Kellogg Company

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

IN THE MATTER OF KELLOGG COMPANY, ET AL.

DISMISSAL ORDER , ETC., IN REGARD TO ALLEGED VIOLATION OF SEC. 5 OF THE FEDERAL TRADE COMMISSION ACT Docket 8883. Complaint, April 1972-Dismi.c;sal Order, Jan. 15. 1982 This order vacates in its entirety, the September 1 , 1981 Initial Decision and dismisses with prejudice the Commission s April 26, 1972 complaint which charged three cereal manufacturers with engaging in practices having the effect of maintaining a highly concentrated noncompetitive market structure in the production and sale of ready-ta-cat cereals. Appearances For the Commission: Anthony Low Joseph, David M. Malone Lawrence B. Berman, Claudia R. Higgins, Louis R. Monacell, Alan J. Friedman and Dennis F Johnson. For the respondents: Frederick P. Furth, Thomas R. Fahrner Bruce J. Wecker, Charles P. Wolff, Michael P. Lehmann and George F Bishop, Furth, Fahrner, Bluemle Mason San Francisco, Calif. and Scott R. Campbell Battle Creek, Mich. , in-house counsel, for respondent Kellogg Company. David C. Murchinson, Edward F Howrey, Ralph ,J. Savarese, John R. Fornaciari, Margaret M. Zwisler and Frank P. Spinella, Howrey Simon Washington, D.C. and Clifford L. Whitehill, Robert J. Fulgency and John F Finn Minneapolis, Minn. , in-house counsel, for respondent General Mills. Robert MacCrate, James E. Akers, Jeffrey I. Zuckerman, Richard J. Rawson and David Lesser, Sullivan Cromwell Washington, D. John F Kovin, Clifford Warnke Washington, D. C. and Robert Y. Fox, Peter J. Deluca and Bruce L. Bozeman White Plains, N. , in-house counsel, for respondent General Foods Corporation. Charles Orlove and Joseph M. Jacobs, Jacobs, Burns, Sugarman Orlove, Chicago Ill., for intervenor American Federation of Grain Millers, AFL/CIO. COMPLAINT The Federal Trade Commission has reason to believe that the party respondents named in the caption hereof, and hereinafter more particularly designated and described, have violated and are now violating the provisions of Section 5 of the Federal Trade Commission Act (Title 15, V. C. 45). Accordingly, the Commission Complaint hereby issues this Complaint stating its charges with respect thereto as follows:

1. Respondents have been and are now engaged in, among other business activities, the manufacture and sale of ready-to-eat (RTE) cereals. RTE cereals are food products made from barley, corn, oats rice or wheat and various combinations of such grains which are flaked, granulated, puffed, shredded or processed in other ways. RTE cereals are eaten primarily as a breakfast food requiring no cooking or heating preparation by the consumer. (2) All of the respondents have been engaged in the cereal business for over 40 years, and in the RTE cereal business for over 30 years. Since 1950 respondents have consistently accounted for over 84 percent of the sales of RTE cereals.

2. A. Respondent Kellogg Company (Kellogg) was founded in 1906. It is a corporation organized and doing business under the laws of the State of Delaware, with its principal office and place business located at 235 Porter St., Battle Creek, Michigan. Kellogg manufactures and sells, among other things, RTE cereals, tea, soup, gelatin, and pudding.

In 1970 Kellogg had assets of $347 million and sales of $614 millon. In 1970 Kellogg ranked 191st in sales among the nation 500 largest industrial corporations.

In 1969 Kellogg s domestic sales of RTE cereals were $300 million and advertising expenditures for RTE cereals were over $36 million. Kellogg is the largest producer of RTE cereals in the United States. B. Respondent General Mills, Inc. (General Mills) was incorporated in 1928. It is a corporation organized and doing business under the laws of the State of Delaware with its principal office and place of business located at 9200 Wayzata Boulevard, Minneapolis, Minnesota. General Mills manufactures and sells, among other things, RTE cereals, flour, toys, chemicals, clothes, and jewelry. In 1970 General Mills had assets over $665 million, and sales were over $1 bilion. In 1970 General Mills ranked 116th in sales among the nation s 500 largest industrial corporations. In 1970, General Mills' domestic RTE cereal sales amounted to $141 million and advertising expenditures for RTE cereals were $19 million. General Mills is the second largest producer of Rfe cereals in the United States. (3) C. Respondent General Foods Corporation (General Foods) was incorporated in 1922. It is a corporation organized and doing business under the laws of the State of Delaware with its principal office and place of business located at 250 North St., White Plains Complaint 99 F.

New York. As the nation s largest food manufacturer, General Foods produces and sells, among other things, RTE cereals, coffee, beverages, frozen food, pet foods, and desserts. In 1970 the total assets of General Foods were over $1.3 bilion and sales were over $2 bilion. In 1970 General Foods ranked 45th in sales among the nation s 500 largest industrial corporations. In 1970, General Foods' domestic sales of RTE cereals were over $92 milion and advertising expenditures for RTE cereals were over $9 millon. General Foods is the third largest producer of RTE cereals in the United States.

D. Respondent The Quaker Oats Company (Quaker) was incorporated in 1901. It is a corporation organized and doing business under the laws of the State of New Jersey with its principal offce and place of business located at Merchandise Mart Plaza, Chicago, Ilinois. Quaker manufacturers and sells, among other things, RTE cereals frozen food, cookies, pet foods, and chemicals. In 1970, Quaker had assets over $391 million and sales of $597 milion. In 1970 Quaker ranked 195th in sales among the nation 500 largest industrial corporations.

In 1970 Quaker s domestic sales of RTE cereal were $56 million. Approximately $9 milion was spent in 1970 to advertise Quaker RTE cereals. Quaker is the fourth largest producer of RTE cereals in the United States.

E. Nabisco, Inc. (Nabisco) is not a respondent herein. It has however, participated in some of the acts and practices alleged herein and has contributed by acquiescence to the noncompetitive structure of the RTE cereal market, as alleged herein. Nabisco was incorporated in 1898. It is a corporation organized and doing business under the laws of the State of New Jersey with its principal office and place of business located at 425 Park Ave., New York, New York. Nabisco manufacturers and sells, among other things, RTE cereals, cookies, candy, and snack foods. (4) In 1970 Nabisco s total assets were over $503 million and sales were over $868 million. In 1970 Nabisco ranked 140th in sales among the nation s 500 largest ine .1strial corporations. Nabisco s domestic sales of RTE cereals were $26 million in 1969 and advertising expenditures for RTE cereals were $3 milion. Nabisco is the fifth largest producer of RTE cereal in the United States.

F. Ralston Purina Company (Ralston) is not a respondent herein. It has, however, participated in some of the acts and practices alleged herein and has contributed by acquiescence to the noncompetitive structure of the RTE cereal market, as alleged herein. Complaint Ralston was incorporated in 1894. It is a corporation organized and doing business under the laws of the State of Missouri with its principal office and place of business located at Checkerboard Square, St. Louis, Missouri. Ralston manufactures and sells, among other things, RTE cereals, pet foods, animal feed, snack foods, and frozen food.

In 1970, Ralston s total assets were over $775 milion and sales were over $1.5 bilion. In 1970 Ralston ranked 71st in sales among the nation s 500 largest industrial corporations. In 1969 Ralston s domestic RTE cereal sales were over $20 million and advertising expenditures were over $4 million. Ralston is the sixth largest producer ofRTE cereal in the United States. 3. In the course and conduct of their business, respondents now ship, and for some time past have shipped, their RTE cereals from their respective production facilities in various States to locations in various other States of the United States, and maintain and at all times mentioned herein have maintained, a substantial course of trade in RTE cereals in commerce, as ucommerce" is defined in the Federal Trade Commission Act.

4. Each of the respondents is in substantial competition with each and all of the other respondents and with other cereal producers in the manufacture and sale of RTE cereals in interstate commerce, except to the extent that competition has been hindered, lessened and eliminated as hereinafter set forth. (5) 5. During the past 30 years the RTE cereal industry has experienced substantial growth. In 1940, 453 million pounds of RTE cereal were produced; 900 milion pounds were produced in 1960; and in 1970 over 1 bilion pounds of RTE cereal were produced. The value of RTE cereal increased from $163 millon in 1950 to over $650 million in 1970.

In 1940 respondents' sales accounted for approximately 68 percent of the RTE cereal market; in 1950, for 84 percent, and in 1970, for 90 percent. In 1969 respondents controlled the following approximate shares of the RTE cereal market: Kellogg, 45 percent; General Mills 21 percent; General Foods, 16 percent; and Quaker, 9 percent. In 1969 Nabisco and Ralston each had an approximate share of four percent of the RTE cereal market.

6. For at least the past 30 years, and continuing to the present respondents, and each of them, have engaged in acts or have practiced forbearance with respect to the acts of other respondents the effect of which has been to maintain a highly concentrated noncompetitive market structure in the production and sale of RTE cereal.

Complaint 99 F.

During this period respondents, in maintaining the aforesaid market structure, have been, and are now engaged in, among others the following acts and practices:

Brand Proliferation, Product Differentiation and Trademark Promotion Respondents have introduced to the market a profusion of RTE cereal brands. During the period 1950 through 1970 approximately 150 brands, mostly trademarked, were marketed by respondents. Over half of these brands were introduced after 1960. In introducing and promoting these new brands respondents have employed intensive advertising directed particularly to children. Respondents have used advertising to promote trademarks that conceal the true nature ofthe product.

Respondents artificially differentiate their RTE cereals. Respondents produce basically similar RTE cereals, and then emphasize and exaggerate trivial variations such as color and shape. Respondents employ trademarks to conceal such basic similarities and to differentiate cereal brands. Respondents also use premiums to induce purchases of RTE cereals. (6) Respondents have steadily increased the level of advertising expenditures for RTE cereals. During the period 1950 through 1970 respondents' aggregate annual advertising expenditures for RTE cereals tripled from $26 million to $81 million. In 1970, respondents advertising to sales ratio for RTE cereals averaged 13 percent. These practices of proliferating brands, differentiating similar products and promoting trademarks through intensive advertising result in high barriers to entry into the RTE cereal market. Unfair Methods of Competition in Advertising and Product Promotion 0) By means of statements and representations contained in their advertisements, respondents:

In advertisements aimed at children, represent directly or by implication, ttat their RTE cereals without any other foods enable children to perform the physical activities represented or implied in their advertisements.

In truth and in fact:

Respondents' RTE cereals do not enable children to perform the physical activities represented or implied in their advertisements. A Complaint child' s ability to perform such physical activities depends on many other factors, including but not limited to general body build exercise, rest, a balanced diet and age.

(2) By means of statements and representations contained in their advertisements respondents Kellogg, General Mils, and General Foods represent, directly or by implication, that consuming R'IE cereal at breakfast:

(a) Will result in loss of body weight without vigorous adherence to a reduced calorie diet (b) Will result in maintenance of present body weight even if total caloric intake increases, or (c) Will result in loss or maintenance of body weight without adherence to regular physical exercise. (7) In truth and in fact:

(a) Consuming RTE cereal at breakfast will not result in loss of body weight without vigorous adherence to a reduced calorie diet. (b) Consuming R'IE cereal at breakfast will not result in maintenance of body weight even if total caloric intake increases. (c) Consuming UTE cereal at breakfast will not result in loss or maintenance of body weight without adherence to regular physical exercise.

(3) By means of statements and representations contained in their advertisements respondent General Mils and Kellogg: (a) Represent, directly or by implication, that failure to eat one of their RTE cereals results in the failure of athletes or others to perform to their full capabilities.

(b) Represent, directly or by implication, that the ingestion of one of their RTE cereals by athletes or others enables them to perform better in their repsective activities.

In truth and in fact:

(a) Failure to eat one of the R'IE cereals of such respondents will not result in the failure of athletes or others to perform to their full capabijities.

(b) The ingestion of one of the UTE cereals of such respondents wil not enable athletes or others to perform better in their respective activities.

(4) The use of respondents of the aforesaid unfair methods of competition in advertising and product promotion has the capacity and tendency to mislead consumers, particularly children, into the mistaken belief that respondents' UTE cereals are different from Complaint 99 F.

other RTE cereals, thereby facilitating artificial differentiation and brand proliferation. These unfair methods of competition have contributed to and enhanced respondents' (8)ability to obtain and maintain monopoly prices and to exclude competitors from the manufacture and sale of RTE cereal.

Control of Shelf Space Kellogg is the principal supplier of shelf space services for the RTE cereal sections of retail grocery outlets. Such services include the selection, placement and removal of RTE cereals and allocation of shelf space for RTE cereals to each respondent and to other RTE cereal producers.

Through such services respondents have interfered with and now interfere with the marketing efforts of other producers of RTE and other breakfast cereals and producers of other breakfast foods. Through such services respondents restrict the shelf positions and the number of facings for Nabisco and Ralston RTE cereals, and remove the RTE cereals of small regional producers. All respondents acquiesce in and benefit from the Kellogg shelf space program which protects and perpetuates their respective market shares through the removal or controlled exposure of other breakfast food products including, but not limited to, RTE cereal products.

D. Acquisition of Competitors During the past. 70 years numerous acquisitions have occurred in the breakfast cereal industry. One of the effects of these acquisitions was the elimination of significant sources of private label RTE cereal. Among them are the following.

In 1943, General Foods acquired Jersey Cereal Company, a Pennsylvania corporation. Before acquisition by General Foods, Jersey Cereal Company was a substantial competitor in the sale of private label and other RTE cereal.

In 1943, Kellogg leased and controlled the manufacturing facilities of Miler Cereal Company, Omaha, Nebraska, a substantial competitor in the sale of private label and other RTE cereal. In 1958, upon termination of the said leasing agreement, Kellogg purchased the assets of Miller. (9) In 1946, General Foods acquired the RTE manufacturing facilities of Campbell Cereal Company, Minneapolis, Minnesota, a substantial competitor in the sale of RTE cereal. Following this acquisition Complaint General Foods dismantled the RTE facilities of Campbell and shipped said facilities to South Africa.

The aforesaid acquisitions have enhanced the shared monopoly structure of the RTE cereal industry.

7. Respondents, and each of them, have exercised monopoly power in the RTE cereal market by engaging in the following price and sales promotion practices, among others: (a) Refrained from challenging each other s decisions to increase prices for RTE cereals, and, in general, acquiesced in or followed the price increases of each of them;

(b) Restricted the use of trade deals and trade-directed promotions for RTE cereals;

(c) Limited the use of consumer-directed promotions for RTE cereals, such as coupons, cents-off deals, and premiums. 8. Respondents' acts and practices aforesaid have had the foHowing effects, among others:

(a) Respondents have, individually and collectively, established and maintained artificially inflated prices for RTE cereals. (b) Respondents have obtained profits and returns on investment substantially in excess of those that they would have obtained in a competitively structured market.

(c) Product innovation has been largely supplanted by product imitation.

(d) Actual and potential competition in the manufacture and sale of RTE cereals has been hindered, lessened, eliminated and foreclosed. (10) (e) Significant entry in the RTE cereal market has been blockaded for over thirty years.

(f) Meaningful price competition does not exist in the RTE cereal market.

(g) American consumers have been forced to pay substantially higher prices for RTE cereals than they would have had to pay in a competitively structured market.

Through the aforesaid acts and practices: (a) Respondents individually and in combination have maintained and now maintain, a highly concentrated, noncompetitive market structure in the production and sale of RTE cereal, in violation of Section 5 of the Federal Trade Commission Act. (b) Respondents, individually and collectively, have obtained, Initial Decision 99 F. shared and exercised, and now share and exercise, monopoly power , and have monopolized, the production and sale of RTE cereal, in violation of Section 5 of the Federal Trade Commission Act. (c) Respondents, and each of them, have erected, maintained and raised barriers to entry to the RTE cereal market through unfair methods of competition, in violation of Section 5 of the Federal Trade Commission Act.

INITIAL DECISION BY ALVIN L. BERMAN, ADMINISTRATIVE LAW JUDGE SEPTEMBER 1 , 1981 PRELIMINARY STATEMENT The Commission s complaint, issued in April 1972, charged Kellogg Company ("Kellogg ), General Mills, Inc. ("General Mils General Foods Corporation ("General Foods ), and The Quaker Oats Company ("Quaker ) with violating the provisions of Section 5 of the Federal Trade Commission Act (15 V. C. 45). Respondents' violations were alleged to have been in connection with their manufacture and sale of ready-to-eat ("RTE") cereals described as "food products made from barley, corn, oats, rice or wheat and various combinations of such grains which are flaked granulated, puffed, shredded or processed in other ways. RTE cereals are eaten primarily as a breakfast food requiring no cooking or heating preparation by the consumer" (Par. 1). Respondents were charged (Par. 6) with engaging in acts of, or with practicing, uforebearance with respect to the acts of other respondents, the effect of which has been to maintain a highly concentrated, noncompetitive market structure in the production and sale of RTE cereal." The acts and practices charged included brand proliferation, product differentiation and trademark promotion, backed by intensive and steadily increasing levels of advertising (Par. 6A); and control of shelf space to the disadvantage of competitors by acquiescing in a Kellogg shelf space program (Par. 6C). The complaint alleged further, as follows: 7. Respondents, and each of them, have exercised monopoly power in the RTE cereal market by engaging in the following price and sales promotion practices among others:

(a) Refrained from challenging each other s decisions to increase prices for RTE cereals, and, in general, acquiesced in or followed the price increases of each of them; (b) Restricted the use of trade deals and trade-directed promotions for RTE cereals; Initial Decision (c) Limited the use of consumer-directed promotions for RTE cereals, (2Jsuch as coupons, cents-off deals, and premiums. I 8. Respondents' acts and practices aforesaid have had the following effects, among others:

(a) Respondents have, individually and collectively, established and maintained artificially inflated prices for RTE cereals. (b) Respondents have obtained profits and returns on investment substantially in excess of those that they would have obtained in a competitively structured market. (c) Product innovation has been largely supplanted by product imitation. (d) Actual and potential competition in the manufacture and sale of RTE cereals has been hindered, lessened, eliminated and foreclosed. (e) Significant entry in the RTE cereal market has been blockaded for over thirty years.

(f Meaningful price competition does not exist in the RTE cereal market. (g) American consumers have been forced to pay substantially higher prices for RTE cereals than they would have had to pay in a competitively structured market. Through the aforesaid acts and practices: (a) Respondents individually and in combination have maintained, and (3Jnow maintain, a highly concentrated, noncompetitive market structure in the production and sale of RTE cereal, in violation of Section 5 of the Federal Trade Commission Act. (b) Respondents, individually and collectively, have obtained, shared and exercised and now share and exercise, monopoly power in, and have monopolized. the production and sale of RTE cereal, in violation of Section 5 of the Federal Trade Commission Act.

(c) Respondents, and each of them, have erected, maintained and raised barriers to entry to the RTE cereal market through unfair methods of competition, in violation of Section 5 of the Federal Trade Commission Act. Nabisco, Inc. ("Nabisco ) and Ralston Purina Company ("Ralston ), neither of which was named as a respondent, were alleged to have participated in some of the acts and practices alleged in the complaint and, by their acquiescence, to have contributed to the noncompetitive structure of the RTE cereal market (Par. 2E F). The complaint further alleged (Pars. 3-4) that the respondents were engaged in commerce in connection with their trade in, and manufacture and sale of, RTE cereals, as "commerce" is defined in the Federal Trade Commission Act.

On June 29 1972, respondents filed their answers to the complaint in which they admitted engaging in commerce, but otherwise denied in substance, the allegations of the complaint. By order of November 16, 1979, the Commission granted the amended motion of the American Federation of Grain Milers AFL- , Th" complaint also contained ctJrtain alltJgations of false and deceptive advertising (Par. 6BJ. These aller.ationsare not being pursued hy complaint counsel ilt thistim " , , Initial Decision 99 F. CIO/CLC for leave to intervene for the purpose of presenting evidence relative to the relief proposed by complaint counsel. It is significant that the complaint does not charge respondents with having conspired to monopolize. The words conspire contract" or " agree" (or variants thereof) are nowhere to be found in the complaint. While respondents' acts and practices are alleged (Pars. 8(a), 9(a), (b)), individually and collectively and in combination, to have brought about certain results, no conspiratorial acts or practices are alleged. In light of the "shared monopoly" theory that is being tested by this case and the relative ease of drafting conspiracy charges when it is desired to do so, it can only concluded that the complaint intentionally did not include the charge of conspiracy. (4) This reading of the complaint was expressly confirmed by the position taken by complaint counsel early in these proceedings. In Reply Of Complaint Counsel To Motions By General Mils And General Foods For More Definite Statement, filed May 18, 1972, we find The complaint is quite clear as to the nature of the collective charge against respondents. It does not aver conspiracy. It is simply an indictment of shared monopoly and the common course of action pursued by respondents to maintain their monopoly (emphasis supplied; at 1-2). This position was affirmed in Opposition Of Complaint Counsel '10 General Mills ' Application For A Determination By The Hearing Examiner That His Ruling Denying Motion For More Definite Stateme';t Involves Reviewable Questions: The original motion, which was denied by the Hearing Examiner, dealt with only five areas of the complaint and asked only one question: Should the complaint be read to charge respondents with having conspired or with having engaged in consciously parallel action in violation of Section 5 of the Federal Trade Commission Act? Une part of this question was, in fael, answered clearly and unequivocally by complaint counsel, i. , conspiracy is not alleged (emphasis supplied; at 3). In their Memorandum In Support Of Opposition To General Foods Motion For Severance, dated May 18, 1972, complaint counsel explained that the "complaint charges all four respondents with engaging in certain interdependent acts and practices in order to achieve a highly concentrated, non-competitive market structure and shared monopoly power" (at 1). After noting that competitors are routinely named co-respondents in conspiracy cases, complaint counsel expressly stated A /though conspiracy is not alleged in this matter the common course of action and the interdependent acts of respondents create a common bond that provides the nexus for joinder in the instant case" (emphasis supplied; at 3). Thus, , , , Initial Decision complaint counsel were justifying the joinder of respondents in a case that did not charge conspiracy.

In Supplemental Memorandum of Complaint Counsel In Opposition To Severance, filed June 19, 1972, at 4, it was stated Although (5)conspiracy is not alleged in this matter there is a common bond that provides the nexus for joinder in the instant case. . . . The effects of respondents' common practices are actually the same as if they had engaged in a conspiracy" (emphasis supplied). Thus complaint counsel were clarifying the complaint to the effect that it charged acts and practices to be unlawful because of their anticompetitive effects, but that a charge of conspiracy was not being made. Complaint counsel confirmed the "no-conspiracy" aspect of the complaint at the very first prehearing conference held on June 5 1972.

MR. LIEDQUIST, First, there is no mention of conspiracy in the complaint, for that matter we have emphasized this in our reply to respondent' s motion. I have already made a statement there is no charge of conspiracy under the complaint. I think this is suffcient.

HEARING EXAMINER IIINKES: In your words, when you use words "joint" interdependent" combination collective acquiesce, your use of those words or any other words that arc used in the complaint, there is no suggestion of conspiracy in the complaint, is that correct? MR. LIEDQUIST: There is no conspiracy as you would normally plead it in an Anti- Trust matter under the Sherman Act or the FTC. HEARING EXAMINER HINKES: Are you saying that when you used the words of joint" and "combination, it was joint without a conspiracy, and a combination without a conspiracy. (6) MR. LIEDQUIST: That is right, Your Honor, and I am saying that they did not meet together (Tr, 17 25-26 29).3 (7) , This position was affirmed in complaint counsel's Reply To The Quaker Oats Company s Motion For Severance, filed June 19 1972.

, Tr. is an abbreviation for Transcript of Proceeding, and is followed by the page numberls). Other abbreviations used herein include the following CP - f'Almplaint Counsel' s Propo Findings of Fact, Conclusions of Law, Order, and Supporting Argument CPF - Complaint Counsel's Propo Finding in CP, followed by its numher(s) (Continued) Initial Decision 99 F. The position that conspiracy was not alleged was reaffirmed at Tr. 70-71 (8) On August 10, 1972, Mr. Liedquist reaffirmed and restated complaint counsel' s position in part, as follows: I said there is no conspiracy. I believe my words also pointed out that there is no conspiracy in the traditional sense of the word. We don t look upon it w; generally KI' - Kellogg s Proposed Findi,,!:;; of !"ad KPF - Kellogg s Proposed finding in KI', followed by ib number(SI KPL - Kellogg s Proposed nmclusions of Law GMP - General Mills' Proposed Findings of Fact, Conclusions uf Law and Order GMPF - General Mills' Proposed VimJing in G;.P, rollowed by iL number(sJ GFP - General o"ds' Pmposed Findings of Fact, Cu"cI",;;ons of Law, Memorandum and Order GFPF - G,'neral !"oods'Proposed Finding in G , followed by its numberls) (I' - Intervenor s Proposed Findings of act, Conclusions of Law, and Supporting Argument IPF Jntervt'nnr I'ropo ed finding in JP, rollow"d by its numberls) CR - Complaint Counsel's Reply CRPF - Complaint coun "l's I'ropo ed Finding in (;1( followed by its numberls) KS - Kellngg SLHreply GMS - Genera! Mills' Surreply GMSPF General Mills' I'mpo ed Finding in GMS, followeri by its number(s) GFS - General Foo ' Surreply SPF - General Foos' Proposed I"finding in GFS , followed by iL, numberlsJ IS - Intervenor s Surreply ex - Commission Exhibit ex K Commission Exhibit Seured From Kellogg CX-GF - Commission Exhibit Secured From G..ner,,1 oods CX-GM Commission ;xbibit Secured From General Mills CX-Q - Commission Exhibit Secured Fmm Quaker CX-R - Commission io;xhibit Secured rom Ralston eX- - Commission Exhibit Secured From Nabisco CX-CI - Commiso;ion ;xhibit Seured From Cere,,1 Institute ex-pc; - Commission Exhibit .secured rom Procter & Gamble CX-ACN - Commission Exhibit. Secured From A.C Nielson Co CX- - Commission Exhibit. Secured I, rom National Commission On Food Marketing - Kellogg :xhibit.

GFX - General Foos Exhibit GMX - General Mills Exhibit COURTX - Court Exhibit QX - Quaker Exhibit AFX - American Federation of Grain Millers Exhibi! Exhibits are referred tohy t.he abbreviations set forth above r"no wed by the appropriate exhibit numberisJ and if applicable, p"gc( J- ;xhibit. abbreviations are not repeated within st.ring citations, nDr are exhibit numbers repeated when more than one page is referred to. Initial Decision recof!nized as a conspiracy. They haven t met behind closed doors. At the same time their behavior hasn t been coincidental behavior. There has been a pattern of behavior, a common course of action that has been followed for thirty years and which amounts to a combination in violation of Section 5 (emphasis supplied; Tr. 104). Notwithstanding the absence of a charge of conspiracy in the complaint and complaint counsel's early concession to this effect, the case was tried under both a conspiracy and a shared monopoly theory. At the very outset of their Introduction And Summary to their proposed findings of fact (CPF 1-3), complaint counsel state In the most traditional antitrust sense, the three respondents have tacitly colluded and cooperated to maintain and exercise monopoly power- power over price' and ' power to exclude' additional competitors." And, at CP 649, complaint counsel begin their legal argument section on conspiracy with a "TACIT CONSPIRACY" tab and the heading, "RESPONDENTS COMBINED AND CONSPIRED TO MONOPOLIZE THE READY-TO-EAT CEREAL INDUSTRY. However, complaint counsel may not be heard to urge that a tacit conspiracy was not included in their previous disclaimers of conspiracy. A tacit conspiracy is a conspiracy normally pleaded under the Sherman and Federal Trade Commission Acts. It is still a conspiracy and all essential elements of conspiracy must be proved notwithstanding the fact that the conspiracy may be shown by evidence (9) other than that of an express overt agreement. See, e.g., United States v. Bausch Lomb Optical Co. 321 U.s. 707, 723 (1944); American Tobacco Co. v. United States 328 U.S. 781, 809-10 (1946). It would serve no purpose to attempt to trace the steps taken by complaint counsel in reversing their original position and construing the complaint to include a charge of conspiracy. Suffice it to say that on February 24, 1974, in an Order Denying Motion Of General Mills Inc., For Summary Decision Dismissing Complaint, Judge Hinkes who was then assigned to this matter, ruled, in part, as follows: Moreover, the complaint does allege that lhe respondents individually and in combination have mainlained a noncompetitive market structure and that individually and collectively have obtained, shared and exercised monopoly power, although complaint counsel concede that their behavior is "not a conspiracy in the traditional sense.

General Mills contends, however, that these allegations of the complaint amount to no more than conscious parallelism which, as noted earlier, is not recognized as within the meaning of the Sherman Act. General Mills docs not address itself to whether or not conscious parallelism is within the meaning of Section 5 of the Federal Trade Commission Act, an Act which has been recognized as going beyond the narrow confines of the Sherman Act. Moreover, General Mils discounts the complaint' clarification expressed by complaint counsel. They explained that respondents engaged in "interdependent actions and decisions" which they defined as "ones taken with the knowledge that the action requires cooperation of each member of a group in , Initial Decision 99 F. order to minimize competition among the group and maximize the joint profits of the group s members.

General Mills equates this definition of the complaint with merely conscious parallelism. It may, however, be more nearly equated with a tacit conspiracy (emphasis in original; at 5).

On March 12, 1975, in an Order On Complaint Counsel's Motion Re Discovery, Judge Hinkes allowed depositions covering the issue of conspiracy or similar agreement, relying in part on his February 24 1974 statement, reproduced above. Re also stated I agree with (lO) complaint counsel that the interrelationship among the respondents by whatever name is a relevant issue and therefore appropriate for discovery " (at 5).

Finally, on August 20 1976, Judge Rinkes, relying on his orders of February 19, 1974 and March 12, 1975, issued an Order Denying Respondent Kellogg s Motion To Preclude Trial Of Complaint Counsel's Conspiracy Claim. Judge Rinkes concluded by stating, The interrelationship among respondents by whatever name is an issue raised by this complaint and by respondents' answers thereto (at 2).

Under the shared monopoly charge of the complaint, which relies upon the concentrated structure of the RTE cereal industry and the conduct of the individual respondents allegedly taken in recognition of their resulting mutual interdependence, evidence of respondents conduct is admissible. Such evidence would not become inadmissible merely because it tended to show an actual conspiracy or agreement express or tacit.

On the other hand, as I have already ruled, the complaint does not allege a conspiracy; and a violation of Section 5 of the Federal Trade Commission Act by reason of conspiracy is not encompassed in the shared monopoly charges that have been made. If the shared monopoly charges fail, by reason of either legal or factual insufficiency, respondents may not, under the instant complaint, be found to have violated Section 5 by reason of conspiracy. To the extent that Judge Rinkes may be deemed to have amended the complaint to include the charge of conspiracy, such amendment would violate Section 3. 15(a)(1) of the Commission s Rules. That section requires the filing of a motion for amendment and limits the authority of the ALJ to allow only such amendments that are reasonably within the scope of the original complaint. Neither requirement has been met here. Further, Judge Rinkes has issued no order of amendment.

The issue of conspiracy may not be considered to have been raised under the concept of conformance to the evidence (Section 3. 15(a)(2)). Initial Decision This, because the issue was not within the scope of the complaint and the issue was not tried by express or implied consent of the parties, two requirements of Section 3.15(a)(2). To the contrary, complaint counsel's efforts to try the conspiracy issue have been contested by respondents every step of the way. Notwithstanding my holdings with regard to conspiracy, in view of the fact that this case was allowed to be tried under a conspiracy theory and in light of the unusually long time it has taken to try this case and the voluminous record that has been compiled, I am going to make all findings called for by the evidence, including those relating to the issue of conspiracy.

The case was initially assigned to Administrative Law Judge Rarry R. Rinkes for trial. Judge Rinkes was the ALJ during the (11) course of all of the pretrial, and he presided during the presentation of complaint counsel's entire case- in-chief. Both the pretrial and presentation of the case-in-chief were extensive. Prehearing conferences are reported at Tr. 1-2750. Formal hearings commenced on April 28, 1976, at Tr. 2753. Complaint counsel rested their case-inchief on January 11, 1978, at Tr. 28 975. At the close of complaint counsel's case-in-chief, all respondents filed motions to dismiss the complaint and to strike certain portions of the case-in-chief. Quaker s motion that the complaint be dismissed as to it was granted on February 24, 1978. Judge Rinkes deferred consideration of the other motions to dismiss until the conclusion of the entire case. By order of February 8, 1979, 1 denied requests of General Foods and General Mills for reconsideration of their motions to dismiss.

Judge Rinkes denied the motions to strike, by order of March 18 1978, ruling that they amounted to, and should be considered together with, the motions to strike, thus effectively postponing the consideration of any such motions unti the final arguments of the parties at the close of the record. By order of February 8 , 1979, I denied General Foods' request for reconsideration of its motion to strike. Respondents, therefore, have been allowed to again raise at this time the motions to strike previously denied by Judge Rinkes. After the consideration of various matters on the record (Tr. 976-29 228), respondents began their defense on April 25, 1978, at Tr. 29 229, with Judge Rinkes stil presiding. On September 7, 1978 Judge Hinkes announced his impending retirement effective the end of that week and his intention to continue presiding in this matter under a special contract (Tr. 34 821-22). Rearings were continued on September 8, 1978 (Tr. 34 942- 043), while Judge Rinkes was still an administrative law judge. Commencing September 12, 1978, and , Initial Decision 99 FTC. continuing through September 28, 1978 (Tr. 35 044- 984), Judge Hinkes heard seven defense witnesses at a time he was no longer administrative law judge. On October 12, 1978, the next date scheduled for hearings, counsel for Kellogg, in light of the pendency of a motion to determine that Judge Hinkes was disqualified from presiding over the hearing, declined to produce Kellogg s next witness (Tr. 35 986-35 990).

On December 11 , 1978, following the Commission s determination on December 8, 1978, that Judge Hinkes became "unavailable within the meaning of 5 VB. C. 554(d) upon his retirement, I was appointed substitute administrative law judge to preside over further proceedings.

On August 9, 1979, complaint counsel and counsel for General Mills and General Foods stipulated that the seven witnesses who testified before Judge Hinkes during the period September 9, 1978 through September 28, 1978 (i. after Judge Hinkes had retired) if recalled, would testify under oath identically as they (12)had already testified, and that their testimony already given be accepted as part of the record. Documents offered and received in evidence during the period in question were also stipulated as accepted as part of the record. Kellogg refused to become a party to the stipulation. Accordingly, on August 27, 1978, I ordered that the stipulation and the testimony and exhibits stipulated to be accepted as part of the record applicable, however, only among the stipulating parties. (13) , Th"testimonysostipu!lItedintotherecordcu""rswitm,sses- Richard H. Trox..l Clr. 35 044-35.273) Rolf 0 HcrgrrL3fi 274-35. 521 Arthur R Schulze (Tr, 35 353- 657) Edward K Bixby (Tr. 35 658- 7181 Robert M. Cameron ('(r. 35 719-- 779J Frank C. Blodgett (Tr. 35.780 35 874) Johns. Livingstone (Tr. 35 875-35 9851 The following exhibits were so stipulaled intuth"r"c:urd GMX 106 (CX-GM 173) GMX 1121CX-GM 21R6i GMX 116ICX-GM 21901 GMX 119 (CX-GM 219::) GMX 124 (CX-GM 2-17-1) GMX 130 (CX-GM 56\) GMX 14!J (CX GM 1291 GMX 158 GMX169 GMX170 GMX171 GMX 172 GMX 173 GMX310-320 GMX;n:J GMX463 GMX463A IC"nti"uedJ Initial Decision When the Commission, on December 8, 1978, ruled that Judge Hinkes was unavailable and directed the appointment of a substitute ALJ, it also ordered the parties to file with the substitute ALJ briefs pertaining to the necessity and desirability of recalling and rehearing witnesses who had previously testified before Judge Hinkes. Respondents uniformly took the position that all witnesses should be reheard.

On February 21 , 1979, complaint counsel filed a statement in support of their position that it was unnecessary to recall any witnesses. In that statement, complaint counsel "submit(tedJ that there are no significant issues in this proceeding which turn crucially upon demeanor" (p. 2). It was stated that prior to Judge Hinkes' retirement on September 9 , 1978, complaint counsel and respondents had presented 64 "fact witnesses" and that five more fact witnesses" were presented by General Mills after Judge Hinkes' retirement (p. 9). With respect to these " fact witnesses complaint counsel took the following position: Furthermore, thirty-six of complaint counsel's fifty-six fact witnesses were employees or former employees of respondents, as were the 13 fact witnesses called by respondents. If anyone is to suggest that the testimony of these fact witnesses should be taken at less than face value, it should be complaint counsel, toward whom most of these witnesses stood as agents of an adverse party. Nevertheless, complaint counsel are willing to rely upon the record evidence as a basis for any judgments of credibility that need to be made. If complaint counsel are willing to rely upon the record with respect to the credibility of these adverse witnesses, it is illogical for respondents toward whom the witnesses were friendly, to insist they be recalled (pp. 10-11). (14) On March 12, 1979, pursuant to my order of February 22, 1979 complaint counsel identified the "fact witnesses" referred to in their February 21, 1979 submittal and the pages of transcript covered by their testimony. ' (15) GMX469 GMX471 GMX472 GMX501 GMX510-512 GMX515--S16 GMX533 GMX 540 GMX54&-551 GMX 380 (CX-GM 2503A-Z23J ex 2205 ex 22()fi ex 2211 , The witness"", and pages oftest.imony so identjfied are as follows Witness Transcript Pages Dr. James R. Green 28R9-3221 Eugene M. Sehlenk 3222- 3339 Dr. James R. Green 3340-3482 (Co"tinued) Initial Decision 99 F. In my order of May 24, 1979, Denying (Respondents ) Motions For Hearing De Novo, I relied upon complaint counsel's concession and stated that HI shall accept the testimony of the (16)fact witnesses at Jack W. Emry 3483-3622 Lawrence W. Corzioe 3623-3770 Dr. ,Jarnc R. Green 3771- 3835 Virginia ve Laird 3836--3871 Betty Jean Dunham 3872-3894 Dr. Alan A. Fisher 3948-5824 Dr. James R. Green 5825-:10 Robert Calvin Bland 6462-7404 Joseph W. Pedersen 7405-7610 William E. Gentry 7611-85 William II Baumann 806181 Biron A. ValieT 8182-8296 Thomas N - Bezick 8297-8328 Henry T. Chandler 8329-t593 Ralph Hocella 85948654 Ben C. !r-in 8655-- 8791 Herman L. Stroup 87928871 Ralph Maron 88728930 Frank J. Kupka 8931-9071 Walter Steven Rubow 9072-9254 Richard F. Hurst 9255-9327 Richard W. Maurer 9328-9440 OJ. !ngraham 10682-10778 Jack W. Emry 10779- 11056 Kenndh C Englert 11219- 11594 Jerry D. Wells 11595-11744 Howard M. List 11745-11919 Robert E Hutson 1l920-12122 Robert L. Nichols 12123- 12743 Robert T. Bland 12744-- 12978 Charles A. Tornabene 12979- 13238 Adolph S. Clausi 1:1:19::- 13684 Wilfred II. IIaur;hey 1:1685-1:84:, EUfienc Raymond Moh!ie 13844-14162 F. Kent Mitchel 14163-14334 Alvin OS5ip 14:J35-14552 L.rry D. Weiss 14553-1460 Max Randall Gould 14601-14884 Charles A. Hinman 14885-14938 John J. McBride 14939-15086 Kenneth Masn 15087- 15248 James E. Legere 15249-15:J44 Robert A. Bowen 15345-15585 Harry E. Nickelson 15586-15651 Richard S. Sheehy 15652-15691 Vernon W. Cafarella 15692--16209 Joseph P. Manfrida 1621016393 Vernon W. Cafarella 16394-- 16551 Robert F. Boeau 16552- 168:J6 Vernon W. Cafarella 16838-17145 Arthur R. Schul7.e 1714f)-17428 John Richard Schneider 17429-17495 Donald S Schnitz 17496-17578 Arthur G. Yates 17579-17614 Arthur R. Schulre 17615-17842 Guy Lalone, Jr 17843-17977 Preston Townley 17978-18038 Bernard J. Hogan 18039- 18077 Alfred Boberg 18078- 18125 (Continued) Initial Decision face value and give it weight according to the overall testimony of the witnesses and the record as a whole" (p. 23). As for the expert witnesses who testified prior to my presiding, I have evaluated their testimony on the basis of the reasonableness and logic of their theories and economic conclusions, as I have done with respect to the experts who testified in my presence.

In so doing, I am in agreement with Kellogg s concluding witness economic expert Dr. Robert Clower, who, when asked how the ALJ could decide which economic model or theory espoused in this case to accept, stated (at Tr. 40 175):

I think this record contains an incredible amount of information for anyone who is seriously interested in comparing the kind of description of basic economic theory that is contained in the record at all levels, with the actual facts, and arriving, at least, at an informed judgment about which model makes the most sensc. In so appraising the expert evidence, I have found no need to evaluate the credibility of any expert witness who did not appear before me in the course of reaching the findings and conclusions that I make. Of course, to the extent any expert witnesses rely upon facts in reaching their conclusions or in giving opinions, which facts have not been established on the record, such conclusions or opinions are being afforded lesser or no weight.

This initial decision is based upon the entire record, including proposed findings of fact and conclusions of law and supporting memoranda filed by the parties as well as their answers, replies and Richard H. Troxel 18125-18421 Dianne H. l' llison 18422-18691 Owen B. Butler 2579:1-2.5887 Toby Ira Schreiber 25888-26070 Rudolf William Hirzel 29229-295:J7 Howard List 29.537-29757 William E. LaMothe 29758-:10132 Richard R. Walters 30133-30294 David E. Kinnisten 30295-30458 Howard L. Ross 32667-32779 Wiliam McKown 32780--33272 Richard B. Troxel 3336833546 Todd S. Johnoon 3354733945 Richard B. Troxel 33946-34208 Michael J. Stevens 34209- 34307 Boyd Sneddon 3430834672 Rolf Q. Berg 35274-35352 Arthur R. Schul 35353-35657 F.dward K. Bixby :J5658-35718 Hobert M. Cameron :J5719-:!5779 Frank C- Blodgett 3578(1-35874 Initial Decision 99 F. surreplies. The undersigned has also taken into account his observation of the witnesses who appeared before him and their demeanor. (l7J Proposed findings not herein adopted, either in the form submitted or in substance, are rejected either as not supported by the evidence or as involving immaterial matters. In light of the 41 000 pages of transcript, 35 bound volumes of public exhibits, 16 binders of camera exhibits and the extensive findings of fact, conclusions of law and supporting memoranda filed by the parties it is literally impossible to expressly and separately address each item of evidence and contention. Nevertheless, because of the salience and importance of this case, I am making factual findings in addition to those upon which I rely for disposition of the case so they may be available to a reviewing authority which may fect they arc important in resolving any issues.

ADMISSIBILITY OF DOCUMENTS PRODUCED FROM THE FILES OF RESPONDENTS Before making any findings, it is necessary to consider respondents' uniform position that documents produced from the files of one respondent should be stricken as against all other respondents. During the course of the trial, Judge Rinkes established the rule that documents produced from the files of any respondent, which were not forgeries, would be admitted into evidence conditionally against all respondents, subject to establishing a connection with those respondents other than the one from whose files the documents were produced. On February 22, 1978, following the close of complaint counsel's case- in-chief, each respondent moved to have stricken as to it all documents procured from the files of other respondents or from the files of any other entity. General Foods identified over 2 000 documents (totalling some 000 pages) which it sought to have stricken. These documents came from the files of Kellogg, General Mils, Quaker, the Cereal Institute, Nabisco, the National Commission on Food Marketing, Ralston, A.C. Nielson Co., A&P and Procter and (lSJGamble. General Mills, without identifying specific documents, moved that all documents originating with the other respondents bc stricken as to it. In light of General Mils' assertion that the vast majority of the o I heard witnesses starting at Tr. 36 307 , The following were ried: CP- 769 pages; KP and KPr 971 pagcs; GMP-554 pagp.s; G P--98 P"f:"S; IP-43 pages; CI--576 f"'li"s: KS-330 pages; GMS- 83 pages; GFS-98 pages; IS-3U pagcs-- a total of 4 752 pages " Note g. my de6sion to make rmdings on the issue of conspiracy, notwithstanding my ruling that the complaint fails to charge Cl'nspiracy Initial Dccision documentary evidence originated with respondents other than General Mils, it may be assumed that General Mils' motion to strike encompassed a larger number than the 2 000 documents covered by General. Foods in its motion. The documents covered by Kellogg s motion to strike are listed on some 50 pages of its motion and include documents secured from General Mills, General Foods, Quaker and Ralston.

As related above, both Judge Hinkes and I refused to consider the motions to strike, ruling that such motions should await the final briefing. Respondents have now renewed their motions to strike. Hespondents' contentions underlying their motions to strike fall in two categories: (1) a lack of connection between the document and the moving respondents, and (2) the hearsay nature of the document inasmuch as the author was not produced and so could not be crossexamined by the respondents.

Connection between the documents and the moving respondents.

The shared monopoly theory of this case is that, given the structure of the RTE cereal industry, the actions (or conduct) of each individual respondent, considered in conjunction with the actions of other respondents and others in the industry, have served to maintain a highly concentrated, noncompetitive market structure to obtain, share and exercise monopoly power and monopolize, and to erect, maintain, and raise barriers to entry. The acts of each respondent which bear upon the allegations of the complaint therefore, are relevant in appraising the acts and practices of the other respondents. Therefore, there is no substance to the general allegation that documents produced from the files of one respondent which describe the conduct of that respondent bear no relationship or connection to the other respondents.

For example, each respondent would have stricken as to it the price lists of the other respondents, for the reason that each company s price lists renect only its own prices. While each price list shows only the pricing conduct of the issuing company, the price lists of all respondents show the aggregate pricing conduct of practically the entire industry. Under the complaint, which in large part relies upon economic theories flowing from the central theory of analyzing the structure, conduct and performance of an industry, each price list is relevant or connected to all respondents. While each respondent is to be tried on the basis of its own conduct, the theory of the Initial Decision 99 F. case requires consideration of that conduct in the light of that of the other respondents. (19) Respondents rely upon the principle applicable in conspiracy cases that there must be independent evidence of the existence of a conspiracy, and the participation of a party in the conspiracy, before declarations of an alleged co-conspirator in the course of executing or furthering the conspiracy may be admitted against another respondent. See, e.g., United States v. Nixon 418 U.S. 683, 701 (1974); United States v. Kessler 530 F.2d 1246, 1256-57 (5th Cir. 1976); Flintkote Co. v. Lysfjord 246 F.2d 368, 378 (9th Cir.), cert. denied 355 S. 835 (1957). The shared monopoly charge, however, is not a conspiracy matter as to which the above principle would apply. Respondents assert that there is no legal basis for trying the charges of the complaint; that the documents secured from one respondent, therefore, are not relevant to the others under any recognizable theory of law. However, the only means by which to ascertain whether the acts and practices and methods of competition of the several respondents constitute unfair methods of competition within the meaning of Section 5 is to admit evidence of the acts and practices and methods of competition of the individual respondents and evaluate them in their aggregate.

Even though particular admissions or declarations against interest of one respondent may not be used directly against other respondents without being connected, they may be admissible against the party from whose fies they were secured to show its individual activity or purport. This, in turn, could be considered in evaluating the environment within which the activities of all respondents are to be judged. Further, in addition to evaluating the overall situation of the industry, the acts of the individual respondents, quite apart from allegations of conspiracy or agreement would be relevant under the theory that respondents have engaged in price leadership in lieu of overt agreement and have otherwise acted in concert or in similar fashion by reason of the their interdependent coordination induced by the structure of the RTE cereal industry.

As noted above, respondents contend that the shared monopoly aspect of this case fails as a matter of law. As I have also previously noted, I believed it inappropriate to consider motions to dismiss apart from a full consideration after completion of the entire case. Similarly, irrespective of my disposition of this matter, I believe it to be important not to strike evidence relevant to complaint counsel's theory of violation so that a reviewing authority may have a full record upon which to appraise that theory. Initial Decision Returning to the conspiracy issue, I have already ruled that the complaint does not encompass the charge of conspiracy but that I shall, nevertheless, make findings on that issue. The cases relied upon by respondents, in asserting that documents taken from the fies of one respondent are inadmissible against (20)the others, deal with the necessity of proving the existence of a conspiracy by independent evidence before declarations of a conspirator may be admitted against his co-conspirators- However, in a tacit conspiracy case, there is no evidence of express, verbal agreement. Conspiracy must be established through the conduct of the several alleged coconspirators. American Tobacco Co. v. United States 328 U.S. 781 809- 10 (1946); United States v. Bausch Lomb Optical Co. 321 U.s. 707 , 723 (1944); United States v. Consolidated Packaging Corp., 575 2d 117, 126-27 (7th Cir. 1978).

The cases cited above, and many others, reveal that there are many varying considerations to be taken into account in determining whether the business activities of individual respondents, as indicated by their own business documents, may be evidence of a tacit conspiracy. Consequently, a blanket ruling covering the thousands of documents objected to by respondents is not possible. It is necessary to evaluate each document in question, to the extent relevant to a particular issue, to ascertain, for example, whether it constitutes an act or recitation, and whether it constitutes an act or activity irreconcilable with independent business judgment of the particular respondent. To the extent such identical or closely related irreconcilable acts and practices are shown as to the several respondents, a conspiracy may be evidenced. In my opinion, as explained below, complaint counsel have failed to establish that respondents conspired to monopolize or to maintain a monopoly. The relevant documents considered in reaching that conclusion are not being stricken.

2. Hearsay Nature Of The Documents Each respondent moves to strike documents produced from the other respondents' files on the ground the documents are hearsay as to it and it has not had the opportunity to cross-examine the person who prepared the document.

Section 3.43(b) of the Commission s Rules of Practice provides for the admission of relevant, material and reliable evidence. It does not exclude hearsay evidence, and hearsay evidence may be received. Philadelphia Carpet Co. 64 F. C. 762, 773 (1964); Phelps Dodge Refining Co. v. 139 F.2d 393 , 397 (2nd Cir. 1943); Hoover v. Beta Initial Decision 99 F. 467 F. 2d 516 (5th Cir.), cert. denied 409 U.s. 1086 (1972). As stated in Lenox, Inc. 73 F. C. 578, 604 (1968), "Clearly documents coming from a respondent's files can be regarded as reasonably reliable absent some countervailing evidence demonstrating their unreliability." In receiving the documents in question in evidence, the administrative law judge has already ruled upon their reliability, subject to the connection with other respondents for purposes of showing relevance and materiality, discussed above. (21) This, of course, leaves the question of the weight to be given a particular document. In deciding that question, it is necessary to consider . such matters as the circumstances surrounding the preparation of the document, its author, and the extent to which its contents are corroborated or are consistent with other evidence. Accordingly, each respondent's blanket request that all of the thousands of documents produced from the files of the other respondents be stricken is being denied. Rather, each such document considered wil be afforded the weight to which it is entitled in the course of making the particular findings that follow. COMPLAINT COUNSEL S THI'ORY OF VIOLATION Complaint counsel's theory of how respondents violated Section 5 of the Federal Trade Commission Act is summarized at CP 1-20 (CPF 50). The following recitation is derived from that summary. Consumers are being overcharged for RTE cereals because they are being deprived of the benefits of competition. Kellogg, General Mills, and General Foods have maintained a pattern of conduct that has enabled them to charge supracompetitive prices and to reap monopoly level profits. The fact each respondent gains these enormous monopoly profits as a result of their conduct demonstrates that they are sharing monopoly power.

In the most traditional antitrust sense, the three respondents have tacitly colluded and cooperated to maintain and exercise monoply power- power over price' and ' power to exclude' additional competitors. " Respondents' conduct has regulated and suppressed competition in the RTE cereal market and has caused consumer prices to be substantially higher than if this were a more competitive market.

Respondents' monopoly power will continue to injure consumers unless greater competition is encouraged. This can be done by bringing additional competitors into the market. (22) Among such documents are many that simply aid in understanding the imJustry by providing industry description or background, Theydo not purport to constitute evi dCllccofconspiracyofofintcrdept,ndenthchaviof Initial Decision The Commission should dissipate respondents' monopoly power because respondents have eliminated competition among themselves and have raised barriers to entry. Therefore, the Commission should issue an order requiring respondents to create five strong new firms from their assets; and should encourage the entry of additional firms into the industry by requiring respondents to grant royalty-free licenses of brands they continue to own after the divestiture. This would encourage competition and prices would fall toward competitive levels.

The complaint charges that the three respondents and Quaker have collectively and individually engaged in acts and practices which maintained a non-competitive, monopolized market in the sale of RTE cereals. The main issues are consumer injury and the exercise of monopoly power by respondents. RTE cereals constitute a large ready-to-eat product market with retail sales of over $740 milion in 1970 and about double that in 1975.

Complaint counsel' s proposed findings are organized according to a structure-conduct-performance framework of analysis. This describes a causal flow from an industry s structure to its behavior or conduct, and ultimately to its economic performance. The RTE cereal industry is highly concentrated and is marked by high barriers to the entry of new firms. There are no barriers to entry unrelated to respondents' conduct. Respondents' conduct therefore, explains the absence of entry by new firms which otherwise would pursue the extraordinary high profits in the market and seek a share ofthe rapid growth thereof. The RTE cereal industry is highly concentrated at the six firm four firm and three firm concentration levels. Since at least 1940 there have only been six significant producers of RTE cereals: Kellogg, General Mills, General Foods, Quaker, Nabisco and Ralston of which respondents have always been the three largest. New firms normally enter markets where profits are above the level that can be earned in alternative investments. Further, new entry is generally more attractive and easier in growing industries. The RTE cereal industry is enormously profiable. This is shown by the accounting rates of return on capital employed by the industry as a whole, as well as the accounting rates of return for Kellogg, General Mils and General Foods individually, compared to the rate of return for the manufacturing sector of American industry. This conclusion is also true when alternative calculations are made which adjust for potential biases in the accounting rates of return. (23) Initial Decision 99 F. The RTE cereal industry has grown more rapidly than the United States economy as a whole. Despite the high profitabilty and rapid growth of the RTE cereal industry, no new firm had entered the industry from 1950 to 1972. The lack of entry into the apparently attractive RTE cereal industry must be caused by barriers to entry. Neither of the traditional nonconduct harriers to entry--conomies of scale and the ownership of specialized resources such as patentsare present here. The only possible explanation for the lack of new entry lies in the respondents' conduct of introducing a large number of intensively advertised, trademarked RTE cereal brands which has had the effect of raising barriers to entry into the industry. In highly concentrated industries, where only a few firms control most of the output, each firm may recognize that its competitive actions may have direct effects on the other large firms in the industry and may cause the others to respond in predictable ways. Recognizing their interdependence, each firm recognizes that it is in their collective best interest to avoid taking competitive actions which wil result in the other firms reacting competitively. Therefore, if each firm manages to communicate or signal its decision to the others not to use certain competitive weapons and to follow the leadership of other firms, the small group of sellers may reach understandings with each other. The understandings can arise by reason of the firms' patterns of conduct without oral or written communication among them. These patterns of conduct can become long standing and become the rules of the game to which each firm wil abide.

The respondents have adopted and adhered to certain rules of the game and, thereby, have suppressed the use of a wide variety of competitive weapons. Whereas the RTE cereal firms were highly price-competitive through the early 1950' , they reached an understanding with others to eliminate some of their more competitive price activities, as well as others, which could have reduced industry prices and profits. Having avoided such competitive action, the respondents channeled their competitive energies into brand proliferation. This raised barriers to the entry of new firms which otherwise would have entered. Thus, respondents have been able to earn monopoly profits for a substantial period of time. Monopoly power is the power to hold prices above a competitive level of costs. Respondents exercised that monopoly power by coordinating their activities so as to suppress price and other forms of competition. They reached an understanding to minimize the use of trade deals, cents-off labels, and in-pack premiums. Kellogg and General Foods acquired two of the larger producers of private label Initial Decision RTE cereal products; and, thereafter, respondents reached an understanding to terminate or (24Jgreatly reduce the sale of private label products. Respondents arranged through the A.C. Nielsen Company to exchange current detailed RTE cereal advertising information whereby they were able to coordinate a reduction in advertising expenditures when sales growth slowed in the late 1960' , and were able to avoid misunderstandings regarding each other s advertising strategies. In the late 1950' , each respondent adopted a shelf space plan based on common principles. The respondents avoided engaging in fortification competition prior to 1970, at which time they came to an understanding that they would fortify all of their RTE cereal products to comparable levels at about the same time.

The respondents avoided list price competition among themselves and tacitly colluded to increase prices by adopting a system of collusive price leadership in which Kellogg was usually the price leader, and General Mils and General Foods raised their price in consonance with Kellogg so as to maintain prices at the monopoly level.

The furnishing of private label products constitutes an important form of price competition. Respondents avoided this form of price competition by refusing to provide private label RTE cereals, despite respondents' available production capacity and a substantial mand by their customers for private label products. The discontinuance of the powerful competitive tools of trade deals, in-pack premiums and cents-off labels could not have occurred without an understanding among respondents. Retail shelf location is an important factor in the success or failure of an RTE cereal brand. By adopting virtually identical shelf-space plans and by reason of General Mills' and General Foods' tacit adoption of Kellogg s control of the retail shelves, so long as Kellogg allotted each of the other two respondents a fair share of the space respondents suppressed competition among themselves and placed smaller firms and potential entrants in inferior and less desirable shelf locations.

Respondents refrained from fortifying most of their RTE cereal brands until they were spurred to do something by public criticism. Then, the respondents met at their industry association, The Cereal Institute, where they reached an understanding as to what each was going to do increase their fortification to comparable levels within a relatively short time.

Respondents suppressed competition among themselves by their exchange of current advertising data through the A.C. Nielsen Initial Decision 99 F.T. Company. Monthly or bi-monthly advertising data by brand and by advertising media were furnished to Nielsen which compiled the (25) data into a comprehensive industry report and distributed the information to each firm within two months of the actual expenditures. This allowed the respondents to coordinate their advertising efforts.

Normally, in view of respondents' monopoly profits and the industry s growth, new firms would have entered the market with competitive impact. However, there was no such entry because of respondents' conduct. By suppressing and avoiding the various forms of competition narrated above and in channeling substantially all of their efforts to increase sales into the introduction of a large number of intensively-advertised, trademarked RTE cereals ("brand proliferation ), respondents have persistently continued their monopoly profits while a new firm entering the industry would expect to operate at a loss. In this manner, respondents erected a barrier to entry.

Respondents' brand proliferation created a barrier to entry because RTE cereal brands are highly differentiated, with each brand having only a few directly competitive other brands. Thus, a price change of a particular RTE cereal brand will directly affect only a few, closely substitutable brands. This localized competition means that in order for a firm to introduce a new RTE cereal, the new brand must take a significant portion of its sales from the few brands with which it must compete most directly. In addition, a new RTE cereal brand must gain a minimum level of sales in order to be successful. This means that if there are a number of existing brands in the segment aimed at by the new product, there is not room for a new product to enter, although the existing brands can continue to make substantial profits. This situation exists in a number of the areas or segments of the RTE cereal market. Existing firms are able to introduce new brands, while new companies cannot, because of certain advantages. One advantage is that an existing firm can preempt a profitable new product opportunity before a new entrant can act. One reason is that a manufacturer of RTE cereals must have from 3.5% to 5% of the market without being at a significant cost disadvanatage relative to other firms (firm economy of scale). Thus, one of the respondents which already has much more than 3.5% of the market, can introduce an additional single brand while it is satisfying its firm economies of scale. An outsider would have to launch several brands at the anticipated level of 1 % (brand economy of scale) before it could reach a firm efficiency scale approaching 3.5%. Initial Decision Potential entrants can reasonably expect that their brand introductions would be subject to more vigorous competition from existing RTE cereal producers than would the brand introductions of respondents. A respondent could not react to a brand introduction of an existing competitor without breaking the rules of the game, and it could then expect a competitive (26)response throughout the cereal industry. On thc other hand, a respondent could react to a single brand entry of a new company in the particular localized area where the new brand was competing while its other products would continue to reap monopoly profits.

Apart from preemption or retaliation, respondents' brand proliferation has made entry into the RTE cereal industry by outsiders less likely because it has significantly increased the costs which a new entrant would have to incur. This is because brand proliferation has decreased the size or share of market enjoyed by any individual brand and has increased the rate of failure of brand introductions. Thus, the outsider could not hope to introduce a single brand which would get enough return to make it worthwhile. It would be required to introduce multiple brands (some successful and some unsuccessful) with the increased costs of having to research and develop, advertise, and promote and market a number of products, each with its own costs. This increase in capital costs would eliminate potential smaller firms which could not raise the necessary capital and would even make the larger potential competitors more cautious before they might venture an entry into the RTE cereal market. Economic performance of the RTE cereal industry is poor. One measure of the poor economic performance is respondents' high monopoly prices and profits. With prices so artificially inflated at monopoly levels, consumers purchase less than they would if prices were at lower competitive levels. Since less RTE cereal is produced fewer resources are devoted to the manufacture and distribution of RTE cereals than would be employed in a competitive market. Because of respondents' high profits, consumers have been inequitably overcharged. If prices were at lower, competitive levels manufacturers would provide RTE cereals to the consumers at lower prices at which consumers would demand, and producers would supply, more RTE cereal. This would provide additional employment.

When firms reap monopoly profits, there is an incentive to increase costs, including advertising expenditures, above those which would prevail in a competitive market. This is because the profit to a firm for selling an additional unit of output is greater than if the prices were at a lower, competitive level. Firms charging Initial Decision 99 F. higher prices, therefore, have an incentive to spend additional amounts on promotional activities in order to sell additional units of the commodities and gain higher returns. Therefore, monopolists are wiling to incur higher costs than competitive firms would be wiling to expend. Further, the channeling of rivalry into brand proliferation has required an increased amount of advertising and promotional expenditures as each new cereal is introduced. If additional price competition were restored to the industry and product (27Jproliferation were decreased, advertising and other promotional expenditures would fall toward competitive levels, as would prices. The exorbitant advertising and promotional costs have been passed on to consumers in respondents' monopoly prices.

Only substantial structural relief holds the prospect of significantly improving the performance of the industry. " Complaint counsel' s proposed order includes provisions requiring divestiture by spin-off, mandatory royalty-free trademark licensing, a ban on future acquisitions, and a ban on shelf space plans. Kellogg would be required to divest itself by spin-off of three viable firms. General Mills and General Foods would each be required to divest itself by spin-off of one viable firm. Each divested firm would be capable of producing at least 5% of industry output and would be granted exclusive rights to manufacture and distribute trademarked cereal brands formerly owned by the parent firm. This divestiture requirement, by increasing the number of firms in the industry, would substantially lessen the possibility that the firms in the industry could tacitly collude to avoid competition. The trademark licensing provision would require that respondents offer to license royalty-free for 20 years any RTE cereal brand remaining in their possession after the divestitures. Any new RTE cereal brand introduced by a respondent would become subject to the licensing requirement five years after it was introduced into national distribution.

The spin-off provision would have at least three important competitive effects. First, it would create the potential for direct competition between the remaining RTE cereal products of the respondents and the licensed versions of those brands. Second, it would lower barriers to entry for new firms, since a new firm by license could get a new product immediately. Third, the five divested firms could use this additional licensing provision to utilize the remaining portion of their productive capacities not utilzed by the products exclusively licensed to them by the parent firms. The proposed ban on acquisitions by the respondents for 20 years Initial Decision would inhibit the reconcentration of the industry following divestiture.

The ban on shelf space programs would prohibit respondents from making shelf space recommendations covering the entire RTE cereal section. Each firm could continue to make recommendations concerning the shelving of its own brands. This provision would restrain respondents from making it more difficult for new entrants and smaller producers to obtain better shelf space. (28) Monopoly power is power over price and power to exclude competitors. The existence of such power finds support in the combined market shares of the respondents. !!However, proof of monopoly power in this case rests primarily on the evidence that each of the respondents gained monopoly profits over a long period of time." Monopoly power is also demonstrated by the respondents exercise of their power to exclude further competition by reason of their brand proliferation.

Respondents are said to have violated Section 5 of the Federal Trade Commission Act because their conduct violates the Sherman Act prohibitions against conspiracies to monopolize. This is evidenced by the overall understanding among respondents to abide by certain rules of the game whereby respondents suppressed and avoided various forms of competition and channeled their activities into brand proliferation which, in turn, led to the erection of barriers to entry.

Conduct may violate Section 5, whether or not it is held to have violated the Sherman Act. Respondents' cooperative conduct is said to violate the policies of the antitrust laws and is, therefore, an unfair method of competition. Respondents have been unfair to consumers in terms of the monopoly overcharges that have resulted from their conduct, and respondents have been unfair to competitors because they have barricaded the market against additional competitive entry.

Complaint counsel have further explained their theory of the case at CP 54-1 (CPF &-1 through &-19) from which the following additional recitation has been derived.

Complaint counsel rely for their evaluation of the RTE cereal industry and the behavior of the major RTE cereal producers in accordance with the structure conduct-performance framework of economic analysis. This demonstrates that the highly concentrated structure of the RTE cereal industry enabled respondents tacitly to collude or cooperate to avoid a wide variety of competitive activities which would have eroded the monopoly profits they enjoyed. Instead respondents introduced a large number of intensively advertised Initial Decision 99 F. trademarked RTE cereal brands, with the effect of creating a brand proliferation barrier to the entry of new firms. This allowed respondents to charge monopoly prices and earn monopoly profits which are the measure of poor economic performance. The structureconduct-performance framework may also be used to evaluate complaint counsel' s proposed remedy.

The basic principle underlying the structure-conduct-performance framework is that the structure of an industry may affect the behavior or conduct of its members which, in turn, (29Jdetermines the quality of the industry s performance. There may also be feedback effects inasmuch as firm conduct may affect structure. Structural characteristics of an industry include the number of sellers and their relative sizes, concentration, and the existence of barriers to the entry of new firms. Important conduct features of an industry include pricing practices and product policies, including the kinds and numbers of products to offer, and how products should be physically differentiated, as well as advertising and research and development policies, and innovation.

The economic performance of an "industry falls into three broad categories: efficiency, equity, and progressiveness. One. Efficiency. This incl udes a determination of whether an industry manufactures and distributes its products at the lowest possible cost and whether the optimal amount of society s resources are devoted to producing the industry s products. An example would be a monopolist which restricts output in order to raise prices and increase its profits. This would result in consumers purchasing less at monopoly prices than they would if there were lower, competitive priges. With less products being produced, the monopolist employs fewer resources than under a competitive environment. Enjoying inDated monopoly profits, the monopolist may not hold its costs at their lowest possible levels as it would attempt to do in a competitive market.

Two. Equity. This includes a determination of whether the prices consumers are paying are higher than necessary to attract capital investment into the industry. When this occurs, there is unnecessary and inequitable transfer of income from consumers to producers.

Three. Progressiveness. This concerns the extent to which suppliers have taken full advantage of scientific and technological opportunities to provide the best possible products made by the most efficient production processes available. In a competitive market one where the structure of an . . .. .

Initial Decision industry is such that there are many sellers, prices and profits will tend to equal costs (costs being defined to include a normal profit). This is good economic performance. When there is a single monopolistic seller, there is a tendency for the seller to raise prices in excess of costs and so earn excess or monopoly profits. This indicates poor industry performance. Between the extremes of pure competition and monopoly, there may be oligopoly. In an oligopolistic industry, the prices may gravitate toward costs (the competitive result) or they may tend to be in excess of cost (the monopolist result). (30) In an oligopoly, where you have few sellers each of considerable size, the sellers often recognize that each seller s actions may materially affect the fortunes of the others; that their fortunes and decision making are interdependent. The oligopolists are likely to realize that, if they work together and coordinate their actions, they may achieve monopoly profits by sharing the monopoly power they possess as a group. If there is a barrier to the entry of new firms, they would be able to sustain monopolistic profit levels without the threat of new competition coming into the industry. There is a great incentive for sellers in an oligopoly to coordinate their activities rather than to engage in aggressive competitive actions which may lead to retaliatory actions by their rivals. Thus sellers in an oligopoly may achieve a monopoly result by tacitly colluding or by pursuing a cooperative policy. Even if they do this, if there is no substantial barrier to entry, new entrants would be attracted to the industry who would add additional competition. However, if there is a barrier to the entry of new firms, the oligopolists would be able to sustain their monopolistic profit levels without the threat of competition from the outside. In . order to determine whether a market is operating in a competitive or a monopolistic manner, or close to one of those extremes, a careful analysis of the structure of the industry and of the conduct and performance of its sellers must be undertaken (CPF 6-16).

Having demonstrated that the RTE cereal market's highly concentrated structure contributes directly to respondents' conduct in tacitly colluding and cooperating to avoid competition and to raise prices to monopoly levels that the respondents' conduct and the ultimate poor performance of the RTE cereal industry flow from the industry s structure, it is clear that an effective remedy requires structural relief. "So long as the structure of the RTE cereal industry is unchanged, it is unlikely that an order directed solely to the firms conduct will result in more vigorous competition and improved economic performance" (CPF &-19).

Initial Decision 99 F. From complaint counsel's own analysis of what they purport to show, it can be seen that they have assumed a formidable burden of proof. Having outlined the elements of complaint counsel's theories of violation, we can now turn to what the record actually shows. (31) FINDINGS OF FACT AND DISCUSSION THE RESPONDENTS AND OTHER PRODUCERS OF RTE CEREAL 1. The relevant product market, as found in the next section, is the RTE cereal industry. The six major producers in the industry are the three respondents and Quaker, Nabisco and Ralston (CX I06A- G).

Kellogg 2. Kellogg, founded in 1906 is a Delaware corporation with its located in Battle Creekprincipal offce and place of business Michigan. Kellogg and its wholly-owned subsidiaries manufacture and sell, among other things, RTE cereal products (its principal business), tea, soup, gelatin, and puddings (Kellogg Answer 3; Tr. 936-38, 12 173-74; CX K 746G). In 1968, RTE cereals accounted for approximately 87% of Kellogg s net sales; in 1970, this figure was about 75% (CX-K 1090Q). Kellogg is the largest producer of RTE cereals. In 1950, Kellogg s share of the total RTE cereal market was (32)between 35% and 37% , on either a pound or dollar sales basis; 1970, it had approximately 44% of the market, on either basis (CX 106A-G).

3. Kellogg has both domestic and foreign subsidiaries. The part of Kellogg that has responsibility for manufacturing RTE cereals in the United States is called Kellogg Company, U.S. The Kellogg Sales Company subsidiary encompasses the field selling force of Kellogg. Kellogg International has responsibilty for the sales of Kellogg RTE cereals in countries other than the United States (Tr. 11 925-26 930-31, 11 939).

10 Findings of fact, for the most part, are made in numbered paragraphs- Dis.u!;ions and applications of findings well as consideration of legal and other matters, appear where deemed appropriate. Some prece or follow particular findings which pertain thereto; others foHow an of the numbered findings.findings whkh appear inunnumbero; paragraphs are, nevertheles, findings. The record cootaim a considerable amount of economic testimony. When I cite the testimony of an economist under findings of fact, it may mean that I am accepting the economic principle. theory or conclusion testified to as a finding in this matter. And when I do so, that necesarily means that I have rejected other contrary or incon"iS1\'nt economic positions. On the other hand, it may simply indicate that an economi t has testified to or espoused a particular economic propoition without my necessarily accepting that economic propoition as a fact. The n8ture of my treatment of such economic evidence will be apparent from the finding. Many of the conclusiuns reached by economic expert are bas upon other evidence in the record. The weight to be given such economic conclusions, of cours, depends upon, and varies with, the reliability of the evidence relied upon as well as the economic witness' accuracy in relating and interpreting the relied upon evidence Initial Decision 4. Until 1939, Kellogg manufactured RTE cereal products only at Battle Creek, Michigan (CX-K 1097 A). In 1943, it acquired the business and certain assets of the Miler Cereal Company in Omaha Nebraska, and leased, with an option to purchase, the manufacturing facilities. In 1958, Kellogg purchased these facilities (Kellogg Answer 11). From 1963 to 1970, Kellogg manufactured RTE cereal at four plants in the United States: Battle Creek, Michigan; Memphis, Tennessee; Omal.a, Nebraska; and San Leandro, California (CX-K 1079A; Tr. 11.926). Today, Kellogg has another plant in Lancaster, Pennsylvania (Tr. 36 583-84). Kellogg also has RTE cereal manufacturing facilities in 17 foreign countries and sells its RTE cereal products in more than 130 countries (CX-K 1090R). 5. In 1970, Kellogg Company had RTE Cereal Operations assets of $179 milion (CX 757B). In 1950 and 1970, Kellogg s RTE cereal sales were $64 922 000 and $308 944 000, respectively. Pound sales for 1950 and 1970 were 258 604 000 and 591 707 000, respectively (CX 430C - Tables, p. 11 "G-Kellogg," received into evidence as GFX 1319).

General Mills 6. General Mills is a Delaware corporation with its principal office and place of business located in Minneapolis, Minnesota: General Mills is a diversified company manufacturing and selling, among other things, RTE cereals, flour, toys, chemicals, clothes and jewelry (General Mills Answer 2B).

7. Prior to 1969, General Mills' RTE cereal business was a part of its Grocery Products Division with several functional areas: a manufacturing division, a sales division, a marketing division, and a comptroller (Tr. 18 145). In 1969, General Mils reorganized its divisions which had responsibilty for RTE cereals and created the following divisions within the general framework of its Consumer Foods Group: (1) the Big G Division, which is responsibile for all marketing of RTE cereals and associated products; (2) the Packaged Food Operating Division, which manufactures all products of the Consumer Food Group with the exception of family flours; (3) the Grocery Products Sales Division, which sells all products of the (33) Consumer Food Group with the exception of institutional products; and (4) the Sperry Division, which is responsible for institutional sales (Tr. 18 132- , 18 135- 140-1; CX-GM 2041). 8. General Mills operates RTE cereal plants in Buffalo, New York; South Chicago and West Chicago, Ilinois; Toledo, Ohio; and Lodi, California (CX-GM 2119; Tr. 18 133). RTE cereal represents , p.

Initial Decision 99 FTC. between 40% to 60% of production for each of these plants (Tr. 133-34; CX-GM 2119).

9. In 1970, General Mills had RTE cereal operations assets of $66.9 million (CX 754D). Its sales of RTE cereal for fiscal year ending " May 1951 were $33 224 000 and 125 481 000 pounds; sales for fiscal year 1970 were $141 775 000 and 269 427 000 pounds (CX 430C - Tahles C-GMI " received into evidence as GFX 1319). General Mills' share of the total RTE cereal market for both 1950 and 1970 was approximately 20% on either a pound or dollar basis (CX 106A- G).

General Foods 10. In 1895, Charles Wiliam Post began to produce Postum Cereal, a grain-based cereal beverage, as a coffee substitute. In 1896 he organized the Postum Cereal Company, Ltd. (a partnership association limited) in Batte Creek, Michigan to market this product. Mr. Post developed Grape Nuts cereal, and introduced it in 1898 (GFX 1234D, E, 1370A, B).

11. In 1899, Mr. Post founded the Batte Creek Paper Company Ltd. to provide cartons and containers for his products. This company evolved into the present Carton & Container Division of General Foods Corporation, which provides packaging for various General Foods products-including RTE cereals as well as for some contract customers (GFX 1226F, 1370B).

12. In 1904, the Postum Cereal Company introduced a corn flake product named Elijah's Manna. In 1907, this product was renamed Post Toasties. In 1922, the Postum Cereal Company introduced its third RTE cereal-Post' s 40% Bran Flakes-which, like the first two, remains in distribution today (GFX 1226F 1234E 1370B). 13. In the late 1920' , the company moved outside the cereal category and expanded its product line to include gelatin desserts flours, puddings, chocolate, coconut, syrup, coffee, baking powder fruit pectin and frozen foods. In recognition of this development of its lines of business, in 1929, the name of the Postum Cereal Company was changed to General Foods Corporation (GFX 1226D, 1370C). 14. In 1943, General Foods acquired the Jersey Cereal Company of Irwin, Pennsylvania, a manufacturer of private label and controlled brand RTE cereal, with plants in Irwin and St. Joseph, (34) Missouri (Tr. 12 143-44 , 13 714). General Foods subsequently closed the Irwin and St. Joseph plants and moved their operations to Battle Creek (CX-GF 121G).

15. General Foods is a Delaware corporation with its principal , p.

Initial Decision office and place of business located in White Plains, New York. General Foods produces and/or sells, among other things, RTE cereals, baking powder and ingredients, coffee, beverages, nuts popcorn, frozen foods, syrups, pancake mixes, pet foods, and gelatin and other desserts (General Foods Answer n 11 (b); GFX 1242- 1370C). Since approximately 1965, General Foods has been the third largest producer of RTE cere which are sold under its " Post" label; before 1965, General Foods had been the second largest RTE cereal producer (CX 106A-G; GFX 1370J).

16. During the period of time covered by the complaint, all of General Foods' RTE cereals were manufactured in Battle Creek Michigan (CX-GF 562). General Foods now has an additional RTE cereal manufacturing plant in Modesto, California (Tr. 36 658-59 966).

17. After an internal reorganization in 1946, General Foods began to manufacture and market its RTE cereals in the United States through its Post Cereals Division. In the early 1970' , as a result of further organizational changes, a newly formed Beverage and Breakfast Division assumed responsibility for General Foods RTE cereal business, and the Post Division was eliminated (Tr. 215-20; CX-GF 167).

18. In 1970, the RTE cereal operations assets of General Foods were over $44.9 milion (CX 752B). General Foods' sales of RTE cereals for fiscal year 1950 were $25 785 000 and 126 000 000 pounds; in 1970, sales were $94 242 000 and 225 730 000 pounds (CX 430 C - Tables C-General Foods " received into evidence as GFX 1319). In 1950, General Foods' share of the total RTE cereal market was approximately 22% based on dollar and pound sales. By 1970, these market shares had declined to approximately 17.7% based on pound sales, and 14.8% based on dollar sales (CX 106A-G). Quaker 19. The Quaker Oats Company is a New Jersey corporation with its principal office and place of business in Chicago, Ilinois. It was incorporated in 1901. Quaker manufactures and sells, among other things, RTE cereals to- cooked cereals, mixes, frozen foods cookies, pet foods, and chemicals (Quaker Answer n 2D; Tr. 15 389- 90). As of 1970, Quaker produced RTE cereals in plants located in Cedar Rapids, Iowa; Depew, New York; Shiremanstown, Pennsylvania; and Danville, Ilinois (Tr. 15 390; CX Q 576). Quaker has both domestic and foreign subsidiaries (Tr. 15 398-99). 20. Quaker s share of the RTE cereal market, based on pound Initial Decision 99 F.TC. sales, was between 4. 1 % and 5.0% in 1950, dropped to approximately (35)2.4% in 1960, and rose to about 7.0% in 1970. Quaker s market share, on a dollar sales basis, went from 6.6% in 1950 to approximately 9.0% in 1970 (CX 106A G).

E. Nabisco 21. Nabisco is headquartered in New Jersey (Tr. 3224). It is engaged principally in the manufacture, processing, and sale of food products in the United States and foreign countries. Its products include biscuits, cookies, crackers, hot cereals, RTE cereals and pet foods (CX GF 3000Z-139). Nabisco s major RTE cereal products are Nabisco s Shredded Wheat and Spoon Size Shredded Wheat (CX 430 C - Tables, p. " Nabisco," received into evidence as GFX 1319). 22. Nabisco purchased Ranger Joe Cereal Company, which had been a regional producer of presweetened puffed wheat and rice, in the mid-1950' s (Tr. 26,494 95; CX-GF 167Z-11). 23. Nabisco s share of the RTE cereal market, based on pound sales, was 9.3% in 1950 and approximately 4.8% in 1970; on a dollar sales basis, these market shares were 6.6% and 3.7%, respectively (CX 106A-G).

F. Ralston 24. Ralston was incorporated in Missouri in 1894 and is headquartered in St. Louis, Missouri. It manufactures and sells, among other things, RTE cereals, pet foods, tuna fish, and snack foods (Tr. 3739, 3839, 8190).

25. Ralston produces RTE cereals in Battle Creek, Michigan; Cincinnati, Ohio; and Lancaster, Ohio (Tr. 3531 3572 722). It sells under the Ralston name and under private label (Tr. 3839-40, 8213 506-8). Ralston s share of the RTE cereal market, based on branded pound sales, grew from approximately 3.0% in 1950-1951 to approximately 5.0% in 1960. It declined to approximately 3.8% in 1970. If Ralston s private label sales were included, its market share would be somewhat greater (CX 106A-G).

G. Other Companies 26. Over the years, there have been a number of manufacturers of RTE cereals. It has been estimated that in 1911, there were over 100 brands of corn flakes being packed in the Battle Creek, Michigan area alone (GFX 1370G). (36) 27. In 1965, the National Commission on Food Marketing sur- Initial Decision veyed 58 cereal producers in preparing its study of this category. Twenty-four of these producers had sales of $200 000 or more (GFX 1370G). .

28. For example, Van Brode Miling Co. manufactured corn flakes and crisp rice and private label cereals dating back to the 1940' s (Tr. 12 136, 13 717, 36 316). Jersey Cereal Company produced a number of RTE cereals, including corn flakes, wheat flakes, bran flakes, rice flakes, rice gems, wheat puffs, and rice puffs under its own name as well as private labels. Jersey s net sales were $2. milion in 1942 (GFX 253F).

29. Colgate-Palmolive marketed Weetabix, a shredded wheat and Alpen, a natural cereal (Tr. 12 573 , 26 492). Carnation acquired the Albers Miling Company, an early manufacturer of corn flakes, and began to market the product as Carnation Corn Flakes. Carnation quit the business in 1963 (Tr. 11 752, 12 911 , 26 334). Pilsbury was manufacturing a granola by 1972 (Tr. 12 914). 30. Specialty Brands, early in 1973, acquired the Vita Crunch Foods, Inc., which produced a granola-type cereal. By 1975, this company s sales amounted to about $5 millon (Tr. 25 916-18). 31. Organic Milling Company entered the RTE cereal business in 1971 with a granola product and, by 1975, sales had reached $1 million. Organic Milling Company acquired the Vita Crunch label in 1975 and, in 1977, earned annual revenues in excess of $2 milion (Tr. 37 295, 37 298).

32. The H.J. Heinz Company entered the RTE cereal business in the 1930's and produced a rice flakes product until some time prior to the late 1950's (Tr. 12 992, 26 334).

33. International Multifoods Corporation acquired the Kretchmer Wheat Germ Company and subsequently manufactured a natural cereal (Tr. 30 039). The company also manufactured a granola, Sun Country Granola (Tr. 26,487). 34. Pet, Inc., by 1972, had sold various types of granolas, including Heartland, and two presweetened cereals (Tr. 12 916 998).

II. THE RELEVANT MARKET A. Relevant Product Market The standards set out in Brown Shoe Co. v. United States 370 U. 294 (1962) (a Section 7 Clayton Act case), by which relevant product markets and submarkets are determined are also applicable to a case involving charges of monopolization and restraint of trade. (37) Initial Decision 99 F.TC. United States v. Grinnell Corp. 384 U.S. 563 , 573 (1966); Borden, Inc. 92 F. C. 669, 782-84 (1978). " As stated in Brown Shoe: The outer boundaries of a product market are determined by the reasonable interchangeability of use or the cross-elasticity of demand between the product itself and substitutes for it. However, within this broad market, well-defined submarkets may exist which, in themselves, constitute product markets for antitrust purposes. United States v. E.I duPunt de Nemours Co. 353 U.S. 586, 593-595. The boundaries of such a submarket ay be determined by examining such practical indicia as industry or public recognition of the submarket as a separate economic entity, the product' s peculiar characteristics and uses, unique production facilities, distinct customers, distinct prices, sensitivity to price changes, and specialized vendors (at 325 footnotes omitted).

1. Recognition By The Major RTE Cereal Producers, Including Respondents, Of The RTE Cereal Market As A Distinct Product Market 35. In a 1969 "(oJrientation presentation for people outside the Company telling about the Kellogg Company" (CX-K 746A), "The Ready-To-Eat Cereal Industry" comprises a separate section. Highlights in this section include "the growth in pound volume of this industry" and Kellogg s share of the RTE cereal "market" (CX- 746G). Another Kellogg product, Pop Tarts, is discussed under a separate section entitled "Toaster Pastry Market" (CX-K 746H). 36. Speeches by Kellogg executives clearly reflect Kcllogg s view that RTE cereals is a distinct product market (CX-K 549K, M, 559C), exclusive of products such as Pop Tarts and hot cereals (CX-K 549D 559D). Kellogg s marketing plans for its RTE cereals (38Jalso contain numerous references to the "RTE cereal market" (CX-K 7176A 7177 A, 7178B, 7179A).

37. General Foods' marketing plans also consistently refer to the RTE cereal market" in their analyses of particular RTE cereal brands and groups of brands (CX-GF 1300B, F, 1303A). Several General Foods marketing plans used the following formats which specifically outlined RTE cereal as the relevant market for analysis: Respoodents a&;ert that the Commission s ruling in Borden is in orrcct and rene ts a misreading of Grinnell. Thc Commission s ruling is oot a mere passiog referen c to languagl' inGrinnell but constitutes a sludiedcunsiderat.ionofthcquest.ion. therefore constitutes a controlling prccedentcovering my determinalion of rell'vant product market , Initial Decision Total Industry A. Total Market Trend During Fiscal ' , the total RTE pound cereal market.

(CX-GF 602Q);

Market Climate J.

Market Profie Volume. Ready-to-eat cereal market pound volume.

(CX-GF 130GB).

38. Similarly, General Mills' marketing plans contain references to the "RTE cereal market" (CX-GM 600A; GMX 106A). The marketing plan for the product Lucky Charms, for example, refers to the "Total RTE Market" when discussing "Cereal Market Trends (CX-GM 2176B).

39. When Quaker purchased and used Nielsen data for sales forecasts for RTE cereals, the information was on the basis of the total market, the presweetened market and the nutritional market and it looked essentially at those two segments within the total market. This is the format in which this Nielsen data was ordered by Quaker (Tr. 14 969-70). The respondents similarly purchased reports from Nielson and other survey service companies which included data on RTE cereal sales (See infra Findings 153, 154). 40. The marketing plans of the major producers of RTE cereals also divide the total RTE cereal market into sub-segments. Kellogg recognized "the presweetened segment of the R- E Cereal market" (CX-K 397 A). A Kellogg marketing plan stated that its new presweetened RTE cereal product Kombos "wil not only add to. . . (its 1 overall sales volume, but will, in addition, protect and increase . . . (its) dominance in an important segment of the RTE cereal business" (Tr. 11 689; CX-K 9031A). General Foods noted that "(t)he bran category (excluding raisin brans) continues to decline in importance to the total R- T -E market" (CX-GF 1302A). General Mills' analyses of " Cereal Market Trends" in its marketing plans include graphic presentations which show and plot "Total RTE Market" pound sales over (39)time, with a "Pre-Sweet Market" pound sales line juxtaposed on the same graph (CX-GM 2176B 2178B, 2185B). A Quaker marketing plan states; "(T)his project (Quisp and Quake) is another major effort on the part of Quaker to establish its position in the pre-sweetened segment of the RTE market" (CX-Q 153B). Several Quaker marketing plans included the following format for market analyses:

. .

Initial Decision 99 F. Analysis Market Trends 1. Total RTE pound sales.

2. Total RTE dollar sales.

(CX-Q 164B 167C 2496B, 2497B, 2630B).

41. Kellogg, General Foods, and Quaker projected and evaluated market shares for particular RTE cereal brands in relationship to the total RTE cereal market (Tr. 11 724, 12 133, 12 150, 13 054 969, 36 397).

42. Both Kellogg and General Mills endeavored to secure shelf space for their RTE cereals at least equivalent to the share their sales bore to the total market sales of RTE cereals (Tr. 7103, 8022- 23). Kellogg s magna board, which was used as part of its shelf space program, made no space accommodations for instant breakfast or toaster pastry products (Tr. 8965), and Kellogg s recommendations included the replacement of instant breakfast from the RTE cereal shelf section to the hot cereal section (Tr. 8915). 43. Major producers of RTE cereals viewed only other RTE cereal products and the firms producing those products as competitors in the RTE cereal industry. These producers examined other RTE cereal products to learn about the competitive environment of which their RTE cereals were a part. Industry witnesses testified that the companies with which they compete are other RTE cereal manufacturers (Tr. 7521, 11 374, 14 717, 17,173-77, 18 015; CX-K 549F, G, H). 44. Kellogg, the largest factor in the RTE cereal industry, did not consider other breakfast foods as being in the same market. Pop Tarts, generally known as toaster pastries, were viewed as constituting a separate market of convenience bakery products (CX-K 129 260C 549D, 740F, 746H). The advent of new products known as instant breakfast drinks did not cause Kellogg to make any changes in its advertising policies, promotional strategies, or pricing policies with respect to Kellogg s RTE cereal products. Neither did the introduction of toaster pastry products cause Kellogg to make any adjustments or changes in its promotional strategies or pricing decisions regarding RTE cereals (Tr. 11 558-0). (40) 45. Kellogg wants instant breakfast drinks to be placed in the milk fortifier section of supermarkets, along with such products as Nestles Quick, Oval tine and Bosco (Tr. 12 702). Kellogg s position is that a separate market exists for instant breakfast drinks, apart from the cereal market:

Instant Breakfast drink is not a cereal! It is a milk fortifier, comparable to: LLUGlj CU. , El AL.

Initial Decision Oval tine Homo Bosco Hershey s Milk Amplifier Nestles Quick Ghiradelli' s Flick (CX-K 7172B).

Instant Breakfast" drink is not a cereal. It was reportedly designed to appeal to people who do not eat breakfast. (Cereal customers eat breakfast so it does not belong in that department.

More than 78% of the population drink coffee. To reach breakfast skippers (especially the "coffee only" kind), place "instant breakfast" drinks in the "Coffee, Tea, and Cocoa" department (CX-K 7172D).

46. Ready-to-eat and hot cereals are two separate types or categories of cereals (Tr. 6558, 9095). This distinction was recognized by a Kellogg official in a marketing strategy address at a sales meeting in November 1960:

I think we can safely say that the ready-to-cat cereal industry has done a better selling job than the hot cereals (CX-K 559D).

And Kellogg separately disseminated market share data on cereals to be cooked (Tr. 12 133).

47. The Cereal Institute is a trade association, with membership limited to all producers of RTE and hot cereals (Tr. 11 866-67 333). While producers of hot cereals were members, in part, for the purpose of funding research into the effects of eating cereals of any kind for breakfast (Tr. 15 333), hot cereals clearly do not belong in the RTE cereal industry. RTE cereals are a revolutionary departure from a cooked cereal (Tr. 29 585). Jewel Food Stores, for (41Jexample, treated hot cereals as a separate category from RTE cereals and shelved hot cereals after RTE cereals along with instant breakfast and toaster pastries (Tr. 9095). As developed throughout this section of the initial decision covering relevant product market hot cereals are not a part of the recognized RTE cereal market. They have no impact on the pricing of RTE cereals, they do not qualify as the convenience food which characterizes RTE cereals, but require cooking or heating, and are not produced on RTE cereal type of equipment. Indeed, Kellogg, the largest producer of RTE cereals, has never produced a cooked cereal (Tr. 12 134). The following RTE cereal manufacturers are members of the Cereal Institute: Kellogg Company General Foods Corporation General Mils, Inc.

The Quaker Oats Company Initial Decision 99 F. Ralston Purina Company National Biscuit Company Malt- Meal Company Van Brode Milling Company, Inc.

(CX-CI 59A; Tr. 11 866-67) Distinct Prices Or RTE Cereal And Insensitivity To Prices Or Other Products 48. When establishing prices for RTE cereal products, Kellogg considered the prices of competitive RTE cereal products (Tr. 11 564- , 11 647, 11 736, 12 927-28). General Mills and General Foods also set their RTE cereal prices to be responsive to, or competitive with those of their major RTE cereal competitors and their particular competitive brands (Tr. 14 208-09, 14 235, 36 403; CX-GM 110A; CX- GF 17L, 485Z-107, 60lF, 1382U, 1410L). In pricing Wheaties, for example, General Mills considered primarily all-family cereals and to a lesser extent, children or adult cereals (Tr. 35 527). 49. The introduction of instant breakfast drinks and toaster pastry products caused no changes in the pricing, promotion, or advertising of Kellogg s RTE cereals (Tr. 11 558- , 12 497). Kellogg never instituted a trade deal on an RTE cereal product in response to a trade deal on toaster pastry products (Tr. 12 535). Representative chain store executives testified that, in pricing RTE cereals, retailers would not consider the prices of instant breakfast drinks, toaster pastry products, eggs, or any other products (Tr. 9003, 9132, 9347). 50. For the period 1960-1972, the actual cost per serving of an RTE cereal ranged from two cents to six cents, and an additional four (42Jto five cents for milk (Tr. 17 682, 30 044--5). Carnation Instant Breakfast cost substantially more-13 cents per serving for the product itself, and 25 cents additional for the milk that is required (Tr. 17 683, 30 044--5; CX-K 9B, 11A). A bacon and egg breakfast is materially higher in price than a bowl of cereal and milk to the extent of there being dollar differentials (Tr. 17 098-99 682-83). Pop Tarts cost more than RTE cereals on a cost per serving basis err. 12 213). RTE cereals are more convenient and less expensive, on a per serving basis, than toaster pastries, frozen waffles and frozen pancakes, and have an economic advantage over such products (Tr. 30 042). While hot cereals are sold at lower prices than RTE cereals, this has not resulted in an increased sale of hot cereals at the expense of RTE cereals (Tr. 30 051). Initial Decision RTE Cereals ' Peculiar Characteristics And Uses 51. RTE cereals have grain as a basic ingredient, and are ready to be consumed as purchased without further preparation (Tr. 12 142). They are processed from corn, wheat, oats, rice or bran, and combinations or blends thereof (CX-K 698; CX-GM 736), and are then either flaked, puffed, granulated, extruded, or shredded (Tr. 7524 , 10 723 , 13 405; CX-GM 736). RTE cereals contain sugar and vitamin additives and some are processed in a flavored syrup (Tr. 723- , 11 482, 11 786, 11 806). These are a unique combination of characteristics.

52. Unlike many other breakfast products, RTE cereal products require little preparation prior to consumption. In fact, preparation requirements arc so minimal that a child can easily prepare his or her own RTE cereal The consumer adds only cold milk, and sugar if desired (Tr. 11 796- , 11 858-0, 12,426, 12 996, 30 041-42). The consumer does not cook or heat the product, nor is boiled water or any other cooked substance added to an RTE cereal prior to its being consumed (Tr. 11 753). Such ease of preparation is a significant and distinguishing characteristic of RTE cereals, just as its name "readyto-eat" describes (Tr. 12 923).

53. Most other foods consumed at breakfast require more than minimal preparation. Bacon and eggs must be cooked pdor to being eaten. Hot cereals lack the convenience and ease of RTE cereal products, requiring cooking or the additjon of boiling water prior to consumption (Tr. 13 220). Producers of toaster pastries and frozen breakfast products recommend that these products be toasted or warmed before being eaten (Tr. 30 043).

54. RTE cereal products are packaged for easy storage in moisture-resistant boxes or bags normally containing half a dozen or more servings, and they can be stored or shelved for relatively long periods of time, often for as long as nine to twelve months (Tr. 6664 569). This is in contrast to many other breakfast (43Jfoods which must be kept refrigerated or frozen, and to some which may be stored only for short periods of time. Frozen pancakes, frozen waffles, and frozen french toast, as their names imply, must be frozen; bacon and eggs must be refrigerated. Perishable products, such as bacon and eggs, must be used within a short time.

55. RTE cereals are designed to be a nutritional breakfast which can be eaten from a bowl. Many cereal eaters prefer eating their breakfast instead of drinking it from a glass, as is done when consuming instant breakfast. Buc Wheats, for example, was introduced by General Mills to provide the nutrition of a bacon and egg Initial Dccision 99 F. breakfast for the many consumers who wanted this nutrition in the convenient form of a cereal (Tr. 17 714- 041-42). 56. Unlike other food products, RTE cereals are consumed predominantly at breakfast (Tr. 12 142, 14 224-25). Instant breakfast drinks, on the other hand, are viewed by Kellogg as "quick energy snacks" which are consumed anytime throughout the day, particularly between meals, at lunch, and before bedtime. Furthermore instant breakfast drinks were designed for breakfast "skippers" (CX- K 71720).

RTE Cereals ' Unique Production Facilities 57. RTE cereal products are manufactured by a number of basic high volume processes which are unlike those used for the production of other breakfast foods. The processing includes such steps as cooking, pelleting, drying, flaking, puffing, extruding, toasting, and coating. Bacon and eggs, for example, are not "produced" with manufacturing equipment. Cereal production equipment, on occasion, has been used to manufacture non-cereal products such as Whistles and Bugles, which are marketed by General Mills as snack foods (CX-GM 736, 2018A, 20190, 20200). It is significant that Kellogg, the largest producer of RTE cereals, has never produced a cooked cereal (Tr. 12 134).

58. It is clear, therefore, that the major RTE cereal producers recognize the RTE cereal industry as a distinct market, with certain submarkets, and plan, compete and conduct their RTE cereal business on that premise; that the prices of RTE cereals are distinct from those of other breakfast foods and are insensitive to the prices of such other products; that RTE cereals have peculiar characteristics and uses and are manufactured on unique production facilities. Under the criteria enunciated in Brown Shoe v. United States, 370 S. 294, 325 (1962), it must be concluded that RTE cereals constitute the relevant product market. (44) Segments Of The RTE Cereal Industry 59. While RTE cereals constitute the relevant product market all RTE cereals do not .compete to an equal degree with each other. There are segments or categories of cereals which compete more strongly with each other because of their similar attributes. There are some cereals that are so similar that they compete with each other on a one to one basis. At the same time, some cereals may have a broader appeal than the particular segment or category they may Initial Decision fall in so that they compete to varying degrees with cereals outside of their particular category.

60. The record does not permit an exact delineation of the segments and of the relative degree of competition within and among segments and individual cereals. It does, however, support the finding that there are such segments and such primary and varying degrees of competition.

61. General foods, for example, when looking for new product opportunities, recognized that competition among brands is, to a degree, segmented or localized (Tr. 14 192- , 14 407- , 14 497- 500).

62. Some brands are more directly competitive with one another than with other cereals (Tr. 7563-71). New brands mostly affect the sales of other cereals in the same category or segment. For example if a new cereal is "very close " to a General Foods brand in meeting some of the same consumer wants, some consumers of the General Foods brand would likely try the new brand (Tr. 14 088-89, 14 220- 27). Sometimes there could be an alternative direct choice to a particular General Foods brand (Tr. 13 537--0). While the introduction of a new RTE cereal would have some effect on the sales of all other RTE cereals, it would have more effect on those products in the same segment, and the most effect on products with which the new product is most closely competitive (Tr. 15 222, 15 754-55). 63. The greater the difference between an existing brand and a new brand, the less the impact from the introduction of the new brand. For example, the introduction of a new chocolate flavored presweet would have greater impact on on existing chocolate flavored presweets than on fruit flavored presweets (Tr. 14 966). 64. Kellogg, when pricing its new RTE cereals, considered the pricing of brands with which its new brand would compete most directly. Kellogg planned to sell its new cereals at prices comparable to those charged for these most directly competing brands (Tr. , 13 200-1). Kellogg generally identified areas of opportunity for new brands by looking at the sales growth of particular brands in particular market areas (Tr. 12 832-33). It estimated sales for its new brands, in part, on the particular brands the new brand would have to compete with. Kellogg looked at the brands the new brand was tested against. The sales estimates assumed that the new brand would be priced "competitively" in relation to those brands it might compete most directly against (Tr. 12 876-79). (45) 65. In pricing each established brand, Kellogg gave special consideration to the prices of the brands that the Kellogg brand competed with most directly. Relatively few brands were considered Initial Decision 99 F. as the directly competitive brands for each Kellogg brand (Tr. 056-9; CX-K 130C-0).

66. A Kellogg marketing plan for seven Kellogg cereals indicates that certain new brands are direct competition" for the Kellogg products, while other new brands are "indirect competition, " The new directly competitive brands were deemed to have affected the sales of the Kellogg products (CX- 397C). 67. In determing whether to use an in-pack premium with a particular brand, Kellogg considered what its competitors were doing on a comparable product (Tr. 12 321). 68. General Foods expected its new brand introductions to take some sales from directly competitive brands in the new brand' category, such as all-family, unsweetened or adult (Tr. 8824-25). 69. General Foods believed that if it introduced brands similar to ones it already had on the market, they would take a fair share of sales from those similar brands; that if it introduced brands dissimilar to ones it already had on the market, they would take fewer sales from its own brands (Tr. 14 19&-99). 70. General Mills sales personnel recognize that certain RTE cereal brands are more directly competitive with one another than with others (Tr. 769&-99, 7709-12).

71. While there is not complete agreement among respondents and other RTE cereal producers (or their representatives) as to the demarcation of the segments and as to which brands are in most direct competition, the following brand categories and individual brand competitors have been recognized by them to the extent indicated:

Corn Flakes 72. Kellogg sees its Kellogg s Corn Flakes as most closely competitive with other corn flakes, including Post Toasties, General Mills' Country Corn Flakes and private label corn flakes (Tr. 7565- 66; CX-K 7181I). Kellogg compares the retail pricing of the corn flakes brands (CX-K 130C, 663).

73. In pricing, General Foods has regarded Post Toasties as directly competitive with Kellogg s Corn Flakes, and chose not to change Post Toasties' price in the absence of a price change by Kellogg (Tr. 13 983-84; CX-GF 4180, 485Z-107, 2024C). General Foods decided that its Post Toasties and Kellogg s Corn Flakes should sell (46Jat identical prices for the same size packages (Tr. 235). General Mils recognized that consumers perceived its Country Corn Flakes as nothing more than another corn flake (Tr. Initial Decision 634 , 17 750). Country Corn Flakes, Kellogg s Corn Flakes, and Post Toasties are direct alternative choices to the consumer (Tr. 538-0).

Wheat Flakes 74. The RTE cereal manufacturers have seen other wheat ilakes as the most direct competitors to their own. Kellogg sees the direct competitors for its Pep as General Mils' Wheaties and Post's Grape Nuts Flakes (CX-K 684 758C).

75. Mr. Schulze, a top marketing official of General Mills, felt that its Wheaties' primary competition was Kellogg s Corn Flakes but that Wheaties' pricing was in parity with Kellogg s Pep and General Foods' Grape Nuts Flakes (CX-GM 280). He testified that Wheaties' most direct competitors were Kellogg s Corn Flakes, Pep, analysis of aand Team Flakes (Tr. 17 18&-89). A General Mills proposed Wheaties price change compared its prices to those of its major flake competitors," Post Toasties, Kellogg s Corn Flakes Grape Nuts Flakes, and Cheerios (CX-GM 287B). A General Mils marketing plan listed General Mils' Cheerios and Kellogg s Corn Flakes, Rice Krispies and Special K as Wheaties 4 Direct Competitors" (CX-GM 2181B).

76. General Mils conducted a "switching" analysis study on a test product called Frosted Wheaties that showed a high rate of switching between the test product and regular Wheaties. General Mills, therefore, decided not to market Frosted Wheaties (Tr. 15 830 191-92).

77. General Foods' strategy was to maintain the same retail price on Grape Nuts Flakes that was charged on Wheaties (CX-GF 418D). Variety Packages 78. Kellogg saw variety packages as direct competitors of each other. Kellogg compared the retail prices of its tray pack to those of Post and of Kellogg s Variety to Post's Tens (CX-K 118E, 130C). If Kellogg were introducing a new variety package, it would look at the price of the competing variety packages already on the market (Tr. 564). (47) 4. Rice Krispies, OK' , Alpha-Bits And Cheerios 79. At times, the RTE cereal producers have compared Kellogg and Rice Krispies, Kellogg s OK' , General Foods' Alpha-Bits, General Mills' Cheerios as directly competing brands. Initial Decision 99 F. 80. Kellogg believed that Rice Krispies competed more with Cheerios than with other cereals. Thus, in pricing Rice Krispies Kellogg was "conscious of the price of Cheerios" (Tr. 30 114-15). Kellogg, in its marketing strategy for Rice Krispies, noted that its sales force had made an effort to bring Rice Krispies retail pricing in line with Cheerios, and that this effort resulted in share and volume increases for Rice Krispies. Kellogg s marketing personnel frequently compared the retail price of Rice Krispies with that of Cheerios (CX-K 125B, 126B-, 130C, 131A, 135B; 7179B). General Mils also compared these two brands. In pricing its Cheerios, General Mills originally concluded that Kellogg s Corn Flakes and Rice Krispies were Cheerios' only " major competitive products" (CX-GM 278A). Subsequently, General Mills added Kellogg s Sugar Frosted Flakes to Cheerios principal competition" (CX-GM 563Z-5, Z-22, 2167 A). 81. Kellogg recognizes that its OK' s competes in the same area of the market as Cheerios. It identified General Mills' Cheerios, an established oat cereal, as being in a growing area of the market, thus offering an opportunity for a new Kellogg brand. The sales success of Cheerios was the principal reason Kellogg developed OK's (Tr. 267-68, 12 832-33). Kellogg sees Cheerios as OK's biggest single oat competition (CX-K 563C-E). " Kellogg s objective for OK's was to establish it in the oat area of the market, occupied exclusively by Cheerios until Alpha-Bits and Life were introduced (CX-K 396A). Kellogg expected that the most important source of OK's sales would be Cheerios and Alpha-Bits consumers switching brands. The next most important source would be other kid oriented cereal consumers; then, consumers of all other cereals (CX-K 396A). 82. General Foods introduced Alpha-Bits in recognition of the opportunity to introduce a new brand in this shaped dough area of the market. Consumers had shown a liking for the expanded, doughbased Cheerios. General Foods determined that some variety in flavor, texture and form might support a new brand in this area. That was the genesis of Alpha-Bits, which included other letters of the alphabet in addition to the Cheerios 0" (Tr. 13 413-18). (48) Nutritional Cereals 83. Nutritional cereals compete more closely with each other than with other cereals. Even among the nutritional cereals, some cereals compete more directly than do others. 84. Kellogg s advertising agency considered Quaker s Life and 12 Cheeri~g and OK'" are primarily oats with some differences. Cheeri~g is entirely compo of D's; OK' g had some O's and K's. Later, the K's were removed (Tr. 756, 29 651) , Initial Decision General Mills' Total to be potent competition for Kellogg s Special K with Total a roadblock to further Special K sales increases. It considered Special K and Total to be in a "head-on-battle " with Special K losing volume to Total (CX-K 7355A). In 1965, the agency reported to Kellogg that "products directly competitive in nutritional appeal to Special K are: Total, Life, Kellogg s Concentrate General Foods' Grape Nuts and National Biscuit Company s Team; and that Total was a threat to Special K (CX-K 7353F). 85. The agency reported to Kellogg a "need for a related product (to Special Kj, specifically the need for a strong Kellogg contender in the nutritional cereal market that can combat the nutritional claims of Total" (CX-K 7353H). The sales growth of Special K and Total in the nutritional segment led Kellogg to identify that segment as presenting an opportunity for profit from yet another brand. Kellogg, therefore, developed Product 19, a similar product (Tr. 396-98, 12 839-40).

86. Kellogg identified two important subsegments of the nutritional market segment-the protein segment, toward which Special K was aimed, and the vitamin segment, toward which Product 19 would be aimed. Other products in the protein segment were Life and Concentrate. The only other product in the vitamin segment before Product 19 was General Mills' Total. The most important requirement for Product 19 was to compete in the Total market (T 006; CX-K 466).

87. Most new products are targeted at those people expected to be most interested in the product. Advertising is then developed to appeal to those people. Product 19's advertising was aimed at those 35 years of age and older. Kellogg directed its marketing plan toward the same audience as that served by General Mills' Total (Tr. 11 483 398-03; CX-K 7176H).

88. There were six logical direct competitors of Product 19: Special K, Concentrate, Nabisco s Team Flakes, Total, Corn Total Life-the nutritional cereals (Tr. 7563-64). 89. Kellogg s marketing plan for Special K compared the retail prices of the six other cereals in the "nutritional category" to that of Special K. Kellogg noted that the prices of Total and Product 19 remain very competitive. " In addition, the new presweetened cereals with vitamins, such as King Vitaman represent indirect . . . competition" (CX-K 7184K). Within the grouping of the seven (49Jnutritional cereals, two were considered as most directly competitive: Total and Product 19 (Tr. 11 483, 12 398-403; CX-K 7176A-M). 90. Kellogg considered taking competitive moves with respect to its nutritional brands if Nabisco s Team Flakes continued to improve Initial Decision 99 F. its sales. It also noted that General Mills had responded to Product 19 by increasing Total's iron fortification level. General Mills advertising campaign on Total included an assertion that it was a calorie controlled hreakfast " following Special K's primary claim for being a "calorie-controlled, complete breakfast" (CX-K 7180E). 91. "Since its introduction, the overall marketing strategy for Product 19 has been to place the brand in direct competition with General Mills' Total." Kellogg recognized that Product 19's most readily identifiable competition has come from within the nutritional category. The other cereals in that category are Total, Special K Post Fortified Oat Flakes, Life, Grape Nuts, and Nabisco s Team Flakes. Total has traditionally been Product 19's primary competitor. Kellogg compared advertising and pricing among these seven cereals (CX-K 607, 674, 7176A, E, F, G).

92. Kellogg s product manager for Special K recommended a coupon on Special K to combat General Mills' use of a promotional insert with Total and Post's Grape Nuts ' sales growth (CX-K 649). Kellogg compared the retail prices of Product 19 and Special K to Total to determine whether they maintained their proper relationships (CX-K 125B 130C).

93. In analyzing the consumption of "health cereals " General Mils concluded that consumers were more likely to switch their purchases among Total, Special K, Product 19, and Fortified Oat Flakes (CX GM 128C 570B). The nutritional category, according to General Mills, included these cereals plus Life and Grape Nuts (Tr. 7699, 7710; CX GM 597F, G). Among these cereals, the most nearly competitive were considered to be Total, Special K, and Product 19 (Tr. 15 746-7; CX-GM 174, 280). General Mills responded to the introduction of Total's direct competitor " Product 19, by increasing the fortification of Total (CX-GM 567).

94. From 1966 to 1970, General Mills believed that Kellogg Product 19 competed more directly with General Mills' Total than it did with other RTE cereals. Mr. Schulze of General Mills was concerned that the introduction of Product 19 would have an adverse effect on the sales of Total, which it did (Tr. 17 196-203). General Mils' product manager and assistant product managers for Total thought that the introduction of Product 19 adversely affected Total's rate of sales growth (Tr. 17 236). 95. General Mills thought Kellogg s Product 19 and Special K and Quaker s Life would likely have the greatest impact on the sales of its Total. General Mills considered Product 19 a "direct competitor" to Total. Special K was said to be "more directly competitive Initial Uecision with Total" after Kellogg increased the fortification level of Special K (CX-GM 564C). (50) 96. Product 19's introduction in Kellogg s Sperry Zone (Western United States) adversely affected Total's market share in that area to(CX-GM 570J, L). General Mills undertook "defensive activity" blunt the inroads of Product 19 on Total's sales (Tr. 17 231-37; CX- GM 570N).

97. While General Mills felt that Product 19 had taken sales from brands other than Total, its impact was felt in lost Total sales (CX-GM 69B, 567A, 570Z- , 60lG, I) General Mills expected increased pressure on Total's sales from other brands to be introduced in the health cereal category, which included Total, Product , Special K, Fortified Oat Flakes, and Life (CX-GM 69C 570Z-13- 14). General Mills' strategy for Total was to "stave off inroads made by Product 19" (CX-GM 570Z-15).

98. In determining its introductory price for Buc Wheats, General Mills considered the prices of Total, Product 19, and Special K. It wanted to make sure that Buc Wheats' pricing was reasonably competitive with those products (Tr. 17 297 , 17 284; CX-GM 592Q, 593U, Z-16, 599S, V).

99. General Mills' marketing personnel concluded that bran and health categories differed. General Mils expected Vital 7, a fortified bran product, to compete in the health category, although it might compete with bran products as well. General Mills expected Vital 7 to compete with other 100% minimum daily vitamin requirement products such as Total and Product 19, but not necessarily with Special K, which was considered quite different (Tr. 17 261) 100. In January 1968, a General Foods consumer study concluded that "three brands form the basic nutrition(al) cereal category- Total, Life, (and) Special K with other brands used substantially less frequently for (their more) specific nutritional attributes" (CX- GF 1348Z-204). A second consumer survey found that the nutritional cereal category included Special K, Total, Product 19, and Grape Nuts (CX-GF 14030, Q). General Foods considered the prices of Special K, Total, and Life when it priced Grape Nuts (CX-GF 485Z- 85).

101. General Foods' market research group explored how large an opportunity there was for a particular new brand by looking at the brand's characteristics, the strength of its consumer appeal, the size of the potential consumer group, the other brands that could satisfy those consumers in the same general market segment, and how those other brands were performing (Tr. 14 188-89). 102. When General Foods introduced Fortified Oat Flakes, it Initial Decision 99 F. concluded that its "direct competition" was Special K, Product 19 Total, and Life (CX-GF 1406G). In exploring how large the opportunity was for Post' s Oat Flakes, General Foods looked at the market segment comprised of the cereals whose primary attribute was nutrition. Those cereals included Grape Nuts, Special K and the wheat based cereals (Tr. 14 18&-89). Kellogg s introduction of Concentrate, marketed as a nutritional cereal, was closely watched by (51JGeneral Foods because of the likely effect on its Oat Flakes (Tr. 220-27).

103. Quaker priced its Life cereal to maintain a "spread relationship" with "its principal competition-Special K and Total" (CX- 46E, H; 47E).

6. Raisin Brans 104. Raisin bran cereals constitute a recognized cluster of cereals that compete more closely with each other than with other brands (Tr. 7565). In marketing Kellogg s Raisin Bran, Kellogg gives special consideration to Post's Raisin Bran advertising (CX-K 7177 A- , D , !) Kellogg compares the retail prices of the Post and Kellogg raisin brans (CX- 130C 663). A Kellogg analysis of its Raisin Bran business considered only Post's Raisin Bran in addition to its own (CX-K 439).

105. In addition to comparing raisin brans, Kellogg also considered some less direct relationships between raisin brans and other cereals. For example, in preparing its marketing strategy for Kellogg s Raisin Bran, Kellogg also watched Post's Bran/Prune Flakes and Ralston s Bran Chex with Raisins. Kellogg also noted the demise of General Mills' Bran with Raisin Flavored Flakes (Tr. 7565; CX-K 7177 A- , 0, H-I).

106. When General Foods priced its raisin bran, it compared only Kellogg s Raisin Bran prices (CX-GF 439C 440J, 2000R). It viewed Kellogg s Raisin Bran as Post Raisin Bran s !!major (and virtually only) competitor" (Tr. 13 537--0; CX- , 440F, 485Z-120). General Foods believed that the two raisin brans should sell at identical prices (Tr. 14 235).

107. General Mills, with its Wheaties Bran with Raisin Flakes attempted to get some of the growing business that was being obtained by the competitive raisin bran products (Tr. 17 995). Whole Bran 108. Kellogg developed Bran Buds in recognition of the opportunity for a new brand in the growing segment of whole bran products . .

Initial Decision consisting of Kellogg s All Bran and Nabisco s 100% Bran (Tr. 844). Kellogg compared Bran Buds with Nabisco s 100% Bran in terms of consumer preferences. Bran Buds was designed to protect Kellogg s position in the whole bran category, while blocking the sales growth of Nabisco s 100% Bran. Kellogg introduced Bran Buds to work with All Bran to "bracket competition. " Kellogg expected Kellogg s All Bran and Bran Buds and Nabisco s 100% Bran to be close competitors (CX-K 409B, 566, 605, 686). (52) 109. It is not likely that a bran product would compete with a presweet flavored product such as Apple Jacks. If there were any competition, it would be very small (Tr. 12,880). 8. 40% Brans 110. The 40% bran cereals constitute a recognized cluster of brands that compete more closely with each other than with other brands. General Foods believed that its Post Bran Flakes "competes most directly with the Bran segment of the RTE cereal market. composed of Bran Flakes and Whole Bran products, excluding Raisin Bran " (CX-GF 1990C); that within that grouping, Post's Bran Flakes are "directly competitive" with Kellogg s 40% Bran Flakes (CX- 485Z-23). General Foods tried to price its Bran Flakes to match Kellogg s 40% Bran Flakes (Tr. 14 235; CX-GF 418D, 1989B, 1991H). Kellogg compares the retail prices of Post 40% Bran Flakes and Kellogg s 40% Bran Flakes (CX-K 87B, 122B, 127 A). 111. General Mils' marketing people believed that the bran market might support another product. General Mils did a simulated sales test, mailing a package flat to a representative sample of bran consumers to determine if they were interested in buying General Mils' new product, Alive (Tr. 15 797-99). 9. Shredded Wheat 112. The shredded wheat segment includes cereals perceived by consumers as meeting similar desires of taste and texture-Quaker Muffets, Quaker Shredded Wheat, Nabisco Shredded Wheat and Sunshine Shredded Wheat (Tr. 15 289). Within the shredded wheat segment, brands were divided into whole biscuit and spoon-sized products (Tr. 15 289-92).

113. Quaker priced its Shredded Wheat relative to Nabisco Shredded Wheat (Tr. 15 289-301).

114. Kellogg placed its Mini-Wheats in the "Shredded Wheat market" where it would appeal first to consumers of shredded wheat products. Other products deemed by it to be in this area of the Initial Decision 99 F. market were Nabisco s regular and spoon-sized shredded wheats Ralston s Wheat Chex and General Mills' Wheat Stax. Mini-Wheats was expected to take sales away from other shredded wheat competitors (CX-K 573C-G, 679). Nabisco s Sweet Wheats was seen as a close competitor to Mini-Wheats. Kellogg reasoned that to cut into Nabisco s Shredded Wheat sales significantly will require a more directly competitive product-an unsweetened" Mini-Wheats (CX-K 7201E). In determining the introductory advertising for its Mini-Wheats, Kellogg considered the introductory advertising (53) expenditures for the most recent brand introductions in the shredded wheat category (Tr. 13 096). However, Mini-Wheats was suffciently differentiated from other shredded wheat products that consumers of Mini-Wheats, over 60% of whom came from purchasers of regular shredded wheats, did not make price comparisons with the regular shredded wheats (CX-K 529A).

115. Kellogg attributed declines in the sales of Kellogg and Nabisco shredded wheat to the success of Sunshine Shredded Wheat (CX-K 647).

116. Kellogg s Krumbles was considered by Kellogg to be a part of the Wheat Shreds market, consisting of six other shredded wheat brands. Krumbles was the only one not in biscuit shape (CX-K 680A C).

10. Presweets 117. Kellogg regarded particular products as representing narrower areas or subsegments within the presweetened segment. It sought to develop products to compete more directly with particular presweet products, such as Cap n Crunch and Trix. Sales growth for these products indicated that this was a narrower area of opportunity within the overall presweet area. Kellogg developed Froot Loops for the "fruit flavored presweet" area (Tr. 12 869-75; CX-K 526, 909). 118. Kellogg compared the retail prices of certain presweets: Kellogg s Sugar Pops and General Foods' Honeycombs; Kellogg Sugar Smacks and General Foods' Sugar Crisp; Kellogg s Froot Loops and General Mills' Trix; Kellogg s Cocoa Krispies and General Mills' Cocoa Puffs; General Mills ' Corn Bursts and Kellogg s Sugar Frosted Flakes (CX-K 33, 130D). These pricing comparisons indicate Kellogg s belief that the brands compared were more directly competitive with one another than with other brands. 119. General Foods considered introducing an unflavored Pebbles in addition to its two flavored Pebbles products (CX-GF 1372L). It did not do so, however, since it was a presweetened unflavored rice Initial Decision product like its Sugar Sparkled Rice Krinkles, and might drive Krinkles out of the market. While there were differences, the products were too similar (Tr. 17 456-58). 120. When Kellogg introduced Apple Jacks, it packaged and priced that product competitively with General Mils' Wackies and Lucky Charms, Post's Honeycombs and certain other new items. Kellogg was concerned that Apple Jacks might take sales from its own existing presweets (CX-K 7342B). General Mils considered Apple Jacks to be one of five major competitors of Lucky Charms (CX-GM 604Z-11, 2176E; Tr. 2824-26). (54) 11. Presweetened, Shaped Oat Cereals 121. Within the overall presweetened segment, presweetened shaped oat cereals compete more directly with each other than with other presweets. Kellogg tested its All Stars against General Mills Frosty O's and General Foods' Alpha-Bits to judge consumers preferences (CX-K 528). All three were presweetened, shaped oat cereals (CX-GM 2). All Stars' introduction was analyzed by Kellogg in terms of its effect on General Mils' Frosty O's and Twinkles (CX- K 604). General Mils thought that the introduction of Kellogg s All Stars adversely affected the sales of General Mils' Twinkles (Tr. 621; CX-GM 2). General Foods endeavored to minimize Crispy Critters' cannibalization of Alpha-Bits (Tr. 14 198-99). OK's and Stars Otoverlapped some " but were not considered by General Foods Director of Corporate Marketing Services to be identical to General Foods' shaped cereals, such as Alpha-Bits (Tr. 14 220-27). 12. Chocolate Flavored Cereals 122. Within the overall presweet segment, the chocolate flavored cereals compete more directly with each other than with other presweets. While the flavored cereals, such as Froot Loops, Cocoa Puffs, Trix, Orange Sugar Crisp, Kream Krunch, Kombos and Krinkles, are the logical direct competitors to Kellogg s Cocoa Krispies, even among these flavored cereals, the chocolate or cocoa flavored cereals, such as Cocoa Puffs, are more direct competitors (Tr. 7563, 7567). Thus, Kellogg compared its Cocoa Krispies to other chocolate flavored cereals (Tr. 12 825). It compared the advertising of General Mills' Cocoa Puffs to that of Kellogg s Cocoa Krispies (CX- 7175N) and the retail pricing of Cocoa Pebbles to that of its Cocoa Krispies (CX- 574C). Cocoa Pebbles was a more direct competitor to Cocoa Krispies than were some other products (Tr. 12 296; CX- 2021D).

, Initial Decision 99 F. 123. Kellogg expected to lose retail support for Cocoa Krispies when it introduced Cocoa Hoots. It noted a declining share trend for cocoa flavored cereals as a category. "It' s possible that the market is over-saturated with cocoa-flavored products." Other cocoa flavored cereals were Cocoa Puffs, Count Chocula and Cocoa Pebbles (CX- 7205D). Kellogg sought to acquire more of the chocolate flavored segment with Chocolate Kombos (CX-GM 2171B). 124. General Mils believed that Cocoa Krispies and Cocoa Puffs were more closely competitive with one another than with other cereals (Tr. 15 747; CX-GM 2171A). It concluded that its Cocoa Puffs should be "priced right with the Kellogg product," Cocoa Krispies (CX-GM 278A) (55) 13. Fruit Flavored Cereals 125. Within the overall presweet cereal segment, the fruit flavored cereals compete more directly with each other than with other presweets. Kellogg identified the fruit flavored presweets as an area that presented an opportunity for a new brand. This was Trix in thisindicated by the sales growth of General Mills' subsegment. Kellogg, therefore, introduced Froot Loops (Tr. 12,875- 76). Kellogg s Apple Jacks is a flavored cereal. Therefore, its logical competitors are other flavored cereals; then, to a lesser degree, other presweetened cereals (Tr. 7566).

126. Anticipating that General Mils' Lucky Charms would be a close competitor to Kellogg s Froot Loops, Kellogg s product manager recommended an increase in Froot Loops' advertising to Htake the steam away" from the introductory advertising for Lucky Charms (CX-K 596).

cereals 127. The Director of Marketing for General Mills' thought that there was a competitive set of fruit flavored cereals competing more directly with General Mills' Trix. Froot Loops was , 16 742; CX-GMTrix s most directly competitive cereal (Tr. 16,599 19B). The competitive environment for Trix was all children cereals, but the product that was watched most closely by General Mils' marketing people was Froot Loops (Tr. 18 015). 128. General Mils did a market test to determine which of two flavor levels of its Trix was superior when tested in the context of a set of fruit flavored cereals. It tested only against Trix s key competitor, Froot Loops because there wasn t anything else really in that category that was a sufficient alternative for the consumers (Tr. 15 814-15).

129. General Foods concluded that its Pebbles "is more likely to Initial Decision attract users of Froot Loops than it is to attract users of other brands" (Tr. 14,461 2; CX-GF 2021D).

130. General Foods determined that, within the overall presweet market, fruit and cocoa flavored cereals were growing at a faster rate than unflavored cereals. General Foods, at the time, did not have any products in these two flavored presweet segments. Subsequently, Fruity Pebbles was tested against Froot Loops and Cocoa Pebbles was tested against Cocoa Krispies head-on-head" (Tr. I 7 ,443--4).

14. Rally and Life 131. Kellogg expected its Rally to compete most directly with Quaker s Life. Kellogg s Rally was aimed at the market in which Life was the leading cereal. Kellogg s estimate of sales for Rally were based on the sales of Ralston s Chex-type products and on Life (CX- 742A- , 7235B, 7239). (56) 132. Mr. Wells, Kellogg s New Products Marketing Manager testified that when he prepared a suggested retail price for Rally, th only directly competitive brand he could recall comparing was Quaker s Life (Tr. 11 735--0).

133. Kellogg recognized Life as Rally s closest, though not its only, competitor, and tested Rally for consumer preference against Life. Kellogg knew that, in competing with Life, it would need to match the marketing efforts, both advertising and promotion, of Life (Tr. 11 700; CX-K 530).

15. Sugar Smacks and Sugar Crisp 134. Kellogg s Sugar Smacks and General Foods' Sugar Crisp are more directly competitive with one another than they are with other cereals. Kellog believed Sugar Crisp to be Sugar Smacks' most direct competitor (Tr. 12 384; CX-K 7175N). When vitamin fortification of Sugar Crisp and inserts caused a softening of Sugar Smacks' sales Kellogg responded by fortifying Sugar Smacks to the same level as Sugar Crisp (CX-K 595, 7175N, 7352B-C; Tr. 7559, 7603 , 12,384 130-33).

135. Sugar Crisp, termed by Kellogg the logical, direct competitor to its Sugar Smacks, was said to have gained sales from Sugar Smacks and Quaker s Cap n Crunch, Quisp and Quake, as a direct result of the insert in the Sugar Crisp package. Kellogg s own insert in Sugar Smacks was expected to result in reducing Sugar Crisp sales to "normal levels" Puffa Puffa Rice was also included in Initial Decision 99 F. Kellogg s analysis of "sweetened puffed products" (CX-K 650A- Tr. 7559).

136. General Foods priced Sugar Crisp in terms of the pricing on the "identical Kellogg brand " Sugar Smacks (CX-GF 418D, 1381S 14JOL).

137. Kellogg had what may be termed a market or brand agency, Leosegmentation study prepared for it by its advertising Burnett (Tr. 12 892-99; CX-K 9012). That study, prepared sometime between 1951 and 1972, came up with six groups or segments as follows:

Group J cereals are Kellogg s Corn Flakes, l ice Krispies, Post Toastics, Special K Whcaties, and Cheerios. Because these cereals "are perceived as having similarities they "are seen as somewhat interchangeable." Rringing these cereals together is their plain " or "bland" taste. (571 They are also seen as the "old, established, traditional brands," The "plain " tasting cereals contrast with flavored cereals. They are good fruit carriers and sugar is added individually (CX-K 90l2G-K).

The Group II cereals are the " kid flavored cereals" including Pebbles, Quisp, Cap Crunch, Lucky Charms, Kaboum, King Vitaman, Froot Loops, Cocoa Puffs, Trix, and Apple Jacks. They are presweetened. They are not bland, but carry a "multitude of flavors." Pebbles, Froot Loops and Trix are seen by consumers in very similar ways. These cereals are defined in terms of their sweet, multi-fruit flavors by those who like them (CX-K 9012M-N).

Group III are presweetened, but not flavored. They include Sugar Pops, Sugar Crisp, and Sugar Smacks. These cereals have rather definite grain tastes, not an added flavor. They are less bland than the Group I cereals. Another similarity is that they arc puffed (CX-K 9012Q.-R).

Group IV are the shredded wheat cereals, spoon size and regular shredded wheat Mini-Wheats, and Wheat Chex. This group is "clearly defined by the texture of its primary ingredient " wheat. There is a distinctive wheat taste, and they are not good fruit carriers (CX-K 90121' , V).

Group V consists of the Kellogg and Post raisin brans alone, with "nothing else quite like them. IRlaisins are the most salient feature of these cereals" (CX-K 9012V). Group VI are the adult nutritional cereals which include Total, the Post and Kellogg 40% bran flakes, Product 19, Fortified Oat Flakes, and to a lesser extent, Special K Grape Nuts, and Wheaties. The first four are soggy and flaked and possess a strong taste. They are "good for you" (CX-K 9012W-X).'3 (58) by 138. The situation just described is termed "localization" " This is a part.ia!sunn"ary of tlw findings of thp. study. SeeCPF 9-94. Except for quoted words and phrases, thisisnntaverbatirnreproductionofllnypartorthestody Initial Dccision complaint counsel and is central to their contention that respondents created a barrier to entering thc RTE cereal industry by their proliferation of products. However, as indicated above, localization is a matter of degree and cereals do compete to varying extents with other than so-called directly competing products and products in their particular category or segment (Tr. 22 794). 139. This is true, to some extent, because individual cereal products possess a' number of attributes and many other cereals also contain a number of these same attributes; and, there is a considerable overlapping of attributes even among cereals not deemed directly competitive.

140. Consumers aged 35 and older consume certain products adult products ), which are not eaten in any significant amounts by children 13 years old and younger. Adult products include Total Special K, Product 19, All Bran, 100% Bran and 40% Bran Flakes. Childrcn consume certain products ("child products ) which are not eaten in significant amounts by adults. Child products are essentially the presweets. Certain products ("all-family products ) are eaten in substantial amounts by consumers of all ages (Tr. 35 367-70). Allfamily products include Cheerios, Rice Krispies and Corn Flakes. Child and all-family products compete with each other (GMX 546A 549A). Adult and all-family products compete with each other. (Tr. 8824- 367-68 372-87; GMX 194, 195, 547 A- , 548A-B). 141. Although only 16.2% of Cheerios' volume is consumed by individuals 35 years of age and over (GMX 547 A), this is a substantial figure. It is such a large volume that Cheerios cannot ignore that group of consumers and advertises and promotes to them (Tr. 35 378-79 386-87).

142. Cheerios does not achieve its volume increases strictly at the expense of products such as Cap n Crunch, Sugar Frosted Flakes or Rice Krispies. It very likely draws from many different consumers who eat many different products (Tr. 35 370-71). " For example, a brand switching analysis shows that of the two-member families who purchased Cheerios in 1969-70 each of 21 other brands accounted for over 1 % of their consumption (GMX 518A). In 1975- , for a similar sample of families, 25 other brands each represented more than 1 of their consumption (GMX 518B).

143. Lucky Charms, a child cereal, need not be concerned with adult cereals; but children who consume Lucky Charms also consume many other child products (GMX 546A, 549A). (59) 144. Total, an adult cereal, competes with many other adult and This is true of other all-fllmily product.(Tr- 791 15,137 392- , 35 367 :J5 384 :J92). Initial Dccision 99 F. all-family products, including Special K, Grape Nuts, Post 40% Bran Flakes and Product 19 (GMX 525A; Tr. 33 730-31 , 33 743-44, 33 751 753 899-900 397-400).

145. Kellogg s Corn Flakes appeals to all age groups and, in a sense, competes with every other cereal on the shelf (Tr. 7566). The brand switching analyses referred to above (Finding 142) show that two member families consuming Kellogg s Corn Flakes had more than 1 % of their consumption accounted for by each of 20 other brands in 1969-1970 (GMX 520A), and by each of 23 other brands in 1975-1976 (GMX 520B).

146. Apple Jacks competes with all flavored cereals and most other presweetened products (Tr. 7566).

147. General Foods' market planners believed that all cereals were competitive with Sugar Crisp, although Sugar Crisp was primarily competitive with other presweetened cereals (CX-GF 4K 141OA).

148. A General Foods market research review of Grape Nuts in 1956 stated: "Because of its distinct flavor, unique form and texture (Grape Nuts) does not fall into any specific category, and competes with al1 RTE cereals for its share" (CX-GF 40). 149. The 1963 Menu Census concludes that there "is a strong usage relationship between Post Raisin Bran and Post presweetened cereals" (GFX 121OZ-5).

150. A FY 70 review by General Foods' Marketing Division states that "Alpha Bits is the only letter-shaped cereal and is primarily competitive to colored, flavored and shaped presweetened brands (CX-GF 1382). The FY 72 Management Summary for Crispy Critters describes that product as a "unique, fun to eat, animal shaped presweetened cereal primarily competitive with other presweetened cereals that possess shapes, additives or flavors" (CX-GF 1417 A). Pink Panther Flakes was considered to be in competition " to al1 children s presweetened cereals" (CX-GF 1439). Cocoa Pebbles frequently interacted" with Cap n Crunch and Frosted Rice, in addition to chocolate flavored cereals (GFX 1317). The Relevant Geographic Market-The United States As A Whole Use by Major Producers of National Price Lists 151. The price lists issued by the RTE cereal companies show uniform national prices for each brand. Kellogg lists a single price for each brand for al1 states within the United States, except for Hawaii (CX-K 828, 846, 856, 861, 866, 869). General Mills' and (60) InjtiaJ Decision General Foods' RTE cereal list prices are each uniform for the entire United States, except for Alaska and Hawaii (CX-GM 440, 452, 469 502; CX-GF 297, 311, 314). Quaker and Ralston each issue price lists which show a single price for each RTE cereal brand throughout the country (CX-Q 430, 432, 434, 436; CX-R 579 580 thru 591). 152. Except when a product is being test marketed, the major RTE cereal companies generally sell their products nationally (CX- 1067; CX-GM 2111, CX-GF 556; CX-Q 2983). During a test market, companies introduce the new product in selected regions and, if the product meets its expected sales volume, the sales regions are expanded in a planned sequence, often termed a ((roll out. When the roll-out has been completed, the product is distributed. on a national level (Tr. 12 537- , 16 009- , 16 035- , 29 773, 29 991). 2. Purchase By Major Producers Of National Market Share Data 153. Each respondent and Quaker subscribed to services provided by the A.C. Nielson Company (Tr. 7728-29, 7732, 8604-D5 , 11 550 802, 14 349, 14 351 , 15,594-96). A typical report purchased from Nielson by these companies contains, among other data, information on the total national sales of RTE cereals, average nationwide consumer prices for individual products, national market shares and advertising of individual brands, national distribution and out-of. stock conditions reflecting retailers' temporary product shortages and estimates of national consumer sales on both a dollar and a pound basis (KX 14). Respondents used Nielson as a medium for exchange of national advertising expenditures on a total RTE cereal basis and on a brand-by-brand basis nationwide (Tr. 11 842-43 233, 14 354, 14 356-58, 15 596-98, 15 600-1; CX-ACN 2; CX- 176).

154. Similar nationwide information was purchased by respondents from Market Research Corporation of America (MRCA), Sellng Area Marketing Information (SAMI), and National Purchase Diary (Tr. 7464 , 11 764 , 12 784 , 14 365, 16 031 , 16 062--6, 29 591, 34 349; CX-GF 1380; CX-K 560; CX-GM 604Z-5).

3. Preparation By Respondents Of Nationwide Marketing Plans For Their RTE Cereals 155. The internal marketing plans prepared by the RTE cereal companies discuss the total RTE cereal market on a national basis the nationwide performance of various segments of the RTE cereal market, and also the nationwide performance of individual cereals Fgm;RAL TRADE COMMISSION DECISIONS Initial Decision 99 F. considered in the context of a national market (CX-K 7175 thru 7179; CX-GM 2170; CX-GF 1300; CX-Q 155). (61) 156. In 1965, Kellogg s marketing plan for its presweetened cereals considered the impact of a limited national roll-out of Cap Crunch and the national introduction of General Mils' Lucky Charms and Crispy Critters (CX-K 7357F). In planning media strategy for Count Chocula and Frankenberry, General Mills decided that "a combination of child network and spot television will be used (in 1972/73) support of in-pack premium promotions to achieve national reach and impact" (CX-GM 2198C-D). A marketing analysis of the raisin bran segment attributed the growth in that segment to a nationwide increase in the absolute number of households consuming raisin bran cereals and to a nationwide increase in the consumption of raisin bran cereals per household (CX-GF 20000). According to Quaker s 1968-1969 marketing plan for Life, among the factors favorable to the growth of the nutritional segment was an increase in the nationwide percentage of better-educated people and wealthier families (CX-Q 155B).

U,e By Major Producers Of Nationwide Network TV And Other Media To Sell Their RTE Cereals 157. Because of the national scope of the RTE cereal market Kellogg uses network television to advertise its products. Network television provides nationwide television coverage for Kellogg advertising efforts (Tr. 11 379-80). Part of Kellogg s 1970-1971 media strategy for Corn Flakes was to use "nighttime TV as the primary medium to provide continuing national support against the primary target audience of younger families with children as well as older adults" (CX-K 7181L). Similarly, television is the primary medium for Kellogg s Raisin Bran because of its national coverage (CX-K 7177G, 7184M, 7187Q).

158. General Foods planned to use "kid network" 3D-second advertisements for Honeycomb in 1971 to take advantage of the medium s ability to generate efficient national reach of children under 12 years of age. Additionally, "kid network allows the brand to effectively tie-in with planned show-oriented promotions" (CX- 1383, 1374N). General Foods admitted making substantial expenditures for network television advertising for its RTE cereals (General Foods Admission of Fact 37).

159. In 1970-1971, General Mils planned to use night network television advertising for Wheaties "to provide a national base of continuity against the target family audience and deliver optimum Initial Decision reach" (CX-GM 2172C; and see, CX-GM 2167, 2171). General Mils has admitted making substantial expenditures for network television advertising for its RTE cereals for the years 1965 through 1970 (General Mils Admissions of Fact 56, 58, 59, 60, 63, 64, 65, 68, 82, 83 90). (62) 160. Quaker also used network television to advertise its RTE cereal products (Quaker Admission of Fact 108; CX-Q 151G , 153I 154J).

161. The above findings compel the conclusion that the relevant geographic market is the United States as a whole. III. STRUCTURE OF THE RTE CEREAL INDUSTRY 162. The RTE cereal market is large and has experienced rapid and substantial growth. Following are the dollar sales of RTE cereals through grocery stores for the years 1952 through 1971, as reported by Nielsen:

Year Total Dollar Sales ($ millons) 1952 237 1953 254 1954 265 1955 280 1956 312 1957 336.4 1958 379.

1959 401.

1960 421.

1961 449.

1962 480.

1963 527.

1964 5775 1965 624.

1966 655.

1967 656.

1968 679.

1969 710.

1970 741.

1971 782 (CX 101F). " (63) " Genera! Mills (GMPF 5-15 thru 18) challenges the reliability of ex 101 find 106 (market share daw) asserting that Nielsen data, upon which they !Ire ba!;ed, are not accurate; that General Mills' Director of Consumer Research testified (Tr. 16 161-63) that Nielsen s fig-res for several product. were not in accord with Genera! Mils own adual gales figures for those products However, Ill! respondents subscribe to Nicben services, including those that furnished sales data (Tr 15 595), and aJl used and relied upon the &8les data so furnished (Tr. ll 489-90, 12 782-83, 13 349- , 15 725, 15 842, 16 063 16, 16:-) While General MiUs' Director of Consumer Research testified to some diffeni'c"s, he exp!lIined that (Continued) Initial Decision 99 F. 163. Total industry sales showing growth in pounds for the period 1940-1975 at five year intervals is as follows: INDEX OF GROWTH IN TOTAL RTE CEREAL INDUSTRY POUND SALES 1940 = 100 Year Ending 12/1 Sales MM Ibs. Total Index 1940 458 100. 1945 594 129. 1950 610 133. 1955 783 171.0 1960 941 205. 1965 1178 257. 1970 1183 258. 1975 1688 368. (CX 100E). (64) 164. The preceding tables show that, until the mid-1960s, there was rapid growth in pound sales of RTE cereals (And see Tr. 26 157). From 1966 through 1970, there was almost no growth. The industry was relatively stable in terms of pounds and had a modest growth in terms of dollars. Rapid growth resumed after 1970. Between 1970- 1971 and 1975-1976, the total RTE cereal market grew 55%. There was very rapid growth, particularly in 1973 and 1974 when the market was growing at an annual rate of 10% or 11 % (Tr. 14 021 684- 158).

165. RTE cereal pound sales grew considerably faster than other sectors of the economy. For the period 1952 through 1966, RTE cereal pound sales grew by 72.8%. Using the Economic Report of the President, which uses constant dollar GNP, the "all goods and services" group grew by 63.9% in real terms over the same period. A comparison of these differences shows that growth of output was considerably faster for RTE cereal than it was for the economy as a whole. For the same period, growth in RTE cereal dollar sales was 176.8%. GNP data from the President' s Economic Report, which is based on current dollars unadjusted for inflation, show that the growth for "all goods and services" was 116.9%. RTE cereal dollar sales grew at an average of about 8.7% a year for the period 1955 to 1959, while the "food and kindred products" growth rate was 3.74% Geneml Milb' figu( covert,,1 point of sales from Genera! Mills whereas Nil'sen s figures refleded sales ilt the store level .' that pipd;".. delays an:ounte for some of the differences (Tr 167). Further, Iw Was not prepared to testify to the accuracy of Nielsen s overall figures (Tr. 16 163); and General MiJ!s failed to iotnxiuce its own sales figures which wereavaiJablc to it KeOogg s Marketing Direcwr testjfi,'(! (Tr- 12 795) that, in his opinion, Nidstm s sales figures were generally accurate and that they were relied on by Kellogg- Under aU these circumstances, the Nielsen sales data are deem,'d reliable.See Bordell. fnc. 92 F. C. 669, 805 n- 40 (1979); and Genferal Foo. CrJrp. 69 F. C. 380 446--47 (1966), for precedent' where the Commission has relied upon Nies')f data. Initial Decision the "nondurables" growth rate was 5.48% and the "al1-manufacturing" growth rate was 6.52%.

166. The RTE cereal industry, therefore, has been experiencing rapid growth. Its growth has been substantially faster than the growth in the value of "all goods and services" and substantially faster than the growth in real output of the economy as a whole. (Tr. 834 158-59; CX 101E, F)."

167. For each year since at least 1940, there have been only six manufacturers of RTE cereal that produced in substantial quantity. The three largest have been Kellogg, General Mils and General Foods, followed by Quaker, Nabisco and Ralston, although not always in that order. From 1943 to 1970, these six firms accounted for at least 89% of industry sales, either on a pound or dollar basis. That 89% figure was in 1943. In 1970, these firms enjoyed over 97% of the market (CX 106A--). " The following chart shows four-firm concentration ratios for the industry from 1962 through 1970: (65) General Mills, for example, favorably compared an RTE cereal annu,,! growth of 7% with the 4% growth of the c!lnnecl 5nup industry. The latter was cOflsid..rer by it to be a strong annual growth (CX-GM Uta) General Mills deemed the lol1g-t.rm growth of cereal sales to be phenomena! (CX-GM 262A), And Kellogg regardt,d cereals to be "one of the few rapidly "xpanding industries" (CX-K 552B). " There has ben no new major entry to challenge the positions of them' six companies 74i!5-36,(Tr 11 749 911- 995) _ _ FEDERAL TRAm: COMMISSION DBC!SIONS Initial Decision 99 FT.C RTE CEREAL INDUSTRY FOUR-FIRM CONCENTRATION RATIOS POUND BASIS DOLLAR BASIS Nielsen Dala Adjusted Co Adjusted Co from General Adjusted Co. Submitted Adjusted Co. Submitted Data Foods and Nielsen Submil1ed Data, Brand Submitted Branded Prod' Year Daa cls Only 1 Source: ex 1 06A 2 Source: CX 106B.

3 Source: CX 106F 4 Source: CX 106G.

5 Source: ex 1060.

I; Source: ex 106E.

(66)168. Since at least 1937, the four-firm concentration ratio based on pound sales has never been below 81%. And Kellogg, General Mils and General Foods together have accounted for over 80% of the pound sales from 1951 through 1971 (CX 106A, B, F, G). In 1950, Kellogg s share of the market was between 35% and 37% on either a pound or dollar sales basis; in 1970, it' s share approximated 44%. General Mills' share of the total RTE cereal market for both 1950 and 1970 was approximately 20%, on either a pound or dollar sales basis. In 1950, General Foods' share of the total RTE cereal market was approximately 22% based on dollar and pound sales. By 1970, its market share had declined to approximately 17.7% based on pound sales, and 14.8% based on dollar sales (CX 106A-G).

169. Quaker s share of the RTE cereal market, based on pound sales, was between 4. 1 % and 5.0% in 1950, dropped to approximately 2.4% in 1960, and rose to about 7.0% in 1970. Quaker s market share, on a dollar sales basis, went from 6.6% in 1950 to approximately 9.0% in 1970. Nabisco s share of the market, based on pound sales, was 9.3% in 1950 and approximately 4.8% in 1970; on a dollar basis its market shares were 6.6% and 3. , respectively. Ralston share of the RTE cereal market, based on branded pound sales, grew from approximately 3.0% in 1950-1951 to approximately 5.0% in 1960. It declined to approximately 3.8% in 1970. If Ralston s private label sales were included, its market share would be somewhat greater (CX 106A-G).

Initial Decision 170. Another measure of industry concentration, the Herfindahl Index, is calculated by summing the squares of each firm s market share. The Herfindahl Index differs from the more commonly used concentration ratio in that it shows a higher figure for industries where individual firm market shares are disproportionately high. For example, an industry in which the four largest firms had 50%, 10%, 10%, and 10% of the market, respectively, would have a higher Herfindahl figure than an industry where the four largest firms each had 20% of the market. Despite this difference in result, there is a correlation between four-firm concentration ratios and the Herfindahl Index. Highly concentrated industries have high Herfindahls. Industries that are not highly concentrated have low Herfindahls (Tr 21 709- 215-21).

171. The Herfindahl Index for the RTE cereal industry has increased from .223 in 1945 to .276 in 1970 (CX 106A-G). " The following chart shows the lIerfindahl Indices for the RTE cereal industry using several available alternate sources: (67) RTE CEREAL INDUSTRY HERFINDAHl INDICES POUND BASIS DOLLAR BASI Nielsen Dala Adjusted Co. Adjusted Co. Irom General Adjusted Co. SubmiUed Adjusted Co. Submitted Data Foods Data Brand SUbmitted Branded Prod' and Nielsen Submitted Year !Qgg Data ProduclsOnly' 1962 270 286 284 305 304 1963 276 1964 0.278 1965 0.273 262 1966 262 266 281 280 0.292 291 1967 269 268 277 275 0286 285 1968 0.268 283 0.282 294 294 1969 274 287 286 293 292 1970 80.276 0.289 288 294 293 1 Source: ex 106A.

2 Source: ex 106B.

J Source: ex 106F.

4 Source: ex 106G.

5 Source: ex 1060.

6 Source: ex 106E.

(68)172. By various alternative methods of comparison, including ran kings of industdes by concentration ratios, averages of concentration ratios, and averages of Herfindahl Indices, the RTE cereal " The Herfndahl Indices for the Itte cereal industry are somcwhal underslated inasmuc), as they Were c,,1culated using market. shares Df only lhe si" largest producers, where data were available. Inc1usi"n of lhe "quaresofthcn\arket.shilresnflhesrnallerllrmswouJdhavemisedthe industry indices Initial Decision 99 F. industry has been one of the most highly concentrated industries in the United States (Tr. 26 166-67).

173. Census Bureau classification Standard Industrial Code 2043 entitled Cereal Breakfast Foods (SIC 2043), includes RTE cereals, hot cereals and baby cereals. As hot cereals and baby cereals comprise a relatively small fraction of the classification, SIC 2043 is a good approximation of the RTE cereal market (Tr. 26.194- , 26 207 672-73). '"

174. In 1958, SIC 2043 had a four-firm concentration ratio of 83%. Only 29 of the 443 industries engaged in manufacturing (the All Manufacturing Group) that year had concentration ratios greater than 80%. Two-thirds of the industries in this group had concentration ratios of less than 50%. Only three of the 42 industries engaged in the manufacture of food and kindred products (the Food Group) had concentration ratios greater than 80%, while two-thirds of those industries had concentration ratios of less than 50% (Tr. 181-82; CX 108D).

175. In 1963, SIC 2043 had a four-firm concentration ratio of 86%, ranking it among the 27 industries in the All Manufacturing Group with concentration ratios above 80%. More than two-thirds of the 417 industries in this group had concentration ratios of less than 50%. Only one industry in the Food Group, in addition to cereals had a concentration ratio greater than 80%. Two-thirds of the 44 industries in the Food Group had concentration ratios of less than 50% (Tr. 26.179-80; CX 108C).

176. In 1967, only 22 of the 354 industries in the All Manufacturing Group had four-firm concentration ratios higher than 80%, one of which was SIC 2043 with a concentration ratio of 88%. It was one of only two industries in the Food Group with four-firm concentration ratios over 80%. Two-thirds of the industries in the All Manufacturing and Food Groups had concentration ratios less than 50% (Tr. 26.178; CX 108B).

177. The four-firm concentration ratio for SIC 2043 in 1972 was 90%. SIC 2043 was one of the 11 most concentrated in the All Manufacturing Group, comprised of 381 industries, and it was the most concentrated of the 43 industries in the Food Group in 1972. Again, two-thirds of the industries in the All Manufacturing and Food Groups had concentration ratios of less than 50% (Tr. 26 167- 72; CX 108A). (69) 178. The simple arithmetic average of four-firm concentration ratios for 1972 was 39% for the All Manufacturing Group and 44% '" Actually, the inclusion of hot cereals and baby cereals would tend to understate the concentration of the RTEcereaJindm;try Initial Decision for the Food Group. The four-firm concentration ratio for SIC 2043 in 1972 was more than twice those averages, 90%. In that same year the eight-firm simple arithmetic concentration average was 520/0 for the All Manufacturing Group and 58% for the Food Group. The eight-firm figure for SIC 2043 in that year was 98% (Tr. 26 185-86). 179. CX lOSJ is a set of tables prepared by the Bureau of the Census which shows average concentration ratios, weighted by industry employment, for the All Manufacturing Group and for the Food Group. The weighted average four-firm concentration ratio for the All Manufacturing Group in the years 1947, 1954, and 1958 ranged from 33.2% to 34. , depending on the base year used to determine the employment weights. The weighted average four-firm concentration ratio for the Food Group for the same years ranged from 30.9% to 33. , again depending upon the base year used in determining the weights. The four-firm concentration ratio for the cereal industry (SIC 2043) was 79% in 1947 88% in 1954 and 83% in 1958 (Tr. 26 186-90; ex 108J).

180. The average concentration ratio, weighted by product value for 184 manufacturing industries which appeared in all four of the Census years 1947, 1954, 1958, and 1963, ranged from 35.3% in 1947 to 38.9% in 1963. The ratios for SIC 2043 again are significantly higher than the averages, ranging from 79% in 1947 to 86% in 1963 (Tr. 26 190-92).

181. The RTE cereal industry is highly concentrated in relation to other industries when comparing Herfindahl Indices of concentration. For 108 four-digit industries in 1947, the average four-firm Herfindahl Index was .0968.'0 During that same year, the four-firm Herfindahl Index for SIC 2043 was . 1609. Of the 108 industries which made up the average Herfindahl Index, only 18 industries had higher Herfindahls than did SIC 2043. In 1954, the average Herfindahl Index for the 93 industries covered was .0921, while the Herfindahl for SIC 2043 was .4026. Of the 93 industries which made up the average Herfindahl Index for 1954, only four had higher Herfindahls than did SIC 2043 (Tr. 21,821- , 26 218-22; CX 106A- , 108S, T). (70) 182. In addition to concentration being consistently high for many years in the RTE cereal industry, the positions and market shares of the leading firms have remained relatively stable (CX 106A-G; Tr. 26 230-31, 26 498-504). Until about 1966 or 1967, the four largest producers in the industry were the three respondents 20 The averag.. Herllndahl Index was calculated from indu",try lIerfindahls compiled by Ralph 1. Nelson using Bure..u of the f'..nsus data (Nelson COnlwllroti'Nj in the Man"faclurilll! Industries of the UniledSiaies 1963). Nelson s bok contains Herfindahl Indices for a numn. r or large industries ror the years 1947 and (Yr.195426 218- 19).

Initial Decision 99 F. and Nabisco, at which point in time Quaker replaced Nabisco as the fourth largest (CX 101A- , 106A-G). Of 44 five-digit product classifications for which necessary information was available covering the years 1954, 1955, 1963, 1967 , and 1972, the RTE cereal industry had the third lowest fluctuation in four-firm levels of concentration. Of 47 such product classifications for which the necessary information was available, the RTE cereal industry had the sixth lowest fluctuation in eight-firm concentration levels (CX lost; Tr. 26 231-33).

183. Various renowned economists have specified certain industry concentration levels as "critical" in the sense that, at or above such levels, sellers become conscious of their interdependence that the action of each seller affects all of the others; and that this recognition of interdependence may lead to a pattern of conduct that causes monopoly power to be exercised. Bain, Meehan, Duchesneau and others have estimated that the critical level is reached when an industry has an eight-firm concentration ratio of 70%. Most studies of economists who espouse the ncritical level" theory estimate the four-firm critical level of concentration to be between 50% and 60% although Stigler found that a 70% four-firm concentration was required to be critical (Tr. 21 711-16). The concentration levels of the RTE cereal industry, as found above, far exceed all of the various estimated critical levels.

The above finding should not be construed as one to the effect that respondents are conscious of their interdependence with the result that they have or are exercising monopoly power by means of tacit collusion or otherwise. It is merely a finding as to the existence of the economic theory of critical levels and that concentration in the RTE cereal industry exceeds the critical levels that have been estimated. Not all economists agree with the critical level concept (Tr. 21 713). (71) IV. COMPETITION Industry Characteristics Complaint counsel assert (CP 136) that respondents exercised monopoly power by avoiding price and non price competition on RTE cereal products. Complaint counsel theorize that in a tightly knit oligopoly, such as the RTE cereal industry, each firm recognizes that its actions and the actions of each of its competitors will have an impact upon the other competitors. It takes into account the other firms' anticipated reactions in every potential competitive decision that it makes. Such firms, in economic parlance, are termed ,. . .

Initial Decision interdependent". In such a situation, the firms recognize that it is to their mutual advantage to avoid competitive activities which will only bring about responsive competitive activities by their large competitors. The firms, therefore, avoid such competitive acts in accordance with understood "rules of the game." In order to so successfully avoid competition, all major firms must participate. These interdependent firms consider unrestrained competition in any dimension to be undesirable because competition of any type would have the potential of breaking down the tacit agreement to limit competition and maximize group profits. Not only must the flrms trust each other, but they must all understand the rules of the game so that no action of any of the participants may be misinterpreted to be other than complying with the rules of the game. 184. In support of their structural approach, complaint counsel rely (CPF 8-74) upon an economic theory set forth by Dr. Jesse W. Markham in The American Economic Review, Vol. XLI (1951), pp. 891-905, in an article entitled "The Nature and Significance of Price Leadership, " A reprint of that article was received in evidence as KX 104. KX 104 is paginated 176-189. Dr. Markham is a professor in the Graduate School of Business Administration at Harvard and was Director of the Bureau of Economics of the Federal Trade Commission from 1953 until the end of 1955. Complaint counsel have stipulated to Dr. Markham s credentials as a recognized expert in the field of industrial organization economics (Tr. 38 253-55). 185. In his article, Dr. Markham advanced the proposition that in industries which possess certain specific features one would expect a priori a type of price leadership. . . inimical to the public interest." He termed this Price Leadership In Lieu Of An Overt Agreement and reasoned that such price leadership could achieve the same parallel action among sellers as if there had been overt collusion to that end (KX 104 at 185-86). Dr. Frederick Scherer, the economist introduced by complaint counsel who sought to apply Dr. Markham s model to the Rre cereal industry, testified that the Markham article "is widely regarded by economists as the leading article on the subject of price leadership in industry. It has been extensively reprinted" (Tr. 27 801). (72) 186. Dr. Markham listed five specific conditions prerequisite to application of his Price Leadership In Lieu Of An Overt Agreement thesis. The first condition is:

Firms must be few in number and each firm must be sufficiently large to be compelled to reckon with the indirect as well as the direct effects of its own price policy (KX 104 at 186).

Initial Decision 99 F. 187. As further elucidated by Dr. Markham, when this condition is met, industry members would be cognizant of the fact that there is a common course of action that would benefit all of them (Tr. 38 426). The high concentration ratios for the relatively few leading firms in the industry, headed by Kellogg with some 40% of the market satisfy this condition.

188. The second condition:

Entry to the industry must be severely restricted if the price set by the price leader is to remain close to a rationalized oligopolistic price for any significant length of time. If the long-run cost curve for the new entrant is substantially the same as those which confront entrenched firms, price rationalization can be only temporary since the rationalized price wil attract new entrants which, in turn, will bid the price down. If however, the time lag between investment decisions and actual investment in the industry is significant, price rationalization for the duration of the lag may suggest itself as a profitable possibility (KX 104 at 186). 189. In discussing restriction of entry, we are not limited to barriers to entry" as that phrase is used by economists. In this industry, we have had no entry except with respect to granola or natural RTE cereals, which is an exceptional situation (infra Findings 661 , 662). This in itself reflects severe restriction to entry. The extent to which this may be attributable to high costs of entry by an outsider, including costly and time consuming or even unsolvable problems in developing an acceptable competitive product (73)(situations not deemed by some economists to constitute barriers to entry). is immaterial1.

190. Expansion through the introduction of new products by existing firms would not be considered as entry here since the expansion would be by firms theoretically engaged in price coordination through price leadership. This disposes of respondents' argument that Quaker s expansion, as well as the introduction of new products by the several respondents, constituted entry. 191. It is concluded, therefore, that the second condition of Dr. Markham s economic theory is satisfied.

192. The third condition:

The "commodity" produced by the several firms need not be perfectly homogeneous but each producer must view the output of all other firms as extremely close substitutes for his own. If this condition is not fulfilled, each producer is likely to view " In the face of long and coslly effort to errect entry. together with the very real possibility that the new product may nol be accepted byc:nsumers, a potenti,.! enlrant may decide against entry. Even lhough ilis not deterred by a "barrier to entry," the deterrence, neverlheless, isre ai. Further, in lhe face of such high cost, long delays and problematical success, a potentia) entrant might we!! decide against entry since it would reali,." that once it effecte substantial entry, it could nol count On a continuation by existing firms of prices that afforded supracompelitive profit . An oligopolislic price, therefore will nol automatically draw new enlry oince this may we!! nol be the price at which lhe new entrant will anticipate competing.

Initial Decision his product as distinctive in character and the "market" will not be characterized by a single price pahey but by several. Examples of such individual pricing policies may he found in the automobile and brand-name men s clothing markets. Where the output of each firm is differentiated to the extent that it is only a moderately good substitute for the output of other firms, price leadership, of course, is meaningless (KX 104 at 186). 193. This condition is not present in the RTE cereal industry. To the contrary, as I have already found (Findings 59-150), the RTE (74) cereal market is segmented into numerous subcategories, with cereals competing primarily against other cereals in their own categories. And, in many instances, two or three cereals within a segment compete primarily with each other. Indeed, this segmentation or localization of cereals is an essential element of complaint counsel' s contention that respondents have created a barrier to entry of new competition into the industry by reason of their proliferation of brands.

194. Some of the segments and competing cereals previously identified are: corn flakes; wheat flakes; variety packages; Rice Krispies, OK' , Alphabets and Cheerios; nutritional cereals; raisin brans; whole bran; 40% brans; shredded wheat; presweets; presweetened shaped oat cereals; chocolate flavored cereals; fruit flavored cereals; Rally and Life; and Sugar Smacks and Sugar Crisp. 195. Among the various RTE cereals distributed by the respondents and Quaker during the period 1950-1972 are the following: Kellogg: Apple Jacks, Bran Buds, Chocolate Kombos, Cinnamon Frosted Mini-Wheats, Cocoa Krispies, Concentrate, Corn Flakes with Bananas, Froot Loops, Kellogg s All-Bran, Kellogg s 40% Bran Flakes, Kellogg s Corn Flakes, Kellogg s Jumbo Assortment, Kellogg s Raisin Bran, Kellogg s Request Pack, Kellogg s Snack-Pack Kellogg s Variety, Krumbles, OK' , Orange Cream Krunch, Orange Kombos, Pep, Product 19, Puffa Puffa Rice, Rally, Rice Krispies Special K, Stars, Strawberry Cream Krunch, Strawberry Kombos Sugar Frosted Flakes, Sugar Frosted Mini-Wheats, Sugar Pops Sugar Smacks, Triple Snack and Vanila Cream Krunch (Request For Admissions of Fact # 1G-34 Kellogg Company, dated December , 1972; Kellogg s Admissions of Fact # 1G-34, dated February 16 1973).

General Mils: Alive, Big G Goodness Pack, Bran Wisps, Breakfast Squares, Buc Wheats, Buttercups, Carmel Puffs, Cheerios, Clackers, Cocoa Puffs, Cornados, Corn Bursts, Corn Total, Country Corn Flakes, Frosty O's, Fun Pack, Hi-Pro, Hot Toasted 40% Bran, Jets Kahoom, Kix, Lucky Charms, Pick- Pack, Protein Plus, Smiles Sugar Cones, Sugar Jets, Sugaroos, Total, Trix, Twinkles, Unsweetened Lucky Charms, Vital 7, Wackies, Wheat Hearts, Wheaties FEDERAL TRADE COMMISSION DBCISJONS Initial Decision 99 F. Wheaties Bran with Raisin Flavor Flakes and Wheat Stax (Request For Admissions of Fact # 9-53-General Mills, Inc., dated December , 1972; General Mills' Admissions of Fact # 9- , filed February 20 1973).

General Foods: Alpha-Bits, Bran Flakes, Bran & Prune Flakes Cinnamon Raisin Bran, Corn Crackos, Corn Fetti, Corn Flakes & Blueberries, Corn Flakes & Peaches, Corn Flakes & Strawberries Count Off, Crispy Critters, Fortified Oat Flakes, Frosted Rice Krinkles, Grape Nuts, Grape Nuts Flakes, Heart of Oats, Honeycomb, Oat Flakes, Pebbles, Pink Panther Flakes, Post Tens, Post Toasties, Puff Corn Flakes, Raisin Bran, Rice Krinkles, Sugar Coated Corn Flakes, Sugar Crisp, Super Orange Crisp, Top 3 and Treat-Pak (Request For Admissions of Fact # 7 -34-General Foods Corporation, dated December 4, 1972; General Foods' Admissions of Fact # 7- , filed March 6 1973). (75) Quaker: Cap n Crunch, Cap n Crunch's Crunch Berries, Cap Crunch' s Peanut Butter Cereal, Diet Frosted Puffed Rice, Diet Frosted Puffed Wheat, King Vitaman, Life, Oat Flakes, Quake Quaker Puffed Rice, Quaker Puffed Wheat, Quaker Shredded Wheat and Quisp (Request For Admissions of Fact # 12-20-The Quaker Oats Company, dated December 4 1972; Qi.aker s Admissions of Fact # 12- , dated March 14, 1973).

196. The failure to establish condition three is apparent from the very existence of these many different brands of cereals, each extensively advertised to extol their differences to the consuming public. And Dr. Markham testified that the degree of differentiation among RTE cereal products was such that his condition three has not been met (Tr. 38 433).

197. As Dr. Markham has testified, his model cannot be used to anticipate price leadership in differentiated product industries. (HJeterogenous products imply a certain heterogeneity in prices; and that in and of itself vastly complicates any process of havingarriving at price patterns so parallel that they would be those one might expect in lieu of an overt agreement." In an industry where products are close substitutes, products will be purchased purely on the basis of price, and competitors are in a position to follow price leaders. In an industry with differentiated products, it is no longer clear what there is to be coordinated. Price can no longer be identified and isolated for particular products as something the sellers can focus on for purposes of coordination (Tr. 38 430-31). 198. The fourth condition:

The elasticity of the market demand schedule for the output of the industry. Initial Decision must not greatly exceed unity. If demand for t.he output of the industry is elastic because the oligopoly is only a segment of a larger monopoJisticaJly competitive m.arket, the prices of closely competing products severely limit or possibly even eliminate the gains to be derived from adopting a price leader. Moreover, if demand for the output of the oligopoly is highly elastic, firms are not likely to adhere to the price leader s price if to do so would result in substantially less than capacity operations, since each firm could still stimulate its own sales considerably by lowering its price, even though all other firms met the new price (KX 104 at 186-87). (76) 199. An inelastic demand means that a price change will not significantly affect the quantity purchased (Tr. 27 809- , 38 435- 37). An industry at unity would be one, for example, where a 5% change in price would affect purchases by about 5% (Tr. 27 809-10). 200. While a coefficient of price elasticity of demand for the industry has not been developed on this record, it would appear that the demand for RTE cereal products is relatively price inelastic (Tr. 814, 28 000- , 28 238). The uniqueness and price insensitivity of RTE cereals in relationship to other products has been described above (Findings 48-56). General Foods decision makers believed that consumer demand for RTE cereals is relatively price-inelastic (GFPF 509).

201. Respondents' contention (KPF 3-69; GMPF 2-44 thru 2-49; GFPF 7-628 thra 7-630) that RTE cereals are price clastic in relationship to all other breakfast foods is not supported in the record. To the contrary, the record establishes that the fourth condition is met.

202. The fifth condition:

Individual-firm cost curves must he sufficiently similar so that some particular price allows all firms to operate at a satisfactory rate of output. If, for example, the industry is composed of several high-cost low-capacity firms and several low-cost highcapacity firms, the resulting conflct in price and output policies cannot be resolved by adopting a price leader so long as all firms remain in the industry. Low-cost firms will not accept the price leadership of high-cost firms since there is a better option in the form of a lower price and a higher rate of output open to. them (KX 104 at 187). 203. The record contains no cost studies. Complaint counsel (CPF 73) rely upon the testimony of Dr. Scherer to the effect that product technology and equipment were equally available to each respondent, raw materials (corn, oats, etc.) were purchased by each on the commodities market at the same quoted prices and each could utilize advertising agencies and procure advertising space and time on substantially equivalent terms (Tr. 27 814-16). Also, when estimating the costs of other producers, respondents assumed that the costs were equal to their own (CX-GM 171A; CX-GF 4039Z-91; CX-K 443A). This inference could be made, however, only when respondents were manufacturing similar products (Scherer 28 633). Initial Decision 99 F. 204. In light of the large number of different products using different basic ingredients, different manufacturing processes and (77)different production lines, there is no basis for complaint counsel' s assumption that condition five has been met. Complaint counsel themselves, in the course of meeting respondents' contentions as to Nabisco s profitability, assert (CRPF 11-62) that there is no basis for assuming Nabisco had the same capital-to-sales ratio as its competitors; that Nabisco could have different technology and older assets with resultant different capital intensities. And complaint counsel argue (CRPF 11 246) from the possibility that General Foods' cost per pound in relation to its competitors may be low. 205. Kellogg (KPF 3- , 3-72) relies upon cost/sales figures derived from KX 97 which would show that General Foods was the high cost producer followed by Kellogg and General Mills in that order. General Mills (GMPF 2 , 2 153) and General Foods (GFPF 7-632) rely on GMX 553 which shows that Kellogg was the high cost producer, followed by General Mills with unit costs as much as 25% lower than Kellogg s, and then General Foods with costs 33% lower than Kellogg s and 8% lower than General Mills 206. Complaint counsel (CRPF 114) would discount these exhibits because of their inconsistent results. The burden of proof as to costs however, rests with complaint counsel and, while inconsistent, the only cost analyses in the record show that respondents' costs differed.

207. Complaint counsel (CRPF 8-115 thru 8-117) would also fault KX 97 and GMX 553, because costs as a percentage of sales shown by these exhibits could be accounted for by respondents having different selling prices. They would also fault GMX 553 because the firms differ in their relative order on costs depending upon whether the ratio used is cost/dollar sales or cost/pound sales. These objections however, merely confirm the fact that respondents are selling different products, having different weights, at different prices; and that complaint counsel's inability to demonstrate that Dr. Markham s condition five has been met is a corollary of the fact that condition three has not been met.

208. Dr. Markham s widely regarded model of price leadership in lieu of overt agreement requires that all five stated conditions be met. The record clearly establishes that condition three (that each producer must view the output of all other firms as extremely close substitutes for its own) is not present and that complaint counsel (78) 22 The President or General Foos Testified that Genera! Foos W!L'; a high cost producer (Tr. 36 369), as did iL Breakr..st Food Manager (Tr. 13 984) '" It cost General Mills more to produce Country Corn l"lakes than it cost Kellogg to produce Kellogg s Corn l"lakes(Tr. 748-49) InlLlal VeClSlOTI have failed to demonstrate the meeting of condition five (the similarity of the firm s cost curves). The RTE cereal industry, therefore, has not been shown to be one where price leadership in lieu of overt agreement is to be anticipated. Dr. Markham s model, well-reasoned as it may be, is, of course merely an economic theory as to what may be expected. Even if all conditions of the model had been met, it would not substantiate a finding that the respondents in fact coordinated prices by reason of price leadership and price followership. Complaint counsel would still be required to prove that situation by the acts and practices of the individual respondents. And they are not precluded from that attempt because the industry does not fit the model. Nevertheless respondents' acts and practices, as evidenced, must be evaluated in light of the characteristics of the industry and the inconsistencies of those characteristics with the conditions of Dr. Markham s model. It is alleged that respondents, under the principles described above, tacitly colluded or arranged to avoid a wide range of competitive activities that would threaten to reduce prices and the resultant monopoly profits which respondents allegedly achieved. We shall now consider the various competitive activities alleged by complaint counsel to have been subverted by respondents to following "tbe rules of the game" (CP 13&-38).

B. Monitoring of Competitors ' Activities 209. Each respondent closely monitored the activities, including pricing actions, of all other RTE cereal producers. Each respondent published and delivered list prices for its RTE cereal products to customers in advance of the actual price increase dates (Tr. 7522 7841--2, 13 221 , 13 705; See CX-K 766 thru 869; CX-GM 361 thru 504; CX-GF 214 thru 320). Through its monitoring efforts, each respondent knew, usually in advance of the effective date of the price increase, when, and the extent to which, another respondent was changing prices (Tr. 7521 , 7887, 14 250). 210. Respondents used several methods for monitoring purposes including their own personnel and market reporting services such as AC. Nielsen and Selling Area Marketing, Inc. (SAM!) (Tr. 7472-81 8075-76; CX-K 7B, llc, 13A, 15A, 16, 20B, 24B, 25B, 26B, 27B, 37B 38A, 48B, 76 88C 916A; KX 14; CX-GM 93, 281; CX-GF 1593). Such monitoring allowed each firm quickly and accurately to obtain detailed information about the actions of the others in such areas as pricing, new product introductions, product changes, advertising, promotional programs and shelf space activities. This information Initial Decision 99 F. along with other considerations, was used by each respondent in planning its future conduct with respect to the development marketing, sales and pricing of RTE cereals (Tr. 7523, 7525- 841-43, 12 319 , 12 321 , 12 730- , 12 737 , 12 927- , 13 017- 045, 13 200, 14 172, 14 231- , 14 270, 14 282, 14 346, 14 364 336-38 010; CX- 944C; CX-GM 81). (79) 211. Such monitoring of competitors' activities is consistent with alleged efforts to coordinate competitive practices, including pricing. It is also consistent with an effort to react in a competitive fashion to the competitive activities and strategies of others in the industry (Tr. 433- , 38 606).

212. For example, General Foods' technical research department monitored competitors' new and established product activities in order to remain aware oftechnical advances (Tr. 13 641-42). 213. In order to continually assess the competitive environment General Foods purchased information on other companies' activities and on consumer preferences from A.C. Nielsen, SAMI, Market Research Company of America (MRCA), and the National Purchase Diary (Tr. 14 349).

214. Data received by General Foods from MRCA and National Purchase Diary gave profiles of just what categories of persons purchased particular products. This allowed researchers to determine the number of families purchasing a product, the number and frequency of purchases, the demography of buyers and an analysis of purchasing patterns over time ('fr. 14 364-65). 215. General Foods carefully monitored and tried to assess the impact of all competitive products on its line of RTE cereals. The introduction of a new product by one of its competitors would inaugurate an in-depth analysis of that product. The technical department would undertake to analyze the cereal as to its ingredients and its claims, and then perhaps run a consumer test. The new product' s price was considered to have a potential impact on the market and was, therefore, compared to established products' prices. General Foods' researchers would ask Nielsen to track distribution of the product in the United States, as well as its sales and pricing trends (Tr. 14 219-21).

Price Competition Complaint counsel assert that elimination of price competition in the industry is shown by the industry s high rate of return and price/cost margins, both of which were substantially higher than those of most other manufacturing industries; and by an analysis of Initial Decision price trends showing that the prices for RTE cereals rose faster than those for other food products and that price decreases for RTE cereals were practically non-existent (CPF 8-9). First, however, we turn to the more direct evidence bearing on respondents' pricing activities.

216. We first consider the methods whereby the respondents set their prices. Top executives of all respondents have denied the (80) existence of any agreement among the respondents concerning price ('fr. 29 918, 35 462, 35 815, 36 552). Such testimony of Kellogg and General Mills officials falls within that which I am bound to accept at face value (see, supra pp. 13-16). And, of course, if Kellogg and General Mils had no such agreements with General Foods, General Foods had no such agreements with Kellogg or General Mills. Unless overcome by more persuasive evidence, I am required to accept such testimony as accurately describing the situation. 1. Pricing Policies (a) Kellogg Since Kellogg is the alleged price leader, we shall first consider its method of pricing.

217. Kellogg has not set prices in order to maintain a profitable price structure for its RTE cereal competitors (Tr. 29 928). 218. Among the most important factors considered in Kellogg pricing decisions were production costs. During the period covered by the complaint, Kellogg conducted pricing meetings attended by senior executives at which it considered the basic problem of what to do about increasing costs, including raw materials, packaging, labor and overhead (Tr. 29 913- , 29 993, 30 102, 30 115). The same considerations entered into Kellogg s determinations of the appropriate prices for new products it was introducing (Tr. 13 200-01). 219. Kellogg had a guideline for gross margins on RTE cereals, as well as a target net profit figure (Tr. 30 060, 30 063). Through experience, Kellogg had learned that maintenance of its gross margin guidelines for individual brands would generate sufficient revenues to cover costs and still meet its net profit guideline for each brand (Tr. 30 006). In these pricing meetings, therefore, it was understood that price increases should maintain profit margins on the products (Tr. 30 073-74).

220. Decisions to increase price generally covered a group of cereals and did not always include those cereals which were experiencing the largest cost increases. The final decision to increase thatprices was based primarily on the "total amount of revenue" Initial Decision 99 F. would be generated if the increases were taken. At the same time, an analysis was made on a product-by-product basis (Tr. 30 103-04). The Kellogg executives considered the impact that the changes in input costs would have on the gross margins for particular brands (Tr. 914). Some brands fell short of margin guidelines and others exceeded them (Tr. 30 013-14). Products chosen for the largest price increases were those that were expected to maintain their sales pattern despite increased prices (Tr. 30 063-69). (81) 221. Both Kellogg s market research department and its comptroller s office prepared memoranda for use at the pricing discussions. The comptroller s memoranda detailed, on a product-by-product basis, the present and proposed Kellogg sellng prices, the present and proposed retail selling prices, the date of the last price increase, the package size, the proposed price increase per dozen, the calculated percentage increase, and the selling prices of other producers' brands; also, the amount of additional revenue that would be generated if the particular price increases were made (Tr. 12 936 070-74 30,103-04).

222. When cost increases were incurred, Kellogg could either maintain its price on a particular brand and accept a reduced margin of profit or raise the price of the brand. " If Kellogg felt that a prospective price increase would have a sufficiently adverse impact on sales, it would not increase price, but would accept a lower margin. On the other hand, if Kellogg felt that a price could be increased without a significantly adverse effect on sales, the price of the brand in question would be raised (Tr. 29 914, 29 917, 30 068). 223. To estimate the potential effect of a price increase on the sales of a particular brand, Kellogg examined the prices of products which competed with that brand. Kellogg recognized that pricing a brand above competition could adversely affect sales (Tr. 29 919). For brands with direct competitors, Kellogg had to exercise care in raising prices. For brands with no direct competitors, Kellogg had greater pricing flexibility (Tr. 30 059). 224. When attempting to estimate the impact of a price increase on sales, Kellogg considered the recent sales trend of the brand in question, sales trends of competitive brands and the recent pricing activity of competitive brands (Tr. 12 933- , 12 941, 30 102). It also considered what competitors might do if Kellogg changed its prices (Tr. 30 056).

225. Kellogg considered the effect that changes in the wholesale prices of its cereals might have on retail prices. Kellogg was " At times, Kellogg wah able to preserve the margin of a particular brand in the face of rising costs by substituting acceptable less costly inputs(Tr 29 915) InlLial t.l::IUII concerned particularly about establishing wholesale prices that would lead to retail prices that would surpass "psychological barriers" (such as fifty cents per box, or one dollar per box), which might present a significant deterrent to sales ('1r. 12 939, 29 917). (82) 226. When sales of a particular brand had fallen or had not grown as expected, Kellogg might consider a price reduction. The deciding factor was whether the price reduction would increase sales volume sufficiently to offset the resultant lower per unit margin. The purpose of any price reduction, like any price increase, was to increase profits (Tr. 29 920).

227. List price decreases, however, were seldom used. Coupons were the most frequently used means by which Kellogg effected short-run price savings on selected RTE cereals (Tr. 13 041). Unlike trade deals, which are short-term price reductions to grocers Kellogg felt assured that coupon price reductions would reach the consumer (Tr. 12 545). Kellogg believed that price reductions did not stimulate sales among regular heavy users of a particular brand whereas couponing effectively granted price reductions to the marginal consumer who might not otherwise purchase the couponed brand (Tr. 12 289, 12 545, 29 926-27).

228. If competitors did not follow a Kellogg price advance and the resultant price differential adversely affected Kellogg s sales, consideration would be given to taking appropriate action ('1r. 30 106-07). 229. In 1966, Mr. Charles Tornabene, then Kellogg s Vice President and Domestic Sales Manager, made a speech at an annual regional meeting of Kellogg offcials. In it, he described Kellogg as the leader in the industry (CX-K 549L) and stated: leader should do all in its power to build and expand its industry sales and avoid any steps which will drag the industry down. A leader must maintain a profitable price structure within the industry-both for its members and its distributors. Only a strong company can afford to exercise restraint when it is needed to keep an irritable condition from deteriorating into a war that no one wins. Only a strong company can set the pace that provides a favorable climate for a strong and growing industry. Only by disciplining ourselves can we set the example for a disciplined industry (CX-K 540M, Q).

He also quoted with approval the comment of one whom he identified as a well-qualified observer: (83) lTJhere is no area in the food business today in which the true qualities of industry leadership are more aptly displayed than in the cereal industry where Kellogg provides strong and consistent leadership in building and expanding the profitable Initial Decision 99 FTC. climate of true growth, virtually free from destructive pricing and promotional practices (CX-K 549P).

230. In addition, Mr. Tornabene stated that " Kellogg has a long history of consistently resisting price . cutting and gimmicks and withstanding competitive pressure in these areas;" and that "where it was necessary to participate-overwhelmingly-in order to put an end to destructive practices," Kellagg has done so and would continue to do so (CX-K 549M).

231. This speech reflects Mr. Tornabene s and Kellogg s belief that, as the industry s leading competitor, it would not serve Kellogg s interest to engage in destructive pricing and promotional wars. It does not indicate the existence of an agreement between Kellogg and its competitors; and even if it did, Mr. Tornabene statement could not be used against the other respondents as evidence of a conspiracy since that statement was not made in furtherance of a conspiracy, and the existence of a conspiracy has not otherwise been established (see supra p. 19). 232. In the face of Kellogg s President's sworn testimony that Kellogg has not set prices in order to maintain a profitable price structure for its RTE cereal competitors ('fr. 29 928), and in the absence of any other evidence that would tend to show that Kellogg acted to help its competitors realize profits, Mr. Tornabene generalization in this area is entitled to little weight. Contrary to the factual conclusion complaint counsel would draw from Mr. Tornabene s statement, both General Mills and General Foods were constrained by Kellogg prices from raising prices on their own prod ucts.

(b) General Mills 233. As with Kellogg, a primary determinant of price, both for new and established General Mills products, was input costs ('fr. 604, 16 609 524).

234. General Mills also considered the prices of competitive products in reaching pricing decisions. This was to avoid pricing its products out of line with competition, which might result in deterring sales (Tr. 16 604- , 35 653-54). (84) 235. General Mills considered the past pricing activities of General Foods and Kellogg. Pricing data considered included pricing histories, analyses of price changes by all three respondents including price change effective dates, brands affected and the amount of revenue generated by increases (Tr. 17 331-34; CX- , 168, 2007B, 2012A).

Initial Decision 236. General Mills had its field personnel promptly report competitive price change information (CX-GM 186, 194, 281 , 283). Price changes were often reported by telegram (CX-GM 199, 201 205 249 250-51).

237. By 1967, General Mills' average price per pound had risen by $. 095 over 1962, the highest absolute increase of the six largest RTE cereal companies. By 1970, General Mills' average price per pound had increased by $. , or 52. 1 %, a much larger increase than that of any other major RTE cereal manufacturer. From 1968 on General Mills' average price per pound was higher than that of any of its competitors except for Quaker (GMX 560A; GFX 1151 , 1152). 238. Because it saw itself in a premium priced position, General Mills was reluctant to initiate desired price increases for fear of substantial sales losses. Reporting to the head of General Mills cereal department in March 1967, Mr. Schulze (then in charge of Adult Cereal Marketing) concluded that General Mills was "not in a position to advance prices" on its adult cereals because of three factors: (1) recent price increases on these cereals; (2) current prices on these cereals were equal to or above their "direct competition and (3) "Kellogg is still the clear leader in the cereal industry, and with our current premium-priced position, we do not feel that we could lead a price advance" (CX-GM 280A). 239. The competitive constraints brought to bear by General Mills' perceived position as a premium priced manufacturer are exemplified in pricing decisions involving Trix and Lucky Charms. These brands were recognized as "premium priced products. " In March 1967, Mr. Bodeau, a General Mills official, recommended against a price increase on these brands because they were already priced higher than most presweets. "We believe increasing our premium difference would be unwise and result in volume losses on these products" (CX-GM 278B).

240. The prices of Trix and Lucky Charms were not increased at that time. The price of Trix was not increased until November 1967 (CX-GM 477); and Lucky Charms' price was not raised unti January 1969 (CX-GM 487). In the interim, General Foods had increased the prices of Honeycombs and Super Sugar Crisp in June 1967 (CX- 304), Kellogg had increased the price of Sugar Smacks and Sugar Pops in August 1967 (CX-K 858), and General Foods had again increased the price of Super Sugar Crisp, along with Alpha-Bits, in September 1967 (CX-GF 305).

241. In analyzing the possibilty of raising prices for its "child cereals" in mid-1967, General Mills compared its prices with (85) those of its competitors and evaluated the "price changing history. Initial Decision 99 FTC It concluded that, because it expected to lose too much volume if it raised prices. while its major competitors did not, there was " price advance possibility unless our major competitors initiate a change" (CX-GM 278A, C-D).

242. It is concluded that the record reflects genuine independent business reasons why General Mills did not originate price changes to a greater degree than it did and followed price increases of its competitors.

(c) General Foods 243. Pricing questions arose at General Foods as a part of the annual budgeting and marketing planning process. Costs and market conditions were evaluated and pricing assumptions were factored into annual product plans. In addition, decisions were often made during the year, as necessary (Tr. 13 984, 36 396, 36 400-1). 244. General Foods, as did Kellogg and General Mills, viewed input costs as a prime determinant in its pricing decisions (Tr. 984, 36 400-01; CX-GF 4039Z-23). As a much smaller producer than Kellogg and General Mills, General Foods believed that it had smaller profit margins and that it was under greater cost pressures and urgencies to increase prices in response to cost increases (Tr. 984). General Foods, therefore, sought opportunities to increase prices (Tr. 36 369, 36 574- , 36 400-Dl). 245. Nevertheless, General Foods could not unilaterally raise its prices on brands which faced direct competition. It believed, for example, that initiating price advances on Post Toasties or Post Raisin Bran which put those products at a price disadvantage vis-avis their direct competitors (Kellogg s Corn Flakes and Kellogg Raisin Bran, respectively) would result in significant share swings away from the Post brands (CX-GF 2022E, 3000Z-23c, 4039Z-79). Despite increasing costs, General Foods had to wait for larger companies to move before it could act to meet its own profit margin goals (Tr. 13 984). Accordingly, a 1967 General Foods Task Force recommended that General Foods "laJssume the position of a follower in pricing action that occurs in the industry recognizing that price increases will not be followed automatically unless such action is justified by cost (emphasis supplied) (CX-GF 4039P). 246. General Foods' overall policy was to "Price directly competitive generic brands Corn Flakes, Bran Flakes equal to prime competition on a price per ounce basis" (CX-GF 4039P). Further, its intent was to price such cereals as high as possible within the constraints imposed by its competitors. It became a follower rather . .

Initial Decision than a leader (CX-GF 17L, 418A, l408E, 2033J, 3000Z-193, Z-216 4033H). (86) 247. The record is replete with instances where General Foods policy to follow rather than initiate pricing action was implemented with respect to particular brands General Foods' Raisin Bran relationship with Kellogg s Raisin Bran (CX-GF 345E, 440J 1996C); General Foods' Alpha- Bits' relationship with Kellogg s Froot Loops and General Mills' Trix and Lucky Charms (CX- GF 1382U); General Foods' Pebbles ' relationship with leading flavored brands (CX- 1420C, 2021H); General Foods' Bran Flakes' relationship with Kellogg s Bran Flakes (CX-GF 1989B); General Foods' Sugar Crisp relationship with Kellogg s Sugar Smacks (CX-GF 141OL). General Foods' Post Toasties ' relationship with Kellogg s Corn Flakes (CX- GF 2022E 2024C); General Foods' Pink Panther Flakes' relationship with "other kid presweet brands" (CX-GF 1435Z-38); General Foods cereal assortments' relationship with Kellogg s cereal assortments (CX-GF 2008B, 2009C).

248. General Foods' policy was to " r r )ecognize through pricing any unique attribute or quality which a Post product offers the consumer higher than average prices will be charged for products which are unique in convenience or benefit" (CX- 1300M). However, even these relatively exclusive brands had to be priced "within a generalized area of reason relative to other cereals (Tr. 36 403). As stated in its 1967 Task Force report, General Foods sought to "lp lrice brands having no direct competitor to maximize gross margins after recognizing the retail price range for brands within the same category" (CX-GF 4039Q).

249. General Foods' 1967 Task Force sought to raise the profit margin of brands such as Post Toasties, Bran Flakes, Raisin Bran and Grape Nuts Flakes, which were subject to direct competition (CX-GF 4039Z-104). However, it was unsuccessful in this effort (CX- GF 3000Z-23c). General Foods, despite its desire to avoid initiating price increases, was forced to initiate a price increase on Raisin Bran in June 1969. Kellogg did not increase the price of its Raisin Bran until December of that year (CX-GF 3000Z-23c). The consequent price disadvantage resulted in a substantial sales swing to Kellogg (CX-GF 3000Z-23d). This experience was cited by the 1971 Task Force as evidence " that brands with direct competitors. witness volume swings with price differentials.. . ." Consequently, the Task Force recommended that, on such products, General Foods should follow the market leaders or direct competitor whenever possible (CX-GF 3000Z-193).

250. General Foods has evaluated the possibility of lowering FEDERAL TRADle COMMISSION DECISIONS Initial Decision 99 F. prices-the issue being whether price reductions would generate sufficient additional sales to offset the lower margin General Foods would receive on each sale (CX-GF 571K, S, Z-1, Z-6, Z-15 thru Z- , 4039Z-106, Z-109). A 1970 report was particularly interested in and analyzed, the pricing alternatives (upward and downward) for Post Toasties (CX-GF 2022E).

251. After an internal assessment, a member of the 1971 Task Force rejected the idea of an across-the-board price reduction, (87) because it would require an unrealisitc increase in sales volume to offset the decreased margins resulting from the price reduction. Given our current margin situation a reduction in price (as a marketing strategy) does not appear to be a realistic opportunity for Post (CX-GF 3000Z-193). Two price increases on Post Grape Nuts, in June 1969 and in November 1970, caused substantial share losses (CX-GF 3000Z-23c Z-180).

252. Categories in which General Foods faced direct competitive pressure produced substantially lower variable gross profit rates than categories in which competition was less direct. A low variable gross profit rate in the corn flakes category was attributed to the reluctance of Kellogg s and Post to make major pricing moves" (CX- GF 3000Z-68), General Foods lowered its price on Post Toasties to meet a Kellogg price decrease on Corn Flakes in July 1971 (CX- 3000Z-121, Z-177).

253. General Foods' overall margins were lower in FY 1973 than in FY 1969 (CX-GF 3000Z-261). This situation was attributed three factors, the first two of which were: (Post ToasticsJ price decline.

Inability to take planned pricing action (CX-GF 3000Z-179). These factors reflect competitive constraints that affected General Foods' pricing decisions (CX-GF 3000Z-261). 254. General Foods' price per pound was less than that of the other respondents (CX 106D, F; GFX 1151-52). 255, It is concluded that the record reflects genuine, independent business reasons why General Foods did not originate price changes to a greater degree than it did and followed price increases of its competitors.

2. Coordination of Prices (a) Brand Price Coordination Initial Decision Dr. James Green, formerly an economist with the Commission, was presented by complaint counsel to testify with respect to price (88)coordination of particular competitive products. He testified to this end for seven days (Tr. 9687- 501). His direct examination extended for five days (Tr. 9687- 227). 256. Dr. Green, on the basis of his own examination of documents in the possession of complaint counsel, selected products for comparison that he thought were highly substitutable (Tr. 9837, 9897 470). While Dr. Green purported to find pricing coordination among a number of products, a study of his analyses reveals that a pattern consistent with coordination was established only with respect to three sets of products-Kellogg s Corn Flakes and General Foods' Post Toasties; Kellogg s and General Foods' Raisin Brans; Kellogg s Sugar Frosted Flakes and General Foods' Sugar Coated Corn Flakes (Tr. 9 833, 10 130-42, 10 170-83; CX 341; see, supra Finding 106).

257. Dr. Green s failure to present evidence for additional sets of products, consistent with coordination, pricing agreement or price leadership and followership, is reflected by a number of factors. 258. Dr. Green loosely defined !'coordination" in terms of price levels and the amount of price change, and the time lags between price moves of the pairs of products being studied (Tr. 9 833-42). In eval uating price changes, however, he disregarded package size changes of products being compared which effectuated changes in cost per ounce of those products (Tr. 9789 , 10 232- , 10 259--0). Also, because of package size changes, the witness abandoned efforts to compare price levels and relied upon the length of time lags (Tr. 082, 10 085, 10 088). The witness at times relied on timing of price changes, irrespective of price levels (Tr. 10 291-92). Dr. Green finally conceded (Tr. 10 343) that his conclusions and testimony as to coordination had to do with the timing of price changes, not with price levels. (89) 259. However, time lags varied greatly in length, from relatively short periods to periods as long as four to six months and even eight to nine months and, in some instances, there would be no response at all (Tr. 9941 , 9964, 9990, 10 061 , 10 064, 10 081 , 10 087 , 10 142 10,154). Price changes varied so in timing that Dr. Green could not " Other products compared by Dr. Green, but where pri"ing coordin..lion was not establishpd, include GenNal Foos' Grape Nuts and Quaker s Lire ('fr. 9893-9947); Genera! Mills ' Cocoa Puffs and Kellogg s Cocn" Krispies ('fr 9949-9967, 10 377 96); Kellogg s Product 19 and Special K and General Mills' Tnta! (1'r. 9867-- 9990, 10 355-76 10,418-21) General Mills' Frosty D's and Gpneml Foods' Alpha- Hits (1'r. 9992- , 10 422-31) Kellogg s All Stars and General FOHlb' Alpha-Bits (1'r. 10 003-9); Genera! Mils' Whe"ties and General Q(ds' Grape Nul (1'r. 016-4 , 10,il9rr-1O); General oods' 40% Bran lakes and Kellogr., s 40% Bran Flakes (1'r. 10 06::.8); Kellogg All Dr..n and Nabisco s 100% Hran efr. 10092- 122); General Mills' 1'rix and Kellogg s Froot Loops (1'r. 10,14:1- 432- 44); Kellogg s Rice Krispies and General Mills' Ch.'erios (1'r. 10, 149-:,5, 10 276-343). Initial Decision 99 F. tell whether a price change on a product was in response to a prior one on a different product, or was the beginning of a new price round (Tr. 9941-45). Dr. Green disregarded situations where an increase in price was accompanied by an increase in the size of the box, as he felt unable to make a judgment in such a situation (Tr. 10 384, 10 395). Neither did he take into account situations where a price change in one product was preceded or followed by a box size change in another, or a size change in one product was not responded to by a competitive product (Tr. 10 426-30).

260. Apart from the above recitation of some of the defects in the witness' analyses, I have carefully studied Dr. Green s presentation on the issue of brand price coordination. Except for Kellogg s Corn Flakes and General Foods' Post Toasties, Kellogg s and General Foods' Raisin Brans, and Kellogg s Sugar Frosted Flakes and General Foods' Sugar Coated Corn Flakes, there is no evidence of price uniformity, maintenance of pricing levels, or pricing responses consistent with a pricing agreement or arrangement among respondents or with price leadership and followership. 261. In addition to what he presented at the hearings, Dr. Green had prepared comparisons of other sets of products which he thought were comparable. These comparisons were not presented at the hearings because, in Dr. Green s opinion, they did not reveal any apparent pricing coordination (Tr. 10 249). 262. The record, therefore, lacks probative evidence of price coordination between particular brands other than the three sets of products identified above. To the contrary, the record tends to evidence a lack of such brand price coordination or parallel pricing movement.

263. The extent to which the record shows price similarity between the two brands of corn flakes, the two brands of raisin bran and the two brands of sugar frosted or sugar coated corn flakes does not establish any coordination in avoiding price competition. Price similarity, even identity, is to be expected for very similar products even under competitive conditions (Tr. 10,462 813 517 642 371- 579).

(b) Pricing Rounds Coordination Complaint counsel contend that "respondents under Kellogg leadership, coordinated their pricing behavior by engaging in a (90) pattern of price leadership in lieu of overt collusion" (CPF 8-10). Starting with Dr. Markham s model which sets forth the conditions under which such a situation may be anticipated (CPF 8-1 thru 8- , 8-2; CRPF 8-7 thru 8-122), which I have already found does , , Init.ial Decision not fit here (supra Findings 186-208), complaint counsel rely primarily on an analysis of pricing rounds in the industry covering 1965 through 1970, prepared and testified to by Dr. Scherer (CPF 8- 83 thru 8-84). As described by complaint counsel A price round is a series (or group) of list price changes that occur when a firm changes the price of two or more regular size branded products" (CPF 8-83). , it is Before examining in detail the 16 price rounds in question appropriate to compare some of the basic concepts involved in price leadership with what Dr. Scherer s price rounds profess to show. 264. We start with condition three of Dr. Markham s analysisthe requirement that the products of the respective firms be viewed as extremely close substitutes." This requirement is obvious since we are dealing with price leadership and it is necessary to have products sufficiently similar so that price is a primary element and can be coordinated by leadership and followership. As already found except for three sets of products, Kellogg s Corn Flakes and General Food' s Post Toasties, Kellogg s and General Foods' Raisin Brans Kellogg s Frosted Flakes and General Foods' Sugar Coated Corn Flakes, there is no evidence of price uniformity, maintenance of pricing levels or pricing responses at a product level consistent with price leadership and followership.

265. The 16 price rounds, as presented by Dr. Scherer and relied upon by complaint counsel, do not take prices into account other Levels andthan for the fact that price changes were made. magnitudes of price changes are ignored. Indeed, the price round presentation does not even demonstrate that "lead" price changes on particular products were followed by price changes on what may be termed directly competing products. To the contrary, the amount of price change varied by product, and the types of cereals involved in one company s price change varied from those in the subsequent price change of other companies. In short, there is no correlation of individual products or individual product prices in Dr. Scherer price rounds. Indeed, individual produds and prices are not even evaluated. There is, therefore, no showing of correlation of any brand prices or price differentials as to particular brands. (91) 266. General Mills' overall prices in relationship to pounds of RTE cereals sold increased far more rapidly than those of the five other largest factors in the industry:

'" Dr. Scherer conceded that, ror most products there seems to be very littl.. pattern in the relationships between ehan!fes in the price orone product relative lochanges in the price of another " (1'r.27 922) ($) _._ Initial Decision 99 F. AVERAGE PRICE PER POUND SOLD General General Year Kellof; Mils Foods Quaker Ralston Nabisco 1962 392 365 321 .485 288 324 1966 .445 .440 398 524 355 355 1967 .463 .460 .401 .551 366 370 1968 .474 482 392 581 355 378 1969 .496 513 .414 604 382 398 1970 .522 555 .426 633 388 .408 Increase 1962- 130 190 105 148 100 084 % Increase 33. 52. 32. 30.4 34.7 25. Neither Kellogg nor General Foods matched General Mills' overall price increases.

267. This disparity in relative price increases for General Mills vis-a-vis, Kellogg and General Foods cannot be wholly attributed to the possibility that, during the period in question, General Mills may have introduced new, more expensive brands and discontinued older less expensive brands. A comparison of pricing levels limited to brands that were in existence in 1961 shows that General Mills increased these prices overall more rapidly than did its competitors (Tr. 31 702; KX 102).

268. None of the 16 "rounds" of price changes involved list price changes of all of any company s brands (Tr. 27 915; CX 1004). Dr. Scherer originally explained the ilnport of his price round presentation as follows err. 27 828-29):

Now, it is my inference, I don t find exact documentary support for this inference but it is my inference that what is going on is a kind of averaging process. That is to say, one goes into a price round with the expectation of raising a certain amount of (92)additional revenue through the price increase. And one chooses, therefore, a group of products that through which by raising the price, one can, if all goes well, raise that additional amount of revenue.

Now, if one is alone in this price increase, then the products who prices have been increased arc liable to suffer some erosion of market volume. nut, if everybody goes up on a package of products, if everybody takes part in the round and increases prices on a group of products, then the following wil happen: After the round some products prices will not have been increased. Some products prices will have been increased. There will be some tendency for the demand to flow towards the products whose prices have notbeen increased. But if everyone goes up by more or less the same percentage of his total volume, the gainers will more or less on average be offset by the losers. That is to say, the products whose prices have been " Derived from GMX 560A KELLOGG CO., ET AL 101 Initial Decision raised will now be relatively high priced relative to the products whose prices have not been raised.

Each firm wi!! have products whose prices have been raised. Each firm will have products whose prices have not been raised, and on the average, the gains and losses of patronage in this type of group price increase approach wjJ average out if, in fact, the companies tend to move on about the same fraction of their product line. Over the average.

269. Dr. Scherer s approach, therefore, was not to confirm or establish a situation of price leadership where price was an essential element of product competition, but rather to attempt to explain why it was not necessary for the parties to engage in such a practice. 270. Except for a period when General Foods was competing on the basis of the overall General Foods name, all respondents have competed on an individual brand basis, particularly against the competing brands in the pertinent segment or, on an even narrower basis, against individual directly competing brands. I cannot accept in the absence of direct proof, Dr. Scherer s inference that respondents, pursuant to agreement or price leadership and followership, are wiling to suffer an erosion of sales on particular (93Jbrands by raising their prices in the absence of price increases of their direct competitors, in anticipation of securing greater sales on brands whose prices are not increased.

271. Brand competition is much too intense and complicated to assume that, by averaging out the price increase to constitute an equal percentage of total volume, competitive price differentials will not affect different products differently. Therefore, I reject Dr. Scherer s unsupported inference as to the nature of the price leadership and followership existent among respondents. In any event, an analysis of the price rounds demonstrates that they are inconsistent with any such percentage agreement or coordination. 272. In order to have price leadership, the parties must understand the purport of each other s actions. As Dr. Scherer has written: Leadership is a means of communicating, and the process may break down if some messages go unheeded, so that one never knows what the reaction to a new price announcement will be (F. Scherer Industrial Market Structure and Economic Performance 168 (lst ed. 1970)).

273, Keeping this principle in mind, it is noted that Dr. Scherer incl uded in his analysis only "price advances on more than one product unaccompanied by package changes" (Tr. 27 903). According- 2. Even if percentage of Hw line change coordination had been established . this would have ben insufficient proof of price coordination in the absence ofa showing of the levI!ls of price change. , . .

Initial Decision 99 F. ly, he excluded all list price changes by any respondent that applied to only one brand (Tr. 27 827),29 as well as price moves that were accompanied by package size changes (Tr. 27 903). Dr. Scherer explained his reasons for such exclusions: I infer from what I have seen of the processing technology that it is a little (94)bit diffcult mechanically to change a package size. So not uncommonly only one product' s package size is changed at a moment in time and then there may be a price increase accompanying it. And by and large I found that in the company monitoring documents not much attention was paid to those single package size changes accompanied by a price change.

I also observed that hy and large not much attention was paid to single price changes, single product price changes even without a package size change, so I came to the conclusion that occasional single product price changes were really not a part of the standard price leadership-followership pattern. And so I have, for the most part excluded single product moves from my analysis (Tr. 27 826-27). I excluded price moves that were solely accompanied by package changes. That is to say, if there were no changes of packages, package sizes, in the price change, then I excluded that as a round The reason I did that i."i this: That very frequently one finds on package changes that it i."i ambiguous whether a price change has been effected or not. A package change will take place, a price change may take place simultaneously on the price of the package. And it is a little bit ambiguous because it depends upon the relationship of the package increase or decrease relative to the price increase-the case price increase or decrease, whether the effective price has been raised or lowered. And so I found in comprLnY documents that the companies had a little bit aftrouble analyzing just what was going on and how to interpret these package changes, and so I excluded them (emphasis supplied) (Tr. 27 830-31). Q. And are you saying that because a competitor may have difficulty understanding the market implications of a price move of a (95)competitor, that is justification for excluding the price move as a round? THE WITNESS: The word I have trouble with is justification. In that question. That implies some universal set of criteria that aren t defined. It was my best professional judgment that, indeed, those kinds of situations provided much more ambiguous signals than other kinds of situations, and so therefore I did so exclude them.

When you say "ambiguous signals " you mean between and among competitors? '0 An analysis that covered such price ch:mges would have ben instructive since it might reflect the competitive situation on a one-tone basis of directlycompetitive products- Of cours, Dr. Green had already failed toshowa cnrrellition in individual brand pricing(supro, Findings 255-1) Initial Decision A. Price leadership is a method of signalling, and presumably you are signallng between and among competitors, yes.

274. Thus, rather than attempting to account for or reconcile pricing actions which appeared inconsistent with the communication process requisite for price leadership and followership among respondents, Dr. Scherer chose to ignore them. 275. Stil another requisite for price leadership in lieu of overt collusion, to be kept in mind when analyzing the price rounds, is that all respondents must cooperate (CPF 8-7). 276. As part of its price leadership theory, complaint counsel assert that Kellogg is the price leader and that a typical price round begins with Kellogg s announcement that it plans to increase prices on a certain group of products on a particular date (CPF 8-10 8-85). These Kellogg price increases, according to complaint counsel, were expected to be, and in fact were, matched by the other respondents so that approximately the same pricing relationship existed between the firms after the increase as had existed before the increase (CPF 8-21 , 8-5). As previously noted, Dr. Scherer s price rounds analysis does not attempt to demonstrate a brand price relationship. But in our evaluation of the rounds, we must keep in mind the assertion that Kellogg is the price round leader and that the other respondents are the followers. The requirement that the followers follow has been expressed by Dr. Scherer:

Whatever the reasons for a firm s acceptance as price leader, ilc; leadership must be (96)followed consistently and with near unanimity if the industry is to make the most of its market opportunities. Leadership is a means of communicating, and the process may break down if some messages go unheeded, so that one never knows what the reaction to a new price announcement wil be (F. Scherer Industrial Markel Structure and Economic Performance 168 (lst ed. 1970)). 277. Still another requisite of complaint counsel's price leadership-price followership theory is that the price followers "change their prices in consonance with those of the leader" and "thereby reduce. . . the duration of any significant price differentials" (CPF 8-11). This condition is obvious because, to the extent there are prolonged delays in reacting to the price leader, there is an absolute failure of price coordination. Also, from the point of view of the leader, the longer the delay, the less assurance it has that the other firms will follow and this can cause a breakdown of the system. Further, from an evidentiary point of view, the longer the delay by the follower in making its price change, the less clear it is that its price change is in response to that of the leader rather than in response to other competitive conditions (Tr. 32 499). Time delays in Initial Decision 99 F.TC. following price changes by the leader, therefore, are important considerations in evaluating Dr. Scherer s price rounds. Analysis of Pricing Rounds 278. Prior to Dr. Scherer s Round 1 , Kellogg, on January t, 1965 changed prices on 23% of its line (CX-GM 168). '0 The price increase was accompanied by a package size change and so was one of those excluded by Dr. Scherer as "ambiguous" (Tr. 27 830, 28 470). Neither General Mills nor General Foods took any action in response to Kellogg s January 1 , 1965, price change (Tr. 28,470). Round 1 279. Prior to April 8, 1965, Kellogg announced price increases effective April 24, 1965, on cereals representing 24% of its RTE cereal sales volume (CX-K 852; CX-GM 168). On April 8, 1965 General (97JMills issued a price list with increases effective that date on 58% of its line, but allowed customers to purchase at old prices until May 7 (CX-GM 168, 464). General Foods then advanced 28% of its product line, effective May 14 or May 17, 1965 (CX-GF 295; CX- GM 168)" Dr. Scherer s analysis for 1965 shows that the "comparable dollars generated" by General Mills' pricing actions were over 120% higher than Kellogg s and 43% higher than General Foods (GMX 105).

280. Thus, there is no coordination as to share of line moved or comparable dollars generated in Round 1.

281. Between Rounds 1 and 2, Kellogg increased its price on a single product, Froot Loops (CX-GM 168) and General Mills, on June , 1965, increased its prices on 50% of its line, but also made a package size change. In accordance with his guidelines, these price changes were ignored by Dr. Scherer (Tr. 38,466-67). Round 2 282. Kellogg, on December 9, 1965, announced that prices on 34% of its product line would increase effective January 22, 1966 (CX-K 766, 853; CX-GM 168). On January 10, 1966, General Foods announced advances on 120/0 of its line to be effective February 14 1966 (CX-GF 298, 571Z-21). By December 13, 1965, four days after Kellogg s announcement, General Mils analyzed the additional '" The volume percentages used lhroughout t.his price round analysis refer to dolllH stiles volume (Tr. 27 861). J, It isa"slImed in all instances th"t each respondent, through il. monitoring activities, was aware of each price change in the industry soon "after il. announCt'ment. Evidence relating to particular instances of acquiring prieechange information, therefore, is not being referenced Initial Decision revenues to be generated by Kellogg s move (CX-GM 26). General Mils, on January 28, 1966, increased prices on 22% of its line effective that date, but with price protection available until February 22 (CX-GM 168, 468, 2487).

283. It is noted, however, that Dr. Scherer did not take into account two price changes by Quaker, in November and December 1965, both before Kellogg s announcement of its price increase (CX 1005). Again, the shares of respondents' product lines covered by the price changes varied considerably.

284. Dr. Scherer ignored a Quaker price move between Rounds 2 and 3-a price increase announced February 21 , 1966, effective March 14, 1966 (CX 1004). (98) Round 3 285. Kellogg, on March 4, 1966, announced that price increases affecting 48% of its line would be effective March 26 (CX-K 768; CX- GM 92, 173). Two and one-half months later, on May 17, 1966 General Foods announced increases on 36% of its line to be effective June 20, 1966 (CX-GF 300, 571Z-13; CX-GM 168). General Mills did not act unti June 30, when it announced price hikes on 37% of its dollar sales volume, with price protection available until July 23 (CX-GM 168 470 2488).

286. Kellogg s move in Round 3 was closer in time (six weeks) to General Mils' move in Round 2 than it was to either General Foods or General Mils' moves in Round 3 (two and one- half and four and one-half months, respectively). General Mils' move is actually closer to Kellogg s move in Round 4. Therefore, it is not dear where the round lines should be drawn and who is the leader and who is the follower. Again, General MiUs' and General Foods' percentage of line covered differed from that of Kellogg.

Round 4 287. On September 2, 1966, Kellogg announced advances effective October 1, 1966, on 61 % of its product line (CX-K 773, 856; CX- GM 168). General Foods announced, on September 28, 1966, that increases on 59% of its line would be effective October 24 (CX- 571Z-8, Z-9; CX-GM 168, 473). On October 24 or November 7 General Mils advanced 53% of its line effective that date, but allowed customers to buy at the old price until November 14 (CX- GM 168, 473). General Mils had previously analyzed Kellogg increases and estimated that they would provide Kellogg with $5. milion in additional annual revenues (CX-GM 175). Initial Decision 99 F. Round 5 288. General Foods advanced prices on 9% of its product line effective June 26, 1967 (CX-GF 304). This action was not followed by any of the other producers. Nevertheless, Dr. Scherer included it because it satisfied his basic criteria for classifying a price move as a round a price advance on more than one product at the same time (Tr. 27 825, 27 830-34). Dr. Scherer, however, ignored a price increase by General Foods on June 1, 1967, on 8% of its product line because it only involved one product (Tr. 38 467). 32 (99) 289. This solitary move by General Foods is totally inconsistent with a Kellogg price leadership premise.

Round 6 290. Kellogg announced on August 4, 1967, that price increases on 20% of its sales volume would be effective on August 26 (CX- 777, 858; CX-GM 168). On August 29, General Foods announced price hikes on 20% of its volume to be effective September 25 (CX- GF 305, 353K, 571Z-2; CX-GM 168). General Mils made no price move (CX 1005).

291. Thus, we have General Mils making no price changes in two consecutive rounds.

Round 7 292. Kellogg, on October 27, 1967, announced that increases on 41 % of its line would be effective on November 25 (CX-K 779, 859; CX-GM 168). General Foods acted first after Kellogg s increase by announcing on November 17 that price advances on 53% of its line would go into effect on December 18 (CX-GF 307, 571Y; CX- 168). General Mils announced on November 20 that increases on 67% of its line would be effective immediately, but customers could buy at the old price until December 9, 1967 (CX-GM 168 477 2491). General Mils had previously analyzed Kellogg s advances and calculated that this increase would provide Kellogg with an additional $3.7 milion in annual revenues (CX-GM 180). 293. Here, we have General Mils making its first price move in over a year after having failed to participate in price Rounds 5 and 6. 0' Dr. Scherer also ignored single product price moves by Kellogg, C'.-meral Foos, Quaker and Nabi!jo (1' 467) , Initial Decision Round 8 294. Kellogg on December 8, 1967, announced price advances on 32% of its product line effective January 6, 1968 (CX-K 780, 860; CX-GM 168). General Mills, on January 8, 1968, announced price increases on products constituting 24% of its product line, effective immediately, but with price protection offered until January 27 (CX- GM 168 478 479 2492). General Mils had previously analyzed this increase and predicted it would provide $3.5 milion in additional annual revenues to Kellogg (CX-GM 181). General Foods did not participate in this price round. (100) 295. Kellogg s December 8, 1967 lead" announcement was shortly after General Mills' and General Foods' price announcements in Round 7 and close to the effective dates of those price changes. Therefore, calling Kellogg s December 8, 1967, announcement a price lead is somewhat arbitrary.

Round 9 296. On April 25 and May 15, 1968, General Foods announced price increases effective, respectively, on May 13 and June 3, 1968. The May 13 increase alone covered products totalling 17% of its line (CX-GF 309, 310, 571U; CX-GM 168). Neither General Mills nor Kellogg participated in this so-called round. 297. Since General Foods' second price announcement followed the effective date of its first, this really constitutes two price leads by General Foods, neither of which was reacted to by General Mills or Kellogg.

Round 10 298. On August 2, 1968, Kellogg announced increases on 45% of its product line effective on August 3, but with price protection to its customers at the old price until August 31 (CX-K 782, 861; GX- 168). General Mills analyzed these price moves and calculated that they would provide Kellogg with an annual increase in margins of about $5.3 million (CX-GM 186, 252, 253). General Mils believed that the important aspects to the advance were: (1) the total dollars generated, (2) the percentage of Kellogg s line that was affected, and (3) the time intervals between price changes (CX-GM 194 , 252). However, General Mils did not participate in this round. Neither did General Foods.

299. This gives us two consecutive "rounds, months apart, each consisting of a price increase by a different respondent, but where Initial Decision 99 F. that respondent acted alone. Round 10 is particularly significant since it consists of Kellogg raising prices o,! 45% of its product line with neither General Foods nor General Mills reacting. Round 11 300. General Foods announced on November 27, 1968, that it was initiating increases on 57% of its line effective December 30 (101) (CX-GF 313, 5710-P; CX-GM 168). Kellogg, on December 3D 1968, announced advances on 57% of its line to be effective on February 1, 1969 (CX-K 786, 863; CX-GM 168). 301. On January 10, 1969, General Mills announced advances on products accounting for 74% of its line, with price protection offered until February 1 (CX-GM 108, 2493, 2494). General Mills' action followed an analysis of the additional revenues Kellogg and General Foods would generate by their price increases, and General Mills personnel noted that its pricing action followed Kellogg price advances in August 1968 and January 1968 and a price advance by General Foods in December 1968 (CX-GM 31, 35, 110A, 168, 187 566!).

Round 12 302. Kellogg, on May 9, 1969, announced that prices on brands representing approximately 25% of its line would be increased effective June 14 (CX-K 787, 865). On May 29, General Mils advanced prices on a number of its own brands effective June 2, but with price protection offered until June 21 (CX-GM 490 , 2495). General Mills had previously analyzed Kellogg s move and estimated what it would yield in additional margins (CX-GM 168, 188, 258). General Foods, on June 13, announced increases on 50% of its line effective June 30 (CX-GF 317, 571G, 1508; CX-GM 168). Round 13 303. On November 11, 1969, Kellogg announced price increases effective December 13 on products constituting 60% of its product line (CX-K 789, 866; CX-GM 168). On November 17, General Mills announced increases on 77% of its product line, effective that day, but with price protection offered until December 6 (CX-GM 168, 288 289, 496, 497, 2498). On November 12, General Mills had analyzed the increase and determined that it would generate $7.8 million in " Actually, Quaker had previously announced price im:rNlscsOn November It!, 1968 (Tr. 38 4681. Initial Dccision additional annual revenues for Kellogg (CX-GM 191 , 355). By November 18, General Foods had also analyzed Kellogg s advance, stated its belief that "parity pricing" could best be established " quickly matching Kellogg s advance " and on November 20 announced increases covering 440/0 of its line, effective December 15 (CX-GF 317, 320, 355, 418, 571E). (102) 304. This round is described by Dr. Scherer as the one "in which perhaps the most typical pattern existed" (Tr. 27 861). Kellogg led off. General Mils followed within six days and General Foods followed just three days later. While the pattern of respondents action in this particular round may be the most consistent with price leadership, the round is atypical when compared with the 15 other rounds that are relied upon by complaint counsel, where we find that Kellogg is not the price leader, or there are long delays in price followership and even instances of non.participation by one or more of the respondents. Even in Round 13, we find a wide variance among respondents with respect to the percentage of their respective lines covered by the price increases-Kellogg, 60%; General Mills 77%; and General Foods, 44%.

Round 14 305. On February 12, 1970, Kellogg announced that, effective March 21 , it was increasing prices on products constituting 240/0 its dollar volume (CX-K 791, 867; CX-GM 168). General Foods, on April 2, 1970, announced increases covering 20% of its line with an effective date of April 6, but with price protection offered until April 18 (CX-GF 315 571C). While General Mils analyzed Kellogg s price increase, including additional revenues that would be generated, it took no pricing action (CX-GM 207, 287).

306. If Round 13 was considered by Dr. Scherer as perhaps the most typical instance of price leadership (Scherer, Tr. 27 861; CPF 8- 101), this very next round fails to qualify by a large margin. Kellogg s price increase stood alone for about a month and one-half before General Foods acted. And General Mils, while it evaluated Kellogg s price increases, and ostensibly was also aware of General Mills' response, took no action.

Rounds 15 & 15A 307. General Mills announced an increase in price on RTE cereals representing 72% of its dollar volume, effective June 1, 1970 (CX-GM 502). On June 25, Kellogg announced increases on 67% of its line, effective July 25 (CX-K 793, 868). General Foods did not , Initial Decision 99 F. participate in what complaint counsel term the first part of Round 15.

308. At the outset of what complaint counsel term the second part of Round 15 (CPF 8-105), General Mills initiated additional increases on several cereals effective September 28 and November , 1970 (CX-GM 2018). General Foods announced, on November 20 1970, that it was increasing several of its own brands effective November 23, but with price protection offered until December (CX-GF 319, 571A). An internal General Foods memorandum dated November 18, 1970, (103Jcomments The recommended increases will re-establish profit margins that have decreased versus the previous four year average and to meet recent competitive price changes. Competitive price increases have recently been taken by Kellogg-7170 and General Mills 170" (CX-GF 434). Kellogg did not participate in the second part of Round 15. 309. General Foods' November 20 , 1970, price increase announcement was effective December 19, 1970, over six and one-half months after the June 1 effective date of General Mills' price increase and almost five months after the effective date of Kellogg s last prior price increase. This falls far short of meeting the expectation of timely responses to a price leadership situation. Round 16 310. Kellogg, on or before January 18, 1971, announced that it intended to raise prices effective February 6 on a number of products (CX-K 7073; CX-GF 449; CX-GM 205). General Mils and General Foods both announced price increases on January 29, effective February 1, with price protection offered until February 20 (CX- 1525, 1699; CX-GM 449, 2018, 2019). Complaint counsel have offered no evidence as to percent of line covered or dollars generated by the price moves in this round. Dr. Scherer s theory that respondents coordinated price increases by covering equal percentages of products cannot be tested.

311. In addition to the failure of complaint counsel' s price rounds presentation to contain the necessary product and pricing information upon which to evaluate an alleged price leadership-price followership situation, the rounds, to the extent they do furnish information, are inconsistent with price leadership-price followership.

312. Kellogg is alleged to be the price leader. Yet, it was first in Initial Decision only 12 of the 16 price rounds (12 of 17 if Round 15 is considered as two separate rounds, as well it may be). Of the 12 rounds in which Kellogg was first, General Mils followed in only nine instances, and General Foods in only 10 instances. General Mills and General Foods together followed in only eight of the 12 rounds in which Kellogg changed its prices first. General Foods led three times, and on two of those occasions neither Kellogg nor General Mils followed. General Mils led once (twice if Round 15 is considered to be two rounds). Kellogg alone followed in Round 15. General Foods alone followed in Round 15A.

313. Without repeating the details listed in the round by round analysis made above, it is clear that the long delays in price (104) followership and the instances where price changes were not followed at all render complaint counsel's price round presentation inconsistent with a price leadership-price followership premise. It is also clear that the wide variations in dollar share of line affected by the separate price changes of the individual respondents is inconsis. tent with Dr. Scherer s theory that overall price uniformity was achieved by coordinating dollar share of line covered by price changes.

314. On December 20, 1977, following four days of direct examination, Dr. Scherer on cross-examination, for the first time, presented a new explanation of how respondents were coordinating prices (a theory of comparable dollars generated per share of the market) (Tr. 519). This new theory was one which had occurred to Dr Scherer just several days before (Tr. 28 576, 28 619, 28 620). What Dr. Scherer did was to divide the additional dollars generated by each firm price increase by that firm s market share (Tr. 28 620). 315. As one basis for his theory, and in order to ascertain dollars generated, Dr. Scherer relied upon CX-GM 168. CX-GM 168, though found in the fies of General Mils, was prepared by an unidentified employee of General Mills and there is no indication of the purpose for which it was used by that company. The accuracy of the data contained in CX-GM 168 is suspect, subject to verification by corroborating evidence (to which complaint counsel have not called my attention), and the document is binding on General Foods and Kellogg only to the extent of evidencing that such a document was prepared at General Mils (See Tr. 10 614-16). 316. In preparing a comparison of dollars generated by price moves, Dr. Scherer did so on an annual basis for the years 1963-1968 H Geneml Mills' pn.ctice of estimating additional profits to be generate by Kellogg s price changes is not inconsistent with a general business interest in evaluating activities ofcompetitorB It does not prove Dr. Scherer theory. particularly in light of respondents' disparate actions. Further, there is no evidence or such analyses being made by the other respondents Initial Decision 99 F. and reflected the results of all price moves, including those excluded in the preparation of his price round analysis (Tr. 28 519- , 28 581- 82). Dr. Scherer conceded that if he had attempted such a correlation on the basis of his price rounds, nis coefficients of correlation would have been much less (Tr. 28 591). For the 1963-1968 period considered by Dr. Scherer, General Mills increased its prices at a rate 30% higher than Kellogg; and in every year other than 1963 and 1966 General Mills' price increases generated at least 30% more dollars than did Kellog s price increases (Tr. 28 595-B08). (105) 317. Dr. Scherer s theory of comparable dollars generated, therefore, bears no relationship to price leadership in lieu of overt collusion. It takes no account of prices on particular brands although it is clear that competition is on a brand basis. It totally ignores the price rounds, which serve no purpose under this theory. His theory abandons parallelism by price move. Dr. Scherer has simply advanced the assertion that, over the years, prices for the several respondents have increased in a similar fashion (although, in fact, General Mills' price have increased at a 30% greater rate). Price leadership in lieu of overt collusion contemplates a simple action and reaction. It would cover a situation where A would raise its price on a particular item by 5 and B would follow and raise its price by about the same amount. The theory, however, would not apply to the complicated situation envisioned by Dr. Scherer where each respondent would calculate the added revenues that would result from price raises on a number of products by another respondent and then, selecting its own and frequently different type products, work out price increases that would accomplish a similar overall profit increase for it-all in relationship to share of market entitlements. This simply is not contemplated by the economic theories advanced by complaint counsel. It is difficult to comprehend how such a complex arrangement could be reached short of working out an express agreement.

318. As Dr. Telser testified, the figures relied upon by Dr. Scherer merely reflect a crude form of RTE cereal price index (Tr. 670-71). One would expect a degree of correlation over time for products that are closely related in terms of condition of supply and demand such as RTE cereals (Tr. 31 674). Dr. Scherer s supposition that this was the result of collusion or price leadership in lieu of overt collusion remains a supposition. It is not established by the record in this case.

319. It is concluded, therefore, that complaint counsel have failed to prove a conspiracy on the part of respondents relative to the establishment, maintenance or change of prices for RTE cereals. Initial Decision Neither have they proved that respondents engaged in price leadership-price followers hip in lieu of overt collusion. Other Forms of Price Competition Complaint counsel assert (CPF &-111):

In addition to avoiding direct compeiition on the list prices of RTE cereals respondents also avoided virtually all other forms of price competition that might have destabilized the market by spiraling into unrestrained competition. Respondents refrained from offering trade deals (106)(discounts) to customers, from producing private label cereals, and from using "cents-off' labels that would provide consumers with the benefits of at least some price competition. Respondents' avoidance of each of these forms of competitive behavior demonstrates their success at restraining all significant forms of competitive rivalry and maintaining high prices and monopoly level profits.

However, as I have already found, complaint counsel have failed to establish that respondents, either by agreement or by means of a price leadership-price followership regimen, fixed or controlled list prices. Any evidence tending to show parallel action with respect to indirect pricing activities such as trade deals, lower priced private label cereals and cents-off labels, therefore, is of no consequence in establishing the existence of support mechanisms to a price fixing situation.

(a) Competition Through Price Promotions (0 Trade Deals 320. Trade deals are payments by manufacturers to retailers usually in the form of discounts off the wholesale case price of the manufacturers' products (Tr. 9267 , 27 956, 33405-06; CX-GM 16A). Trade deals are offered to retailers either in exchange for some performance promoting the manufacturer s products or they may have no performance requirement. The promotional services required of retailers may include setting up special display shelves featuring a cereal brand at a reduced price, placing coupons redeemable at the retailer s store in the retailer s weekly advertisements, or stocking a new product or package size (Tr. 33 405-06). 321. Both before and immediately following World War II, trade deals were prevalent throughout the RTE cereal industry (CX-GF 18 , 4039Z-108). During the mid to late 1950' , the use of trade deals declined considerably (CX-NCFM 500 at 162; CX-GF 5071). Thereafter, for some ten years, the trade deals used by the RTE cereal producers were primarily introductory allowances. Such allowances . . .

Initial Decision 99 F. are used when new products are introduced to compensate retailers for additional expenses incurred in adding the new products to the retailers' shelves for warehousing and adding to accounting and computer systems (Tr. 7990, 11 341, 17 306-7 , 27 957-58). 322. General Mils felt that trade deal activation by any of the major producers would lead to retaliatory trade deals or some other price reaction by the others (CX-GM 16, 17, 135B, C, 140, 142A). It (107Jbelieved that if trade deals were activated, "major competitors would retaliate immediately to protect their present share of (the) market" (CX-GM 135B, C). It also believed that neither Kellogg nor General Foods was "anxious to try an activation (trade deal) war (CX-GM 17D). It reasoned that, in order to pay for a trade deal program, Kellogg and General Foods would have to "(a) cut advertising, (b) cut profit, lor) (c) raise prices." General Mils could not "see either competitor deciding to cut profits." If Kellogg and General Foods "were to increase their prices-to layout additional trade money- , in turn, following competition, could do the same and retain, if not improve, our profit margins" (CX-GM 17D). 323. General Mils' advertising agency, which had been closely associated with it for years (CX 443), stated that the cereal industry has resisted pressures to enter into the allowance (trade dealJ battles which most other package goods categories wage continuously. The reason has been that once this type of activation is made to work to one company s overall advantage competition must retaliate and it soon becomes more of a defensive than offensive marketing device" (CX-GM 16B).

324. General Foods also was aware that any initial sales benefit that it could gain from the use of trade deals might be wiped out by responsive actions by others (CX-GF 76C). It believed that "The highly price competitive nature of the cereal market in the early 1950' s (price off, trade deals, pack-ins) . . . suggests that price maneuvers may be at most a short term advantage" (CX-GF 76C 4039Z-108). Thus, we find that "As a general rule, Post cereals. . . Lwould) not utilize trade deals" except on new products or new sizes or major product changes (CX-GF 601K).

325. When asked by a grocery store chain about straight case allowances vis-a-vis allowances with performance requirements Kellogg s General Sales Manager told the chain that Kellogg did not want to get store premium types of trade deals started again; that it did "not want to give case allowances just to establish everyday low prices" (CX-K 7144A).

326. The foregoing recitation of evidence relied upon by complaint counsel reflects an awareness by each individual respondent Initial Decision of what competitive practices, including trade deals, are taking place in the industry. It additionally evidences each respondent' s individual desire not to precipitate a price war by giving trade allowances other than those offered in connection with the introduction of new products. List prices having been arrived at after considering costs competitive factors and profit objectives, it is reasonable that respondents would not want to sell at lower prices by reason of trade deals which could establish a permanent discount at which retailers could stock up on all their needs. It is to be anticipated, therefore that respondents would be selective in offering trade deals. (108) 327. Respondents evaluated each other s potential to resume large scale trade deal activities and, while they did not anticipate such a resumption, they each stood ready to respond by competing in like fashion should another start.

328. The foregoing does not reflect an agreement with respect to competition by offering trade allowances. No respondent was waiting for a signal from any other respondent to resume large scale trade allowance activities, nor is there any evidence that an agreement or action by a leader precipitated the decline of trade deals in the 1950' 329. While the record shows an evolvement over time from heavy use of trade allowances prior to the middle or late 1950's to relatively little use until the end of the 1960' , when the practice was resumed there is no evidence of an abrupt shift or of other timing consistent with agreement of leadership.

330. Kellogg has employed trade deals on some of its most popular established brands since 1966. Kellogg employed trade deals with Product 19 each year from 1966 to 1970. It gave trade allowances on both Corn Flakes and Shredded Wheat in 1968 and 1969, on Special K in 1969 and on Raisin Bran in 1970 (CX-K 1072). 331. General Mils employed trade deals with most of its established cereal brands in 1969 and 1970 (KX 203C). 332. By 1971, each respondent offered trade deals on the majority of its brands. General Mils enlarged its trade deal activity in 1969 to (CX 409C), stepped up itscarry trade deals on 11 of its 15 brands trade deal activity in 1970, and then reached an even higher level in 1971. Kellogg waited until 1971, at which time it extended its trade deal usage to 18 brands of its 20 brand product line (CX 409B). General Foods also waited until 1971 to act. During that year, it used trade deals with nine of its 14 brands.

Initial Dccision 99 F. TRADE DEAL USE, 1966-1971 Kellogg General Mils General Foods No- of No. of No. of No. of Brands No. of Brands No. of Brands Deals Affected Deals Affected Deals Affected 1966 (not available) (KX 201-04) (109) 333. Thus we see that in 1969, General Mills (not Kellogg, the alleged price leader) was the first to reinstitute the wide use of trade deals, and that Kellogg and General Foods waited until 1971 before they engaged in trade deals as a general practice. The sequence and timing of respondents' resumption of trade deals negates complaint counsel's contention that respondents coordinated their utilization of trade deals so as not to upset the equilibrium of their alleged list price arrangements; and there is no evidence demonstrating the sequence or timing of respondents' earlier abandonment of trade deals sometime in the 1950' 334. Complaint counsel rely upon the fact that, unlike the situation for RTE cereals, trade deals were used frequently to omote other grocery products from 1960 through 1972 (Tr. 8556 8903, 9267-B9, 18 001-02; CX-GM 16B, 142A). RTE cereal sales however, are not as responsive as other food products to promotional stimuli such as trade deals. Since consumer RTE cereral purchases do not change much seasonally, unlike products such as Kool-Aid there is less need to use trade deals to provide rapid inventory buildup to cover surges in consumer demand (Tr. 36 848). In General Mills' opinion, the additional volume of RTE cereals moving to retailers under trade deals caused shelf space and inventory problems for retailers, inasmuch as RTE cereals were considered (tJoo bulky for display purposes " and inventory overflow caused floor and back room space problems" (CX-GM 140B). 335. General Foods has established that its utilzation of trade deals over the years was responsive to, and consistent with marketing and business factors and problems which it had to face. 336. Trade deals on RTE cereals were perceived by General Foods as effective primarily in prompting grocers to promote particular products, but not in securing retail price decreases (Tr. 847-49). Advertising has been perceived by General Foods as the Initial Decision most effective and efficient marketing tool for making consumers aware, and inducing their trial, of General Foods cereal products (CX-GF 3000Z-7&-77).

337. General Foods' fluctuating use of trade deals must be viewed against the background of: (a) the dramatic growth of television in the middle 1950's just as its cereals were losing market share, and General Foods' shift of marketing emphasis from promoting individual cereals to advertising its full line of products, (b) General Foods efforts to introduce new products in the early 1960' , which directed marketing resources away from existing products and led to a deemphasis of the line approach, and (c) the sudden leveling off of RTE cereal sales in the late 1960' , which led General Foods to increase its use of promotions. (110) 338. Prior to FY 1958, which began in April 1957, the Post Division engaged in "individual brand promotion where each Post product was sold on its own merits in competition with both other Post brands and other company brands" (CX-GF 16A). However, the share of total RTE eereal sales represented by Post cereals declined from over 27% in early 1952 to less than 23% in late 1956 (GFX 203B).

339. The management consulting firm of McKinsey and Company ("McKinsey ) was retained to study the problems of the Post Cereals Division and to rccammed a remedial program. In its July 1956 report to General Foods, McKinsey recommended that "the Division should emphasize the promotion of the Post ' line (GFX l235Z-4).

340. Shortly thereafter, in September 1956, a new "Marketing Philosophy" (GFX 1288) was adopted by the Division calling for advertising and selling a line of cereals as opposed to individual brand marketing" (emphasis in original) (CX-GF 16A). 341. As part of this change, the Division s management decided to increase consumer-directed advertising. To implement this, the Division greatly increased its use of what was then a new marketing tool-television (Tr. 36 355-61).

342. During FYs 1952-1955 (April 1951-March 1955), television advertising accounted for approximately 25% of General Foods cererals' consumer- directed marketing expense (GFX 203J). This grew to nearly 50% in FY 1956, over 60% in FYs 1957-1959 and over 80% in FY 1960 (GFX 203J; see also CX-GF 4T-U). 343. Thus, after the FY 1956 introduction of the line approach the emphasis was shifted away from those marketing tactics that are most effective with respect to individual products-e. trade-directed activities-into line-oriented advertising, especially television. Initial Decision 99 F. 344. By 1960, there was significant improvement in consumer perception of General Foods' cereals (CX-GF 1300G). Therefore, in planning for FY 1961, the Post Division s management decided that (a)after three years of intensive line promotion " it was time to capitalize fully on the high consumer awareness of the Post line theme" by developing "more individual brand sell in (RTE cereals product advertising" (CX-GF 324K).

345. In FY 1962, the return to individual product promotion accelerated, but with advertising continuing to playa more important role than it had in the early 1950' fWJe plan to continue the strong individualistic approach in each of our product commercials. This approach identifies the product as one of the (111)family of Post Cereals but proceeds to sell the product on its own individual merits (CX-GF BE). 346. In February 1965, General Foods was of the opinion that money was better spent on consumer advertising than on trade deals:

Post believes that consumer advertising is a more effective and efficient lung term benefit to cereal success-so trade deals will not be employed. The only exception is the introduction of a new product (and in some instances, a new size) where a trade deal may be employed as an aid in establishing broad-scale distribution (emphasis in original) (CX-GF 485Z-7).

347. The advertising emphasis on individual products continued through FY 1966 (CX-GF 485L, 1300E-G).

348. In FY 1962, the managers of the Post Division had decided to undertake "a substantial new product program" based upon " number of innovative products and concepts on hand" (CX- 4039Z-17). This program was undertaken (GFX 1370H-I). 349. The Division s marketing managers believed "that during periods of new product introduction, advertising and promotional support for established brands could be reduced without impairing the vitality of their franchises" (CX-GF 4039Z-18); "that the halo effect of the new product would compensate for the lack of (advertising and promotions support on the established brands" (CX- GF 4039Z-19). Accordingly, from FY 1962 through FY 1966, the Division s managers followed a marketing strategy of "holding volume on established brands by according them the minimal advertising and promotional support consistent with that objective (CX-GF 4039Z-17).

350. From 1950 through 1965, there had been continuous and substantial growth in sales of RTE cereals. Then, pound sales of RTE cereals grew by only .4% in 1966 and declined by 2.1 % in 1967 (CX Idle).

_._ Initial Decision 351. By late 1966, both established Post cereals and new products were considerably below their planned volume and profi goals (CX- GF 4039B). In November 1966, a special Post Cereal Task Force was organized "to complete an in-depth examination of all areas appro priate to the cereal business and recommend a three year operating strategy for the cereal business" (CX-GF 4039B). 352. In its 1967 report (CX-GF 4039), the task force recommended that "each advertising message (be devotedJ exclusively (112Jto the individual brand rather than to line sale" (CX-GF 4039P). It recognized that consumer promotion was Han effective means of generating short-term incremental sales in the cereal business" (CX- GF 4039Z-111). It, therefore, recommended that promotional expenditures be raised "above the levels of recent years" (CX-GF 4039Z- 114).

353. By the following year, the Division was fully committed to using promotions to " (sJecure short-term volume increases" (CX- 602G), and in the reallocation of marketing resources from advertising into promotions (CX-GF 3000Z-23i).

354. Trade deals have consistently played a relatively small role in the marketing of General Foods' cereals: TRADE DEALS ON GENERAL FOODS' CEREALS FY 1952-1961 (in thousands of dollars) Trade Deals as a Percentage Fiscal Total Marketing Trade of Total Marketing Year Expense Deals Expense 1952 385 277 92% 1953 16,431 200 1.21% 1954 958 310 1.94% 1955 706 526 34% 1956 626 490 94% 1957 562 332 00% 1958 850 35% 1959 630 221 02% 1960 776 223 25% 1961 977 29% 355. While there was a clear decline in General Foods' use of trade deals beginning in FY 1958 from 2% to .35%, this is explained by the change in the Post Division s marketing strategy from an emphasis on individual products to an overall line approach, the belief that promotional efforts on new products would have a halo effect on established products, and the switch to television as a preferred method of promotion. The subsequent switch back to trade Initial Decisjon 99 F. deals reflects the poor performance of both new and established General Foods products and the decision that it was necessary to use promotions, such as trade deals, to secure short-term increases in sales volume.

356. Complaint counsel do not contend that, when utilized respondents did not attempt to gain competitive advantages by their (113Jtrade deals. Thus, CX-GM 17, a December 1970 General Mills memorandum, recommended a !'Plan B" trade deal, providing for extra retailer displays, features and price reductions on Wheaties because it was felt that General Mils' sales force could " better handle activation than theirs IKellogg s and General Foods') (CX- GM 17E), and that General Mills could profitably gain consumer sales by using the trade deal (CX-GM 17). 357. Respondents made extensive use of trade deals throughout the complaint period in connection with the introduction of new products (Tr. 7990-91 , 11 341-42, 12 306, 13 094, 17 306-07, 27 956- , 36 934). Introductory trade deals were considered by respondents to be particularly useful to secure retailer acceptance of new products and thus "as an aid in establishing broadscale distribution for new products (CX-GF 485Z-7). It is immaterial whether an introductory trade discount on a new product is considered to constitute a lowering of price from an established list pdce or as establishing an initial price below list. In either event, the product is being offered at a relatively low competitive price; and that is price competition between the new product and products already on the market.

(ii) Cents-Off Deals 358. Complaint counsel (CPF 8-111, 8-164) assert that respondents avoided the use of cents-off labels as part of their avoidance of all forms of price competition practices that might disrupt their established pricing structure.

359. A "cents-off label" is an announcement printed on a package indicating that the product contained in that package wil be sold for a specified amount less than the listed retail price (Tr. 17 246). 360. Respondents used "cents-off' deals in the early 1950' , but not thereafter (Tr. 17 246, 30 111-12; CX-K 1073C, 7144; CX- 485Z-, 4039Z-108). The record, however, does not indicate the extent to which this type of promotion had been engaged in, nor is there any evidence showing the sequence or timing of the cessation of this practice by respondents. As with the case of trade deals, there , Initial Decision is no evidence of an agreement among respondents not to engage in this practice.

361. Kellogg believed that if "one company starts it (cents-off deals), the others follow suit" (CX- 139C). Although General Mils believed that "cents-off' promotions were powerful enough to " get immediate positive results" if used by competitors, it did not use such methods (CX-GM 135B, C; Tr. 17 246). General Foods believed that price maneuvers" such as cents-off labels Hmay be at most a short term advantage" (CX-GF 4039Z-108); that respondents had avoided this type of rivalry because of the "bitter taste" of previous (114)competition and the danger that such actions would result in a dilution of impact/ineffcient investment of marketing funds" (CX- GF 76C). It stated that it would not utilize off-label packs unless competitive activity forces the issue" (CX-GF 60lL). A General Foods' marketing plan summarizes its attitude toward cents-off labels:

The nature of the Cereal business (in terms of multiplicity of brands and categories with little, if any, brand loyalty) is such that judgment indicates it would be an ineffective and inefficient expenditure of a brand' s "lean" marketing funds to engage in a price-cutting battle via cents-off consumer promotion. Money so employed would be a risky drain on funds available to advertise brands' basic consumer benefitswhich Post believes is the key to long-term cereal business success. Cents-off promotions wil never be used (CX-GF 48fiz-). 362. The facts developed with respect to cents-off deals reflect a situation similar to that involving trade deals-an awareness by respondents of competitive practices and reasons why this practice has not been utilized by the individual respondents, apart from an agreement or an attempt to fix prices by price leadership. More particularly, with respect to General Foods, they reflect an implementation of its line marketing strategy at a time its products were not being promoted on an individual basis. (iii) In-Pack Premiums 363. In-pack premiums are small items, such as toys or games that may be placed in boxes of RTE cereals to promote sales to consumers (CX-K 1073E; CX-GM 2115). Complaint counsel charge (CPF 8-171 8-175) that respondents reached an express agreement to limit the use of in-pack premiums.

364. There is no direct evidence of such an agreement. As was the situation with respect to price fixing, top executives of each (115) os This means of competition is being nmsidered as a type of price promotion, since the offering of II premium along with thecerelll at no additional cost acts as a price reduction to the edent of the value of the premium 122 FEDERAL TRAm: COMMISSION DECISIONS Initial Decision 99 F. respondent have testified to the effect that decisions regarding inpack premiums were made independently; that there was no express or tacit agreement among respondents regarding their use (Tr. 29,726 944 768 35,462 815, 361-62). Again, the testimony of Kellogg and General Mills officials falls within that which I am required to accept at face value. And if Kellogg and General Mils had no such agreements with General Foods, it follows that General Foods had no such agreements with Kellogg or General Mills. Unless overcome by more persuasive evidence, I am required to accept such testimony as accurately describing the situation. Complaint counsel assert (CPR 175) that the alleged explicit agreement is demonstrated by (1) the importance respondents placed on the marketing effectiveness of in-packs; (2) the chronology of respondents' decisions to decrease their usage of in- packs; and (3) the expressed and implied recognition by respondents with respect to an understanding among themselves to limit in-packs. 365. Before 1957, respondents inserted in-pack premiums in many of their brands at the same time (Tr. 11,228, 13 023, 13 025; CX 1006; CX-GF 371). In 1956, Kellogg carried 36 in-packs in nine brands representing a range of approximately 12% to 23% Kellogg pound volume over the year. General Mills carried 11 inpacks in five brands for a 6% to 12% volume share, and General Foods used in-packs in 13% to 23% of its pound volume. From April 1955" until January 1957, Kellogg used in-packs in cereals that volume; Generalaccounted for 9.5% to 19.4% of its pound sales Foods used them in 9.5% to 22.4% of its sales volume; and General Mills used them in 6.2% to 15.1% of its sales volume (CX 1006; CX- GM 2115; CX-GF 371 372).

366. Ralston, not one of the respondents, was the first major firm to discontinue putting premiums in packages. It did so by February 1956 (Tr. 28 753-55; CX-NCFM 500, at p. 177). General Mils started to decrease its use of in-pack premiums in October/November of Kellogg nor1956. This was some eight months later. Neither General Foods decreased its use of in-pack premiums at that time. On the contrary, Kellogg increased its use of in-packs during February/March and April/May of 1957; and General Foods continued to increase its use of in-packs from February/March until June/July 1957. Just after reaching their peak of in-packs usage Kellogg and General Mills began a precipitous decline in mid-1957. By August 1957 (General Mils), February 1958 (Kellogg) and March '" The ",cord doe not rcfJectdetailed data on the use of in- p"cks prior to 1955 g., Initial Decision 1958 (General Foods), respondents were inserting in-packs in less than one half of one percent of their product volume (CX 1006). (116) 367. Thus, the chronology of respondent' s change in their usage of in-packs does not support complaint counsel's assertions as to agreement. With regard to the importance respondents placed on the marketing effectiveness of in-packs, complaint counsel have failed to demonstrate that respondents acted against their best individual interests when they curtailed this promotional device. To the contrary, the record tends to demonstrate significant business and marketing circumstances to which respondent' s actions were reasonable individual responses.

368. In-pack premiums are used in the RTE cereal industry to provide short-term sales increases for the specific brands so promoted. In-packs, generally, are not credited with beneficial sales results extending beyond their term of use or beyond the particular brand with which the premium is employed (Tr. 12 435- , 36 351- 52; CX-GF 4, 485Z-; CX-K 487, 546K, 650A). Television advertising, on the other hand, is regarded as uniquely effective in the RTE cereal industry in establishing long-term sales expansion (Tr. 29 930 352 361; CX-GF 4852-7).

369. Television advertising appeared as a significant factor in the growing RTE cereal market in the early 1950' , and its employment rapidly expanded immediately prior to the decline of in-packs. Kellogg spent $706 000 in 1950 on television advertising, representing 8.3% of Kellogg s advertising budget. By 1957, Kellogg s television advertising had expanded to $13 537 000, which was 78.3% of total advertising expenditures. For the same period, General Mills television advertising increased from $548 000 to $6 533 000. This was an increase from 10% to 83% of General Mills' total advertising expenditures. General Foods' television advertising jumped in similar fashion from $773 000 (12% of its total advertising expenditures) in 1951 to $5,863 000 (85.4% of total advertising expenditures) in 1957. RTE cereal television advertising by respondents continued to expand through the 1950's and early 1960's (GFX 1319). 370. Thus, it is reasonable to believe that each respondent independently reacted to the advent of television as an effective (117)advertising medium by shifting from promotional expenditures C--mplaint counsel would support their asertion that respondents placed "relit importance on the marketing effeetivenef\ of in-packs by a number of respondent documents extolling the benefits of in-packs I see, CX- 37A 47D, 650A, 651; CX--GM 38A; CX..F 76B), The"" documents, however, were prepared in the last half of 196 and in early 1970, when market conditions had changed and respondents were about to use, or were using, in-pack premiums, They do not indicate that re"l'ondent. curtailed their um.. of in-pack premium" in the face ofeompctitive reason why they should have been continued . . .

124 FEDl;RAL TRADle COMMISSION DECISIONS Initial Decision 99 F. such as in-pack premiums to television advertising. Kellogg decided to make the shift (Tr. 11 235-36). So did General Foods (Tr. 14 181). 371. General Foods decided to replace in-pack promotions with television advertising as part of its strategy, initiated in 1957 , to market Post cereals collectively as one line of cereals rather than as individual brands (Tr. 36 352, 36 356-62; CX-GF 5A, 18T; GFX 406A). The new line strategy emphasized television advertising as a means of increasing sales of the Post line of ready-to-eat cereals and relegated consumer promotions to a supplemental role (CX-GF 5B). Post sought to replace the prior marketing approach, which emphasized individual brand promotions such as in-packs, with heavy investment in advertising in order to create long-term expansion of Post line sales (Tr. 36 351-52, 36 360-61). The change in marketing philosophy resulted in a shift of marketing funds from premium promotions to advertising (CX-GF 16A, 406B). In 1952, General Foods' television spending on RTE cereals amounted to 30% of total marketing expenditures. In 1957, it had risen to over 60% (CX- 507I).

372. General Foods, in August 1967, referred to the previous inpack rivalry as "fierce and unprofitable competition" (CX-GF 76A) and stated:

We do not believe that any manufacturer will benefit-on a long term basisreverting to the premium wars of the 1950's (CX-GF 75). General Foods " recognized. . . that too intensive utilization(of in-packsJ is certain to bring strong competitive retaliation. This would not serve the best long-term interests of Post, and could even be deleterious in the short-term" (CX-GF 4039Z-114). General Foods believed that the respondents avoided this method of competition because of the "bitter taste" of previous competition in this area. However, it was prepared to reinstitute the use of in-packs if Kellogg or General Mills started first (CX-GF 76A C). 373. Kellogg had an "arsenal of package inserts ready to go" if another respondent resumed the practice (CX- 546K). General Mills believed that if Kellogg got "the in-pack fever " General Mills would "probably have to follow" (CX-GM 38A). 374. The common evaluation by the several respondents as to the undesirability of resuming extended in-pack competition, and their stated readiness to respond if another competitor reinstituted the practice, does not evidence the existence of a conspiracy some ten years before, or a continuation thereof.

375. This brings us finally to several documents produced from the files of General Foods which refer to an industry guideline or Initial Decision (1l8Jrule. The first is a memorandum dated August 3, 1967, from a Mr. R. S. Braddock to a Mr. R. P. Brubaker (CX-GF 76). Mr. Braddock was an assistant or associate product manager under Mr. Brubaker, who was a product manager for a number of items. Mr. Brubaker was subordinate to Mr. Cobb, who was in charge of marketing (Tr. 14 593-94). In that memorandum of Mr. Braddock we find the following:

Subject Defensive Premium Plan This memo will . analyze the likeihood of competition breaking the pack-in premium guideline; Summary If competition packs premiums in several presweet brands (thus breaking the industry "guideline ) and we take no action until they arrive at retail, we risk loss of at least$1 100M in PBT.

Discussion Rationale for Plan To date, the three major manufacturers in the cereal business have been respecting an "unwritten rule " stemming from fierce and unprofitable competitioIl in the early and middle '50s, that they have retail exposure with only one pack-in premium in one brand at a given time. . (CX--F 76A). In summary, it is likely that Kellogg and General Mills will read the favorable Alpha-Bits experience and will react by putting greater emphasis in pack-ins but that they will do so within the existing guidelines--ne premium in one (119Jbrand at a time (emphasis in original) (CX-GF 76C).

376. On August 11, 1967, Mr. Brubaker apparently forwarded Mr. Braddock's Defensive Premium Plan to Mr. Cobb. In Mr. Brubaker s forwarding memorandum to Mr. Cobb (CX-GF 75), he stated:

Subject Defensive In-Pack Premium Plan In view or the persistent rumors that Kellugg is planning to pack premiums in several cereals, it is advisable for us to have a defensive plan prepared. Such a recommended plan is attached (emphasis supplied), This recommended plan would enable us to strike back with four brands at retail containing premiums in a reasonably short period of time-for the modest investment of $10 000.

Initial Decision 99 YT. We recommend this investment in order to be prepared to launch such a program; however, we feel very strongly that if we learn that another manufacturer is packing premiums in multiple brands, our first step should be to do everything possible to ascertain the long term intent of the manufacturer in this area. (emphasis in original).

377. Thus, we find Mr. Brubaker not only forwarding Mr. Braddock' s plan, but personally expressing concern if another manufacturer (possibly Kellogg) should pack premiums "in multiple brands." On the other hand, the reference to "rumors that Kellogg is planning to pack premiums in several cereals" indicates that the parties were not operating under explicit agreements, as complaint counsel contend.

378. In a May 1968, memorandum to Mr. Cobb from a Mr. P. A. Schweitzer, another assistant or associate product manager (Tr. 594-95), Mr. Schweitzer analyzed possible competitive reaction to a planned in-pack in Alpha-Bits. In that memorandum, we find: (120) Being rational they will realize that if we were seriously going to pursue the multiple in-pack premium route, we would use a much stronger premium and it would be supported much more heavily.

In view of the above point.c;, breaking a "magic guideline" with this paper pack-in wil not force competition into reacting (CX-GF 26B). So again, we find comment concerning a guideline which refers to the multiple use of in-packs.

379. If, in fact, there was a conspiracy to limit in-packs to one product at a time, the statements of Messrs. Brubaker, Braddock and Schweitzer would have been made during the course of and in furtherance of that conspiracy. Without these memoranda, however there is no prima facie evidence establishing such a conspiracy. The documents, therefore, may not be considered against General Mills or Kellogg for purposes of establishing a conspiracy or agreement as to in-packs (supra p. 19).

380. Even if the documents could be used against all parties standing alone it is unclear just what they would evidence. The reference to a gentlemen s agreement or guideline, coupled with the specific limitation of only one in-pack promotion at a time, does indicate the existence of an actual agreement. However, the period of time encompassed by that agreement would remain unsettled. The documents prepared in August 1967 and May 1968 purport to recite the situation then existing. They would not establish Wat the parties mutually abandoned the use of in-packs some ten years earlier, as complaint counsel contend. The evidence showing independence of action at that prior time (supra Findings 365-7) is much too strong , Initial Decision to be overcome by these documents. The documents, therefore, at most are consistent with a later understanding, around 1967 and 1968, not to resume in-packs on a multiple seale. And the reference to rumors concerning Kellogg s intention to utilize multiple in-packs reflects General Foods' belief that Kellogg did not consider itself bound by any such understanding.

381. General Mills' analysis, in December 1967, was that the use of in-packs would favor Kellogg s medium sized brands; that without in-packs, General Mills' established brands had outperformed those of other companies even during the year in which Post went back to using them on several brands," The memorandum concluded Let' s hope Kellogg does not get the in-pack fever because if they do ll probably have to follow and suffer" (CX-GM 38). This indicates that General Mills, in December 1967, was not using in-packs because of its own business judgment. It also evidences that General Foods, during one unidentified year, had in-packs on several brands- This would be inconsistent with an alleged agreement not to have multiple in-packs. (121) 382. Going one step further and assuming, arguendo that the parties had a gentlemen s agreement in 1967 and 1968 not to use multiple in-packs, this would constitute the only promotional arrangement shown to exist among respondents. In-packs are a relatively expensive promotional device (Tr. 11 232; CX-GM 233; CX-GF 40138; CX-K 546K) and are unique in that they put respondents "into the toy and gimmick business rather than the cereal business" (GFX 406A). A gentlemen s agreement limited to curtail this practice would be insufficient to establish an overall agreement not to compete pricewise or in any other fashion. 383. In any event, respondents resumed more extensive use of inpacks beginning in 1968, and increased utilization through 1972 (Tr. 906, 21 937; GMX 531; CX-GM 2115; CX-K 1073). This was in response to the lull in overall market growth and the respondents ostensibly independent decisions to promote individual brands by short-term means such as in-packs (CX 101E, CX--F 3000I, X, Z- 100; GMX 241, 531A- , 552A, 555A, 556A; GFX 1319). General Foods' resumption of in-packs was part of the abandonment of its full line advertising policy in favor of a brand promotional strategy (CX- GF 3000Z-11, Z-12, Z-20; 4039 0, Z-1l4, Z-1l5). (iv) Coupons 384. Respondents' couponing activities negate complaint counsel's contentions that respondents avoided price related promotions. g., Initial Decision 99 F.T.C. Dr. Scherer, one of complaint counsel's economic experts, testified that respondents ' couponing remained at a fairly significant level during the 1960s" and was "widespread" (Tr. 27,967-68). Since 1965 is available, eachwhen data on coupons for all respondents respondent used coupons on both established and new products (GMX 475-96; KX 78, 20IC thru 204C; GFX 1336-44). 385. Kellogg distributes coupons inside cereal packages, in newspapers or through the mail, offering discounts worth usually five to 15 cents redeemable at retail stores. Retailers are reimbursed by Kellogg for the face value of the coupon, plus handling costs of two cents per coupon (Tr. 29 924-25).

386. From 1966 to 1972, Kellogg used coupons on products representing 75 to 80% of its line (KX 79A- , 80; GFX 1319). Kellogg s yearly marketing plans show its strong reliance on coupons for promoting sales (e. CX-K 7178K, 7179G, 7180H, 7193J, 7194M 7195M, 7196K, 7198F, G, 7201G, I, 7207H, M). 387. General Mills used two types of coupons to promote its products. "Manufacturer coupons" are coupons whkh General Mills distributes to the consumer by direct mail, by distribution at shopping areas, through magazines and newspaper advertisements, or by (122Jpacking them in boxes of cereals or other General Mills products. The consumer redeems these coupons at the store when purchasing cereal. "Retailer coupons, denominated Plan F and Plan , are coupons which General Mills induces a retailer to place in its own food advertisement. The consumer also redeems these at the store when purchasing cereals. General Mills reimburses the retailer for the cost of the coupon and gives it an additional allowance (Tr. 33,405, 33 425). Every General Mils product in national distribution " or "manufac-since June 1968 has been promoted by either "retailer turer" coupons or a combination of both. From 1965 to 1968, when data for "manufacturer" coupons only is available, General Mills used coupons on products representing over 800/0 of its pound volume in every year but one (GMX 211, 225, 271, 564E-1). 388. General Foods similarly used coupons with the majority of its brands every year from 1965 to 1972. In 1965, all General Foods established cereals carried coupons except Sugar Crisp, Bran Flakes Rice Krin.kles and Raisin Bran. In 1966, coupons for Raisin Bran and Bran Flakes were added and, in 1967, coupon.s were distributed for Sugar Crisp and Rice Krinkles. From 1968 to 1970, General Foods used coupons with all of its cereal brands other than Crispy Critters and, in 1971, with all products except Crispy Critters and Honeycombs (GFX 1337 thru 1344).

389. Complaint counsel assert (CPF 8-165 thru 8-169) that Initial Decision coupons were an ineffective means of promoting RTE cereal sales. However, the price incentive afforded by cents-off coupons amplified Kellogg s sales volume by inducing trial of cereal brands by nonusers and by stimulating additional purchase by current buyers (Tr. 926-27).

390. Further, Kellogg emphasized their use in the early 1970' , at a time when other methods of promotion such as in pack premiums and trade deals (conceded by complaint counsel to be very effective competitive devices (CPF 8-176 thru 8-179, 8-112 thru 8-115)) were available and widely used by Kellogg and the other respondents. For example, Kellogg s 1972-1973 Special K promotional strategy featured coupon promotion because of its effectiveness in promoting RTE cereal sales:

Continue the pattern of in-pack couponing. This cDuponing can encourage increased consumption among current buyers and increase tria) and re-trial (CX- 7194M).

Kellogg s Mini-Wheats 1972-1973 "Promotion Strategy" likewise featured coupons:

Stimulate trial and repeat purchase through the use of Mini-Wheats' in-pack coupons (CX-K 7201G), (123) 391. Consumers used over. . . of all Kellogg coupons distributed over the 1970-1972 period (derived from KX 80D , F in camera). This is a very high figure when it is realized that newspaper and mail offerings reach many people who are not potential customers of the cereal covered by particular coupons.

The pay back to consumers and cost to Kellogg from Kellogg coupon promotions was substantial each year from 1966 to 1972. 392. General Foods' 1967 Cereal Task Force recognized that couponing was "known effective:

(a) To generate trial among non or infrequent users. (b) To provide an incentive for current users to immediately repurchase or to purchase in extra quantity (loading).

(c) To provide the Sales Force with an effective vehicle to sell against (CX 4039Z-1l6).

The task force credited Post's 1967 national coupon campaign with generating substantial gains in market share for General Foods (CX- GF 4039Z-112). (124) 393. A consumer survey conducted for General Foods' Post Division in 1964 revealed that:

. Chart t.itled "KF.LLOGG COUPONS ANI) CASH REFUND REDEMPTIONS \966-1972" not reproduced hccC!in, because data isin camera. (KX BOA, 1: in ca.mera) g., Initial Decision 99 F. Among a number of different types of premiums listed, housewives overwhelmingly selected as the one which would be most likely to cause them to buy a cereal: " coupon allowing you to make your next purchase more cheaply" (GFX 569E). 394. Not only was the use of couponing by respondents substantial, but it appears, as indicated by the following chart " not to have been coordinated.

As shown above, there is no continuity of pattern for any individual respondent or correlation among respondents on coupon and cash refund redemptions from anyone year to the next. (125) 395. Complaint counsel (CPF &-167, &-168) would downplay the importance of coupons. They assert that the redemption of 2.6% of Kellogg coupons distributed in newspapers, 3% of those distributed in magazines, 8% of those distributed by direct mail and 6% distributed inside cereal packages, and the 5% and 6% redemption rate anticipated by General Mills of a coupon being distributed through newspapers (CX-K 576B-E; KX 78; CX-K 5A; CX-GM 14C) is insignificant. I cannot agree. The respondents have apparently engaged in mass distribution of coupons and not all coupons are noticed. Further, not every recipient wants RTE cereals or the particular cereal involved. And many recipients will receive duplicated coupons through different media.

396. In assessing the amount of price reduction involved, complaint counsel would divide overall sales by the value of the coupons redeemed. Respondents, on the other hand, would divide the price of a particular cereal (say $.50) by the value of the particular coupon (say $.07) and reach a 14% price reduction. In Finding 394 above, I have presented the price reduction resulting from coupons in the manner advocated by complaint counsel. However, to the person who redeems a $.07 coupon on a $.50 item, the price reduction is 14%. And respondents have made this type of price reduction readily available.

(b) Private Label RTE Cereals Complaint counsel allege (CPF &-124, &-137) that respondents tacitly agreed to avoid private label business; and that, as a result, they reduced private label product production and so limited their pricing discretion and removed a threat to the stability of their agreement to avoid price competition.

397. The term "private label RTE cereal, " in its strictest sense, refers to RTE cereal sold under a retailer or wholesaler label (e. Safeway Corn Flakes), in contrast with cereal sold under the .. Chart titled "COUPQr\ AND CASH REFUND REDEMPTIONS AS A PERCENTAGE OF DOLLAR SAU.;S 1965 1972" not reproduced herein, bfausI' data isin comero., (GMX 561A, in comemJ Initial Decision manufacturer s label (e. Kellogg s Corn Flakes) (Tr. 11 549, 14 896). Private label RTE cereal generally is not advertised by the manufacturer and is sold to wholesalers and retailers at much lower prices than branded cereal sold under a manufacturer s label (Tr. 17 508- 11.

398. "Controlled brand" RTE cereals are cereals sold under a manufacturer s label with no advertising support from the manufacturer (Tr. 13 697- , 17 511). Controlled label cereals are distributed on a limited basis, generally to retailers too small to carry their own private label, and are priced at levels similar to private label (Tr. 700, 13 722-23; CX-GF 124A). As used here, the term "private label" encompasses private label and controlled brand RTE cereals. 399. The lack of advertising of private label RTE cereals enables manufacturers to price them lower than their branded counterparts. These lower prices are passed on by retailers in substantial measure to the consumer (Tr. 9001-02, 9338-39, 13 724, (126)17 505-26). Thus, private label products compete pricewise with branded products. 400. Ralston generally priced its private label products to give the consumer about a 10% saving over branded products. Retailers normally paid less than the price paid for the comparable branded product. Thus, retailers enjoyed about a 6% higher gross margin than on branded products (Tr. 17 509-11). General Foods priced its private label cereals by deducting the advertising and direct selling expenses of the comparable advertised brand. It then added the selling and brokerage expenses (CX-GF 121I 401. Retailers attempt to capitalize on the price competition offered by private label RTE cereals by placing them close to their branded counterparts, so that their customers may choose the private label products on the basis of their lower prices (Tr. 9001- 9I27- 9338).

402. Private labeling in the RTE cereal industry is much less extensive than in many other food product categories (Tr. 21 873-74; CX-CI 103). RTE cereal producers frequently use the "All Other category reported by A.C. Nielsen as a basis for estimating the sales and market share of private label RTE cereals (Tr. 7414, 11 550-51; 947; CX-GF 121M). In 1943, the "All Other" category was 10. of total pound sales (CX 106A). This category declined to as low as 1.5% in 1967-1968 and never exceeded 3% after 1956 (CX 106). 403. Kellogg has never sold a private label RTE cereal product (Tr. 11 549, 12 655--9, 13 153- 927). 404. In August 1943, Kellogg leased the facilities of Miller Cereal ," Consistent with this policy, Kellogg njeded requests from retailers during the p"riod covered by the complaint (1950-1972) for private label RTf' cereal products (Tr. 9016-19 660-8 J3,153-54; CX- K 7054 , 7148) Initial Decision 99 F. Company, which was primarily in the business of manufacturing and selling private label RTE cereal products. Miller produced several types of RTE cereals, including corn flakes, wheat flakes, 40% bran flakes and puffed wheat (CX-K 619). Kellogg discontinued Miler private labels and began packing Kellogg brand cereals in the Miler facilities. The lease agreement contained an option to buy within 15 years, which Kellogg exercised in 1958. As a result, Miller was eliminated as an independent producer of RTE cereal products (Tr. 873; CX-K 444B, 619).

405. Respondents, including Kellogg, are alleged to have engaged heavily in price competition prior to 1950. Kellogg s consistent (127) avoidance of private labels dating back over 40 years prior to 1950 including its discontinuance of private labels in the Miller facility in 1943, does not become suspect as a step in carrying out an alleged tacit agreement around 1950 not to engage in private label competition.

406. Like Kellogg, General Mills has never produced private label cereals (Tr. 15 933). Therefore, this practice, which long predated 1950, does not fit a time pattern from which it can be argued that there was tacit agreement not to engage in private label competition.

407. Both Kellogg and General Mills have maintained operations at a high level of capacity (Tr. 26 364, 27 155, 27 443, 28 236; CX 203). To the extent that excess capacity occurred, each company, as a matter of business policy, elected to utilize that capacity with newly developed branded products (Tr. 17 357 , 17 866-67, 29 973, 33 134). Complaint counsel have failed to show that either Kellogg or General Mills had that degree of excess capacity over an extended period of time that would be necessary to enter into long-term commitments to provide private label products (See Tr. 13 787 550-51; CX-GF 159B).

408. General Foods has produced private label RTE cereal products since at least 1937, when it sold about 100 000 pounds ($10 000) of private label Rfe cereal (GFX 1370K). 409. In 1943, General Foods acquired the Jersey Cereal Company, with manufacturing plants in Irwin, Pennsylvania and St. Joseph, Missouri. Jersey produced a number of RTE cerealsincluding corn flakes, wheat flakes, bran flakes, rice flakes, rice gems, wheat puffs and rice puffs-some under its own name, but primarily under private labels (GFX 253F 1370K). 410. General Foods purchased Jersey in order to acquire certain '" Th'-H' is no evid,'ncc, or col tention . that Kellogg uCLJuired Miller with the intent of destroying private label competition Initial Decision patent rights, equipment and manufacturing knowledge, and to expand its private label business (Tr. 13 715-16; CX-GF 167Z-1O, Z- 11).

411. In 1947, after encountering problems in production, marketing and distribution, the private label business of General Foods showed an operating loss of $141 000 (CX-GF 121H). 412. In order to eliminate duplicative expenses and quality control problems, the plants at Irwin and St. Joseph were closed and their operations were transferred to Battle Creek (CX-GF 121G). 413. General Foods' private label business suffered further losses (CX-GF 121L). By 1953, General Foods had lost over $2 million (without full allocation of overhead costs) on its private label business since acquiring Jersey (GFX 1370L). 414. In order to save money in the marketing of its private label cereals, General Foods moved the responsibility for private (128) label sales from brokers to the General Foods sales division. Since this additional responsibility interfered with the sales force s marketing of General Foods' trademarked brands, responsibility for selling private label was shifted back to brokers (CX-GF 121F). 415. Around 1956, General Foods discontinued its business with what it considered ((borderline private label cereal customers" whose volume did not justify the expense involved in supplying them (Tr. 729). This cut-back improved General Foods' profit situation on its private label products (CX-GF 121P).

416. General Foods continued to supply private label cereals to three principal customers-A&P, Kroger and American Stores. The Jersey control brand name was dropped in favor of private label sales (Tr. 13 730).

417. General Foods subsequently decided to discontinue its private label cereal operations as it was not considered to be a profitable business (Tr. 36 590). General Foods' 1967 Task Force had recommended discontinuance of private label operations, anticipating that the future demand for branded products would require the time, space, machinery and efforts then being devoted to private label (CX-GF 120B, 133A).

418. General Foods utilized its excess capacity in the production of new products (Tr. 13 656, 14 119 591-93). 419. General Foods' private label cereal sales declined from a high of about 7 millon pounds in 1955 to under 5 milion pounds in 1960 and under 2 million pounds in 1970 (GFX 1370L). In 1966 General Foods' private label business was less than one- half of its 1949 level (CX-GF 121H.., 152B). Its 1965 level was approximately 134 FEDERAL TRADE COMMISSION m;CISIONS Initial Decision 99 F. the same as its 1944 level of private label production (including Jersey) (CX-GF 121M).

420. Thus, General Foods, unlike Kellogg and General Mills, did make an effort to compete by offering private label products. For a time, it increased its private label production beyond the point previously reached by the company it had acquired, in part, for the purpose of producing private label. There is no evidence, or contention, that the acquisition was made to remove private label competition.

421. Relying upon a number of requests by retailers for private label products (CPF &-154), Jersey s profitabilty in 1943, General Foods' profitability in the mid 1950' s and Ralston s profitability on private label business (Tr. 17 501- , 17 515-26), complaint counsel assert (CPF &-156) that "respondents' excess capacity on production systems, the significant retailer demand for private label, and the profitability of producing private label, are factors that should have encouraged respondents to engage in private labeling activity. The absence of private labeling activity in the face of these incentives strongly indicates that respondents tacitly agreed to avoid competition by restricting output of private label products." (129) 422. As found above, however, neither Kellogg nor General Mills have been shown to have had suffcient excess productive capacity to engage in private label competition; and General Foods tried, but was unable to generate the returns deemed appropriate. 423. Complaint counsel' s effort to second-guess the business judgment of respondents is made despite a failure to show that the sales and profit potential for private label products is greater than for branded products. It also flies in the face of an obvious disinclination of any company to private label its own products and so create competition for its own brands (Tr. 26 684, 28 365), as well as the difficulties in simulating the branded products of others (Tr. 9173 556 987-89 857).

424. It is concluded, therefore, that complaint counsel have failed to establish that respondents' activities with respect to private label RTE cereals were in accordance with any agreement, tacit or otherwise; that respondents' activities reflected anything other than independent business decisions made in an effort to further legitimate business interests.

Nonprice Competition Complaint counsel assert (CPF &-170) that "(rJespondents did not limit their tacit agreement to avoid competition to pricing and Initial Decision related areas. Rather, respondents also avoided other actions in the form of non price competition that might have led to price competition or that potentially might have threatened their goals of maintaining marketplace stability and maximizing profits at the expense of customers. " Here, complaint counsel include in-pack premium competition, competition for shelf space allocation at the retail level, competitive advantages by fortifying products and the exchange of data covering current advertising expenditures, as means of reducing marketplace uncertainties which otherwise might result in more active competition.

Respondents' activities with regard to in-pack premiums have been considered above as a price-related activity (Findings 363-83). We shall now consider the exchange of advertising expenditure data, shelf space competition and product fortification. The additional, but related, subject of competition in the introduction of new products will also be considered.

Advertising 425. Since before 1950, all three respondents and Quaker Ralston, and Nabisco each submitted to A.C. Nielsen Co., usually through their advertising agencies, their brand-by-brand advertising expenditures for network and spot television, network radio, (130) newspaper, and magazine advertising for the two months prior to submission (Tr. 11,842-43, 15 596-97, 15 600- , 15 641; CX-ACN 2; KX 14). Using this data, Nielsen prepared a bi-monthly report which detailed total advertising expenditures broken down by media for each brand of each company and, within seven weeks of the end of the bi-monthly period covered by the report, supplied it to the respondents and the other participating RTE cereal producers (Tr. 232- , 14 354, 14 356-58 , 14 953- , 15,597-99; CX-GF 40l0Z- , Z-36; CX-GM 176)"0 426. Complaint counsel assert (CPF 8-238) that this exchange of advertising information has served to allow respondents to prevent an expensive advertising war."41 However, there is no evidence that the respondents utilized the information to curtail or otherwise coordinate their advertising efforts. To the contrary, the record shows that industry members competed very strongly against each other in their advertising endeavors.

.0 General Mills did 'lot participate in 1950 IInd for some six years thereafter (Tr- 15 615-16). ., This assertion is inconsistent with the furtherasserlion (CPF ) 1-183, 11-205) that respondents have engaged io exe,,"siveand wasteful advertising. Initial Decision 99 F.T.C. 427. Total advertising expenditures for the five largest producers increased from $23 million in 1950 to $84 million in 1971." addition to this absolutely large amount, advertising also has been substantial in comparison to sales. The advertising to sales ratio in the RTE cereal industry exceeded 10%, sometimes by a substantial amount (Tr. 21 905 , 27 676; CX 513). This is substantially higher than for most other industries. Of more than 320 manufacturing industries, the advertising-to-sales ratio in the RTE cereal industry was among the two highest (Tr. 21 904, 27 678). 428. Total advertising for the three respondents went from $19. million in 1950 to over $86 million in 1967. During that (131)period, Kellogg s yearly advertising expenditures grew from $8.5 million to more than $40 milion. General Mills' went from $5.4 million to $25.7 million. General Foods' went from $5. 6 million to $19.5 million (GFX 1319; CX-GF 324J).

429. Advertising expenditures of those firms which furnished advertising data to Nielsen have varied widely (GMX 555). During the period 1958-1972, General Mills' advertising expenditures as a s in 10percentage of dollar sales were over 30% higher than Kellogg 200/0 higher in four of the remainingof the 15 years, and were over five years. On the other hand, General Foods' advertising expenditures varied from being 33% less than General Mills' in 1961 to 23% higher than General Mills' in 1970. Quaker s advertising expenditures were twice as high as Kellogg s in 1971 and were never lower than Kellogg s; Ralston s advertising expenditures were never lower than 28% above Kellogg s; and Nabisco s varied from 57% higher to 14% lower. None of the firms followed Kellogg, the alleged industry leader. The following table, derived from GMX 555, depicts usingadvertising expenditures as a percentage of dollar sales, Kellogg s expenditures as the benchmark (at 100%). " Them' expenditure . in millions or dollars. are as follows (CX 5Q(;Bi. 1950 1971 Kellogg '-5 32- General Mills 24.

General Foos 16.

Quaker 1.8 Nabisco 1.2 Total 2:J. 84_ '" These figures area!! advt'rti;;ing expenditures as a percentage "rdo\l arsiJles Initial Decision INDEX OF ADVERTISING EXPENDITURES AS A PERCENTAGE OF DOLLAR SALES 1958-1972 General General Year Kellogg Mils Foods Quaker Ralston Nabisco 1958 100. 1?6. 110.4 149. 144.5 157. 1959 100. 130. 92.4 113. 139.4 100. 1960 100. 107 91. 127. 182. 86. 1961 100. 131. 88. 182. 191. 99. 1962 100. 121. 967 168. 189. 104-7 1963 100. 135. 100. 140.7 170. 126. 1964 100 132. 111.6 132.4 152. 135. 1965 100. 134. 112. 131. 127.4 131.3 1966 1000 1260 133.5 1282 138. 140. 1967 100. 133. 142. 129. 141. 112. 1968 100. 142. 152_ 174. 134.4 122_ 1969 1000 132. 127. 133. 145.7 86. 1970 100. 124. 154. 155. 174.4 88.4 1971 100. 130. 150. 200. 115. 1972 100. 157. 146. 139. 138.4 NA - Data not available (132) 430. The following table shows advertising expenditures as a percentage of dollar sales for the three respondents: ADVERTISING EXPENDITURES AS A PERCENTAGE OF DOLLAR SALES 1958-1972 Year Kellogg General Mils General Foods 1958 1457 18. 16. 1959 17. 22. 15. 1960 16.26 17.43 14. 1961 15.42 20. 13.49 1962 15. 18.43 14_ 1963 15. 20.57 15. 1964 15. 20. 16. 1965 16. 22. 18. 1966 15. 19. 20. 1967 15. 20. 22. 1968 12. 17. 18.43 1969 12. 16. 15. 1970 10. 13. 16. 1971 9.74 12. 14. 1972 14. 13. (GMX 555A) 431. As the above tables demonstrate, Kellogg, General Mils and General Foods' expenditures on advertising followed markedly different and varying patterns during the period 1958-1972. (133) 432. Advertising expenditures on individual products have also Initial Decision 99 FTC. varied greatly. An examination of pairs of brands that Dr. Green studied in looking for brand pricing coordination fails to show parallel or coordinated advertising spending by respondents (See GMX 242-49). This is demonstrated by the following examples: ADVERTISING TO SALES RATIOS FOR SELECTED PRODUCT PAIRS A/S Ralio A/S Ralio A/S Ratio A/S Ratio General General Kellogg Mills Kellogg General Kellogg General Kellogg Mils Product Cocoa Cocoa Mils Rice Mils Froot Total 'M_ Puffs Krispies Cheerios Trlx Loop 281 1.81 132 137 188 278 353 147 139 141 144 0316 159 142 170 160 185 161 151 147 426 102 130 148 140 196 136 160 138 157 124 159 134 170 211 155 152 132 158 157 182 627 162 106 152 115 168 094 189 336 143 087 148 126 060 169 308 102 113 123 125 091 097 156 176 104 073 119 094 112 062 182 200 109 051 099 095 096 045 (GMX 244, 245, 247, 248) (134) 433. Complaint counsel, purporting to rely on Mr. Glassman and Dr. Schmalensee (Tr. 22 469- , 27 131-33) contend (CRPF 8-283) that such comparisons are not persuasive; that advertising-to-sales ratios would have to be adjusted for the introduction rate of new products before comparisons would be meaningful. Such an adjustment might well affect the comparison of overall advertising-to-sales ratios. However, it is compliant counsel's speculation that advertising expenditures may have been coordinated. And so it is complaint counsel' s burden of proof that has not been met. In any event complaint counsel's objection is inapplicable to the comparisons of advertising for the individual pairs of products selected for comparison by complaint counsel's own witness, Mr. Glassman. 434. Those economic experts who addressed themselves to an evaluation of advertising expenditures (both those introduced by complaint counsel and those introduced by respondents) agreed that there was no evidence of parallel or coordinated advertising endeavors by respondents (Tr. 22 467-BB, 27 129- , 38 565-B6, 38 570 622).

435. Complaint counsel assert (CPF 8-237) that "the mutual exchange of detailed advertising data permitted the respondents to make. . . (reductions in advertising expenditures in the late 1960' Initial Decision without fear of losing market share because each was able to coordinate its reduction with the others;" and that (CRPF 8-276) without such an exchange, respondents may not have been successful in sharply reducing advertising expenditures in the late 1960' This, however, is pure conjecture.

436. Respondents' executivies testified that no such arrangement was made (Tr. 29 725, 29,727, 32 768, 35 462-63, 35 815). And there is no evidence tending to establish the existence of any such agree- . ment. The tables indicating lack of coordinated action and the evaluative testimony of expert witnesses to the effect there was no such coordination, referred to above, covered the late 1960's period in question. Further, there is no basis for complaint counsel' conjecture with regard to an exchange of advertising information instituted some 20 years before the downturn in advertising. 437. While total advertising expenditures in the RTE cereal industry declined in the late 1960' , this was largely in response to the slowing of RTE cereal market growth. RTE cereal sales reached 1.177 billon pounds in 1965, but then leveled off, reaching only 1.190 bilion pounds by 1971 (CX Idle).

438. In 1968, General Foods reversed its long-term trend of advertising expansion and sharply reduced its advertising spending by more than $3 million (GMX 241). This was after the 1967 General Foods Task Force had recognized the decline in market growth and anticipated the continuing absence of RTE cereal market growth in planning its late 1960's marketing strategy (CX-GF 4039V). Further General Foods was shifting its marketing emphasis away from advertising to individual product promotions as part of its (135) withdrawal from a full line advertising strategy to that of pushing individual brands (Tr. 27 369, 36,405-06; CX-GF 4039Z-114, Z-115). Overall advertising and promotional expenditures of General Foods for 1966 through 1970 showed no decline (GFX 24). 439. Advertising is expected to decline in a competitive industry as sales growth stops (Tr. 27 368-9, 38 566-7). There is nothing suspect, therefore, in the fact that Kellogg s and General Mils advertising had begun to decline by 1968, and that advertising of Quaker, Ralston, and Nabisco also declined between 1967 and 1969 (GMX 241; GFX 1319)." Apart from a uniformity in decreasing the percentage of advertising expenditures in 1968, the following table taken from GMX 241, shows a disparity of action on the part of respondents both before and after that year: .. A. previously found (Findings 33Q.33, 349, 352, 353, 383) respondenl ' use of various product promotions increas as their advertising decreased Initial Decision 99 F.TC. ANNUAL ADVERTISING EXPENDITURES BY COMPANY AND PERCENT AGE CHANGE FROM PREVIOUS YEAR 1964-1972 Kellogg General Mils General Foods (SOOO) Change (SOOO) Change (SOOO) Change 1964 363 19,751 886 1965 39,071 10. 017 26. 115 1966 40,212 22,404 10.4 419 + 20. 1967 40,877 619 14.4 486 1968 33,847 17. 21,029 17. 463 15. 1969 190 085 100 14. 1970 32,740 18,750 11. 399 1971 853 130 098 4.5 1972 32,988 018 32. 076 440. General Mills examined the marketing expenditures on a competitive product and its sales volume in order to evaluate what might be working Of not working for a competitor in comparing the sales results of General Mills' own advertising efforts. There was no attempt to match spending levels (Tr. 15,945- , 17 651- , 17,656). Advertising spending levels for individual products were based on the recent and projected sales performance of tbe product, its recent advertising levels, its responsiveness to advertising and its existing or targeted consumption profile (Tr. 17 651, 17 303, 17 391). (136) 441. Kellogg, General Foods, and Quaker also used such advertising information to correlate advertising efforts with sales results and so evaluate advertising efficiencies- Quaker and General Mills used sales and advertising data to project their own sales volumes (Tr. 12 804- , 14 233, 14 953, 14 976 771). 442. While competitors' levels of advertising expenditure were considered by respondents in setting advertising budgets for their cereal brands, they were not a controllng factor (Tr. 12 367, 15 243 174, 651- 656 29,712).

The complaint charges respondents with having maintained and utilized monopoly power. The exchange of advertising data has not been alleged as a violation of law. Therefore, the failure to show that such an exchange was utilized for monopolistic purposes constitutes a failure or proof under the instant complaint. 443. It is concluded that respondents engaged in very heavy advertising as a means of competing in the sale of RTE cereals. It is concluded further that, while respondents exchanged advertising data, such data was employed for legitimate business purposes, and that there is a lack of evidence to support complaint counsel's initial Decision speculation that the data exchanged was used to parallel or coordinate advertising endeavors.

Competition For Retail Shelf Space 444. Complaint counsel assert (CPF 8-170, 8-189, 8-204) that respondents had a tacit agreement to avoid competition for shelf space in retail stores; that this was accomplished by Kellogg formulating and implementing a shelf space allocation plan and General Mills and General Foods acquiescing in that plan. General Mills and General Foods executives have denied the existence of any such agreement or conspiracy (Tr. 32 712- , 35 462, 35 815, 36 527). The record furnishes no basis for not accepting this testimony. 445. Normally, the RTE cereal section in a store comprises most if not all, of one side of an aisle (CX-K 553J; Tr. 6523, 6560-1 , 7291 7607- 8308-11 8818-22 9062-65 9070-71). As of November 1961 there were over 100 brands and sizes of RTE cereal products from which the retailer selected about 75 to 90 to be displayed (CX- 553J; KX 3). In shelving these products, decisions must be made on (i. thewhere to locate each RTE cereal and how many "facings space needed to place one RTE cereal package broadside on the shelf with the front panel of the package displayed) each is to be given. Additional packages are placed in a row behind each front facing (Tr. 8090- , 8346-47).

446. Apart from preventing out-of-stocks, there is a degree of relationship between sales volume and the amount of space an RTE (137Jcereal product receives on the shelf (Tr. 8817). The location of the RTE cereal product on the retail shelf is also a factor in 135C, 547H-promoting sales (CX-GF 4039Z-12; see also CX-K 95A J; CX-GM 148, 704; CX-GF 324C , S, 325P, 0, L, 560-560A; QX 21 C). Traffic patterns areaka CX-K 108; GMX 7B, C aka CX-GM 152B considered in selecting preferred locations (Tr. 9048-56; CX-K 556D E). Either end of the aisle (Particularly the end reached last) is considered less desirable, and respondents usually preferred the center of the gondola where visbility to a consumer is normally the best. Also, a center location permits an expansion of space in either direction (Tr. 6520- , 7210-11, 7283, 7291- , 7695- , 8233- 8313- 8750 8820 8949-50; CX-GM 148C). 447. Preferred locations would frequently vary from store to store, depending upon traffic patterns (Tr. 7684- , 8121-24). General Foods tried to have its products shelved next to Kellogg with its Post Toasties next to Kellogg s Corn Flakes (Tr. 8820). 448. Many of respondents' officials believed that grouping a Initial Decision 99 F. company s products together increases the manufacturer s opportunity to capture the customer s attention and, therefore, his impulse purchase (Tr. 6503, 6521 , 6558, 7646-8, 8100, 8113- , 8819). 449. There is a conflict in the evidence with respect to the importance of impulse purchases. General Mills executives testified that most RTE cereal purchases are planned and that the space allocated for a cereal may affect the time spent by the consumer searching for it, but will not have an impact on the product' s sales (Tr. 7829- , 8022, 8097, 817 77). On the other hand, General Foods' 1961 marketing plan contains a reference to recent studies showing that 71.8% of all cereal purchases at the store level are made on impulse (CX-GF 324S). Since the studies are not identified and respondents in 1976 (when the exhibit was received) had no opportunity to cross-examine any witness with respect to the unidentified pre-1961 studies, the 71.8% estimate cannot be accepted as an accurate figure. Further, if RTE cereals were primarily impulse purchase items, respondents would not be making the immense advertising and promotional expenditures that they are to presell their products ( see, supra, Findings 427, 428, 430, 439); and Dr. Scherer (and other economic experts) would not have felt required to make adjustments for advertising persistence rates when converting accounting rates of return to economic rates of return ( see, infra Findings 682, 685-7, 700, 768-91). 450. Before Kellogg developed its shelf space plan in the 1950' retail grocers varied in their methods of shelving RTE cereal products. Some would group cereals by grain corn-based products together, wheat-based products together, oat-based products together, rice-based products together, etc. Some retailers would group cereals by types, presweetened cereals together, unsweetened cereals together, etc. Other retailers used other methods of shelving, including grouping by manufacturers (Tr. 6504-D5, 6514- , 8018, 8084, 8113, 32 675, 32 722). Also, most RTE cereal brands were accepted by most large retail outlets and few RTE (138Jcereal products were allocated more than a minimum number of facings on the shelf (CX-K 553I, J, K). This meant that Kellogg, whose products were the largest sellers, was not getting its proportionate share of shelf space and that there were out-of-stocks of Kellogg products. 451. As grocery stores were changing from over-the-counter to selfservice, Kellogg salesmen in particular, as representatives of by far the largest cereal producer, were called upon by retailers to offer shelf placement recommendations (Tr. 36 818). In 1956, Kellogg developed a standard Kellogg approach to shelf placement recom- Initial Decision mendations. Space according to sales and company grouping were the two key elements of the Kellogg program (Tr. 6474, 6501). 452. Starting in 1956, and thereafter, Kellogg made recommendations to retailers about how to organize their cereal shelves using these two principles. The recommendations covered the products of all major manufacturers because retailers needed to see how the whole section would look (Tr. 6474, 6501, 6503- , 6514). The same two general principles governed Kellogg s use of its "Magnaboard" (a mock-up of a cereal shelf with movable parts representing individual cereal brands) introduced in the early 1960's (Tr. 6542--3). 453. Sometime in the late 1960's, Kellogg computerized its shelf space plan. Where the plan was accepted, the retailer supplied Kellogg with its own sales data for each product (Tr. 6657). The computerized Kellogg shelf space plan, called Computerized Space Allocation ("CSA"), consisted of two phases. In Phase I, Kellogg recommended a change in the case size ordered for each cereal. The recommendation was usually to lower the case size pack of competitive products, which resulted in a reduction in facings for such products (Tr. 6943, 7114; CX-K 1063).

454. As part of Phase I, Kellogg recommended which items should be discontinued. It generally recommended that a product be discontinued when its share was less than .5% of all RTE cereal sales, or when the product was selling only four to five packages per week. However, the particular cut-off for discontinuance of a product, or of a size of a product, varied by account in accordance with what would minimize Kellogg s losses and maximize competitive losses. In all cases, an attempt was made to maintain distribution of Kellogg products. Discontinuance of a Kellogg product would be recommended only if it could take a number of competitive products with it (Tr. 6807-08 7124-25). Phase II of the Kellogg program allocated the shelf space among the remaining cereals by volume (CX-K 1065).

455. Kellogg s shelf space efforts and strategies were designed to create more sales for Kellogg at the expense of its rivals (CPF 8-187 8-191, 8-196, 8-204; Tr. 6501-03). The primary purpose underlying Kellogg s shelf space program was to reduce Kellogg out-of-stocks and get more retail shelf space for its RTE cereal products. Kellogg share of shelf space and inventory historically (139Jhad been less than its share of sales. Thus, the principle of sales according to shelf space enhanced the placement of Kellogg s products at the expense of those produced by its competitors (Tr. 7095 7103-06 7285, 8758 8769-70).

456. Kellogg recommended that retailers shelve RTE cereal Initial Decision 99 F. products in company groupings so that Kellogg s lower volume products would be next to its high volume products and gain increased sales as as result of impulse purchases (Tr. 6501- , 6521- 7662- 7744 7821- 8090-92).

457. Kellogg, therefore, unilaterally instituted a shelf space program which was designed to afford it a competitive advantage over other RTE cereal manufacturers, including the other respondents. There is no reason to believe that General Mills or General Foods was party to the institution or implementation of Kellogg plan.

458. Kellogg s shelf space program was one of making recommendations to the retailers. The retailers made their own decisions on how to shelve RTE cereals (Tr. 6527 , 6762, 7121, 7146, 7149- , 7208 7210, 7669, 7687, 8266-67, 8274, 8494, 8840- , 8893- , 8910- 8917- , 9039, 9092- , 9112, 9115, 9364- , 29 960- , 29 964 342-43). To the extent retailers adopted Kellogg s recommendations, they did so because such recommendations served their own profitability and efficiency interests (Tr. 8904- , 9032- , 9136 9377-78).

459. Kellogg s shelf space allocation principles were followed in a majority of retail outlets (CX-K 553Q, 556G-G), although utilization of Kellogg s CSA computerization service was not as universal (KX 2). The principles appear to have been adopted because they were consistent with, and in the best interests of, the retailers. 460. The principle of space according to sales ensured that the retailers would avoid out-or-stocks and over-stocks, increase their efficiency and profitability and reduce labor costs (Tr. 6540-42, 6694 6722- , 7151 , 7206- , 8904-7, 8906-07, 9032- , 9136, 9377- 746, 36 813). Consequently, the retailer would achieve a better utilization of its capital and a better return on its investment (Tr. 7109 7606-7 700-01).

461. Kellogg did not invent the principle of allocating shelf space by sales volume (Tr. 7213). Sales volume is, and has been, the basic method of space allocation throughout grocery stores. A retailer profitability is directly related to sales turnover. The allocation of shelf space according to sales volume reduces the likelihood that a product will be out of stock, maximizes turnover and return on investment and minimizes lost sales and lost profit both to the retailer and the man'1facturer (Tr. 6540- , 7085- , 8010- , 8021- , 8091 , 8838- , 8904-05, 9032- , 9135- , 9195- , 9377- 745-46; CX-NCFM 500 at 222).

462. As Kellogg s Director of Marketing Research testified (Tr. 7085): (140) Initial Decision WJrofitability in the operation of the cereal department revolves around turnover, the amount of stock that has to be carried, labor costs in placing the items on the shelf and, to be more specific, profitability, if you will ake, for example, the item turnover by allocating space according to sales, what that tends to do is to reduce the day supply on the slower selling items and increase it on the faster selling items, and consequently increases the sales on those faster selling items because it decreases the out-or-stocks.

So what you have got is a better return on investment because your merchandise is turning Dvcr faster in total In the back room stock, I noted earlier that back room stocks have diminished over the years. In fact, today, the hackrooms don t anywhere near resemble what they used to be in the 1950's. Very little hackroom stock is given for any product on the shelf and especially cereals. By reducing the back room stock the retailer has tended to reduce his investment Therefore he is selling more goods today than he ever did before with less stock and therefore he is making more money.

463. Shelf placement by manufacturer was commonly used by retailers for a wide variety of goods other than RTE cereals (Tr. 8762- , 8847- , 9036-39, 9098, 32 727-29). Retailers believed that grouping products by manufacturer served both the consumer s best interests as well as their own. It was the most efficient arrangement from the standpoint of the consumer s ability to locate quickly the RTE cereals he or she was looking for (Tr. 8 113, 8888, 9032- 9035- , 9039, 9136, 9377- , 32 727). This system facilitated inventorying and ordering because the stores' order guides were arranged by manufacturer (Tr. 32 726).

464. To the extent that other arrangements, such as stocking by grain type, have been tried, they were unsuccessful (Tr. 9039, 9139- 40). Each RTE cereal has many characteristics important to the consumer and no single one can be effectively used as a basis for grouping. Grain, for example, is not a workable basis for grouping because there are many multigrain cereals and because grain is not always the most important feature to the consumer (Tr. 12 684- 694 717 727). (141) 465. Faced with Kellogg s shelving program which advocated space 'according to sales and grouping by manufacturers, which principles were logical and advantageous to retailers, as well as ones to which retailers were accustomed, it is not surprising that General Mils and General Foods advocated the same guidelines as they competed for shelf space (Tr. 7765- , 7781, 7789- , 8020-31; See CX-GM 148; GMX 7, CX-GF 560, 4039Z-86).

466. While all three respondents generally advocated shelving by manufacturer and allocation of space according to volume, they each stil competed for all the space it could get. This included efforts to get more than a "fair share" of shelf space if its credibility would not be impaired by doing so. And each recommended the discontinuance Initial Decision 99 F. of slow moving products of its competitors (Tr. 6548, 6565, 6808 7126 7279 7644 7653- 7660- 7669- 7721- , 7745, 7659- 7670 8003- 8038- 8132- , 8376-77, 8619- , 8841-42, 8623- , 9145-46, 9171- , 32 699-701 , 32 707- , 32 713- , 32 720 736- 526-27 820- 892-94).

467. General Mills recommended space allocation by sales volume only where this would lead to an increase in shelf space for General Mills at the expense of its competitors (Tr. 8004- , 32 698- 99). Generally, it tried to take the additional facings needed from Kellogg and General Foods, because products of those companies presented the heaviest competition to General Mills (Tr. 32 699-702). 468. In retail accounts where General Mills had a share of total RTE shelf space which was greater than or equal to its share of sales in the region, it recommended space allocation according to sales volume only for those products and sizes which had a share of shelf less than share of sales (Tr. 8005, 8054- , 32 705-06). 469. Recommendations to increase facings beyond share of sales could not be justified as being in the economic self-interest of the retailer, and would not sit well with him. Such recommendations would have an adverse impact on the salesman s credibility (Tr. 8918, 9081 , 9179-80). Thus, a respondent's failure to recommend space beyond that indicated by volume is not an avoidance of competition, but is totally consistent with competition and the respondent' s economic self-interest.

470. For its items which already had shelf space equal to or greater than sales, General Mills aggressively used other strategies such as promotions and displays, or attempts to get the retailer to accept a second size of the product (Tr. 8005, 32 704-05). If successful such efforts would increase sales. Once sales increased to the point where share of market was greater than share of shelf space General Mills would recommend an increase in facings based upon the principle of space allocation according to sales volume (Tr. 8005 704-05).

471. General Foods' salesmen sought to have its cereals shelved where they would get the most exposure, preferably some distance down (l42Jthe aisle in order to afford customers easier access (Tr. 8814 797). At the same time, however, a salesman could not make ludicrous proposals that would destroy rapport with the grocer and his credibility as an advisor that Kellogg s cereals should not be shelved at all, or should be shelved at an inconvenient location despite their relatively greater consumer popularity (Tr. 8918, 9179- 791-92).

472. In 1968, General Mils developed a more formalized shelf Initial Decision space program (CX-GM 148). It was entitled Cereals Organized for Profit ("COP"). General Mils' COP kit was developed as a partial response to the Kellogg computerized shelf space program, which Kellogg came out with in the late 1960's (Tr. 32 717-18). The main selling point for COP was that it used regional market share figures to allocate facings on the cereal shelf as opposed to share data based solely on sales by the particular account, as the Kellogg plan did. For example, if Cheerios enjoyed a 5% share in a region, but only a 4% share in a particular grocery chain, General Mills would urge the retailer to increase Cheerios' facings to its regional share level of 5% (Tr. 718-19).

473. The COP kit was a sales aid which allowed salesmen to simulate an RTE cereal aisle. The kit, as distinguished from the use of sales and shelf facings data, was generally used only when a complete reset of the cereal section was contemplated. This was usually only once a year, because it required a lot of preparation time on the part of the General Mils salesmen. The kit was abandoned as a clumsy sales aid, but General Mils continued to vie for shelf space (Tr. 32 719, 32 717-22).

474. From 1955 to 1965, Nielsen market share data was utilized by General Foods to recommend space allocation to retailers. Toward the end of that period, warehouse withdrawal figures were used in place of Nielsen data when requested by retailers (CX-GF 560). Thus, the General Foods plan, unlike Kellogg s program, recommended that a manufacturer s products be placed on the shelves in proportion to that manufacturer s market share rather than its share of sales in the particular store.

475. In 1955, all General Foods cereal packages were redesigned to achieve a "family resemblance " and line packaging was emphasized (GFX 1255 0, 1256W, 1288F). Accordingly, General Foods salesmen were advised to try to get all General Foods cereals shelved together to create a "billboard effect" and to stress the line of products (Tr. 36 346, 36 809).

476. By 1956, the Post Planned Shelving Program had taken shape. It advocated: (a) space according to sales as measured by A. Nielsen; (b) shelving by company; and (c) General Foods cereals (143) in the middle of the cereal section with Post Toasties next to Kellogg s Corn Flakes (CX-GF 4039Z-86).

479. While the Planned Shelving Program continued to provide the basic framework for General Foods' shelf activities, in April 1964, General Foods introduced "Compact Packages" (GFX 1304A). .. At One point in.the 1960', General Foos experimented with Tfxommending groupiog RTE cereals by grain. However, the plan was a failure (Tr. 8630, 8814, 9140). g., Initial Decision 99 FTC. Compact Packages were resized versions of existing cereal packages designed to hold equivalent quantities of cereal in smaller boxes (Tr. 841).

480. These new packages were intended to appeal both to consumers and to the grocery trade (Tr. 36 386-87). The anticipated consumer benefits were: (a) a better fit on home cupboard shelves; (b) greater stability; and (c) easier handling. The anticipated trade benefits were: (a) better space utilization more cereal in the same amount of shelf space; (b) greater package stability; and (c) better pallet patterns (GFX 1304B C).

481. General Foods anticipated increased sales and profits arising from (a) consumer preference for, and therefore purchase of, the new packages relative to the other manufacturers' old-style packages, and (b) increased distribution of General Foods cereal brands by placing additional items in the section of cereal shelves where its cereals were placed ('fr. 36 391- , 36 837). General Foods also anticipated a benefit from being perceived by the trade as a leader in innovative cereal shelving, and thereby being able to secure greater acceptance for its shelving recommendations (Tr. 36 841). 482. After an advertising campaign and promotional efforts (Tr. 391, 36 393, 36 839-40), General Foods' compact packages made an initial favorable impact, and it gained some additional facings. Over time, however, the program was a failure, and General Foods ended up losing shelf space (Tr. 36 392- , 36 841-42; GFX 1301). 483. In 1965, General Foods terminated the Compact Packages program and returned to larger packages (Tr. 36 831; GFX 1310 1324).

484. In 1971, General Foods developed and introduced A.S.S. (Customer Oriented Method of Profitability' and Sales Service), a computerized system for making shelf space allocation recommendations on the basis of product profitability (GFX 1371). (144) 485. This effort was undertaken because some of the Post Division s most popular products (e. Post Grape Nuts, Tang breakfast drink, Gaines dog food) were packaged in relatively smaller containers when compared with competing products in the same category. Therefore, they were assigned relatively few linear feet of shelf space by retailers, although their profitability per linear foot was relatively high (GFX 137JB).

486. Preliminary presentations of the C. S.s. system '" Grocers, instead Dr stacking addition a!' General Fuud cereals in lhe newly available pacc, continued to allocate facings according toa!es ,md so placed other ma!",facturers' products in SOm,' of the space where General l"ooscen,,,I had beenlTr. 36.8421.

Initial Decision were made to a few selected major accounts (GFX 1371C). The accounts' reactions to the presentations varied (GFX 1371D). 487. The Post Division staff concluded that the trade s reaction to S. justified further development. During the summer of 1971, the program was further developed and a C. A.s.s. Manual was distributed to Post Division salesmen (GFX, 1236 1372D).

488. Efforts with C. S. continued throughout 1971 and 1972 (GFX 1371E). However, attempts to use the C. A.S. program as a sales tool to gain additional RTE cereal shelf space for General Foods products were largely unsuccessful, as few retail accounts were willing to provide the data necessary to permit a A.S.S. analysis of their RTE cereal sections. It was the consensus among the Post Division sales personnel directly involved in the development of C. S. that three principal factors explained the low level of interest among retailers in using it for their RTE cereal sections: (a) grocery retailers were generally satisfied with the results of their allocating cereal shelf space on the basis of movement, (b) some retailers had their own internal information systems which provided the same information as C.O. S.s., and (c) many retailers were not sophisticated enough to understand the concept underlying, or the potential benefits of using, C. A.s.S. (GFX 1371E).

489. Respondents' shelf space programs do not preclude new products from being allotted space. Procter and Gamble was not concerned over its ability to secure shelf space for an RTE cereal product it might introduce ('1r. 25 836-37, 25 864-65). When natural cereals became popular, adequate space was available for these products (Tr. 25 922- , 37 324- , 37 312-15). Kellogg s program provided for two facings for a new product (Tr. 7156-59). Slower moving products often were allotted a larger share of shelf space than their share 0f the market called for ('1r. 8017- , 9057- 9177- 9379, 32 737-38). There is no evidence that position on the shelf is of such ilIportance that it would stop a new entrant. While some positions are better than others, there is not a truly bad spot in the total RTE cereal aisle (Tr. 9175-76). 490. It is concluded, therefore, that the record does not support complaint counsel's assertion that respondents had a tacit agreement to avoid competition for shelf space in retail stores. Kellogg independently formulated a shelf space allocation plan that incorporated principles which were in accord with retailers' preferred methods of doing business. The other respondents, faced (145Jwith the same requirements of retailers, responded with plans that Initial Decision 99 F.TC. incorporated the same basic principles. While each respondent competed for the most favorable shelf location and the most space it could get, it was constrained not to push for more than a reasonable share in order to maintain rapport and credibility with retailers. Nevertheless, both General Mils and General Foods did present shelving alternatives and variations in an effort to gain competitive advantages.

Fortification of Cereals 491. Complaint counsel assert (CPF 8-170): "Widespread product fortification was avoided until outside pressure forced the respondents to fortify their cereals. Even then, fortification occurred only as a result of coordinated activity, thereby ensuring that no one of the respondents would gain a competitive advantage by introducing fortified products before others were prepared to fortify their cereals." And (CRPF 8-264): "(RJespondents agreed upon fortifying their RTE cereal lines in concert." The record does not support complaint counsel's assertions of agreement and coordination. 492. The addition of vitamins, minerals or other nutrients to cereal products is described as "fortification" or "restoration. Restoration is the replenishment to whole-grain levels of naturallyis theoccurring nutrients lost during processing. \' Fortification" addition of nutrients in amounts greater than those occurring in the whole grain, or of nutrients not naturally found in the grain (Tr. 714 35,807).

493. The Council of Foods and Nutrition of the American Medical Association issued standards on the addition of nutrients to foods as early as 1939. These standards were to the effect that there were both nutritional and economic problems associated with the supplementation of food with specific nutrients, since supplementation might be wasteful and might have deleterious effects. This position was reaffirmed in 1946, 1953, and 1961. The last two times the position was announced in conjunction with the Food and Nutrition Board of the National Research Council. In May 1961 , the two groups issued a joint policy statement which recommended that nutrients be added to foods only to the extent of restoring those which were contained naturally in the food item, but which had been lost by cooking during part of the manufacturing process. It was specifically recommended that the addition of nutrie.,ts to RTE cereals be limited to the restoration of the natural occurring nutrients (GMX 540).

494. Consistent with these standards, Kellogg has added thiinitial Decision amine, niacin, riboflavin and iron to restore its cereals to wholegrain levels ever since 1941 (Tr. 11 806-7 , 29 714, 29 948). Similarly, General Foods restored most of its RTE cereals to so-called wholegrain levels with essential B vitamins and iron in the middle 1940' (Tr. 14 115 051; GFX 1370,)). (146) 495. By 1941, Kellogg had fortified its Corn Flakes with vitamin , and its Pep brand with vitamin D and vitamin B-1 (Tr. 29 715-17). Later in the 1940' , its Raisin Bran was fortified with iron and its Corn Soya had extra protein (Tr. 29 716). 496. Kellogg introduced Special K in 1956, fortified with seven vitamins and enriched with 20% protein (Tr. 29 716, 29 948; CX- 564C) and, in 1959, it introduced Concentrate, also vitamin fortified and containing 40% protein (Tr. 29 716). Vitamin D fortification in Kellogg s Corn Flakes was increased to 100% MDR" in 1956 (CX- 457).

497. Product 19, fortified with 100% MDR of eight vitamins and iron, was introduced in 1967. (Tr. 29 663), and Frosted Mini-Wheats was introduced into test-market with fortification to the one-third MDR level in 1969 (Tr. 11 786-87; CX-K 573F). Kellogg fortified its Sugar Smacks with 33% MDR of six vitamins about 1968 (CX- 7175S 7352C). Kellogg s 40% Bran Flakes received 100% iron fortification in 1966 (CX-K 487), as did its Raisin Bran by 1969 (CX- K 415, 7177F).

the 498. When General Foods developed Oat Flakes in 1962, product consisted of oats, soy, and wheat, and initially rice. However since the product was called an oat flake, it was fortified to the point of having the protein quality and quantity, the nutritional micronutrients, the vitamins and the minerals of whole oats (Tr. 37 051-52). In 1966, one-quarter of the daily requirement of a series of vitamins and minerals was added to the product and the name was changed to Fortified Oat Flakes (Tr. 13 616-17).

499. After fortifying Oat Flakes, General Foods considered fortifying its entire cereal line (Tr. 37 059). In 1967, Sugar Crisp was fortified to one-third MDR and reintroduced as Super Sugar Crisp (Tr. 37 052).

500. According to complaint counsel, respondents avoided widespread fortification prior to 1971 , at which time, pursuant to agreement, they fortified their RTE cereal lines in concert. However the following fortified RTE cereal products were in the market by 1970:

MDR" refers to the minimum daily requirements fot certain vitamins and minerals as set by the United Stal.sgovernrnent.

Initial Decision 99 F. KELLOGG: Kellogg s Corn Flalws fortified with vitamin D by 1941, increased to 100% MDR, 1956 (Tr. 29 716-17; CX-K 457); Pep, fortified with vitamins D and Bl by 1941 (Tr. 29 715- 17); Special K fortified (147)with seven vitamins and protein, 1956 (Tr. 29 716 948; CX-GM 564C); Concentrate vitamin fortified and protein enriched, 1958 (Tr. 29 716); 40% Bran Flakes fortified with 100% iron, 1966 (CX-K 487); Product, fortified with 100% MDR of eight vitamins and iron, 1967 (Tr. 29 663); Sugar Smacks fortified with X MDR of vitamins and iron, 1967 (CX-K 7135S, 7352B C); Mini- Wheats fortified with X MDR of vitamins and iron, 1969 (Tr. 11 786- 87; CX-K 533); Raisin Bran fortified with 100% iron by 1969 (CX- 415, 7177F).

GENERAL MILLS: Hi-Pro fortified with seven vitamins and iron 1958 (GMX 180); Total fortified with 100% MDR of eight vitamins 1961 (GMX 179B); Corn Total fortified with 100% MDR of eight vitamins, 1966 (GMX 180B); Vital, vitamin fortified, 1967 (Tr. 162); Alive fortified with extra Vitamin B2, Niacin and Iron, 1968 (Tr. 16 979; GMX isle); Kaboom fortified with 100% MDR of eight vitamins and iron, 1969 (CX-K 765E; GMX 174B); Buc Wheats fortified with X MDR of eight vitamins, as test marketed August 1970 (CX-K 765E; GMX 179B).

GENERAL FOODS: Bran Flakes fortified with iron, 1966 (CX- 487); Fortified Oat Flakes fortified with six vitamins and minerals 1967 (Tr. 37 052, 37 059; CX-GF 1406C); Sugar Crisp, fortified with X MDR of vitamins and minerals, 1967 (Tr. 37 052); Alpha Bits vitamin fortified to X MDR, 1969 (CX-GF 601H); Honeycomb vitamin fortified to X MDR, 1969 (CX-GF 601H). QUAKER: Quisp and Quake fortified with X MDR of vitamins and iron, 1968 (Tr. 15 044-45); King Vitaman fortified with 100% MDR of vitamins and iron, 1970 (CX-K 765C).

501. In 1970, vitamin fortified brands accounted for some 17% of industry pound sales, and cereals fortified with iron accounted for another 6% of the market (CX 434). Respondents' fortification efforts prior to 1970 achieved competitive results-for example, the competitive impact of Kellogg s Product 19 on General Mills' Total (CX- 7176A; CX-GM 567 A), the fortification rivalry between Kellogg Product 19 and Special K and General Foods' Fortified Oat Flakes (CX-GF 34A), and between General Foods' Sugar Crisp and Kellogg Sugar Smacks (CX-K 595A, 7352B, C, I).

502. This fortification activity prior to 1970 took place not only in the climate of the pronouncements of the Council on Foods and Nutrition of the American Medical Association and the Food and g., Initial Dccision Nutrition Board of the National Research Council, related above (Finding 493), but also in the facc of an adverse position taken by the Food and Drug Administration (FDA).

503. FDA policy adverse to the fortification of RTE cereal products was a major obstacle to RTE cereal fortification. During (148)the mid-1960' , the FDA publicly opposed fortifying RTE cereals beyond allowable limits, on the ground that the availability of vitamins to consumers from other sources made cereal fortification unnecessary (Tr. 29 949- , 37 054-59). 504. In 1962, the FDA commenced rulemaking proceedings concerning regulation of processed food fortification (27 Fed. Reg. 5815 (June 20, 1962)). In 1966, it proposed a rule, Part 80.2 of which would have prohibited the addition to RTE cereals of vitamins and minerals other than niacin, thiamin, riboflavin and iron. Minimum and maximum limits covering the use of these four nutrients would also have been established (31 Fed. Reg. 8525-26 (June 18, 1966)) (KX 101; Tr. 29 718, 37 056). Part 80.2 embodied the prevailing FDA policy that restoration of vitamins to whole-grain levels constituted the maximum appropriate vitamin supplementation (Tr. 37 056). FDA held protracted hearings regarding the proposcd regulation (Tr. 718; CX-CI 80A).

505. If adopted, the FDA regulation would have precluded the high levels of fortification subsequently adopted by respondents in 1971-1972, and would have required the reformulation of certain established fortified cereals (KX 101; see, e. GMX J 74 through 181; CX-GF 102K).

506. At the time of the FDA's rule proposal, respondents anticipated serious impact on their fortified cereals (Tr. 29 949). For example, in 1966, General Mills believed that the regulation threatened the future of its fortified Total: "(T)he probability of passage creates a serious potential threat to the brand" (CX- 564C). General Mils' 1969-1970 Total marketing plan stated: Possible future FDA regulations could force elimination or massive reformulation of Total" (CX-GM 567 A). Kellogg delayed increasing the iron content of Raisin Bran from 80% to 100% MDR until after it was advised that such action would not cause a government reprimand (CX-K 439B).

507. The effect of the FDA policy was to delay fortification of many cereals until the 1970' s (Tr. 13 141, 29 717- , 35 812, 37 053- 55).

508. In addition to the FDA, American Medical Association and National Research Council opposition to fortification of foods and perhaps as a result of such opposition, there was limited consumer Initial Decision 99 F. demand for fortified RTE cereals during the 1960's and only limited temporary success for products that were fortified (CX-GF 102G; CX-K 487; Tr. 15 043-44, 16 980-82). The record fails to indicate that respondents' individual competitive efforts prior to 1970 in the field of product fortificaton were not fully commensurate with the public demand. There is no evidence that indicates that respondents reached agreements concerning, or coordinated, their pre-1970 fortification conduct.

509. Throughout the pre-1970 period and thereafter, respondents individually engaged in extensive research to overcome technical (149)problems involved in cereal fortification. There have been many such problems including workable methods of fortification application, uniformity of product, unacceptable taste, odors and appearance, deleterious chemical reactions, maintenance of vitamin potency through the cereal processing procedure, maintenance of proper product moisture levels and shortened shelf life. These problems varied product by prodl and by the particular nutrient and combination of nutrients involved (Tr. 12 367 , 13 004 , 13 339-401 670- , 13 668-69, 14 150, 16 982, 29 280- , 29 951 , 32 911 916-17, 35 808, 36 670- , 37 002, 37 052- , 37 062-B3; GMX 373).

510. Some of the problems have been very difficult to overcome. For example, General Mills has never been able to add vitamins A and D to Cocoa Puffs or to add minerals other than iron to any cereals because of bad taste problem (Tr. 35 803, 35 813). Considerable expenses were incurred in researching and implementing product fortification (Tr. 16 982, 29 952-53; CX-GF 477). 511. General Foods' research into beneficial product additives was not limited to vitamin and mineral fortification. Starting about the end of 1959, General Foods promoted research on the possible inclusion of phosphates to inhibit dental caries. Much of this was done in conjunction with the Indiana University Foundation, as well as outside specialists and statisticians. Because of what appeared to be questionable research procedures, Indiana University and General Foods agreed to terminate the joint research arrangement in 1972. General Foods continues to perform animal research with phos phates (Tr. 37 110-11).

512. General Foods' unilateral investment of substantial sums in research into the prevention of tooth decay is inconsistent with complaint counsel' s allegation that General Foods, in conjunction with other respondents, avoided having one respondent acquire a competitive advantage over the others. Respondents' overall, vigorous competition in the introduction of new products (infra Findings Initial Decision 530-02) is also totally inconsistent with the charge that they conspired in the particular area of fortification. 513. In the late 1960's and early 1970' , there was a dramatic change in the national attitude toward the fortification of cereals. During 1969, the President had convened a White House Conference on Nutrition, which issued a report in December 1969 (GMX 501). That report was to the effect that there were significant nutritional deficiencies in the diets of large segments of the population. It was recommended that the proposed FDA regulations barring the fortification of breakfast cereals not be adopted, because the widespread acceptance and consumption of breakfast cereals made them effective carriers of essential nutrients. It attacked the view that all needed nutrients were obtained from ordinary diets and recommended strong food fortification programs (Tr. 35 811-12; GMX 501Z94-96, Z120-22, Z-253). Consequently, the FDA abandoned its proposed rule to prohibit food fortification (Tr. 29 718 057-58). (150) 514- The Conference increased consumer interest in vitamin fortification and provided impetus to the fortification programs of each respondent, which resulted in whole-line fortification in the early 1970's (Tr. 13 140, 29 719- , 35 057--8, 35 810-11). 515. At the time of the White House Conference, congressional hearings were being held, but there was no resolution of the matter at that time (Tr. 37 058).

516. Subsequently, in July 1970, Mr. Robert B. Choate, a civil engineer, in testimony before a congressional committee, criticized the lack of nutrients in RTE cereals. Mr. Choate s testimony further increased industry and consumer interest in vitamin fortification. His testimony was widely publicized, and the majority of RTE cereal consumers were aware of it (Tr. 37 061-62; CX-GF 3000Z-105). 517. Choate rated RTE cereals by name as to their nutritional value. As a result, certain fortified brands benefited from Choate highly publicized testimony (Tr. 29 951-52). Sales of four fortified cereals, Fortified Oak Flakes, Total, Special K and Product 19 improved after Choate testified (CX--F 477C, 1429B; KX 4; CX- 340A). Purchase of certain nonfortified cereal brands, Wheaties Cheerios, Rice Krispies, and Grape Nuts, appeared to decline (CX- GF 1429B; see CX-GM 16A, 17 A). 518. Kellogg had commenced development of the systems necessary for extending fortification at the one-third MDR level to all of its cereals in the late 1960' , at which time Kellogg had decided to fortify all products (Tr. 29 278-1). Kellogg, at the time of Choate testimony, had already begun to install the equipment required to 156 EDERAL TRADE COMMISSION DECISIONS Initial Decision 99 F. apply increased levels of eight vitamins called for in its expanded fortification effort. This enabled it to begin production of cereals fortified at the new, higher vitamin levels within several months of Choate s testimony (Tr. 29 952-53; see CX- 7187R). Kellogg decision to fortify its entire line was given additional impetus by the 1969 White House Conference on nutrition (Tr. 29 719-20). The decision had already been made to fully fortify prior to Choate testimony (Tr. 29 719-20).

519. Kellogg s Sugar Pops, Froot Loops, Apple Jacks, Sugar Frosted Flakes, Cocoa Krispies and Puffa Puffa Rice were fortified by late 1970 or early 1971 (CX-K 7187R). Two of Kellogg s largest selling brands, Corn Flakes and Rice Krispies, were not fortified until 1972; Corn Flakes was not fortified until September of that year (CX-K 7192F, 7193G, 7209E).

520. General Mills decided to fortify all its cereals to the X MDR level of seven vitamins and iron on September 17, 1970 (Tr. 35 814; GMX 373). It fortified the majority of its cereals from August 1971 to January 1971 (GMX 174 thru 182). Fortification of Cheerios and Wheaties was in August 1971, and General Mills' established presweetened cereals were fortified from October 28, 1971 , to January 1972 (GMX 174, 175, 177, 179). (151) 521. General Foods ' decision to fortify its entire line of cereals preceded Choate s testimony (Tr. 37 052, 37 062). By February 1970 General Foods was prepared to fortify its entire line of cereals (at X MDR vitamin and 100% MDR iron) subject to "Business Manager approval" (CX-GF 2022E). Choate served to accelerate implementation of General Foods' decision (Tr. 37 062). By late 1970 or early 1971, General Foods had completed its plan for such fortification and set a schedule to fortify its brands through 1971 and 1972 (CX- 477). General Foods planned to fortify its cereals in stages-first the presweets by September 1971 , and then the remaining cereals by March 1972 (CX-GF 477C). The project was actually completed in 1971 (Tr. 37 059).

522. General Foods, through advertising and other promotional activity, attempted to secure a competitive advantage for its fortified products (CX-GF 340B 477C). Kellogg recognized that it had been disadvantaged by General Foods having fortified its Raisin Bran before Kellogg did so (CX-K 7198D, E).

523. Kellogg introduced its newly fortified line of cereals in 1971 with an aggressive advertising and promotional campaign (CX- 765R, S, U, W, Y, Z-2). However, since Kellogg had lagged behind its competition by not fortifying its most popular brand, Corn Flakes .

"'VJ""'''''''- Initial Decision until September 1972, it IOHt sales to General Mills' Cheerios and Wheaties and to General Foods' Post Toasties (CX-K 7192F, 7209E). 524. The foregoing recitation of respondents' activities in the area of product fortification reveals that they were fully consistent with individual, competitive responses to stated public policy and consumer interest and demand.

Complaint counsel, however, assert (CPF &-227 thru &-233; CRPF 264 thru &-269) that respondents convened meetings of the Executive Committee of the Cereal Institute on July 27, 1970, and August 21, 1970, in order to agree on how to respond to the attacks of Choate and others on the cereal industry for its failure to provide nutritious food, and that respondents there reached agreement on how to fortify their RTE cereals. This agreement is evidenced according to complaint counsel, by the contemporaneous actions of respondents to fortify to the X MDR level. 525. The July 27, 1970, meeting of the Executive Committee of the Cereal Institute was called "to consider the impact of the recent testimony of Robert Choate in which he attacked the nutritional value of cereals" and to decide upon "what action if any should be taken by the Institute on behalf of the industry to introduce proof of the nutritional value of cereals before the Subcommittee and otherwise to repair the damage done by the unjustified statements and charges of Mr. Choate" (CX-CI 78B). The only action taken at the meeting, that was evidenced, was that of authorizing the Institute to arrange for a (152)leading nutritionist to testify before the Senate subcommittee "as to the nutritional value of breakfast 78B).cereals and their place in the American diet" (CX-CI 526. On August 21, a special meeting of the Board of Directors of the Cereal Institute was held to discuss the testimony by Institute and industry witnesses which had been presented to the Senate subcommittee on August 4, and to discuss further efforts to educate the public on the role RTE cereals played in a nutritional diet (CX- CI 80). It was noted in the minutes of the meeting that Senator Moss of the Senate Subcommittee had suggested that the cereal industry should eliminate differences in t.he nutritional content of breakfast cereals and make greater progress in educating the public of nutritional facts concerning breakfast cereals; and that Mr. Paxton the Cereal Institute s legal counsel, had advised that "despite the senator s suggestion, the elimination of product differences might not be an appropriate matter for concerted action" (CX-CI 80A). 527. The record contains no evidence that respondents discussed any plans for action regarding fortification at. either of the two Cereal Institute meetings. Representatives of Kellogg and General .. , Initial Decision 99 F. Mils who were present at the meetings testified that there were no agreements regarding fortification (Tr. 29 720-21 , 35 463, 35 806 815). And the record contains no evidence to overcome this testimony. Nor is there any evidence that the respondents otherwise communicated regarding their fortification activity, or that any respondent had advance knowledge of the fortification plans of the others. General Mils became aware of competitors' fortified products only when they appeared on retail shelves (Tr. 35 814). 528. Not all cereals were fortified to the X MDR level (Tr. 29 663 095; CX- 765C). To the extent they were, this is not surprising inasmuch as breakfast is one of the three usual daily meals, and respondents were being responsive to recommendations of the White House Conference regarding fortification levels (GMX 501Z-101, Z- 121, Z-197- , Z-231, Z-232). General Mills was unaware of Kellogg s and General Foods' plans when it set fortification levels (Tr. 808, 35 814).

529. Complaint counsel do not challenge the right of respondents to belong to the Cereal Institute. The two meetings of the Institute relied upon by complaint counsel for their hypothesis of agreement have not been shown to have been conducted for other than legitimate purposes." There is no basis for an inference that (153) fortification activities, which were most reasonable in the light of ongoing events, were in response to an otherwise unproved agreement rather than the ongoing events.

4. Introduction of New Products Complaint counsel assert that there is a barrier to entry into the R'IE cereal industry and would place responsibility for the existence of the barrier upon respondents. The cornerstone of the barrier to entry theory, which theory will be considered in the next section, has been termed by complaint counsel Hbrand proliferation. " It is complaint counsel's position that respondent's avoidance of competi tion by other means led them to turn to brand proliferation the introduction of a large number of differentiated, highly advertised trademarked brands" (CPF 1- , 9- , 9- , 9-35; CRPF 9- , 9- , 9- 11).

530. Complaint counsel and their expert witnesses have conceded that respondent's brand proliferation is vigorously competitive, not predatory and not in itself unlawful (Tr. 22 607, 22 614- , 22 622 Th" Cereal In titute was founded in 1941 for the purpose of advancing public under8tandin of the nutrition offered by cereal proof\lcts(Tr. 11 863-72). Initial Decision 629, 22 865, 22 906, 23 264, 23 304, 23 678, 27 267; GFX 1166A- 1167).

531. Respondents engaged in unrestrained and uncoordinated competition in the introduction of new products. Such competition was intense (Tr. 21 922, 22 056, 22 605-08, 22 905- , 22 615), and there is no evidence of a conspiracy or intent to deter entry by means of new product introductions (Tr. 22,109, 26 693, 27 028, 28 284 518- , 30 538).

532. Respondents, therefore, may not be held responsible for the results of this legitimate method of competition unless it was the proximate result of their having otherwise limited their competitive efforts as charged. However, as I have already held, complaint counsel have failed to prove those charges. Further, even if respondents had conspired or otherwise unlawfully coordinated their other competitive efforts, new product introduction would stil have remained as a legitimate vehicle of competition. There is no causal relationship shown between the alleged avoidance of other kinds of competition and competition by brand introduction. Not only is there no showing of proximate cause, but, if respondents had conspired to fix prices or had engaged in price leadership-price followership in lieu of overt collusion, competition by introduction of a large number of differentiated products would have been avoided as the antithesis of such coordinated behavior (see, supra findings 192-97). As found above (Findings 229- , 238-39, 245-47, 250- , 323-26), respondents did not want to engage in price wars and the record does not evidence strong price competition among them. This is consistent with Professor Schum peter s theory that firms in oligopolistically-(153)structured industries would tend to pursue competitive strategies which could not easily be matched by their rivals; that price competition, for example, would give way, among other things to the development of new products, whereby a company could secure an extended competitive advantage (Tr. 38 276). As Dr. Scherer has written Industrial Market Structure and Economic Performance 342 (1st ed. 1970):

(AJny fool can match a price cut but an ingenious promotion campaign is hard to counteract.

This may well explain the emphasis by respondents on competition through new product introduction rather than price competition. But respondents may not be held accountable for any results flowing from their individual choice to pursue this lawful means of competition.

The following findings, therefore, are not necessary to my disposi- Initial Decision 99 FTC. tion of the issue of the introduction of new products, but are included for the use of a reviewing authority in the event it might take a different view.

533. Consumers desire variety for breakfast (Tr. 14 421 , 14 446 191 , 17 398-401 , 22 751 , 35 367, 35 400, 35 447, 36 372; CX- 103Z-64; GFX 1153Z-71). Such a desire is responsible in large measure for the differentiation of RTE cereals (Tr. 22 751). A firm in the RTE cereal industry must introduce new products in order to remain profitable and compete for market share (Tr. 38 520, 38 797- , 38 830; CX-CI 103Z-5).

534. Kellogg s policy has been to rely primarily on its proven brands, but to build on top of them with new products having good potential (CX-K 397D 549C 7358F). It believes that if it has a product with wide appeal, it must introduce it or someone else will; that it is better for Kellogg to continue to expand its products even if it is taking some business away from other Kellogg products than for a competitor to do so (Tr. 13 046- 954 683-84). 535. "Although Kellogg does not agree that all profitable opportunities in the ready-ta-cat cereal market have been exploited Kellogg is doing its best to continue to exploit those additional opportunities. . ." (KPF 5-155).

536. General Mills stressed new product introduction at increasing levels as a major competitive effort and sought to outdo competitors in this regard (Tr. 17 353-66; CX-GM 38A, 608F, 609M BLOW). It believed it to be imperative to continue to introduce competitive new products (CX-GM 263A). General Mills was concerned at the inroads on its own absolute sales volume and market share that new products of its competitors might make (CX-GM 3D E). Both General Foods and Kellogg recognized General Mils' policy of stressing new product introduction (CX-GF 4039Z-1; CX-K 553H). (155) 537. General Foods also recognized a competitive requirement to introduce new products. As early as the 1950-1952 three-year plan for the Post Cereal Divisions, we find (CX-GF 167Z-10): The need for new products which could augment our volume, help carry our overhead and at least potentially contribute to profits has long been recognized. Since the total cereal business is at best stationary, and since therefore our principal chance to increase our volume is to take business away from competitors, new products are vitally important. It is not easy to increase the share of business done by our older established products.

538. General Foods, from the early 1960' , introduced new products in an attempt to maintain total volume and to make up for declining sales of established General Foods brands. New products Initial Decision were also viewed as the key to growth (Tr. 14 140 371; CX-GF 4A 17D, 324A, 602K, 2044D, 4039Z-53).

539. General Foods, in 1967, decided to remain in the Rfe ccreal industry by placing a high priority on developing new products (CX- GF 4039D, G). In its proposed marketing plan for FY' s 1968-1970, we find (CX-GF 4039Z-59):

The underlying assumption for Post new product strategy is that over the next three to five years continued competitive new product pressures, compiled with a static market will force volume losses on brands currently being marketed. . Therefore, Post' s new product program will be designed over the next three years to provide new products to hold or slightly grow total Post volume. 540. Changes in American society in the 1950's and 1960' , with resultant changes in consumer demand, contributed in large mea. sure to the introduction of new RTE cereal products. The "baby boom" significantly affected the Rle cereal industry. Since RTE cereals are so convenient, the increase in the number of children offered a great opportunity for producers to develop new products to appear to them (Tr. 29 621- , 29 786-87; CX-GM 736A). Because of the increased pace of modern living and the increased number of women in the lahor force, there was an even greater demand for RTE cereals (156)which children could eat without parcntal assistance (Tr. 29 623- , 29 787). Other shifts in demand, including the call for nutritious and natural cereals, also impelled respondents to introduce new products (Tr. 26 256, 29 680- , 29 787). 541. The advent of television enabled respondents to visually impact consumers with the claimed benefits and attributes of new products. This ability to have a direct, nationwide impact on potential consumers facilitated the sale of new products and provided an incentive to respondents to develop new products (Tr. 100-01 , 29 624, 29 702-07, 29 780-81; CX-GM 736A; CX-GF 4U; CX-K 563D).

542. New products were introduced by each respondent in order to compete against other RTE cereal manufacturers. Kellogg, for example, introduced a new product, Puffa Puffa Rice, so that it would not be preempted by a Quaker product, Tin Tin, that was being successfully marketed in Canada ('' r. 12 380- , 12 891- 965--6, 22 131 , 23 081-82, 30 706-07; CX- 7163A). Kellogg introduced Product 19 to compete with General Mills' Total (Tr. 396-98, 12 839-40; CX-K 7353H), OK's to compete with General Mills' Cheerios, Froot Loops to compete with General Mills' Trix, and Initial Decision 99 F. Cocoa Krispies to compete with General Mills' Cocoa Puffs (Tr. 907- , 12 684- , 12 875- , 13 545- , 29 661--2; CX-K 502C). And Kellogg introduced its granola in response to entry into that segment by competitors (Tr. 13 087-88), as did General Mils (Tr. 802).

543. General Foods, during the early to mid 1960' , spent several million dollars to introduce a line of corn flakes with fruit (GFX 416D). It considered this to be a very exciting opportunity to secure a real competitive advantage, to the point of overtaking Kellogg (Tr. 380-81; GFX 1297). Even with respect to new products that were essentially variations of existing ones, General Foods sought to capture a small but profitable share of the total RTE cereal business (CX-GF 6T).

544. While respondents' new product introductions, to some extent, expanded the RTE cereal market by appealing to additional consumers and inducing consumers to eat more cereal products (CPF 286; Tr. 7573, 11 432, 13 010, 15 223- , 17 681- , 29 678-79 29,780, 35,410, 35 417 , 35 421-22; CX-K 560C; CX-GF 4039Z-53), they did have a competitive impact on respondents' other RTE cereal products, either by a reduction of sales or by adversely affecting sales growth (Tr. 7551, 7559- , 8824- 198-99 088-89 220-27 966, 15 222, 15 754-55; GMX 71 , 73, 97, CX- 397C; CX- 14551).

545. Kellogg did not believe that every new product had to be profitable as long as the whole line showed a profit (CX-K 5651, L). Kellogg introduced new products when necessary to stop competitors from making inroads into Kellogg s business (Tr. 11 316; CX- 686B). Kellogg believed that the introduction of two products at the same time reduced product trial (Tr. 12 965-70; CX-K 604A-C). Kellogg increased its advertising on Froot Loops while General Mils was (157)seeking to introduce Lucky Charms (Tr. 22 133-34). It also increased its advertising of Special K when General Foods introduced Fortified Oat Flakes (CX-GF 34).

546. General Mils introduced products with relatively low life potential in order to boost its overall line share of the market and counter a potential loss of customers to competitors' new products (CX-GM 3D, E). It introduced new brands as defensive moves to help prevent competitors' new brands from taking hold (CX-GM 2A, C 262B, 276 2171C, 2176F). It sought to keep its own products on the shelves, even though it recognized they had no future, until it could introduce still more of its own products to replace them (CX- 285). General Mils increased its advertising, use of coupons, and use Initial Decision of samples with Total during Kellogg s introduction of Product 19 (Tr. 17 620-29 075; CX-GM 21, 123A, B, 570N, Z-15, 720A). 547. General Foods introduces a new product only when it is believed to meet a perceived consumer demand and a reasonable profit can be anticipated (Tr. 36 372). It has never introduced a product that it did not think would succeed (Tr. 13 666, 36 435), although it did not expect its brand introductions in the 1960's to be as long lived as older brands (CX-GF 4039Z-59). It has introduced line extensions of particular cereals (e. flavor variations) to induce consumers to stay with its products rather than try RTE cereals of competitors (Tr. 17 481-82; CX-GF 1455C, 2029B). 548. Kellogg believed that the growth of sales in an area meant that there might be an opportunity for a new brand in that area. Its general practice, therefore, was to identify areas of opportunity for new brands by looking at the sales growth of particular brands in particular market areas (Tr. 12 832-41).

549. Kellogg has many sources for its new product ideas. These include brainstorming sessions involving Kellogg employees and members of its advertising agency (Tr. 29 977); outside consulting firms (Tr. 30 034); and observations of the marketplace (Tr. 16,534- , 16 871- , 29 795, 37 014- , 37 030; GFX 1299). For example Kellogg continually monitors the products of its competitors to determine whether an opportunity for an improved product exists (Tr. 12 269, 12 404, 29 801-02). Kellogg s General Sales Manager urged all Kellogg product marketing managers to watch for the product innovations of small manufacturers. He advised marketing managers to (158)"keep especially alert for successful locally marketed products that could be duplicated and mass-produced for national marketing-ispecially those compatible with our existing product lines" (CX- 676B). Once such an idea was obtained Kellogg, because of its technological capabilities, could improve on it and make the Kellogg product available nationwide (Tr. 12 184 609-11).

550. Almost every department at Kellogg has some responsibilty for the development of a new product, including the research and development group, the marketing research group, the process development, packaging development and quality control groups, the administrative group, the controllers and the purchasing people (Tr. 774, 37 032-33). The product development coordinator, whose position was established in 1958, is responsible for facilitating a .. Fat example; the growth of General Mills' Tota! led Kellogg to develop Product 19 (Tr. 29 795- 96); Froot Lops were introduced to take advantage of an opportunity to appeal to users of the first fruit-navored presweet Trix, another Geneml Mills product (Tr. 12 875-76). Initial Decision 99 F.TC. product' s orderly progression through its developmental stages into test market and general distribition (Tr. 29 773-75). 551. Kellogg made extensive use of marketing research tests to determine consumer reaction to various characteristics of products. This included blind paired comparison tests to determine what consumers deemed to be significant differences between products (Tr. 11 638- , 12 957-62, 29 846). This was to prevent the offering of a me-too product (a product without a significant difference), which is recognized to be a likely failure in the marketplace (Tr. 9184- 965 , 14 516-25, 14 584- , 14 966-67, 15 228, 15 847, 15 851- 952 453 627- 634 650 722 549; CX-K 396B). 552. Thus, the development of a new Kellogg product calls for expensive testing, including consumer panel tests, concept testing, blind paired comparison tests, in-home tests, central location tests test marketing, as well as taste testings which expert respondent employees are continuously engaged in (Tr. 15 981- 046, 29 800 843-46, 32 970-62, 37 039-44).

553. Although each new product may vary as to the steps in development and problems encountered, Product 19 affords some perspective of what can be involved in the development of a new product (Tr. 29 794). Kellogg observed the success of General Mills Total, as well as vitamin and mineral supplements like One- Day vitamin pills and Geritol. Kellogg analyzed and tested Total and decided that that product indicated an area of opportunity for Kellogg (Tr. 29 795-96).

554. Kellogg employees from all sections of the company then gathered to discuss the characteristics, technological needs, and competitive potential of the desired product. Upon establishment of a product objective, research personnel were put to work. They requisitioned the use of a pilot plant and laboratories for the manufacture of a product prototype. Once work on the product had begun, the group met constantly to review the prototype and its progressive transformations (Tr. 29 796-99). (159) 555. After the manufacture of what was believed to be a successful prototype, Kellogg employees conducted internal taste tests in which they compared t.he Kellogg product with Total. Generally, after a prototype has elicited positive responses from an internal taste panel, Kellogg turns it over t.o members of the process development department, the packaging department and the research department for further improvements. Meanwhile, Kellogg continues testing the product in order to insure consumer acceptance. If the results of these tests are encouraging, the company hires professional testing organizations to conduct panel tests among Initial Decision consumers. Because the results of the new product' s consumer tests were positive, Kellogg decided to go ahead with it, Product 19 (Tr. 800 835- 846-8 887-88).

556. Once Kellogg decided to market Product 19, its production required technology that had never been used under manufacturing conditions. In order to test this technology, Kellogg established a manufacturing pilot plant with actual-sized equipment, but on a smaller scale. This plant cost $550 000.00. Product 19 was test marketed in 1966 throughout the Pacific West Coast, in Florida and in Texas. Although the product lost money during test marketing, Kellogg had expected the loss and considered it an investment (Tr. 84&-52 246-7; CX-K 7368B).

557. After being test marketed for 18 months, Product 19 proved to be a success. Kellogg then established a full-scale production facility which cost approximately $4 milion. This investment did not return any profits unti Product 19's third year of production. It took seven years to recoup the product's early operating losses (Tr. 854- 858).

558. New product ideas at General Mills are generated from several sources, such as research and development, marketing research and interviews with consumers. New product ideas are first concept tested with consumers. A concept test is based on a pictorial representation and a written or oral description of the proposed new product. New product ideas which are well received by consumers are then guidance tested with consumers. Guidance testing is a sequence of tests in which groups of 50 to 75 consumers taste the product and respond to written questions concerning its attributes. The product is modified after testing in response to consumer input. A product that is rejected by consumers is not pursued. Products that survive guidance testing are then put through large scale evaluative testing. Here, the product is placed with a representative sample of 300 to 600 consumers to determine its level of satisfaction (Tr. 966-71 97&-79 983- 999).

559. General Foods periodically analyzed the RTE cereal market by looking at segments of people and brands to try to find an unoccupied space or gap in the market. Brands were arrayed in relation to each other based on how people perceived them, and how they were rated with respect to particular attributes. People indicated what brand they would substitute if they could not find a (160)certain product. New product ideas were suggested by the study reports (Tr. 14 192- , 14 407-16).

560. General Foods' market research department regularly conducted product testing, advertising research information gathering g., g., 166 EDERAL TRADE COMMISSION DECISIONS Initial Decision 99 F. on marketplace performance, test marketing research, attitude studies, packaging tests and other types of research (Tr. 14 344; see, GFX 584).

561. The market research department also collected information about the types of people who were buying different productsconsumer profiles." Studies to learn what was happening in particular markets were also conducted (Tr. 14 345-46). 562. General Foods' and competitors' new products and their performance were evaluated (Tr. 14 346; GFX 552). Market Research conducted product quality tests in which consumers were asked to compare General Foods brands with other cereal products (Tr. 472-74; GFX 547, 549).

563. Information was secured from syndicated types of services such as A.C. Nielsen Company, SAM! and MRCA and from market research organizations (see, e. Tr. 14 349; GFX 289, 537, 547; CX- GF 1348).

564. General Foods' market research department thus was able to obtain information concerning the incidence of cereal purchases the demography of the consumer groupings and the frequency of purchases (Tr. 14 365).

565. An effort was then made to come up with products that answered people s wants so identified and measured (Tr. 14 370). 566. Market research would track a new product's performance in terms of volume, share of market and consumer feedback. Such tracking would aid in the development of new product ideas (Tr. 368).

567. Consumer testing continued long after the introduction of a brand. General Foods tried to determine how its products were performing in the marketplace and to evaluate new opportunities (see, e. GFX 535, 552).

568. New product ideas came from many sources-personal experience, marketing sources, marketplace sources, but primarily from technical sources (Tr. 37 015). Some ideas were generated in brainstorming sessions (see, e. GFX 1299). 569. Following a brainstorming session, representatives from technical research, marketing research and marketing would meet to select those ideas which appeared to have merit (Tr. 37 032). (161) 570. These ideas would then be presented to potential consumers in an "omnibus test " for group discussion and evaluation. This is also known as "concept testing" (Tr. 14 417 035 037). 571. After omnibus testing, highly regarded product concepts would be considered for feasibilty by technical personnel who would Initial Decision develop prototype products reflecting particular ideas (Tr. 37 037 223).

572. A prototype would be exposed to the consumer through market research. Research and development personnel would consider feedback and modify the product accordingly (Tr. 13 412). 573. Further consumer testing would be conducted, exposing prototype products to hundreds or even thousands of families for their rcactions (Tr. 37 040).

574. If the product still looked promising, it would be test marketed to indicate what might be expected on a national scale (Tr. 440, 37 043).

575. If the product performed well in test market-if it appeared that its potential volume was sufficient to constitute a viable business-it would be introduced nationally. Not all products that were test marketed went into national distribution (Tr. 37 043-44). 576. Performance of a product in national distribution usually lagged behind performance in the test market. If a product failed to meet the goals set for it, it was withdrawn (Tr. 37 044-45). 577. Each product concept was evaluated several times as it was developed to determine whether it justified additional investment. Many concepts would be discarded; others would move forward toward the marketplace (Tr. 36 375-76).

578. General Foods had a New Product Committee, composed of senior managers in the RTE cereal business, to evaluate ideas and to decide whether they justified further investment (Tr. 36 375). 579. Utilization of excess capacity was one of the considerations in evaluating new product development opportunities (Tr. 13 481; GFX 423C).

580. Several of General Foods' new products were offshoots of products made by the Jersey Cereal Company, the small RTE cereal producer that General Foods had acquired (CX-GF 167Z-11). 581. In 1950, 26 brands were in distribution beyond test market. During the next 23 years, the largest six firms introduced 84 brands beyond test market, of which respondents accounted for 60, and withdrew 30 of the 84 introduced. Thus, the number of RTE cereal 1973 (162)brands in the market increased from 26 in 1950 to 79 in (Tr. 22 024, 22 029-31; CX 405). Broken down into five year periods we find the following:

168 FEDE1tal TRADE COMMISSION DECISIONS Initial Decision 99 F. Brands Introduced Beyond Test Market 1950- 195640 1961- 1 966-70 1971- (CX 407) 582. Because of the influx of new brands, established brands lost market share. The top eight brands accounted for 47.3% of the market in 1964, a drop from 56.9% in 1954 (CX-PG 6B C). Corn flakes, which accounted for 33% of total RTE cereal sales in 1940 had only a 10.8% share in 1972 (CX-K 7054A 7148C 7209D). From 1950 through April 1970, average market share per RTE cereal brand declined from about 4% to about 1.3% , and the average pounds sold per RTE cereal brand declined from around 22 to 18 million pounds. This was despite a substantial growth in total sales (Tr. 22 030-31; CX 409A).

583. RTE cereal pound sales increased by 78% between 1955 and 1972. Products introduced prior to 1955 accounted for only slightly more than 10% of this increase, while products introduced after 1954 accounted for almost 90%. By 1971 , about 36.9% of RTE cereal sales consisted of products introduced in the prior 16 years (GMX 564). Respondents' products that existed beyond test market distribution prior to 1955 increased in aggregate pound sales by 107 125 000 from 1955 through 1970 (GMX 564E- , I). 584. The parties differed in the degree of product introduction activity during the period. Whereas Kellogg was introducing products throughout the 1950' , General Mills increased its introductions in the mid-1950' , and both General Foods and Quaker increased their introductions in the late 1950's or early 1960's. After General Foods introduced Krinkles and Corn-Fetti in 1950 and 1951, it introduced only one more new RTE cereal product before 1959 (GFX 1370H).

585. Between 1950 and April 1972, Kellogg introduced 24 brands into test market or beyond (CX-K 1067); General Mils introduced 34 (CX-GM 2049, 2111); and General Foods introduced 21 (CX-GF 556 18690).

586. Following General Foods' decision to emphasize its new product activities in FY 1962 (CX-GF 4039Z-17), it introduced nine (163)new RTE cereal products in the next five years (GFX 1370H- Quaker did not introduce a new RTE cereal product until 1961. Initial Decision Kellogg introduced products at a fairly constant level throughout the complaint period, with slightly more activity in 1959 (CX 409B, E). 587. The respondents' success with new products differed. During the complaint period, Kellogg introduced nationally 16 new RTE cereal products. All but three found widespread consumer acceptance and were still on the market at the close of that period (Tr. 600-01; CX-K 1067, 7173; CX 434). The other respondents and Quaker all had varied success, with General Foods being the least successful (Tr. 27 009, 27 434, 38 331-32; CX-GF 4039Z-4). General Foods marketed fewer brands of RTE cereal in 1973 (15) than it did in 1964 (17) (GFX 1370).

588. The firms have also had differing success with respect to their already established products. Kellogg s older products have had a good deal of durability. Their sales increased by about 25% since 1958. Sales of General Mills' older products were flat; General Foods sales of older products declined; and Quaker s sales of older products declined to a little more than half of their 1958 levels (Tr. 38 335-36; GMX 565).

589. Based on market share, the cereal industry became business of relatively small brands (CX-GF 17 A, D, 601Z-5, 602B; CX- 7342C; CX-GM 2178D). An analysis of the peak market shares achieved by RTE cereal brands introduced from 1950 through 1972 shows that only seven of the 84 introductions had market shares exceeding 2% (CX 434):

YEAR OF PEAK SHARE BRANDS INTRO. AND YEAR Kellogg s Sugar Frosted Flakes 1953 3% 1971 Kellogg s Special K 1956 1% 1970 & 1971 GF' s Alpha Bits 1957 7% 1959 GM' s Total 1961 6% 1967 Quaker s Cap n Crunch 1965 4% 1965 Quaker s Life 1961 0% 1966 through 69 GM' s Jets 1954 1% 1955 Of those seven, only Sugar Frosted Flakes, launched in 1953 achieved a 50/0 market share or better; only one acheived 40/0 better; and the remaining five achieved between 2% and 3% at their peaks. No brand introduced after 1956 and before 1972 achieved or more of the market. Respondents generally considered a new brand a success if it could sustain a market share in the 1 % to 1.5% range (Tr. 11 723- , 15 964 , 17 746; CX 434J; CX-GM 2A, 179E 276A, 591F, 603B, 700B, 2198B; CX-GF 20A-C, 30A, B, 470H , I, K 1372F, H, 1389E, 2018A, C; CX-K 742A, 995). (164) 590. However, as indicated above, not all new product attempts g., Initial Decision 99 F. were successful. For example, from 1968 through 1971, six new products were introduced in the presweet segment of the RTE cereal market. Of these, at least five were definite failures (CX-GF 40101 Z-39, Z-0).

591. As Quaker analyzed the 195B-1969 period (CX-Q 177H): Since 1958, 76 brands of RTE cereal have been introduced into either test market or national distribution. Ofthese, only 54% are stil on the market. Of the 36 RTE cereal brands on the market before 1958 34 are still on the market. The failure rate for new brands is even higher than indicated by these data since a considerable number of brands stil on the market do not have a large enough share to qualify them as real successes. By dividing the study period into halves it appears that the number of introductions and the proportion of failures increased in the more recent years.

592. From 10 failures of 25 introductions in the 195B-1963 period, the failure rate increased to 35 of 51 introductions in the 1964-1969 period (CX-Q 177M). Altogether, 50% of the RTE cereal brands introduced beyond test market between 1950 and 1972 failed (Tr. 26 406, 26 416-17, 26 75B-59; CX-GF 40101, Z- , Z-0; CX- 177H, M; CX-K 547B; ex 435).

593. Of 26 brands introduced prior to 1950 that were in the market in 1950, 23 were stil available at the end of 1972. Of nine brands introduced in the five year period 1950-1955, eight were stil available at the end of 1972. Comparable figures for subsequent five year periods are: 1956-1960, 10 of 21 still available; 1961-1965, 12 of 37 still available; 1966-1970, 15 of 42 still available (CX 435). These figures which show the number of brands that were withdrawn after introduction, do not reflect the numerous and costly efforts which did not culminate in products worth introducing, the cavity preventive RTE cereal effort of General Foods (supra, Finding 511). 594. General Foods, in particular, experienced difficulty during the 1960's in its introduction of new brands of cereal and was repeatedly unsuccessful (GFX 1370). It had costly failures in its attempts to develop and introduce cereals with freeze-dried fruits (Tr. 046-7).

595. The fixed costs necessary to launch a new individual brand include those of research and development of the brand, market research, production equipment and plant and introductory advertising (Tr. 21 964--5, 21 969-70). (165) 596. The fixed costs are significant. The major RTE cereal producers spend on average about $180 000 to research and develop Initial Decision and test consumer acceptance of each brand they launch (Tr. 21 969 408; CX I 103Z-7; GFX 1153Z-47; CX-NC 500 at p. 34). '0 The major RTE cereal producers also spend substantial amounts for new plant and production equipment for many of their new brands (Tr. 970). Kellogg spent some $4 million to $6 million for new equipment for Product 19 (Tr. 13 203). General Mills planned expenditures of $1.1 milion and $1.8 milion, respectively, on new equipment for a new health and a new puff cereal (CX-GM 607E). The Danville pla;'t of Quaker, used initially for King Vitamin, cost more than $6 million.

597. The most significant fixed cost of launching a new RTE cereal is introductory advertising (Tr. 21 969-70). Newly introduced brands require disproportionately heavy advertising to achieve market penetration (Tr. 12 809; see CX 508A). 598. Substantial introductory advertising expenditures are necessary in order to persuade enough consumers to try new brands (Tr. 466-9 , 14 435- , 15 038, 15 771 , 15 777- , 17 301-03; CX- 4010Z-2, 4039Z-56; CX-GM 2180C). Unless it intensively advertises a new brand, an RTE cereal producer cannot get many retailers to place the new brand on their shelves. Retailers wil not place a new RTE cereal brand on their shelves unless the manufacturer provides or promises to provide sufficient introductory advertising to "presell" the new brand to consumers (Tr. 8919, 9185, 9348-50, 12 722- 24). The high advertising level of the many RTE cereal brands already in the market (the high "noise level") requires high introductory advertising expenditures for new RTE cereals (Tr. 097, 14 437- , 15 038, 15 243-44; CX-GM 557H). Kellogg s Marketing Director has stated:

It is not unusual for a new product introduction to involve an initial outlay for advertising and promotion money that far exceeds the total dollar volume of sales for the first year s introduction. Unless the product succeeds in an unusual way-it may become virtually impossible to ever recover (166Jthe cost of introducing the new product (CX-K 552G) (emphasis in original), 599. By late 1961 or early 1962, General Foods concluded that introductory advertising expenditures of around $3.5 milion were necessary to launch a new brand expected to achieve a 1% market share (CX-GF 17 A, D). General Mills stated that the heavy introductory advertising it planned for Smiles ($4.1 millon) was essential for success (CX-GM 603E, H, W). The respondents and Quaker spend 'u General FOos spent between $150 00 and $300 00 on resarch and development for Alpha Bits during its firstthreyearstTr. 436 635-37J.

., Advertising during the first 12 months a product is placed in national distribution is considered introductory advertising.

Initial Decision 99 F. substantial sums on introductory advertising for each brand they introduce. Between 1956 and 1972, they spent on average some $3. million per new brand (Tr. 26 399-404; CX 206, 508; CX-Q 177Z-20). 600. Not only are the fixed costs of developing and introducing a new product very high, but there is an extended leadtime in going from the drawing board to national distribution and in reaching a break-even point. Dr. Schmalensee, complaint counsel's economic expert, agreed that it takes approximately four years to get a new product off the drawing board and into national distribution and another three years to achieve a break-even point on the product (Tr. 413-14).

601. General Mills' development time for Mr. Wonderful's Surprize and Golden Grahams, products that required new technology, was in excess of 10 years (Tr. 32 990). For products developed primarily on existing technology, General Mils' development time averaged from five to six years (Tr. 17 237, 32 989). General Mills began experimental work on Buc Wheats in 1964, but did not introduce the product Buc Wheats until 1971 (Tr. 32 914, 32 989). 602. The foregoing demonstrates that respondents engaged very heavily in new product competition; that this method of competition was expensive and risky and that there was a long period of time before even a successful new product venture would payoff. It has also been demonstrated that, as a result of respondents' new product competition, individual cereal brands on average accounted for smaller shares of the market and lower poundage of sales. IMPEDIMENTS TO NEW ENTRY INTO THE RTE CEREAL INDUSTRY A. Brand Proliferation Complaint Counsel's Theory Complaint counsel (CPF 9-1 thru 9-338; CRPF 9-1 thru 9-112) would place responsibility upon respondents for the lack of new entry into (167)the RTE cereal industry. The following is a summary of complaint counsel's theory of the existence of a barrier to entry and respondents' responsibilty therefor.

The RTE cereal industry has enjoyed supracompetitive profits (an issue which wil be dealt with later in this initial decision) and was growing rapidly. This should have attracted new entry into the industry. The absence of such entry indicates the existence of a barrier to entry. Respondents' brand proliferation conduct raised effective barriers to the entry of new RTE cereal producers and provides the complete answer to this absence of entry. While brand Initial Decision proliferation is not in itself unlawful, respondents must be held responsible for its deterrent effects upon entry, since respondents turned to this method of competition as a result of their mutual avoidance of other means of competition.

As I have already found (supra Findings 529-31), brand proliferation in this case is nothing more than the introduction of new brands as a legitimate means of competition and did not result from any other activities of respondents. Thus, respondents cannot be held legally responsible for the impact upon potential new entrants of their introduction of new products. However, for purposes of providing complete findings, I shall continue with complaint counsel's theory of respondents ' brand proliferation as the sole barrier to entry into the RTE cereal industry.

The proliferation is said to be of highly-differentiated, intensively advertised, trademarked, new RTE cereal brands. The brands actually differed to varying degrees or, by reason of advertising, the consuming public was led to believe they differed. In this manner competition among RTE cereal brands was "localized" the first of three basic conditions under which brand proliferation will cause a barrier to entry.

When competition is not localized, a change in price, advertising or promotion of one brand would equally affect all other brands, and industry-wide reaction could be anticipated. The introduction of a new brand would similarly affect all other brands in the industry. A potential entrant would look to the market as a whole for its source of sales and would anticipate being able to take comparable percentages of sales from all existing brands. Any reaction to the new entry by existing competitors would impact all brands in the market.

On the other hand, when competition is localized, a change in price, advertising or promotion of one brand would affect primarily the other products in the segment in which it is located, and reaction would be anticipated only from the other brands in the segment. The introduction of a new brand would affect significantly only the brands in the segment into which the new brand had been introduced. The potential entrant could look only to the segment (168)into which it was considering entry for its source of sales. Existing competitors would react only in the segment involved without impacting other brands in the industry. In sum, the new entrant would have a more limited area from which it might secure sales and it would anticipate more direct competitive reaction from the existing brands in the segment it had entered. The second basic condition, under which complaint counsel assert Initial Decision 99 F. that respondents' brand proliferation has created a barrier to entry, is termed "brand immobility." This means that the attributes of existing brands are so firmly established in the minds of consumers that the brands cannot be repositioned to appeal to those who want different attributes. Consequently, when a new brand is entered into a segment, existing brands cannot avoid competition by changing their appeal but must meet the competition of the new brand head on. A potential entrant, therefore, would anticipate this head-on competition from existing brands in the segment considered for entry and so might be deterred from entry. The third basic condition of complaint counsel's assertion that respondents' brand proliferation has created a barrier to entry is the existence of substantial fixed costs associated with the development and introduction of a new product. These are production, marketing and distribution costs which do not vary directly in proportion with the amount of the item produced.

Respondents are said to have so crowded the market with their products that the introduction of additional, profitable brands by new entrants has been foreclosed.

As the demand for cereals grew, opportunities for new brands did occur. Except for the granola or natural segment of the market these opportunities or "holes" in the market were filed by the six existing major producers. This is because the existing majors were already operating at an overall efficient scale, so that it was profitable to add an additional product even though the volume of that product alone would not cost justify an outsider entering the RTE cereal industry. An outsider would need several simultaneous successful product entries of the size that can be anticipated in this industry before it could operate at minimum efficient scale and so not be at a cost disadvantage to respondents. This would be most unlikely since it is difficult to develop successful brands; and by the time an outsider could be ready with several brands, an existing firm could fill any particular hole or opportunity that might exist.

In sum, complaint counsel assert that there is a product efficiency of scale equal to about 1 % of the market, whereas firm effciency of scale is not reached until sales of 3.5% to 5% of the market are achieved. Respondents, all of whom are already operating (169)at or above firm efficiency of scale, can introduce a new product whenever a 1 % opportunity appears, whereas an outsider cannot. The outsider is faced with the insurmountable task of finding simultaneously three or more opportunities in order to enter the market at Initial Decision minimum firm economy of scale so as not to be at a cost disadvantage to respondents.

Complaint counsel assert that an outsider would have to find larger individual holes or opportunities than respondents since respondents do not react to each other in a competitive fashion whereas they would react very strongly to an outsider s entrance in a particular, limited segment.

The requirement to achieve entry by multiple brands is also said to have deterred entry since the entrant would have to expend large capital costs on each of the several brands to be introduced. Complaint counsel assert (CP 9-29):

The conclusion that respondents' brand proliferation practices have increased a potential entrant' s capital costs is based on the previous analysis demonstrating that a potential entrant can no longer expect to attain entry at efficient operating scale with a single brand. The necessity of a multi-brand strategy means that the entrant must fund several research and development efforts, several introductory advertising programs, and possibly several production lines. The record shows that it is much more expensive to develop and introduce five brands whose total sales will fall in the range of from 3.5% to 5% of the market, than it is to develop and introduce a single brand to achieve the same market share, The additional costs of entry via a multibrand strategy are so great that many smaller firms simply cannot raise the necessary capital. The magnitude of the required capital does not prevent entry by the largest firms, but it does result in their acting more cautiously. If the cost of entry were significantly lower, the attractiveness of the industry would result in more frequent gambles by many firms. But as the costs grow, fewer gambles are taken and firms are likely to exercise such great caution that by the time they have decided to enter, their original brand ideas may have been preempted by one or more brand introductions by respondent-c;. (170) Localization 603. Localization or segmentation of brands is the first of the three necessary conditions under which, according to complaint counsel, brand proliferation wil cause a barrier to entry. As I have previously found (Findings 59-150), there are segments or categories of cereals within which brands compete more strongly with each other than with other brands because of their similar attributes, and there are some cereals that are so similar that they compete with each other on a one-to-one basis. The record, however, does not permit an exact delineation of the segments and the degrees to 52 Consumers' desires for variety in breakfast are responsible in large measure for the differentiatiol1 of RTE cereals Isupm, Fil1ding 532). Respondents, for example, may be said to have reacte to children s desires for presweetened and l1avored cereals, tv many consumers' desires for natural cereals and tv many other consumers desires for fortiried cereals. Respondenu;, therefore, may not be held responsible for locali7.ation of this nature. Loalization, to the extent it has come about by what complaint counsel term "brand proliferation" is the result of competition by means of introducing new producu; to appeal to the varying desires of differef1t categories of Consumers Initial Decision 99 FTC. which competition among cereals is confined to those cereals in particular segments. The necessary corollary to this description of localization, as shown to exist in the RTE cereal industry, is that some cereals have a broader appeal than the particular segment or category in which they fall so that they compete to varying degrees with cereals outside of their category.

604. For example, adult products, such as Total, Special K Product 19, All Bran, 100% Bran and 40% Bran Flakes, not only compete with each other, but also compete strongly with all-family products, such as Cheerios, Rice Krispies, and Corn Flakes. And children s products, which arc essentially the presweets, not only compete with each other but compete strongly with the all-family products (Tr. 35 367-70; GMX 194, 195, 546A, 547 A- , 548A- 549A). While some presweets compete more directly against other presweets, all 30 or so presweets compete to some degree with each other (CPF 9-172; CX-GF 4K, 1410A, 1439, 30002-95; 'lr. 22 778-81). Grape Nuts competes with all RTE cereals for its share (CX-GF 40), while Post Raisin Bran competes strongly with General Foods presweetened cereals (GFX 121OZ-5).

605. Twenty-one other brands each accounted for over 1 % of the RTE cereal consumption of the two-member families that purchased (17J)Cheerios in 1969-1970. For 1975-1976, 25 other brands each accounted for more than 1% of their consumption (GMX 518A-B). Two-member families consuming Kellogg s Corn Flakes had more than 1 % of their consumption accounted for by each of 20 other brands in 1969-1970 and by each of 23 other brands in 1975-1976 (GMX 520A-B).

606. Many consumers of RTE cereals are interested in many product attributes such as puffing, flaking, specific grains, shapes textures, flavors and degrees of sweetness. The number of directly competing brands is determined by the number of attributes relevant to each consumer s purchasing decisions. The number of direct brand competitors increases as the number of attributes relevant to particular consumers increases (CPF 9-185, 9-186, 9-188; Tr. 21 953- , 22 078- , 22 771- , 26 253-54). 607. It is uncontested that RTE cereal products compete on the basis of varying numbers of in-common characteristics that differ in importance, but that no analysis was made to quantify the numbers of direct competitors as a result of these in.common characteristics (Tr. 22 753- , 22 777 , 22 794, 23 892-93). Inasmuch as localization is a necessary predicate to complaint counsel's barrier to entry by proliferation theory, the question arises whether 10caJization exists to the degree required. As an abstract Initial Decision theory, it is logical to assume that a new product' s success would be limited by the size of the area or segment within which it is to compete. However, to the extent that it may draw customers from outside of its primary segment, its success potential becomes less limited.

608. While the extent to which localizaton prevails in the RTE cereal industry is an unknown, it exists to a sufficient degree to have the type of impact theorized by complaint counsel." Respondents believed that the effect of increasing the number of brands on the market was to increase segmentation (CX-GF 17 A, D), fractionalization (CX-K 7342C; CX-GM 2178D), or fragmentation (CX-GM 601Z- 5). "The cereal industry has become a business of small brands. Kellogg believed that "a lot of products on the market in 1970 wil account for less than one percent as a result of the continued introduction of new brands. This would result in fewer brands with large market shares (i. 5%) (emphasis in original) (CX-K 565J). (172) Brand Immobility 609. Brand immobility is the second of the three conditions under which complaint counsel contend that respondents' brand proliferation has caused a barrier to entry. While I would not expect that a manufacturer of an existing product would, even if it could 54 the conditionreposition a product to accommodate a new entrant, of brand immobility has been established on the record. 610. It is difficult to reposition a product from an idea or concept that people have gained. It may be hard to get people to accept the product' s new position if they have associated the product over time in its former position. There is a good deal of risk and expense involved in trying to change people s established concept of a product (Tr. 14 212- , 14,466-68). General Mills has learned that when you try to make radical changes, you sometimes lose the entire market. It doesn t make sense to reposition an RTE cereal product after it has gone national (Tr. 15 762-63, 17 707-11; CX-GM 2476F-H). 611. Attempts at repositioning have failed. Kix was originally an all-family cereal. It had always had some consumption by children. As Kix s sales declined, General Mills attempted to move the brand more toward children. General Mills changed the package, used '" Since localization is an elf'ment in evaluating crowding, the commellts at the end of my findings on crowdingUn(m, Finding621)applyequallyhere. .. With the large number of product. inthc market catering to fill nature of consumer want., there would be no point to t.he mflnufadurer of a successful existing product to pick up and run from a new entrant and attempt. to compete in a different area. There is 00 evidence that. an existing brand hah ever bceo rcposilionedt.oaccommodate a new brand Initial Decision 99 F. Kix s first in.pack premium and urged retailers to place Kix in a presweet shelf position. These efforts at repositioning were not successful. The result was a lowering of Kix sales. Former adult consumers ceased buying it, while it failed to win children s support (Tr. 17 316-18, 17 323, 17 662--5; CX-GM 124B, 2173). General Mils attempted to reposition Trix by changing its shape, texture sweetness, package, and name. The attempt destroyed the identity of the product and Trix sales decreased significantly (Tr. 17 708-11; CX-GM 2476F-H). An attempt by Kellogg to reposition Cocoa Krispies towards younger consumers was unsuccessful (Tr. 12 887- 88).

612. When introduced, Clackers' advertising and its premium offer positioned it as a child cereal. Because of its low repeat sales General Mills considered repositioning the product. However, (173) General Mils believed it would be very diffcult to persuade adults to purchase the product since the original image had been established so strongly. Instead, the product was withdrawn from the market (Tr. 16 039-41, 15 758-0; CX-GM 2049A).

Fixed Costs In Development And Introduction Of A New Product 613. The third prerequisite, under complaint counsel's premise that respondents' brand proliferation has created a barrier to entry into the RTE cereal industry, is the existence of substantial fixed costs associated with the development and introduction of a new product. This requirement has been met. As I have previously found (Findings 595-99), there are significant fixed costs necessary to launch a new product including the costs of research and development, market research, production equipment and plant and introductoryadvertising.

Crowding of the Market 614. A market would be crowded when existing firms have packed it so densely with their brands that another brand cannot enter and generate enough sales to earn a normal return. Sales opportunities available to an entrant must be sufficient to cover the costs of introducing new brands and yield at least a normal rate of return. Fixed costs establish a certain minimum volume required to break even. Crowding deters entry because an entrant is prevented from attaining the break-even point in sales output (Tr. 21 987-88). Complaint counsel assert (CPF 9-221) that respondents' brand proliferation activities have crowded the market so as to reduce the ...

Initial Decision opportunity for the introduction of new brands; that "(i)n some areas (or segments) of the market crowding has proceeded to the extent that the remaining sales opportunities, or holes, are so small that no one, including established firms and potential entrants, could launch a brand and expect to achieve sales sufficient to become profitable. 615. As found above (Findings 530-31, 534-89), each respondent has made a planned and concerted effort to find and seize upon every opportunity for the introduction of a profiable new product; and Kellogg and General Mils have found it to be to their respective competitive advantages to introduce new products even when they may not be individually profitable. New product introduction has outdistanced market growth, as demonstrated by the reduced market share and poundage sales of the average RTE cereal product. (174) 616. In the face of high fixed costs, opportunities for profitable new brand introductions declined as the number of brands on the market increased (Tr. 12 424-25; CX-K 552G, 565I- , 604A, B; CX- GM 2178D, E, G; CX-GF 1455C). The increased number of cereals on the market correspondingly increased the possibility of new brand failures (Tr. 11 465 672-74). Kellogg believed that the sodium free corn flake segment which consisted of another seller s brand, was too small to support a profitable entry by Kellogg (Tr. 12,840), and that the market is over saturated with cocoa-flavored products" (CX- 7205D).

617. General Mills concluded that the market is so "highly fractionated (that) opportunities to successfully introduce unique new cereal products. . . are limited" (CX-GM 2178D); and that (t)he successful introduction of new cereal products has become increasingly difficult in a market which has reached maturity" (CX- GM 2178G). General Mills conducted a market research test which showed that the bran segment of the RTE cereal market would not support another bran product, Alive/Bran Wisps. General Mills decided not to introduce that product into the national market (Tr. 798-04; CX-GM 2049).

618. General Foods observed that "(t)he presweet segment is crowded with over 30 brands competing for 21 share points, for an average SOM (share of market) ofless than 0.7%" (CX-GF 2021C). In 1972, it found that "subsegments of the presweet category are becoming saturated" (CX-GF 4010Z-10). For example, the unflavored, presweetened rice area would not support a new General Foods cereal, an unflavored Pebbles (Tr. 17 456-58). 619. Quaker concluded that a major inhibiting growth of the presweet cereal market was consumer "satiety" with uniquely Initial Decision 99 r' T.C. flavored products. So many differently flavored cereals were on the market that opportunities for new brands were limited (CX-Q 2631B). It also believed it was unlikely that an opportunity remained in the spoon-sized shredded wheat segment for an additional brand capable of reaching the 1.24% market share that Quaker estimated was necessary for profitability (CX-Q 183A-B). 620. Complaint counsel's theory that the introduction of al1 of the products that can profitably compete in an area wil dissuade the entry of stil more products is a logical one, and the situation clearly has obtained in the RTE cereal industry. This, however, is just another way of saying that competition will deter entry and the more vigorous the competition, the more likely it is that new entry wil be deterred. Here, complaint counsel have merely evidenced and analyzed how competition by existing firms in the form of new products will deter entry by new firms.

621. While crowding has undoubtedly deterred new entry into some areas, the areas so precluded, their economic significance, and the (175)time periods of preclusion have not been identified (CPF 9- 221, 9-270, CRPF 9- , 9- , 9-49, 9-52). This limitation of proof weakens complaint counsel's position that respondents' brand proliferation constitutes the sale reason why there has been a failure of new entry into the RTE cereal industry.

Preemption Faced with the fact that established firms have introduced numerous brands during the period covered by the complaint complaint counsel agree that crowding is not the complete explanation of why new firms did not enter. It is necessary to go one further step. That step is preemption (CPF 9-289 thru 9-291). As explained by complaint counsel (CPF 9-291):

It is obvious that in a changing market. entry deterrence does not result solely from the crowding effect of brand introductions. Crowding cannot be the complete explanation of entry deterrence because crowding will affect existing firms as well as new firms. However, existing RTE cereal firms and potential entrants differ in ways that account for the ability of existing firms to launch new brands while new firms cannot. One difference is that respondents' brand proliferation strategies have resulted in a more crowded market with reduced average brand market shares, thus making it more difficult for new firms to achieve efficient firm size, in the range of 5% to 5% of the market, without entering with several successful brands. Potential entrants, therefore, are more likely to adopt an entry strategy based on multiple brands. Respondents, having already achieved efficient firm size, do not need to locate multiple brand opportunities. They can launch single brands as the opportunities develop. A second difference is that new entrants must have different expectations as to the reactions their brand introductions would provoke from the established firms. Initial Decision That difference would lead prospective entrants to be more uncertain and more cautious, causing their brand development ideas to be preempted by respondents brand introductions (Schmalensee, Tr. 22 647). These differences lead to preemption they operated to deter potential entrants while respondents are able to introduce new products CSchmalensee, Tr. 22 113-18). (176) (a) Preemption By Reason Of Economies Of Scale 622. Complaint counsel assert (CPF 9-292) that economies of scale are not reached at the firm level until the firm acquires from 5% to 5% of the R'IE cereal market. A potential entrant would not wish to enter at less than minimum efficient size because it would then have higher costs and lower profits than the respondents and would be at a considerable competitive disadvantage (Tr. 22 115 409-11). (This finding is subject to the comments contained in the paragraph following Finding 624.

623. However, the opportunities for developing a product that can achieve sales of 3.5% to 5% of the market are practically nonexistent; and respondents, who are already operating at or above minimum efficient scale, are constantly searching for opportunities to introduce products at the 1%-1.5% level, a profitable level for them (supra Finding 589). Thus, existing firms are in a position to preempt the entry of potential entrants by introducing individual products into perceptable market holes before potential entrants can develop and introduce the multiple products necessary to reach the 5% to 5% minimum efficient firm size. Even if a potential entrant were to consider attempting to develop a product that might achieve 5% to 5% of the market, respondents, by repeatedly introducing smaller volume brands, would remove the possibility that an opportunity for a 3.5% to 5% brand might come about (Tr. 22 116- , 22 930-31 , 26,409-10). (This finding is subject to the comments contained in the paragraph following Finding 624. 624. In summary, respondents can seize every individual new product opening as it appears, whereas an outsider would have to wait for an unusually large size or for a number of smaller normal size openings; and the respondents, by seizing each individual opportunity as it appears, preempt outsiders from ever having the opportunities they require to enter. (This finding is subject to the comments contained in the following paragraph. Thus, potential entrants would be disadvantaged in their efforts to introduce a new product into the RTE cereal industry. The extent of that disadvantage, however, depends upon the minimum efficient firm scale facing a potential entrant. Findings 622, 623 and 624 therefore, are conditioned upon the establishment of a 3.5%- 182 FEDERAL TRADE COMMISj;ION m;CISIONS Initial Decision 99 F. share of market as the minimum efficient firm SIze for a viable competitor in the RTE cereal market.

625. Economies of scale are those effciencies related to volume of output that enable larger firms to produce and distribute their goods at lower average costs than smaller firms. As firm size (177) increases, average costs per unit of production decline until a firm reaches an output level called "minimum efficient scale." Unit costs beyond that level generally stay constant. Minimum efficient scale is that level of output at which all relevant economies of scale are achieved and at which unit costs attain their minimum value. A firm with this level of output would have costs as low as any firm in the industry. A firm with an output level below minimum efficient scale would find itself operating at a cost disadvantage relative to firms with higher levels of output. If a company were to enter an industry with a le el of output below minimum efficient scale, it would have costs greater than existing firms in the industry and, therefore might not be able to earn a competitive rate of return (Tr. 21 846-7 991 319; CX-CI 103Z-38, Z-39).

626. Complaint counsel assert (CPF 7-126) that "A study entitled The Structure, Conduct and Performance of the Breakfast Cereal Industry 1954-1964,' written by Robert S. Headen and James W. McKie in 1966 ('Headen- McKie Study, CX-CI 103), . . . provides a reliable basis for determining firm economies of scale in the RTE cereal industry." The Headen-McKie Study was commissioned by the members of the RTE cereal industry through the Cereal Institute, as an industry response to an anticipated report of the National Commission on Food Marketing, and was ultimately submitted to that Commission (CX-CI 59, 60, 62, 63, 64, 75, 170R- , 172A- 173E-F, 177B-). The study examined the behavior, the character and nature of competition, and the performance of the breakfast cereal industry as it existed during the period 1954-1964 (CX-CI 103C).

627. The study, among other things, considered production scale economies as a condition of entry. In consideration of the rate of technological progress in the industry, no empirical estimates of average costs at different output rates were derived, nor were any statistical cost studies made. The only information secured was by interviewing cereal production managers (CX-CI 103Z-38-39). The study reports (CX-CI 103Z-0):

The estimates obtained in discussion revealed that: 1. It would be "economically feasible" fof a new entrant to enter the market with production facilities built to supply about 1 % of the national market or approximately Initial Decision 12 milion pounds of cereal. By "economically feasible" it is presumed that production managers felt that a plant built to this scale would not be as efficient as a larger one but that the cost disadvantage would not be so great that the potential entrant could not earn satisfactory profits. (178) 2. It was felt that a ruby effcient cereal plant could be built to a scale such that would supply about 5% of the national market or about 60 million pounds per year. It is presumed then that the production cost disadvantage of a plant of this size would be negligible.

And at CX-CI 103Z-5, it is stated:

It was estimated in the above section on possible production scale economics that an effcient cereal plant would supply between 1 % and 5% of the market. 628. Apart from production costs, the study considered media availability and costs, ability to get retailers to carry a new product cost differentials in having a company sales force as opposed to using brokers, patent barriers, raw material availability and advantages of vertical integration; and found that these did not constitute significant barriers to entry (CX-CI 103Z-29-Z-38, Z-0-Z-5). 629. The Headen-McKie conclusions on firm economies of scale have no analytical or other substantive support. As conclusions drawn from conversations with production managers, and limited to production costs, they are at best a rough estimate that firm economies of scale are not fully realized until a 50/0 market share is reached. However, the report does not indicate the difference in production efficiencies at the 1 % and 5% levels or for points between. Accepting the report at full value, it may be that, while firm economies of scale are fully realized at the 5% level, there is not a significant difference in firm efficiencies at the 1 and 5% levels. 630. Apart from the Headen-McKie report, the record contains no study or other direct evidence bearing upon economies of scale in the (179)RTE cereal industry. Complaint counsel attempt to support and further refine the conclusions reached by Headen and McKie by drawing inferences from other matters in the record. 631. First, complaint counsel (CPF 7-129 thru 7-133, 7-141) rely on the contention that RTE cereal manufacturers (Kellogg, General Mills and General Foods) with market shares above 5% were not more profitable than manufacturers (Quaker and Ralston) with market shares approximating 5%56 and that there was no signifi- .. These considerations appear to have been made on a cursory basis (Tr. 26864-65) and do not take inw account other basic costs such as distribution and mluketing. I cannot accept the Headen-McKie study as establishingthattheonlysignific"nteconomiesofscaleareinprodudion. .. Complaint counsel's I1Q,umentis premise On rate of return figures which have ben found to bI' unreliable ( Sl?e, infra Findings 676-799), Contrary cnmp!aint counsel's contention, Quaker s rate of return was considerably lower than those of Kellogg and Genera! Mills(infra indings 792, 797), Further, costs per dollar ofsa!es would be a more appropriate consideration in evaluating economies of ale than returns on capital (Tr. 3U 618), Initial Decision 99 F.'l'. cant disparity between market share movement of the firms in the two groups. It is contended that, if the larger firms had a cost advantage over the smaller firms, t.hey would either have been more profitable or they would have expanded their sales and market shares at the expense of their smaller rivals. 632. There are many factors in addition to economies of scale that account for profits and market shares. These include differences in management decisions as well as differences in the products being manufactured. It may well be that the products of a particular manufacturer are selling at maximum consumer demand, that others are losing favor in the marketplace and that efforts to develop , under comnew products are meeting with varied success. Thus plaint counsel's own figures (CPF 7- 132), General Mils' return on capital was well above the industry average, Kellogg s was slightly above average and General Foods' was below average. And, among the firms operating near the 5% share of market figure, Quaker return was below average and Ralston s was above. These disparities using complaint counsel's own figures would show that there are many clements that impact profits other than share of markets. 633. There is simply no reliable way, on the basis of this record in which company profits and market shares can be related to economies of scale to the exclusion of numerous other market factors (Tr. 23 799-801).

634. Complaint counsel (CPF 7-141 thru 7-144) next rely on instances where firms which were operating at the 1 % to 3. market share level took steps to increase their market share to above the (180)3.5%-5% level. The contention is made that they did so in order to achieve economies of scalc.

635. Complaint counsel (CPF 7-143) rely upon the fact that when Quaker dropped into the 1 %- 5% range in the early 1940' , it expanded its operations and remained above 3.5% into the 1950' and that when its share again fell below 3.5% in the 1950' , it introduced new brands and undertook expenditures on plant equipment and product development with the result that it brought its share to over 7% by 1971. A similar recitation is made of competitive efforts on the part of Ralston on the several occasions its share fell below 3.5%. Actually, Ralston continued to operate at below the 3.5%-5% rate during the periods 1943-1963 and 1966- 1971 (Tr. 23 515).

636. All that complaint counsel have demonstrated is that when Quaker s and Ralston s business slipped, and profits correspondingly fell, they took steps to increase their business. None of the various Quaker and Ralston documents relied upon by complaint counsel Initial Decision refer to or discuss economies of scale. They discuss sales volumes and profits to be realized through greater sales. Quaker wanted to grow overall and accelerated new product development in most of the major businesses it was already in including RTE cereals, hot cereals, Aunt Jemima mixes, frozen foods, Quaker corn goods Kennel Ration dog food and Puss n Boots cat foods (Tr. 15 111-14). There is no evidence that this growth effort was stimulated by considerations of economies of scale in any particular area. 637. Complaint counsel (CPF 7-145 thru 7-157) next point out that, generally, individual plants of Kellogg, General Mils, General Foods, Quaker, Ralston, and Nabisco have capacities to supply over 4% of the market; that those which fell below that figure were either expanded or were closed. This, it is contended, supports the reliability of the Headen-McKie estimate and the conclusion that economies of scale in the RTE cereal industry are in the 3.5%range.

638. The expansion of a sma1J plant may -tell reflect increased demand for the particular products made there. The closing of a very small plant may be because it is no longer needed in the business. There may be all manner of reasons for these actions unrelated to economies of scale. Dr. Scherer, one of complaint counsel's expert economists, assigns very little weight to this technique (Tr. 26 934- (181)35). Dr. Glassman, through whom complaint counsel sought to develop the relationship between plant closings and expansions and economies of scale, conceded that motivations quite apart from minimum efficient scale could have been involved (Tr. 26 945- 983, 26 975-76). Plant closings and expansions, to the extent developed in this record, therefore, do not provide probative evidence on the question of firm efficiencies of scale. 639. Complaint counsel (CPF 7-159 thru 7-162) next contend that plant capacity utilization tends to be lower for smaller firms than for larger ones. This, according to complaint counsel, means that smaller companies build plants in excess of initial requirements in the hopes of expanding production to the point of reaching a 5% minimum efficient scale. There is no substance to this argument. Any competitor has hopes for a flourishing and growing business and would build a plant on the basis of projected future demand. Further, complaint counsel's assertion as to the relative utilization of capacity of larger firms vis-a-vis smaller firms is not supported by the records (Tr. 26,971- , 27 001-02). General Foods' capacity " Ally firm realizillg a normal or higher return Of! a part.icular volume or sales would consider increasing its sa!esir wit.houtunn'3sonablerisk it.couldn'ali7A,thesamerat. of return on the increilsed saj"s "" or course, to lhe extent a part.icular plant is operatinv,below capilcity, the firm bf,inRSubj ect to constant overhead cost.s,CQuid be operating helow minimum effcient scale. Initial Decision 99 F. utilization was relatively low and dropped as it lost market share (Tr. 26 364).

640. Finally, complaint counsel assert (CPF 7-163) that the evidence shows Procter and Gamble was considering entry into the RTE cereal industry, but that it would not enter unless it could acquire between 4% to 10% of the market; and that this confirms the reliability of the 5% Headen-McKie estimate. Mr. Butler executive vice-president and vice-chairman of the board of Procter and Gamble, contrary to complaint counsel's contention, did not testify as to the market share Procter and Gamble would require to enter the RTE cereal industry; and there is no probative, reliable evidence on this issue elsewhere in the record. 641. In summary, under Headen-McKie, which does provide a rough estimate of production scale economies, 59 it would be economically (182)feasible to enter the RTE cereal market with production facilities, capable of supplying about 1% of the national market. While the ntrant could earn "satisfactory" profits at that level of production, it would not achieve full production economies of scale unless it supplied about 5% of the market. The study, however does not indicate the degree of disadvantage that a firm would be under at various levels of production below 5% down to 1%.60 Even if a firm required 5% of the market to achieve minimum efficient scale, it would enter at a smaller volume if the cost disadvantage was not too great. It is impossible, therefore, to evaluate the extent to which preemption of new entrants by reason of their inability to achieve economies of scale has acted to deter entry. 642. As noted above, Headen-McKie estimated production economies of scale, and in a cursory manner stated that other economies of scale were insignificant. As marketing and distribution fixed costs would appear to be considerable, it appears that Headen-McKie underestimated economies of scale in areas other than production. However, there is no reason to believe that the volume required to reach production economies of scale would not also satisfy the economies of scale of the other fixed costs. Further, the companies most likely lo enter the RTE cereal industry would be food distributors and possibly large retail food chains which already have '" Whileexpress dint.nmsofml\rketsh"re ec()nnmiesofscalearedetermined by unit. volume irrespedive of market share that volume account for (Tr. 23,024, 23, 038-39, 2:J 506, 2.1 508091. The I % and 5% market share rigures in He..den-McKie were defined as 12 milion and 60 mi!lionpounds per year, respectively (CX-CI 103Z-0), These market share figures remain a rough approximation, as RTE cereal industry sales increased only 14% from 1961 , the dat.e rellPch,d by the Headen-McKie report, to 1972 (GFX 1.16). Economies of 5""le could also vary bet.ween firms tu the ext.ent. their products, ingredi.'Ots and technology may vary (Tr,, 9931 00 While various ''x pert witnesses atu,mpt.ed to evall. mte the ..xu'nt of economies of scal" at various levels below 5% (see e.g, Tr 2.1027 29, 2:J 0:Hi- , 23,517, 26 820- , 26 827- , 26 863, 26 87G-711, there is no record or ot.her referenced basis for such evaluclt.ions, other than an effort to construe t.he Headen-McKie Report. KELLUlJU CU. , t;T AL. 'D' Initial Decision their marketing and distribution systems (Complaint counsel's 1976 Trial Brief, Vol. I, nn 201.41 , 201.43; Tr. 25 810, 25 848, 26 813 836-2, 26 873, 36 839-40). Inasmuch as the Headen-McKie production economies of scale estimates were in terms of size of plant (CX-CI 103Z-0J, it is unclear whether or not a food distributor could reach economies of scale by combining RTE cereal production with the production of other food products in the same plant. This reflects stil further upon complaint counsel's failure of proof in establishing economies of scale for the RTE cereal industry. (183) This failure carries over to the question of relief in the event of a finding of violation, where complaint counsel propose spin-offs to create firms capable of producing at least 5% of total industry output.

(b) Preemption By Reason of Anticipation of Retaliation Complaint counsel (CPF 9-300 thru 9-307) assert that a potential entrant may be preempted from introducing a new product because of the belief that existing firms would be in a position to react more strongly against an outsider s new product than against new products of each other; that an outsider would seek a product having a larger share of the market in order to better resist such anticipated retaliation; and that while the potential entrant is delaying in seeking a larger opportunity, existing firms would preempt his entire endeavor by introducing their own products. 643. Complaint counsel' s theory requires localization to the extent that an existing company s reaction can be limited to a narrrow segment of the market. As previously found (Findings 59- 150, 603), localization exists to an undefined, variable degree. The theory also requires significant differences between costs at the level of entry available to new firms and the level at which firm economies of scale can be achieved. This has not been established since there has been a failure of proof with respect to firm economies of scale. 644. Further, complaint counsel' s theory is just that-a theory. There is no evidence of outside entry into a segment in which respondents have been located. Therefore, there is no evidence of reaction to such entry by an outsider; nor is there evidence of planned reaction to such an entry. The only contemplated entry into a segment occupied by existing companies evidenced by the record is that of Procter & Gamble, and that company expected no stronger reaction to its entry into the RTE cereal industry than to an entry into any other industry (Tr. 25 859).

Initial Decision 99 F. 7. Proliferation By General Foods 645. The introduction of new brands at a gradual rate which is sufficient to offset the gradual decline in shares of older brands is not anticompetitive and does not constitute brand proliferation. General Foods' pattern of brand introductions, in the face of a steadily declining market share since 1953, constituted that kind of product introduction activity aimed at recapturing a declining share of market. It did not constitute brand proliferation (Tr. 27 429- 468-9 572-73; GFX 28 1366). (184) Deterrences Other Than Brand Proliferation 646. As defined by complaint counsel's expert, Dr. Schmalcnsee There is a barrier to entry if there is a difference or a lack of symmetry between the positions of existing firms and potential entrants, such that the existing firms can earn profits while the potential entrants would suffer losses" (Tr. 22 345). The definition was subsequently restated, I'Barrier to entry is a factor or condition that makes possible, or more precisely, a factor or condition that makes the situations of existing firms in an industry different from the situation of potential competitors in such a way that the existing firms can earn excess or'IDonopoly profits persistently, and yet entry will not be attractive." (Tr. 22 381).

647. Economists disagree as to what conditions of entry are properly classified as barriers (Tr. 22 388, 23 408). There are other conditions of entry which make entry more difficult but which both incumbent firms and potential new entrants are required to overcome. Because these conditions of entry do not present an asymmetry or difference between incumbents and potential entrants, they are not regarded as barriers to entry (Tr. 23 407 8). Even barriers to entry do not bar entry. They represent obstacles to be overcome, but entry deterrence is not inevitable (Tr. 22 384). We need not concern ourselves with the differing opinions of economists as to which deterrents to entry are properly classified as barriers. We are concerned with the extent to which any conditions of entry explain the lack of new entrants into the RTE cereal industry.

Complaint counsel have contended that respondents' brand proliferation is the sale explanation for lack of entry in that industry. As I have found, complaint counsel have espoused a logical theory why brand proliferation (which is another term for competing by introducing new brands) would act to deter the introduction of products by new entrants. The extent to which lack of entry in the Initial Decision RTE cereal industry may be so explained, however, is an unkown quantity because of a failure of proof of two elements upon which the theory rests. These are: (1) minimum firm efficient scale of entrance below which the new entrant would be at a competitive disadvantage to existing firms and (2) the degree of localization of RTE cereal products.

To the extent these two elements may have an impact on a potential entrant, brand proliferation may well exert a deterrence on entry. It is not, however, the only deterrence. (185) 648. Because of the high capital costs associated with entry, only large firms are realistic potential entrants into the RTE cereal industry (Tr. 17 559-60 857- 676 155 258-60 022- , 27 009- , 26 603; CX-CI 103Z-52; CPF 9-319). Based on complaint counsel's position as to production requirements to achieve minimum efficient scale, initial costs for plant, equipment and marketing would range from $50 million to $150 milion (Tr. 21 857 039).

649. Of those large firms, potential entrants would be further limited to those already in the field of grocery production and supply. This is so because Firms look for opportunities where their established abilities can be utilized, not merely as an exercise of existing capability, but in order to cut down their overall costs of entering and operating in the RTE cereal industry (Tr. 25 809- 259; CX-CI 103Z-52). For example, a grocery firm with a diverse line could add RTE cereal using its existing sales force and distribution system (Tr. 22 998- 000, 26 595- , 26 813). 650. Even the limited number of potential entrants would be cautious about entering because of the high capital costs and the long lead times in recovering capital investments (CRP 7-65; Tr. 155). The existence of such long lead times to develop, test and market successful products has previously been found (Findings 552- 568-77 600-01).

651. Obviously, faced with a long lead time for showing profits and recovering very high investments, a potential entrant must come up with a potentially overwhelmingly successful product, one that promises a relatively high rate of return (Tr. 22 420- , 26 624; CX-CI 103Z-52-53). The problems in coming up with such a product are considerable, costly to overcome and time consuming (supra Findings 552, 596, 650). They may well be insurmountable. Procter and Gamble, the only company evidenced to be interested in entering the RTE cereal industry (other than the granola segment), began by 1960 to consider whether to enter. It started its efforts to develop a product in 1964. By 1970, it had spent between $500 000 to Initial Decision 99 F. 000 000 only to determine that the product it had developed was not suitable. While still interested, Procter and Gamble has not introduced an RTE cereal product to this date (Tr. 25 803-05, 25 833- , 26 898, 26 903; CX-PG 60).

652. Companies enter an industry only in anticipation of profits. However, a new entrant could not be assured that its product would receive consumer acceptance. As previously found (Findings 590-94), the failure rate of new products is high and a number of other products which remained on the market failed to earn satisfactory profits (CX-CI 103Z-52, Z-53, Z-54). While the industry rate of return and the rate of return of individual competitors is a complex matter (to be addressed in the next section) and a potential (186) entrant would be unaware (except for Kellogg)" of a company s rate of return on RTE cereal (Tr. 22 669-79), a potential entrant would be aware of a disparity of success among the various RTE cereal competitors. It would know, for example, that General Foods introduced 22 products from 1950 to 1977, but that only seven such products were still on the market in 1978 (Tr. 38 136). It would know also that General Foods' share of market grew only 1 % from 1958 to 1966 compared to a gross national product growth rate of 4.9% (GFX 27). It would have observed General Foods' costly failure in its attempt to produce and market RTE cereals with frozen dried fruits (Tr. 36 382, 36 384, 37 181; GFX 1211, 1212; CX-GF 571). It would have observed General Foods' decline in market share from 25. 3% in 1951 to 17. 1 % in 1973 and Nabisco s decline in market share from 12.5% in 1945 to 4.9% in 1973 (CX 106B). A potential entrant would not normally anticipate emulating the apparent success of the leading company, Kellogg (Tr. 26 554), but would be forewarned of risks by the experience of other competitors. And General Foods growth rate and experience would indicate a degree of risk and lack of attractiveness for new entry (Tr. 27 420-21). 653. This observation of high risk would deter a potential entrant from entering the RTE cereal industry, unless it could foresee a rate of return higher than normal in order to compensate for the risk it would be incurring (Tr. 22 389- , 23 237, 26 584, 38 517). 654. Entry must be made with a product that is perceived by consumers to be meaningfully different. "Me-too" products, products without significant differences, are likely to be failures (Tr. 9184- 965, 14 516-25 , 14 584- , 14 966-7, 15 228, 15 847, 15 851- 952, 17 453, 17 627- , 17 634, 17 650, 22 720, 26 549; CX-K 396B). Since the introduction of presweets in 1950, nutritional cereals and 0' Since Kellogg was almost entirely in the RTE cereal bu ines.s, i rf!ported profjts would reflect its SUCCe!' thatiodustry Initial Decision then natural cereals, no products encompassing new basic concepts have entered the market (CX-GF 401OZ-2). Therefore, it takes a considerable amount of research and development, market research and basic know-how to come up with an acceptable product. With six major companies already experienced in the field constantly seeking to fil every new product opportunity, the difficulties of a wouldentrant to come up with an acceptable product are apparent. Witness the so-far unsuccessful efforts of Procter and Gamble. The extent to which companies already in the market have advantages over would-be entrants adds to the impediments to entry. The advantages vary in degree but, taken together, they may constitute a significant impediment to new entry.

655. The incumbents have the advantage of being in place with products that overall constitute a full line or lines (Tr. 26 554, (187) 760 576). Not only do the existing products have an advantage over new products by reason of brand loyalty, but that brand loyalty carries over to new products of existing companies (Tr. 15 098-99 15,165 584-85).

656. Another advantage is the experience and technological know-how of insiders, not only as reflected in the individual capabilities of their employees, but as developed as a team or group effort." New entrants do not have all of the skills of existing firms. It takes time to devclop those skills. This is referred to as the learning curve. As cumulative experience is gained under the learning curve phenomenon, costs go down (Tr. 23 488, 27 321 495- , 30 542-43). This is a disadvantage to new entrants which require time to develop necessary skills (Tr. 27 321, 30 543- 44).

657. Even existing firms lack the know-how to produce certain categories of products and to duplicate some of the products of their competitors (Tr. 22 987- , 32 995-96, 33 133, 33 209). This lack of know-how would be an even greater handicap to a new firm attempting to produce a targeted product. In building its second RTE cereal plant at Modesto, General Foods attempted to replicate the equipment it had in place at Battle Creek. Battle Creek engineers participated in the design of the Modesto plant. A Battle Creek foreman supervised training. Key Modesto personnel were sent to Battle Creek to work in that plant and gain as much experience as 02 Product and equipment devdopment and maintenance, as well as production itself. at Genen.J Mills is accomplished by teams of personnel working together (Tr. 29 304- 786-- , 33 ()(H . 3. 2191. Team leaders ar.. promoted from the inside, butnnly frotn those who have had at Jeast eight yea rsofexp"rience(Tr. 997 999- :13 , 33 002--3) "' Employees of General Mills IHlve been able . over time. to incrcas," their productivity and reduce costs (Tr 664 670, 36 6f!2-84) g., Initial Decision 99 F. possbile. Battle Creek technicians supervised the Modesto start-up (Tr. 36 691-93). After nearly two years of operation at Modesto General Foods obtained an output on the Raisin Bran/Bran Flakes line of 55.5% of the raw material input. This compares unfavorably with an 80% yield at Battle Creek (GFX 1347). 658. After three years of production, Ralston s Lancaster plant had a lower yield on flakes than Ralston s Battle Creek plant. Ralston attributes this to the lower level of experience and knowledge on the part of the Lancaster plant personnel (Tr. 10 693, 10 755- 56).

659. The costs to Ralston in 1969 of producing its Chex products at its Cincinnati plant, after ten years of operation, exceeded the (188)costs of producing those products at Battle Creek, a plant built in 1928, by from 1.9 to 3.8 cents per pound. Yields were higher in Battle Creek than in Cincinnati (Tr. 10 693 , 10 755- , 23 489- 897). Ralston s Director of Production would attribute this difference to " experience, length of operation and personnel knowledge (Tr. 10 756). Ten years would seem to be long enough to overcome all learning curve requirements. While I do not find that the learning curve phenomenon explains the lower efficiency of Ralston s Cincinnati plant, we do have these three instances where more recently built plants are less efficient than older plants of the same companies.

660. Manufacturing multiple products as they do, and with fluctuations in the success of particular products including failures respondents frequently have excess capacity that they can use to manufacture new products (see, e. Tr. 13 481; GFX 423C). Thus whereas an insider need only consider operating costs, a new entrant would have to build capacity and cover both fixed and operating costs (Tr. 13 057-t1, 13 481- , 13 656, 13 994, 17 832- , 23 090- 862 , 29 890, 35 829-30; CX-GF 461, 4039G, Z-3 thru Z-4, Z-61 thru Z-63).

661. The only entry into the RTE cereal industry during the period covered by the complaint has been into the granola or natural cereal segment. Complaint counsel contend that entry was possible here since respondents had overlooked this segment, so that there was no brand proliferation which otherwise would have imposed a barrier to new entry.

662. Lack of existing competition certainly created a greater degree of opportunity for new entrants (Tr. 25 915-18). Entry into the granola segment, however, cannot be equated with entry into the remainder of the RTE cereal industry. Granola is a simple product to Initial Dccision produce. Very small manufacturers can produce granola on Unsophisticated machinery,64 without any research and development and with very low capital investments (Tr. 21,858, 23 471- , 37 298 300). "

663. By 1970, the granola segment accounted for 9% to 10% of the RTE cereal market (Tr. 29 678, 36 622). Competition existed among many companies, including Kellogg and General Mills which entered in response to the entry of others (Tr. 13 087- , 17 802). (189) 664. Among those that entered the RTE cereal industry with natural cereal products were Pet, Pillsbury, Colgate and International Multifoods (Tr. 25 912- , 26 487, 26 491-92). Pet was the most successful with its Heartland product which, at one time, attained a 1 % market share (Tr. 22 934). These firms, after some initial success, all withdrew except for Pet which has become an insignificant participant (CX 434D). Carnation had been in the RTE cereal industry for some 20 years, but exited in the 1960's (KX 35). Pillsbury had previously exited in the 1930's and then again exited from the natural cereal segment in the 1970's (Tr. 26 334, 26,492). H. J. Heinz was in the industry in the 1930's and exited sometime prior to 1958 (Tr. 12 992, 26 489; KX 35). Potential entrants would have been dissuaded from entering the RTE cereal industry by the noted inabilities of these major grocery product companies to compete successfully (Tr. 26 511-13).

665. Respondents are charged with responsibility for lack of entry since 1950 when, it is alleged, they turned from price competition to brand proliferation. Y ct, the industry was experiencing growth and respondents were flourishing in the 1940's and there was no entry during that period (Tr. 26 112; GFX 1366). Thus, no change in entry was effected by respondents' alleged change in competitive activity around 1950.

666. A growing market is normally an attractive one for entry since this reflects opportunity for a new product to share in the growth rather than simply try to take business away from existing products. Therefore, entry into such a market is less likely to provoke strong retaliation (Tr. 26 156-57). 667. For the overall period 1952-1966, RTE pound sales grew by 72.8% whereas the "all goods and services" Gross National Product ("GNP") grew by 63.9% (Tr. 26 158). However, for the period 1958- 1966, RTE cereal pound sales increased 32.9% while the GNP went a, The m""ufadure of granol" is sO Rimp!", that it Can be accomplished using the facilities of an existing bakery(Tr.37 277).

,." Organic Milling initj"lly''''lppndcd $1,000 and then e;.pended "nnua! sales to $1 millio" On "dditiunaJ expenditUff'H of$1O oo. Sales reached $3 miJJion with total c"pital expenditures ofsom.' SH,5 thol"'''IJd g., Initial Decision 99 F.T. up 42.1 % (GFX 25). Further, the RTE cereal industry experienced a period of "no growth" from 1966 to 1970, which would have made entry into the industry relatively unattractive (Tr. 26 157-58). Summary 668. Respondents have engaged in intense, uncoordinated competition in the introduction of new products. This competition is not unlawful nor was it induced by other unlawful activity. 669. Obviously, the more successful new products introduced by respondents and other incumbents, the more saturated the market and the less requirement and opportunity for the introduction of new products by outsiders. By the very nature of differences in demand (e. presweets, natural, fortified, bran, flavored), products are to be a degree localized, and a new product would to a degree be limited as to the segment from which it could attract its users. (190) Individual products account for lower market shares and smaller poundage of sales than formerly. The industry has become one of relatively small volume brands.

670. Incumbents are at an advantage over potential entrants in developing and marketing acceptable new products. They can utilize existing research and development, market research and other expertise in locating opportunities and developing products to meet perceived demand. They can also utilize unused capacity for production of a new brand, whereas a new entrant would have to build that capacity.

671. Respondents and other incumbents are not only capable of finding and taking advantage of an opportunity before a potential new entrant, but, because of economies of scale, are in a position to take advantage of smaller opportunities.

672. Would be entrants are faced with substantial fixed costs in research and development, market research, plant production equipment and introductory advertising. To the extent the requirement exists to introduce multiple products, the costs would multiply. Potential entrants, therefore, are limited to large firms, primarily those already producing and supplying grocery products who can utilize their existing expertise and so minimize costs of entry and operation.

673. The limited number of potential entrants would exercise caution in actually entering because of high capital costs, long lead times in developing and marketing acceptable products, extended periods even after entry in reaching levels of profitable operation and recovering capital investments, and the high risk that a product Initial Decision may prove unacceptable at various stages up to national entry or may fail after entry. Indeed, the problems of developing an acceptable product with which to enter may be insurmountable. 674. While potential entrants would be aware of the publicly reported profitability of Kellogg, they could not hope to emulate the most successful company. Their desire to enter would be tempered by their observation of others in the industry. They would hesitate knowing of the failure of General Foods in marketing its cereals with fruit and its loss of market share, and of Nabisco s decline in market share. Potential entrants would also hesitate because of the observed inability of Pillsbury, Colgate, International Multifoods, and H. Heinz to remain in the market and the limited success of Pet. As stated at the very beginning of this section, complaint counsel's theory of the existence of a barrier to entry into the RTE cereal industry is premised on the assertion that the industry has enjoyed supracompetitive profits and rapid growth; and that the ahsence of entry under these conditions indicates the existence of a barrier to entry.

675. While there was overall substantial growth from 1952-1966, growth from 1958-1966 was below that of the GNP and the industry was (191)in a period of no growth from 1966-1970. This leaves for determination the very first premise of complaint counsel's barrier to entry theory-the enjoyment by the industry of supra competitive profits. I have left this opening premise until last since it falls within a very complex, highly contested set of issues involving industry performance. This wil be dealt with in the next section. VI. PERFORMANCE A. Economic Theory of Performance Espoused By Complaint Counsel As previously noted (pp. 22, 28-29), complaint counsel are relying upon a structure-conduct-performance analysis of the RTE cereal industry in their effort to show a violation of law. Findings on industry structure and respondents' conduct have already been made. This leaves an analysis of performance to which we now turn. The following analysis of performance is derived from complaint counsel' s introductory definition and explanation of that subject (CPF 11-1 thru 11-17). It shows preliminarily the burden of proof they have assumed.

Performance. This is the degree to which an industry serves the consumer and can be defined in terms of three major goals (Tr. Initial Decision 99 F.T. 661). These three goals are: (1) to maximize the efficiency with which an industry and the economy overaJi operate; (2) to charge prices low enough so that excessive profits are not reaped from consumers ("equity ); and (3) to maximize producer progressiveness (Tr. 27 652, 27 661).

Efficiency. Competitive industries maximize efficiency in two ways. First, firms in an efficient industry will minimize their costs and so avoid waste of production resources (Tr. 21 693, 27 655). The second form of efficiency is termed "allocative' efficiency." When an optimal level of allocative efficiency is obtained, prices are kept at cost (including a normal rate of return), consumers demand and are able to purchase the greatest quantity of the good, and an optimal level of resources is devoted to production of the good (Tr. 21 695- 153, 27 652-54).

The economic law of demand is that the higher the price, the lower the quantity demanded of a good (Tr. 27 992-93). Whenever an industry prices above the competitive level, consumers demand less than they normally would, output is restricted and, thus, resources which should have been devoted to production of the good will be diverted to other goods. A misallocation of resources results etr. 692, 21 695, 21 707, 26 153, 27 652). Where there are excess profits, resources are being misaJlocated (Tr. 21 814). With free and open competition and prices kept at competitive levels, supply (192) and demand reach appropriate levels and consumers spend appropriate amounts for the good. In the absence of such free and open competition, consumers will spend lesser or greater amounts for the good and, correspondingly, will spend more or less on other choices. Thus, resources would be allocated- The resultant losses when resources are misallocated are called "deadweight" or "welfare losses (Tr. 27 999- 000).

Consumers lose due to an industry s cost inefficiency. With competitive pressure, firms within a competitive industry strive to minimize their costs. When all firms minimize their costs, the industry s average level of costs goes as low as possible and the quantity of the product dcmanded by consumers would be at the greatest level that would permit firms to earn a competitive rate of return. However, in monopolistic industries, costs may be unduly high. For example, firms may be inefficient due to poor management, yet be able to survive because they are insulated from competition.

Equity. If an industry is performing equitably, neither prices nor profit levels will exceed those needed to attract the required amount initial Decision of capital to the industry. When prices, and therefore profits, are held unnecessarily high, income is being transferred unfairly from consumers to producers, consumers are injured, and performance is poor (Tr. 27 658-60, 27,991-92). Excess profits are earned whenever consumers must pay more for a product than is necessary to cover all the costs of distribution and production, including a "normal" return to the capital invested (Tr. 21 692 , 21 707, 27 991-92). Persistent excess profits are evidence of poor performance and monopoly power (Tr. 21 790- , 26 100-01, 27 991- , 37 934). Excess profits are also called "supranormal profits supracompetitive profits" or simply "monopoly profits" (Tr. 21 707, 26 090 991- , 37 934). An industry that reaps monopoly profits for a persistent period of time is said to be exercising "monopoly power (Tr. 21 707, 26 081, 37 934). Monopoly profits show that consumers are being overcharged (Tr. 21 707, 27 991-92). Progressiveness. Industries that perform well are progressive in taking advantage of new technological developments in order to increase efficiency. They also strive to offer superior new products and improvements of existing products to consumers (Tr. 27 655). Underlying both equity and effciency is the principle that the lower the price of a good, the. greater will be the quantity demanded by consumers (Tr. 27 992-93).

Profits As A Measure of Performance Performance is initially appraised by an analysis of profits. Economists commonly analyze profitability to determine an industry s level of performance (Tr. 21 692, 26 105). Monopoly power is the (193)ability of firms in an industry to hold prices above competitive levels (Tr. 21 707 , 26 100- , 27 934). Profitability is frequently examined to determine the existence of monopoly power, because if firms have been able to avoid competition, it is likely that profits will exceed competitive levels and will do so persistently over time (Tr. 100-1). Persistent high profitability indicates that competition has not been vigorous (Tr. 21 790-91).

Respondents take issue with many of the principles expounded and developed by complaint counsel that have been recited above. There is basic agreement, however, that the degree of respondents and other industry members' profits is an important issue in this case, preliminary to reaching the other issues referred to above. . . . . . . Initial Decision 99 F. B. Profits 1. Accounting Vis- Vis Economic Rates of Return 676. Complaint counsel have elected to measure and evaluate profi in terms of rates of return on capital employed. This is an appropriate means of measuring profits for purposes of this case and is preferable to other measures, such as return on equity or return on sales (Tr. 19 069- , 21 75B-59, 31 288, 31 290). 677. Capital employed is the sum of all capital supplied to a firm by stockholders, in the form of equity or stock purchases, plus debt such as bond issues; total assets minus current liabilities plus the debt included in current liabilties (Tr. 19 160, 19 169-70, 20 172). Since the investment base in the accounting rate of return includes capital supplied by both equity and debt, the income measure or numerator of the ratio includes income or payments to both forms of investment. The income measure, therefore, is net income to the firm after taxes (the monies available for distribution to stockholders) plus interest on the debt included in capital employed (Tr. 070). This corresponds to a return on net assets. 678. Economists and other business analysts have traditionally relied on accounting rates of return to measure both firm and industry profitability (Tr. 19 071, 21 698, 22 301, 31 373, 38 224-25). However, more recently, economists and financial analysts have (194)come to recognize serious shortcomings in the use of such accounting measures to judge firm performance, especially in certain industry settings (Tr. 19 045 , 19 333-34, 26 106-9, 31 213, 37 954). 679. All of the economic experts who testified in this case agreed that accounting rates of return could be unreliable and misleading by overstating the profitability of a firm or industry (Tr. 13 336-1 049, 19 332 , 19 352, 19 361--2, 19 35B-59, 19 987, 21 700 705-D6, 25 288-9, 26 116-18, 31 177- , 37 493-94, 37 951- 182, 38,47B-79). Dr. Stauffer, who was retained by the Commission to update its profitability analysis (Tr. 26 114-15), and who was its primary witness on the issue of profiabilty, was of the opinion that:

rate of return as conventionally computed by accountant." is seriously in error, generally having an upward bias and thus overstating the economic rate of return of the firms (Tr. 987; KX 17, p. 97). He believed that accounting rates of return can be quite imperfect .0 This t.nd to uodersl.l. Kelloggs ast bas to a slight degree and 11 bias Kellogg s rate of return slightly upward relative to most other firms, since Kcllogg is able to finance its operations through nonintercst bearing liabilities ('fr. 31 28).

Initial Decision and significantly misleading by overstating profitability (Tr. 19,352 358-59, 19 361); and that thc biases would be most substantial in overstating rates of return for industries and firms that engaged in a lot of research and development and advertising and also where there were substantial time lags between investments and the receipt of returns on the investments (Tr. 19 336-1 , 19 361, 25 288- 89). As previously found, all of these conditions exist in the RTE cereal industry.

680. Recognizing the shortcomings of accounting rates of return Mr. Michael Glassman, then Assistant Director for Economic Evidence of the Commission s Bureau of Economics, arranged for the employment of Dr. Stauffer who developed the economic rates of return relied upon by the Commission in this case (Tr. 26 101-D2 114-15).

681. The fundamental difference between an accounting and an economic rate of return concerns timing (Tr. 19 069, 19 332- 177, 37 951, 37 490). Timing relates to when moneys are spent by a firm, and when the incomes associated with those expenditures are received. The economic value of dollars received two years after investment is greater than dollars received five years later ('' 332- , 19 336-1, 19 386 478-79). Generally, the more lagged the cash flow of receipts relative to the cash flow of expenditures, the higher will be the disparity between the accounting rate and the economic rate (Tr. 31 178). Nevertheless, a dollar is recorded as a dollar in the books of account of a company regardless of the length of time between the investment and the receipt of money that is generated (Tr. 37 951).

682. Accounting and economic practices differ with respect to timing of both expenditures and receipts (Tr. 19 332-33). These (195) timing differences frequently are associated with the different practices of accountants and economists with respect to advertising, research and development and depreciation (Tr. 19 069, 19 333). Timing differences can result in an accounting rate of return that is not an accurate measure of the economic rate of return (Tr. 31 177- 78).

683. The difference between accounting and economic rates of return is illustrated by an example in the record of a firm which produces whiskey, which it ages for eight years. If it is assumed that it costs the company $1 to prepare a barrel of whiskey which the firm sells eight years later for $2, the firm makes $1 of accounting profits at the time of the sale. The accounting rate of return for the transaction is 100%, based on an original investment of $1 and the profit of $1. An economist views the transaction differently. An Initial Decision 99 F. economist would ask at what interest rate one could have invested the $1 to yield $2 eight years later. The economic rate of return is that interest rate. Thus, in this example, the economic rate of return is 9.05%, because if the distiller had invested $1 at 9.05%, he would have had $2 at the end of eight years.

684. In the example, the economic rate of return was 9.05% because that interest rate would yield $2 on a $1 investment in eight years. More technically, the economic rate of return is defined as the rate of discount (interest rate) under which the discounted present value of all cash inflows over the full life of a firm s investments is equal to the discounted present value of all cash outlays (Tr. 19 340- 171- 489- 944-46).

685. There may also be differences between accounting and economic rates of return caused by the accountant's treatment of research and development, advertising and depreciation (Tr. 19 333). 686. Accountants treat all research and development ("R&D") as a current expense (Tr. 19 342, 19 504). This means that, in computing profits, all expenditures on R&D are deducted as expenses from the firm s total revenues in the year in which the money is spent. R&D is similar to fixed assets, however, in that it generates revenues for the firm over future periods of time (Tr. 19 342-43). The accounting treatment of R&D may cause a distortion in the true rate of return because today s income is being charged for expenses that produce benefits in the future (Tr. 19,342-43).

687. Advertising expenditures also may cause differences between the accounting and economic rates of return (Tr. 19 333). Advertising has some effects which occur quickly, but it also has effects that are longer lived (Tr. 19 359--0). Some of the effects of advertising wil be realized in the same accounting period as that in which the expenditures are made (Tr. 19 359--0). However, a portion of the benefits wil occur in future accounting periods (Tr. 19 360 106-7). Accountants treat all advertising as a current expense charging it all to the year in which it is incurred, even though it may benefit future years (Tr. 19 360). To the extent that the revenues or benefits generated by advertising do not occur (196)in the same year as the advertising expenditure, an accounting rate of return does not provide an accurate measure of the economic rate of return (Tr. 361). "

688. Depreciation practices are another source of difference between accounting and economic rates of return (Tr. 19 333, 19 362- .' In addition to R&D and advertising expenses. portions of expenditures relating to market reflearch, test marketing, promotions, establishment of a distribution system and quality control, are of a qu;mi-capital nature and their treatment as 100% immediate expen ;es 00 books of account tends to overstate true rate of return (Tr. 322- 332-33 574- , 29,Sfj2, 37 622- 37.712). Initial Decision , 21 701- , 31 191-95). Depreciation is the cost to a firm associated with "using up" an asset purchased by the firm to generate future income. This means that if a firm purchases a machine at a given point in time it must subtract from its revenues a charge for using up (depreciating) that investment over some period of time (Tr. 363).

689. Accounting depreciation is determined by accounting conventions which are based on convenience, uniformity or other considerations apart from economic significance. Economic depreciation is a function of future cash flows (Tr. 19 362--3). Since accounting conventions do not necessarily (or even usually) reflect the actual timing of the benefits received from an investment in fixed assets, a bias in the accounting rate of return results (Tr. 195).

690. Accounting rates of return distort true rates of return because of the accounting conventions associated with depreciation. Accountants have conventions for stating the estimated life of investments which mayor may not correspond to the true or economic life of the investment. They then amortize the amount of the original investment over its estimated life in form of depreciation expense. The economist, on the other hand, is interested in the true investment life and the amount and pattern of cash flow generated by the investment over time (Tr. 19 362; CPF 11-66). Cash flow does not include a deduction for depreciation (Tr. 19 366, 19 373 376, 19 394 , 21 701--2). Only under very special circumstances would accounting and economic depreciations be the same (Tr. 362).

691. One reason for the difference results from the rules used for tax depreciation purposes. Firms have been permitted by the IRS to take more rapid depreciation for tax reporting than they would for normal bookkeeping purposes. Firms use various accelerated depreciation formulae for tax purposes. In general, firms use accelerated depreciation because it is in their interests to use the most rapid depreciation schedule allowable under the tax code (197)(Tr. 19 363 409-10). Again, this accounting convention of understating the true life of a fixed asset overstates the economic rate of return on investment.

692. Taking particular note of the large capital investments, the extremely high rate of advertising, the extensive amount of R&D and the long delay in securing a return on investment " I find that .. The larger the magnitude ofsuch expeTlditures ami th..loTlger thp delay in seclIring a return, the larger the bias in the accounting rate of return the more the accounting rate ()freturn is greater than the economic rat!: of return (Tr. 25 363 621J.

Initial Decision 99 F. economic rate of return, not accounting rate of return, is the appropriate measure to be used in appraising profits enjoyed in the RTE cereal industry and in making comparisons among respondents with other companies and with other industries. Methodology Of Computing Economic Rates Of Return 693. The economic rate of return for a project can be measured if one knows how much has been invested in it and can identify and measure the cash flows or benefits associated with the investment. Thus, if one knows, for example, that a company has invested $100 in a particular project and that the cash flow generated by that project is $30 a year for ten years, the economic rate of return can be computed from tables or by calculation. The formula for the economic rate of return for a single project is widely used by bankers, economists and insurance brokers (Tr. 19 366-72). 694. The determination of the economic rate of return for an ongoing firm, which may be viewed as a collection of projects, is a much more complicated matter. Dr. Thomas R. Stauffer has devised a formula to make such a determination. This formula is a pioneering contribution in the field and was the subject of Dr. Stauffer s doctoral dissertation in economics at Harvard University. Other doctoral dissertations in economics at Harvard have applied his formula in the analyses of specific industries (Tr. 19 373, 19 378- , 19 385, 21 759-64). This formula was applied by Dr. Stauffer in the instant case.

695. Dr. Stauffer s formula starts with an accounting rate of return and from it calculates an economic rate of return. It is necessary to know the investments a firm has made and identify the cash flows associated with those investments. However, as noted above, there are accounting conventions that govern the way a firm accounts are set up, as well as accounting rules which govern tax liabilities which, in turn, affect the firm s accounts. (198)Dr. Stauffer s formula takes into account various accounting conventions and tax rules and makes certain assumptions, including that of the pattern of cash flows (Tr. 19 374-79; CX 702B). 696. Dr. Stauffer s calculation of the economic rate of return involves an equation which has four terms, the sum of which equals the accounting rate of return. Each term of the equation takes into account, and adjusts for, potential biases in the accounting rate of return. There isa ubase" term which deals with the firm investment in fixed assets and working capital, terms relating to advertising and research and development, and a term which Initial Decision corrects for the difference between corporate and tax accounting rules (Tr. 19 379-82, 19 485-87; ex 702C, M). The Base Term 697. The base term in the equation reflects the contributions to the income of the firm generated by investments in fixed assets, such as plant and equipment, and working capital such as cash and inventories (Tr. 19 491). Estimates were made as to how long assets would earn income for the firm. These estimates were necessary because they relate to the timing of revenues associated with investments. This !!service lifetime" is included in the base term. The average service lifetime of assets purchased by a firm was calculated by Dr. Stauffer from data available from each firm s books of account (Tr. 19 411- , 19 442-49, 19 451-54; ex 702G-H). 698. It was also necessary to assume the average time interval between the investment in fixed assets and inventory, and the first generation of revenues from the investment (Tr. 19 406, 25 576). This lead time assumption is the average for all types of assets, including assets with short lead times such as inventories, automobiles and typewriters (Tr. 19 406-9, 25 576, 25 580). The Research and Development Term 699. The second term of Dr. Stauffer s economic rate of return equation relates to the contribution of research and development to income. It is intended to correct for the bias in the accounting rate of return caused by the accounting treatment of R&D as a current expense (Tr. 19 342-43, 19 504-05, 19 508; CX 702 0). As part of the R&D correction, it was necessary to make an assumption concerning the time interval between the expenditure of funds for R&D and the first receipt of revenues associated with those expenditures (Tr. 424-25 510-11).

The Advertising Term 700. The third term in the economic rate of return equation relates to advertising. It reflects the contribution of advertising (199) to accounting income (Tr. 19 520, 19 485-6; CX 702C, M). This correction was necessary to account for the revenue and income for the firm generated by advertising for a period extending beyond the year in which the expenditures were made and were deducted from accounting income (Tr. 19,360).

Initial Decision 99 F. The Depreciation Term 701. The fourth term in Dr. Stauffer s formula adjusts for differenccs caused by the accounting method of treating depreciation (Tr. 19 380 , 19 540; CX 702C, Q). The term takes into account the economic value of the timing difference which results from the use of a more rapid depreciation schedule for tax purposes than the schedule used for book purposes. The formula corrects for the effect on the accounting rate of return of the difference between tax depreciation and book depreciation (Tr. 19,412, 19 544-45). The Cash Flow Profile 702. Since the pattern and timing of cash inflows earned by a firm plays an important role in calculating the economic rate of return, in order to calculate the economic rates of return, Dr. Stauffer made assumptions concerning the actual pattern of cash inflows earned by each company (Tr. 19 386, 24 942-44). The assumed pattern of cash inflows is called the cash flow profile assumption and is included in each of the four terms of the equation (i. 1. base term; 2. R&D term; 3. advertising term; and 4. depreciation term) (Tr. 19 495, 19 514, 19 525, 19 541). 703. Two alternative shapes were assumcd by Dr. Stauffer for the cash flow profile: the triangular and the rectangular (Tr. 19 387 944-47). These names reflect the shapes of the graphs of the annual cash flows for the two assumptions (CX 702D). The two shapes were chosen as upper and lower bounds for the cash flow profile. The actual cash flow profile, according to Dr. Stauffer, lies somewhere in between these two extremes (Tr. 19 386-87, 19 394- 944-47, 24 954). The underlying assumption for a rectangular cash flow profile is that the cash flow produced by an asset is the same in each year for the entire life of the asset. The assumption underlying the triangular cash flow profile is that the cash flow produced by a new asset starts at a high level, and declines steadily over the life of the asset (Tr. 19 394- , 24 944; CX 702D). The shape of the cash flow profile had to be assumed, rather than calculated because Dr. Stauffer could not directly ascertain its shape from available data (Tr. 19 397).

While respondents uniformly agree that Dr. Stauffer s formula for reaching economic rates of return is much superior to relying upon an (200)accounting rate of return analysis, they do object to a number of Dr. Stauffer s inputs into his formula. The first objection is to the accounting data used by Dr. Stauffer. Initial Dccision (a) Accounting Dc.\t.a Allocations 704. The first, and very cdtical, step in reaching an economic rate of return is to segregate and allocate the accounting data (assets, liabilities, costs and expenses) that pertain to the RTE cereal business of the company being studied (Tr. 20 166). 705. General Mills, General Foods, Quaker and Ralston are engaged in various lines of business in addition to RTE cereals. Expenses were incurred and assets employed in the joint or common support of two or more of these lines of business without allocation by those companies to the particular businesses. For example, the records of General Mills show only operating profits based upon operating costs which include such activities as manufacturing, distribution, advertising and selling. These operating profits are before any charges for interest, taxes and a group of costs called unallocated corporate expenses. Major items included in unallocated corporate expenses are salaries of the top officers of the company, executive incentive payments to such officers, contributions and airplane operations and maintenance (Tr. 18 154, 18 173, 18 265 275; CX-GM 2486D). Similarly, these firms did not maintain records showing the portion of balance sheet values for all assets associated with their RTE cereal business (Tr. 18 178-89, 18 301- 305; CPF 11-21).

706. Identifying the accounting data that is applicable to RTE cereal is important because the accounting data chosen is one of the most influential inputs to Dr. Stauffer s economic rate of return calculations. Inasmuch as General Foods, General Mils, Quaker and Ralston have substantial joint assets, liabilities, costs and expenses which are not broken down by business endeavor '" it was (201) necessary to allocate to the RTE cereal business an appropriate portion of those items. Those allocations had to be made by means of estimates (Tr. 18 148, 166).

707. The following discussion of allocations will be limited t.o General Mils and General Foods since Kellogg, Ralston, and Nabisco s accounting data were not analyzed by Dr. Mellman and 0' Similarly, since General OOdH pays taxes on a company-wide basis, it has no need to develop such figures as accflH'd taxes payable" Or " provisions for income tax" solely for its RTf: cereal operations ('11'- 16 299). Yet, such figures are fwceSS'HY to obtain RTE ccrcal1in" of business data whicJl is as complete and accurat" as possible for cornp3rison with other companies.

10 For example. in 1972, corporate interest lwd unallocflted corporate cltpenscs amounted lo 42% of Genem! Mills' total corporale consolidated earnings before !.,x (CX--GM 2486) The rate of return estimates, therefnre, are highly sensitive to the amount of these expenses alloc"ted to the RTE cereal operation. Kellogg s bus;ness during the complaint period was virtually 1111 RTE cereals. Thus, allll:ations of joint costs and assets were unnecessary to arrive at its RTE cereal accounting rate of return (Tr. 12 036-37). The minor adju6trnenL that were made to Kellogg s ""counting data by complaint counsel's accounting expert, Dr. Mdlman are nol chal!.mged (Tr 20 171- 72: KPF 5-71) , Initial Decision 99 F. allocations of Quaker s figures amounted to only .03% on its accounting rate of return for years 1956-1972 (CX 701M). 708. There are no uniformly accepted accounting principles for allocating joint and common costs, expenses, assets or liabilities to individual segments of a business (Tr. 20 599). The President of the National Industrial Conference Board has concluded It is evident from the multiplicity of plans uncovered by this year-long study that there is no one best method of allocating central expenses. What is desirable in one situation sometimes proves to be inadequate or misleading in another" (Foreword to C. Baumes Allocating Corporate Expenses National Industrial Conference Board, Inc. (1963)). Among those in common use are sales dollars, sales units and costs (Tr. 38 024-25; GFX 1332Z-132).

709. The Federal Trade Commission s Bureau of Economics has indicated that it wil use a variety of allocation bases in processing line of business data:

In processing the LB data, the Pic staff will apply plausible, alternative nontraceable cost allocation rules- , allocation according to sales, assets, payroll, and contribution margin (profitsl-and test the sensitivity of calculated operating income figures to those alternatives (Bureau of Economics Staff Memorandum - 1974 Form LB Revision, at p. 5).

710. When detailed information on the make-up of a joint cost is unavailable, allocation methodologies are necessarily somewhat arbitrary (Tr. 21 036-52, 28 715-17). (202) 711. All parties agree that the standards published by the Cost Accounting Standards Board (CASB), Section 403 (4 C. R. 403), which relate to accounting standards for government contracting, constitute an appropriate guideline for use here (Tr. 20 162--4 907-D8; CPF 11-34; KPF 5-78, 5-79; GMPF 3-8). 712. CASB Section 403 provides that any home office expense that is directly identifiable with particular segments should be allocated directly to those segments. For the residual, the allocation should be based, to the maximum extent possible, on cause/effect relations. If after making that allocation, there is still a residual that residual should be allocated on the basis of a surrogate for causal or beneficial relationships. If after that allocation, there stil remains a residual, the standard provides that a three-factor formula should be used to allocate it. The three factors consist of payroll dollars, operating revenue (sales dollars), and tangible " A Bureau of Economics staff memorandum accompanying the Commission s Line of Business repurting rorm suggest use ofCASB standard 403 (Bureau of Economics Staff Memorandum-1974 Form LB R..vision, at p 5).

KELLOGG CO., ET AL. 207 Initial Decision capital assets plus inventories (Sections 403.20(a), 403.40(a)(I); Tr. 907-D9, 35 923-26). CASB 403.50(b)(IH2) (4 C.F.R. Section 403.50(b)(IH2)) provides in part:

. The allocation of centralized service functions shall be governed by a hierarchy of preferable allocation techniques which represent beneficial or causal relationships. The preferred representation of such relationships is a measure of the activity of the organization performing the function.. . . (2) Where neither activity nor output of the supporting function can be practically measured, a surrogate for the beneficial, or causal relationship must be selected. Surrogates used to represent the relationship are generally measures of the activity of the segments receiving the service. . . . Any surrogate used should be a reasonable measure of the services received and, logically, should vary in proportion to the services received. 713. The CASB has recognized that a cost basis is an appropriate surrogate for the activities managed, including general and (203) administrative expenses, under certain circumstances (CASB 410 421; 4 C. R. 410.106-10. , 421).

714. It is clear, therefore, that there are a number of alternative cost accounting methods for allocating joint assets, liabilities, costs and expenses, and that none of the methods has been authoritatively declared to be superior to any of the others. Certainly, none has been dictated for use to the exclusion of the others. Of course, in individual cases under particular circumstances, it may be demonstrated why a specific method of allocation is superior to alternatives or should be used to the exclusion of all others. This has not been demonstrated here; and the burden of doing so was on complaint counsel. All that the parties have done is to advance general reasons, not related to the specifics of this case, why one method should be preferable to another. All of these reasons ostensibly have been advanced to and considered by CASB (as well by the National Industrial Conference Board) and that board has seen fit to issue general guide lines which, in my opinion, allow all of the methods of allocation advocated by the various parties here. 715. Under these circumstances, there is a lack of proof that the allocations advocated and introduced by respondents are improper; and respondents are entitled to have those allocations utiized in arriving at accounting rates of return for their RTE cereal businesses. These allocations, in turn, are the ones to be used in Dr. Stauffer s formula for arriving at economic rates of return. More detailed findings covering the several allocation methodologies of the different parties follow.

70 'This formula could not be used. here beause of insuffcient payroll and tangible sales dollars is recognize by CASH as having II controllng impact ast data- In any event Initial Decision 99 FTC. Dr. Mellman s Method (The Method Contended For By Complaint Counsel) 716. Dr. Mellman did not reallocate costs and expenses, or assets and liabilities, where the RTE cereal company s original assignment was direct and not allocated, because there can be no improvement on a direct assignment (Tr. 20 169). Thus, for example, advertising expenses were not reallocated because advertising was charged directly to the brand for which the advertisements were run. 717. Dr. Mellman recognized that to determine the profits of a business segment, such as the RTE cereal segment of a diversified corporation, an effort basis is the best way to allocate expenses to segments where the information is available; that expenses should be allocated in proportion to how efforts were actually applied because it reflects the way resources were used by that segment; and that effort can be measured on a number of different bases. For example. in allocating such expenses as the salary of a division manager who is responsible for products in more than one segment an effort basis might allocate the associated expenses according to the amount of time spent in supervising the segments under his control. (204)Thus, if the manager spends 40% of his time supervising one of two segments, 40% of the expenses associated with the manager, such as his salary, would be allocated to that segment (Tr. 158-0). Respondents would agree with Dr. Mellman up to this point.

718. When detailed data on effort are unavailable, an alternative basis of allocation must be used. Dr. Mellman was of the opinion that the most reasonable approximation of an effort basis, in the absence of detailed information on effort, is generally cost (Tr. 20 158 164). The cost basis allocates expenses to the segments in the same ratio as direct expenses have been incurred by those segments (Tr. 20 161).

719. An example would be a division manager supervising two segments, A and B, and the absence of records which show the actual time spent (the actual effort) supervising the two segments. Assume that 40% of the total direct costs of activities managed by the division manager are generated by segment A. These costs include materials used, labor, overhead, marketing costs, promotions, selling expenses, and so forth. In the absence of detailed information on the time spent working on each segment, 40% of the division manager cost would be allocated by Dr. Mellman to segment A. The assumption underlying this method of allocation is that the division manager would allot his time in accordance with the relative Initial Decision activities of segments and the 40% share of costs attributed to segment A would reflect its share of total activity supervised by the division manager (Tr. 20 161-62).

720. Professor Mellman also used a cost basis to allocate assets and liabilities. He recognized that, ideally, allocations of assets would be based on actual usage of the assets allocated. Thus, for example, a warehouse might be allocated based on the square feet of space used by each segment. However, such a basis would require detailed information about the way in which an asset was actually used (Tr. 20 167). In the absence of such detailed information, Dr. Mellman used a cost basis for the same reasons he used a cost basis for allocations of expenses. In the illustrative example of the division manager, the assets such as the office used by the division manager would be allocated in the same proportions as his expenses. Thus, if 40% of his expenses were assigned to segment A, 40% of the assets he used would be assigned to the segment (Tr. 20 167-68). 721. Applying the same principles, a cost basis was used in the assignment of liabilities, whereunder liabilties are allocated according to ratios of the expenses related to that liability. For the division manager described above, there would be a liability shown on the company s books for salary owed to him but not yet paid. Under a cost principle, that liability would be allocatcd in the way in which the division manager costs were allocated, inasmuch as liabilities are costs and expenses that are incurred but not yet paid (Tr. 20 169- 70).

722. Of course, division costs do not necessarily reflect corporate activity and the utilization of corporate assets. (205)Relatively costly modern equipment may relieve the division manager of time consuming exercise of responsibility, and the cost may not at all reflect the share of warehouse space occupied. 723. The cost of effort or activity that is to be allocated is that at the corporate level, not at the division level. No necessary relationship exists between expenses at the corporate level and costs at the division level. The two levels involve different groups of people doing different things (Tr. 35 907-15).

724. Costs at the division level include outlays for raw materials. At General Mills, for example, roughly 75% of the costs of RTE cereal sales were ingredient costs (Tr. 18 272). Changes in the price or composition of raw materials are unrelated to changes at the corporate level. Depreciation is a significant cost item at the division level. There is no plant and equipment at the corporate level (Tr. 912-13).

725. Executive incentive payments are an item of unallocated Initial Decision 99 F. corporate expense. Such costs are inversely related to operating costs such as manufacturing, distribution and marketing inasmuch as the amount of executive incentive payments increases with the amount of corporate profits. Thus, as costs rise, profits decline and incentive payments would decline (Tr. 18 267, 18 269, 18 342-43). 726. Professor Mellman allocated cash and marketable securities on the basis of total segment operating costs plus interest, taxes and unallocated corporate expense, on the gound that cash is used to pay for costs and expenses (Tr. 20 549, 20 799-803). However, cash and marketable securities, as they appear on the General Mils balance sheet as of the end of its fiscal year, for example, are a function primarily of profits. They would appear to be inversely related to costs inasmuch as the higher the costs, the lower the cash on hand and the cash temporarily placed in short-term marketable securities (Tr. 18 326, 35 920-22).

727. Application of Dr. Mellman s costs basis allocation formula resulted in a range of General Mills' ratio of allocation of from 10% to 18% from year to year (Tr. 20 794-97), a substantial variation without apparent explanation. It would also result in an exceptionally low ratio between current assets and current liabilities for General Mils' RET cereal segment. This is one measure of the financial soundness of a firm. Professor Mellman allocation so little of the unallocated corporate assets to General Mils' RTE segment that the segment has an average current ratio of only 1.4 for the period 1958-1972 and only 1.0 in 1965, 1.2 in 1971 and 1.1 in 1972 (Tr. 20 925-30; GMX 58, 472). These General Mils ratios are substantially lower than those for Kellogg, General Foods, Quaker and the Internal Revenue Service group of manufacturing firms which compliant counsel use as a profit " benchmark." Indeed, the General Mils ratios are lower than the ratio for the firms in the IRS sample that suffered net earning losses (GMX 58, 472). Such ratios would be unrealistic for a General Mils RTE cereal segment which is alleged by complaint counsel to have been very profitable (Tr. 11I). In addition, Dr. Mellman s allocation to RTE cereal of (206) General Mils' corporate unallocated expenses decreased by onethird from 1967 to 1972 despite the fact that sales and profits increased by 30% (Tr. 21 1I0).

728. The above analysis of a cost allocation approach and its application to General Mills' and General Foods' operations indicates why such a method of allocation cannot be found to be the only Qr a superior method of allocation to be utilized in this case. Dr. Mellman has conceded that his cost allocations may well vary from Initial Decision what an actual measurement of the breakdown of a joint effort might show (Tr. 20 893, 20 905, 21 524).

General Mills' Method of Allocation 729. General Mils' expert, Mr. Troxel, used profits as the primary basis for allocation (Tr. 18 266-72, 18 302--5, 18 325- 342-43), although some items were allocated on the basis of sales (Tr. 20 830-33; CX-GM 2486). As previously indicated, there does appear to be a causal relationship between profits and incentive payments since incentive payments are normally based on profits (Tr. 18 266-7, 18 269, 18 342-43, 35 909). The same causal relationship appears to exist between profits and cash and marketable securities (Tr. 18 325-26, 35 920-22). Inasmuch as taxes are a direct result of profits, the allocation of accrued taxes on a profit basis also appears appropriate (Tr. 18 323).

730. Just as with Dr. Mellman s cost basis of allocation, Mr. Troxel's profit-sales basis has its obvious failings. Since sales figures include an element of profit, allocations based on sales reflect, in part, a segment's ability to bear costs rather than the effort or resources expended. Furthermore, there are numerous factors that affect selling price, and there is not necessarily a relationship between selling price and cost or usage (Tr. 20 164-65). 731. For reasons similar to those applicable to allocations based on sales dollars, a profit basis reflects the ability of a segment to pay or bear the allocated costs, rather than effort or activity (Tr. 20 165). The most obvious failing of a profit margin basis of allocation however, is that, if a segment has zero profits, it would get no allocation of corporate expense; and if it had a loss, it would actually receive a negative allocation, despite the fact that corporate headquarters might have been devoting an inordinate amount of work in an effort to turn the unprofiable segment around (Tr. 20 166 168-9). Further, a profit basis of allocation would result in allocation revisions reflecting changes in profits, irrespective of the flow of services (Tr. 20 793). " (207) 732. Nevertheless, sales and profit bases of allocation are acceptable under statements issued by the Commission s Bureau of Economics and under CASB standards, and are no less acceptable than the cost basis utilized by Dr. Mellman. 733. Further, in the Bureau of Economics' staff memorandum attached to the Commission s Supporting Statement to its 1974 Line " The Jack of complete ,'eliability of a profit basis ofal!ocation is thus demorJstr'lte; IInd it is immaterial that '10 segment of General Mils showed II loss on Uw basis of II five-year moving average of profits (Tr. 35 927-29) Initial Decision 99 F. of Business reporting form, it was stated "that the reporting companies are in a better position than the FTC staff or anyone else to make them (allocations)" (at p. 5). Normally, substantial weight would be given to allocation procedures if they were utilitzed by a company in the normal course of its business. While the allocations relied upon by General Mils were prepared by an outside consultant for purposes of this case (Tr. 18,150), they are nevertheless entitled to the same presumptions of correctness they would be entitled to in Line of Business reporting.

General Foods' Method of Allocation 734. General Foods' established cost accounting system assigned any expense or revenue that could be directly identified with a particular product or brand directly to that product or brand. Residuals were allocated based on formulas prescribed throughout the corporation. Within divisions that contained homogeneous units the allocations were based on unit sales. Where not homogeneous the residual was normally allocated on the basis of a fifty-fifty average of unit sales and gross dollar sales (Tr. 16 277, 16 333 020).

735. The figures used were taken from the basic accounting records of General Foods, used by it for operations and decisionmakiog purposes in the regular course of its business. The basic data was not prepared for purposes of this litigation. Existing records were utilized and methodologies used by General Foods in the ordinary course of its business were employed (Tr. 16 230 016-17). 736. Under the previous discussion of criteria for an allocation system, General Foods' method certainly qualifies as an acceptable method and there is no reason to believe that any alternative method would be preferable (Tr. 37 924- , 38 016-17). 737. Not only does the presumption of preferabilty for a company s own method of allocation apply, as stated in the Bureau of Economics' staff memorandum accompanying the Line of Business reporting form, but General Foods' method was one used in the regular course of business and so is entitled to the presumptions normally accorded such business procedures. 738. General Foods' cost accounting system has already been the subject of hearings In the Matter of General Foods Corporation Docket No. 9085. In that case, Professor John Dearden, a professor of accounting at the Harvard Business School, was called as an expert witness by complaint counsel (Tr. 38 021-22). In analyzing profit (20S)calculations, Professor Dearden reviewed General Foods ' ac- Initial Decision counting and financial manuals and concluded that General Foods has an excellent cost accounting system and that there was nothing he would, as a cost accountant, want to change (Tr. 38 023-24; GFX 1332Z-1O, Z-11, Z-131).

739. Professor Dearden s opinion was that General Foods' allocation procedures, while a matter of judgment, were " not capricious that they constitute a good and sound way of making allocations; and that there is no better way (Tr. 38 025 26; GFX 1332Z-220, Z-221). 740. Thus, General Foods' cost accounting system has been presented by the Commission s staff as highly reliable in the other case involving General Foods. Complaint counsel cannot merely turn their backs on this position.

741. Complaint counsel (CRPF 11- , 11-29) agrue that Professor Dearden s testimony in Docket 9085 should be limited to General Foods' Maxwell House Division which was directly involved in that case; that "(i)f Professor Dearden had testified in the RTE cereal case, he might have explained that there are differences in the situations that make his testimony in the 'Coffee' case inapplicable to the RTE cereal case." An examination of Professor Dearden testimony, however, indicates that his commendation of General Mills' cost accounting system went to the system as a whole and was not inapplicable to RTE cereals. If complaint counsel were of the opinion that Professor Dearden would have limited and distinguished his testimony had he testified in this case, it was complaint counsel's obligation to have called him as a rebuttal witness. They did not.

Respondents assert (GFPF 5- , 5- , 5-85 thru 5-128; KPF 5- 93; GMPF 3-51) that Dr. Mellman has made a number of errors in applying his allocation formula. These assertions include alleged utilization of allocation ratios when direct assignments were possible, utilization of improper and inconsistent ratios, improper "grossing-up" from segment figures to corporate figures, improper and unsupported assumptions and failure to make certain adjustments. Complaint counsel (CRPF 11-38 thru 11-42) defend Dr. Mellman grossing-up procedure" as reasonable, and otherwise purport to justify Dr. Mellman s procedures. Apparent errors are asserted to be insignificant (CRPF 11-43 thru 11-59).

Inasmuch as respondents are entitled to rely on the more favorable allocation procedures utilized by General Mills and General Foods, it would serve no purpose to resolve the numerous issues raised regarding alleged errors in application of Dr. Mellman methodology. Complaint counsel make no claim that General Mills Initial Decision 99 F. or General Foods' computations are not in accord with their respective methodologies or contain any errors. 742. In summary, each of the allocation procedures proposed in this case provides a rough approximation of how joint costs ought to (209)be assigned to the RTE segments of General Mils' and General Foods' overall businesses. And each is subject to the possibility of a degree of error. Each method falls within accepted accounting procedures and, in the absence of a showing that the methodology advocated by complaint counsel is superior for purposes of this case to those followed by General Mills and General Foods, the parties are entitled to rely upon the procedures most favorable to them. This is particularly true in the case of General Foods where the procedure has been followed in the normal course of its business. (b) Cash Flow Profile 743. A second critical element of Dr. Stauffer s formula is the cash flow profie. The formula relates the investment in plant property and equipment to the associated cash flow generated by that investment. Cash flow equals the gross receipts that are associated with a product produced using the plant, property and equipment. From such receipts, costs such as labor, raw material and supplies are subtracted. Thus, cash flow is the net money flowing in and does not include deductions for depreciation, research and development, or advertising which reflect outlays of cash for investment, not deductions from return on investment. Net income would equal cash flow minus these additional items and taxes (Tr. 354, 19 394, 37 499-500).

744. As previously explained (Findings 702-03), a rectangular cash flow profile assumes that the cash flow produced by an asset is the same for each year of the life of the asset; and a triangular cash flow profile assumes that the cash flow produced by an asset starts at a high level and declines steadily over the life of the asset. Inasmuch as cash flow data is not readily available, the profile of the flow had to be assumed. In the absence of actual knowledge, Dr. Stauffer applied his formula utilizing two alternative profie assumptions (rectangular and triangular) in order to establish lower and upper bounds of economic return. Dr. Stauffer made no effort 1,0 14 Of course, it would be preferable have G..nera! Mills' and Genewl fo'oods' joint costs allocated by the Same method, which is not the case, except for complaint counsel's cost method of allocation. But, inasmuch as the cost method may not be imposed, we are left with General Mills' and General Foos' joint cosls being allocated by diffcrentmethods " Kellogg s witness, Dr. Solomon, agreed that, in the absence of actual data, it is appropriate tD work with two outside extreme assumptions with the further asumption that the correct answer lies somewhere in between (1'r 216).

Initial Decision determine the true profile. The two ' profiles used were merely assumptions to establish bounds within (210)which the true profile would lie. Dr. Stauffer was of the opinion that in the real world, the result would lie somewhere between the rectangular and triangular extremes, but that he could not be more specific than that (Tr. 386-87 395- 074).

745. The economic rate of return calculations may be affected significantly by the choice of cash flow profile (Tr. 19 394, 38 186). Each respondent has a substantially higher economic rate of return under Dr. Stauffer s formula when a triangular cash now profile is used than when a rectangular cash now profile is used (Tr. 38 051; CX 701Z-21, Z-22; GFX 1368).

746. Under circumstances where the rectangular cash flow profile or lower bound is a distinct possibility, I am precluded from applying against respondents any cash flow profile assumption which would result in the computation of higher economic rates of return. As detailed below, complaint counsel have failed to establish that the rectangular cash flow profie may not be a reasonable approximation of the cash flow experienced by respondents and others in the RTE cereal industry. Therefore, I shall not consider and apply against respondents any Stauffer type computation of economic rate of return which does not embody a rectangular cash flow profie.

747. Complaint counsel assert (CPF 11-84 thru 11-87; CRPF 11- 103) that a declining cash now profile "close to the triangular " isprofile" or "closer to the triangular than the rectangular profile the appropriate cash flow profie to use in computing the RTE cereal companies' rates of return. This position, however, does not indicate just how close to the triangular the true profile is alleged to be. Complaint counsel ground their position on the general assumptions (1) that as assets age, the costs of maintenance increase and there are longer periods of downtimes, (2) as assets age, they tend to become obsolete as newer and more efficient equipment becomes available and older assets are placed on partial use or standby status, and (3) since a firm has a variety of assets with differing lifetimes, depiction of each asset by a box, with the shortest box on top followed by each longer box below with the longest box on the bottom, would result in a triangular diagram.

748. The cash now profile varies from industry to industry and individual industry assessments are required to attribute specific profies to particular industries. (Tr. 20 000, 20 029, 31 207-D9 215-21; CX 703A). Complaint counsel have made no such assessment for the RTE cereal industry, but rely primarily on the Initial Decision 99 F. testimony of Dr. Stauffer. Dr. Stauffer, however, had made no study of the cash flow in the RTE cereal industry and had no specific familiarity with the industry (Tr. 19 051 , 19 993). In any event, Dr. Stauffer did not testify that the true profile would be "close" to the triangular profile, only that, in his opinion, the actual profile would be closer to the triangular rather than the rectangular profile (Tr. 592). And this opinion was based on his intuition, not on any study of the industry (Tr. 19 395, 19 592). Dr. Stauffer s intuitive expectations were based upon the same general considerations asserted by complaint counsel-that output would (211)decline as assets aged and became obsolete and that there would be increased maintenance costs (Tr. 19 395 954-55). 749. However, there is no record evidence to support Dr. Stauffer s theory; and the record evidence tends to refute it. 750. The RTE cereal industry involves sufficiently complex processes that productivity increases over time and costs decrease due to the learning curve phenomenon. People learn to do their jobs better and engineering personnel find and remove bottlenecks (Tr. 963..5 , 27 235- , 27 289, 28 147 , 31 252, 33 018-19, 36 664 064..5, 38 483-84). For example, newer plants of Ralston had lower yields and greater production problems than older plants because of lesser personnel experience (Tr. 10 693, 10 723, 10 742-43 755- , 17 561); and Kellogg upped its production capability with respect to Product 19 by overcoming various production problems over time and because of the increase in efficiency over time of its personnel (Tr. 29 341- , 29 364).

751. Maintenance costs for General Foods are highest during the start-up period of a new product manufacturing line. Thereafter through an extensive preventive maintenance program, maintenance costs are lowered over the life of a project. The number of unscheduled break-downs are reduced and the durability of replacement parts is increased so that periods between planned maintenance shutdowns are lengthened err. 36 677-87). Kellogg s equipment is also durable and requires little in the way of maintenance (Tr. 29 439). Its preventive maintenance program decreases the number of breakdowns and smooths out operations over time (Tr. 670- 752. General685).Foods is often able to increase productivity and decrease unit production costs over the life of its assets. The components of a production process normally have differing capacities. By finding ways, usually with little or no additional investment to increase the capacity of the limiting component, the capacity of the entire line can be increased. Production is also increased and Initial Decision costs decreased through direct application of sophisticated engineering techniques (Tr. 36 665-70).

753. General Foods' production equipment is long lived. For example, ovens installed in 1927 to bake Grape-Nuts are still being used effciently today, in part because of the regular upgrading of equipment that accompanies preventive maintenance (Tr. 36 686- 87). Rather than finding that its older equipment was obsolete General Foods found, in planning its Modesto facility, that duplicating its Battle Creek equipment would provide it with the most current technology (Tr. 36 690-91).

754. In any event, obsolescence is primarily a matter of longevity of assets, another aspect of Dr. Stauffer s formula. When an asset is retired because of obsolescence, this does not impact the cash flow during the life of the asset. (212) 755. To the extent that sales increase over time, cash flow would increase (Tr. 31 209). In an article in Drug Development and Marketing, published in 1973 or 1974, entitled "Profitability Measures in the Pharmaceutical Industry," Dr. Stauffer described the process of choosing an appropriate profile for use in computing the profitability of the pharmaceutical industry. Dr. Stauffer plotted the gross sales of successful pharmaceutical products and observed a product life-cycle profile with sales rising, then level and finally declining (Tr. 38 098-99).

756. It was Dr. Stauffer s opinion, expressed in this published work, that this product life cycle profile could produce an economic rate of return lower than that obtained by using a rectangular profile. Dr. Stauffer used a rectangular profile for his analysis of the pharmaceutical industry, explicitly recognizing that its use could overstate the economic rate of return (Tr. 19 990- 38,099). 757. Products accounting for over 50% of total RTE sales of General Foods and Kellogg enjoyed increased sales over time (Tr. 246-51 , 38 101-04; GFX 1360). Productivity of equipment is increased as demand increases (Tr. 36 662-64). As in the pharmaceutical industry, this would indicate that the cash flow profile for the RTE cereal industry was at least rectangular, if not rising ramp, which would result in an even lower accounting rate of return (Tr. 246-7 099 105-8).

758. To the extent cash flow would have declined due to any rising costs, it is to be expected that respondents would have raised prices. Further, the extent of inflation over the period would have moved even a triangular cash flow profile toward the rectangular (Tr. 31 233- , 38 482).

759. The purpose of Dr. Stauffer s formula is to correct the extent Initial Dccision 99 F. to which the accounting rate of return overstates the economic rate of return (Tr. 19 360, 19 987). Yet, when a triangular cash flow profile is used in the formula with a rapid decay rate of advertising,76 the result is an economic rate of return higher than the accounting rate of return (Tr. 38 081-83; GFX 1358). Further, with a triangular cash flow profile in Dr. Stauffer s formula, a firm would continue to invest heavily in advertising despite a negative cash flow over the last several years of the life of the investment (Tr. 34 782- , 34 79G-91 , 34 864-5, 37 782-87, 38 092-97; GFX 1357; GMX 456). One would not expect a firm to continue to produce and heavily advertise an RTE cereal product for years while it continued to lose money.

760. As previously noted, complaint counsel contend (CPF 11-87) that, since the plotting on top of each other of blocks of varying (213) lengths depicting the different length of lives of different assets forms a declining triangular shape, the overall profile for the company must be triangular. This contention has no merit. What complaint counsel are arguing, contrary to the basic testimony of Dr. Stauffer, his original dissertation and his pharmaceutical study, and the testimony of all of the other experts who testified in this case that there is no such thing as a rectangular cash flow profie. For every firm has assets of varying longevity and every firm s assets could be depicted in the triangular fashion relied upon by complaint counsel.

761. The manner in which Dr. Stauffer defines cash flow profile for use in his formula has nothing to do with the triangular depiction of different lived assets. If the individual underlying assets produce rectangular cash flo.ws, it is immaterial how you stack one on top of the other by depicting boxes so that they come out in a triangular fashion. So long as the individual underlying assets produce rectangular cash flows, Dr. Stauffer s formula must use a rectangular cash flow profile (Tr. 35 146-7).

762. The overall economic rate of return for a firm is the result of the returns earned by each of its investments. If each investment (whether the same or of unequal lie) earns at a rate of 5%, the company overall earns at a rate of 5%. If each investment earns its 5% through a rectangular cash flow, a rectangular cash flow profile is the appropriate profile to characterize the firm as a whole (Tr. 146-7).

763. To demonstrate this point, Dr. Troxel created a hypothetical firm with investments of different lives. Each was given a rectangu- ,. This relates to the advertising correction which wi!! be considered below , Initial Decision lar cash flow profile resulting in an economic rate of return of 10%. Thus, the company in total would show a 10% economic rate of return. The data on the firm were inserted into Dr. Stauffer s model using, alternatively, a rectangular and triangular cash flow assumption. Only the formula using the rectangular assumption produced the known correct answer of 10%. The formula using the triangular assumption computed a return of 12.3% (Tr. 35 127-49; GMX 463). 764. As Dr. Fox has explained, if a project has necessary components of differing lives, the entire process would stop and cash flow would fall to zero as soon as the shortest-lived asset gave out. The project would then have a rectangular profile with a service life equal to that of the shortest-lived asset. If, instead of allowing the project to die, the shorter-lived assets were replaced over the life of the longest-lived asset then, graphically, there would be a series of short boxes on top, a series of longer boxes on the next level and so forth; and the sum of the profiles would be rectangular (Tr. 38 075- 77).

765. Complaint counsel, in their March 1976 trial brief and in their July 1976 revised trial brief, took the position that the appropriate cash flow profile in this case was rectangular (GFX 21 (214)22A-C). This obviously reflected the position at that time of complaint counsel's expert in the matter, Dr. Stauffer. At his April 1976 disposition, Dr. Stauffer took the same position. He said decided early on to use a constant (rectangular) cash flow" instead of other profies (Tr. 25 088-89, 25 231); that "there are sound reasons for using the rectangular profile" (Tr. 25 091). He explained that the real" profie (CX 702D) was most likely to have a product life cycle shape e., one which rises gradually, plateaus and then declines gradually (Tr. 19 395- , 25 114). He had run an "extensive series of simulations" and had found that the estimates using the product life cycle profile were ((very close to the result obtained using a rectangular profile" (Tr. 25 095-98). He rejected the triangular profile. The assumption associated with a triangular cash flow profile that cash flow "rises rapidly" and then "dwindles away quickly" was not "empirically accurate" (Tr. 25 230-31). In estimating economic profits for the drug industry, Dr. Stauffer had used the rectangular profile assumption as a proxy for the "real product life cycle profile" (Tr. 19 988-90). He concluded that "the rectangular profie is, broadly speaking, appropriate" (Tr. 25 096). He also testified at the April 1976 deposition that his work in computing economic rates of return was completed (Tr. 25227). 766. I read Dr. Stauffer s deposition testimony, given in April 1976, as unequivocally expressing his opinion that the rectangular Initial Decision 99 F. cash flow profile is the appropriate one for use in this case. Having read his testimony in hearings in October 1977, I fail to find any reasonable basis for accepting his later opinion in lieu of that previously given. Dr. Stauffer s original considered opinion as to the appropriateness of the rectangular cash flow profile is joined in by economic experts, Dr. Fox and Dr. Markham (Tr. 38 108, 38,482 484).

767. While respondents have advanced arguments in support of using a rising ramp profile (KPF 5-112, 5-113, 5-128; GMPF 3-68; GFPF 5-224), which would result in lower economic rates of return than by using the rectangular, respondents' uniform position is that a rectangular cash flow profile is appropriate (KPF 5-114, 5-127, 5- 128; GFPF 5-153, 5-160; GMPF 3- , 3-76). On the basis of all of the evidence recited above, it is concluded that complaint counsel have failed to establish that any cash flow profie less favorable to respondents' position than a rectangular one may be utilized. (c) The Advertising Adjustment 768. A third critical input into Dr. Stauffer s formula is the advertising decay rate. As previously explained, accountants treat all advertising as current expense even though it may benefit future years' incomes. To the extent that advertising generates sales in future years, an adjustment must be made in order to arrive at the true or economic rate of return. The rate at which past advertising continues to benefit sales is its persistence. Correspondingly, the (215)rate at which its benefit falls off is termed the decay rate. Inasmuch as advertising expenditures are very high in the RTE cereal industry, the expensing of advertising substantially overstates the true or economic rate of return and the advertising adjustment is very important.

769. As part of his overall formula for computing economic rates of return, Dr. Stauffer incorporated what is known as the Koyck model to account for how the effects of advertising wear off over time (Tr. 19,415 169-70).

770. The Koyck model assumes that the effects of advertising wear off at the same percentage rate each year. This percentage is known as the advertising decay rate (Tr. 19 416-18, 22 310-11; CX 702F). A decay rate of 80%, for example, means that if advertising generates $100 of sales in year one, the sales generated by the original advertising will have worn off or decayed to $20 in year two " Th.. decay rate can beviewed as a proxy for dOHay in the intangible asset, goowill. Advertising is viewed as an investment in goowill (Tr. 22 :\1;J) Initial Decision and the sales will once again be 80% lower in year three than in year two, or $4.

771. The Koyck model has been used extensively in the economic literature to represent the relationship between advertising and sales. Virtually all studies that have adjusted profitability to take account of the advertising bias have employed the Koyck model (Tr. 415, 22 170-71). Dr. Stauffer obviously used the Koyck model for lack of anything else, for his opinion of the model, previously stated in his pharmaceutical study and reaffirmed at this hearing, is as follows (Tr. 19 988-89 989-90):

The assumed functional relation between sales and advertising, while plausible in its general features, is clearly grossly oversimplified if not arbitrary, although it has been widely used in other economic analyses. That choice has been inciuded faute de mieux in the detailed mathematical model described in the appendix with no allusion as to its accuracy. The contribution of advertising to the rate of return discrepancies for the (pharmaceutical) cases analyzed here is modest. So the DvcraJI conclusions are not affected by the imprecision in the economic description of advertising It is useful to include even thi.c. crude model for sales response to advertising in order to better ilustrate the direction of the necessary corrections, but it is important to note the caveat that in other applications where advertising places a mure (216)important role in the financial results, the estimates for the real rates of return might prove quite sensitive to the particular specification of the advertisinff sales response function (emphasis supplied).

Here, of course, advertising plays a most important role in the financial results.

772. While widely used in economic jiterature and studies, it is highly questionable whether the Koyck model provides a sufficient degree of precision to permit an adjudication of facts in a litigated matter where the proposed remedy is to require the respondents to divest themselves of assets and offer royalty-free trademark licenses of their remaining products. Nevertheless, since the Koyck model is an element in converting accounting rates of return into more accurate economic rates of return, I shall continue to consider Dr. Stauffer s conversion formula, including the Koyck model. 773. Because of the admitted imprecision of the Koyck model however, which in large part flows from the difficulty in selecting an appropriate decay rate, I am compelled to accept the lowest rate (the one most favorable to respondents' position) which appears to be possibly correct.

774. While Dr. Stauffer used his own formula, which incorporated the Koyck model, he used alternative advertising decay rates of 35% and 800/0 which were estimated as lower and upper bounds by .- Dr, Butters har described t.he Koyck mode! as ext.remely rough and imperfectITr .'7636) , , pp. , Initial Decision 99 F. Dr. Schmalensee and Mr. Glassman (Tr. 19 420; CX 701Z-21-Z-25). Dr. Schmalensee had the primary responsibility to develop decay rates for purposes of presenting complaint counsel's case. Mr. Glassman relied upon the same data Dr. Schmal en see relied upon in reaching similar conclusions.

775. Dr. Schmalensee and Mr. Glassman relied in part upon several published economic studies for their estimates of the advertising decay rate for RTE cereals (Tr. 22 172- , 26 135-38). Professors Parsons and Bass studied the effects of advertising on sales (said by Dr. Schmalensee to be Kellogg s Corn Flakes and Sugar Frosted Flakes). Their analysis indicates that they believed the advertising decay rate for these brands to approximate 80% (Tr. 173 135-36). Professor Darrel Clark, in an article which (217) surveyed more than 70 studies of the decay rate of advertising, concluded that the decay rate for advertising is very rapid, approximately 90% per year (Tr. 22 176-77, 26 137- , 37 787). Professor Ayanian estimated that the advertising decay rate for RTE cereals was 350/0.

776. Dr. Schmalensee testified (Tr. 22 176) that the Parsons and Bass study appeared to be a sound piece of work; that it used modern econometric figures with monthly data and, on the basis of methodology, was superior to the Ayanian paper. However, neither Dr. Schmalensee nor any other witness called by complaint counsel described the methodology used by any of the studies relied upon. No indication was given as to the supposed preciseness of the estimates. 777. The only witness who testified as to the methodology utiized in the studies relied upon by Dr. Schmalensee and Mr. Glassman was Dr. Butters. Dr. Butters testified that, whereas the Koyck model purports to utilize a decay rate on total sales, the literature relied upon by Dr. Schmalensee and' Mr. Glassman does not reflect calculations of decay rates on total sales, but only on marginal sales (Tr. 37 652- 663-64). 778. The literature dealt with on-going products and thus was based on data involving relatively small changes in advertising from one period to the next, as opposed to data involving cessation or drastic reductions in advertising. Since the analyses included in the articles showed only the sales effect of small or "marginal" changes " Frank M. Ha.% and I..onard .J. Par ons Simultaneous-Equation Regression Analysis of Sales and Advertising. Applied &onomics 1969. pp 103-24 "" DaneJ Clark &onometric Mea.urement of the DLJration of Advertisinli Effect on SaJ"3 Journal or Morkeli'I;;Reseorch Novernbir1976 345--57. ., Robut Ayanian Advertisinr; and Raws of Return Journal or Law and ECUfwmics October 1975, pp. 479- Initial Decision in advertising, the decay rate literature analyzed only the impact on marginal sales (Tr. 37 642-43).

779. Since marginal sales (the increase in sales associated with a modest increase in adverising over a pre-existing level) would be, to a large degree, the result of "whim" purchases, rather than from the core of loyal users, such sales could be expected to decline more rapidly once advertising returned to its pre-existing level. It is more likely that marginal units of advertising would attract additional sales which, when the marginal units of advertising are withdrawn would decay more rapidly than the rate for all sales (Tr. 37 647 664).

780. The decay rate literature, which measures only the decay rate on marginal sales, is not useful for estimating a decay rate on total sales. A biased estimate results from this literature, because for use in Dr. Stauffer s advertising term, it is not (218)measuring the right thing (Tr. 37 642-65). The measurement of the decay rate on marginal sales fails to measure the level of continuation of all sales if all advertising were stopped.

781. Complaint counsel assert (CRPF 11-147, 11-148) that the studies in question purported to measure the total decay rate on advertising, not marginal rates. Irrespective of what was purported to have been studied, Dr. Butters was the only witness who described what in fact was considered. And his testimony, which is unrebutted clearly establishes that the studies were limited to analyses of marginal decay rates.

Dr. 782. In addition to the economic articles on decay rates, Schmalensee and Mr. Glassman relied upon a study conducted by Dr. Schmalensee of RTE cereal brands which had had all advertising support withdrawn. Initially, Dr. Schmalensee calculated decay rates for seven products for which advertising was withdrawn and the products were subsequently withdrawn from the market prior to December 1972, the end of the time period he examined. Subsequently, Dr. Schmalensee added four products to his study for which advertising had been cut off, but which had not been withdrawn from the market as of December 1972 (Tr. 37 666-7; GFX 1329 1330). Dr. Butters located and caused decay rates to be calculated for an additional five products which were still in national distribution as of December 1972 (Tr. 37 676-3; GFX 1329). 783. The mean decay rates expressed in percentages for the 16 products so selected are as follows:

Initial Decision 99 F. Products in National Products Withdrawn From Distribution Distribution Puffa Puffa Rice Puffed Corn Flakes Concentrate - 1 Corn Flakes w/Strawberries Pep Corn Flakes w/Blueberries Krumbles - 4 Bran & Prune Flakes Grape Nuts Flakes - 3 Clackers Krispy Critters (All) Stars Treat Pak - 2 OK' Muffets Sugar Jets 784. Two facts immediately stand out: (1) There is a tremendous variation in decay rates among products; and (2) Products which were withdrawn from national distribution prior to December 1972 generally had a much higher decay rate than those which remained in distribution after that date.

785. Among products remaining in national distribution, the variation in decay rates ranged from -2% to 420/0; and for products withdrawn from distribution, the decay rates ranged from 23% to 76%. (219) 786. The negative decay rates for four of the products are plausible because market growth or a shift in consumer tastes can lead to increased sales even without advertising. Also, sales of some brands may be influenced by advertising for related brands. For example, sales of Grape Nuts Flakes may be affected by advertising for Grape Nuts (Tr. 37 685-86).

787. Brands differ in decay rates depending upon a number of factors, including the quality of the product. Other factors could include the length of time the product had been advertised, the amount of advertising of competing products, as well as institutional advertising or overall RTE cereal advertising by the company owning the brand in question. There is no single best way to arrive at a single number reflecting the various differing figures (Tr. 23 307- , 26 574- , 37 637-39).

788. The first seven brands selected by Dr. Schmalensee were withdrawn from the market by 1972. This indicates that they were failures in the marketplace and that their sales decay was not so much related to the withdrawal of advertising as to the fact they were failing brands. Also, a large part of the decay may have been due to a loss in distribution, not to a decision of consumers not to buy the product (Tr. 37 674-75). The other nine brands, while having greater consumer acceptance, as reflected by the fact they were not withdrawn from the market, were still not typically successful , Initial Decision brands. Otherwise, advertising support would not have been withdrawn (Tr. 37 676).

789. The large majority of advertising is associated with RTE cereals which have been successful and have been nationally advertised and consumed for 10, 20, or 30 or more years (GFX 1319). The 11 products studied by Dr. Schmalensce, by and large, had been nationally advertised for much shorter periods, but the decay rates were lower for those products which had been advertised over a longer period of time (Schmalansee Revised Estimates of Annual Advertising Decay Rates Based on Sales of RTE Cereal Brands with Advertising Withdrawn " Paper Presented for the FTC Bureau of Economics, April 1977, p. 4 (Table 2)) (GFX 1318; KX 25; CX- 680C). So again, the products studied were not those typically associated with prolonged advertising. However, for brands on the market for a considerable period of time for which advertising support had been withdrawn, the decay rates were low, ranging from 005 to .14 (Tr. 23 310-11).

790. Of the nine brands which had greater consumer acceptance and less of a failing product bias, as reflected by the fact they were kept in national distribution, seven had advertising decay rates below 10%, and all but one rate were substantially below that figure. 791. The evidence, therefore, does not support Dr. Schmalensee and Mr. Glassman s conclusions (Tr. 22 171- , 22 177- , 22 316- , 26,136-38; GFX 1329, 1330) that the upper and lower bounds for the decay rate were 80% and 35% , respectively, and more likely to be closer to 80% than 35%. On the other hand, Dr. Butters' estimate (Tr. 37 687) that the decay rate could be as low as 10% is not (220) unreasonable in light of the studies made of individual RTE cereal products where advertising was discontinued, but the products remained in distribution beyond 1972. As explained above, I am required not to apply any decay rate higher than the lowest rate which appears to have a reasonable possibility of being correct. Extent To Which Rates Of Return Have Been Established 792. Applying Dr. Stauffer s formula and utiizing a rectangular profile and a 10% decay rate for each of the respondents, together with General Mils' and General Foods ' methods of allocating joint accounting data to their RTE cereal business segments, the following economic rates of return are arrived at:

Initial Decision 99 F. Kellogg General Mils General Foods 1958-1970 11.7 10. 1954-1972 13.2 10. (ex 701Z-22; GMX 466; GFX 1354) 793. I am not finding that the above figures constitute respondents' economic rates of return, but simply that there is a lack of proof of any higher figures.

794. Respondents assert that the rates of return are stil overstated by reason of alleged omissions or errors in Dr. Stauffer calculations."' The alleged errors are either not established or are offset by other aspects of the formula, and are minimal in nature particularly in com parisian to the three major adjustments I have just ruled upon. The figures stated above, therefore, are sufficiently (221)accurate upon which to proceed in the appraisal of respondents profis.

795. Quaker s accounting rate of return for the period 1958-1970 was 9% (CX 701A)." The purpose of Dr. Stauffer s formula is to adjust for those items which improperly are expensed, such as advertising and research and development, and to otherwise adjust for the time sequence of returns on investment in order to reduce the overstated accounting rates of return. As a company which was expanding with new products, Quaker obviously had large research and development and advertising costs which necessitated a downward adjustment. Indeed, Quaker s advertising to sales ratios were consistently higher than Kellogg s and were usually higher than General Mills' and General Foods ' (CX 513). Nevertheless, applying Dr. Stauffer s formula with a rectangular cash flow profie and a 10% advertising decay rate, Quaker s 1958-19709% accounting rate of return did not go down but increased to a 12.4% economic rate of return. Normally, the higher the advertising decay rate, the higher the resultant economic return. In the case of Quaker, however, just the opposite result occurred. Using a 350/0 decay rate, its economic rate of return went down to 10%, and with an 80% decay rate, its economic rate of return went stil lower to 8% (CX 701Z-22). ., All three respondents contend that Dr.Stauffer erred with respect to the timjn of tax benefits connette with the expensing of advertising and research and development (KPF 5-102; GMPfo' 3-102; GJo' 5-334 thru 5-39). General Mills and General Foos assert that additional downward adjustments should have ben made by reason of the quas;-(apital nature of such items as promotion, m..rket research and certain selling expenses, as well as product quality control, preventive maintenance and personnel training (GMPF 3-97 thru 3-. 101; GFPF 5-324 thru 5-33). General Foos asserts that Dr. Stauffer has underestimated the service life of iL investment PF 3-104 105) and has incorrectly adjuste perceived differ"m:es between tax depreciation lifetime!) and bok depreciation lifetimes(GFPF 3-106, 3- 107).

., Kellogg agrees (KPF 5-74) that Quaker has submitted all accounting data needed to calculate rates of return Initial Decision 796. As discussed above, utilization of a triangular cash flow profie consistently results in a higher accounting rate of return than when a rectangular profile is used. In the case of Quaker however, just the opposite occurred. Using a triangular cash flow profie, Quaker s economic rate of return was 10.2% at the 10% decay rate, .89% at the 35% decay rate and minus 34.4% at the 80% decay rate (CX 701Z-22).

797. It has been suggested that the unusual results with respect to Quaker flow from the fact that its rate of growth exceeded its rate of return (CPF 11-123; CRPF 11-249; Tr. 31 259, 38 089-90). Whatever the reason, it is clear that Dr. Stauffer s formula is inappropriate to measure Quaker s economic rate of return (Tr. 259, 38 089-92), and there is no record evidence of that rate. It may be assumed, however, that it was well below its 9% accounting rate of return.

798. Dr. Stauffer, on the basis of accounting data submitted by Ralston, made certain allocations and assumptions (Tr. 19 156-6 670-76) and, using a rectangular cash flow profile and an advertising decay rate of 10%, arrived at an economic rate of return of9.9% for 1958-1970 and 10.7% for 1954-1972 (CX 701Z-22). 799. The data submitted by Ralston differed from those supplied by the respondents and Quaker and were somewhat less comprehensive (222)(Tr. 19 157). It was necessary, in order to approximate RTE cereal figures, to assume that the relationships between various accounting data for the RTE cereal segment were comparable to such data for the total Ralston corporation (Tr. 19 664, 25 175-80). In 1970, about 1 X% of Ralston s total net sales were of RTE cereals (Complaint Counsel's Trial Brief, Vol. 1 111.21). In light of the small portion of overall corporate endeavor accounted for by RTE cereals, the assumption could well be in error. The significance of any such error is conjectural. In any event, Ralston s rate of return is not substantially higher than the benchmark or normal rate of return contended for by complaint counsel 84 No calculations were made for Nabisco as no asset data were secured from that company.

The Benchmark Comparison 800. Having determined the extent to which rates of return have been established for respondents and others in the industry, it becomes necessary to evaluate those rates to see if they are above ,. This subject is considered next Initial Decision 99 F. normal, or excessive. There are two definitions of a "normal" rate of return. They are consistent with each other. (1) Economists often define a normal rate of return as that rate which is necessary attract appropriate capital into an industry (Tr. 21 706, 31 274- 356, 38 164). (2) A normal rate of return is also defined as one that is necessary to cover all costs of production and distribution including a "normal" or competitive return to capital invested (Tr. 692 707 991-92).

801. Rates of return significantly above a reasonable, competitive benchmark rate of return may indicate poor economic performance (Tr. 21 698-700, 21 719- , 26 119-20). Complaint counsel rely upon the average rate of return earned by finns in the manufacturing sector of the United States economy as an estimate of the normal or competitive rate of return. The data relied upon to calculate this average rate of return for all manufacturing was compiled by the United States Internal Revenue Service ("IRS") in the "Statistics of Income the IRS annual summary volume covering income tax fiings by all United States firms classified by the IRS in the manufacturing sector of the United States economy (Tr. 19 074-75). (223) 802. The average accounting rate of return on capital employed for all firms in the manufacturing sector was 8.9% (CX 701A). Using his formula, Dr. Stauffer converted the accounting rate of return to an economic rate (Tr. 19 683 et seq. 720-21). There was no great problem of whether to use a rectangular cash flow profie or a triangular cash flow profile, or as to the appropriate advertising decay rate. This is because the benchmark figures overall do not include relatively large expenditures for research and development and advertising, the two principal bases for correction from an accounting to an economic rate of return (Tr. 25 099, 25 438-39). Therefore, there was not a very large correction to be made. 803. Thus, for the period 1958-1970, the benchmark economic rate of return was 8.3% using a rectangular cash flow profile irrespective of the decay rate. For the same period, the rate was 8.4% using a triangular cash flow profile, irrespective of the decay rate. For the period 1954-1972, using the rectangular cash flow profile, the rate of return was 8.2% with an 80% or 35% decay rate, and 1 % with a 10% decay rate. And for this longer period, using a triangular cash flow profile, the rate of return was 8.4% with an 80% or 35% decay rate, and 8.3% with a 10% decay rate (CX 701Z- "" The His data set is the largest and most cornpmhensive sOLirce of sw:h data available Hnd hab been used frequently by industri"luq;aniwtion e.:notnisL studying indw;try p"rfo'mance (Tr. 19 079-0). ., ACl"untingd"tB had air""dy bencorrected with respect to "ccelerateddepreeilltion(Tr. 25 438) Initial Dccision , Z-22). Inasmuch as I have used a rectangular profile and a 10% decay rate in reaching respondents' and Ralston s economic rates of return, I shall use the same assumptions for the benchmark. This gives us the following comparisons:

General General Kello Mils Foods Ralston Benchmark 1958-1970 11.7 10.

1954-1972 13.2 10. 804. The respondents and Ralston are large, established firms which have shown profits over a number of years. The benchmark on the other hand, reflects the activities of 200 000 manufacturing corporations, 40% of which have either no net income or negative income (Tr. 31 275, 31 334, 38 167--8). Based on 1970 figures, when firms which earned no income are excluded from the benchmark, the benchmark rate of return is increased by two points (Tr. 24 922- 345; GMX 78). Respondents' profis should not be evaluated on the basis of firms which had no income or negative income Crr. 38 489- 90). Therefore, a modification of the benchmark upward by two points is appropriate. This gives us the following: (224) General General Kellogg Mils Foods Ralston Benchmark 1958- 1 970 11. 10. 10. 1954-1972 13. 10. 6.7 10. 10. 805. There are other biases in the benchmark which cannot be measured. For example, the sample is constantly changing. Many firms go bankrupt each year, while many are just starting up (Tr. 278, 38 165--6, 38,489-90). In 1954, there were 120 896 firms in the sample. In 1970, there were 197 807 (Tr. 38 165). New firms frequently have initial periods of very low rates of return or even losses until they get established. This biases the sample downward when it is being used in comparison with established companies whose start-up cost were incurred in the distant past (Tr. 38 166 488-90). In Finding 804, I have adjusted the benchmark to exclude firms with no income. This still leaves firms with low rates of return attributable to their just starting up.

806. The large number of small firms in the IRS sample biases the benchmark average downward because the owners of such firms have a tax incentive to take the firms' profits in the form of high salaries. This results in an understatement of their profitabiliy Initial Decision 99 F. relative to large, publicly-held corporations (Tr. 31 278- , 38 166- , 38 489-90). Some 80% of the IRS sample is composed of firms with total assets of under a half million dollars (Tr. 38 166-67). 807. When firms acquire others, they normally write up the assets acquired to reflect values at the time of acquisition (Tr. 314- , 22 302, 24 914). Since the assets of the respondents have not been restated to reflect current values, the benchmark understated to some extent in comparison with respondents. 808. On the other hand, the IRS data biases the benchmark upward to the extent it includes the return to those firms that are earning monopoly profits (Tr. 21 796, 21 798-99, 26 104, 27 990). There are two additional factors that must be considered before meaningful comparisons may be made between RTE cereal companies' rates of return and the benchmark. These are (1) inflation and (2) risk.

(a) Inflation 809. Dr. Stauffer made no adjustments for inflation, since he was not trying to compute absolute levels of profitability, but (225)rather to compare relative levels (Tr. 19 890-92). Thus, for both the RTE cereal companies and the benchmark, the fixed asset portions of capital investments were valued at cost at the times of purchase and were not adjusted upward to reflect current, inflated values (Tr. 314- , 22 302, 35 887). Inflation, however, has a material effect on profit measurement. The issue is whether inflation has affected the individual RTE cereal firm computations arid the IRS benchmark to the same degree (Tr. 19 891- , 21 702, 21 707, 31 214 231- 244 254 887-900, 38 482- , 38 522-24). 810. The impact of inflation varies from firm to firm (Tr. 35 888- , 35 896-900). Firms that have particularly old stocks of fixed assets wil have their accounting statements affected by inflation to a relatively greater degree (Tr. 21 707). The bias on profits due to inflation is aggravated when computing the accounting rate of return by the expensing of quasi-capital expenditures, such as exploration, research and development and advertising (Tr. 25 126- 27; KX 17, p. 13). The RTE cereal firms would appear to have older fixed assets than the average firm (CPF 12-62; Tr. 29 439). They engage in a considerable amount of research and development and clearly advertise much more than the average firm. Consequently, it must be concluded that the failure to account for inflation has ., However, Dr. Schmalense, One of complaiot counsel's economist. , has testified that any such bias would be small(Tr.2I 799).

LLU\J\J "', 1 lHJ.

Initial Decision caused a greater upward bias on the computation of profits of the RTE cereal firms than on the benchmark.

(b) Risk 811. A higher than average return is stil considered competitive or normal if the amount above the average is compensation for above.average risk. A normal rate of return for a business with above-average risk would be greater than for less risky businesses. And if an industry is risky, a potential entrant would require a higher than average return before it would consider entering (Tr. 706, 21 720- , 23 233, 26 582- , 38 486-87, 38 491- , 38 517). 812. RTE cereal companies must introduce new products in order to remain profitable and compete for market share, and competition by introduction of new products has been intense (supra Findings 530- 533).

813. As I have already found (Findings 652-53), after evaluating the high capital costs of entry, the long lead times required to recover capital investments, the costly and time consuming and often insurmountable problems of developing an acceptable product (with Procter and Gamble s experience cited as an example), the high failure rate of new products, and General Food's and Nabisco drastic decline in market shares: (226) . . A potential entrant would not normally anticipate emulating the apparent success of the leading company, Kellogg, but would be forewarned of risks by the experience of other competitors. And General Foods' growth rate and experience would indicate a degree of risk and lack of attractiveness for new entry. This observation of high risk would deter a potential entrant from entering the RTE cereal industry, unless it could foresee a rate of return higher than normal in order to compensate for the risk it would be incurring (citations omitted). I then went on (Finding 664) to note the inability of Pilsbury, Colgate, International Multifoods, Carnation and H.J. Heinz to remain in the industry, a further indication of the risky nature of the industry.

814. Risk is further reflected by General Foods' drop in rates of return from 7.7% for the 1954-1970 period to a minus .1% for the period 1966-1970 (Tr. 38 125-26);" also by Nabisco s low, but highly variable, sales return on capital. 89 .. The rate of return was calculate with a rectangular profie and a decay rate of 10%, but u oo Dr. Melman '5 allocations rather than those contended for by respondents. Year Percent ,Ot, ,Ot.

(Continued) Initial Decision 99 F. 815. The benchmark derived from IRS data for the manufacturing sector cannot be deemed to draw an absolute line of demarcation between normal and monopoly profits. While such IRS data have generally been relied upon by economists, there is concededly some degree of inexactitude (Tr. 19 930, 25 534, 25 640, 25 646, 26 233; and note the various biases enumerated above). Further, the benchmark is simply an average of firms' earnings below the average and firms earnings above the average. It cannot be said that every firm earning above average, whose rate of return contributed to reaching the average, is earning monopoly profits. Finally, while (227)not quantified, considerations of inflation and risk establish a normal rate of return for the RTE cereal respondents at a somewhat higher level than an average benchmark.

816. In consideration of the above, while the benchmark may be used as a rough guideline to aid in evaluating rates of return, a rate of return cannot be said to be monopolistic or supracompetitive unless it is substantially in excess of the benchmark. That is not the case here. General Foods' rate of return is substantially below the benchmark. Quaker s return may be assumed to be below it (Finding 796). General Mills' and Ralston s rates of return approximate it. Only Kellogg is shown to have a rate of return in excess of the benchmark; and, in consideration of the variables and biases discussed above, its rate cannot be said to be substantially in excess. 817. Even if Kellogg s rate of return were deemed to be substantially in excess of the benchmark, this would not evidence a violation under the complaint. The complaint alleges a shared monopoly and Kellogg would not be shown to have shared its allegedly monopolistic return with any other RTE cereal manufacturer. Kellogg was not charged with enjoying an individual monopoly nor could it with, at most, 44% of the market (supra Finding 168). 818. Since the benchmark average is reached by combining a variety of rates of return which range above and below it, it is not suggestive of improper industry structure to find a distribution in the RTE cereal industry of one company (Kellogg) having a rate of return above the benchmark, two (General Mills and Ralston) with Initial Decision rates of return approximating the benchmark, and two (General Foods and Quaker) with rates below the benchmark. Again, I want to make it clear that I have made no findings of what respondents' and Ralston s economic rates of return have been. I have simply found those rates above which there is a lack of proof and I have compared those rates with the benchmark. General Foods' Profitability 819. The record establishes that General Foods did not earn above-normal (supracompetitive) profits.

820. Complaint counsel (CPF 11-122), while contending that General Foods earned profits in excess of the competitive rate of return, recite that General Foods suffered poor management committed costly errors in brand introductions and possessed an inferior production plant. More specifically, complaint counsel assert that General Foods had problems with its management and advertising agency due to a lack of continuity in personnel; that it had extraordinary failures in product development, losing over $12 milion alone on just three projects from 1966 to 1970; that General Foods recognized that its failures in new product development had a (228)major negative effect on total cereal business profitability; that it operated its plant at only 60% of capacity and had the least flexible, most dated and expensive to operate production facility of the major manufacturers.

821. Other matters, not recited by complaint counsel, include the following. General Foods, since at least 1952, considered that its RTE cereal profit margin was so low that it did not merit reinvestment in the business and did not provide suffcient contribution to overall corporate results (Tr. 36 369).

822. From 1952 to 1966, while the overall RTE cereal market grew 4% per year and the adjusted gross national product grew 3. per year, General Foods' RTE cereal annual growth was only 1.1 (GFX 1321). General Foods' market share over the period 1962 to 1970 declined from 19.7% to 14.0% (CX 106A, C). Its new product activities were overall costly failures (CX-GF 4039Z-17). 823. In light of its poor performance, General Foods organized a task force in November 1966 to study its RTE cereal business. One of the assignments of the task force was to decide whether General Foods should remain in the RTE cereal business (Tr. 23 617- 404; CX-GF 4039D-E). The task force recommended that General Foods remain in the business, in large part, because of large fixed costs already associated with its plant (CX-GF 4039E). Initial Decision 99 F. 824. A second task force was organized in 1971 because management again was greatly concerned whether General Foods had the ability to compete (Tr. 13 884-5).

825. The 1967 task force reported that new cereals as a group had yet to show a profit, and the 1971 task force reported a $13.1 million loss on new products since 1967 (Tr. 38 137-38; CX GF 3000Z-167 4039&-). In light of the importance of new products to effective competition in the RTE cereal market, this in itself reflects unsatisfactory profits.

826. It is hard to believe that General Foods, after some 70 years in the RTE cereal industry, would have considered getting out and absorbing the losses this would have entailed if it were earning a normal profit. A decision to close a business permits an inference that below-normal profits are being earned (Tr. 26 512-13). General Foods' consideration of quitting the RTE cereal industry allows a similar inference.

827. Complaint counsel (CPF 11-122; CRPF 11-238) would attribute General Foods' relatively low rates of return to poor management, costly major errors in brand introductions and inferior production plant. Without passing judgment on General Foods management' s expertise or on the wisdom or lack of wisdom in going forward with the new products that it developed (which I certainly am not qualified to do on the basis of this record), this would appear to reflect the risk facing a competitor in the RTE cereal industry. Further, to the extent General Foods' relatively low rate of return (229)can be explained away by poor management decisions and unsuccessful products, Kellogg s and General Mills' relatively higher rates of return may reflect superior management and superior products.

The COMPUSTAT Comparisons 828. COMPUSTAT is an organization which provides financial data on some 1 000 corporations. The COMPUSTAT data tape is a computer tape which records the financial information of such corporations. The data is taken from several sources, primarily the firms' 10-K fiings with the Securities and Exchange Commission (Tr. 19 163, 19 750-52, 19 755).

829. Dr. Stauffer computed an average economic rate of return for some 400 firms on the COMPUSTAT tape for which he had sufficient information, and arrived at arate about .5% higher than the IRS benchmark (Tr. 19 758). The firms listed on COMPUSTAT are relatively large and successful (Tr. 24 771). They account for Initial Decision approximately 75% of all the assets in the IRS Statistics of Income (Tr. 19 755).

830. Dr. Stauffer had considered using COMPUSTAT data to arrive at his benchmark. He did not, but utilized the IRS data as his only benchmark as he needed figures reflecting a larger and more representative sample of corporations (Tr. 24 755-58). The benchmark derived from IRS data, as modified above, is more reliable. In any event, the inflation and risk factors would stil justify a higher rate of return for the RTE cereal industry than for the average COMPUSTAT rate.

831. Dr. Stauffer also did a dispersion analysis whereby ranked firms covered by COMPUST A T according to their profitability. Dr. Stauffer calculated economic rates of return for each of the firms in the COMPUSTAT data base for which there was sufficient data and then ranked the firms according to their rates of return. These included, for particular years, between 500 and 1 000 firms which are listed on the New York and American stock exchanges (Tr. 19 750- , 19 755, 21 800; CX 701Z-23 thru Z-26). Dr. Stauffer prepared a graph covering the period 1958-1970 showing the percentage of COM PUS TAT reported companies that earned various economic rates of return (CX 701Z-23, Z-24, Z-25). Each cited exhibit shows the same graph for the COMPUSTAT companies. On each graph was plotted rates of return of the three respondents Quaker and Ralston and a composite of those five cereal companies under different cash flow profile and decay rate assumptions. The only exhibit to which further reference will be made is CX 701Z-25 as that purports to plot the economic returns of Kellogg, General Mils, Quaker, Ralston, and the composite of five cereal companies (the four mentioned plus General Foods) under a rectangular cash flow profie and a 10% decay rate.

832. Of all the companies plotted, only the entries for Kellogg and Ralston reflect what I have found to be the extent to which their rates of return have been established. Kellogg, with a rate of (230) return of 11.7%, falls among the top 16% of the most profitable firms in the manufacturing sector of the United States. Ralston, with a 9% rate of return, falls around the 30% mark. While Quaker is listed on the exhibit as being among the 12% most profitable companies with a 12.4% rate of return, I have already found that complaint counsel have failed to establish an economic rate of return for Quaker, but that Quaker s economic rate of return may be assumed to be well below 9% (Finding 797). Further, CX 701Z-25 uses the cash flow profile and decay rate which results in the highest rate calculation for Quaker among the various assumptions. General , Initial Decision 99 F. Mils is plotted as having a 13% rate of return, whereas complaint counsel have failed to prove a rate any higher than 10.5%. This would place General Mills among the 23% most profitable companies rather than among the top 10% as depicted by complaint counsel. General Foods, whether at the 7.7% rate computed by complaint counselor at the 7.3% limit I have found, has a rate of return so low that it is not depicted on the chart. The chart indicates that the 40% most profitable corporartions have a rate of return of 9% or more. The composite of five cereal companies which shows a weighted return of 11.6%, which would place that figure within the returns earned by the 16% most profitable corporations, is in error since it is composed in part of overstated returns for General Mills and General Foods and a clearly overstated return for Quaker for which no return has been established.

833. Further, the significance of the dispersion of rates of return and the ranking of firms in profitability categories is weighted by the fact that a firm earning a rate of return equal to the IRS benchmark of a normal rate of return (10.3%) would be among the 26% most profitable corporations in the manufacturing sector. This is because the firms listed by COMPUST A T are large and successful and account for 75% of all of the assets in the IRS Statistics of Income. 834. As stated in the prior section Impediments To New Entry In The RTE Cereal Industry," complaint counsel's theory of the existence of a barrier to entry into the RTE cereal industry is premised in part on the assertion that the industry has enjoyed supracompetitive profits. As I have found in this section, this premise fails for lack of proof.

Innovation As A Measure of Performance Complaint counsel assert (CPF 11-128 et seq. that respondents have not been innovative in their efforts to bring new and better products to consumers; that their new product developments have been essentially variations of existing RTE cereal products rather than innovations; and that respondents resisted widespread vitamin fortification even though they believed it would provide consumers with new and better products. (231) 835. This last stated aspect of complaint counsel' s assertion has already been found to be without substance. Respondents were innovative both in restoring their RTE cereals to whole-grain levels and in fortifying their products with vitamins and minerals (Findings 491-524).

836. Complaint counsel (CPF 11 129 thru 11-171) rely upon ":ula.I\.nA '-... .uA .n-'.

Initial Decision evidence that tends to show the similarity of many RTE cereal products that were introduced to products already on the market. From this they argue that respondents have performed poorly in that they failed to be innovative in their introduction of products. Respondents, on the other hand (KPF 5-158, 5-159, 5-161 thru 5- 163; GMPF 2-166 thru 2-169 , 2-171 , 2-175; GFPF 7-68 thru 7-139), rely on evidence which points up the differences' between products and purports to establish the significance of those differences. 837. Upon a review of the evidence, including many product formulas, it does appear that a number of RTE cereal products are quite similar without "earth shaking" differences. However, it also appears that complaint counsel are raising a question of the consumer welfare significance of new product introductions; and this is not a question for my judgment in this case. Any product that expands the choices of a consumer is an innovation. The signififance of an innovation is determined by its success in the marketplace (Tr. 264-B5, 23 391, 26 689-90). Whenever Kellogg introduced a product similar to that of a competitor, Kellogg believed that its product was superior (Tr. 29 805). During the complaint period, Kellogg nationally introduced 16 RTE cereal products, all but three of which were stil on the market at the end of the period (Tr. 29 600-01; CX- K 1067, 7173; CX 434). By 1971, some 37% of all RTE cereal sales were of products introduced during the prior sixteen years (GMX 564).

838. I have already made findings relative to respondents competition by the introduction of new products (Findings 530-B02). Pertinent to the present issue are the findings that respondents have engaged in intense, unrestrained and uncoordinated competition in the introduction of new products; that each respondent attempted to seize every new product opportunity before its competitiors, but, at the same time, tried to introduce similar products to compete with particular products of competitiors; that each company was fully organized to recognize, evaluate and develop all perceived product opportunities.

839. The extent to which complaint counsel have been able to point out relatively insignificant product differences does not (232) evidence a failure to be significantly innovative. To the contrary, it demonstrates the extent to which respondents try to take advantage of every product opportunity. If respondents find it competitively expedient to beat their rivals to relatively insignificant product holes, it is all the more important that they vie for more significant 8" Much of this evidence is in the nlt.ure of t.stimony of witnesses whose testimony I am required to accept at face value Initial Dccision 99 F. innovations. There is no evidence that they have not done so. And so we find that all respondents have competed by restoring and fortifying cereals with vitamins and minerals (supra Findings 492- 528); and that General Foods experimented with the inclusion of phosphates to inhibit dental caries (supra Finding 511) and experienced costly failures in its efforts to develop and introduce cereals with freeze-dried fruits (supra Finding 594). 840. There is no reason to believe that respondents have not attempted to come up with all reasonable innovations consistent with anticipated consumer acceptance. Neither complaint counsel nor I are in a position to pass judgment on what, if any, additional significant innovations the industry could or should have come up with. It cannot be found that something more dramatic should have been developed, or that there are areas where innovativeness was neglected. The only area specifically alleged by complaint counsel to have been neglected is that of product fortification, and that allegation has been found to be unsubstantiated. 841. In addition to new product introductions, respondents have innovated production changes. Kellogg, for example, has improved its knowledge of how to work with grains over time and has continually made changes in product formulation, production and packaging which it considers to be improvements (Tr. 29 318, 29 790- , 29,896, 29 974). It has been able to increase the running time capabilities of some of its product lines (Tr. 29 897-98). Certain products arc made by different methods which are improvements over how they were made some years ago (Tr. 29 270-72, 29 278 298-302).

842. General Mills spent over seven years in researching and developing the continuous puffing gun which replaced the batch gun. This increased product uniformity and eliminated problems that occurred with the start-up of each operating period, Each continuous puffing gun replaced six batch guns. Labor effciency was increased fourfold. The shift-over resulted in less maintenance, a reduction of start-up times and a need for less floor space ('lr. 33 016-23 33,127- 30).

843. From 1961 to 1973, General Foods made 115 product changes on 17 of its trademarked brands. Eighteen were process improvements, 26 were formula changes designed to affect the nutritional value of (233)the products and 71 were other formula changes. Thirty-five of the 115 product changes were intended to lower costs without noticeably changing customers' acceptance of 91 Procter and Gamble, after six years and up to $1 million in expenditures, was unable to develop a product that it rell would receive consumer acceptance(supra. finding 651) Initial Decision products. In General Foods' current view, the purpose of the other 80 product changes was to improve the quality of the product (GFX 13701).

Wastefulness In Advertising Expenditures 844. Since monopoly power is the ability to hold prices above competitive costs (Tr. 21 707, 26,l00-D5), the operation of businesses with excessive costs may manifest monopoly power and poor performance. Efficient firms will minimize their costs (Tr. 21 693 655). Complaint counsel assert (CPF 11- 185) that advertising cost levels in the RTE cereal industry are excessive and are imposed on consumers as a result of the existence of monopoly power. The following table shows advertising expenditures as a percentage of dollar sales for the years 1950 through 1972 for the six largest RTE cereal companies.

SIX FIRM ADVERTISING-TO-SALES RATIOS General General Kellogg Mils Foods Quaker Nabisco Ralston 1950 131 166 105 1951 165 1954 127 175 138 186 065 1955 175 1956 176 1957 124 157 177 1958 146 185 161 218 229 211 1959 172 226 159 196 173 240 1960 163 174 148 207 140 296 1961 153 202 135 279 152 294 1962 152 184 147 256 159 288 1963 152 206 153 .214 192 258 1964 152 202 170 202 206 232 1965 166 224 186 218 218 212 1966 157 198 210 202 222 218 1967 156 207 221 201 174 220 1968 121 173 184 210 148 162 1969 121 161 154 162 105 176 1970 106 132 163 164 094 185 1971 097 127 147 195 112 1972 094 148 138 131 130 NA - Not available.

(CX 513) (234) initial Decision 99 F. 845. This shows that the ratio of Kellogg s advertising-to-sales (A/S) ranges from over 9% in 1972 to over 17% in 1959. From 1950 to 1972, the Kellogg ratio exceeded 10% in 16 of the 18 years for which there is data. From 1951 through 1972, General Mills A/S ratio exceeded 12% in each of the 20 years for which there is data. At its peak, in 1965, General Mills A/S ratio exceeded 22%. From 1954 through 1973, General Foods A/S ratio exceeded 13% in each of the 17 years for which there is data. General Foods ' highest A/S ratio was reached in 1967 when it exceeded 22%. 846. The most information on advertising and sales for other industries exists for the years 1963 and 1967 (Tr. 27 676-78). The RTE cereal six company total advertising divided by the six company sales figure is . 175 for 1963 and . 172 for 1967. 847. A/S ratios for many manufacturing industries are reported in Stanley Ornstein s book Industrial Concentration and Advertising Intensity, at pages 60-61 (1977). The mean A/S ratio calculated by Dr. Ornstein for 87 consumer goods industries was 3.8% in both 1963 and 1967 , while the RTE cereal A/S ratio exceeded 17% in each of those years (Tr. 27 676-78).

848. The appendix to Dr. Ornstein s book separately reported A/S ratios for more than 320 manufacturing industries. Among the reported industries was Census Industry # 2043, Cereal Preparations, which is somewhat more inclusive than the RTE cereal industry, but is the closest category to RTE cereal for the purpose of making an A/S ratio comparison. Cereals had the second highest A/S ratio of all manufacturing industries in both 1963 and 1967 (Tr. 677-78).

849. Of the A/S ratios of 42 industries listed by advertising economists William Comanor and Thomas Wilson n only two were double-digit, one of which was of the RTE cereal industry. 850. Kellogg recognized that its advertising made it the most heavily advertised brand name on the American scene (CX-K 565K). 851. Respondents' high A/S ratio does not establish that their advertising has been wasteful. As previously found (Findings 425- 443), respondents independently set levels of advertising for each individual product based upon their own judgments of required (235) advertising levels to reach or maintain projected sales levels. This record provides no basis upon which to substitute my judgment for that of respondents.

852. My feeling is comparable to that expressed by Dr. Schmalansee in Brand Proliferation and Entry Deterrence: The Ready- To-Eat ., COffanor, William S., and Wilson, Thomas A. Advedi.ingand Marke! Power, (1974) Initial Decision Cereals Case, Report, Bell Journal of Economics, February 1977, at p. , that it has not been established "that the existing level of advertising in RTE cereals is above the socially optimal level." Further, I do not believe it to be appropriate to mandate a "socially optimal level" of advertising in place of one reached as a result of competitive considerations. OUf concept of a free competitive system does not envision imposition by government of permissible levels of advertising.

853. Advertising performs a necessary and Jegitimate function of advising prospective customers of the attributes of products offered for sale (Tr. 12 809, 22 475 , 31 509, 26 665-67; CX-K 456; CX- 4039Z-94). Advertising may be used in lieu of alternative methods of promotion and its success may increase sales volumes and so reduce unit costs (Tr. 22 490-91, 28 557, 29 911). Advertising is particularly important in the RTE cereal industry where there are many nonhomogeneous products and competition is carried on in large measure by the introduction of new products. 854. The large number of nonhomogeneous products, each with its requirement for separate advertising, dictates a relatively high A/S ratio. This requirement is further impacted by the high incidence of new product introduction.

855. Introductory advertising is designed to inform people of the existence of the new product and its important attributes. Introductory advertising levels are usually higher than subsequent levels because, in order to best get a return on the product, the length of time spent penetrating the market must be as short as possible (Tr. 809, 31 509). Speed is desirable in order to establish a sales base quickly so that both the manufacturer and retailer can determine whether there is a good market for the product and enough people will buy it frequently enough for each to make money (Tr. 15 242). The advertising behind a product is important to the retailer because it shows whether the manufacturer believes in the item strongly enough to support it. Heavy advertising programs are required to help persuade retailers to stock new products (Tr. 9185, 29 911). Also it may be necessary to increase the advertising on existing products to help combat the introduction of new products by others (supra Findings 126, 545-46).

856. Complaint counsel themselves have contended (CRPF 8-283) that A/S ratio comparisons between respondents are not meaningful without adjustments for the introduction rate of new products. It follows that A/S ratio comparisons between the RTE cereal industry and other industries are also not meaningful without taking into account the heavy incidence of new product introduction in the RTE , Initial Decision 99 F. (236Jcereal industry. Not only is there the incidence of heavy advertising in connection with the sales of successful new RTE cereal products, but the heavy advertising of products that fail also increases the overall company A/S ratio. General Foods, for example, was concerned with wasted advertising expenditures on unsuccessful products (CX-GF 4039Z-22).

857. There are still other differences between industries which do not permit a meaningful comparison of A/S ratios. The nature of the product may differ to the point of justifying different levels of advertising expenditures, or electing between different methods of advertising (with different costs) or choosing between advertising and other forms of promotion (Tr. 22 464-67, 28 143-45). For example, advertising expenditures in the automobile industry exceed the level in RTE cereal, but the dollar volume of sales is so much greater that the advertising-sales ratio is lower (1972 Census of Manufacturers, Special Report, SR 2-6; SR 2-144). 858. Thus, a comparison of A/S ratios between industries without more, does not allow an evaluation of whether the advertising in an industry is inefficient or is above an economic optimal level (Tr. 31 595).

859. Complaint counsel would substantiate their premise that respondents' high A/S ratios reflects advertising inefficiency and waste by reliance upon the Dorfman-Steiner principle. The Dorfman-Steiner principle is composed of two parts. First, firms try to maximize their profits in choosing, among other things, the level of their advertising expenditures. Second, there are diminishing marginal returns to advertising, as firms increase their advertising, the additional impact on sales decreases (Tr. 28 004-07). For example, an additional $1 000 expended for advertising might cause the sale of an additional 5 000 units, but the next additional $1 000 expended for advertising might cause the sale of only 2 000 moreunits. A point would be reached where the marginal return on the additional volume of sales would not equal the cost of the additional advertising that brought about those sales. 860. These two principles lead to the conclusion that the level of advertising is affected by the difference between sellng price and marginal costs (the gross margin). The firm s profi-maximizing level of advertising will be higher when gross margins are higher, because the marginal returns to that advertising wil be higher. Thus, the higher the price, all else including unit production costs being equal "' Uorfman, Robert and Steiner, Peter Optimal Advertising md Optimal Quality, The American &unomic Reuiew Dec- 1954.

!;_ Initial Decision the larger will be the expenditure on advertising (Tr. 28 004-12; CX 1008). (237) 861. The gross margin, or price-cost margin, is simply sales minus manufacturing costs divided by sales. More specifically, it is calculated as the value of shipments less payroll and materials costs divided by the value of shipments PCM = Value of hip.ments - (vavroll cost." + materials costs) Value of shipments (Tr. 27 698, 27 714).

862. Price-cost margins were calculated directly from the indus- 1972 Census oftry statistics contained in Table IB of Volume 2 of the Manufacturers and were verified later with accounting data supplied by Kellogg and General Foods. The Census of Manufacturers published approximately every five years and is publicly available (Tr. 703-06).

863. The cereal industry is classified as Standard Industrial Classification Code (SIC) 2043 by the Bureau of Census. The industry classification includes hot cereal and some baby cereal. The Census data shows that the PCM for the cereal industry is 48%. This was the ninth highest price cost margin of 451 manufacturing industries (Tr. 719-20). The same tables of the 1967 Census of Manufacturers show that the cereal industry was tied for the sixth highest price-cost margin of the 412 industries reported (Tr. 27 720). 864. PCMs are not reliable measures of profitability, such as are the economic returns on capital employed, the measures actually relied upon by complaint counsel in this case. PCMs, for example do not take into account distribution and selling costs, capital intensity. research and development, risk and various timing factors. In the absence of proof that any firms in the RTE cereal industry were enjoying supracompetitive economic returns, and the established fact that General Foods and Quaker were not, it cannot be held that any RTE cereal advertising was caused by monopoly profits. (238) 865. The Dorfman-Steiner principle is just that-a principle or theory of how gross margins wil be expended on additional advertising. It does not take into account the extent to which .. Complaint counsel, in .a actinn dealing with price competition (CPF 8-14) f!&;crt that. thf' RTE cereal industry is extrem,'ly profituble as shown by its price-ost margins. Thii; is an isolated reference and th" contention is not develope or even referred to in the sections of their brief dealing with the issue of,'xcess profits (CPF 11thru 11-127). PCMs constitute an incomplete view of corporate activity and arc not reli!lble indicators of the state of competition in an industry (Tr. 26 100, 31 581J Initial Decision 99 F. advertising expenditures wil be dictated by competitive requirements. And, as previously found, competitive requirements in the RTE cereal industry necessitate high advertising expenditures. The Dorfman-Steiner principle would not apply to the advertising of products which turn out to be failures; and it may well be that the costs of introductory advertising together with those of introductory promotions could exceed the gross return on new products for some time, so that there is no price-cost margin to consider expending for additional advertising.

866. Thus, the Dorfman-Steiner principle cannot be applied with the surety that advertising levels are coordinated with PCMs. This is apparent from an examination of the A/S ratios and gross margins of firms in the RTE cereal industry. Here, we find that General Mills A/S ratios are has a higher gross margin than Quaker. Yet, Quaker consistently equal to or higher than those of General Mills. Kellogg and General Foods' gross margins are close, but General Foods has much higher A/S ratios:

COMPARISON OF GROSS MARGINS (GM) AND ADVERTISING-SALES (A/S) RATIOS 1966-70 General General Kellogg Mils Foods Quaker A/S A/S A/S A/S 1966 .44 1967 .44 .20 .42 1968 .45 .40 1969 .47 .40 1970 .49 1966-70 .40 (Source: GMX 239, 553A).

867. Complaint counsel (CPF 11-200) assert that the pricing and advertising of Kellogg s Corn Flakes in the late 1960's and early 1970' s evidences that the Dorfman-Steiner principle has operated in (239)the RTE cereal industry. Sales of Kellogg s Corn Flakes were going down coincident with Ralston s private label price competition. Kellogg maintained its list prices through 1970. It reduced advertising expenditures between 1966 and 1972 , as well as its A/S ratio, but granted case allowances in 1968, 1969, and 1970 and reduced list ., For example, firms increase the advertising on particular products in anticipation of the introduction of competitive producl (Tr. 22 133-34) Initial Decision prices in 1971 (Tr. 17 528, 28 014-16; CX-K 856, 868, 1072, 7073B 7083B, 7192G-H; GFX 1318, 1319).

868. It cannot be said to what extent the pre-1966 advertising levels reflected advertising requirements or, if at all, an application of the Dorfman-Steiner principle. The same may be said of Kellogg actions taken after 1965.

869. The Corn Flakes situation is an isolated instance which may reflect a myriad of business considerations and cannot be relied upon to characterize respondents' advertising practices overall. Their practices overall are more properly evidenced by the comparison of gross margins and A/S ratios which has been considered above. In any event, advertising is engaged in to produce sales and profits, and there is nothing amiss for a man ufacturer to increase advertising to the point of realizing all marginal returns possible. In the absence of a showing that respondents were earning supracompetitive profits their advertising levels, even to the extent thay may be based on marginal returns, cannot be termed inefficient or wasteful. 870. In summary, I start with the basic assumption that advertising is a viable and legitimate method of competition. And complaint counsel have failed to demonstrate that the large volume of advertising in the RTE cereal industry does not have a competitive basis or that it reflects the existence of supracompetitive profits. A firm advertises in order to generate or increase sales and so increase profits. At the same time, once profits are realized, or in anticipation of profits, the firm may increase advertising up to the point of realizing all possible marginal returns. This, however, is pure theory and no findings can be made as to the extent respondents have increased normally competitive advertising levels to take advantage of the Dorfman-Steiner principle.

871. While the Dorfman-Steiner principle is logically sound there is stil another basic principle; and that is that, at a lower price, there may be a greater potential to sell more products, so that a firm with lower prices may increase its advertising to reach its potential (Tr. 31 590, 31 595).

Complaint counsel (CPF 11-205 thru 11-208) assert that respondents have engaged in wasteful advertising by misdirecting consumers to products which do not have the attributes or advantages over other products claimed in their advertising. Complaint counsel have abandoned their claims that respondents' advertising was false and misleading (Tr. 29 073; Complaint Counsel' s Answer To Respondents Motions To Dismiss, February 24 1978, at A- , A-15), and the issue cannot be resurrected at this time. (240) . , Initial Decision 99 F.T. Waste and Misallocation of Resources In the introductory portion to their section on performance (CPF 11-12), complaint counsel explain an aspect of consumer loss when an industry prices above cost. "This is due to the law of demand the higher the price, the lower the quantity demanded of a good. Whenever an industry prices above the competitive level, consumers demand less than they normally would, output is restricted, and thus, resources which should have been devoted to production of the good will be diverted to other goods and a misallocation of resources results. . ." (Tr. citations omitted). Complaint counsel then quote from economist, Clair Wilcox (C. Wilcox, Public Policies Toward Business, 11 (1955)), in part, as follows: The mobility characteristic of competition thus tends to achieve the allocation of resources that consumers desire. Monopoly, by contrast, frustrates such an allocation. The monopolist is likely to increase his profit by raising his price. He will then limit hi." output to the quantity that the market will take at the price that he has fixed. Consumers who would be willing to purchase larger quantititcs of his product at lower price arc left, instead; to buy goods that are wanted less. Resources are thus diverted from those things which the community prefer to those which are, at best, a second choice (emphasis supplied), Complaint counsel then conclude (CPF 11-13): The loss that results when resources are misallocated is called the "deadweight" or welfare" loss. (Scherer, Tr. 27 999- 000). "It is a loss of value that the consumers would have realized if they could have consumed at prices equal to cost. So consumers lose it. It is an inefficiency that cumes from the restriction of output by monopolists causing consumers willing to pay a price greater t.han the cost. not to be able to make those purchases" (Scherer, Tr. 28 000) (emphasis supplied). 872. While this aspect of the economic law of supply and demand is stated in general terms, complaint counsel obviously intend it to be descriptive of the RTE cereal industry. Complaint counsel however, have failed to show how the law of supply and demand has (241)impacted the cereal industry or its customers. The record does not support complaint counsel's assumption that there have been high prices in thc RTE cereal industry to the extent that demand has been impacted and producers have curtailed supply. 873. The cereal industry has experienced rapid and substantial growth, indeed substantially faster than the growth of all goods and services and faster than the growth in real output of the economy a whole (supra Findings 162-66). RTE cereal prices have been maintained at a level substantially lower than those of most other breakfast foods (supra Findings 49-50), and price differentials in favor of hot cereals have not impacted the sales of RTE cereals Initial Decision (supra Finding 50). The RTE cereal industry is less price sensitive than other commodity industries (Tr. 23 113, 27 134), and the cost of RTE cereals is not a very important component of the total food budget (Tr. 27 142).

874. Consumers are being offered a large number of RTE cereal products at various prices from which to choose. At the lower edge of the pricing spectrum, we find the leading RTE cereal product, corn flakes. Consumers who may be interested in the lowest available prices may choose from Kellogg s Corn Flakes, General Foods' Post Toasties, General Mils' Country Corn Flakes and private label corn flakes (supra Finding 72; Tr. 16,611, 17 750, 25 815- , 28 821). The presence of private labels in this segment would mean the availability of prices even lower than those of the branded products. 875. There is no showing that any consumers are being priced out of the market. And, as Dr. Scherer testified, if buyers are able to purchase a product at a price they are wiling to pay, even though the price exceeds the manufacturer s costs, allocative efficiency is being staisfied (Tr. 27 652).

876. On the other side of the coin, there is no evidence that respondents have limited supply. Each company has attempted to come up with new products to replace volume attrition of existing brands and to increase its market share (supra Findings 533-40). While General Foods was relatively unsuccessful in its efforts, both Kellogg and General Mils have maintained their operations at a high level of capacity (supra Finding 407; Tr. 26 726). 877. Economist Michael Glassman, introduced by complaint counsel, has suggested that Kellogg, for example, could sell more if it were to lower its prices (Tr. 26 569), and that it could meet that increased demand by building more capacity (Tr. 26 726). This suggestion is rejected. Kellogg has the right to exercise its own judgment with respect to operating under current production capacity. Kellogg, which enjoys some 40% of the market, may well believe that it is already supplying all of its cereals that the consuming public wants, and that it would not be economically feasible to make additional capital investments. (242) 878. In summary, there is no showing that consumers have reduced their overall consumption of RTE cereals because of alleged supracompetitive prices; nor is there any showing or reason to believe that, if prices were lowered, demand or supply would be increased.

." I am not finding that the RTE cereal industry is immune to the impad of the economic law of supply and demand; Rimply that there has been no showing that extant prices are so high a. to have reduced demand or supply Initial Decision 99 F. Consumer Overcharge Calculation Complaint counsel (CPF 11-209 thru 11-232) utilize several alternative methods to calculate the amounts by which consumers have been overcharged by reason of respondents' alleged supracompetitive profits. Having found a lack of proof of supracompetitive profits, there are no overcharges to calculate. Nevertheless, some of complaint counsel' s methods of computing the asserted overcharges merit comment. Complaint counsel recite (CPF 11-212): In July 1971, Kellogg reduced the list price of its Corn Flakes by 16% , from $6. for a case of 24 12-ounce boxes to $5.80 for a case of 24 12-ounce boxes (CX-K 7192G 7083B, 7073B; Scherer, Tr. 28 015). In order to "still show an impressive price spread" between Ralston s private label Corn Flakes and Kellogg s Corn Flakes, Ralston reduced its prices from $5.80 a case to $5.50 (CX- 1516A). From this, complaint counsel argue (CPF 11-213) that the price of Kellogg s Corn Flakes had been infiated by at least 16% duc to monopoly power; and that it is likely that all RTE cereal prices have been inflated by a comparable amount. Complaint counsel (CPF 11- 214) then take 16% of the total sales of branded products and arrive at a $1 037 980 000 overcharge for the three respondents for the years 1958-1972 and a $1 223 135 000 overcharge for the respondents and Quaker, Ralston, and Nabisco for that period. The figures include a $207 817 000 overcharge attributed to General Foods even though, as found above (Findings 819-26), General Foods did not earn supracompetitive profits.

879. Kellogg s 16% price reduction on a single product in the year 1971 cannot be interpreted as a reflection of 16% monopoly profits on that product. It certainly cannot be interpreted to mean (243)that Kellogg was earning 16% monopoly profits on all products from 1958 through 1972; and most certainty it cannot be interpreted to mean that the five other largest firms in the industry were earning 16% monopoly profits on all of their branded items. 880. Complaint counsel next (CPF 11-215 thru 11-221) would apply the 3.8% advertising to sales ratio found by economist Stanley Ornstein for 87 consumer goods industries (see supra Finding 847) to the RTE cereal industry, and conclude that all advertising in excess of the figure was wasteful. Having already found (supra Findings 847-77) that it would be inappropriate to compare the RTE cereal industry A/S ratio with the 3.8% figure and that there is a lack of proof that advertising in the RTE cereal industry is wasteful complaint counsel's calculations are meaningless. 881. Finally, complaint counsel's attempts (CPF 11-222 thru 11- 230) to calculate the extent of alleged supranormal profits on the Initial Decision basis of its already rejected accounting and economic rates of return in relationship to a benchmark are similarly rejected. (244) SUMMARY AND FURTHER DISCUSSION Complaint counsel advance two grounds for asserting that respondents have violated Section 5 of the Federal Trade Commission Act: (1) Conspiracy; and (2) Acts and practices of respondents and economic performance of respondents and the RTE cereal industry, under a shared monopoly industry structure. The complaint as issued failed to charge a conspiracy and was never amended to encompass that charge. While the ALJ before whom the case was then being heard allowed complaint counsel to proceed under a conspiracy theory, respondents never consented to this. The issue of conspiracy, therefore, many not be deemed to have been raised under the concept of conformance to the evidence under Section 315(a)(2) of the Commission s Rules of Practice, which requires that the issue be within the scope of the complaint and be tried with the consent of the parties. Respondents, therefore, may not be found to have violated Section 5 by reason of conspiracy. Notwithstanding this holding, in consideration of the fact that this case was allowed to be tried under a conspiracy theory and in light of the unusually long time it has taken to try this case and the voluminous record that has been compiled, I have made all findings called for by the evidence, including those relating to the issue of conspiracy.

In any event, complaint counsel's case fails under either of its theories for lack of proof.

The factual issues, to a large extent, are common to both the conspiracy and the ushared monopoly" theories. Under the conspiracy theory, it was necessary for complaint counsel to establish that the respondents acted in particular noncompetitive fashion pursuant to agreement, express or tacit. Under the "shared monopoly" theory, many of the same acts and practices are asserted to be demonstrative, and an exercise, ofresponrlents' alleged monopoly power. While the oligopolistic structure of the RTE cereal industry is an essential element of complaint counsel's "shared monopoly" theory, respondents are not charged with violating Section 5 simply on account of the structure of the industry. Complaint counsel recognize that in an oligopolistic industry the members may be competitive or they may operate together in a monopolistic manner, so that, in addition to the structure of the industry, "a careful analysis. . . of 250 FEDERAL TRADE COMMISSION DBCISIONS Initial Decision 99 F. the conduct and performance of the sellers must be undertaken (CPF 6-17).

Complaint counsel assert that respondents avoided price and non price competition on RTE cereal products. Complaint counsel assert that each respondent is aware that it is to their mutual advantage to avoid competitive activities which can only bring about responsive competitive activities by their large rivals. The firms therefore, avoid such competitive acts in accordance with understood rules of the game. (245) On the issue of the avoidance of price competition, and in support of their structural approach, complaint counsel rely upon a generally accepted economic theory of Dr. Jesse W. Markham. That theory is to the effect that, in industries having certain characteristics, one would expect price leadership in lieu of overt collusion. One of the conditions required under Dr. Markham s theory, is that the commodity produced by the several firms be viewed by all of the firms as extremely close substitutes for each other. This condition is not met in the RTE cereal industry. To the contrary, there are many different types and categories ofRTE cereals. Price leadership" among clearly differentiated products is meaningless. It is necessary to have products sufficiently similar so that price is a primary element and can be coordinated by following a price leader. In this industry with differentiated products, it is not clear what there is to be coordinated. Price cannot be identified and isolated for particular products as something the sellers can focus on for purposes of coordination.

The RTE cereal industry, therefore, has not been shown as one in which price leadership in lieu of overt agreement is to be anticipated. Respondents' unlimited product competition with non- homogeneous products is inconsistent with a desire or effort to coordinate their activities and eliminate competition. To the extent they have proliferated products, they have engaged in a manner of competition that hinders and restrains coordination.

The record evidence pertaining to respondents' alleged price coordination is what might be expected from the above analysis of Dr. Markham s economic theory. A pricing pattern consistent with coordination was established with respect to only three sets of products-Kellogg s Corn Flakes and General Foods' Post Toasties; Kellogg s and General Foods' Raisin Brans; and Kellogg s Sugar Frosted Flakes and General Foods' Sugar Coated Corn Flakes. Other than for these three sets of similar products, there is no evidence of ,n Another condition not established on this record is that individuul firm cost CUrVI;S must he suffciently similarsothat.'omeparticularpriceallowsalJfirrnstoopcrau,atasatisfactoryratcofoutput. , Initial Dccision price uniformity, maintenance of pricing levels, or pricing responses consistent with a pricing agreement or arrangement among respon- For all (246)otherdents or with price leadership and followership.98 products, the record tends to evidence a lack of brand price coordination or parallel pricing movement. The evidence complaint counsel rely on to show a pattern of price leadership in lieu of overt collusion, with Kellogg being the price leader, consists of an analysis of pricing rounds in the industry covering 1965 through 1970, prepared and testified to by Dr. Scherer. As described by complaint counsel, Ita price round is a series (or group) of list price changes that occur when a firm changes the price of two or more regular size branded products" (CPF &-83). The lack of probative value of this analysis has been found as follows: Finding 265. The 16 price rounds, a"; presented by Dr. Scherer and relied upon by complaint counsel, do not take prices into account other than for the fact that price changes were made. Levels and magnitudes of price changes are ignored. Indeed, the price round presentation docs not even demonstrate that "lead" price changes on particular products were followed by price changes on what may be termed directly competing products. To the contrary, the amount of price change varied by product and the types of cereals involved in one company s price change varied from those in the subsequent price change of other companies. In short, there is no correlation of individual products or individual product prices in Dr. Scherer s price rounds. Indeed individual products and prices are not even evaluated. There is, therefore, no showing of correlation of any brand prices or price differentials as to particular brands. Dr. Scherer has conceded that, for most products there seems to be very little pattern in the relationships between changes in the price of one product relative to changes in the price of another " (Tr. 922).

In addition to the failure of the price rounds analysis to contain necessary product and pricing information, the rounds are inconsistent with price leadership-price followership. Kellogg, the alleged price leader, was not always first in announcing price (247)changes; and price changes by one respondent were not always followed by the others. Also, when a price change was followed, it was often only after a long delay.

Kellogg set prices for RTE cereals on the basis of a guideline for gross margins and a target net profit figure. It did not establish prices in order to maintain a profitable price structure for Its RTE cereal competitors. The record reflects genuine, independent business reasons why General Mills and General Foods did not originate ,. &onomisL ,'xpcr:t price similarity, even identity, for very similar products under competitive condition(l!; Price uniformity alone is insufficient to "stablish a price fixing: conspiracy, since their may be the normal result of competition among similar products,See, e. , Murlun Salt Co. United Slale" 2:15 F2d 573 I 10th Cir. 1956); FT' v. Lukem; Steel Cu. 454 F,Supp. 1182, 1190 ID. c. 1978) Initial Decision 99 F. price changes more frequently than they did and why they followed price increases of their competitors to the extent that they did. Having failed to establish coordination on the part of respondents relative to the establishment, maintenance or change of prices for RTE cereals, complaint counsel have failed to prove (1) a price fixing conspiracy, or (2) noncompetitive pricing activity under a shared monopoly theory.

Complaint counsel assert that respondents have supported their pricing arrangement by refraining from indirect forms of price competition which might have spiraled into unrestrained price competition. These include trade deals, the use of cents-off labels, the insertion of in-pack premiums, and the production of private label cereals. Complaint counsel also assert that, while respondents offered discount coupons, their use was insignificant. However, complaint counsel have failed to establish that respondents, either by agreement or by means of price leadership-price followership, fixed or coordinated list prices. There is, therefore, no price fixing arrangement shown, to be supported by refraining from indirect forms of price competition. In any event, the record fails to establish that respondents' activities with regard to indirect pricing activities were as a result of agreement or arrangement or flowed from the structural nature of the market. To the contrary, the record shows that, to the extent respondents acted similarly, each respondent reacted individually to common market conditions and the advent of television as a marketing device. In addition to individual reactions to common marketing and advertising factors, respondents made independent decisions on the basis of their own situations. General Foods, for example, went through a period when it stressed a full line promotional approach. General Foods, accordingly, cut down on individual item promotions (including trade deals, cents-off labels and in-pack premiums). On the matter of production of private label products, neither General Mills nor Kellogg had the degree of excess capacity over an extended period of time that was required to enter into long term commitments to provide private label products. General Foods, on the other hand, had produced private label RTE cereal products since at least 1937 and attempted to expand its private label business. It discontinued private label only after it had suffered financial reverses and was unable to generate appropriate returns. Contrary to complaint (248)counsel's attempt to downplay respondent' s utilization of coupons, couponing was widely used by respondents and they were redeemed by consumers at significant levels. Complaint counsel have failed to establish that respondents Initial Decision indirect pricing activities were in accordance with any agreement tacit or otherwise, or that respondents' activities reflected anything other than independent business decisions made in an effort to further legitimate business interests.

Complaint counsel assert that respondents avoided acts of nonprice competition that might have led to price competition or that potentially might have threatened their goals of maintaining marketplace stability and maximizing profits. Respondents exchanged advertising data through the A. C. Nielsen Co. Complaint counsel assert that this exchange of advertising information allowed respondents to prevent an expensive advertising war. However, there is no evidence that the respondents utilized the information to curtail or otherwise coordinate their advertising efforts. To the contrary, the record shows that industry members competed very strongly against each other in their advertising endeavors. Respondents' advertising expenditures followed markedly different and varying patterns both overall and on individual brands.

The complaint charged respondents with having maintained and utilized monopoly power. The exchange of advertising data was not alleged as a violation of law. Therefore, the failure to show that such an exchange was utilized for monopolistic purposes constitutes a failure of proof under the instant complaint. Complaint counsel assert that respondents had a tacit agreement to avoid competition for shelf space in retail stores; that Kellogg formulated and implemented a shelf space allocation plan and General Mills and General Foods acquiesced in that plan. The assertion has no record basis.

Kellogg s shelving program advocated space according to sales and grouping by manufacturers. The program was unilaterally instituted by Kellogg to afford it a competitive advantage over other RTE cereal manufacturers, including the other respondents. There is no reason to believe that General Mills or General Foods was party to the institution or implementation of Kellogg s plan. Retailers adopted Kellogg s recommendations because such recommendations served their own profitability and efficiency interests and were considered the most reasonable way of stocking RTE cereals.

Faced with Kellogg s shelving program which advocated space according to sales and grouping by manufacturers, which principles were logical and advantageous to retailers, as well as ones to which retailers were accustomed, it is not surprising that General Mills Initial Decision 99 F. and General Foods advocated the same guidelines when they competed for shelf space. (249) While all three respondents generally advocated shelving by manufacturer and allocation of space according to volume, they each competed for all the space it could get. This included efforts to get more than a "fair share" of shelf space if its credibility would not be impaired by doing so. And each recommended the discontinuance of slow moving products of its competitors.

General Foods developed Compact Packages" which were resized versions of existing cereal packages, designed to hold equivalent quantities of cereal in smaller boxes. General Foods thought this would enable it to acquire facings for additional itcms and would also add to its prestige as a leader in innovative shelving and so secure greater acceptance for its shelving recommendations. The program was a failure and General Foods lost shelf space as a result. In 1971, General Foods developed and introduced C. S.s. (Customer Oriented Method of Profitability and Sales Service), a computerized system for making shelf space allocation recommendations on the basis of product profitability. However, retailers were not interested in the program.

Finding 490 summarizes the situation with respect to shelf space competition as follows:

It is concluded, therefore, that the record does not support complaint counsel' assertion that respondents had a tacit agreement to avoid competition for shelf space in retail stores. Kellogg independently formulated a shelf space allocation plan that incorporated principles which were in accord with retailers' preferred methods of doing business. The other respondents, faced with the same requirements of retailers responded with plans that incorporated the same basic principles. While each respondent competed for the most favorable shelf space location and the most space it could get, it was constrained not to push for more than a reasonable share in order to maintain rapport and credibility with retailers. Nevertheless, both General Mills and General Foods did present shelving alternatives and variations in an effort to gain competitive advantages.

Complaint counsel assert that respondents avoided widespread product fortification until outside pressure forced them to fortify their cereals; that fortification then occurred as a result of coordinated activity to ensure that no one of the respondents would gain a competitive advantage by introducing fortified products before the others. The record does not support complaint counsel's assertions of agreement and coordination. (250) Prior to 1970, the Council on Foods and Nutrition of the American Medical Association and the Food and Nutrition Board of the National Research Council both recommended that nutrients be added to foods only to the extent of restoring what was lost during Initial Decision the manufacturing process. During the mid-1960' , the Food and Drug Administration publicly opposed fortifying RTE cereals beyond allowable limits, on the ground that the availability of vitamins to consumers from other sources made cereal fortification unnecessary. The FDA proposed a rule and instituted rulemaking proceedings to so limit fortification.

Notwithstanding the pronouncemcnts of the Council on Foods and Nutrition of the American Medical Association and the Food and Nutrition Board of the National Research Council and the adverse position of the FDA, respondents engaged in considerable fortification activity prior to 1970.

However, there was limited consumer demand for fortified RTE cereals during the 1960's and only limited, temporary success for record fails to indicate thatproducts that were fortified. The respondents' individual competitive efforts prior to 1970 in the field of product fortification were not fully commensurate with the public demand. There is no evidence that indicates that respondents reached agreements concerning, or coordinated, their prc-1970 fortification conduct.

In the late 1960's and early 1970's, there was a dramatic change in the national attitude toward the fortification of cereals. A White House Conference on Nutrition issued a report in December 1969 to the effect that there were significant nutritional deficiences in the diets of large segments of the population. It recommended that the proposed 1966 FDA regulations barring the fortification of breakfast cereals not be adopted, because the widespread acceptance and consumption of breakfast cereals made them effective carriers of essential nutrients. It attacked the view that all needed nutrients were obtained from ordinary diets, and recommended strong food fortification programs. Consequently, the FDA abandoned its proposed rule to prohibit food fortification. In July 1970, Mr. Robert B. Choate, in testimony before a congressional committee, criticized the lack of nutrients in RTE cereals. Mr. Choate s testimony was widely publicized and further increased industry and consumer interest in vitamin fortification. Following the White House Conference on Nutrition report, the FDA reversal of position and the Choate testimony, the respondents became very heavily engaged in product fortification. Respondents activities, however, were fully consistent with individual, competitive responses to stated public policy and consumer interest and demand.

While the respondents attended two meetings of the Cereal Institute shortly after Choate s testimony, respondents did not (251) Initial Decision 99 F. discuss plans regarding fortification at either meeting. No agreements were there made regarding fortification. There is no evidence that the respondents otherwise communicated regarding their fortification activity, or that any respondent had advance knowledge of the fortification plans of the others. There is no basis, therefore, for an inference that respondents fortification activities, which were most reasonable in the light of ongoing events, were in response to an otherwise unproved agreement rather than the ongoing events.

Complaint counsel assert that the RTE cereal industry is marked by high barriers to the entry of new firms and that there are no barriers to entry unrelated to respondents' conduct. The conduct so targeted by complaint counsel is brand proliferation which is asserted to provide the complete answer to lack of entry. It is complaint counsel' s position that respondents' avoidance of competition by other means led them to turn to brand proliferation the introduction of a large number of differentiated, highly advertised trademarked brands; that, while brand proliferation is not in itself unlawful, respondents must be held responsible for its deterrent effects upon entry, since respondents turned to this method of competition as a result of their mutual avoidance of other means of competition.

Brand proliferation is nothing more than the introduction of new brands which is a legitimate means of competition. Respondents brand proliferation is vigorously competitive and, as conceded by complaint counsel, is not predatory and not in itself unlawful. Respondents engaged in intense, unrestrained and uncoordinated competition in the introduction of new products. There is no evidence of a conspiracy or intent to deter entry by means of new product introductions.

Consumers' desire for variety for breakfast is responsible, in large measure, for the differentiation of RTE cereals. A firm in the RTE cereal industry must introduce new products in order to remain profitable and compete for market share.

Respondents, therefore, may not be held responsible for the results of this legitimate method of competition, unless it was the proximate result of their having otherwise limited their competitive efforts as charged. However, complaint counsel failed to prove those charges. Further, even if respondents had conspired or otherwise unlawfully coordinated their other competitive efforts, new product introduction would still have remained as a legitimate means of competition. No causal relationship has been shown between the alleged avoid- Initial Decision ance of other methods of competition and competition by brand introduction.

This fully disposes of complaint counsel's effort to hold respondents liable under Section 5 by reason of having engaged in new (252)brand competition (proliferation). However, in order to provide complete findings for a reviewing authority, I examined complaint counsel's analysis of how brand proliferation allegedly created a barrier to entry and made numerous findings with respect to various steps in complaint counsel's theory. Without here repeating all of the elements of complaint counsel's theory and all of the findings following is a reference to some of the more dispositive findings. One of the requisite elements of the proliferation theory is that the market be localized or segmented so that each cereal competes almost exclusively in its own limited segment. There are segments or categories of cereals which compete more strongly with each other because of their similar attributes. There are some cereals that are so similar that they compete with each other on a one to one basis. At the same time, some cereals may have a broader appeal than the particular segment or category they may fall in so that they compete to varying degrees with cereals outside of their particular category. The record, however, does not permit a delineation of the segments and the degrees to which competition among cereals is confined to cereals in particular segments or extends beyond such segments. The extent of the impact of localization, therefore, is unknown.

Complaint counsel assert that proliferation has crowded some segments to the extent that there is not sufficient opportunity for a new product. While crowding has undoubtedly deterred new entry into some areas, the areas so precluded, their economic significance and the time periods of preclusion have not been identified. Complaint counsel' s assertion that the introduction of all of the products that can profitably compete in an area wil dissuade the entry of stil more products is not at all disturbing. It is just another way of saying that competition wil deter entry and the more vigorous the competition, the more likely it is that new entry wil be deterred. Complaint counsel have merely evidenced and analyzed how competition by existing firms in the form of new products wil deter the introduction of still more products by new entrants. There is a product efficiency of scale equal to about 1 % of the market. Complaint counsel assert that firm efficiency of scale is not reached until sales of 3.5% to 5% of the market are achieved. It is argued that respondents, all of whom are already operating at or above firm efficiency of scale, can introduce a new product whenever Initial Decision 99 F. a 1% opportunity appears, whereas an outsider cannot; that an outsider is faced with the insurmountable task of finding and expensing three or more opportunities in order to enter the market at minimum firm economy of scale, so as not to be at a cost disadvantage to respondents.

Evidence of probative value on the issue of firm efficiency or economies of scale was found to be limited to a " Headen-McKie study. Findings 629 and 641 pertain to the Headen-McKie report: (253) 629. The Headen-McKie conclusions on firm economies of scale have no analytical or other substantive support. As conclusions drawn from conversations with production managers, and limited to production costs, they are at best a rough estimate that firm economies of scale are not fully realized until a 5% market share is reached. However, the report does not indicate the difference in production effciencies at the 1 % and 5% levels or for points in between. Accepting the report at full value, it may be that, while firm economies of scale are fully realized at the 5% level, there is not a significant difference in firm effciencies at the 1 % and 5% levels. 641. In summary, under Headen-McKie, which does provide a rough estimate of production scale economies, it would be economically feasible to enter the RTE cereal market with production facilities capable of supplying about 1 % of the national market. While the entrant could earn "satisfactory " profits at that level of production, it would not achieve full production economies of scale unless it supplied about 5% of the market. The study, however, does not indicate the degree of disadvantage that a firm would be under at various levels of production below 5% down to 1 %. Even if a firm required 5% of the market to achieve minimum efficient scale, it would enter at a smaller volume if the cost disadvantage was not too great. It is impossible, therefore, to evaluate the extent to which preemption of new entrants by reason of their inability to achieve economies of scale has acted to deter entry (footnotes ommitted).

While brand proliferation may well exert a deterrence on entry, it is not the only deterrence. The situation has been summarized as follows in Findings 668-74:

668. Respondents have engaged in intense, uncoordinated competition in the introduction of new (254)products. This competition is not unlawful nor was it induced by other unlawful activity.

669. Obviously, the more successful new products introduced by respondents and other incumbents, the more saturated the market and the less requirement and opportunity for the introduction of new products by outsiders. By the very nature of differences in demand (e. preswcets, natural, fortified, bran, flavored), products are to a degree localized, and a new product would to a degree be limited as to the segment from which it could attract its users. Individual products account for lower market "' As noted ahove. the extent to which lack of entry in the RTE cereal industry may be explained by brand proliferation i an unknown quantity, IJe:ausf' of a failure of proof of two of the eement upon which the theory rests- These are: (lJ minimum firm efficient scale of entrance below which the TI'W entrant would be at a significant competitive disadvantage Lo existing firms and (2) the degree oflocali?.ation of RTE cere..1 products Initial Decision shares and smaller poundage of sales than formerly. The industry has become one of relatively small volume brands.

670. Incumbents are at an advantage over potential entrants in developing and marketing acceptable new products. They can utilize existing research and development, market research and other expertise in locating opportunities and developing products to meet perceived demand. They can also utilize unused capacity for production of a new brand, whereas a new entrant would have to build that capacity. 671. Respondents and other incumbents are not only capable of finding and taking advantage of an opportunity before a potential new entrant, but, because of economies of scale, are in a position to take advantage of smaller opportunities. 672. Would be entrants are faced with substantial fixed costs in research and development, market research, plant production equipment and introductory advertising. To the extent the requirement exists to introduce multiple products, the costs would multiply. Potential entrants, therefore, are limited to large firms, primarily those already producing and supplying grocery products who can utilize their existing expertise and so mjnirnize costs of entry and operation. 673. The limited number of potential entrants would exercise caution in actually entering because of high capital costs, long lead times in developing and marketing acceptable products, extended periods even after entry in reaching levels of profitable operation and recovering capital investments, and the high risk that a product may prove unacceptable at various stages up to national entry or may fail after entry. Indeed, the problems of developing an acceptable product with which to enter may be insurmountable.

674. While potential entrants would be aware of the publicly reported profitability of Kellogg, they could not hope to emulate the most successful company. Their desire (255)to enter would be tempered by their observation of others in the industry. They would hesitate knowing of the failure of General Foods in marketing its cereals with fruit and jts loss of market share, and of Nabisco s decline in market share. Potential entrants would also hesitate because of the observed inability of Pillsbury, Colgate, International Multifoods, and H. T. Heinz to remain in the market and the limited success of Pet.

Basic to complaint counsel's shared monopoly theory is the allegation that respondents have maintained a pattern of conduct that has enabled them to charge supracompetitive prices and to reap monopoly level profits; and that this demonstrates that they are sharing monopoly power. Complaint counsel have stated However proof of monopoly power in this case rests primarily on the evidence that each of the respondents gained monopoly profits over a long period of time." Complaint counsel's theory of the existence of a barrier to entry into the RTE cereal industry is also premised, in part, on the assertion that the industry has enjoyed supracompetitive profits.

Complaint counsel have elected to measure and evaluate profits in terms of rates of return on capital employed. This is an appropriate means of measuring profits for purposes of this case and is preferable to other measures, such as return on equity or return on sales. There are significant differences between accounting and economic rates of return. Upon an analysis of those differences and the Initial Decision 99 F.T. nature of the RTE cereal industry, it has been found that economic rate of return, not accounting rate of return, is the appropriate measure to be used in appraising profits enjoyed in the RTE cereal industry and in making comparisons among respondents, with other companies and with other industries.

Dr. Thomas R. Stauffer has devised a formula to determine the economic rate of return for an ongoing firm- This formula is a pioneering contribution jn the field and was the subject of Dr. Stauffer s doctoral dissertation in economics at Harvard University. Other doctoral dissertations in economics at Harvard have applied his formula in the analyses of specific industries. The formula was applied by Dr. Stauffer in the instant case. Dr. Stauffer s formula starts with an accounting rate of return and from it calculates an economic rate of return. In order to determine General Mils' and General Foods ' accounting rate of return for RTE cereal, it was necessary to segregate and allocate portions of overall company accounting data covering multifacets of those companies businesses that pertained to the RTE cereal segment of the company. Once this was done, there were various unknowns in the formula used to convert accounting rates of return to economic rates of return which had to be estimated. The results were then compared with a benchmark-the average economic rate of return on capital (256Jemployed for all firms in the manufacturing sector of the United States.

After applying allowable allocations to general company accounting data to reach accounting rates of return for RTE cereal, utilizing allowable estimates for the unknown factors in Dr. Stauffer return toconversion formula, and adjusting the benchmark . rate of exclude data for manufacturing companies which had no income, the following economic rates of return were arrived at: General General Kellogg Mils Foods Ralston Benchmark 1958-1970 11.7 10. 10. 1954-1972 13.2 10. 10. 10. This is not a finding of what respondents' and Ralston s economic rates of return have been, but rather that there is a lack of proof of any higher returns.

Probably because of the fact that Quaker s rate of growth exceeded its rate of return, Dr. Stauffer s formula was inappropriate to estimate Quaker s economic rate of return; and there is no record Initial Decision evidence of that rate. There is record basis, however, for assuming that it was well below 9% well below the benchmark. The record also independently establishes that General Foods did not earn above-normal (supracompetitive) profits.

There are a number of considerations which bias the benchmark downward understate the average profits of the manufacturing sector when comparing them with the profits of RTE cereal companies. Two important factors that must be considered before meaningful comparisons may be made between RTE cereal companies' rates of return and the benchmark are (1) inflation and (2) risk. Dr. Stauffer made no adjustments for inflation, since he was not trying to compute absolute levels of profitability, but rather to compare relative levels. However, the failure to account for inflation caused a greater upward bias on the computation of profits of the Rre cereal firms than on the benchmark.

A higher than average return is still considered competitive or normal if the amount above the average is compensation for above average risk. A normal rate of return for a business with above average risk would be greater than for less risky businesses. RTE cereal is a relatively high risk industry. Thus, it has been found:

Finding 815. The benchmark derived from IRS data for the manufacturing sector cannot be deemed to (257)draw an absolute line of demarcation between normal and monopoly profits. While such IRS data have generally been relied upon by economists there is concedely some degree of inexactitude (citations omitted; and note the various biases enumerated above). Further, the benchmark is simply an average of firms earnings below the average and firms' earnings above the average. It cannot be said that every firm earning above average, whose rate of return contributed to reaching the average, is earning monopoly profits. Finally, while not quantified, considerations of inflation and risk establish a normal rate of return for the RTE cereal respondents at a somewhat higher level than an average benchmark. Finding 816. In consideration of the above, while the bencr-mark may be used as a rough guideline to aid in evaluating rates of return, a rate of return cannot be said to be monopolistic or supracompetitive unless it is substantially in excess of the benchmark. That is not the case here. General Foods' rate of return is substantially below the benchmark. Quaker s return may be assumed to be below it (Finding 796). General Mills' and Ralston s rates of return approximate it. Only Kellogg is shown to have a rate of return in excess of the benchmark; and, in consideration of the variables and biases discussed above, its rate cannot be said to be substantially in excess.

Thus, complaint counsel's factual assertion basic to its shared monopoly theory, that respondents and others in the RTE cereal industry realized supracompetitive profits, fails for Jack of proof. Complaint counsel assert that respondents have not been innovative in their efforts to bring new and better products to consumers; Initial Decision 99 FTC. that their new product developments have been essentially variations of existing RTE cereal products rather than innovations. Complaint counsel are thus raising a question of the consumer welfare significance of new product introductions. This is not a matter for my judgment in this case. Any product that expands the choices of a consumer is an innovation. The significance of an innovation is determined by its success in the marketplace. As stated in Finding 840:

There is no reason to believe that respondents have not attempted to come up with all reasonable innovations consistent with anticipated consumer acceptance. Neither complaint counsel nor I are in a position to pass judgment on what, if any, additional significant innovations the industry could or should have come up with. It cannot be found that (258)something more dramatic should have been developed, or that there are areas where innovativeness was neglected. The only area specifically alleged by complaint counsel to have been neglected is that of product fortification, and that allegation has been found to be unsubstantiated (footnote omitted). Complaint counsel assert that advertising cost levels in the RTE cereal industry are excessive and wasteful. While the advertising-tosales ratio for the RTE cereal industry is almost at the top of all manufacturing industries, RTE cereal industry advertising cannot be found to be excessive.

Advertising performs a necessary and legitimate function of advising prospective customers of the attributes of products offered for sale. The large number of nonhomogeneous RTE cereal products each with its requirement for separate advertising, dictates a relatively high A/S ratio. This requirement is further impacted by the high incidence of new product introduction which necessitates high levels of advertising.

It may well be that levels of advertising, to some extent, are being raised to achieve additional returns depending on the gross profit margins being realized (the Dorfman-Steiner principle). However, it is not improper for a manufacturer to increase advertising to the point of realizing all possible marginal returns. In the absence of a showing that respondents were earning supracompetitive profits their advertising levels, even to the extent they may be based on marginal returns, cannot be termed inefficient or wasteful. As summarized in Finding 870:

. . . I start with the basic assumption that advertising is a viable and legitimate method of competition. And complaint counsel have failed to demonstrate that the large volume of advertising in the RTF. cereal industry does not have a competitive basis or that it reflects the existence of supracompetitive profits. A firm advertises in order to generate or increase sales and so increase profits. At the same time, once profits are realized, or in anticipation of profits, the firm may inc/ease advertising up to the point of realizing all possible marginal returns. This, however, is pure theory . . .

Initial Decision and no findings can be made as to the extent respondents have increased normally competitive advertising levels to take advantage of the Dorfman-Steiner principle. It is not appropriate to mandate a "socially optimal level" of advertising as a substitute for one reached as a result of (259) competitive considerations. Our concept of a free competitive system does not envision imposition by government of permissible levels of advertising.

To establish the existence of a conspiracy to monopolize, the following elements are necessary: (1) the existence of a combination or conspiracy; (2) overt acts done in furtherance of the combi.nation or conspiracy; (3) an effect upon a substantial amount of interstate commerce; and (4) the existence of specific intent to monopolize. Cullum Electric Mechanical Inc. v. Mechanical Contractors Association of South Carolina 436 F .supp. 418, 425 (D. C. 1976), aff'd 569 F. 2d 821 (4th Cir. 1978), cert. denied 439 U.S. 910 (1978). Conspiracy is an agreement by two or more persons to pursue a common goal or objective having an unlawful purpose. Standard Oil Co. of California V. Moore 251 F.2d 188, 196 n. 3 (9th Cir. 1957), cert. denied 356 U.s. 975 (1958). However, a conspiracy need not be demonstrated by evidence of an express agreement. Comfort Trone Air Conditioning CO. V. Trane Co. 592 F. 2d 1373 (5th Cir. 1979); FTC V. Lukens Steel Co. 454 F.Supp. 1182 (D. C. 1978). Direct evidence is not required because of the general recognition that: seldom are the conspiratorial villains so devoid of cleverness as to broadcast their oral agreements or publicly circulate the written memos which describe their plan. Rutledge V. Electric Hose and Rubber Co. 327 F. Supp. 1267, 1274 (C. D. Ca. 1971), aff'd 511 F. 2d 668 (9th Cir. 1975). While the essence of conspiracy is agreement or a "meeting of the minds " tacit understanding can be found by mere acquiescence or acceptance of an invitation to participate in a plan, if each competitor knows that cooperation is essential to the success of the plan. Interstate Circuit, Inc. v. United States 306 U.s. 208 (1939); Wall Products CO. V. National Gypsum Co. 326 F. Supp. 295, 316 (N. D. Ca. 1971). If people with knowledge give adherence to a plan acquiescence is sufficient to establish a tacit agreement. Tacit agreement can be inferred from circumstantial evidence of conduct or a course of dealing even though there is no evidence of the acts by which the conspiracy was formed. Interstate Circuit, Inc. United States 306 U.s. 208 (1939); Overseas Motors, Inc. V. Import Motors Limited, Inc. 375 F. Supp. 499 (E.D. Mich. 1974), cert. denied 423 U.S. 987 (1975).

The use of circumstantial evidence to infer a conspiracy presents . . . .

Initial Decision 99 F. problems of proof. The evidence must create more than a suspicion of wrongdoing. Admiral Theatre Corp. v. Douglas Theater Co. 585 F. 877 , 884 (8th Cir. 1978). As stated in Overseas Motors, Inc. V. Import Motors Limited, Inc., 375 F. Supp. at 531: (260) There is a limit, however, to the degree of indirection and innuendo which the law will tolerate. Where. the plaintiffs case is based entirely on such circumstantial evidence, the court must be especially vigilant to insure that liberal modes of proof not become the pretext for unfounded speculation. Therefore, courts are cautious and consider the circumstantial evidence as a whole rather than viewing it in isolation. American Tobacco CO. V. United States 328 U. S. 781 (1946); Continental Ore CO. V. Union Carbide Carbon Co. 370 U.s. 690, 699 (1962). Consciously parallel conduct is circumstantial evidence that competitors have acted pursuant to a tacit agreement. Standing alone, however, parallel business behavior does not constitute violation of the antitrust laws. The Supreme Court has stated that: this Court has never held that proof of parallel business behavior conclusively establishes agreement or, phrased differently, that such behavior itself constitutes a Sherman Act offense. Circumstantial evidence of consciously parallel behavior may have made heavy inroads into the traditional judicial attitude toward conspiracy; but conscious parallelism" has not yet read conspiracy out of the Sherman Act entirely (footnote omitted).

Theatre Enterprises, Inc. v. Paramount Film Distributing Corp., 346 U.s. 537, 541 (1954).

The fact that conduct is parallel or occurred near in time may be an independent response to a common set of market factors. The coincidence of parallel decision making does not justify inference of tacit agreement if action is based on sound business judgment. Schoenkopfv. Brown Williamson Tobacco Corp. 483 F. Supp. 1185 1191 (KD. Pa. 1980), aff'd 637 F.2d 205 (3rd Cir. 1980); Hunt Mobil Oil Corp. 465 F. Supp. 195, 22B-29 (S. Y. 1978), aff'd, 610 2d 806 (2nd Cir. 1979); Harlem River Consumers Cooperative, Inc. Associated Grocers of Harlem, Inc. 408 F. Supp. 1251, 1278 (S. 1976). As stated in Independent Iron Works, Inc. V. United States Steel Corp. 177 F. Supp. 743, 747 (N. D. Ca. 1959), aff'd, 322 F.2d 656 (9th Cir. 1963), cert. denied 375 U.S. 922 (1963): The antitrust laws were not meant to prohibit businessmen from adopting sound business policies (261)merely because competitors had already adopted the same or similar policy.

As with other parallel conduct, uniformity of price, without more is not evidence of collusion. See United States V. FMC Corp. 306 F. Initial Decision Supp. 1106, 1117 (RD. Pa. 1969). Mere leadership or price followership based on independent decisionmaking violates no law. United States v. International Harvester Co. 275 U.S. 693, 708-09 (1927). Thus, it has been held that price uniformity alone is insufficient to establish an antitrust violation because it may be the normal result where a product is standardized or fungible, even though there is no agreement between competitors and the costs for the participating companies are not the same. See FTC v. Lukens Steel Co. 454 F. Supp. at 1190. Where there is an oligopoly involved in the production of a standardized product, such as salt, for which the demand is stable, it is inevitable that pricing policies of one company would influence the other. Therefore, other factors should be considered in addition to the parallel pricing. See, e. , Morton Salt Co. v. United States 235 F. 2d 573 (loth Cir. 1956).

Because evidence of parallel business behavior alone is not sufficient to support an inference of conspiracy, courts have required additional factors to show that the parallel decisions were interdependent. See Levitch v. Columbia Broadcasting System, Inc. 495 F. Supp. 649, 674 (S. Y. 1980). The additional circumstances which support the inference are: (1) a showing that the acts by defendants are in contradiction to their own economic interests and (2) a satisfactory demonstration of a motivation to enter an agreement. Venzie Corp. v. United States Mineral Products Co. 521 F.2d 1309 131'4 (3rd Cir. 1975).

Action against apparent individual self-interest is a strong indication of interdependence and tacit agreement among competitors. Modern Home Institute Inc. V. Hartford Accident and Indemnity Co. 513 F.2d 102, 111 (2nd Cir. 1975). Parallel conduct which is complex original, unanimous or corresponds exactly, especially where it is an abnormal reaction to market stimuli and inconsistent with each company s economic self-interest, strengthens the inference. Overseas Motors, Inc. V. Import Motors Limited, Inc. 375 F. Supp. at 535; Trist V. First Federal Savings and Loan Association of Chester 466 F. Supp. 578, 581 (E.D. Pa. 1979).

Motive to enter an agreement can be inferred from a pattern of conduct and can reasonably be drawn from the facts presented. First National Bank of Arizona V. Cities Services Cv. 391 U.S. 253, 287 (1968); American Tobacco CO. V. United States, 328 U.S. 781 (1946). Proof of motivation is evidence which makes Uthe inference of rational, independent choice less attractive than that of concerted action. Bogosian V. Gulf Oil Corp. 561 F.2d 434, 446 (3rd Cir. 1977), cert. denied 434 U. S. 1086 (1978). (262) The ultimate determination is whether the evidence reveals Initial Decision 99 F. conduct that can be explained by rational, independent business behavior or could only make sense in the context of the behavior of others. If the letter is true and the behavior only makes sense if each respondent' s competitors behave in a similar fashion, then the inference of concerted action may be warranted. FTC v. Lukens Steel Co. 454 F. Supp. at 1191. However, circumstantial evidence is not sufficient to establish a conclusion where the circumstances are merely consistent with such a conclusion o where they give equal support to inconsistent conclusions. Pevely Dairy Co. v. United States, 178 F.2d 363 , 370 (8th Cir. 1949), ccrt denied 339 U.S. 942 (1949).

If the proof supports the inference of tacit agreement, the burden then rests on respondents to explain it away or to contradict it. Interstate Circuit, Inc. v. United States, 306 U. S. 208 (1939); Milgram v. Loew's, Inc. 192 F.2d 579 , 584 (3rd Cir. 1951), cert. denied 343 U. 929 (1952). The inference can be overcome where there is direct evidence that the action was taken unilaterally for sound business reasons. See, e. , Feminist Women s Health Center, Inc. v. Mohammad 586 F.2d 530, 549 (5th Cir. 1978), cert. denied 444 U.s. 924 (1979).

In considering the implications of the respondents' behavior, it is importnnt to stress the Supreme Court's admonition that: . . . each case arising under the Sherman Act must be determined on the parti ular facts disclosed by the record, and that the opinions in those cases must be read in the light of their facts and of a clear recog-nition of the essential differences in the facts of those cases, and in the facts of any new case to which the rule of earlier decisions is to be applied.

Maple Flooring Manufacturers Ass V. United States 268 U.s. 563 579 (1925).

In the instant case, there has been a total failure to demonstrate pricing coordination among respondents. There is, therefore, no coordinated pricing activity concerning which the possibility of tacit agreement may be considered.

Complaint counsel have failed to prove the alleged coordination or agreement among respondents with respect to trade deals, cents-off deals, in-pack premiums, private labeling, advertising, competition for retail shelf space or fortification of cereals. In any event respondents have established that their actions were independent responses to similar economic and market conditions or to particular conditions facing an individual respondent. (263) The only coordinated activity alleged and proved was the exchange of advertising data. However, the data so acquired were not utiized in a noncompetitive fashion. The exchange of non price information Initial Decision alone is not a violation of the Sherman Act. FTC v. Lukens Steel Co. 454 F. Supp. at 1191-92. See Maple Flooring Manufacturers Ass United States 268 U.S. 563, 582 (1925).

It has been concluded, therefore, that respondents did not violate Section 5 of the Federal Trade Commission Act by reason of conspiracy.

Complaint counsel have presented their shared monopoly theory under a structure-conduct-performance approach, and have concedconstitute aed that an oligopolistic structure alone docs not violation of Section 5; that it is also necessary to show the existence and exercise of monopoly power.

Structure consists of two elements: (1) degree of concentration and (2) extent of barriers to entry. The RTE cereal industry is one of the most highly concentrated industries in the United States. However there has been a failure of proof that the industry members have been enjoying monopoly profits. a condition under which it may be argued that there would have been entry if there were no barriers to entry. Further, it has been found that respondents may not be held responsible for the lack of entry into the RTE cereal industry. As for conduct, there has been a lack of proof that respondents have engaged in the coordinated conduct alleged as part of complaint counsel's theory of violation.

This brings us finally to performance where complaint counsel have failed to prove the monopolistic performance alleged as part of their theory of violation that respondents earned supracompetitive or monopoly profits; that respondents have not been innovative in new product development; that respondents have been wasteful in advertising expenditures; and that respondents' prices have been so high that there has been a lessened demand for and supply of RTE cereals with a resultant waste and misallocation of society s resources.

Section 5 of the Federal Trade Commission Act was designed to reach unfair methods of competition which do not achieve the level of Sherman Act violaions. FTC v. Motion Picture Advertising Service Co. 334 U.S. 392 (1953); Fashion Originators ' Guild of America, Inc., v. FTC 312 U. S. 457 (1941); FTC v. Cement Institute 333 U.S. 683 (1948).

The scope of Section 5 is not bound by Section 2 of the Sherman Act. Therefore, conduct which threatens an incipient violation of (264)the Sherman Act can amount to a Section 5 violation. See, e. .. .. .. . . 268 FEDERAL TRADE COMMISSION m:CISIONS Initial Decision 99 FTC. G. Balfour Co. v. FTC, 442 F. 2d 1 (7th Cir. 1971); Borden, Inc., 92 C. 669, 781 n. 4 (1978). The Supreme Court has stated that the Federal Trade Commission has "authority to consider public values beyond simply those enshrined in the letter or encompassed in the spirit of the antitrust laws. FTC v. Sperry Hutchinson Co., 405 S. 233, 244 (1972).

The Supreme Court has set some parameters as to what kinds of commercial conduct can constitute a Section 5 violation, declaring: The point where a method of competition becomes ' unfair' within the meaning of the act wil often turn on the exigencies of a particular situation, trade practices, or the practical requirements of the business in question. FTC v. Motion Picture Advertising Service Co. 344 U.s. , 396 (1953).

Conduct has been found to be unfair in violation of Section 5 if it has the effect of increasing monopoly power to the detriment of competition. See, e.g, Sugar Institute, Inc. v. United States, 297 U.s. 553 , 598 (1936); Maple Flooring Manufacturers Ass v. United States 268 U.S. 563, 586 (1925). As Administrative Law Judge Brown has recently stated:

Section 5 of the Federal Trade Commission Act can be invoked to effect structural changes in an industry only where it is clearly demonstrated that the competitive disequilibrium is the result of some conduct that could be designated as ' unfair.' If the challenged conduct is not unreasonable and not the cause of the trend toward monopoly power, no violation of Section 5 exists, merely because . the effects upon competition may be undesirable from an economic point of view. 1. Dupont de Nemours 3 Trade Reg. Rep. n 21 613 at 21 750 aff'd by the Commission, n 21 770 (1980).

In Boise Cascade Corp. v. FTC 637 F.2d 573 (9th Cir. 1980), the court considered whether industrywide use of a delivered pricing system was a collusive practice to fix prices and whether it could be condemned under Section 5. The court refused to resolve the question whether conscious parallelism might ever support a Section 5 violation. However, it did find that: (265) in the absence of evidence of overt agreement to. avoid price competition, the Commission must demonstrate that the challenged pricing system has actually had the effect of fixing or stabilizing prices. Without such effect, a mere showing of parallel action will not establish a Section 5 violation. Id. at 577.

Finally, the court noted that:

where the parties agree that the practice was a natural and competitive development in the emergence of the southern plywood industry, and where there is a , .

Separate Statement complete absence of evidence implying overt conspiracy, to allow a finding of a Section 5 violation on the theory that the mere widespread use of the practice makes it an incipient threat to competition would be to blur the distinction between guilty and innocent commercial behavior.

Id. at 582.

It would serve no purpose to consider in a vacuum what factual showing, if any, of industry structure conduct and performance would constitute a ' shared monopoly" violation of Section 5, and might justify an order restructuring an industry. Such a showing has not been made here.

FINAL CONCLUSIONS OF LAW 1. Respondents Kellogg Company, General Mills, Inc. and General Foods Corporation were, at all times material herein, corporations engaged in commerce, as "commerce" is defined in the Federal Trade Commission Act.

2. The complaint does not encompass a charge of conspiracy. Counsel supporting the complaint have failed to sustain, the burden of establishing that respondents, or any of them, have violated Section 5 of the Federal Trade Commission Act. This finding of failure of proof relates inter alia to the issue of conspiracy, which was tried though not covered by the complaint. (266) ORDER It is ordered That the complaint in this proceeding be, and it hereby is, dismissed as to all respondents. ORDER DENYING APPEAL AND VACATING INITIAL DECISION The Commission has determined not to hear further appeal of this matter. The Commission has also determined that the Initial Decision shall not become the final decision of the Commission. Thus It is ordered That, the Initial Decision in Docket No. 8883 be vacated in its entirety, and the Commission s complaint in this matter be, and it hereby is, dismissed with prejudice. Commissioner Pertschuk dissented as to denial of the appeal. SEPARATE STATEMENT OF COMMISSIONER CLANTON This case raises important issues concerning the application of Section 5 of the FTC Act to oligopolistic conduct. Because of these Separate Statement 99 FTC. issues and the present posture of this case-Administrativc Law Judge Berman s decision in favor of respondents coupled with the Bureau Director s decision not to appeal-it seems highly desirable for the Commission to determine now whether a full briefing on the merits is warranted.

Of course, even in the absence of an appeal, the Commission has the right under Section 3.53 of the Rules of Practice to undertake a thorough review of the recorq. However, I believe the circumstances of this case justify an exception to that practice. Given the theories of liability and proposed relief under consideration, it is entirely proper for us to see if there is a likely basis for issuing an order, even if the facts conform closely to what complaint counsel contend. After having reviewed the parties' most recent submissions, in response to the Commission s order of December 18, 1981 , as well as other parts of the record, I cannot find a basis for continuing the case. In its most succinct form, complaint counsel urge that liability be premised on the basis of two related but distinct theories. The first is a traditional conspiracy to monopolize based upon the principles contained in Section 2 of the Sherman Act; the second is a shared monopoly theory under Section 5 of the FTC Act, a theory which does not depend upon a showing of collusion. Under either of these theories, complaint counsel argue that the only effective form of relief would be a divestiture order, including royalty-free licensing of respondents' cereal trademarks.

As to the first theory, I agree with AU Berman that a conspiracy to monopolize was not properly pled. As for the separate shared monopoly theory, I do not believe such a theory, however characterized, can serve as a predicate for the Commission to restructure an industry, at least in the absence of clear predatory behavior, which is not claimed here.

I do want to emphasize, however, that Section 5 may well provide the Commission with suffcient authority to attack non-collusive behavior that contributes to or enhances anticompetitive conduct and which is without compelling business justification. In such circumstances, the principal remedial tool for dealing with this kind of behavior would be a conduct order. (2) Before elaborating further on these points, I would offer a comment about the characterizations, or mischaracterizations, that have been advanced in the past by critics of this case. An awful lot of rhetoric has been spilled on this subject, with some critics claiming that the case is just the first step in a broad-based attack on concentrated industries. Others have accused the agency of attacking competitive forms of behavior, such as product differentiation Separate Statement, and brand proliferation; while still others have derided the "shared monopoly" concept, suggesting that this is evidence alone of the Commission s confused thinking since the description itself is a contradiction in terms' Suffice it to say, I do not share al1 of the views of the critics of this case. In issuing the complaint, I think the Commission sought to address a legitimate concern, not about oligopolies per se but rather about oligopolistic behavior that is uniquely anticompetitive. Respected antitrust commentators of different persuasions-such as Professor Posner and Professors Areeda and Turner-have advocat ed different approaches for dealing with collusive-type behavior among oligopolists.' Whether the theories of relief proposed by complaint counsel are proper is one thing, but it is clear to me that the Commission was not attempting through this case to challenge structure or bigness per se.

Conspiracy to Monopolize As one of their principal prongs ofliability, complaint counsel now contend that an implied conspiracy to monopolize can be inferred from respondents' course of dealing over the past twenty years. It is contended that respondents have consistently eschewed various forms of price competition and channelled their energies instead into promotional activites and brand diversification. Whether this conduct gives rise to a conspiracy to monopolize can be addressed only if a conspiracy in fact was properly tried. (3) In finding that a conspiracy theory was not part of the case, ALJ Berman looked to the language of the complaint and statements made by complaint counsel during the prehearing conference stage of the proceedings. In particular, he found the complaint ambiguous since it didn t mention specifically that either an agreement or conspiracy was at issue. Rather, the complaint charged, inter alia that respondents "individually and collectively, have obtained , Se P. Are"da & 11 TUrf1 , HI Anliln.. 1 Law Ch. 8E (1978); R. Posn An(ilr-!1. t Law: An Economj,. J'erspedive 39- 77(1976) " Compaint crmnsel's conspiracy case depends primarily upon" showing that respondents' roughly parallel behavior on several fronl institutinli pri e char-ges topping trade de"J , eliminating" in-pack premiums refusinr;to sell to private labelers, fortification veratwentyyearperiodcannot bee"plained aWlIyasa coincidence of indcpendcntjudgmenb. While it is unnecessary to delve inlQ this atissueany great length, since the conspiracy theory, as di cussed hereinaftN, i not properly before us. 1 would observe that this kind ofevid€rlli"ry approach to proving collusion is not unprecedented. C""ws suchAmericanas Tobacco Co. v. Uniled Siaks. :!28 US 781 (1946) and Wall Produr:ts, v. National Gypsu.m Cu. 326 F. SUpp. 295 (N.D, Cal 1971), indicait quite cleady that conspiracy Clln be inferr from proof of parallel actions coupled with economic conditions that "appear to negate ,m inference of innocent, independent business conduct. This point is not to suggest that the case before us unquestionably involves a conspiracy, for, among other things, the evidem:e of imultaneity of lwhaviDr.(e,c" price leadership) lwre is perhaps not as stronf( as it was in th' aforementioned rases. What it does suggest isthllt a con piracy theory is nut impotent to d aj with o1igopojj tic behavior that exhibits more than a casual trend tow,nd int"rdependent.condud Separate Statement 99 F. shared and exercised. . . monopoly power in, and have monopolized the production and sale of RTE cereal. . ." (Complaint, n9B). More importantly, ALJ Berman noted that complaint counsel had emphatically denied on several occasions early in the proceedings that the complaint contained a conspiracy count. (LD. at pp. 3-10) Thus although the first ALJ (Judge Hinkes) subsequently allowed the case to be tried on a tacit conspiracy theory, ALJ Berman found that procedure deficient since the complaint was not amended by the Commission and respondents did not acquiesce to that procedure. (See Rules of Practice Section 3. 15) (4) A review of the record indicates that, at best, the conspiracy pleading issue is murky. Obviously, we should not engage in excessive nitpicking of pleadings or overly technical interpretations but the pleadings are not a model of clarity, especially in light of the shared monopoly theory being alleged. Words such as "combination and "collectively" are used in the complaint, but those words are also consistent with a shared monopoly theory that requires no showing of collusion. Nevertheless, the pleadings might have been adequate to encompass a conspiracy to monopolize theory had it not been for complaint counsel's insistent denials that a conspiracy was at issue in the case.

During the prehearing conferences in 1972, ALJ Hinkes on more than one occasion sought further clarificaton of the pleadings from complaint counsel in view of their simultaneous denial of conspiracy and assertion that the case focused on joint, interdependent conduct. (Tr. 17, 25, 71, 104) After several tries, complaint counsel finally indicated that the case could encompass a tacit agreement (Tr. 106- , 163--4), but only after having previously indicated that no traditional conspiracy was being alleged.' Of course, as ALJ Berman noted, traditional conspiracies, under either Section 1 or Section 2 of the Sherman Act involve both express and tacit agreements. Thus, it seems hard to conclude that the case as originally envisioned, encompassed a traditional Sherman Acttype conspiracy. Yet, that is precisely what complaint counsel now contend has been established in this case-a traditional Section 2 Sherman Act conspiracy to monopolize. (Complaint Counsel's Proposed Conclusions of Law, Vol. IV at 655- , Sept. 30, 1980) Even giving the benefit of the doubt to complaint counsel, it is hard to understand why it was necessary to dance step around the conspiracy issue if the case clearly covered it. After all, the major case now j Although the p!eadirIg issue was not before the court, it is interesting to note that the D-C. Circuit Court of Appeals, in II subpoena enforcement action brought four years after the case was filed, observed that a conspiracy was not allo;gcd in t.hO; complaint. FTCv. Lvnning, 539 F. 2d 202, 204, n, 3 mc. Cir, 1976) Separate Statement relied upon by complaint counsel American Tobacco Co. v. United States 328 U.S. 781 (1946), which was decided twenty-five years before this case was brought, involved an implied conspiracy to monopolize. While there have not been many implied conspiracy cases brought under Section 2, there certainly have been numerous such cases under Section 1, and complaint counsel, in fact, cite to such precedent in support of their case. (5) Perhaps, in denying that a conspiracy was at issue in the case, complaint counsel may have been attempting to distinguish between Section I-type conspiracies and Section 2 conspiracies. There is however, no evidence of that and, in any event, such a distinction would appear to be unnecessary since the type of proof needed to establish a conspiracy would appear to be much the same regardless of which provision of the Sherman Act was being pleaded. See Areeda & D. Turner supra note 1, n839. Alternatively, complaint counsel may have felt that the kind of tacit conspiracy they had in mind would not rise to the Sherman Act level, but if that is the case then the issue is little different from the pure Section 5 shared monopoly theory discussed below. There can hardly be two kinds of conspiracies, one suffcient to establish Sherman Act liability and another sufficient for Section 5 purposes. To be sure, there have been suggestions that the scope of the conspiracy theory under the Sherman Act could be broadened to give greater emphasis to the use of economic evidence in determining whether a tacit conspiracy exists. See R. Posner supra note 1, at 71- 77. Still, the issue is whether a conspiracy exists, not whether some other theory, by a different name, is viable. Although AW Hinkes ultimately allowed the conspiracy charge to be tried, it is clear that the parties did not agree on this procedure and that no amendment to the complaint was sought from the Commission. Accordingly, I do not believe the Commission can hinge liability on a conspiracy theory.

Section Shared Monopoly Even if a conspiracy count is not present, that does not end the matter. Complaint counsel alternatively argue that even absent a conspiracy the conduct is sufficiently like one to justify a finding of liability under Section 5 of the FTC Act. It is quite clear, of course that Section 5 can reach anticompetitive behavior that is not covered by the Clayton or Sherman Acts. And, I believe such authority Separate Statement 99 F. extends to non-collusive, marketwide behavior that may not involve traditional forms of predation. Presumably, this could include behavior that would not be ilegal for a single firm to engage in but due to the industrywide nature of the practice, could lead to significant anticompetitive effects. (6) Various commentators have also urged that the antitrust laws are flexible enough to deal with shared monopolies or oligopolistic behavior outside the context of a traditional conspiracy. ' In highly concentrated industries competitors may learn to react to each others' moves in a fashion that is closely analogous to the workings of a cartel. Firms will recognize that it is not in their self-interest to chart an independent course because other competitors will be able to quickly detect and match their moves, thereby leading to lower profits for the industry as a whole. The degree to which this phenomena occurs, and its success, are obviously subjects of great debate and the identification of markets in which firms are operating in a closely interdependent fashion is admittedly complex. Because of the difficulties in proving collusion, Professors Areeda and Turner have advocated a different approach to the problem of single firm monopolization and shared monopolies. In their recent treatise, they suggest that evidence of persistent monopoly performance in a market, whether exhibited by a single firm or a small group of firms, should be sufficient to justify sweeping relief in the form of divestiture or other like remedies. Of course, they would require fairly strong evidence that the market is performing badly and that structural remedies would not lead to inefficiencies. They also believe that such actions should be limited to government initiatives and not allowed in private suits. In essence, their proposal does not depend on improper conduct for establishing liability and is akin to various no-fault monopoly proposals that have been advanced in the past.

But even Areeda and Turner recognize that the kinds of markets that might warrant intervention under such an approach are limited, even more so in the shared monopoly area than for single firm monopolies.

Complaint counsel, of course, are not advocating a no-fault approach in this case. Instead, they have attempted to develop their case in a way that focuses on the extent to which respondents interrelated behavior has exacerbated competitive problems in the RTE cereal industry. Their theory depends less on the unreason . Se, e. Boise Ca. cade Corp. , el a/. 91 F. C. 1 (1978), rely ,m ulher gro!1nw,'. Ruise Cascl.de rrJrp. v, FTC 637 2d573(9their. 1980).

, E.g. P. Areed" & D. Turner supra note 1; L. Sullivan Handbook ofthe Law of A"tilruM Section 125 (1977). . Se p, Areeda & D. Turner supra note 1 845. Separate Statement bleness of specific forms of behavior than it does on the totality of the conduct. (7) Nevertheless, even under complaint counsel's theory, one must recognize the implications of using such an approach to restructure an entire industry. As complaint counsel acknowledge, the kind of theory and relief they are seeking require extensive proof of industry structure, performance and conduct. While that kind of analysis is highly commendable, it provides a less than certain guide as to what kinds of conduct or market conditions would be subjected to antitrust attack. In addition, such an approach, of necessity, dramatically limits the number of instances where market intervention is warranted and, even then, it does not fully remove the risks associated with developing a structural remedy for an industry. Thus, absent collusion or clear evidence of predatory behavior, I believe it would be unwise for the Commission to seek dissoultion of an industry on the basis of the cumulative effects of multi-firm behavior. That does not mean, however, that such behavior would go unaddressed. Rather, it means that the kind of relief soughtnamely, conduct remedies-would reflect two realities about the oligo polis tic market context: (1) the lower probability that serious anticompetitive problems wil exist for long, and (2) the potentially greater costs of attempting to restructure an industry. Because of the complexity involved, it may take many years to recover the costs of obtaining and implementing a successful dissolution order. Those costs may be worth incurring where we can be fairly confident that the market behavior under attack cannot be justified. The rigor of conspiracy analysis can help to provide that assurance, and strong evidence of predatory behavior may also provide the necessary predicate for divestiture.

But to pursue structural relief in less compelling circumstances carries with it too great a risk of wrong or imperfect judgments. The alternative, conduct relief, obviously has its limitations. Such a focus, however, enables more precise judgments to be made about the reasonableness of particular behavior without the risk of overkill.' To (8)be sure, conduct relief may not be feasible in all oligopolistic market settings, but neither is divestiture. After all, antitrust deals primarily with probabilities. Where the probable benefits of improving competition are very high, as they are in breaking up a , Apart from their suggested approach for dealingwith per istent shared monopoly, Areeda and Turner also have expressed the view that the antitrust laws even more clearly can reach exclusionary behavior engaged in by shared monopolists. Such Ii theory would, they believe, amount to a logica! extension of attempted mO!1opolization since the prim!lry missing ingredient would thebe absence of a dangerous probability of a single firm monopoly. P. Areeda & TUrfer supra note 1 856-1. Whether the courts woud be willng to extend Section 2 of the Sherman Act this far is not clear, hut, in any event, such an approach would appear to be within the scope of 5 of the FTC Act Separate Statement 99 F. horizontal price-fixing conspiracy, there is little risk that the imposition of harsh sanctions will chill desirable competitive behavior. But, as the difficulty of distinguishing between harmful and beneficial conduct increases, so does the danger of imposing maximum remedies. Therefore, in my view, remedial restraint is called for, even in the kind of tightly concentrated market presented to us here, where profitability is good, market shares are stable and new entry has been minimal.

Although I am opposed to structural remedies, I reiterate my belief that the Commission can reach non-collusive, industrywide behavior under Section 5 of the FTC Act. An example of such an approach is the Commission s decision in Boise Cascade 91 F. C. 1 (J 978), rev, Boise Cascade Corp. v. FTC 637 F.2d 573 (9th Cir. 1980). There, the Commission found that the industrywide use of an artificial freight factor contributed to price stability and could not be justified by market exigencies. Although the Ninth Circuit Court of Appeals felt that the Commission did not make a satisfactory showing of anticompetitive effect, the appellate decision did not foreclose the possibility that the Commission could employ Section 5 to reach unjustified forms of non-collusive behavior which are practiced on a marketwide basis. It should also be emphasized that the Commission in Boise Cascade believed that it could fashion an effective conduct order that would not be highly regulatory in nature. If conduct relief can be easily evaded, or requires extensive government intrusion in ongoing market operations, there is little reason to press ahead. And, the absence of effective relief may suggest that there are offsetting competitive justifications for the practices under scrutiny.

As reflected in the. Boise Cascade decision, the Commission may address multi-firm behavior that facilitates non-competitive conduct as well as more exclusionary forms of conduct. On the other hand simply refusing to compete, by itself, probably would not be subject to attack-to do so would be analogous to going after a passive monopolist. However, passivity in an oligopolistic context is harder to maintain without resort to collusion or other practices that may contribute to a stable market.

In this instance, assuming complaint counsel's case were to be established, several practices might be singled out for (9)possible action. For example, a central issue in the case is brand proliferation. Complaint counsel argue that respondents have engaged in excessive product differentiation as a less disrupti-;e form of competition than price competition. It is claimed that by carving up the market into smaller and smaller product segments, respondents Separate Statement have made it extremely difficult for new firms to enter, since an entrant would have to offer several brands to achieve minimal scale economies. Moreover, a new firm could not expect established competitors to give much ground, as might be the case if limit pricing were being practiced, because of the difficulty of reposition ing brands.

Without getting into an extended discussion of these allegations there does seem to be considerable evidence that brand proliferation has made entry more diffcult. But that is not the end of the analysis. Whatever the social value of these products, we are not dealing with the kind of product design or change that is introduced primarily as a blocking device to discipline competitors." In other words, we are not talking about predatory conduct that serves little, if any, legitimate competitive ends. For the most part, the myriad cereal brands on the market are self-sustaining and they appear to generate significant consumer demand. Even if we would prefer to see fewer brands and more price competition, it would be extremely difficult to distinguish between legitimate and ilegitimate brand proliferation. Certainly, it would be quite inadvisable and impractical to attempt to limit advertising expenditures or new brand offerings. Thus, an order provision directed to this practice does not seem very promising.

A second practice that might be susceptible to correction concerns respondents' shelf space recommendations. to grocery retailers. These plans rely largely on past market shares as the benchmark for allocating space. Complaint counsel contend that these recommendations, which many retailers have adhered to in principle, tend to stabilize competition among existing competitors and make it more difficult for new entrants to get shelf space. Indeed, this is the only practice for which complaint counsel have sought conduct relief. (10) It could be argued that an order restricting or preventing respondents from making shelf space recommendations would help to inject more competitive pressures into an important area of nonprice competition, without intruding unnecessarily into respondents day-to-day business judgments. On the other hand, this type of activity is undoubtedly normal commercial behavior that is engaged in by many other food manufacturers, although it is not clear whether the nature and pattern of recommendations in the cereal industry are followed in other markets.

" Allegations of this nature have he.'IJ made in several recent cases involving:the communications and comput"r industries.Northem;/ern Tel. Co.v, American TPI. Tel. Co.. 497 F. Supp. 230, 244 (D. Conn. 1980),rev 651 1".2d 76, 94-95 (2d Cir. 1981);Berkey Photo, Inc.v, Easlman Kodak, Cu. 60:1 F.2d 263, 2Hf, (2d Cir. 1979), cert. denied 444 U.s 1093 (1980); Cui. Computer Pmduds v. I.BM. Corp. 61:1 F.2d 727, 744 (9th Cir. 1979); Tran. omerit' Computer Co. Inc. I.HM. Corp. 4811". Supp. 965, 1006---8 (N . C,,)1979). , Separate Statement 99 F.T.C. However, regardless of whether a workable remedy could be crafted on this subject, it is questionable whether the issue is all that important from a remedial standpoint. Dr. Schmalensee, one of complaint counsel' s expert witnesses in the case, expressed the view in a separate article that shelf space plans were probably not powerful deterrant devices. Schmalensee Entry Deterrence in the Ready-to-Eat Breakfast Cereal Industry," 9 Bell J. Econ. 305, 307, n.4 (1978). A third area for possible relief involves the exchange of recent advertising expenditure data among respondents through the vehicle of a third party reporting service. It is not entirely clear whether this particular practice, specifically with respect to the accuracy and currency of the data exchanged, is unique to the cereal industry or occurs frequently in other industries. While this exchange makes it easier to monitor the actions of competitors it may be less sensitive than an exchange of price information, since the quantity of advertising expenditures does not necessarily reveal the effectiveness of those expenditures. Of course, the effectiveness of restricting this practice depends on how easily and quickly respondents could obtain similar data through other means. But, even if an order it is restriction would make that task substantially more diffcult, not at all clear that it would inject a very significant destabilizing force into the market. To be sure, advertising is a major factor in respondents' non-price competition, but inducing more rivalry in this area is not necessarily calculated to produce similar spin-off effects in the pricing of RTE cereals.

Another candidate for reform is respondents ' fairly consistent refusal to supply private brand cereals to retailers. While increased private brand competition could bring about more price competition an order requiring respondents to supply such product is fraught with all kinds of problems, and could easily lead to a highly intrusive regulatory-type order. (11) Finally, with respect to other allegedly anticompetitive conduct, such as respondents' refusal to offer trade deals or other off-list discounts, it is quite clear that the Commission cannot mandate respondents to compete. Such behavior, if it does not involve collusion, represents the kind of passive noncompetitive behavior noted above for which there is probably no practical enforcement remedy.

Thus, a review of possible avenues for conduct relief suggests that assuming a Section 5 case can be made out, the available remedial alternatives are either intrinsically undesirable or hold little promthis task is made ise for producing beneficial results. To some extent, , ! Dissenting Statement more diffcult by the fact that there is no record for determining which, if any, conduct restrictions would be useful, since the case from the outset focused almost exclusively on structural remedies. Nevertheless, despite this limitation, I am inclined to agree with complaint counsel that if any relief were suitable in this case, it should probably be structural in nature. However, for the reasons discussed previously, I am firmly opposed to seeking dissolution in a case of this kind without collusion, or possibly predation, as a legal predicate. While we might desire a better mix of price and non-price competition in the RTE cereal industry, the potential costs associated with a divestiture order, not to mention the difficulty in getting a court to approve such an exercise of our remedial discretion, lead me to reject this approach. I am also simply not persuaded that the class of cases reflected here is sufficiently large to warrant pursuing this kind of complicated, time-consuming remedial avenue. Even if this is the one case in a thousand that might justify such an approach, I do not feel that we ought to apply our Section 5 powers in ad hoc lawthis way, for what would be an essentially one-time enforcement initiative.

Accordingly, it is my belief that the Commission should not pursue this case further.

DISSENTING STATEMENT OF COMMISSIONER PERTSCHUK The Commission today takes an unprecedented step in refusing to hear the appeal of this matter. This decision raises serious implications for the integrity and propriety of Commission adjudicatory procedures. The complaint in this matter was issued in 1972 and it has taken nine years to complete pretrial procedures and the trial itself. The case raises difficult and unanswered legal questions as well as vigorously disputed factual controversies. Legal conclusions about the allegations in the complaint would have important ramifications for the applicability of the antitrust laws to concentrated industries which do not operate competitively. Thus, the case is precisely the kind of matter that warrants full-scale review by a responsible Commission, charged by Congress with adjudicatory determinations. I cite no less an authority than the Commission itself: "In this as in any other case . . . it is the agency s function not the Examiner, to make the findings of fact and select the ultimate decision, and where there is substantial evidence supporting each result it is the agency s choice that governs. , Order Denying Motion for Dismissal of Complaint Counsel's Appeal and for Adoption of Initial Decision Docket No. 8883 (Nov. 3, 1981), citing Greater Bos/on Tdeuision Corp.v. FCC 444 F.2d 841 , 853 (D.GCir. 1970). Dissenting Statement 99 F. Yet the Commission has precipitously determined that it will not grant full-scale review of this matter, apparently on the grounds that it would be a waste of resources. The resources at stake apparently are the costs of a round of briefs concerning the administrative law judge s opinion, an oral argument, and the preparation of an opinion-not trivial, I grant, but not significant compared to the length, complexity and importance of this matter. It cannot be ignored that this case has been controversial and that Congress has expressed concern about it. The respondents have attempted to make much of the fact that the Commission, at one point, offered a contract to the original trial judge to continue the matter after his intended retirement. In my view, this step by the Commission raised no significant question of impropriety, and the respondent' s zeal in pursuing it has essentially been an aggressive legal and political maneuver. Respondents have also engaged in (2) intensive lobbying efforts in Congress to accomplish the premature demise of this case. I do not question their right to pursue either of these tactics. I do question, however, the propriety of Congressional intervention in any matter before it has run its course of proper adjudicatory procedures. I also question whether Congress has ever been furnished with a complete analysis of this difficult case and the legal arguments that have been raised in it. I view the prior Congressional intervention in this matter, the vigorous efforts by respondents to heighten controversy about it, and the spurious characterizations of the Commission s effort to carry out its administrative responsibilities by offering a contract to the original trial judge as an unfortunate backdrop to this decision by the Commission to stop review of this matter in midstream. This case was argued on the basis of two theories-I) a conspiracy based on traditional Sherman Act Section I principles and 2) a theory of interdependent behavior in a highly structured industry with poor competitive performance and where industry members have engaged in exclusionary conduct, what has come to be known in somewhat misleading shorthand as a "shared monopoly." If the case had been appealed, I would, of course, have carefully considered the conspiracy argument, which if proved, would certainly violate longheld principles of the antitrust laws. Based upon my tentative review, I am inclined to believe a finding of a traditional conspiracy could not fairly serve as a basis of liability.' I emphasize this conclusion is only tentative, and I would have wished to have heard further arguments by the parties on this issue. I base this conclusion on my interpretation of the complaint, which doe s n(Jt appear to allege a conspiracy. rather than the evidentiary record, which may well show a conspiracy existe Dissenting Statement As to the shared monopoly theory as initially pled, it is important to note what the theory is and what it is not. It is not an allegation of Conse-an agreement, either an expressed or implied conspiracy.3 quently, it does (3)not fi the conventional analysis applied in Sherman Act Section I-type matters. Nor is it an allegation of practices by individual companies which, under the circumstances of the industry, facilitate uniform pricing or other anticompetitive behavior. Finally, it is not an allegation that each firm engaged in exclusionary practices which independently violate Section 5 by unfairly raising barriers to new entry or driving out equally efficient competitors. Although there are aspects of these theories which were pled in the complaint, it is more appropriate to view the shared monopoly theory as standing apart from them and not requiring the same elements necessary to show a violation under these other theories. It is important to note, however, that these theories, which like the shared monopoly theory, are attempts to define ilegal behavior in oligopolistic markets, are important to effective antitrust enforcement and are not undermined in any way by today s action by the Commission.

The shared monopoly theory, as reflected in the Commission complaint in this matter, was predicated upon an allegation of high concentration, as evidenced by a three-firm concentration exceeding 80%; poor competitive performance, as measured, for example, by sustained high profits and the absence of price competition; and high barriers to entry caused by exclusionary conduct of industry members 4 as evidenced, for example, by the absence of significant new entry since 1950. Thus, the theory of the case does not condemn the (industry) structure itself' as the Bureau Director statement of December 11 supposes.

Such a theory is supported by scholarly commentary, including that of Professors Areeda and Turner ' Professor (4)Sullvan' and others. Thus, it is not the case that "the theory has. . . utterly failed to enter the mainstream of economic thought " as the respondents claim. Rather, this case represents a serious, carefully thought out , A charge of conspiracy to monopotizecould ofcourse bebased upon t acit(:oJlusion where joint action stems from assent by actioo rather than words.Intp state Circuli v.US. 306 U.S, 208 (1939). Consequently, it would not have been unreasonable to plead a conspiracy based on the racL alleged in the complaint, though asstatedearlier IdonotbdicvetheCornmissiondidso . Exclusionary practices need not.in themselves be indefH'ndcntly unlawful or predatory Loconstitute acts of monopolization, See, e,g, Us. Y. Griffith 334 U.s. 100, 105 (1948); American TobmnJ Co. v, u.s. :,28 U.s: 781 , 809 Hl4 (19461. Thus, each of the respondents need not have engaged in unfair predation under Section 5 or an attempt Lo monopolize under Section 2 of the Sherrnan Act in order to have bf-en found to h..ve engaged in monopolizing We do not deal here with an allegilt;on that there has been monopoly power shared by three firms without rnonopo!izingconduct.

, See 3 p, Areeda and D, Turner Antitrust Law 359-390 (1978) , &eL.Sullivan Antitn,st3fil-3fi6(1977J.

!;.

Dissenting Statement 99 F. attempt by a no-nonsense Republican-led Commission in 1972 to deal with the problem of a tight oligopoly and a poorly performing industry.

Today, the Commission turns its back on this attempt, not wishing to deal with the difficult but necessary task of spelling out whether and under what circumstances the antitrust laws reach this problem. Such a step by the Commission is a significant one, with major ramifications for government antitrust policy. We should make no mistake about it: the problem of high concentration-industries operated by a few giant companies with poor competitive performance, as indicated by the absence of meaningful price competition and the absence of significant entry of new competitors over a long period-is not going to disappear from our economy in the coming decades. Our economy is now made up of a number of highly concentrated industries without meaningful price competition and if the merger laws are not to be enforced vigorously, this situation will become more frequent, not less.

I for one believe that Section 5 of the Federal Trade Commission Act does reach a situation where an industry is highly concentrated; the performance of the industry as measured by profit levels, lack of price competition or other factors, is poor; effective barriers to entry are created by exclusionary conduct on the part of the firms; and a government-ordered remedy can be shown to be likely to improve competition. I also believe that it is possible for this Commission and for the courts to identify, after careful study, which industries are appropriate for restructuring in order to deal with the problem, and which industries are not.' But I also conclude that the prospect for some future Commission effectively to apply this theory is highly unlikely. It is not that there will not be farsighted and courageous Commissions in the future, nor certainly that there will be an absence of careful economic analysis capable of identifying (5) industries which should be addressed; nor do I view this decision by the Commission today in any way as a legal precedent which deserves to be followed by a future Commission or by the courts. Rather, I view today s decision as confirmation of the political inability of a Commission to see such a case through to the end. As our political system provides, the Commission reflects, to a large extent, the prevailing political attitudes and the economic philosophy of the current administration. And, quite properly, future Commissions will reflect the then-prevailing political philosophy. Unfortunately, an attempt by the inherently lengthy process of , In Some situ"tions. II conduct:-oriented remedy "lone mlly be ade'1u,de by prohibiting certain e"ell1 ionary pr"dice Statement litigation to deal with tr.e oligopolistic problem I have described requires a political consensus that an independent commission is legitimate and competent to carry out the task, and a political environment which gives it the room and time to carry it out. Today s decision seems to me to tell us that such a consensus unlikely. Therefore, I believe strongly that Congress, not this one perhaps, but some future one, should brace itself for the task of spelling out in careful, responsible legislation what government' role is in dealing with the problem of oligopoly. And I emphasize again that it is a problem which is destined to become more, rather than less, significant for our society.

STATEMENT OF COMMISSIONER PATRICIA P. BAILEY This controversial case was formally launched on January 24 1972, when Chairman Miles Kirkpatrick and Commissioners Rand Dixon and Mary Jones voted to issue the Commission s complaint. I Commissioners David Dennison and Everette MacIntyre opposed this action, and filed dissenting statements. Commissioner MacIntyre was prescient with regard to the ultimate procedural course of this matter, predicting "much litigation and little reform " and further suggesting that the resolution of certain " untried issues could take "perhaps years . A Congressman from Michigan was prescient about the political course of this matter: the then-Representative from Battle Creek denounced the issuance of the complaint almost immediately.

The Commission s complaint (paragraph 9) charged three species of violations of the FTC Act: First, that the respondents had individually and collectively maintained a "highly concentrated, noncompetitive market structure." Second, that the respondents had individually and collectively shared and exercised monopoly power. Third, that the respondents "erected, maintained, and raised" barriers to entry of new competition through unfair methods of competition. The methods of competition in question representing the exercise of monopoly power, as detailed in paragraphs six and seven of the complaint, included: 1) brand proliferation through differentiating similar products and promoting trademarks through intensive advertising, 2) misleading advertising of the value of cereals in regard to childrens' health, weight control and athletic , 5)prowess, 3) control of shelf space, 4) acquisition of competitiors mutual restraint (2)in challenging price increases, 6) restricted use , This was pursuant to the Commissinn s old Part II ("opportunity for settlement") Rules. The Part III adjudicative complaint issued in April 1972. Statement 99 F.

of trade deals or trade-directed promotion, and 7) limitations on the use of otconsumer directed promotions, such as coupons, cents off deals, and premiums." According to paragraph 8, as a result of all this, high barriers to entry successfully forestalled new competition; the introduction of new products was curbed; cereal prices and profits were higher than they would be in a competitive market. Much has been made, since the inception of this case, over the fact that it represents a test of the so-called "shared monopoly" theory. The complaint is subject to being so construed, depending on one interpretation of paragraph 9(b).' It is certainly true that press reports and even academic journals have from the first characterized this case in terms of a shared monopoly theory. ' The respondents in this matter have sought for years irrevocably to pin this badge on complaint counsel, who, for their part, have raced across the legal plain seeking to avoid the shared monopoly stigma, trying first the conspiracy gambit, falling at last across the finish line with a paper that mentions only " tacit collusion, and not shared monopoly even once. The Director of the Bureau of Competition believes that (3)only a shared monopoly theory was ever at issue in the case. The Administrative Law Judge details the history of the theory of thc case, beginning with early pretrial statements by complaint counsel that conspiracy was specifically not a part of the complaint. Later the original Law Judge assigned to this matter permitted complaint counsel--ver strenous objections by respondents to put into the record evidence of tacit collusion.' Judge Berman, though believing collusion not to have been pled as a part of this case, nonetheless based his ruling on consideration of both collusion and shared monopoly theories.

Judge Berman has dismissed this case for failure of proof on either of those theories.

The Director of the Bureau of Competition withdrew the earlier notice of intent to appeal filed in this matter because he believes a shared monopoly theory "is not consistent with the public interest is "an unwarranted expansion of the law " and has the prospect of punishing success among the competing cereal producers. Withdrawal of an appeal of a Federal Trade Commission complaint by those , 1 objected, joined by Commissioner Pert.chuk, to the singullH characterization of this Case as a shared monopoly matter in conne(tion with the issuance of the press release date December IH, 1981 , announcing o\lr3- 1 vote to solicit staff views On the propriety of the appeal of this matter. We felt that the characterization ofthe case was one of the main issues before us; hut, in any event, the press release was issued notwithstanding our expressedcnncerns The Cereal Case: Opening Shot. in FTC War on ' structual' Shared- Monopoly ur Ati.ack on ' markl'ing irregularities l" 5 Antitrust Law and Econ.Rev. 71 (lB71) "Oligopolies, Cereals, and Section Vive of the F'ederal Trade Commission Act " 61 Geo. J. 1145 (1973) . Orders uf .J\ldlje Hinkes dated February 24 , 1974 , March 12, 1975, and August 20 1976. . , Statement charged formally with its prosecution is virtually without precedent. The fact is that this is the Commission s-and not the staffscomplaint. The Bureau s involvement in this matter is in the nature of a stewardship, and the Bureau itself is but an administrative creation of the Commission to facilitate the accomplishment of tasks delegated to it by the Commission, including the litigation of Commission complaints. (4) Nonetheless, the issue has been joined, and I do not fault the Director of the Bureau for candidly stating his views of which theories, in his opinion, the Commission should and should not be litigating.

My concern, expressed in our Order of December 18, 1981 following the Commission s receipt of the Bureau Director s views of December 11 , 1981, was only that our Order of December 3, 1981 directing the submission of the views of complaint counsel be complied with. That has now belatedly been accomplished. As already noted, the attorneys and economist who have handled this matter for the past decade do not describe the case in terms of shared monopoly but in terms of tacit collusion, violative of the standards set out in American Tobacco Co. v. United States 328 U.S. 781 (1946). In the most traditional antitrust sense, the three respondents have tacitly colluded and cooperated to maintain and exercise monopoly power- power over price' and ' power to exclude' additional competitors." According to the staff The case does not challenge the industry s structure itself " (Staff at 11), but shows a pattern of sophisticated conduct that has raised barriers to entry for new competition, and facilitated higher than competitive price levels. I am aware of that body of commentary that believes the American Tobacco case adds to the law of monopolization by extending the analysis into an oligopoly setting. Professor W.H. Nicholls, stated as long ago as 1949:

The Tobacco case is clearly a legal milestone in the social control of oligopoly. By permitting the inference of illegal conspiracy from detailed similarity of behavior. . the (5)courts have at last brought oligopolistic industries within the reach of successful prosecution under the antitrust laws. He went on to say that the inference of conspiracy in the case was based on the assumption "that a few dominant firms wil perhaps independently and purely as a matter of self interest, evolve nonaggressive patterns of behavior. . . " Professor A.D. Neale disagrees with this interpretation of the case. In response to Nicholls he stated: o W. Nichols The Tobacco Case of 1946 " :19 American Economic Review, 296 (1949), cited as A.D. Neale The Antitrust Laws o(Thc U.S.A. 165 (2J Ed" 1970) , Statement 99 FTC (Il)e (Nicholls) is attributing to the court and the jury more economics than they would own: what the jury found and the Court confirmed was precisely that the conduct of the firms could not be accounted independent action. Rightly or wrongly, the Kentucky jury felt able to infer a true ' meeting of the minds' from the evidence and the case really adds nothing new to the law of conspiracy. This is confirmed by Mr. ,Justice Burton s summary of this aspect of the case: ' The essential combination or conspiracy in violation of the Sherman Act may be found in a course of dealings or other circumstances as well as in any exchange of words. . where the circumstances are such as to warrant a jury in finding that the conspirators had a unity of purpose or a common design and understanding, or a meeting of the minds in an unlawful agreement, the conclusion that a conspiracy is established is justified. Despite complaint counsel's use of the terminology " tacit collusion " I believe such an argument amounts to allegation of a conspiracy, even under complaint counsel's cited cases. (6) I am also constrained to agree with Judge Berman that a conspiracy charge was not a part of the Commission s complaint and that complaint counsel so stated in pretrial filings made on May 18 and 22 and June 19, 1972, and in pretrial statements made by. complaint counsel Robert Liedquist on June 5 and 8 and August 10, 1972. The absence of a conspiracy charge was also asserted at least twice in federal court proceedings. The admission of evidence on conspiracy made by the original AL.I, Judge Hinkes, amounted to a transgression of Commission Rule 3.15(a)(1) which prescribes the correct procedures for amendments to complaints. Nor does Rule 15(a)(2), providing for the admission of evidence "reasonably within the scope of the original complaint or notice of hearing," provide solace, since that rule requires both sides to acquiesce, and respondents have fought the notion of a collusion theory from the inception of these proceedings.

Viewing this case as a shared monopoly matter does not trouble me. Professors Areeda "nd Turner have postulated that If, Sherman Act Section 2 permits a government action in equity against a substantial and persistent single firm monopoly that has not behaved improperly, it also permits a similar action against the substantial and persistent shared monopoly. As compared with single-firm monopoly, however, there are important additional diffculties in identifying cases suitable for intervention, identifying the proper defendants, formulating appropriate remedies, and supporting judicial innovation, The authors go on to state that evidence of shared monopoly power e Neale, Op. Cit. 166. Violations of the Sherman Act ure also violatiuns of Section 5 uf the l'J'C Act FT Raladam Co. 283 U.S. 643 (1931).

1 Contemporary rel'vrts of the Commission s action took special note of the absence of a conspiracy charge Note Oligopolies, Cereals, and Section Five of the Federal Trade Commission Act " 61 Geo. 1145. 1 149 (1973) " III Areeda and Turner Antilnt. t Law 360 (1978). , Statement should be coupled with evidence of "exclusionary conduct (7Jhaving a significant causal relation to shared monopoly power. "9 For the purposes of my analysis, I am willing to assume that the exclusionary conduct alleged by complaint counsel has occurred, and that the respondents, three of which control 79% of cereal sales, share monopoly power in this $740 million market. The question then would be one of remedy. Indeed, the paramount difficulty with this case has always been the question of remedy. For whatever it is that Kellogg and the other respondents may have done, the proposed solution-to carve new cereal companies from the hides of existing ones and to force the licensing of successful trade names to the newly created competitors-is both draconian and manifestly uncertain to achieve the relief complaint counsel postulates that it will.

Areeda and Turner have stated:

Quite apart from statutory limitations, even a czar would consider restructuring only where it is likely to improve net economic performance substantially; and we say substantially' to take account of the costs of the process, including the risk of erroneous judgments. 11 Thus, assuming that the appropriate substantial and noncompetitive market structure required for a shared monopoly is present, the key issue becomes whether relief is available significantly to improve economic performance without sacrificing such economically worthy goals as substantial economies of scale. 12 (8) The difficulties possibly attendant to divestiture relief make it less than clear that improved industry performance is the inevitable consequence. Respondents and intervenors (the grain millers union) view the industry restructuring proposal as "an unprecedented and unworkable experiment in industry reorganization" (Respondents at 5). Dismantling of existing cereal plants, including those that now make other non-cereal products as well, may cause substantial interruptions or reductions in production. Existing lahar-management harmony may be disrupted. Trademark licensing may result in excess capacity and stifle product development, and disrupt economies of scale in production, distribution and sales. Of course, all of . ld.

10 I recognize that the AU did not find this industry to be ch!lraderized by monopoly profits. I also recognize complaint counsel's strong assertion that the AU' s methodology Wa. faulty on this score. " Id. 372 " For an interesting discussion of the primacy of economies of scale in !Jligopoly antitrust analysis, see Sherman and Tollison Public Policy Toward Oligopoly: Dissolution and Scale Economics " 4 Antitrust Law and Econ. Rev. 77 (1971), The evidc!1ce in this record on economies ofsca)e is in conflict. Complai!1t counsel believe that entry into this market by II single cereal bnmd line is impnlctical, and that i!1dostry participation at fmm 3.5% to 5% is !1ecessllry to achieve firm eco!1omies of scale, This dispute bears both on the issue or barriers to e!1try as well % the issue or relief. Only an appeal "fthis matter would IIllow the Commission to assess the conflicting evidence. , 288 FEDERAL TRAm COMMISSION DECISIONS Statement 99 F.T.

this may be the conjecture of self-serving private interests, but the complex dismantling of a long-existing industry is sufficiently clouded with doubts to give one pause. Indeed, even complaint counsel's own expert has stated that the remedy proposed may "for reasons unforeseen bring about the opposite result or may impose debilitating losses upon the big three.

Moreover, it is also a concern of mine that an administrative agency, operating under a tightly supervised legislative mandate undertake industry restructuring under a theory that clearly represents an extension of Sherman Act analysis. While Areeda and Turner believe that a viable shared monopoly theory exists, and that it can--ven now-be entrusted to the government antitrust authorities, they acknowledge the difficulties of such a case. Concentration ratios and market performance tests, determination of markets appropriate for attack, and whether (9)there should be an assessment of substantial scale economies proved by an accused oligopolist all present merely the threshold questions in approaching a shared monopoly prosecution.

Although not necessarily insuperable, these difficulties may suggest that the courts should refrain from taking so grave and novel a step without a new mandate from Congress. The issue is a serious one.

I come ultimately to the view that industry restructuring, such as is proposed here, is essentially a legislative concern, and as an agency that fairly can be characterized as an arm of the Congress, we should not undertake to restructure an industry under Section 5 of the FTC Act without a clear supportive signal from the Congress. In this case, the signals are, for the present, quite to the contrary-as they were not so apparently in 1972 when this complaint issued. The Federal Trade Commission has from time to time commented favorably on various legislative proposals amounting to industrial restructuring,14 None of these proposals has taken root as a preferred route for industrial market reorganization. I do believe that if the Congress were to endorse a shared monopoly approach to restructuring an oligopoly, this agency has the power to effect this sort of (10)change under a viable and respectable theory. " But the use of the power of divestiture or divorcement under Section 5 of the U III Arcea and Turner, 380. Professor Turner s uwn views seem to favor the !er;isl"tive approach; see: Turner The Scope of Antitrust and Other Economic ReguIal.ry Policies " 82 liar. L. Rev. 1209 (1969J. " Se for example: Statement of i"ederal Trade Commission before the National Commission for the Review of Antitrust Laws and Procedurl'S. urging legislation affecting antitrust actioo against some monopolies, February '. For a complete summary of the IT view Do this poiot: Testimony of Commissioner Robert Pitofsky un the Heflin Amendment before the Senate Commerce Committe, Novemhf'r 30 , 1979. "Certainly the Commission has ben restrained and circumspet in the extreme in imposing strudural relief, We have ben able to find only six cass since 1963 where structural relief was sought Statemen t FTC Act is the ultimate exercise in administrative authority. It should be used only to achieve a Congressionally endorsed result, or at least not to defy a clearly expressed congressional animus, as exists here. I note, too, that even candidates of both major political parties in the last national election denounced this case as ill-advised and contrary to the public interest.

Why all this concern has risen to the level that it has is difficult to explain, in light of the fact that the Commission is only mid-stream in- this case, and both Commission and court review-not to mention potential congressional action lie ahead to safeguard against any precipitous or unwarranted action in this matter. The issue here is larger than the Kellogg case, to be sure. Professor Joseph Brodley reflected in some depth on the practical, as well as the philosophical problem of exercising prosecutorial discretion to its ultimate end in cases of this sort:

In law, as in politics, public policy must often be the art of the possible. For this, as well as for other reasons, I put to one side the proposal that existing concentrated oligopoly firms be broken up . . . (A)ny large or even moderate scale attack on existing industrial concentration would run into congressional stormwaters of imposing magnitude. (11) In part, this may simply reflect an ambivalence of attitude in United States antitrust laws. The British writer, Neale, has noted the tendency of Americans " take a romatic view of the achievements and effciency of large industrial organizations even while they take a suspicious view of their power." (citation omitted) Such an attitude has made the remedy of divestiture rare even in Sherman Act cases. Perhaps, more basically, in a nation so thoroughly pragmatic as this one, there is an understandable reluctance to push an economic theory, however web founded, to the extreme conclusion of causing drastic rearrangements of large sections of American industry.

The paradox we are left with is that while there may be a legitimate concern about the anticompetitive effects of the exercise of oligopoly power, it is rarely true that these concerns wil mandate an administrative agency decision to restructure an industry, short of a legislative warrant to that effect. Therefore, I will vote that this appeal be terminated, not for the reasons relied upon by the Administrative Law Judge, but because the promulgation of relief by this agency will not, in any eventuality, conceivably lead to a restructuring of the cereal firms.

As may be evident, I do not intend the Initial Decision in this matter to become the final decision of the Commission. The Commission s Rules provide that if an appeal of an Initial Decision is 10 Brodley, "Oligopoly Power Under the Sherman and Clayton Acts - From Oligopoly Theory to Legal Poliq, 19 Stan. L. Rev. 285, 344-5 (1967) Statement 99 F.

not forthcoming, the Initial Decision becomes the decision of the Commission. According to complaint counsel, (12)the Initial Decision is riddled throughout with major procedural errors, and does not fairly give weight to certain of the evidence. I do not know whether this is true or not, and since I could not resolve these conflicts except through the process of appeal, it is my intention that the Initial Decision have no precedential or even persuasive authority for any proposition whatsoever. Thus, I will vote to docket the Initial Decision for the sole purpose of vacating it in its entirety arid substituting for it a simple order of dismissal. A few final words:

The fortitude and determination ofthose attorneys and economists who were charged so long ago by the Commission to prosecute this matter is a source of pride and admiration. No more difficult matter ever was considered to this point by this institution. No matter has been subjected to more criticism. No more difficult duty has been undertaken without sufficient thanks or recompense. In the face of everything, the staff of this agency has done the job it was ordered to , and acquitted itself ably and with honor. The fact that now the Commission abdicates its commitment to see this matter through is a responsibility that rests uniquely upon our shoulders and not upon theirs.

I very specifically dissassociate myself from what the Chairman had to say about this matter in his statement of December 18, 1981. Neither do I join in Commissioner Clanton s statement of views on this matter nor in Commissioner Pertschuk's (13)clarion, but quixotic, call to battle. " I speak, only for myself, and I wilingly take public and lasting responsibility for what I do today. " Rule3.51i").

'" Commissioner Pert.chuk's participation in thi matter has 1"",0 challenged; any !legal ability to advance tne appeal of this rnaltcr pursuant to Reorganization Plan No. 4 of 1961 would surely be challenged-adding only and inevitably LO the recnrd of delay of this case, diverting foradditional ye arstheCommis i()n sauilitytoevaluatethis ca5!onlhemerit.

, , 291 Complaint