THE COMPETITION TRIBUNAL IN THE MAT IER OF an Application by the Director of Investigation and Research under section 79 of the Competition Act R.S.C. 1985 c.C-34 as amended.
AND IN THE MATIER OF certain practices by The D & B Companies of Canada Ltd.
BETWEEN: THE DIRECTOR OF INVESTIGATION AND RESEARCH Applicant
-and-THE 0 & B COMPANIES OF CANADA LID. . Respondent . ·
-and-INFORMATION RESOURCES, IN
REBUITAL EXPERT AF.FIDA VIT OF PROFESSOR FRANK MATHEWSON (on behalf of the Respondent)
COMPETITION TR:BUNAL ,,TRIBUNAL DE 0 CONCURRENCE File ~fo. . U -94/; ~"'JL';,".'' ';A) J f2~ - I I ~E-;hibil 1 No. _:_ L - /() .3 ( fL) No. Je l;i. piece · /:J.. . , ! J . filed on ~ ' ,),; fTl.:.t.!uJ~ 08pd•l• M1~ Re91~tra: ~ • Gremer /
AFFIDAVIT OF PROFFSSOR FRANK MATHEWSON I, Professor Frank Mathewson, in the City of Toronto, in the Province of Ontario, MAKE OATH AND SAY: 1 I delivered an expert report in this proceeding on September 20, 1994. 2 Attached hereto as Exhibit "A• to this affidavit is a true copy of my Rebuttal Expert Report prepared in response to the Expert Reports delivered on behalf of the Director and . IRI.
Sworn befcn me at the ) City of Toronto in the ) )J~J~ Province of Ontario ) this 4th day of October 1994. ) Professor Frank Mathewson dvw w- Ere:J:-----A Commissioner, etc .
.... a . t . • , • , _ Dlllf. I C1•111ll1,_., & 1111 GI llllnlPOlftlll Toranto fol ,_,. ..,, ........... SOllcltal'L ........... , .. 1. ..
EXHIBIT A REBUTI'A L EXPERT REPORT OF FRANK MATHEWSON IN THE MA TIER BETWEEN THE DIRECTOR OF INVF.STIGA TION AND RESEARCH AND NIEUEN MARKETING RFSEARCH
Frank Mathewson Imtitute for Policy Analym and Department of F.conomlcs University of Toronto
140 St. Georp Street Toronto Ontario M5S lAl
October 4, 1994
Table of Contents Page Number 1 The Affidavit of Dr. Ralph Winter (a) Summary of Dr. Winter 3 (b) Comments on Dr. Winter's Affidavit 4 2 Comments on the Affidavit of Donald N. Thompson 8 3 Comments on the Affidavit of John P. Gould and Andrew M. Rosenfield 10
1. The Affidavit of Dr. Ralph Winter (a) Summary of Dr. Winter 1.1 To begin, I summarize the central claim in Dr. Winter's report: exclusivity in scanner based data contracts inevitably leads to a monopoly outcome. (The reason is that monopoly profits always dominate oligopoly profits -- see Appendix 1 of Dr. Winter's Affidavit.) In this outcome, major grocery distributors in Canada are organized through a single intermediary, such as Nielsen or IRI, whose position is enforced through exclusive contracts. The intermediary then sells a scanner based national market tracking service at a monopoly price to manufacturers who need such a service to market successfully their products in Canada. In this claim, the monopoly outcome does not restrict the production of scanning data but does restrict the use of that data by setting monopoly prices for its use. The ensuing. monopoly rents are somehow divided between the retailers and the intermediary. (See conclusions 8, 9, 10, and 11 on pages 4 and S of Dr. Winter's Executive Summary.)
1.2 To continue with Dr. Winter's claim: If exclusivity were to lead to two intermediaries in the market, for example each holding a collection of data, perhaps with regional representation, this too would be inefficient: products would be artificially differentiated, leaving products less valuable, and increasing prices. (See conclusion 12.)
1.3 There is still more to Dr. Wmter's claim. Scanner data is a public good that should be made available to all. (See conclusion 13 on page S.)
4 (b) Comments on Dr. Winter's Amdavit 1.4 In contrast, my position is that exclusivity of retailer contracts was not and still is not anticompetitive.
1.5 Dr. Winter claims that the only motive for exclusivity in the contracts with the major grocery distributors is Nielsen's desire to achieve a monopoly. But there are other efficiency reasons why retailers may prefer exclusive arrangements. For example, the evidence, as I understand it, indicates that the retail grocery distributors have indicated a preference for exclusive agency for reasons of confidentiality of their data and the need for only a single source of processed scanner data for their internal management needs.
1.6 Importantly, exclusivity did not prevent entry by IRI or anyone else into the production and sale of market tracking services in the Canadian market place over the period 1986 to the present. Had it chosen to do so subsequent to 1986, IRI could have chosen to bid in a sequential manner on the contracts of the major grocery distributon as each was renegotiated. Had it chosen to do so, IRI could have entered producing non-scanner based market tracking services. IRI chose to do neither of these.
1. 7 Market entry by IRI or anyone else sequentially securing Canadian grocery distributon as their Nielsen contracts expired was not and is not precluded by Nielsen's existing contracts. Furthermore, regional or retailer-specific success by IRI would provide a valuable asset. Any firm such as IRI that had regional or specific retailer representation
5 would have bargaining power in any subsequent attempt to assemble a national market tracking service for users. This means that a successful entrant would have something valuable to trade for data. from other parts of the Canadian grocery distribution network. Of course, the request for assembling a representative national sample could come from a user such as a manufacturer. And that manufacturer would have access to either one of the intermediaries with a partial sample, or to another outside firm to assemble a national sample.
1.8 Dr. Winter's theory, however, precludes such an arrangement. He claims that even if his monopoly scenario were incorrect, two intermediaries using exclusive contracts and securing some subsample of the grocery distributors, would leave the market with artificially differentiated products that were less valuable, and increased prices.
1.9 The available facts contradict this claim. Paragraph 2.4.6 of my affidavit posits a hypothetical arrangement. The assumptions posited there, however, tit the Australian market as I understand arrangements in this market. In Australia, there are two intermediaries (Nielsen and South and Walker), each with exclusive rights to one or more of the major grocery distributors. (For example, South and Walker represents Coles; Nielsen represents Woolworths). Each intermediary then produces key account data on behalf of those grocery distributors that it represents. A demander of a scanning based national market tracking service can select Nielsen or South and Walker or a third party such as IRI to assemble and collate the data from the various distributors into a representative national sample. There is
• 6 competition among intermediaries for the business of marketing individual retailer data and for the business of assembling and collating representative national and regional reports. This appears to be a stable market outcome, that is not dominated by a single monopoly bid as Dr. Winter's theory imagines.
1.10 The facts from Australia, as I understand them, further illustrate entry potential with exclusive contracts. An entering intermediary could begin with an exclusive regional key account for a grocery distributor that operates only in selected states. There is demand for such state-specific key account data. In fact, IRI entered the Australian market with an exclusive agreement to broker the data on behalf of a major grocery distributor.
1.11 There is nothing inherent in the contracting practices of Nielsen which prevented or prevents entry and a similar market outcome from occurring in Canada.
1.12 Dr. Winter points out a public good feature of these data: in particular, its use by one firm does not increase the cost of its use by another. The claim is that market efficiency requires that the full set of scanner data be available to any competitor. 11te claim is a claim about a-post efficiency in my view. I have two problems with this claim.
1.13 The first problem involves initial fixed costs and a ante efllclency. Once the data have been collected, cleaned, processed and uwmhted, the additional cost of another user may be small. But the initial costs of collecting, cleaning, processing and assembling the
7 data are significant. And they are large as Nielsen discovered when it entered the business of developing a scanner based market tracking service in 1986. If the data were sold on an incremental cost basis to any party, who would have an incentive to undertake the investment that is required to generate the product in the first place? The obvious answer is no one. Products are developed and launched into the market with the expectation that the accompanying revenue stttam will yield a risk-corrected competitive rate of return. If, however, the product is made available to all on only an incremental cost basis, this anticipated revenue stream will disappear. Without an adequate return on any critical investment, the product or service could be absent from the market. What is critical in my view is this notion of ex ante efficiency, or the incentives in a market for some agent to create a product or a service.
1.14 The second problem involves property ri&bts. The owners of these data are the grocery distributors themselves, not the intermediaries such as Nielsen. And these distributors are concerned about the wide and indiscriminate dissemination of these data and the potential strategic use of these data by their competitors. 'Ibis potential governs the terms of trade and restrictions placed on the use of these data by their ultiJpate owners, the grocery distribut.on. 'Ibis means that, in contrast to the simple ex post notion of a public good, the ue of the data by competitors has a potentially harmful external effect on the profitability of the owner of the data, the firm that generated it in the first place. Recognition of this effect will dictate the conditions of the sale and distribution of the data by its owner.
8 1.15 I have one other concern with the monopoly claim in Dr. Winter's affidavit. In any conventional monopoly claim, the demanders of the monopoly output are passive buyers who are contract takers accepting the announced price of the monopolist. This is implicit in the monopoly claims made by Dr. Winter. The bulk of the buyers of the market tracking services of Nielsen and the users of decision support services of Nielsen and others, however, are not passive buyers but large and sophisticated manufacturers. The evidence is that the contracts with manufacturers are negotiated, and not presented on an take-it-or-leave-it basis. This 'countervailing' power of the manufacturers alone reduces any monopoly potential in the sale of any inputs to these manufacturers, including market tracking services.
2 Comments on the Amdavlt of Donald N. Thompson 2.1 On page 33 of his report, Dr. Thompson claims that there is a 'striking slowness' by Nielsen in its introduction of new scanner based services in Canada relative to the US. In particular, Dr. Thompson cites national scanner-based market tracking and key account services. The cited lags are respectively S and 3 yms. In my view, these lags do not reflect any lack of competitive discipline on Nielsen in C-anada. Rather, the lags follow from the 'slowness' of major Canadian, relative to American, grocery distributors in adopting scanner technology in retail outlets and in producing usable data. Decisions by major Canadian grocay distributors are not under the control of Nielsen. If anything, the facts are that Nielsen promoted, rather than discouraged, the introduction of scanner technology by Canadian grocery distributors and the development of usable scannina data. Nielsen invested effort and resources in improving these impediments to a scanner-based market tracking
service.
2.2 On pages 62 and 63 of his report, Dr. Thompson claims that there are currently high barriers to entry into Canada for a firm such as IRI because of Nielsen's current exclusive contracts. In a hypothetical, Dr. Thompson considers the entry strategy for IRI were IRI to bid on the Loblaws' contract. The first point is that IRI could have bid sequentially and independently on the contracts for various grocery distributors as they terminated over the period from 1986 to the present. I claim that this is an entry option which IRI did not exercise. The Australian facts, as I understand them, illustrate the potential for entry using such a sequential strategy.
2. 3 The second point is that IRI does not need to wait five years for an opportunity to bid on all other major Canadian grocery distributor contracts before competing for Loblaw's market research needs. The Australian facts, as I understand them, again provide a counterexample. Had IRI been successful in securing th~ position as a data intermediary for Loblaw, it could bid on market research needs for Loblaw in a manner open to all compctiton. Suppose that this is the only contract that IRI bad won (the rest going to Nielsen); suppoee further that Loblaw data are critical to a scanner-based market tracking service. Similar to the arrangements in the Australian market, IRI could have entered into a conttact with Nielsen to assemble and collate the data to produce a representative sample. Others have this potential as well. IRI could then offer its scanner-based market tracking service in competition with all others who had assembled the national data to offer a market
9
tracking service.
3 Comments on the Amdavit ol John P. Gould and Andrew M. Rosenfield 3.1 Ors. Gould and Rosenfield claim that Nielsen's exclusive contracts are the source of higher prices and lower quality services in Canada relative to the US. The claim is that these effects are the result of a lack of competition in Canada with IRI absent from bidding on retailer contracts to secure scanning data.
3.2 In the report of Dn. Gould and Rosenfield, price comparisons between Canada and the US are casual without a definition or examination of the accompanying market tracking service offered in both countries. Furthermore, there is no examination of the cost structure for market tracking services in Canada and the US.
3.3 My response to the claim that the development of market tracking services in Canada lagged behind that of the US is outlined above in paragraph 2.1.
3.4 Dn. Gould and Rosenfield claim that revenues accruing to grocery retailers are tantamount to an ex-post windfall pin as the incentives to install scanning equipment and to produce usable data capable of being assembled, processed and agpepted with ease into reliable regional or national market tracking services are complet.ely determined by the internal business needs of the grocery distributon. Once created for intanal business purposes, these data are available for all to use. In this sense, the claim is that scanning data
10
11 have a public-good feature. This claim runs counter to the facts of the development and launch of a scanner based market tracking service in Canada. Considerable investment was required of Nielsen and significant lags were experienced before reliable and usable scanner data were consistently produced on a national basis. Furthermore, as I understand the facts, Nielsen was instrumental in promoting the development of scanner technology in Canada. As a consequence, it is reasonable to claim that considerable incentives were offered by Nielsen to grocery distributors to quicken and enhance their use of scanner technology. Nielsen, therefore, has an investment in these historical data, even if the property rights to them remain with the individual grocery distributors. My additional comments on the applicability of a 'public-good' feature to the scanning data of grocery distributors are above in paragraphs 1.12, 1.13, and 1.14.