Over a century ago, workplace injuries that accompanied the new era of increasing industrialization were placing substantial pressures on the ability of the tort system to handle the challenge of compensating the victims of these injuries. Eventually, the interests of labor and management came together, and a new administrative compensation system, imposing strict financial responsibility on employers for work-related injuries to their employees, was substituted for the complicated and inadequate tort remedy that had been in force.

This system of workers’ compensation is still the most far-reaching tort reform ever adopted. The system both promotes safety and compensates for injuries more effectively than tort did at the time, or would do now. Workers’ compensation has its flaws, but there is no significant desire on anyone’s part to go back to tort.

A New Era of Auto Liability

We are on the verge of another new era, requiring another new legal regime. This time, it is our system of transportation that will be revolutionized.

In the coming years, manually-driven cars are going to be replaced by automated vehicles, which will become the dominant, though still not exclusive, mode of roadway transportation. Accidents attributable to driver negligence—and to a lesser extent, contributory negligence by other parties, such as wayward bicyclists or darting pedestrians—will become a minor safety (and accident) concern.

This new era of automated vehicles will eventually require a legal regime that properly fits the radically new world of auto accidents. The new regime should more effectively and more sensibly promote safety and provide compensation than the existing tort doctrines governing driver liability for negligence and manufacturer liability for product defects.

Like labor and management a century ago, auto manufacturers, consumers, and the public at-large—often currently at odds about the tort system—will need to have their interests come together if the new era of automated transportation is to be governed by an adequate legal regime.

Any new approach will:

  • have to deal with the long and uneven transition to automated technology;
  • impose substantial but appropriate financial responsibility for accidents on the manufacturers of highly automated vehicles; and
  • provide satisfactory compensation to the victims of auto accidents in the new era.

Should liability for driver negligence and manufacturer liability for product defects be retained in this new era? In our view, substantial considerations militate against continued reliance on negligence and products liability once a vehicle mix threshold has been crossed and the roadways contain a substantial proportion of highly automated vehicles (HAVs) that retain very limited driver control over the vehicle’s operation or eliminate driver control altogether.

Given the greatly heightened complexity and sophistication of the computerized control systems in highly-automated vehicles, judicial and jury assessment of the acceptable limits of engineering capability for alleged design defects in HAVs—through reliance on expert assessments of risk-utility analysis—will come to be needlessly contentious and costly. From a systemic perspective, once the contribution of third parties to HAV-related accident scenarios becomes virtually de minimis, contests over blameworthiness will be replaced by examination of esoteric alleged engineering failures that can best be regarded—both from the vantage points of administrative cost and administrative feasibility—as simply having arisen out of the operation of a motor vehicle. Hence, retaining tort would ignore the substantial efficiencies that could be achieved by eliminating hotly contested issues of reasonable technological expectations.

Indeed, at this point, the underlying premise of risk-utility analysis in the context of HAVs warrants rethinking. As largely uniform software becomes pervasive, the concept of a reasonable alternative design (RADs, under Products Liability Restatement terminology) is likely to become increasingly indeterminate. Moreover, the confounding effect of technological innovation will perhaps be most evident as HAV software increasingly incorporates machine learning—that is, the compilation of operational data on the roadways through trial-and-error leading to continuous revision and updating of algorithm-based driving instructions.

Judge and jury examination of these esoteric, computer-based design differences—which might, in addition, be subject to trade secret protection—would impose overwhelming stress on the premises of conventional analysis. Relatedly, as software-related failures become not just the prevailing, but the near-exclusive source of HAV-related injuries, a limitation on the consumer expectations test (an alternative to the risk-utility test) along conventional lines—that is, disavowing the test in cases of complex products—will not be a rational basis for determining liability.

Nonetheless, auto accidents, though at a much-reduced level, will continue to occur. The needless cost and anachronistic quality of design defect litigation will consequently call into question not only how responsibility for these accidents should be allocated, but also how compensation should be measured and awarded. In this regard, a reformulation of responsibility standards for HAV-related accidents would highlight for reconsideration the case-by-case “make whole” approach to tort compensation as well. We now turn to such a reformulation by detailing our proposal for the adoption of what we call “Manufacturer Enterprise Responsibility (MER).”

Reconfiguring Responsibility for Roadway Accidents in the New Era: Manufacturer Enterprise Responsibility

Once HAVs become a significant and frequently-adopted mode of transportation, retention of the fault-oriented standards of negligent driving and of liability for defective products would become outmoded or inconsequential as these standards pertain to HAVs. The current driver-focused liability system will become a thing of the past, and there will be very few occasions for drivers to be negligent, because there will be very little “driving” by people.

Auto manufacturers will still be making vehicles, however, and their vehicles will be the cause of most remaining accidents. But as mentioned above, the system of auto manufacturer liability for product-related accidents will become inordinately difficult to apply and anachronistic.

As we describe it below, our conception of a new approach—MER—is that it would be a manufacturer-financed, strict responsibility bodily-injury compensation system, administered by a Fund created through assessments levied on HAV manufacturers.

MER will necessitate a uniform system enacted at the federal level and applicable throughout the country. Auto manufacturers are national, indeed international businesses. Their HAVs will carry the same safety features and the same risk in whatever state they are driven, independent of where state boundaries are located. Variation in products liability standards and the unpredictability of their application from state to state, as much as the scope of potential liability, have been the major dissatisfaction of auto manufacturers with the current system. A single national approach would solve this problem.

An HAV should be defined as one that satisfies the definition of Level 4 and Level 5 vehicles under the Society of Automotive Engineers (SAE) classification system. These categories are for vehicles that incorporate “high automation” and “full automation,” respectively. Such vehicles may be available for sale, and potentially involved in accidents, in less than a decade.

The MER system should go into effect when 25 percent of all vehicles registered in the U.S. are HAVs. The 25 percent threshold that we propose is not set in stone. MER should be introduced as soon as the percentage of HAV vehicles is substantial. Still, as noted, roadway accidents by definition will at times involve HAVs exclusively; at times, HAVs and conventional vehicles (CVs); and of course, in some instances, CVs alone.

Hence, once HAVs are on the road, there will be two new types of accidents, in addition to accidents involving CVs only (for which the current liability regime should be retained): “pure” HAV accidents, whether involving one or multiple HAVs (and sometimes involving HAVs and pedestrians or other third parties); and “mixed” accidents involving a CV and an HAV (and sometimes pedestrians or other third parties as well). Our proposals for the legal treatment of these two new types of accidents, and for the use of MER in connection with them, follow.

Pure Cases: HAV-Only Accidents and the Scope of MER

HAV occupants and specified third parties involved in HAV-only accidents would be entitled to MER, provided automatically as one of the terms of sale of all HAVs. All bodily injury “arising out of the operation” of the vehicle would be covered, up to the specified benefit limits, except for injuries caused by the HAV owner’s own negligence. For example, MER would not cover injuries caused by the negligent failure to upload software updates, negligent tweaking of software, or negligent maintenance. Nor would MER cover negligent “drivers” for those limited tasks that they might still perform. For example, we envision the possibility that a SAE 4 vehicle could be manually driven a few feet in order to set it at a particular angle to a wall or to place it on a hydraulic lift for maintenance.

MER would be a victim’s exclusive remedy for the injuries to which it applied; there would be no tort cause of action for injuries caused by allegedly defective features of an HAV. Benefits would be provided for out-of-pocket losses (such as medical bills and lost wages), plus scheduled benefits for substantial pain and suffering.

1. Covered Individuals and Manufacturer Incentives

MER would provide compensation to the HAV’s occupants, as well as to pedestrians, bicyclists, motorcyclists, and other third-party bystanders, for bodily injuries “arising out of the operation” of an HAV. This would include injuries suffered in accidents involving two or more HAVs. When a third-party, not an occupant, is injured by convergence of two HAVs, to promote simplicity and low administrative costs, the third-party should be able to choose between HAVs and there should be no contribution. Statistically, this will end up being random for manufacturers and therefore should not be unfair in the aggregate. Less randomly, there may be injuries arising from safety breakdowns in the category of vehicle-to-vehicle communication (V2V communication). Here, too, we would opt for third-party choice to promote simplicity and low administrative costs.

It is also conceivable that on rare occasions an injured pedestrian or bicyclist will have been so substantially responsible for an accident that his or her injuries would not be considered to have “arisen out of the operation“ of an HAV and therefore would be non-compensable.

MER would not cover property damage, suffered either by an HAV owner or occupant, or by a third-party. In most instances, this would produce no hardship or unfairness. HAV owners still would likely purchase conventional auto insurance (sometimes called “collision” and “comprehensive” coverage) covering damage to their vehicles—for example, protection against fire and theft. HAV occupants, pedestrians, and bicyclists still would be able to purchase property insurance (usually under Homeowners or Renters policies) covering them against loss related to personal property. And owners of real property damaged on rare occasion by an HAV would have similar coverage (if, for example, an HAV damages a garage).

The manufacturer will be in the best position to decide what to invest in designing the system, by comparing (among other things) the cost of compensating losses for which it is responsible—through targeted contribution to the Fund, discussed below—with the cost of including features that will help to avoid additional accidents. Bearing financial responsibility for HAV losses will give the manufacturer an incentive to determine what to invest in research into ways of avoiding currently-unavoidable accidents. Because the manufacturer will bear the above-mentioned costs, the anticipated cost of HAV-related accidents will be reflected in the price of the vehicle. As a result, HAV manufacturers will internalize the cost of HAV accidents, the purchase price of HAV’s will reflect that cost, and an “excessive” number of HAV vehicles will not be on the road.

As to this last point, some observers have argued that the threat of even conventional products liability would impede the development of HAVs. They can be expected to argue that adoption of MER would generate an even greater disincentive to the development of HAVs. In our view, however, this effect is likely to be minimal. HAV technology promises transportation and safety advances of such magnitude, and profits that will inevitably accompany these advances, that the threat of financial responsibility for HAV-related losses is unlikely to deter research and development or the marketing of HAVs in any significant way. This is especially true given that the cost of funding MER is likely to be more predictable and steady than the threat of conventional products liability.

Virtually all studies suggest that the advent of HAVs will radically reduce the incidence of auto-related injuries, because such a high percentage of these accidents now result from driver error. For the residue of HAV accidents that still occur, the quid pro quo given to claimants in return for the elimination of their potential tort claims is the expansion of the right to compensation to include all injuries arising out of the operation of the vehicle, even if not caused by a defect in the HAV. MER would provide a swift and in most instances automatic source of full compensation for bodily injuries.

Accidents caused by defects in the non-autonomous features of an HAV are likely, at least over time, to arise less frequently than might be supposed. Presumably the incidence of defect-caused accidents would decline, because at least some such accidents, and perhaps most, would be avoided by the operation of the HAV. For example, features of the vehicle involving defective crashworthiness would much less frequently cause harm, since there would be fewer crashes. Similarly, operational mishaps resulting from malfunctions of component parts such as tire blowouts or the sudden failure of headlights or signaling capacity would often be handled more effectively and with fewer accidents by HAVs than they would have been handled by an active driver. This would reduce the number of conventional products liability actions that could have been brought after HAVs become dominant, even if such liability had been preserved.

The exclusive remedy approach would cleanly and effectively render the manufacturer the focal point for the creation of incentives that would otherwise have to be addressed through complex rules regarding component-part maker responsibilities for funding MER or for defects in the parts they supplied to the manufacturer. Thus, we see no reason to involve the makers of component parts in the MER system. Correspondingly, they should also be immune from conventional tort liability. The result will be a significant saving on transaction costs.

This does not mean, however, that makers of components parts will be relieved of the incentive to make safe products. These parties are now and will continue to be part of a chain of contracts that run through the HAV manufacturer. Contractual indemnities tailored to each individual situation, in anticipation of, and in reaction to, the manufacturer’s MER responsibility, will much more simply create safety incentives for component part makers than their involvement in the MER system. This is also true for ride-sharing enterprises, which are already prominent—and are widely predicted to become a dominant feature of roadway traffic in the driverless era.

Nor will there be any need to address what has come to be called the “trolley problem” in the philosophical and legal literature. This is the split-second choice that HAVs will have to be programmed to make: Whether and when to reduce safety to HAV occupants at the cost of endangering occupants of other HAVs or pedestrians. Should the vehicle collide with a utility pole or swerve to avoid it and strike pedestrians instead? HAV manufacturers will internalize the costs of whatever programming choices they make regarding such dilemmas, because the manufacturers will bear MER responsibility for the consequences of their choices.

In short, making MER the exclusive remedy in virtually all cases will minimize complications and uncertainty about liability, and it will optimize safety incentives by focusing financial responsibility on the manufacturers of HAVs, who will be in the best position to make the decisions that influence safety levels.

2. Benefit Levels

For these purposes, our working assumption is that unlimited medical expenses and up to $1 million in compensation for wage loss, indexed for inflation, should be available to each eligible MER claimant. Benefits would be paid periodically. In addition, up to $500,000 should be available, according to a schedule of noneconomic losses, for specified permanent or long-term injuries. The schedule should be modelled on, but substantially more generous than, the lump sums available under state workers’ compensation regimes for permanent disabilities. Finally, up to $1 million for wrongful death should be payable to the heir or heirs-at-law of a person killed in an HAV accident. This also should be paid pursuant to schedule that is a function of the heir or heirs’ age and relationship to the decedent.

In order to maximize manufacturers’ safety incentives, MER should have primary responsibility for medical expenses and lost wages. Other sources such as health insurance, sick leave, disability insurance, and workers’ compensation should have only secondary responsibility. An arrangement analogous to the difference between in-network and out-of-network health insurance might also be workable, under which the amount of cost-sharing by patients is greater under the latter than the former. These devices should help to manage MER medical costs, while still imposing responsibility for a very high percentage of those costs on HAV manufacturers.

Mixed Accidents

During the long period when both CVs and HAVs are on the road, a substantial percentage of all multiple-vehicle accidents are likely to be mixed. Here, two accident types should be distinguished: those in which a CV driver or passenger is injured and an HAV is also involved, and those in which an HAV occupant is injured and a CV or third-party is involved. Accidents between CV plaintiffs and bicycles, motorcycles, pedestrians, and other CVs would continue to be handled, as they are now, by current state tort law rules on negligence and products liability, and by auto no-fault statutes, if applicable. Similarly, accidents in which the responsible party’s conduct (including HAV manufacturers) was sufficiently egregious to warrant punitive damages would remain subject to applicable state tort rules on punitive damages recovery.

1. CV Plaintiffs

There are three possible approaches for handling injuries that a CV plaintiff incurs when an HAV is also involved:

  • (a) retention of the current rules governing tort liability;
  • (b) permitting CV plaintiffs access to the HAV owner’s MER; and
  • (c) enactment of mandatory first-party no-fault insurance covering such losses by CV plaintiffs.

Consistent with our desire to minimize the complications that would be involved in tort litigation in the new era, we think that either the second or third approach is to be preferred to the first.

2. HAV Plaintiffs

Should HAV occupants have a right to recover from CV drivers in tort for injuries caused by the latters’ negligence? We prefer the complete abolition of such negligence liability, for the same reasons that we referenced above. As much as possible during the long transition to the time when there will be only HAV vehicles on the road, we should be moving HAV accident victims into an exclusively MER system.

Funding, Administration, and Resolution of Disputes

During the initial phase of MER, manufacturers should be assessed a presumptive charge based on annual market share. Once there is statistically adequate data, however, assessments should be based on the frequency and severity of payouts for each manufacturer’s HAVs. This targeted experience-rating of assessments should maximize each manufacturer’s incentive to optimize the safety of its HAVs.

Although HAV owners will be relieved of most tort liability, for the foreseeable future they still will need to purchase auto insurance. As suggested earlier, pockets of liability exposure will remain; for example, for injuries resulting from negligent failure to maintain the vehicle, for shoddy repairs, and—to the extent that HAV technology makes limited manual operation possible—for any liability that may remain for aspects of the vehicle that still can be operated manually. In addition, auto owners still will be purchasing insurance covering property damage to their own vehicles, whether that property damage results from an accident or not.

Consequently, it will be a simple matter for each owner’s auto insurance policy to carry a statement or endorsement indicating the scope of MER that automatically accompanies the insured vehicle, and to give the policyholder notice that claims are to be submitted to the insurer. The insurer can then serve as the conduit for MER claims. This is the way that most federal flood insurance is currently purchased, and claims for coverage made—through a property owner’s own homeowners’ or commercial property insurer. The U.S. government is the flood insurance risk-bearer, but private insurers are its processing agents, for which insurers receive a commission. A similar approach is followed in California for earthquake insurance, where the state government is the actual risk-bearer but private insurers administer the system.

The MER Fund, created by assessments on HAV manufacturers, would operate the same way, although there would be no need for MER coverage to be purchased: it would have been incorporated into the purchase price of the vehicle and would run with the vehicle even after change of ownership. But the MER Fund created by assessments levied on HAV manufacturers would disburse claim payments through insurers, upon submission by the claimant’s auto insurer. Each insurer would receive a commission paid by the Fund for processing a claim, whether or not the claim was paid in full.

The great advantage of this approach is that owners of HAVs will have a pre-existing relationship with a particular insurer, will know with whom to file a claim, and will not be forced to file a claim and then deal with proof of claim through communications with a distant government bureaucracy. The vast majority of non-owner passengers will be family members of the HAV owner and therefore effectively have the same relationship with the owner’s insurer. Third-party claimants entitled to access to an HAV owner’s MER would communicate with the HAV owner at the initial stage, and would be given the owner’s insurance information, in the same manner in which drivers of CVs involved in accidents now exchange insurance information. At that point the claimant would deal directly with the HAV owner’s insurer.

As we envision it, an HAV’s auto insurer would be the agent for receipt of claims and submit them to the MER Fund, along with accompanying documentation. The process would be roughly analogous to making a workers’ compensation claim through an employer, to the employer’s insurer. The MER Fund would then pay in full, pay in part, or deny the claim. The Fund would be a small agency or division within a Cabinet Department (most likely the Department of Transportation), with personnel that would manage the process of assessing manufacturers, performing other administrative functions, receiving and evaluating claims, and disbursing claim payments.

If the Fund denied the claim or paid only in part, disputes would be resolved by an individual serving effectively as an administrative law judge (ALJ) in each state. The proceedings before the ALJ would, when possible, be summary and based on claim documentation alone. The ALJ would make de novo findings of fact and, if necessary, conclusions of law. Appeals from an ALJ decision to a federal district court could be reversed only if they were arbitrary or capricious or not in accord with existing law.


The current system of driver-focused liability will eventually disappear on its own. But before that disappearance, there will be a lengthy period of mixed accidents involving various combinations of CVs, HAVs, and third parties. For the present, it makes sense to begin addressing such mixed accidents with the doctrinal tools in tort that we have at hand. Over time, however, as HAVs become substantial contributors to roadway traffic, the deficiencies of this approach will become apparent. As that occurs, a legal regime will be needed that better fits the new world of accidents than our current negligence and product defect liability system. We have argued that MER, as we have outlined it, can do this job. Just as our forebears did a century ago when they proposed and then enacted workers’ compensation, we believe that a break with the tort system, dramatic and unsettling though it may be to some, will prove to be a desirable approach.

Author Information

Kenneth S. Abraham is David and Mary Harrison Distinguished Professor of Law at the University of Virginia School of Law.

Robert L. Rabin is A. Calder Mackay Professor of Law at Stanford Law School.

This article is based on a longer article to be published in Volume 105 of the Virginia Law Review in March 2019. A version of the full article is currently available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3159525.