Uber and Lyft driver sitting in traffic. Photo by Yesica Prado, SF Public Press, in October 2021

Leer en español

San Francisco, the city that spawned Uber and Lyft, has also suffered the most in emissions and congestion from them, according to a first-of-its-kind, statewide study of ride-hailing activity released by the San Francisco County Transportation Authority on Friday. 

The city clocked the highest concentration of rides, at 500 times more trips per square mile than the rest of the state, from September 2019 to August 2020, the period covered in the report.

That year, CO2 emitted by Uber cabs in California reached an estimated 494,000 metric tons, numbers comparable to the 2020 Caldwell Fire in northern California, which burned 81,000 acres of land. 

About 30 percent of the CO2 emissions were generated with no passenger in the car, as drivers waited for ride requests or were on their way to pick up passengers. 

The study fills in a gap in a field where, normally, “it’s very hard” to get data on ride-hailing activities for both government agencies and academic studies, said Tilly Chang, the executive director of the County Transportation Authority. Shey says the only exception is New York City, which gets regular reports from Uber and Lyft.

The facts of the ride-hailing industry have a long history of appearing opaque, not just to drivers, but even to those who study it — or regulate it. For the first time, the report, “TNCs 2020: A Profile of Ride-Hailing in California,” offers San Francisco a panoramic view into the real cost of the industry and its impact on the city’s environment, public safety and, most important, its people. TNC is an abbreviated term for Transportation Network Companies such as Uber and Lyft. 

A map of California
San Francisco boasted the highest concentration of rides at 500 times more trips per square mile than the rest of the state, According to “TNCs 2020: A Profile of Ride-Hailing in California,” a study of statewide ride-hailing activity released by the San Francisco County Transportation Authority.

The data shows that San Francisco also suffered from an increased level of fine particle emissions, known as PM2.5 emissions, that create haze and health problems. This resulted in a concentration that is 340 times higher than in the rest of the state.

According to the study, Uber accounted for five percent of all PM2.5 fine particle emissions produced by passenger vehicles and light duty trucks in the city. (Lyft did not report any data to the state that could be used to estimate emissions.)

Around 64 percent of all Uber and Lyft rides in California between September, 2019, and August, 2020, took place in three counties — San Francisco, Los Angeles and San Diego — which collectively make up five percent of the state’s land, according to the study.

Pre- and post-Covid fluctuations

Looking at data from millions of trips, the study is able to tell the story of ride-hailing across the state. Before the pandemic, trips in California were regularly increasing from Monday to Friday, peaking on Saturday and bottoming out on Sunday. 

Unsurprisingly, that exacerbated congestion. The highest volumes of ride-hailing activity occur in the morning and early evening, when congestion is at its worst. Also, “Within San Francisco, trips are further concentrated within the downtown core, on the city’s most congested streets where the city prioritizes sustainable, space-efficient modes of travel, such as transit, bicycling and walking,” says the study.

Lyft, the platform that now plans to lay off 1,200 of its roughly 4,000 employees to reduce costs, reported more driver hours and driver days than Uber before covid, and fewer driver days and driver hours during covid.

In the first six months of the pandemic, trips declined by 80 percent on both platforms.

More than half of the wheelchair-accessible trip requests went unserved. And of the 47 percent of wheelchair requests that were fulfilled, 107,752 out of 229,540, nearly all of them ended up in Ubers. Uber provided 16 times as many wheelchair rides as Lyft.

Data is inconsistent and redacted

The study is based on an analysis of annual reports filed by Uber and Lyft to the California Public Utilities Commission. But that data is incomplete: The report states that the data the County Transportation Authority received under a Public Records Act request has “been highly redacted” by the state.

A graphic showing that different report sources will lead to different numbers for the same metric
Even the total trip numbers vary a lot from different report sources. Graphic by Chuqin Jiang.

Additionally, Uber and Lyft may have used different definitions when reporting their data, making comparisons difficult.

Consequently, “Uber’s and Lyft’s data is internally inconsistent,” according to the report. Lyft reported 36 percent of the required data, while Uber reported 99.99 percent of the required data. Due to a lack of raw data, Lyft’s PM2.5 and CO2 emissions were missing from the report, for instance. 

Even for the items that both companies provided, each seemed to be reporting differently. For instance, Lyft reports three times more total public-safety incidents per trip than Uber, and 30 times more assaults and harassment incidents per trip, possibly the result of a different definition of “public-safety incident.” In addition, Lyft suspended drivers at more than 11 times the rate of Uber.

A graphic showing four types of public safety incidents reported by Uber and Lyft.
Uber and Lyft reported public safety incidents under different standards. Graphic by Chuqin Jiang.

“We’re just confused,” said Chang. “The commission has repeatedly found that the data should be unshielded — the actual commission. But then the staff have not been implementing that.” 

Chang said it was unclear if the discrepancies were due to poor reporting, or redactions by the state.

“Our wish is for the full data set to be made available, transparent,” Chang continued. “But maybe these were how the companies submitted them, we don’t know. So if the company submitted them fully, then I don’t know why we would not have the benefit of seeing them. So this is a question for the CPUC.”

The Transportation Authority presented its report today at San Francisco’s Transportation Authority Board meeting. The Transportation Authority’s 11 members are the San Francisco Board of Supervisors.

Supervisors Dean Preston and Myrna Melgar urged legal actions regarding the incompleteness of the data at the meeting.

“The CPUC doesn’t understand its job, or is not interested in performing its functions of regulating this industry. And here we are, with a mess on our streets,” said Preston. “How is Lyft not facing a lawsuit right now? You basically tell regulators you’re not going to turn anything over that you’re required to provide. This data is a joke.”

“We don’t exactly know where the problem starts. Is it the companies, or the CPUC?” Joe Castiglione, deputy director for technology, data and analysis at the San Francisco County Transportation Authority, said. “I would agree, though, that there is clearly an enforcement failure here.”

“I just want to encourage both the County Transportation Authority, as well as the city attorney, to explore what the legal options are. This has gotten to the point of ridiculousness that you see this extent of redactions,” said Preston.

Uber officials pushed back on the study’s findings.

“Broad conclusions should not be made based on old, outdated data,” Andrew Hasbun, Uber’s Head of Safety Communications, wrote in an email to Mission Local.

He said the data went to the state “with the information they requested in the required format,” but added that some data were withheld as confidential to protect rider and driver safety and privacy.

He blamed the inconsistencies in responses on “evolving definitions and instructions.” It wasn’t until June, 2022, he said, that the state “began implementing uniform definitions for reporting sexual assault and sexual harassment.”

As of publication time, neither Lyft nor the CPUC replied to requests for comment.

The County Transportation Authority’s last report on ride-hailing activity in San Francisco was “TNCs Today” in 2017, and in 2018, the department’s “TNCs and Congestion” report analyzed how Uber and Lyft had affected roadway gridlock. 

Within the current report, the County Transportation Authority says it will follow up — when the state releases  “properly redacted” sets of data.  

This story has been updated to include comments from today’s hearing from the County Transportation Authority Board.

Follow Us

REPORTER. Yujie Zhou is our newest reporter and came on as an intern after graduating from Columbia University's Graduate School of Journalism. She is a full-time staff reporter as part of the Report for America program that helps put young journalists in newsrooms. Before falling in love with the Mission, Yujie covered New York City, studied politics through the “street clashes” in Hong Kong, and earned a wine-tasting certificate in two days. She’s proud to be a bilingual journalist. Follow her on Twitter @Yujie_ZZ.

Join the Conversation


Please keep your comments short and civil. Do not leave multiple comments under multiple names on one article. We will zap comments that fail to adhere to these short and very easy-to-follow rules.

Your email address will not be published. Required fields are marked *

  1. I do think this needs to be studied but rideshares aren’t the primary cause of gridlock it is CARS that cause gridlock. If all rideshare rides were replaced with personally owned vehicles we’d have the same problem just with worse parking. It also seems incorrect to include the stat that the major cities make up “5% of the land” of California… what percentage of the population is it? Thats a better reflection of the car usage.

  2. What are the emissions per passenger mile traveled for rideshare? What are the emissions per passenger mile traveled for Muni?
    Compare apples to apples…

  3. It’s just because of more trips vs. the previous taxi status quo, because people want to take those rides. Taxis drive around between rides too. But the apps make hailing a ride easier and more predictable, especially if you’re starting or ending some place without a lot of taxis. If Uber and Lyft disappeared, and everyone hopped to Flywheel or similar, the carbon situation wouldn’t change, provided Flywheel was just as easy and cheap. Of course cheap is also a factor – if driver comp and thus ride prices were higher, then rides would decrease too. But you need a replacement, e.g. lots more public transit (fast, frequent, safe) and bike/ped infrastructure. Or they all go electric and reduce non-ride drive time. Otherwise, you’re just telling people to take longer to get where they want to go, or not go at all. That’s not to say we can’t regulate some of these issues – driver comp, safety, app transparency, , even vehicle type, etc. But it’s not like the apps force people to ride, anymore than taxis do.

  4. Wrong and stupid comparisons. SF is the second more dense city in the U.S., and it’s being compared to the rest of the state of California? People need to get around. Uber and Lyft provide excellent service– fast, safe, efficient, and reliable. Unlike Muni and BART. And people are disincentivized from using their private vehicles because the City has taken away many parking spaces. Bicycles, which are used by only a tiny percent of people, are awful in inclement weather, and dangerous at night. And no, Uber and Lyft are not the dominant factor in congestion and gridlock– the City’s stupid projects like putting largely unused bus lanes down Van Ness are much more responsible. It is also irresponsible to imply that buses create much less pollution than cars. About 20% of a bus’s runs are empty– deadheading to or from the bus barn. At these times they’re creating nothing by pollution, and causing congestion because of their size. City environmentalists have also opposed Google buses, even though each Google bus takes about 48 cars off the road. Transportation planning and thinking needs to be much more realistic, including considering the needs, wants, and desires of real people. If you want to reduce pollution and congestion, replace timed traffic lights with demand lights, and permit right turn on red. This would cut idling significantly.

    1. My taxi friends were kicked to the curb so a few investors could get rich quick. But… it’s been ELEVEN YEARS and Uber and Lyft have lost BILLIONS. Guess what: Uber gave up and asked to install their dispatch in San Francisco taxis for a cut. Too bad for taxi drivers already driven to poverty and poor health as they struggled to survive. Dear departed Mayor Lee allowed Uber and Lyft to freely operate w/o restrictions. His daughter was a Lyft executive by the way.

  5. “…such as transit, bicycling and walking,” isn’t realistic since most people prefer a private, door to door trip, for various reasons. Uber and Lyft have become a seamless necessity in our lives.

  6. This morning on the way to work I saw about 5 totally empty self-driving cars contributing to traffic as well. They don’t emit greenhouses gases on our streets, but their carbon footprint is definitely being felt by the locations that provided manufacturing and raw materials.

    1. … and the traffic jams they cause as they, “learn,” are pretty dangerous too.

      Nothing in this article surprises me at all, it is just all incredibly disappointing.

      In the olden days, not 10 years ago, we just took Muni or rode our (non battery operated) bikes around the city. Owned a beater car for longer trips. Even once we had kids, for goodness sake.

      Now, starting with millennial-aged tech workers, people have built up their feelings of entitlement to a fevered pitch, and spend money on having essentially a private driver or delivery person on demand — and they think they are not part of the problem.

    2. Brake dust and tire wear is still a big problem with EV/AVs. Especially since they are so heavy compared to ICE vehicles. Please just take me back to 2010 before Uber and Lyft but keep Flywheel. Let me order a regulated Taxi on my phone, driven by an old angry man. Ban everything else.

    3. The carbon footprint is also contributed to by the electricity generated to power these AI cars.