Fact check: Uber and Lyft in Prop. 22 ads say they reduce DUIs, but what do studies show?

Steven Jones, a rideshare driver from Moreno Valley, thinks he’s saving lives in his job.

Jones’ mother was killed by a drunk driver. His wife was hit head-on by a drunk driver a few years ago. By picking up drunk riders for services like Uber and Lyft, he thinks he can save a family from going through the same pain he went through.

That perception is at the heart of one of the arguments over Proposition 22, the California ballot initiative that would exempt Uber and Lyft from a new state labor law that requires companies to provide employment benefits to more workers.

Uber and Lyft say the labor law would raise the cost of doing business, and cause them to reduce service and raise prices. They say that would lead to fewer people calling ride-hailing services, and potentially increase instances of drunken driving.

But setting aside the fact that the initiative’s critics say they don’t believe the cost of rideshare services will go up drastically under Prop. 22, studies have not come to a consensus as to whether Uber or Lyft has decreased DUI injuries and fatalities. Here’s what the studies show:

Claim: “Prop. 22 saves lives by preserving access to rideshare services across California, helping keep drunk and drugged drivers off our roads,” Mothers Against Drunk Driving National President Helen Witty said in a campaign ad for Yes on 22.

Rating: Mixed

Details: Witty points to a 2018 study from the University of California at Davis and Moll Law Group.

That study found that DUI arrests dropped significantly in the years after Uber launched in big California cities. In Sacramento, DUI arrests dropped 16% between 2012 and 2015 (Uber launched in Sacramento in 2013).

A 2017 study from Western Carolina University researchers found that using county-level data from 2007 to 2015 across the country, introduction of Uber has led to some declines in arrests for DUIs. And a 2015 study from George Mason University and Temple University researchers found that the entrance of low-cost UberX resulted in a 3.6% decrease in alcohol-related fatalities per quarter in California.

“Rideshare is a piece of the puzzle,” Witty said. “It’s a part of the puzzle that helps people make good choice if their plan includes alcohol.”

But other research has contradicted those findings.

“Literature hasn’t converged toward the consensus,” said Noli Brazil, an assistant professor at UC Davis. “It’s a relatively new topic.”

A 2018 study of U.S. cities from researchers at Washington University in St. Louis and Rice University says deaths related to drunk driving didn’t change significantly to offset a general increase in the number of vehicle fatalities since the introduction of rideshare services.

A recent study from researchers at UC Davis and the University of Oxford found no link between Uber and the number of alcohol-related fatalities, defined as involving at least one driver who tested positive for alcohol. The study looked at traffic data from the 100 largest metropolitan areas in the U.S. from 2009 to 2016.

Brazil, one of the researchers, said his study found alcohol-involved fatalities actually increased in dense urban areas following the introduction of Uber. It may be because rideshare services have put more cars on the road, increasing congestion and risk for accident, he hypothesized.

Brazil also said it’s possible that Uber is simply replacing other forms of public transportation such as buses. In other words, people taking an Uber after a night out may had gotten into a bus or a ride from a designated driver if the rideshare service hadn’t existed, he said.

“Uber may not be naturally attracting the audience” needed to reduce the fatality rates, he said.

In Brazil’s earlier study looking at the same issue, he also hypothesized that someone contemplating drunk driving may not be rational enough to think about an Uber ride.

“It is also possible that many drunk drivers rationally conclude that it is too costly to pay for an Uber ride (or taxi) given that the likelihood of getting arrested for drinking and driving is actually quite low.”