Brad Greenwood had a bit too much to drink at a wedding, and he needed a ride home.
“Uber X took me home,” said Greenwood, an assistant professor of management and information systems at Temple. “I emailed myself that night, and I was like, ‘Uber was helpful.’ That’s the actual genesis of the paper. Now, it’s situated in a much larger body of work.”
Greenwood’s experience with Uber, a ride-sharing service, sparked his already existing interest in studying the effects of the app on fatalities related to drunk driving. Greenwood teamed up with his colleague Sunil Wattal, another assistant professor of management information systems, and in January the two published their study, “Show Me the Way to Go Home: An Empirical Investigation of Ride Sharing and Alcohol Related Motor Vehicle Homicide.”
Greenwood and Wattal found that from 2009 to 2014, the introduction of Uber X, Uber’s low-cost option, in California reduced alcohol-related vehicular fatalities by 3.6 to 5.6 percent.
“We’re pretty excited by these findings,” Wattal said. “At some level, we think the findings are also what you would expect. One of the reasons that people drive drunk is that there are not many low-cost options to get home.”
Greenwood and Wattal think that’s why Uber X is connected to fewer alcohol-related fatalities, while Black Car, Uber’s premium service, has not had the same effect.
Wattal said he and Greenwood tried to gather government data for their study from states including California, Texas and Florida, but California was the only state that had adequate data.
“Uber was introduced into many different places, but it has the most installations in California cities,” Wattal said. “California had 34 or more cities with Uber.”
All the data in the study was gathered from the California Highway Patrol Statewide Integrated Traffic Records System. Then, Greenwood said he and Wattal used the difference-in-differences estimation method to “try and establish causal relationships in [their] secondary data.”
“The reason that you would use something like that is because it mimics an experimental setup when you don’t actually have an experiment, and you can’t actually manipulate things,” Greenwood said. “So, we can see how the change in the trend of drunk-driving and alcohol-related fatalities is different in locations that get [Uber] from locations that don’t get [Uber].”
Greenwood added this method also helps account for the unobserved characteristics of the California cities, like their sizes and their differing rates of DUI arrests.
Wattal said that although this study was conducted using data from California, to an extent, it is still applicable to Philadelphia.
“The theoretical motivation for the study, like why people drive drunk, those factors are likely to be very similar across the country,” Wattal said. “So we expect the results would hold across the country.”
Greenwood said there are still many questions left unanswered, and he hopes other researchers continue to study the effects of Uber on society.
“Who are these lives that Uber is saving?” Greenwood said. “Are they affluent lives, lives that would’ve taken a taxi anyway? Are they lives of the socioeconomic lower class? … There’s so much left to do. I hope that people work toward unlocking these mysteries.”
Michaela Winberg can be reached at email@example.com or 215.204.7416.