Chadha et al. v. Bayer Inc. et al. [Indexed as: Chadha v. Bayer Inc.]

63 O.R. (3d) 22

[2003] O.J. No. 27

Docket No. C37224

Court of Appeal for Ontario

Austin, Rosenberg and Feldman JJ.A.

January 14, 2003

*Application for leave to appeal to the Supreme Court of Canada dismissed with costs July 17, 2003 (Gonthier, Major and Arbour JJ.). S.C.C. File No. 29651. S.C.C. Bulletin, 2003, p. 1152.

Civil procedure -- Class proceedings -- Certification -- Class action by home purchasers -- Purchasers alleged to have suffered damages from price-fixing conspiracy -- Representative plaintiff alleging that defendants conspired to fix prices which affected cost of concrete brick and paving stones used in construction of homes -- Certification refused -- Plaintiffs providing insufficient evidentiary basis to satisfy test for class proceedings -- Class Proceedings Act, 1992, S.O. 1992, c. 6.

The appellants, who had purchased a new home that contained coloured bricks and paving stones, were representative plaintiffs in a proposed class action against the respondents, who were the major manufacturers and suppliers to the Canadian market of the iron oxide pigments used to colour concrete brick and paving stones. The appellants alleged that between 1985 and 1991, the respondents engaged in a price-fixing scheme, thereby illegally increasing the price of bricks and paving stones. Although the appellants were not direct purchasers of iron oxide from the defendants, they alleged that as a result of the price-fixing activities of the defendants, they overpaid for their home. On a motion for certification, the record disclosed that if the appellants could prove their claim and show that the increased cost of coloured bricks was passed through to them as home buyers, the amount of each claim would be between $70 and $112 on a $150,000 home. Sharpe J. certified the plaintiffs' action as a class proceeding under the Class Proceedings Act, 1992 ("CPA"). However, on appeal to the Divisional Court, a majority of the court held that liability could not be a common issue and that the class action process was not the preferable procedure. Disagreeing with the motions judge, a majority of the Divisional Court held that the actual losses could not be proved on a class-wide basis or on the basis of statistics. The majority concluded that proof of loss could only be established on an individual basis and that, in this case, the procedures in the Competition Act, R.S.C. 1985, c. C-34 were better suited to the goal of behaviour modification. The Divisional Court set aside the certification order, and the appellants appealed to the Court of Appeal.

Held, the appeal should be dismissed.

The Divisional Court was correct in concluding that the issue of liability, including proof of loss, could not be a common issue. The appellants assumed but did not show what method of proof could be used to establish loss on a class-wide basis. The motion judge did not address the complexities of proving the extent to which the participants in a chain of purchase bear the higher price caused by an illegal conspiracy to fix the price of an ingredient product, and he erred by relying on the expert evidence filed by the appellants as the basis for the certification order. That evidence did not address the issue of what method could be used at trial to prove that all end-purchasers of buildings using materials with the respondent's iron oxide pigment overpaid for the buildings as a result. Rather, the appellants' expert assumed that higher costs of products would have been passed on to end-users. However, the assumed issue was the very issue that the court must be satisfied is provable by some method on a class-wide basis before the common issue can be certified as such. In the immediate case, the appellants presented no evidence from industry representatives to explain how the manufacturers and [page23] distributors of bricks and the developers of new homes price their products, and, in particular, whether there is a direct pass-through of the price of every component into the sale price of all homes, the relevance of the value of the land, and how other factors such as the real estate market and the individual bargaining process of the purchaser and vendor affect the price. The evidence presented by the appellants on the motion did not satisfy the requirements prescribed by the Supreme Court of Canada in Hollick v. Metropolitan Toronto (Municipality) of providing sufficient evidence to support certification.

On the issue of preferable procedure, the motion judge did not have the benefit of the Supreme Court's decision in Hollick and it is unlikely that he would have certified the class action had he not viewed liability as a common issue. He discounted apparent problems of self-identification of potential plaintiffs who might have large claims and who might therefore want to opt out of the class procedure. This was a problem of under-inclusion. There was, however, a problem of both under-inclusion and over-inclusion of parties in the class. These problems were masked by accepting proof of loss as a common issue. The appellants' argument that the goal of behaviour modification would be defeated if a class action was not certified was not compelling given the other deficiencies and the fact that the Competition Act provides criminal sanctions to achieve that goal. The appellants' argument that a rejection of class actions would bar class actions in price- fixing cases was not correct. The question of whether consumers should be able to use class proceedings to obtain relief from price fixing remained an open question. In the immediate case, the appellants did not provide the evidentiary basis to justify a class proceeding.

The Divisional Court was also correct in concluding that the class definition was in error because the definition was not objective but turned on the outcome of the litigation or the merits of the claim. Accordingly, the appeal should be dismissed.

APPEAL from the judgment of the Divisional Court (Somers, Thomson JJ., O'Driscoll J. dissenting) (2001), 2001 28369 (ON SCDC), 54 O.R. (3d) 520, 200 D.L.R. (4th) 309 (Div. Ct.), setting aside an order of Sharpe J. (1999), 1999 14812 (ON SC), 45 O.R. (3d) 29, 36 C.P.C. (4th) 188 (S.C.J.) certifying an action under the Class Proceedings Act, 1992, S.O. 1992, c. 6.

Cases referred to Bogosian v. Gulf Oil Corp., 561 F. 2d 434 (3d Cir. 1977); Caputo v. Imperial Tobacco Ltd. (1997), 1997 12162 (ON SC), 34 O.R. (3d) 314, 148 D.L.R. (4th) 566, 13 C.P.C. (4th) 163 (Gen. Div.); Hanover Shoe Inc. v. United Shoe Machinery Corp., 392 U.S. 481 (1968); Hollick v. Metropolitan Toronto (Municipality), 2001 SCC 68, [2001] 3 S.C.R. 158, 56 O.R. (3d) 214n, 205 D.L.R. (4th) 19, 277 N.R. 51, 24 M.P.L.R. (3d) 9, 13 C.P.C. (5th) 1; Illinois Brick v. Illinois, 431 U.S. 720, 97 S. Ct. 2061, 62 L. Ed. 2d 707 (1977); In re Linerboard Antitrust Litigation, 305 F.3d 145 (3d Cir. 2002), revg 203 F.R.D. 197 (E.D. Pa. 2001); Newton v. Merill Lynch, Pierce, Fennerd Smith, Inc., 259 F. 3d 154 (3d Cir. 2001); Robertson v. Thomson Corp. (1999), 1999 14768 (ON SC), 43 O.R. (3d) 161, 171 D.L.R. (4th) 171, 85 C.P.R. (3d) 1, 30 C.P.C. (4th) 182 (Gen. Div.), supp. reasons (1999), 1999 14780 (ON SC), 43 O.R. (3d) 389, 43 C.P.C. (4th) 166 (Gen. Div.); Taub v. Manufacturers' Life Insurance Co. (1999), 1999 19922 (ON SC), 42 O.R. (3d) 576n (Div. Ct.), affg (1998), 1998 14853 (ON SC), 40 O.R. (3d) 379 (Gen. Div.); VitaPharm Canada Ltd. v. F. Hoffmann-LaRoche Ltd. (2000), 4 C.P.C. (5th) 169 (Ont. S.C.J.)

Statutes referred to Class Proceedings Act, 1992, S.O. 1992, c. 6, ss. 5, 23, 24, 26 Clayton Act, 15 U.S.C. 15, s. 4 Competition Act, R.S.C. 1985, c. C-34, s. 36(1) Sherman Act, 15 U.S.C. 1

Rules and regulations referred to Rules of Civil Procedure, R.R.O. 1990, Reg. 194, rules 21, 39.01(4) [page24 ]

Authorities referred to Page, W.H., "The Limits of State Indirect Purchaser Suits: Class Certification in the Shadow of Illinois Brick" (1999) 67 Antitrust L.J. 1

Paul J. Pape, for appellants. J.L. McDougall, Q.C., and Kent E. Thomson, for respondents.

The judgment of the court was delivered by

[1] FELDMAN J.A.: -- The appellants are representative plaintiffs in a class action. The respondents were the major manufacturers and suppliers to the Canadian market of iron oxide pigments used to colour concrete bricks and paving stones which were incorporated into the construction of homes, buildings and landscaping. The bulk of their sales were for the construction of new homes. It is alleged that during the period between 1985 to 1991, the respondents engaged in a price-fixing scheme, thereby illegally increasing the price of concrete bricks and paving stones coloured by iron oxide pigment.

[2] The appellants purchased a new home during that period from a new home developer. Their home contains some coloured concrete bricks and paving stones. They believe that they were indirect purchasers of bricks and stones containing the respondents' iron oxide pigments. The appellants allege that they suffered damage by overpaying for their home.

[3] The issue under appeal is the propriety of certification of the class action. The issue turns on the efficacy and method of proof of whether all indirect purchasers of the respondents' product overpaid for their homes as a result, and thereby suffered damage. The majority of the Divisional Court held that damage, a necessary component of the cause of action of each plaintiff, could not be proved on a class-wide basis; rather, damage must be proved individually for each plaintiff, making the class action process not the preferable process. The dissenting judge essentially adopted the reasons of the motion judge. For the reasons which follow, I would uphold the conclusion reached by the majority of the Divisional Court and dismiss the appeal. [page25]

I. Facts and History of the Proceeding

[4] The motion for certification is based on s. 5(1) of the Class Proceedings Act, 1992, S.O. 1992, c. 6 which provides:

5(1) The Court shall certify a class proceeding on a motion under section 2, 3 or 4 if,

(a) the pleadings or the notice of application discloses a cause of action;

(b) there is an identifiable class of two or more persons that would be represented by the representative plaintiff or defendant;

(c) the claims or defences of the class members raise common issues;

(d) a class proceeding would be the preferable procedure for the resolution of the common issues; and

(e) there is a representative plaintiff or defendant who,

(i) would fairly and adequately represent the interests of the class,

(ii) has produced a plan for the proceeding that sets out a workable method of advancing the proceeding on behalf of the class and of notifying class members of the proceeding, and

(iii) does not have, on the common issues for the class, an interest in conflict with the interests of other class members.

[5] The Statement of Claim of the appellants describes the details of a conspiracy by the respondents who were manufacturers and distributors of iron oxide pigments for use in making bricks, paving stones and other building materials, and who, between 1984 and 1992, held between 90 per cent and 100 per cent of the Canadian market for iron oxide. The alleged conspiracy to fix and thereby raise the price of iron oxide pigments in Canada was carried out between 1985 and 1991. The appellants plead that they purchased a new house in Ontario which was built in 1988 with bricks containing iron oxide pigment supplied by the defendants. The pleading alleges that as a result of the respondents' price-fixing conspiracy, the purchase price of products containing pigments was increased over what it would have been had there been an open competitive market, and that the appellants and the rest of the members of the class suffered damages as a result by overpaying for their homes. There is no dispute by the appellants that their cause of action includes proof of damage. Under the heading "Effects of the Illegal Activities", para. 28 of the Statement of Claim reads:

The Plaintiffs plead that as a result of the Defendants' illegal actions, the Plaintiffs and class were harmed by having to pay higher prices and were deprived of the benefits of a free and open competition for the purchase of products containing pigments. [page26]

[6] The members of the class are described in the Statement of Claim as follows:

All persons in Canada who have suffered loss or damage as a result of the Defendants' agreement to wrongfully increase or maintain the price of iron oxide and black pigment and otherwise unduly lessen competition, and in general restrict and inhibit competition in the pigment market; in particular all persons who purchased either directly or indirectly, bricks or other construction products containing iron oxide pigment or black pigment manufactured or distributed by one or more of the Defendants (or, where applicable, their corporate predecessors), between 1985 and 1992.

(Emphasis added)

[7] The appellants subsequently revised the definition of the proposed class to exclude direct purchasers of the respondents' products, and to include only the ultimate end-users of the products after they had been incorporated into construction, particularly home owners. In his Order, the motion judge certified the class with the following description:

All homeowners or other end users in Canada who have suffered loss or damage as a result of the Defendants' agreement to wrongfully increase or maintain the price of iron oxide and black pigment and otherwise unduly lessen competition, and in general restrict and inhibit competition in the pigment market; in particular, all home owners or other end users of bricks, interlocking or other construction products containing iron oxide pigment or black pigment manufactured or distributed by Bayer Canada and Northern Pigment Company or where applicable, their corporate predecessors between 1985 and 1992.

[8] The respondents first moved before Sharpe J. under Rule 21 of the Rules of Civil Procedure, R.R.O. 1990, Reg. 194, to strike the claim on the basis that the pleading disclosed no cause of action. Because the appellants, as indirect purchasers, did not buy any iron oxide pigment directly from the respondents, but only purchased an end-product that incorporated bricks made with the respondents' product, it was argued that the appellants could not be the object of an unlawful conspiracy by the respondents to affect the price of their product. The Rule 21 motion was dismissed on the basis that, although the cause of action was a novel one, it was not plain and obvious that it could not succeed.

[9] The appellants then brought their motion for certification of the action as a class proceeding. On that motion, affidavit material was filed by both sides. The record disclosed that if the appellants could prove their claims and show that the increased cost of the coloured bricks was passed through to them as homebuyers, [page27] the magnitude of each claim would be between $70 and $112 on a $150,000 home.

[10] The respondents again raised the issue, this time under s. 5(1)(a) of the Class Proceedings Act, that the appellants' claim raised no cause of action, relying on authority from the United States Supreme Court (Illinois Brick v. Illinois, 431 U.S. 720, 97 S.C.T. 2061 (1977)), which precludes class action claims by indirect purchasers for damages for conspiracy to fix prices. The motion judge held, as he had on the earlier motion, that the pleading did disclose a cause of action, rejecting the American authorities as inapplicable in Canada.

[11] The second issue before the motion judge was whether there was an identifiable class. The problem raised by the respondents was that potential members of the class could have great difficulty self-identifying because, as home-owners, they would not know whether their home was built with materials which contained the iron oxide pigments. The motion judge described the problem as follows: "[t]he question is whether the impracticality or inefficiency of applying the definition to actually identify the members of the class on an individual basis renders it unacceptable." He approached the issue by considering the three important objectives of the Class Proceedings Act: (a) judicial economy; (b) improved access to the courts for actions that may not otherwise be asserted; and (c) behaviour modification for actual or potential wrongdoers. He held that in this case, the primary object of certification was behaviour modification and that because of the small size of any individual award, compensation for the appellants and therefore access to justice for individual claims would be a secondary goal, making the ability of potential plaintiffs to self-identify of less concern to the court.

[12] The third issue addressed by the motion judge was identifying and articulating three common issues. The first common issue was whether the respondents had entered into a price-fixing agreement. He defined the remaining common issues as follows:

Are the defendants liable to the members of the plaintiff class for conspiracy to fix the price of iron oxide, and if so, what is the appropriate measure of damages?

[13] The fourth matter addressed by the motion judge under s. 5(1) of the Class Proceedings Act was whether a class proceeding was the preferable procedure "for the resolution of the common issues". The motion judge concluded that it was. It is that conclusion and the basis for that conclusion, combined with the identification of liability as a common issue, that form the main focus of the appeals to the Divisional Court and to this court. The [page28] majority of the Divisional Court reversed the decision of the motion judge, concluding that liability could not be a common issue and that a class action was not the preferable procedure for determining the issues between the respondents and the members of the plaintiff class.

[14] Finally, the motion judge held that the appellants were representative plaintiffs. Although the appellants could not confirm that the bricks in their home contained the respondents' iron oxide pigment, or that they had overpaid for their home as a result, the motion judge found that there was a sufficient factual basis to qualify them as representative plaintiffs.

III. The Appeal to the Divisional Court

[15] The majority of the Divisional Court found that the motion judge had erred in his interpretation of s. 5(1)(d) of the Class Proceedings Act in the determination of whether a class action is the "preferable procedure". The motion judge had stated that ". . . the specific wording of s. 5(1)(d) . . . requires only that a class action be the preferable procedure for the resolution of the common issues." Somers J., writing for the majority of the Divisional Court, rejected limiting the preferable procedure analysis to the resolution of the common issues. Instead he took a broader approach, including considering the individual issues and whether a class action is the preferable procedure to advance the interests of all the parties in accordance with the objectives of the Class Proceedings Act.

[16] The Divisional Court then focused on the nature of the causes of action asserted by the appellants and the requirements for establishing those causes of action. The appellants' causes of action are based on a breach of s. 36(1) of the Competition Act, R.S.C. 1985, c. C-34, as well as the common law torts of conspiracy and infliction of economic injury by unlawful means. All causes of action require that the appellants establish that they suffered actual loss. In particular, s. 36(1) provides:

36(1) Any person who has suffered loss or damage as a result of

(a) conduct that is contrary to any provision of Part VI, or

(b) the failure of any person to comply with an order of the Tribunal or another court under this Act,

may, in any court of competent jurisdiction, sue for and recover from the person who engaged in the conduct or failed to comply with the order an amount equal to the loss or damage proved to have been suffered by him, together with any additional amount that the court may allow not exceeding the full cost to him of any investigation in connection with the matter and of proceedings under this section.

(Emphasis added) [page29]

The appellants' claim is that they suffered loss by paying higher prices for their houses, built using bricks containing the respondents' product.

[17] The Divisional Court disagreed with the motion judge that such a loss could be proved on a class-wide basis. The court concluded that proof that the appellants had suffered such loss by overpaying for their houses could only be established on an individual basis. The appellants would have to prove that any overcharge by the respondents to the direct purchasers of the iron oxide pigment was passed on through the chain of manufacture and distribution of the bricks to the ultimate purchaser of a home which was built using those bricks. The majority of the Divisional Court also focused on the multitude of variables that can affect the price of a building, including regional differences and delivery costs, and the fact that ". . . the product in question, iron oxide, is used merely as a small component in another product or series of products and the alleged overcharge is only a trivial part of the purchase price of residential or commercial buildings, which are highly individualized end products" (para. 23), as well as many subjective factors, such as the relative bargaining skills of the purchasers and vendors. The court also rejected the concept that statistical evidence could be used to prove the fact of loss, as opposed to the quantum of the loss. As a result, the Divisional Court concluded that the issue of liability could not be a common issue in the proceeding.

[18] The Divisional Court further found that establishing the price-fixing conspiracy would not advance the litigation in a legally material way, because the balance of the action would be an unmanageable series of individual trials of the "pass-on" issue. Therefore, certifying the class action would not further the purposes of the Class Proceedings Act, particularly judicial economy, but also access to justice and behaviour modification, the purpose singled out by the motion judge. The Divisional Court was of the view that the procedures in the Competition Act were better suited to the goal of behaviour modification in this case, and noted that an investigation under the Competition Act had been conducted and discontinued by the Director of Investigation and Research of the Competition Bureau.

[19] The Divisional Court found the following further errors in the Order that had been made below:

(a) the class definition is circular as it defines the class in terms that depend on the merits of the individual claims; [page30]

(b) the affidavit of a proposed representative plaintiff should have been rejected as it did not state that his home construction included paving stones which actually contain iron oxide manufactured by the defendants, and should have been amended to excise the portion stating his belief that the cost of his home had been artificially inflated when he did not state the basis for his belief as required by rule 39.01(4) of the Rules of Civil Procedure;

(c) the motion judge erred in concluding that s. 24 of the Class Proceedings Act could be used to assess damages on an aggregate basis.

IV. Issues

(1) Was the Divisional Court correct to conclude that the issue of liability, including proof of loss, could not be a common issue?

(2) Was the Divisional Court correct that a class action is not the preferable procedure for the conduct of the action?

(3) Is the class definition, as formulated by the motion judge, in error because it defines the class in terms of those who have suffered damages and not in objective terms, and therefore turns on the outcome of the litigation or the merits of the claim?

V. Analysis

(i) Common issue

[20] The difference in approach between the motion judge and the appellants on one hand, and the majority of the Divisional Court and the respondents on the other, turns on whether this is a case where all end-purchasers paid a higher price for their homes and therefore the loss can be proved on a class- wide basis, or whether each individual end-purchaser of a building that contains, as one component, bricks made with iron oxide pigment from the defendants, may or may not have had the inflated price of the iron oxide pigment passed through as part of the purchase price of the home they bought.

[21] The motion judge concluded that liability was a common issue and that it could be proved on a class-wide basis. Based on that premise, he turned to the issue of preferable procedure.

[22] The motion judge based his decision on preferable procedure on his view of the correct interpretation of s. 5(1) (d), which [page31] requires the plaintiff to prove that "a class proceeding would be the preferable procedure for the resolution of the common issues." His interpretation was that the section requires only that the class action be the preferable procedure for resolution of the common issues, not for the remaining individual issues. However, having already concluded that liability was not an individual issue, the spectre of an unmanageable judicial proceeding for determining liability on a plaintiff-by-plaintiff basis was not a concern for him.

[23] The critical finding of the motion judge on preferable procedure is at para. 22 of the reasons where he states:

If the plaintiffs are successful in establishing a price fixing conspiracy and in establishing that damages from such conspiracy flowed through to the ultimate owners of buildings containing the pigments supplied by the defendants, it will be for the trial judge to determine whether it is necessary to have individual hearings to assess and distribute damages. As I have already indicated, the Act contains provisions which contemplate damage assessment and distribution in cases of this kind without such individual hearings. In any event, for the purposes of the preferable procedure test, I have no difficulty in finding that a class proceeding is the preferable procedure for resolution of the common issues. This is not a case like Abdool v. Anaheim Management Ltd. (1995), 1995 5597 (ON SCDC), 21 O.R. (3d) 453 (Div. Ct.); Mouhteros v. DeVry Canada Inc. (1998), 1998 14686 (ON SC), 41 O.R. (3d) 63 (Gen. Div.); or Rosedale Motors Inc. v. Petro-Canada Inc., 1998 14721 (ON SC), [1998] O.J. No. 5461 (Gen. Div.); where it was simply not possible to resolve the common issues without scrutinizing the individual circumstances of each member of the proposed class. Here, there is an allegation of a general price-fixing agreement which is alleged to have a price impact upon the ultimate consumers of the product in question. If those are to be litigated at all, it seems apparent that a class proceeding is the preferable procedure. It would advance the goal of modification of behaviour as discussed earlier.

(Emphasis added)

And he concluded at para. 24:

While there is no doubt that this will be a complex action involving the claims of a large number of individuals, as the claims have at their core significant common issues which can be readily dealt with on a class action basis, it is my view that the complexity of the proceeding favours rather than detracts from a class proceeding.

[24] In other words, the motion judge contemplated the possibility of individual hearings only for the assessment of damages, but not for proof of loss as a component of liability. On that basis, he concluded that the class action is the preferable procedure.

[25] The difference in the fundamental premises of the two sides is reflected in the way the appellants formulate in their factum the two issues for decision on this appeal. The issues as stated assume, rather than raise for decision, that proof of loss is a common issue which will not require the participation of any [page32] class member to prove. The issues are set out in the appellants' factum as follows:

(1) Should the case be certified where liability and damages can be proven at a trial of the common issues without the participation of any class member?

(2) Should the case be certified where the damages suffered by each class member can be assessed and distributed to them without the participation of the defendants?

It is clear from this formulation that the appellants assume that the issue of the loss component of liability can be proved on a class-wide basis. The difficulty is that the question of what method of proof could be used to establish loss on a class-wide basis has not been addressed, and it is the major subject of dispute on the certification motion.

[26] Although the appellants recognize and acknowledge that "[l]oss is a critical component of the causes of action pleaded", they rely on the finding by the motion judge that the loss would be proved at the trial of the common issues through proof of two components: (1) an overall assessment of damages on the basis of the net gain realized by the defendants as a result of their allegedly unlawful agreement; and (2) a measurable price impact upon the ultimate consumer of the building products containing the iron oxide pigments. The appellants argue further that proving these two components of loss would not require the participation of the plaintiff class. Finally, and critically, they state that there was "probative evidence" before the motion judge to support as reasonable the judge's conclusion on proof of loss through the two proposed components.

[27] The motion judge's finding in that regard is at para. 11 of his reasons:

Third, the parties filed expert evidence from economists as to the effect of a price increase at the manufacturing stage on the ultimate consumer of the product. The defendants' expert deposed that it is not possible to trace the impact of such prices through to the consumer. The expert retained by the plaintiffs disagreed with the defendants' expert and deposed that there would be a measurable price impact upon the ultimate consumer of the building products containing the iron oxide pigment. The plaintiffs' expert also deposes that it would be possible to determine an over-all assessment of damages on the basis of the net gain realized by the defendants as a result of their allegedly unlawful agreement. While I am not to be taken as holding that it was necessary to adduce such evidence at this stage of the proceedings, the conflict on the evidence only highlights the point that the issue will have to be resolved at trial, rather than on the pleadings.

(Emphasis added) [page33]

[28] Although the motion judge expressed reservations about the need for the appellants' expert evidence at this stage of the proceedings, it is only on the basis of that evidence that any determination can be made as to whether loss can be proved on a class-wide or an individual basis, and therefore whether it can be a common issue. In Hollick v. Metropolitan Toronto (Municipality), [2001] 3 S.C.R. 158, 2001 SCC 68 at para. 22, McLachlin C.J.C., writing for the court, clarified the role of evidence at the certification stage:

The 1990 Report of the Attorney General's Advisory Committee . . . suggests that "[u]pon a motion for certification . . . , the representative plaintiff shall and the defendant may serve and file one or more affidavits setting forth the material facts upon which each intends to rely" [emphasis added in Hollick]: see Report of the Attorney General's Advisory Committee on Class Action Reform, supra, at p. 33. In my view the Advisory Committee's report appropriately requires the class representative to come forward with sufficient evidence to support certification, and appropriately allows the opposing party an opportunity to respond with evidence of its own.

[29] The Supreme Court also noted that this evidentiary scheme represented the existing practice in Ontario, referring to the case of Caputo v. Imperial Tobacco Ltd. (1997), 1997 12162 (ON SC), 34 O.R. (3d) 314, 148 D.L.R. (4th) 566 (Gen. Div.) at p. 319 O.R., where the court held that the adequacy of the record on a certification motion was of "primary concern". The Supreme Court also quoted with approval from the case of Taub v. Manufacturers Life Insurance Co. (1998), 1998 14853 (ON SC), 40 O.R. (3d) 379 (Gen. Div.). The relevant passage in Hollick reads as follows [at para. 24]:

The court wrote (at pp. 380-1) that "the CPA requires the representative plaintiff to provide a certain minimum evidentia[ry] basis for a certification order". While the Class Proceedings Act, 1992 does not require a preliminary merits showing, "the judge must be satisfied of certain basi[c] facts required by s. 5 of the CPA as the basis for a certification order".

(Emphasis added in Hollick)

McLachlin C.J.C. concluded [at para. 25]:

In my view, the class representative must show some basis in fact for each of the certification requirements set out in s. 5 of the Act, other than the requirement that the pleadings disclose a cause of action.

[30] In my view, with respect, the motion judge erred by relying on the expert evidence filed by the appellants as the basis for the certification order. That evidence does not address the issue of what method could be used at a trial to prove that all end-purchasers of buildings constructed using some bricks or paving stones that contain the respondents' iron oxide pigment overpaid for the buildings as a result. Rather, the appellants' expert effectively assumes that higher costs of products containing the [page34] respondents' iron oxide pigment would have been passed on to end-users, reasoning that they would have been willing to pay the higher cost because the amounts in question were so minimal. He made it clear that he did not know how willing end-purchasers would be to pay higher costs and that he had not had sufficient time to do any analysis to determine the response of the marketplace. He then went on to postulate a conceptual model for calculating the damages "to the extent that buyers of homes or other buildings made of construction materials using iron oxide pigment incur the damages of the conspiracy". The expert's models are based on the assumption of a full pass-through of the price increase of the iron oxide to the homebuyers. However, it is that assumption that is the very issue that the court must be satisfied is provable by some method on a class-wide basis before the common issue can be certified as such.

[31] The motion judge relied on the opinion of the appellants' expert that "there would be a measurable price impact upon the ultimate consumer of the building products containing the iron oxide pigment". However, the fact that any price impact may be "measurable" goes only to the issue of how the damages can be calculated and distributed, not whether the inflated price charged to the direct buyers of the product was passed through to all of the ultimate consumers. The issue of whether there would be a price impact on all ultimate consumers of iron oxide coloured products, i.e., a pass-through to the class members of the inflated price charged by the respondents to their direct buyers, was what the expert assumed, but he did not indicate a method for proving, or even testing that assumption.

[32] The critical importance of the issue of whether, and if so, by what method, loss is provable on a common basis in class action anti-trust suits was canvassed in detail in the recent decision of the U.S. Federal Court of Appeals Third Circuit in In Re: Linerboard Antitrust Litigation, 305 F.3d 145 (2002). The case involved an alleged conspiracy by linerboard (i.e., corrugated cardboard) manufacturers to reduce the industry inventory of linerboard and then, once supply was limited, to implement price increases. The linerboard manufacturers challenged a lower court ruling that had certified two classes of plaintiffs on the basis that they would be able to prove "common impact". [See Note 1 at end of document] The two classes were both direct purchasers, one of corrugated sheets [page35] and one of corrugated boxes. Although the plaintiffs purchased the product directly from the defendants, the issue of whether the conspiracy affected the price of these products on a class-wide basis was a live one because corrugated cardboard was only a component of the finished product.

[33] The manufacturers' challenge to the lower court decision was two-pronged. The main argument was that the lower court should not have relied on a legal presumption of impact, that it failed to apply rigorous scrutiny of the plaintiffs' impact evidence and that the existence of injury required an individualized inquiry. Second, the manufacturers argued that a question of fraudulent concealment also raised individualized issues. The parallel to the case at bar lies in the court's approach to the first prong, and its analysis of the type and strength of evidence required at the certification stage to satisfy the court that there is a method in a price-fixing case by which impact on the plaintiff class can be proved as a common issue.

[34] The lower court in Linerboard based its decision that loss could be proved on a common basis for all members of the class on two types of evidence and analysis. The first was expert evidence that the price-fixing conspiracy can be presumed to have a common impact on all purchasers. The relevant portion of the lower court judgment on the presumption of common impact is approved at p. 152 F.3d of the appeal reasons and reads:

Plaintiffs have shown that they plan to prove common impact by introducing generalized evidence which will not vary among individual class members. For example, plaintiffs contend that even though prices may have varied among regions, the alleged conspiracy caused these prices to rise throughout the country. Although the prices for corrugated sheets and boxes may have increased due to demand, because defendants allegedly conspired to reduce production of linerboard, the price was higher than it would have been under competitive conditions. Such allegations, supported by the evidence presented, are of the kind contemplated by the Third Circuit in [Bogosian v. Gulf Oil Corp., 561 F.2d 434 (3d Cir. 1977)] and [Newton v. Merill Lynch, Pierce, Fennerd Smith, Inc., 259 F.3d 154 (3d Cir. 2001)]. See also [Lumco Indus., Inc. v. Jeld-Wen, Inc., 171 F.R.D. 168, 173 (E.D. Pa. 1997)].

The Court recognizes that defendants dispute plaintiffs' allegations. However, at the class certification stage, "the Court need not concern itself with whether Plaintiffs can prove their allegations regarding common impact; the Court need only assure itself that Plaintiffs' attempt to prove their allegations will predominantly involve common issues of fact and law." Lumco Indus., 171 F.R.D. at 174. "Plaintiffs need only make a threshold showing that the element of impact will predominantly involve generalized issues of proof, rather than questions which are particular to each member of the plaintiff class." Id. (citing In re Disposable Contact Lens Antitrust Litig., 170 F.R.D. 524 (M.D. Fla. 1996)). Therefore, the Court concludes that plaintiffs' [page36 ]allegations regarding impact, like their allegations regarding conspiracy, will focus the inquiry on defendants' actions, not on individual questions relating to particular plaintiff class members.

(Emphasis added)

[35] According to the court of appeals, the lower court's ruling represented a sound application of the concept of "presumed impact" approved in an earlier decision of the same appeals court in Bogosian v. Gulf Oil Corp., 561 F.2d 434 (3d Cir. 1977). The foundational essence of this approach to determining impact is that the focus of the evidence will be on the actions of the defendants and not on individual questions relating to particular plaintiffs. In applying the concept of presumed impact, the court takes notice of the laws of economics as support for a theory that an individual plaintiff can prove the fact of damage simply by proving that the free market prices would be lower than the prices actually paid by the plaintiff. In the Linerboard case, a deliberate cut in supply was alleged. A reduction in supply will cause prices to rise. The concomitant rise in linerboard prices in the relevant market, on the presumed impact theory, represents the laws of supply and demand at work.

[36] In addition to relying on the presumed impact theory, the lower court in Linerboard relied on the extensive empirical investigations that had been undertaken by the plaintiffs' experts. The plaintiffs' experts' testimony was that advanced economic models could be prepared to establish class-wide impact. In the evidence before the lower court, the experts supported their opinions with charts, studies, company records, industry data and articles from leading trade publications. The key issue on which the experts focused was whether the variations in purchasers, products, regions, etc., precluded common impact. Taking the variations into account, they concluded that all purchasers would have paid a higher price because of the conspiracy. As a result, the fact of loss was common. Only the quantum of loss would vary. One expert stated categorically: "[b]ased on my analysis of the pricing data and company records, I conclude that the alleged unlawful conduct to raise linerboard prices would have impacted all members of the proposed class through higher corrugated sheet prices" (p. 154 F.3d). The court of appeals noted that this conclusion was supported by relevant data.

[37] The court of appeals in Linerboard approved of what it referred to as the lower court's use of a "belt and suspenders rationale" (p. 153 F.3d) by relying on both the presumed impact theory together with the expert evidence, to form the evidentiary [page37 ]basis for its conclusion that loss as a component of the cause of action could be proved on a class wide basis.

[38] The defendants in Linerboard also argued that the Linerboard case was exactly comparable to the case of Newton v. Merill Lynch, Pierce, Fennerd Smith, Inc., 259 F.3d 154 (3d Cir. 2001), where the same court of appeals overturned the lower court's decision to certify. In Newton, the complaint was that brokers in stock transactions had not obtained the best price for their clients. The evidence on the certification motion showed that the computer system used by the defendant brokers obtained variable prices, some of which were the best, and some of which were not. Consequently, some stock purchasers suffered loss while others did not. The appeals court rejected the comparison with the Newton case, essentially because in Newton, not all members of the putative class had suffered loss. In Newton, the court of appeals noted that "[w]hether a class member suffered economic loss from a given securities transaction would require proof of the circumstances surrounding each trade, the available alternative process, and the state of mind of each investor at the time the trade was requested" (p. 187 F.3d). Based on the evidence before the court on the Linerboard motion, no similar individual inquiries were required to prove loss to every member of the class of linerboard purchasers.

[39] A useful comparison can be made between the evidentiary record in the Linerboard certification motion and the record in this case. In Linerboard, the court had both evidence based on economic theory as well as industry evidence that formed the basis for expert opinions of what actually occurred in the market when corrugated cardboard was sold as sheets and boxes. The evidence demonstrated to the certification court that it could be proved at trial that the plaintiffs did in fact suffer loss on a class-wide basis. The certification of the plaintiff class in Linerboard did not, of course, mean that there would not be contrary evidence and substantial opposition to the plaintiffs' position at trial. It did mean that, on a preliminary basis, there was a sufficient record to support a decision to certify based on liability as a common issue.

[40] In this case, the appellants presented no evidence from industry representatives to explain how the manufacturers and distributors of bricks and the developers of new homes price their products, and in particular, whether there is a direct pass-through of the price of every component into the sale price of all homes, the relevance of the value of the land component, and how other factors such as the real estate market and [page38 ]the individual bargaining of the purchaser and vendor affect the price. The evidence on the issue of loss to the members of the plaintiff class came only from the affidavit of an expert economist who did not address those issues. In his affidavit, the expert does not suggest that he consulted any industry records or other data which would substantiate a pass- through analysis.

[41] Finally, in the Linerboard case, the defendants asked the court to apply the decision of the U.S. Supreme Court in Illinois Brick, supra, which the respondents in this case also seek to rely on. Illinois Brick was an anti-trust treble damages action, brought under s. 4 of the Clayton Act, 15 U.S.C. 15, by end-purchasers of structures built using concrete blocks. They alleged that the manufacturers of the concrete blocks had engaged in a conspiracy to fix prices in violation of federal antitrust legislation (Sherman Act, 15 U.S.C. 1). The issue was whether indirect purchasers, as opposed to the intermediaries who had purchased directly from the manufacturers, could sue based on the alleged overcharge. In its earlier decision in Hanover Shoe Inc. v. United Shoe Machinery Corp., 392 U.S. 481 (1968), the U.S. Supreme Court had held that where the plaintiffs were direct purchasers from the defendants, the defendants could not use the defence that the alleged overcharge had been passed through to the ultimate consumer and therefore that the direct purchaser had suffered no damage. In Illinois Brick, the same bar was applied to indirect purchasers as plaintiffs: they could not sue for treble damages for price fixing because allowing claims by both direct and indirect purchasers would create the risk of double recovery and make the process of determining who had suffered what proportion of the price overcharge too complex, thereby undermining the effectiveness of the remedy.

[42] In the Linerboard litigation, both classes of plaintiffs were direct purchasers from the defendant manufacturers. Because both the sheets and the boxes contained linerboard only as a component, so that the price-fixed product formed only one ingredient of the product purchased, the defendants argued that the class members were akin to indirect purchasers, and that the Illinois Brick prohibition should apply. The appeals court rejected the analogy and held that the class members were direct purchasers and "entitled to recover the full amount of any overcharge" (pp. 159-60 F.3d).

[43] In his reasons in the case at bar, the motion judge declined to follow Illinois Brick, which, as discussed above, bars actions by indirect purchasers for price fixing as a matter of law [page39] in the U.S. [See Note 2 at end of document] However, in so doing he did not address the underlying reasoning of the U.S. Supreme Court that led to the result in Illinois Brick and in Hanover Shoe, namely, the complexities of proving the extent to which the different players in the chain of purchase bear the higher price caused by the illegal conspiracy to fix the price of the base product.

[44] The complexity of the "pass-through" problem was recognized by the Divisional Court. The court referred with approval to the following passage from pp. 742-43 of Illinois Brick:

. . . "in the real economic world rather than an economist's hypothetical model," the latter's drastic simplifications generally must be abandoned. Overcharged direct purchasers often sell in imperfectly competitive markets. They often compete with other sellers that have not been subject to the overcharge; and their pricing policies often cannot be explained solely by the convenient assumption of profit maximization. As we concluded in Hanover Shoe, 392 U.S., at 492, attention to "sound laws of economics" can only heighten the awareness of the difficulties and uncertainties involved in determining how the relevant market variables would have behaved had there been no overcharge.

[45] The Divisional Court noted the many problems of proof facing the appellants with respect to the pass-on issue, including the number of parties in the chain of distribution and the "multitude of variables" which would affect the end- purchase price of a building. The appellants would have to show that the price increase (or a part of it) was passed through from the respondents to the building materials manufacturer and distributor, to the builder, to the purchaser and on to any subsequent purchaser. If the price increase was absorbed at any point, the chain would be broken. The problem is compounded by the fact that the iron oxide pigment forms such a minimal part of the whole structure and therefore a similarly minimal portion of the purchase price of a building.

[46] As noted above, neither the variables nor the issue of how to prove the flow of the price increase through the distribution [page40] chain were addressed by the appellants' expert in his evidence. Nor does he discuss the effect of the market on real estate prices and the relative effects on the purchase price of (a) the market, (b) the value of the land, (c) the value of the building, and (d) how one assesses the value of the component parts of the building at any particular point in time, remembering that the proposed class members are not only purchasers of new homes, but of resale homes as well, and that not all homes were constructed using the impugned materials.

[47] In my view, this latter point -- that not all buildings built and sold during the relevant period contained the respondents' materials -- highlights a significant aspect of the problem with the appellants' theory and their position. The appellants' expert alludes to the issue of whether the increase in price of iron oxide pigmented materials may have had the effect of causing an increase in all substitute materials as well. In other words, the effect of the increase in price of coloured bricks might have been to cause the price of all bricks to rise, or looking at it another way, the homebuilders may have increased the prices of all houses built with bricks regardless of what bricks they used, in order to accommodate and at the same time to take advantage of the price increase of the iron oxide pigmented bricks. The expert suggests that if that were true, then "all homebuyers or other end-users would have been damaged by the iron oxide conspiracy to a greater or lesser extent", not just buyers of homes containing the defendants' product. However, he goes on to say that he did no analysis to determine whether this in fact occurred, and opined that there was no practical relevance to the issue.

[48] To the contrary, from the point of view of proof of loss to homebuyers as a class based on a pass-through of the price increase, if it could be shown that all home prices were artificially inflated as a result of the use of both iron oxide pigmented and non-iron oxide pigmented building materials, that could well have formed the basis for concluding that proof of loss could be presented on a class-wide basis as a common issue.

[49] The Divisional Court also rejected several other methods referred to by the motion judge for arriving at class-wide proof of loss. First, the Divisional Court held that s. 24 of the Class Proceedings Act, which deals with an aggregate assessment of monetary relief, cannot resolve the problems of proving loss on a class-wide basis. I agree that s. 24 of the Class Proceedings Act is applicable only once liability has been established, and provides a method to assess the quantum of damages on a global basis, but not the fact of damage. [page41]

[50] Second, the Divisional Court rejected the suggestion that the variables in circumstances of different plaintiffs could be addressed by the creation of plaintiff sub-classes. Again, the creation of subclasses is an evidentiary matter. I note that the appellants have not relied on this solution before this court, nor is there any evidentiary foundation to suggest that different sub-classes of plaintiffs can be formed which will effectively create a basis for commonality on the issue of proof of loss.

[51] Finally, the Divisional Court concluded that s. 23 of the Class Proceedings Act, which contemplates the use of statistical evidence to determine the amount or distribution of a monetary award, would not allow the issue of liability to be proved through otherwise inadmissible statistical evidence. I do not adopt this comment by the Divisional Court. In the American cases, as discussed above, expert evidence that includes an analysis of statistical data has been used to establish loss on a class-wide basis. The admissibility of any such evidence will have to be considered when the issue arises.

Conclusion on common issue

[52] In my view, the motion judge erred in finding that liability could be proved as a common issue in this case. The evidence presented by the appellants on the motion does not satisfy the requirement prescribed by the Supreme Court in Hollick of providing sufficient evidence to support certification. The evidence of the appellants' expert assumes the pass-through of the illegal price increase, but does not suggest a methodology for proving it or for dealing with the variables that affect the end price of real property at any particular point in time. The motion judge focused on the expert's opinion that the loss could be measured, rather than on how any such loss could first be established on a class-wide basis.

(ii) Preferable procedure

[53] In reaching his conclusion under s. 5(1)(d) of the Class Proceedings Act that "a class proceeding would be the preferable procedure for the resolution of the common issues", the motion judge specifically limited his focus to whether the class action procedure was preferable for resolution of the common issues only, and not of any other individual issues that would have to be resolved in the course of the litigation. In so doing, the motion judge did not have the benefit of the Supreme Court of Canada's decision in Hollick, supra, where the court specifically dealt with the breadth of the inquiry under s. 5(1)(d) of the Act. After noting that the section [page42] only speaks about the preferable procedure for resolving the common issues, McLachlin C.J.C. stated [at para. 29]:

I would not place undue weight, however, on the fact that the Act uses the phrase 'resolution of the common issues' rather than 'resolution of class members' claims.

[54] The Chief Justice concluded that "[t]he question of preferability, then, must take into account the importance of the common issues in relation to the claims as a whole", and quoted with approval [at para. 30] the statement of the Chairman of the Attorney General's Advisory Committee that the class representative must

"demonstrate that, given all the circumstances of the particular claim, [a class action] would be preferable to other methods of resolving these claims and, in particular, that it would be preferable to the use of individual proceedings".

(Emphasis added in Hollick)

[55] Regardless of whether the motion judge was correct in concluding that a class action would be the preferable procedure when there were three major common issues, as I have concluded based on the record, that proof of loss as a component of liability cannot be a common issue, the only common issues are the price fixing conspiracy, and possibly, the measure of damages, if the scope of liability can be determined.

[56] I do not believe the motion judge would have certified the action as a class action had he not viewed liability as a common issue. The number of potential plaintiffs in this case is very large, estimated at 1.1 million. Clearly, if individual trials are needed to establish loss and therefore liability, the action will be unmanageable.

[57] Together with the issue of proof of loss is the question whether a home purchased by any plaintiff actually contains bricks or paving stones coloured with the respondents' iron oxide pigment. As part of the proof of loss, a massive record- tracing exercise will be required to establish the inclusion of the respondents' product in any particular structure. The respondents point out that the period over which records must be obtained spans 17 years. The respondents also point to the many intermediary parties from whom those records, if they exist, must be sought.

[58] In his discussion regarding the class definition and the potential difficulties of identifying class members because of this extensive tracing exercise required, the motion judge discounted apparent problems of self-identification of potential plaintiffs who might have large claims and would therefore want to opt out [page43] of the class procedure. He did so because all potential plaintiffs' claims appeared to be very minimal, so that the goal of certification would not be compensation of individual plaintiffs but rather behaviour modification. Consequently, the motion judge concluded that possible under-inclusion of potential plaintiffs was not a serious matter.

[59] The motion judge did not consider how the same problem of identification of class members can also arise in the context of possible over-inclusion of parties in the class. The potential problem was masked by making proof of loss a common issue. However, with liability as an individual rather than a common issue, identification and proof of those actually affected is required, with all of the difficulties referred to above.

[60] The appellants also base their argument on ss. 24 and 26 of the Class Proceedings Act. The expert evidence filed by the appellants opined that the illegal profits gained by the respondents could be calculated on an aggregate basis. The appellants argue that the loss to the class is equal to the gain of the respondents based on their illegal conspiracy. The appellants suggest that by finding that individual trials are needed to prove a loss in each case, the Divisional Court confused the process of assessing damages on an aggregate basis with the process of distributing damages to the class members. Section 24(1)(b) of the Class Proceedings Act provides:

24(1) The court may determine the aggregate or a part of a defendant's liability to class members and give judgment accordingly where,

(b) no questions of fact or law other than those relating to the assessment of monetary relief remain to be determined in order to establish the amount of the defendant's monetary liability; . . .

[61] It is clear to me that the Divisional Court made no such error. Section 24(1)(b) is a mechanism for assessing damages where there is no issue of liability. Section 26 also deals with the distribution of the damage award. The Divisional Court focused its analysis on the mechanism of proving the loss necessary to base liability. By seeking to equate the respondents' gain with the class members' alleged loss, the appellants effectively skip over the process of determining who in the chain, beginning with the direct purchasers from the respondents, absorbed the loss. In the U.S., this problem has been resolved by limiting the civil remedy of treble damages for price fixing to direct purchasers only and attributing the entire loss to them. The effect of the appellants' approach is to attribute the entire loss to the indirect or end-purchasers rather than to determine whether those [page44] parties suffered loss as required by s. 36(1) of the Competition Act and as part of the common law causes of action.

[62] The appellants complain that if the action is not certified, this will effectively end the litigation, and thereby defeat the goal of behaviour modification which would be accomplished through the action and which is one of the three goals of the Class Proceedings Act. However, where access to justice through compensation of individual plaintiffs who have suffered a loss is not a significant goal (because the amounts in issue are so minimal, and most plaintiffs do not know if or that they suffered any damage), and where judicial economy would be undermined, not enhanced, by certifying the action, the circumstances requiring behaviour modification would have to be extremely compelling to allow that single goal to overcome the other deficiencies. In this case, the Competition Act provides criminal sanctions to achieve that goal.

[63] The appellants argue further that if certification is not allowed in this case, the effect would be a complete bar on all class actions by consumers in price-fixing cases. They argue that the civil action under the Competition Act, used together with the Class Proceedings Act, is an important tool for preventing and effectively punishing price-fixing activity. It must be able to be used along with the enforcement mechanisms in the Competition Act in order to effectively regulate and discourage anti-competitive behaviour in the marketplace. The appellants point to the limited resources of the Competition Bureau for carrying out comprehensive enforcement and therefore the necessity that class actions exist as an effective threat to potential anti-competitive market behaviour.

[64] Although the civil remedy for anti-competitive behaviour can be an important component of the enforcement of the goals of the Competition Act, it is only one component. In appropriate cases, the Competition Act and Class Proceedings Act can work together as a valuable tool against price-fixing. However, in this case, the obstacles to an effective class proceeding override its potential benefits.

[65] In my view, the question of whether and how consumers will be able to use class actions to obtain relief from price fixing by suppliers and manufacturers remains an open one in this jurisdiction. The appellants were unsuccessful in this case because they did not present the evidentiary basis for a certifying court to be satisfied that loss as a component of liability could be proved on a class-wide basis. Whether such evidence could have been obtained is not clear. [page45]

[66] The Divisional Court's approach suggested that it could not: that the variables in house purchase prices were such that the type of evidence that would have been required to show "pass-through" on a class-wide basis would not have been available in this case, in large part because of the nature of real estate and the individualized pricing factors on each sale. The Divisional Court's concerns follow the U.S. approach as defined in the Illinois Brick and Hanover Shoe cases.

[67] The difficulties with proving pass-through of price increases on a class-wide basis are illuminated in an article that discusses Illinois Brick. (William H. Page, "The Limits of State Indirect Purchaser Suits: Class Certification in the Shadow of Illinois Brick" (1999) 67 Antitrust L.J. 1). The author essentially concludes that indirect purchaser litigation for price-fixed goods is not a viable method of achieving behaviour modification against anti-competitive behaviour. At pp. 36-37 the author states:

Thus, only a highly artificial subset of indirect purchasers of price-fixed goods will ever be compensated by class actions. Moreover the denial of certification is largely unrelated to the merits of the underlying claim. Most of the factors that preclude certification of classes of indirect purchasers have little to do with whether a price fixing conspiracy actually existed or whether indirect purchasers bore an overcharge. The number of levels of intermediate purchasers between the price fixers and plaintiff class is unrelated to the success of the conspiracy. Similarly, whether plaintiff intermediate purchasers alter or add value to the product, or use it as an ingredient in another product, has nothing to do with whether price fixing has occurred upstream, or even whether the overcharge was passed on. Yet these factors may preclude certification because they make it impossible to establish harm to each class member by any kind of common proof.

Thus, in many cases, a price-fixing overcharge will simply dissolve into the currents of the channels of distribution. Eighty years ago, Justice Holmes noted the "endlessness and futility of the effort to follow every transaction to its ultimate result," even though "in the end the public pays the damages in most cases of compensated torts." Now, as then, it may well be that an overcharge is passed on but the legal system cannot identify its incidence. Common proof is impossible and individualized proof would be more costly than the amount of the harm. The emerging reality of the indirect purchaser class action offers no realistic mechanism for accomplishing compensation for remote purchasers of price- fixed goods. If the indirect purchaser class action is only available to a small subset of indirect purchaser injuries, even among price-fixing conspiracies that are actually detected, it is not fulfilling its stated purpose.

(Internal citations omitted)

[68] In this jurisdiction it remains to be determined whether in a particular case a sufficient evidentiary record can be brought before a certifying court to satisfy that liability can be proved as a common issue. Whether it can be done is a question left open for future cases. [page46]

(iii) Definition of the class

[69] Leave to appeal to the Divisional Court was granted by Lane J. on the basis that the motion judge erred in his definition of the class. As part of its decision, the majority of the Divisional Court held that the class definition was in error because the definition is not objective, but turns on the outcome of the litigation or the merits of the claim. I agree with that conclusion. As Sharpe J. stated in another case, (Robertson v. Thomson Corp. (1999), 1999 14768 (ON SC), 43 O.R. (3d) 161, 171 D.L.R. (4th) 171 (Gen. Div.) at p. 169 O.R.):

I agree with Winkler J. in [Bywater v. Toronto Transit Commission, [1998] O.J. No. 4913 (Gen. Div.)] and with [H. Newberg and A. Conte, Newberg on Class Actions, 3rd ed. (West Group, 1992)] at p. 6-61, that the class should be defined in objective terms, and that circular definitions referencing the merits of the claim or subjective characteristics ought to be avoided. Such definitions make it difficult to identify who is a member of the class until the merits have been determined. Definitions based upon the merits of the claim also violate the statutory policy that the merits are not to be decided at the certification stage.

[70] As discussed earlier, when the motion judge was considering the class definition, he rejected the problems of self-identification of potential class members because he did not consider under-inclusion as a problem. However, by defining the class as those who suffered damage, he in effect defined away any potential over-inclusion that could occur if proof of loss is a common issue. In my view, the two errors were linked in this case.

V. Conclusion

[71] In light of my conclusion that the action cannot be certified as a class action because a class action is not the preferable procedure given the limited common issues, it is not necessary to address the propriety of the motion judge's reliance on the appellant's affidavit.

[72] I would dismiss the appeal.

[73] The respondents have provided their bill of costs. The appellants shall have ten days from release of these reasons to provide the Senior Legal Officer with their submissions as to costs. The respondents may respond within seven days thereafter.

Order accordingly.

[page47]

Notes

Note 1: The lower court decision is reporte at In Re: Linerboard Antitrust Litigation, 203 F.R.D. 197 (E.D. Pa. 2001).

Note 2: The same approach has been followed, for example by Cumming J. in VitaPharm Canada Ltd. v. F. Hoffmann-Laroche Ltd. (2000), 4 C.P.C. (5th) 169 (Ont. S.C.J.) where he named lead counsel for the plaintiff classes in a class action involving both direct and indirect purchasser classes of plaintiffs affected by a worldwide conspiracy to increase the price of vitamins to both wholesale and retail users. The ability of the different classes to prove loss did not appear to be a disputed issue in the motion. The issue of concern was the allocation of the global loss among the various groups. The class action procedure did not fail based on the potential difficulty of allocating damages. In the Divisional Court decision in the case at bar, Somers J. specifically left open the possibility that there could be indirect purchaser anti-trust claims advanced by way of class proceeding, and at para. 44, referred to the VitaPharm case.