A Mere $100 Billion More

The California High-Speed Rail Authority recently released a new draft business plan saying that it needs only $100 billion more to finish the project. The plan admits that the agency expects to spend more on the 171 miles between Merced and Bakersfield than the $33 billion it had projected the entire 463-mile project would cost when voters approved it in 2008. Even with a recent federal grant, the agency only has about $25 billion for the project, most of which it has already spent.

Click image to download a 17.8-MB PDF of this plan.

As shown on page 65 of the plan, the current projection is that the final cost of the project will be between $89 billion and $128 billion, with $106 billion supposedly being most likely. It pairs this with a projected cost of $211 billion “that would be necessary to construct the equivalent highway and air passenger capacity.” However, this is entirely bogus. It assumes, for example, that the only way to increase airline capacities is by building new airports; increasing the size of planes flying between LA and San Francisco is somehow impossible. It also assumes that new freeway lanes would have to be constructed the entire distance between LA and the Bay Area, even in places that aren’t expected to be congested in the future.

In any case, the authority is trying to make $106 billion sound cheap, when it is anything but. Moreover, $106 billion or even $128 billion are almost certainly optimistic; the final cost will probably be well over $150 billion if the thing is ever completed.

Page 88 shows that the Merced-to-Bakersfield segment is expected to have a “base” (50-percent probability) cost of $33.8 billion, which is $198 million a mile. This is based on the amount the authority is currently spending on a 119-mile segment between Madera (23 miles north of Fresno) and Shafter (19 miles north of Bakersfield), which is almost $20 billion or $165 million per mile.

All of this land is flat while the miles between the Central Valley and both LA and the Bay Area will have to go over mountain ranges, requiring lots of cuts, fills, and tunnels. The authority has no experience in this type of terrain and so it is likely to have underestimated costs there as much as it underestimated costs on the flat lands.

The agency thinks it will be able to build the 88 miles from San Jose to the Central Valley for $19.6 billion, or $223 million a mile, and the 120 miles from Bakersfield to Burbank for $33.9 million or $283 million per mile. Those are rather small premiums over the flat-land average of nearly $200 million per mile. I would expect the per-mile cost of building through the mountains to be twice that of building across the Central Valley. If so, this would add at least $30 billion to the business plan’s projected total.

Meanwhile, in light of the pandemic, the agency has modestly reduced its ridership expectations from 35 million trips per year to 27.6 million. This is still more unrealistic than the cost projections. In 2019, Amtrak’s high-speed Acela carried less than 3.6 million riders in a corridor with a higher population than the LA-SF corridor. While California is promising higher speeds than the Acela, the Northeast Corridor has the advantage of really being two corridors anchored by New York, America’s largest city. The California corridor has no similar mid-point metropolis; the Fresno urban area has fewer than 725,000 residents compared with New York’s 19.5 million. Incidentally, to the extent that the 27.6 million turns out to be too high, the supposed savings from not having to expand road and airline capacities dwindles.

Few recognize that there is an important qualitative difference between highways/airlines and high-speed rail, which is that highways and air services can be built incrementally. If a few people are driving, a dirt or gravel road is sufficient (and the U.S. had more than 2 million miles of such roads before the first autos appeared). If more use the road, they can generate enough revenue out of tolls, fuel taxes, or other fees to pave it. If a lot more people are driving, expressways and freeways can be funded.

Similarly, the only infrastructure needed for air travel is an air field (or a body of water for float planes) and a small terminal for ticket sales and baggage handling. High-ceilinged waiting rooms, moving walkways, jet ways, baggage conveyor systems, and other facilities only need to be build as demand warrants (and are generally funded out of airline ticket fees).

In contrast, high-speed rail systems must be built at great expense before a single ticket is sold. To achieve truly high speeds (which trains like Florida’s Brightline don’t do), trains must be built extra light, which means they can’t operate on the same tracks as freight trains because what might otherwise be a trivial freight derailment could demolish an ultra-lightweight high-speed train. This means trains can’t be run at high speeds on existing rail lines.

California is a great and terrible experiment to see if spending $100 billion, or $150 billion, will ever pay off. Relying on its 27.6-million ridership projection, the agency still estimates that ticket sales will earn an operating profit. Whether it truly believes this or is only saying so because the 2008 law authorized it to build only if it can cover its operating costs is difficult to say. But based on the projected profits, the agency says that it hopes it can find private investors willing to contribute to the $100-billion-plus cost.

On one hand, any private investors doing a due diligence will not be naive enough to hand their money over to the state. On the other hand, after spending $100 billion or so of their own money, California taxpayers will probably not be too keen to see the rail line and any profits it makes turned over to a private company that made, say, a $5 billion contribution.

If the California line ever is completed, its boosters will claim it to be a great success even if it carries only 5 million riders a year and doesn’t come close to covering its operating costs. This business plan is already preparing the groundwork for that by claiming that the line is generating billions of dollars of “economic output” and supporting hundreds of small businesses (all of which could be done by spending the same money digging holes and filling them up).

Such claims of success will lead to political demands to build similar lines between, say, Portland and Seattle, Chicago and Minneapolis, St. Louis and Kansas City, and Richmond and Charlotte. High-speed rail doesn’t make any sense in those corridors, but it didn’t make sense in California either. Taxpayers will end up spending trillions on high-speed rail lines that are slower than flying, less convenient than driving, and more expensive than both. Since fares won’t come close to covering operating costs, much less maintenance, this will eventually result in another infrastructure crisis.

As Reason magazine says, the only sensible solution is to pull the plug now. Better to have wasted $20 billion on a project that never got completed than to waste trillions of dollars on projects we don’t need.