Focus on transit: How to save BART without raising taxes
San Francisco’s “beloved” BART system is “on the brink of collapse,” says the New York Post, and the only thing that will save it is for voters to agree to raise taxes for it in November. While I question how “beloved” BART is to anyone other than the bureaucrats and union workers who are paid very well to run it, this kind of scare tactic is just what will lead voters to support more subsidies for a transit system that is doing less.
In fact, it is easy to design a plan to keep BART running with no new taxes. The data show that, between 2019 and 2024, BART fare revenues declined by $264 million (55%). Meanwhile, flush with federal COVID relief funds, BART allowed its operating expenses to grow by $228 million (34%). As a result, operating subsidies grew by a whopping 258% from $190 million in 2019 to $681 million in 2024.
BART is treating this as a shortage of funds, but the real problem is the huge increase in expenses when ridership was down. Although it carried 10% more riders in 2025 than 2024, ridership is not going to fully recover anytime soon. BART’s most important job, then, should be to reduce expenses.
It can do this with a combination of reduced train frequencies, shorter trains, increased numbers of operating employees and reduced numbers of administrative employees. Fewer trains mean lower operating costs and less vehicle and track maintenance. Shorter trains won’t save money on vehicle drivers, but drivers make up only 14% of employee costs, and shorter trains saves money on everything else.
More employees are needed because the agency currently spends 14% of its operating costs on overtime. New employees will also have lower seniority and so will cost less per hour than existing employees even without counting the overtime.
Finally, BART’s bureaucratic overhead grew by 64% between 2019 and 2024, which was almost three times more than inflation. The number of hours worked by administrative employees grew by just 16% (though why it should have grown is questionable), so three-fourths of this 64% represented pay increases to bureaucrats. Voters and taxpayers should not tolerate this kind of bloat.
I propose a 25-25-10-40-25 plan:
- 25% fewer trains per hour;
- 25% fewer cars per train;
- 10% more operating employees; and
- 40% less administrative overhead, either fewer administrators or less pay per administrator.
Adding 10% more employees should save around $60 million in overtime a year after accounting for the cost of the new employees. Reducing the number of trains per hour should reduce the remaining costs by about $210 million per year. Many BART trains are eight cars long; reducing the average numbers of cars per train by 25% should save another $136 million a year. Reducing overhead to 2019 levels will save $90 million a year. That’s a total of $436 million a year.
That still leaves a $56 million deficit that I would close with the final 25: a 25% increase in fares. Between 2019 and 2024, fares grew by less than 6%. Since inflation during that time was 23%, that’s a 14% decrease in real fares. Since there has been more inflation since 2024, I propose to raise fares by 25%, which will increase revenues by $60 million. Increasing fares runs the risk of losing some riders, but the continued economic recovery of downtown San Francisco combined with BART’s efforts to increase the safety of its riders should offset that loss.
Altogether, the 25-25-10-40-25 plan would save $496 million per year. Since BART’s 2024 subsidies were $491 million more than in 2019, this should be enough to completely eliminate BART’s deficit. If fare increases are too objectionable, BART could cut train service and the length of its trains by 30% instead of 25% and achieve the same result. If these cuts overshoot the mark, or as ridership increases, BART could increase the frequencies and the length of its trains.
BART’s own plan calls for completely closing 15 stations, including all of the stations on the Dublin line, most of the stations on the San Francisco Airport line, all of the eBART and Oakland Airport lines, two of the stations on the San Jose line, and two other stations. I suspect BART is counting on the people who use those lines and stations to protest.
In contrast, my plan recognizes the reality that ridership, even though it has grown since 2024, is still less than half of 2019. It makes perfect sense to ask this reduced number of riders to live with shorter and less-frequent trains; in fact, reducing frequencies by only 25% or 30% is generous considering the 50% reduction in riders.
Unfortunately, BART is unlikely to consider this plan or one like it. It would rather scare the voters by claiming that failing to deal with reduced ridership by increasing subsidies will lead to some kind of apocalypse. But if voters are smart enough to reject the tax increase in November, BART may have to implement a plan like this one.