VIA Rail

[TerryJohnson]

With income for the proponent dependent on operating revenue, not having the proponent operate misaligns incentives needed to ensure success.

With O in there with revenue risk exposure, we have the opportunity for the proponent to invest in additional frequencies, additional classes, potentially even shorter travel time.

What would the optimal implementation model be in your mind?

Transport Action's position, based on member and supporter polling, is that DBFM infrastructure is reasonable but that passenger rail operations belong in the public sector. There's slightly less aversion to private sector operation of a completely new Calgary-Edmonton project than to outsourcing VIA's existing corridor. The level of concern around the future of VIA Rail and services outside the corridor if corridor operations are outsourced is also very high.

--

Personally, I believe infrastructure procurements should be priced based on tangible deliverables.

Let's say we put out an RFP a for work package which amounts to : "New railway between Havelock and Glen Tay, built and maintained to Transport Canada standards, able to support train frequency of A with average speed B and with availability C, designed, financed, built and maintained for 30 years."

Apart from the small problem of TC absurdly not having published a Canadian equivalents of FRA class 6 or 7 in its 2021 track standards update, this makes for a straightforward and tightly specified job that it should be pretty hard to screw up.
The private sector bidder has:

• The incentive to design and build for cost effective long term maintenance.
• The ability to innovate and find the lowest cost yet durable construction methods
• Some freedom, subject to the Impact Assessment process, to choose tweaks to the alignment to deliver the design speed.

In the DBFM scenario, where the public sector continues to operate the trains and collect the fares, the project risks transferred to the private sector are all within the ability of a project management team to wrangle, and payment is based on a simple set of measurable metrics, so the risk premium and financing costs can stay low.

The Assistant Deputy Minister for HFR cities the new Champlain Bridge project as a P3 success story, although there are complexities: it was DBFOM to start with, and then the government decided to cancel tolling it, making it more like a DBFM and removing a lot of revenue risk from the consortium.

What will happen with a HFR DBFOM, according to the ADM, is that they do not expect the P3 consortium to include a general contractor or a rolling stock manufacturer, so it would subcontract such work packages anyway. However, the financing is bundled and the risk aggregated, including operations and revenue which brings in a whole menagerie of externalities over the which government still has more control over than the consortium, so the principle of "risk transfer" breaks down.

One of the risk factors is the regulatory framework for operations over Metrolinx, Exo, CN and CP infrastructure. The JPO report says nothing useful about this crucial issue. It needs a complete overhaul, or the new operator will be just as thoroughly hexed as VIA Rail is now. My thinking would be that the model of buying a 50% interest in a corridor, as used by Virginia and CSX, could be applied to tracks to and from the new HFR alignments. A revised form of TSA, with performance guarantees and phased capacity enhancement plans built in, might work elsewhere. CN and CP will insist upon managing the projects on their infrastructure, so third-party cost monitoring would have to be arranged in order to prevent a reprise of the Kingston Sub upgrade snafu. There's nothing a HFR consortium can do about track access that couldn't be done in the public sector.

Long term ridership and travel demand in the corridor is going to be more dependent on public policy than the operational details of HFR, assuming the stated travel times and frequencies are delivered. The only certainty is the public policy on issues like housing is going to have to change. How will the government handle road pricing or congestion charging once the gas tax is gone? What will immigration policy be in 2040? Will the integration with public transit to make complete journeys seamless happen? The consortium can not arrive at anything better than a wild-ass guess, even with good data the best of intentions. There's no "secret sauce" that a private proponent can bring to the table to do this better.

That's why franchising in the UK was already failing and being dismantled, even with shorter concession periods to predict for, and before the pandemic consigned everyone's ridership projections to the dumpster. At least in the UK, with multiple franchisees, the "operator of last resort" could take over underbid franchises and restructure them. Here, with a single concessionaire, they'd be "too big to fail" and that's very bad news for taxpayers because we keep all the downside risk while losing most of the upside.

What started as a fairly straightforward $4B plan (In my opinion. I have colleagues in Transport Action who thought it too large even then.) is now, two election cycles later, a ~$12B megaproject and not likely to happen before another election kills it... yet again. That's ridiculous and depressing. I owe a big thank you to the people who've convinced me over the last few weeks that Canada's passenger rail network is still worth fighting for.

Terry

 

[ManyQuestions]

The government can adjust the upfront cash cost of $NPVQ, but whether the money flows at the start, or over the term, the liability (and asset) is reflected on the government's books similarly as the project is meant to cover its operational costs with revenue.

The REM project is a good case study for this. It's being financed largely by the CDPQ on the condition that the ARTM pay over time. The city governments will likely be be charged about $0.70 per passenger kilometer traveled. I'm not an expert but this seems to imply a higher operating subsidy then most metro systems.

https://www.cdpqinfra.com/sites/cdpqinfrad8/files/2019-10/note_financiere_ang_vf_0.pdf

 

[Urban Sky]

'm not an expert but this seems to imply a higher operating subsidy then most metro systems.

It's absolutely obscene, especially when considering that the public pays for more than half of the construction costs directly and the operating subsidy payments are guaranteed over a period of - effectively - 198 (!) years...

 

[rbt]

I'm not an expert but this seems to imply a higher operating subsidy then most metro systems.

It is about 3x higher than the Montreal metro operating subsidy BUT it also includes the capital portion of REM. Where the Metro required taking on debt via bonds to fund construction, CDPQ is also financing that portion.

For the first 30 year period, they're basically equal. Starting around 2050 though, after the capital component is paid down, REM will turn an astonishing profit if the agreement remains unchanged. Highway 407 revenue will look like kids selling lemonade by comparison.

At the same time, CDPQ is a department of the Quebec government, so it doesn't really matter. If the transit bill becomes too high they'll be ordered (via legislation if necessary) to renegotiate the contract provided pension payments are not at risk.

 

[crs1026]

What started as a fairly straightforward $4B plan (In my opinion. I have colleagues in Transport Action who thought it too large even then.) is now, two election cycles later, a ~$12B megaproject and not likely to happen before another election kills it... yet again. That's ridiculous and depressing. I owe a big thank you to the people who've convinced me over the last few weeks that Canada's passenger rail network is still worth fighting for.

My strength is not in business economics or finance, but it sure seems that a project which was largely a matter for civil engineering and project management has become a food ticket for investment advisors, finance specialists, and consultants. The seemingly straightforward need to convince a lender that there will be ROI (the CIB being mostly a proxy for a lender, and not a real one) has led to a much more thoughtful, but perhaps overthought, tangled governance.

The Quebec-Montreal and Toronto-Peterboro sections are very conventional rail rebuilding and upgrading packages. I have utter confidence that these could be competitively bid among the existing contractors in Canada, who already do lots of this kind of work. The only higher order design skill might be for the signalling system. The risks (supply chain logistics, material costs, labour availability and cost, weather and other execution barriers, physical unforeseens) are pretty well understood and managed by the construction industry. The processes for dealing with risk and non-performance are pretty easy to build into tenders and contracts.

The approaches to and routing through Montreal are a more complicated puzzle, but largely tied to CN/CP and their control over what happens on their row's . Metrolinx in Toronto pretty much mirrors CN/CP in how it treats VIA these days. As you note, that's something the JPO simply hasn't spoken to.

The Havelock-Smiths Falls section is a bigger package, perhaps, but again railway building even from unprepared land is pretty well understood.

Putting such an elaborate fiscal envelope around the financing and risk packages seems to be just polish on a cannonball. Over 30-50 years, I'm not sure it guarantees much. And while the course corrections for a failed public venture are not pleasant (Ministers get lambasted in the press, Auditors General generate adverse findings, CEO's and DM's get fired, organizations get restructured, taxpayers write nasty tweets), these tools are preferable (faster, more directly applied, more line of sight to the root causes, more comfort to the taxpayer) versus what is needed to redirect a P3 - where litigation may be the only permissible path forward. #Crosslinx, #TTCFlexity

I am not sure that the revenue/earnings levers in a DBFOM are that much more effective in derisking or securing project success, and they come at the expense of a very complicated contractual arrangement that may limit transparency. I'm just not a fan.

- Paul

 

[Darwinkgo]

Transport Action's position, based on member and supporter polling, is that DBFM infrastructure is reasonable but that passenger rail operations belong in the public sector. There's slightly less aversion to private sector operation of a completely new Calgary-Edmonton project than to outsourcing VIA's existing corridor. The level of concern around the future of VIA Rail and services outside the corridor if corridor operations are outsourced is also very high.

How do you account for design choices which reduce costs for the builder but raise costs for the operator, or operational choices which lower costs for the operator (or increase revenue for the operator, like additional frequencies) but raise costs for the maintainer?

Having operations as part of the package aligns the incentives and means the government doesn't have to attempt to optimize ahead of time, or set nearly as extensive provision standards.

 

[Darwinkgo]

The REM project is a good case study for this. It's being financed largely by the CDPQ on the condition that the ARTM pay over time. The city governments will likely be be charged about $0.70 per passenger kilometer traveled. I'm not an expert but this seems to imply a higher operating subsidy then most metro systems.

https://www.cdpqinfra.com/sites/cdpqinfrad8/files/2019-10/note_financiere_ang_vf_0.pdf

One of the reasons for P3s is to make the lifecycle availability cost transparent to the government investor. We just don't commonly calculate a comparable number for most projects, so can't really say whether it is high or not. Having the number then lets the government decide whether the project is worthwhile and aviod white elephants.

The payment for movement volume is typically called a 'shadow toll'. It is a way to incentivize design which maximizes use. Can couple it to a congestion charge to prevent under building and ensure capacity upgrades without additional agreeements, such as clawing back double the shadow toll for users passed by full trains, or reducing the shadow toll if average speed on a highway is lower than 85 kph.

REM is also weird in that it has tax increment financing for station areas as a supplemental revenue source—a revenue source subject to a whole lot of government risk. It does provide a huge incentive to serve areas that can become TODs, but depending on the nuts and bolts of the agreements I doubt it reduced the shadow toll by much.

ALSO: the REM project is indefinite term, with no transfer, no lease. The government doesn't get to carry the asset on its book, nor does it carry the liability: it is only buying a service. Since the pension fund under accounting rules is rolled up in the government net asset calculation, the asset and liability are still on the books, just only very distantly.

An interesting thought experiment is what if instead of creating VIA, the government had just offered CN and CP a passenger/mile subsidy with a speed proviso. Would we be in a better place today? I don't know. Just interesting to think about.

 

[Bordercollie]

It is about 3x higher than the Montreal metro operating subsidy BUT it also includes the capital portion of REM. Where the Metro required taking on debt to fund construction, CDPQ is also financing that portion.

For the first 30 year period, they're basically equal. Starting around 2050 though, after the capital component is paid down, REM will turn an astonishing profit if the agreement remains unchanged. Highway 407 revenue will look like kids selling lemonade by comparison.

At the same time, CDPQ is a department of the Quebec government, so it doesn't really matter. If the transit bill becomes too high they'll be ordered (via legislation if necessary) to renegotiate the contract provided pension payments are not at risk.

I don't think there is a single railway in the world build without subsidy that turns a profit on their entire network.
Even Japan has this problem with trains in rural areas not having enough funding being replaced by busses.

 

[rbt]

I don't think there is a single railway in the world build without subsidy that turns a profit on their entire network.
Even Japan has this problem with trains in rural areas not having enough funding being replaced by busses.

Few have a 72 cent per passenger per km cost to operate.

Lets say your a student and commute 5 days a week via metro to Edouard-Montpetit, then transfer to the REM to McGill and do the reverse on the way home. The government will pay $124.80 to REM for that short portion of the trip out of the $54 from your student metropass. I've yet to see a document describing what STM will receive for that trip or what kind of revenue split might exist. STM top-ups have typically been a %age of farebox revenue up until this point, will it go straight to a cost model (they get paid to run service and no longer worry at all about revenue?).

 

[Bordercollie]

Extremely few have a 72 cents per passenger per km cost to operate.

Lets say your a student and commute via metro to Edouard-montpetit, then hop the REM to mcgill 5 days a week and do the reverse on the way home. The government will pay $124.80 to REM for that short portion of the trip out of the $54 from your student metropass. I've yet to see a document describing how STM revenue will be impacted by REM fees.

That's not bad actually. I think the Northlander was costing $400 per rider in subsidy.

 

[rbt]

That's not bad actually. I think the Northlander was costing $400 per rider in subsidy.

Yes, I suppose I should have said high-capacity or mass-transit system. Various light-duty systems for those with disabilities also approach $1/km/passenger.

But even then, Northlander was a 1000km trip, so about 40 cents per km per passenger for those going from end to end; nor was a typical per-trip fare under $3.

 

[Darwinkgo]

Few have a 72 cent per passenger per km cost to operate.

Lets say your a student and commute 5 days a week via metro to Edouard-Montpetit, then transfer to the REM to McGill and do the reverse on the way home. The government will pay $124.80 to REM for that short portion of the trip out of the $54 from your student metropass. I've yet to see a document describing what STM will receive for that trip or what kind of revenue split might exist. STM top-ups have typically been a %age of farebox revenue up until this point, will it go straight to a cost model (they get paid to run service and no longer worry at all about revenue?).

You're equating availability and operation cost with operation cost. Who do you think pays for the capital cost for a student traveling to York University? It (and the rolling stock) was debt financed just the same. The levelized cost of availability is just hidden.

As for the second part - 'not worrying about revenue'. Revenue comes from use. Use is what they are worried about because they are worried about revenue. What does it matter how the money gets from one account to another? It makes sense to not have them directly take revenue from the farebox, due to concessionary passes like you describe.

Anyways, this is far off topic, but it has really shown what barriers HFT will have in public acceptance over the next 10 years.

 

[rbt]

You're equating availability and operation cost with operation cost. Who do you think pays for the capital cost for a student traveling to York University? It (and the rolling stock) was debt financed just the same. The levelized cost of availability is just hidden.

In the contract REM costs are forever (inflation adjusted), surviving many many decades after the capital debt is paid down. Capital debt in most systems largely disappears after 30 years, at which point government can afford to do another expansion cycle again, even with high interest rates.

But I addressed that difference already using the 2050 example year. Also addressed how the government can force CDPQ to change the terms of the contract; they don't need to get stuck like Ontario with Highway 407 terms when they realize they're still paying over 50% of their total transit budget (capital + operations) on REM in 2060.

REM as a capital project is awesome. REM as a legal/fiscal structure for implementing transit is one of the worst possible; not a model for other governments (like the federal government) to follow.

 

[TerryJohnson]

How do you account for design choices which reduce costs for the builder but raise costs for the operator, or operational choices which lower costs for the operator (or increase revenue for the operator, like additional frequencies) but raise costs for the maintainer?

Having operations as part of the package aligns the incentives and means the government doesn't have to attempt to optimize ahead of time, or set nearly as extensive provision standards.

This is a fair question, but there would still be a project team at VIA Rail that would draft the 100+ pages of detail that go into an RFP and/or negotiations with shortlisted bidders.

There's nothing the infrastructure provider could do, if they deliver a train path availability based spec, that would significantly vary operating cost because the majority of the operating costs are equipment hours and crew hours.

If I was writing it, the spec would accommodate the service plan already developed, but with future-proofing - for example new structures to be built to double track clearances - or even better costed-out options for additional frequencies.

However, in my model, what happens over at Toronto Union Station - a big negotiation between agencies controlled by two levels of government over platform availability, train paths, dwell times, etc, etc - doesn't bleed into pricing the Havelock to Glen Tay infrastructure. The risk of the government not being able to gets its act together stays with the government, giving the government he incentive to get its act together.

In the menagerie of project risks for HFR as a whole , most of the elephants are "government fails to get its act together about X" which would really worry me if I was trying to write a DBFOM bid because Transport Canada's record is not great. They've failed on track access regulation for decades. Yet the redacted "top risks" section of the JPO report is only half a page, which is less than we see in VIA's corporate reports.

 

[Bordercollie]

This is a fair question, but there would still be a project team at VIA Rail that would draft the 100+ pages of detail that go into an RFP and/or negotiations with shortlisted bidders.

There's nothing the infrastructure provider could do, if they deliver a train path availability based spec, that would significantly vary operating cost because the majority of the operating costs are equipment hours and crew hours.

If I was writing it, the spec would accommodate the service plan already developed, but with future-proofing - for example new structures to be built to double track clearances - or even better costed-out options for additional frequencies.

However, in my model, what happens over at Toronto Union Station - a big negotiation between agencies controlled by two levels of government over platform availability, train paths, dwell times, etc, etc - doesn't bleed into pricing the Havelock to Glen Tay infrastructure. The risk of the government not being able to gets its act together stays with the government, giving the government he incentive to get its act together.

In the menagerie of project risks for HFR as a whole , most of the elephants are "government fails to get its act together about X" which would really worry me if I was trying to write a DBFOM bid because Transport Canada's record is not great. They've failed on track access regulation for decades. Yet the redacted "top risks" section of the JPO report is only half a page, which is less than we see in VIA's corporate reports.

At this point it would be such a faliure if they don't deliver this project.