Image by the author.

Over the holiday, Matt Yglesias at Vox drew up a fantasy idea — crayonista — to reroute the Yellow and Red Lines through central DC. He argues his plan would be better and cheaper than WMATA's official loop plan to build a new Yellow/Blue subway. Yglesias' proposal is elegant on the surface, but gets a number of details wrong and probably doesn't make sense in the real world.

The loop, and Yglesias' critique

The big problem with Metro's layout is how multiple suburban lines converge onto the same tracks downtown. The loop plan solves that by adding another downtown tunnel, so Blue and Yellow don't share with Orange/Silver or Green.

The plan makes it possible to run much more frequent trains on every Metro line except Red, which doesn't share with another line. Here's an explainer with more detail.

Metro's loop. Complicated but operationally sound.  Image by WMATA.

Yglesias argues that Metro's loop plan is too expensive, and that building new subway tunnels close to existing ones doesn't add service to any new parts of DC except Georgetown. And, he says, it would result in some riders making awkward transfers between lines, such as riders on the loop heading in from Pentagon who might change to a Green Line train for a faster ride downtown.

The Yglesias/Levy proposal

To counter those criticisms, which are indeed real challenges, Yglesias proposes the following re-think:

  1. Turn the Blue Line into a shuttle that just goes back and forth between Rosslyn and Pentagon without going farther south or into DC at all. De-coupling Blue from Orange and Silver would allow Metro to run more Orange and Silver trains on existing tracks.
  2. Call trains going to Franconia-Springfield a branch of the Yellow Line.
  3. Use the existing Amtrak/VRE/freight tracks between L'Enfant and Union Station to send Yellow Line trains to Union Station. De-coupling Yellow from Green would allow more trains on both.
  4. Have this new Yellow Line replace the Red Line's eastern leg, running north from Union Station.
  5. Construct a new subway under H Street NE and run the Red Line from there to the Benning Road leg of the old Blue Line, replacing the portion of Blue east of the Anacostia River with Red.

This map shows how it would work.

Yglesias' proposal. Circles show problems areas. Image by the author.

Yglesias is building off of a similar idea floated by transit expert Alon Levy, who proposed the Blue Line shuttle and Red/Yellow flip on his excellent blog Pedestrian Observations.

Problems with this proposal

This idea looks elegant and there is some good logic behind it. New subways are tremendously expensive, so building them in places that don't provide a lot of extra service is not very efficient. And in fact, I proposed the same Blue Line shuttle back in 2008.

Unfortunately, there are big problems with every part of this proposal. Each one seriously undermines Yglesias' claim that this idea can do more for less money. Overall, the obstacles and costs are far greater than Yglesias gives credit for, and the benefits are less.

This is way more expensive than Yglesias thinks

Yglesias' plan hinges on the assumption that we can do three things for free or at least cheap: Turn trains around at Rosslyn and Pentagon, send the Red Line east of the Anacostia River, and use the L'Enfant-to-Union intercity tunnel for Metro instead. All three would actually be tremendously expensive.

First, turning Blue trains around at Rosslyn and Pentagon only saves train capacity on the Orange and Yellow Lines if Blue trains aren't sitting still on the same tracks while they turn around. Even if the turnaround were instantaneous (it wouldn't be), you'd be taking away train slots. You need a third track to make it work, at least at Rosslyn and maybe at both. And since these stations can't fit a third track inside their existing tunnels, that really means the Blue shuttle would need at least one if not two entire new subway stations. That's probably a billion-dollar proposition.

You cannot fit a turnaround track inside Rosslyn station for free. Image by Matt' Johnson.

Second, sending the Red Line east of the river to take over the Largo leg of the Blue Line has at least two big costs: Building an underground track switch at Union Station (already the most complex spot for multi-level tracks in the region), and building a new parallel Metro bridge over the Anacostia itself. The parallel bridge would be necessary to keep Orange/Silver and formerly-Blue-now-Red trains fully separate, since if they share even one inch of track anywhere the existing bottleneck still exists and this entire exercise is moot.

Third, those existing L'Enfant-to-Union Station tracks that Yglesias wants to repurpose for Metro have a lot of trains on them already. Repurposing them would not only require a very expensive retrofit to add electric power and new underground track connections at either end, it would also necessitate completely ending all VRE service, as well as all Amtrak from DC south. Freight trains might also have to stop, depending on the details. Good luck with that.

VRE would have to stop running.  Image by Steve Petrucelli licensed under Creative Commons.

Levy's plan would leave the VRE/Amtrak/freight tracks alone and dig a new Metro subway instead. That's preferable, but it would cost billions of dollars.

Far from Yglesias' claim that this can be done on the cheap, it's actually hugely expensive. Probably less than the entire WMATA loop, yes, but hugely expensive on its own terms. Levy more accurately describes the costs.

This hurts Metro capacity by exacerbating imbalances

In a strictly mathematical sense, Metro already has more than enough tracks and trains to handle demand. Averaged over the entire system, there's enough. The problem, of course, is that some places are bigger destinations than others. For example, at rush hour everyone wants to go downtown, so trains leading into downtown burst at the seams while trains heading away are comparatively empty.

Yglesias' plan ignores this completely.

First, by routing the Yellow Line to miss downtown DC, this plan forces thousands and thousands of peak period Yellow Line riders to transfer to other lines at exactly the time of day when capacity on those other lines is stressed the most. Yglesias criticizes WMATA's loop for making some downtown trips easier with a transfer, but his plan requires it even more.

Second, by sending the Red Line to Largo, Yglesias introduces a new ridership imbalance into the Metro system. Here's how that works: Ideally, opposite legs of each Metro line should have approximately equal demand, so you run the right number of trains on each leg. If ridership on a line west of downtown necessitates a train every 3 minutes, but lower ridership east of downtown only needs a train every 15 minutes, you either have to waste money running extra trains to the east, or provide bad service to the west, or do an inefficient and expensive downtown turnaround.

Metro's existing lines are set up to try and match the demand on opposite legs as much as possible. The opposite ends of lines are neither random nor interchangeable.

The Shady Grove leg of the Red Line is the busiest in the Metro system. Thus, it has to be paired with a busy opposite leg. Glenmont is a pretty good match, but Largo isn't. Thus, rerouting the Red Line towards Largo would create a big operational problem and hurt the system's overall efficiency.

WMATA's loop plan, by contrast, is designed not only to get every line downtown and thus avoid transfers, but also to efficiently pair suburban legs of the system.

It's not just that Yglesias' plan costs way more than he says. It also quite possibly makes Metro's crowding problems worse instead of better.

Swapping H Street for Georgetown doesn't add much more coverage

Finally, Yglesias seems to think that connecting H Street NE counts as expanding Metro's footprint a lot, but connecting M Street in Georgetown doesn't. That's a fairly silly distinction. They'd both be nice, but it's probably about a wash to switch them. They're in the same ballpark of new coverage.

Meanwhile, Yglesias overplays his attacks on the H Street streetcar. The streetcar absolutely took too long to build and absolutely should've had dedicated lanes, but it's faster, more consistent, and more accessible than the X2 bus, and it's beating ridership expectations. There's plenty about it to rightly criticize, but Yglesias' uncompromising hatred for it veers into the sensational, and is one more example among many in his article of the dangers in reductionist thinking.

The big problem is capacity

The fact that Yglesias' crayonista has problems doesn't mean that WMATA's loop idea is perfect. Yglesias' criticisms of it are true. It is tremendously expensive. And although it allows Metro to run many more trains to existing destinations, it doesn't add very many new ones.

But Metro is going to need more core capacity, and that's a fundamentally different goal than adding new destinations. WMATA recommended the loop after studying rider demands and track options in 2013, and although it's possible there's a better idea out there somewhere, or that non-track changes like rearranging seats inside railcars could help more affordably, it's dangerous to think crayonista that doesn't take into account all the issues is going to find a magic bullet.

The loop is going to be the biggest transit lift in the Washington region since Metro was originally built. It's natural to talk about alternatives, and WMATA is going to have to do much more detailed studies before convincing the region's leaders get on board with funding. But beware anyone who says they can do it for free.