no flow charts from February 3 to March 31 (RX 1207K) and from April 28 through September 22 (Tr. 7200-2). Thus, for eight of the 12 months depicted in Syracuse, General Foods did not produce any records that showed its knowledge of Folger s deals at the time they were offered. During the first quarter of Folger s entry into Pittsburgh, there were also no contemporaneous documents showing General Foods' perceptions of Folger s deals (RX's 1044C, G; Tr. 7172).

489. In addition, these documents rely on a concept called "pyramiding -which, while it reflects Mr. MacDonald's perception of Folger s offers which the Maxwell House Division was meeting, does not in my opinion, reveal the perception of Division personnel who responded to the Folger offers when they were actually made. 490. "Pyramiding" occurs when a producer "double(s) up one deal on top of another" (Tr. 7142). For example, a producer might offer a $2 trade deal in one period, and offer a $1 count and recount in the next. A retailer, aware of both offers, would purchase in one period and move the product from his warehouse to his stores in the other (Tr. 7142-43) so that he could receive a $3 discount. 491. Since offers are made continuously by producers, it is possible for one who is reconstructing dealing events to "pyramid" them and prove" that a response which appears at (129) first blush to beat 204 Initial Decision competition did not do so because it was intended to meet "pyramided" offers.

492. I do not mean to imply that "pyramids" do not exist in the coffee industry. They undoubtedly do. The problem which I have with Mr. MacDonald' s reconstruction is that he is a little too ingenious in calling pyramids which benefit General Foods. Thus, his judgment is so subjective that I simply cannot accept his reconstruction. 493. For example, in Pittsburgh, on three pound coffee in September, Mr. MacDonald pyramided Folger s deals (Tr. 7429). For this pyramid, he claimed that retailers would buy heavily on a $3. buying allowance and a $.60 count and recount, and then would pyramid that $3.64 rate on to a $2.04 count and recount and $.53 retailer coupon in September, with a resulting dead net of$1.955 (RX 1207G). However, this pyramid could not have taken place. A count and recount is an offer under which the grocer gets paid for moving product from his warehouse to his store (Tr. 6770). It is impossible for a retailer to pyramid two count and recounts in a row, as Mr. MacDonald did with these deals. A retailer cannot collect on two count and recounts on the same coffee, since he already moved the product from the warehouse into the store on the first count and recount, and therefore cannot perform on the second count and recount. Mr. MacDonald suggested on cross-examination that some fictitious retailer might have moved the product from the warehouse to the store and then moved it back to the warehouse so that he could ship it back to his store (Tr. 7596), but there is no evidence that any retailer would do this. Also, Mr. MacDonald conceded that there would be no substantial volume moved on this type of pyramid since the Division would never pay a grocer in the situation where there are two count and recounts on the same can of coffee (Tr. 7597-98). 494. Although the potential for pyramiding by Regular Maxwell House is as great as that for Folger, Mr. MacDonald appears to have ignored many ofthem in RX 1207. For example, in Cleveland on two pound coffee in March, Maxwell House was running a $2.40 advertising or display allowance (RX 1207B). In April, that allowance was reduced to $1.20 (RX 1207C). According to Mr. MacDonald's own definition of a pyramid, grocers would have purchased heavily on the $2.40 allowance in March and featured with Maxwell House s 32 cents retailer coupon in April (RX 1207C). Mr. McDonald did not pyramid the earlier allowance, although the situation involved a reduction in case rate and although, in a similar situation, he called a pyramid for Folger (RX 1207B, C; Tr. 7353). 495. Complaint counsel' s reply findings reveal that this was not unusual and that in other instances Mr. MacDonald did not call a Initial Decision 103 F. Regular Maxwell House pyramid when he should have (CRF 5-31). (130) 496. Other problems exist in Mr. MacDonald's reconstruction. On RX 1207 J, he reports a $6.24 buying allowance and a $3.70 advertising/display allowance for Folger running from September 30 through December 2 (RX 1207J). However, Mr. Salesman testified that according to the flow chart which he prepared for this area, Folger only had a $4.32 buying allowance and a $3.70 performance allowance upon entry. He testified that he did not learn that Folger added a $1.92 allowance until December (Tr. 6192-93). Folger reduced revenue recommendations46 confirm his perception. They show that Folger entered Syracuse with a combination of allowances that added to $4. 32 plus a $3.70 merchandising allowance (RX 1185A). These allowances expired on November 2, 1974 (RX 1185A). It was not until November 4 that Folger added the additional $1.92 allowance (RX 1185J). Thus, Mr. MacDonald ignored the testimony of a General Foods' witness, as well as corroborating evidence from Folger s business records.

497. RX 1207 shows the dates on which deals became effective, not when they were announced to the grocery trade, but Mr. MacDonald testified that Folger invariably announced its deals first. For example, in Cleveland on the two and three pound sizes, both Folger s and Maxwell House s dead net prices were effective on March 4. Mr. MacDonald stated that it was his considered opinion that Maxwell House waited for Folger to announce first (Tr. 7348). In rebuttal complaint counsel introduced CX's 1392, 1397, and 1398 which challenge the conclusion that Maxwell House invariably announced its deals after Folger. These exhibits show that in many instances when the dead net prices were identical, Maxwell House announced its deals before Folger, and that Maxwell House could not have been responding to Folger s deals.

498. After the introduction of complaint counsel' s rebuttal exhibits Mr. MacDonald introduced a revised set of meeting competition charts. His revision, which shows that General Foods was meeting different dead net prices than those shown on its first set of charts suggests that neither of the charts are entitled to serious consideration.

499. On RX 1292, Mr. MacDonald entered a key that supposedly indicated the Folger dead net price to which Maxwell House was responding. These assignments of Maxwell House s deals to Folger prices were derived from Mr. MacDonald' s (131) experience, but they '" Reduced revenuc recommendations record Folger s perception of competitive uffer and jt. ponse (Tr 27fia) 204 Initial Decision are unsupported by documents (Tr. 12 208, 12 314). 47 In effect, Mr. MacDonald took the dates from complaint counsel's exhibit and arranged his key in a manner most favorable to General Foods. Whenever the dates indicated that Maxwell House announced a contemporaneous price before Folger, Mr. MacDonald structured his second chart to show that Maxwell House responded to some earlier Folger deal that had already expired.

500. During his first four days of testimony, Mr. MacDonald claimed that Regular Maxwell House s deals were responsive to Folger s deals in effect at the same time. He continually spoke in terms of "rounds" of deals. In each quarter, there would be two or three rounds of deals (Tr. 7223-24). An examination of his first testimony reveals that he analyzed the first set of charts in terms of Regular Maxwell House s response to Folger s contemporaneous rounds of deals. For example, he claimed that its first round of deals was responsive to Folger s firstround of deals (Tr. 7450-53, 7231, 7385-6, 7417- 7445-6 7484 7697-98). However, on the second set of charts, Mr. MacDonald changed his approach and claimed that Regular Maxwell House was responding to Folger deals well after they took effect. 501. For example, during his first testimony, Mr. MacDonald stated that in Cleveland on the three pound size, Regular Maxwell House waited to announce its March dead net price of $1.92 until Folger announced its March $1.92 dead net price (Tr. 7348). However, Regular Maxwell House announced its $1.92 dead net price on January 11, while Folger did not announce its dead net price until January 21 (CX 1392B). Mr. MacDonald then changed his position. On RX 1292, he claimed that Regular Maxwell House s March dead net price was responsive to Folger s dead net price that expired in December (RX 1292C, deals 5D, 15H, key 2).

502. In another case, Mr. MacDonald claimed that in Syracuse in September, the $3.51 dead net price for Regular Maxwell House on three pound was responsive to a $3.51 dead net price for Folger (RX 1292I, deals 7, 17, key 7). For the same time period, he claimed that the $2.36 dead net price for Maxwell House on two pound was responsive to the $2.36 dead net price for Folger (RX 1292H, deals 5, 14, key 7). On b,)th the two and three pound deals, both the Folger and Maxwell House announcement dates were August 26. Despite the exact same date, Mr. MacDonald claimed that Maxwell House was responding to Folger. However, Folger s contemporaneous records show that it (132J was responding to Maxwell House. In fact, Folger had obtained a Maxwell House deal form that specifically identified the Maxwell House dead net price during the period (CX 1379Z-32, Z- 17 Tbe only document to which Mr. MacDonald referred was RX 301, but this does not support his delailed reconstruction.

Initial Decision 103 F. 120). Mr. MacDonald' s unsupported reconstruction simply does not outweigh the contemporaneous documents demonstrating that Folger was responding to Maxwell House.

503. Complaint counsel have fashioned a chart (following CRF 5-53) which I will not reproduce; however, it presents solid evidence that Mr. MacDonald assigned dead net prices in a manner favorable to General Foods without any consistent or logical justification. This chart lists the deal assignments made by Mr. MacDonald. The dates for Maxwell House s and Folger s deals show the dates on which the deals for both competitors became effective. The lines demonstrate the key used by Mr. MacDonald.

504. The chart reveals that when the prices and dates of contemporaneous deals favored General Foods, Mr. MacDonald aligned the deals vertically (vertical arrow). He indicated that Maxwell House was responding to the Folger deal in effect at the same time. However when the dates or prices did not favor General Foods, Mr. MacDonald changed approach and claimed that Maxwell House was responding to a Folger deal well after the Folger deal became effective (rightslanted arrow). For example, the chart for Pittsburgh shows that the two pound dead net price for Maxwell House beginning on February 18 was responsive to the Folger dead net price that began on November 19, three months earlier. However, the same chart shows that the Maxwell House dead net price beginning on May 14 was responsive to the Folger dead net price beginning at the same time. 505. Complaint counsel' s criticisms of Mr. MacDonald's charts are well founded, and I am satisfied that his use of pyramids and his assignment of deals on RX 1292, while creative and ingenious, are so subjective that they do not prove the perception of Folger s deal by General Foods employees during the relevant time periods. Thus despite the evidence of general intent, I could not find, if it were necessary, that General Foods' has satisfied its burden of proving that it met, but did not beat, Folger s deals in Cleveland, Pittsburgh and Syracuse. (133) III. CONCLUSIONS OF LAW A. The Relevant Geographic Market For Regular Coffee Is Nationwide 1. Introduction Complaint counsel' s claim that General Foods attempted to monopolize the HFNI "markets" (CLA, p. 1) requires a finding that the relevant geographic market for regular coffee is no larger than a sales district, for they do not allege that in any larger area-either the 204 Initial Decision entire HFNI or nationwide-id General Foods possess the market power to succeed in an attempt to monopolize the sale of regular coffee.

I cannot make such a finding, for the pertinent evidence on the issue reveals that the relevant geographic market for regular coffee is nationwide. Much of the evidence offered by complaint counsel relates to the marketing activities of wholesalers and retailers, but this is irrelevant in a determination ofthe market for the production and sale of regular coffee by roasters.

2. Market Areas And Sales Districts Are Not Relevant Geographic Markets (a) Transportation Costs The relevant geographic market for a product is the area to which customers can practicably turn for alternative sources of supply. Tampa Electric Co. v. Nashville Coal Co. 365 U.s. 320, 327 (1961). A recent Commission case, Pillsbury, Inc. 93 F. C. 966 (1979) applied the Tampa Electric test to a grocery product. The issue in Pillsbury was whether that company s acquisition of Fox Deluxe Foods violated Section 7 of the Clayton Act. The parties agreed that the United States was a relevant geographic market, but complaint counsel also argued that regional submarkets existed. The Commission found that transportation costs posed no significant barrier to the distant shipment of frozen pizzas and held that geographic submarkets did not exist: (134J The test fot mea-suring geographic market is where consumers (in this Care retailers) can practicably turn for an alternative source of supply. Here the record is clear that frozen pizza manufacturers could sell virtually throughout the United States from a single plant with no significant cost disadvantage. Thus, the power of any given group of sellers serving a city or region at a given time to raise prices is limited by the capacity of virtually all other domestic manufacturers to compete on practically an equal footing in that city or region-an economic situation which requires a finding of a national market and the elimination of geographic submarkets. Id. at 1030. Complaint counsel point to the Commission s statement in a footnote that "special factors, like slight economic barriers, could produce submarkets (ld. n. 8; CLA, p. 108), but this record reveals no economic or legal barriers to the nationwide shipment of regular coffee (Findings 105-16).

Another Commission Section 7 case involving a grocery product emphasized the significance of transportation costs in the determination of the relevant geographic market. The complaint in Golden Grain Macaroni Co. 78 F. C. 63 (1971), aff'd 472 F.2d 882 (9th Cir. 1972), cert. denied 412 U.S. 918 (1973), alleged that the relevant . . .

314 EDERAL TRADE COMMISSION DECISIONS Initial Decision 103 F. geographic market for the production and distribution of dry paste products was a four-state region in the Pacific Northwest. The Commission stated that:

The appropriateness of the Pacific Northwest as a geographic market in which to determine whether respondents possess monopoly power turns on the significance of transportation costs to outside producers. Id. at 158. While the Commission found that the Pacific Northwest was the relevant geographic market because "freight costs constitute a barrier suffciently high to forestall entry by firms with production facilities outside" that area Id. at 159, it refused to accept complaint counsel's argument that there were (135) also three geographic submarkets. The reason was that "the record fails to show that transportation costs between one and another ofthe three areas is so great in relation to the overall cost (of dry paste J as to make these areas separate submarkets. Id. at 159.

The Commission in Golden Grain remarked on the same phenomenon that exists in this case-the shipment of products into the alleged submarkets by outside producers:

both Major and respondents, although operating exclusively out of plants in Seattle sell a significant amount of their products to retailers in the three alleged submarkets. Id. at 159.

Other cases dealing with the issue of relevant geographic market stress transportation costs. In United States v. Hammermill Paper Co. 429 F. Supp. 1271 (W.D. Pa. 1977), the government argued that addition to a nationwide market for paper manufacturing, New England was a relevant market because an acquired paper merchant was the largest firm of its kind operating in that area. The court rejected this argument:

Paper manufacturers located in New England compete effectively in the Midwest against manufacturers with mils located there, and mills in the Midwest compete for sales in New England. Paper manufacturers in New England do not sell primarilyin New England but sell throughout the United States. Paper manufacturers do not find it necessary to build mils in New England to participate in the New England trade. Id. at 1278.

In RSR Corp. 88 F. C. 800 (1976), the Commission held that the fact that secondary lead manufacturers "can and do frequently ship products into regions far distant from their plants" demonstrated substantial regional interdependence even in the face of relatively high transportation costs. Id. at 885. See also International Telephone 204 Initial Decision & Telegraph Corp. v. General Telephone Electronics Corp. 518 F. 913 , 937 (9th Cir. 1975). (136) (b) Complaint Counsel's Argument Relies On Irrelevant Facts Complaint counsel' s emphasis on local demand factors injects irrelevant considerations into the issue of relevant geographic market. The error in their reasoning was described accurately several years ago in The Market: A Concept In Antitrust 54 Colum. L. Rev. 580 598-99 (1954):

Underlying much of this discussion is the premise that it is essential to distinguish between markets at different levels in the chain of sellers. This has not always been clearly discerned by the courts. An example of the resulting confusion is found in United States v. National City Lines, Inc. where the defendants were charged with conspiring to monopolize the sale of busses and bus supplies in forty-five cities by means of exclusive dealing contracts which excluded others from selling to the transportation company that operated the busses in those cities. One defense was that Section 2 ufthe Sherman Act applied only to monopolization of a geographic market. Holding against the defendants, the court stressed the fact that the operating company had a monopoly in the forty-five cities, and that this constituted the geographic market which the suppliers had monopolized by excluding competitors. The court apparently regarded the forty-five cities as a geographic market for the sale of bus supplies, an absurd view in light of the fact that bus products are sold on a nationwide basis. The area in which the buyers of bus supplies do their purchasing-in this case, the whole nation-should be the geographic market for the sellers of these supplies. The court in Hammermill made the same point: The arguments that paper merchants in New England "cannot and do not compete with paper merchants located in other sections of the country" does not establish that New England is a separate section of the country for measuring the anticompetitive effect ofl137) Hammermil' s acquisition, because the same criteria apply to all paper merchants in any section of the country. The principal function of a paper merchant anywhere is to serve the local printing trade with a variety oflines, locally warehoused and available for quick delivery in a large range of quantities. We cannot find the facts establishing the close relationship between paper merchants and printers to be relevant to the question of the effect of this acquisition to lessen competition among manufacturers of printing and fine papers in the sale of their products to paper merchant.o;. 429 F.Supp. at 1278-79. Here, too, complaint counsel' s claim that the retail distribution of regular coffee is contained within sales districts-whether true or not48-ignores the fundamental point. While this says much about the geographic market for the sale of regular coffee by retailers, it is simply irrelevant where the issue is the relevant geographic market for the sale of regular coffee by producers. Furthermore, industry recognition that demand differs from region 16111 fact. this confinement is breaking down (Findings 99- 104). Initial Decision 103 F. to region and that deals must be tailored to these regions says nothing about which producers retailers and wholesalers can practicably turn to for their supplies: It merely indicates that producers who compete over a much broader area than a sales district often must take into account the different desires of customers located in different regions of the country.

Not all coffee producers compete in every city throughout the United States, and their market shares differ from city to city (Findings 117-18), but where there are no significant transportation costs, this does not mean that regional submarkets exist: There is considerable evidence that retail frozen pizza manufacturers often target, or (138) even confine their marketing regionally. Thus, in any particular city, only Pillsbury, Fox, and three or four other frozen pizza manufacturers might be selling at any given time. In such local markets, Pillsbury s and Fox s market shares of course would be high and perhaps suffcient to indicate anticompetitive effects under Section 7. But (w)e do not believe the pie will slice so thinly," at least not on this record. The test for meafmring geographic market is where consumers (in this case retailers) can practicably turn for an alternative source of supply. Pillsbury, 93 F. C. at 1030. See also Golden Grain 78 F. C. at 159-tO.

Complaint counsel also argue that Regular Maxwell House product differentiation" advantage creates barriers to entry into the sales districts which, by virtue ofthese barriers, are relevant markets (CLA, p. 101; CRLA, pp. 24-25).

Complaint counsel's argument is a theoretical construct which finds no support in the record. Hill Bros.' entry into the HFNI in the 1960' s and other roasters' expansion within the HFNI even after Folger s entry (Findings 322-24) establishes beyond question that Regular Maxwell House s "product differentiation" advantage if it exists, has not raised even modest entry barriers to other coffee roasters.

The price data relied upon by complaint counsel Can be interpreted in several ways and it can be rejected as an indicator of separate geographic markets for that reason alone (Finding 130, n. 11), but it suffers from an even more fundamental defect. A relevant geographic market contains all producers and purchasers whose interactions affect price, but complaint counsel presented evidence of price differences in the sale of a single brand of coffee, Regular Maxwell House. Thus, this price data does not reflect the supply and demand forces that operate with respect to the sale of regular coffee.49 (139) 49 Dr. Elzinga agreed that it is appropriate to analyze Reguar Maxwell House s market power rather than that of al 01" General Food ' regular coffee brands (Footnote 23), but that power can only be analyzed in relationship to the relevant geographic market in which it is sold, and that market contains all regular coffee producers operating within it1confines.

204 Initial Decision (c) Conclusion In conclusion, complaint counsel' s argument that the relevant geographic market for regular coffee is no larger than a sales district is based on irrelevant facts. The most significant fact in this record is that the relative transportation costs for regular coffee are so low that producers are able to ship their coffee anywhere in the United States. These producers compete or, if they so choose, can compete, throughout the nation, and the relevant geographic market for regular coffee is nationwide.

B. General Foods Did Not Attempt To Monopolize The HFNI Regular Coffee "Markets 1. Introduction Since complaint counsel do not claim that General Foods, through its Maxwell House Division, attempted to monopolize the relevant market-the nation-the Section 5 count could be dismissed without analyzing their claim that General Foods monopolized the sale of regular coffee in what they claim are relevant geographic markets the Division s market areas and sales districts. However, so much below costeffort has been devoted to the market power and sales issues that the parties are entitled to my views on the evidence offered by complaint counsel in support of their argument. 2. Elements Of An Attempt To Monopolize Section 5 of the Federal Trade Commission Act, 15 U. C. 45, empowers the Commission to prohibit unfair methods of (140) competition, and this power has been construed to reach conduct that violates the prohibitions of the Sherman Act. Times-Picayune Publishing Co. v. United States, 345 U.S. 594, 609 (1953). The elements of the offense of attempted monopolization were described in Swift Co. v. United States, 196 U.S. 375, 396 (1905): Where acts are not suffcient in themselves to produce a result which the law seeks to prevent-for instance, the monopoly-but require further acts in addition to the mere forces of nature to bring that result to pass, an intent to bring it to pass is necessary in order to produce a dangerous probability that it will happen. The three elements of this offense were recently restated by the Commission in E.I Dupont de Nemours Co. 3 CCH Trade Reg. Rep. 770 at 21 97G-71 (Oct. , 1980) (96 F. C. 653): . . . the attempt offense includes three principal elements: (1) specific intent to control prices or destroy competition, (2) exclusionary or anti competitive conduct, and (3) a dangerous probability of success. Id. at 21 970. Initial Decision 103 F. The Commission in Dupont recognized that while all three elements must be satisfied:

These criteria, however, are not mutually exclusive but rather are interrelated to the extent that evidence of conduct may shed light on intent and the probability of success; conversely, evidence of a respondent's purpose may reveal the extent to which there are legitimate business justifications underpinning the respondent' s conduct. ld. 971 (141) 3. Specific Intent To Control Prices Or Destroy Competition Read with an uncritical eye, some of the statements by General Foods' employees and agents might lead one to believe that they were looking forward to Folger s entry into the HFNI as an opportunity to increase Regular Maxwell House s market share and profits in that area.

The reality is different. The Maxwell House Division was aware at the time of Folger s entry into the HFNI that Procter & Gamble was a formidable competitor (Findings 218-27), and it did not look forward with satisfaction to the impending struggle for regular coffee share in the HFNI since there was no real hope that Folger s could be excluded from any area that it chose to enter (Findings 231-32). Despite this serious threat, the Maxwell House Division s initial response to Folger s Cleveland introduction was modest (Finding 149), and it was only after management became seriously concerned that the "defend now" strategy was adopted (Findings 154-66). In fact even as late as Folger s Syracuse entry, the Division did not match its trade rates (Findings 193-99).

While the "defend now" strategy projected losses in Year I in the HFNI as Folger s marketing expenditures were met by the Division it was, realistically, the only strategy which had any chance of success. The adoption of alternative strategies would certainly have resulted in significant share and profit losses (Findings 228-30). Thus, despite some suggestive statements in internal documents the weight of the evidence specially taking into account the nature of the competitor which the Maxwell House Division was facingestablishes that the Division s overall intent was to engage in an honest competitive struggle with the new entrant. Since the Maxwell House Division s actions were "predominantly motivated by legitimate business aims " I cannot find that it had a specific intent to monopolize the sale of regular coffee, whether the relevant geographic market is nationwide or circumscribed by the boundaries of a market area or a sales district. Times-Picayune Publishing Co. v. United States 345 U.S. 594 , 627 (1953); see also Lektro- . .

204 Initial Decision Vend Corp. v. Venda Co. 1981-2 CCH Trade Cases 11 64 258 at 74 089 (7th Cir. 1981). (142) The inevitable consequence of General Foods' decision to defend vigorously its regular coffee market share and profits was losses by the new entrant and a few of its smaller competitors (Findings 240- 73), but the competitive struggle inevitably involves winners and losers, and the intent to compete and succeed is not anticompetitive: More than an intent to win every sale, even if that would result in the demise of a competitor, is required before it can be concluded a deffmclant has the type of exclusionary intent condemned by the antitrust laws. Transamerica Computer Co. v. IBM, 481 Supp. 965, 1010 (N.D. Cal. 1979).

See also William Inglis Sons Baking Co. v. ITT Continental Baking Co. 652 F.2d 917, 932 (9th Cir. 1981): "Direct evidence of intent to vanquish a rival in an honest competitive struggle cannot help to establish an antitrust violation International Air Industries, Inc. , 723 (5th Cir. 1975), cert.American Excelsior Co. 517 F.2d 714, 719 denied, 424 U. S. 943 (1976); II Antitrust Law 822a, at 314 (1978). Complaint counsel answer, however, that the Maxwell House Division s actions were not honestly competitive, for the Division sold below average total cost in the Cleveland and Pittsburgh market areas and the Syracuse sales district, introduced a "fighting brand" into Syracuse, retaliated against Folger s in the West and copied its Mrs. Olsen" campaign (CLA , p. 33).

In Transamerica the court stated: A firm that prices its product at levels above marginal or average variable cost is not necessarily engaged in clearly exclusionary conduct. A firm that prices its products below those levels is. 481 F.Supp. at 989.

Complaint counsel rely on this case and statements by Areeda & Turner for the proposition that regardless of actual intent, pricing below reasonably anticipated average variable cost should be characterized as a per se attempt to monopolize because such pricing: (143) (1) list totally unrelated to competition on the merits. . (2) clearly implies the presence or prospect of some degree of durable market power ; (anda (3) has potentially significant exclusionary em cts in the generality of cases. . . III Antilrust Law n 820, at 313.

Furthermore, Areeda & Turner:

. . . would not permit a monopolist to price below marginal cost in order to meet the lawful price of a rival. 111 Antitrusl Law n 717 , at 178. Initial Decision 103 F. On the other hand, the Commission held in Golden Grain 78 F. at 165 that, with respect to the pricing of a firm which one could characterize as dominant in the relevant market: even if we should find that respondent' s sales were below cost, if we find further that they acted defensively in reaching such a pricing level, their conduct wil not be illegal because ofthe absence ufthe requisite predation. See also United Fruit Co. 82 F.T.C. 53, 162 n. 43 (1973), Although predatory intent may be inferred from sales below cost, this inference does not arise when there is evidence that the below-cost level was reached defensively. The Maxwell House Division did sell below cost in the Cleveland and Pittsburgh market areas and the Syracuse sales district (Finding 423), but I cannot infer from this fact alone a specific intent to monopolize since it ignores the reason for the Division s actions. (144) The Division sold below average variable cost as a direct result of the significant threat posed by its much larger rival, Procter & Gamble. The Division could have, as complaint counsel suggest, responded less vigorously to Folger s entry, but this would have been disastrous. The Division, in my opinion, chose the only possible respons&-headto-head competition. And, although General Foods was unable to satisfy me that it could establish a meeting competition defense to the Robinson-Patman charge, I am convinced, once the more rigorous requirements of Section 2(b) are set aside, that while there may have been instances ofleading Folger s in promotions, the Division s general intent was to meet, and not beat, Folger s prices and promotions (Findings 476-86).

Futhermore, Procter & Gamble s proven track record-in addition to providing the impetus for the Division s vigorous respons&-and the absence of market power establishes that there was no likelihood that the Division could use its Regular Maxwell House brand to successfully monopolize a market area or sales district (infra). Thus, this is not "the generality of cases" which Areeda & Turner refer to in their per se rule, and I wil not infer specific intent to monopolize from the Division s defensive, below cost pricing for if I were to do so, it would amount to a holding that a large firm (as opposed to a monopoable to com- list or one with substantial market power) would not be 5o See United Fruit, supra, 82pete with equally large or larger firms. C. at 163:

Certainly the logic of complaint counsel's argument would require companies con- 5( Complaint counsel argue that under the rationale Borden,of Inc. 92 F. C. 669, 798-02 (1978), the Maxwell House Division s prices cannot be defended on the ground t.hatit was meeting competition (CLA, p. 62), but Borden involved a monupolist. Regular MaxweU HOllsc s shares in its sales districts never approached monopoly proportions (F'finding 377).

204 Initial Decision cerned with competitive threats in the marketplace to sit back and wait until they become a reality before taking action. This is surely not the aim and purpose of our competitive marketplace nor the role of competition which this Commission was created to promote. (145) 4. Exclusionary Or Anticompetitive Conduct In Dupont the Commission held that the conduct element of the offense of attempted monopolization is satisfied by proof that "unreasonable" means were used to control prices and restrict competition. 3 Trade Reg. Rep. at 21 972. Other cases suggest that conduct which is "unfair or unreasonable Northeastern Telephone Co. v; American Telephone Telegraph Co. 651 F.2d 76, 85 (2d Cir. 1981) or which is without legitimate business purpose Janich Bros., Inc. v. American Distilling Co. 570 F.2d 848 , 853 (9th Cir. 1977), cert. denied 439 U.s. 829 (1978), is exclusionary or anticompetitive. Even monopolists are allowed some freedom to respond to competitive threats. In SuperTurf, Inc. v. Monsanto Co. No. 80-1484 (8th Cir. Oct. 2, 1981), Monsanto, a monopolist in the artificial turf market, was alleged to have priced below cos ,n an effort unlawfully to eliminate competition. Evidence indicated that Monsanto engaged only in competitive defensive pricing and that SuperTurf, Inc. had initiated the price warfare. The Eighth Circuit held that: Even if Monsanto is a monopolist, it was within its rights to respond to the lower prices of its competitors while still pricing above its marginal costs. Slip Op. at 14. Of course, Regular Maxwell House has not had nor does it have monopoly or market power in the relevant market-the nation-or in any of the sales districts or market areas. It has never even approached such power in these areas, contrary to complaint counsel's claim (CLA, pp. 109-10).

Because it has never had monopoly power or even approached it in the sales districts or the nation, the sale of Regular Maxwell House below average variable cost in two market areas and one sales district (again, assuming that they are relevant geographic markets) was not anti competitive or exclusionary; it was, instead, legitimately defensive behavior. Golden Grain, supra; United Fruit, supra. While the use of the Horizon brand in Syracuse had no purpose other than to respond to Folger s entry (Findings 425-27), and in that sense was not "legitimate, I do not believe (146) that the use of a fighting brand by a company lacking monopoly power in one sales district against a larger rival was inherently anti competitive. Compare United Statesv. American Tobacco Co. 221 U.S. 106 (1911). The Initial Decision 103 F. Division s western retaliation (Findings 428-30) and its "Cora" campaign (Findings 431-35) were legitimate defensive acts. My conclusion with respect to Horizon and the western retaliation is not altered by the claims that Horizon was sold below cost (CPF 4-35 to 4-37) and that Regular Maxwell House s trade rates equalled contribution margin in Kansas City and Dallas (CPF 5-20 to 5-26), for Horizon was a new entrant into Syracuse, and Regular Maxwell House s western retaliation was similar to Folger s expensive Cleveland entry.

5. Dangerous Probability Of Success Whether a dangerous probability of success exists requires a definition of the relevant geographic market in which the challenged conduct took place, and an assessment ofthe accused's power to exclude competition or control prices within the market. Richter Concrete Corp. v. Hilltop Basic Resources, Inc. 1981-1 CCH Trade Cases n 947 at 75 891 (S. D.. Ohio 1981); General Communications Engineering, Inc. v. Motorola Communications Electronics, Inc. 421 F.Supp. 274 292 n. 42 (N.D. Cal. 1976).

The cases dealing with this issue generally look to a company market share to determine whether it has the ability to succeed in attaining monopoly power in the relevant market. See, e. , Hiland Dairy, Inc. v. Kroger Co. 402 F.2d 968 , 974 (8th Cir. 1968), cert. denied 395 U. S. 961 (1969). See also, II Antitrust Law n 831, at 336: Perhaps as an aid to predicting dangerous future probability, the cases also require that the defendant possess a measure of present proximity to completed monopoly-a measure of power in a relevant market. As with monopoly, the requisite market position is normally measured through an analysis of market share. Although precise definition uftbe requisite position is no easier here than in the case of monopoly itself it is clear that the basic thrust afthe classic rule is the presumption that attempt does not occur in the absence of(147J a rather significant market share. The defendant need not have present market power, but its position must be suffciently "proximate " to monopoly that the challenged conduct threatens success. Moreover, the presumption resting on market share may be overcome on a showing that the market is so narrowly defined, or entry so free, that "success " would not yield the substantial market power needed for the offense of completed monopolization (emphasis in original). No particular market share can be said to evidence the market power needed to succeed in an attempt to monopolize: Complaint counsel suggest that a 40% absolute market share is enough, citing, for example Broadway Delivery Corp. v. United Parcel Service, 1981- CCH Trade Cases n 64 068 at 74,469 (2d Cir. 1981). Since Regular Maxwell House s share exceeded this figure in some HFNI sales districts, they argue that there was a dangerous probability that the Maxwell House Division would acquire monopoly power (CLA, p. 113). 204 Initial Decision General Foods naturally disagrees, citing Hoffman v. Delta Dental Plan 1981-1 CCH Trade Cases IT 64 138 at 76 838 (D. Minn. 1981), which held that a 40% share of market was insuffcient as a matter of law to permit the inference of dangerous probability of success. Given the divergence of opinion on this issue, I cannot find that Regular Maxwell House s market share in the nation, the HFNI or in any of the HFNI sales districts was so large that, as a matter oflaw it possessed the market power which would make it dangerously probable that the Division could succeed in monopolizing the sale of regular coftee in those areas (Finding 377).

Other facts of record beside market share convince me that the Division could never have succeeded in acquiring monopoly power in the HFNI sales districts-that is, the power to control prices or exclude competition United Statesv. Grinnell Corp. 384 U.S. 563, 570- 77 (1966); United Statesv. E.I Dupont de Nemours Co. 351 U.S. 377 391 (1956).

Barriers to entry into the HFNI sales districts are minimal (Findings 321-24), and if the Division attempted to raise prices to the monopoly level, new entrants would be attracted into the districts (Finding 440). (148) Futhermore, because of a decline in demand, the regular coffee industry had substantial excess capacity (Findings 20-24), and the Maxwell House Division could not, therefore, restrict its output and raise prices in the sales districts, for other roasters would simply increase their output. See II Antitrust Law IT 501 , at 322; Transamerica Computer Co. v. IBM, 481 F.Supp. 965, 974-75 (N.D. Cal. 1979). Finally, Folger s greater staying power must be considered. Considering all of these facts, I find that the Division s sales below average variable cost in the Cleveland and Pittsburgh market areas and the Syracuse sales district were not predatory. III Antitrust Law IT 711b, at 151-52:

. . . predation in any meaningful sense cannot exist unless there is a temporary sacrifice of net revenues in the expectation of greater future gains. Indeed, the classifically feared ca. e of predation has been the deliberate sacrifice of present revenues for the purpose of driving rivals out of the market and then recouping the losses through higher profits earned in the absence of competition. Thus, predatory pricing would make Jittle economic sense to a potential predator unless he had (1) greater financial staying power than his rivals and (2) a very substantial prospect that the losses he incurs in the predatory campaign wil be exceeded by the profits to he earned after his rivals have been destroyed.

As for the first prerequisite, it should, of course, be recognized that predation cannot be successful, and therefore is unlikely to occur, when the predator s rivals possess resources comparable to his own. Even when an alleged predator has greater staying power, however, attention must also be given to the second prerequisite, which is less likely to occur. Although a predator may drive competitors into bankruptcy, their , Initial Decision 103 F. durable assets may remain in the market in the hands of others. Moreover, a firm can anticipate monopoly profits for only so long as its monopoly prices do not attract new entry. (149J Losses incurred through predation could be regained in markets with very high harriers to entry. In many markets, however, and especially in those having a number of small rivals, entry barriers may be non-existent or at least too low to preclude entry. Admittedly, a demonstrated willingness to indulge in predatory pricing might itself deter some smaller potential entrants, but it is unlikely to inhibit firms with resources comparable to those of the predator. Complaint counsel argue, nevertheless, that the Division s ability to achieve high profits for Regular Maxwell House despite the decline in demand reveals that brand' s market power (CRF's 4-27 and 4-28). The problem with this argument is that I cannot accept Professor Dearden s computations and their comparison with Dr. Mann benchmark as evidence that Regular Maxwell House s profits were higher than normal. Furthermore, even assuming that Regular Maxwell House s profits were high, there might well be reasons for this other than market power. See Transamerica, supra at 981: . . . the inference that a defendant that enjoys healthy profits only does so because of an unhealthy market structure is not a strong one. Good management, superior effciency and differences in accounting provide explanations that are just as plausible and none of those explanations is inconsistent with an effective competitive market. In Boise Cascade Corp. 91 F. C. 1 (1978), rev d on other grounds, 637 F.2d 573 (9th Cir. 1980), the Commission stated that: "The persistence of supra-normal profits over some period of time may signal the existence of anti competitive conduct. .." but Recognizing the uncertainties that are likely to attend the use of profit data as evidence of anticompetitive .conduct, and the additional uncertainties as to the particular data introduced in this case" placed little reliance on the data. Id. at 97. Similar uncertainties exist in this case (Finding 376). Thus, it is simply inconceivable, given Procter & Gamble s marketing expertise and its financial resources and Regular (150) Maxwell House s lack of market power, that the Division could succeed in monopolizing the sale of regular coffee in the HFNI sales districts. In fact, Folger is now the number one regular coffee in the United States and has succeeded in reducing Regular Maxwell House s share in the HFNI sales districts (Findings 276-79), and other companies have been able to survive and prosper in the HFNI. Nifty Foods Corp. Great Atlantic Pacific Tea Co. 614 F.2d 832, 841 (2d Cir. 1980); Lektro- Vend Corp. v. Vendo Corp. 500 F.Supp. 332, 356 (N.D. Ill. 1980); aff'd 1981-2 Trade Cas. (CCH) n 64 258 (7th Cir. 1981); Dia- \1 Complaint counsel !\tate that the !\UCCCf\S of one company, Chock Full O'Nuts, occurred after tlh alleged predation in Syracuse (CRLA, p. 11, n. 9), but there is no reason why the succes. or failure ofaJlegedly predatory taclics must be mca ured solely during the time uch pract.ice upposedJy occurred. See &ktro- Vend, supra 087.

204 Initial Dccision v. Walterhoefer 289 F.Supp. 550, 578 (D.mond International Corp. Md. 1968). See also, Berkey Photo, Inc. v. Eastman Kodak Co. 603 F. 263 273 n. 11 (2d Cir. 1979), cert. denied 444 U.S. 1093 (1980) (noting that jury could consider significance of decline in market share). 6. Conclusion Assuming that the Maxwell House Division s HFNI sales districts are relevant geographic markets, its sales of Regular Maxwell House below average variable cost in the Cleveland and Pittsburgh market areas and the Syracuse sales district, and its use ofthe Horizon brand in that sales district do not reveal specific intent to control prices or destroy competition, nor was its conduct exclusionary or anticompetitive. In view of Regular Maxwell House s lack ofthe necessary market power, the Division s conduct was, instead, legitimately defensive. Furthermore, the lack of market power, the ease of entry into the sales districts, the declining demand for regular coffee, and the nature of its chief rival for market share in the HFNI sales districts establishes that there was no dangerous probability that the Division could succeed in monopolizing the sale of regular coffee in those districts. In conclusion, General Foods did not, through its Maxwell House Division, attempt to monopolize the sale of regular coffee in the nation, the HFNI or in the HFNI sales districts, nor were its acts " inherently anticompetitive that they constitute unfair methods of competition" (CLA, p. 142), for they are lawful under Sherman (151) Act standards, they do not amount to an incipient violation of the antitrust laws FTC v. Motion Pictures Advertising Service Co., 344 U.s. 392, 394-95 (1953), and they do not violate the "spirit" of any antitrust law or important public policy. FTCv. Sperry Hutchinson 405 U.S. 233, 244 (1972). Therefore, the Section 5 count of the complaint will be dismissed.

C. General Foods Did Not Violate The Robinson-Patman Act 1. RECU's Are Not Discounts From List Price According to complaint counsel, the Robinson-Patman Act price must include a subtraction for all offers made available to retailers v. General Motors Corp. 337(CLA , p. 152). Kapiolani Motors, Ltd. Supp. 102 (D. Hawaii 1972), one ofthe cases they cite, defines Robinson-Patman price as "the actual amount paid to the supplier for goods furnished. Id. at 104.

This case does not however, support the claim that RECU's are an element of price, for it is clear that RECU's do not allect the price paid by a retailer to the Maxwell House Division, and the Division does not make RECU' s "available to retailers." They are, rather, made avail- Initial Decision 103 F. able to the consumer through retailers, for they differ in no significant way from consumer coupons which complaint counsel agree are not an element of Robinson-Patman price (Findings 447-49). Thus, I agree with General Foods that the recent decision in Indian Coffee Corp. v. Procter Gamble Co. 482 F.Supp. 1104 (W.D. Pa. 1980) is persuasive here, for even though it dealt with consumer coupons rather than with RECU' , no "principled basis exists" for distinguishing between them (RLA , p. 154).

In Indian Coffee a local coffee producer charged Procter & Gamble with unlawful price discrimination arising from Folger s entry into the HFNI. The court agreed with Procter & Gamble, which argued that consumer coupons were not reductions in price under Section 2(a):

flat is clear to us that Folger s consumer coupons reduced the price solely to the ultimate consumer and not to Folger s customer, who under 2(a) are the retailers, since those (152) retailers received absolutely no price concession and served merely as redemption agents for Folger (emphasis in original). Id. at 1109. The Commission s decision in Fred Meyer, Inc. 63 F. C. 1 (1963), rev in part and enforced 359 F.2d 351 (9th Cir. 1966), reu d on other grounds 390 U.s. 341 (1968), which the court in Indian Coffee refused to follow, is, nevertheless, cited by complaint counsel as support for their claim that retailer coupons are price reductions. The decision does so hold, but the facts which led to the Commission s decision are different from those of the present case.

In Fred Meyer respondent, a retailer, had been conducting since the mid 1930's annual coupon book promotion under which it published and sold books to consumers which contained some 72 coupons featuring an article sold by Fred Meyer. Consumers could redeem those coupons and purchase the featured items at a reduced price. Fred Meyer s suppliers participated in its program by selling to it at specially reduced prices a quantity of merchandise necessary to cover expected redemption, or by giving free merchandise, and/or by paying respondent a fee of $350. " 63 F. C. at 12. The Commission held that these redemption payments were price reductions: The only possible "service" that respondents performed in return for the $4 814 is that they resold the goods at the same 1/3 price reduction that they had received from the supplier. (They actually agreed to do so.) But this "passing on" of the discriminatory lower price is the very worst of the vices involved in price discrimination. . . . Thus, it would be a strange result indeed if we were to hold that a buyer, by passing on to his own customers a price discrimination he has received from his supplier, has merely performed a "promotional service " for that supplier! 63 F. C. at 33. An administrative law judge s decision which was not appealed to 204 Initial Decision the Commission also found that a program similar to General Foods involved a price reduction. Purex Corp. 51 (153J F. C. 100, 103 (1954). See also National Tea Co. 46 F. C. 829, 833-34 (1950), order modified 47 F. C. 1314 (1951) Despite the language cited by complaint counsel, there are important differences between Fred Meyer and this case. Fred Meyer induced payments from its suppliers and set up the ground rules under which they were made, and it is apparent that the Commission viewed the whole arrangement as a sham designed by that company to induce price concessions:

Where money or something of value is given by a seller to a buyer without even the contemplation of promotional services by the purchaser there has been no payment ' compensation or in consideration' for such services and Section 2(d) is therefore not applicable. 63 F.T.C. at 33.

The Maxwell House Division s RECU's had a different inception and a different purpose. There is no evidence that any customer induced the Division to develop its RECU program. The Division rather than its customers, sets the conditions under which it will redeem its RECUs, and the purpose ofthe Division s coupons, whether they are RECUs or consumer coupons is identical-to promote its coffees, not to give retailers a hidden price concession. One final fact convinces me that RECU's are not price concessions, and that is their reception by retailers. One would be surprised if Fred Meyer, after arranging its coupon program with suppliers, failed to take full advantage of it, for its purpose was to benefit Fred Meyer. On the other hand, retailers do not always participate in the Division s RECU programs, a convincing demonstration that they do not view them as price reductions (Finding 456). I therefore find that the Maxwell House Division s RECU' s are not deductions from the price paid by retailers to the Division and are therefore not cognizable under Section 2(a).52 (154J 2. Performance Allowances Are Not Discounts From List Price Complaint counsel argue that the Maxwell House Division s performance allowances are price reductions because the performance required is much less than the benefit received by the retailer. The Fred Meyer Guides state:

Allowances that have little or no relationship to cost or approximate cost ufthe service provided by the retailer may be considered to be in violation of Section 2(d) or subject '" TIlis finding is not inconsistent mywithdecision in the Section 5 part of the e findings that RECD's affect the "economic price " for that price is the revenue received by the Division, and the Division s revenue is reduced to the extent that RECD's or consumer coupons are redeemed by retailers , p.

Initial Decision 103 F. to the prohibitions of Section 2(a) of the amended Clayton Act, such as an allowance of $1 per case of goods purchased if the retailer furnishes a display or provides shelf space. . 16 C. R. 240.9 n. 2 (19S1).

There is no dispute that the Maxwell House Division requires its customers to perform ifthey receive promotional allowances (CLA 156), and the case which gave rise to the Guides emphasized that because this element was lacking, the allowances in question were price reductions. Fred Meyer, supra, 63 F. C. at 33. Complaint counsel argue nevertheless that the Division s promotional allowances should be treated here as offering discounts from its list prices because the Division s customers retain much more of the allowance than they spend in performance.

While General Foods argues that complaint counsel have not met the burden imposed on them by the Fred Meyer Guides (RLA, p. 160), I am willing to concede that many retailers receive a substantial benefit from the monies which are left after their cost of performance is deducted from the allowances they receive. I also concede that some authorities have chosen to view allowances similar to the Division as price reductions. See Continental Baking Co. v. Utah Pie Co., 349 2d 122 , 130 (155) (10th Cir. 1965), rev d on other grounds 386 U.s. 685 (1967); Borden, Inc. 92 F. C. 669, 797 (1978). Despite the cases cited above, I believe that the Division s promotional allowances should not be treated as price concessions, but as Section 2(d) payments. I agree that promotional allowances which are promotional" in name only should be treated as price concessions. One must not forget that the Commission in Fred Meyer held that Section 2(d) was not applicable because the payments were made without even the contemplation of promotional services by the purchaser" 63 F. C. at 33, and it is with this in mind that one must read the subsequent Fred Meyer Guides. Thus, Rule 9, footnote 2 of the Guides was designed to eliminate from serious consideration as Section 2(d) allowances those programs which required trivial performance-i. , programs which are not truly promotional. The Division s advertising allowances were designed to promote the resale of its products (see Rickles and Kirby, 54infra), not to give hidden price concessions to its customers. 6J On lhe other hand, since Borden was not a Robinson-Patman case, its precedenJial effect is questionable 54 Compare the Maxwell House Division s "buying allowances" where the only performance required is the purchase uf coffee. This is clearly a price concession under Fredthe Meyer rationale. . .

204 Initial Decision 3. Treating RECU's And Performance Offers As Discounts From List Price Would Subject Manufacturers To Unwarranted Liability For Price Discrimination Assuming that the Maxwell House Division offers its RECU's and performance allowances to all competing customers on proportionally equal terms, Fred MeyerGuides 16 C. R. 240. 7 (1981), treating them as discounts from list price could result in liability for secondary line injury, for customers who did not choose to participate in these offers would be paying a higher "price" than those who did (RLA, p. 164). This suggests, contrary to complaint counsel's argument, that a true advertising allowance should not be subject to scrutiny under Section 2(a). See Rickles, Inc. v. Frances Denney Corp. 508 F.Supp. 4 CD. Mass. 1980), in which the court dismissed allegations that promotional payments and services were Section (156) 2(a) price reductions because the payments were provided in connection with projected resale ofthe product: "The alleged discriminations in this case properly are subsumed exclusively under Sections 2(d) and 2(e), rather than Section 2(a). Id. at 6.

In Kirbyv. PR. Mallory Co., Inc. 489 F. 2d 904 (7th Cir. 1973), cert. denied 417 U. S. 911 (1974), the court rejected the position now being advocated by complaint counsel, stating that the "theory that Sections 2(d) and 2(e) proscribe acts which are themselves prohibited by Section 2(a) is not supported by either the legislative history or scheme of the (Clayton) Act. Id. at 910. In holding that promotional allowances covered by Section 2(d) were not elements of price for purposes of Section 2(a) price discrimination, the court focused on the different statutory standards governing price discrimination and promotional discrimination, and the fact that the payments and services under consideration were provided in connection with " projected resales Congress. imposed stricter standards oflegality respecting promotional discriminations than price discriminations. Price discrimination is lawful if it can be justified under several exculpatory provisos or has no effect on competition. In contrast, promotional discrimination is ilegal per se, irrespective of competitive impact and without resort to statutory justification. Id. at 910. The court went on:

(plaintiffs) argument would have us collapse the distinction in schemes and standards and would have us find that the two sections are mere surplusage. This we decline to do. In view of the strict standards of *9 2(d) and 2(e), which focus on resale, it appears quite clear that Congress carefully considered the deficiency in the original law proscribing price discrimination in the supplier-customer sale and drafted 2(d) and 2(e) Initial Dccision 103 F. to apply exclusively to promotional discriminations like (157J those alleged in this case. Id. at 910-11.

Complaint counsel argue that neither Rickles nor Kirby are controlling here since they are secondary line cases (CRLA, p. 36, n. 33), but I know of no authority for the proposition that, or any logical reason why, "price" should be defined differently in primary and secondary line cases. The court in Indian Coffee agrees. 482 F.Supp. at 1109, n. 11.

Finally, treating reductions in the producer s revenue (such as are caused by RECU redemptions) as price discounts is not dissimilar to the old "mill net" theory. 55 While I have accepted "mill net", or revenue received by the seller, as a price which can be used to analyze the Section 5 charge, this definition of price is no longer acceptable in Robinson-Patman cases. See National Lead Co. 49 F. C. 791 (1953):

. . . the amended complaint, in Count II, alleges that each ofthe respondents discriminates in the price of its lead pigments as between different purchasers, in violation of section 2(a) of the amended Clayton Act. Certain of the discriminations, it is alleged occur as a result of the use of the zone method of pricing and sellng to com peting customersin the same zone. As stated in the conclusion appended to the findings, the Commission is ofthe opinion that on this phase of the case the complaint fails to state a cause of action. This is because the allegations are that each of the respondents sells its products in accordance with a delivered pricing- system, but the alleged discriminations (158J occur as a result of differing net prices received by each of the respondents at its factory. Thus, the complaint does not show that the alleged unlawful discriminations as between purchasers located in the same zone occur as the result of differences in actual prices at which the respondents' products are sold. Id. at 881-82. More recently, in Kapiolani it was argued that Section 2(a) price should be defined as "the amount the seller actually and ultimately receives from the sale net receipts rather than money initially paid out by the buyer for the product." 337 F.Supp. at 104. The court rejected this argument, holding that there is no authority to support a "net receipts" approach to Section 2(a). According to the court, price under Section 2(a) means "the actual amount paid to the supplier for the goods furnished." 337 F.Supp. at 104 citing Corn Products Refining Co. v. FTC, 324 U.s. 726 (1945). And in Robbins Flooring, Inc. Federal Floors, Inc. 445 F.Supp. 4 (E.D. pa. 1977), the court defined price for purposes of Section 2(a) as "the amount actually paid or laid out for goods by the buyer. Id. at 8. Since neither RECU's nor performance allowances are elements of Section 2(a) price, complaint counsel have not proved that General Foods has, through its Maxwell House Division, discriminated in \' Complaint coungp.l argue that they are vot espou!iing a return to the "mill net" theory since they do not deduct tran portation cOSUJ (CLA p. 161 , n. 570), but their proposed findings on the Section 5 price empha jze that RECl)'s reduce t.he Division s revenue: "General Foods sponsors the coupons, absorbing the financial loss" (CPF 3-2U6). General Fouds, the scher, thus experiences a reduction in the price it receives" (CPF 3--226). 204 Initial Decision price in the sale of its regular coffee and the Section 2(a) count wil be dismissed.

4. Assuming That General Foods Discriminated In Price, It Has Not Established That It Was Meeting Folger s Competition In Good Faith If complaint counsel's price definition is accepted, the Maxwell House Division s response to Folger s entry injured some of its smaller competitors in Cleveland, Pittsburgh and Syracuse (Findings 255-73), and I do not believe that General Foods could successfully meet its burden of proving that the prices it charged in those cities were offered in good faith to meet the equally low prices of Folger The Division s overall approach to Folger s entry was not to aggress and I believe that it intended not to beat Folger s competition. Howev- , General Foods has a greater burden than (159) proving intent; it must satisfy me with reasonable accuracy that its intentions were carried out, and Mr. MacDonald's reconstruction does not provide that kind of proof, for he did not have personal knowledge of each transaction he charted, and much of his reconstruction was based upon unproven assumptions, such as the times of deal announcev. FTC, 324ments (Findings 500-05). See Corn Products Refining Co. S. 726 (1945):

The only evidence said to rebut the prima facie case made by proof of the price discriminations was given by witnesses who had no personal knowledge of the transactions, and was limited to statements of each witness' assumption or conclusion that the price discriminations were justified by competition. Id. at 741. IV. SUMMARY A. General Foods, through its Maxwell House Division, has sold and is selling regular coffee in interstate commerce and the Commission has jurisdiction over the acts and practices of General Foods which are challenged in the complaint.

B. The relevant market for the production and sale of regular coffee by roasters to wholesalers and retailers is nationwide. C. General Foods has not, through its Maxwell House Division attempted to monopolize the production and sale of regular coffee in the relevant market, the HFNI, or in any of its market areas or sales districts, nor has it engaged in any other unfair methods of competition in those areas.

D. The Maxwell House Division s RECUs and performance allowances are not elements of Robinson-Patman price and General Foods has not, through the Division, sold regular coHee to purchasers at different cognizable prices. (160) Opinion 103 F.

V. ORDER It is ordered That the complaint be, and it hereby is, dismissed. OPINION OF THE COMMISSION By MILLER, Chairman:

I. Introduction General Foods Corporation ("General Foods ) was charged with violations of Section 5 of the Federal Trade Commission Act' and Section 2(a) ofthe Robinson-Patman Act in connection with the marketing of Regular Maxwell House Coffee. The complaint alleged that General Foods-by sellng its regular ground coffee below cost or at unreasonably low prices, by engaging in excessive promotional activities, by using a "fighting brand" to impede entry, and by discriminating geographically in prices and promotions-monopolized or attempted to monopolize the sale of regular ground coffee in a national market and in several regional submarkets. After almost two years of hearings, Administrative Law Judge Lewis F. Parker dismissed the complaint. He concluded that complaint counsel had failed to establish the existence of any submarkets smaller than the United States as a whole. On the national level, and even in the proposed submarkets, he determined that General Foods market share was too small to admit of a reasonable probability of successful monopolization. He found nothing inherently anticompetitive about a company lacking market power pricing below average variable cost or using a so-called fighting brand in response to competition. The AW also rejected the Robinson-Patman Act price discrimiand performancenation (2) charge, finding that the coupons allowances, which formed its basis, were not elements of price. Complaint counsel appeal the initial decision on essentially three grounds. First, they urge that the AW erred in concluding that the evidence on interregional price variations failed to establish the existence of relevant regional submarkets in which respondent held market power. Second, they contend that sustained sales below average variable cost and selected use of a "fighting brand" to deter local entry and preserve profitability constitute at least unfair methods of competition, if not attempted monopolization. Finally, they argue that certain coupons and performance offers should be calculated as price differences under Section 2(a) of the Robinson-Patman Act. 115UB. 45.

215 v. c. 13a.

204 Opinion For the reasons discussed below, we reject these arguments and affrm the Initial Decision.

II. The Regular Ground Coffee Industry A. The Product The parties stipulated the relevant product market to be all regular ground coffee packaged for sale at retail, including caffeinated, decaffeinated and extended ground coffee products. This definition excludes instant coffees and institutional sales (those to offces and hospitals, for example), as well as tea and other possible coffee substitutes. (3) B. The Major Sources of Supply 1. General Foods - Maxwell House General Foods is one of the leading manufacturers of processed foods, with over 30 major brand names in distribution.3 (IDF 1. For many years General Foods' Maxwell House Division has been the leading seller of regular ground coffee in the United States, marketing it under the Regular Maxwell House (RMH), Y uban, Sanka, Brim Max-Pax and Mellow Roast labels. In the most recent twelve-month reporting period in the record, fiscal 1980, the Maxwell House Division accounted for almost 32 percent of national ground coffee sales. Its Regular Maxwell House brand accounted for 22 percent of national sales, drawing most of its strength from the eastern United States. (RX 1279-80).

2. Procter & Gamble - Folger The Procter & Gamble Company became the nation s second largest seller of regular ground coffee with its purchase of The Folger Coffee Company in 1963. The Folger s brand was then the best sellng coffee in the West, but had limited sales east of states bordering the Mississippi River. (RX 1106.) By fiscal 1980, Folger s had expanded throughout the East and had surpassed RMH to become the leading national brand with 26 percent of (4) sales. (RX 1279-80.) Procter & Gamble also sells a decaffeinated coffee under the High Point label. (IDF 10. J The foJJowiQg abbreviations arc used in thig opinion Initial Decision Page Numher IDF Initial Deci j()n Finding Number 'f. Transcript of Testimony Page Numb Comphlint Counsp-I's Exhibit Nwnber - Respondents' Exhibit Nurnher CAB - Complaint Cuunse1's Appeal Brief l'a/i€ Number 334 FEDF;RAL TRADE COMMISSION DECISIONS Opinion 103 F.

3. Copersucar - Hils Bros.

The Copersucar Company is a Brazilian conglomerate that owns the Hils Bros. Company. (IDF 10.) At the close of the record, Hils Bros. coffee was sold in all but the southeastern portions ofthe nation and accounted for approximately 7 percent of national ground coffee sales. (RX 1117.

4. Miscellaneous Suppliers The remaining roughly 35 percent of national sales went to regional roasters' brands and the stores private label" brands. (RX 1107.) Among the major regional brands in the East are Savarin, Chock Full Nuts (CFON), and Chase & Sanborn (owned by Nabisco Brands formerly Standard Brands). (IDF 10 and 12; CX 1072.) The private labels, also called "chains-own-brands" (COB) or "controlled group, include Kroger, Safeway and A&P, among others, and have been more successful in the eastern United States than elsewhere. (Compare Tables 1 and 2 below.) Table 1 presents a year-by-year tabulation of national market shares by brand since 1971. (5) 19B1 31. (21.9)25. NA!6J , 19BO (22.26. -"cr 1979 31. (22.26.

14.

.

197B 33. (2421.

16.

1977 32. .. (24.21. .. CoffeeYears Leading 17. FIscal 1976 33.4 22. 13, 1971 Ground 19B1' (23. ,h, the to Table of States: M '" 1971 1975 35. Shares Regular (25.22. 1.1 1970 of United 11. Market the April Brands In 1974 34. &'m (2422. run. ,, .

g. 1973 33. (24.21.3 """' Y"' ""'nd year 1972 31. (24.20. 12.4 pre",,", oftbefiscl 1971 29. (2419. 10. ,'th, month 14. April seven 13).'n 11- lirat CPFb,,,,, for Bros. FUll Brands (Sey' are ... (RegularMaxellHOuse) TotalGeneralFoods "",.,, ForgerHils ChaseSanborn Chock A Data NutsSaVarinStore Group. +13.DaccOrdI18toRX1107A12.5accordingtoRXl107A 0 Opinion 103 F.

C. Competition in the Regular Ground Coffee Industry, 1971-1980 The events giving rise to this case began with Procter & Gamble campaign to bring Folger s into the Maxwell House eastern sales area-that portion ofthe nation roughly east of a line extending from the Ilinois/Indiana border on the north to the Mississippi/Alabama border on the south. (IDF 34; CX 1077 A.) Procter & Gamble chose Cleveland for its first test in October 1971. Cleveland, where RMH held nearly 45 percent of sales, was one of Maxwell House s most successful markets. (IDF 150, 235.) Procter & Gamble s goal was a sustainable share of20 percent by the end ofthe year. (CX 493B & C. Folger s employed a marketing mix of media advertising, incentives for retailers' promotions, consumer coupons, assorted in-pack gifts and home-delivered free samples to introduce its coffee to Cleveland consumers. (Zurcher, Tr. 6226-7; CX 491-499, 528A.) Maxwell House responded with 50 cent mailed coupons to combat Folger s 35 cent coupon, and in-pack coupons beating Folger s inserts by 4 cents to 10 cents per can. Maxwell House (7) also increased its promotional incentives for retailers, but stopped short of giving away samples. (See IDF 175.

In its first year Folger s garnered 15 percent of Cleveland sales while RMH grew by five share points to 50 percent. (RX 1111-12.) By early 1973 Procter & Gamble detected that General Foods' defensive efforts in Cleveland had subsided, but found that aggressive discounting by Hills Bros. was limiting further gains. Once again, Procter & Gamble increased its consumer coupon activities, began to "compete directly with Hils for (grocers ) feature support " and, as a result increased its Cleveland share to 18 percent for the last six months of 1973, displacing Hils as the second leading brand. (CX 528B. In the Spring of 1972, executives in the Maxwell House Division concluded that Folger s was "disturbingly successful" in Cleveland and that similar successes could easily recur throughout the East. (IDF 149-155; CX 122.) The Maxwell House division sought authority from General Foods to mount a strong defense to Folger s eastern expansion. According to Maxwell House a so-called "defend now strategy-immediate discounts, coupons and promotional allowances in each test area ould limit Folger s share in the East to 10 percent while increasing RMH's market share from 39 to 42 percent. Otherwise, Maxwell House predicted Folger s would ultimately gain 20 4 Folger had cnltJr",dome e;,stern areas prior to its purchase by Procter & Gamble. (CX 11Z- 28.) The brand held substantial shares of Cincinnati, Indianapolis find JOicbonville. Procter & Gamble s delay in resuming Fulger eastern expansiun was partly the result. of a Commission investigation begun slJOrtly after Procter & Gamble purchased The Folger Coffn Company io 1963 and 1967 consent decree which forbade Procter & Gamble from any promotional price discriminat.ion. for a five year period.(See ex 570J); RX 1106 , ,md Procler Gamble Company, 71 F. C, 135, 147 (1967).) Until 1974 Folger avoided the Baltimore/Washington area out of concern for possible Commission responses (CX 522, 507C; Hunter, Tr. 3085-(;) 204 Opinion percent of the entire eastern region and RMH would decline by 9 share points. (IDF 157, 165; CX 130C.) General Foods concurred in the strategy with two restrictions: sales of RMH were not to fall below reasonably anticipated variable costs, and consumer (8) promotions were not to exceed Folger s levels. (IDF 163- , 166, 173. In February 1973, 16 months after Folger s Cleveland entry, Procter & Gamble introduced Folger s into Philadelphia.5 One month later Folger s came to Pittsburgh. In both cases Procter & Gamble emphasized free samples to consumers and discounts to retailers. (IDF 181- 85.) The Maxwell House Division responded again with advertising coupons and promotional allowances, matching some, but not all, of Folger s discounts to retailers. (IDF 183.

Maxwell House also introduced its Horizon brand, which had been developed to appeal to consumers preferring a "milder" blend such as Folger s, into a portion of the Philadelphja district. It was packaged in a can resembling Folger s (CX 109F - , 115B, 120A.), but failed to have a significant impact. (CX 541B.) In April 1974, Procter & Gamble noted that two and one-half years after Folger s introduction into Cleveland and just over a year following its Philadelphia and Pittsburgh introductions the brand was " solid number 2" in each area, but below objectives in the latter areas and "only now approaching the going 20 percent share objective in Cleveland." (CX 528A.) Folger s had (9) gained 9 percent of sales in Philadelphia and 11 percent of sales in Pittsburgh. (CX 651G; RX 1113.

For its entry into the Syracuse district jn the fall of 1974 Folger switched from free samples to retajler discounts and promotional incentives (such as those used by Maxwell House). (CX 538A, B.) Its retailer discounts ranged from four to five times the traditional allowances. Maxwell House did not match Folger s trade deals, and soon discovered the consequences. Retailers offered Folger s at a lower price than RMH. In less than a year RMH dropped from 41.5 percent to 30 percent of area sales, while Folger s captured 27 percent. (IDF 191, 198; RX 1114A.) Not until Maxwell House decided to meet Folger s discounts "dollar for dollar" did RMH sales begin to recover. (IDF 199; CX 1072F.) Nevertheless Folger s was able to maintain ! A Brazilian frost in July 1972 was an "overruding lsicJ preemptive condition " regarding Folger s d",ci ion not to go ahe;;d with a simultaneous expansion ("roll out ) of the remaining eastern area at that lime. (lIunter, Tr 3103, Procter & Gamble attributed Folger s shortfall in Pennsylvania toeveral factor . The brand had struggled for the first nine months in Philadelphia, as local retailers sold the coffee as a premium price brand. (CX521R.) Likewi, Folger suffered from weak efforts by Pittsburgh retailers to feature the brand in Rales. (CX 528B.) In both areas Folger perceived an "unprecedented level of Maxwell House defensive activity," in terms of media advertsing and consumer coupons (CX 528A.) There was also increased competition from local roasters and the store brands- (CX 105D K) Finally, both Maxwell House and Folger independently concluded that Folger s sales had suffered from inadequate emphasis on retailer promotions, and repeat.purchase incentives- (CX JQ5D, 535E 540D- Opinion 103 FTC.

roughly 20 percent of the Syracuse area s sales throughout the latter 1970' s. (RX 1279.

In addition to Folger s new sales strategy, Procter & Gamble attributed its "very strong Syracuse progress" to a "less intense (10) Maxwell House defensive activity." (Hunter, Tr. 3075-77; CX 541A- B.) A November 1975 Folger memo concluded:

We think it is realistic to anticipate similar restrained defensive activity in the remaining northeastern expansion area. We wil no longer be testing, so there will be no motivation to cloud our test results and discourage our expansion. Further, Maxwell House wil be under considerable economic pressure from the breadth of our introduction into its high share territory. (CX 541B.) However, Folger s expansion plans into the remainder ofthe East were suspended when crop damage from a winter freeze in Brazil increased costs and created an unstable pricing environment. (Hunter Tr. 3078; CX 538B.) Folger s "roll out" was begun in October 1977 being completed in February of 1978. (RX 1l07B.) The results are displayed in Table 2. (11) 'S.

+ (12)CIDI! 1981 39. (31.15.

l072A.FfDr ex 1980 40. (32.17. in entr the 1979 41.4 (33.16. frm taken are 1978 44. (36.4)10. NuUI Q' Full 2.4 Chock 1977 43. (36.10. 21. fot Sales Data Coffee Leading 1976 46.4 (37.4) 15. 1279-80. EssternYears the Ground of 1117, Table HouseFiscal and 1971101981 1975 49. (39. 12. l107A SharesRegular of MaxwellArea: Market the Brands In 1974 48. (39. 12. 1980-81-RX and 1279-80;

1973 47. (38. 13. RX and B-G, 1972 43. (38. 14. 1977-CXI072 1971 42. (38. 16.4 1972- months. 11-13).seven CPFfiI' for Full ar Brands 1971--XI072A; (S Bros, (RegularMaxellHouse) Nuts Data TotalGenera!Foods FolgerHils ChaseSanbornChock SavarinStore Source:Group.-; Opinion 103 F.

By fiscal 1980 Folger s accounted for just over 17 percent of sales in the Maxwell House eastern sales area and had become the nation best-selling brand. (See Table 1.) Folger s share of sales in the eight Maxwell House sales districts it entered since October 1971, ranged from just under 12 percent in New York City to just under 22 percent in Syracuse. (RX 1279.) Meanwhile RMH's share of its total e3Btern area fell from 38 percent in fiscal 1971 to 32. 5 percent by fiscal 1980. All General Foods' coffees combined fell from 43 to 40 percent of sales. D. The ALJ's Findings The ALJ found that following Folger s introduction into Cleveland RMH W3B sold there below average variable cost? for 4 of 12 consecutive quarters, dropping below unit cost by 1.4 cents per pound (1.6 percent of price) from October 1973 to September 1974. (IDF 384.) Following Folger s introduction into Pittsburgh, the ALJ concluded, RMH prices dropped below average variable cost for 7 of8 consecutive quarters, by 7.1 cents per pound (in effect, 9.5 percent of price) from April 1973 to March 1974 and by 2.4 cents (13) per pound (2. percent of price) from April 1974 to March 1975. (IDF 385.) For Maxwell House s entire Youngstown sales district, which contains Cleveland and Pittsburgh, the ALJ found that RMH was sold at a price 2.7 cents per pound (3.5 percent) below average variable cost from April 1973 to March 1974. (IDF 385.) And he found that following Folger s introduction into Syracuse RMH W3B sold below average variable cost for 7 of 9 consecutive quarters, by 4.3 cents per pound (5.2 percent ofpriceJ from October 1974 to September 1975 and by 3. cents per pound (2.4 percent of price) from October 1975 to September 1976. (IDF 386; CX 1389.) However, the ALJ found that Maxwell House generally failed to match the levels of Folger s discounts and promotions in each of the areas where RMH W3B priced below cost. (IDF 175-201.

The ALJ also found that Maxwell House employed its Horizon ,rand as a "fighting brand" in Philadelphia and Syracuse. He found hat Horizon was introduced, at prices far below average variable ost, solely for the purpose of disrupting Folger s introductory efforts. :DF 424 to 427.

The ALJ, nonetheless, found the evidence of below-cost sales insufcient to support liability for an antitrust violation. Maxwell House ,icing in Cleveland, Pittsburgh, and Syracuse could not have been The ALJ adopted complaint counf;J's method of computing the relevant revenues and ost.. Specifically, he Iputed net revenue as gross revenue Jess cash discounts and aJl product promotions t.o retailers as well as JUIefS; he computed variable cost as the sum of the casU! of matc-days (valued at current prices), labor lsportation, warehousing and media advertising, which he expensed in the period incurred; he computed fixed as the sum of the cost.of selling, plant depr!Jciation, administration and central corporate overhead. (lDF 147 379--13, 115.) For the reasons discussed in Section lite.infra we cannot accept the judge s findings of oslsalesjD these areas GENERAL FOODS CORP. 341 204 Opinion predatory, he found, since Maxwell House had no prospect of driving the experienced and financially strong Folger s out of these areas during the periods of below-cost pricing, and could not block the entry of others during any subsequent periods of recouping profits. (IDF 440.) He (14) concluded that despite Maxwell House s below-cost pricing there was no evidence of intent to monopolize, exclusionary behavior, or a dangerous probability of successful monopolization. (ID at 150.

II. The Elements of an Attempted Monopolization Offense A. Introduction The issue raised on appeal under the Sherman Act is whether Maxwell House is guilty of attempting to monopolize interstate trade or commerce.8 The nature of the charged violations requires us to determine which occurred-rivalry that benefited consumers, or conduct that threatened competition.

Making the distinction necessitates a careful examination of the facts. To err in either direction--ither by permitting anticompetitive conduct or by chiling the rivalry that is the essence of dynamic competition-would be to stifle the very competition the antitrust laws are intended to foster.

The Commission recently set out the legal elements of an attempted monopolization offense in E.I Dupont De Nemours Co., 96 F. 653, 725 (1980). Finding the courts to be virtually unanimous on the issue, the Commission has held that a violation consists of three elements:

(1) specific intent to control prices or destroy competition; (2) exclusionary or anticompetitive conduct; and (15) (3) a dangerous probability of success.

We consider each in turn, then proceed to an analysis of the facts of this case.

B. The Intent Element To be liable for attempted monopolization, an offender must b, shown to possess specific intent to achieve monopoly power by preda tory means.'o As the Commission made clear in Dupont this elemen is not satisfied by ambitious and aggressive plans to compete, eve with the goal of taking business from competitors or vanquishing o The monopolization charge was abandoned on appeal. 96 F. C. at 725. The Ninth Circuit, which, the Commission nored, had appeared to depart from the requirem' of a dangerous probability of succe, has reiterated the neces.qity of the third element of the attempt offer William Inglis Sons Baking Co. u. lit Continental Baking Co. 668 F.2d 1014 (9th Cir. 1981),cerl denied S-Ct. 58 (1982) 10 Times PI'cayenne Publishing CO. United States 345 U.S. 594, 626"27 (1953) !! ); &; ); Opinion 103 F.

troublesome rival.11 The antitrust laws provide no protection from such designs, where the means to effectuate them amount to no more than vigorous competition.

Direct evidence of intent to vanquish a rival in an honest competitive struggle cannot help to establish an antitrust violation. 12 (16) To conclude otherwis would contravene the very essence of the competitive marketplace which is to prevail against all competitors. Thus, inherent in the concept of intent is the need to consider the means used to achieve the goals. Moreover, the ambiguities associated with motives, combined with the diffculties inherent in proving subjective state of mind, have further limited the independent significance of the intent element.14 The Commission has, accordingly, narrowly circumscribed the context within which intent may be considered:

Intent is a barren issue without consideration of the means contemplated for acquiring monopoly power.15 Thus, while essential to a finding of attempted monopolization, the element of intent inevitably entails the element of conduct. (17) C. The Conduct Element The issue in cases of attempted monopolization that has drawn by far the most attention from adjudicators and commentators is the boundary between permissible aggressive competition and anticompetitive predatory conduct.!6 While there may be reasonably obvious cases of objectionable means toward monopoly-enforcing a patent fraudulently obtained 'or example-most such activities present a diffcult task ofline drawng. Aggressive pricing is just such an activity. l Dupont DcNemoun; 96 F. C. 653, 72&-7 (1980). 12 Willam Inglis & SUIIS . lit Continental Baking Co., 668 F.2d 1014, 1028 (9th Cir. 1981),cert. denied 103 Ct. 58 (1982).

Id. at 1028 11.7 (quoting Blair Foods, Inc. o. Ranchers Culton Oil 610 F.2d 665. 670 (9th Cir. 1980). Accord ..yes v. Solomon 597 F.2d 958, 977 (5th Cir. 1979), , denied 444 U.S. 1078 (1980) (threat. to r.ompete, and to ve out of business if necessary in the absence of U,fair, anticompetitive or predatory conduct. is not enough /A:ktrr- Vend Corp.lJ. Venda Co. 660 F.2d 255, 273 (7th Cir. 1981) ("We agree with the Fifth Circuit's innatioJ1 inHayes v. Solomon (just quoted) . " AgresheU, Inc. II. liammons Products Co. 479 F.2d 269, 285 1 Gir.J, cert. denied 414 U.S. 1022 (1973) (Court attaches little significance to defendants' assertion that plflintiff d nu right" in a market that defendant claimed as "my domain MCI CommuniCllions Corp. v. American Td. & Tel. Co. 708 F.2d 1081, 1113nal (7th Cir. 1982),cat. den;ed erican Tel. Tel Co. u. MCI Cummunicutiuns Corp. 104 S.Ct. 234 (1983) (Blaming exces.'Jive reliance on the lent for burdening litigation and encouraging inconsistent results) E.I. DaPant DI-NI-mours Co. 96 FTC. 653, 727. Id. at 726-'J .

204 Opinion Predatory pricing has long been condemned by the courts,1 while at the same time its anticompetitive potential has been questioned by economists,!8 More recently, debate over the issue has centered around a rule proposed by Professors Areeda and Turner, who advocate that predation be conclusively presumed when a monopolist' price falls below "reasonably anticipated" average (18) variable cost. Conversely, their argument holds that a price at or above that level would prove lack of predatory pricing. Areeda and Turner reason that no manager trying to maximize the net worth of a firm would voluntarily incur those costs that could be avoided by halting production unless the sacrifice generated future profits above the competitive norm. Simple economic models of the firm suggest that the most likely explanation for such behavior is the expectation that setting price below average variable cost wil eliminate enough competition to allow supra-competitive pricing later. In a dynamic setting, however, where competition involves more than the selection of one product's price and quantity for the current period, both the measurement and meaning ofprice-(19)cost comparisons can be far more diffcult to interpret. An obvious problem is the allocation of joint costs among different products in the multi-product firm.2! Where a firm produces several brands of a product from the same plants, with the same workers and with some of the same raw materials, the precise allocation among brands of even the variable production costs can be arbitrary." Similar diffculties attend the 17 A classic treatment of predatory pricing is containedStandardin Oil Co. of N.el v. United Slates 221 U. 1(1911).

lB See, e, McGee Predatory Price Cutting' The Standard Oil (NJ.) Case J .J, Law & Ecan. 137 (1958) 19 Areeda & Turner Predatory Pricing and Related Prar:ices under Section of the ,'-,herman Act 88 lIarv L. Rev. 697 (1975). Among the other significant contributions to the commentary, are Posner Antitrust Law 184- (1976); R Bork The Anlitru.t Parado:r144-60 (978); the collected articles Strategicin Predation and Antitrust Analysis (S. Salop ed. 1981); .Joskow & Klevarick A Framework for Analyzing Predatory Pricing Policy, 89 Yale L.J, 213 (1979); and McGee Predatory Pricing; Revisitl'd,23 J.L. & Econ. 289 (1980). Calvani and Lynch summarize literature that has become tuu voluminous to described here. SeeCalvani and Lynch,Pred"tory Pricing Under the Robinson Putman and Sherman Acts: un Intr(J/action 51 Antitrust L.J, 375 (1982) !(1 While Areeda & Turner speak in their treatise of a munupolist's pricing practices, they also think their rule relevant to attempted monopolization Predatory priciItg is IJ particularly good example, for such pricing is itself proof that thenot final,alreadyif possessing a degree of markd power, anticipates oht.aining it. Otherwise it would not be able to recoup its losses by monopoly profits earned aller the target lirms "re p.xtinguished,AnlilrustIII Law at 353 1 As the Ninth Circuit stated inTrunsAmcril' Computer Co., Inc. v, IBM Corp. 698 F.2d 1377, 1387 (1983) The unn"rUlinty and imprecision inherent in determining "custs" counsel against basing conclusive presumptions on the relation hdween prices and costs Assessing those relations for the products of a multi-product finn requires allocating known IJnd estimated costs and revenues among various products. While accounting problems do not warrant ignoringcost. figures completely, they do make it unwise to rely exclusively on such ligures Sec also George J. Bcnst.n Accountin!; Numbers and Economic Values, Tllc Antitrust Bulletin (Spring 1982), pp 161215 :i" The diffculty in allocating respond.-mt' s marketing out.lays is perhaps best evidenced by a close correlation between sales of inst.ant and regular ground coffee for both Folger and Maxwell House-which means that promotion of une type of coffee can generat.e revenues for others. (CX 640M. , !! Opinion 103 F.

allocation of costs and revenues related to advertising and promotion that span different geographic areas.

Even more diffcult is the challenge we face in this case: the allocation of costs attributable to and revenues derived from advertising and promotion over time. Promotional outlays or reduced prices that cause current accounting losses may represent an investment in longlived information and goodwill that wil (20) payoff with enhanced future revenues.23 If so, the investment component should be amortized over the life cycle for which respondent expected it to endure. Given the correct measurement of costs and prices, it is possible to find that pricing below average variable cost presents no threat to competition. Substantial discounting or giving away samples may be the most effcient way for a new firm to establish goodwill that might sustain greater sales in the future. Such reductions might also be justified where it is less costly for existing firms to incur current losses than to close and later reopen.25 These are examples of a general proposition that is already well accepted. As even Areeda recognizes: If the market prerequisites for antisocial predation are absent, then prices below cost-however defined-would be serving a different and presumably competitive function.26 (21) In short price" below !'cost"--specially as nprice" and \least" are conventionally estimated an be an ambiguous signal. The limitations ofthe Areeda & Turner rule have not escaped the attention of the courts, where it has gained but qualified acceptance. Judges have recognized the advantage of the rule as providing a more objectively measurable standard to guide pricing decisions than the traditional standards of " unreasonably low below cost " or "ruinous competition" that have been employed in the past. However, the trend in court decisions is drawing away from relying on price/cost comparisons as conclusive. Prices above average variable costs raise a strong, often conclusive, presumption oflegality. Conversely, prices See Nagle Do Adverse in(J-Profitability Studies Really Show that Advertising CreatesBarrier to Entry?24 J. Law- Leon. 333 (1981).

2. The ALJ , in computing GeneralFoods' revenues and costs, treated all such prnmotional incentives as straight price reductions, without recognizing any investment value. Likewise, he expensed al outlays for advertsing in the period incurred and mllde no assessment of their long-lived effects on conswner goodwill. The record does not support that approach- (IDF 379 to 387;see note 52 below. 25 Areeda, 1982 Supplement to Antitrust Law 120-21. 150-2 (1982). Costs of reentry can include investments in goodwil as well as investnuw-ts in plant and equipment: each can resultin several periods oflosse8 before yielding compensating revenues 2" Jdat 149--50.

See, e. 0. Hommel Co- u. Ferro Corp.659 F.2d 340 (3d Cir- )981),cert. denied 102 S.Ct. 1711 (1982) (inclined to accept the basi premise); Northeastern Telephrme v. AT&T, 651 F.2d 76 (2d Cir. 1981),cert. denied 102 S.Ct. 1438 (1982) (adopting marginal costs);William Inglis Sons v. 117' Continental Baking Co., 668 F.2d 1014 (9th Gir. 1981),cert. denied 103 S.Ct. 58 (1982) (price-cost relationships rais. preswnptions of legality or illegality) 2B See e. , MCI Communications v. American Tel. and Tel Cu- 708 F-2d 1081. 1113 (7th Cir. 1981),cert. denied, American Tel. Tel Co- lJ- MCI Communications Corp_ 104 S-Ct. 2.11 (1983) , 204 Opinion below average variable costs can raise a presumption of predation, but that presumption is rebuttable.

Thus, the courts are leaving one principle undisturbed: low prices alone, including prices below some measures of variable (22) cost, do not by themselves mean the seller has violated the antitrust laws. D. The Dangerous Probability Element The third element of the attempted monopolization offense is the dangerous probability of success. However objectionable the intent or conduct may be, there must also exist a dangerous probability that the defendant wil achieve the power to control price or exclude competition in the alleged market(s).3o The courts have attempted to quantify the probabilty of success by measuring the market share of the alleged offender. Market shares in the range of 40 to 60 percent if corroborated by other evidence, have been found close enough to monopoly to support an (23) inference that success was dangerously probable.31 Nevertheless, the decisions have consistently held that market shares standing alone, even in this range, do not provide an adequate springboard to The Commission is onmonopoly.32record as rejecting a narrow marketshare approach:

(O)ur disposition of this matter should not depend on a showing that Dupont' market position exceeded some magic market power (as measured by market share) criterion.

The test for market power depends on all the relevant characteristics of a market: the strength and capacity of current competitors; the potential for entry; the historic intensity of competition; and the impact of the legal or natural environment, to name just a few. Accordingly, courts view evidence on the dangerous probabilty of success " 29 See Calvani and Lynch supru note 19, at 306; see also Hurwitz and Kovacic Judicial Analysis of Precktion and the Emerginf: Trends 35 Vand. L. Rev. 63, 113-9 (1982). 30 Within the range from pure competition to unfettered monopoly, it is necessary to define that degree of freedom from competitive pressure which amounts to legally cognizable monopoly power. The Supreme Cour has held that the requisite power must be substantial: lWJe have monopolistic competition in every non-standard commodity with each manufacturer having power over the price add production of his OWI product. However, this power that, let us say, automobile or sofldrink manufacturers have over their trademarked product. is not the power that makes an ilegal monopoly. United States v. E.I. DltPonl de Nemour. Co. 351 U.S. 377, 393 (1956) (footnote deleted). On the other hand, the Court har noted that. one need not possess "unettered power to control the price " of goods or services to cross the ilegal thrcshold.United States v. Grinnell Corp. 384 U.S. 563, 574 (1966) 3J , Kearny& Trecher Corp. v. Giddings Lewis, Inc. 452 F.2d 579, (7th Cir.1971), cert denied 405 L".S. 1066 (1972) (33 percent, with patent impressive 12 E. , uktro Vend Corp. v. Venda Co.660 F.2d 255, 270-71 (7th Cir. 1981) (33 percent, without entry barriers, insufcient); Nifty Foods Corp. V. Great Atlantic& Pacific Tea C"- 614 2d 832, 841 (2d Cir. 1980) (54.5% dropping to 33% insuffcient); United Stotes V. Empire Ga. Corp. 537 F2d 296, 305 (Ath Cir 1976),rert. denied 429 U.s 1122(1977)(50% insuildent).

33 E. 1 Dupont de Nemours & Co., 96 F. C. 653, 726, n. l6 (1980) j, . g., 346 FEDERAL TRADE COMMISSION DRCISIONS Opinion 103 F.

a whole" to determine "the firm s actual or threatened impact on competition (24) in the relevant market. 34 The Commission wil follow this precedent.

E. Our Analytical Approach As our review of the legal elements reveals, most of the purported improvements to the law of attempted monopolization have not involved rejection of basic rules. Rather, the evolution of the law has advanced through their clarification.

Cost-based rules of conduct are not new, but the developments inspired by the Areeda & Turner proposal have permitted a more consistent and logical application of the standard for conduct in predatory pricing cases. Intent remains an element of the offense, but the capricious condemnation of any strategy to succeed has ended. Similarly, recent improvements in our understanding of competition have provided the setting for proposals to refine the analysis of competitive effects.

Nevertheless, an explicit methodology has failed to emerge from the decisions dealing with alleged attempts to monopolize. The reported cases show as many starting points as there are (25) elements. Some courts have begun with intent.37 Others, eschewing the intent element as an unreliable signal, have commenced with conduct. The Commission, however, is among those tribunals that have approached attempted monopolization by addressing the probability of success before proceeding to the other two elements.'9 Our approach is guided by two related objectives: the desire to improve the accuracy of OUT decisions and the concern for effcient use of prosecutorial and adjudicative resources. Improvements in the effciency of enforcing a rule need not sacrifice the accuracy of adjudication. Indeed, it is possible to enhance simultaneously the confidence in and effciency of decision making by observing some basic priorities. Where the allegations raise a number of issues, the inquiry should focus first on those that may permit an early and effcient resolution of the case.

For potentially complex predatory pricing cases, this principle sug- 31 Lektro- Vend Corp. u. Vend" Cu-, supra 660 F.2d at 271, quotingK('arny& Trecker Corp. u. Giddinfis& l wis, Inc. upm 452 F.2d at 598 35 See e.g.Arccda & Turner supra note 19 at 709- 13; Posner supra nute 19; Joskow & Klevorick supra note S(' e e. Posner supra note 19 at 182-98; Arceda &, Turner III Alltilmst Law 353-4 (1978); Joskow & Klevorick supra note 19 31 See, e.g, Wilham Inglis Sons v. 11'1 Continental Ruking Co. 668 F.2d 1014 (9th Cir, 1981),eer/. denied. 103 S, Ct 58 (982); United Stotes v. Empire Gas Corp. 537 F.2d 296 (8th Cir. 1976),cert, denied 429 U,S. 1122 (1977) 38 See, e.g., ChWicothe Sand Gravel u. Martin Muriel/a Corp" 615 F.2d 427 (7th Cir. 1980) Northeastern Tel Co. u. American Tel Tel. Co" 651 F.2d 76 (1981),cerl. denied 455 US. 943 (2d Cir. 1982). 1" EJ. Dupont de Nemours Co. 96 F. C, 653 (i980).See Lektro- Vend Corp- v. Venda Co"660 F.2d 255 (7th Cir. 1981) Nifty FOOlL, Corp. u. Gre,,! AtlanticPacific Tea Co. 614 F.2d 832 (2d Cir. 1980). 204 Opinion gests certain possibilities for preliminary consideration. The evidence may be clear and free of ambiguity (26) with respect to a particular element. It may become readily apparent, for example, that the alleged predatory price remained above the relevant measure of cost. Alternatively, there may be a complete failure of proof with regard to intent or effect. Accordingly, it is appropriate, as a preliminary test, to seek satisfaction that the record contains suffcient evidence at least to raise material questions with respect to all issues. If a more detailed factual inquiry is required, the next step should be to examine the competition in the alleged market. It may be relatively clear, for example, that resources of the alleged predator are no match for the endurance of its competitors. Only if the examination confirms the vulnerability of a market to predation should the inquiry proceed to the conduct and intent of the alleged predator, the final steps of the analysis.

This approach shares the advantages of narrowing the initial inquiry and perhaps most importantly, starting out on ground well suited to the kind of objective analysis familiar to the (27) Commission. A methodology grounded in analysis of competitive elects promises simultaneously to reduce the probability of incorrectly condemning vigorous competition and to save the time and resources that might otherwise be wasted on an unfocused inquiry. The more elusive assessment of conduct and intent can be avoided where no dangerous probability of successful predation is present. Moreover the prohibited conduct and intent, as we have stated, are firmly rooted in the expectation of success. Thus, where a high probability is found, the analysis of these elements can proceed on a much firmer basis.

IV. The Attempted Monopolization Charges At the outset we note that this record does not reveal any obvious grounds for preliminary disposition. Evidence of each material fact is genuinely disputed. Accordingly we shall proceed to an assessment of 'D Areerla & Turner describe thi approach as an "intermediate position" in thp.ir treatise. An intermediate pusition would seek to identify easily proved fClct without which predation i most unlikely to occur and then to addres such facts before the antitrust tribunal is forced to embark upon the r.omplcx inquiry into price-cost relationships- IIIAntitrust Low at 354 Joskow & Klcvorick call the first step a "structuralist approach" tdthough t.he inquiry tl1",y dp.stribc goes far beyond the simple concentration calculations typir.ally associated with t.he "structural" approach upranote at219 "Th(' basic approach of proceeding from the competitive sdting to the challenged conduct has been standard ill other antitrust adjudications. The prevailing characterization ofthe monopolizlItion offense calls for a two stage analysis: first determining the exis1fnce of monopoly power, and then examining for the exercise of it. Joskow & Klevorick supra nlJte 19 at 265, citing United SCates v. Grinnell Corp384 US. 563 (1966) Similarly, in merger cases market definition is a necessary first step of the competitive analysis-United States v. Marine Banc:rpIJration 418 U.S. 602, 618 (1974). Examination lJfthe strengths and weaknesses of competition in markets affected by a merger focuses the competitive analysis and reduces the uncertainty of drawing the correct conclusions.See generally, Joskow & Klevorick supra note 19 at 213.- 269 Opinion 103 F.

Maxwell House prospects of success in monopolizing the alleged markets. (28) A. Definition of the Relevant Market To determine whether a firm has the means to raise price and maintain it above a competitive level e. monopoly power, antitrust analysis generally begins with the definition of the relevant market. According to complaint counsel, the relevant market in this case consists of Maxwell House s individual sales districts-areas roughly centered around metropolitan areas such as Cleveland, Pittsburgh Philadelphia and Syracuse.

The Commission, in previous cases has made clear what kinds of evidence we consider most valuable in the definition of a relevant market. Most direct, but rarely available, are reliable measures of supply and demand elasticities.42 Of the indirect evidence, especially probative is the level of entry barriers surrounding a market. also have recognized the inferential value of evidence revealing price disparities, transportation costs, and transshipments between locations, as well as the perceptions firms have about the competitive threat posed by outsiders.

With regard to entry barriers, we find the record consistent and clear. There is virtually no support for the contention that significant economic boundaries exist among the sales districts alleged to be relevant markets. Evidence on transportation (29) costs, shipping patterns (including transshipments), and the perceptions of competitors all point to markets that reach far beyond the geographical divisions Maxwell House drew for the purpose of marketing coffee. Perhaps most persuasively, the record shows repeated examples of successful entry by other firms into district after district. By comparison, the evidence on which complaint counsel rely to prove markets and monopoly power-price disparities, profits, advertising and product differentiation-is ambiguous and incomplete. 1. Evidence of Transportation Costs and Shipping Patterns There is no dispute over the insignificance of transportation costs compared to the value of the product in this case. For example, from 1971 to 1977 transportation costs for all Maxwell House Division ground coffee averaged 1.2 percent of wholesale price for distances of 500-1 000 miles, 2.2 percent for distances of 1 000 to 1 500 miles and 7 percent for distances exceeding 1 500 miles. (RX 1224.) On aver" The Grund Union Carnpuny, III Trade Reg- Rep- (CCH) n22 050 at 22 703 (1983) 43 Federal Trade Cumri88ion Po/iq Statement on Horizontal Mergers (June 14, 1982) SeeId.

)).

204 Opinion age, the transportation cost for all distances amounted to 1.2 percent of retail price. (RX 1224.) That transportation costs imposed no more than trivial burdens even on long distance shipments, is reflected in the shipping patterns of the major manufacturers. Maxwell House produced coffee at only four plants, located in New Jersey, Florida, Texas and California. (IDF 105.) Shipments from those plants were not confined to their surrounding areas; the Florida facility produced 50 percent of the RMH distributed in the United (30) States, 60 percent of all ground coffee, and 100 percent of the Max-Pax brand (IDF 106, 107.) Mellow Roast was produced only in New Jersey and distributed nationally. Maxwell House s other coffee brands traveled equally far. (IDF 108.) Folger s has only three regular coffee plants-in California, Louisiana and Missouri. For its entry into the East, Folger s added no new capacity, (IDF 111.), and there is no evidence that transportation costs figured in its selection of any specific targets. Hils Bros. used two plants, and sometimes only one, to produce coffee shipped coast to coast. (IDF 114-15.) Chase & Sanborn had been shipping to all its sales areas from only one plant, in New Orleans since the early 1970's. (IDF 113.) Even the tiny Euclid Coffee Company of Cleveland serviced accounts in several of respondent's sales districts-in Ohio, western Pennsylvania, western New York and throughout the State of Michigan. (Repak, Tr. 2602.) Presented with similar evidence in a recent case, the Commission rejected regional geographic markets.

The test for measuring geographic market is where consumers (in this case retailers) can practicably turn for an alternative source of supply. Here the record is clear that frozen pizza manufacturers could sell virtually throughout the United States from a single plant with no significant cost disadvantage. Thus, the power of any given group of sellers serving a city or region at a given time to raise prices is limited by the capacity of virtually all other domestic manufacturers to compete on practically an equal footing in that (31) city or region an economic situation which requires a finding ofa national market and the elimination of geographic submarkets.46 lcitations omitted. The record also reveals patterns of transshipment and trade flow that belie the existence of significant barriers separating the markets complaint counsel attempt to prove. Trade flow is the shipment of product from a retailer or wholesaler s warehouse located in one sales district to a retail outlet located in another, while transshipment is (L. Nel- the cross-district movement of product among warehouses.'7 '5 Complaint counsel's witness, II. Michael Mann, testified that transportation costs are "negligible" jn terms of the optimal location of Tracting planL'I. (Tr. 3568-72. 46 Pillsbury, Inc.93 F.Tc. 966, 1030 (1979). 47 Tran hipping is not only done by the grocery trade, but there are people that we have referred to as "bandit brokers They know what the going trade (rate) is, and they can buy that coffee in that area and stil pay the freight and make a profit on it. (Salesman, Tr. 5934 Opinion 103 F.

son, Tr. 5729.) Such movements are evidence of arbitrage, prompted by price differences that offer profits from the shipment oflow price coffee among regions.

Trade flow had been a problem for RMH before Folger s entry into the east. (CX11-Z18.) Maxwell House referred to the Boston, New York and Philadelphia sales districts as "the complex" because beginning in the late 1960's trade flow made it diffcult to limit deals to retailers to just part of the area. The complex has grown in recent years to include part of the Syracuse district and now runs from Maine to Virginia. (IDF 103, J. Mann, Tr. 6454-5; L. Nelson, Tr. 5730-1; see also CX 705B.) With similar problems throughout the East, Maxwell House took (32) such cross-district product movement into account when structuring its promotional incentives in each district.

Neither Folger s nor Maxwell House was immune from arbitrage as Folger s moved east. Soon after coming to Cleveland, Folger s impact on supply began to spread beyond the areas in which it was introduced. Folger s coffee appeared on the shelves in Kroger s and several independent stores in Pittsburgh as a result of "spil-over from (the) Cleveland" trade.48 (Cleveland and Pittsburgh are alleged to be in separate markets, although both are in Maxwell House Youngstown sales district.) (CX 919F; Zurcher, Tr. 6288.) Similarly trade flow required Folger to offer the same promotional allowances in Cincinnati and Columbus-two areas outside the Youngstown district-that it offered coffee retailers in Cleveland. (Hunter, Tr. 3094- 5; CX 497 A; see also, IDF 101, 104. The Maxwell House Division s response to Folger s entry into both the Philadelphia and Syracuse sales districts was constrained by the prospect oftrade flow. (CX l05E, 192F, 634C, 63-Z11, 705B.) In Philadelphia, fears that discounts would spread throughout the eastern complex kept Maxwell House s promotions and discounts much smaller than Folger s. (CX 649F; Laster, Tr. 6982.) Maxwell House at first took the risk of allowing higher shelf prices than Folger s in Syracuse, rather than seeing discounted RMH cut into sales in other districts. (CX 649G.) (33) When substantial sales losses prompted a price reduction in the Syracuse district after all, fears of trade flow problems were confirmed. (IDF 99.) Maxwell House had introduced a fresh-lock lid in that area for monitoring purposes, and cans with the new lid showed up in significant quantities as far West as Portland and Los Angeles and as far south as Jacksonville. (Salesman, Tr. 5933-6; see also Keller, Tr. 6584- ) Folger, which had chosen to do its test 1R Retailer cuupons, which can be limited morcefTedively tu specific geographic areas, were used more extensively in t.he Youngstown sales district than elsewf,cr",-ex(See86H, 87C, 89B, 187Z, 710A.) Their USf' in Cleveland failed to prevent arbitrage ).

204 Opinion expansion in Syracuse partly in order to avoid trade flow (eX 539B), also experienced transshipment out ofthat area. (CX 541A-B.) Overall, the record shows a trend towards increased trade flow and transshipment throughout the East in the period following Folger s 1975 entry into Syracuse. (J. Mann, Tr. 6450-51 , 6462-4; IDF 98-104.)49 The implication of this evidence is inescapable. Ifan economic barrier exists between any domestic coffee producer and any other producer s marketing area, that barrier is not distance. The cost of transportation is unlikely to deter either a producer of ground roast coffee or a reseller, no matter where (34) located, from considering profitable opportunities in distant regions. Accordingly, the putative monopolist of a sales district must look beyond transportation costs for the insulation from competition that permits monopoly power. 2. Evidence of Price Differentials Complaint counsel claim there are other signs of entry barriers that outweigh the evidence of shipping patterns. The best indicators of geographic markets, argue complaint counsel, are price differences among regions. (CAB 9-13.) We recognize that significant, persistent differences can indeed be persuasive. However, the quantity and quality of evidence showing price difterences on this record is far from substantial.

First, as the ALJ observed, the record is virtually devoid of any evidence on the overall wholesale or retail price of regular ground coffee. As a result, definitive comparisons among brands or across areas are impossible. The only company for which cross-district price data are ofiered is Maxwell House. Even there, however, actual transaction prices at which Maxwell House sold coffee were not introduced. Rather, Maxwell House s prices were calculated by deducting from total annual revenues, cash allowances and expenditures for allowances and promotions other than advertising. (CX 1081. This method of calculating prices results in some level of distortion because some consumer promotions and trade incentives have both price-reducing and advertising effects. A year-by-year comparison of RMH prices in the Youngstown district with those in (35) the Syracuse district between 1971 and 1977 illustrates this problem. If price is calculated by deducting the cost of all retailer and consumer incentives from gross revenues, the price differences between the two dis- 49 Although Maxwell House was able ill some instances to mibgiite trade fiow problems by i)Signng price. according to historical resale patterns among areas (Keller, Tr- 6670-71), there is no evidence of any such ability to control the prices oftnmsshippcd coffee by the wholesale arbitragers kolorfully called "bandit brokers (See Salesman, Tr- 5934- , J. Mann, Tr. 6168-9.) Moreover t.he very reluctance to reduce wholesale prices in some areas for fear of trade f1uw into others is itselfinconsistent with a complete ability to prevent trade flow from equaljzinj; prices among areas. For example, the Maxwell House response to Folger s Philadelphia entry wa constrained by fears that wholesale price reduction1s there would spreCld to the New York, Boston1, Baltimore lmd Washington areas owing to "trade now throughout (thej eastern complex," (CX 70Sll) Opinion 103 F.

tricts range from 2.6 percent of price in one year to 18.1 percent of price in another. (CX 1081.) If only non-performance retailer incentives (those requiring no advertising or display function) are deducted the range is from 0.2 to 10.8 percent of price. (RX 1141A, 1142A; Elzinga, Tr. 9464.

Complaint counsel argue that deducting both performance and nonperformance incentives from total revenues gives a more accurate calculation of price. Respondent contends that performance incentives should not be deducted, as they are not price reductions, but promotional costs. Neither party offers suflcient evidence to support either extreme. 50 Second, to find that a specific area is a separate market there must be a demonstration ofthe ability to raise prices above those prevailng in contiguous areas without attracting substantial entry. Complaint counsel offer no evidence of any such ability by respondent or any other roaster in the alleged markets. At best, their evidence suggests the ability to lower (36) prices in periods of intense price competition below those of surrounding areas. For example, both before and even more significantly, after the alleged predation in the Youngstown district, the wholesale price ofRMH (as calculated by complaint counsel) was lower than in the contiguous districts and than the average price for the nation as a whole. (CX 1081.) Thus, Maxwell House prices were, on average, the lowest in Youngstown of any of its eastern districts, despite the tact that it was in Youngstown that Maxwell House had the largest market share. (RX 1279-80. Third, leaving aside the issue of which costs to subtract and how much, we have just described the transshipping and trade flow that occurred repeatedly during Folger s expansion to the eastern regions. One would expect outbreaks of transshipment when price differences between regions were large, and this is indeed what happened here. Substantial reshipment wil tend to reduce or even eliminate such differences, as the retail trade in the higher-price areas buys coffee from the trade (retail or wholesale) in the lower-price areas. The average revenue figures ofiered by complaint counsel, which were taken from Maxwell House s records, do not account for those trades. Some quantities of lower-price coffee that were reshipped to higher-price areas are erroneously attributed in complaint counsel' evidence to the lower-price area. Accordingly, prices that were actually paid by the retail trade in the allegedly high-price areas are overstated (and quantities understated), while the opposite holds in the 50 There is no dispute that .\axweJl House enforce the requirement that retailer1 provide services in order to obtain a performance allowance. (lDF 462-474.) The question concerns the extent to which some allowances may exceed the cost of performances. No direct evidence was offered on this issue. (IDF 464.) TIIU:!, we cannot justify allotting any specific portion of the allowances to price reductions. Complaint counsel' s argument, that none ofthe cost of performance he consider, is clearly contrary to the evidence. 204 Opinion alleged areas of (37) predation. It is impossible to determine the extent to which reshipments caused the actual price prevailing in an area to vary from the calculated average revenue on Maxwell House books. Most likely the effect was the greatest when the revenue figures show the largest disparities, because that' s when incentives to reship are at their peak.

For these reasons we must agree with the AU' s findings on the meaning and relevance of the price comparisons to the geographic market. With insuffcient evidence of overall ground coffee prices or of supra competitive RMH prices, with price differences that change substantially between respondents' and complaint counsel's calculations, and with the failure of the evidence to account for arbitrage, the price data in this record are "so complex and ambiguous that (they are) of little practical utility in resolving the issue of relevant geographic market." (IDF 130, citing Elzinga at Tr. 11 992. 3. Evidence of Product Differentiation Complaint counsel contend that product differentiation-by respondent and by others--recte, ross-district entry barriers. Specifically, they argue that different levels of(38) product promotion and consumer acceptance among the competing brands constitute barriers that establish each sales district or specified area therein (e. the Cleveland and Pittsburgh portions of the Youngstown district) as a separate market. (CAB 16-19.) The argument suffers both from logical inconsistences and a basic misunderstanding of entry barriers. The contention that product promotion can create differential brand preferences among areas runs squarely into the claim that advertising and promotion are expensable items that should not be amortized over extended periods of time. If promotional effects are quickly dissipated, then advertising should not delay entry or raise entrants' costs above those of established firms. Each firm presumably would begin from the same position, with no goodwil remaining from prior promotions, at the start of each period. If the claim is that the effects are only gradually dissipated, which finds more support in the record 52 then complaint counsel's price-cost calculations (39) er- 51 Complaint counsel commissioned a study Qftheextent of trade f10w for aJ1grocery prouuct.' combined among Maxwell House s sales districts in calendar 1971 and 1973. (CAB 14.) (IDF 73-4 , ex l029A-C; Lysaker, Th. 1762 1795.) Since it ignores trans.'hipping, does not break out coffee flows and does not cover the period after 1973 in which both trade flows and transshipping ofRMH and Folger s were increasing, we must agree with the ALl that it teJla us little about the actual extent of arbitrage of coffee across districts as Maxwell House respollded to Folger entry 52 The record in this case distinctly reveals an investment component in the cost.! of advertising and promotion. Respondent expected the benefits from its advertising and promotional dcfense to occur over a per:ud of many years. (IDF 157- 171.) For example, a memo dated Beven months after Fulger s initial expansion into the East from the Maxwell House president to the General Foods executive in charge of that division indicawd the expectation that respondent's defensive efforts would generate additional cost. 'I over a two-year period aod additional benefits over a subsequent three.year period. (CX 130 F ) Other Maxwell House documents utili7.e a ten-year period for estimating the benefit. 'I of stepped up advertising and promotional outlays. (JDF 168; RX 518 e.t;; ex 640 M, 6,16Q.) Opinion 103 F.

roneously expense advertising in the year incurred and overstate the relevant costs.

Accepting that real or perceived differences among products can build a loyal following of consumers does not amount to an acknowledgement of entry barriers. There is no dispute that the effective use of every element ofthe "marketing mix including the maintenance of consistently desirable product quality as well as the selection of a proper balance of the various types of advertising and promotioncontributes to the identity and success of a brand. Indeed, a company may reap the benefits of both marketing as well as superior quality by charging retailers a higher price than its less successful competitors. (See below, Section IV. ) But, in and of itself, this fact is no more a barrier to entry than the requirement that every coffee producer must make an effcient investment in plant and equipment and must roast an acceptable mix of green coffee beans in order to remain profitable. We cannot accept product differentiation and price disparities as entry barriers without an explanation of how product differences bar entrants. The question is whether product differentiation imposes substantial non-recurring outlays that raise the threshold entry level of potential competitors so high as to enable respondent and/or any of the established firms in any of the alleged markets to (40) restrict output and raise prices for a significant period of time.

We can find no substantial evidence of any non-recurring costs of advertising and promotion that enabled respondent or any of the established firms to raise consumer prices in any of the alleged markets without attracting significant entry. As we have found, Folger sales responded immediately and dramatically to introductory promotions. In fact, Folger was able to achieve nearly 3 percent of sales in both the Charlotte and Atlanta districts with minor marketing efforts. (CX 539C- , lO72F; RX 1107 A.) Numerous local store brands were able to take hold and compete in each area with minimal marketing efforts. In Philadelphia, during the intense competition following Folger s entry, store brands doubled their sales share to just over 30 percent of the total in four years. (CX 1072 D ) For the entire Maxwell House eastern sales area, the sales share of those brands grew from less than 15 percent in 1972 to just over 21 percent in 1977-a 40 percent gain. (Table 2.

Moreover, the record reveals that changes in the relative intensity SJ The relevant costs would he understated if, contrary to the record (Compare RX 1141A and E, RX 1142A and ) there had been a heavy increase in respondent' s promotional ou.trays before Folger s entry OInd a reduction thereafter. The record shows no such aUempts to anticipate Folger s introduction by heavy investment in advertis- \; As Posner succinctly defines it, an entry barrier is a "condition that imposes higher long n. costs. on a new entrant than arc borne by competitors. Antilrust Law, (I97fi) 204 Opinion of promotion among coffee brands did not insulate any of the established firms from the threat of entry. Rather, such changes resulted in sharp and sudden swings in sales shares among them. (See below Section IV. B.2. & C.) In short, support for (41) a connection between product differentiation and entry barriers is lacking in this record. 4. Actual Entry The failure of the record to establish entry barriers into Maxwell House s eastern sales districts suggests that the evidence might show instances of successful entry. Indeed, that is so, with the most obvious example being Folger, whose successful entry into the alleged markets precipitated this case. In addition, both Chock Full O'Nuts and Savarin were expanding into the Pittsburgh portion of the Y oungstown sales district as the .record was closed. Indeed, Chock Full 0' Nuts, which was expanding generally throughout the East, extended its distribution to include the entire Syracuse sales district where it increased its share of sales from 3 percent in fiscal 1971 to 10 percent in fiscal 1980. (CX 1072A; RX 1278; Salesman, Tr. 5993; CPF 11identifies Chock Full 0' Nuts as " CJosed Group" in CX 1072A.) There was also significant entry before 1971. The most notable example came in 1962, when Hills Bros. Co. began a marketing campaign to enter the eastern sales regions of RMH. Like Folger nine years later, Hils Bros. began within Cleveland. General Foods responded there, and in every other area Hills Bros. entered, with stepped up promotional and advertising expenditures on its RMH brand. (CX ll-Z118.) In spite of these efforts, which spawned rounds ofgeneraJ discounting, Hills Bros. first-year penetration of each ofthe (42) five areas it entered from 1962 to 1967-Cleveland, Pittsburgh Syracuse, Boston and Philadelphia--xceeded 10 percent. (CX 8-Z48 ll-Z144, 1070 G- ) According to the 1967 Maxwell House "marketing plan I-EUs has probably become a major factor. . . for the foreseeable future. . . . If(they are) willing to spend in future introductions as they have in Boston, it wil be extremely diffcult for RMH to prevent them from becoming a major market factor (i. over 10 percent share), (CX 8-Z4 7,) As of fiscal 1971, the year preceding Folger s entry into the Cleveland area, Hills Bros.' shares of Maxwell House s sales districts stood at 10 percent in Boston and Philadelphia, 16 percent in Youngstown (Cleveland plus Pittsburgh) and 17 percent in Syracuse. (CX 1072A. Hils viewed its eastern expansion as successful (Toy, Tr. 2130.), as did Maxwell House. A Maxwell House memo dated October 1971 reviewed the "results" of Hils' eastern expansion just as Folger s was starting its entry into the Cleveland area: Opinion 103 F.

Significantly, Hils has been able to maintain respectable shares in all ufihe eastern areas, except for the New Yark district. This could certainly provide ample encouragement for Folgcr s-especially in view ufthe fact that Hills is strictly a coffee company, with considerably less in the way of resources or marketing expertise than Folger s or its parent, Procter and Gamble. (CX 642B.) Such expansion left a distinct impression on manufacturers as well as retailers. They perceived the eastern sales districts (43) as vulnerable to the entry of the other roasters, especially Folger. It is clear, therefore, that the eastern sales regions alleged by complaint counsel to be distinct markets--leveland, Pittsburgh, Philadelphia and Syracuse-had experienced significant entry before Folger s successful introductions there. In addition, there was substantial expansion of local and regional roasters following Folger entry. As the ALJ, concluded, all this taken together is "convincing evidence that the Maxwell House Division could not exclude competitors from" the eastern sales areas. (IDF 321.) 5. Conclusion In conclusion, virtually every characteristic of the market for ground roast coffee contradicts complaint counsel's contention that barriers to entry exist in the eastern sales districts of RMH. The low cost of transportation put every area in the country within the economic reach of a single production plant. Price disparities between areas prompted the transshipment of coffee from discounted regions to premium-priced locations. Two major brands, Folger s and Hills Bros., (44) successftilly entered each of the areas in which attempted monopolization is alleged. Folger s became established in some areas with minimal promotional support. Regional roasters and supermarket chains also entered some areas and successfully expanded their shares of sales in others. Thus, the evidence upon which we typically rely to define relevant markets indicates that those markets are larger than complaint counsel claim they are.

B. Indices of Monopoly Power A firm whose unilateral output decisions substantially affect the prevailing price ofa product is said to possess monopoly power. Ifsuch a firm restricts its output, the result wil be a permanent increase in price. This can only happen if other firms do not respond by expand- 55 The testimony afthe head of Maxwell House s Young town district (which includes Cleveland and PitU!hurgh) maybe most descriptive The Cleveland trade said that they Wolger sJ arc coming in and they arc going tu take number one-the number one coffee and the number on" brand OIt our expense. . . . (Zurcher, Tr. 6293. 5E.Complaint counsel do not allege attempted monopolization in any mOirket larger than a sales district. Lacking 'Ouch an allegation, we are constrained on review from liDding that respondent attempted to monopolize some larger market 204 Opinion ing their output to make up the shortfall. If, however, other firms do respond to attractive prices and fill in the shortage, there is no monopoly power.

Complaint counsel rely on evidence we have not yet addressed to argue that Maxwell House possessed or could attain monopoly power in certain of its sales districts. While our findings with respect to relevant markets cast serious doubt on whether monopoly power can be shown in the sales districts, the (45) possibility remains.57 Therefore, we wil review the allegations going directly to Maxwell House monopoly power.

1. Market Shares Complaint counsel point first to Maxwell House s high sales shares in the alleged markets. Maxwell House accounted for the largest portion of sales in each of these districts, with shares exceeding 40 percent during the period. Such high shares, according to complaint counsel, establish that Maxwell House possessed market power. Ifnot already within its grasp, the successful monopolization that can break the law was dangerously close to Maxwell House s reach. (CAB 19- 44-45.

While the definition of monopoly power has gained general acceptance among legal and economic practitioners, the measure of that power has not. The gauge most frequently used has been the market share that firm commands. This measure has the advantage of being easily quantifiable, albeit after the appropriate units of account (revenue, volume, capacity, etc.) have been determined and the relevant markets (geographic and product) have been drawn. The disadvantage is that this gauge does not directly measure market power, but is at best a rough proxy for it. That proxy has been increasingly questioned on both theoretical and (46) empirical grounds for its correspondence to actual monopoly power.

At best a high market share can be said to be a necessary but not a suffcient condition of monopoly power, since it is unlikely that a firm with a small share can affect market output and price significantly by adjusting its own output. Even with a large market share, however, a firm cannot safely presume that its own attempts to restrict output wil go unanswered by competitors in the market or potential entrants outside the market. Thus, as we have already noted, market concentration indices standing alone, tell us little about 57 A firm, for example, may possess market power throughout a market that completely encompa,lRes an area incorrectly identified as the r levant market. 51 A critical review of economists' attempts to document the correspondence can be found in Goldschmidt, Mann & We.'to!1, Indu t.jaj Concentration: The I\cw Learning (1974).See also Ethyl Corp. D. 9!28 (Dissenting Opinion ofChainnan Miler), rf!IJ d sub Mm. E. 1. Dupont de Ne"'Ollrs Company, Docket os. 83-4102, 83-1106 (2d Cir. Feb. 23, 1984) );

Opinion 103 F.

competition or monopoly power.

Our previous discussion of the relevant market undermines the competitive significance of complaint counsel' s evidence on market share within a district. Since, as we have found, the relevant market is larger than a sales district, a firm with as much as 100 percent of the sales in a district could still lack monopoly power. This is because it is inherent in our finding as to the relevant market that customers in the district have access to suppliers outside it. The record reveals that this is true with respect to the Maxwell House eastern sales districts. (47) Retailers have access to supplies from roasters, wholesalers and warehouses outside each area.

A more fundamental flaw further diminishes the utiity ofevidence on market shares here. Even supposing insuperable barriers to entry by new firms, the existing firms' shares of sales in each district do not reveal the full pressure they exerted on Maxwell House or any other producer. No less important than current shares is the ability of the competition to augment its output in response to a price increase. This ability derives from the productive capacity of competing firms, not merely their current production. Sales may sometimes serve as a proxy for capacity, but when the relationship between sales and capacity breaks down-for example when production falls short of capacity or a firm outside an area ships in only a portion of its production-current sales shares lose their significance. Excess capacity has haunted the coffee industry throughout the 1970' s. (See above, Section II.A. and IDF 22.) With each individual sales district accounting for a minor portion of sales, undoubtedly the national firms and probably the larger regional firms could expand their shipments enough to serve every (48) customer in any sales district.61 For the major firms like Folger, doubling sales in one dis- 59 SC" '"ction HID slJpra wThis principle has been recognized ill both the cour'- add the commentaries.See e. , United Stotes v. General Dynamics Corp- 415 GS. 486, 502 (1974) (" A more significant indicator ofa company s puwer effectively to compete rthan its sales) lies in the state of a company s I.committed reserves (i.e. capacityJ" Ta.mpa Electric Co. !! Nashvile Coal Coo 365 C.8, 320 (1961); Lande & Posner Market Power in Antitrust Cases 94 Harv. L. Rev. 937 963-64; Arceda & Turer, II Antitrusl ww 11523a (1978). G\ The fate of tile smaller roasters provides BOfie indirect evidence of their immunitythe intenseto competition in various soles districts.The record cont.ains no evidence that any of respondent s competitors were unable to operate their roasting plants at any time because of respondent s activilies. None of those cited by comploint counsel has departed any of the cited areas, let alone halted operations; even the least successful Dc(PawLima and Eudid) have remained viable Still others, including Chock Full D'Nuts and Savarin as well as Folger expanded into additional Easten' areas.(See above, Section IV.A. Although for some of the roasters allegedly injured by respondent' s actions the sale of regular ground coffee in specific "areas did become unprofitable for a time, none ever decided to cease plant operatio.os. For example, Paw De Lima, a local roaster in the SynlCllse area, narrowed its regular ground coffee marketing radius from the Greater Syracuse area to the city itself in response to the increased competition. However, it then extended it. marketing efforl to include institutional and offce buyers and in 1979 reported an overall profic(Dc Lima, Tr 2535 and 2590.) Moreover, it later reexpanded to its original marketing radius of regular ground coffee. (Salesman Tr. 5941.) Euclid, a roaster based in Cleveland, also distributed in western New York and Pennsylvania and all of Michigan, Ohio and West Virginia, so that respondent' s activities in Cleveland were unlikely to have threatened it. existence. Tills, the r cord strongly suggcsw that sales of ground roast coffee arc unlikely to be disciplined by ao outbreak of intense competition, price or non-price, within any Maxwell House sales district. (De Lima, Tr. 2535 and 2590J 204 Opinion trict while entering others was accomplished without building new plants. Under these circumstances, it is doubtful whether market shares, however measured, can yield any useful information about the monopoly power of a firm or the probability of achieving it in any of General Foods' sales districts. (49) 2. Prices and Profits Complaint counsel maintain that high market shares in the alleged markets distinguished RMH from other brands and facilitated higher prices and profits for General Foods-a sign of market power. (CAB 16-18.) We have already noted our reservations regarding the evidence on prices. Whether the source is higher wholesale prices, lower promotional costs, or some combination of the two, the record does show that Maxwell House enjoyed greater profits in sales districts where it was more popular. (IDF 283-302.) Whether we can infer market power from this evidence depends upon RHM' s power over ultimate consumers. Product differentiation does not permit the restriction of output and maintenance of prices above competitive levels unless the producer can restrain ultimate buyers from turning to other brands. If General Foods enjoys monopoly prices and profits in coffee sales, it should be visible at the retail counter in the price of RMH.

The evidence in the record indicates otherwise. Retailers view coffee as a "board item " a product which is among the "most price sensitive items that we carry. . .." (Engel, Tr. 1707.) Sales volumes and relative shares in the RMH sales districts fluctuated widely as consumers responded to price changes and promotional features of various brands. There are (50) numerous examples of this extreme share sensitivity to relative prices in every area in which predation is alleged and it is corroborated in testimony by witnesses for both parties. Prior to Folger s October 1971 entry into the Youngstown sales district (Cleveland and Pittsburgh areas), the first focus of Folger eastern expansion, RMH average retail prices had previously tended towards dose parity with its major competitor, Hills Bros. (IDF 318- 20; CX 89D; Toy, Tr. 2149-2150; Trone, Tr. 1462-3; Engel, Tr. 1648 12067- , 12070, 12095-6; Epstein, 12161- ) But the average parity obscured frequent difierences among brands depending on which one was being featured by local retailers. As described in Maxwell House documents, Youngstown was a "hot" features area. First one brand and then another would reduce price or increase promotions in efforts to gain share. The moves proved successful-but only for temporary In that regard, a differentiated brand may offer 11 higher lp.vel of cunsumer satisfaction and hcn e a higher level of output than a lower price brand. In that case the higher output and not market power explains the higher price. See Bork supra note 19 at 313.

Opinion 103 F.

sales gains; in the 1960s and early 1970s RMH penetrations had varied from 33 to 52 percent, Hills from 14 to 37 percent, and Chase and Sanborn from 6 to 15 percent. (CX 85B , 86F; see also Engel, Tr. 12070.

Folger s was repeatedly sold at a retail price above that of RMH in Youngstown (eX 528B, F.) Then, in 1976, for the first time since Folger s entry, retailers began to sell the brand at lower prices than RMH. (CX 689A.) In March of that year Folger s temporarily gained 10 percentage points as RMH lost 16 points. (eX 683B.) Over the course of the year Folger s gained and RMH lost 3 points apiece. Thus, in the most successful area for RMH, where it held a (51) to-l advantage over the number two brands, its fortunes fluctuated with its relative retail price. When comparable brands were priced at a discount compared to RMH, it lost sales. When RMH regained a pricing advantage, it gained sales. The same held true for Folger Hills Bros., and, earlier, Chase & Sanborn. When Folger entered its second Maxwell House sales district, Philadelphia, in February 1973, RMH was being priced as much as 3 to 4 cents per pound below its closest major competitors, Hills and Chase & Sanborn. (CX 89A- ) Folger s was priced for three months at a premium of 3 to 5 cents per pound above the other major brands. (CX 521A.) At that time, a Folger s document noted the extreme price sensitivity of coffee consumers, callng its pricing disparity with RMH and others the major problem with its introductory efforts. The Folgs memo said:

We view our Philadelphia pricing disadvantage as a potentially crippling problem which may well preclude our attaining a reasonable franchise. OUf overall experience in the coffee business is that consumers are keenly price sensitive. So much so, in fact that parity pricing is a basic strategic objective in the Folger area. Adherence to that objective is regarded as critically important, despite our market leadership position. (52) In fiscal 1977, RMH prices in the Philadelphia district rose above Folger s and those of its other major competitors, whereupon RMH lost 15 per cent of its share of sales. (CX 691B-C; CX 1072.) With rapidly escalating coffee bean prices that same year, RMH also lost significant ground in the Boston and New York sales areas (not yet entered by Folger s) to Chock Full O'Nuts and the store brands, as re! ex 521B. The memo noted two similar expcricnc'-S outside the eflstern area" In two jn tances in the recent paBt in which any major brand har been out of line in the Folger area, both underscore coffee pricing sensitivity- Tn our PorUand district Maxwell House gained a (3-5 cents per pound) advantage over Folger in the 2-!b- size in early 1972 (and within 6 months hadn . increased its 2.1b. shar by 13 percent, while Folger declined by to percent. lln Kansas City a 4 cents per pound MaxweU House advantage on the 3-lb- size caused a similar shin to RMH.J It should be noted that both these districts are long-established Folger districts in which we are the dominant brand with shares of36 percent and 54 percent respl ctjvely, vs. axwell House at 6 percent and 16 percent respectively- (CX 521IJ) , 204 Opinion RMH prices rose more rapidly than those of its competition. (CX 687 A, 689B, 691B, 192B.) Folger s on the other hand, got the benefit of a lower retail price in Syracuse the first expansion (district) where Folger achieved a shelf price advantage (i. a lower retail price than RMH). (CX 20-Z94; see also CX 528F). Within two months of Folger s entry, it had attained a 31 percent share against 36 percent for RMH. (CX 7l0E.) Two quarters after its entry, Folger s was at a 22 percent share and RMH had declined to 34 percent or 8 points below its base level. (CX 709D ) Two years later a Maxwell House memo recommended that RMH match a Folger s price cut in order to retain retail shelf price parity and thereby avoid "a loss of 10 share points." (CX 160A.) The RMH Fiscal 1977 marketing plan stated:

Consumers have been educated to buy ground coffee on price. Ground coffee shares are dramatically responsive to changes in price levels vs. competition. (eX 20-Z38; see also CX 193B.

In sum, there is no evidence that indicates either RMH, Folger, or any other brand could sustain retail prices above the major branded competition without incurring "dramatic" share (53) losses in any area of the country and in the nation as a whole. (See also IDF 302-329.)64 To the contrary, retail price premiums appear to have had the same debilitating impact regardless oflength of a brand' s establishment or level of success in an area. No major brand could increase price without losing substantial retail sales.6s Nowhere does the evidence support a finding that any brand held or was threatening to gain the kind of market power that gives rise to antitrust concerns."6 In sum, the extreme consumer sensitivity to coffee prices caused a strong tendency towards retail price parity among the leading brands in every area, with diversions from such parity . occasioning sharp swings in relative sales. We cannot find evidence of a competitive breakdown at the consumer level. The evidence at the wholesale level suggesting a positive correlation (54) between prices, profits, and sales shares does not, therefore, demonstrate either Maxwell House monopoly power or the probability of achieving it in ground roast '" Thi much has largely been conceded by complaint courmel. (SeeCPF 10- 13. , Thu, while there may be a loyal consumer following for each major brand, the record suggcst2 that these dedicated consumers arc not the Diles determining the prevailing retail price ofcQfTee, A major portion of each brand' s franchisc apparently compriscs conswners ready to take advamagc of a relatively mioor discount in another brand; and the loyalty uf even the dedicated COnsumersef1ms less than absolute.See Bark supra note 19 at 312-13; Nutter The Plateau Demflnd Curve, and Utility Theory, J. Pol. Econ. 525 (1955), As we noted earlier the law does not demand "perfect" price competition, The abilty of some roasters to compete on the basis of quality or reputation, and charge a higher price for their efforts, does not signify unhealthy market power.Tv the contrary, such competition meets a demand and will persi t only aD long as il results are valued by consumers. United States v. E. I Du Pont de Nem(Jurs Co. 351 U,S, 377, 391 (1956) Opinion 103 F.

coffee at the wholesale level. The source of that correlation must lie elsewhere.

C. Efficiency and Profits Crucial to the competitive motivations of respondent and its rivals is the existence of larger gross profits generated by the leading regular ground coffee brand in an area. This phenomenon derives from the frequent use of regular ground coffee as a loss"leader by the grocery trade to generate store traffc. Except in periods of extreme price escalation or of price controls, coffee is valued very highly by grocers as a loss-leader, because it results in substantial store tramc. (CX 191B, 193B, 205A; Toy, Tr. 2027; Salesman, Tr. 5937-9; Epstein, Tr. 12187.) Indeed, nationally, as much as 50 percent of coHee is purchased on a reduced price basis as a result of store features, manufacturer s coupons, or some other promotion. (CX 191B.) The record also contains many examples ofthe disproportionate use of the leading coffee brand as a loss leader. Prior to the July 1972 Brazilian frost, respondent' s documents state, for example, that in the eastern area RMH was the only coffee brand priced at retail below retailer s cost. (CX 69M, 109H, ll1C.) Before Folger s February 1973 entry into the Maxwell House Philadelphia sales district, RMH was being sold substantially below wholesale cost throughout the eastern complex (55J-including the Boston and New York City districts as well as Philadelphia.67 (CX 89B.) This was also the case in part ofthe Syracuse district prior to Folger s entry in October 1974, where RMH had been the only brand that grocers were selling below cost. (CX 539C.

In the western area, the roles were reversed, but the story was the same. RMH was considerably less successful there, and Folger s was the leading brand. Maxwell House found it necessary to offer larger incentives to retailers than Folger s did (i. lower wholesale prices) in order to achieve retail pricing parity. (CX 3-Z33; CX 8Z. In short, the proximate source of superior profits for leading regular coHee brands in an area was their exceptional ability to generate store tratIc when sold at a discount from normal retailer profit margins.6s Our concern with this phenomenon is whether it ultimately results from output-enhancing or output-restricting behavior. The record shows that in the regular ground coHee industry, differences in the degree of promotion and quality (or its consistency) among brands explain the greater wilingness of merchants to feature the leading brands in an area and to pay more for that opportunity. 67 Even after Folger s entry in Philadelphia, R:MII was sold at lower retailer margins than Folger s. (CX !06P 52!A 532F.

68 Put another way, h ;Jdjng brands carry a lower cost of distribution than their less successful cumpetitors. That lower cost is equal to the value retailers attach to the trOiffc-generating ability of the leading brand. 204 Opinion Consumers respond not only to changes (56J in relative prices, but also to changes in the intensity of advertising, the kind of product promotion and the level of perceived quality or taste. Like price changes, the mix of marketing elements appears to have an extremely important influence on sales. For example, one of complaint counsel's retailer witnesses testified that at the time of Folger expansion, RMH was a more valuable item to him than either Hills or Chase & Sanborn because of its "greater customer acceptance, which he attributed to Maxwell House s superior "quality and promotion against the ultimate consumer." (Engel, Tr. 1714-16.) Quality and taste depend upon a number offactors, including the kind of bean and the grinding and roasting processes used.

In part, at least, one leading brand suffered from what suppliers and retailers regarded as an inconsistent and confused marketing effort. (IDF 271.) There is also substantial evidence of quality problems for a number of other roasters. As early as 1968, Maxwell House s documents indicated that the quality of one of its major competitors was "highly variable" and (57J "poor compared to Maxwell House, Folger s and Chase & Sanborn " attributing this result to a periodic mixing-in of cheaper, poorer quality coffee bean blends. (RX 973- ) One of complaint counsel's retailer witnesses confirmed the existence of a serious consistency problem for that product in the early 1970's-a problem so severe as to preclude continued success in the Cleveland area. (IDF 271, 272; Engel, Tr. 1698-1700.) Similar quality problems troubled other major brands. (Engel, Tr. 1713; Graham Tr. 10570.

Although there is no evidence of any problems with the quality element of Folger s marketing methods, there is evidence of an ineffcient initial choice among its advertising and promotion policies. Both Folger s and Maxwell House attributed Folger s superior results in the Syracuse district, its third district target, in part, to a major shift in its marketing strategy (from sampling and retailer coupons to direct mail consumer coupons and straight cash allowance incentives to retailers.) (CX 540A, D, 71OA , F; see also, Metzger, Tr. 11 870-79. In short, the record indicates that the leadership position of the RMH brand in the eastern sales areas stemmed in large part from respondent' s superior effciency in advertising, promoting and maintaining the quality ofthat brand in those areas. Through these efforts coffee consumers gained valuable information from Maxwell House 69 Six montns after Folger s 1971 Cltweland area entry, a 1972 Maxwell House document indicated: The only major competitor in the mild segment (now in the EastJ is Hills Bl'thcrs. There are several indications thjs brand is in trouble, and is highly lJulnerableto competitive introduction. . (IJt has evolved into primarily a price brand, with no soJidly established consumer franchise. With an inferior product (loses significantly to RMII) and no consumer image (six campaignsin past several years) Hills will be &'verely hurt by the first major entry into the 'mild' segment. (CX 109 F , emphasis in original; seealso Laster, Tr. 7076- 364 l"EDERAL TRADE COMMISSION DECISIONS Opinion 103 F.

promotions and attention to quality. Retailers capitalized on consumer acceptance of the brand by using it as an advertising tool to build store (58) traffc. For these features, retailers were wiling to pay a premium to Maxwell House (or other leading brands). But it was not passed on to the coffee consumer. Nor was it safe from capture by other rivals. Hence, we cannot find respondent' s attempt to maintain or enhance its sales position to be anticompetitive by effect or design. D. Conclusion We agree with the AW that the evidence in this record indicates the relevant markets are, at the very least, larger than those alleged by complaint counsel. In the face of the substantial entry shown in the record, neither price differences, transportation costs, nor different levels of product acceptance by consumers among any of respondent's sales districts or portions thereof are suffcient to establish the requisite entry barriers for a finding that they are separate markets. We need not, however, reach the question of whether the record supports the United States as the relevant market.

We agree with the AW that complaint counsel have not shown respondent to possess any ability or prospect thereof to control price or exclude competition from any ofthe alleged markets. Respondent' sales, as well as those of its competitors, are simply too responsive to changes in price and product promotions to support an inference of monopoly power. We also conclude that at least a major source of the superior profitability ofthe more successful firms in the alleged markets has been their superior effciency in marketing regular ground coffee and in giving retailers and consumers what they want. (59) For all these reasons, we must find that Maxwell House did not come dangerously close to gaining monopoly power as a result of any of its challenged conduct in any ofthe alleged markets. Quite to the contrary, its actions were output-enhancing and procompetitive-the kind of conduct the antitrust laws seek to promote. Therefore, we dismiss complaint counsel's charge of attempted monopolization under Section 2 of the Sherman Act.

V. Unfair Methods of Competition Charge Complaint counsel contend that, even if Maxwell House did not violate Section 2 of the Sherman Act, the company s responses to Folger s eastern expansion were unfair methods of competition under Section 5 of the Federal Trade Commission Act. (CAB 46-49. By this argument, pricing below cost and employing a "fighting brand" are anticompetitive tactics that can injure competition when employed by a firm with substantial market power. Section 5, according to complaint counsel, was intended to be prophylactic in its efiect , 204 Opinion and to prevent antitrust violations in their incipiency; therefore under Section 5, the degree of market power need not reach the Sherman Act threshold to establish a violation. In short, we are asked to expand the reach of the prohibition against attempted monopolization in the Sherman Act by condemning less offensive conduct under the purview of the Federal Trade Commission Act. It is true that the broad language of Section 5 of the Federal Trade Commission Act permits the Commission to supplement (60) the more specific terms of the antitrust laws.7o Exactly how far that authority extends, however, is an issue the Commission should treat cautiously. While Section 5 may empower the Commission to pursue those activities which offend the "basic policies" ofthe antitrust laws 71 we do not believe that power should be used to reshape those policies when they have been clearly expressed and circumscribed. Senator Cummins, a principal sponsor of the Act, explained the words unfair competition " to his colleagues as follows:

It wil be the duty ufthe Commission to apply those words in the sense precisely as it is now the duty ofthe court to apply the words "undue restraint oftrarle" in the sense in which we commonly understand that phrase. 51 Congo Rec. 13048 (1914). The record in this case does not offer a rationale for using the Federal Trade Commission Act to graft an extension onto Section 2 of the Sherman Act. We have dealt at length with evidence of the charge that Maxwell House ilegally attempted to monopolize the regular ground roast coffee market. Complaint counsel refer to the same evidence to support their claim of an unfairness violation of Section 5. We believe our earlier analysis applies with equal force now. (61) The qualitative differences among the products and the consumers loyalty to certain brands in the market for ground roast coffee do not indicate the kind of market power the antitrust laws were intended to address. The failure of Folger s to reach the local sales volumes of RMH is not an injury to the competitive process, but a result of healthy competition. Complaint counsel's assertion that Maxwell House managed to postpone Folger s introductions into eastern markets is not supported by the evidence. Finally, we do not think that the use of a product design or advertising theme which bears a resemb1ance to that of a competitor, without more, rises to a violation of Section 5. There is no evidence of consumer confusion or deception under any accepted legal standard that could form the basis for a finding of unfair competition.

10 FTC u. Sperry Hutchinson Co. 405 FS. 233 (1972); FTC v. Brown Shoe Co 384 U.S. 316, 320-21 (1966). 11 Id. 381 C.S. at 321 Opinion 103 F.

What the record does reveal is an episode of intense competition that allowed consumers to purchase the kinds of coffee they prefer at attractive prices. There is no evidence that this advantage came with the risk offuture price premiums through a monopolist's recoupment of lost profits. Accordingly, we reject complaint counsel's argument that Maxwell House violated Section 5 of the Federal Trade Commission Act. As the Commission observed in United Fruit Co. et al. Certainly, the logic of complaint counsel's argument would require companies concerned with competitive threats in the marketplace to sit back and wait until the threat becomes a reality before taking action. This is surely not the aim and purpose of our competitive marketplace nor the role of competition which the Commission was created to promote. (82 F. C. at 163,) (62J The proscription against attempted monopolization in Section 2 of the Sherman Act does not require a showing of monopoly power or injury to competition-a dangerous probability is suffcient. We do not believe this standard should be changed when a case is brought under Section 5. To distinguish between an attempt to monopolize and an incipient attempt on the basis of potential market power is to engage in such fine distinctions as to challenge the legal philosopher let alone the competitor trying to conform its conduct to the law. If the conduct at issue here cannot reach the early threshold of doubt under the Sherman Act, we wil not condemn it under the Federal Trade Commission Act.

VI. Price Discrimination Charge Complaint counsel argue that Maxwell House violated Section 2(a) of the Robinson-Patman Act by countering Folger s eastern moves with discriminatory prices. (CAB 49-56.) Specifically, confining RMH coupons and performance allowances to selected areas is alleged to have resulted in price differences that are prohibited by the Act. The ALJ concluded that incentives and coupons for consumers to redeem are not elements of the wholesale price of coffee. Accordingly he dismissed the Robinson-Patman count for failure of proof. (63) We need not examine the record for support of discriminatory prices, however defined, for we have already found a dispositive failure of the evidence on other grounds.

Section 2(a) ofthe Robinson-Patman Act prohibits price discrimination when its effect "may be substantially to lessen competition or tend to create a monopoly in any line of commerce, or to injure destroy, or prevent competition with any person. . . ." (15 V. C. 13(a). The requirement that price discrimination pose potential harm to 72 Becaus" we based our dcdRion on the finding ofprohability of success, it i unnecessry to determine whet!",r Section 5 would require t.he same proof of int.p.nt as does the Sherman Act. , 204 Opinion competition has been upheld consistently by the courts. As the Ninth Circuit recently stated (S)ection 2(a) does not prohibit mere price discrimination. . .. (Plaintiff must prove that the) price discrimination produced a requisite effect on competition."73 The Supreme Court has defined this effect as "a reasonable possibility that a price difference may harm competition.

The requirement of proving competitive effects has recently provided for a reconciliation between Section 2(a) of the Robinson-Patman Act and Section 2 of the Sherman Act in cases involving attempted monopolization. The same facts that reveal a dangerous probability of successful monopolization wil indicate a lessening of competition or a tendency to create a monopoly. Since both statutes are directed towards the same goal-the (64) protection of competition-it follows that the inquiries under each should be the same. We agree with the Ninth Circuit that:

In primary-line Robinson-Patman Act cases, slich as this one, the distinction between vigorous, but honest, price competition and predatory assaults on the competitive process is just as important as it is to Sherman Act cases brought under its Section 2. Under these circumstances the analytical standards should be no different.5 We have just conducted an extensive analysis of the competitive effects of General Foods' activities. We were unable to find any prospect of injury to competition from the events described in this record. Indeed, our reading ofthe evidence indicates that competition in General Foods' market is healthy and virtually invulnerable to the assaults of anyone firm. The principal effect of the heated rivalry between Folger and Maxwell House has been to reduce the prices consumers pay for coffee and promote product quality. Competition in this industry has thrived, as both firms, and the other producers have battled for the customers' franchise. 76 While there may be a lower threshold of competitive harm under the "reasonable possibili- " standard ofthe Robinson-Patman Act than under the "dangerous probability" (65) standard of the Sherman Act, it is clear that the evidence in this record satisfies neither one. Healthy competition does 73 William Inglis Sons Raking Co. v. ltt Continental Rukin/: Co. 668 F.2d 1014, 1040 (9th Cir., 1981), cerl. denied S. 103 S,Ct 58 (1982) rCitation omittedl 7. Falls City Indu.stries, Inc. v. Vanco Bevemge, Inc 103 S.Ct 1282, 1288 (1983) (CitingCorn Products Refining Co. u. F''C324 U.8. 726, 742 (1945)) 7S Wiliam Inglis Sans Baking CD. v. 117' Continental Baking COo 668 F.2d 1014 , 1042 (9th Cir. 1981),cert. denied US. 103 S.Ct. 58 (1982); Accord, D.E. Rogers Associates, Inc. v. Gardner D n"er Co. 718 F.2d 1431 (6th Cir. 19R:1) 70 The fact that sume finns have gained ground and some have lost in the battle does not determine injury lo competition. "Certinly the mere fact that (plaintifiJ suffered losses and eventually ceased operations isnot uffcieDt to establish a section 2(a) Robinson-Patman violation. William Inglis Sons Baking Ca. u- lit Continental Baking Co., supra at 1042; Accord, Anheuser-Busch, lnc u. F1'C, 289 F-2d 835 (7th Cir. 196f) (other competitors losing market share does not demonstrate injury) , Concurring Opinion 103 F. not violate the Robinson-Patman Act.?7 CONCURRING OPINION OF COMMISSIONER MICHAEL PERTSCHUK I concur in the Commission s decision to dismiss the complaint in this matter. On this record I cannot conclude that respondent General Foods engaged in predatory pricing or other conduct in violation of the antitrust laws.

The essential element of predatory pricing that is missing in this case is the "dangerous probability of success. E.I Dupont de Nemours Co. 96 F. C. 653, 725 (1980). While I conclude that General Foods possessed suffcient market power! in at least some geographic markets to engage in effective predatory pricing, I am unable to conclude, on the basis of the facts of this case, that General Foods did engage in ilegal predatory pricing.

A necessary element of predatory pricing is the likely effect of a pricing policy "that somehow restricts competition by driving out existing rivals or by excluding potential rivals from the market. Joskow and Klevorick A Framework For Analyzing Predatory Pricing Policy, 89 Yale L.J. 213, 219 (1979). It would theoretically be possible for a large, well-funded competitor to price below some measure of its costs for the (2J period of time necessary to drive a smaller competitor out of the market or to deter one from entering the market. The larger company would then be free to raise prices and gain market share. This scenario is only possible, however, where the larger competitor has a reasonable expectation of driving a smaller competitor out ofthe market or discouraging new entrants in the market. In other situations, pricing below some measure of cost will only result in aggressive price competition. Predatory pricing, like aggressive price competition, benefits consumers in the short run. The difference between them is that in the long run consumers are the losers when predatory pricing occurs.

The situation the Commission faces in General Foods has more in common with aggressive competition than predatory pricing. Even though General Foods priced below some measure of costs and had market power to influence prices in geographic markets considerably smaller than the entire U. 2 it still did not have the ability to succeed 77 Because the requi ite effect on competition has not hefm shown in this case, wC need not decide the questions surrounding the proofof dj criminatory prices- We should record, however, our reservations over the suffciency of that evidence.

1 General Foods' market share in its sales districts over the periud 1971-77 ranged from a low uf34.9% to a high of 61.4% , averaging nearly 50%. IDF 278. 1 have no trouble in concluding that silch market power could result in illegal predatory pricing in other circumstances I disagree with the ALJ' s conclusion that the relevant geographie market is the entire United States. TIle ALJ dismissed substantial evidence of General Foods' supra-competitive profits (fDB' 330-333) in areas significantly sIUOIU'-r than the entire UB. as web as substantial barriers to entry sucb as advertising cost and comp..tition for scarce shelf space. Furthermore, there is substantia! evidence that General Foods' pricing and marketing policies OIssume theit distin.ct geogrOlphic market. much smaller thOin the entire US. exist (IDF 28:J302) !:. , ); ), , 201 Concurring Opinion at monopolization when dealing with a competitor as formidable as Procter & Gamble. As the Administrative Law Judge concluded is simply inconceivable, given Procter & Gamble s marketing expertise and its financial resources. . . (that General Foods J could succeed in (3) monopolizing the sale of regular coffee in the (relevant J sales districts." Initial Decision at 149-50. See also IDF 231-32. I believe this conclusion is required, not because the activities of General Foods could not constitute predatory pricing, but because the record does not demonstrate a dangerous probability that General Foods could succeed at driving a determined and well-funded competitor such as Procter & Gamble out of the market. Nor is there any credible evidence that General Foods' pricing activities created a dangerous probability of driving any smaller competitors out of the market or deterring their entry.

While much ofthe text ofthe majority s opinion is mere dicta, I am compelled to register my disagreement on several important points. First, the majority s opinion suggests that pricing below Heast" is an ambiguous signal " (Maj. Op. at 21) and that "prices below some measures of variable cost do not by themselves mean the seller has violated the antitrust laws." (Maj. Op. at 22). But prices at that level are highly suspect if not per se predatory.3 More importantly, howev- , prices above some measure of average variable cost can also be predatory.' (4) Courts are increasingly examining the effect of pricing in the market in which it occurs rather than merely adhering to strict cost/price accounting principles to determine whether pricing is predatory.5 Some of the factors that might lead the Commission to conclude that illegal predatory pricing exists even when pricing is above average variable cost include the existence of significant barriers to entry, the use of limit pricing, the existence of excess capacity in the market, the use of non-price predation, and other competitive factors.6 The Commission and the Courts should retain the flexibility See e. , Northeastern Td Co. v. AT&T, 651 F.2d 76, 88 (2d Cir. 1981), errt. denied 455 VB. 943 (1982);William Inglis Sons Baking Co. v. IITContincntal Baking Co. 668 F.2d 10\4, 1036 (9th Cir. 1981), cerl. denied 103 S.Ct. 58(1982).

. See gcncraUy, BrodJey and Hay, "Predatory Pricing; Competing Economic Theories and the Evolution of Legal Standards," 66 Cornell L. Rev. 738 (l981); JORkow and KJevorick A Framework for Analyzing Predatory Pricing Policy, " 89 Yale 1. J. 213 (1979) 5 See Intematirmul Air Industries Inc. u. American R:xnlsior Co. 517 F.2d 714, (5th Cir 1975),eert. denied, 424 S, 943 (1976) Superturf Inc. v. Monsanto Co. 660 F.2d !275 (8th Cir. 1981);Transamerica Computer CO. U. IBM Corp. 698 F.2d 1377, (9th Cir. 1983); Chillicothe Sand Gruvel CO. V. Martin Marietta Corp. 615 F.2d 427, (7th Gir. 1980);Pacific Engineering Prod CO. U. Kerr-McGee Corp. 551 F,2d 790 (10th Gir. cert. denied 134 U.s. 879 (1977), California Compllter Produ.cts V. IBM 6131".2d 737 (9th Gir. 1979);ILC Peripherals Leasinf! Corp. U. IBM 1458 F.Supp. 423 (N.D, Gal 1978) affd. 636 F.2d 1188 (9th Cir. 1981) See Tmn. arnerica ComputerCr'" Inc. v. IBM Corp. 698 F.2d 1377 1387 (9th Cir. cert denied104 Ct. 370 (1983) prices exceeding average total cost might nevertheles.s be predatory in some circumstances" including limit pricing and temporary reductions by II monopolist to a level "above average total cost but below the profitmaximizing price whenever a new entrant appears ready to enter the market. " IEC Peripherals LeMing Corp. V. IBM, supra (barriers to entry should be considered in addition to price/cost analysis);William Inglis Sons Baking Co. v. lit Continental Buking Cu., Illc., supra 668 F.2d at 1035 (standard for detennning illegality of predatory pricing not based on "rigid adherence to a particular cost-based rule , Concurring Opinion 103 F. to condemn predatory pricing as ilegal where an examination of a firm s pricing and other actions in the competitive environment demonstrate the requisite predatory intent, market power, and dangerous (5) probability of successful monopolization even if prices are above the level of average variable cost. This accords with the judicial trend to employ a more flexible test than that proposed by Professor Areeda and Turner, a test which has the distinct disadvantage of holding dominant firm pricing per se legal."

Second, I disagree with the implication in the majority opinion that the standard of liability for attempted monopolization under the Sherman Act is the same as the standard for price discrimination under Section 2(a) of the Robinson-Patman Act. (Maj. Op. at 62-65). Under Section 2(a) of the Robinson-Patman Act the Commission need only prove that the effect of price discrimination "may be substantially to lessen competition." This should not be confused with the need to demonstrate a dangerous probability of successful monopolization to prove predatory pricing under the Sherman Act. Similarly, while the majority pays lip service to the notion that Section 5 of the FTC Act has a broader reach than the Sherman Act (Maj. Op. at 59-62), I think it is important to state unequivocally that Section 5 gives the Commission important powers to prohibit conduct that does not fall squarely within either the Sherman Act or the Robinson-Patman Act. For example, the use of strategic, (6) below cost pricing which harmed competition by significantly delaying or permanently deterring entry of potential competitors could constitute an unfair method of competition even if General Foods was unlikely to achieve a monopoly position.

CONCURRING OPINION OF COMMISSIONER PATRICIA P. BAILEY I concur in the dismissal of this case, but perhaps on narrower grounds than the majority and certainly without endorsing much of the overbroad and unnecessary theory which is introduced into the opinion.

My reason for dismissing the case is simply a demonstrated lack of anti competitive effects. In all the proposed submarkets General Foods lost market share by the end of the alleged periods of predation. Moreover, no competing roasters were forced to halt operations or were driven from the market; some even expanded business. As to the appropriate geographic market definition, I agree that the price data are not clear enough to support complaint counsel's chosen 7 Brodley and Hay. supra 66 Cornell L. Rev. at 793. See William Inglis Sorts Baking Co. v. lit Cuntinental Baking Ca., supra 668 F.2d at 1042 See e.g., FTC v. Brown Shoe Co., 384 U.S. 316 (1966); Grand Union u. FTC: 300 F.2d 92 (2d Cir. 1960),See also HandlcT and Stever Attempts to Monopolize and No-Fault Mooopoljzation 129 l'a. 1. Rev. 125 , 177-180 (1980). g., , 204 Concurring Opinion submarkets of General Foods' sales districts. Moreover, evidence on significant levels of transshipments tends to disprove those specific markets. On the other hand, I would emphasize that the transshipment evidence, standing alone, could not establish anything close to a national coffee market. For one thing, evidence on product differentiation as an entry barrier and market-definer is a good deal stronger than the opinion acknowledges. All the major coffee wholesalers promotional strategies are tailored to meet the differing demand characteristics of some local area. (IDFs 38-7) The trouble is, the borders ofa market can be defined by grocery retail and distributional patterns, by television broadcast limits, or by (2) coupon redemption limitations. (e. IDFs 69- , 82 , 88-89) No one ofthese variables stands out as the consistent and direct equivalent of General Foods sales districts; nor is it possible, on this record, to use these factors to draw more accurate local markets. This does not mean, however, that product differentiation can never be an entry barrier ! merely that in this instance there has been an inexact showing of its scope and effect. I must therefore disassociate myself from that part of the opinion which so strongly suggests that the term "entry barriers" can never be applied to any cost faced by both current market participants and potential entrants. This "Stiglerian" definition is flawed by failure to recognize that firms already in the market often spend far less to get the same result as the entrant, whether that result be scale economies, know-how, promotional and distributional effciencies, or consumer acceptance. Established firms build on previous investments in market knowledge whereas their new rivals face significant information costs. See, e. Demsetz Barriers to Entry, 72 Amer. Econ. Rev. , 50 (1982); Schmalensee On the Use of Economic Models in Antitrust: The Realemon Case, 127 U. ofPa. L. Rev. 994, 1021 (1979). Thus in this case, over the last twenty years, before Procter & Gamble there were only two major entrants into General Foods' eastern sales districts, and (3) much evidence linking their success to substantial promotional efforts. Procter & Gamble s own experience shows how successful entry turns on stimulating local consumer demand, and that such efforts are both diffcult and expensive. The nation s second largest coffee producer found it neither quick nor easy to erode General Foods' dominance of the east coast and any submarkets therein. I applaud the time-saving analytical approach taken in the opinion with its emphasis on doing no more than is necessary to resolve the case. I regret, therefore, that having mapped out such an effcient course, the opinion then immediately strays into a wholly unneces- See, e. Caves and Porter From Entry Barriers to Mobility Barriers: Conjectural Decisions and Contrived Deterrence to ?\ew Competition 91 J. Ecan. 241 (1977); Spence, "Entry, Capacity, Investment, and Oligopolistic Pricing," 8 Bell J. Ecall.534 (1977) Concurring Opinion 103 F. sary discussion of the correct cost standard for predation. I will reserve my full comments on this subject until they are relevant to a Commission decision.2 I wil note briefly, however, that a majority of the Commission apparently feels that pricing above average variable cost is protected by a strong presumption oflegality whereas pricing below average variable cost produces a fairly weak presumption of ilegality. I believe the relative weights of the presumptions should be reversed.

I believe the opinion does not suffciently emphasize that there is a lower threshold for competitive harm under the Robinson-Patman Act than under the Sherman Act. Although the Sherman Act and Robinson-Patman Act march together on proof of( 4) anticompetitive conduct and intent from below cost pricing, they part company when it comes to proof of injury to competition. The offense of attempted monopolization requires proof of a dangerous probability of success while Section 2(a) of the Robinson-Patman Act requires only a showing that a price discrimination "may" substantially lessen competition. By phrasing the prohibition prophylactically, Congress intended to prevent the results of full scale anticompetitive behavior by catching price discrimination in its incipiency and preventing its growth. FTC v. Morton Salt Co. 334 U.S. 37, 43-47 (1948). Moreover, while Section 2 of the Sherman Act requires attempted monopolization of a "part" of commerce, and therefore looks to general competitive conditions in the line of commerce affected, Section 2(a) ofthe Robinson-Patman Act requires only an impermissible effect upon competition among the price-discriminating seller competitors or customers. William Inglis Sons Baking Co. v. ITT Continental Baking Co. 668 F.2d 1014, 1042 (9th Cir. 1981, as amended on denial of rehearing and rehearing en bane 1982), cert. denied 103 S.Ct. 58 (1982). Taken together, I interpret these two provisions ofthe Robinson-Patman statute to allow somewhat more latitude for a determination of competitive injury that does the Sherman Act. This does not mean that the injury standard is inevitably satisfied by mere diversion of business from a competitor, or (5) even by one competitor s exit from the market due to alleged predatory pricing. On the other hand the standard can be met by a showing of below-cost pricing which has severely weakened a class of competitors and augmented the predator s market power, albeit not to the very high levels which may be needed to prove a dangerous probability of success. In this case, however, I agree that injury to competition-in any quantum-has notbeen2 I have discussed predation previously,shown.in my February 28 1983, dissnting statement of the fimd order in Borden, Inc. Docket Nu. 8978- f18 FR 9023 at 9030 (March 3, 1983)1 See generally, Marasco, Tracing an Antitru t Injury in Secondary Line Price Di (Tjminatjon Cases, 50 r' ordham L. Rev. 909, 911-18 (1982) 204 Final Order FINAL ORDER This matter has been heard by the Commission upon the appeal of complaint counsel from the initial decision and upon 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 sustain the initial decision.

It is ordered That the complaint is dismissed. Complaint 103 F.