delivered the opinion of the court:
The plaintiff in this appeal alleges that she was injured by the drug diethylstilbestrol (DES), which her mother ingested during pregnancy. She seeks relief against defendant DES manufacturers. The issue is whether, in a negligence and strict liability cause of action, Illinois should substitute for the element of causation in fact a theory of market share liability when identification of the manufacturer of the drug that injured the plaintiff is not possible. The trial court granted defendants’ motion for summary judgment as to various counts of the complaint, including a negligence count, but denied the defendants’ motion as to the strict products liability count and adopted the market share liability theory developed in Sindell v. Abbott Laboratories (1980), 26 Cal. 3d 588, 607 P.2d 924, 163 Cal. Rptr. 132. The appellate court affirmed the trial court’s holding as to the products liability count, but reversed the trial court’s holding as to the negligence count, holding that the market share theory should apply to both the plaintiff’s negligence and strict liability counts. (173 Ill. App. 3d 1.) However, it rejected the Sindell rule which the trial court had adopted and instead adopted the market share liability theory as it was recognized in Martin v. Abbott Laboratories (1984), 102 Wash. 2d 581, 689 P.2d 368. The appellate court affirmed the trial court’s entry of summary judgment for defendants on the other counts. We granted defendants’ petition for leave to appeal (107 Ill. 2d R. 315).
I. FACTS
The plaintiff, Sandra Smith, was born on July 13, 1953, in Chicago, Illinois. In 1978, she was admitted to the Ravenswood Hospital in Chicago, where she underwent a dilation and curettage, cervical biopsy, and an ex-cisional biopsy of the vaginal wall. The biopsy revealed that plaintiff had a form of cancer known as clear cell adenocarcinoma of the vagina. She was then transferred to the University of Illinois Hospital, where she underwent extensive surgery. Plaintiff alleges that the DES prescribed for her mother while plaintiff was in útero caused the cancer.
Elizabeth Smith, plaintiff’s mother, had a history of difficulty with pregnancy before she gave birth to Sandra. Therefore, in early 1953, when she learned of her pregnancy, she went to the Field Clinic in Chicago and consulted with Dr. Jack E. Davis regarding her condition. The doctor gave Mrs. Smith a prescription to be filled at the clinic pharmacy for “Tab 98.” The practice at the Field Clinic was to store and dispense drugs by number, rather than by name. The record establishes that Tab 98 designated 25 milligram tablets of DES. Mrs. Smith continued to take DES tablets daily up until the time she gave birth to Sandra. Dr. Davis attended Mrs. Smith throughout her pregnancy but at the time this suit was filed he was deceased.
The records recovered from the Field Clinic identify numerous companies which supplied drugs to the clinic, some of which were also suppliers of DES, but these are insufficient to match the company with the drug dispensed to the plaintiff’s mother. The person responsible for purchasing the pharmaceutical products stocked at the clinic’s pharmacy had also died by the time this suit was filed. Therefore, although the plaintiff knows the color, size and dosage of the drug her mother took, she is unable to identify the specific manufacturer of the product.
In 1980, plaintiff filed her initial complaint naming as defendants 138 drug companies. According to an affidavit of John Kraas, an employee of defendant Eli Lilly & Company, there were 81 companies which marketed DES in 25 milligram tablets between 1952 and 1953. This information was derived from medical and pharmaceutical industry references. Of the 81 potential manufacturers of the DES taken by plaintiff’s mother, 63 were not named in the complaint. Of the 138 companies named, 70 filed appearances. Motions were filed by a number of named defendants attacking jurisdiction, asserting changes in corporate structure or ownership that would bar successor liability, or charging error of identity. Twenty companies remained in the suit after these motions were resolved.
In November 1982, plaintiff filed a second amended complaint consisting of 11 counts. Counts I through VI sound in negligence, strict liability, breach of express warranty, fraud, breach of implied warranty, and violation of the Federal Food, Drug, and Cosmetic Act. Counts VII and VIII sound in conspiracy and pray for assessment of damages on various bases of concerted action, joint and several liability and joint enterprise liability. Counts IX and X allege theories of negligence and strict liability, respectively, and invoke market share as the means for apportioning damages. The thrust of these causes of action is that the drug companies failed to properly test DES and to adequately warn of its dangers. The last count is a tort action against the Field Clinic. There are apparently no motions pending on this count.
Following completion of discovery, 14 defendants filed a joint motion for summary judgment. Some of these defendants and a number of others filed individual motions for summary judgment on the ground that the plaintiff’s mother did not use their products. Twelve defendants were able to exculpate themselves on the basis that they could not have manufactured the DES that plaintiff’s mother took because their product either was not of the same dosage, color or type, or was not sold to the Field Clinic. The remaining eight defendant manufacturers included Abbott Laboratories, Eli Lilly & Company, Premo Pharmaceutical Laboratories, Inc., Carroll Dunham Smith Pharmacal Company, William H. Rorer, Inc., S.E. Massengill Company, Harvey Laboratories, Inc., and Boyle & Company.
The trial court granted the joint motion for summary judgment on counts I through IX of the second amended complaint. However, the court denied the motion with respect to count X, the strict liability action, and adopted market share liability, based on the theory that the California Supreme Court articulated in Sindell v. Abbott Laboratories (1980), 26 Cal. 3d 588, 607 P.2d 924, 163 Cal. Rptr. 132. The defendants appealed the denial of their motion for summary judgment as to count X and plaintiff cross-appealed the trial court’s grant of summary judgment as to counts I through IX. As noted above, the appellate court likewise adopted a theory of market share liability, although different from that applied by the trial court, and extended its application to the negligence count. The appellate court affirmed the dismissals of the other counts. The defendants filed a petition for leave to appeal to this court based on the denial of their motion for summary judgment on counts IX and X, and we granted leave to appeal (107 Ill. 2d R. 315). The plaintiff has not cross-appealed from the dismissal of the other counts.
II. HISTORY OF DES
The history of the development of DES and its marketing in this country has been repeatedly chronicled, especially in cases which address the issues of conspiracy and concert of action. (See Ryan v. Eli Lilly & Co. (D.S.C. 1981), 514 F. Supp. 1004; Martin, 102 Wash. 2d 581, 689 P.2d 368; Sindell, 26 Cal. 3d 588, 607 P.2d 924, 163 Cal. Rptr. 132; Lyons v. Premo Pharmaceutical Labs, Inc. (App. Div. 1979), 170 N.J. Super. 183, 406 A.2d 185; Note, The DES Causation Conundrum: A Functional Analysis, 32 N.Y.L. Sch. L. Rev. 939 (1987); Comment, Market Share Liability: A New Method of Recovery for D.E.S. Litigants, 30 Cath. U.L. Rev. 551 (1981).) Nevertheless, before getting into the market share issue, we believe that a brief account of the drug’s history will be helpful.
Diethylstilbestrol is a synthetic substance which duplicates the activity of estrogen, a female sex hormone crucial to sexual development and fertility. Professor E.C. Dodds and his associates first synthesized the drug in England in 1937. The drug was not patented by Professor Dodds, but was left available for general production by pharmaceutical companies. In 1940, a number of pharmaceutical companies in the United States sought the approval of the Food and Drug Administration (FDA) to market DES in up to 5 milligram doses to treat vaginitis, engorgement of the breasts, excessive menstrual bleeding and symptoms of menopause. Standard procedure at the FDA required the filing of a new drug application (NDA), which included clinical data establishing the drug’s safety, its chemical composition, methods of manufacture, the proposed uses of the drug and proposed labeling. In an effort to avoid duplication of time and effort in determining the sufficiency of the documentation presented, the FDA requested that the drug companies withdraw their NDAs and submit their data jointly in a master file. Accordingly, a working committee of four companies was formed which collected all the data, prepared the master file and submitted it to the FDA. In September 1941, the FDA approved the production and marketing of DES for the requested uses, none of which were for problems related to pregnancy.
The first supplemental NDAs seeking FDA approval for DES as a miscarriage preventative were filed in 1947. These NDAs were filed separately and relied on clinical studies published in medical journals which attested to the safety and effectiveness of DES for this purpose. The FDA subsequently approved these applications. In 1952, the FDA declared that DES was no longer a new drug within the meaning of the Federal Food, Drug, and Cosmetic Act, and was therefore considered safe for general use. This declaration meant that any manufacturer could market the drug without submitting additional data to the FDA concerning its safety and effectiveness. Between 1947 and 1952, approximately 85 companies manufactured DES. By the end of 1952 up to 191 companies were manufacturing and distributing DES.
In 1971, two medical studies suggested that there was a statistically significant association between the outbreak in young women of clear cell adenocarcinoma, a form of cancer, with the maternal ingestion of DES during pregnancy. Later that year the FDA banned the marketing of DES for use by pregnant women. It has been estimated that at the time of this ban as many as 300 companies had produced DES for sale. Many of these companies are no longer in existence, having merged with other concerns or gone into liquidation. Although DES is no longer used during pregnancy, it is still prescribed as an estrogen replacement in cases of hormone deficiency, for treatment of unusual menopausal symptoms, and for treatment of certain kinds of cancers of the breast and prostate, and is a major ingredient in the “morning after” pill, a post-coital contraceptive.
Beginning in the 1970s, hundreds of lawsuits were filed against manufacturers of DES by the daughters of women who took the drug while pregnant. These plaintiffs are commonly referred to as the “DES daughters.” The seriousness of the injuries they suffer cannot be questioned and the hysterectomy required for Sandra Smith was not unusual. (See Bichler v. Eli Lilly & Co. (1982), 55 N.Y.2d 571, 577, 436 N.E.2d 182, 184, 450 N.Y.S.2d 776, 778; Namm v. Charles E. Frosst & Co. (App. Div. 1981), 178 N.J. Super. 19, 25-26, 427 A.2d 1121, 1124; Lyons v. Premo Pharmaceutical Labs, Inc. (App. Div. 1979), 170 N.J. Super. 183, 189, 406 A.2d 185, 188; see also Enright v. Eli Lilly & Co. (Supp. 1988), 141 Misc. 2d 194, 195, 533 N.Y.S.2d 224, 226.) The defendants before us, however, point out that statistics regarding DES daughters have not shown a high incidence of cancer and that it is not widely accepted that the injuries suffered are the consequence of the maternal ingestion of DES (see, e.g., Sindell, 26 Cal. 3d at 620-21, 607 P.2d at 942, 163 Cal. Rptr. at 150 (Richardson, J., dissenting) (incidence of cancer is estimated at one-tenth of one percent to four-tenths of one percent)), though the plaintiff contests these assertions. Whether or not there is a correlation sufficient to establish a cause of action is an issue properly for the finder of fact. We have before us only the legal issue of the viability of the causes of action.
III. SUBSTANTIVE TORT PRINCIPLES
A fundamental principle of tort law is that the plaintiff has the burden of proving by a preponderance of the evidence that the defendant caused the complained-of harm or injury; mere conjecture or speculation is insufficient proof. (Schmidt v. Archer Iron Works, Inc. (1970), 44 Ill. 2d 401, 405-06; M. Polelle & B. Ottley, Illinois Tort Law 422-23 (1985); Annot., 51 A.L.R.3d 1344, 1349 (1973) (“it is obvious that to hold a producer, manufacturer, or seller liable for injury caused by a particular product, there must first be proof that the defendant produced, manufactured, sold, or was in some way responsible for the product”).) In a negligence action this causation-in-fact requirement entails a reasonable connection between the act or omission of the defendant and the damages which the plaintiff has suffered. (See Ney v. Yellow Cab Co. (1954), 2 Ill. 2d 74, 79; W. Keeton, Prosser & Keeton on Torts §41, at 262 (5th ed. 1984).) The theory of strict liability is that one who sells a defective product unreasonably dangerous to the user is liable for the resulting injury. (Suvada v. White Motor Co. (1965), 32 Ill. 2d 612, 621, 623; Restatement (Second) of Torts §402A (1965).) Likewise, to recover under strict liability the plaintiff must establish some causal relationship between the defendant and the injury-producing agent. See M. Polelle & B. Ottley, Illinois Tort Law 581 (1985).
In Schmidt v. Archer Iron Works, Inc. (1970), 44 Ill. 2d 401, a defective pin used to attach a metal chute to a construction tower failed, allowing the chute to fall and strike the plaintiff. Plaintiff sued Archer and evidence established that the defective pin was similar in color to the pins manufactured by defendant, but several other manufacturers produced similar pins. We affirmed the trial court’s grant of summary judgment, reasoning that:
“The plaintiffs’ evidence failed to establish sufficient connection between the admittedly defective pin and Archer. *** [The] evidence shows. no more than that Archer was one of several possible manufacturers which could have supplied the pin.” (44 Ill. 2d at 405-06.)
The identification element of causation in fact serves an important function in tort law. Besides assigning blameworthiness to culpable parties, it also limits the scope of potential liability and thereby encourages useful activity that would otherwise be deterred if there were excessive exposure to liability. Fischer, Products Liability — An Analysis of Market Share Liability, 34 Vand. L. Rev. 1623, 1628-29 (1981).
The plaintiff before us alleges that after extensive discovery she has been unable to identify the manufacturer of the DES her mother ingested. A number of circumstances contribute to the barrier in establishing causation in fact in DES cases. The effects caused by prenatal exposure to DES usually do not manifest themselves until at least after the child reaches puberty, and more years may pass before the cancer is linked to DES. During this long lapse, whatever records the doctor, pharmacy or manufacturer maintained have often been lost or destroyed and the memories of the persons involved have faded. Further exacerbating the problem is the fact that during the 25 years that DES was used to treat pregnancy-related problems, as many as 300 companies manufactured the drug. The manufacturers were only required by law to maintain records for five years and many manufacturers have either gone out of business or destroyed their records or have only partial records available.
Although proof of causation in fact is ordinarily an indispensable ingredient of a prima facie case, the plaintiff points out that competing tort interests have compelled courts to create exceptions to the causation requirement. These exceptions to the rule have allowed a plaintiff to shift to a defendant or a group of defendants the burden of proof on the causation issue. Included within the exceptions are “enterprise liability,” “alternative liability” and “market share liability.”
In addition to market share liability, most plaintiffs in the DES cases have argued that enterprise liability or alternative liability, as well as a concert of action and a conspiracy theory, should apply to extend liability to a group of defendants.
The criteria necessary for a cause of action based on enterprise liability have been summarized to include: “(1) The injury-causing product was manufactured by one of a small number of defendants in an industry; (2) the defendants had joint knowledge of the risks inherent in the product and possessed a joint capacity to reduce those risks; and (3) each of them failed to take steps to reduce the risk but, rather, delegated this responsibility to a trade association.” (Emphasis omitted.) (Burnside v. Abbott Laboratories (1985), 351 Pa. Super. 264, 285, 505 A.2d 973, 984; see also Hall v. E.I. Du Pont De Nemours & Co. (E.D.N.Y. 1972), 345 F. Supp. 353.) Alternative liability may apply when two or more defendants act tortiously toward a plaintiff who, through no fault of her own, cannot identify which one of the joined defendants caused the injury. The burden of proof shifts to each defendant to prove his innocence. (Restatement (Second) of Torts §433B(3), at 441-42 (1965); Summers v. Tice (1948), 33 Cal. 2d 80, 199 P.2d 1.) Concert of action applies when a tortious act is done in concert with another or pursuant to a common design, or a party gives substantial assistance to another knowing that the other’s conduct constitutes a breach of duty. (Restatement (Second) of Torts §§876(a), (b), at 315 (1979).) A civil conspiracy involves two or more persons who combine for the purpose of accomplishing by concerted action either (1) a lawful purpose by unlawful means, or (2) an unlawful purpose by lawful means. M. Polelle & B. Ottley, Illinois Tort Law 389 (1985); see also Montgomery Ward & Co. v. United Retail, Wholesale & Department Store Employees of America (1948), 400 Ill. 38, 52.
Though the market share liability theory has received some acceptance, in nearly every instance, the other theories have been soundly rejected. (See, e.g., Ryan v. Eli Lilly & Co. (D.S.C. 1981), 514 F. Supp. 1004 (civil conspiracy, concert of action, alternative liability and enterprise liability rejected); Burnside v. Abbott Laboratories (1985), 351 Pa. Super. 264, 505 A.2d 973 (no conspiracy, concert of action or enterprise liability); Collins v. Eli Lilly Co. (1984), 116 Wis. 2d 166, 342 N.W.2d 37 (no conspiracy, concerted action, enterprise liability or alternative liability); Martin, 102 Wash. 2d 581, 689 P.2d 368 (concerted action, alternative liability and enterprise liability properly dismissed); Zafft v. Eli Lilly & Co. (Mo. 1984), 676 S.W.2d 241 (no alternative liability, concert of action or enterprise liability); Sindell, 26 Cal. 3d 588, 607 P.2d 924, 163 Cal. Rptr. 132 (concert of action, enterprise liability and alternative liability rejected); Bichler v. Eli Lilly & Co. (1982), 55 N.Y.2d 571, 436 N.E.2d 182, 450 N.Y.S.2d 776 (because DES manufacturer made no motion to dismiss the complaint for failure to state a cause of action, concerted action theory became controlling law of case), overruled, Hymowitz v. Eli Lilly & Co. (1989), 73 N.Y.2d 487, 508, 539 N.E.2d 1069, 1076, 541 N.Y.S.2d 941, 948; contra Abel v. Eli Lilly & Co. (1984), 418 Mich. 311, 343 N.W.2d 164 (allowed concert of action).) Plaintiff in our case either included these causes of action in her amended complaint or argued them in her briefs. The circuit court dismissed all but the market share liability theory, and the appellate court affirmed application of that theory. In this appeal, plaintiff has chosen not to cross-appeal the dismissal of the claims based on concert of action, civil conspiracy, enterprise liability and alternative liability. Instead, we have before us only the narrow legal issue of whether to adopt market share liability in negligence and strict liability actions filed by a DES daughter. Currently, four States have adopted some form of this theory when confronted with the issue of imposing liability on drug manufacturers for injuries caused to women whose mothers ingested DES while pregnant. However, none of these States agree on the remedy or its application.
IV. JUDICIALLY PROMULGATED MARKET SHARE THEORIES
A. California
In Sindell v. Abbott Laboratories (1980), 26 Cal. 3d 588, 607 P.2d 924, 163 Cal. Rptr. 132, the California Supreme Court rejected the plaintiffs three bases for her cause of action and instead modified the alternative liability theory, thus fashioning its form of market share liability. In reaching this conclusion, the court reasoned that in a contemporary complex industrialized society, advances in science and technology create fungible goods which may harm consumers and which cannot be traced to any specific producer. It then went on to give three policy reasons for developing market share liability. First, as between an innocent plaintiff and a manufacturer of a defective product, the manufacturer should bear the cost of the injury. Second, it believed that the manufacturer was in a better position to bear the cost involved in an injury. Third, because the manufacturer is in the best position to recognize defects in products and to guard against them, holding the producer liable for these defects would provide an incentive to product safety.
Under the remedy as fashioned in Sindell, the plaintiff must first join as defendants the manufacturers of a “substantial share” of the DES which her mother may have taken, and must prove a prima facie case on every element except identification of the direct tortfeasor. After joining the manufacturers, the burden of proof shifts to defendants to demonstrate that they could not have manufactured the DES that caused plaintiffs injuries. If a defendant fails to meet this burden, the court fashions a market share theory to apportion damages according to the likelihood that any of defendants supplied the product by holding each defendant liable for the proportion of the judgment represented by its share of that market. The intended result of the rule is that each manufacturer’s liability for an injury is approximately equivalent to the damages caused by the DES it manufactured.
The Sindell court realized that the rule was not flawless and, in Brown v. Superior Court (1988), 44 Cal. 3d 1049, 751 P.2d 470, 245 Cal. Rptr. 412, the California Supreme Court resolved some of the ambiguities. Brown held that liability in the market share theory is not joint and several, rather it is only several. Furthermore, in cases in which all manufacturers in the market are not joined, liability will be limited to the market share represented, resulting in a less than 100% recovery for a plaintiff.
From its inception Sindell has not been widely accepted. In Sindell, Justice Richardson, joined by two other justices, dissented, arguing that the majority was abandoning a traditional tort requirement for the creation of a new, modified, industry-wide tort. Justice Richardson argued that the theory will result in imposition of liability on pure conjecture and that it rewards the plaintiff who, unlike the ordinary plaintiff, no longer has to take the chance that the responsible defendant cannot be reached or is unable to respond financially. Therefore, “it is readily apparent that ‘market share’ liability will fall unevenly and disproportionately upon those manufacturers who are amenable to suit in [those few jurisdictions which adopt some form of the theory].” (Sindell, 26 Cal. 3d at 617, 607 P.2d at 940, 163 Cal. Rptr. at 148 (Richardson, J., dissenting).) The dissent stressed that the opinion has the effect of making pharmaceutical companies insurers of their industry and that because of the sweeping effect of market share liability, the policy decision to introduce and define it should rest not with the court but with the legislature.
Other than the overall concept of market share liability, which will be addressed later in this opinion, the rule as specifically developed in Sindell has been extensively criticized, and as of this date only one Federal district court has adopted it in the same form. (McElhaney v. Eli Lilly & Co. (D.S.D. 1983), 564 F. Supp. 265, 270-71 (applying what it thought would be South Dakota law).) Criticisms include that the court failed to identify the relevant market for purposes of determining a particular defendant’s market share, i.e., local, countywide, statewide or national, and a manufacturer’s liability will vary widely depending on which market is used. This uncertainty undermines the court’s claim that market share liability approximates each defendant’s responsibility for the injuries caused by its own products. (Fischer, Products Liability—An Analysis of Market Share Liability, 34 Vand. L. Rev. 1623, 1643-44 (1981).) The court also failed to define what constitutes a “substantial share” of the market, one which is sufficient to shift the burden of proof to the defendant. A law review article that influenced the court suggested that plaintiff join 75% to 80% of the manufacturers (Comment, DES and a Proposed Theory of Enterprise Liability, 46 Fordham L. Rev. 963, 995-96 (1978)), but the court rejected this as too high and held that only a substantial percentage is required (Sindell, 26 Cal. 3d at 612, 607 P.2d at 937, 163 Cal. Rptr. at 145). Moreover, Sindell failed to specify how the market for DES can be allocated fairly when DES has been prescribed for uses other than as a miscarriage preventative. (See also Miller & Hancock, Perspectives on Market Share Liability: Time for a Reassessment?, 88 W. Va. L. Rev. 81, 88-91 (1985); Note, The DES Causation Conundrum: A Functional Analysis, 32 N.Y.L. Sch. L. Rev. 939, 959-61 (1987).) Further attestation to the flaws in the specifically developed procedure in California is the fact that it was rejected by the three other State supreme courts which have recognized some form of market share liability, as well as by our own appellate court in the case now before us. 173 Ill. App. 3d 1, 18; Hymowitz v. Eli Lilly & Co. (1989), 73 N.Y.2d 487, 511, 539 N.E.2d 1069, 1077-78, 541 N.Y.S.2d 941, 949-50; Collins v. Eli Lilly Co. (1984), 116 Wis. 2d 166, _, 342 N.W.2d 37, 48-49; Martin v. Abbott Laboratories (1984), 102 Wash. 2d 581, 600, 689 P.2d 368, 380.
B. Washington
The theory most closely paralleling the Sindell rule is the “market share alternate liability” theory which the Washington Supreme Court adopted in Martin v. Abbott Laboratories (1984), 102 Wash. 2d 581, 689 P.2d 368. As in Sindell, Martin found unavailing the enterprise and alternative liability, concert of action and civil conspiracy theories. It further believed that the California approach was insufficient because the court failed to define what constituted a “substantial” share of the market and, the Martin court mistakenly believed, it distorted liability by providing that the “substantial market share” bears joint responsibility for 100% of plaintiff’s injuries. The Washington court nonetheless did not outrightly reject the market share theory. It reasoned that each defendant contributed to the risk of injury to the public and consequently to the risk of injury to the plaintiff. Thus, each defendant shares in some measure a degree of culpability in producing or marketing DES.
The market share alternate liability theory that the Washington court formulated allows the plaintiff to bring suit against only one defendant. The plaintiff must prove that her mother took DES; the DES caused subsequent injuries; the defendant produced or marketed the type of DES taken by plaintiff’s mother; and the production and marketing of DES breached a legally recognized duty to the plaintiff. The burden then shifts to the defendant to prove by a preponderance of evidence that it did not produce or market the type of DES taken by the mother; did not produce or market DES in that geographical area; or did not produce or market DES at that time. The defendant or defendants unable to exculpate themselves become members of the plaintiff’s DES market. In George v. Parke-Davis (1987), 107 Wash. 2d 584, 592, 733 P.2d 507, 512, the court clarified that the geographic market area ideally should be defined on a local level; however, where such evidence is unavailable, county, State, or even national figures may be admissible to determine the market share.
Defendants are presumed initially to have an equal market share and are liable on a pro rata basis. They may rebut this presumption by proving their actual market share and are then only liable for that percentage of the damages. The presumptive share of the defendants that are unable to establish their actual market share is adjusted upward so that 100% of the market is accounted for. If all defendants are able to establish their actual market share and the percentage of the market represented is less than 100%, plaintiff’s recovery is limited to that percentage of the market which is actually represented. Our appellate court in this case and a Federal district court in Massachusetts have subsequently adopted this theory. 173 Ill. App. 3d at 18; McCormack v. Abbott Laboratories (D. Mass. 1985), 617 F. Supp. 1521, 1527 (court held the theory was consistent with statements the State supreme court made in Payton v. Abbott Laboratories (1982), 386 Mass. 540, 437 N.E.2d 171).
The Martin alternative was formulated in part on the erroneous belief that Sindell created joint and several liability and that it increased the share of each defendant found liable by the share attributed to nonjoined manufacturers. (Martin, 102 Wash. 2d at 601-02, 689 P.2d at 380-81; but see Brown v. Superior Court, 44 Cal. 3d at 1075, 751 P.2d at 487, 245 Cal. Rptr. at 428.) Our appellate court was also under the same misconceptions. (173 Ill. App. 3d at 18.) Defendants Boyle and Massengill contend that instead of refining Sindell so that liability will more closely equate the harm caused, this theory has the realistic potential of creating liability well in disproportion to a manufacturer’s market share. First, under Sin-dell the plaintiff must at least bring a substantial number of potential defendants before the court, whereas Martin only requires plaintiff to sue one defendant. If that sole defendant is a small contributor to the DES market, such as Boyle and Massengill, it possibly could shoulder complete liability without proof of its being the cause in fact for the injury. Second, a smaller company which no longer has records of its actual market share will be given a presumptive share under Martin which is equal to that portion of the damages unattributed. Thus, the small company could be responsible for 50% or more of the damages when common sense dictates that it surely could not have distributed such a high percentage of the DES used in the market. Also, defendants assigned presumptive shares are held liable for the share of the market attributable to companies no longer in business or not otherwise amenable to suit. This makes those defendants insurers of products which others made. Therefore, this theory can be substantially unfair to any company that is unable to prove its market share, especially if that company is small.
C. Wisconsin
The Wisconsin Supreme Court addressed the DES liability issue in Collins v. Eli Lilly Co. (1984), 116 Wis. 2d 166, 342 N.W.2d 37. The court rejected unalloyed market share liability, concluding that it “does not constitute the most desirable course to follow in DES cases because the theory, while conceptually attractive, is limited in practical applicability.” (116 Wis. 2d at 189, 342 N.W.2d at 48.) It found that defining the market and apportioning the market share is nearly an impossible task if it is to be done fairly and accurately and that a second mini-trial to determine market share would be a waste of judicial resources. Therefore, Wisconsin formulated its alternative, commonly referred to as the “risk contribution theory.” The public policy grounds articulated in support of this theory were that:
“Each defendant contributed to the risk of injury to the public and, consequently, the risk of injury to individual plaintiffs ***. Thus each defendant shares, in some measure, a degree of culpability in producing or marketing *** a drug with possibly harmful side effects. Moreover, as between the injured plaintiff and the possibly responsible drug company, the drug company is in a better position to absorb the cost of the injury. *** Finally, the cost of damages awards will act as an incentive for drug companies to test adequately the drugs they place on the market for general medical use.” (Emphasis in original.) 116 Wis. 2d at 191-92, 342 N.W.2d at 49.
Under the theory, the plaintiff must also allege that her mother ingested DES and this caused plaintiff’s injuries; that defendant manufactured or marketed the type of DES ingested; and that the defendant’s conduct constituted a breach of a legally cognizable duty to the plaintiff. The plaintiff need only sue one drug company and that company need not constitute a substantial share of the market. Once the plaintiff has proven a prima facie case under negligence or strict liability, the burden shifts to the defendant to prove by a preponderance of the evidence that it did not produce or market DES for the prevention of miscarriage during the relevant time period or in the relevant geographical market area. If only one company is sued and no others are impleaded, that company is liable for all the damages if it cannot exculpate itself. If more than one defendant is joined or impleaded, damages are determined according to the jury’s assignment of liability under Wisconsin’s comparative negligence statute. The court included a number of factors for the jury to consider in apportioning damages, such as the market share of the defendant, whether the company conducted safety tests on DES, the role the company played in seeking FDA approval of the drug, and whether the company issued warnings.
No court, other than the Wisconsin court, addressing the DES causation issue has gone so far as to impose total liability on a defendant merely for creating a risk of harm. It has been said that this theory contravenes the fundamental tort principle that a mere possibility is insufficient to satisfy causation. (See Note, The DES Causation Conundrum: A Functional Analysis, 32 N.Y.L. Sch. L. Rev. 939, 965-66 (1987) (imposition of liability under this approach requires a substantial reduction in the degree of proof traditionally required, and thus threatens over-deterrence and inequity).) It has been contended that Collins does not resolve its perceived errors in market share liability, but rather further exacerbates them. (Miller & Hancock, Perspectives on Market Share Liability: Time for a Reassessment?, 88 W. Va. L. Rev. 81, 99-101 (litigation costs will increase, there is a risk of overwhelming the jurors with evidence, and the Sindell mini-trial is transformed into a maxi-trial on a plethora of issues).) Furthermore, it is possible that liability will far exceed the probability that a defendant caused the injuries. Comment, Torts — Products Liability — Where a Plaintiff Cannot Identify Which Drug Company Manufactured the DES Ingested, a Cause of Action Exists Under the Market-Share Alternate Theory of Liability, 55 Miss. L.J. 195, 210 (1985).
D. New York
The court of appeals of New York recently declined to accept Wisconsin’s risk contribution theory, believing that it would only be feasible on a limited scale. (Hymowitz v. Eli Lilly & Co. (1989), 73 N.Y.2d 487, 539 N.E.2d 1069, 541 N.Y.S.2d 941.) The court was wary “of setting loose, for application in the hundreds of cases pending in this State, a theory which requires the fact finder’s individualized and open-ended assessment of the relative liabilities of scores of defendants in every case.” (Hymowitz, 73 N.Y.2d at 511, 539 N.E.2d at 1077-78, 541 N.Y.S.2d at 949-50.) It concluded that the injustice resulting from delays in recoveries and inconsistent results militated against adoption of the theory. New York also criticized and rejected the California and Washington versions of market share liability, noting the difficulty of determining market share. However, the Hymowitz court did develop its own version of market share liability.
New York’s theory utilizes a national market. The court did this realizing that a national market could not provide a reasonable link between liability and the risk created by a defendant to a particular plaintiff. Instead, this theory apportions “liability so as to correspond to the over-all culpability of each defendant, measured by the amount of risk of injury each defendant created to the public-at-large.” (Hymowitz, 73 N.Y.2d at 512, 539 N.E.2d at 1078, 541 N.Y.S.2d at 950.) A defendant can exculpate itself only through proof that it did not participate in the marketing of DES for pregnancy use. The rule also provides that liability is several only, and is not to be inflated if all the manufacturers are not before the court.
Though it is too early to determine how Hymowitz will be received, it certainly is the most radical in its departure from established tort principles and it is admittedly flawed in that it cannot equate liability to actual harm caused. (Hymowitz, 73 N.Y.2d at 512, 539 N.E.2d at 1078, 541 N.Y.S.2d at 950; see Shackil v. Lederle Laboratories (1989), 116 N.J. 155, 197, 561 A.2d 511, 532 (O’Hern, J., dissenting) (recognizing Hymowitz as perhaps the most controversial of the market share decisions); Note, Hymowitz v. Eli Lilly: New York Adopts a “National Risk” Doctrine for DES, 25 Tort & Ins. L.J. 150 (1989); but see Twerski, Market Share — A Tale of Two Centuries, 55 Brooklyn L. Rev. 869 (1989).) Just as the previous theories have not been embraced by subsequent courts, it is unlikely that New York’s theory will receive broad acceptance.
V. COURTS WHICH HAVE REJECTED MARKET SHARE LIABILITY
Other than these cases, the concept of market share liability has not received strong support. The supreme courts of two of our sister States have outrightly rejected its application in DES daughter cases.
The Iowa Supreme Court rejected the doctrine “on a broad policy basis.” (Mulcahy v. Eli Lilly & Co. (Iowa 1986), 386 N.W.2d 67, 75.) Mulcahy equated the theory with a court-constructed insurance plan which requires manufacturers to pay for injuries their product may not have caused. It recognized market share liability as a radical departure from traditional tort concepts and it rejected allowing “ ‘negligence in the air’ ” to serve as a substitute for causation in fact. (386 N.W.2d at 76, quoting F. Pollock, The Law of Torts 455 (11th ed. 1920).) The court concluded “that awarding damages to an admitted innocent party by means of a court-constructed device that places liability on manufacturers who were not proved to have caused the injury involves social engineering more appropriately within the legislative domain.” Mulcahy, 386 N.W.2d at 76.
The Missouri Supreme Court agreed with the arguments of the drug manufacturers that market share liability was unfair, unworkable and contrary to Missouri law and violated the State’s public policy. (Zafft v. Eli Lilly & Co. (Mo. 1984), 676 S.W.2d 241, 246.) The court found that Sindell had not sufficiently articulated the concepts involved. Furthermore, there was too great a risk that the actual wrongdoer was not before the court and the rule exposed those who were joined to liability greater than their responsibility. Zafft rejected the arguments that as between an innocent plaintiff and negligent defendant, the latter should bear the cost of the injury, and that defendants can better absorb the costs. It noted that the requirement of proving causation had not been altered by the development of products liability law in Missouri. Thus, shifting the burden of proof to the defendants would significantly alter existing rights and liabilities of the litigants. It also reasoned that strong public policy arguments militated against market share liability. The court was concerned that liability of this type would “discourage desired pharmaceutical research and development while adding little incentive to production of safe products” because all companies face potential liability regardless of their safety efforts. (676 S.W.2d at 247.) The court concluded that there was insufficient public policy justification to support abandonment of so fundamental a concept of tort law as the requirement that a plaintiff prove, at a minimum, a nexus between wrongdoing and injury.
Most of the Federal courts which have addressed the issue of applying market share liability in a DES case have declined to adopt such -a radical departure from the common law of the State in which each sits without a clearer direction from that State’s supreme court. In Tidler v. Eli Lilly & Co. (D.C. Cir. 1988), 851 F.2d 418, the court reasoned “that the theory that plaintiffs would have us ‘construct’ requires that we build on a new foundation, not on the structural underpinnings of the traditional common law of torts.” (851 F.2d at 424.) Neither the highest court of Maryland nor of the District of Columbia had addressed the issue, and the Tidler court held that such a marked deviation from the common law was beyond the authority of a Federal court. In Mizell v. Eli Lilly & Co. (D.S.C. 1981), 526 F. Supp. 589, the district court found that according to conflict of law principles, California substantive law, and thus the Sindell rule, was the appropriate choice of law. However, the court refused to apply California substantive law because it would violate the public policy of the forum. The court concluded that “[m]arket share represents a radical departure from the body of products liability law that has been developed in South Carolina” and has the potential for placing liability on defendants who bear no responsibility for the defective product. 526 F. Supp. at 596; see also Morton v. Abbott Laboratories (M.D. Fla. 1982), 538 F. Supp. 593, 599 (“market share theory unquestionably represents a radical departure from the traditional concept of causation” and there was no indication that Florida would abandon such a fundamental principle); Pipon v. Burroughs-Wellcome Co. (D.N.J. 1982), 532 F. Supp. 637, 639 (there is no indication that the New Jersey Supreme Court would deviate from the causation requirement), aff’d (3d Cir. 1982), 696 F.2d 984; Ryan v. Eli Lilly & Co. (D.S.C. 1981), 514 F. Supp. 1004, 1019.
Plaintiffs have pursued the application of market share liability with minimal success in areas other than DES cases. The plaintiff in Shackil v. Lederle Laboratories (1989), 116 N.J. 155, 561 A.2d 511, became severely retarded as a result of a diphtheria, pertussis' and tetanus (DPT) vaccine. Unable to identify the specific manufacturer, plaintiff sued a number of manufacturers who potentially could have produced the vaccine she was given and argued for adoption of a market share liability theory. The court determined that to adopt market share liability in a DPT “case would frustrate overarching public-policy and public-health considerations by threatening the continued availability of needed drugs and impairing the prospects of the development of safer vaccines.” (116 N.J. at 158, 561 A.2d at 512.) The court’s decision was further influenced by the fact that Congress had already established legislation to compensate vaccine-injured plaintiffs. (National Childhood Vaccine Injury Act of 1986, 42 U.S.C. §§300aa — 1 through 300aa — 34 (Supp. V 1987).) The court also addressed the plaintiff’s argument that there was a trend in New Jersey to relax the causation requirement. It noted that a trial court in Ferrigno v. Eli Lilly & Co. (1980), 175 N.J. 551, 420 A.2d 1305, held that alternative liability based on a percentage-share apportionment was permissible in a DES case. However, that complaint was dismissed following the appellate court opinion in Namm v. Charles E. Frosst & Co. (App. Div. 1981), 178 N.J. Super. 19, 427 A.2d 1121, which refused to adopt alternative liability or enterprise liability in a DES action. Upon further review of New Jersey law, it found that there was no trend toward wholesale adoption of market share liability.
The Oregon Supreme Court rejected use of the theory against two DPT manufacturers in the context of a design defect. (Senn v. Merrell-Dow Pharmaceuticals, Inc. (1988), 305 Or. 256, 751 P.2d 215.) The court claimed that the “adoption of any theory of alternative liability requires a profound change in fundamental tort principles,” which is more properly the domain of the legislature. 305 Or. at 271, 751 P.2d at 223; see also Chapman v. American Cyanamid Co. (11th Cir. 1988), 861 F.2d 1515 (child died after receiving a DPT vaccine; parents could not proceed against three manufacturers on an alternative liability theory); Poole v. Alpha Therapeutic Corp. (N.D. Ill. 1988), 696 F. Supp. 351 (Federal court would not recognize market share liability in an action against manufacturers of a type of blood product from which plaintiff contracted AIDS); Griffin v. Tenneco Resins, Inc. (W.D.N.C. 1986), 648 F. Supp. 964 (court determined that under North Carolina law manufacturers of benzidine congener dyes could not be held liable based on market share theory); Sheffield v. Eli Lilly & Co. (1983), 144 Cal. App. 3d 583, 192 Cal. Rptr. 870 (rejecting application of market share liability against manufacturers of Salk polio vaccine); but see Morris v. Parke, Davis & Co. (C.D. Cal. 1987), 667 F. Supp. 1332 (applying market share liability against manufacturers of DPT based on allegations of industry manufacturing defects).
Other than in actions against drug manufacturers, the major area of cases in which plaintiffs have attempted to impose market share liability has been asbestos litigation. The success rate in these cases is considerably less than in DES cases. In Goldman v. Johns-Manville Sales Corp. (1987), 33 Ohio St. 3d 40, 514 N.E.2d 691, the Ohio Supreme Court rejected its application in an action against suppliers and manufacturers of products containing asbestos by the wife of a person who died allegedly due to asbestos exposure. The court reasoned that market share liability is inappropriate “in an asbestos litigation case, especially where it cannot be shown that all the products to which the injured party was exposed are completely fungible.” (33 Ohio St. 3d at 50, 514 N.E.2d at 700.) Moreover, the risk the manufacturer created is not accurately reflected in its market share because many products contain different degrees of asbestos, and the largest asbestos supplier, Johns-Manville, was not amenable to suit. Instead of adopting such a divergent theory, the court concluded that the problem was more in need of a legislative solution. See Case v. Fibreboard Corp. (Okla. 1987), 743 P.2d 1062, 1067 (“the public policy favoring recovery on the part of an innocent plaintiff does not justify the abrogation of the rights of a potential defendant to have a causative link proven between that defendant’s specific tortious acts and the plaintiff’s injuries”); Mullen v. Armstrong World Industries, Inc. (1988), 200 Cal. App. 3d 250, 246 Cal. Rptr. 32; Nutt v. A.C. & S. Co. (Del. Super. 1986), 517 A.2d 690, 694 (rejecting market share liability and recognizing that such a change in tort law should be left to the legislature); Bateman v. Johns-Manville Sales Corp. (5th Cir. 1986), 781 F.2d 1132; Thompson v. Johns-Manville Sales Corp. (5th Cir. 1983), 714 F.2d 581; Marshall v. Celotex Corp. (E.D. Mich. 1987), 651 F. Supp. 389, 393 (“asbestosis litigation is an inappropriate area in which to extend market share liability”); Starling v. Seaboard Coast Line R.R. Co. (S.D. Ga. 1982), 533 F. Supp. 183; see also Cummins v. Firestone Tire & Rubber Co. (1985), 344 Pa. Super. 9, 495 A.2d 963 (rejecting theory in action against manufacturers of multipiece tire and rim assemblies because products are not sufficiently similar to be considered identical or fungible); Bixler v. Avondale Mills (Minn. App. 1987), 405 N.W.2d 428 (cotton flannelette not fungible product).
VI. ANALYSIS OF MARKET SHARE IN ILLINOIS
Each of the four courts which have adopted some form of market share liability has criticized and ultimately rejected in whole or in part the theory as developed in the other jurisdictions. Our appellate court also recognized that its theory may be flawed but accepted this and believed that subsequent opinions could eventually resolve the uncertainties which develop, but which other courts apparently have been unable to resolve in the past decade. (173 Ill. App. 3d at 22.) We conclude that market share liability is not a sound theory, is too great a deviation from our existing tort principles and should not be applied in cases brought by plaintiffs who were exposed to DES while in utero.
In addition to the criticisms already expressed by the courts which recognize market share liability, we see numerous problems with its adoption. A major flaw, in regards to DES cases, is that there is only a small amount of, or in some cases no, reliable information available to establish the defendants’ percentages of the market. As mentioned earlier, no party can be blamed for this fact. It is, in part, the result of the laws in effect regarding maintenance of records and factors relating to the long lapse in time from the sale of the drug to the filing of the lawsuit. The lack of available records is evidenced in this case by the fact that after extensive discovery plaintiff was unable to identify the responsible manufacturer. Many of those defendants who have been named are no longer in business or have filed motions challenging jurisdiction and for these companies especially it is unlikely that records will be available to establish their share of any market.
The courts which have adopted market share liability have done so while ruling on pretrial motions and have not had the benefit of first having heard evidence on the availability of market share data. (See Sindell, 26 Cal. 3d at 613, 607 P.2d at 937-38, 163 Cal. Rptr. at 145-46 (“We are not unmindful of the practical problems involved in defining the market and determining market share, but these are largely matters of proof which properly cannot be determined at the pleading stage of these proceedings”).) Unlike these courts, we have the benefit of the experiences of the trial courts in California which have been instructed to apply the market share liability theory. In San Francisco, the trial court determined that the only logical or practical definition for “market” would have to be on a national scale because the parties were unable to present data on a more narrow market, which the State supreme court directed it should attempt to do. The trial judge in Los Angeles expressed exasperation with the task of attempting to formulate market shares after spending over four weeks examining the DES market. (Stapp v. Abbott Laboratories (Super. Ct. Los Angeles County), No. C 344407 (“The harsh blunt fact that the evidence has shown is that that information and data is just not available” and “when the Supreme Court, *** without having any evidence says that you can determine what the [sales are] as to a particular manufacturer, it’s just, just not there. That data doesn’t exist”).) Plaintiff here argues that the difficulty the trial judge in Los Angeles experienced could be attributed to the uncooperativeness of the defendants. The transcripts clearly refute this assertion. The judge began his analysis of the situation by thanking the parties for all their cooperation and for the highly professional manner in which the case was presented. The judge then went on to criticize those who developed the market share theory because of their obvious lack of trial experience or knowledge as to what would go into proving a case based on the theory.
Acceptance of market share liability and the concomitant burden placed on the courts and the parties will imprudently bog down the judiciary in an almost futile endeavor. This would also create a tremendous cost, both monetarily and in terms of the workload, on the court system and litigants in an attempt to establish percentages based on unreliable or insufficient data. See Fischer, Products Liability — An Analysis of Market Share Liability, 34 Vand. L. Rev. 1623, 1657 (1981) (“The legal fees and administrative costs arising from litigation of this magnitude easily could rival the cost of the plaintiff’s judgment”); Comment, Market Share Liability for Defective Products: An Ill-Advised Remedy for the Problem of Identification, 76 Nw. U.L. Rev. 300, 323-26 (1981).
If we were to allow courts and juries to apportion damages when reliable information is not available, the clear result would be that the determinations will be arbitrary and there will be wide variances between judgments, without sufficient explanation as to these differences. The unfairness inherent in apportioning damages without adequate evidence is increased for a number of reasons. It is likely that the defendant who actually sold the product is not before the court. For example, in this case defendants have presented evidence that 63 of the potential 81 manufacturers were never before the court. Other defendants either were not served, have gone out of business, have merged with other companies and due to successor liability laws cannot be held liable for the sale of DES, or are not amenable to suit in Illinois. To impose liability when it is quite possible that the defendant is not before the court is too speculative. Ryan v. Eli Lilly & Co. (D.S.C. 1981), 514 F. Supp. 1004, 1007, 1018.
Moreover, it is unrealistic to say that a true percentage of the market can be established by the defendants before the court. Throughout the history of the use of DES as a miscarriage preventative, hundreds of manufacturers produced the product and it is impossible to bring them before our courts. The defendants who do appear will have a difficult enough time to establish their market shares. Those who cannot meet this task but desire to reduce their potential liability will have the difficult burden of establishing the shares of manufacturers not before the court. (See George v. Parke-Davis, 107 Wash. 2d at 597, 733 P.2d at 514 (in order to inhibit defendants from randomly impleading insolvent corporations to reduce their share of presumptive liability, defendants are required to establish the actual market share of impleaded defendants).) The result would likely be that, even though common sense dictates that other companies are responsible, market share liability makes those companies which are unable to establish their market share liable for a wholly speculative and disproportionate amount of the damages. (See Sindell, 26 Cal. 3d at 616, 607 P.2d at 939-40, 163 Cal. Rptr. at 147-48 (Richardson, J., dissenting) (it is highly speculative that defendant’s liability equals the harm actually caused); Fischer, Products Liability — An Analysis of Market Share Liability, 34 Vand. L. Rev. 1623, 1645-47 (1981) (Sindell rule inherently distorts defendant’s liability).) If the goal of this theory is to attribute damages in accordance with the harm caused, a truly accurate market should be limited to those manufacturers of DES which was used as a miscarriage preventative and their share of that narrower market because the drug was and is safe for the other purposes for which it was sold. (Cf. Hymowitz, 73 N.Y.2d at 512, 539 N.E.2d at 1078, 541 N.Y.S.2d at 950 (realizing that liability will not, over the run of cases, approximate causation).) Surely, this specific market would be nearly impossible to establish.
Market share liability also has the potential to treat plaintiffs who cannot identify the specific manufacturer responsible for the DES maternally ingested more favorably than one who can. In a typical tort case the plaintiff takes the risk that the defendant will be unable to assume financial responsibility for injuries caused. However, with the market share theory, liability is spread throughout members of the industry, reducing the risk that plaintiff will be without a solvent defendant. The theory thus punishes plaintiffs who can satisfy the identification element, while creating an incentive not to locate the particular manufacturer. Comment, Overcoming the Identification Burden in DES Litigation: The Market Share Liability Theory, 65 Marq. L. Rev. 609, 632-33 (1982) (arguing further that the theory exposes defendants to double liability, first to plaintiffs who can identify them as the causal party, and again to plaintiffs who cannot); Comment, The Application of a Due Diligence Requirement to Market Share Theory in DES Litigation, 19 J.L. Reform 771, 782-83 (1986) (without a due diligence requirement there will be little incentive for a plaintiff to identify the causal manufacturer); but see McCormack v. Abbott Laboratories (D. Mass. 1985), 617 F. Supp. 1521, 1528 (rejecting a due diligence requirement).
The appellate court supported its conclusion that market share liability should be adopted based in part on analogies to two other causation exceptions. In res ipsa loquitur and alternative liability the burden of identifying the culpable defendant is shifted from the plaintiff to defendants. We recognized res ipsa loquitur in Kolakowski v. Voris (1980), 83 Ill. 2d 388, where the plaintiff was injured while unconscious during surgery and could not identify the negligent party. We stated that the purpose of res ipsa loquitur is to allow plaintiff to establish negligence by circumstantial evidence when the direct evidence concerning the cause of injury is primarily within the knowledge and control of defendant. In order to take advantage of the inference, plaintiff “must show that he was injured (1) in an occurrence which would not have occurred in the absence of negligence, (2) by an instrumentality or agency under the management or control of the defendant, and (3) under circumstances which were not due to any voluntary act or negligence on the part of the plaintiff.” (83 Ill. 2d at 394. But see Dyback v. Weber (1986), 114 Ill. 2d 232, 239 (in light of comparative negligence, plaintiff’s freedom from contributory negligence is no longer a requirement in order to make out a prima facie case under res ipsa loquitur).) In Summers v. Tice (1948), 33 Cal. 2d 80, 199 P.2d 1, the theory of alternative liability was recognized in an action where the plaintiff was injured when two hunters negligently shot in his direction, but the plaintiff could not ascertain which hunter’s bullet injured him. The court-developed exception was codified in Restatement (Second) of Torts, section 433B, and reads as follows:
“Where the conduct of two or more actors is tortious, and it is proved that harm has been caused to the plaintiff by only one of them, but there is uncertainty as to which one has caused it, the burden is upon each such actor to prove that he has not caused the harm.” (Restatement (Second) of Torts §433B(3), at 441-42 (1965).)
As noted, every court which has addressed the issue has held, for a number of reasons, that alternative liability does not apply to DES cases.
On the surface, cases utilizing these concepts and market share liability seem similar in that the plaintiff in each lacks evidence to establish the identity of the responsible defendant and as a result the court shifts the burden of proof to the defendants. However, though there exist some similarities, the analogy is too tenuous to rely on res ipsa loquitur and alternative liability as a sound basis for adopting the theory. Comment, Market Share Liability for Defective Products: An Ill-Advised Remedy for the Problem of Identification, 76 Nw. U.L. Rev. 300, 307-12 (1981).
In res ipsa loquitur and alternative liability situations, all parties who could have been the cause of the plaintiff’s injuries are joined as defendants. This helps to preserve the identification element because liability will surely fall on the actual wrongdoer. By contrast, market share liability merely requires the plaintiff to name as defendants either a substantial share of those in the market or, in some theories, only one manufacturer who was in the market. As a result, there is a real possibility that the defendant actually responsible for the injuries is not before the court. Second, in res ipsa loquitur and alternative liability, burden-shifting is considered equitable because defendants are typically in a better position than the plaintiff to determine who caused the harm. Market share liability shifts the burden to defendants without regard to whether plaintiff is better able to identify the defendant responsible or without regard to defendants’ ability to identify who among them is actually responsible. As is clearly demonstrated in these DES cases, the manufacturers are in no better position than the plaintiffs to identify the culpable party. (Kroll, Intra-Industry Joint Liability: The Era of Absolute Products Liability, 687 Ins. L.J. 185, 194 (1980).) Third, in the earlier exceptions the burden is shifted to parties who bear some culpability for causing plaintiff’s injury. In alternative liability, each defendant was at least negligent toward the plaintiff, and in res ipsa loquitur at least one defendant caused the injury and the others are intimately connected to the activity and instrumentality that caused the harm. But with market share liability the named defendant need not have been directly connected with the activity or instrumentality that caused the harm. Indeed, it is inevitable that some defendants wholly innocent of wrongdoing towards the particular plaintiff will shoulder part or all of the responsibility for the injury caused.
Plaintiff further attempts to support her position by contending that market share liability should be applied because she has maintained a “sufficient connection” between each of the named defendants and the form of the DES which caused her condition. This link is allegedly established because of the joint industry efforts in obtaining FDA approval to sell DES. Plaintiff points out that in 1941 a “small committee” was created to gather FDA-required data and the committee jointly submitted an application for approval to market DES for non-pregnancy-related purposes. The efforts of this committee formed the basis of subsequent FDA approval for the manufacturing of DES by other companies. The information utilized at that time was also influential in securing approval in 1947 for use of DES to prevent miscarriages. Plaintiff contends she has brought before the court “virtually all” of the companies which comprised the small committee, thus she has all the parties responsible for making DES available for use as a miscarriage preventative. Moreover, she claims to have narrowed the number of potential defendants to only a few and that liability may be imposed upon them.
We believe that plaintiff’s “link” is insufficient to create what is commonly understood as the connection between a potentially responsible defendant and the injury-causing product. No court which has addressed the issue has found that the actions in 1941 constituted a joint, concerted or conspiratorial effort by defendants to market DES. (See, e.g., Hymowitz, 73 N.Y.2d at 506-07, 539 N.E.2d at 1074-75, 541 N.Y.S.2d at 946-47; Collins, 116 Wis. 2d at 183-88, 342 N.W.2d at 46-48.) It was upon the request of the FDA that the small group convened to present a master file of data on behalf of the rest of the industry. Thus, it is unwarranted to make each responsible for the others’ products based on some type of enterprise liability theory.
The connection between this committee and a plaintiff who suffers injuries as a result of DES’ being used to prevent miscarriages is even weaker because the FDA did not approve of that use of the drug until 1947 and under different circumstances than in 1941 and upon submission of additional information. (See, e.g., 173 Ill. App. 3d at 28 (our appellate court recognized that the joint submission of clinical data in 1941 was distinct from 1947 application process).) Reliance on the facts comprising the 1941 filings is unwarranted unless plaintiff is able to prove that one of the companies in the “small committee” manufactured the DES maternally ingested. Moreover, plaintiff overestimates her narrowing of the likely defendants. The record discloses a much greater number of potential defendants than the eight which are before this court. To say that the true defendant has been located is speculative and improper conjecture. (See Schmidt, 44 Ill. 2d at 405; Tiffin v. Great Atlantic & Pacific Tea Co. (1959), 18 Ill. 2d 48, 60.) The fact that over 300 companies sold a similar product for similar purposes cannot fairly be held to have created a sufficient nexus such that each company can be responsible for the injuries caused by the others’ products, even under the unique facts surrounding the approval of the manufacturing of DES.
Plaintiff next claims that certain underlying principles of products liability laws dictate that we should impose liability on the manufacturers. (See 173 Ill. App. 3d at 15-16 (detailing the policy reasons for products liability).) We agree with the idea that liability based on tort law should be shouldered by the responsible manufacturer or manufacturers. However, we do not believe that we should abrogate a fundamental precept of tort law to reach this goal and ignore the effects of adopting market share liability.
Other courts have looked to these underlying principles when reaching their conclusion of whether or not to recognize market share liability. One of the bases relied upon in adopting the theory is that the drug companies are better able to insure against liability and to pass the costs On. In its opinion, the appellate court concluded that the pharmaceutical drug companies were in solid financial condition and would be able to insure against drug-related costs. 173 Ill. App. 3d at 26.
Defendants have strongly contested the figures and conclusions regarding their financial status and the implications regarding other participants in the drug manufacturing industry. They further argue that the expansions in tort law are having the perverted results of eliminating production of certain useful and necessary drugs, and dramatically increasing insurance costs such that some companies either can no longer obtain insurance or cannot pass the costs on to consumers.
In support of these assertions they cite examples of drugs which no longer are being produced because of potential liability as well as areas where the Federal government has had to intercede to protect from liability and insure availability of a drug. (See generally Note, A Question of Competence: The Judicial Role in the Regulation of Pharmaceuticals, 103 Harv. L. Rev. 773 (1990).) The New Jersey Supreme Court recently declined on policy grounds to impose market share liability on manufacturers of DPT because of the crippling effect potential liability has had on the industry. (Shackil v. Lederle Laboratories (1989), 116 N.J. 155, 561 A.2d 511 (DPT industry was diminished to one supplier and the Federal government had established a fund to deal with liability claims).) It has also been necessary for government intervention to insure availability of the swine flu and polio vaccines.
We do not believe that in this case it is necessary to become embroiled in the “insurance crisis” or to speculate as to the relative financial security of the participants in the prescription drug industry. However, we note that market share liability will surely broaden manufacturers’ liability exposure because they will need to insure against losses arising from the products of others in the industry as well as their own. (See Fischer, Products Liability — An Analysis of Market Share Liability, 34 Vand. L. Rev. 1623, 1654 (1981) (adoption of a market share theory will dramatically increase liability exposure and it may discourage development of new products); Comment, Market Share Liability for Defective Products: An Ill-Advised Remedy for the Problem of Identification, 76 Nw. U.L. Rev. 300, 321-23 (1981) (increasing liability has prompted insurers to dramatically increase premiums and they are reluctant to insure particularly risky industries; this in turn has resulted in higher prices).) This added potential for liability will likely contribute to diminishing participants in the market as well as research and availability of drugs. (See Woodill v. Parke Davis & Co. (1980), 79 Ill. 2d 26, 37 (“This Court is acutely aware of the social desirability of encouraging the research and development of beneficial drugs”); Zafft, 676 S.W.2d at 247 (there are legitimate concerns that market share liability “will discourage desired pharmaceutical research and development while adding little incentive to production of safe products”); Payton v. Abbott Laboratories (1982), 386 Mass. 540, 574, 437 N.E.2d 171, 189-90 (“Imposition of such broad liability could have a deleterious effect on the development *** of new drugs, especially those marketed generically”).) It is tempting in this case to impose liability based on the fact that these companies profited from the sale of the type of drug which may be responsible for the plaintiff’s injuries, regardless of the manufacturers’ ability to cover these costs. However, this is not a strong enough reason to adopt a theory which would alter our tort law significantly while only providing a markedly flawed alternative with unclear future ramifications. (Sindell, 26 Cal. 3d at 618, 607 P.2d at 941, 163 Cal. Rptr. at 149 (Richardson, J., dissenting) (imposition based on defendant’s perceived wealth is an unsound principle and creates a two-tiered system of justice); Kroll, Intra-Industry Joint Liability: The Era of Absolute Products Liability, 687 Ins. L.J. 185, 195 (1980) (it is an unsound principle to impose liability based on the perceived wealth of defendant and its ability to obtain insurance); Comment, Market Share Liability for Defective Products: An Ill-Advised Remedy for the Problem of Identification, 76 Nw. U.L. Rev. 300, 328 (1981) (it is an unfair system to impose liability solely due to ability to pay and subsequently spread the costs).) Perhaps, as a number of other courts and commentators have suggested, this change is most appropriate for the legislature to develop, with its added ability to hold hearings and determine public policy. Goldman v. Johns-Manville Sales Corp. (1987), 33 Ohio St. 3d 40, 51, 514 N.E.2d 691, 701; Mulcahy v. Eli Lilly & Co. (Iowa 1986), 386 N.W.2d 67, 76; Note, Market Share Liability: A Plea for Legislative Alternatives, 1982 U. Ill. L. Rev. 1003.
Another underlying principle of products liability law is to enhance safer production of goods. (See Prosser, The Assault Upon the Citadel, 69 Yale L.J. 1099, 1119 (1960).) It is argued that adoption of market share liability in this case will provide incentive to produce safer generic drugs. However, we are not convinced that utilization of market share liability in suits against manufacturers of DES will have such an effect, though we recognize some courts and commentators believe market share liability is necessary for promotion of this safety goal. (See, e.g., Collins, 116 Wis. 2d at 192-93, 342 N.W.2d at 49-50; Sindell, 26 Cal. 3d at 611, 607 P.2d at 936, 163 Cal. Rptr. at 144; Robinson, Multiple Causation in Tort Law: Reflections on the DES Cases, 68 Va. L. Rev. 713, 741 (1982); Comment, The DES Manufacturer Identification Problem: A Florida Public Policy Approach, 40 U. Miami L. Rev. 857, 867-70 (1986).) First, it is not clear that the drug industry needs this even further amount of encouragement to produce safer drugs, above and beyond the incentives that products liability and negligence laws provide. (Fischer, Products Liability— An Analysis of Market Share Liability, 34 Vand. L. Rev. 1623, 1657 (1981) (market share liability may result in over-deterrence); Note, A Question of Competence: The Judicial Role in the Regulation of Pharmaceuticals, 103 Harv. L. Rev. at 780-83 (1990).) Second, it is unlikely that an overall safety incentive could result from imposition of market share liability 40 years after the undesirable behavior occurred and almost 20 years after the potential harm was discovered and the product removed from the market. (See Zafft, 676 S.W.2d at 247 (theory adds little incentive for production of safe products).) Third, market share liability imposes potential liability on all manufacturers in the particular industry; thus there may not be an incentive to produce safer products if liability could still be imposed as a result of the negligence of others in the industry and if the manufacturer knows that others in the industry will absorb the damages resulting from its negligence. (Miller & Hancock, Perspectives on Market Share Liability: Time for a Reassessment?, 88 W. Va. L. Rev. 81, 103-04 (1985).) Moreover, this theory is being accepted in a limited number of jurisdictions and is only being applied to manufacturers of DES or similar products. Thus by our adopting market share liability, the goal of warning manufacturers to produce safer products likely will not reach a wide array of producers.
Similarly, we find unavailing the appellate court’s conclusion that market share liability will encourage manufacturers to maintain more detailed records which will enable plaintiffs to identify the culpable party. (Robinson, Multiple Causation in Tort Law: Reflections on the DES Cases, 68 Va. L. Rev 713, 734-35 (1982) (this reason is scarcely dispositive in determining whether to shift the burden).) Due to the fungible nature of the product, after it leaves the control of the manufacturers they have very little ability to keep track of in what market and by whom the drug will ultimately be used. Defendants point out that when the drugs leave their plant they are identified, but it is along the chain of distribution that the goods become commingled and less traceable. For instance, in this case part of the problem in identification may be attributed to the Field Clinic for its labeling of the drug as only “Tab 98” and to the laws which require maintenance of records for only a short period of time. We do not believe that this or the supposed safety incentives provided are sufficient to adopt the theory.
Plaintiff also argues that the drug manufacturers are liable for their breach of duty to a foreseeable plaintiff. Under plaintiffs interpretation of duty, manufacturers of products for human consumption have a special responsibility and any manufacturer of DES can be held liable because it breached a duty owed to her. The appellate court accepted plaintiffs notion of duty, reasoning in part that drug manufacturers “owe a special duty of care to the public.” (173 Ill. App. 3d at 23.) In determining that a duty existed the appellate court also relied on Renslow v. Mennonite Hospital (1977), 67 Ill. 2d 348, in which this court held that a child not conceived at the time negligent acts by a doctor and hospital employees were committed against its mother could sue for that negligence. (67 Ill. 2d at 358.) However, in Kirk v. Michael Reese Hospital & Medical Center (1987), 117 Ill. 2d 507, we recognized that Renslow’s extension of duty was limited. We stated that the duty in Renslow arose “from a special relationship between either the defendant and the other party or the third-party plaintiff and the other party.” (Emphasis in original.) Kirk, 117 Ill. 2d at 528 (Renslow represents a “limited area of transferred negligence”).
The plaintiff and appellate court have too broadly interpreted the duty of a drug company and to whom it owes that duty. Both negligence and strict liability require proof that defendant breached a duty owed to a particular plaintiff. (See, e.g., Rowe v. State Bank (1988), 125 Ill. 2d 203, 215; Cunis v. Brennan (1974), 56 Ill. 2d 372, 374 (defendant is under an obligation for the benefit of a particular plaintiff).) Each manufacturer owes a duty to plaintiffs who will use its drug or be injured by it. However, the duty is not so broad as to extend to anyone who uses the type of drug manufactured by a defendant (see Kirk, 117 Ill. 2d at 519-21 (duty does not extend to unknown and unforeseeable non-users of drug company’s products)), and the fact that a duty is owed does not abrogate the requirement that the plaintiff maintains the responsibility of identifying the defendant who breached the duty.
The concept that liability may be imposed based merely on a breach of duty, without causation being established, has long been rejected in American tort law. (See Washington v. Atlantic Richfield Co. (1976), 66 Ill. 2d 103, 108-09; Palsgraf v. Long Island R.R. Co. (1928), 248 N.Y. 339, 341, 162 N.E. 99, 101 (negligence in the air is not a basis for imposing liability).) Though it has been suggested that the DES manufacturers’ creation of risk is sufficient to impose liability (Robinson, Multiple Causation in Tort Law: Reflections on the DES Cases, 68 Va. L. Rev. 713, 755 (1982) (damages imposed under market share liability will equal the risks created)), we have held that creation of risk or breach of a duty alone is not sufficient in imposing liability. (Board of Education v. A, C & S, Inc. (1989), 131 Ill. 2d 428, 442-43; Stallman v. Youngquist (1988), 125 Ill. 2d 267, 277; Note, Market Share Liability: A Plea for Legislative Alternatives, 1982 U. Ill. L. Rev. 1003, 1021 (adoption of market share liability may lead to assessing damages based only on wrongful conduct).) These principles should not be ignored merely because the defendants are members of the drug industry.
Abrogation of these concepts would also result in violating the principle that manufacturers are not insurers of their industry. In Woodill v. Parke Davis & Co. (1980), 79 Ill. 2d 26, this court held that in a strict liability action based on a failure to warn of a danger the plaintiff must allege and prove that defendant knew or should have known of the danger and this is tested on knowledge existing at the time of production. The holding was “justified because a logical limit must be placed on the scope of a manufacturer’s liability ***. To hold a manufacturer liable for failure to warn of a danger of which it would be impossible to know based on the present state of human knowledge would make the manufacturer the virtual insurer of the product, a position rejected by this court in Suvada. [Citation.] Strict liability is not the equivalent of absolute liability.” 79 Ill. 2d at 37; see also Coney v. J.L.G. Industries, Inc. (1983), 97 Ill. 2d 104, 111 (“imposition of strict liability was not meant to make the manufacturer an absolute insurer”); Suvada v. White Motor Co. (1965), 32 Ill. 2d 612, 623.
The market share liability theory disregards these precedents and turns manufacturers into insurers of their own products and products made by others in the industry. (Mulcahy v. Eli Lilly & Co., 386 N.W.2d at 76; Kroll, Intra-Industry Joint Liability: The Era of Absolute Products Liability, 687 Ins. L.J. 185, 194-97 (1980).) As illustrated in this case, the majority of plausible defendants have not been or cannot be brought before the court. Those who are present have the difficult burden of establishing their share of a market. The companies which cannot prove their share will be made to pay the unattributed portion of the damages, thus paying the damages which rightfully belong to companies which are insolvent, not amenable to suit in the jurisdiction or for some other reason are not before the court. Sindell justified its ruling in part on the belief that over the run of the cases a company’s liability would approximate the harm it caused. (Sindell, 26 Cal. 3d at 612, 607 P.2d at 937, 163 Cal. Rptr. at 145.) However, this is a purely illusory assumption, as recognized in Hymowitz, 73 N.Y.2d at 512, 539 N.E.2d at 1078, 541 N.Y.S.2d at 950. Justice Richardson, writing for the dissenters in Sindell, argued that market share liability makes the entire drug industry “an insurer of all injuries attributable to defective drugs of uncertain or unprovable origin, in-eluding those injuries manifesting themselves a generation later, and regardless of whether particular defendants had any part whatever in causing the claimed injury.” (Sindell, 26 Cal. 3d at 621, 607 P.2d at 942-43, 67 Cal. Rptr. at 150-51 (Richardson, J., dissenting).) We agree with his conclusion that such a solution is an unreasonable over-reaction in attempting to achieve what is perceived as a socially satisfying result.
The plaintiff contends that by not recognizing a market share liability theory we will be abdicating our responsibility in the development of Illinois common law. We have not in the past been hesitant to develop new tort concepts; however, in this instance we decline to do so because of the infirmities in the proposed theory. Furthermore, this is too great a deviation from a tort principle which we have found to serve a vital function in the law, causation in fact, especially when market share liability is a flawed concept and its application will likely be only to a narrow class of defendants.
Accordingly, we reverse the judgments of the appellate and circuit courts, and remand this cause to the circuit court of Cook County for further proceedings consistent with this opinion.
Reversed and remanded.