Westreco, Inc. v. Commissioner

WESTRECO, INC., Petitioner v. COMMISSIONER OF INTERNAL REVENUE, Respondent Westreco, Inc. v. Commissioner Docket No. 24078-88 United States Tax Court T.C. Memo 1992-561; 1992 Tax Ct. Memo LEXIS 583; 64 T.C.M. (CCH) 849; September 23, 1992, Filed *583 Decision will be entered for petitioner. Petitioner is a corporation whose income is taxable in the United States. It performs research and development services for Nestec, a Swiss parent corporation which is not taxable in the United States. Both corporations and other corporations are subsidiaries of Nestle, S.A., a publicly traded multinational corporation located in Switzerland. The fees paid to petitioner by Nestec were paid pursuant to a written contract under a cost-plus formula. The Commissioner determined that the fees paid to petitioner were inadequate under sec. 482, I.R.C. 1954, and thus determined deficiencies in petitioner's income tax. Held, respondent abused her discretion under sec. 482 by allocating additional fee income to petitioner because the fees which petitioner charged Nestec clearly reflected income within the meaning of sec. 482, I.R.C. 1954. For Petitioner: Joel V. Williamson, Thomas C. Durham, Alexander Spitzer, Thomas L. Kittle-Kamp, Joseph R. Goeke, David A. Hyman, William A. Schmalzl, Mark I. Levy, Arthur R. Miller, Gregory L. Barton, Stephen M. Haracz, Elizabeth A. Levy, Roger J. Jones, Gary P. Kirschenbaum and Daniel A. Dumezich. For Respondent: Jill A. Frisch, *584 Jane Beaver Wilson, Raymond A. Kahn, and Diane P. Heller. GOFFE GOFFE MEMORANDUM FINDINGS OF FACT AND OPINION GOFFE, Judge: The Commissioner determined deficiencies in petitioner's Federal income tax liabilities for the following taxable years: Taxable Year Deficiency [1978] $ 1,520,238 [1979] 1,713,878 [1980] 1,714,266 [1981] 1,712,071 [1982] 2,169,397 The deficiencies determined by the Commissioner result from her reallocation of additional income earned by Westreco (petitioner) from Nestec, Ltd. (Nestec), Westreco's parent corporation, pursuant to section 482. Unless otherwise indicated, all section numbers refer to the Internal Revenue Code in effect for the taxable years 1978, 1979, 1980, 1981, and 1982, and Rule numbers refer to the Tax Court Rules of Practice and Procedure. The sole issue for decision is whether the fees paid to petitioner by Nestec for research and development (R and D) services clearly reflected income under section 482. We hold that such fees were within a range which are accurately characterized as arm's length and, accordingly, clearly reflected income. FINDINGS OF FACT Some of the facts have been stipulated and are so found. The stipulation*585 of facts and the accompanying exhibits are incorporated by this reference. Petitioner is a Panamanian corporation with its principal office in Marysville, Ohio. Background Petitioner is in the business of conducting research which is indirectly controlled by its ultimate parent, Nestle, S.A. (Nestle). Nestle is a publicly held multinational corporation headquartered at Vevey, Switzerland. Nestle had a humble beginning. Over a century ago, Dr. Henri Nestle, a Swiss pharmacist, had a sister who could not breast-feed her baby. Dr. Nestle set out to develop an infant formula that the child could tolerate. After consulting with pediatricians, Dr. Nestle developed such a formula, which consisted of concentrated cow's milk and carbohydrates. Dr. Nestle incorporated the Nestle corporation to commercialize this infant food formula and subsequent developments. During the years in issue, the Nestle group engaged in worldwide research, development, manufacturing, and marketing of processed food products. Among Nestle's product lines were instant milk, coffee, tea, and agglomerated 1 cocoa products. Nestle's specific products included roast and ground coffee, chocolates, milk, *586 frozen foods, culinary products, food ingredients, soups, fruit drinks, and baby food. Nestle also engaged in the hotel, restaurant, pharmaceutical, and cosmetics businesses. Nestle owned the trademarks for such well-known products as Nescafe and Taster's Choice (coffee), Nestea (instant tea), Quik and Nescoa (agglomerated cocoa products), Cailler, Frigor, Toll House, Nestle Crunch and Gala (chocolate products), Chambourcy (refrigerated milk products), NIDO (instant milk products), Findus (frozen foods), Maggi and Crosse & Blackwell (culinary products and food ingredients), Cerelac (baby food), and Beringer (wine). Among Nestle's principal competitors in the manufacture and sale of food products were Unilever, BSN, Mars Inc., Hershey Foods Corp., Cadbury Schweppes, PLC, and Corn Products Corp. To facilitate Nestle's worldwide network of operations, Nestle *587 created and operated through at least 25 subsidiaries. It conducted manufacturing operations through approximately 285 plants located in 55 countries and experienced between $ 11.5 and $ 13.5 billion in annual sales. It employed a total of 146,000 people, of whom 26,000 were employed in North America. Corporate Structure Nestec was a Swiss subsidiary of Nestle engaged in food-related R and D and technical assistance in the manufacture and marketing of Nestle products around the world. Westreco was and is a subsidiary of Nestec. Its principal place of business is in the United States. Nestle Holdings, Inc. (NHI) was and is a subsidiary of Nestle. NHI is the holding company for the Nestle operating subsidiaries located in the United States. Nestle Foods Corp., Inc. (TNCo) was and is a subsidiary of NHI engaged in the manufacture and sale of processed food products primarily in the United States. During the years in issue, this corporation's name was "The Nestle Company, Inc."; however we will, for convenience, refer to it as "TNCo". TNCo was made up of the United States operating companies of Nestle to which Westreco provided assistance. The revenue which Westreco received*588 in exchange for these services was not reallocated by the Commissioner and is, therefore, not at issue in this case. Nestle in the U.S. Market TNCo was responsible for the manufacture and marketing of certain Nestle products in the United States marketplace. TNCo had the following net sales: Year Ended Net Sales July 1, 1978 $ 1,306,823,000 June 30, 1979 1,376,266,000 June 28, 1980 1,356,085,000 January 2, 1982 (78-week period) 2,071,641,000 Among TNCo's significant products during the years in issue were instant coffee (e.g., Nescafe and Taster's Choice), chocolate products (e.g., Nestle Chocolate Morsels, Nestle Crunch, and Nestle Quik), and instant tea (e.g., Nestea). Chocolate was a major product of TNCo, accounting for 30 to 40 percent of its sales volume. Instant coffee sales equaled 30 to 40 percent of its sales. During the years in issue, TNCo experienced problems in its sales of instant tea and instant coffee. Instant tea sales declined from $ 122,972,000 for the year ended June 30, 1979, to $ 91,922,000 for the year ended January 1, 1983. Instant coffee sales declined from $ 570,385,000 for the year ended June 30, 1979, to $ 407,255,000 for*589 the year ended January 1, 1983. Both Nestle and TNCo management were concerned about the declining sales, low profitability, and overall declining market share of Nestle's instant coffee products in the U.S. market. Nestle and TNCo management determined that the reasons for the decline in sales of its instant coffee were threefold: (1) Consumption of coffee in general (roast/ground and instant) was declining; (2) instant coffee was losing its market share to roast and ground coffee among coffee drinkers; and (3) the market for instant coffee was increasingly competitive. TNCo manufactured and marketed three primary brands of instant coffee products in the United States during the years in issue: Taster's Choice (a freeze-dried 2 instant coffee), Nescafe (a spray-dried 3 instant coffee), and Sunrise (a coffee/chicory 4 spray-dried blend). Taster's Choice was the most successful of Nestle's three instant coffee products marketed in the United States. Nescafe, while popular on the East Coast, was less well accepted in other parts of the U.S. market. It was generally viewed as a low-price generic coffee. Sunrise was accepted only when first introduced in 1977 because it enjoyed*590 a price advantage over 100-percent coffee. The general decline in coffee consumption during the years at issue was attributable to several factors. At least one important factor was price. In the mid-1970's, Brazil, a major coffee bean producer, suffered a bad freeze. As a result, fewer coffee beans were harvested, and the price of coffee increased from 60 cents per pound to $ 3.30 per pound. Thus, the increased raw material cost of coffee beans caused the price of a jar of Taster's Choice to nearly double. Another factor causing a slowdown in coffee consumption was the increased consumption of cold soda beverages, such as Coca Cola, which many young people, in particular, chose over coffee. Also, there was a general perception that coffee was the cause of a variety of health problems. Instant coffee*591 offered the convenience of immediacy over traditional coffee. However, instant coffee met a new and innovative challenger with the advent of the "automatic-drip" coffee maker, such as the Mr. Coffee machine. These faster coffee makers seriously eroded instant coffee's advantage of speed. Instant coffee had another strike against it. Many consumers viewed instant coffee as an artificial product, created with chemicals, while roast and ground enjoyed the perception of being a pure and natural coffee. Thus, in the United States, roast and ground coffee took away instant coffee's market share as follows: [1984] Percentage of sales for instant coffee 40.8 40.6 38.7 37.3 35.8 34.8 Percentage of sales for roast and ground coffee 59.2 59.4 61.3 62.7 64.2 65.2 As the market for instant coffee began to shrink in the United States, advertising became the weapon of choice for the big food companies eager to position their products to take advantage of whatever remained of the instant coffee market. Two of TNCo's major competitors, General Foods and Procter & Gamble, had advertising budgets larger than TNCo to spend on both their ground/roast coffee*592 and instant coffee products. In fact, the heavy advertising by these competitors on their ground and roast coffee had a beneficial spillover effect onto the instant coffees marketed by these two food giants. General Foods marketed an instant freeze-dried coffee under the label Maxim and instant spray-dried coffee under the label Maxwell House. It also sold a decaffeinated coffee under the Sanka label. Sanka was so popular, in fact, that "Sanka" became a generic term for decaffeinated coffee. Procter & Gamble was a relatively recent entrant to the coffee market. In the early 1970's, it acquired Folger's, which at that time was a strong regional brand in the area of roast and ground coffee. After establishing a national identity for Folger's roast and ground coffee through massive advertising, Procter & Gamble brought out an instant coffee under the Folger's label. Procter & Gamble used a unique marketing tactic by developing shiny particles in its Folger's coffee which it claimed were "flavor crystals". These competitors were successful in taking away TNCo's instant coffee market share as illustrated: Instant Coffee - Sales (in millions of pounds) Year U.S. Total Market Total TNCo [1979] 180.3 53.4 [1980] 178.5 48.7 [1981] 169.9 44.8 [1982] 163.1 41.6 *593 In light of these developments, the management of Nestle and TNCo decided that the way to regain market share for their instant coffees was to improve the overall quality of the products. The strategy for regaining market share had two phases: A short-term period involving the relaunch of Nestle's existing products with improved taste and aroma and a second long-term method whereby Nestle would segment 5 and market its instant coffee products. Nestec and the Importance of Research and Development The Nestle corporation was founded upon food-related R and D. This fact was a key component of Nestle's overall philosophy during the years in issue. Technological development was and is vital to the success and growth*594 of Nestle's business. It must rely upon its technological prowess to retain and expand market share, improve reputation, and, in general, to compete in the world market. In 1926, Nestle created Nestec to provide R and D, technological development, and technical assistance to its manufacturing operations throughout the world. These three operational areas of Nestec may generally be described as: (1) Basic R and D in the fundamental sciences to acquire new knowledge, including such sciences as toxicology, nutrition, and molecular biology; (2) technological development, which is the transferring of the results of basic R and D into industrial-scale development, e.g., new factory design and construction, plant and process improvements; and finally, (3) technical assistance, which consists of the details of marketing, advertising, packaging, and sales of either a newly developed or existing product. Nestec's R and D expenditures were equal to approximately 1 percent of Nestle's sales during the years in issue. The specific expenditures break down as follows: Nestec R and D Expenses (in 000s of Dollars) Basic Technological Technical Year Research Development Production Total [1978] $ 11,527 $ 54,691 $ 41,739 $ 107,957 [1979] 12,609 61,412 39,222 113,243 [1980] 14,650 70,535 51,587 136,772 [1981] 13,900 69,018 43,013 125,931 [1982] 16,044 76,000 41,275 133,319 *595 Nestec is organized into six basic divisions; three of these divisions, Basic Research, Technological Development (T-Dev), and Technical Production (T-Production), were referred to as the Technical Divisions. The fourth division, Strategic Market Planning (D-Produits), provided Nestle operating corporations with global marketing strategies and organizational marketing advice. The fifth and sixth divisions, Finance et Controle and Etats Majors, respectively, controlled the financial, legal, and administrative functions of Nestec. Nestle, through Nestec, underwrote basic R and D because scientific knowledge was and is the foundation for new and improved products and processes developed and commercialized by Nestec T-Dev and Nestec T-Production. Nestle also paid for such basic R and D to ensure that the products manufactured and marketed were safe for human consumption. Nestec T-Dev conducted its operations through 17 technological development centers spread over 10 countries in diverse parts of the world. Most of these centers used the suffix "reco" (for research company) and, with a few exceptions, were located near a manufacturing plant to efficiently transfer technology*596 from R and D to actual operation. Nestec T-Dev, utilizing scientific knowledge developed by Nestec Basic Research, engaged in the technological development of new and improved products and manufacturing processes by using these technological development centers (Recos) throughout the world. This point in the development stage may be called the "pilot" or testing stage, whereby a product or process is not yet at a full scale commercial production level. Nestec Basic Research -- Background Fundamental scientific knowledge was developed by Nestec Basic Research. This is the knowledge upon which Recos build and develop in order to design new and improved products and processes. Recos also relied upon publicly available knowledge obtained from scientific literature. Nestec Basic Research's facilities were located at La Tour-de-Peilz and Orbe, Switzerland. All of Nestle's basic research facilities were located in Switzerland because this was where the company historically began. It was also important to centrally locate basic research because such research was extremely expensive in terms of capital investment and personnel. In addition, it afforded Nestle and Nestec management*597 better control and direction of using the R and D to create marketable products instead of simply pure, but unprofitable, scientific inquiry. Finally, centralization permitted a concentration of scientists from different disciplines in one Swiss location that encouraged consultation and "cross-pollination" of information which was essential to successful basic scientific research. Nestec Basic Research was organized into four departments: (1) Food Sciences, (2) Biological and Nutritional Sciences, (3) Toxicological Sciences, and (4) Fundamental Sciences. Employees were not evenly divided among the departments. The Nutritional Sciences department was the largest with about 80 people, and the remaining departments had about 50 people each for a total staff of about 280 people. Nestec Basic Research -- Food Sciences Department Food sciences is the study of the three main raw materials that exist in food, namely, proteins, lipids, 6 and carbohydrates. The Food Sciences Department of Nestec Basic Research studied food ingredients, food composition, and food raw materials from a chemical, physical, and biological standpoint. It also studied the effect of unit operations in*598 food processing. 7 The Food Sciences Department communicated its research findings quite frequently with the several Recos. In addition to knowledge of fundamental sciences, this department was staffed by professionals with expertise in technological trials, pilot plants, sensory evaluation, and chemical engineering as well as in electronics and various other engineering disciplines. Nestec Basic Research -- Biological and Nutritional Services This department of Nestec Basic Research studied the nutritional value of Nestle products and the effect that processing had on food value. This department was staffed by professionals in nutrition, metabolism, biochemistry, clinical chemistry, immunology, psychology, fermentation, and medicine. Nutrition was the origin of Nestle products. The milk-based*599 food product developed by Dr. Henri Nestle was the first infant food formula. In addition, this department conducted studies of population groups to determine how the environment affected human health. These studies included food intake surveys that accumulated knowledge about the needs and habits of different populations. Nestec Basic Research -- Toxicological Sciences Department Nestec Basic Research was also responsible for evaluating the nutritional value of products and the effect which processes had upon nutrients. This was important to determine whether any toxic substances were present in Nestle's products or were produced as a result of processing during manufacture. It was very important for Nestle to have an "in-house" toxicological department to allow it to become familiar with the corporation's products and to anticipate potential problems with the safety of those products. The toxicological department in Nestec's Basic Research could also perform defensive research, i.e., research to address charges frequently raised by the scientific community or government regulatory agencies that a product is unsafe. A contract research firm, such as Westreco, can be*600 employed to test products for safety; however, these firms lack expertise and familiarity with any particular corporation's products, and therefore must surmount a steep learning curve before they can be effective. Significant amounts of time and money would have to be expended before a contract research firm was sufficiently educated to effectively perform the role of Nestec Basic Research's toxicological department. Nestec Basic Research -- Fundamental Sciences Department This department studied the physical and chemical properties of food and food ingredients. It was staffed by professionals in organic and inorganic chemistry, biochemistry, biophysics, physics, mathematics, microscopy, and electronics. The Recos Petitioner and the other Recos, engaged in contract research activities at Nestec's direction. Nestec provided information such as scientific and other technical data developed by Nestec or one of the other Recos for use in specific Reco's activities. Nestec coordinated communications between the Recos and Nestle operating corporations. Recos communicated directly with Nestec and with one another regarding their respective technological development *601 activities. The Recos had two main objectives: First, the development of new products and processes; second, new methods of manufacture and improvement of products and processes already in existence in light of the latest technology. The various 17 Recos developed specific skills by working in certain production lines, e.g., milk products, infant and dietetic foods, instant drinks, chocolate and confectionery, culinary products, frozen foods and ice creams, chilled products, soft drinks, cakes and biscuits, and packaging studies. Each of the Recos specialized in a particular group of products but, in particularly important areas, several Recos shared various tasks. For example, both petitioner and the Reco named Linor located in Orbe, Switzerland, performed similar and, in some areas, identical tasks. The Recos contained impressive personnel to carry out their mission. Total employment at the Recos consisted of about 2,000 men and women with extensive training and academic credentials in more than 40 professions, including food technology, chemical engineering, mechanical engineering, industrial engineering, chemistry, biochemistry, organic chemistry, microbiology, mathematics, *602 agricultural engineering, and consumer research. The Recos were generally made up of three separate units: An experimental kitchen, a pilot plant, and a development laboratory. The kitchen was used to develop recipes and evaluate the finished product. The pilot plant was a small-scale factory equipped to carry out the production processes of various projects. The laboratories provided scientific support to the development groups by evaluating and controlling the properties of raw materials and finished products. The following is a list of the Recos, their locations, and some of their respective areas of technological development: Areas of Reco Location Technical Expertise Hispareco Badajoz, Spain Frozen foods, sterilized products, agriculture Londreco Hayes, Sterilized products, United Kingdom refrigerated culinary products, cold sauces, pickles Nordreco Bjuv, Sweden Frozen foods, refrigerated culinary products, dietetic products, agriculture Westreco Marysville, Coffee, tea, Ohio U.S.A. agglomerated cocoa Westreco New Milford, Dehydrated culinary Connecticut U.S.A. products, sterilized products, aromas, cooking aids, fruit juices, drinks Latinreco Quito, Ecuador Cereal-based products, dehydrated culinary products, agriculture Eastreco Singapore Fermentation products, products based on soy and other oil-yielding plants, dehydrated culinary products Linor Orbe, Switzerland Coffee, cereal, dietetic cereal-based products, agglomerated cocoa-based drinks, fermentation products Alpura Konolfingen, Long-life dairy products, Koreco Switzerland soy-based products, dietetic dairy products Vitoreco Kempttal, Dehydrated culinary Switzerland products, bouillons, seasonings, cooking aids, pasta products Chocolate Broc, Chocolate and Technology Switzerland confectionery Packaging Orbe, Packaging technology Switzerland Food La Tour-de-Peilz, Textured products, bakery Technology Switzerland and pastry products, technical applications of microwaves, experimental kitchens Dereco Ludwigsburg, Herb teas, coffee sur- Weiding, Fed. Rep. rogates, cold sauces of Germany (condiments, mustard, mayonnaise, etc.), refrigerated dairy products, baby food in jars Francereco Beauvais, La Meaufe, Frozen foods, dehydrated and Bezier, France culinary frozen foods, dehydrated culinary products, ice-creams, soft cheese, special oils, milk-based products, catering products Novareco Robbio, Italy Hard cheese and processed cheese Venreco Venray, Netherlands Potato research, storage and preservation of potatoes, dehydrated potatoes *603 The interrelationship between the Nestle operating corporations, Nestec, and the Recos influenced the direction of technological development and technical assistance. Nestec conducted regular conferences on the various product areas to determine where future time, money, and effort should be spent. Nestec, by virtue of contracts with the employees of Nestec and the Recos, was assigned all ownership rights to the patents issued to any Nestec or Reco employee. During the years in issue, 91 patents were issued and assigned to Nestec in the United States. Westreco As part of the organizational structure of Nestec, Westreco was incorporated and began operations in 1957 in Marysville, Ohio. A second branch of Westreco, Westreco-New Milford, was started in New Milford, Connecticut, in 1981. Petitioner contracted to provide R and D services to Nestec and, ultimately, Nestle. Petitioner communicated directly with Nestec and sometimes with other Recos regarding its technological development activities. Nestec issued manufacturing instructions to petitioner. These instructions prescribed procedures for manufacturing the products made and marketed by Nestle operating companies. *604 Nestec also issued laboratory instructions to petitioner which detailed appropriate methods for performing onsite quality assurance and quality control testing. Westreco-Marysville Petitioner operated a pilot plant and laboratory at its facilities in Marysville, Ohio. This location permitted petitioner to be in close contact with Nestle's manufacturing plant of TNCo., in nearby Sunbury, Ohio. Petitioner's employees also travelled to TNCo.'s coffee and tea facilities in Ripon, California; Freehold, New Jersey; Fulton, New York; and Granite City, Illinois. Petitioner's technological development at this location concentrated on instant coffee, instant tea, and the manufacturing processes associated with these products. In particular, petitioner specialized in texturization, 8 extraction, 9 flavor recovery, 10 aromatization, 11 spray drying, freeze drying, and decaffeination 12 for instant tea and coffee products and agglomeration for chocolate drinks. It also analyzed current and potential Nestle products and Nestle's competitors' products, and provided results obtained in its laboratories to Nestec and TNCo. In addition, petitioner conducted limited technological*605 development with regard to a few culinary products, such as soy-based food products. During the years in issue, petitioner was divided into the following departments: analytical basic support, new products, protein/protein products, engineering services, engineering development, process development/pilot plant, and basic technical development. These departments included professionals in food technology, chemical, engineering, mechanical engineering, food science, chemistry, biochemistry, and consumer research. Petitioner had approximately 212 to 218 employees, *606 of whom approximately 49 had college and/or post-graduate degrees in scientific or technical fields. The heart of petitioner's Marysville location was the pilot plant, where 70 to 75 employees, a third of the location's staff, worked. This plant operated 24 hours a day in order to produce product prototypes. The pilot plant's operation was supported by the analytical department. This department was a group of 20 to 23 chemists who analyzed and monitored the results of product trials in the plant. The analytical department did not develop its own analytical testing methodologies but rather relied in large measure on techniques developed by Nestec Basic Research. About 80 to 90 percent of the analyses performed by the analytical department was characterized as moisture content analysis or oil content analysis. Petitioner was not capable of commercializing new and improved products and processes at this location. In order to bring new products and processes to commercial fruition, petitioner relied upon Nestec T-Production. Petitioner did, however, assist in the industrialization of new and improved manufacturing processes for instant coffee and tea at nearby operating companies. *607 Westreco-New Milford Petitioner's New Milford location performed technological development and technical assistance services principally in the area of culinary products. Petitioner employed 72 to 75 people here, of whom 45 to 46 had college and/or post-graduate degrees in scientific or technical fields during the years in issue. In 1981 and 1982, the last 2 years of the years in issue, the Westreco-New Milford location was in a setup process and was not at full capacity. The pilot plant was not fully assembled but it was able to produce retorted and dehydrated products, pastas, and beverages, and perform filling trials. Also, some small-scale equipment was operated in the plant for drying and evaporation work. In 1981, Westreco-New Milford had a capital budget of approximately $ 1 million, which it began investing in key equipment. It received considerable technical assistance from Nestec with regard to the purchase of this equipment. Financial Information Nestec assigned a project to a Reco on the basis of its competence in that particular product line. Nestec operated on a global scale and worked in 10 currencies, which made budgeting on one particular currency*608 impossible. Therefore, Nestec used man-hours as its common denominator for allocating priority to its projects among the Recos. Nestec maintained budgetary control over capital investment expenditures, staffing and salaries at petitioner. Nestec approved petitioner's capital investment budgets annually, but any special requests or nonbudgeted extraordinary capital investment expenditures were subject to review by the management of Nestle. Nestec compensated petitioner for its R and D services on a cost-plus basis. Nestec was contractually obligated to reimburse petitioner for certain expenses plus a profit equal to 7.5 percent on the first $ 350,000 of such expenses, 5 percent on the next $ 1,500,000 of expenses, and 3.5 percent of petitioner's expenses thereafter. The reimbursable expenses included all salaries, rent, consulting fees, raw materials, equipment, and administrative expenses, but not taxes. The contract could be terminated by either party on a year-by-year basis, subject to 3 months' written notice. Approximately 50 percent to 60 percent of the cost of operating a Reco was attributable to salary expenses. Westreco employees filled out a daily timesheet which*609 recorded the amount of time spent on various projects. Each employee was required to account for a minimum of 8 hours per day and was required to record any overtime in excess of that amount. Westreco had three primary categories for employee time reported on timesheets. The first category was for technological development and technical projects. Each separate technological development or technical assistance project was assigned a separate account number. The second category of Westreco's timesheets was for administrative duties performed by the various departments throughout Westreco. Some of these departments contained a significant number of administrative personnel. For example, the pilot lab included a full-time timekeeper, a janitorial staff, and maintenance mechanics who charged all of their time to an administrative account. The pilot lab was Westreco's largest department. Also included in Westreco's administrative account were the salaries for two secretaries in its front office and for the director of research. Other administrative accounts included salaries for the general maintenance staff, the staff of the carpentry shop, the painting shop, and the plumbing*610 shop. Some of the administrative departments which charged 100 percent of their time to administrative accounts included the personnel department, accounting department, the shipping and receiving department, the purchasing agent, and the night janitorial staff. Other members of Westreco's technical personnel charged some of their time as administrative to account for such duties as salary reviews or attendance at professional seminars. A third category for Westreco's timesheets was the fringe benefit accounts. The fringe benefit accounts did not include any costs other than salaries for paid absences. These accounts included time charged by employees for sick pay, vacation, personal leave, jury duty, workers' compensation leave, military leave, death in the family, and holiday pay. Westreco's employees maintained timesheets in order to reflect the cost of technological development and technical assistance projects. The allocation of these labor expenses served as the basis for allocating overhead expense to technological development and technical assistance projects. Westreco reported the cost of each project to Nestec for payment under its contract; accordingly, Westreco*611 placed a high priority on the accurate maintenance of time records. In fact, if timesheets were not filled out and collected on a timely basis, Westreco employees were not paid. Westreco translated the information received in timesheets into labor costs by accumulating the hours which an employee worked on each project. These hours were then multiplied by the employee's hourly salary and the resulting amount was allocated to the particular project. Westreco paid overtime to its nonexempt employees when they worked more than 8 hours per day, or on Saturdays, Sundays, or holidays. The amount of overtime attributable to an employee's normal hourly rate was charged to the appropriate technological development or technical assistance account or administrative account in accordance with the normal procedures for salary cost allocation. The amount of the overtime pay in excess of that employee's normal hourly rate was referred to as "premium overtime" and was charged to a fringe benefit account. The following is an allocation of Westreco's salary expense for the years 1978 through 1982: [1982] Fringe benefit: Marysville $ 423,255 $ 449,191 $ 513,744 $ 566,001 $ 642,345 New -- -- -- 154,159 186,849 Milford Subtotal 423,255 449,191 513,744 720,160 829,194 Administra tive: Marysville 971,574 1,106,152 1,150,241 1,282,577 1,575,422 New -- -- -- 408,640 468,584 Milford Subtotal 971,574 1,106,152 1,150,241 1,691,217 2,044,006 Projects: Marysville 2,487,569 2,793,738 3,281,327 3,564,467 3,753,334 New -- -- -- 1,210,000 1,353,488 Milford Subtotal 2,487,569 2,793,738 3,281,327 4,774,467 5,106,822 Overtime: Marysville 129,517 130,280 129,981 132,996 129,550 New -- -- -- 21,046 18,265 Milford Subtotal 129,517 130,280 129,981 154,042 147,815 TOTAL 4,011,915 4,479,361 5,075,293 7,339,886 8,127,837 Salaries per tax return 4,011,915 4,459,785 5,075,293 7,339,886 8,127,839 *612 In the statutory notice of deficiency, the Commissioner reallocated additional fees of $ 19,515,542 from Nestec to Westreco as follows: Taxable Year Ended Amount December 31, 1978 $ 3,186,209 December 31, 1979 3,655,649 December 31, 1980 3,616,368 December 31, 1981 3,716,995 December 31, 1982 5,340,321 The statutory notice did not explain the method used by the Commissioner to allocate additional income to petitioner. The method used for the statutory notice was, in fact, a multiplier applied to salaries paid by petitioner to its employees. At trial, respondent did not formally abandon the multiplier method but she offered no evidence or explanation to support the determination made in the statutory notice. Rather, respondent relied upon fees paid to purportedly comparable businesses to support reallocation of fees to petitioner. OPINION This is the final opinion in this case. We decided certain procedural matters in our prior opinion which was filed on September 20, 1990, and reported as T.C. Memo. 1990-501. Petitioner raised additional procedural matters prior to trial and during the trial. We find it unnecessary to decide the procedural*613 questions presented because we hold that petitioner has proved that the methodologies used by respondent both in her statutory notice of deficiency and at trial were defective and further that the proof offered by petitioner clearly sustained petitioner's burden of proving that the fees paid to petitioner by Nestec were reasonable within the parameters of section 482. Section 482 provides: In any case of two or more organizations, trades, or business * * * owned or controlled directly or indirectly by the same interests, the Secretary may distribute, apportion, or allocate gross income, deductions, credits, or allowances between or among such organizations, trades, or businesses, if he determines that such distribution, apportionment, or allocation is necessary in order to prevent evasion of taxes or clearly to reflect the income of any such organizations, trades, or businesses. The purpose of section 482 is to prevent the artificial shifting of the true net income of controlled taxpayers by placing controlled taxpayers on a parity with uncontrolled, unrelated taxpayers. Commissioner v. First Security Bank, 405 U.S. 394, 400 (1972); Bausch & Lomb, Inc. v. Commissioner, 92 T.C. 525, 581 (1989),*614 affd. 933 F.2d 1084 (2d Cir. 1991); sec. 1.482-1(b)(1), Income Tax. Regs. Such allocations are, of course, appropriate when one entity is subject to taxation in the United States and the related entity is not. The fees received by petitioner are taxable in the United States but Nestec, which pays the fees, is not taxable in the United States. Respondent's authority to make allocations under section 482 is broad. Bausch & Lomb, Inc. v. Commissioner, supra; Edwards v. Commissioner, 67 T.C. 224, 230 (1976); PPG Industries, Inc. v. Commissioner, 55 T.C. 928, 990-991 (1970). Her determination of a deficiency is presumptively correct and must be sustained absent a showing that she has abused her discretion. Rule 142(a); Bausch & Lomb, Inc. v. Commissioner, supra; Paccar, Inc. v. Commissioner, 85 T.C. 754, 787 (1985), affd. 849 F.2d 393 (9th Cir. 1988). The taxpayer thus bears the heavier than normal burden of proving that respondent's allocations under section 482 are*615 arbitrary, capricious, or unreasonable. Your Host, Inc. v. Commissioner, 489 F.2d 957 (2d Cir. 1973); G.D. Searle & Co. v. Commissioner, 88 T.C. 252, 359 (1987). In cases where the subject matter of the reallocation involves the performance of services between controlled or related parties, an "arm's-length" standard is used to calculate the correct amount that the seller should charge and the buyer should pay. Section 1.482-2(b)(3), Income Tax Regs., defines the standard as the "amount which was charged or would have been charged for the same or similar services in independent transactions with or between unrelated parties under similar circumstances considering all relevant facts." Petitioner's principal business activity is performing research services for Nestec. Accordingly, there is no particular methodology which must be used in order to determine the appropriateness of respondent's determination, but rather we must consider all the particular facts and circumstances of petitioner's arrangement to decide if the amount it received from Nestec was equal to an arm's-length amount. Sec. 1.482-2(b)(7)(ii)(a), Income*616 Tax Regs. The pertinent part of section 1.482-2(b)(7)(ii)(a), Income Tax Regs., provides: Such facts and circumstances may include the time devoted to the rendition of the services, the relative cost of the services, the regularity with which the services are rendered, the amount of capital investment, the risk of loss involved, and whether the services are in the nature of supporting services or independent of the other activities of the renderer. Petitioner's burden is to show that respondent's section 482 allocations are arbitrary, capricious, or unreasonable. Sundstrand Corp. v. Commissioner, 96 T.C. 226, 353 (1991); Eli Lilly & Co. v. Commissioner, 84 T.C. 996, 1131 (1985), affd. in part, revd. and remanded in part 856 F.2d 855 (7th Cir. 1988). We must examine all of the facts and circumstances to determine whether respondent has abused her discretion. Eli Lilly & Co. v. Commissioner, supra; American Terrazzo Strip Co. v. Commissioner, 56 T.C. 961, 971 (1971); Woodward Governor Co. v. Commissioner, 55 T.C. 56, 65 (1970).*617 The Commissioner assigned Walter Trepashko, an engineer, to examine petitioner's returns. Mr. Trepashko's assignment was to determine the appropriate arm's-length charge for the services provided by Westreco to Nestec. The services provided to TNCo were not readjusted by the Commissioner. We have not considered whether the amounts between petitioner and TNCo were or were not at arm's length for purposes of section 482. Petitioner called Mr. Trepashko to testify as to the method he used to reallocate additional income to petitioner. Respondent objected to the examination of Mr. Trepashko and the method he used because that would require the Court to "look behind" the statutory notice of deficiency under the rationale of Greenberg's Express, Inc. v. Commissioner, 62 T.C. 324, 327-329 (1974). Generally, we refuse to examine the Commissioner's conduct in arriving at her determinations. Capitol Federal Savings & Loan v. Commissioner, 96 T.C. 204, 214 (1991); Vallone v. Commissioner, 88 T.C. 794, 806 (1987). In a section 482 reallocation, however, petitioner bears the admittedly heavy burden*618 of showing respondent abused her discretion in making the determination. Petitioner argues that it is entitled to examine the manner in which respondent exercised her discretion to ascertain whether she abused her discretion. We have found an abuse of discretion when such scrutiny has revealed a fundamental defect in the methodology used to determine a deficiency under section 482. American Terrazzo Strip Co. v. Commissioner, supra at 971-973 (reports and testimony of revenue agent admissible); PPG Industries, Inc. v. Commissioner, 55 T.C. at 993 (revenue agent's use of the Source Book of Statistics of Income without consideration of the nature of the taxpayer's operation was arbitrary and unreasonable); see also Branerton Corp. v. Commissioner, 64 T.C. 191, 200-201 (1975) (taxpayer permitted to inquire into Commissioner's method of calculating bad debt deduction in order to show abuse of discretion). In Capitol Federal Savings & Loan v. Commissioner, supra at 214, we stated: * * * we have held that a review of an exercise of discretion which is *619 necessary to determine the merits of the Commissioner's substantive determination of deficiency does not constitute "looking behind" the notice of deficiency. Estate of Gardner v. Commissioner, 82 T.C. 989, 1000 (1984). We have recognized that a somewhat broader scope of inquiry is permitted where taxpayers have the heavy burden of demonstrating that the Commissioner has abused his discretion.* * * The language of the statutory notice was as follows: It is determined that income should be reallocated to Westreco, Inc., from Nestles, [sic] Ltd. for the taxable years ended 12/31/78, 12/31/80, 12/31/81, and 12/31/82 in the amounts of $ 3,186,209.00, $ 3,655.00, $ 649.00, $ 3,616.00, $ 368.00, $ 3,716,995 and $ 5,340,321.00 respectively, pursuant to I.R.C. Sec. 482 and the regulations promulgated thereunder including specifically Treas. Reg. Sec. 1.482-2(b). This reallocation has been made to reflect arms [sic] length charges for the provision of engineering and research services to Nestles, [sic] Ltd. This arms [sic] length charge has been determined to reflect the amount which would have been charged for the same or similar services in independent*620 transactions between related parties under similar circumstances. The statutory notice does not explain the methodology used. The deficiencies which Mr. Trepashko proposed in his report are identical with the deficiencies determined in the statutory notice. We have no evidence that anyone else in the Internal Revenue Service proposed deficiencies based upon a different methodology. Mr. Trepashko's methodology, therefore, represents the methodology which the Commissioner utilized in exercising her discretion under section 482 in the statutory notice of deficiency. We hold that under the limited circumstances of this case, petitioner is entitled to examine Mr. Trepashko at trial as to his methodology. Mr. Trepashko, respondent's engineer, recommended an upward adjustment to petitioner's income by using what is called the salary multiplier method. He took the total salaries which petitioner reported on its income tax returns for each of the years in issue and multiplied these salaries by a factor or "burden" of 3 in order to redetermine petitioner's fees from Nestec. He then calculated the difference between his purportedly "correct" revenue amount and the amount actually reported*621 as income on petitioner's income tax returns. Petitioner alleges several errors in the salary multiplier method applied by Mr. Trepashko. First, Mr. Trepashko used all of petitioner's salary expense in his calculation, including amounts paid employees for administrative duties, vacations, sick days, holidays, and other paid absences. Second, the salary multiplier method is a compensation method used in large engineering and construction firms, not for small R and D firms such as petitioner. Petitioner called as its witness Charles A. Parthum, senior vice president with Camp, Dresser & McKee, Inc., an engineering firm in Cambridge, Massachusetts. Mr. Parthum holds a bachelor of science degree in civil engineering from Northeastern University and is a registered professional engineer in more than 15 States. Mr. Parthum has been in the civil engineering profession for more than 40 years. Mr. Parthum's numerous activities and contributions to the field of civil engineering include his membership to the American Society of Civil Engineers (ASCE). The ASCE publishes a manual titled "Consulting Engineering: A Guide for the Engagement of Engineering Services" (the manual). Mr. *622 Parthum served on the ASCE Committee of Standard of Practice which prepared and updates the manual. In the 1988 revision, Mr. Parthum prepared several chapters for the manual. The purpose of the manual is to educate clients as to what an engineer does, what services are offered by engineers, and the methods by which engineers bill for their services. Engineers base their fees to clients upon the time directly devoted by professional employees to the client's project, similar to the manner in which lawyers and other professionals bill clients. The salary multiplier, a method recognized and accepted in the engineering industry, is commonly used when the project is not definite in scope, nor in time. The multiplier can be based upon either salary costs or basic wages. Salary costs are defined as direct payroll (including sick leave, vacation, holiday, and incentive pay) plus fringe benefits. These costs averaged an additional 35 percent of salary. This percentage is based upon the theory that for every $ 100 paid to an employee as salary, it costs the firm an additional $ 35 to account for such items as holidays, sick time, retirement plan costs, medical and life insurance. *623 Mr. Parthum testified that the multiplier would range between 2 and 2.4, presumably to provide for a sufficient profit to the engineering firm, but provide flexibility to remain competitive. The manual also recommends a higher multiplier for projects of short duration or small size and also for engineering firms which possess extensive experience or special knowledge in a particular area. If an engineering firm bills according to basic wages, then a multiplier of 3 to 3.3 would be average. This is because basic wages should not contain anything but time billed to a project for a client. That is, basic wages do not contain the costs for such items as sick leave, vacation, holiday, incentive pay, etc., so the higher multiplier is used. Both multipliers are designed to produce the same result from either base, for example: Basic Wages Multiplier Revenue $ 100 3 to 3.3 $ 300 to $ 330 Salary Cost Multiplier Revenue $ 135 2.3 to 2.5 $ 310 to $ 337.50 When using the basic wage base, the employees working on a particular project accumulate the hours spent on timesheets and the multiplier is applied to the total time. If the higher multiplier is used, *624 certain types of salaries should not be included in the total, such as support services, e.g., administrative and personnel staff or general support for the firm such as maintenance or janitorial services. Also, certain overhead costs which would not be billed to a client should not be included in the basic wage base, e.g., payment to professional and support staff during down time between projects. Mr. Parthum's testimony was candid, forthright, and credible. He reviewed Mr. Trepashko's engineering report and concluded that the use of the 300-percent multiplier was much too high because Mr. Trepashko included 100 percent of petitioner's salary expense. Mr. Trepashko's calculations would result in double billing clients and reflect an unrealistically high income figure for Westreco. This was characterized by Mr. Parthum as a "fundamental error" in Mr. Trepashko's report. We have difficulty in believing Mr. Trepashko's testimony. Despite repeated admonitions of the Court, he continued to be evasive and could not or would not respond to simple and direct questions regarding his use of the salary multiplier method. For example, petitioner introduced a letter from respondent's*625 counsel listing the contents of the administrative file which was prepared by Mr. Trepashko. The list included the manual from the ASCE containing the methodology for using the salary multiplier. Mr. Trepashko testified that he had never seen the manual, although he admitted that he was the only engineer working on petitioner's audit. We do not believe his testimony that he never looked at the ASCE manual. The statutory notice of deficiency in this case did not, in any way, modify the engineering report prepared by Mr. Trepashko. Mr. Trepashko states in the report that his application of the salary multiplier is derived from his experience while at Foster Wheeler Corp. (Foster Wheeler) as a project manager and at Mobil Oil Corp. (Mobil Oil) as a project engineer. Mr. Trepashko testified that he worked at Mobil Oil from 1950 to 1957 and he thinks he worked at Foster Wheeler for around 8 months in 1959. He could not recall whether Foster Wheeler billed clients for staff time spent on vacations or for salaries attributed to administrative personnel. He admitted that he did not know whether Foster Wheeler included indirect salaries or overhead costs in the salary base when applying*626 the 300-percent salary multiplier. With regard to his employment at Mobil Oil, Mr. Trepashko stated that he coordinated the services of Fluor Corp. (Fluor) and The Bechtel Group, Inc. (Bechtel) for new plant development design, construction, startup, and operation. He admitted that he had no knowledge of whether those corporations billed Mobil Oil for employee time attributable to paid leave such as vacations and administrative duty. Further, Mr. Trepashko testified that he would not have approved a bill which included such charges. In using the 300-percent salary multiplier to calculate petitioner's "corrected income", Mr. Trepashko contradicts his own experience in this area. Mr. Trepashko stated that he knew that a portion of petitioner's income was derived from technical assistance services provided to TNCo. The income from TNCo assistance was not at issue in respondent's section 482 determination, yet Mr. Trepashko did not allocate petitioner's salaries between services performed for Nestec and those performed for the TNCo. Mr. Trepashko used total Westreco salaries as reported on petitioner's Federal income tax returns as the salary base. This is an obvious error in *627 Mr. Trepashko's calculations. Petitioner produced abundant evidence to establish that the salary multiplier, as applied to petitioner's total salary amount, was not an appropriate method of determining the fees which petitioner should have charged Nestec. Mr. Trepashko based petitioner's income adjustments on the billing methods of Fluor and Foster Wheeler. These two corporations are primarily involved in the design and construction of large-scale operating plants and industrial complexes, e.g., oil refineries. Petitioner, on the other hand, is engaged in process and product development at the pilot-plant level, and is not involved in such large-scale construction projects. They are not comparable to petitioner. Petitioner presented calculations at trial, using the salary multiplier method, and it corrected respondent's flawed salary base and reallocated the salary figures. Petitioner began by adding premium overtime wages (which had not been allocated to specific projects) to the "projects" salary amount: [1982] Projects $ 2,487,569 $ 2,793,738 $ 3,281,327 $ 4,774,467 $ 5,106,822 Overtime 93,139 93,328 96,244 113,750 105,563 $ 2,580,708 $ 2,887,066 $ 3,377,571 $ 4,888,217 $ 5,212,385 *628 These salary figures represent Westreco's total salary expenditures for work performed for both Nestec and TNCo. The amounts billed for services to TNCo were not reallocated by respondent and, accordingly, are not in issue. It is, therefore, necessary to allocate these labor costs to the Nestec contract. Petitioner contends that the appropriate salary base to which the salary multiplier should be applied is as follows: [1982] Westreco expenses related to Nestec agreement $ 7,835,167 $ 8,585,990 $ 10,396,112 $ 14,370,635 $ 16,738,158 Total Westreco expenses $ 8,517,539 $ 9,468,677 $ 11,183,363 $ 17,645,762 $ 18,364,549 Ratio of Nestec expenses to total expenses 91.99% 90.68% 92.96% 81.44% 91.14% Salary $ 2,373,958 $ 2,617,928 $ 3,139,807 $ 3,980,943 $ 4,750,769 base The question now becomes what multiplier should be used against the salary base computed by petitioner. Petitioner contends that a multiplier of 3, which Mr. Trepashko used, is too high. Petitioner argues that the purpose of the salary multiplier is to generate revenue sufficient to cover nonbillable costs. If a firm has a lower amount of nonbillable*629 costs than other firms, its salary multiplier should be adjusted accordingly. One nonbillable item enjoyed by Westreco was the lack of down time between projects. As a result of Westreco's contract with Nestec, Westreco had a steady stream of work and its employees were always engaged. Mr. Parthum, petitioner's expert, observed that Westreco should be able to use a multiplier of less than 3, which he concluded would be the more appropriate salary multiplier. Petitioner also presented other arguments to support a lower salary multiplier, but chose instead to use the multiplier which respondent's agent used to demonstrate the erroneous result in Mr. Trepashko's report and the statutory notice. The salary multiplier of 3, when applied to the correct salary base, results in the following amounts of revenue for Westreco: [1980] Salary base $ 2,373,958 $ 2,617,928 $ 3,139,807 X 3 $ 7,121,874 $ 7,853,784 $ 9,419,421 [1982] Salary base $ 3,980,943 $ 4,750,769 X 3 $ 11,942,829 $ 14,252,307 The actual revenue which Westreco received from Nestec under their agreement was as follows: [1982] $ 8,140,648 $ 8,917,749 $ 10,791,226 $ 14,904,857 $ 17,355,243 *630 Thus, if the salary multiplier method had been applied to the correct salary base and with a multiplier of 3, Mr. Trepashko would have concluded that petitioner overpaid its income tax for the years in issue because it reported substantially higher revenue figures from Nestec on its income tax returns. Petitioner maintains that these conclusions merely demonstrate that the salary multiplier method does not accurately result in correct revenue figures for a unique firm like Westreco. For example, Mr. Parthum, petitioner's expert, testified that engineering firms that use the salary multiplier method usually require reimbursement of direct nonsalary expenses associated with a client's project. Normally, an engineering firm has no expenses for raw materials. However, Westreco spends a large amount of money for raw materials to be used in its research. Because Westreco is not comparable to traditional engineering firms in this respect, the conventional method for billing direct nonsalary costs may not result in accurate figures. Petitioner offered the testimony of its expert, Dr. Clark Chandler, to support its position that no reallocation of fees was justified. Dr. Chandler*631 is the vice president of Economic Consulting Services. He holds a B.A. degree from Dickinson College and M.A. and Ph.D. degrees in economics from the University of Michigan. Dr. Chandler compared petitioner with four corporations to determine if petitioner's pricing practices and financial results were similar to those of the comparable corporations. The four corporations which Dr. Chandler used were: 1) Arthur D. Little, Inc. (ADL), a publicly held corporation that provided contract research, engineering and management consulting services to clients. ADL also had a division involved in food and beverage research during the years in issue. ADL generally provides its clients with reports and laboratory products and processes rather than large scale prototypes or specific machinery. 2) Artisan Industries, Inc. (Artisan), a closely held corporation that did customized research and development for clients. These services included a substantial amount of work on distillation and evaporation process and design, and the construction of large pressure vessels. Artisan usually markets the results of its work through the sale of processing systems or machinery. 3) Gulf Machinery*632 Company, Inc. (Gulf), a closely held corporation that designed and built machinery used in citrus processing plants for citrus processing companies. 4) Knechtel Research Sciences, Inc. (Knechtel), a closely held corporation that provided contract research in the food science area, particularly with respect to confectionery products, such as large candy companies. It develops new products along with the prototype process technology used to produce such products. Dr. Chandler based his comparison on a number of factors: (1) Business relations with clients, (2) financial comparability, (3) economic and business risks, e.g., client risk, general business risk, research risk, and downstream market risk. (1) Business Relations with Clients Knechtel and ADL generally act as pure contract research firms for their clients in which the client identifies the project and pays for the results, whether commercially successful or not. The clients in this arrangement own the rights to the technology developed and realize all the profit, if any. This is similar to petitioner's arrangement with Nestec. Gulf and Artisan, on the other hand, develop technology for their own accounts*633 and, consequently, own the proprietary technology which they develop. Artisan on occasion operated as a contract research organization for some clients. (2) Financial Comparability Knechtel is significantly smaller than Westreco, while ADL is significantly larger and more diversified. Westreco had a higher ratio of fixed assets to total assets because of its larger investment in fixed plant and equipment; however, its ratio of sales to both total and operating assets is approximately the same as that of the four comparable corporations, as illustrated: 1978-82 Five-Year Averages Net Revenues Divided By Net Total Operating Total Operating Revenues Assets Assets Assets Assets ADL: Consol. $ 156,425,600 $ 79,019,600 $ 69,036,400 1.98 2.27 Food/agri. 3,270,019 [2] Artisan 8,626,414 4,055,510 3,902,748 2.13 2.21 Gulf 16,011,500 8,937,473 8,489,166 1.79 1.89 Knechtel 524,910 239,155 191,280 2.19 2.74 Unweighted [2] 2.02 2.28 avg. Weighted avg. [2] 1.97 2.22 Westreco 13,528,173 6,262,535 6,177,017 2.16 2.19 *634 (3) Economic and Business Risks All of the corporations, including Westreco, faced some client risks because if the quality of the work fell below standards demanded by the client or if costs became excessive, the client might take the business elsewhere. Westreco was somewhat insulated from day-to-day risk because it did not have to regularly solicit new clients for work or existing clients for additional work. As a result, there was no guarantee that the four corporations used as comparables would be able to bill sufficient hours to cover their fixed labor and overhead costs. The four comparable corporations also faced higher general business risk because Nestec was contractually committed to compensate Westreco for all its expenses except taxes. The other corporations faced the possibility of general increases in costs which could not be passed on to clients, cost overruns on specific projects that a client may not be willing to absorb, and changes in the scope of work required which could lead to additional costs that could not be recovered, as well as the inevitable bad debt risk. Similar to petitioner's arrangement with Nestec, Knechtel and ADL faced only limited*635 research risks, in that their clients were generally committed to pay for the research and development projects regardless of whether they led to commercially viable products and/or processes. Artisan and Gulf, although trying to minimize this risk by performing work only after negotiating a contract with a client, nonetheless put their own money into certain projects without a guarantee that these products would become commercially viable. Westreco was particularly comparable with the four corporations in the area of downstream market risks. The corporations were compensated under a purchase price or fee arrangement rather than through licensing, royalties, or profit-sharing arrangements. Dr. Chandler, satisfied that the four corporations were comparable to petitioner's operation, performed several calculations to compare and evaluate financial performance between petitioner and the comparable corporations. His financial computations consisted of four ratios: (1) Operating income 13 as a percent of net sales; (2) operating income as a percent of operating assets; 14 (3) pretax income 15 as a percent of net sales; and (4) pretax income as a percent of total assets. Dr. Chandler's*636 results were as follows: 1978-82 Five-Year Averages Operating Income As Pre-tax Income As A Percent Of A Percent Of Net Operating Net Total Company Revenues Assets Revenues Assets Consolidated results: ADL 7.2% 16.2% 7.1% 14.1% Artisan 4.3% 9.6% 4.0% 8.5% Gulf 5.7% 10.7% 5.3% 9.5% Knechtel 6.0% 16.3% 9.4% 20.6% Unweighted avg. 5.8% 13.2% 6.5% 13.2% Weighted avg. 6.9% 15.3% 6.8% 13.4% Segment results: ADL -1.6% [1] -1.6% [1] Artisan 4.3% 9.6% 4.0% 8.5% Gulf 5.7% 10.7% 5.3% 9.5% Knechtel 6.0% 16.3% 9.4% 20.6% Unweighted avg. 3.6% [2] 4.3% [2] Weighted avg. 4.4% [2] 4.2% [2] Westreco 3.1% 6.9% 2.9% 6.2% Westreco results after IRS adj. by Revenue Agent Trepashko 24.8% 70.1% 24.6% 68.5% *637 Operating Income as a Percent of Net Revenues 1978-82 Company [1982] ADL: Consolidated 8.6% 7.9% 6.6% 7.2% 6.1% Food/agribusiness [1] -2.1% -7.8% -7.8% 9.2% Artisan 7.1% 7.6% 4.9% 1.1% 1.5% Gulf 11.7% 6.7% 4.5% 4.2% 1.2% Knechtel 10.0% 7.0% 6.9% 4.6% .6% Unweighted avg. 9.3% 7.3% 5.7% 4.3% 2.3% Weighted avg. 8.8% 7.7% 6.3% 6.6% 5.7% Westreco 3.2% 3.1% 3.3% 3.3% 2.8% Westreco results after IRS adj. by Revenue Agent Trepashko 28.8% 29.4% 26.3% 19.7% 24.1% Dr. Chandler concluded, from the data analyzed, that Westreco's profits were likely to be somewhat lower than those of the four comparable corporations because two of them (Artisan and Gulf) developed commercially viable technology on their own account and owned the rights to that technology. Westreco's profits were relatively more consistent from year to year than other comparables because it faced lower risks. The comparable corporations would *638 be expected to realize higher average profits to compensate for their higher risks. Accordingly, Dr. Chandler opined that the fees received by petitioner from Nestec were comparable to such income from services performed for unrelated corporations. Respondent, at trial, did not rely upon the method used in the statutory notice. Instead, she argued that comparable fees required reallocation of the fees paid to Westreco by Nestec. Respondent's new method of allocating the fees was presented by her expert witnesses Dr. Dennis Carlton and Dr. Richard Leftwich and their written report. Dr. Carlton holds a B.A. in mathematics and economics from Harvard University and an M.S. and Ph.D. from the Massachusetts Institute of Technology (MIT). He is vice president of Lexecon, Inc. (Lexecon). Dr. Leftwich has a bachelor of commerce degree from the University of Queensland, Australia, an M.S. in business administration, and a Ph.D. in economics and finance from the University of Rochester. Drs. Carlton and Leftwich posit that Westreco should have earned from Nestec a "plus" over its costs of 13.17 percent. The contract provides that Nestec will pay Westreco an amount equal to the reimbursable*639 costs incurred plus an additional markup (or plus) based upon reimbursable costs. Reimbursable costs include all salaries and wages, all fees paid to consultants and advisers, all rents, all raw materials, all supplies, all cable and postage expenses, stationery, and all office equipment and maintenance. The reimbursable costs do not include any fee, franchise tax, income tax, or any other tax paid or to any government authority within the United States. The contract obligated Nestec to pay Westreco 7.5 percent on its first $ 350,000 of costs, 5 percent with respect to its next $ 1,150,000 of costs, and 3.5 percent on costs over this amount. Drs. Carlton and Leftwich found that the appropriate plus fell between 10.76 percent and 15.57 percent, based upon their financial analysis of corporations which were "comparable" to Westreco. The midpoint of that figure, 13.17 percent, is thus the proper "benchmark" plus which Westreco should charge Nestec. They concluded that if Westreco receives a plus percentage which is more than this amount, then its revenue is too high and, consequently, a plus below this percentage is too low. Drs. Carlton and Leftwich arrived at their 13.17-percent*640 benchmark by selecting 15 "comparable" corporations from Standard & Poor's Compustat Services, Inc. database. This is a commercial database of publicly traded corporations which are indexed according to Standard Industrial Classification (SIC) codes. They began with the code that Westreco used in filing its income tax returns, IRS business code 7389 (which is for "business services, other than advertising"). They translated this to the closest SIC codes: Number 8711 (for engineering services) and number 8731 (commercial physical and biological research). 16 The earliest year in which Compustat listed companies for these SIC code numbers was 1981. The 15 companies which they selected as comparables to petitioner are: (1) Baker (Michael) Corp. - *641 this corporation's services include planning, design, and construction phase services in the heavy civil, mechanical, electrical, and building design area. The market for this corporation's services are governments and quasi-governmental agencies. (2) COAP Systems, Inc. - this corporation has between 1 to 3 employees and is engaged exclusively in research and development of the atri-rotor engine. (3) Comarco, Inc. - this corporation provides computer-based systems, and support of computer systems, to the Federal government and prime contractors and some private industry. (4) EG&G, Inc. - this corporation's products include radiation depictions, nuclear instrumentation modules, and computer based instrumentation systems, as well as polarographic and instrumentation systems and equipment for oceanic exploration and calibration equipment for pressure gauges. (5) Gilbert Associates, Inc. - an independent employee-controlled engineering firm which develops and markets voice communications systems for electric generating stations and other industrial and mining operations. (6) Hazelton Laboratories Corp. - this corporation is divided into 3 segments: *642 (a) biological and chemical research for specified physiological effects of various chemical, pharmaceuticals to other products, (b) manufacture and sale of equipment for use in biological research, human orthopedics, and in the clinical care of animals, and (c) importation, breeding, and sale of research primates and canines for use by biological research facilities. (7) Impell Corp. - this corporation provides engineering, management, and technical support services to the electric power generation industry, principally to nuclear power segment. (8) International Research & Development Corp. (IRDC) - an independent research laboratory engaged in safety evaluation of drugs and other chemical compounds, primarily to assess the hazards of use of such products as drugs, cosmetics, coloring agents, food additives, and pesticides. (9) The Parsons Corp. - this corporation's services include oil and gas production and refining, as well as petrochemical and petroleum product development. (10) Polymer Research Corp. of America - this corporation provides research and development for industrial equipment, tire manufacturing, packaging materials, and plastic film manufactures. *643 (11) Quadrex Corp. - this corporation provides a wide range of engineering and technical services to the nuclear power industry, including all phases of planning, design, construction, startup, maintenance, and operation of nuclear power plants. (12) STV Engineers, Inc. - this corporation's market is all types of government agencies. It provides civil, structural, mechanical, electrical, and architectural services in the area of urban planning and land development, highway and bridge engineering. (13) Systems Planning Corp. - a corporation involved exclusively in public and private urban planning, including aerial survey and photogrammetric mapping services (e.g., topographic and planimetric maps). (14) TSC Corp. - this is a multi-disciplinary consulting engineering design firm for architects, developer-builders, planners, and industrial clients. (15) VTN Corp. - this corporation provides comprehensive and integrated engineering, architectural, planning, surveying, and environmental services, including project management in the areas of airport planning, recreational facilities design, and water resources engineering. After calculating Westreco's*644 ideal return on assets from the financial data of these corporations, Drs. Carlton and Leftwich calculated the "corrected" plus percentage which they argue Westreco should earn in total, which they calculated to be 9.3 percent on all revenue. Then they subtracted non-Nestec revenue sources (TNCO, the domestic Nestle operating companies which Westreco assisted) and were left with Nestec revenues upon which they divided that amount by the reimbursable cost to arrive at the plus for the Nestec portion of revenue. To check their work, Drs. Carlton and Leftwich used the Capital Asset Pricing Model (CAPM), a well-recognized financial model designed to predict the rate of return, adjusted for risk. Applying the CAPM, they arrived at a benchmark rate of return on assets for Westreco of 11.4 percent. They considered this to be a confirmation of the accuracy of their ideal benchmark of 13.17 percent. The mathematical computations performed by respondent's experts are not difficult to follow. Their 13.17 percent benchmark is based upon their range of three separate figures: An operating income to operating cost ratio, a return on asset percentage, and a rate of return calculated with the*645 CAPM. They weighted these figures to make up for the disparities in the size of the comparables. They then averaged the three numbers to arrive at 13.17 percent. Petitioner attacks the assumptions made in the report and the methodology. We agree with petitioner but also find the fundamental premises upon which these figures are based to be defective. Respondent's experts began with the assumption that there was something amiss with Westreco's rate of return on its assets based upon comparables that they examined. They attributed this low rate of return to a low "plus" figure used by petitioner in its contract with Nestec. Dr. Carlton testified that a low rate of return results from having an unsuccessful business. After a visit to the facilities of Westreco and a meeting with its management, Dr. Carlton testified that he came away with the impression that Westreco's business operation was successful and management was happy with the performance of the firm. Thus, Dr. Carlton concluded that petitioner should use the average plus derived from the analysis of the comparables. It is obvious why petitioner's return would be lower when matched with these purported comparable corporations. *646 Drs. Carlton and Leftwich fail to consider key differences between these corporations and Westreco. There was very little, if anything, comparable in the corporations they selected and Westreco. As explained by Dr. Chandler, petitioner's expert, at least one reason for the reduced revenue which Westreco earned was the fact that there was a great deal less economic and business risk for Westreco. In fact, research and downstream risk was virtually nonexistent because Westreco was paid regardless of success or failure on its projects. It is a fundamental economic principle that firms which do not take risks can expect to earn a lower rate of return. This concept can readily be seen from the standpoint of a consumer as the difference between a passbook savings account (a safe but low return) and a junk bond (speculative but a much higher return). Moreover, as Dr. Chandler noted, it is doubtful that Westreco would have been able to bargain for a higher profit, regardless of the sophistication or commercial value of its technology, in light of its weak bargaining position with Nestec. The successful commercialization of Westreco's technology required at least: (1) Basic research, *647 which was conducted by Nestec, not Westreco, (2) applied research and pilot plant development, which was conducted by Westreco but could be shifted to another Reco, notably Linor, which also did research in the instant coffee and tea area, (3) engineering and process design work to develop commercial scale production processes, which Nestec and TNCo controlled, (4) marketing research and evaluation, also the purview of Nestec and TNCo, and (5) the successful introduction of the product into the marketplace, which was the sole responsibility of TNCo. The most important fallacy in the report of Drs. Carlton and Leftwich is the lack of comparability between petitioner and the 15 enumerated corporations. Unfortunately, like the proverbial house of cards, nothing that follows from their analysis can be relied upon because appropriate comparables are the linchpin of the report's accuracy. After identifying the corporations on the basis of SIC numbers, Drs. Carlton and Leftwich did nothing to ensure that the functions of the corporations in reality approximated petitioner's business operations. Dr. Carlton, at trial, emphatically stated that comparability is assured when relying *648 solely upon SIC codes because the codes "naturally group firms that do similar things together." We cannot see how architectural services, nuclear power plant construction and operation, map making and oceanographic services, tire making, breeding of research primates and canines, and packaging material development can be compared with petitioner's food research. We have held that allocations based blindly on statistics without consideration of specific aspects of a taxpayer's operation is, by definition, arbitrary. PPG Industries, Inc. v. Commissioner, 55 T.C. 928, 993 (1970). At trial, respondent's expert testified: Q. [By petitioner's counsel]: Did any of your comparables merge during the years '78 to '82 -- 1978 to 1982, to the best of your knowledge? A. [By Dr. Carlton]: I don't believe they merged with each other, to the best of my knowledge, no. Q. I'd like to explore just a minute, if I could, your criteria for comparability. As I understand it, your criteria for comparability was SIC Codes; is that correct? A. That was certainly one of the bases, yes. Q. And you didn't examine your individual comparables from the particular*649 line of business that they were in; did you? A. The individual firm? Q. Yes. A. I -- as long as the individual firm was in the SIC Code, I -- and it says -- Q. You didn't explore the matter any further as to what their business would happen to be? A. That's -- I did read it -- a description, but if they were in my sample, they were in, yes. Q. You didn't look at what unique risk each of these comparables might have faced; did you? A. That would be correct. Q. Okay. You didn't look at what particular intangible assets each one of these comparables owned; did you? A. I didn't take that into account in my analysis. Q. You didn't look at what level of trade they operated in; did you? A. Again, I looked, but it didn't affect my analysis. I mean, I might have -- I've read their 10-Ks. Q. Dr. Carlton, are you telling us that you determined the level of trade which each of your comparables functioned at? A. No, I'm not saying that. Q. You have no knowledge about that, really; do you? A. No. Q. And you didn't examine the exact functions that each of your comparables performed; did you? A. Not more than reading the 10-Ks. *650 Q. Well, you made no independent inquiry that would have given you specific knowledge of the individual functions they performed; did you? A. Not beyond the 10-Ks, right. Q. Now, you didn't examine who their individual customers might be; did you? A. That would be correct. Q. You didn't examine whether they were involved in technological development or not; did you? A. Again, my answers to all of these are the same: other than reading the 10-Ks, I didn't do any further exploratory work. Q. If you would just answer my question; you didn't determine whether your 15 sample-able companies were involved in technological development, did you? A. Well, yes, I read the 10-Ks which indicates that sometimes. THE COURT: You didn't eliminate any; did you? THE WITNESS: No, that's correct. THE COURT: Isn't it fair to say that your statistical sample of comparables is strictly statistical? THE WITNESS: Strictly based on SICs. THE COURT: Strictly statistical; those 15 comparables are strictly statistical comparables -- no other factors? THE WITNESS: Yes. The factors SIC Code, '89 and the same one in '81. THE COURT: If it's SIC, they're in, and*651 you don't care what else is -- you don't care about any other comparable factors beyond that? THE WITNESS: That would be correct. Respondent's experts ignore fundamental differences between their selected comparables and petitioner. For example, under the contract between Westreco and Nestec, Nestec is the owner of all patent and other proprietary rights created and developed by Westreco. In contrast, EG&G, Inc., a comparable used by respondent, owned over 80 patents during the years in issue. On its Form 10-K filed with the Securities and Exchange Commission for its fiscal year ending December 28, 1980, EG&G, Inc. reported: Patents and Trademarks While the Company's patents, trademarks, and licenses are cumulatively important to its business, the Company does not believe that the loss of any one or group of related patents, trademarks, or licenses would have a material adverse effect on the overall business of the Company or on any of its business segments. Dr. Carlton testified that he did not attempt to determine which of his proposed comparable corporations owned patents, if any, and whether any proprietary value should be attributed to them. Another comparable *652 used by respondent, Quadrex Corp., filed a Form 10-K with the Securities and Exchange Commission for the fiscal year ending January 21, 1980, stating that it obtained several design patents and patents were pending with respect to improving certain proprietary codes. The Form 10-K further states: "Registrant [Quadrex Corp.] believes these patents are important and will play a significant role in [Quadrex Corp's] business." Dr. Carlton testified that he made no attempt to quantify the relative income of EE&G, Inc. or Quadrex Corp. by virtue of their patent ownership. Patents are intangible assets which can be quite important in business operations. In light of petitioner's contractual arrangement with Nestec requiring petitioner to convey all patents and proprietary rights to Nestec, it is obvious that respondent's experts should have taken the absence of patent ownership by petitioner into consideration when they identified comparables. Further, as indicated above, a large number of respondent's comparable corporations are under contract to supply products and/or services to the United States Department of Defense (Comarco, Inc. for computer software and systems used in the evaluation*653 of air strike attacks and for logistical tracking; EG&G, Inc. for nuclear weapons technology; and, STV Engineers, Inc. for defense systems and military hardware development engineering). Some of respondent's comparables have markets identified as "Federal, State, and municipal government agencies" without specifying the department with which they contracted. Petitioner introduced a study prepared by the General Accounting Office (GAO) which states that there are many economic indicators which establish "that Government contractors have consistently earned a higher rate of return than commercial manufacturers." Respondent's experts, however, did not attempt to modify their data to accommodate these established differences in rates of return: Q. [By Petitioner's counsel]: Do you have any reason to doubt the GAO's conclusion that defense contractors were substantially more profitable than commercial firms? A. [By Dr. Carlton]: I don't have any reason to doubt that particular statement. Q. Nevertheless, you used defense contractor data, even though the GAO concluded that defense contractors were more profitable than commercial firms, for purposes of comparison to Westreco? *654 A. Yes, we used the defense industry as a, in our literature review. Q. Simultaneously, you had access to the GAO study which bore on the relationship between defense and commercial contractor profits? A. We had access to this study. Yes. By using comparables with higher income, petitioner's income obviously appeared artificially low. The choice of comparables by respondent's experts is so flawed that we would hardly classify them as "comparables". Respondent's experts did not appear to fully understand petitioner's relationship with Nestec. Petitioner's attorney asked Dr. Carlton whether in "the contract between Nestec and Westreco, is Westreco guaranteed a certain profit?" to which Dr. Carlton stated "I don't believe so." Petitioner's attorney attempted to clarify Dr. Carlton's understanding of petitioner's relationship with Nestec: Q. [By petitioner's attorney]: The parties have stipulated that Nestec pays Westreco a profit each year, depending upon the sliding scale of plus, no matter what? A. [By Dr. Carlton]: I agree with that statement. Q. You don't disagree with that stipulation, do you? A. I agree they get paid a plus. Q. Would you agree*655 that that plus is a profit? A. It is an operating profit on their assets. Q. The answer to my question is yes, then. A. Yes. This colloquy demonstrates the lack of familiarity which respondent's experts displayed with respect to petitioner's operations. Even if we were to follow the theory of respondent's experts and find that the 15 enumerated corporations were comparable to petitioner, other defects in the report render it unpersuasive. It utilized the operating income of the 15 corporations from the Compustat database calculated before interest and taxes. Dr. Carlton admitted at trial that he used Westreco's operating income after State income taxes. Petitioner's counsel asked Dr. Carlton: Q. And logically, if your report calculates Westreco's income, operating income, after state income taxes, there's a problem of comparing apples and oranges with your sample. Is that correct? A. [Dr. Carlton]: Well, I would certainly want to adjust the Westreco numbers to make them comparable. I don't know how large -- you know, if it's a trivial amount, it wouldn't make much difference, but to be consistent, I would make the adjustment, yes. When the difference*656 is corrected and Westreco's proper (after state income tax) operating income figure is used, Dr. Carlton acknowledged that Westreco's corrected rate of return is 7.54 percent. Thus, when Westreco is placed among the 15 sample corporations, Dr. Carlton admitted that "around a third -- 30 percent of the sample had lower [rate of returns] and 70 percent, roughly, had higher [rate of returns], or two-thirds." Petitioner's counsel asked Dr. Carlton: Q. [Westreco's rate of return is] in the range of observed arms-length [sic] dealings. A. [By Dr. Carlton]: Yes * * * -- they are in that broad range. Accordingly, it is apparent by the admission of respondent's own experts, that petitioner is within an acceptable range when the measurement is made using publicly traded corporations which are presumably dealing at arm's length. We agree with this admission of respondent's experts and the conclusion of Dr. Chandler (petitioner's expert) that petitioner's fees from Nestec were comparable to fees between uncontrolled corporations. As a result, no recollection is warranted, and, thus, there is no deficiency in petitioner's Federal income tax. For the reasons we have given above, *657 we have concluded that the fees which Nestec paid petitioner for its R and D services clearly reflected income under section 482. We come to this conclusion after weighing all of the evidence offered by both parties. To reflect the foregoing, Decision will be entered for petitioner. Footnotes 1. Agglomeration is the process of forming many small particles into a larger particle. Nestle uses this process to create its product called "Nestle Quik".↩ 2. This is a process by which water is removed from frozen coffee extract. ↩ 3. A process by which hot air is used to remove water from coffee extract. ↩ 4. A root-like plant used in combination with coffee beans.↩ 5. Segmentation is a marketing technique by which a producer of a product takes an otherwise nondescript item, e.g. sugar, shampoo, salt, or coffee, and develops an identifiable "line" readily identifiable to the consumer, e.g., using a trademark unique to the product such as "Sanka" or "Nescafe".↩ 6. Lipids are in the family of organic compounds, in this instance, fats. ↩ 7. Unit operations in food processing, in this context, means the effect, if any, of drying, heating, sterilization, and microwaving on food.↩ 8. This is a process which makes instant coffee have a more appealing texture. ↩ 9. A commercial method of brewing coffee. This is a step in manufacturing instant coffee. ↩ 10. This involves the recovery and concentration of volatile flavor substances, which are subsequently added back before drying. ↩ 11. Volatile components are added back to the product so that instant coffee powder smells like roast and ground coffee. ↩ 12. A process in which caffeine is removed.↩ 1. Not available. ↩ 2. Not applicable.↩ 13. Operating income is comprised of sales, less cost of sales, less selling, general and administrative costs, and excludes income from financial assets. ↩ 14. Operating assets are comprised of total assets less assets which generate nonoperating income. ↩ 15. Pre-tax income includes all income, including financial income and other nonoperating income.↩ 1. Not available. ↩ 2. Not applicable. ↩ 1. Not available.↩ 16. The expert report states that "12 [corporations] came from SIC code 8711, Engineering Services, and 3 [corporations] from SIC Code 8731, Commercial Research." However, 11 corporations are engineering firms and 4 corporations are research and development firms.↩