A Comparison of Costs for Modern Subway Construction in North America
Spadina and North Yonge Cost Breakdown
TTC Spadina Extension Cost Estimates
TTC Yonge North Extension Cost Estimates
Subway Yard Costs
Rolling Stock Costs
Property Acquisition Costs
Cost Estimates for Stage One of Construction: Cityplace (Spadina) to Wynford (Eglinton)
Cost Estimates for Stage Two of Construction: Cityplace (Spadina) to Dundas West
Cost Estimates for Stage Three of Construction: Wynford (Eglinton) to Sheppard/Don Mills
Cost Estimates for Stage Four of Construction: Dundas West to Pearson Airport
Combined Cost for All Stages
Subway construction is understood to be a costly venture, but there are various factors that contribute to the cost. Beyond the obvious costs of labour and materials, costs can vary depending on methods of construction, as well as geology and other local factors. Thus, costs for similar projects can vary dramatically. The following is a list of projects that have been constructed or proposed in the last ten years. Due to a difference in labour laws and regulations in Europe, the data has been limited to recent North American construction, where although laws may differ slightly, no major variations occur.
Two cities stand out when compared to the cost of projects in Toronto, which have more than doubled in cost estimates for the past decade. Montreal: and New York.
On the cheaper end of construction, Montreal should be studied as to why it is able to deliver subway at a substantially lower cost per kilometre than we have been able to in Toronto.
In Montreal, we have the benefit of seeing both a recently constructed project (the orange line to Montmorency in Laval) as well as a number of proposed projects. Costs range from $143 million to $173 million per kilometre, a cost half that of modern subways in Toronto such as the Spadina extension and the North Yonge extension. Unlike several other cheaper projects that have been constructed or proposed in North America, Montreal’s subway construction is entirely tunnelled, with no sections at grade. The Laval extension in particular had the challenge of tunnelling under the Rivière des Prairies, and still was able to complete construction at a cost of $143 million/km. As nice as it would be to be able to build the Downtown Relief Line for this sort of cost, unfortunately it is not feasible. As well as being of a different geology than the City of Toronto67, a more important factor for Montreal is the width of their trains. Due to thinner trains, Montreal is able to construct their system using one large bored tunnel with multiple tracks rather than being forced to construct two smaller tunnels54. This can reduce tunnelling costs substantially with half as much tunnel required for construction.
On the other end of the spectrum, New York suffers from significantly higher tunnelling and construction costs, with the MTA stating the following in regards to the costs of the Second Avenue Subway.
Our partners in the contracting community tell us that they are forced to charge more for our projects to protect against the perceived risks of working with the MTA and to compensate for the difficult environment of working around service and the complexity of each project. In addition, construction industry practices in New York tend to increase costs. As an example, tunnelling for the expansion projects has cost between three and six times as much as similar projects in Germany, France and Italy162
Because other cities may be substantially cheaper or costlier than Toronto, their estimates are not as relevant to the costing of the Downtown Relief Line, and are simply provided for the sake of comparison. Because we have recent construction cost estimates in Toronto, we can use those as a base for estimations of the cost of the Downtown Relief Line with much greater accuracy.
The Spadina Subway has begun construction on an extension to Vaughan Metropolitan Centre that is projected to cost $2.6 billion148. In 2008, the TTC released a report breaking down costs of the line. Though the total at the time was $2.1 billion, the details allow for a number of cost estimates to be derived. Most of these costs can be broken down to a per kilometre rate, which can be used as a baseline for Downtown Relief Line estimates.
TTC Spadina Extension Cost Estimates149
The estimates of the Yonge subway extension have also been released, though in less detail.
TTC North Yonge Extension Cost Estimates173
Since the initial release of the Spadina extension, the cost of the stations has been re-estimated, and has gone up.
Though it may appear that stations ballooned over their expected cost by 73.5%, it isn’t quite that simple. Initial cost estimates were made prior to station design. In the revised estimates, it establishes what was actually budgeted for in the most recent $2.6 billion estimate. The total costs of the budgeted stations is $610.2 million, making the stations cost 32% over what the specific designs were budgeted for.
These increases are not insignificant, and cast severe doubts upon the estimates of the Yonge line. Using the Spadina Line’s higher station costs then, we learn that station costs can be broken down as such:
To determine the cost of above ground and elevated stations, the easiest method is to compare with the proposed Chicago extensions, where the high end estimate for at grade stations is approximately $25 million for a terminal station and $31 million for inline elevated stations. To be on the safe side, we will estimate a cost of $30 million for an average at grade station and $40 million for an elevated station, not including the cost of parking.
Though cost estimation can fluctuate depending on other factors such as the depth and complexity of the station, these figures can provide a good base for the determination of station costs for the Downtown Relief Line.
Though station costs are a large part of the project, calculating the cost of building tunnels, laying track, and providing the necessary support structures and systems necessary for operation also play a major role.
In the Spadina line, the combination of factors that go into track building totalled $76.3 million per km. The Yonge line has it projected at $88.2 million per km. When considering that the Spadina line has escalated in cost from $2.1 billion to $2.6 billion, approximately $200 million of which can be attributed to stations, there is still $300 million unaccounted for. As a rough estimate, split amongst the remaining costs of the subway, it raises costs by 13%. This brings the cost to $86.2 million per km, very close to what the predictions for the Yonge extension estimate which includes difficult track configuration where crossing the East Don River. Important to note is the cost of structure without tunnelling using the same method of increasing costs proportionately on the Spadina estimate, which results in a cost of $22.6 million per km for at grade sections of construction, or approximately one quarter the cost of tunnelled track. This echoes the 2004 work of the International tunnelling Association, which studied projects to determine that in subway construction projects, median at grade construction cost 29% of what tunnelled construction did, with the average at 35%89.
Because we cannot derive cut and cover cost savings as a result of the TTC’s two extensions that have been studied, the estimate for that must come from the International tunnelling Association document. It has the median of cut and cover construction costing 68% of tunnelling, and the average cost being 84% of tunnelling. The reason there is a significant difference between the median and the average here is outlined in the document, noting that where land prices are high, tunnelling can actually be less expensive. Though relative to cities in the International tunnelling Association document Toronto’s land values are comparably cheap, much of the line will be constructed through the downtown core, and thus using the average rather than the median is probably closer to the expected costs of cut and cover in Toronto. Similarly, elevated track consistently costs approximately half of what tunnel bored construction costs. Because few sections of the line will be built in this manner, this selection of number will not be a major factor in the determination of costs for the line.
Major bridge crossings, such as would be required over the Don Valley, are much more difficult to determine the price of due to no similar bridges being required by the TTC, and bridge construction and length playing a major factor in the cost determination. The gap across the Don Valley is approximately 500m in length, which is comparable to the Saint John Harbour Bridge, of which a replacement was recently estimated at $150 million to construct28. A new bridge over the Ottawa River is estimated at $400-500 million for 1.5 km41. Probably closest in construction to the crossing of the Don Valley is the Park Bridge in British Columbia crossing the Kicking Horse Canyon, which is 404m in length, 80m high, and was constructed at a cost of $143 million118,95. It is therefore not unreasonable to expect a crossing of the Don Valley to register around $175 million. This estimate is higher than what those of the other bridges would infer, but it is intentionally so in order not to severely underestimate the costs of the project. Smaller crossings, such as a reconstruction of the Overlea bridge, would not be as significantly expensive.
Finally, other bridges, such as over small creeks or rivers, are much harder to determine information on how much track would cost. It would likely be similar to elevated tracks, but in order to not underestimate costs I will be using a figure of $100 million per kilometre for this document. This is based on the idea that as a worst-case scenario, a tunnel under the river that would likely be more expensive than regular tunnels due to engineering could be used. Since these bridges are so short and make up such an insignificant part of the overall line, even overestimating by double the cost would put the total estimates off by no more than 1% or so, well within the contingency of 25%.
The costs of constructing the tracks between stations in Toronto is as follows:
Though certain segments of the line should be relatively easy to construct, there is no doubt there will be a number of segments that will pose significant engineering challenges. Although the cost of engineering used by the Yonge line is almost three times higher than the Spadina line, rising costs on the Spadina line combined with the relative complexity of the Downtown Relief Line mean that engineering on the DRL will likely be somewhat pricey in certain segments. On others, however, tracks can run at grade and engineering is a lot less difficult. We will therefore use $80 million per km as a baseline for engineering costs for the Downtown Relief Line, with costs being adjusted slightly above or below that cost based on the complexity of the construction stage being evaluated. At grade segments will cost significantly less per km than complex tunnels.
Subway Yard Cost
A new subway maintenance and storage facility would need to be constructed as a result of any significant expansion in Toronto’s subway. In a November 2009 report, the TTC projected the cost of a new maintenance and storage facility slightly larger than the current Wilson yard for the Yonge line at $632 million175. Given the land value and availability around Yonge Street, it is doubtful that cost would be any higher for the proposed location of a subway yard at the former Lever Brothers factory at base of the Don River. For the purposes of this cost estimation, $632 million will be added for the provision of a new subway yard.
Using the Spadina extension’s determination of 56 new subway cars for an 8.6 km line means that roughly 6.5 cars per kilometre of new subway are required. One of the new Toronto Rocket cars currently costs $3 million22. Thus, $19.5 million becomes the price per kilometre necessary for rolling stock.
Property acquisition is required for stations and when the track runs in an above ground right-of-way. It is not needed when the track passes below property at a depth that it does not cause disturbance to the buildings above105. Higher property costs for stations in the downtown can be offset by the lack of need for bus terminals, and the ability to convince property owners of the value of a direct entrance from the subway to their building, integrating the entrances into existing or proposed properties rather than needing a property for only the station itself.
A contingency cost equal to 25% of the total will be added to cost of the estimation, not including property or rolling stock.
Cost Estimates for Stage One of Construction. Cityplace (Spadina) to Wynford (Eglinton)
Stage One of construction will be the most expensive, posing the greatest challenges from an engineering perspective, having only a minimal part of track at grade level, requiring the crossing of the Don Valley and very deep tunnelling through the downtown core. Additionally, the cost of the subway yard will be included in the cost of Stage One, while it will not be necessary for future stages.
Stage One consists of thirteen stations. From southwest to northeast, they are as follows:
The positives in terms of station cost estimation are that Wynford and Cherry are the only stations that would require further expenditure in terms of a bus platform and commuter parking in the case of Wynford, and a streetcar transfer platform in the case of Cherry. All other stations would serve passengers through a walkup model with transfers potentially required if a smart card system still is not operational.
The expensive side of these stations comes in terms of the three most southeasterly stations, which must be extraordinarily deep by Toronto’s standards in order to tunnel under other subways, utilities, the PATH network, and building foundations. In terms of depth, they may reach the point at which it becomes more economically feasible to construct them using tunnel boring machines rather than the traditional method of cut and cover for stations, as stations deeper than 30m underground become cheaper to construct with a tunnelling method36. Between the standard depth of 15m and 30m, cost increases by $1.9 to 2.2 million per metre of depth. Specifically, the Financial District station must not only be deep, but additional costs will occur due to the likely necessity to upgrade PATH tunnels that connect Union, St. Andrew and King stations to allow for transfers.
From the southwest, the track will begin at Cityplace station from around Bathurst and the railway tracks. From there, it will be constructed deep underground using a tunnel boring machine for 2.3 km to St. Lawrence station, reaching a maximum depth of 30-35 metres underground west of Financial District station. From St. Lawrence, it will be constructed as cut and cover for 700m along The Esplanade, where again tunnelling will be used for 1 km to a point just west of the Don River, where the track will come above ground, cross the Don River on a small new bridge, and run at grade for 1.9 km to just past Gerrard, where it will again enter a tunnel. This tunnel will run for 3.4 km to the intersection of Pape and Donlands, where a new bridge of 500m will need to be constructed to carry the subway across the Don Valley. Upon reaching the other side of the Don Valley, construction again will tunnel under Thorncliffe Park where it will continue for 700m to the western edge of the Don Valley, behind the towers on the eastern side of Thorncliffe Park. Track will run at grade for 400m before joining Overlea Blvd. and recrossing the valley on a rebuilt Overlea Blvd bridge 200m in length. From that point, construction will remain tunnelled to the end of the line 1.6 km away, at Wynford station. Tail tracks are not included in this estimate as they are included in the costs of the terminal stations. Bridges may need to be widened when following the railway corridor to accommodate the extra tracks.
Stage One Cost
The cost of Stage One is therefore estimated to be as follows:
This estimate of $5.95 billion for the line, or $470 million per kilometre, may seem high but is probably in line with the difficult engineering involved with routing a subway downtown that does not always follow major roads.
The main problem is that although these estimates may appear expensive, the cost will only continue to climb for a piece of infrastructure which is undoubtedly necessary for the long-term health of a city that as of 2010 has one of the worst commutes on the planet69.
Cost Estimates for Stage Two of Construction. Cityplace (Spadina) to Dundas West
Stage Two of construction will be relatively cheap, being able to stay at grade for most of the construction so long as the potential development of the Downtown Relief Line is accounted for when the Georgetown Rail Corridor is reconstructed.
Stage Two consists of eight stations. From southeast to northwest, they are as follows:
Though running through the rail corridor, many of these stations must be below ground as a result of the Georgetown Rail expansion.
From the southeast, the track will begin where the tailtracks from Cityplace station left off from around Bathurst and the railway tracks. From there, the tracks will tunnel 600m under the GO Lakeshore Line and continue as a bored tunnel for 2.9 more km to a point slightly beyond MacDonell station. There, they emerge in the railway corridor where more space exists than further south to continue 1.5 km to Dundas West Station. Tail tracks are not included in this estimate as they are included in the costs of the terminal stations.
Stage Two Cost
The cost of Stage Two is therefore estimated to be as follows:
$1.85 billion, or $372 million per km, will get us a western section of the Downtown Relief Line to allow passengers traveling from the west a quicker journey downtown, as well as provide important service to Parkdale and Liberty Village. Though this estimate may be high, it is expected that station costs must be higher than average in order to accommodate the construction of them without using cut and cover construction, to avoid disruption to the rail corridor above.
Cost Estimates for Stage Three of Construction. Wynford (Eglinton) to Sheppard/Don Mills
Stage Three of construction will be relatively cheap, being able to stay at grade for most of the construction so long as the potential construction of the Downtown Relief Line is accounted for when the Georgetown Rail Corridor is reconstructed.
Stage Three consists of four stations. From south to north, they are as follows:
Stations not already constructed (The Donway and Moatfield) will have a small bus terminal. A terminal already exists at Sheppard/Don Mills as does commuter parking.
Track will be tunnelled the entire 5.7 km length of the line from the tail tracks at Wynford to Sheppard/Don Mills Station. Tail tracks are not included in this estimate as they are included in the costs of the terminal stations.
Stage Three Cost
The cost of Stage Three is therefore estimated to be as follows:
$1.9 billion, or $333 million per km, will connect the Downtown Relief Line with the Sheppard Subway, the Sheppard East LRT, and potentially the Finch West LRT, forming a major transit hub at Sheppard/Don Mills. This will give a huge boost in terms of network connectivity and taking even more riders off the Yonge subway.
Cost Estimates for Stage Four of Construction. Dundas West to Pearson Airport
Stage Four of construction will finish the Downtown Relief Line, providing an alternate route to Eglinton to the airport which brings rapid transit closer to the dense and underserved communities of Weston and North Etobicoke, while providing a more direct path to the airport along Dixon Road, hitting some important trip generators.
Stage Four consists of thirteen stations. From southeast to northwest, they are as follows:
Stations along this part of the line vary in complexity, but none will require expansive bus bays or extreme engineering techniques. Some will be at grade, others below ground, and a few elevated.
From the southeast, the track will begin where the tailtracks from Dundas West station left off. From there, the tracks will remain at surface for approximately 800m to a point just beyond Junction station. From there, they will enter a brief 700m tunnel to get underneath the railway junction, after which they will emerge back at grade to travel for 1.7 km to Black Creek Dr., where a bridge will carry the tracks 250m before reemerging at grade for 3.25 km to just beyond Weston station. From there, they will enter a brief tunnel at John St, curving west to Fern Ave. and Weston Rd 600m away, where the tracks will briefly appear above ground to bridge over the Humber River for 200m. The tracks then enter back underground, using cut and cover along the length of Dixon 4.5 km until leaving Highway 27 station, just east of Highway 27. There the tracks become elevated and run that way 3 km to Pearson Airport. Tail tracks are not factored into the calculation as they are already included in terminal stations. Cut and cover is used along Dixon to save costs due to the width of Dixon and minimal businesses making cut and cover construction minimally disruptive. Elevated track is chosen west of Dixon to further decrease costs and due to a lack of residential housing and airport related land uses which minimize development potential.
Stage Four Cost
The cost of Stage Four is therefore estimated to be as follows:
$3.2 billion, or $215 million per km, will get us a route to the airport, a subway through some of the highest density areas of the city, and much closer transit to Weston and North Etobicoke.
The total cost of all stages of this proposal is $12.9 billion for 39.9 km of revenue track. If the City of Toronto builds one kilometre a year on average, the cost would be $323 million per year to make a significant change in our subway system by 2050. Speeding up the implementation to 2030 to try to avoid even further cost escalation and get these important routes constructed means a yearly cost of $646 million. A notable portion of these costs will be required regardless of whether or not the Downtown Relief Line is built in order to relieve congestion on the Yonge subway. The longer we wait the more these costs continue to grow, which is why it is imperative that construction on the Downtown Relief Line begin as soon as possible.
From top to bottom: