Laura Bliss is a staff writer at CityLab, covering transportation and technology. She also authors MapLab, a biweekly newsletter about maps (subscribe here). Her work has appeared in the New York Times, The Atlantic, Los Angeles magazine, and beyond.
With a rare look at trip data from the ride-hailing giant, a UCLA researcher finds promising equity results.
In the eight short years since the first “UberCab” pick-up in San Francisco, ride-hailing apps like Uber and Lyft have upturned old transportation networks and created unprecedented demand for travel.
But have their benefits extended to communities long left behind by the taxi industry, and that need car services most? For decades, racial discrimination by cab drivers has left black riders, in particular, waiting longer for pick-ups, having their destinations refused, and flat-out ignored, studies show; a 2013 investigation in Washington, D.C., found taxis were 25 percent less likely to pick up a black rider than a white rider. This plays out on a spatial level—outer-urban neighborhoods that are predominantly home to people of color are often “redlined” by taxi companies, for various reasons. Earlier research has shown some of the same practices persist in the new apps.
But a dissertation from UCLA’s Institute for Transportation Studies paints a more promising portrait of access to Lyft in Los Angeles County. Contrary to the belief that ride-hailing primarily serves the affluent, it appears neighborhoods with low rates of car ownership—which tend to be populated by people of color—actually see more pick-ups and drop-offs than others. But on the individual level, bias against certain passengers still persists.
Alongside a team of graduate researchers, Anne E. Brown (who received her Ph.D. from UCLA this year) analyzed trip-level records from more than 6.3 million Lyft journeys made within L.A. County in the fall of 2016. Previously unavailable to scholars or policymakers, this data was carefully negotiated upon with Lyft.
The most basic finding is striking: Virtually no neighborhood in the country’s most densely populated urban area has been left unpenetrated by Lyft. The company’s drivers serve 99.8 percent of the population of L.A. County. That in itself suggests that communities aren’t being systematically excluded.
Earlier studies have found that Uber and Lyft riders tend to be on the higher-end of the income spectrum. Brown found that Lyft riders are indeed disproportionately concentrated in wealthier neighborhoods, all else being equal. But she also found that users living in low-income areas made more Lyft trips per person, compared to middle- and high-income communities.
Why was that? The most important factor explaining how often someone used Lyft wasn’t income—it was whether they owned a personal vehicle. On the neighborhood level, Brown found that every 10 percent increase in the portion of households without a car was linked to a 7 percent increase in the number of Lyft trips an individual rider there made.
The figure above puts it another way: When the local vehicle ownership rate is held constant (darker bars), wealthier groups appear to ride more often. But when it’s not (lighter bars), lower-income people clearly ride the most.
So Lyft does appear to be providing vehicle access to areas “where its substitute—the household car—is least available,” Brown wrote. “Users living in low-income neighborhoods ... may have low—or zero—personal car access and therefore use Lyft to provide rather than supplement auto-mobility.”
That’s is a score for equity, given that 80 percent of car-free households in the U.S. point to financial constraints, not any personal choice, as the reason. That low car-ownership neighborhoods are getting this much Lyft service is doubly encouraging because, as Brown found, majority-black neighborhoods have the highest share of zero-car households in Los Angeles, compared to any other racial group. Many of these communities were previously underserved by taxi companies. Trips from majority-black neighborhoods were also more likely to be shared rides through the company’s Lyft Line carpool option; fascinatingly, Brown discovered that mixed-race neighborhoods saw fewer shared rides.
Ride-hailing also appears to be reducing the pernicious effects of discrimination at the passenger level. In addition to analyzing the Lyft data, Brown and her colleagues conducted an audit of Lyft, Uber, and taxi rides. Field researchers requested and rode vehicles to gather observations about how drivers responded to requests by riders of different races.
Brown discovered the experience of “hailing while black” is about as bad as it’s always been in the yellow cab industry. Black passengers waited for taxi rides 52 percent longer (between about 6 and 15 minutes) than white riders. But the disparity in wait times was much reduced on Uber and Lyft, with black riders waiting between 11 seconds and 1 minute, 43 seconds longer than white riders. No meaningful differences among white, Asian, and Hispanic riders were observed.
These findings contrast with earlier studies in Boston and Seattle, which found black ride-hailing riders wait longer and get canceled more often than white riders. Brown’s research used a different methodology and focused on a different city; it wouldn’t be right to extrapolate her findings beyond Los Angeles without a comparable audit in other places. Furthermore, for the first part of the analysis, the data she used came from Lyft alone and represented only a limited period of time. The findings might not be generalizable to the ride-hailing industry in general.
On the other hand, this is also one of the first studies (if not the first, outside of New York City) to make use of an otherwise proprietary dataset containing a full population of ride-hailing trips. Other studies have had to rely on surveys or ad-hoc data-scraping methods. So these findings are significant.
They are also a feather in the caps of both Lyft and Uber; the oft-embattled companies have been making efforts to stress their social value in the face of growing evidence to the contrary. Ride-hailing, studies show, has been putting more cars to the road, increasing vehicle-miles traveled in cities, and likely adding—not diminishing, as once promised—greenhouse gas emissions. Lately, both companies have recalibrated their messaging: Now, they say, they’re all about reducing the need for personal car ownership. For those who can’t own vehicles due to financial limitations, research like this suggests that ride-hailing is indeed a lifeline.
But it doesn’t address a much bigger question: Can Uber and Lyft serve, systematically, as chosen substitutes for the personal car, still among the most revered class symbols in American society? The answer to that one lies further down the road.