Don’t Talk About the North-South Rail Link without Context

Boston’s North-South Rail Link is in the news again. And while blog readers know I think the project should have been finished a century ago, and anyone who follows me on Twitter knows I’m a big fan of the idea, I’m a little frustrated with the way dialogue about it seems to be going.

The current round of NSRL discussion was set off by former Massachusetts governors Michael Dukakis and William Weld, who penned an op/ed in the Boston Globe arguing against the Baker administration’s plans for expanding North and South Stations and for building NSRL instead. Perhaps Dukakis and Weld are working through some of their own culpability for not getting the project built as part of the Big Dig, as originally planned; regardless, I certainly agree with them that expanding the dead-end terminal stations is utterly foolish and wasteful.

I also, however, agree with some of Gabrielle Gurley’s rejoinder in Commonwealth Magazine, which argues that Dukakis and Weld’s piece failed to factor in the realistic politics of taxation and spending in Massachusetts, or present a realistic plan for building the very expensive tunnel. Gurley’s piece is far from perfect–she lumps NSRL, a project that would massively increase system efficiency, with South Coast Rail, a boondoggle that continues in the T’s long tradition of overengineered, efficiency-sucking outward expansions–but she does have a point that the politics of spending on large infrastructure projects are tricky at best. Boston has had some traumatic experiences with massive infrastructure projects, and the T’s problems with project management are both well-known and ongoing.

Dukakis is nothing if not determined, though, and he wrote back to Commonwealth, arguing the Gurley’s math had failed to take operational savings and new ridership into account:

And third – a point your piece totally misses – because the thousands of new passengers will produce approximately $120 million dollars a year in new passenger revenue and $80 million in maintenance savings. And that doesn’t even factor in the “value added” that will come from private leasing and development along the route, some portion of which can be captured as additional revenue.

Those combined funds will support a 20-year bond issue that can pay for the project. That is why it will be far easier to win public support for the project and why your column today totally misses this key point.

Dukakis’ numbers (he seems to assume a project cost of $4.4 billion) may or may not be accurate; it seems to me that he states them with far too much certainty for a project that hasn’t been extensively studied in at least 10 years. What bothers me is that both he and Gurley treat NSRL as if it’s a a standalone project. It’s not, and it can’t be.

Currently, the entire MBTA commuter network–even the Providence line that runs under Amtrak-owned catenary–runs traditional (for America) commuter rail trains consisting of a diesel locomotive and towed coaches. Diesels won’t work in a long tunnel, for obvious ventilation reasons, so any NSRL will of course be electrified. Some past studies have assumed that the MBTA and Amtrak would switch to using dual-mode locomotives on some or all lines once a tunnel is built, but that’s kind of nuts; while New Jersey Transit uses dual-modes in regular service, they’re heavy, slow to accelerate, and inefficient. Realistically (and desirably in terms of service efficiency, quality, and reliability), large-scale or complete electrification of the MBTA network must accompany NSRL.

And that’s a good thing! Electric trains are faster, quieter, and more reliable. Electrifying the whole network and treating it like a regional rapid transit system would be a huge boost to mobility for the Boston area. But it also adds considerable expense to the NSRL project. Alon ballparked a figure of $1.5 billion for complete electrification in his post on the topic. Amtrak’s New Haven-Boston electrification was contracted at $2.3 million per double-track route-mile, though Paul (and others) think the final number was higher. $3 million per route-mile seems reasonable; you might save some on single-track sections (although the underlying infrastructure is the expensive part), and save in other places. Using the $3m/mile number, and excluding the already-electrified Providence Line (though that line, as F-Line points out, would need upgrades to the electrical infrastructure because of Amtrak’s cheapness), I come up with a number lower than Alon’s, just under $1 billion: (information from here)

Line	Route-Miles	Cost @ $3 m/route-mile
Newburyport/Rockport	34.9	104.7
Gloucester Branch	16.8	50.4
Haverhill	32.1	96.3
Lowell	24.5	73.5
Wildcat Branch	3	9
Fitchburg	48.8	146.4
Worcester	44.3	132.9
Needham	9.3	27.9
Franklin	27.4	82.2
Stoughton	4.6	13.8
Fairmount	9.2	27.6
Middleboro	35	105
Greenbush	17.3	51.9
Kingston/Plymouth	24.2	72.6
Total	331.4	$994.2

At a minimum, we can say that complete electrification will add $1 billion to the cost of NSRL, and quite likely more. There are ways to reduce that number; most transit pundits think the Needham Branch should be converted to an extension of the Orange Line, and you might not bother electrifying Kingston and Greenbush because of low ridership. On the other hand, Amtrak might want to electrify the all the way to Portland and convert the Downeasters into Regional extensions.

Lest anyone accusing me of scaremongering, I will clarify that electrification of the MBTA system is something that should happen regardless. This is especially true of the South Side network, which persists in the bizarre situation of running aging, slow, unreliable diesels under catenary on the Providence line. It’s about bringing an aging system into the last century. But it needs to be part of the conversation on NSRL–both because of the additional costs and because making the case for NSRL should be only part of making a comprehensive case for modernizing the whole MBTA system. 

Electrification, of course, comes with additional costs. The biggest is rolling stock. By my count, based on Wikipedia, the current MBTA commuter rail fleet has almost 67,000 seats split among 457 or so coaches. Depending on configuration, Metro-North’s M8 EMUs seat about 110 people, meaning you’d need over 600 equivalent units to replicate existing MBTA fleet capacity. At $2.54 million per unit for the initial M8 order, that’s over $1.5 billion to buy EMUs. Things could get a little cheaper; Metro-North is clearance-limited by the Park Avenue Tunnel, whereas MBTA uses a large number of bilevel coaches and NSRL would presumably be built to clear them. The newest (only?) US bilevel EMUs, Metra’s Highliners, cost $3.6 million per unit, and seat 128 (not that great a capacity improvement over single-level coaches, because Metra’s gallery-car setup sucks–MBTA bilevels seat up to 185). MBTA could also choose to buy electric locomotives to haul the existing coach stock.

The final major expense that might accompany NSRL is the cost of building high-level platforms at all, or almost all, MBTA stations. Many already have them; it’s just good practice. Most of the MBTA network sees little or no freight traffic, so clearances really should not be considered an issue. Having level boarding is especially important at outlying stations when all trains are running through a constrained central segment such as a tunnel, and thus must keep to exacting slots. I count 132 stations on MBTA lines that might need high-level platforms; I’m too lazy to look up how many already have them like I did for the Morris & Essex Lines, and how many stations have 1 as opposed to 2 platforms. Cost estimates for high-levels vary; somewhere in the ballpark of $5 million per platform seems reasonable, while MBTA’s fairly simple new Fairmount Line stations cost $6.9-$9.4 million each. If each of the 132 stations costs $5 million for a high-level platform, the total cost would be $660 million; if it’s more like $9.4 as on Fairmount, it would be $1.2 billion. The actual number would likely be somewhere in the middle. 

For the record, Dukakis’ estimate of $4.4 billion in tunnel construction costs seems reasonable if (a HUGE if) the project is managed correctly. Though he and Weld failed to get the tunnel included in the Big Dig, despite it being included in the initial plans at various points, the construction did leave “slurry walls’ underneath the freeway tunnel that make future tunneling easy. The real complication is the need to construct portals for various lines, winding them between many layers of infrastructure. Including my estimates here, the costs of a modernized MBTA rail system would be:

$4.4 billion for the NSRL tunnel

+

$1 billion for electrification

+

$1.5 billion for new rolling stock

+

$900 million (roughly splitting the difference) for high-level platforms

=

a total of $7.8 billion for complete modernization. 

That might be optimistic; one cannot count on the T to manage projects well, and the projected electrification costs may well be low. But it should be fairly comprehensive.

I’ve just spent 1,300 words talking about how expensive it would be to build NSRL and the other improvements it requires. The final number is truly massive, almost twice what Dukakis seems to assume in his recent writing. And yet, I’m doing this as a fan of the NSRL project. Why? Because not acknowledging the true price of the project also risks obscuring its true benefits. Seeking to win over opponents by minimizing the projected cost of NSRL isn’t going to work; Americans are too paranoid about government spending for that.

It’s very easy for skeptics to point to a multibillion dollar price tag and compare the NSRL project to disasters like the Big Dig. But there’s a key difference: the Big Dig represented a choice for path dependence, for continuing Boston’s dependence on highways at massive cost. It was the natural next step for the highway network, representing a relatively minor improvement in functionality–and a major improvement in aesthetics. But it didn’t provide a new function the road network had never provided before. By contrast, NSRL would be revolutionary, not evolutionary. It’s not “just a tunnel,” and if you’re having the discussion on the grounds that it is, you’ve already lost. Proponents of NSRL need to talk about it in terms of its potential to truly transform the entire system, and that means detailing all of the additional functionality that NSRL can provide–and acknowledging its cost.

Advertisement