In 1984, Justice Patricia Boyle of the Michigan Supreme Court wrote, "When the societal goal of holding manufacturers accountable for the safety of their products has been threatened by the interpretation of technical rules of law, it has been the rules that have gradually given way." [42]

Michigan has been more honest than most states in announcing this policy. We have already seen how rules relating to privity and warranties were knocked down in the 1960s. Finally, in 1970, Judge Levin, writing for the Court of Appeals in Cova v. Harley Davidson Motor Co., urged Michigan to quit playing with legal fictions, or trying to force opinions allowing recovery into modes consistent with prior judicial doctrine, and to instead recognize a theory of "products liability" which would include elements of duty and negligence, warranty, and ultrahazardous activity. [43] Wrote Levin, "The ‘product liability’ of the manufacturer... is simply the liability which... the law imposes on the manufacturer... for the loss suffered by reason of a defective product attributable to that manufacturer." [44] It is fair to say that this philosophy, though not then explicitly endorsed as Michigan law by the State Supreme Court, was the standard by which Michigan courts operated throughout the 1970s, until 1978, when Levin’s suggestions were enacted into statutory law by the state legislature. [45]

The fundamental premise of enterprise liability, as the doctrine proclaimed by Traynor in Escola is widely known, is that causation is a hopelessly tangled web which is beyond the court’s ability to determine. Therefore, the basis for deciding a tort claim is the ability of the defendant to pay damages. Michigan courts claim to leave steadfastly refused to adopt Traynor’s form of "strict liability" [46] However, the risk-spreading rationale of enterprise liability has nonetheless had its effect.

Much of the movement to eliminate archaic legal rules was legitimate and beneficial. But for all the problems they sometimes caused, rules such as privity and contractual concepts of warranty did serve valuable purposes. The emphasis on striking down barriers to recovery has inevitably lead to a breakdown of all restrictions on recovery, including the fundamental element of causation.

As we examine just a handful of Michigan cases decided under the new rationale, the reader will note the extent to which causation and contract leave been rendered meaningless before this judicial onslaught.

In Rutherford v. Chrysler Motors Corp., [47] the court determined a manufacturer had a duty to build a "crashworthy" car. Thus when the plaintiff’s car skidded on ice and smashed into a bridge abutment at a speed of approximately 35 miles per hour, through no malfunction of any kind on the part of the car, the plaintiff was still allowed recovery on the grounds that the manufacturer was negligent in not making the car safer.

This use of an indefinite standard, "crashworthy," creates a situation in which an insurer or self-insurer cannot gauge the risk created, because the extent of the manufacturer’s duty, and what constitutes his negligence, is unpredictable. One would expect, perhaps, that the duty would be less when manufacturing a small car, which most everyone knows is more dangerous in a crash than a full size auto. Yet if the duty is to build a crashworthy car, why should the standard ever vary? Further, what are we to make of competing desires, on the part of both auto consumers and the public at large, for more fuel efficient, and thus typically smaller, less crashworthy, cars? For less expensive cars? Must every car be as safe, and expensive, as a Rolls Royce? If the test is simply to become one of balancing competing goals, why should the plaintiff be allowed to substitute the ex post facto judgment of a jury regarding the appropriate balance for his own ex ante judgment of the proper mix of safety, fuel efficiency, cost and style at the time of purchase? Once again, the voluntary world of contract, with its maximization of choice, has been replaced by the coercive world of tort, where one style fits all.

From a risk selection standpoint, the idea of "crashworthy" cars begs for adverse selection. The manufacturer, expected to absorb this cost by the court, must charge all buyers of a particular model of car the same premium. When applied to sports cars or off-road vehicles, buyers who drive in a high risk fashion – at high speeds, or bouncing through off-road terrain, will continue to buy such cars. Those who drive more safely, but simply prefer the feel of a sports car or the occasional usefulness of four wheel drive, may find indulging their preference is not worth the added cost. Thus only riskier drivers are left in the risk pool, and insurance costs rise still higher.

Where general passenger sedans are concerned, the higher insurance premium that must be added to inherently less "crashworthy" small, fuel efficient cars narrows the price differential between small and large cars. Some lower income buyers will be forced out of the new car market and into less safe used cars. Higher income buyers may decide to spend a bit more for a larger, but less fuel efficient car. The tort system is unable to evaluate such trade-offs.

A legal question similar to Rutherford was posed in the 1980 case of Chaney v. Whiting Corp., [48] The plaintiff was burned by an explosion of molten metal released from the defendant’s properly functioning machine. The machine operator had failed to sound a warning alarm on the machine before release, and the plaintiff claimed that the machine should have been designed with an interlock system that would have made the release impossible while persons were on the opposite side of the firewall from the operator. The court ruled that because it was foreseeable that the operator would forget to sound the warning alarm, the manufacturer was negligently liable. The loose causation definition and the abolition of privity combine to relieve the employer and the machine operator of responsibility for their acts, leaving the, manufacturer liable, even though it was the employer who controlled the machinery, the workplace environment, and the training of the operator.

From a risk standpoint, Chaney asks the insurer not only to calculate the inherent risk of the product, but the odds that others will behave irresponsibly. If the manufacturer can be held liable for the unsafe behavior of others, it may be impossible to predict the risk and find insurance for the product.

In Koski v. Automatic Heating Service, [49] the defendant manufacturer was held liable even though the disease caused by fungal growth on his air conditioner was unknown to scientists or doctors at the time of manufacture. The court felt that this was nonetheless a foreseeable risk, thus creating a duty, negligently breached by the manufacturer, to protect the plaintiff. Of course, the defendant could not protect against an unknowable harm. But the court reasoned that the defendant should have known that fungal growth might spread some disease, and should have found a way to prevent it. Imagine the perplexed look on the face of the underwriter asked to make an insurance bid on that risk.

The gist of this new thinking on causation in products liability is seen in the case of Moning v. Alfono, [50] in which the manufacturer was held liable for marketing a slingshot which could be bought by children. Plaintiff sued when another child hit his son with a rock hurled from one of the defendant’s slingshots. In this case, no one, using common language, would claim that the manufacturer hit or injured the plaintiff. Still, one might find causation if it were determined that the plaintiff, by marketing the product to children, had created an inherently dangerous situation in which otherwise harmless activity might cause injury.

But the Court did not address the issue in this fashion. Instead, it delivered a rambling opinion attempting to weigh the relative social costs and benefits of slingshots. It is not entirely clear what this issue, important as it may be, has to do with causation of the plaintiff’s injury. If a manufacturer can be found liable for an injury it did not cause, based on a court’s ex post facto decision that society – though not necessarily the plaintiff – doesn’t benefit from the defendant’s legal activity, is there any method by which an insurer can predict future liabilities with any degree of certainty?

Yet this thinking, based not on causation but on an analysis, with the benefit of hindsight, as to who best could have prevented the accident – not the same as who caused it – is the law of Michigan.

In reaffirming the doctrine of "products liability" in Prentiss v. Yale Manufacturing, Co., the court noted, "The term ‘defect’ in design cases is ‘an epithet – an expression for the legal conclusion rather than a test for reaching that conclusion." [51] What the court is saying is that products liability is not a prospective standard under which insurers and self-insured manufacturers can evaluate risk with confidence, but a retroactive determination that a plaintiff should be compensated by a defendant.

Despite the Prentiss court’s declaration that, "Courts have never made manufacturers insurers", [52] this is precisely what was urged in Traynor’s Escola concurrence and what has occurred in Michigan law. For example, in Bleeda v. Hickman-Williams & Co., the Court of Appeals found liability, "...because [the manufacturer] is better able to absorb [costs], and to distribute them, through prices, rates, or liability insurance, to the public, and so to shift them to ... the community at large." [53]

It is difficult to imagine what rationale, other than insurance, could have been lurking in the mind of the court in cases such as Moning v. Alfono or Koski. The court rationalized its Koski decision by noting that the air conditioner was placed where it was more likely to grow fungi, and thus it was foreseeable that a disease then unknown to modern medicine might result. But in that respect, it is always "foreseeable" that a product may contain an "unforeseeable" defect, regardless of where installed. Liability on such grounds can only be called insurance.

Then there is the very difficult case of Abel v Eli Lilly Co.. [54] Eli Lilly was one of the myriad of DES cases flooding U.S. courts in the early part of this decade. DES was a drug used to prevent miscarriages in the immediate post-World War II period. It was eventually found to contribute to vaginal and cervical cancer in the daughters of women who had taken the drug. Because plaintiffs were unable to determine from which of the many DES manufacturers they had obtained the drug decades earlier, proving causation on the part of any particular defendant was problematic. Further, as in Koski, there was no sign of negligence, or even that the problems resulting from DES were in any way medically foreseeable at the time of manufacture and sale.

The Court’s resolution was to change the burden of proof in this one case. Whereas traditionally the accuser must prove his charge, the Court held here that the DES manufacturers would be held liable unless they could prove that their product had not harmed the plaintiff. The Court did so with the knowledge that such proof was unobtainable. This creation of a "DES-unique" theory [55] to allow recovery can only be justified as an insurance measure, one that, "raises the possibility that the ‘DES-unique’ theory of alternative liability may be extended to create other product-unique solutions to class action product liability claims." [56]

Such a development would be the natural evolution of a system that is based on how to compensate rather than whether to compensate. If the Prentiss court meant to say that Michigan has rejected the enterprise liability theories of Fleming James and Justice Traynor, the case law doesn’t show it.