Product Liability and Innovation: Managing Risk in an Uncertain Environment (1994)

PETER W. HUBER

There seem to be limits to the total uncertainty that any undertaking can bear, or so business experience suggests. Sellers of old products with established markets can sometimes shoulder the risk that attends operation within an unpredictable and sometimes capricious regime of liability. It is the new venture with the unfamiliar product that can least tolerate the extra measure of instability from a legal environment that does not provide predictable results. Insurance provides an antidote, if it can be obtained at a reasonable price.

In the past three decades, U.S. liability law has changed in many ways that have undermined insurance, and thus innovation. That effect was unintended. The architects of the modern U.S. liability system fully expected that insurance would continue to provide a broad financial umbrella over the expanding new tort system. But changes in liability law were implemented with little understanding of how insurance markets operate. The result has been a sharp increase in demand for liability insurance, but a marked decline in supply.

The most dramatic adjustment came in the early 1980s, with an insurance crisis eerily reminiscent of the endless gas lines during the Arab oil embargo. The insurance industry goes through regular cycles of ''harder" and "softer" markets, and the market is currently softer (i.e., supply is greater) than it was some years ago. But the overall, long-term trends are not in serious dispute. Insuring high-litigation-risk products and services—obstetrics, contraceptives, light aircraft, vaccines, morning sickness drugs, and so on—is far more difficult today than it was 20 years ago.

A prudent insurer must know, first of all, just who it is insuring. It will

write one contract for a driver with a history of drinking at the wheel, another (at a quite different price) for a well-run municipality, and yet another for a car manufacturer. But U.S. courts have steadily expanded principles of "joint" liability, which sweep together dozens, sometimes hundreds, of defendants in a single courtroom. "Several" liability then allows the full costs of an accident to be channeled to the wealthiest, inevitably the best-insured, defendant.

A second essential ingredient of intelligent insurance is accurate timing. How much risk an insurer is covering depends not only on when the policy begins and ends, which the insurer itself can control, but also on when legal claims are born and expire. In pricing a policy, the insurer counts on earning some investment income between the time when premiums are collected and the time when claims are paid. A reasonably predictable and quick clock on insurance claims also allows the insurer to identify better and worse risks. It will not do to sell liability insurance at a flat price to all comers for 20 years, and only then discover that some in the group are careful and others (as finally judged by the tort system) are grossly careless. But U.S. courts have been steadily dismantling principles of ripeness and limitation that once kept litigation on a fairly strict and predictable timetable. Lawsuits today can look forward and backward for decades, or even generations, with no one knowing which of the dozens of policies a customer might purchase during that period will then be called into play.

An insurer also needs a reliable yardstick for pricing injury. A policy speaks of accidents and such, but the final accounting is always in cash. A stable rate of exchange between injuries and dollars is therefore essential. The conversion is not hard to make when the injury is a broken leg that must be set or lost wages that must be replaced, which is why you can easily buy first-party health or disability insurance to cover such contingencies. But the rate of exchange is quite indeterminate for pain and suffering, loss of society, or criminal fines and penalties. For precisely this reason, first-party insurance never covers such things. In U.S. courts, as elsewhere, it used to be that tangible economic losses, like medical costs and lost wages, always accounted for most damages ultimately paid. In recent decades, however, there has been great expansion in U.S. awards for such things as punishment, pain, suffering, cancer phobia, and loss of society.

A last and perhaps most essential ingredient of rational insurance is knowing just what risk is being covered. If an insurer sells a policy to a vaccine manufacturer, it must plan to cover the risks of vaccines, not the general health problems faced by young children. A policy priced to cover injuries caused by a spermicide cannot also cover birth defects originating

in quite independent genetic accidents or a mother's drinking habit. Here too, in U.S. courts as elsewhere, the legal rules used to be fairly accommodating: claims of cause and effect were usually tested skeptically, and the courts declined to speculate about causes remote in time and place. But U.S. courts have gradually come to accept ignorance, or at least substantial uncertainty, about the risks of such things as toxic wastes as sufficient reason for setting the judicial machinery in motion. The upshot has been waves of litigation over "phantom risks,"1 with marginal science peddled with impunity from the witness stand.

The legal disaster of the pertussis (whooping cough) vaccine unfolded in much the same way. The vaccine has virtually ended the 265,000 cases of pertussis and 7,500 pertussis-related deaths recorded in the years before 1949, the year the vaccine was first licensed. But a 1984 English study, serious and cautiously phrased in itself, suggested that the vaccine's use (extrapolated to the U.S. population) might be causing 25 cases a year of serious brain damage.2 American lawyers responded with an avalanche of litigation, blaming the vaccine for brain damage, unexplained coma, Reye's syndrome, epilepsy, sudden infant death, and countless other afflictions.3 Horrified pharmaceutical companies bailed out, and at one point it appeared that the last U.S. manufacturer of the product would be leaving the market.4 More solid scientific evidence slowly accumulated.5 Then, in March 1990, a report of a huge study of 230,000 children and 713,000 immunizations concluded that the vaccine had caused no serious neurological complications of any kind, and no deaths.6 "It is time for the myth of pertussis vaccine encephalopathy to end," declared an editorial in the Journal of the American Medical Association. "We need to end this national nonsense."7

As experience now richly demonstrates, the incentives for lawyers today are simple and compelling. If the consensus in the scientific community is that a hazard is real and substantial, the trial bar will trumpet that consensus to support demands for compensation and punishment. If the consensus is that the hazard is imaginary or trivial, the bar will brush it aside and dredge up experts from the fringe to swear otherwise. Even when lawyers pursue certifiably real hazards, there will be a strong incentive to stretch claims to the margins of validity and beyond, to reach not just dangerous IUDs but also safe ones, not just serious exposures to asbestos but also trivial ones. If the law allows a lawyer to put just about anybody on the witness stand, she is going to search far and wide for any expertise, real or otherwise, that is congenial to her case.

It is a deep irony that the one place the law tolerates this sort of thing is in court. The professional seated in the witness box, alone among all other obstetricians, engineers, chemists, or pharmacologists, is above the rules. Or, to be more precise, he is often not subject to any rules at all. Malpractice

by scientific and medical professionals is not only tolerated but encouraged, so long as it is solicited by lawyers themselves.

The law has always been ready enough to impose standards of competence on quacks outside the courtroom. Negligence law requires every doctor to "have and use the knowledge, skill and care ordinarily possessed and employed by members of the profession in good standing." If there are contending schools of thought in the profession, a malpractice defendant may be given the benefit of the doubt if he favors one school rather than another. But as a leading legal treatise hastens to add, this does not mean

The designer of a product is likewise expected to have used "state-of-the-art" materials and technology.9

Often the law demands even more. In a famous 1932 decision concerning a tugboat called The T. J. Hooper, which (like most other tugs of the day) was not equipped with a radio, Judge Learned Hand sonorously declared that "in most cases, reasonable prudence is in fact common prudence; but strictly it is never its measure."10 A court may thus require tugboat operators or doctors to surpass even accepted industry standards and consensus norms of the profession.11

From 1923 until the mid-1970s, the Frye rule made some attempt to hold expert witnesses to similar standards (Frye v. United States, 293 F. 1013 [D.C. Cir. 1923]). Certainly not to anything better than mainstream scientific norms, but the rule did at least refer to competent science as defined by the consensus views of a profession. Under Frye, the expert witness could report only learning that was "generally accepted" in his scientific discipline. Negligence, incompetence, irresponsibility, reckless disregard for professional standards, and every other variation on professional malpractice were as unacceptable on the witness stand as they were anywhere else.

One might suppose that this sort of symmetry would be a matter of fundamental fairness. If an obstetrician is to be judged guilty of malpractice, it will be on the say-so of some other doctor sitting in the witness box. A similar showdown between professionals decides every challenge to the design of a contraceptive or a Cuisinart. The jury must choose between yesterday's expert, who designed the morning-sickness drug or delivered the baby, and today's, who claims that the job was botched. Incredibly, many courts today enforce serious standards of professional competence only against defendants, not against their accusers.

The standards for medical witnesses are more biased still; the hermit clinician can usually testify to anything if he holds an M.D. and is willing to mumble some magic words about "reasonable medical certainty." Malpractice by mouth from the witness stand is thus not subject to any control at all. Any old résumé qualifies someone to be a witness, but only those who comply with the mainstream standards of professional medicine are good enough to escape liability.

The fringe theories and fanciful methods used to condemn experts in malpractice and product-design cases surface in other cases to explain disease. In the Alcolac litigation, for example, Zahalsky and Carrow, who testified for the plaintiff, based their chemical-AIDS conclusions on tests conducted with monoclonal antibodies.12 The antibodies, however, were not approved by the Food and Drug Administration for any diagnostic purposes at all. Used in a clinic as a basis for treatment, such methods would have been actionable malpractice.

Our pursuit of incompetence among scientific and medical professionals is now often led by incompetents from the fringes of those same professions. Courts too eager to chase after distant and mysterious causes willingly attend to far-out, pseudoscientific mystics. In our eagerness to suppress inept, irresponsible, or fraudulent practice everywhere else, we have embraced inept, irresponsible, or fraudulent practice on the witness stand.

The impacts are felt throughout the economy, but most heavily on products, procedures, and technologies that are new and unfamiliar. Burt Rutan, the pioneering designer of the Voyager, used to sell construction plans for novel airplanes to do-it-yourselfers. In 1985, fearful of the lawsuits that would follow if a home-built plane crashed, he took the plans off the market. Meanwhile, the U.S. general aviation industry, once the unquestioned world leader, has almost ceased operations because of liability problems. In 1977 small-plane manufacturers in the United States paid a total of $24 million in liability claims. By 1985 their pay out was $210 million. Companies like Beech, Cessna, and Piper sharply curtailed or completely suspended production; they quickly discovered that the new-model planes, carrying a 50 percent surcharge for liability insurance, could no longer compete with used planes already on the market. The market is moving steadily to manufacturers in Europe and South America.

For similar reasons, Monsanto decided in 1987 not to market a promising new filler and insulator made of calcium sodium metaphosphate. The material is almost certainly safer than asbestos, which it could help replace in brakes and gaskets. But safer is not good enough in today's climate of infectious litigation.

Liability fears have likewise caused the withdrawal of exotic drugs that the Food and Drug Administration deems safe and effective, including some for which no close substitute is known. Carl Djerassi, developer of the oral contraceptive, sees reform of the U.S. liability system as the most urgent priority in resurrecting the U.S. contraceptive industry.13

Some of the most regressive effects have been felt precisely where aggressive commercialization of new products is most urgently needed. U.S. liability today is highly, but often indiscriminately, contagious, which means that the introduction of new products is undercut the most in the markets already swept up in a hurricane of litigation—contraceptives, vaccines, obstetrical services, and light aircraft, for example. Yet it is often in such markets that new products are most urgently needed.

The commercialization of new products in the United States is thus discouraged on two sides. The availability of insurance in particular industries, and for particular products that have proved most attractive to litigants, has declined, and prices have increased, as the courts have adopted rules that discourage rational underwriting and systematic selection and differentiation of risks. At the same time, U.S. courts are effectively demanding that all goods and services come wrapped in a special-purpose insurance contract. The availability of insurance, most especially modestly priced insurance, depends largely on an accumulation of accident experience. That is something that established technologies always have and truly innovative ones never do. Insurance is easiest to find when a good has been used by many people for many years, so that the caprices of tort liability have been as far as possible washed out and the statistics of experience speak for themselves. Innovation necessarily starts without an established market, and so is often condemned to start without insurance as well. For the prudent businessperson, a start without insurance is often worse than no start at all.

How can a manufacturer commercialize a new but somewhat risky product absent affordable insurance, in a legal environment that always holds out the possibility for large verdicts, numerous lawsuits, and substantial expenses even in defending nonmeritorious claims?

One possibility, which was often suggested as a means for bringing the new French abortifacient RU-486 into the U.S. market, is to create a thinly capitalized, single-product firm, operate without insurance, siphon off profits quickly, and use the specter of immediate bankruptcy to deter litigation.14 One small-plane manufacturer has attempted to operate in this way, and the option appears to be increasingly attractive for other small start-up companies. This is not much of solution, however. Most new

products of any consequence require large, up-front investments to create a chain of distribution, publicize the product, and for such things as drugs, to shepherd the product through the very expensive and lengthy regulatory process.

A second possibility is to build a case-hardened litigation team and adopt a scorched-earth litigation policy to deter attack. Every large manufacturer does this to some extent, though the smart money seems to have concluded that this strategy is simply infeasible with certain classes of products used, for example, by pregnant women and young children. Tobacco companies are the only ones that have used this approach with almost complete success, and they have little in the way of new products to commercialize.

Yet another approach for the U.S. manufacturer is to establish a foreign base for the early stages of commercialization. This is by no means a complete solution, because it is becoming easier to pursue manufacturers back to U.S. shores, wherever products may be sold or used. Foreign manufacturers do have one important advantage. They can effectively limit their total liability for their U.S. operations by the cold-blooded expedient of keeping few of their assets in this country. Transnational recognition of judgment rules is still cumbersome enough to provide an effective shield against the wholesale export of U.S. liability verdicts to collect against the foreign-based assets of a foreign firm, no matter where its products may have been marketed.

In 1993 the U.S. Supreme Court took an important step in the right direction in ruling on the case of Daubert v. Merrell Dow Pharmaceuticals . The Court squarely held that district judges must play an important role as "gatekeepers," and admit only "reliable" and scientifically "relevant" evidence. The Court linked reliability directly to scientific validity, and stated that trial judges must make a preliminary determination that expert scientific testimony reflects "scientific knowledge." The two dissenting Justices reproached the majority for setting up judges in the role of ''amateur scientists."

To be sure, Daubert will not immediately put an end to bad science in federal court. The Supreme Court emphasized the trial courts' "flexibility" in screening scientific evidence and did not designate failure to meet any particular criterion as automatically dispositive. This will undoubtedly tempt some judges who are uncomfortable in the gatekeeping role to engage in only cursory screening of scientific evidence. Daubert is a pragmatic, not dogmatic decision—and understandably so, given the wide practical sweep of the rules of evidence.

That said, the overall impact of Daubert should be favorable. The Court listed all the things that should be considered in screening scientific evidence. The opinion orders trial judges to determine whether a theory or technique can be (or has been) tested, and is therefore, as Karl Popper put it, "falsifiable." Peer review, the Court added, is an important, though not dispositive, factor to weigh. The Court also directed judges' attention to determining the known or potential rate of error of the technique in question, as well as the existence and maintenance of standards controlling the technique's operation. Moreover, general acceptance of the methods and theories at issue bear on the inquiry: the Court noted that "[w]idespread acceptance can be an important factor in ruling particular evidence admissible."

Most judges applying these factors will dismiss junk science claims before they reach trial. The trend toward stricter scrutiny of scientific evidence began in the late-1980s; in the aftermath of Daubert it will accelerate.

The four factors enumerated in Daubert as relevant to the admissibility of scientific evidence are likely to gather the lion's share of judicial attention. Attorneys faced with junk science claims, however, should recognize that the opinion gives them some easily overlooked but potentially powerful weapons.

The Court, for example, stated that proposed expert scientific testimony must have a valid scientific connection to the issue before the court to be admissible. As the Court emphasized, scientific validity for one purpose is not necessarily scientific validity for other purposes. High-dose animal studies, for example, arguably are relevant for risk research and could therefore be admissible in the context of litigation over Environmental Protection Agency regulations. The same studies, however, have questionable relevance for actually predicting harm to humans from low-dose exposure, much less in establishing that the particular substance at issue was more probably than not the cause of human injury. Such studies should be excluded if they are being used to prove causation.

The Court's focus on scientific relevance makes it clear that it is not sufficient that a claimant have a causation theory and an expert who is willing to testify at trial on the general subject at hand. Rather, the claimant must have an expert who can present testimony that will provide scientific support for the theory of causation being advanced.

Given that ruling, and the Court's emphasis on district judges as "gatekeepers," Daubert should encourage courts to issue appropriate pretrial orders to screen scientific evidence. An example of such an order has been routinely used by Federal District Judge J. Foy Guin, Jr.:

Junk science is offered on both sides of the courtroom aisle. Good science sometimes favors plaintiffs, sometimes defendants: what Daubert stands for is the proposition that judges have a major role to play in distinguishing accurate, reliable science from untested speculation, transparent error, or outright fraud.

No one can say whether the courts' single-minded pursuit of scientific truth will yield more universal justice than other approaches. Universal justice, justice not only for the parties but also for society as a whole, is, after all, a difficult thing to measure. But as the majority declares at the end of Daubert, judges, unlike scientists,

How else might policy be changed to make liability a less serious obstacle to the commercialization of new products? One approach is to increase the supply of insurance. The other, to reduce the need for it.

Government has granted immunity, indemnification guarantees, or alternative insurance coverage to providers of various niche products and services, ranging from childhood vaccines to nuclear weapons. Similar approaches have been considered for "orphan drugs," which may be kept from market because the costs of insurance may exceed any potential profit. In other areas, however, immunities from liability have been cut back rather than expanded. Charities and the government itself used to enjoy blanket immunities from liability, but no longer do. In waiving its own absolute immunity from suit, however, the U.S. government carefully insisted on trial by judge, not jury, forbade punitive damages, and denied liability for any "discretionary function," the government equivalent of the judgment calls that a private company routinely makes in designing a new product.

More recently, and much less fruitfully, have been attempts to expand supply or curtail the cost of insurance by direct legislative decree. When

insurers have canceled policies of their highest-risk customers, many states have passed laws limiting the insurer's freedom to cancel or refuse renewal. When coverage for particular lines has become entirely unavailable, some states have created joint underwriting associations, which force all private insurers doing any business in the state to offer coverage collectively, through the pool, for the otherwise uninsurable. When insurance prices have then risen still faster, several other states have turned to price controls. When insurers have attempted to flee, a few states have even challenged their right to depart. It seems likely, however, that such efforts will quite quickly reduce, rather than increase, the supply of insurance.

The second possible approach is to curtail demand, the need for insurance. This means changing the legal rules, so that cases are less frequent and judgments more modest. One of the blunter and more controversial approaches is to establish dollar caps on awards of such things as pain and suffering, or to limit punitive awards to some fixed multiple of the compensatory judgment. Many states have placed limits of one sort or another on joint liability, to reduce the increasingly magnetic properties of insurance, which otherwise cause the best insured to attract the most lawsuits. In addition to the 1993 Daubert ruling, there have been various other efforts to improve the quality of scientific testimony in the courts, so that factually unfounded claims cannot be used for their speculative value before a jury that may not understand the science in question. Other sensible guidelines for how the law can be changed to limit the chilling effects of open-ended liability can be found in the Federal Tort Claims Act, and in modern U.S. libel law.17

In the end, the search must be for rules that allow society to say yes to new and better products, with the same conviction and force that an open-ended liability system can say no to old and inferior ones. In many areas of policy, the answers given depend largely on the question asked. For several decades, U.S. policymakers in the courts and elsewhere, have asked: What is unduly risky? And how can risk be deterred? But an equally important pair of questions is: What is acceptably safe? And how can safety be embraced? What indeed are our legal vehicles for getting to yes?

The first is private contract, which restores to individual buyers and sellers the power to make binding deals—deals addressing safety and risk along with other matters of risk—that will then be enforced, as written, in the courts. Current legal policy generally refuses to enforce such contracts, at least insofar as they address safety matters, on the assumption that consumers have unequal bargaining power, inadequate information, or simply too little patience to consider these matters.

But such rules could be changed. When we deal with the essentially private risks, of transportation or recreation for example, fair warning and conscious choice by the consumer must be made to count for much more than it does today. Not because individual choices will always be wise—they surely won't be—but because such a system at least permits positive choice and the acceptance of change.

A second vehicle for getting to yes is through the positive choice made by surrogates. Some safety matters will always remain too complex or far-reaching to be collapsed into the world of purely private agreement. Informed consent by the individual is never going to take care of such things as chemical waste disposal, mass vaccination, or central power generation. These are and obviously must remain matters to be delegated to expert agencies acting for the collective good. But if they are to be useful at all, agents must be able to buy as well as to sell. For safety agencies, this means not just rejecting bad safety choices, but also embracing good ones. Yet the long-standing rule, still rigidly applied in U.S. courts, is that even the most complete conformity to applicable regulations is no shield against liability.

The courts should be strongly encouraged to respect the risk and safety choices made by expert agencies. It is politically unrealistic to propose that liability be cut off entirely for any activity expressly approved by a qualified regulatory agency. But liability could at least be firmly curtailed in such circumstances. Some states, like New Jersey, have already adopted reform along these lines for pharmaceuticals and other products comprehensively regulated by federal agencies. At the very minimum, complete compliance with a comprehensive licensing order should provide liability protection against punitive damages—awards that have a strong, depressive effect on innovative experiment and change. It has always been true that ignorance of the law is no excuse. At present, knowledge of the law is no excuse either. It should be.

We may end, then, where we began, with the issue of insurance. Insurance remains the best and most robust market instrument for making the random predictable. It is possible to insure against misadventure in the courts, just as it is possible to insure against wind and wave on the high seas. But only up to a point. When the legal rules are such that insurance itself attracts litigation, when the rules change too quickly for actuarial assessment, and when systematic biases creep in against new technologies and innovative undertakings, insurance will rise in price and narrow in coverage. No other country in the world has gone through liability insurance shocks like those experienced in the United States. The reason is that

no other country has transformed its liability system so completely and abruptly. It is within the direct power of U.S. courts to restore a climate that is healthy for private insurance markets. The courts thus have a very direct role to play in shaping policies to renew the spirit of enterprise in the United States, and recreate a social climate that welcomes, indeed encourages, the commercialization of new products.

This paper is adapted from P. W. Huber, Galileo's Revenge: Junk Science in the Courtroom (Basic Books 1991); Phantom Risk: Scientific Inference and the Law (K. R. Foster, D. Bernstein, P. W. Huber, eds. MIT Press 1993); and D. Bernstein and P. Huber, 'Daubert' Plaintiffs Won Technical Battle But Plainly Lost the War, 21 Product Safety & Liability Reporter 16 (BNA Special Report Summer/Fall 1993).

For the views of a plaintiff's lawyer who disagrees with me on virtually everything, see Chesebro, Galileo's Retort: Peter Huber's Junk Scholarship , The 42 Am. Univ. L. Rev. 1637 (Summer 1993).