Nothing brought the issue of older drivers into sharper focus than the 2003 crash at the Santa Monica Farmer’s Market, in which a 86-year-old man who had confused accelerator for brake killed nine people and injured scores more. As often happens in the media, something that had once basked outside of the light of attention suddenly became an “epidemic,” and untold numbers of stories warned us of the specter of the aging Baby Boom behind the wheel. This is, undoubtedly, a real issue, and the problems of the older driver do merit societal attention, but it is also likely that the circumstances of the Santa Monica event may have helped skew the actual risk posed: It was novel, it represented something out of our control, and, compared to most traffic fatalities at least, a large number of people were involved.
A new paper by Bryan Tefft, a researcher at the AAA Safety Foundation, published in the latest edition of the Journal for Safety Research, tries to put the older driver risk question into context, addressing some shortcomings of previous studies, most of which have not, as he notes, “analyzed responsibility for — as opposed to mere involvement in—crashes that kill other road users in relation to driver age, and none has done so while taking the amount of driving done by drivers of different ages into account.”
It is a truism of road safety research that a kind of u-shaped curve exists, in which the riskiest drivers are found at both ends of the age spectrum, as shown in the chart below, which comes from a Rand study (more on that later).
But this leaves out various, but important, parts of the risk equation, including: Which drivers (if any) bore a greater responsibility for the crash (an admittedly “noisy” bit of data), and the risk posed by certain classes of drivers to other drivers.
In any case, Tefft, using data from the FARS database, as well as “exposure” data from the National Household Travel Survey (and he notes using two different sources is a limitation of the work), comes to some interesting conclusions.
As other studies have found, he notes older drivers do have a greater risk of being involved in a fatal crash, but that this fatality risk is largely to themselves, as they are more likely, owing to increased “fragility,” to die in a crash than a younger driver. But as Tefft notes, “the degree to which older drivers’ risk to other road users is elevated depends strongly upon whether risk is measured on a per-driver, per trip, or per-mile basis.”
As an example, he writes, “if a randomly-selected driver in his or her thirties and a randomly selected driver aged 85 or older were to drive equal numbers of miles, the older driver would be over 1500% more likely than the younger driver to be responsible for and die as a result of a crash, and about 220% more likely than the younger driver to be responsible for a crash fatal to an occupant of another vehicle or a non-motorist.”
But, of course, most older drivers don’t drive as much as younger drivers, and they drive differently (i.e., they modulate risk based on their ability by choosing only certain roads, or certain times of day to drive, they may drive more slowly — insert Florida joke here — etc.). And so, while “a randomly-selected driver aged 85 or older is about 720% more likely than a randomly selected driver aged 30 to 39 to die in a crash, but only about 0.8% more likely to be responsible for a crash fatal to an occupant of another vehicle or a non-motorist, over the course of a year.” Per trip, the risk older drivers pose to others is “not statistically different” from drivers 30 to 39.
For the greatest source of risk from “without,” then, we need to look at the other end of the age spectrum. Tefft writes: “Drivers under age of 20 are responsible for more than twice as many deaths of occupants of other vehicles and non-motorists as are all drivers aged 70 and older.”
Tefft’s findings are supported by another paper — which uses another methodology (based on a technique created by Steven Levitt and J. Porter in this paper, which uses the drivers involved in fatal crashes as a surrogate for exposure) — namely, the Rand study mentioned above, “Regulating Older Drivers: Are New Policies Needed?,” by David S. Loughran, Seth A. Seabury, Laura Zakaras. They conclude: “In summary, we find that older drivers are only slightly likelier than other drivers to cause an accident but are considerably likelier to be killed in one. Younger drivers, on the other hand, are considerably likelier than other drivers to cause a crash, drive much more frequently than older drivers, and are less susceptible to fatal injuries than older drivers are.”
There are good reasons to be concerned about older drivers, but the news stories suggesting the greatest threat to our safety might come at the wrinkled hands of aging drivers seems somewhat misplaced. In strict terms of cost and benefit, it would seem wiser (assuming older drivers continue to do less driving than their younger peers), rather than rolling out new “mandatory retesting” programs in state DMVs (granted, one must leave open the possibilities these already existing programs have impacted the older driver crash rate), to ratchet up GDL programs at the other end — it really is shocking that driver’s licenses, in states like North Dakota, can still be had at age 14. Or simply cracking down on the most risky drivers, regardless of age, rather than blithely allowing people with clear patterns of dangerous driving to inhabit the roads (and, by the way, please don’t start on the sob story, ‘but it’s the U.S., you can’t function without a car…’). Demonizing older drivers may also subtly suggest to younger drivers, in their 30s say, that they have much less to worry about, when as Teftt’s per trip numbers indicate, may not be the case.