The Increasingly Exorbitant Cost of (Transit) Infrastructure

[DirectionNorth]

This is something I've been thinking about for a while, both in UT posts and outside of it. Additionally, I've seen lots of allusions to this in many of our transport threads, so I thought I'd start a dedicated place to debate this.

In the last few years, we've seen the cost of transit infrastructure in Toronto explode.

Surface LRTs that cost in the region of $100 million/km (Finch West) a few years ago are now double that or more (Hamilton, Hurontario).^ TYSSE cost $450 million/km and we all derided that project as overbuilt. Now, SSE and YNSE are running towards $700 million/km. Eglinton West costs double what ECLRT cost, ($500 million/km vs. ~$280 million/km), but while the project's wholly tunneled, there's no underpinning or ROW difficulties to deal with. John Tory's always idiotic Smarttrack is at $1.4 billion for five stations. European cities can build an entire tramway network for that. My understanding is that GO RER is also overpriced for the work involved, but there aren't many comparisons and there's no final price yet.

Perhaps worst of all, the Ontario Line's capital cost is $19 billion, which works out to over $1 billion/km. This is a line that skimps on alignment: there's significant at-grade track, and tunnelling is only used where it has to be (under Queen and Pape).

This is in comparison to downtown European projects, such as Berlin's U5 from Alexanderplatz, which cost maybe $400 million/km (CAD); Paris's RER E to La Défense, which was a similar amount; Milan's Line 4, for $200 million/km; Barcelona's L9/10, a little over $220 million/km; even Amsterdam's Noord-Zuid Subway, built in the Dutch soil and taking 14 years, cost maybe $700 million/km.

(aaaand you get the point)

What gives?

I'm trying to read on it, but I can only scratch the surface of this issue, and its complexity means that I can't read everything and my knowledge base isn't that deep.

My best understanding of the situation is that there's poor oversight of designing projects. Our 1980s-2010s transit building gap means that expertise that IIRC built the YUS/BD subways for $100 million/km or less has left or retired, leaving inexperienced people in charge. Design is poorly done, and there's political meddling, so costs are higher, which makes for more politics. Meanwhile, we tunnel under existing surface ROWS (EWLRT) and build at-grade in the downtown (OL). We contract in large packages, so smaller firms can't compete; hide everything, making transparent pricing impossible (Metrolinx is very, very guilty); we delay because of political uncertainty (see: Hamilton LRT); and at the end of the day we end up with worse results.

(I'm not informed, humour me here)

Do the good people of UT have any insights? Is there some big thing I missed? And what is being done/can we do to ensure that our future projects will remain be reasonably priced?

^all prices come with inflation built in, as according to the BOC's inflation calculator

 

[crs1026]

I'm not well read either, but I have a few theories.

One is that we do load a lot more into designs than we ever did before.... ie more exits from a subway station, more stairways and escalators, deep bore stations that demand a lot more work to excavate and construct. There are few "vanilla" designs or structures. Even our LRT platforms have more cubic feet of concrete and tonnes of steel than anything we built previously. And there are more systems, more wiring, more everything in them.

A second is that we allow much greater contingencies, and we pad schedules more as a way of assuring that different contractors or trades don't overlap. That lengthens project timelines, which amounts to greater financing burden. Project teams last longer.

A third is that we time projects to a political schedule and not to an optimised workload for contractors or workforces. A couple of decades back, the theory was, we should build x miles of subway per year. Mike Harris killed that idea and we've never recovered to an "optimal" but steady pace.

A fourth is that our work methods are more risk averse. I marvel at how projects are protected by jersey barriers and fencing where in past years there might only have been sawhorse barriers or pylons. The effort required just to site prep a project is clearly greater.

Further, I wonder if we are putting too many eggs in international expertise pools. Maybe doing things with our own local design and construction forces wasn't all wrong.

And lastly, labour rates have gone up. If you read the old newspaper articles about the labour unrest during the University subway construction....it's apparent that that project utilised a lot of cheap exploited immigrant labour (to the point that when the tunnelling landed at Queens' Park, then Premier Frost had to appoint a commission to buy peace for the workers who were literally under his feet).

All just me spitballing.

- Paul

 

[Northern Light]

This is something I've been thinking about for a while, both in UT posts and outside of it. Additionally, I've seen lots of allusions to this in many of our transport threads, so I thought I'd start a dedicated place to debate this.

In the last few years, we've seen the cost of transit infrastructure in Toronto explode.

Surface LRTs that cost in the region of $100 million/km (Finch West) a few years ago are now double that or more (Hamilton, Hurontario).^ TYSSE cost $450 million/km and we all derided that project as overbuilt. Now, SSE and YNSE are running towards $700 million/km. Eglinton West costs double what ECLRT cost, ($500 million/km vs. ~$280 million/km), but while the project's wholly tunneled, there's no underpinning or ROW difficulties to deal with. John Tory's always idiotic Smarttrack is at $1.4 billion for five stations. European cities can build an entire tramway network for that. My understanding is that GO RER is also overpriced for the work involved, but there aren't many comparisons and there's no final price yet.

Perhaps worst of all, the Ontario Line's capital cost is $19 billion, which works out to over $1 billion/km. This is a line that skimps on alignment: there's significant at-grade track, and tunnelling is only used where it has to be (under Queen and Pape).

This is in comparison to downtown European projects, such as Berlin's U5 from Alexanderplatz, which cost maybe $400 million/km (CAD); Paris's RER E to La Défense, which was a similar amount; Milan's Line 4, for $200 million/km; Barcelona's L9/10, a little over $220 million/km; even Amsterdam's Noord-Zuid Subway, built in the Dutch soil and taking 14 years, cost maybe $700 million/km.

View attachment 579834View attachment 579840

View attachment 579839View attachment 579843

View attachment 579859View attachment 579857
(aaaand you get the point)

What gives?

I'm trying to read on it, but I can only scratch the surface of this issue, and its complexity means that I can't read everything and my knowledge base isn't that deep.

My best understanding of the situation is that there's poor oversight of designing projects. Our 1980s-2010s transit building gap means that expertise that IIRC built the YUS/BD subways for $100 million/km or less has left or retired, leaving inexperienced people in charge. Design is poorly done, and there's political meddling, so costs are higher, which makes for more politics. Meanwhile, we tunnel under existing surface ROWS (EWLRT) and build at-grade in the downtown (OL). We contract in large packages, so smaller firms can't compete; hide everything, making transparent pricing impossible (Metrolinx is very, very guilty); we delay because of political uncertainty (see: Hamilton LRT); and at the end of the day we end up with worse results.

(I'm not informed, humour me here)

Do the good people of UT have any insights? Is there some big thing I missed? And what is being done/can we do to ensure that our future projects will remain be reasonably priced?

^all prices come with inflation built in, as according to the BOC's inflation calculator

You've got some good thoughts as does @crs1026

But lets give you some light bedtime reading on the subject, LOL.

https://rccao.com/research/files/RCCAO-STATION-TO-STATION-REPORT-APRIL2020.pdf

I will put it down to a lot of what Stephen has to say above, but I will add greater emphasis on a couple of things....

P3s. This isn't an ideological question for me, its simply the note that

a) If you contract out financing to a non-government entity (you don't have to do with a P3 but its common here), the private entity on raw cost will borrow at slightly more than double what the government could. That's a number in the billions on a large project. That's before realizing, they will not do this for you as a favour, they will charge a fee on top of their cost of financing.

b) If you nominally shift all the risk onto a consortium, they will, duh, charge you a risk premium to cover themselves, if they don't, watch out for the problems of penny-pinched sourcing (see Bombardier losing money on every Flexity they sold the TTC, and the substandard work on same performed in Mexico.)

*****

Also a problem is that these things tend to go to randomly formed consortia (that is to say, the players haven't necessarily worked together before and don't have teams that know and understand each other's needs, see Ottawa)

*****

Finally, I'd add..........what else is the project paying for......... I've seen the estimates coming in for the new GO Stations and for the Rail Path (both being built by Mx for the City).....there is 100% stuff being billed to the City that should not in my opinion be part of the contract cost.

ie. adding a crash wall along the Rail Path is debatable in necessity, but certainly is a choice for Mx, not a design mandate for a bike trail. In K-W, there is active freight rail still running through downtown, on-street/lanes shared with pedestrians.

 

[DirectionNorth]

@crs1026 @Northern Light

This has been a rabbithole of reading while I was abroad the past few weeks.

Thanks for the responses!

 

[DirectionNorth]

I saw this study from UofT today, and I think it's pretty relevant.

PDF: UNDERSTANDING THE DRIVERS OF TRANSIT CONSTRUCTION COSTS IN CANADA

Some key quotes, IMO.

Risk and NIMBYs:

Canadian transit agencies tend to overbuild in multiple ways, with larger and deeper tunnels and stations – a result of greater risk aversion.

(something something fire standards, which I understand is controversial)

The second aspect of risk aversion that is in play is letting external stakeholders drive design. To build political support and avoid the risk of litigation, many Canadian transit agencies allow external stakeholders (e.g., community groups, municipal governments, and business interests) to extract concessions from transit projects.

Methods:

Alignment selection: Low- and medium-cost jurisdictions minimize the use of tunnelling (e.g., by selecting elevated, at-grade, or trenched alignments) and reuse existing rights-of-way within project alignments wherever possible. Where tunnelling is strictly needed, low-cost methods (like cut and cover) are prioritized over more expensive methods (tunnel boring or mining).

Station depth and size: stations make up the largest proportion of hard costs ... Low- and medium cost jurisdictions reduce unnecessary back-of-house space and build underground stations as shallowly as possible. Where deep stations are absolutely necessary, the volume of excavation is minimized by eliminating full-length station mezzanines, or maintaining access exclusively through elevators.

Standardization: Low-cost jurisdictions standardize within a given project, and also implement national guidelines and best practices for design and construction. This lowers not only hard costs (through economies of scale), but also soft costs, since the burden of design and engineering work can be reduced.

The Consultant Enrichment Scheme™:

Like many North American transit agencies, Metrolinx relies extensively on external consultants for project delivery – not just for professional services, but also for management-level positions. The Ontario Auditor General found that “25% to 30% of all staff positions in the Capital Projects Group…[including] 25% of management positions… are filled by external consultants."

The abnormally high proportion of professional services does not seem to stem from a lack of inhouse personnel. Rather, the heavy reliance on external consultants impedes the retention of knowledge and expertise within the organization, leading to a scenario where there is minimal learning and an excessive managerial focus among public servants.

Lower-cost cities like Paris, Milan, and Istanbul, on the other hand, generally enter bidding only when the public sector has developed 30% to 70% of thedesign.³⁶ Allowing designs to reach more detailed stages before procurement also solidifies the scope of the entire project, a factor critical to reducing construction risk and costs for private sector bidders.³⁷Even when contract models with greater private-sector involvement are used, a competent public sector leads project management, issuing smaller contracts with clear deliverables and limited scope.

Risk:

However, excessively large contingencies can lead to “budget laxism,” where the incentive to control costs over a project life cycle is low and money, having been allocated, is unnecessarily spent.

The Metrolinx cost percentages in Figure 7 illustrate a habit of loading budgets with several layers of contingencies. The Metrolinx cost percentages in Figure 7 illustrate a habit of loading budgets with several layers of contingencies. Not only is project risk accounted for (“contingency,” 19.3%), there is also allowance for the risk of inflation (“escalation,” 13.7%). All told, risk provision accounts for 33% of the budget –more than the total soft-cost proportion of the Italian cases (9%).

the way Canadian agencies perceive and manage risk is at odds with the concept of contingency itself. Theoretically, contingencies should never be entirely spent, and should shrink as project risks are managed through the design and engineering process. However, the continuously escalating per-kilometre costs in Canada indicate a persistent trend of overpadding, where budgets are regularly inflated due to 1) prescriptive risk management practices and 2) the absence of resistance to increasing costs.

That special Metrolinx brand of transparency:

In addition, low-cost jurisdictions emphasize public transparency, enabling more cost-effective planning and decision-making. They avoid lump-sum contracts that make costs opaque and change orders more difficult to track.

In comparison, high-cost jurisdictions like New York and Toronto consider cost estimates to be akin to trade secrets, citing their commercial sensitivity to prevent public disclosure. In fact, Ontario explicitly “preserve the confidentiality of… construction cost estimate,” including “risks, costs, … schedule,… unit prices, and bid prices” as part of its “information sensitivity” guidelines, according to the provincial Ministry of Transportation.

However, low-cost jurisdictions have shown that the opposite is necessary ...

Grab-bag:

Factors like the need for additional third-party agreements, political micromanagement, high labour costs, and lack of competition in the construction sector have been shown to affect project delivery. In Ontario specifically, the volume of housing and infrastructure projects has pushed the province’s construction capacity near its limit, resulting in higher materials and labour costs. Political micromanagement has also been well documented throughout Toronto’s history, from the cancellation of the Eglinton West subway to the advent of Transit City. Although these influences are somewhat outside of the control of agencies, they are critical for establishing the context in which to examine a project’s delivery.

And some graphs. Cost history:

Metrolinx vs. Italy:

Contingencies:

It's, in my opinion, well worth a read. Someone tell Oshawa that a suburban subway shouldn't cost $1.3 billion/km.

 

[Northern Light]

I saw this study from UofT today, and I think it's pretty relevant.

PDF: UNDERSTANDING THE DRIVERS OF TRANSIT CONSTRUCTION COSTS IN CANADA

Some key quotes, IMO.

Risk and NIMBYs:


Methods:


The Consultant Enrichment Scheme™:


Risk:


That special Metrolinx brand of transparency:



Grab-bag:


And some graphs. Cost history:
View attachment 617191

Metrolinx vs. Italy:
View attachment 617188

Contingencies:
View attachment 617189

It's, in my opinion, well worth a read. Someone tell Oshawa that a suburban subway shouldn't cost $1.3 billion/km.

Excellent post. I have bookmarked the study for some light reading later.

Let me blow a couple of those slides up so that people can more easily read them:

 

[innsertnamehere]

No budget for escalation (i.e. inflation) in Italy? is part of it possibly to do with different accounting practices?

I mean look at Ontario - ECWE and YNSE are phased purposefully to be completed later with spending drawn out further, so their escalation costs are substantially higher.

Escalation is inflation, not real dollar impacts.

 

[nfitz]

Fascinating paper!

No budget for escalation (i.e. inflation) in Italy? is part of it possibly to do with different accounting practices?

Likely. TTC and early Metrolinx never included inflation in the budget. Simply that the numbers were in something like $2010. It waas always understood they'd be higher based on inflation.

I'm curious at what inflation rate was used in the calculations on that graph from 1950 to the current. Often people mistakenly use the Consumer Price Index - which reflects all sorts of bizarre stuff, like the move of production from Canada to Asia (etc.). This suppresses the inflation rate.

It would be interesting to see what that graph would look like if one used the local Construction Price Index instead. The Construction Price Index has exceed the Consumer Price Index for decades - massively at time. Personally, when I'm doing costing I tend to use 3% for the consumer price index, and 5% for the construction price index. While 3% might be a bit conservative (rate too high), 5% is (if anything) too low for many periods! But people start freaking if you take the last 10 years, and project it on the next 10 years!

Perhaps a question for @Stephen Wickens.

 

[Northern Light]

No budget for escalation (i.e. inflation) in Italy? is part of it possibly to do with different accounting practices?

I mean look at Ontario - ECWE and YNSE are phased purposefully to be completed later with spending drawn out further, so their escalation costs are substantially higher.

Escalation is inflation, not real dollar impacts.

I can't speak to Italy specifically, but there are other ways to budget for inflation.

Frankly, a lot of work is fixed price tender and its the responsibility of the contractor to account for where their wages are likely to be in 2, 3, or 5 years.

You can forward budget at the level of builder/client for inflation, its just an integrated number.

Materials can be procured early at fixed prices or with upset increase limits tied to commodity indexes or CPI etc.

A big thing as well though is smaller contracts structured differently.

We (Ontario) have a tendency to scope an entire line all at once, which by the nature of the scale will take longer to complete than a phased continuous build and open project.

Madrid notably tends to pursue continuous build......add one station at a time to the network.....(but don't stop adding), and you have projects with a nominal 3-year time horizon, in which fixed price with no inflation would be common.

*****

Aside from large contracts with huge time horizons, Ontario suffers from putting stuff out at 30% design to the private sector, with a whole slew of costs not fixed, because they aren't known.

If you take a project to 60% or better still 90% design before tender, you have a much clearer picture on costs and risks and the contingency can be reduced accordingly.

 

[DSC]

Some additional info/feedback at https://www.thestar.com/news/gta/po...cle_0274c036-b31b-11ef-9f14-4796fb315e37.html

Poor budgeting and too many consultants behind soaring cost of Toronto subway projects, study reveals​

 

[nfitz]

I can't speak to Italy specifically, but there are other ways to budget for inflation.

Frankly, a lot of work is fixed price tender and its the responsibility of the contractor to account for where their wages are likely to be in 2, 3, or 5 years.

I can be done in many ways (including that one). The payments could be tied to a particular index, on a particular day. Which is how many tax rates and minimum wage is done these days.

I had a contract once, where the index we were tied to, jumped an astronomical 18% one year, mostly due the price increases in diesel, steel, and concrete. But our input was mostly labour (the index selection was at the choice of our client). We could have taken the 18%, but instead we proactively contacted the client and said - this rate is just insane. And offered something that was more appropriate, agreeing to use something else in the future!