At Face Value
Five numbers from ALTO’s economic report are now in wide circulation. Each one is either the top of a range or the middle of one. In every case the range exists. In every case it was not printed.
$49.5 billion in benefits · $24.5 billion in added GDP · 1.1 per cent of Canada’s economy · 50,000 jobs · 24 million riders a year
None of these figures is invented. Every one comes from real modelling work by real economists, and each is accurate on its own terms. But each describes something narrower, or more conditional, than it sounds — and in every case the report prints one number where the work behind it reports a range.
$49.5 billion is labelled an upper estimate — the good end of a range whose other end is nowhere in the document. $24.5 billion is a central estimate from a range its own modeller published as $14.8 billion to $41.0 billion. 24 million riders is prefixed “up to.” 50,000 jobs sits in a table headed “upper estimate” — and ALTO’s own appendix states those figures are not net economic gains.
Meanwhile the report declines to publish a benefit-cost ratio — the single number that would tell you whether the benefits exceed the costs.
The point of this brief is not that ALTO’s numbers are wrong. It is that they are being used as though they were settled, when the documents behind them show they are not.
“$49.5 billion in benefits”
| What you’ll hear | What it actually is |
|---|---|
| ALTO delivers $49.5 billion in benefits to Canadians. | The upper estimate of benefits, added up over 60 years, adjusted downward for the fact that most of them arrive decades from now. Nearly four-fifths of it is saved travel time. |
Economists assume a benefit arriving in 2085 is worth less to us today than the same benefit arriving next year — the same reason $100 now beats $100 in twenty years. So future benefits get shrunk before they are added up. This is standard, required, and correct.
It also does a great deal of work here, and the Initiative’s brief Hours Are Not Dollars sets it out in full. The report says ALTO saves 9.3 billion hours of travel time, and values an hour at $22.32. Multiply those together and you get $207.6 billion. The figure that appears in the benefit table is $38.4 billion. Discounting removes about 82 per cent of the face value, because most of those hours are saved by people who have not been born yet.
The benefit table is headed “upper estimate.” Every narrative figure is prefixed “up to” — up to 24 million riders, up to 9.3 billion hours, up to 400 lives saved. An upper estimate is one end of a range. The other end does not appear anywhere in the 83 pages.
There is one more thing about this number worth knowing. The report opens by making traffic congestion the problem — Highway 401, Pearson airport, journeys that are too slow and too unreliable. In the benefit table, easing congestion is worth $570 million out of $49.5 billion: about 1.2 per cent. Cleaner air is worth $27 million, roughly one part in two thousand.
The problem the report leads with and the benefit the report counts are almost entirely different things.
“$24.5 billion in added GDP”
| What you’ll hear | What it actually is |
|---|---|
| ALTO adds $24.5 billion to the Canadian economy. | The middle of a range running from $14.8 billion to $41.0 billion, produced by a different model, in a different year’s dollars, and measured as an annual figure rather than a 60-year total. |
| Can it be added to the $49.5 billion? No — the report says so on page 9 | |
The Initiative examined this figure in detail in Two Point Two Trillion. It is the number ALTO’s website states most confidently: the analysis “concludes that Alto will permanently uplift Canada’s GDP by 1.1%.” The study behind it is more careful, and it publishes its range.
It comes from a computable general equilibrium model — a simulation of the whole Canadian economy that works out what happens to wages, prices, trade and output when you change one thing. These are legitimate, widely used tools. Governments run them all the time.
The thing being changed here is a single assumption: that firms in Toronto, Montréal, Ottawa and Québec City become three per cent more productive because the train exists. That one assumption produces $21 billion of the $24.4 billion — 86 per cent of the total.
The modellers say so openly, and deserve credit for it. Studies in Germany and Sweden found productivity gains of two to four per cent. So, they write, “it seems reasonable to consider a baseline scenario” of three per cent.
They also tested two per cent and five per cent. Two per cent gives the $14.8 billion; five per cent gives the $41.0 billion. There is no scenario in which the productivity gain fails to appear at all — even the pessimistic case assumes a two per cent uplift across four city economies.
One further detail. Because nobody yet knows where the stations will go, the modellers applied that productivity boost to the entire metropolitan areas of Toronto, Montréal, Ottawa and Québec City — standing in for the 30-kilometre radius around a station that the research actually supports. Peterborough, Laval, Trois-Rivières and Kingston appear nowhere in that calculation.
“1.1 per cent of Canada’s GDP”
| What you’ll hear | What it actually is |
|---|---|
| ALTO raises Canada’s GDP by 1.1 per cent. | 1.1 per cent of the Canadian economy as it was in 2019 — the last pre-pandemic year, used because 2020 and 2021 were distorted. |
The modeller says this plainly: the gain is “roughly 1.1% of Canada’s 2019 GDP.” Canada’s Moment drops the year. The website drops the year and adds the word “will.”
Taken across the published range rather than the midpoint, the same calculation gives roughly 0.6 per cent at the low end and 1.8 per cent at the high end. One of those three numbers is in circulation.
“50,000 jobs and $86 billion”
| What you’ll hear | What ALTO’s own appendix says |
|---|---|
| Building ALTO creates 50,000 jobs and adds $86 billion to GDP. | These figures “represent spending-supported economic activity rather than net economic gains” and are “therefore not included in the benefit-cost ratio.” |
If you spend $60 billion on anything — a railway, a bridge, a very large hole — people get paid to do it, and those people spend their wages locally. Counting that as a benefit of the project would mean any spending is a benefit, which cannot be right: the money had to come from somewhere, and would have employed someone else.
ALTO’s appendix says this outright, and adds that the model used “does not account for potential constraints in the economy, such as labour shortages or capacity limits.” Most promoters present numbers like these as benefits and say nothing. ALTO explicitly refuses to — on page 80. The 50,000 jobs appear on page 5.
The caveat is genuine and creditable. It sits in an appendix seventy-five pages behind the figure it qualifies, and it has not travelled with the number. The Initiative examines this figure in full in Where Do 50,000 Jobs Come From?, which rebuilds it from the annual spending and workforce figures HS2 and the Réseau express métropolitain both publish, and finds that roughly 18,000 of the 50,000 are people working on the railway.
“24 million riders a year”
| What you’ll hear | What it actually is |
|---|---|
| ALTO will carry 24 million passengers a year. | “Up to” 24 million a year by 2055 — the output of scenarios and sensitivity tests whose range the report describes but does not show. |
ALTO’s own methodology appendix explains exactly what it should have published. Sensitivity testing, it says, gives decision-makers “a range of plausible outcomes rather than relying on a single forecast.” The flowchart’s stated output is “a range of plausible ridership outcomes.”
The report then prints one number.
That number matters more than it looks, because four of the seven benefit lines depend on it — car running costs, road safety, congestion and greenhouse gases all flow from an estimate of 90.1 billion kilometres of driving avoided. And how much driving is avoided depends on who the new passengers are. Someone who switches from a car takes kilometres off the road. Someone who switches from a plane takes none. Someone making a brand new trip takes none. That breakdown is published nowhere.
There is no benefit-cost ratio
The standard test of whether a public investment is worth making is simple: divide the benefits by the costs. Above one, it pays. Below one, it does not.
Canada’s Moment does not publish that number. Page 62 explains why: the cost estimate is too early-stage for the ratio to be meaningful. The same absence runs through ALTO’s public benefits page, examined in Many Benefits, One Missing Number.
ALTO names a Metrolinx appraisal manual as the authority for two of its key figures — the subject of It Left the Rules Behind. That manual treats early-stage uncertainty as the reason to test and publish ranges — not as a reason to withhold them. Its instruction for the earliest project stage is to conduct sensitivity testing to understand the level of uncertainty.
The same manual also requires early-stage rail costs to be marked up by 64 per cent before being compared with benefits, because rail megaprojects are systematically undercosted. That would put ALTO’s $60–90 billion into the comparison at roughly $98–148 billion. No such uplift is applied, and the concept is not mentioned.
So the report does not divide the benefits by the costs, while placing $49.5 billion in benefits and $60–90 billion in costs on the same spread.
It has been done before — on this corridor, at this stage
In December 2021 the Joint Project Office — a body formed by VIA Rail Canada and the Canada Infrastructure Bank — completed a business case for High Frequency Rail, the slower and cheaper predecessor to ALTO along the same corridor. It was at a comparable point in its development.
That document published a benefit-cost ratio.
It was not published at the time. It became public in November 2025, when the Canada Infrastructure Bank released it under the Access to Information Act — almost four years after it was written.
A benefit-cost ratio of 1.00 means a project returns exactly what it costs. Above 1.00 it pays for itself in economic terms; below 1.00 it does not.
A ratio of 0.13 means that for every dollar spent, about thirteen cents of measurable benefit came back. On the wider basis, which counts effects that are harder to measure, about forty cents. The Joint Project Office published those figures anyway, alongside the subsidy the project would need.
0.13 is not ALTO’s ratio, and it is not an estimate of ALTO’s ratio. High Frequency Rail was a different project: slower, at $27.71 billion rather than $60–90 billion, assessed over thirty years rather than sixty. The Joint Project Office described its own results as preliminary. None of its figures transfers to ALTO by arithmetic, and this brief does not offer them as a forecast of anything.
What the document establishes is narrower, and harder to answer: a benefit-cost ratio can be produced for a project on this corridor at this stage of design — because one was. Immaturity did not prevent it then, on a cost estimate roughly a third the size.
There is one further detail worth recording. The same document, released under a different access request, comes back with one section withheld in full, the capital cost and revenue sentences cut off mid-clause, and the subsection headings of its Economic Case not shown in the table of contents. No exemption provision is marked against any of these.
The benefit table cannot be audited by a reader
Each row of the main benefit table gives you a quantity and a dollar value. The natural thing to do is divide one by the other and see what price has been put on an hour, a tonne of carbon, or a life. You cannot.
The dollar figures are discounted. The quantities are not. And the prices that would connect them are referenced by source but never stated — only the $22.32 hourly value appears anywhere.
One row does not appear to add up at all. The report says ALTO avoids up to 400 deaths and 26,000 injuries, valued at $610 million. Working backwards, that is roughly $1.2 billion before the time adjustment — which 26,000 injuries alone would exhaust at $50,000 each, a low figure by Canadian standards, leaving nothing for the 400 lives. Either the casualty counts and the money cover different periods, or the values placed on a life and an injury are far below what Canadian governments normally use. The report does not publish enough to say which.
The same thing has happened to every number
This is the finding. Not that any figure is fabricated — none is. Not that the modelling is incompetent — the underlying studies are careful, and say so about their own limits. But wherever the underlying work reported a range, the document carries a single figure from the optimistic end of it, and it is that single figure which has reached councils, newspapers and the public.
The ranges are not secret. Most of them are in ALTO’s own commissioned studies, sitting on ALTO’s own research page. They simply did not make it into the document that everyone reads.
None of these figures should be repeated at face value, in either direction. They are the optimistic end of work that its own authors describe as uncertain — and anyone quoting them, for or against the project, should say which end of the range they are quoting.
What this brief is not saying
Credit where it is owed
And the limits of what we show
Five questions for anyone quoting these figures
1. Is that the top of the range, or the middle?
For four of the five headline numbers, it is one or the other.
2. What is the lower estimate?
The benefit table is explicitly labelled an upper bound. Every range has another end.
3. What is the benefit-cost ratio?
Not published. It is the number that puts every other number in proportion — and one was produced for this corridor in 2021, so “too early” is a choice rather than a constraint.
4. What price was put on a life, or a tonne of carbon?
Referenced by source, never stated. Without them, no line of the benefit table can be checked.
5. Where do the new passengers come from?
Cars, planes, existing trains, or trips that would never have happened? A fifth of the benefits depend on the answer.