During a taxi ride late one night a few years ago, a cabbie started berating me for asking him to drive the 25 minutes from downtown D.C. to my home in Maryland. At the time, I thought to myself, At least he took me. In the past, some drivers had simply refused. And, like many other cab riders, I came to expect that his cab’s credit-card machine would be mysteriously “broken” when it came time to pay.

All that has changed. Nowadays, with a few taps on my phone, an Uber or Lyft pulls up, takes me where I need to go, and then automatically charges my credit card and emails me a detailed receipt. While taxis used to face little competition for on-demand point-to-point transportation, today they find themselves in a cutthroat business. The number of taxi trips in New York City decreased by about 8 percent from 2012 to 2014, and one recent report found that Uber provides almost half of all paid rides in major U.S. markets.


Annual Percentage Change in the Number of New York City Taxi Trips

Calculations based on data provided by the New York City Taxi and Limousine Commission (ATLAS / Scott Wallsten)

Any business facing new competition will do its best to keep its existing customers from straying. In regulated industries, like the taxi business, some of those efforts will involve pressuring politicians and regulators to keep competitors out of the market. In France, that opposition has turned violent and even resulted in the arrests of two Uber managers.

But in addition to the lobbying and protests, taxis—according to the rules of the market—should be expected to respond by trying to make their service more attractive to consumers.

In the face of new competition, the first thing an industry would likely see is a decrease in prices. In this case, however, taxi fares tend to be regulated—so prices change slowly and an individual driver has little control over them. It’s also expected that new competition will cause existing firms to increase quality. Over time, taxi companies might improve service by modernizing their fleets and introducing new technologies. And, in fact, some cities have introduced apps to hail taxis.

But the quality of a taxi ride depends on much more than just the fare and the condition of the vehicle. As my angry D.C. cabbie demonstrated, taxi drivers have some control over the quality of the ride. This raises an important question: Are taxi drivers themselves trying to improve quality in the face of competition?

Last fall, a writer for DNAinfo in Chicago quoted a driver who said that cabbies were “trying to behave themselves,” and noted a decrease in the number of taxi complaints. But quotes are anecdotal, and the number of complaints will fall as the number of taxi trips fall, regardless of competition. In order to determine empirically whether taxi drivers are responding to competition from the likes of Uber and Lyft, I set out to assemble a dataset that would help answer it. In a new paper based on data from New York City, Chicago, and Google search trends, I present evidence that cab drivers do seem to have started offering better service.

Like Uber and Lyft themselves, my analysis is possible thanks to data sources and tools that did not exist—or, at least, were not inexpensively available—just a few years ago. New York maintains a publicly available database of taxi complaints online going back to 2010, and cities are putting statistics online through various “open data” projects, making other large datasets available upon request. Through Freedom of Information Act requests, the New York City Taxi and Limousine Commission provided me with data on every taxi ride between 2009 and 2014 and the city of Chicago gave me a similar set of complaint data. For New York City, I was able to look at data for more than one billion rides.

The data showed that in New York, complaints per trip to the Taxi and Limousine Commission have decreased as Uber has grown—even controlling for factors unrelated to Uber such as weather.


Number of Complaints Submitted to the New York City Taxi and Limousine Commission

Scott Wallsten

That decrease in the relative level of complaints is consistent with improvements in taxi quality, but it might also be due to dissatisfied customers switching to Uber rather than bothering to file complaints. But data from Chicago, where Uber also entered the market in 2011, suggest that at least some of the decrease in complaints is related to drivers making an effort to improve the quality of the ride.

Unlike New York, Chicago collects data not just on complaints, but on the nature of the complaint. As the total number of complaints has fallen since 2012, as one would expect given the falling number of taxi rides, so have the types of complaints that one would expect if drivers were making an effort to offer better customer service. In particular, complaints about messy cabs and talking on cell phones fell as a share of total complaints in Chicago. Complaints about cab climate and “broken” credit-card readers, which can be controlled by drivers and the companies that maintain the cabs, also decreased as a share of the total. (The spike in complaints in 2012 can be chalked up to a new rule requiring taxis to display a bumper sticker that read “How’s my driving? Compliments or Concerns, Call 311.”)


Taxi Complaints in Chicago Between 2012 and 2015, by Complaint Type

Based on data provided by the city of Chicago (Scott Wallsten)

In most lines of work, improved quality can generate more business, either by attracting more customers or repeat business from the same customers. But the taxi industry is different: An interaction between a cabbie and a passenger is typically a one-off event. The passenger sticks his or her hand in the air and gets into the first taxi that stops; a cabbie can’t easily signal to potential passengers that he or she won’t be talking on the phone. The one-off nature of the interaction will dilute the incentive to improve quality that a passenger can’t observe prior to the trip. The incentive, then, might come either from a desire for a bigger tip, to help offset decreased demand, or (less likely) a shared concern to avoid hurting the reputation of taxis overall.

It’s obvious that consumers who use Uber, Lyft, or other ride-sharing companies prefer them to regular taxis. If not, they wouldn’t be flocking to those services in ever-larger numbers. But my analysis suggests that even consumers who stick to traditional taxi services are reaping some of the returns. Policymakers would be wise to take that into account when fielding complaints from taxi companies about their new competitors, since eradicating ride-sharing might result in taxi drivers going back to their old ways.