Executive Summary

This study analyzes the financial performance of 85 California transit operators for fiscal year 2023, revealing a significant gap between revenue and expenses that necessitates over $10 billion in annual taxpayer subsidies. The report identifies long-term cost escalation, driven by what economist William Baumol termed “Cost Disease,” as a core issue. It concludes that without significant policy changes, such as embracing automation and more cost-effective transit models, these subsidies will continue to grow.

Key Financial Findings

• Overall Deficit: In FY 2023, California’s transit agencies generated $897 million in fares against $8.76 billion in operating expenses. When including non-fare revenue and capital costs for maintenance and replacement, the total taxpayer subsidy required to break even was $10.31 billion.
• Farebox Recovery: The statewide farebox recovery ratio was a mere 10.25%. Performance varied dramatically, from the California Vanpool Authority (CalVans) at 96.77% to the City of Los Angeles at 1.31%, largely due to its fare-free policies.
• Cost Per Mile: Operating expenses per passenger mile were lowest for CalVans ($0.13) and the San Diego Metropolitan Transit System ($0.91), and highest for paratransit services like Access Services ($4.83) and bus-heavy systems in high-cost areas like the Alameda-Contra Costa Transit District ($4.35).
• Largest Subsidies: The largest systems required the biggest subsidies in absolute terms, led by the Los Angeles County Metropolitan Transportation Authority ($2.45 billion), San Francisco’s BART ($1.39 billion), and San Francisco’s Muni ($1.09 billion).

Factors Driving High Costs

The study attributes the financial unsustainability of California transit to several factors:

• Baumol’s “Cost Disease”: Transit is a labor-intensive industry where productivity gains are slow. However, wages must keep pace with other sectors of the economy, leading to a continuous rise in operating costs that outpaces inflation.
• Low Ridership Density: Many systems, particularly in suburban and rural areas, operate long routes with few passengers, driving up the cost per trip.
• Inefficient Operations: Slow bus speeds due to traffic and frequent stops, high administrative overhead, and costly capital projects that add few riders contribute to poor financial performance. For example, AC Transit’s budget works out to roughly $15 for a typical four-mile ride.
• Fare Policies: Fare-free programs and lax enforcement, while intended to promote equity, significantly reduce revenue and can lead to safety and cleanliness issues that deter riders.

Policy Recommendations

The report argues that California transit’s impact on climate change is negligible and that policy should therefore focus on financial sustainability. The key recommendations include:

1. Embrace Automation: Transitioning to automated trains and buses can significantly reduce labor costs, which are the primary driver of expenses. The study points to successful automated systems in Vancouver and Honolulu as models.
2. Support Micromobility: E-bikes, e-scooters, and neighborhood electric vehicles (NEVs) are more cost-effective alternatives for personal mobility that can solve the “first-mile/last-mile” problem and replace short transit trips.
3. Focus on Cost-Effective Models: Prioritize investments in proven, efficient modes like vanpools, Bus Rapid Transit (BRT), and app-based rideshare services. The high cost and long timelines of rail projects make them a less attractive option.
4. Consolidate Agencies: Merging smaller, inefficient transit agencies into larger, regional ones can reduce administrative overhead and improve service coordination.

Introduction

We analyzed financial data reported to the federal government by 85 California transit operators for fiscal year 2023, the latest year for which complete National Transit Database data is available. Most systems reported large gaps between fare revenues and operating costs, necessitating large taxpayer subsidies. When capital expenditures and deferred maintenance are considered, the revenue shortfall is even greater.

An ongoing post-COVID ridership recovery may narrow the losses somewhat in FY 2024 and FY 2025, but transit costs were spiraling even before the pandemic and the need for taxpayer subsidies can be expected to increase going forward in the absence of policy change and adoption of emerging transit technologies.

While transit operations in California were a losing proposition overall, there was wide variability across operators. In this analysis, we look at the best and worst performing operators as well as best practices outside California to see what works. We conclude with policy recommendations.

Overall Picture

California transit agencies reported a total of $897 million of fare revenue and $8.76 billion in operating expenses during fiscal year 2023, resulting in an overall “farebox recovery ratio” (fare revenues over operating expenses) of just 10.25%.

But when evaluating the overall impact of transit on taxpayers now and in the future, factors other than fares and operating expenses need to be considered. In FY 2023, transit operators also collected $427 million in non-fare revenues from advertising, parking, and other services but they spent $2.88 billion on maintaining, improving, and replacing their capital stock. This latter amount excludes capital investments on service expansions. Factoring in these additional revenues and costs, we estimate that California transit needed $10.31 billion in taxpayer support to break even.

FY 2024 data should be available by December. To preview what the data may show, we also reviewed FY 2024 audited financial statements filed by some of the largest systems. These statements suggest some improvement over FY 2023 with aggregate 2024 statewide deficits likely to be in the vicinity of $9 billion.

Ranking Individual Systems

NTD provides data for 85 California systems overall, but many of these are quite small. In the rankings below, we focus on the 38 systems that had at least 2 million Unlinked Passenger Trips in FY 2023 (by “unlinked” we mean that transfers during a single trip may be counted as separate trips). We have published summary data on all 85 agencies in a shared Google Sheet at this link.

The tables in this section show the top five and bottom five agencies for key financial metrics. We also provide commentary for selected agencies with the rankings and include commentaries about larger agencies which do not appear in the rankings.

Farebox Recovery

A common metric for comparing transit system cost-effectiveness is the farebox recovery ratio, which is the quotient of passenger fare revenues and operating expenses reported to the National Transit Database. The next two tables show the large California transit systems with the highest and lowest farebox recovery ratios, respectively.

Table 1 – Highest Farebox Recovery Ratios, Large Systems

The California VanPool Authority (CalVans) has the strongest farebox recovery ratio and performs well on other metrics. Its vanpool aggregator model tightly aligns revenue with cost and minimizes subsidy. Vanpool drivers are volunteers. Like the passengers, they are using the van trips to commute. Riders (often with employer or pre‑tax commuter benefits) pay full, predictable monthly charges that cover vehicle lease, fuel, insurance, and administration, yielding near‑100% farebox recovery. On the cost side, CalVans records very little beyond administration, while vanpools generate long, commute‑length trips, producing exceptionally high passenger‑miles per trip.

Another large system with a high farebox recovery ratio is the Anaheim Transportation Network (ATN). ATN’s ART buses are designed for the Anaheim Resort, Convention Center, and Disneyland area where demand is steady and time-sensitive. That market supports higher effective revenue per rider than typical local transit. Schedules are built around resort opening/closing times, which concentrates loads and improves productivity relative to cost.

San Francisco BART enjoys a well-above average farebox recovery ratio because the system charges relatively high fares. BART uses distance-based pricing rather than flat fares, so longer trips contribute more revenue toward costs. The network is designed for regional, long‑haul travel; BART reports an average trip length of ~15 miles and a pre‑pandemic role carrying over half of the Bay Area’s transit passenger‑miles. Those two factors — long trips and distance pricing — mean each mile traveled is supported by more fare revenue. The Peninsula Corridor Joint Powers Board, which operates Caltrain rail service between San Francisco, San Jose, and Gilroy, also reported relatively high farebox recovery attributable to higher, distance-based fares. Although BART and Caltrain have high farebox recovery ratios relative to other California transit systems, their ratios are well below pre-pandemic levels subjecting the two systems to large annual deficits as COVID-era federal support is exhausted.

Table 2 – Lowest Farebox Recovery Ratios, Large Systems

Most of the City of Los Angeles’ transit services were fare free during FY 2023. Indeed, Los Angeles Department of Transportation’s (LADOT’s) DASH buses were fare-free for nearly all of the FY2021–FY2024 period, and even after fares resumed in February, 2025, they were set at $0.50 (with many categories still riding free). That keeps fare revenue close to zero while operating costs remain substantial. LADOT reinstated and expanded free‑ride programs alongside fare collection (e.g., DASH to Class/GoPass, Cityride, LIFE, Access, Metrolink/EZ pass holders). These erase or reduce the fare paid by many riders.

Urban transit agencies that go fare free or offer lax fare enforcement are at increased risk of encouraging vagrants and criminals to ride their systems. The resulting deterioration in cleanliness, orderliness, and safety can drive passengers to alternative transportation modes.

Although not among the lowest five agencies, Santa Clara Valley Transportation Authority (VTA) has posted below average farebox recovery ratios for many years — about 9% in FY 2019 compared to 6.04% now. Its light-rail productivity is weak with long crosstown trips that are uncompetitive with driving, depressing ridership and fare revenue. The California State Auditor also flagged VTA’s capital spending that yields minimal added riders, underscoring structurally modest fare generation relative to costs. On the revenue side, VTA participates in Clipper START, giving 50% fare discounts to eligible low-income adults—which may be a compelling equity policy but one that lowers average fare per trip. Employee costs have kept spiraling over the last decade even as the number of trips have reduced. Bus speeds and reliability improvements are in progress (e.g., transit-priority lanes), but until ridership density and service speed materially improve, operating costs will outpace fare revenue, keeping VTA among California’s lowest farebox-recovery large agencies.

Operating and Capital Expenses

For those who prefer fare-free transit or those who would just rather de-emphasize revenue collections, a more useful metric is Operating Expenses (OE) per Passenger Mile Travelled (PMT). Tables 3 and 4 show the large transit systems with the lowest and highest operating expenses per passenger mile.

Table 3 – Lowest Operating Expense Per Passenger Mile, Large Systems

Table 4 – Highest Operating Expense Per Passenger Mile, Large Systems

Access Services, which administers paratransit on behalf of other transit providers in Los Angeles County, reported high operating expenses and low farebox recovery. The agency contracts with private providers to offer on-demand, door-to-door shared ride service. The agency is disadvantaged on financial metrics because the Americans with Disabilities Act (ADA) requires it to provide origin‑to‑destination service countywide, with tight on‑time and eligibility standards. That produces very low productivity—few passengers per vehicle‑hour and short, circuitous trips. As a result, passenger‑miles per hour are low while labor, fuel, and contractor overhead keep climbing. With so few passenger miles over which to spread costs, OE per PMT is high. On the revenue side, ADA fares are capped at no more than twice the base fixed-route fare and cannot be distance‑priced, so farebox recovery is chronically low regardless of demand. All that said, providing mobility to those with physical challenges may be seen as a public service particularly worthy of taxpayer support, so there is an argument for not assessing Access Services using the same financial yardsticks applied to agencies serving the general public.

San Mateo County Transit District’s (SamTrans’) high OE per PMT results from the fact that it largely provides short bus trips in an expensive area with moderate population density. The county’s main thoroughfare, El Camino Real (ECR), has long lights which slow buses. ECR is the system’s spine and carries a large share of riders, but it’s congested with frequent stops (¼–½‑mile spacing). ECR consumes ~20% of the bus operations budget and is “very long” and challenging to operate—amplifying the cost impact of low speeds. Coverage obligations on lower‑density Pacific Coastside and suburban segments add miles with modest PMT returns, further worsening the ratio. SamTrans has launched speed/reliability measures on ECR (signal priority, stop consolidation, targeted lanes) but those are yet to be fully implemented. Increased compensation and lower operator productivity have worsened SamTrans’ metrics. A Metropolitan Transportation Commission audit found that vehicle service hours per full-time equivalent employee decreased from 1187 in FY 2018 to 855 in FY 2023

Alameda-Contra Costa Transit (AC Transit), which serves Oakland, Berkeley, Richmond, and adjacent cities in the East Bay, faces structural and administrative issues that result in high unit costs. The network is overwhelmingly served by traditional bus service with frequent stops. Only a small share of passenger‑miles is served by faster submodes (e.g. Tempo Bus Rapid Transit). That means the system average reflects slow, stop‑dense urban bus corridors rather than rail/BRT speeds. Peak operating speeds around 12–13 mph and closely spaced stops suppress passenger‑miles per vehicle hour. Riders also take short trips (≈3.5 miles on average)—so even when buses are full, each boarding produces limited PMT. On the cost side, AC Transit faces a high cost per revenue hour (≈$260) and only ~17 UPT per Vehicle Revenue Hour (VRH), a combination that necessarily yields a high OE per PMT. The district’s budget works out to roughly $15 for a typical four-mile ride—evidence of unusually heavy back-office spending. A board-approved $500k “sweetheart” consulting payout for departing General Manager Michael Hursh raises concern about AC Transit Board’s financial stewardship. The Bay Area’s labor‑intensive, high‑wage environment compounds this: labor is the dominant expense share.

Operating expense is an incomplete measure of the cost of providing transit. Capital must also be considered. There are various ways to measure the cost of transit capital on a per unit basis, none of which is perfect. One option is to add depreciation expense to the operating expenses, but NTD does not include depreciation expense. This value can be obtained from each entity’s financial statements, but some transit providers do not report the depreciation expense related to their transit operations. NTD does collect capital expenditure data. But a large portion of transit agency capital spending is driven by expansion projects, which arguably should not be attributed to the cost of providing transit to current users. In the tables below, we only include transit agency capital expenses not related to expansions, i.e., those needed to replace and repair equipment and facilities. (This is admittedly an imperfect measure because, for example, vehicle replacements occur irregularly, and we look forward to finding and using more accurate metrics in future research.)

Table 5 – Lowest Operating Expense Plus Non-Expansion Capital Expenditure per Passenger Mile, Large Systems

California Vanpool Authority’s long trips keep Operating Expense per Passenger Mile Traveled quite low compared to other operators. Capital intensity is also low: most vehicles are leased, so renewal needs are modest and largely embedded in participant charges. Scale further compresses overhead per vehicle.

Another system with very low OE per PMT is the San Diego Metropolitan Transit System (MTS). Its Mid-Coast (UC San Diego) Blue Line extension added mileage and new destinations, lifting average trip length and spreading fixed costs over more passenger‑miles. Extensive BRT services with limited stops, transit‑signal priority, queue jumps, and targeted lanes (e.g., along I‑15) to maintain higher cruising speeds and reliability, producing more passenger miles travelled (PMT) per vehicle‑hour. MTS also re-timed and restructured bus routes around the County’s rail spine, trimming duplicative mileage and focusing service where demand is strongest. Service frequency increases, including later‑evening Trolley headways and a new overnight border–Downtown link, improved load factors without proportionate cost growth. The long‑haul Blue Line from San Ysidro to Westfield UTC in La Jolla concentrates PMT, while a modern fare system (PRONTO) reduces boarding friction and helps vehicles spend more time moving.

San Diego Metro’s low costs contrast with the neighboring Orange County Transportation Authority (OCTA), which is in the middle of the pack on operating cost and subsidy metrics among larger California systems. Because OCTA relies on traditional bus service in a largely suburban county, it lacks routes that serve large numbers of passenger miles. OCTA is moving into rail service, but the new addition is unlikely to help. The new OC Streetcar, now slated to debut in spring 2026, will only span 4.15 miles and has ten stops. The system is opening after years of delays and escalating costs that have drawn criticism from the County’s civil grand jury.

Table 6 – Highest Operating Expense Plus Non-Expansion Capital Expenditure per Passenger Mile, Large Systems

The Central Contra Costa Transit Authority, doing business as The County Connection, is another suburban bus system that has high OE per PMT due to its heavy reliance on short, low-speed trips. The Authority purchased 28 diesel buses in FY 2023 as replacements for older vehicles, accounting for the large size of its Non-Expansion Capital Expenditure per PMT.

Sacramento Region Transit’s (SacRT) operating expense plus non-expansion capital cost per PMT is not among the five highest, but at $4.77, is well above the statewide average for large California agencies for two main reasons. First, the district is in the middle of a light-rail modernization that adds temporary operating costs before ridership fully materializes. Second, SacRT’s former was unusually costly at over $47 per ride, which raised the average cost per passenger trip. SacRT discontinued SmaRT Ride at the end of 2024, replacing it with the SacRT Flex program, which is open to fewer riders (seniors, teens, and low-income individuals) and uses smaller vehicles.

Taxpayer Subsidies

The next two tables capture the overall cost of operating each system that must be covered by taxpayers. We estimate the taxpayer subsidy as the total of fare and non-fare revenue less operating and non-expansion capital costs. Because we are showing total subsidies rather than subsidies per mile or per trip, the rankings are driven by the relative size of the systems.

Table 7 – Largest Taxpayer Subsidies

Although the list of most-subsidized systems largely mirrors the list of California’s largest systems, the San Diego Metropolitan Transit System is notably absent despite providing the third highest number of passenger trips and passenger miles statewide. As discussed above, this system’s heavy use of rail and BRT reduces its costs per PMT.

Los Angeles Metro is California’s largest transit operator and the second largest in the U.S., spanning bus, BRT, rail, and paratransit. Its scale is underwritten by dedicated county sales taxes (Props A/C, Measures R/M), which together translate into well over $5.5 billion annually with an annual operating shortfall of almost $2.5 billion. LA Metro posts one of the state’s lowest farebox-recovery ratios (low single digits post-pandemic) and middling operating-efficiency metrics versus large peers, reflecting slow bus speeds, complex service patterns, and heavy safety/cleanliness demands on the network. There have been significant concerns on safety, cleanliness, fare evasion, and reliability, which constrain ridership and revenue. LA Metro’s response has been operational rather than fiscal: the NextGen Bus Plan and an expanding network of bus-priority lanes to lift speeds; higher entry gates to improve fare compliance; Transit Ambassadors and enhanced security (including a transition toward an in-house police department) to improve the customer environment. Since some of these measures had not been fully implemented in FY 2023, LA Metro could see enhanced revenue in subsequent fiscal years.

The City of San Francisco’s Muni system has a high taxpayer burden in absolute terms and one that is particularly high on a per PMT basis. Fare revenue per mile is low and operating speeds are slow. San Francisco permanently expanded Free Muni for Youth (all riders ≤18 ride free), and maintains multiple low‑/no‑fare programs (people experiencing homelessness can ride free)—choices that meet equity goals but depress fare receipts. Meanwhile, most service operates in mixed traffic with dense stop spacing, yielding slow average bus speeds and short trips.

Table 8 – Smallest Taxpayer Subsidies

With its low expenses and strong revenue, CalVans almost pays for itself despite serving 3.6 million passenger trips and providing 105 million passenger miles of travel during FY 2023.

Potential Costs of Deferred Maintenance

For agencies that provide financial statements, it is possible to make some estimate of their deferred maintenance burden. Independent researcher Gregg Dieguez has applied a methodology originally developed by Fitch Ratings to measure the degree to which a transit system’s assets have been exhausted. The Life Cycle Ratio is the system’s average asset age divided by their average lifetime as implied by capital and depreciation data on its financial  statements. The higher the ratio, the closer the system is to being obliged to replace its assets.

Table 9 – Lowest Life Cycle Ratios (Capital Stock Furthest from Replacement Age)

San Francisco BART’s appearance in this list may be surprising to passengers who have recently faced delays due to track and electrical system failures. Two structural features push BART’s lifecycle ratio down relative to its peers. First, the system’s core fixed assets—Transbay Tube, tunnels/structures, traction power, train control—have very long planned lives (often a half-century or more). Even at ~50 years old, many components are still within their design lives; recent retrofit work further extends service life, keeping lifecycle ratio modest. Second, BART just replaced its rail car fleet: the Fleet of the Future reached the initial 775-car milestone in 2024 and continues with Core Capacity cars, sharply lowering rolling-stock average age. Meanwhile, train-control modernization (CBTC) and Measure RR rebuild packages keep renewing power, track and structures in large blocks.

San Francisco’s ferry operator, the Water Emergency Transportation Authority (WETA), has a relatively new asset base that is frequently refreshed. In just the past few years, WETA has opened or advanced new terminals—Richmond (2019), Seaplane Lagoon/Alameda Point (2021), and Mission Bay which is now moving into final design and construction—resetting terminal ages. On the fleet side, WETA has commissioned multiple new high-speed vessels (e.g., MV Dorado, 2022; “Karl,” 2025) and is proceeding with all-electric ferries under its Bay Ferry 2050 Service Vision, further lowering average fleet age as these deliveries arrive. The agency has also secured state and federal grants for expansion and electrification, accelerating replacement cycles rather than stretching assets to end-of-life.

Table 10 – Highest Life Cycle Ratios (Capital Stock Nearest to Replacement Age)

Riverside Transit Agency’s (RTA) lifecycle ratio is elevated largely because the agency is timing fleet replacement to its hydrogen transition. Since late 2024, RTA has also been building a new hydrogen fueling station and staging zero-emission procurements, a strategy that often lets legacy buses run closer to their Useful Life Benchmark (ULB) while infrastructure and grants are finalized. Once the fueling facility is online and orders arrive, the average age of assets will fall. As a bus-only system with no rail assets to dilute the average, the ratio reflects bus fleet age more than facility condition.

Zero-Emission Bus (ZEB) transitions appear to have also slowed bus replacements at Long Beach Transit’s (LBT) and SamTrans before and during FY 2023. The California Air Resources Board requires all bus operators to transition to fully zero-emission fleets by 2040, a policy which may spike bus system capital expenditures in the near future.

Small Systems

We devote less attention to smaller systems in this study because their impact on taxpayers is limited. But the following table is intended to capture the most inefficient small systems. It lists the six agencies with operating expenses and non-capital expansion expenses of more than $10 per passenger mile. These are the systems that should receive the strongest consideration for consolidation or replacement by private alternatives.

Table 11 – Costliest Small Systems Per Passenger Mile (Non-Expansion Capital Included)

Notably, three of these systems absorb taxpayer subsidies of more than $10 million, so non-trivial savings would be possible by replacing or reorganizing them.

Kern Transit’s financial efficiency is very low because it delivers service across a very large, rural service area with long intercity corridors and significant demand-response obligations. The agency operates 12 fixed routes plus Dial-A-Ride across 906 square miles, linking small communities to Bakersfield and distant hubs like Lancaster, Ridgecrest, Lake Isabella, Delano, and Mojave. These long, low-density runs result in inferior financial metrics. (Kern Transit’s PMT is an estimate since only UPT was reported to the National Transit Database).

Similarly, Tulare County Regional Transit Agency (TCRTA) covers a large, rural service area with long deadhead miles (i.e., miles over which the vehicles are operating but not taking passengers). As a relatively new multi‑jurisdictional consolidation, it also carries integration and start‑up costs as well as inherited fleet maintenance requirements—pushing costs up despite modest PMT.

The City of Fairfield’s bus system suffers from low post‑pandemic ridership on its commuter lines, resulting in long runs with light loads. Its paratransit responsibilities also add cost pressure. Fairfield is in Solano County which is also served by a countywide operator, Solano Transit (Soltrans). Administrative savings and better coordination might be achieved by merging the city and county bus systems. Soltrans is itself the result of a merger of two other Solano County municipal transit agencies (based in Vallejo and Benicia) in 2010, so this type of merger has a local precedent.

Although the City of La Mirada’s bus system is experiencing losses that are an order of magnitude smaller, its dilemma is worthy of comment. The system’s short trips, limited scope, and low fares yield minimal fare revenue and PMT, so unit costs and subsidy per mile are high. The service is contracted out to a private company, MV Transportation, but the terms do not seem favorable to the City. La Mirada may benefit from adopting the contracting model now being used by Huntington Beach, a much larger city that is spending less money annually for a private provider, Circuit Transit, that offers an app-based on-demand rideshare service.

Policy Analysis and Recommendations

Having established that California transit is heavily subsidized overall and having explored the best and worst systems from a financial perspective, we now turn to policy analysis. We begin with a discussion of the economic effect driving California’s transit cost spiral.

 An Economist’s Insight Captures Transit’s Dilemma

Writing in the 1960s, Economist William Baumol first described an effect that plagues labor-intensive industries such as the performing arts, education, and healthcare. Baumol’s “Cost Disease” describes a phenomenon where sectors with inherently slower productivity growth experience rising costs relative to more technologically advanced sectors such as manufacturing. This occurs because wages in low-productivity sectors must still rise to compete for talent with high-productivity sectors, even if their own output per worker has not increased at the same rate. Since these sectors cannot offset higher wages with increased productivity, their unit costs inevitably rise, leading to persistent upward pressure on prices or required subsidies.

As Javier Morales-Sarriera discussed in a 2016 study, public transit also suffers from Baumol’s Cost Disease. The operation and maintenance of vehicles and infrastructure require a significant human workforce. Unlike manufacturing, where automation can drastically increase output per worker, a single transit operator typically manages one vehicle, and tasks like track maintenance remain highly dependent on human labor, making significant productivity gains difficult to achieve. While wages for transit workers tend to rise in line with economy-wide wage growth, their output per worker does not increase proportionally.

For transit, the effect has been exacerbated by declining transit utilization. As average wages rose in the decades after World War II, many commuters switched to cars, and, more recently, some started commuting to their workplace fewer days per week if at all.

Figure 1 shows average operating cost per US transit trip in constant 2023 dollars since 1947.

Inflation-adjusted operating costs per trip rose 744% between 1947 and 2019 before spiking during the COVID pandemic. In 2022 and 2023, costs started to fall back toward their pre-COVID trendline.

Although we do not know when this normalization process will end, we can be confident that the long-term trend toward higher costs per trip will resume given the enduring nature of Baumol’s Cost Disease. As long as the transit sector remains labor-intensive and its inherent productivity growth lags that of the broader economy, the cost of delivering transit services will continue to outpace general inflation.

California policymakers hope to bend the declining utilization curve by concentrating the state’s population near train stations and bus routes. By also pursuing policies that raise gasoline and electricity costs, they aim to arrest the trend toward the greater affordability of driving.

While these policies may stabilize and even raise ridership in the coming years, transit unions and management may be able to continue pushing up overall operating costs. So California should continue to follow the national trend toward higher costs per ride, although possibly at a slower rate.

But this picture could change if transit operations are automated.

Automated Transit

Waymo has proven that automated driving at scale is possible. By March 2025, the service was providing 700,000 monthly trips in the Bay Area and Los Angeles. The fact that Artificial Intelligence can reliably handle all the challenges of city driving provides convincing evidence that more transit automation should be possible in California.

Relative to driving, it is much easier to automate train service on a fixed route, especially on systems like BART that have no at-grade crossings. Comparable systems elsewhere have already adopted automation. One of the earliest to do so was Vancouver’s SkyTrain in 1986. More recently, the Paris Metro has been both adding new automated lines and converting existing lines to driverless operations. In the U.S., Honolulu’s Skyline became the first driverless metro system in 2023, but airport train systems have been offering automated service around the country for decades.

Buses are more challenging to automate, but progress is being made outside the U.S. Japanese startup Boldly, a subsidiary of SoftBank, is operating driverless bus pilots around the country. Among the more ambitious pilots is a 2.5-mile shuttle service between the train station and airport in Komatsu. Most of Boldly’s service is provided by minibuses built in Estonia and France, but its premiere vehicle is a full-sized bus built in China and retrofitted with a package of radar, LIDAR, and software provided by Tier IV, a Japanese company that offers open source autonomous driving solutions.

The Japanese government is making a major investment in driverless buses due to a shortage of drivers, which is expected to get worse as the country’s population ages.

In the U.S., Jacksonville is pioneering autonomous bus technology. The Jacksonville Transportation Authority’s NAVI (Neighborhood Autonomous Vehicle Innovation) vans serve twelve downtown stops along a 3.5-mile route. Since the June 30, 2025 launch date, NAVI’s nine passenger vans have human safety monitors for now, but JTA plans to remove them in late 2026. The service is operated by Beep, a U.S.-based autonomous transit startup, and uses mobility software from Oxa, a U.K.-based company.

In California, the state’s transit agencies can leverage advancements in Japan and Florida to provide riders with more frequent bus service at a lower cost. The Contra Costa Transportation Authority has been offering autonomous shuttle pilots in the East Bay, but major transit operators have yet to get on board.

Transit automation in California has begun to meet resistance from public employee unions. One labor backed group recently proposed including anti-automation language into a draft 2026 Bay Area transit tax measure.

Bikes, E-Bikes, E-Scooters and EVTOLs

Micromobility—encompassing conventional bicycles, electric-assist bicycles (e-bikes), and electric scooters (e-scooters)—has grown rapidly in California cities and suburbs. The relationship between micromobility and transit is complex. On one hand, shared bikes and scooters address the “first-mile/last-mile” (FM/LM) problem, acting as a feeder system for public transit. On the other hand, they often serve as a direct competitor, particularly for short trips. Like cars, micromobility options require no waiting and enable the user to get from origin to destination without switching transportation modes.

The growth of micromobility is being propelled by two parallel trends. The first is a boom in the consumer market for personally owned vehicles, most notably e-bikes. This surge in ownership is fueled by technological advancements, rising fuel costs, and state-level incentive programs.

The second trend is the maturation of shared micromobility systems, the fleets of dockless and station-based bikes and scooters that have become common in California cities.  After a period of volatility, these systems are experiencing a strong, e-bike-led ridership renaissance.

Micromobility options are becoming available to a wider range of users. Three-wheeled e-trikes make micromobility available to individuals with weight and balance concerns. E-trikes range from cruisers with large cargo baskets to heavy-duty fat-tire versions and even recumbent styles with backrests for enhanced support. Specialized designs are also emerging, such as “reverse trikes” with two wheels in the front for more nimble steering and electric rickshaws capable of carrying passengers.

One step above the electric rickshaw is the neighborhood electric vehicle (NEV), which may be legally operated on California roads that have speed limits up to 35 mph. Laguna Woods, a city in Orange County with a high concentration of retirees, facilitates widespread usage of NEVs as well as golf carts. As California’s population ages and NEV technology improves, this alternative may spread to more communities.

Micromobility options allow individuals or small groups to move around in vehicle footprints that are much smaller than traditional cars and SUVs. They thus address an objection transit advocates have against personal vehicles: that they consume too much scarce road space.

Another technology that promises to provide small group mobility while conserving road space is the eVTOL, Electric Vertical Take-off and Landing aircraft. Archer Aviation’s Midnight eVTOL can transport four passengers up to 50 miles making it suitable for downtown to airport connections. Archer is partnering with the Los Angeles Olympic Committee to provide eVTOL service during the 2028 summer games.

Is California Public Transit Needed to Stave Off the Climate Catastrophe?

Transit spending is often framed in terms of the state’s climate change policy. But because California produces such a small share of global greenhouse gas emissions and Californians’ propensity to use transit is so low, investments in California transit cannot meaningfully impact the trajectory of global warming.

In the calendar year 2024, Californians took 955 million transit trips covering about 5.1 billion passenger miles in aggregate. If California transit agencies were all abolished and all of these trips were replaced by travel in fossil fuel passenger vehicles, greenhouse gas emissions would have increased by 204,000 metric tons. Given that global greenhouse gas emissions in 2024 were approximately 54 billion tons, abolishing California transit would have resulted in a mere 0.0004% increase in emissions last year.

Therefore, while advocates for adding rail extensions and maintaining bus lines often couch their support in terms of the need to avert a climate catastrophe, any individual service change has an immaterial impact on the trajectory of greenhouse gas emissions. The best argument for California climate policies is that they set an example for other states and other countries, but most other jurisdictions cannot afford to lose $10 billion annually on transit, so the example we set is unlikely to be followed.

Recommendations

While California transportation policy has no material impact on global climate change, political leaders and policymakers will maintain their preference for greener transportation alternatives.

But they can decarbonize transportation by supporting the transition to electric robotaxis, electric autonomous buses, e-bikes, e-scooters, and neighborhood electric vehicles. These modes are likely to be more cost-effective than traditional public transportation modes because they do not require paid operators, whose compensation is often maximized by public employee unions. These alternatives also do not require the same level of administrative staff, fixed assets, and costs necessary to operate and maintain traditional transit systems.

In addition to downsizing public transit overall, policymakers should encourage what works best within the industry. Vanpools, app-based rideshare, and bus rapid transit are proving to be relatively cost-effective. While rail often outperforms local buses in terms of cost per PMT, the high construction costs and long time frames associated with rail projects militate against increasing reliance on trains.

Finally, to reduce administrative costs and improve schedule coordination, the State should encourage transit agencies to merge.

Conclusion

Overall, California transit providers experienced $10 billion of losses in their 2023 fiscal year. While some improvement may occur in FY 2024 and 2025, the longer term trend is for losses to grow as fare revenues fail to keep up with growing labor and capital asset replenishment costs.

While transit imposes a substantial burden on California taxpayers, it will not save us from the worst effects of climate change. Transit policies should thus be assessed from a financial perspective.

Transit agency losses can be staunched by reducing our reliance on paid human operators. Just as office building operators saved money by transitioning away from human elevator operators for providing vertical mobility in the 20th century, California cities and suburbs should transition toward self-service and automated operations to lower the cost of horizontal mobility in the 21st century.

Marc Joffe is a Visiting Fellow at California Policy Center and President of the Contra Costa Taxpayers Association.

Athan Joshi is a research assistant at California Policy Center. He is a regular contributor of opinion pieces on international trade and public finances to moneycontrol.com.