San Francisco County Grand Jury
2025-2026
Findings & Recommendations
5 findings
F1:
SFPD Fails to Adequately Enforce Traffic Laws ........................................................ 34
Related Recommendations (1)
R1:
1 By September 30, 2025, the mayor should direct the chief of police to develop a plan to specify the level of traffic enforcement activities needed to achieve safer streets. The plan should describe the appropriate frequency of traffic enforcement operations, the extent to which officers should engage in traffic enforcement as part of their duties, any barriers to traffic enforcement, and what active training may be required to support enforcement activities. The plan should be completed by March 31, 2026.
F2:
Traffic Enforcement Is Hampered by Outdated Technology .................................... 35
Related Recommendations (1)
R2:
1 By December 31, 2025, the mayor should direct the chief of police to audit SFPD traffic-specific equipment and software and determine how to expedite deployment of improved software and hardware for citation processing, collision reporting and any other traffic-related administrative duties. SFPD should report back on the proposed implementation of improvements recommended by their audit by December 31, 2026.
F3:
SFMTA Lacks Alignment on Proactive Vision Zero Planning ................................... 35
Related Recommendations (1)
R3:
1 By March 31, 2026, SFMTA should update its planning and site selection process for traffic calming. The process should prioritize proactive intervention selections, informed by data about where the greatest safety benefits are likely to be and where they are needed to support other street safety goals, over reactive selections driven by individual citizen applications.
F4:
Sustained, Community-Focused Outreach is Lacking .............................................. 36
Related Recommendations (1)
R4:
1 By March 31, 2026, SFMTA should assign a point of contact for outreach within each supervisory district, allowing SFMTA to work more closely with members of the board of supervisors; and to build and maintain relationships with key stakeholders in the community on all roadway and transit issues.
F5:
Education Fosters Citizens’ Commitment to Safe Streets ....................................... Required and Requested Responses ............................................................................................. Methodology .................................................................................................................................... Glossary and Acronyms .................................................................................................................. Background Motor vehicle collisions kill more than a million people per year worldwide, including more than 40,000 per year in the United States. Vision Zero is a branding term for a broad policy concept: that all these traffic fatalities should be treated as systems design failures and prevented through systemic fixes. Vision Zero was first adopted as national policy in Sweden in 1997. Sweden’s approach was multi-pronged, including engineering safer roads, slowing speeds, improving vehicle design, educating people and enforcing laws. The results have been remarkable: Sweden’s traffic fatality rate has declined by more than 65% since adoption of this approach, and it currently has the second lowest fatality rate in the Organization for Economic Cooperation and Development (OECD).1 While San Francisco had been taking measures to improve street safety for many years, the full embrace of Vision Zero began in 2014, when a board of supervisors’ resolution, adopted unanimously, urged the mayor, the chief of police and the director of the San Francisco Municipal Transportation Agency (SFMTA) to “implement an action plan to reduce traffic fatalities to zero in the next ten years through better engineering, education, and enforcement.”2 Vision Zero was adopted by the city that same year, with a commitment to eliminate traffic fatalities and greatly reduce severe injuries by 2024. By 2017, 17 department heads signed onto the Vision Zero plan, signaling a citywide commitment to achieve the Vision Zero goal.3 1 OECD, “Road accidents,” For year 2022, Deaths per 1 000 000 inhabitants, Accessed Apr 2025 2 Board of Supervisors, “Implement a Vision Zero three point plan: engineering, education and enforcement,” Resolution 91-14, Mar 2014 3 Office of the Mayor, “Mayor Lee celebrates Vision Zero progress,” April 9, 2015; Vision Zero SF, “Vision Zero San Francisco two-year action strategy 2017-2019,” Accessed Apr 2025 More than a decade later, not only have citywide traffic fatalities not gone to zero, they have not even decreased. The Jury inquired into the root causes of this failure, seeking to understand what a realistic improvement to street safety in San Francisco might look like and where the city has fallen short. Street design and traffic laws are contentious political issues in San Francisco, as in many places. The Jury’s members came to this investigation with different perspectives, but we found common ground on this: while zero traffic fatalities may be an unrealistic near-term goal, the city could and should be doing a lot more to improve street safety. All three pillars of the Vision Zero strategy—enforcement, engineering, and education—need attention. To inform the findings of this report, we need to examine the context—legal and physical—of street safety in San Francisco. That context poses many obstacles to improvement. We’ll also need to discuss the varying causes, consequences and frequencies of different types of traffic collisions. To do that we need to define two key terms: • Vulnerable road users in this report only include the most vulnerable: pedestrians, bicyclists, scooter riders and other users of small, low-powered devices.4 • Vehicle-only collisions are collisions that involve one or more vehicles, but no vulnerable road users (for example, two vehicles colliding or a single vehicle hitting a tree).5 4 Vision Zero SF, “A system of safety,” Accessed May 2025 5 While motorcyclists are often considered vulnerable road users, motorcycles are included as vehicles in the vehicle collision database. Thus, in this report, motorcycles are included in vehicle-only collisions and are not broken out separately; DataSF, “Traffic crashes resulting in injury,” Open Data Portal, Accessed Jan 2025 The headwinds San Francisco is facing strong headwinds in its effort to reduce traffic fatalities and severe injuries, with significant challenges beyond city control. While these factors should not be excuses for inaction, acknowledging them is essential to making progress. Traffic fatalities in San Francisco are not isolated from national trends. Traffic deaths in the U.S. had been declining since the 1990s but began rising in 2014—contrary to continued declines in other wealthy countries. Since 2014, U.S. traffic fatalities have increased by about 26%. The New York Times estimates that if the U.S. had matched the continued decline in Europe, 25,000 fewer Americans would die annually.6 Figure 1. Trends in traffic fatalities per 1 million residents (by the New York Times, Source OECD)7 While road design and enforcement—factors largely within local control—play a role, a major reason for the disparity lies in vehicle design, a factor outside of San Francisco’s control. Compared to Europe, U.S. passenger vehicles are on average heavier (29%), larger (17%), and more powerful (77%).8 Trucks and SUVs make up 73% of U.S. new vehicle sales, compared to 6 David Leonhardt, “The rise in U.S. traffic deaths,” New York Times, Dec 11, 2023 7 ibid 8 ICCT, “European vehicle market statistics,” Pocketbook 2023/24, Accessed Apr 2025, pg. 44 (year 2022); EPA, “The 2023 EPA automotive trends report,” Dec 2023, pg. 36 (year 2022) 50% in the European Union (EU).9 These trends increase the likelihood and severity of all collisions.10 Speed remains the primary factor in crash severity. More powerful vehicles accelerate faster and, along with heavier vehicles, are more likely to speed excessively.11 Faster and heavier vehicles take longer to stop—at 40 mph, stopping distance is two and a half times longer than at 20 mph.12 A pedestrian struck at 30 mph is up to eight times as likely to die as one struck at 20 mph.13 Larger vehicles also have higher, blunter front ends and wider roof supports, which reduce visibility of all other road users and increase severity of collisions, especially for pedestrians.14 Compounding the issue, U.S. vehicle safety regulations lag those in Europe. The EU first issued a directive requiring pedestrian-friendly vehicle designs in 2003. The U.S. only mandated automatic braking in 2023. Pedestrian-specific testing requirements were not even proposed by the U.S. National Highway Traffic Safety Administration (NHTSA) until 2024 and still have not been adopted. 9 ibid, pg. 12 of ICCT and pg. 17 of EPA 10 Michael Anderson, “Safety for whom? The effects of light trucks on vehicle fatalities,” Journal of Health Economic, July 2008; Insurance Institute for Highway Safety (IIHS), “Supersizing vehicles offers minimal safety benefits—but substantial dangers,” Feb 2025 11 Anne McCartt & Wen Hu, “Effects of vehicle power on passenger vehicle speeds,” Traffic Injury Prevention, May 2016; Aviv Steren et al., “Is vehicle weight associated with risky behavior? Analysis of complete national records,” Mark Lett, Mar 2025 12 NHTSA, “Speed-measuring device operator training,” Participant manual, 2018; Samir A.E. Mohamed, et al., “Safe driving distance and speed for collision avoidance in connected vehicles,” Sensors, Sep 2022 13E. Pasasen and H. Salmivaara. “Driving speeds and pedestrian safety in the City of Helsinki,” Traffic Eng. Control, 1993; Erik Rosén and Ulrich Sander, ”Pedestrian fatality risk as a function of car impact speed,” Accident Analysis & Prevention, May 2009; U.S. DOT, “Chapter 5: Risk factors other than exposure,” Federal Highway Administration, Accessed Apr 2025 14 Keith Barry, “The hidden dangers of big pickup trucks,” Consumer Reports, Aug 19, 2024; Wen Hu, at al., “The association between passenger-vehicle front-end profiles and pedestrian injury severity in motor vehicle crashes,” J. of Safety Res., Sep 2024; Wen Hu & Jessica Cicchino, "Relationship of pedestrian crash types and passenger vehicle types,” J Safety Res., Sep 2022 State and federal laws and regulations offer further obstacles to local safety efforts. SFMTA doesn’t have full authority over many major streets, like 19th Avenue, because they double as state or federal highways.15 Until Assembly Bill (AB) 43 came into force in 2022, the city couldn’t even adjust speed limits on local streets to protect pedestrians.16 While red light cameras have been legal in California since 1995, speed cameras were only approved by AB 645 in October 2023 after a decade of attempts and advocacy.17 Finally up and running in 2025, the five-year pilot allows only 33 speed cameras in San Francisco and strictly limits penalties and fines, as well as revenue collection and spending.18 In September 2024, Governor Newsom vetoed Senate Bill 961, which would have required new vehicles to alert drivers when exceeding the speed limit by ten miles per hour.19 A last point here: the California Department of Motor Vehicles (DMV) allows many reckless drivers to remain licensed.20 A potentially brighter future San Francisco residents certainly have noticed the driverless vehicles on our streets. These vehicles can make some mistakes uncommon among human drivers, but they reliably stop at stop signs and red lights, yield to pedestrians, and comply with speed limits.21 And because of their wider vision and faster reaction times, recent analysis suggest that some kinds of autonomous vehicles have been able to avoid many collisions, including those that would have been the fault of others.22 However, not all self-driving technologies are created equal.23 15 SFMTA Board of Directors, “Resolution No. 170516-062,” May 2017 16 Assembly Committee on Transportation, “AB 43—as amended March 22, 2021,” Memo, Apr 19, 2021 17 Heather Knight, “California allows cities to catch speeding drivers with automated cameras,” New York Times, Oct 2023 18 California Legislative Information, “AB-645 Vehicles: speed safety system pilot program,” Oct 16, 2023 19 Soumya Karlamangla, “Newsom vetoes bill requiring cars to warn speeding drivers,” New York Times, Sep 28, 2024 20 Robert Lewis, “License to kill,” CalMatters, Apr 2025 21 Richard Cano, “Yes, there are more driverless Waymos in S.F.,” San Francisco Chronicle, Jul 24, 2024 22 Di Lillo, et al., “Do Autonomous Vehicles Outperform Latest-Generation Human-Driven Vehicles? A Comparison to Waymo's Auto Liability Insurance Claims at 25.3M Miles,” 2024 23 NHTSA, “Additional information regarding EA22002,” Apr 25, 2024 Automated safety technologies on human-controlled vehicles, such as lane assist, automated braking, and intelligent speed assistance, are also becoming much more widespread. Estimates of the potential impact of these technologies range from a 6.7 percent reduction in fatal and injury collisions to a 22 percent reduction in fatalities.24 Automated braking is highly effective in protecting against collisions with other vehicles when traveling at moderate speeds, but is still less reliable in protecting pedestrians, especially at night, and does not protect against reckless driving.25 There is strong consistent evidence that speed cameras reduce vehicle speeds, lower crash rates and save lives; the Jury is encouraged by the San Francisco debut of 33 new speed cameras.26 24 Nordic Road and Transport Research, “Safety impact of advanced driver assistance systems in Europe,” Mar 19, 2024; AAA, “Advanced driver assistance systems,” 2025 25 IIHS, “Pedestrian crash avoidance systems cut crashes—but not in the dark,” Feb 3, 2022; AAA, “Progression of automatic emergency braking technology,” NewsRoom.AAA.com, Oct 2024 26 Wilson et al., “Speed cameras for the prevention of road traffic injury and death,” Cochrane Database of Systematic reviews, Nov 10, 2010 Analysis Drawing upon collision trends, research and interviews, our analysis suggests the severe drop in enforcement, beginning in 2016, has increased risks to all road users. SFMTA’s engineering interventions have followed the opposite trend. Increasing over the Vision Zero decade, SFMTA has provided some protection for vulnerable road users, even with the headwinds the city faces. The analysis presented below begins with an examination of collision trends over the Vision Zero period. The collision databases from the DataSF Open Data Portal demonstrate that total injuries (and the subset of severe injuries) from all vehicle collisions were higher from 2016 to 2019, but they dropped in 2020 when COVID reshaped traffic patterns.27 In 2021 injuries from vehicle-only collisions quickly returned to almost peak levels, while injuries from vehicle collisions with vulnerable users remained lower. Collision trends are followed below by Jury analysis for each of the three Es: enforcement, engineering and education. While traffic safety researchers have called for more comprehensive frameworks—such as the safe systems pyramid, which incorporates socioeconomic factors, vehicle design, and more—the three E approach that served as the starting point for San Francisco’s Vision Zero policy continues to fit the city’s implementation.28 It might not be sufficient to achieve the original zero fatality goal, but the Jury found that improving the effectiveness of the three Es has the power to make San Francisco streets substantially safer. We have come to think of the three Es as a three-legged stool: if any leg is removed, the stool ceases to function. DataSF, “Traffic crashes resulting in injury,” Accessed March 2025 28 David Ederer et al., “The Safe Systems Pyramid: A new framework for traffic safety,” Transportation Research Interdisciplinary Perspectives, 21(2023): 100905 Collision trends Vehicle-only collisions are responsible for 65 percent of the total injuries over 2014 to 2024. Vehicle collisions with pedestrians account for only 20 percent of the total, but these collisions are much more deadly. Over this period, 23 of every 1000 injuries in vehicle-pedestrian collisions were fatal, in comparison to 4 out of every 1000 injuries in vehicle-only collisions. Total injuries from vehicle collisions dropped by 30 percent in 2020; in the next year, vehicle- only collisions rebounded quite a bit, while vehicle collisions with vulnerable users have stayed lower (Figure 2). Vehicle-only Vehicle & vulnerable user Figure 2. Injuries from vehicle-only collisions and collisions between vehicles and vulnerable road users. Note differences in scale. (DataSF, 2025) While a goal of zero was only established for fatalities, Vision Zero targets both fatal and severe injuries. Unfortunately, severe injuries get less attention in official Vision Zero documents than fatalities. While fatal injuries were not clearly impacted, severe injuries for all vehicle-related collisions experienced a substantial drop in 2020 due to the pandemic lockdown. • Severe injuries from vehicle-only collisions quickly rebounded to pre-2020 levels, with fatal injuries trending slightly higher over the decade. • Severe injuries in vehicle collisions with vulnerable road users largely maintained the 2020 drop until increasing in 2024 to 2015 levels (still lower than 2016 to 2019). Fatalities were higher for both categories in 2024 than in any other Vision Zero year. In total, fatalities in 2024 were almost 50 percent higher than the average over the decade.29 As traffic patterns continue to adjust to the post-pandemic world, it is unclear if the high fatalities of 2024 are an outlier or a warning of what is to come. SFPD’s Traffic Collisions Investigative Unit (TCIU) assigns fault for almost every severe and fatal collision.30 Looking at all fault-assigned collisions, drivers were assigned fault for almost 80 percent of the severe and fatal injuries overall and 60 percent of collisions between vehicles and vulnerable users (Figure 3). Figure 3. At-fault in all severe and fatal collisions involving vehicles, 2014-2024 (DataSF, 2025) 29 DataSF, “Traffic crashes resulting in fatality,” Open Data Portal, Accessed Apr 2025 30 Some collisions are not assigned fault and or do not have a fault listed in the public database. Enforcement is absent The SFPD does not meaningfully enforce traffic laws. As Supervisor Rafael Mandelman said on this topic, “[i]f it seems like there isn’t any traffic enforcement in San Francisco, that’s because there largely isn’t.”31 Statistics underlying the supervisor’s conclusion are stark. The number of traffic citations issued in 2024 was 87 percent lower than the number issued in 2014. At the four public hearings of the Public Safety and Neighborhood Services Committee of the San Francisco board of supervisors on the drop in traffic enforcement after the adoption of Vision Zero, SFPD identified several factors impacting the decline in citations, described below.32 The evidence suggests that, while the factors the SFPD cites do affect enforcement efforts, their aggregate impact is on the margins. Through numerous interviews, the Jury found that the primary reason for the stunning drop in citations is that police leadership does not prioritize traffic enforcement and does not hold officers accountable for performing what has historically been a part of an officer’s day-to-day job. Over time, this lack of prioritization and accountability has metastasized into a seemingly broad acceptance within the SFPD culture that traffic enforcement is not a valued part of an officer’s job. The result is the virtual abdication by SFPD of its essential role in keeping our streets safe. Enforcement saves lives What is intuitive to most people is also supported by research. Traffic enforcement decreases the severity of injuries and fatalities.33 When there are increases in citations issued, there are 31 “Supervisor Mandelman announces hearing announces hearing on San Francisco police department’s plan to restore traffic enforcement,” Press Release, Sep 12, 2023 32 These hearings were held on October 13, 2022, September 28, 2023, April 24, 2024 and December 12, 2024. James Davis et al., “Aggressive traffic enforcement: A simple and effective injury prevention program,” The Journal of Trauma: Injury, Infection, and Critical Care 60(2006): 972-977 decreases in vehicle collisions and injuries, with a bigger impact on fatalities than on less serious injuries.34 Larger increases in enforcement have a bigger impact than small increases.35 Enforcement of a range of serious violations has a greater impact than targeted enforcement on one type of violation, e.g., speeding.36 The converse is also true: when enforcement falls off, fatalities and severe injuries increase.37 San Franciscans have witnessed a spectacular decline in the number of traffic citations issued by SFPD in the 10 years since Vision Zero was adopted. During that time, the number of issued citations dropped by 95 percent, from more than 120,000 citations in 2014 to a little over 4,000 in 2022, before recently creeping back up to 15,500 in 2024. The initial decline in enforcement (2016–2019) corresponds to a period of increased serious injuries from vehicle collisions. Factors cited by SFPD for the decline in traffic citations At the September 28, 2023 hearing of the Public Safety and Neighborhood Services Committee of the board of supervisors, the SFPD described four primary factors as having impacted citation totals over the past decade: officer staffing; expanding administrative workload; pandemic-related changes; and educational enforcement.38 The SFPD spokesperson publicly acknowledged that these factors do not fully account for the gravity of the decline. Below, we consider each of the factors. Officer staffing Sworn officer staffing is down 9 percent from 2014. The traffic company, a division of the SFPD whose responsibilities include traffic enforcement as well as escorts, major events, 34 Rune Elvik, “A comparative analysis of accident modification functions for traffic law enforcement,” Accident Analysis and Prevention, 195(2024): 107415 35 ibid 36 ibid 37 ibid; City of Austin, “Special Report on the Effects of Traffic Patrol on Road Safety,” City of Austin Office of the City Auditor, Nov 2023 38 Board of Supervisors, “Hearing of the Public Safety and Neighborhood Services Committee,” Granicus, Sep 28, 2023 traffic collisions, and investigations, is down by 40 percent, to 46 in 2023 from 77 sworn officers in 2014. Citations have, however, not historically correlated with officer staffing, as shown in Figure 4.39 Citations had their first big drop in 2016, as staffing increased by 6 percent. As citations declined each year after, staffing levels only dropped below 2014 levels in 2022. This divergence strongly suggests that the problem is one of management emphasis, not lack of personnel. Figure 4. Traffic citations and officer staffing (SFPD 2025 & 2024) SFPD also has asserted that officers are required to spend more time responding to 911 calls from citizens and have less “discretionary” time to do other police work, such as traffic enforcement. However, as shown in Figure 5, police-dispatched 911 calls have dropped by 36 percent from 2015 levels, a larger decrease than staffing declines.40 39 SFPD, “Traffic Violation Reports,” Accessed March 2025; SFPD, Traffic enforcement updates, Dec 12, 2024 40 DataSF, “Law enforcement dispatched calls for service by year,” Open Data Portal, Accessed Mar 2025 Figure 5. 911 calls that were dispatched to the SFPD Administrative workload Administrative requirements have increased since 2014, including reporting for the Racial and Identity Profiling Act (RIPA) of 2015 and body-cameras protocols. While SFPD claims these added 10–15 minutes to each stop, our investigation suggested that the actual increase is closer to 5 minutes, once officers are in the habit of doing the work. Regardless, the increased administrative burden could only account for a portion of the dramatic drop in enforcement. Pandemic-related changes The Jury learned that SFPD explicitly deprioritized traffic enforcement in 2020 in order to minimize transmission of COVID from the public to officers. In the September 2023 hearing of the Public Safety and Neighborhood Services Committee referenced earlier, the SFPD explained that due to the reduction in traffic, the city was also safer and fewer citations were needed.41 Previously cited fatality and severe injury data suggests the contrary. Traffic patterns did indeed change during the pandemic, and fewer collisions with vulnerable users in the years 41 Board of Supervisors, “Hearing of the Public Safety…” that followed suggest that SFMTA interventions did improve street safety. But injuries from vehicle-only collisions rebounded to pre-2020 levels very quickly and vehicle-only fatalities continued to increase, while enforcement levels continued to drop. Educational enforcement SFPD notes that citation data does not include traffic stops that improve traffic-related behavior but do not involve citations. Those are categorized as “educational enforcement” in which officers stop vehicles and give drivers advisements and warnings. According to SFPD traffic stop data, available for mid-2018 through 2023, warnings for moving violations add 32 percent more stops than traffic citations alone.42 Because the citation figures are already so low, this adds little to an already small number. And the citation to warning ratio for 2023 is very similar to pre-COVID years; so, the drop would be coming from an even higher level if warning stops were incorporated into the analysis. A lack of commitment to safer streets The Jury found the dramatic decline in SFPD traffic enforcement during the Vision Zero decade cannot be fully explained by staffing or other cited factors. Instead, it reflects a deeper cultural issue within SFPD: traffic enforcement is not prioritized or expected by leadership, and officers face no accountability for neglecting it. New officers received little traffic stop training or experience during and after COVID and now view enforcement work with reluctance. The Jury heard some officers are wary of citizen complaints and frustrated by the limitations of SFPD General Order 9.07, which limits enforcement for certain infractions to reduce racial disparities. The General Order, however, does not in any way impede enforcement against the most dangerous behaviors. The Jury also came to understand that making traffic enforcement a regular practice helps officers get 42 DataSF, “Police department stop data,” Open Data Portal, Accessed Mar 2025 in the habit of interacting with the public and that the body camera is good protection against complaints. The Jury came to understand that some officers are more productive in enforcing traffic safety, while others make few or no stops. For the traffic company, citations per officer fell from 500 per year in 2014 to 80 in 2022-2023, rebounding to about 160 in 2024. Among the precincts, citations per officer went from almost 60 per year in 2014 to between one and two in 2022–23, before rebounding to seven per year in 2024. With current staffing,43 citations could be restored almost to 2016 levels if active traffic company officers averaged approximately three citations per shift (fewer than the average five citations per day of the more productive traffic company officers) and other active officers averaged approximately one per working week. To return to a level of enforcement that is protective against severe and fatal traffic collisions, SFPD leadership would have to set clear goals and expectations around what level of enforcement is needed to make streets safe and how their officers will meet this need. Regular planned traffic enforcement operations can reacquaint patrol officers with the practice of traffic enforcement. To be most effective, enforcement operations should target a range of dangerous violations, such as driving under the influence (DUI), not yielding to pedestrians in the crosswalk, running red lights and speeding.44 To help rebuild a culture of traffic enforcement, SFPD would have to take ownership of street safety. Reporting regularly and at the highest level on the violations issued for traffic fatalities and severe injuries, as well as the numbers of citations issued, would help the SFPD make traffic enforcement into a key performance metric. Just like trends in crime, trends in traffic deaths and severe injuries are a police responsibility and a meaningful indicator of police effectiveness. Assuming 1200 active officers in the precincts and 18 active officers in the traffic company. Elvik, “A comparative analysis of…,” The de-prioritization of traffic safety extends to understaffing of key roles, including support for automated enforcement of red-light violations and investigation and analysis of severe and fatal traffic collisions. Our investigation found that there is a problem staffing the TCIU. Because the staff is so small, team members are all on call every day and night. Preparation for these roles takes several years and there are no replacement officers in the queue to backfill officers approaching retirement. The SFPD needs to plan for how it will provide continuous and sustainable staff coverage for these key traffic safety functions. Outdated technology wastes staff time As staffing levels have indeed declined in both the precincts and the traffic division, SFPD needs to do its work more efficiently. Instead, the Jury’s investigation revealed that SFPD has failed to plan and invest in an infrastructure to facilitate robust traffic enforcement. Technology for issuing traffic citations is cumbersome and unnecessarily slow. Because the e- cite system often can't print citations, many officers avoid it altogether, instead issuing handwritten tickets. These tickets must then be hand logged into SFPD reports--and eventually transcribed by the courts back into a database years later. The mobile interface for the reporting required by RIPA is also unwieldy and slow, further discouraging enforcement activities. For highly complex collision reporting, SFPD uses a fillable PDF technology that is prone to errors. The poor technology wastes valuable officer and staff time, discourages traffic enforcement, and reduces data transparency and quality. The Department of Technology and the Digital Services Team could provide support to the SFPD to audit their traffic-specific equipment and software toolchain to prioritize time-saving improvements. Although ideally the state would develop a robust collision reporting tool, the city should not wait.45 Collaboration with SFMTA 45 State of California, “Traffic collision report,” CHP 555, Accessed Apr 2025 and the San Francisco Department of Public Health (SFDPH) is essential to ensure any new system to report injury collisions also improves tracking efficiency and data quality. The engineering approach The theory behind the engineering approach to traffic safety is that engineers should design roads with the expectation that human drivers will make mistakes. This design-driven approach has natural limits: even if it stops ordinary mistakes from resulting in injury or death, it cannot adequately prevent the kinds of egregiously reckless behavior which enforcement helps to deter. Yet, as enforcement fell off over the last decade, SFMTA shouldered much of the burden of Vision Zero, and clearly feels the weight of that responsibility. SFMTA has made real progress. To reduce fatalities and make further reductions in severe injuries, SFMTA needs to be more efficient and timely in identifying and implementing much-needed safety improvements for city streets. Tools for safer streets SFMTA uses a number of engineering tools to make streets safer. The general principles of these interventions include increasing visibility of vulnerable road users, physically separating vehicles from vulnerable road users, keeping vehicles from speeding, and slowing vehicle turns. Slower speeds, i.e. not exceeding the established speed limit, both reduces the severity of collisions and gives drivers more time to react in order to avoid collisions. Interventions include the following: • Improved traffic lights include leading pedestrian intervals, which improve visibility for pedestrians by letting them out into the intersections before vehicles are allowed to turn; pedestrian countdowns; accessible pedestrian signals; and longer walk times allowing pedestrians to cross safely. • Crosswalk improvements include high-visibility crosswalks, recessed stop lines, and flashing beacons. These are all tools to increase the visibility of pedestrians. • Daylighting removes parking within 20 feet of a marked or unmarked crosswalk in the direction of vehicle travel to improve visibility of pedestrians. State Assembly Bill 413 makes parking in these zones illegal in California. SFMTA plans to implement the law by painting curbs red to identify the no-parking zones.46 • Hardened daylighting adds safety zones to visually and/or physically narrow the street, along with physical barriers such as concrete blocks, planters, bike corrals, plastic delineators, curb extensions, bulb-outs and pedestrian islands to reduce the crossing distance for pedestrians and slow vehicles through intersections and turns. • Road diets remove excess capacity on multi-lane streets in order to discourage excessive speeds. Narrowing overly wide travel lanes also helps to reduce speeding. • Turn calming uses rubber speed bumps, paint and plastic flex-hit posts to make vehicles take wider, slower turns at intersections. • Turn restrictions, including no turn on red or no turn during the bike or pedestrian signal, reduce interactions between vehicles and vulnerable road users at intersections. • Traffic calming includes bumps in the road to slow vehicles—speed humps, speed tables, speed cushions, and raised crosswalks are some of the numerous varieties—as well as other measures to slow traffic, primarily on single lane, lower volume streets. • Protected bike lanes can be separated from traffic by parked vehicles, concrete barriers, planters, or flex-hit posts. They both dissuade drivers from blocking the lanes and protect users from moving vehicles. • Slow streets reduce and slow vehicle traffic, making the street safer and more comfortable for vulnerable road users. They use a variety of design interventions 46 California Legislative Information, “Assembly Bill No. 413,” October 12, 2023 described above, along with turn restrictions and traffic diverters to restrict through- traffic. • Quick builds are corridor improvements that can use a variety of the paint and plastic approaches outlined above, along with transit boarding islands and adjustments to traffic signal timing and parking and loading. • Speed limit changes, signage, and radar speed signs can help reduce vehicle speeding. Since AB 43 passed in 2021, SFMTA has reduced speed limits on 46 miles of streets.47 SFMTA also oversees automated enforcement: red-light and illegal-right-turn cameras at 19 intersections mostly installed in 2020 and 33 speed cameras installed in 2025. Violations captured by the red-light cameras result in official citations issued by SFPD, including license penalties tracked by the DMV, while the speed cameras issue municipal citations assigned to the car, more like a parking ticket. Given the great number and variety of SFMTA interventions, it is difficult to provide a clear timeline for implementation. By the beginning of 2025, along with many more interventions, SFMTA had implemented leading pedestrian intervals on 1035 crosswalks, used turn calming on 223 intersections, painted 2,461 high visibility crosswalks, protected 51 miles of bikeways, completed 39 quick builds, installed many pedestrian islands and put a few roads on a diet.48 Several key safety interventions, such as red-light cameras, speed limit changes, and speed cameras required changes to state law that took five to ten years to pass and implement. SFMTA has provided data to demonstrate the collision reduction benefits of some types of these interventions, along with data on vehicle speeds, pedestrian or bike and vehicle interactions, close calls, vehicles yielding to pedestrians, and other safety metrics.49 The 2023 47 SF.gov, “Vision Zero benchmarking: street network speeds,” Accessed May 2025 48 SFMTA, “Vision Zero safe streets progress,” Accessed Apr 2025 49 SFMTA, “Neighborhood traffic calming projects,” Accessed Apr 2025; SFMTA, “2023 Slow streets evaluation report,” Accessed Apr 2025 evaluation of 18 corridor projects from 2017 through 2022 found that the total collision rate was reduced by 16 percent, bike collisions by 25 percent and pedestrian collisions by 35 percent.50 Safer streets require a proactive approach SFMTA prioritizes improvements along the high injury network (HIN), a map of streets and intersections where the majority of severe and fatal injuries have taken place.51 Past interventions, such as car-free spaces in Golden Gate Park and speed limit and traffic light changes on Sunset Blvd., have successfully removed portions of the roadway from the HIN, while many parts of the city remain stubbornly on the map (Figure 6).52 2015 2017 2022 Figure 6. High Injury Network maps for 2015, 2017, and 2022 (Vision Zero SF, Maps and Data) While an important tool, the HIN is by design backwards looking and reactive. SFDPH is now working with DataSF, with the support of SFMTA, to build a predictive model to understand where fatal and severe injuries are more likely to happen in the future.53 Such a model could 50 SFMTA, “Safe streets evaluation program 2023 annual report,” Sept 1, 20224 51 Another update is expected to be released sometime in 2025. Ricardo Olea, “2017-2022 San Francisco traffic crashes report,” May 2023 53 One untapped set of data to seed this model is the Waymo open-sourced road usage data. This data set may be too complex for city engineers to be able to use, but SFMTA could ask Waymo to pull out specific data points, such as density of pedestrian traffic, turn speeds of vehicles at different types of intersections, and incidence of high speed and reckless driving. help the city predict the impact of street design proposals and make better prioritization decisions, allowing for more efficient resource allocation.54 The problem remains that SFMTA is not well organized to quickly implement proactive and integrated interventions. The clearest example is the traffic calming division, which is reactive by design. That program responds to applications from residents and then carries out extensive surveys to determine what interventions are warranted. This favors wealthier neighborhoods and limits implementation speed. Mayor Breed stated that over 700 traffic calming devices had been installed over the Vision Zero decade, an average of 70 per year.55 By contrast, Boston recently moved to a proactive prioritization model—taking into account collision frequency, percentages of older residents, children, and people with disabilities, and racial and socioeconomic indicators—and is on track to install 1,500 speed humps in three years.56 SFMTA has just finalized the Biking and Rolling Plan, which is a very broad, city-wide, plan for safer streets.57 This is important, as fear of injury from vehicles is one of the most important barriers to biking and 80 percent of surveyed San Franciscans said they would bike more if they felt safer.58 In addition, research shows that protected bike infrastructure makes the city safer for all road users.59 As SFMTA considers how to prioritize implementation of the plan, there is an opportunity to consider each corridor in context and improve the safety benefits beyond the routes themselves. Miguel Álvarez, “Predicting traffic accident hotspots with special data science,” Carto, Dec 2, 2020 55 Office of former Mayor London Breed, “Mayor Breed and city officials mark 10 years of Vision Zero and next steps in street safety,” SF.gov, Mar 28, 2024 56 City of Boston, “How did you choose speed hump locations?,” Making Neighborhood Streets Safer, Accessed Apr 2025 57 SFMTA, “San Francisco’s biking and rolling plan,” Mar 2025 58 Lauren Pearson, et al., “Adults’ self-reported barriers and enablers to riding a bike for transport: a systemic review,” Transport Reviews, 43 (2023): 356-384; Rebecca Ashton-Dziedzan, “Adopting the biking and rolling plan: A safer, more connected San Francisco,” SFMTA News, Mar 12, 2025 59 Wesley Marshall and Nicholas Ferenchak, “Why cities with high bicycling rates are safer for all road users,” Journal of Transport & Health, 13 (2019): 100539 SFMTA generally considers one corridor, one street, sometimes even one intersection, at a time, which can raise fears that major changes will drive traffic and collisions to neighboring streets and crossings. An area-based approach would focus on improving road safety within a specific geographic area by implementing a range of interventions from the tools listed above, along with other street-use changes such as parking management, bus lanes, traffic signal timing, etc. By taking into account the overall street network, SFMTA can create a more comprehensive and effective strategy. This approach would benefit from a collaborative process—involving community stakeholders and SFMTA planners and engineers all working together for safer streets—to ensure that the designs are technically sound and responsive to the community. While broadening the scope may increase complexity, it could also widen the safety impact and broaden neighborhood support for projects, thus increasing the rate of implementation. Another important opportunity for SFMTA to take proactive measures to improve safety is in implementing California’s Daylighting Law (AB 413), which eliminates parking within 20 feet of a crosswalk to improve pedestrian visibility. Unfortunately, simply painting curbs red may not be sufficient to get real safety benefits. According to a January 2025 study by the New York City (NYC) Department of Transportation (DOT), just removing parking at intersections, without taking additional measures, can increase collision injuries.60 However, when coupled with complementary interventions, newly daylit intersections reduced risk of injury, and even more so for severe and fatal injuries. The most beneficial safety interventions assessed in an earlier study by NYCDOT are included in Table 1. Table 1. NYCDOT reductions in pedestrian injuries for specific safety treatments (DOT, 2022)61 Safety Treatment Pedestrian Injury Change Pedestrian Severe & Fatal Injury Change Protected Bike Lanes -17.8% -29.2% Pedestrian Islands -10.2% -29.9% 60 NYC DOT, “Daylighting and street safety,” 2025 43 NYC DOT, “Safety treatment evaluation (2005-2018),” 2022 Road Diets -12.5% -31.7% Turn Calming -17.5% -32.7% Leading Pedestrian Intervals -18.1% -34.3% Curb & Sidewalk Extensions -16.5% -44.7% The lesson from NYC is clear—to get the full intended benefits, daylighting must be hardened through other interventions to narrow the roadway and slow travel through the intersection. SFMTA has used each of these interventions in many locations throughout the city and has collected data to demonstrate safety benefits.62 SFMTA has also identified at least some non- hardened daylighted intersections where collisions have increased.63 SFMTA currently plans to comply with AB 413 by painting daylit zones red to give guidance for vehicle parking, as requested by Resolution 632-24 of the board of supervisors.64 By revisiting this plan to consider incorporation of hardening practices, SFMTA could ensure that this important law achieves the sought-after safety benefits. Building trust through neighborhood-based outreach In West Portal, after the horrific, deadly crash outside the public library in March 2024, there was a strong push to make the intersection of West Portal and Ulloa safer. The SFMTA straw proposal released for community feedback was taken by one group of residents and business owners as a done deal that would kill small businesses.65 To gather input from trusted neighborhood leaders and representatives, District 7 Supervisor Melgar helped to convene a working group to engage with SFMTA planners and engineers. One of the big topics discussed by the group was touched upon above—if changes are made 62 SFMTA, “Left turn safety,” Accessed Apr 2025 63 SFMTA, “Tenderloin daylighting project,” Accessed Apr 2025 64 Board of Supervisors, “Urging Implementation of Red Zones for Daylighting at All Crosswalks and Intersections,” Resolution No. 632-24, Dec 20, 2024 65 Bay City News, “West Portal business owners lambast city over proposed traffic plan,” Kron 4, Apr 25, 2024 to this one intersection, won’t all the problems just move down West Portal to another intersection? It turned out that many of the intersections along and just beyond West Portal had been asking and waiting for traffic calming for a long time. By broadening the scope of the plan, they were able to provide safety benefits all along the corridor, while still maintaining almost all the benefits of the original straw proposal. By working directly with neighborhood stakeholders, SFMTA was able to implement important safety interventions on a timely basis. The Jury’s investigation revealed that community input and outreach is critical to SFMTA success in promoting street safety. The Jury also heard that community involvement takes a huge amount of time. We heard that SFMTA outreach events often seek open-ended input, rather than clearly explaining needs and narrowly defining queries. As a result, there have been wild swings on high profile projects and there can be a lot of mistrust of SFMTA as a partner on street safety. SFMTA’s lack of effective community outreach substantially slows down progress on safer streets. A proactive outreach approach could help build trust and expedite and improve projects. A dedicated outreach contact per supervisorial district could foster long-term relationships with both the supervisor and key neighborhood stakeholders who are committed to safer streets. This SFMTA contact person could integrate community input into safe street projects, speeding implementation and improving outcomes. Expanding community involvement through tangible activities to improve street safety, such as installation of planters or benches as part of daylighting or slow streets, could further bolster community support and trust. A strict neighborhood focus could, however, leave out street users that are traveling to or passing through a neighborhood. Broader input on such things as perceived street safety or inefficient light timing could help inform future projects or lead to spot fixes to improve overall street safety and comfort. By adding a street safety routing to the existing city services tool, 311, SFMTA could provide residents with a channel to offer feedback on street safety and traffic flows. Outreach also is hindered by the negative focus of the key Vision Zero metric. While zero fatalities is an important aspirational goal, setting a deadline for zero is a set-up for failure. Instead, SFMTA needs clear, common-sense metrics to demonstrate progress on safer streets. Tracking the collision reduction benefits of different interventions, following NYCDOT’s example, would help increase trust. Additional metrics such as incidence of excessive speeding, average turn speed, and yielding to pedestrians could be used to transparently track and communicate progress, helping to increase public support for safer streets. More education, please! When it launched in 2014, education was one of the three pillars of the Vision Zero initiative.66 Enforcement and education work well together—education can extend the reach and reinforce the messaging of enforcement. Education is also needed to help communicate the whys and hows of engineering decisions that result in street changes. SFMTA generally has taken the lead on education efforts, ranging from broad public campaigns to classes for specific user groups. The Jury heard from several interviewees that education was a missing link in street safety. Broad public outreach not only teaches people how to navigate streets safely but also forges a shared commitment to street safety. Hands-on training is by far the most effective method, and the city should build upon existing programs to get kids and families safely walking, biking and taking transit to school. Finally, the Jury noted important gaps in knowledge of Vision Zero among different city agencies. Board of Supervisors, “Implement a Vision Zero three point plan: engineering, education and enforcement,” Resolution 91-14, Mar 2014 Broad public campaigns Early Vision Zero public education campaigns by SFMTA were impactful but have declined in recent years. This despite evidence that sustained outreach efforts are essential for maintaining safety benefits.67 In January 2025, SFMTA launched a highly visible citywide campaign to announce speed cameras, similar to its 2021 campaign about lower speed limits in the Tenderloin (Figure 7). Figure 7. Educational campaigns from 2025 and 2021 regarding speed enforcement and limit changes Vision Zero’s early public education campaigns worked closely with project partners such as the SFPD and SFDPH, as well as community partners like Walk SF and the San Francisco Bicycle Coalition (SFBC). These campaigns were comprehensive, including different outreach modes, and often incorporated an enforcement component. Past notable examples include: • Be Nice, Look Twice was launched at the start of the Vision Zero initiative (Figure 8). SFMTA found that, “After the campaign, nearly 80 percent of San Franciscans reported knowing that pedestrian safety in the city is an issue that needs to be tackled.”68 67 Vision Zero, “Safe Speeds SF High Visibility Enforcement Campaign Findings,” SFMTA, Nov 2019; Sara Waring, et al., “Examining the effectiveness of an education-based road safety intervention…,” Traffic Psychology and Behavior, 105(2024): 336 68 SFMTA, “Be Nice, Look Twice Pedestrian Safety Campaign,” Accessed Mar 2025 • It Stops Here addressed drivers’ failure to yield to pedestrians at crosswalks.69 • Safe Speeds SF Campaign combined outreach with enforcement, issuing 1800 speeding tickets along 11 corridors over 12 months.70 Figure 8. Be Nice, Look Twice campaign Broad public education campaigns can inform the public about how to use streets more safely, what enforcement measures to expect, and how to navigate major roadway changes, such as traffic diversions and slow streets. In addition to reducing risky behavior, education can help build a shared vision of safer streets and increase buy-in for street safety improvements. Training the next generation of safe road users Traffic-related dangers are identified as a key barrier to walking to school.71 SFMTA co-leads a Safe Routes to School program, in partnership with the San Francisco Unified School District (SFUSD) and the San Francisco County Transportation Authority (SFCTA).72 This program helps families safely walk, bike and take MUNI to school, encouraging bike trains and walking school buses—groups that bike or walk together for safety, community-building, and exercise. Several non-government organizations (NGOs) also provide valuable street safety education: 69 SFMTA, “Safe Streets Campaign,” Accessed Mar 2025 70 Vision Zero, “Safe Speeds SF High Visibility Enforcement Campaign Findings,” SFMTA, Nov 2019 71 John Omura et al., “Prevalence of children walking to school and related barriers—United States, 2017,” Preventative Medicine, 118 (2019): 191-195 72 “Safe Routes to School San Francisco,” Accessed Mar 2025 • Y-Bike (YMCA) partners with the SFUSD and afterschool providers, funded by the SFMTA to teach kids safe biking. • The SFBC offers free classes and multilingual resources to teach kids, families and adults how to bike safely in the city.73 • Walk SF leads a Safe Routes to School program, which provides training and tools in multiple languages to help families start walking school buses.74 • SF Bike Bus is a grassroots movement to help communities form and maintain bike buses to schools.75 The Jury learned that kids being able to safely walk, bike and take MUNI to school was a big priority for the city.76 To help achieve this goal, street safety could be more fully incorporated into the SFUSD curriculum. Getting the messaging right is important for education to be effective.77 As SFMTA expands their traffic-calming program to focus more on schools, there also is an opportunity to lean into Safe Routes to School to increase outreach around street safety improvements.78 Training specific user groups SFMTA also offers and supports specialized training and classes for specific users. In collaboration with SFPD, free in-person motorcycle safety classes are offered to the public.79 For large urban vehicles, SFMTA offers a video and training curriculum, which is required for all city contractors.80 In 2024, two of the 42 Vision Zero fatalities were the result of collisions 73 SFBC, “Resources,” Accessed Mar 2025 74 Walk SF, “Safe Routes to School,” Accessed Apr 2025 75 “SF Bike Bus,” Accessed Mar 2025 76 SFMTA, “SFMTA school safety programs,” Accessed May 2025 77 Sara Waring, et al., “Examining the effectiveness of an education-based road safety intervention…,” Traffic Psychology and Behavior, 105(2024): 347 78 Rebecca Ashton-Dziedzan, “Making trips to school safer across the city: A new daylighting program,” SFMTA, Apr 17, 2025; SFMTA, “School safety programs,” Accessed Apr 2025 79 SFMTA, “Motorcycle Safety,” Accessed Mar 2025 80 SFMTA, “Large Vehicle Urban Driving Safety Program,” Accessed Mar 2025 involving city employees, leading the Jury to consider what more could the city do to align these high-profile and high-mileage users around safe street goals.81 The State of California Commission on Peace Officer Standards and Training (POST) requires all police officers to take a driver training class every two years, as part of their Perishable Skills Program.82 While the training covers all the vehicle codes that have to do with liability of public employees, it does not cover the toll of traffic collisions or the highest risk violations of the traffic code. The SFPD could benefit from additional department training to help officers align traffic enforcement with the highest risk behaviors and understand the critical role of the SFPD in improving street safety. In support of Vision Zero goals, the city has required use of devices to track driving behaviors such as speeding, hard acceleration and hard braking, known as telematics, in city vehicles since 2014.83 Although most non-exempt vehicles are now equipped with the systems, the original deadline of 2020 has slipped to the end of 2025.84 Pairing telematics data with training can greatly reduce serious collisions.85 The pieces are largely in place to make use of telematics to reduce city vehicle-involved collisions, but there is no citywide training guidance on use of telematics to help improve driver behavior. Gabe Greschler et al., “Cyclist killed after crash with city vehicle identified by officials,” The San Francisco Standard, Jun 17, 2024; Eleni Balakrishnan, “S.F. police driver strikes, kills man with car in Bayview,” Mission Local, Dec 14, 2024 82 POST, “POST Perishable Skills Program (PSP),” May 2021 83 Budget and Legislative Analyst’s Office, “Vehicle telematics update,” Board of Supervisors, Aug 19, 2020 84 Noah Baustin, “‘It’s freaking horrific’: City workers are running over pedestrians at alarming rate,” The San Francisco Standard, Dec 18, 2024 85 Leverson Boodlal and Kun-Hung Chiang, “Study on the impact of a telematics system on safe and fuel-efficient driving in trucks,” U.S. Department of Transportation, Apr 2024 Findings and Recommendations The Jury made the following findings and recommendations regarding the city’s process of advancing Vision Zero street safety goals.
Related Recommendations (1)
R5:
1 By March 31, 2026, SFMTA should develop a new street safety education campaign that builds on earlier efforts. The new campaign should focus on ways all road users can prevent the traffic violations that result in the most severe and fatal injuries.
Quick View
Full Details →
Findings & Recommendations
5 findings
F1:
SFPD Fails to Adequately Enforce Traffic Laws ........................................................ 34
Related Recommendations (1)
R1:
1 By September 30, 2025, the mayor should direct the chief of police to develop a plan to specify the level of traffic enforcement activities needed to achieve safer streets. The plan should describe the appropriate frequency of traffic enforcement operations, the extent to which officers should engage in traffic enforcement as part of their duties, any barriers to traffic enforcement, and what active training may be required to support enforcement activities. The plan should be completed by March 31, 2026.
F2:
Traffic Enforcement Is Hampered by Outdated Technology .................................... 35
Related Recommendations (1)
R2:
1 By December 31, 2025, the mayor should direct the chief of police to audit SFPD traffic-specific equipment and software and determine how to expedite deployment of improved software and hardware for citation processing, collision reporting and any other traffic-related administrative duties. SFPD should report back on the proposed implementation of improvements recommended by their audit by December 31, 2026.
F3:
SFMTA Lacks Alignment on Proactive Vision Zero Planning ................................... 35
Related Recommendations (1)
R3:
1 By March 31, 2026, SFMTA should update its planning and site selection process for traffic calming. The process should prioritize proactive intervention selections, informed by data about where the greatest safety benefits are likely to be and where they are needed to support other street safety goals, over reactive selections driven by individual citizen applications.
F4:
Sustained, Community-Focused Outreach is Lacking .............................................. 36
Related Recommendations (1)
R4:
1 By March 31, 2026, SFMTA should assign a point of contact for outreach within each supervisory district, allowing SFMTA to work more closely with members of the board of supervisors; and to build and maintain relationships with key stakeholders in the community on all roadway and transit issues.
F5:
Education Fosters Citizens’ Commitment to Safe Streets ....................................... Required and Requested Responses ............................................................................................. Methodology .................................................................................................................................... Glossary and Acronyms .................................................................................................................. Background Motor vehicle collisions kill more than a million people per year worldwide, including more than 40,000 per year in the United States. Vision Zero is a branding term for a broad policy concept: that all these traffic fatalities should be treated as systems design failures and prevented through systemic fixes. Vision Zero was first adopted as national policy in Sweden in 1997. Sweden’s approach was multi-pronged, including engineering safer roads, slowing speeds, improving vehicle design, educating people and enforcing laws. The results have been remarkable: Sweden’s traffic fatality rate has declined by more than 65% since adoption of this approach, and it currently has the second lowest fatality rate in the Organization for Economic Cooperation and Development (OECD).1 While San Francisco had been taking measures to improve street safety for many years, the full embrace of Vision Zero began in 2014, when a board of supervisors’ resolution, adopted unanimously, urged the mayor, the chief of police and the director of the San Francisco Municipal Transportation Agency (SFMTA) to “implement an action plan to reduce traffic fatalities to zero in the next ten years through better engineering, education, and enforcement.”2 Vision Zero was adopted by the city that same year, with a commitment to eliminate traffic fatalities and greatly reduce severe injuries by 2024. By 2017, 17 department heads signed onto the Vision Zero plan, signaling a citywide commitment to achieve the Vision Zero goal.3 1 OECD, “Road accidents,” For year 2022, Deaths per 1 000 000 inhabitants, Accessed Apr 2025 2 Board of Supervisors, “Implement a Vision Zero three point plan: engineering, education and enforcement,” Resolution 91-14, Mar 2014 3 Office of the Mayor, “Mayor Lee celebrates Vision Zero progress,” April 9, 2015; Vision Zero SF, “Vision Zero San Francisco two-year action strategy 2017-2019,” Accessed Apr 2025 More than a decade later, not only have citywide traffic fatalities not gone to zero, they have not even decreased. The Jury inquired into the root causes of this failure, seeking to understand what a realistic improvement to street safety in San Francisco might look like and where the city has fallen short. Street design and traffic laws are contentious political issues in San Francisco, as in many places. The Jury’s members came to this investigation with different perspectives, but we found common ground on this: while zero traffic fatalities may be an unrealistic near-term goal, the city could and should be doing a lot more to improve street safety. All three pillars of the Vision Zero strategy—enforcement, engineering, and education—need attention. To inform the findings of this report, we need to examine the context—legal and physical—of street safety in San Francisco. That context poses many obstacles to improvement. We’ll also need to discuss the varying causes, consequences and frequencies of different types of traffic collisions. To do that we need to define two key terms: • Vulnerable road users in this report only include the most vulnerable: pedestrians, bicyclists, scooter riders and other users of small, low-powered devices.4 • Vehicle-only collisions are collisions that involve one or more vehicles, but no vulnerable road users (for example, two vehicles colliding or a single vehicle hitting a tree).5 4 Vision Zero SF, “A system of safety,” Accessed May 2025 5 While motorcyclists are often considered vulnerable road users, motorcycles are included as vehicles in the vehicle collision database. Thus, in this report, motorcycles are included in vehicle-only collisions and are not broken out separately; DataSF, “Traffic crashes resulting in injury,” Open Data Portal, Accessed Jan 2025 The headwinds San Francisco is facing strong headwinds in its effort to reduce traffic fatalities and severe injuries, with significant challenges beyond city control. While these factors should not be excuses for inaction, acknowledging them is essential to making progress. Traffic fatalities in San Francisco are not isolated from national trends. Traffic deaths in the U.S. had been declining since the 1990s but began rising in 2014—contrary to continued declines in other wealthy countries. Since 2014, U.S. traffic fatalities have increased by about 26%. The New York Times estimates that if the U.S. had matched the continued decline in Europe, 25,000 fewer Americans would die annually.6 Figure 1. Trends in traffic fatalities per 1 million residents (by the New York Times, Source OECD)7 While road design and enforcement—factors largely within local control—play a role, a major reason for the disparity lies in vehicle design, a factor outside of San Francisco’s control. Compared to Europe, U.S. passenger vehicles are on average heavier (29%), larger (17%), and more powerful (77%).8 Trucks and SUVs make up 73% of U.S. new vehicle sales, compared to 6 David Leonhardt, “The rise in U.S. traffic deaths,” New York Times, Dec 11, 2023 7 ibid 8 ICCT, “European vehicle market statistics,” Pocketbook 2023/24, Accessed Apr 2025, pg. 44 (year 2022); EPA, “The 2023 EPA automotive trends report,” Dec 2023, pg. 36 (year 2022) 50% in the European Union (EU).9 These trends increase the likelihood and severity of all collisions.10 Speed remains the primary factor in crash severity. More powerful vehicles accelerate faster and, along with heavier vehicles, are more likely to speed excessively.11 Faster and heavier vehicles take longer to stop—at 40 mph, stopping distance is two and a half times longer than at 20 mph.12 A pedestrian struck at 30 mph is up to eight times as likely to die as one struck at 20 mph.13 Larger vehicles also have higher, blunter front ends and wider roof supports, which reduce visibility of all other road users and increase severity of collisions, especially for pedestrians.14 Compounding the issue, U.S. vehicle safety regulations lag those in Europe. The EU first issued a directive requiring pedestrian-friendly vehicle designs in 2003. The U.S. only mandated automatic braking in 2023. Pedestrian-specific testing requirements were not even proposed by the U.S. National Highway Traffic Safety Administration (NHTSA) until 2024 and still have not been adopted. 9 ibid, pg. 12 of ICCT and pg. 17 of EPA 10 Michael Anderson, “Safety for whom? The effects of light trucks on vehicle fatalities,” Journal of Health Economic, July 2008; Insurance Institute for Highway Safety (IIHS), “Supersizing vehicles offers minimal safety benefits—but substantial dangers,” Feb 2025 11 Anne McCartt & Wen Hu, “Effects of vehicle power on passenger vehicle speeds,” Traffic Injury Prevention, May 2016; Aviv Steren et al., “Is vehicle weight associated with risky behavior? Analysis of complete national records,” Mark Lett, Mar 2025 12 NHTSA, “Speed-measuring device operator training,” Participant manual, 2018; Samir A.E. Mohamed, et al., “Safe driving distance and speed for collision avoidance in connected vehicles,” Sensors, Sep 2022 13E. Pasasen and H. Salmivaara. “Driving speeds and pedestrian safety in the City of Helsinki,” Traffic Eng. Control, 1993; Erik Rosén and Ulrich Sander, ”Pedestrian fatality risk as a function of car impact speed,” Accident Analysis & Prevention, May 2009; U.S. DOT, “Chapter 5: Risk factors other than exposure,” Federal Highway Administration, Accessed Apr 2025 14 Keith Barry, “The hidden dangers of big pickup trucks,” Consumer Reports, Aug 19, 2024; Wen Hu, at al., “The association between passenger-vehicle front-end profiles and pedestrian injury severity in motor vehicle crashes,” J. of Safety Res., Sep 2024; Wen Hu & Jessica Cicchino, "Relationship of pedestrian crash types and passenger vehicle types,” J Safety Res., Sep 2022 State and federal laws and regulations offer further obstacles to local safety efforts. SFMTA doesn’t have full authority over many major streets, like 19th Avenue, because they double as state or federal highways.15 Until Assembly Bill (AB) 43 came into force in 2022, the city couldn’t even adjust speed limits on local streets to protect pedestrians.16 While red light cameras have been legal in California since 1995, speed cameras were only approved by AB 645 in October 2023 after a decade of attempts and advocacy.17 Finally up and running in 2025, the five-year pilot allows only 33 speed cameras in San Francisco and strictly limits penalties and fines, as well as revenue collection and spending.18 In September 2024, Governor Newsom vetoed Senate Bill 961, which would have required new vehicles to alert drivers when exceeding the speed limit by ten miles per hour.19 A last point here: the California Department of Motor Vehicles (DMV) allows many reckless drivers to remain licensed.20 A potentially brighter future San Francisco residents certainly have noticed the driverless vehicles on our streets. These vehicles can make some mistakes uncommon among human drivers, but they reliably stop at stop signs and red lights, yield to pedestrians, and comply with speed limits.21 And because of their wider vision and faster reaction times, recent analysis suggest that some kinds of autonomous vehicles have been able to avoid many collisions, including those that would have been the fault of others.22 However, not all self-driving technologies are created equal.23 15 SFMTA Board of Directors, “Resolution No. 170516-062,” May 2017 16 Assembly Committee on Transportation, “AB 43—as amended March 22, 2021,” Memo, Apr 19, 2021 17 Heather Knight, “California allows cities to catch speeding drivers with automated cameras,” New York Times, Oct 2023 18 California Legislative Information, “AB-645 Vehicles: speed safety system pilot program,” Oct 16, 2023 19 Soumya Karlamangla, “Newsom vetoes bill requiring cars to warn speeding drivers,” New York Times, Sep 28, 2024 20 Robert Lewis, “License to kill,” CalMatters, Apr 2025 21 Richard Cano, “Yes, there are more driverless Waymos in S.F.,” San Francisco Chronicle, Jul 24, 2024 22 Di Lillo, et al., “Do Autonomous Vehicles Outperform Latest-Generation Human-Driven Vehicles? A Comparison to Waymo's Auto Liability Insurance Claims at 25.3M Miles,” 2024 23 NHTSA, “Additional information regarding EA22002,” Apr 25, 2024 Automated safety technologies on human-controlled vehicles, such as lane assist, automated braking, and intelligent speed assistance, are also becoming much more widespread. Estimates of the potential impact of these technologies range from a 6.7 percent reduction in fatal and injury collisions to a 22 percent reduction in fatalities.24 Automated braking is highly effective in protecting against collisions with other vehicles when traveling at moderate speeds, but is still less reliable in protecting pedestrians, especially at night, and does not protect against reckless driving.25 There is strong consistent evidence that speed cameras reduce vehicle speeds, lower crash rates and save lives; the Jury is encouraged by the San Francisco debut of 33 new speed cameras.26 24 Nordic Road and Transport Research, “Safety impact of advanced driver assistance systems in Europe,” Mar 19, 2024; AAA, “Advanced driver assistance systems,” 2025 25 IIHS, “Pedestrian crash avoidance systems cut crashes—but not in the dark,” Feb 3, 2022; AAA, “Progression of automatic emergency braking technology,” NewsRoom.AAA.com, Oct 2024 26 Wilson et al., “Speed cameras for the prevention of road traffic injury and death,” Cochrane Database of Systematic reviews, Nov 10, 2010 Analysis Drawing upon collision trends, research and interviews, our analysis suggests the severe drop in enforcement, beginning in 2016, has increased risks to all road users. SFMTA’s engineering interventions have followed the opposite trend. Increasing over the Vision Zero decade, SFMTA has provided some protection for vulnerable road users, even with the headwinds the city faces. The analysis presented below begins with an examination of collision trends over the Vision Zero period. The collision databases from the DataSF Open Data Portal demonstrate that total injuries (and the subset of severe injuries) from all vehicle collisions were higher from 2016 to 2019, but they dropped in 2020 when COVID reshaped traffic patterns.27 In 2021 injuries from vehicle-only collisions quickly returned to almost peak levels, while injuries from vehicle collisions with vulnerable users remained lower. Collision trends are followed below by Jury analysis for each of the three Es: enforcement, engineering and education. While traffic safety researchers have called for more comprehensive frameworks—such as the safe systems pyramid, which incorporates socioeconomic factors, vehicle design, and more—the three E approach that served as the starting point for San Francisco’s Vision Zero policy continues to fit the city’s implementation.28 It might not be sufficient to achieve the original zero fatality goal, but the Jury found that improving the effectiveness of the three Es has the power to make San Francisco streets substantially safer. We have come to think of the three Es as a three-legged stool: if any leg is removed, the stool ceases to function. DataSF, “Traffic crashes resulting in injury,” Accessed March 2025 28 David Ederer et al., “The Safe Systems Pyramid: A new framework for traffic safety,” Transportation Research Interdisciplinary Perspectives, 21(2023): 100905 Collision trends Vehicle-only collisions are responsible for 65 percent of the total injuries over 2014 to 2024. Vehicle collisions with pedestrians account for only 20 percent of the total, but these collisions are much more deadly. Over this period, 23 of every 1000 injuries in vehicle-pedestrian collisions were fatal, in comparison to 4 out of every 1000 injuries in vehicle-only collisions. Total injuries from vehicle collisions dropped by 30 percent in 2020; in the next year, vehicle- only collisions rebounded quite a bit, while vehicle collisions with vulnerable users have stayed lower (Figure 2). Vehicle-only Vehicle & vulnerable user Figure 2. Injuries from vehicle-only collisions and collisions between vehicles and vulnerable road users. Note differences in scale. (DataSF, 2025) While a goal of zero was only established for fatalities, Vision Zero targets both fatal and severe injuries. Unfortunately, severe injuries get less attention in official Vision Zero documents than fatalities. While fatal injuries were not clearly impacted, severe injuries for all vehicle-related collisions experienced a substantial drop in 2020 due to the pandemic lockdown. • Severe injuries from vehicle-only collisions quickly rebounded to pre-2020 levels, with fatal injuries trending slightly higher over the decade. • Severe injuries in vehicle collisions with vulnerable road users largely maintained the 2020 drop until increasing in 2024 to 2015 levels (still lower than 2016 to 2019). Fatalities were higher for both categories in 2024 than in any other Vision Zero year. In total, fatalities in 2024 were almost 50 percent higher than the average over the decade.29 As traffic patterns continue to adjust to the post-pandemic world, it is unclear if the high fatalities of 2024 are an outlier or a warning of what is to come. SFPD’s Traffic Collisions Investigative Unit (TCIU) assigns fault for almost every severe and fatal collision.30 Looking at all fault-assigned collisions, drivers were assigned fault for almost 80 percent of the severe and fatal injuries overall and 60 percent of collisions between vehicles and vulnerable users (Figure 3). Figure 3. At-fault in all severe and fatal collisions involving vehicles, 2014-2024 (DataSF, 2025) 29 DataSF, “Traffic crashes resulting in fatality,” Open Data Portal, Accessed Apr 2025 30 Some collisions are not assigned fault and or do not have a fault listed in the public database. Enforcement is absent The SFPD does not meaningfully enforce traffic laws. As Supervisor Rafael Mandelman said on this topic, “[i]f it seems like there isn’t any traffic enforcement in San Francisco, that’s because there largely isn’t.”31 Statistics underlying the supervisor’s conclusion are stark. The number of traffic citations issued in 2024 was 87 percent lower than the number issued in 2014. At the four public hearings of the Public Safety and Neighborhood Services Committee of the San Francisco board of supervisors on the drop in traffic enforcement after the adoption of Vision Zero, SFPD identified several factors impacting the decline in citations, described below.32 The evidence suggests that, while the factors the SFPD cites do affect enforcement efforts, their aggregate impact is on the margins. Through numerous interviews, the Jury found that the primary reason for the stunning drop in citations is that police leadership does not prioritize traffic enforcement and does not hold officers accountable for performing what has historically been a part of an officer’s day-to-day job. Over time, this lack of prioritization and accountability has metastasized into a seemingly broad acceptance within the SFPD culture that traffic enforcement is not a valued part of an officer’s job. The result is the virtual abdication by SFPD of its essential role in keeping our streets safe. Enforcement saves lives What is intuitive to most people is also supported by research. Traffic enforcement decreases the severity of injuries and fatalities.33 When there are increases in citations issued, there are 31 “Supervisor Mandelman announces hearing announces hearing on San Francisco police department’s plan to restore traffic enforcement,” Press Release, Sep 12, 2023 32 These hearings were held on October 13, 2022, September 28, 2023, April 24, 2024 and December 12, 2024. James Davis et al., “Aggressive traffic enforcement: A simple and effective injury prevention program,” The Journal of Trauma: Injury, Infection, and Critical Care 60(2006): 972-977 decreases in vehicle collisions and injuries, with a bigger impact on fatalities than on less serious injuries.34 Larger increases in enforcement have a bigger impact than small increases.35 Enforcement of a range of serious violations has a greater impact than targeted enforcement on one type of violation, e.g., speeding.36 The converse is also true: when enforcement falls off, fatalities and severe injuries increase.37 San Franciscans have witnessed a spectacular decline in the number of traffic citations issued by SFPD in the 10 years since Vision Zero was adopted. During that time, the number of issued citations dropped by 95 percent, from more than 120,000 citations in 2014 to a little over 4,000 in 2022, before recently creeping back up to 15,500 in 2024. The initial decline in enforcement (2016–2019) corresponds to a period of increased serious injuries from vehicle collisions. Factors cited by SFPD for the decline in traffic citations At the September 28, 2023 hearing of the Public Safety and Neighborhood Services Committee of the board of supervisors, the SFPD described four primary factors as having impacted citation totals over the past decade: officer staffing; expanding administrative workload; pandemic-related changes; and educational enforcement.38 The SFPD spokesperson publicly acknowledged that these factors do not fully account for the gravity of the decline. Below, we consider each of the factors. Officer staffing Sworn officer staffing is down 9 percent from 2014. The traffic company, a division of the SFPD whose responsibilities include traffic enforcement as well as escorts, major events, 34 Rune Elvik, “A comparative analysis of accident modification functions for traffic law enforcement,” Accident Analysis and Prevention, 195(2024): 107415 35 ibid 36 ibid 37 ibid; City of Austin, “Special Report on the Effects of Traffic Patrol on Road Safety,” City of Austin Office of the City Auditor, Nov 2023 38 Board of Supervisors, “Hearing of the Public Safety and Neighborhood Services Committee,” Granicus, Sep 28, 2023 traffic collisions, and investigations, is down by 40 percent, to 46 in 2023 from 77 sworn officers in 2014. Citations have, however, not historically correlated with officer staffing, as shown in Figure 4.39 Citations had their first big drop in 2016, as staffing increased by 6 percent. As citations declined each year after, staffing levels only dropped below 2014 levels in 2022. This divergence strongly suggests that the problem is one of management emphasis, not lack of personnel. Figure 4. Traffic citations and officer staffing (SFPD 2025 & 2024) SFPD also has asserted that officers are required to spend more time responding to 911 calls from citizens and have less “discretionary” time to do other police work, such as traffic enforcement. However, as shown in Figure 5, police-dispatched 911 calls have dropped by 36 percent from 2015 levels, a larger decrease than staffing declines.40 39 SFPD, “Traffic Violation Reports,” Accessed March 2025; SFPD, Traffic enforcement updates, Dec 12, 2024 40 DataSF, “Law enforcement dispatched calls for service by year,” Open Data Portal, Accessed Mar 2025 Figure 5. 911 calls that were dispatched to the SFPD Administrative workload Administrative requirements have increased since 2014, including reporting for the Racial and Identity Profiling Act (RIPA) of 2015 and body-cameras protocols. While SFPD claims these added 10–15 minutes to each stop, our investigation suggested that the actual increase is closer to 5 minutes, once officers are in the habit of doing the work. Regardless, the increased administrative burden could only account for a portion of the dramatic drop in enforcement. Pandemic-related changes The Jury learned that SFPD explicitly deprioritized traffic enforcement in 2020 in order to minimize transmission of COVID from the public to officers. In the September 2023 hearing of the Public Safety and Neighborhood Services Committee referenced earlier, the SFPD explained that due to the reduction in traffic, the city was also safer and fewer citations were needed.41 Previously cited fatality and severe injury data suggests the contrary. Traffic patterns did indeed change during the pandemic, and fewer collisions with vulnerable users in the years 41 Board of Supervisors, “Hearing of the Public Safety…” that followed suggest that SFMTA interventions did improve street safety. But injuries from vehicle-only collisions rebounded to pre-2020 levels very quickly and vehicle-only fatalities continued to increase, while enforcement levels continued to drop. Educational enforcement SFPD notes that citation data does not include traffic stops that improve traffic-related behavior but do not involve citations. Those are categorized as “educational enforcement” in which officers stop vehicles and give drivers advisements and warnings. According to SFPD traffic stop data, available for mid-2018 through 2023, warnings for moving violations add 32 percent more stops than traffic citations alone.42 Because the citation figures are already so low, this adds little to an already small number. And the citation to warning ratio for 2023 is very similar to pre-COVID years; so, the drop would be coming from an even higher level if warning stops were incorporated into the analysis. A lack of commitment to safer streets The Jury found the dramatic decline in SFPD traffic enforcement during the Vision Zero decade cannot be fully explained by staffing or other cited factors. Instead, it reflects a deeper cultural issue within SFPD: traffic enforcement is not prioritized or expected by leadership, and officers face no accountability for neglecting it. New officers received little traffic stop training or experience during and after COVID and now view enforcement work with reluctance. The Jury heard some officers are wary of citizen complaints and frustrated by the limitations of SFPD General Order 9.07, which limits enforcement for certain infractions to reduce racial disparities. The General Order, however, does not in any way impede enforcement against the most dangerous behaviors. The Jury also came to understand that making traffic enforcement a regular practice helps officers get 42 DataSF, “Police department stop data,” Open Data Portal, Accessed Mar 2025 in the habit of interacting with the public and that the body camera is good protection against complaints. The Jury came to understand that some officers are more productive in enforcing traffic safety, while others make few or no stops. For the traffic company, citations per officer fell from 500 per year in 2014 to 80 in 2022-2023, rebounding to about 160 in 2024. Among the precincts, citations per officer went from almost 60 per year in 2014 to between one and two in 2022–23, before rebounding to seven per year in 2024. With current staffing,43 citations could be restored almost to 2016 levels if active traffic company officers averaged approximately three citations per shift (fewer than the average five citations per day of the more productive traffic company officers) and other active officers averaged approximately one per working week. To return to a level of enforcement that is protective against severe and fatal traffic collisions, SFPD leadership would have to set clear goals and expectations around what level of enforcement is needed to make streets safe and how their officers will meet this need. Regular planned traffic enforcement operations can reacquaint patrol officers with the practice of traffic enforcement. To be most effective, enforcement operations should target a range of dangerous violations, such as driving under the influence (DUI), not yielding to pedestrians in the crosswalk, running red lights and speeding.44 To help rebuild a culture of traffic enforcement, SFPD would have to take ownership of street safety. Reporting regularly and at the highest level on the violations issued for traffic fatalities and severe injuries, as well as the numbers of citations issued, would help the SFPD make traffic enforcement into a key performance metric. Just like trends in crime, trends in traffic deaths and severe injuries are a police responsibility and a meaningful indicator of police effectiveness. Assuming 1200 active officers in the precincts and 18 active officers in the traffic company. Elvik, “A comparative analysis of…,” The de-prioritization of traffic safety extends to understaffing of key roles, including support for automated enforcement of red-light violations and investigation and analysis of severe and fatal traffic collisions. Our investigation found that there is a problem staffing the TCIU. Because the staff is so small, team members are all on call every day and night. Preparation for these roles takes several years and there are no replacement officers in the queue to backfill officers approaching retirement. The SFPD needs to plan for how it will provide continuous and sustainable staff coverage for these key traffic safety functions. Outdated technology wastes staff time As staffing levels have indeed declined in both the precincts and the traffic division, SFPD needs to do its work more efficiently. Instead, the Jury’s investigation revealed that SFPD has failed to plan and invest in an infrastructure to facilitate robust traffic enforcement. Technology for issuing traffic citations is cumbersome and unnecessarily slow. Because the e- cite system often can't print citations, many officers avoid it altogether, instead issuing handwritten tickets. These tickets must then be hand logged into SFPD reports--and eventually transcribed by the courts back into a database years later. The mobile interface for the reporting required by RIPA is also unwieldy and slow, further discouraging enforcement activities. For highly complex collision reporting, SFPD uses a fillable PDF technology that is prone to errors. The poor technology wastes valuable officer and staff time, discourages traffic enforcement, and reduces data transparency and quality. The Department of Technology and the Digital Services Team could provide support to the SFPD to audit their traffic-specific equipment and software toolchain to prioritize time-saving improvements. Although ideally the state would develop a robust collision reporting tool, the city should not wait.45 Collaboration with SFMTA 45 State of California, “Traffic collision report,” CHP 555, Accessed Apr 2025 and the San Francisco Department of Public Health (SFDPH) is essential to ensure any new system to report injury collisions also improves tracking efficiency and data quality. The engineering approach The theory behind the engineering approach to traffic safety is that engineers should design roads with the expectation that human drivers will make mistakes. This design-driven approach has natural limits: even if it stops ordinary mistakes from resulting in injury or death, it cannot adequately prevent the kinds of egregiously reckless behavior which enforcement helps to deter. Yet, as enforcement fell off over the last decade, SFMTA shouldered much of the burden of Vision Zero, and clearly feels the weight of that responsibility. SFMTA has made real progress. To reduce fatalities and make further reductions in severe injuries, SFMTA needs to be more efficient and timely in identifying and implementing much-needed safety improvements for city streets. Tools for safer streets SFMTA uses a number of engineering tools to make streets safer. The general principles of these interventions include increasing visibility of vulnerable road users, physically separating vehicles from vulnerable road users, keeping vehicles from speeding, and slowing vehicle turns. Slower speeds, i.e. not exceeding the established speed limit, both reduces the severity of collisions and gives drivers more time to react in order to avoid collisions. Interventions include the following: • Improved traffic lights include leading pedestrian intervals, which improve visibility for pedestrians by letting them out into the intersections before vehicles are allowed to turn; pedestrian countdowns; accessible pedestrian signals; and longer walk times allowing pedestrians to cross safely. • Crosswalk improvements include high-visibility crosswalks, recessed stop lines, and flashing beacons. These are all tools to increase the visibility of pedestrians. • Daylighting removes parking within 20 feet of a marked or unmarked crosswalk in the direction of vehicle travel to improve visibility of pedestrians. State Assembly Bill 413 makes parking in these zones illegal in California. SFMTA plans to implement the law by painting curbs red to identify the no-parking zones.46 • Hardened daylighting adds safety zones to visually and/or physically narrow the street, along with physical barriers such as concrete blocks, planters, bike corrals, plastic delineators, curb extensions, bulb-outs and pedestrian islands to reduce the crossing distance for pedestrians and slow vehicles through intersections and turns. • Road diets remove excess capacity on multi-lane streets in order to discourage excessive speeds. Narrowing overly wide travel lanes also helps to reduce speeding. • Turn calming uses rubber speed bumps, paint and plastic flex-hit posts to make vehicles take wider, slower turns at intersections. • Turn restrictions, including no turn on red or no turn during the bike or pedestrian signal, reduce interactions between vehicles and vulnerable road users at intersections. • Traffic calming includes bumps in the road to slow vehicles—speed humps, speed tables, speed cushions, and raised crosswalks are some of the numerous varieties—as well as other measures to slow traffic, primarily on single lane, lower volume streets. • Protected bike lanes can be separated from traffic by parked vehicles, concrete barriers, planters, or flex-hit posts. They both dissuade drivers from blocking the lanes and protect users from moving vehicles. • Slow streets reduce and slow vehicle traffic, making the street safer and more comfortable for vulnerable road users. They use a variety of design interventions 46 California Legislative Information, “Assembly Bill No. 413,” October 12, 2023 described above, along with turn restrictions and traffic diverters to restrict through- traffic. • Quick builds are corridor improvements that can use a variety of the paint and plastic approaches outlined above, along with transit boarding islands and adjustments to traffic signal timing and parking and loading. • Speed limit changes, signage, and radar speed signs can help reduce vehicle speeding. Since AB 43 passed in 2021, SFMTA has reduced speed limits on 46 miles of streets.47 SFMTA also oversees automated enforcement: red-light and illegal-right-turn cameras at 19 intersections mostly installed in 2020 and 33 speed cameras installed in 2025. Violations captured by the red-light cameras result in official citations issued by SFPD, including license penalties tracked by the DMV, while the speed cameras issue municipal citations assigned to the car, more like a parking ticket. Given the great number and variety of SFMTA interventions, it is difficult to provide a clear timeline for implementation. By the beginning of 2025, along with many more interventions, SFMTA had implemented leading pedestrian intervals on 1035 crosswalks, used turn calming on 223 intersections, painted 2,461 high visibility crosswalks, protected 51 miles of bikeways, completed 39 quick builds, installed many pedestrian islands and put a few roads on a diet.48 Several key safety interventions, such as red-light cameras, speed limit changes, and speed cameras required changes to state law that took five to ten years to pass and implement. SFMTA has provided data to demonstrate the collision reduction benefits of some types of these interventions, along with data on vehicle speeds, pedestrian or bike and vehicle interactions, close calls, vehicles yielding to pedestrians, and other safety metrics.49 The 2023 47 SF.gov, “Vision Zero benchmarking: street network speeds,” Accessed May 2025 48 SFMTA, “Vision Zero safe streets progress,” Accessed Apr 2025 49 SFMTA, “Neighborhood traffic calming projects,” Accessed Apr 2025; SFMTA, “2023 Slow streets evaluation report,” Accessed Apr 2025 evaluation of 18 corridor projects from 2017 through 2022 found that the total collision rate was reduced by 16 percent, bike collisions by 25 percent and pedestrian collisions by 35 percent.50 Safer streets require a proactive approach SFMTA prioritizes improvements along the high injury network (HIN), a map of streets and intersections where the majority of severe and fatal injuries have taken place.51 Past interventions, such as car-free spaces in Golden Gate Park and speed limit and traffic light changes on Sunset Blvd., have successfully removed portions of the roadway from the HIN, while many parts of the city remain stubbornly on the map (Figure 6).52 2015 2017 2022 Figure 6. High Injury Network maps for 2015, 2017, and 2022 (Vision Zero SF, Maps and Data) While an important tool, the HIN is by design backwards looking and reactive. SFDPH is now working with DataSF, with the support of SFMTA, to build a predictive model to understand where fatal and severe injuries are more likely to happen in the future.53 Such a model could 50 SFMTA, “Safe streets evaluation program 2023 annual report,” Sept 1, 20224 51 Another update is expected to be released sometime in 2025. Ricardo Olea, “2017-2022 San Francisco traffic crashes report,” May 2023 53 One untapped set of data to seed this model is the Waymo open-sourced road usage data. This data set may be too complex for city engineers to be able to use, but SFMTA could ask Waymo to pull out specific data points, such as density of pedestrian traffic, turn speeds of vehicles at different types of intersections, and incidence of high speed and reckless driving. help the city predict the impact of street design proposals and make better prioritization decisions, allowing for more efficient resource allocation.54 The problem remains that SFMTA is not well organized to quickly implement proactive and integrated interventions. The clearest example is the traffic calming division, which is reactive by design. That program responds to applications from residents and then carries out extensive surveys to determine what interventions are warranted. This favors wealthier neighborhoods and limits implementation speed. Mayor Breed stated that over 700 traffic calming devices had been installed over the Vision Zero decade, an average of 70 per year.55 By contrast, Boston recently moved to a proactive prioritization model—taking into account collision frequency, percentages of older residents, children, and people with disabilities, and racial and socioeconomic indicators—and is on track to install 1,500 speed humps in three years.56 SFMTA has just finalized the Biking and Rolling Plan, which is a very broad, city-wide, plan for safer streets.57 This is important, as fear of injury from vehicles is one of the most important barriers to biking and 80 percent of surveyed San Franciscans said they would bike more if they felt safer.58 In addition, research shows that protected bike infrastructure makes the city safer for all road users.59 As SFMTA considers how to prioritize implementation of the plan, there is an opportunity to consider each corridor in context and improve the safety benefits beyond the routes themselves. Miguel Álvarez, “Predicting traffic accident hotspots with special data science,” Carto, Dec 2, 2020 55 Office of former Mayor London Breed, “Mayor Breed and city officials mark 10 years of Vision Zero and next steps in street safety,” SF.gov, Mar 28, 2024 56 City of Boston, “How did you choose speed hump locations?,” Making Neighborhood Streets Safer, Accessed Apr 2025 57 SFMTA, “San Francisco’s biking and rolling plan,” Mar 2025 58 Lauren Pearson, et al., “Adults’ self-reported barriers and enablers to riding a bike for transport: a systemic review,” Transport Reviews, 43 (2023): 356-384; Rebecca Ashton-Dziedzan, “Adopting the biking and rolling plan: A safer, more connected San Francisco,” SFMTA News, Mar 12, 2025 59 Wesley Marshall and Nicholas Ferenchak, “Why cities with high bicycling rates are safer for all road users,” Journal of Transport & Health, 13 (2019): 100539 SFMTA generally considers one corridor, one street, sometimes even one intersection, at a time, which can raise fears that major changes will drive traffic and collisions to neighboring streets and crossings. An area-based approach would focus on improving road safety within a specific geographic area by implementing a range of interventions from the tools listed above, along with other street-use changes such as parking management, bus lanes, traffic signal timing, etc. By taking into account the overall street network, SFMTA can create a more comprehensive and effective strategy. This approach would benefit from a collaborative process—involving community stakeholders and SFMTA planners and engineers all working together for safer streets—to ensure that the designs are technically sound and responsive to the community. While broadening the scope may increase complexity, it could also widen the safety impact and broaden neighborhood support for projects, thus increasing the rate of implementation. Another important opportunity for SFMTA to take proactive measures to improve safety is in implementing California’s Daylighting Law (AB 413), which eliminates parking within 20 feet of a crosswalk to improve pedestrian visibility. Unfortunately, simply painting curbs red may not be sufficient to get real safety benefits. According to a January 2025 study by the New York City (NYC) Department of Transportation (DOT), just removing parking at intersections, without taking additional measures, can increase collision injuries.60 However, when coupled with complementary interventions, newly daylit intersections reduced risk of injury, and even more so for severe and fatal injuries. The most beneficial safety interventions assessed in an earlier study by NYCDOT are included in Table 1. Table 1. NYCDOT reductions in pedestrian injuries for specific safety treatments (DOT, 2022)61 Safety Treatment Pedestrian Injury Change Pedestrian Severe & Fatal Injury Change Protected Bike Lanes -17.8% -29.2% Pedestrian Islands -10.2% -29.9% 60 NYC DOT, “Daylighting and street safety,” 2025 43 NYC DOT, “Safety treatment evaluation (2005-2018),” 2022 Road Diets -12.5% -31.7% Turn Calming -17.5% -32.7% Leading Pedestrian Intervals -18.1% -34.3% Curb & Sidewalk Extensions -16.5% -44.7% The lesson from NYC is clear—to get the full intended benefits, daylighting must be hardened through other interventions to narrow the roadway and slow travel through the intersection. SFMTA has used each of these interventions in many locations throughout the city and has collected data to demonstrate safety benefits.62 SFMTA has also identified at least some non- hardened daylighted intersections where collisions have increased.63 SFMTA currently plans to comply with AB 413 by painting daylit zones red to give guidance for vehicle parking, as requested by Resolution 632-24 of the board of supervisors.64 By revisiting this plan to consider incorporation of hardening practices, SFMTA could ensure that this important law achieves the sought-after safety benefits. Building trust through neighborhood-based outreach In West Portal, after the horrific, deadly crash outside the public library in March 2024, there was a strong push to make the intersection of West Portal and Ulloa safer. The SFMTA straw proposal released for community feedback was taken by one group of residents and business owners as a done deal that would kill small businesses.65 To gather input from trusted neighborhood leaders and representatives, District 7 Supervisor Melgar helped to convene a working group to engage with SFMTA planners and engineers. One of the big topics discussed by the group was touched upon above—if changes are made 62 SFMTA, “Left turn safety,” Accessed Apr 2025 63 SFMTA, “Tenderloin daylighting project,” Accessed Apr 2025 64 Board of Supervisors, “Urging Implementation of Red Zones for Daylighting at All Crosswalks and Intersections,” Resolution No. 632-24, Dec 20, 2024 65 Bay City News, “West Portal business owners lambast city over proposed traffic plan,” Kron 4, Apr 25, 2024 to this one intersection, won’t all the problems just move down West Portal to another intersection? It turned out that many of the intersections along and just beyond West Portal had been asking and waiting for traffic calming for a long time. By broadening the scope of the plan, they were able to provide safety benefits all along the corridor, while still maintaining almost all the benefits of the original straw proposal. By working directly with neighborhood stakeholders, SFMTA was able to implement important safety interventions on a timely basis. The Jury’s investigation revealed that community input and outreach is critical to SFMTA success in promoting street safety. The Jury also heard that community involvement takes a huge amount of time. We heard that SFMTA outreach events often seek open-ended input, rather than clearly explaining needs and narrowly defining queries. As a result, there have been wild swings on high profile projects and there can be a lot of mistrust of SFMTA as a partner on street safety. SFMTA’s lack of effective community outreach substantially slows down progress on safer streets. A proactive outreach approach could help build trust and expedite and improve projects. A dedicated outreach contact per supervisorial district could foster long-term relationships with both the supervisor and key neighborhood stakeholders who are committed to safer streets. This SFMTA contact person could integrate community input into safe street projects, speeding implementation and improving outcomes. Expanding community involvement through tangible activities to improve street safety, such as installation of planters or benches as part of daylighting or slow streets, could further bolster community support and trust. A strict neighborhood focus could, however, leave out street users that are traveling to or passing through a neighborhood. Broader input on such things as perceived street safety or inefficient light timing could help inform future projects or lead to spot fixes to improve overall street safety and comfort. By adding a street safety routing to the existing city services tool, 311, SFMTA could provide residents with a channel to offer feedback on street safety and traffic flows. Outreach also is hindered by the negative focus of the key Vision Zero metric. While zero fatalities is an important aspirational goal, setting a deadline for zero is a set-up for failure. Instead, SFMTA needs clear, common-sense metrics to demonstrate progress on safer streets. Tracking the collision reduction benefits of different interventions, following NYCDOT’s example, would help increase trust. Additional metrics such as incidence of excessive speeding, average turn speed, and yielding to pedestrians could be used to transparently track and communicate progress, helping to increase public support for safer streets. More education, please! When it launched in 2014, education was one of the three pillars of the Vision Zero initiative.66 Enforcement and education work well together—education can extend the reach and reinforce the messaging of enforcement. Education is also needed to help communicate the whys and hows of engineering decisions that result in street changes. SFMTA generally has taken the lead on education efforts, ranging from broad public campaigns to classes for specific user groups. The Jury heard from several interviewees that education was a missing link in street safety. Broad public outreach not only teaches people how to navigate streets safely but also forges a shared commitment to street safety. Hands-on training is by far the most effective method, and the city should build upon existing programs to get kids and families safely walking, biking and taking transit to school. Finally, the Jury noted important gaps in knowledge of Vision Zero among different city agencies. Board of Supervisors, “Implement a Vision Zero three point plan: engineering, education and enforcement,” Resolution 91-14, Mar 2014 Broad public campaigns Early Vision Zero public education campaigns by SFMTA were impactful but have declined in recent years. This despite evidence that sustained outreach efforts are essential for maintaining safety benefits.67 In January 2025, SFMTA launched a highly visible citywide campaign to announce speed cameras, similar to its 2021 campaign about lower speed limits in the Tenderloin (Figure 7). Figure 7. Educational campaigns from 2025 and 2021 regarding speed enforcement and limit changes Vision Zero’s early public education campaigns worked closely with project partners such as the SFPD and SFDPH, as well as community partners like Walk SF and the San Francisco Bicycle Coalition (SFBC). These campaigns were comprehensive, including different outreach modes, and often incorporated an enforcement component. Past notable examples include: • Be Nice, Look Twice was launched at the start of the Vision Zero initiative (Figure 8). SFMTA found that, “After the campaign, nearly 80 percent of San Franciscans reported knowing that pedestrian safety in the city is an issue that needs to be tackled.”68 67 Vision Zero, “Safe Speeds SF High Visibility Enforcement Campaign Findings,” SFMTA, Nov 2019; Sara Waring, et al., “Examining the effectiveness of an education-based road safety intervention…,” Traffic Psychology and Behavior, 105(2024): 336 68 SFMTA, “Be Nice, Look Twice Pedestrian Safety Campaign,” Accessed Mar 2025 • It Stops Here addressed drivers’ failure to yield to pedestrians at crosswalks.69 • Safe Speeds SF Campaign combined outreach with enforcement, issuing 1800 speeding tickets along 11 corridors over 12 months.70 Figure 8. Be Nice, Look Twice campaign Broad public education campaigns can inform the public about how to use streets more safely, what enforcement measures to expect, and how to navigate major roadway changes, such as traffic diversions and slow streets. In addition to reducing risky behavior, education can help build a shared vision of safer streets and increase buy-in for street safety improvements. Training the next generation of safe road users Traffic-related dangers are identified as a key barrier to walking to school.71 SFMTA co-leads a Safe Routes to School program, in partnership with the San Francisco Unified School District (SFUSD) and the San Francisco County Transportation Authority (SFCTA).72 This program helps families safely walk, bike and take MUNI to school, encouraging bike trains and walking school buses—groups that bike or walk together for safety, community-building, and exercise. Several non-government organizations (NGOs) also provide valuable street safety education: 69 SFMTA, “Safe Streets Campaign,” Accessed Mar 2025 70 Vision Zero, “Safe Speeds SF High Visibility Enforcement Campaign Findings,” SFMTA, Nov 2019 71 John Omura et al., “Prevalence of children walking to school and related barriers—United States, 2017,” Preventative Medicine, 118 (2019): 191-195 72 “Safe Routes to School San Francisco,” Accessed Mar 2025 • Y-Bike (YMCA) partners with the SFUSD and afterschool providers, funded by the SFMTA to teach kids safe biking. • The SFBC offers free classes and multilingual resources to teach kids, families and adults how to bike safely in the city.73 • Walk SF leads a Safe Routes to School program, which provides training and tools in multiple languages to help families start walking school buses.74 • SF Bike Bus is a grassroots movement to help communities form and maintain bike buses to schools.75 The Jury learned that kids being able to safely walk, bike and take MUNI to school was a big priority for the city.76 To help achieve this goal, street safety could be more fully incorporated into the SFUSD curriculum. Getting the messaging right is important for education to be effective.77 As SFMTA expands their traffic-calming program to focus more on schools, there also is an opportunity to lean into Safe Routes to School to increase outreach around street safety improvements.78 Training specific user groups SFMTA also offers and supports specialized training and classes for specific users. In collaboration with SFPD, free in-person motorcycle safety classes are offered to the public.79 For large urban vehicles, SFMTA offers a video and training curriculum, which is required for all city contractors.80 In 2024, two of the 42 Vision Zero fatalities were the result of collisions 73 SFBC, “Resources,” Accessed Mar 2025 74 Walk SF, “Safe Routes to School,” Accessed Apr 2025 75 “SF Bike Bus,” Accessed Mar 2025 76 SFMTA, “SFMTA school safety programs,” Accessed May 2025 77 Sara Waring, et al., “Examining the effectiveness of an education-based road safety intervention…,” Traffic Psychology and Behavior, 105(2024): 347 78 Rebecca Ashton-Dziedzan, “Making trips to school safer across the city: A new daylighting program,” SFMTA, Apr 17, 2025; SFMTA, “School safety programs,” Accessed Apr 2025 79 SFMTA, “Motorcycle Safety,” Accessed Mar 2025 80 SFMTA, “Large Vehicle Urban Driving Safety Program,” Accessed Mar 2025 involving city employees, leading the Jury to consider what more could the city do to align these high-profile and high-mileage users around safe street goals.81 The State of California Commission on Peace Officer Standards and Training (POST) requires all police officers to take a driver training class every two years, as part of their Perishable Skills Program.82 While the training covers all the vehicle codes that have to do with liability of public employees, it does not cover the toll of traffic collisions or the highest risk violations of the traffic code. The SFPD could benefit from additional department training to help officers align traffic enforcement with the highest risk behaviors and understand the critical role of the SFPD in improving street safety. In support of Vision Zero goals, the city has required use of devices to track driving behaviors such as speeding, hard acceleration and hard braking, known as telematics, in city vehicles since 2014.83 Although most non-exempt vehicles are now equipped with the systems, the original deadline of 2020 has slipped to the end of 2025.84 Pairing telematics data with training can greatly reduce serious collisions.85 The pieces are largely in place to make use of telematics to reduce city vehicle-involved collisions, but there is no citywide training guidance on use of telematics to help improve driver behavior. Gabe Greschler et al., “Cyclist killed after crash with city vehicle identified by officials,” The San Francisco Standard, Jun 17, 2024; Eleni Balakrishnan, “S.F. police driver strikes, kills man with car in Bayview,” Mission Local, Dec 14, 2024 82 POST, “POST Perishable Skills Program (PSP),” May 2021 83 Budget and Legislative Analyst’s Office, “Vehicle telematics update,” Board of Supervisors, Aug 19, 2020 84 Noah Baustin, “‘It’s freaking horrific’: City workers are running over pedestrians at alarming rate,” The San Francisco Standard, Dec 18, 2024 85 Leverson Boodlal and Kun-Hung Chiang, “Study on the impact of a telematics system on safe and fuel-efficient driving in trucks,” U.S. Department of Transportation, Apr 2024 Findings and Recommendations The Jury made the following findings and recommendations regarding the city’s process of advancing Vision Zero street safety goals.
Related Recommendations (1)
R5:
1 By March 31, 2026, SFMTA should develop a new street safety education campaign that builds on earlier efforts. The new campaign should focus on ways all road users can prevent the traffic violations that result in the most severe and fatal injuries.
Quick View
Full Details →
Findings & Recommendations
4 findings
F1:
Concerns over the potential risks of AI have led to an overly cautious approach toward emerging technology. The city risks missing opportunities to harness new technology to improve governance and delivery of services to citizens.
Related Recommendations (1)
R1:
1 By September 30, 2025, the mayor should direct DT to produce a comprehensive AI strategy — to be published by June 30, 2026 — outlining near- and long-term implementation targets for incorporating AI into city systems and services. The strategy should include guidance on infrastructure, data sharing, ethics, pilot programs and performance evaluation, training and human resource needs.
F2:
Governance of technology in the city is hindered because of a federated management structure across departments. Such hindrance has slowed or impaired the ability of the city to efficiently identify, pilot, test, and deploy emerging technologies.
Related Recommendations (1)
R2:
1 By December 31, 2025, the mayor and city administrator should adopt a plan for unifying more technology-related organizations within DT, including digital services and other technology functions under the city administrator.
F3:
Procurement of technology in the city is hindered because of a federated management structure across departments. This hinders the ability to find and implement useful, scalable AI and emerging technology solutions, and presents risks to enforcing quality, standardization, privacy and interoperability.
Related Recommendations (1)
R3:
1 By June 30, 2026, the mayor and CIO should jointly conduct a detailed review and adopt new procurement guidelines for city department technology purchasing such that technology that meets certain criteria (cost, strategic relevance, overall risk level) should be prioritized, purchased and implemented through DT in accordance with the ICT plan, as affirmed by DT. The CIO and emerging technologies director should have the ability to definitively reject purchases deemed incompatible with ICT policy or vendor strategy, and/or propose alternative purchases that are better aligned with ICT strategy. Purchase orders with vendors deemed not compatible with ICT objectives should be cancelled.
F4:
The Committee on Information Technology (COIT) is comprised mostly of non-technical leaders and has insufficient authority and influence over departments’ technology plans. As a result, it is falling short of its objective to streamline ICT policy and roadmapping in San Francisco, which threatens current and emerging technology initiatives alike.
Related Recommendations (1)
R4:
1 By June 30, 2026, the city should enact an ordinance amending the Administrative Code to eliminate COIT and centralize a replacement advisory body under DT. This ordinance could be enacted through the customary legislative process established in the Charter. In the alternative, by December 31, 2025, the mayor and the board of supervisors should each recommend to the Commission Streamlining Task Force (established by Proposition E, November 2024) that it include COIT in an ordinance the Task Force would introduce to eliminate certain commissions.
Findings & Recommendations
4 findings
F1:
Concerns over the potential risks of AI have led to an overly cautious approach toward emerging technology. The city risks missing opportunities to harness new technology to improve governance and delivery of services to citizens.
Related Recommendations (1)
R1:
1 By September 30, 2025, the mayor should direct DT to produce a comprehensive AI strategy — to be published by June 30, 2026 — outlining near- and long-term implementation targets for incorporating AI into city systems and services. The strategy should include guidance on infrastructure, data sharing, ethics, pilot programs and performance evaluation, training and human resource needs.
F2:
Governance of technology in the city is hindered because of a federated management structure across departments. Such hindrance has slowed or impaired the ability of the city to efficiently identify, pilot, test, and deploy emerging technologies.
Related Recommendations (1)
R2:
1 By December 31, 2025, the mayor and city administrator should adopt a plan for unifying more technology-related organizations within DT, including digital services and other technology functions under the city administrator.
F3:
Procurement of technology in the city is hindered because of a federated management structure across departments. This hinders the ability to find and implement useful, scalable AI and emerging technology solutions, and presents risks to enforcing quality, standardization, privacy and interoperability.
Related Recommendations (1)
R3:
1 By June 30, 2026, the mayor and CIO should jointly conduct a detailed review and adopt new procurement guidelines for city department technology purchasing such that technology that meets certain criteria (cost, strategic relevance, overall risk level) should be prioritized, purchased and implemented through DT in accordance with the ICT plan, as affirmed by DT. The CIO and emerging technologies director should have the ability to definitively reject purchases deemed incompatible with ICT policy or vendor strategy, and/or propose alternative purchases that are better aligned with ICT strategy. Purchase orders with vendors deemed not compatible with ICT objectives should be cancelled.
F4:
The Committee on Information Technology (COIT) is comprised mostly of non-technical leaders and has insufficient authority and influence over departments’ technology plans. As a result, it is falling short of its objective to streamline ICT policy and roadmapping in San Francisco, which threatens current and emerging technology initiatives alike.
Related Recommendations (1)
R4:
1 By June 30, 2026, the city should enact an ordinance amending the Administrative Code to eliminate COIT and centralize a replacement advisory body under DT. This ordinance could be enacted through the customary legislative process established in the Charter. In the alternative, by December 31, 2025, the mayor and the board of supervisors should each recommend to the Commission Streamlining Task Force (established by Proposition E, November 2024) that it include COIT in an ordinance the Task Force would introduce to eliminate certain commissions.
Findings & Recommendations
4 findings
F1:
Capacity Building Efforts are Insufficient .................................................................. 40
Related Recommendations (1)
R1:
1 By December 31, 2025, the Mayor should request that the Controller prepare a report on the level of competitiveness in city social services grant procurements and the obstacles to increased competitiveness. The report should be published by June 30, 2026.
F2:
Approval Delays Undermine Grant Effectiveness ..................................................... 41
Related Recommendations (1)
R2:
5, R3.1, R3.3 Controller’s Office F4 R4.1, R4.2, R4.3,
F3:
Skill and Tooling Shortages Impede the Grant Process ........................................... 42
Related Recommendations (1)
R3:
1 By June 30, 2026, the City Administrator’s Office should implement an internal education program for city employees whose job requires them to navigate city procurement processes that apply to social services grants (and if appropriate, other grants and contracts). The internal education program should provide clear guidance on the entire end to end procurement process and on the proper use of preapproved standards to expedite procurements. It should also clarify ethics obligations around making procurement decisions with integrity, including guidance on detecting and reporting biased decisions.
F4:
Monitoring Doesn’t Lead to Timely Corrective Action .............................................. 43
Related Recommendations (1)
R4:
2, R4.3, R4.4 Board of Supervisors F2, F3, F4 R2.3, R3.2, R4.1 The Jury requests responses to the findings and recommendations shown in Table 3 within 60 calendar days. Table 3: Requested responses Respondent Findings Recommendations City Administrator’s Office F2, F3 R2.2, R2.3, R2.4,
Quick View
Full Details →
Findings & Recommendations
4 findings
F1:
Capacity Building Efforts are Insufficient .................................................................. 40
Related Recommendations (1)
R1:
1 By December 31, 2025, the Mayor should request that the Controller prepare a report on the level of competitiveness in city social services grant procurements and the obstacles to increased competitiveness. The report should be published by June 30, 2026.
F2:
Approval Delays Undermine Grant Effectiveness ..................................................... 41
Related Recommendations (1)
R2:
5, R3.1, R3.3 Controller’s Office F4 R4.1, R4.2, R4.3,
F3:
Skill and Tooling Shortages Impede the Grant Process ........................................... 42
Related Recommendations (1)
R3:
1 By June 30, 2026, the City Administrator’s Office should implement an internal education program for city employees whose job requires them to navigate city procurement processes that apply to social services grants (and if appropriate, other grants and contracts). The internal education program should provide clear guidance on the entire end to end procurement process and on the proper use of preapproved standards to expedite procurements. It should also clarify ethics obligations around making procurement decisions with integrity, including guidance on detecting and reporting biased decisions.
F4:
Monitoring Doesn’t Lead to Timely Corrective Action .............................................. 43
Related Recommendations (1)
R4:
2, R4.3, R4.4 Board of Supervisors F2, F3, F4 R2.3, R3.2, R4.1 The Jury requests responses to the findings and recommendations shown in Table 3 within 60 calendar days. Table 3: Requested responses Respondent Findings Recommendations City Administrator’s Office F2, F3 R2.2, R2.3, R2.4,