Table 4 names the seven shocks and states what each one assumes. This appendix is the working behind them: what actually happened, where the number comes from, and what bounds it. Nothing here changes any figure in Table 5. A reader who accepts Table 5 does not need this appendix; a reader who wants to argue with it should start here.
The sections below follow the order of the evidence rather than the order of the table, because several of the shocks share a single source. The cross-references in Table 4 point to the right one.
The sections below are grouped in Table 4 by what causes them, because the three groups behave differently and it matters which one a given shock belongs to. Group 1, the terms of separation, is settled by negotiation and fixed on the day it is agreed. Group 2 is the market’s reaction to a country without a track record, and it fades as the record is built. Group 3 is the commodity cycle, which arrives whether or not Alberta is independent and which has reversed every time it has occurred. Groups 1 and 2 together — C$13.5 billion a year — are what independence itself costs on the least favourable assumptions in this paper.
K.2 The Oil Price Slump — energy prices (Group 3, the commodity cycle)
The worst case for energy prices is Alberta Budget 2026’s own downside case, because it maps directly onto Alberta revenue. For context on how far a long-run path could fall, the Canada Energy Regulator’s Lower scenario holds Brent at US$55 a barrel in constant 2025 dollars, flat from 2030 through 2050 [CER-26].
The International Energy Agency’s Net Zero pathway is lower still at US$25 by 2050, but that is a normative backcast rather than a forecast, and propagated through the light-heavy differential it implies a Western Canadian Select price below the operating cost of most of the oil sands. This case would be so disastrous globally as to be seen as meaningless, which is why it has been discontinued. [IEA-25].
The United States Energy Information Administration discontinued its Low Oil Price case in 2026; its last published low case put Brent at US$45 to US$48 [EIA-25].
Alberta publishes its own downside, and that is the figure carried into the stress test because it maps directly onto Alberta revenue. Budget 2026 models West Texas Intermediate at US$51.50 rather than US$60.50 and puts the cost at C$6.5 billion of combined tax and resource revenue [AB-26]. The lower Canada Energy Regulator and Energy Information Administration paths above are shown as context for how far a long-run price path could fall; they are not substituted into the fiscal model.
Royalties are price-sensitive by construction — they fall with prices rather than standing as a fixed obligation, so the revenue line and the ability to pay move together.
Contributions to the Heritage Fund are discretionary and are the first thing a government would reduce as opposed to cutting services to citizens. It is important to note that resource royalties are C$13.2 billion against combined revenue of C$129 billion; they are about a tenth of the case, not its foundation.
For scale, the worst resource-revenue year in Alberta’s recorded history was 2015-16, at C$2.8 billion — C$3.5 billion in constant 2024-25 dollars [GoA-HFS]. That was one year at the bottom of a cycle, not a decade, and it has never been sustained.
K.4 The Head Office Exodus — head offices and the corporate tax base (Group 2, the reaction to a new country)
The right comparator is Quebec, because it is the only advanced economy where a credible separatist government held power and corporate flight was actually measured. The best-documented count is twenty-eight major relocations between 1979 and 1982, costing roughly 6,500 head-office jobs [CPQ-82].
Over the longer run, Montreal lost twenty-four per cent of its manufacturing head offices between 1976 and 1997 against four per cent in Toronto [SC-VIN]. The 1995 referendum, which failed by about one per cent of the vote, produced no documented second exodus at all.
Alberta has 423 head offices employing 40,417 people — 1.6 per cent of Alberta employment [SC-110]. A loss on Quebec’s scale, twenty-four per cent over two decades, would be about 9,700 positions, or four tenths of one per cent of the workforce.
Worst case adopted in this document: the corporate tax base contracts fifteen per cent, −C$3.2 billion a year. That is half again the ten per cent a moderate scenario would imply, and it is deliberately set above what the Quebec record would support.
What bounds it, and this is the part worth stating plainly. Alberta and Canada will be independent tax jurisdictions. Alberta will tax the incomes earned in Alberta and Canada will tax incomes earned in Canada. Oil sands plants, pipelines, refineries, transmission and farmland cannot be moved, so the income they generate stays substantially taxable where the production happens: in Alberta.
If any head offices move, only people and operations who move will no longer be taxed in Alberta
Much of the workforce is far less mobile than the letterhead. Fifty-four per cent of Alberta’s oil and gas employment is field, trades and production work; outside Calgary that rises to seventy-one per cent and the corporate share falls to thirteen per cent [SC-455]. Engineering and operations staff have to be near the plant. What is genuinely mobile is the smaller, higher-paid executive and finance function.
Catalonia is the clean demonstration: some five thousand companies shifted their registered office within months of the 2017 vote and almost none moved a plant or a payroll.
Additionally, an executive contemplating the move has to answer where to. Presuming the destination is Canada, a jurisdiction with higher personal and corporate rates and heavier regulation, the executive would be asking Albertans to relocate to a higher tax, more regulated environment and the uptake for such moves is plausibly expected to be poor, suggesting material impact to corporation’s capabilities and effectiveness. Quebec is again instructive on the long run: it retains 22.9 per cent of Canada’s head-office employment, and between 2012 and 2021 its head-office employment fell 2.9 per cent against 5.0 per cent nationally and 6.3 per cent in Toronto [SC-110]. The companies that wanted to leave left, and the sector rebuilt.
K.1 The Household Squeeze — the personal income tax base (Group 3, the commodity cycle)
Alberta does not need a hypothetical here; it has lived the worst case twice. Primary household income fell 14.7 per cent in real terms between 2015 and 2020, including 11.9 per cent in 2016 alone, and in 2024 was still 1.9 per cent below its 2015 real peak [SC-224]. That is the deepest contraction in the Alberta personal tax base in forty years.
Worst case adopted: the personal tax base contracts fifteen per cent, −C$7.5 billion a year. That is three times the five per cent a moderate scenario would imply, and it is the largest single adverse item in this analysis.
What bounds it. Alberta’s population has not fallen year over year since 1971, through every episode in that record. The worst sustained out-migration Alberta has recorded ran from 1982 to 1988 and totalled about 123,000 people, some five per cent of the population — and the population still grew throughout [SC-015].
As this paper lays out, if anything, an independent Alberta is a fiscally stronger, lower taxed jurisdiction and coupled with its cultural bias against taxes and regulation, it will more than likely grow. People move in no small part for economic reasons - the amount of money they can take home, keep in their pockets and use for their own purposes. None of that is banked in this paper. The model assumes no growth dividend and no migration dividend at all, and tests the opposite: a sustained contraction in the personal tax base.
If part of that fiscal room were used for tax relief or for productivity-enhancing investment, it could raise after-tax incomes and strengthen Alberta’s pull as a destination rather than a departure point. No such dividend is assumed in the base case, which is why this element is treated as a cyclical contraction rather than an exodus.
As can be seen in Table 12, the tax rates can plausibly go down in a relatively short period of time.
K.5 The Wider Downturn — consumption taxes, fees and the wider economy (Group 3, the commodity cycle)
The worst peak-to-trough contraction in Alberta real gross domestic product in the available record is 8.6 per cent, between 2014 and 2020, taking eight years to regain the previous peak [SC-222]. Consumption taxes and fees are assumed to fall with it.
Worst case adopted: consumption taxes and fees down eight per cent, −C$1.9 billion a year.
K.3 The Full Debt Share and The Larger Settlement (Group 1, the terms of separation)
Three items are carried at the top of their stated range rather than the middle. The federal debt settlement is taken at the full population share, with no credit for Alberta’s decades of net contribution and no offset for federal assets in Alberta.
An Indigenous settlement above continuation funding is taken at C$2.8 billion.
Defence initial capitalisation was carried in earlier drafts as a stress of C$1.5 billion a year, on the premise that the whole equipment programme might be charged as an operating cost in year one. It has been withdrawn for two independent reasons. The first is physical: the equipment is not sitting on a shelf. Major platforms take years to build and longer to crew, so a single-year charge describes an event that cannot occur, and an extraordinary event is not a scenario. The second is arithmetic: the build-out is C$10.0 to C$16.5 billion spread over ten to fifteen years, an average of C$0.7 to C$1.6 billion a year, and the mature C$5.0 billion defence budget already in the ledger carries C$0.9 to C$1.1 billion a year for acquisition and replacement. Adding C$1.5 billion on top counted most of the same money twice. The build-out is costed in Appendix E and carried in the ledger, where it belongs, rather than in the stress test.
K.6 The New Country Premium — sovereign borrowing costs (Group 2, the reaction to a new country)
A new sovereign has no credit history of its own, and lenders price that. The allowance carried in this paper is C$0.5 to C$1.3 billion a year, and the stress table uses the top of that range.
What bounds it. The premium is charged on the stock of debt that is actually rolled over, not on the whole stock at once, so it arrives gradually rather than on Day 1. Alberta’s taxpayer-supported debt is about 23 per cent of output, which is low by any international standard and is itself the main thing a lender looks at. And the premium is not permanent by nature: it is the price of an absent record, and it falls as the record accumulates. Every sovereign now regarded as safe began without one.
This is also the one shock in the table that is partly within Alberta’s own control. A credible fiscal framework, published accounts and a funded reserve are the standard means by which new issuers shorten the period over which they pay a premium.
