The consequential decisions are not whether a sailing ran. They are whether a fare rise is justified, whether a vessel should be approved, whether a route is worth what it costs. Those turn on twenty years of demand, capacity, cost and subsidy, per route, which exists only as several hundred PDFs. This page is one route, assembled from two decades of them.
Route 12, Brentwood Bay to Mill Bay, is the interesting case precisely because it looks contradictory in the filings: among the most heavily utilised in the system, and among the smallest in absolute revenue. A page that set out to show waste would stop there. The regulator's question is different and duller, and it is the one asked here.
What are ferry users and taxpayers buying with this route, and how would you know if it stopped being worth it?
Clause 2(d) of the contract sets the capacity standard as sufficient to carry the previous year's traffic, which is backward looking, so a route running at record utilisation on record demand is the case in which that standard is doing least work.
Demand, capacity, utilisation and service come from the Operations Summary Report in the annual reports, and from the quarterly returns where those carry the same panel. Revenue, expenses and the transportation fee come from the Route Statements. Three years are not from their own year's report: fiscal 2011 and fiscal 2020 have no readable report of their own, so their figures are taken from the prior-year column the following report prints beside its own, and fiscal 2020's traffic from the March 2020 quarterly return. Those are marked in the source column of every row rather than smoothed over here. One year is blank for one reason in the documents rather than the reading: fiscal 2011's report was filed as a scanned image with no text layer. Dollar figures are as printed, in thousands and nominal; where the page compares years it says so and deflates.
Cost per unit is the figure everyone reaches for, and the unit chosen decides the conclusion. Both are shown because on this route they disagree, and a reader is better served seeing that than trusting whichever one I picked.
Cost per vehicle carried is a productivity measure. It improves whenever traffic grows against a fixed service, so it rewards a route for demand it did not choose and penalises one that exists to be available. Cost per round trip is a supply measure. It isolates what it costs to operate the service and is indifferent to how full the vessel was. A route can become cheaper per vehicle and more expensive per sailing at the same time, which is exactly what this one did, and the two support opposite arguments in a fare case.
Only routes with both an expense figure and a traffic figure in the comparison year appear, which is 24 of the 25 that have a Route Statement: route 13 files passengers but no vehicle equivalent, so it appears in the chart below and not in this table. Northern routes dominate the per-vehicle column for a structural reason worth stating rather than hiding: they carry very little against a long, staffed crossing, so any per-vehicle figure for them is large by construction and says little about how well they are run.
The same year, every route on two axes: how many passengers it carries, and what it earns or loses on each of them. Earnings from operations is revenue less operating expenses, before the cost of capital and before the transportation fee, so it is what a route does on its own account rather than what it costs the public purse.
Both axes are compressed and are labelled as such. Passenger counts run across three orders of magnitude and earnings per passenger from minus a thousand-odd dollars to plus fourteen, so on linear axes twenty-two of the twenty-five routes would sit on top of one another in a single corner. The horizontal axis is a log scale; the vertical one is signed-log, which keeps the zero line where it belongs and lets a loss of $1,488 and a profit of $14 appear on the same chart.
Nothing here establishes whether the route should exist, and the numbers cannot. Route 12, Brentwood Bay to Mill Bay, is a Malahat bypass; part of what is bought is the option of a crossing when the highway is closed, and an option's value does not appear in a cost-per-vehicle figure in the years it is not exercised. Neither does resilience, nor the cost of the alternative.
What the assembled record does do is make the operator's own case testable. A capital or fare application resting on rising demand can be checked against twenty years of demand; one resting on high utilisation can be checked against whether utilisation rose because traffic grew or because capacity was withdrawn; one resting on cost pressure can be checked against which denominator it used.
The value is not a verdict on the route. It is that the assumptions underneath an application stop being unfalsifiable.
Every figure carries the file and page it was read from. Columns are found by looking for the arithmetic each row asserts about itself, and a row whose own arithmetic does not close is refused rather than estimated. The Route Statements are gated on four identities running from revenue through expenses and cost of capital to earnings before fees; 7,084 figures appear in two separate annual reports and agree, and 653 differ because the operator restated them, mostly around the transition to IFRS.
If any of this is already on your desk in another form, the useful outcome of a short conversation is finding out which part, before more is built on the wrong assumption. Rich Vickers, Delta BC, rich@salish-ops.ca.
Salish Ops is an independent project, not affiliated with or endorsed by British Columbia Ferry Services Inc., the BC Ferry Authority or the Office of the BC Ferries Commissioner. Figures are read from published filings; errors are mine.