A recent news article entitled “Big Cars Kill” caught my attention with some startling conclusions from recent research. While driving a large vehicle like an SUV is usually safer for those inside the vehicle than a small vehicle would be, it has significant negative safety effects on others:
A study from the University of California, San Diego … found every life saved in a large vehicle came at the expense of 4.3 dead pedestrians, motorcyclists and car drivers.
In essence, of all the consumer choices Canadians will make in their life, buying an unnecessarily large car is the one most likely to maim or kill a stranger. [1]
The basic idea is that while it is usually “safer” to drive a large vehicle, at least from the point of view of passengers inside the vehicle, this increased safety is bought at the price of imposing a disproportionately large increased risk of death or injury on non-passengers of the large vehicle, ie. passengers of other cars, bicyclists, and pedestrians. This is intuitive: in a crash between a large truck and a small car or bicycle, I think we can all guess who is more likely to suffer serious injury or death. As the article points out, the result is a kind of vehicle arms race, with consumers buying larger and larger vehicles to make themselves safer, even at the cost of making society as a whole less safe. Assuming the study’s conclusions to be correct, they raise some interesting questions as to how the law should respond.
As pointed out to me by my friend Leonid Sirota (of Double Aspect blog fame), this is a classic case of an economic externality. Consumers obtain a benefit (decreased risk of injury or death) by imposing a cost on third parties (increased risk of injury or death) without their consent. How might the law work to internalize this externality? One answer might be through tort liability and the market for compulsory third party liability insurance for motor vehicles. If certain types of vehicles (like, say, Hummers) are more likely to cause significant injury or death to third parties than other vehicles (e.g. Smart Cars), and if the tort system allows the victims to recover against drivers’ insurance plans, then one would think that the insurance market would place a higher price on liability insurance for the more “dangerous” vehicles, in recognition of the increased chance that the insurance company would have to pay a damages award. This would tend to deter consumers from buying a large vehicle unless they were willing to pay extra for the risk that their vehicles impose on other drivers.
I have no idea whether car insurance markets (or government insurers in some provinces) price liability insurance in such a way as to internalize the risks posed by large vehicles. But my guess is that they do not fully internalize these risks. The most obvious reason for this (and I can think of a few) is that the most serious accidents involving large vehicles result in fatalities, and the law places strict limits on what kinds of damages can be awarded for a fatality.
The traditional position at common law was that accidents that resulted in the death of the injured party were effectively non-compensable. The mindset was that since the victim of the wrong was no longer around to bring a claim, the claim could not proceed. This position has been altered by statute in Canadian common law jurisdictions to allow for actions by the dependents of a deceased person, but only for limited heads of damages. In Alberta, damages are available to dependents for loss incurred by dependents (like lost income), as well as funeral costs, costs of care for the dying person and other out-of-pocket expenses leading up to his/her death, costs of grief counselling for survivors, as well as predetermined sums for grief and loss of guidance, care and companionship.[2] In most Canadian common law jurisdictions, the estate of a dead person is not able to sue for lost expectation of life.[3] The amounts likely to be awarded under the allowable heads of damages therefore do not take into account the value of the life in question to the person whose life is lost, and thus almost certainly underestimate the negative effects of a fatality. Since the tort system thus implicitly values the effect of death to the person who dies at zero dollars, the corresponding liability insurance market for vehicles will also tend not to take this effect into account in pricing insurance. The result is likely an incompletely internalized externality.
Assuming that the value of a human life to the person whose life is lost is greater than zero, what is the correct way to place a monetary value on a life so as to internalize the risk of hazardous activities like driving a large vehicle? As one might assume, this is a controversial and difficult topic. At first blush, it seems impossible to assign any monetary value at all to a human life. As observed by Richard Posner, if one were simply to ask people what how much they would accept in return for dying, the answer would often be an infinite amount of money .[4] This is not a helpful answer in a world where practically every conceivable activity entails some risk of injury or death, and decisions have to be made about what activities are worthwhile, and with which precautions.
However, if one instead approaches the issue from the point of view of risk allocation (which is what the liability insurance market is all about, after all), it is possible to come up with a reasonable estimate for the value people place on their lives when making decisions about risk. Using data on things like wage premiums for dangerous occupations, economists have attempted to estimate the statistical value of a human life in the context of decision-making about risk. The going estimate based on labour markets in the United States is between $4 and 9 million per statistical life.[5] If we take the midpoint of $6.5 million, this means that, at the margin, a worker in the US labour market is willing to accept a 0.01% increased chance of being killed in the workplace in exchange for $650 ($6,500,000 * 0.0001).
In labour markets for hazardous occupations, employers have to pay workers in order to get them to accept an increased risk of death. But when I buy a large vehicle and impose a risk of death on everyone in my community, such a market-based solution is not possible. I can’t negotiate with each person in my city or province to arrive at an amount that each is willing to accept in exchange for the miniscule risk that my decision imposes on each person individually. In instances like this, we might normally rely on the tort system (along with liability insurance markets) as a way of making people pay for the risks they impose on others. If you hurt a stranger through negligent conduct, you can be made to pay for it. But as we have seen, the tort system as presently structured does not require people to pay for the most significant cost associated with a fatality – the lost life itself. Should it? It is not clear to me that it should, because doing so would seem to violate the basic structure of tort law as we know it.
Tort law as we normally conceive of it is about a bilateral relationship between the doer and the sufferer of harm.[6] The doer of harm is made to pay for the wrong, thus rebalancing the scales of justice. A case in which the sufferer of harm is dead falls outside this basic structure. The sufferer is no longer around to receive a benefit as compensation for the harm suffered. It seems wrong for a family member to be able to bring an action for the value of the life lost (as opposed to funeral expenses, loss of companionship, etc.), because it is not the family member who actually lost the life in question. A $6.5 million award in recognition of the statistical value of the lost life, for example, would basically be a windfall for the surviving family member. That said, maybe if a deceased person’s estate brings the action for the direct effects of the fatal accident, in the name of the deceased person, some of the conceptual difficulties are overcome. Apparently the Northwest Territories, unlike other Canadian jurisdictions, actually does allow a deceased person’s estate to sue for lost expectation of life.[7] This seems a bit bizarre according to normal tort principles, since the sufferer of harm doesn’t ultimately get the compensation, but maybe this is justified in terms of some of the efficiency considerations set out in this blog post, along with the fact that the “windfall” received by the deceased person’s heirs may at least reflect the deceased person’s wishes as set out in a will.
Perhaps, rather than relying on actions brought by family members, one could instead imagine a new kind of system that required those who negligently cause fatal auto accidents to pay into a government fund. The amount could be based on some statistical estimate of the value of a human life. Normally, it would end up being one’s liability insurer that paid the amount into the fund. If the required sum were in the neighbourhood of $6.5 million per fatality, car insurance would likely cost a lot more, but it would also vary more widely based on the likelihood that one’s vehicle would cause a fatality. Liability insurance for larger vehicles would likely end up costing a lot more than for smaller vehicles, thus forcing people to pay more for the increased risk their vehicles impose on others. This might put a stop to a vehicle arms race that makes us all less safe, and force people to think harder about whether they really need a large vehicle.
Whatever the virtues of a system like the one I have just sketched out (and I think it may have some, though I’m not sure), one should bear in mind it would be structurally nothing like the tort system. It would not be based on a correlative relationship between the doer and sufferer of harm, but rather on a relationship between the government and citizens. It would be public law, then, not private law.
One last point: All of this may soon turn out to be moot if autonomous and networked (ie. driverless) cars become a reality, as seems likely in the near future. In addition to reducing the chances of serious accidents occurring, potentially almost to zero, such vehicles could (and probably should) be equipped with software that values all human lives equally, whether or not they belong to passengers of the vehicle in question. In other words, in the event of a possible crash, the software of all vehicles would act to save as many lives as possible, regardless of whether one is riding in a big SUV or a compact car. In a potential crash between a large vehicle with one passenger and a compact vehicle with a whole family on board, the software might opt to sacrifice the life of the lone passenger in the large vehicle, for instance by driving it off the road, rather than risk a collision with the compact car that could result in the death of the entire family. If autonomous vehicles become ubiquitous (or compulsory), it might not matter so much from a safety perspective whether one has a large or small vehicle, though the prospect of driving a car that might kill you in order to serve the greater good raises its own ethical questions, as pointed out to me on Twitter by robot law expert Kristen Thomasen.
[1] Tristin Hopper,“Big cars kill: ‘Monster’ vehicles may make Canadians feel safer, but they’re more likely to cause fatal collisions”, National Post (31 July 2015) http://news.nationalpost.com/news/canada/larger-vehicles-may-make-canadians-feel-safe-on-the-road-but-heavier-cars-are-proven-to-cause-more-fatal-collisions.
[2] Fatal Accidents Act, RSA 2000, c F-8, ss. 3(1), 7, 8.
[3] See, e.g., Trustee Act, RSO 1990, c T.23, s 38(1).
[4] Richard Posner, Economic Analysis of Law, 8th ed (New York: Aspen, 2011) at 250.
[5] W Kip Viscusi & Joseph E Aldy, “The Value of a Statistical Life: A Critical Review of Market Estimates throughout the World” (2003) 27 J Risk & Uncertainty 5 at 6.
[6] See Ernest Weinrib, “Corrective Justice in a Nutshell” (2002) 52 UTLJ 349.
[7] Trustee Act, RSNWT 1988, c T-8, s. 31(1), cited in Robert M Solomon et al, Cases and Materials in the Law of Torts, 8th ed (Toronto: Carswell, 2011) at 674.
