Here in Los Angeles, the first thing visitors learn is that our public transportation system won’t get them anywhere. The second thing they learn is that, according to legend, Henry Ford is the reason why. Car manufacturers (tire manufacturers in some retellings) tore up all of the train lines in Los Angeles in the early 20th century to make way for more streets and highways, and the City’s public transportation has never recovered. That’s started to change in recent years, with more ambitious investments going towards expanding LA’s meager train system.

But we wonder: might history repeat itself, not just in LA but everywhere? This next time, it might be Sergey Brin or Elon Musk playing Henry Ford, as autonomous cars turn existing public transportation into outdated technology.

Yes, autonomous cars are coming soon, and, really, they can’t come soon enough. These cars will neutralize the distractions, emotions, and other behaviors that impair drivers, while handling inclement weather significantly better than the average driver.

Eventually, with enough autonomous cars on the roadways, automakers will be able to connect their vehicles through what is being called the “intelligent grid,” allowing for a synchronized flow of traffic and levels of speed that make car travel safer and faster. If most of the cars on the road know where other cars are heading, they can work together to create the greatest commuting efficiency, adjusting routes or speeds as needed. This coordination across cars should mean that fleets of autonomous cars will be able to go much faster and at much closer following distances than fleets of existing cars.

A 2012 study found that while a human-operated car needs 115 feet of following distance to safely stop at a speed of 75 miles per hour, an autonomous car would need only 16 feet of following distance at that speed. Highway capacity is currently around 3,000 cars per lane per hour for cars moving 75 miles per hour, but if a fleet of cars were 100% autonomous, communicating via an intelligent grid, that capacity would increase to nearly 12,000 cars per lane per hour (Clarke Bowling, “Autonomous Cars Could Nearly Quadruple Highway Capacity, Says Study.” New York Daily News. Sep. 6, 2012). This increase in roadway capacity could significantly reduce traffic congestion.

A 2014 study by FP Think found that traffic flow would improve by an estimated 25%-35% if 75% or more of the fleet of cars on the road were autonomous. With regulations allowing more aggressive following patterns, traffic flow could improve by 45% (Jane Bierstedt, Aaron Gooze, Chris Gray, Josh Peterman, Leon Raykin, and Jerry Walters, “Effects of Next Generation Vehicles on Travel Demand and Highway Capacity.” FP Think. Jan. 2014). A study by the National Highway Traffic Safety Administration is even more optimistic, finding that vehicle-to-vehicle and vehicle-to-infrastructure communication could reduce freeway and arterial congestion by 75% or more. A steep reduction in accidents, which are the cause of up to 25% of traffic jams, would be the biggest contributor to decreased congestion (Lucas Mearian, “Self-Driving Cars Could Save More than 21,700 Lives, $450B a Year.” Computerworld. Oct. 24, 2013).

Of course, even autonomous driving will have downsides—potentially significant ones. First, although the total number of cars might go down if most (or all) cars are autonomous, car usage will go up. Autonomous cars will open up ridership to people who would otherwise be unable or would have difficulty driving themselves or using taxis or driver services, and/or those who cannot afford to own a car or do not have space for one. Autonomous cars will also be able to run errands, such as making pickups or deliveries, during hours they are now left sitting. These are all positive things, but they will increase car usage.

The FP Think study claims that the availability of affordable ride sharing and an improved driver experience from autonomous cars could increase total vehicle miles traveled by 35% over current levels (see FP Think reference above). Of course, the more miles we drive, the more we tear up existing roads. In Los Angeles, road construction and repairs are already terribly underfunded, and local roads are some of the worst of any major city. Last year, the Los Angeles City Council proposed a half-cent sales tax hike to repair some of the most damaged roads in the City, but the proposal was pulled before ever reaching the ballot (see here).

The State of California has already set rules for using autonomous cars that will encourage Californians to become early adopters of the technology. Commercially available autonomous cars will likely begin to spread from California outward. This could mean that Los Angeles will need to invest heavily in road repair and construction in the years ahead to keep up with the increase in miles traveled due to autonomous car technology. This investment is overdue, but it will still be a new burden on the City’s taxpayers.

Also, with more and more people using cars, what will it mean for the billions of dollars invested in long-term public transportation repairs and upgrades? Investment in buses, subways, light rail, or high-speed rail depends heavily on expected future usage. Many of the reasons people take public transportation would effectively go away if autonomous cars become widely available. People would not need to own a car to cheaply commute or to go out. They would still be able to be productive during a drive, just as they are on a bus or a train—but the bus or train ride would suddenly seem a lot slower and less comfortable by comparison.

For fiscal year 2016, the capital expenditures budget for LA Metro is expected to reach $2.1 billion, of which $1.7 billion comes from Measure R (FY16 Budget Development: Preliminary Capital Program. Los Angeles County Metropolitan Transportation Authority: Finance, Budget, and Audit Committee. February 18, 2015), a half-cent sales tax increase across Los Angeles County that took effect in 2009. Measure R will not expire until 2039 and is anticipated to generate $40 billion in sales tax revenue during that time (see here). These long-term investments will become much less palatable to taxpayers if autonomous cars are able to reduce traffic congestion by a sizable margin.

In the long run, this could mean that fewer ‘30-year-or-longer’ capital projects pass the vote. Most likely, public expenditures will go toward maintenance of existing public transportation while more and more dollars will go toward roadway maintenance and new construction. Local leaders may be less willing to put heavy investment into public transit upgrades. With the introduction of autonomous cars to the roadways, public transportation might not suffer in the short run—repairs will still be made—but it seems much less likely that these systems will be expanded.

Investment in high-speed rail is a more mixed picture. The budget for the Los Angeles to San Francisco high-speed rail system is $68 billion. A three-hour rail ride between the cities is forecast to cost between $80-$90 in 2030. Even today, that is not a very big savings relative to simply taking a 90-minute flight at a state-estimated average cost of $140 and sometimes much lower (see here). Autonomous cars will not be able to reach speeds anywhere near the 200 mph high-speed rail line, but they will likely be able to go significantly faster than the current speed limit. Nonetheless, high-speed rail will save hours relative to driving, even in an autonomous car. But planes are faster than high-speed rail and just as productive for the passenger. Ultimately, for longer trips, if given the option of flying, riding in a high-speed train, or riding in an autonomous car, the best choice for passengers will likely be to fly.

Investments that seem like a good idea today, such as expanding train routes across Los Angeles, might look less appealing in the near future, while more suspect investments, such as California’s high-speed rail, might well seem outlandish in retrospect. But it’s wise not to be too forward-thinking when there are real problems to address in the present day. We need to continue investing in some public transportation projects to deal with current congestion.

At the same time, autonomous cars are well past the theoretical stage – we’ll be using them ourselves before long. State and local policymakers and transportation planners need to seriously reconsider what commuting will look like in as little as 10 or 20 years before locking California or any locality into major infrastructure spending.