In 2014, San Francisco made a “Vision Zero” pledge to eliminate traffic fatalities within 10 years. Instead, it went in the opposite direction. The 42 people killed on SF streets last year were the decade’s dreadful high.
While deaths garner headlines, non-fatal injury collisions happen nearly every day, often with debilitating consequences. The Vision Zero era also did little to reduce those annual totals.
Now the pledge has expired, and San Francisco is scrambling for a key to unlock the secret: How not just to slow down drivers, but to make the city’s streets safer?
On Tuesday, Sup. Myrna Melgar is introducing legislation that calls for several city agencies to collaborate on new measures to “get the results we all want — safe, comfortable, calm, and vibrant streets,” says Melgar’s aide Emma Hare.
It’s common knowledge that speed kills. A widely cited Automobile Association of America study shows that a person struck by a car going 40 mph or faster has a three in four chance of severe injury or death.
If that car is going 25 mph or slower, the odds shrink to one in four. As speed increases, a driver’s field of vision decreases, while the stopping distance increases — a deadly combo.
A May report from the San Francisco Municipal Transportation Agency (SFMTA) cited unsafe speed as the most common traffic violation tied to injury and fatal crashes in SF from 2020 to 2024.
To slow drivers’ roll, SF has a growing arsenal including speed bumps, new automated cameras, daylighting intersections, and old-fashioned cops-on-the-street enforcement.
But each has problems. The speed cameras just arrived and will be slow to roll out — only 33 cameras in a five-year pilot. The city’s speed bump program just hit, yes, a big bump when the SFMTA, in charge of streets, paused new applications due to a huge backlog. (The agency just received an extra $7 million from a county fund just to get through projects approved through 2023.)
As for the cops: in June, a civil grand jury report said SFPD’s well-documented lack of enforcement was a major reason for Vision Zero’s failure. But an SF Standard investigation cast doubt upon the grand jury’s conclusions.
The city has another tool to use, about as low-tech as it gets: lower speed limits. In 2021, California began letting cities lower their own speed limits on some streets. SF immediately put 20 mph limits in place in the Tenderloin and has since reduced speeds on more than 80 other street segments.
SFMTA has examined the effect of some of these reductions on traffic speed, but to date it has not evaluated their effect on injury collisions. To start filling this gap, The Frisc has analyzed 12 major street segments with dangerous histories where officials dropped limits to 20 mph in recent years. Here’s what we found.
Is 20 plenty?
The Frisc looked at injury collisions along a dozen street segments where the city has lowered the speed limit to 20 mph. We chose the segments because they are on the city’s high-injury network (which is due soon for an update); they provide good geographic representation; and the speed-limit reduction took place far back enough to provide enough time for analysis.
Most segments are less than one mile long, but a few stretch just beyond that.
We asked two street safety experts, one with the city, one independent, to examine our data. They provided some caveats, which we will note later, but both said the subset of streets we examined were a good start toward a fuller picture.
Eight of the 12 segments we analyzed saw significant drops in injury collisions, as much as 50 percent better. Three in the Tenderloin were among the best.
Golden Gate Avenue, from Market Street to Van Ness Avenue, topped our list with a 56 percent improvement. Two other Tenderloin streets, Taylor (43 percent) and Turk (29 percent), also showed big drops in collisions. All these streets were engineered post-World War II as fast one-way thoroughfares to move traffic from downtown to outer neighborhoods or vice versa.
Not all recent changes in the Tenderloin helped. Eddy Street saw a 81 percent increase in injury collisions after the speed limit was lowered. SFMTA traffic engineer Ricardo Olea noted the discrepancy and checked it himself. (It checked out.) He couldn’t explain the increase but has flagged it for further study.
Unlike the other Tenderloin streets we examined, Eddy is two-way; there’s currently work in progress on a bulb-out to shorten the crossing distance at Larkin Street.
We also examined Mission Street, San Francisco’s longest. The city has lowered speed limits on four different segments, because “Mission in downtown is different than it is in the Excelsior,” says Olea.
The two segments downtown were a mixed bag of results. Between The Embarcadero and Beale Street, injury collisions were down 34 percent. But a few blocks away, between First and Third Streets, those crashes were up 40 percent.
Quantifying the benefits of traffic calming is hard. We don’t have laboratory conditions.
San Jose State university professor Marcel Moran
In outer neighborhoods there was also a split on Mission. Between 14th Street and Cortland Street — the length of the Mission District — injury crashes were up 8 percent. Between Silver Avenue and Foote Street in the Excelsior, crashes were down 14 percent.
“Quantifying the benefits of traffic calming is hard — we don’t have laboratory conditions,” says San Jose State assistant professor of planning, policy, and environmental studies Marcel Moran, who also examined our data. Changes like lower speed limits don’t happen in a vacuum, and many streets we analyzed have had more than one traffic control treatment, such as no-turn-on-red in the Tenderloin.
On Mission Street between 14th and Cortland, there is a red transit-only lane, effectively reducing the number of lanes for cars. This and the 20 mph limit would be expected to improve safety, yet this two-mile stretch saw a slight increase in injury collisions.
We also examined main corridors in four neighborhoods: Columbus Avenue in North Beach, Third Street in the Bayview, Upper Market Street, and 16th Street in the Mission. Only Third Street showed an increase in injury collisions.
Overall the data show speed limit reductions are “leading to meaningful improvements in traffic safety in tandem with restrictions like no-turn-on-red,” says Moran.
SFMTA’s Olea cautions about drawing conclusions from a small sample of segments. A more thorough analysis would show a fuller picture — one that doesn’t exist yet for San Francisco. But it does for other cities.
Safer Seattle, middling Minnesota
In 2016, Seattle lowered default speed limits to 20 mph on residential streets and 25 mph on arterials. In 2020, city transportation officials analyzed five areas and found that fatal crashes dropped 22 percent, injury crashes fell 18 percent, and “egregious” speeding was cut by more than half.
The overall median speed — where half drive above the limit and half below — was reduced by 10 percent. City engineers made no physical changes to the streets except for signs to remind drivers of the new limits, often spacing them no more than a quarter mile apart.
Cities lowering speed limits should not wait too long to modify signs to remind drivers.
from a study of seattle traffic by the Insurance Institute for Highway Safety
The frequent signs seemed to help, according to an independent study that also corroborated the safety findings: cities lowering speed limits “should not wait too long to modify signs to remind drivers … to maximize the safety benefits.”
A smaller city’s study was not as comprehensive and not as conclusive. In the Minneapolis suburb of St. Louis Park, population about 50,000, city officials in 2021 lowered the limit for most city streets to 20 mph, down from 30 mph.
University of Minnesota engineer Gary Davis directed a study of 28 two-way, two-lane roads in the town. The study only looked at traffic speed, not collisions.
On average, speeds were 1 to 2 mph lower after the change — both on streets with new limits and on streets that did not receive new limits.
Davis wrote that these findings were consistent with similarly modest results from the United Kingdom (Bristol) and North America (Montreal, Edmonton, and Boston), where lower speed limits were implemented without other physical changes to the streets or any “vigorous enforcement.”
More data should be coming soon. New York City is rolling out lower speed limits in 250 locations, thanks to a state law passed last year after several attempts.
Update, 8/1/25: SF transit officials said July 30 that the city’s first speed cameras were having an effect on driver behavior in the first two months of use. Of drivers who received warnings for going at least 11 mph over the speed limit, 70 percent “did not speed past the cameras again,” said SFMTA director of streets Viktoriya Wise. SFMTA said the average daily number of speeders across all camera sites dropped 31 percent. Wise said SFMTA is adding larger speed limit signs and stenciled warnings on the pavement near camera sites.
San Francisco’s only previous study of speed limit effectiveness also came from the SFMTA and was limited to vehicle speed. (It did not evaluate collisions.) The study looked at nine “business activity districts,” including Haight Street and San Bruno Avenue, where the speed limit dropped from 25 mph to 20 mph in 2022.
SFMTA’s Olea says studies using speed as a proxy for safety allow the agency to gather data more quickly. The agency found “no change in typical daily vehicle speeds,” which essentially meant there was “a decrease in compliance” — fewer drivers observing the speed limit.
Remedies not memorials
Because of the 2021 state law, the city for now can only lower speeds in commercial districts, along high-injury streets, and on street segments with high populations of vulnerable residents, including seniors and small children.
SFMTA engineer Olea says the agency chooses its spots according to need and geographical distribution.
Safety advocacy group Walk SF, which has used its own speed gun to measure driver speeds around the city, is calling for a 20 mph limit citywide. “We need 20 mph everywhere so we could have more consistency,” says Walk SF communications director Marta Lindsey.
SFMTA understands that not just one method will make streets safer. Melgar’s proposed legislation, the San Francisco Street Safety Act, outlines to-do items for six city agencies, including SFMTA, Public Works, and the fire department, along with deadlines for completing work. For instance, it calls on SFMTA to deliver by 2026 a plan to redesign high-injury streets that includes traffic-calming measures and physical barriers to separate pedestrians from traffic.
Many safety advocates say the fire department holds up street changes through long reviews. Melgar’s bill sets a 90-day maximum for SFFD review of certain projects, for instance.
The bill also calls for the Department of Public Health to create a new street risk map that factors not just fatal and injury crashes but also close calls. Emma Hare, aide to Sup. Melgar, says the point is to reduce conflicts over road use between bicyclists, pedestrians, cars, and public transit: “We want to identify high-conflict areas to put in a remedy instead of having to host a memorial.”
The Lurie administration says more efficient government will help solve small business and homelessness problems. But SF has a spotty track record on collaborations for complicated street projects. To build safer streets faster, fewer eyes on the street won’t make Vision Zero a reality without figuring out what works.
This story has been updated with SFMTA’s July 30 announcement about the effect of speed cameras.
Just because the speed limit is lowered to 20 MPH doesn’t mean cars follow the new speed limit. I don’t see that acknowledged in this article. Ocean Ave. was lowered to 20 MPH and very few cars follow it and may not even be aware of it. Who looks at speed limit signs? Many don’t even notice them.