Road safety, health inequity and the imminence of autonomous vehicles
Although personal vehicle ownership facilitates economic mobility, negative externalities persist. In the United States, motor vehicle (MV) fatalities represent almost four out of every ten ‘unintentional injury’ deaths — deaths that do not result from old age, disease, homicide or suicide. Globally, MV injuries are estimated to be the eighth leading cause of death for all age groups and the leading cause of death for children and young people 5–29 years of age.
Autonomous vehicles (AV) technologies promise relief. By shifting responsibility for higher-order control functions from humans to machines, these systems are envisioned to reduce road fatalities by eliminating driver-related errors blamed in most automobile crashes. From the vantage point of public health, the overarching goal of AV technology is to transform the current approach to automotive safety from injury reduction post collision to complete collision prevention. Doing so would — by some accounts — rank among the most transformative public-health initiatives in human history.
The ensuing benefits are particularly significant for low-income households as MV fatalities are disproportionately concentrated among individuals with low socioeconomic standing. Lowerand middle-income countries are overrepresented in MV fatality statistics relative to the number of vehicles registered within their respective borders. Moreover, the ‘death disparity’ between high and low socioeconomic groups has — over the decades — widened.
The reasons for this disparity are multifaceted. Inadequate road infrastructure and a dearth of local trauma centres in low-income neighbourhoods contribute to increased MV fatality risk for local populations. Particularly relevant to AV discourse are vehicular choice constraints imposed by low socioeconomic standing. This standing compels procurement of older vehicles; autos that have lower crash-test ratings and often lack advanced safety systems (for example, adaptive cruise control, automatic warnings and rear-facing cameras) that offer more protection against human errors implicated in most MV crashes.
Because low-income households disproportionally incur the public-health costs associated with MV use, they also stand to disproportional gain from AV’s envisioned safety benefits. Yet realizing this gain depends on adoption costs. The financial proposition offered by switching to AVs must be as attractive (if not more) than the status quo. The alternative risks pricing out groups that would benefit the most from AVs’ life-saving benefits.
AV developers typically cite three pathways towards realizing an attractive cost proposition. We discuss them below and explore impediments that may inhibit their effectiveness.
Economies of scale
Industrial learning curves predict fractional decreases in cost proportional to fractional increases in production volume. As volume surges, production becomes more efficient and consequently — on a per unit basis — cheaper. Such reasoning has proven valid across a myriad of industries. The first generation of cellular phones, released in 1983, cost US$3,995, which is over ten times more, when adjusted for inflation, than what consumers pay for comparable devices today. Solar photovoltaic modules have followed a similar cost trajectory.
Yet technology adoption is not solely dictated by cost declines. The magnitude of this decline also matters as does the resulting financial proposition relative to the status quo. Where AVs are concerned, of relevance isn’t just whether costs will decrease, but also how these costs compare to using non-AVs. Publicly available data suggests cost parity between AVs and non-AVs is unlikely, with AV hardware costs alone — estimate at scale — to be three times higher than vehicles currently used by low-income households.
Mobility as a service
AV procurement cost concerns are habitually assuaged by offering the technology for hire, a setup that — like the existing taxi industry — distributes expenditures over a large number of consumers, thereby making AVs more affordable. The ‘robotaxi’ model purportedly offers further consumer savings owing to the antiquated need to employ (and hence pay) human taxi drivers. Robotaxis feature prominently in existing AV-related discourse as a pathway towards realizing a fairer, more equitable mobility ecosystem.
However, this realization faces challenges. Supply–demand matching inefficiencies — rather than capital and labour expenditures — are the principal influencer of taxi fares. Moreover, while AV technology can reduce labour dependence and, consequently, labour costs, regulation necessitates human oversight for safety-critical systems, regardless of automaticity magnitude. Hence, the robotaxi model is unlikely to yield large consumer savings. This is significant for low-income households who demonstrate higher price sensitivity.
The process whereby multiple fare-paying riders — travelling, ideally, to the same destination — share the same vehicle can offer an improved cost proposition. Redistributing costs — on a per trip basis — over multiple riders represents the least effortful means of cost reduction, relative to the alternatives surmised thus far. The result is a cost proposition that is at least as attractive as existing vehicle ownership models. Existing discourse suggests that that complete realization of AVs’ public-health benefits may necessitate high occupancy travel.
However, consumers show strong aversion to this proposition, an effect influenced by higher travel time uncertainty and often longer travel times. Uncertainty is significant for low-income households whose occupations generally offer little room for unpunctuality. Similarly, conditioning a favourable AV cost proposition on ride-pooling risks widening socioeconomic inequality given the strong relationship between longer travel times and economic mobility. Research finds longer travel times impede low-income families’ chances of escaping poverty.
Existing AV discourse has — thus far — overlooked equity concerns. Yet, if technology does not work for society’s most vulnerable, then it does not work at all. Failure by policymakers to acknowledge and address AV cost impediments risks further exacerbating racial, social and economic inequality.
Image credit: Jacky