Procured, and Then?
ALTO commissioned the outside view. Whether it changed anything is the one question the record does not answer.
The standard fix for over-optimistic infrastructure forecasts is to check them against what comparable projects actually cost and carried, rather than trusting the project’s own bottom-up numbers. That check is called reference-class forecasting, and ALTO commissioned one. It hired the firm founded by the researcher who developed the method.
That is to ALTO’s credit. But commissioning a check and acting on it are different things, and only one document would show which happened: a comparison putting ALTO’s own published figures beside the ones the check produced. The Initiative asked for that record. The response was extended to 18 September 2026, with notice that a third party would be consulted — a step the Act provides for where an institution intends to release records that may contain a supplier’s commercial information.
Meanwhile, in June 2026, ALTO published two studies putting large dollar values on the project’s benefits. Neither sets those benefits against what the line would cost. This brief looks at all three documents and asks what they show about how the project’s numbers are being assembled — and what a single unredacted release would settle.
ALTO commissioned the outside view
Chapter 1 of the HPR Research Report sets out the method this brief relies on, so it is only summarised here. Large infrastructure forecasts miss in a consistent direction: costs come in high, benefits come in low. The established corrective is to stop treating a project as unique and instead compare it against the recorded outcomes of projects like it. The technique has a name — reference-class forecasting — and a literature behind it.
In 2024 ALTO issued an advance contract award notice, PAS240625-002-00, for reference-class forecasting, should-cost and should-schedule modelling, and a series of Challenge Boards. An advance contract award notice is the instrument used when a department intends to award without competition, on the basis that only one supplier can do the work. The named supplier was Oxford Global Projects, the consultancy founded by Bent Flyvbjerg and Alexander Budzier.
This is worth stating plainly, because it cuts against the easy criticism. ALTO did not ignore the outside view. It went out and procured it, from the people who developed it.
Buying the instrument is not the same as letting it bind
Reference-class forecasting corrects a forecast only if the number it produces is permitted to move the decision. A should-cost that is commissioned, delivered and then filed next to an unchanged inside-view estimate has not corrected anything. The method’s own literature is explicit that the failure mode is not the absence of the outside view but its subordination — the number produced, and then declined.
So the decisive record is not the existence of the forecast. It is the comparison: does ALTO’s published capital cost reflect its own reference-class should-cost, or diverge from it? One document would answer that — the inside view and the outside view set side by side.
This yields something better than a claim about anyone’s conduct: a prediction that can be checked. If the commissioned reference-class figures are more conservative than the numbers ALTO has published, the outside view was procured but not applied. If they match, the Initiative’s cost critique weakens accordingly.
We do not know which. Nothing in this brief asserts that ALTO set the analysis aside. The point is that the question is answerable, that a single document answers it, and that the document exists.
The record will arrive after the decision has moved on
The Initiative requested the reference-class records under access to information — the workbook, the should-cost and should-schedule outputs, and above all any document setting the inside view beside the outside view. Request A-2026-0004 was met in June 2026 with a ninety-day extension carrying the response to 18 September 2026, together with a notice invoking third-party consultation under section 27.
Section 27 consultation is a routine step, and it is worth being precise about which way it points. The section applies where the head of an institution intends to disclose a record that may contain a third party’s commercial information: the notice tells that third party of the intention to release and gives it twenty days to make representations against disclosure, and invoking the section is what permits the response time to be extended. The notice on A-2026-0004 therefore records that Alto has turned its mind to releasing the reference-class records and has given Oxford Global Projects the opportunity to object. It is not a signal that the material will be withheld.
What remains is a question of timing rather than intent. The third party may object and the institution may then withhold some of the figures; equally it may not. What can be said is the sequence: the record capable of testing the decision will arrive after further commitment has been made. What it contains, the disclosure itself will settle.
An outside-view check disciplines a decision only while the decision is still open. Once enough money is committed, the arithmetic changes: the cost of stopping is subtracted from the cost of continuing, and a project can show better value for money the more has already been spent on it. Britain’s High Speed Two reached exactly that point — the National Audit Office found in June 2026 that the ratio for completing the programme had risen even as the programme grew more expensive, because the estimated cost of cancelling had more than quadrupled.
The cheapest moment to apply the test is before that crossover, not after it.
Two studies, no cost side
In June 2026, two months after the consultation closed, ALTO released two commissioned studies. A computable general equilibrium assessment by Aviseo Consulting reports a national real GDP gain of about $24.4 billion a year. A corridor tourism study by CPCS with HDR adds up to $3.9 billion in GDP and 43,000 jobs.
Neither nets a cost. The macroeconomic study excludes construction and operating expenditure by design; the tourism study has no cost side to exclude. Both are benefit totals unaccompanied by the outlay required to obtain them. Both, to their credit, describe their outputs as illustrative and order-of-magnitude rather than forecasts, and make the largest figures conditional on tourism policy the railway itself does not deliver.
The scenario range has a floor and no ceiling on the downside
Each study is built as a fan of scenarios, from pessimistic to optimistic. In both, the entire fan sits above zero. The macro study reports welfare increasing in every scenario; the tourism study’s weakest case is still $177 million and two thousand jobs. The modelled question is how large the gain is, never whether there is a loss.
Adverse mechanisms are identified but do not reach the total
The tourism study acknowledges that faster trains shorten stays and convert overnight visits into day trips, and shows length of stay falling in several cities. The aggregate rises regardless.
The two studies disagree, and each resolves the disagreement upward
The macro study omits domestic tourism on the ground that it is largely substitution from other household spending, with little net effect on national output. The tourism study builds most of its $33.7-billion base, and most of its headline uplift, from precisely that in-corridor domestic travel — counted through gross multipliers that assume no such displacement. The two treatments diverge, and in each case the treatment adopted is the one that yields the larger figure for that study.
The studies import the literature’s upside but not its realisation record
Both studies draw their benefit magnitudes from the international high-speed rail literature — the same comparison set the Initiative uses. What they import is the size of the upside. What they do not import is that literature’s record on realisation: rail benefits arriving at about two-thirds of forecast, and passenger numbers overstated by roughly a hundred per cent.
Each of the four observations above is a description of what the documents contain. Taken together they describe a benefit case in which every point of divergence has resolved in the same direction — which is the pattern the forecasting literature says to look for, and the reason an independent outside-view comparison matters more, not less, once numbers of this size are in circulation. The same two studies are examined in detail in the Initiative’s briefs Two Point Two Trillion and At Face Value.
Publish the comparison
The Initiative’s recommendation is narrow and does not require anyone to accept a word of its own analysis.
It requires one document to be made public. The framework behind the request is set out in full in Chapter 1 of the HPR Research Report; what ought to be built instead is the subject of the chapters that follow it.
Every figure attributed to Alto, Aviseo, CPCS, the National Audit Office or a published paper is quoted from the source listed below and can be checked there. Nothing else here is a calculation of ours: the argument rests on what the documents contain and on the sequence of dates, not on a competing estimate.
Where a record has not been released, this brief says so rather than inferring its contents, and makes no claim about why any extension was taken or any figure was or was not published. The prediction in section 02 is stated in both directions and will be settled by the disclosure, not by us.