Sydney’s new light rail: not the world’s most expensive after all

With a total construction cost of $2.1 billion, Sydney’s planned CBD and Eastern Suburbs Light Rail (CESLR), it is not surprising that the project has been dubbed as the world’s most expensive light rail line. As my series of articles on Australian light rail projects demonstrated, at $175 million a kilometre it will certainly be Australia’s most expensive. On the world stage however, claims that the CESLR has this dubious honour have just been well and truly demolished.

While we await confirmation from a higher authority (Guinness World Records?), the new world “champion” of expensive light rail projects is almost certain to be Boston’s Green Line extension. Before I discuss the dollars involved, a little background: while the Green Line runs underground through the city’s CBD and is referred to in Boston as part of the city’s subway system, it is actually a light rail line which runs on the surface outside the CBD, either in on-street reservations, in dedicated corridors (mainly converted heavy rail lines) or even in mixed traffic.

The underground section is North America’s oldest subsurface urban rail line. It opened in 1897 and was originally constructed to reduce streetcar (tram) congestion in Boston’s CBD. Some other US cities also constructed underground lines for similar reasons. Decades later when the mass closure of streetcar systems occurred swept across the US, Australia and other countries around the world, most of these underground sections were also abandoned or in a few cases (and as happened elsewhere in Boston) converted into heavy-rail subways.

Boston Green Line underground station

Boston Green Line underground station (author’s photograph)

A small number of cities like Boston and San Francisco retained key underground sections, along with some of the surface lines connected to them, as light rail and tram lines. In the case of the Green Line the extent of the underground section and the level of general traffic congestion made it difficult to replace with buses; it also avoided conversion to a full subway because the underground section has a number of tight curves and the stations are in some cases only a few hundred metres apart, factors which made it easier to continue the light rail service.

Today the system has four above-ground branches starting in the city’s west and south west which join together to form a single underground line running eastwards and then north. The line has nine underground stations including North Station at the north east edge of the CBD, where it surfaces before crossing the Charles River on a bridge. The line terminates over the river at Lechmere station, just south of the densely populated Somerville and Medford areas.

Boston Green Line station in a street reservation

Boston Green Line station in an above-ground street reservation (author’s photograph)

The planned extension involves 6.9 km of new track including a 5.5 km main branch with five stations within the existing Lowell Line commuter rail right-of-way, running north west from a rebuilt Lechmere Station to the Tufts University campus at Medford. A 1.4 km second branch will follow the existing Fitchburg Line commuter rail right-of-way westwards to a single station at Union Square. The light rail would run on its own track within the commuter rights-of-way as the existing heavy rail lines and services to North Station will be preserved (a more detailed map is available here).

Green Line Extension Map

Green Line Extension Map (from Wikipedia)

Towards the end of 2014 (around the time the $2.1 billion estimate for Sydney’s light rail was released) it was announced that the Green Line project is to be jointly funded by the Federal Government with a grant of US$996 million and the Commonwealth of Massachusetts, which will provide the balance of the project’s total US$2.29 billion cost. In Australian terms this translates to a rather mind-boggling AU$2.99 billion, or nearly $900 million more than the CESLR.

I have not been able to track down an economic analysis for the line but even if the estimated ridership on the extension of around 37,900 a day (roughly 18 million annually) is correct, it is hard to see that a project this expensive would have a great cost-benefit ratio. In the context of Boston’s rather convoluted transport and environmental politics, however, it does make some sense.

To cut a very long story short, although commuter rail corridors run through Somerville and Medford, only the periphery of the area is now served by rail (ironically there once were several stations on the commuter lines in this area but these are now closed).  To settle a lawsuit brought by an environmental law group seeking the implementation of measures to mitigate the environmental impacts resulting from The Big Dig (an ambitious program of motorway tunnel construction in inner Boston), the Massachusetts government agreed to extend the Green Line north into Somerville and Medford. Thus the rationale for the line is legal and political as much as it is environmental and economic.

Neither this nor inclusions such as stabling facilities and 24 new light rail vehicles (LRVs) explain why the project will cost a staggering A$433 million per kilometre to build. This is despite it being constructed in existing commuter rail corridors, which means that there will be minimal land acquisitions. This compares to a relatively modest $175 million per kilometre to build Sydney’s 12 km CESLR project which also includes stabling facilities and 30 LRVs (and which also has minimal land acquisition costs).

In other words, Boston’s Green Line extension will cost almost two-and-a-half times more per kilometre than Sydney ‘s CESLR. It will also carry fewer passengers than the CESLR, though to be fair the level of utilisation on a per-kilometre basis is likely to be similar, given the Green Line extension’s shorter length.

The anticipated high levels of usage on both systems will also test other public transport infrastructure and connecting services in their respective cities. In the case of Boston the extension will feed into the busy CBD section of the Green Line (which is the busiest light rail system in the US), while in Sydney the line will have interchanges for buses at its eastern termini, with yet another bus interchange to be located on the edge of the CBD at Rawson Place where inner west bus passengers will be transferred to already-crowded trams (I will have more say on international patronage comparisons in a future post).

To an Australian observer the first thought on seeing the Green Line extension plans is that it would be much simpler and cheaper to convert part or all of the existing diesel commuter line to electric suburban operation and construct extra stations through the Medford-Somerville area. Part of the reason this approach wasn’t adopted may be that electrified suburban rail services are a comparative rarity in the States. Many cities have high-frequency subways servicing the CBD and inner suburbs but few have electric suburban and inter-urban services (New York’s Long Island Railroad is a notable exception).

Most US commuter rail systems are like Boston’s – diesel loco-hauled, low-frequency services which are geared largely to the morning and evening commute. Many cities don’t have any commuter rail services at all. This contrasts with Australia where most mainland state capitals have reasonably frequent and relatively modern electric suburban services  (on the other side of the ledger, no Australian city at present has a dedicated subway/metro).

A more practical objection to electrification of the Fitchburg and Lowell commuter lines is that passengers would still have to change at Boston’s North Station to access subway services to travel further into the CBD, a situation not unlike that in Sydney were passengers have to change from interurban trains at Central Station. Nonetheless electrification would appear to the outside observer to be much more cost-effective than the light rail extension and the money saved could be used to electrify additional commuter lines and provide extra stations throughout the network. Alternatively any spare funds could go towards providing the missing direct rail connection between Boston’s North and South Stations, ideally with an intermediate underground station which electrified commuter trains could use.

Of course there are other options that could be considered for Medford and Somerville such as using tram-trains on the commuter lines or even constructing a heavy-rail subway line, but as an outsider it is probably unwise to attempt to provide detailed advice on overseas urban transport plans unless one knows the cities very well. It is equally unwise to make superficial comparisons between cities as these often mask major differences.

It is probably fair to say however that while Boston’s Green Line extension and Sydney’s new light rail line are both good projects that will bring rail access to some middle and inner-ring suburbs with high population densities currently reliant on crowded buses and congested roads, they are approaching the limits of light rail both in terms of cost and passenger capacity. While both projects are now underway and I believe they should be supported, the future assessment of similar light rail proposals at these limits should include consideration of whether extra investment in heavy rail or metro/subway infrastructure would be more appropriate, especially in terms of catering for future growth in demand.


This entry was posted in Infrastructure, Public Transport, Sydney metro area, Transport and tagged , , , , . Bookmark the permalink.

10 Responses to Sydney’s new light rail: not the world’s most expensive after all

  1. Tony Bailey says:

    You can’t just compare one system to another without specifying exactly what is included.
    That Sydney price includes vehicles and operations and maintenance for the period of the contract, not just construction costs.


    • Alex says:

      Thanks for the comment. My aim was first, to question the assumption that the CESLR is the most expensive LR project in the world and second, to point out that both this and the Boston projects are very expensive to the point where you have to wonder whether other heavy rail alternatives would be more cost-effective. I wasn’t attempting to make a detailed comparison though perhaps I should have made this clearer.

      Obviously it is very difficult to take into account all inclusions and exclusions in a blog post but I did point out that the price for both systems does include vehicle purchase. I did overlook the fact that the CESLR price also includes operations and maintenance, though including these actually flatters the CESLR costing in relation to the Boston Green Line extension. I didn’t include the breakdown for the Green Line which has been published but it clearly does not include any operational or maintenance costs.


  2. Tony says:

    You’re comparing with real “light rail” lines – basically heavy rail with tram vehicles on them. CSELR is a street tramway, not “light rail” in the American sense. Compared with street tramways, particularly in Europe where they know how to build them efficiently, it will be by far the most expensive.


  3. Alex says:

    Actually many North American systems combine separate rail corridors, dedicated on-street reservations and street running in mixed traffic, for example Boston, San Francisco and Philadelphia. While these are older systems some of the newer ones also have sections with on-street reservations.

    My comparison in this case was specifically with Boston which is very much a traditional light rail line with all three elements. One of the western branches is a converted heavy rail line; the other three involve on-street reservations and in one case operations in mixed traffic. The CBD underground section while obviously grade-separated was built to light rail and not heavy rail standards and was not subsequently converted to heavy rail (unlike some of the other underground light rail sections).

    In any case I question the claim that the CESLR is a “street tramway” rather than “light rail”. While much of the CESLR will run on-street this will be mostly within dedicated reservations and not mixed traffic. This is directly comparable to elements of many American and European light rail systems (including elements of the Boston system), but not I believe to “traditional” street tramways like those in Melbourne or Toronto (or parts of Europe).

    In addition the CESLR will incorporate significant sections of completely separate corridor such as the tunnel under Anzac Parade and Moore Park and the bus (and former tram) corridor alongside Anzac Parade and Alison Road as they pass Moore Park and Centennial Park respectively. Finally the stops are further apart than most tramways and similar to the distances commonly found in light rail systems and again the plan to run coupled units carrying up to 450 passengers is more like light rail than tramway.

    I agree that the CESLR will be pretty expensive; in fact my point was that it wasn’t the world’s most expensive but that it was also certainly up there. I don’t know much about the costs of European systems apart from the fact that the Edinburgh system cost close to a billion pounds including financing costs but that was a very troubled build. I’m sure that in other places they have been built much more cheaply but I still think the Boston comparison has some validity.


    • Tony says:

      A street tramway (whether in its own lanes or not) is subject to road traffic management such as traffic lights. This is CSELR. Light rail is grade-separated and subject to its own safeworking regime. IWLR in Sydney is light rail from Dulwich Hill to Darling Drive and a street tramway from Darling Drive to Central. Yes, a lot of lines are a mixture but you need to be comparing like with like in terms of that mixture.


      • Alex says:

        I think we’ll have to agree to disagree on this.

        I was struck by how expensive both projects are and therefore wanted to make some ballpark comparisons rather than undertake a detailed analysis. I know they are very different in some respects but they have enough in common – for example, patronage per kilometre figures are both very high for light rail and broadly similar, the cost estimates both include vehicle purchase (which are themselves also similar) but on the other hand land acquisition costs are minimal – to make an interesting comparison.

        While the approaches to signalling and operation may differ I also don’t think these factors have that big an impact on costs. Don’t forget that while the Green Line extension will probably have wayside signalling similar to the underground section and the D-Riverside branch, the line itself is being built as light rail within the heavy rail corridor (ie it is not taking over the heavy rail lines). In this regard it will probably be similar to the Anzac Parade/Alison Road section of the CESLR. In some respects the CESLR will also be very train-like in operation, running double-coupled LR vehicles with up to 450 passengers at four-minute intervals.

        In any case as I said I was much more interested in the broad point of why both projects and in particular the Boston extension are so expensive and I was looking at this in the context of a blog post rather than a peer-reviewed article. I also wanted to make the more mundane point that the claim made by many of its critics that the CESLR is be the world’s most expensive light rail system may be misplaced.

        And if you say that somehow CESLR really is still the world’s most expensive (whatever it is) since it falls into a different category to Boston’s Green Line, then just how many categories are there and how can you compare any systems anyway when no two lines are exactly alike?


  4. Tony says:

    Even if we agree that there are no differences between types of system, you’ve only been able to find one example that’s more expensive. CSELR is still at least two to three times as expensive as similar tram projects in continental Europe. I specify continental Europe because in UK, like USA and Australia, costs of these projects seem to be grossly high. Mainly down to lack of experience and expertise and naive governance.


  5. Alex says:

    I certainly agree with you about the cost of the CESLR and I did discuss this in an earlier post. I also agree about how unfavourably the costs of light rail in Australia (particularly the CESLR), UK, Australia and I suspect Canada compare with those in Europe.

    While it is difficult to find consistent data and make realistic comparisons for all the reasons we’ve discussed, I did have a look at some stats, especially for France. I concentrated on Paris – while the earlier LR lines there mainly involved the repurposing of existing rail lines, the extensions and new lines planned appear to be mostly street-based. In addition the land acquisition and construction costs involved on the periphery of Paris are probably more broadly comparable to those in inner Sydney than the projects planned in some smaller French and cities (which of course are even lower).

    Some of these costs can be found in this article. The upshot is that at current exchange rates the average cost including rolling stock purchase is $74.7 million per kilometre. In other words, at $175m per km the CESLR is costing 2.3 times the average price of the planned LR lines in Paris. As I indicated the costs in smaller French and German cities are even lower, from around $25 to 45 million per km.

    By the same token, I still think that because of the capacity and other issues I discussed in my earlier posts on the CESLR a metro-based alternative would have been a better solution. However with the signing of the contract this is a moot point, as is I’m afraid the total cost.

    I’ll follow this up and acknowledge some of the issues you raised in a future post, but for the moment I must stop work-avoiding!


  6. Dudley Horscroft says:

    You state:

    “The underground section is North America’s oldest subsurface urban rail line. It opened in 1897 and was originally constructed to reduce streetcar (tram) congestion in Boston’s CBD. Some other US cities also constructed underground lines for similar reasons. Decades later when the mass closure of streetcar systems occurred swept across the US, Australia and other countries around the world, most of these underground sections were also abandoned or in a few cases (and as happened elsewhere in Boston) converted into heavy-rail subways.”

    The Tremont Street Subway (its original name) was opened in September 1897 as a tram tunnel, taking trams from southern suburbs and northern suburbs. At the same time, the Boston Elevated Railway was finalizing its plans for the new elevated lines. Legislation prevented new elevated lines in the CBD and steam power was not allowed, so the heavy rail system was connected to the tram tunnel. Luckily, or perhaps not, the southern portion from the ramps from the street trackage had been built with four tracks, as had the northern section, but there was a short two track section in between from Park Street to Scollay Square (now Government Center). Originally trams on the outer tracks ran through, the trams on the inner tracks turned at these stations.

    The Main Line Elevated (original name, then the Washington Street Tunnel, now the Orange Line) was to operate from Dudley in the south (extended to Forest Hills in 1906) to Sullivan Square in the north, each of which were very large interchange stations with the numerous tram routes that terminated there. On June 10, 1901, the Main Line Elevated started running, and took over the outer tracks of the Tremont Street Subway, leaving the trams on the inner tracks, without through services, much to the displeasure of the passengers. This is hardly “Decades later”!

    As usual, the patronage was underestimated and work began to extend short stations to cater for four car trains instead of three (all stops were extended to take five car trains from early 1905, and six car trains from 1908). By 1903 it was realized that a separate tunnel for the heavy rail trains was needed, and plans eventually agreed on the current Washington Street tunnel.

    On Monday 30 November 1908, the new Washington Street Tunnel was opened and the heavy rail trains no longer ran througn the Tremont Street Subway.

    The above details are from “Boston’s Main Line El: The Formative Years 1879-1908” by George Chasson, Jr. Headlights, 1987.

    Now to the cost of the Green Line extension. The existing double track main line railway runs in a generally wide reservation, enough for four tracks. There are no stations between West Medford (further out) and the junction with the Green Line. I understand that it is proposed to shift the existing railway to one side of the reservation, with the tram tracks on the other – this was preferred to simply added one tram track on each side of the existing rail tracks. This means that the cost is not just for a 6.9 km tram line, but for at least 13.8 km of new track, and no doubt the new commuter railway track will be to the very highest standard (including completely new signalling and communications for the heavy rail). In addition, there will no doubt be flyovers to separate the light rail and heavy rail tracks at the crossing of the Union Square and Medford branches. Nothing in Boston is done cheaply – all same TfNSW! It is logical that there should be a new commuter rail station at College Street and Union Square for interchange between the Green Line and the heavy rail – no doubt as expensive as they can get away with!

    In short, while the Green Line extension looks more costly than the CSELR, the difference given the extra work included means it may in fact be the cheaper. Sydney may yet retain the title!


  7. Pingback: What happened in StrategicMatters in 2015? | StrategicMatters

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s