As Research Shows Ride-Hail Decreases Drunk Driving Fatalities, Can Autonomous Cars Help?
UC Berkeley economics professor Lucas Davis has always had an interest in the intersection of transportation safety and public health. That’s why his curiosity was piqued when he came across a 2017 article in the New York Times, about alcohol-related traffic fatalities dropping in major U.S. cities over the last decade. Despite empirical evidence suggesting a decline in drinking and driving, the body of academic studies at that time drew shaky conclusions about the correlation between the reduction of alcohol-related traffic fatalities and the rise of ride-hailing services. Nor did they say anything about the potential impact of autonomous driving.
After reading and discussing the article, Davis and his colleague Michael Anderson began seeking out research on the linkage between ridesharing services and the reduction of alcohol-related accidents. Most of what they found was inconclusive. “It seemed a little weird to us,” says Davis. “The fact that we had this literature with results all over the place. It made us more interested in the question. Couldn’t we do better? Couldn’t we get a more definitive answer to this?”
So, they formulated the question they sought to answer: What is the effect of ride-sharing on alcohol-related deaths? In a breakthrough study published this July by the Nation Bureau of Economic Research, Anderson and Davis found that Uber alone had reduced alcohol-related traffic accidents in 2019 by 6.1%, which equates to over 200 lives saved that year. Moreover, they deduced that Uber decreased overall traffic deaths by 4% in that year. That data is just the result of examining one ride-sharing service.
The researchers examined data from two sources, the National Highway Traffic Safety Administration’s (NHTSA) Fatality Analysis Reporting System and rideshare activity data from Uber. The ability to access proprietary anonymized data from Uber was hugely advantageous to their findings, according to Anderson, and it took the researchers years of convincing to get a hold of it. Their analysis of the data included road type, geographic location, date and time of incidents, and of course, the involvement of alcohol. The swath of research covered nearly two decades of data from NHTSA, ranging from 2001 to 2017, and five years of Uber data from 2012 through 2017. This analysis allowed for them to extrapolate their findings to approximate the average effect of ride-sharing through 2019.
Given what he’s seen from the Uber data, Davis is encouraged by the potential for autonomous vehicles to further improve on declines in the incidence of alcohol related accidents, especially as AVs become integrated within ride-hail services. Earlier this month, Argo AI announced a new partnership with Lyft that will accelerate the deployment of Ford driverless vehicles for ride-hailing in two major markets: Miami and Austin. This collaboration, the first of its kind in the self-driving industry, could make its own impact on the frequency of alcohol-induced accidents.
“I think automated vehicles could reduce [fatalities] to a fraction of the amount we have now,” the researcher says. And the reason for this? It’s simple, Davis says. Robots aren’t as unpredictable as humans—and they certainly don’t suffer from the slowed reaction times and loss of visual and mental acuity that comes from drinking and driving. Davis will soon be able to test this hypothesis with the era of autonomous ride-hail dawning, as the cars are designed to maximize safety. For now, the data on ride-hail’s effect on alcohol-related incidents seems clear.