.· 5ENT BY:~cCARTHY TETRAULT TOR : 5-29-90 ; 15:14 t\E<;~otJ]Wf -~ £-, v~ {d, 1:4. uJ~ THE COMPETITION 'l~RlBUNAL lN THE MATTER OF an Application by the_. ..... --------------~1;2.._J Director of Investigation and Research for an order pursuant to section 75 of the Competition Act, R.S. 1985, c. C-34, ae amended, requiring that the Respondent accept the Exdos Corporation as a customer for the supply of a proouct B E 1' W E E N: THE DIRECTOR OF INVESTIGATION AND RESEARCH
.. and -XEROX CANADA INC. Respondent
REPLY _AFFIDA.YIT OF __ L~:!'.!'T~~D_ !'!!Y.ER~A~ I, LEONARD WAV!RMAN, Of the City of Toronto ir1 the J ··" :iicial District of York, MAKE O.ATH AND SAY AS FOLLOWS:
Attached hereto and marked as Exhibit 11 A" to this rry affidavit is a true copy of my reply to the Affidavit of ~homas ~- Wilson sworn the 14th day of May, 1990. SWORN BEFORE ME at the City ) ) of Toronto, in the Judicial ) I ) ~istrict of York, this 29th ~,JJ/ ~--·--·~,~ ) . LEONARD WAVERMAN ) )
416j621812~ 61? 957 3170:# 3 COMPnmoN TRIBUNAL T!?'BUNAL DE LA CONCURRENU c.r-n/r ~ l ' M A¥ f A I 29 1990 .r . ( -Y "'l Q A ) D c j OTTAWA, ONT. _Ji_ __ _ Applicant
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:14 4163621812~ 613 957 3170:# 4 In the Matter of the Director of Inve•tiqation an4 aaaearch and Xerox Canada Inc.
Raply to th• Report prepare4 by Professo~ Thoma• ~. Wilaon for the Director of Investiqation and Research
Statement of: Leonard Wavannan Centre for International Studies university ot Toronto Toronto, Ontario Date: May 29, 1990
This ls r~·.h.....i.. ~t .'....t.. f~l~ red lo ln th~ aff1dtl1:,·t of. .. . . ' r .. . ~"-'~,, ···~·-~ .ti:worn befor m his ............... ... ,. ........... . day at........... . .. .. ...................... ,D,qQ,
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:15 4163621812 ... 613 957 3170:# 5 1 SUJDll&D' 1. Below I rebut in detail the specifio assertions made by Professor Wilson. In my view, Professor Wilson's Report is mainly argument, with little factual basis to back up hypotheses advanced. The hypotheses and arguments themselves are based on an exceedingly narrow perspective of economic efficiency and market power. Therefore, Professor Wilson's conclusion in support of an Order to supply E~dos rests on suspect eeonomic reasoning and a lack of understanding of the reproqraphios market.
2. Professor Wilson's Report begins with definitions ot economic efficiency and market power which set unreasonable standards aqainst which to judqe any firm's actions. The basis of Professor Wilson's definitions are that firms $hould obey the textbook perfectly competitive model. Firms should produce small quantities of homoqeneoua products, where price equals marginal cost. In such cases, statie neoclassical economic effieieney is maximized and firms have no above normal profits or, what is equivalent to Professor Wilson, market power.
3. It is obvious that the reproqraphics industry and most real world industries would be judged harshly aqainst this standard of homoqeneous products and perfect competition. The
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:15 4163621812-+ 613 957 3170:# 6 competition in the reproqraphics industry, in all high - tech industries and in mo•t other indu•tri•• as well, is via new technology, new products, new services, i.e. product differentiation. Because the reproqraphics market consists of many firms competinq through research and development (R&D) to provide better and lower oost mean• of deliverinq imaqed copies, Professor Wilson argues that larqe firms would possess "non-trivial market power. 0 (paragraph 11, p. 6) :rn my opinion, this assertion is unwarranted, is derived from the assumptions mada and ignores the nature of competition in the reprographies market. ~here is no factual basis that appeara in the Direotor's experts' affidavits to back up this assertion.
As Benjamin Klein has recently written ("Tha Use of Economics in Anti-Trust Litiqation: Realistic Models of the Competitive Proeesa 11 , -rhe Law and Economics ot competition Policy, Fraser Institute, Vancouver, 1990): It is important not to confuse the existence of economic powar in thia aenae ot a neqatively eloped demand curve with the anti-trust concept of market or monopoly power. A firm has economic power only in the ••n•• that it is not a perfeotly competitive firm, that is, that it is not the type of firm assul!led to exist in the perfectly competitive model. However, every firm in the economy, except possibly the wheat farmers ot economic principles textbooks, faces a neqatively sloped demand ourve. This is a natural reault of the lack of product homoqeneity and imperfect consumer information that exists in the real world ••• Nevertheless, although the conditions of the abstract, perfectly competitive model are not present, this does not i~ply a
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:15 4163621612"" 613 957 3170:# 7 3 condition of competitive concern or the necessity for anti-trust remedy. (p. 425) 4. I disagree with the Competition Policy principle presented in Professor Wilson's Report. In Para;raph 47 he states; in essence, that a firm cannot change a bu•iness practice "unless the increased costs impoaad on buyers and their customers are offset by actual or prospective efficiency qains .••• 11 (p. 12). In my view, it is incorrect to review individual practices under such a policy principle. Moreover, the facts in this case do not support Professor Wilson's conclusions.
In my opinion, firms should be allowed to choose the means by which they do business unless those means are anti-competitive. surely there must be a difference between contract law, which mediates the riqhts of parties to an agreement, and competition low, which examines the competitive conditions in the market. The effect of the Director's experts arguments is that the ~ompetition Act would be used to ensure the ••rights" of a single firm (in this case, Exdos). My analysis rests on the principle that the 'ompetition Act ia grounded in the principles of competition.
5. Finally, Professor Wilson's recommendation that an Order be granted in thi• and similar cases is incorrect, even under the principle of "economic efficiency" that he articulates.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:17 4163621812""' 613 957 3170:# 8 4 Professor Wilson is implying, irrespective of any eontractual terms (or in this case, even purchases outside a contract), that the buyer b• qiven by the Tribunal an indefinite right to purchase with only the buyer being able to terminate. This is a hiqhly inefficient form of oontractin9. It would result I in firms with Professor Wilaon's definition of market power (i.e. a significant market.share) and Professors Wilson's and Gillen' a definitions of the relevant market (that firm 1 s production of proprietary parts) deciding to vertically integrate even when that was not an efficient choice.
Xerox Canada Inc. faces competition in the relevant end-market (repro9raphic services). XCI is alone in the best position to know which structure delivers its broad product line in the best manner to consumers and makes XCI as competi tiv aly efficient as possible.
What follows is orqanized under· the headings in Professor Wilson's Report.
lQ.onomic lffiai1npy and it1 11i•vano• an4 Market l9•tr and Etf iciency (Wilson paragraphs 2-lB, pages 1-6) 6. Professor Wilson defines economic eff ieiency a.s a firm•• minimizing unit costs and pricing at marqinal costs. In this definition, etticiency occurs only when the industry mimics
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:17 4163621812-; 613 957 3170:# 9 !5 the textbook model of perfect competition and. when "a norm.al rate of return on invaatmants (including investments in intanqibla capital such as advertising and research and development [R&DJ) ••• 11 is earned (paragraph 4, p. 2). As a corollary, market power ia defined aa the absence of perfect competition1 market power •xi•t• to Profe•aor Wilson when the firm has the ability to maintain prices above lonq-run costs. Professor Wilaon ties efficiency and the definition of market power toqether. "Market power is generally incompatible with complete economic ef:ficiencyu (paragraph 12, p. 4).
In my opinion, Professor Wilson uses unrealistic and thus artificial atandards for judqin9 market power and welfare losses (deviations from complete economic efficiency). Any lon;-run deviation from the idealized notion of perfect competition and price equalinq marginal eost he sees as evidence of market power and a reduction in welfare. Under thia standard, virtually every industrial firm would have welfare reducinq market power, since most firms can price above marginal costs. A. definition of welfare reducing market power that includes most firms is inadequate.
Professor Wilson equates long-run above normal rates of return with market power: and market power, he says, i• incompatible with economic efficiency. The fact that lonq-run, persistent, above normal returns can reflect competitive superiority is
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:15 4163621812~ 613 957 3170:#10 6 iqnored. Competitively superior firms earn rates of return (rents) on their •uperiority, and therefore the rate of return above the industry norm need not be axoessive. There ia a wide distribution of return on capital in most industries, indicatinq that soma firm• are more efficient than others.
Contrary to Professor Wilson'• views, above normal returns are not incompatible with competition and economic efficiency. Firms and individuals do, in fact, earn above normal returns in highly competitive markets, so equating normal returns with competitive markets is inconsistent with actual market competition in the real world (see Harold Demsetz, "Industry Structure, Market Rivalry and Public Policy 11 , ~ournal ct Lmi and Economigs, April 1973)'. Differential firm profitability moves resources from lower to higher valued uses, improving economic efficiency. Moreover, some firms 9row relative to rivals dua primarily to competitive superiority. With superiority and qrowth comes larger market ahare. Therefore, large market share need not imply economic inefficiency, as Professor Wilson suggests, but the opposite.
Professor Wilson quotes Landes and Posner in their definition of market power. He does not, however, quota Landes and Posner on the misuse of market share data to conclude that market power exists when in fact little exists. "Suppose firm 'i' has 80' of the market, there are no good substitutes, and existinq fit'lttS are
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:18 41 6 3 6 2 1 812 ... 613 957 3170:#11 7 currently operatinq at full capacity, but entry is relatively easy. It miqht be a miatake to conclude that firm • i' had market power. Suppo•• that in th• previous decade there had been both a rapid expansion in demand and a lot of entry into the industry. Assume further that thi• entry was responsible for a tall in firm 1 i 1 s 1 market share from an oriqinal level of nearly lOOt to ita present Sot level. This suqqasts a high supply elasticity of the competitive frinqe ••• Yet there is no ready adjustment to th• ma~k•t share measure of sot that would show that firm 'i 1 lacked. market power. Since in these circumstances market share is not a qood measure of market power, •• ai;ht want a rul• that a tin4in; of •i;nifi oant recent entry an4 output ea:pan•ion zaegataa aza infarezace of •arket power ,,•••4 on market ahar• alone.11 (Landes and Posner, p. 950, emphasis added) The. Fraser Institute of Vancouver published last weak the proceedings of a conference held at the University of Toronto in 1988 (The Law and ~conomics gf Competition Policy).
In one of the papers in this volume, Herbert Hovenkamp ("The Measurement of Market Power: Policy and Science") states: No markets are perfect; as a result, most firms in the real world maximize their profi ta at prices above marginal cost. A oompetition policy vigorously dedicated to eliminatinq market power in society would be so costly that it would drive us back in~o the Stone Age. (p. 4~) Mart1t1 and Compat!tio; 7. Professor Wilson describes the end market (the reprographics market) a• "a differentiated oligopoly with an active competitive trinqe 11 (paragraph 16, p. 5). He also states that "the evidence presented indicates that there obviously is
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:19 4163621812 ... 613 957 3170:#12 8 competition in the end product market, but it ie insutfieient to warrant the conclusion that xerox ha• little market power" (paragraph 16, p. 5).
I agree that the repro9raphics market is a "differentiated oligopoly. . , simply meaninq that product• are not homogeneous and that less than the perfectly competitive infinite number of firms exist. I categorically reject the implicit pejorative meaning in the words "differentiated oliqopoly" that any oligopolistic firm has s\ll:)etantial market power.
It is important to analyze the means by which firms differentiate their products in the reproqraphics market. Competition between products is based on machine-specific features (speed; two-sided copying ability; maqnifioation; document handlin;1 ate.): reliability of the vandorr distribution; costs and price; and technoloqy. Photocopiers also differ in their ability to provide image sharpness, copy uniformity, background whiteness, and resistance to smud9in9. Suyers are generally knowledgeable, espeoially for the higher speed, more expensive photocopiers. Firm• compete through service, price, differentiation, and R&D. In 1989, there were 113 new copiers introduced to the u.s. market, "the highest number to data" (Oataqu.est, CDIS Research Newsletter, 1990-91, p, 1). Competition in raprographic services is not as in the simple textbook perfectly competitive model where firms
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:20 4163621812""' 613 957 3170:#13 9 with known technology, perfect information and perfect foresight compete to minimize static production costs. Competition in reproqraphic service• is competition in providing new taohnoloqy, new feature•, and better and lower cost means of delivarinq imaged copies. Differentiation in reproqraphic services ia not a nieans of increasing :market power, but a :meana of competition.
Since Professor Wilson has perfect competition as the criterion against which to judge an industry, R&O and product differentiation are implicitly market power enhancinq and welfare reducing. Professor Wilson's view implies that any deviation from homogeneous products is harmful, just as any long-run deviation from normal returns he sees as potentially harmful. Superior goods and services, by virtue ot innovation in such factors as quality, design, cost, distribution, technological advances, etc., and the returns they generate, are not inefficiencies but the very things th~t a competitive market system is designed to foster. 1 In his paper in the Fraser Institute volume, "Th• Use of Economics in Anti-Trust Litigation", Professor Benjamin Klein has written much which is directly relevant to this caee. However, some of the assumptions upon which the perfectly competitiv.:t model is built are extremely 1 SH tl'le dl•cwaalon by bwen Br9Mer; llM•r~tt Powetz lmovetlons •nd Antf·Trust 11 fn lh1..t!.!! •Cid Eeonom!ct of Canpetltlen policy, Freaet Instft~te 1990.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:20 4163621812"" 613 957 3170:#14 10 unrealistic and often inappropriate for anti-trust analysis.,. Althou;h firms in the real world differ, for example, in location, types of product• and services, and perceived quality (reputation), the assumption of product homogeneity in the perfectly competitive mod.al lead.a to tha "efficient" result that price will equal marginal cost. This result, that the price of a product will equal the opportunity cost in term.a of the f ore9one real resources needed to produce an additional unit of the product, has desirable properties regarding economic efficiency. Howayer, tbil does npt imply thot th• qreater the deviation from the perfect~ compatitiya model. the l•ss ef(icient tht result. The mod@l i1 merely an ab1~ract econgmic con1truct. not a criteria for governmental intervention in the real world mark1t;lace. (pp. 420-421, emphasis added) 8. The recent history of entry and exit indicates that the market in which XCI operates in is competitive (as Professor Wilson aqrees) and that XCI has little market power (where Professor Wilson disagrees). My rebuttal consists of examining entry and exit.
world and 0 1 1. Karg1ts In 1972, Canon introduced its liquid toner process, which was licensed to Addresso9raph/Multiqraph (AM), Saxon, Riech and Copyer. In 1972, ISM and Litton entered the plain paper copyinq market. In 1975, Xerox Corporation siqned. a consent decree with the U.S. Federal Trade Commission allowinq certain of its copier patents to be licensed by other manufacturers. The Eastman Kodak company entared the medium and high volume segm1nts in 1975. The 3M Company also entered. in this period.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:21 4163621812'"' 613 957 3170:#15 11 In 1975, the Stanford Research Institute designed a liquid toner transfer teahnoloqy for Savin, who awarded manufacturing riqhts to Ricoh, Koniahiroku (Konika), Toshiba, Sharp and Minolta -- all of whom entered the North American market post 1975. Van Dyk, SCM, and A.S. Dick al•o entered. A Dutch firm, Oce van d•r Gruntan, developed a plain paper copier technoloqy in th• early 1970s, which was marketed in North America by Pitney Bowes (OCE now is its own distributor). By 1975 there were approximately 20 plain paper manufacturers in the world. A number of vendors purehaaed :manufactured copi11rs and placed their brand names on the machines. In the early 1980s, a number of tirm.s exited from the plain paper copier market -- Van Dyk, Dennison, Apsco, SCM, Saxon and AM. In 1981, IBM ended its research into a low volume copier and in 1988 IBM exited totally, selling its installed base to Kodak. 3M reorqanized its reproqraphics business as a joint venture with the Harris corporation in 1986: and exited in ~une, 1989, wholly sellinq out to the Harris corporation. New entrants, however, appeared in the 1980s -~ Matsushita, Kyotah Ceramic (Cybernet) and Sanyo. In 1ga3, canon launched th• desktop convenience copier. In 1983, there were more than 30 firms supplyinq plain paper copiers.
New tachnoloqies continue to proliferate. new eoloux- copying process -- cycolor -- licensed to Brother, Gestetner, Savin and Seiko Mead.
Canon developed a which has been canon,
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:22 4163621812-+ 613 957 3170:#16 12 Panasonic, Savin and Sharp have introduced dual-function (colour as well as black and white) maohin••· As noted in my affidavit, Xerox has introducad its Remote Interactive Diagnostics proc•••· Many supplier• are researchinq full diqital technoloqy machines, where the document to be copied t is received as a digital input and nreprographic services" then entails a convergence of computer, photocopier and fax machine -· an intelligent information production machine.
Industry maqazines and reports discuss the hiqhly competitive nature ot markets. For example, the March 1990 Dataquest CDIS Research New~letter states: "Copier vendors are f illinq out their product lines by moving into high-end, highly featured, and. specialty copier markets to meet the increasing competition the copier industry has been facinq. 11 (p, 1)
An article in the Rochester Times-Union of May 2, 1990, discussed the ailinq fortunes of Kodak's business system operations. Analyst Michael Ellmann of Wertheim Schroder & Co. estimated Ko~ak's profit margin on copiers to be 5\ in 1989 and that "Japanese competitors auch as canon, Sharp, Mita, and Ricoh -- formidable, low-cost producers that once confined their efforts to the low end of the copier market now are launching an assault on the high-volume, high-dollar arena." Ellmann said this escalating competition may keep the squeeze on marqins indefinitely.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:22 4163621812"" 613 957 3170:#17 13 Louis E. Slawetsky, president of Rochester-based Industry Analysts Ino., is quoted as sayinq, "[Kodak] looks across the street and sees Xerox continuing to introduce innovative technoloqies, then looks overseas and aaes no fewer than 10 Japanese competitors... There's aiqnificant competition no matter where you look." (Bgghester Times-Union, May 2. 1990) A November 8, 1989, Report by CAP International Inc., eatitled "US Copier Battleground -- a candid Perception", states: Although Japanese manufacturers are compelled to move upstream with more productive and reliable copiers in the u.s. marketplace, they continue at a startling rate to produce new models to replace existinq machines. This intense competition not only considerably shortens product life cycles, out is a drain on U.S. profitability in the lower to middle sector of the market •.• (p. 9) and The average convenience copier unit price fell at the rate of 9% per year between 1975 and 1995. (p. 11)
cana\1,1 Profe saor Wilson augqests that XCI may have market power in the end market if: a) cross price elasticities of demand are lowr b) it is a lower cost competitor; c) entry barriers are hight
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:23 4163621612-1 613 957 3170:#16 14 d) cross price elasticitie• of supply with rivals are low. (pp. 5-6)
Professor Wilson concludes that "in the absence of this information, we cannot conclude that Xerox has little market power." Professor Wilson, however, dQ•• come to a conclusion. "Since the products are differentiated and R.'D is an important element of competition, I would anticipate that a producer with a large ahare in this market would possess non-trivial market power." (p. 6)
Photocopiers are not currently manufactured in Canada. The 15 or so Canadian suppli•ra of reprographie services (the Appendices of my initial Report) all import photocopiers and supply service, parts and supplies.
Once distribution in Canada is set up, increased supply is easy as product has only to be delivered. As noted earlier, users demand photocopyinq services and can substitute different brands to satisfy demand. The sales data indicate the ability of firms to penetrate the market. Exhibit 2 (Confidential) to my original affidavit provided Canadian photocopi•r placement data by firm for th• 1983 to 1988 period as compiled by Dataquest. XCI'• market aha.re was above 40% in 1983 in three market seqments (#3, #5, #6), but only in two market seqments in 1988 (#5, #6). XCI's market •hares are
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:23 4163621812 ... 613 957 3170:#19 15 clearly tar below what they ware in 1975 (the year of xerox•• Consent Decree with the FTC) in all aa;mants but number 6.
Profeaaor Wilson examines the evidence contained in my atfidavi t and find• "non-trivial :market power" •••minqly based on the raasoninq that I provided no evidence of high cross price elaaticit iea of demand and aupply, no evidence of similar costs, and no evidence of low entry barriers. Sut Professor Wilson provides no evidence that suqgasts that the relevant factors point to market power. There im not one faot or one reference in Professor Wilson's Report. Instead, the existence of 11 non-triv ial market power" is inferred from firms having significant market share in a differentiated oligopoly.
I would suqgest that such reasoning and such lack ot evidence should not be the basis of a substantiation of a charge of anti-competitive behaviour under the Competition AQ.t. It is incumbent that a lack ot competition in the relevant market ~ - the end-market be shown. This has not been done in the affidavits produced by the Director to date.
In my opinion, the history of entry and exit and the prospect• of new entry into the r•prographioa mark•t indicate that elasticities are high and •ntry barriers not a deterrence.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:24 4163621812"" 613 957 3170:#20 16 one begins with t.ndea and Posner, "a findinq of siqnificant rieoent entry and output expansion negate• an interance of market power baaed on market share alone" (quoted on page 7 above).
Landes and Posner'• dictum is that one can infer elaaticities
of demand and supply from the facts without actually measuring elasticity values. Where entry and exist occur, products must be substitutable by users and between suppliers.
Significant recent entry has occurred in the low and medium speed aeqmants where Exdos operates. This entry and turnover of market •hare between firms is evidence to me of competition and a lack of significant market power for any firm, including XCI.
XCI has a very larqe share of the high volume market and new entry other than by Kodak Canada ha• as yet not occurred. Professor Wilson concludes ••xerox would appear to have a dominant position (90t mark•t shara) in the high end of the market ••. 11
XCI cannot, however, price in the hiqh volume end without takinq account of the intense competition in high volume copiers from Kodak Canada, the competition from lower volume machinaa, the oom:peti tio n from substitutes such as offset
j
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:25 4163621612-+ 613 957 3170:#21 17 printing and importantly, prospective entry. Aa I noted in my affidavit and above, siqnificant antry would appear to be imminent in the hiqh volume &e9Jll•nt.
Moreover, XCI 1 • market share in the hiqh end is due to ite i technological axcellanoe in producinq machines which reliably produce 100 copies per minute (cpm) or more. Thia market share is not evidence ot anti-competitive behaviour or entry barriers, but Xerox and XCI'• innate ability.
XCI earned its market share in by innovation, hiqh product quality, reliability, and superior service. Rivals have been unable to duplicate, let alone aurpasa, XCI's technical superiority in hiqh volwne copiers. XCI, as a licensee of Xerox Corporation, lost its patent protection in 1975, so rivals have been free to compete against XCI for yaars and many have entered all seqments of the market. In fact, in 1976 a number ot firms entered with high volume copiers, includinq SCM, AM, A.B. Dick, and Royal (Timothy F. Bresnahan, "Poat Entry Competition in the Plain Paper Copier Market 11 , American Econom~c Rayiew, May 1985). The lack of auccess of these entrants, except Kodak, in high volwne copiers is due to the superiority of Xerox products. Thus, product differentiation along these lines indicates XCI's competitive superiority over copier rivals, not anti-competitive market power.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:25 4163621812-+ 613 957 3170:#22 18 Zhe B1l1yant Ma;ket 9. Professor Wilson stat•• that a fi1"1D typically operates in more than one market, but that Xerox copier parts is a relevant sub-market where XCI has market power (paraqraph 18, p. 6). I
My reply2 to Profes•or David Gillen•• Raport 3 concentrated on why the market relevant to purposes of the competition Act was the end~market not the parts market, and not the Xerox parts market.
Professor Wilson views Xerox copier parts as a relevant produet market, apart from the end market for reprographic services. since XCI controls lOOt of its parts supply, it is a monopolist under Professor Wilson's and Professor Gillen'• market definition. If parts :for production in a manufacturing process, such as in the production of automobiles, television sets, computers, refrigerators, ate., constitute a separate market, than virtually all manufacturers who produce their own parts are monopolist• in their own parts and subject to investigation for purposes of the competition Act under this logic. According to Professor Wil•on, lack of competition produces this monopoly condition. Apparently, he believes 2 Attached to Ill'>' eHf~vft elated May 14, 1990. 3 Att•ched to hi• •fffdevit ef April 24, 1990.
jtNI ~Y:McCARTHY TETRAULT TOR 5-29-90 15:25 4163621812"" 613 957 3170:#23 19 there should be nu.meroua producers of parts in every production process. A remedy would be to force manufacturers to have multiple suppliers of their own parts, so other parts buyers, such aa downatream·tsoa, would not be dependent solely on one supplier. However, it it were more efficient to not be vertically inteqrated, but to contract with multiple outside vendors for parta, then competition would force firms to do so. When firms are, in fact, vertically integrated and self-sufficient in parts, it is more efficient than relying on outside suppliers. Under Professor Wilson's approach, this competition produced efficiency in production and organizational form is equated with monopoly, and should be ended.
The relevant market is, however, t.he end-market reproqraphic services -- for it is there that market power must exist for it to be exercised, That end-market, as I have shown, and as Protaaeor Wilson aqrees, is competitive.
Impact pf aefuaal tQ supply on \~• gperating lffigi1noy of lx401 uo it• cu1toa1u 10. Professor Wilson holds that Exdos customers will unambiguously lose if XCI will not supply Series •10• parts (paraqraphs 22 and 23, p. 7). This is unclear if Professor Wilson's def ini ti on ot the market is adopted. Usinq Professor Wilson 1 • market definition, XCI is a monopolist in its own parts. If
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:26 4163621812-1 613 957 3170:#24 20 this is the case, it ma~es no difference whether XCI sells parta to Exdos or directly to final and users. If XCI is a monopolist it will charge a monopoly price, whether it sells to Exdoa or to final customer•, •o end uaera will not by harmed by higher parts prices in the absence of Exdos. It the market is not aa defined by Professor Wilson, l:>ut rather consists of the joint. supply of copiara and service as I contend, then XCI cannot charge a monopoly price for parts and service since it faces competition from numerous other copier manufacturers.
Impact of B1fus11 to Bupply on xarox•a Operating Costs 11. ln my initial Report I stated that vertical integration between photocopiers and ••rvie• was t.h• common form of product delivery in the reproqraphics market, and was an efficient (La. low cost) means ot providing repro9raphic services. Professor Wilson ignores thia feature of oompeti ti on and iqnores the fact that third-party leasing oompaniaa insist on vertical integration (by requiring that the l••••• utilize the service of the manufacturer and not ISO'a). Professor Wilson inatead auggaata that many of the advantages ot vertical intagration (information flows for improved products and R&D -- externalities) are not diaturbed by having one small ISO and that, in any event, XCI could price ita aervice and parts properly to ISO•s to account for
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:27 41 63 6 2 181 2-1 613 957 3170:#25 21 any externality. Professor Wilson devoted paragraphs 25 to 45 to this expo•ition.
Professor Wilson ia correct in one point -- I did overstate tha need for 1oot vertical integration to maintain proper information flows (a larqe random sample would provide the information). Professor Wilson is, however, incorrect in the main thrust of his remarks. Exdos does free ride, and the relevant "externalities" cannot be properly priced. Vertical integration is an efficient form of competitive distribution.
In the remainder of this section, I deal with Professor Wilson's arguments paragraph by paragraph.
para. 'J.7 Professor Wilson admits that vertical intaqration can increase the use of preventive maintenance relative to ISOs. But he also claims that this doea not justify any refusal to supply Exdos. Instead, in hia view XCI should drive Exdoa out of business by superior efficiency. Professor Wilaon•a argument iqnores the free riding advantages of Exdos and the costs to XCI of having Exdos neqlect preventive maintenance. Th• cost consequences to XCI of lax preventive maintenance accumulates over many years. In the interim period, XCI suffers a reduction in brand name capital, as it• machine• fail more often, and a reduction in the present value ot tuture sales.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:2e 4163621812~ 613 957 3170:#26 22 para. 29 Profeaaor Wilson's plan for eliminating th• hiqhar costs ot lumpy ISO orders is to provide price incentives for ISOa to order more continuou•ly. Such pricing incentives entail costs for XCI, inoluding price discounts, lumpy orders, and inefficiant inventory levels.
In addition, Profaeaor Wilson's proposed solution of wait-listing ISOs 1 part• ordara in the event of supply shortages has the potential to impose even further costs on xcr. Any parts etook•out and downed copier maohine time rebounds to XCI, reducinq its reputation for reliability. Hence, the suggestion is no solution to the incompatibility of ISOs with XCI's inventory control.
para. 30 For the raaaona qiven above, the per unit coat of inventory maintenance for XCI increases if tsoa exist. Moreover, the issue is not the division ot total inventory coats between two parties, but the coats ot the antira distribution system --Exdos plus XCI.
para. 31 Professor Wilson concludes that the evidence he reviewed shows that refusal to supply will not reduce XCI's operating costs.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:28 4163621812-1 613 957 3170:#27 23 However, he provides no •upporting evidenee to discount the impact of an Order on -- disturbing potential economies of scale and of scope, inventory co•t change•, brand name reputation effects, and impacts on future XCI sales.
Im;licationa of R1fu111 to P••l for Incantiyea ••• para. 33-35 Professor Wilson contends that Exdoa presents no free rider problems for XCI and has never enqaqed in any manner of free riding on XCI. Professor Wilson does not address the free rider problem properly. on complex, reputation dependant machines, requiring strong service support, outside service companies have an incentive to lower costs by takinq short cuts on required aervioe. since customers have difficulty detecting whether maintenance problems are due to XCI or an ISO, and the ISO can shift blame to XCI by faulting Xerox, an !SO can invest less than XCI and fraa ride on XCI'a reputation for high quality. Alternatively, the ISO can reduce preventive maintenance service, causing photocopiers to be less useful and less valuable in the future. Hence, an ISO has an incentive to increase its wealth at the expense of XCI, who has no incantive to take cost-cutting, quality and reputation reducinq ahort cute.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:29 4163621812"' 613 957 3170:#28 24 para. '.35-36 Prof esaor Wilson argues that Exdos is alao hurt it Exdos provides poor service. The point is correct, but does not come to term• with the essential economic issue. The question is not whether the ••rviaa provider suffers some economic loss from poor sarvioe, but whether it auffer• All the economic loss. Under vartical integration, it XCI offers poor service, XCI bears .Ill the related coats in terms of loss of machine value, loss of reputation, etc. If Exdos provides poor service on an XCI machine, Exdos does not ~ear that loss that falls on XCI. Exdoa imposes an externality on XCI (and XCI in making poor machines an externality on Exdos) which vertical integration internalizes.
Professor Wilson also arques that Exdos has invested in its own reputation as a service provider. It is clear from the evidence (see Responses to Undertakings Nwnber s, May 7, 1990) that the brochures and proposals to customers by Exdoa reflect and play upon the XCI exparienoe of Reid and the service representatives employed by Exdos. The brochures used by Exdos are mere reproduction• (verbatim) of the XCI product brochures, with the only addition beinq that Exao1 indicataa it will be the seller or lessor of th• equipment and that it can also provide service. There is little evidence that Exdoa
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:29 4163621812 ... 613 957 3170:#29 25 has invested in its own reputationJ instead, it is free riding on its aaaociation with XC?.
Professor Wilson also states that used equipment is more likely to suffer breakdown and require repair when compared to new aquipment. This statement ahowa a lack of understandinq ot the reprOCJraphica industry and reproqraphics equipment in particular. Reproqraphiea equipment is designed to be maintained and repaired on a regular basis from the time it comes ott the manufacturing line and once installed in euatomer locations. The standard of maintenance applied ~Y XCI, or any authorized distributor of a brand of equipment, to newly manufactured f irat-of f-the-line equipment and that refurbished and installed 'in customer locations for a number of yaara i• the same, provided it is current teohnology. In addition, the basic models 90 through a maturity cycle from the time of their introduction. As service/maintenance experience is gained while installed in ouatomer locations, deaiqn changes/improvamenta are made at the manufaeturinq level which incorporate the experience so gained. The installed units benefit from such improvementa by having changes inoorporated into tha installed units by authorized service representatives. There is, therefore, no difference in the service required between newly manufactured and installed units of current·technoloqy. Por older technoloqy, there may be a di~ferent standard. For example, for currently
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:30 4163621812-1 613 957 3170:#30 26 marketed 11 10" aeries unita, the standard is the same; for pre-111011 series, the standard of maintenance may be different with the result that the pre-"10" series equipment may require more repairs.
In ••••nee, the distinction between "new" and 11 used" is not apt here. The batter distinction is between current technoloqy as "new" and old technology as "used"·
para. 37 Professor Wilson claims that providing service of last resort will not harm XCI. It is hard to imagine how it could de otherwise. XCI • • reputation suffers every time Exdos is unable to service a photocopier properly and then calla on XCI to fix the machine. Exdos can easily shi:tt blame to XCI, claiming defects in Xerox parts or design that XCI must repair. The lonqer a photocopier is down due to Exdos' failure to repair, the greater the harm to XCI's reputation and future salas. Again, poor service by XCI is also costly, but no ahiftin9 of the blame (or the losses) is possible.
Nor would correet pricing aolve the problem. The higher XCI raises the time and materials price to account for externalities, the lon9er will ISO• wait before calling XCI -- the qreater is the reputation loss to XCI.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 1s:31 41 63 6 2 181 2"" 613 957 3170:#31 27 para. 38 Cream skimming is a problem of unregulated markets that causes market failure and thus can bring about requlaticn. See the book by William J. Baumol, John c. Panzar and Robert o. Willig, Contestable Ma,k1t1 and the Theory of Industry Structure, (Harcourt, Brace, Jananovich, New York, 1982).
para. 41-42 Professor Wilson claims that it ISOs are to be terminated on the grounds of their relative inefficiency, then, instead of termination, XCI should simply compete them away with superior efficiency. Professor Wilson can come to this conclusion only by ignorinq the serious problem of free ridinq. Exdos can free ride in many areas. Exdos can hired experienced XCI repair personnel and not incur the costly process of soreenin9 and traininq job candidates. Exdos does not have to engage in the large-scale advertising' and product promotion expenses, as doea XCI, free ridinq on XCI 1 a investment in brand name capital. Exdos can en;aqe in short cuts by ignorinq preventive maintenance, contrary to XCI who has a strong incentive to promote preventive maintenance. Since XCI is the lessor on most of its copier placements, its incentive is to maintain their future value. Exdos has no incentive to maintain the resale value of newer XCI copiers.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:31 4163621812-1 613 957 3170:#32 28 Professor Wilson's arCJUD1enta seem to suqqest that XCI's interests are in conflict with ita end clients. This is incorrect. If ISOe were an efficient delivery system for the majority of ita clients, I presume XCI would use them. If • mi~ed distribution system were efficient (vertical integration and ISOa) from XCI 1 s perspective, aqain I presume XCI would use this mixed system.
para. 44 Professor Wilson argues that failure of this Tribunal to compel supply will discourage small firms from entering into exclusive relationships with larqe sellers since small firms will have no protection. against subsequent termination. Professor Wilson ia stating that such buyers should have contracts in perpetuity and that only the buyer should be allowed to terminate the contract.
What is being called for is the equivalent of lifetime employment, with only the smaller firm tree to terminate. This is a hi9hly inefficient form of contractinq. Parties must be free to terminate for cause as lonq aa it is not for malicious reasons. we do have courts which examine contract rights. When firms contract exclusively with one supplier they know tha risks of sinqle source dependency. Presumably, they are adequately compensated or they would not enter such contracts. Moreover, Professor Wil•on•s claim that smaller
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:32 4163621812-; 613 957 3170:#33 29 firms have no protection ac;rain1t an opportunistic future refusal to supply once a contract expire•, i9nores contract remedies and the coats to opportunistic firm• from such behaviour. Contract terms ean be da•iqned to ;uard against opportunism. Mor• importantly, a larc;re supplier dealing with a network of ISO• ha• a strong ineentiva to not terminate without proper cause. Opportunistic termination will breed retaliation by the remainin; ISOs and lower greatly the quality of future ISOa seeking to deal with the larger fir111. It will severely weaken the ISO network, harming the large supplier. Hence, there are also stronq non-contractual incentives to never terminate opportunistically.
para. 46-47 To this point in his Report, Professor Wilson had suggested the pQssibility of inefficiencies because the reproqraphica market doea not mimic the ideal of perfeet competition and because I presented no evidence on the values of the efficiencies.
In parac;raph 47, Profa111aor Wilson concludes ooata imposed on buyers and their customers are not offs et by actual or prospeotive •ffici•ncy qains to Xerox and others.,."
This conclusion ia unfounded.
11 that the increased
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:33 4163621812-; 613 957 3170:#34 30 Firat, no facts ara presented to back up this assertion. Second, an economic principle ia uaad in an usual •anner, as a Competition Policy principle. Prof•••or Wilson beqan his paper by axamininq pareto optimality (para9raph 2, p. 2) ; hare ha is suqqestin9 that individual business practices be conditioned by a "Pareto optimum" an unworkable rule. Third, as I indicated above, the facta show tha opposite of th• assertion ..,_ there are efficiency gain• from vertical integration.
QRDQlu1ion In conclusion, Professor Wilson argues that vertical integration into servicing by XCI is inaff icient because XCI has market power in parts (and copiers) and will achieve no added efficiencies. However, XCI does face competition and potential competition in copierar it has no long-run anti-competitive market power in parts.
Professor Wilson's position on vertical inteqration and compelling continued supply ia illoqioal. .Pirat, if XCI has a monopoly position in parts then forcing it to supply Exdoa will not end its monopoly position. seeond, if XCI cannot choose vertical integration in servicinq under Section 75, then it should not be free to choose vertical integration in any of its activities, much as sales, parts manufacturing, R&D, etc. Usinq Professor Wilson's reasoning, anyone who wants to sell XCI copiers or manufacture XCI
SENT BY:McCARTHY TETRAULT TOR 5-29-90 15:33 4163621612-1 613 957 3170:#35 31 parts should be free to do so, since XCI has a monopoly over its parts aanufacturing and copier •ales. Thus, XCI should be forced to use outside parts and sales firms in order to eliminate its monopoly power. Suoh a •olution would, of couraa, destroy XCI. It opens up XCI to all the opportunism that vertical intaqration serves to internalize. In ahort, it denies the whole basis for the existence of firms and their orqanizational boundaries. Whatever the potential pro-competitive merits of Section 7~, Professor Wilson's interpretation of it has decidedly anti-competitive consequences.
As stated in my affidavit and reply, a firm ahould be able to choose that competitive distribution system which maximizes its own profits. Unless it can be demonstrated that gmnpatition sutters because of a distribution practice, the losses of a particular distributor is irrelevant to Compatit1on Policy.
The recently circulated "Draft for Discussion Only on Predatory Pricing" (April 20, 1990) by the Director ot Investiqation and Research discusses when the Director would view a firm'• pricing praetioas as warranting investiqation. In that Draft, short-run market power is suggested as occurring when a firm has a market shar@ threshold of 35 per cent and ia at least twice the size of its next largest competitor.
SENT BY:McCARTHY TETRAULT TOR 5-29-90 16:34 416362~812~ 613 957 3170:#36 32 The Bureau'• criteria •how that XCI has no short-run market power in the low and medium •peed segment•. The evidanoe in industry reports and surveys shows that XCI has no lonq-term market power in the reprographios market.