Google claims its driverless car could save almost 30,000 lives each year on U.S. highways and prevent nearly 2 million additional injuries. Leave aside the myriad of legal issues that might be posed by driverless cars and all the implementation challenges that might lie in the way of their becoming as ubiquitous as cell phones. Let's assume that we reach a point where Google's life-saving claims turn out to be true. In that case, should the Google Car be included on the list of Essential Health Benefitsthat must be covered by the Affordable Care Act?
This may seem like a ludicrous question, but why? The individual mandate requires adults to purchase "minimum essential coverage" costing thousands of dollars a year per individual. For example, a typical employer-based health plan, annual premiums just for employee (single) coverage alone were $5,615 last year and for family coverage such premiums were $15,745. Premiums in the non-group market are typically even higher than this. Admittedly, families may elect to instead just pay a penalty that by 2016 will reach $695 per uninsured person or 2.5 percent of family income (whichever is greater). But in principle, the government expects Americans to spend a sizable amount to secure minimum essential coverage and indeed is prepared to use taxpayer funds to subsidize a considerable portion of that premium burden for those at the lowest end of the income scale.
No one knows what a Google Car will cost, though with current technology, it apparently might cost as much as a Ferrari. But just as portable phones moved from being only affordable to the very rich 30 years ago to being nearly ubiquitous in the U.S., let us imagine that the price of Google Cars falls to within 33% of the average cost of a new car (purportedly about $30,000). Assuming a sale price of $40K, an average life expectancy of 10 years (which is actually slightly lower than the average age of the current U.S. fleet) and financing costs of 6 percent a year (surely on the high side of reality), the cost to put a Google Car into the hands of every adult in the U.S. would be $5,400 a year per adult. The actual cost obviously would be lower than this since not everyone drives and some households with multiple adults make do with one car etc. Averaged across the entire population (since children too would benefit from the life-saving potential of such cars), the annual cost would be about $4,100 apiece.
Put another way, it would cost over $1.2 trillion a year to ensure that every adult had a (new) Google car to drive. Would that be price worth paying to save 30,000 lives? That would be in excess of $41 million per life saved-a figure that might make even the EPA blanch (though not necessarily). It greatly exceeds the $8.3 million value of human life used by EPA in calculating the benefits of environmental regulations.
But, of course, no one buys a car purely for its life-saving potential. Since people buy $30,000 cars anyway for their transportation value, then this amount should be deducted from the cost of a Google Car to determine whether the car's extra costs are worth its extra benefits-i.e., life-saving potential. This would drop the cost per life saved to just over $10 million-still on the high side of what government regulators would view as reasonable. However, were the price of the Google Car to exceed that of the average car by only $8,300, then having Obamacare cover this incremental cost as a preventive health benefit is no more obviously unreasonable than imposing a cost on society of $8.3 million to save one life through regulation.
However, using the cost per life saved is a pretty crude way of estimating the value of lives saved from averting auto accidents since the median age of death from such accidents is far lower than for deaths in general. Teens who die in automobile accidents lose far more years of life expectancy than 80-somethings who die at the median age of death. Assuming years of life are equally valuable, we should attach a much higher dollar value to saving 67 years of life expectancy for a 16-year old than saving 7 years of life expectancy for an 86-year old.
A statistical life valued at $8.3 million is worth about $386,000 per added year of life. This is far higher than the threshold of less than $50,000 per added year of life used by the UK's National Institute for Clinical Excellence in determining what health services to recommend for coverage by the National Health System. It may strike some readers as absurdly high, yet that is the threshold being used today by U.S. regulators in determining whether the benefits of life-saving regulations exceed their costs.
If we tote up all the years of life expectancy due to auto accidents that could be averted if everyone drove Google Cars, it amounts to 668,000 lost life years in present value terms. Averaged over the entire U.S., this means that each year, the Google car would save .79 days of life expectancy per person. This may seem abysmally small, but if we use the $386,000 figure as the value of a statistical year of life, it means that society ought to be willing to pay $1,058 per person annually for a technology that increases (discounted) life expectancy by .79 days. So for a family of 4, having Obamacare subsidize $4,200 of the annual purchase costs of a Google Car would be functionally equivalent to imposing a new regulation that imposed costs of $8.3 million to save a statistical life.
I'm not seriously proposing that Obamacare pay for Google Cars, but I hope this exercise helps readers think more clearly about what Obamacare is doing. At the margin, the law is diverting resources into medical care that otherwise could be spent on other non-medical ways of saving lives. As it turns out, Google Cars may not be particularly cost-effective as a life-saving intervention: it all depends on their incremental cost relative to less safe cars with equivalent weight/performance etc. But there are large numbers of life-saving technologies, such as media campaigns to encourage seatbelt use, public pedestrian safety information and improving traffic safety information for schoolchildren in grades K-12 that can save years of life at a cost that is at least 10 times less expensive than the cost per added life year of most medical interventions. Thus, whether it occurs at the government level or household level, Obamacare may well be diverting resources away from the most cost-effective ways of saving lives. Not only is this a waste of resources, but it also means that Obamacare may well be costing lives. This is just one more reason we ought to think twice about plunging ahead with the implementation of this misguided law.
 Some readers may object that I've included the full premium, not just the share paid directly by the worker. However, most evidence suggests (and the Congressional Budget Office assumes the same whenever it estimates the impact of various health reform plans) that the employer share of premiums ultimately is paid by workers in the form of lower wages or fringe benefits than they otherwise would receive from that employer.
 Just under one quarter of the population is under age 18, so $5,400 x .755 = $4,077. Even if we gave Google Cars to everyone 16 and older, the per capita cost would rise only slightly, to $4,377, which does not materially affect my argument.
 Historically, the EPA has calculated benefits from life-saving regulations using a value of $7.4 million (2006 dollars), which would be about $8.3 million in 2012 dollars, used the GDP price deflator to make this adjustment. However, an analysis of the EPA's rules for hazardous waste listing for wood-preserving chemicals shows that the rule effectively imposes a cost in excess of $10 billion per life saved.
 The median age of death in the U.S. exceeds 80 years (Fig. 4a), i.e., half of deaths are below this age, and half above. The median age of automobile deaths apparently is under 45, calculated from this tabulation.
 These are the official life expectancy figures used by the IRS (Table 1).
 Most estimates of the value of a statistical life (VSL) are drawn from wage trade-off studies that estimate how much extra workers are paid to work in risky jobs. Based on the average remaining life expectancy of workers in these studies (35 years: footnote 2), we can use Excel's PMT function, PMT(0.03,35,8.3m.)=$0.386m. to calculate the annual payment over a period of 35 years at a long-term discount rate of 3 percent that would yield a present value of $8.3 million.
 This is based on data showing the distribution of auto accident deaths by age, using the midpoint of each age interval to estimate remaining years of life expectancy (Table 1). Just as a dollar today is worth more than a dollar in 10 years, these future years are discounted back to a present value using a long-term discount rate of 3 percent. The rationale for this discounting is explained in detail here. Discounting rather substantially reduces the nominal benefits of life-saving. That is, the 72.7 years of life expectancy for a 7 year old auto accident victim become only 29.4 years in present value terms.
 A systematic study of more than 500 life-saving interventions found that the median medical intervention cost $12,000 per added year of life, whereas all three public safety initiatives I list cost well under $1,000 per added year of life. See Appendix.