Mr. Irving Williamson
Chairman, Section 301 Committee
Office of the United States
The Winder Building
600 17th Street, N.W.
Washington, D.C. 20508
Dear Mr. Williamson:
On March 20, 1996, Kodak filed with USTR the results of a survey it commissioned of Japanese retail film outlets. This new submission purports to document Kodak's claims of "systematic anticompetitive activities"*1 in the Japanese color film market. Specifically, Kodak argues that its survey demonstrates that Kodak is not widely available in Japan -- allegedly because an anticompetitive "distribution bottleneck" is preventing Kodak from getting on the shelf in retail outlets. In addition, Kodak claims that its survey shows that Kodak, when available, is generally not "competitively priced" -- allegedly because of vertical and horizontal price collusion that prevents Kodak from getting off the shelf.
A careful analysis of Kodak's submission reveals that Kodak -- once again -- has failed to prove its case. Kodak has not even attempted to rebut critical survey findings provided by Fujifilm that demolish beyond repair its "distribution bottleneck" theory. Furthermore, Kodak's survey results regarding both retail availability and "competitive pricing" are undermined by fatal methodological flaws. Although Kodak has twisted the numbers cleverly, in the end this attempt to document its claims cannot withstand close scrutiny.
A. Kodak Does Not Even Attempt To Rebut The Survey Of Tokuyakuten Customers
In its December 21, 1995 submission rebutting Kodak's "distribution bottleneck" theory, Fujifilm provided USTR with the results of two separate surveys on Kodak availability. The first was a survey of the customers of Fujifilm's tokuyakuten ("tokuyakuten customer survey"); the second was a survey of 1,970 randomly selected retail outlets to determine Kodak's coverage of the retail market ("retail coverage survey").
Kodak's March 20, 1996 submission claims to rebut the findings of the retail coverage survey. To that end, Kodak commissioned its own survey of 2,028 randomly selected retail outlets, and now asserts that its figures show that Kodak is not available in outlets accounting for 66 percent of total film sales volume. See Kodak's March 20, 1996 submission at tab 1, page 1.
In the next section of this letter we will identify the glaring methodological errors that undermine the findings of Kodak's survey. With respect to the "distribution bottleneck" issue, however, the most important thing about Kodak's new submission is what it does not include. Specifically, Kodak has not even tried to challenge the results of the tokuyakuten customer survey.
As discussed in detail in Fujifilm's December 21, 1995 submission, Fujifilm commissioned a survey of 4,772 retailers and dealers that buy directly from Fujifilm's four major tokuyakuten. These customers account for approximately 95 percent of the tokuyakuten's total film sales. The survey shows that 77.3 percent of the tokuyakuten's customers (by volume) already carry Kodak film, and another 10 percent have an active business relationship with a Kodak supplier. Thus, nearly 90 percent of the tokuyakuten's customers either buy Kodak already or have ready access to it. See Fujifilm's December 21, 1995 submission at 3-9.
The results of the tokuyakuten customer survey are devastating to Kodak's "distribution bottleneck" theory. Kodak argues that the tokuyakuten are an "essential facility" for selling film in Japan, and that without relationships with the tokuyakuten Kodak cannot achieve wide distribution of its film in the Japanese market. Yet Kodak's claim is belied by the results of this survey, which show that Kodak is already carried by the overwhelming majority of the tokuyakuten's customers. Relationships with the tokuyakuten would not gain Kodak anything; the vast bulk of the tokuyakuten's customers already buy Kodak or have established business relationships with Kodak suppliers.
In light of the crucial significance of the tokuyakuten customer survey, Kodak's failure even to try to challenge the results is striking. This omission by Kodak must be read as a tacit admission that its "distribution bottleneck" theory is baseless.
Accordingly, even if Kodak's survey results on retail availability were reliable, they would lend no support to Kodak's charge that it is blocked from the retail store shelf by anticompetitive activity. Assuming for the sake of argument that Kodak is not widely available in Japan, the unrebutted results of the tokuyakuten customer survey demonstrate that this situation has nothing to do with lack of access to the tokuyakuten. There is thus no causal connection between any alleged lack of availability and anticompetitive conduct by Fujifilm. Kodak's availability on the retail store shelf -- however good or bad -- is simply a function of retailers' independent business decisions in response to consumer demand. Accordingly, Kodak's low share of the Japanese market merely reflects its lack of serious business efforts to stimulate consumer demand for its products.*2
Kodak's tacit admission, in turn, renders a whole host of other allegations completely irrelevant. Among the now moot issues are:
In raising all of these issues, Kodak was attempting to demonstrate that Fujifilm exerted indirect "control" over the tokuyakuten "essential facility." Since, however, Fujifilm has demonstrated -- and Kodak has implicitly conceded -- that the tokuyakuten are not an "essential facility," whether or not Fujifilm "controls" its distributors is no longer relevant to Kodak's access to the Japanese retail store shelf. See Fujifilm's March 13, 1996 submission at 5-6.
B. Kodak's Survey Of Retail Availability Is Undermined By Fatal Methodological Flaws
In actuality, Kodak film is widely available in Japan. Kodak's new survey purports to show otherwise, but its findings are attributable to irremediable methodological flaws.*7 Significantly, the raw data of Kodak's survey and the raw data of the retail coverage survey commissioned by Fujifilm present very similar results. The final results of the surveys diverge only because of different methodologies used to interpret the raw data. A comparison of methodologies makes plain that the approach taken by the retail coverage survey is correct, and Kodak's approach yields a grossly distorted picture of the marketplace.*8
For the major retail outlet types surveyed, both Kodak's survey and the retail coverage survey show similar figures for Kodak's availability on a percentage of surveyed outlets basis. Thus, for photo shops, Kodak's survey shows that Kodak film is available in about 50 percent of the outlets. Kodak's March 20, 1996 submission at tab 1, page 2. The figures presented by Fujifilm are comparable: 52.3 percent for the local cities Niigata, Shizuoka, and Okayama; 57.0 percent for Osaka and Kyoto; and 66.7 percent for Tokyo. See Fujifilm's December 21, 1995 submission at Exhibit 8. For all other outlet types (e.g., supermarkets, discount stores, convenience stores, kiosks), Kodak's survey shows that Kodak is available in 21 percent of the outlets. Kodak's March 20, 1996 submission at tab 1, page 2. Again, the retail coverage survey's figures are similar: the weighted average availability for other outlet types is 16.6 percent for the local cities, 17.6 percent for Osaka and Kyoto, and 30.2 for Tokyo.*9
Furthermore, both Kodak and Fujifilm combined outlet type-specific availability figures into overall average availability figures using similar weight-averaging techniques. Both Kodak and Fujifilm relied on figures published in Photo Market 1995 that show the percentage of total film sales accounted for by each outlet type.*10 Thus, since photo shops account for 50 percent of film sales, Kodak's survey allocated 50 percent of its sample to photo shops. Similarly, Fujifilm calculated overall average availability figures by weight-averaging its outlet type-specific numbers in accordance with the Photo Market 1995 percentages. See Fujifilm's December 21, 1995 submission at Figure 10 (attached as Exhibit 4).*11
What Kodak failed to do, however, is to account for differences in relative sales volume within a given outlet type. See Kodak's March 20, 1996 submission at tab 2, page 3, endnote 2. This crucial omission completely undermines the validity of Kodak's figures. The problem is that there are enormous variations in sales volume within a given outlet type. For example, some very small photo shops may sell only a few hundred rolls per year, while large shops may sell over 1,000 times as much film. Yet Kodak's survey counts all photo shops equally regardless of size. Thus, if Kodak's availability is not randomly distributed, but rather is concentrated in large volume outlets, then Kodak's survey understates the actual extent of Kodak's market penetration.
The retail coverage survey commissioned by Fujifilm avoided this methodological pitfall by estimating the sales volume of each outlet surveyed and weight-averaging the raw availability data accordingly. See Fujifilm's December 21, 1995 submission at 27. The results show that Kodak is indeed concentrated in large volume outlets, and thus its market penetration is much better than the raw availability data would indicate. For example, even in the local cities where Kodak's market penetration is lower, Kodak is available in 52.3 percent of surveyed photo shops, and those outlets account for 77.4 percent of surveyed photo shop film sales volume. For supermarkets and department stores, the figures are even more dramatic: Kodak is available in only 22.9 percent of the outlets, but those outlets account for 62.6 percent of total volume. See Fujifilm's December 21, 1995 submission at Exhibit 8 (a copy of the relevant page is provided in Exhibit 5).
Thus, the failure of Kodak's survey to take account of relative sales volume within a given outlet type yields a distorted and unreliable picture of the Japanese marketplace.*12 Kodak's finding that Kodak is available in only 34 percent of surveyed outlets does not tell the whole story. It is unimportant that Kodak is available in only a minority of outlets if the rest of the outlets do not sell very much film. And this, indeed, is precisely the case, as the retail coverage survey presented by Fujifilm shows. By measuring the sales volume of each surveyed outlet, the retail coverage survey reveals that Kodak is available in outlets accounting for between 56.4 and 77.4 percent of total sales volume, depending on the region. See Fujifilm's December 21, 1995 submission at Figure 10 (attached as Exhibit 4). Thus, contrary to its claims, Kodak has succeeded in winning shelf space in those outlets that account for the healthy majority of film sales.
Beyond its failure to yield a true volume-based picture of Kodak's availability, Kodak's survey methodology is further flawed by the geographic distribution of its sample. Kodak has explained that its survey visited 2,028 outlets in 144 randomly selected cities. In so designing its survey, Kodak's apparent purpose was to distribute its sample evenly across Japan.*13 This survey design, however, is flawed, because film sales are assuredly not distributed evenly across Japan. Of Japan's 47 prefectures, just two of them -- Tokyo and Osaka -- account for 30.29 percent of total film sales. Another four prefectures -- Fukuoka (including Fukuoka city), Kanagawa (including Yokohama), Aichi (including Nagoya), and Hyogo (including Kobe) -- comprise another 18.62 percent of total film sales. Id. See Photo Market 1995 at 204 (a translated copy of which is provided in Exhibit 6).
As the above figures indicate, film sales in Japan are concentrated in areas with major cities. And as Kodak itself admits, Kodak's market penetration is best in Tokyo and other "larger cities such as Osaka, Fukuoka, Nagoya, and Yokohama." Kodak's March 20, 1996 submission at tab 1, page 4. Kodak's survey, by not concentrating its survey sample where film sales are concentrated, once again produces skewed results. And given that Kodak sales are concentrated in major cities, the results are skewed toward underrepresenting Kodak's true penetration of the Japanese market.*14
Thus, Kodak's attempt to document its alleged exclusion from the retail store shelf is unavailing. First, by failing to challenge the results of the tokuyakuten customer survey, Kodak has effectively conceded that its availability at retail outlets has nothing to do with any supposed "distribution bottleneck." Furthermore, Kodak cannot even show that -- for whatever reason -- its market penetration is limited. Kodak of course has tried to make this point, but the figures it presents are simply the reflection of a flawed methodology, not marketplace reality. By contrast, the retail coverage survey commissioned by Fujifilm avoids Kodak's methodological problems, and shows that Kodak is actually widely available in Japan.*15
C. Kodak's Findings Regarding Relative Retail Prices Provide No Evidence Of Market Barriers
Kodak claims that its survey shows that its penetration of the Japanese market is being stymied by retailers' pricing practices. According to Kodak:
Despite Kodak's claims, its survey provides no evidence of price collusion. First, its findings are marred by the fundamental methodological flaws discussed in the previous section. Furthermore, even if the findings were reliable, they provide no evidence of collusive pricing. As Fujifilm has already demonstrated in its March 28, 1996 submission, an objective review of the evidence shows that Kodak's charges of price fixing are completely baseless.
Kodak's conclusion that it is not "competitively priced" in between 60 and 66 percent of the outlets in which it is available reflects not marketplace reality, but simply a flawed survey methodology. As discussed in the previous section, Kodak's survey fails to take account of differences in relative sales volume within a given outlet type -- a critical omission that produces huge distortions in the survey's findings. In addition, Kodak's survey is further flawed by its failure to concentrate its survey sample in those geographic areas where film sales are concentrated -- i.e., large cities. The most that Kodak's survey tells us is that Kodak and Fuji brand film are priced similarly in a large number of retail outlets; it tells us nothing, however, about how small or large a percentage of total film sales is accounted for by those outlets.*16
Even if the survey results were accurate, they provide no evidence of price collusion. As discussed in Fujifilm's recent rebuttal of Kodak's price fixing allegations, we believe that Kodak's wholesale prices are -- in general -- somewhat lower (but not dramatically lower) than wholesale prices for Fuji brand film. This situation gives retailers three options: (1) reduce Kodak prices to the level that gives the same markup as that earned on Fujifilm products, thereby maximizing Kodak sales volume; (2) mark up Kodak until prices are roughly equivalent to prices for Fuji brand film, and earn higher unit margins on Kodak sales; or (3) do something in between. All of these options reflect normal, pro-competitive business strategies; in particular, for many retailers the ability to earn a larger markup on Kodak film is undoubtedly what persuades them to carry such a slow moving product.*17 The evidence presented by Fujifilm -- and even the flawed data presented by Kodak -- show that retailers are pricing all along this continuum. See Fujifilm's March 28, 1996 submission at 16-17, 22-23.
Kodak has provided no evidence whatsoever that Japanese retailers are refusing -- much less refusing in collusion with one another -- to pass through low wholesale prices to consumers. Kodak has not provided any documentation of a single real-life example of a retailer that has been asked to lower excessive markups but has refused. Indeed, the evidence is overwhelming that -- at least until recently -- Kodak has not even been trying to gain market share by underselling Fujifilm. In particular, it is hard to see how Kodak could complain about Kodak's being sold at the same price as Fujifilm given that, for the past decade, Kodak's manufacturer's suggested retail prices have been identical to or higher than Fujifilm's. See Fujifilm's March 28, 1996 submission at 4-17.
The available evidence shows clearly that Japanese retailers are willing to sell film at low prices. Multipack film at very low prices now accounts for roughly 40 percent of total Japanese film sales. And private and dual brand film -- including film manufactured by Kodak -- is sold at rock bottom prices at over 10,000 outlets in every prefecture of Japan. If Kodak wants to undersell Fujifilm, all it needs to do is reduce its wholesale prices substantially. Such a move would allow retailers to earn bigger unit margins and undersell Fujifilm sufficiently to encourage consumers to switch brands. See Fujifilm's March 28, 1996 submission at 18-20.
Once again, Kodak's "evidence" in support of its allegations turns out to be much less than meets the eye. Despite its impressive-looking presentation, Kodak's new survey boils down to a combination of cooked numbers and non sequiturs. Fujifilm is confident that any neutral, objective review of the facts will vindicate its position that the Japanese consumer photographic market is open and competitive.
William H. Barringer
Counsel to Fujifilm