Do ride-sharing apps discriminate against black customers?

Each month, the NBER Digest summarizes several recent NBER working papers. These papers have not been peer-reviewed, but are circulated by their authors for comment and discussion. With the NBER’s blessing, Making Sen$e is pleased to feature these summaries regularly on our page.

The following summary was written by the NBER and doesn’t necessarily reflect the views of Making Sen$e.

The advent of the ride-sharing industry is rapidly changing the marketplace for transportation. New services are crowding the traditional taxi industry, which is subject to an array of local regulations, including strict anti-discrimination laws designed to prevent taxi drivers from offering differential services to potential passengers from different age groups, racial groups or other groups. A new study explores whether drivers in the ride-sharing industry differentiate among potential customers.

In “Racial and Gender Discrimination in Transportation Network Companies,” Yanbo Ge, Christopher R. Knittel, Don MacKenzie and Stephen Zoepf report the findings of field experiments in two cities, Seattle and Boston. Their results suggest that drivers for ride-sharing services are prone to discriminate against African Americans, making blacks wait longer for rides when they can identify the race of the ride-hailer and frequently cancelling rides when alerted to African American-sounding names. The disparities are particularly pronounced for black males. The researchers also find that ride-sharing drivers take female passengers on longer rides.

In each of the cities in which they fielded their experiment, the researchers measured the performance of ride-sharing services via field experiments in which research assistants — whites and blacks, males and females — were randomly dispatched into the field, at varying times of the day and to varying locations, to order, wait for and ride in ride-sharing vehicles. The research assistants carefully monitored and recorded predetermined performance metrics for every ride they took, including how long it took drivers to accept ride assignments, how long passengers had to wait until drivers arrived and how long and expensive rides were for each passenger. In all, research assistants conducted nearly 1,500 individual rides in Seattle and Boston. In each city, the research assistants summoned rides from several ride-sharing firms.

In Seattle, the main finding was that African Americans had considerably longer waiting times for rides — as much as 35 percent more. In Boston, the researchers could measure the cancellation rates of the drivers from some services after they had preliminarily accepted ride assignments. They modified their experiments so that some students hailing rides used “white-sounding” names while others used “African-American-sounding” names. They found more frequent cancellations — roughly twice the level for white-sounding names — when the students used African American-sounding names. Male passengers requesting a ride in low-density areas, such as in the country or suburbs, were nearly four times more likely to have their trips canceled when they used an African American-sounding name than when they used a white-sounding name.

The researchers note in conclusion that changing the information about potential customers that ride-sharing services provide to their drivers might affect some of the patterns that they observed.

— Jay Fitzgerald, National Bureau of Economic Research