What ERA Announced

An essay from the same shelf of documents as The TIER Files. Sources at the bottom.

In The Float, I left an open file. ERA’s published lifetime commitment, $1.17 billion toward 352 projects as of June 2026, lined up almost exactly with the money it has ever paid plus the money it still owes, which meant the headline was a net number: everything announced and later cancelled had already been washed out of it. The gross total the program has announced over seventeen years had to be larger. I said I was reconstructing it from the press archive. That work is far enough along to publish what the public record now shows.

It shows three things at once. First, ERA’s own project database keeps the dead projects on the shelf, tagged, and countable. Second, the public project count is the database with those dead projects removed. Third, the lifetime dollar figure in the About boilerplate on the press releases sometimes goes down between consecutive announcements, which is what a self-cleaning stock looks like when you watch it in real time.

The database keeps score

ERA’s website runs on WordPress, and the project profiles are a proper content type with a status taxonomy. I pulled every record. There are 469 of them. The statuses, in ERA’s own labels, break down like this:

The shelf, including the funerals.

Add Never initiated, Cancelled, and Terminated and you get 116 dead projects: 24.7 per cent of the shelf. One in four.

Now do the subtraction the headline already did for you. 469 total minus 116 dead leaves 353. ERA’s June 2026 boilerplate says 352. Off by one, which is the kind of gap you get when a status flips between a scrape and a press release, or when a micro-program sits on a different ledger line. It is close enough that the public count is confirmed as the database net of the dead. Part 4 inferred that arithmetic from sixteen years of financial statements. The project database is the same arithmetic with receipts.

Listed funding across all 469 rows sums to about $1.185 billion. Strip the dead and you get about $1.147 billion on the living rows, against the published $1.17 billion. A $23 million difference on a seventeen-year stock is rounding, timing, and the odd commitment that has not yet been given a project page. Same stock, two faces.

The headline is the database after the funerals.

The taxonomy is the tell

I want to be fair about how unusual this is, in a helpful direction. ERA puts “Never initiated,” “Cancelled,” “Terminated,” and “On hold” on the public site as filterable statuses. “On hold” exists as a term even when the live count is zero; the FY2025 statements had parked six projects there after year-end, worth $25.7 million, and those appear to have moved on before I scraped. A granting agency that labels its wreckage in plain language is doing something the communications instinct usually forbids. The candour is real.

It is also the mechanism. Once dead is a first-class status, the lifetime project count can stay tidy by definition: count everything except the dead statuses, publish that number, and the failures never enter the sentence that starts with “Since 2009.” The taxonomy is public evidence of self-cleaning. Screenshot it before someone tidies the tidy.

Dead dollars, two measures

The database understates how much announced money died, because most of the dead rows list $0 in the funding field. All sixty Never initiated and all thirty-one Cancelled show blank amounts. Only the twenty-five Terminated rows still carry dollars, about $38 million among them, led by names that once had press cycles of their own: Enlighten Innovations (Field Upgrading), Purlucid, Rokk, a TransCanada Energy solar-plus-storage file, Carbon Corp., a City of Calgary electric-bus charger deployment, and so on. Useful as a cemetery directory. Useless as a valuation of the funerals.

For that, the audited statements remain the better instrument. In fiscal years 2022 through 2025 alone, ERA cancelled or terminated contribution agreements worth about $179 million: $19.9M, then $22.6M, then $53.8M, then $82.7M. Every one of those dollars was, at some point, an announcement. Many came with the minister and the hard hats. When the agreements died, the running headline absorbed the loss and moved on.

$179 million of agreements died in four years. The project pages still blank most of that money to zero.

The number that shrinks

If you still prefer the press archive to the auditor’s notes, the archive cooperates. Nearly every ERA media release carries a stock paragraph in the About block: since establishment in 2009, ERA has committed $X toward N projects. I walked that paragraph through the live releases (and Wayback snapshots of the older ones). The stock rises when new awards land. It also falls.

Selected drops, from their own boilerplate:

A lifetime commitment figure that can lose nearly fifty million dollars and thirteen projects between late summer and mid-autumn is doing something specific. It is revising the past so the present sentence stays clean. By June 10, 2026, the same boilerplate had climbed back to $1.17 billion toward 352 projects, which is the number you will still find if you open a recent release today.

I also added up the dollar figures in funding-round headlines across the archive I could reach. That naïve sum lands around two billion dollars. Do not treat that as the gross. Program envelopes get announced before awards, the same challenge gets a second release, ceilings go unfilled, and expansions restate money already counted. The sum is evidence that large figures have been put in headlines many times. It is a terrible balance sheet.

Gross, with bands

So what has this program actually announced?

Start with what is hard. Take the current net, $1.17 billion, and add the $179 million of agreements the FY2022-FY2025 statements say were cancelled or terminated. That gives a floor of about $1.35 billion. Everything below that floor is arguing with audited notes and the published stock at the same time.

Above the floor, the band widens because the earlier years disclose washouts more thinly, the database blanks the funding field on most dead rows, and the press drops only show the net revision rather than a per-project ledger of what left. A working central estimate, for readers who want a range rather than a shrug, is roughly $1.4 to $1.6 billion of lifetime announcements: the floor, plus incomplete pre-2022 washouts and the zero-dollar dead shelf. Treat anything sharper than that as false precision. The soft ceiling is poorly constrained; the headline sum of press-round dollars is a curiosity.

The gap above $1.17 billion is the product the announcements sold and the headline forgot.

Part 4 already had the architecture: paid out about $650 million over the program’s life, outstanding commitments about $375 million, headline about $1.17 billion, cancellations accelerating to the point where fiscal 2025 un-funded more than it funded. This essay adds the project-level census and the press timestamps. The float still holds the cash. The taxonomy holds the dead. The About paragraph holds the eraser.

The part where I’m fair about it

Milestone payment remains prudent. Innovation portfolios kill projects; a funder whose shelf never showed Cancelled or Terminated would be a funder taking no technological risk worth the name. Publishing those statuses in plain language is real disclosure practice. None of the figures above required a leak. They required reading the site and the statements the corporation already publishes.

The concessions leave the communications asymmetry and the definitional choice standing. Awards arrive with ministers, podiums, and estimated tonnes. Deaths arrive as status tags and note disclosures. The lifetime sentence the public is invited to remember is built to forget. Seventeen years is long enough to publish a gross announced total beside the net, or a simple table that matches each press release to paid, outstanding, or dead. The architecture that exists instead is a headline that can only look healthy, because its failures are deducted before you see it.

If you were inside one of those 116 files when the photograph was taken and outside it when the status flipped, the inbox is still open: [email protected]. Paperwork beats a scrape.

If I have misread a status, a boilerplate figure, or a cancellation total, the documents are linked below; show me and I will correct it.

Sources

1. ERA WordPress project database, full scrape of 469 project records and project_status taxonomy (Active, Cancelled, Completed, Contribution Agreement, Never initiated, On hold, Terminated), July 2026. Working extract held with the investigation notes. Live browse: eralberta.ca.
2. Status funding sums and dead-share arithmetic (116 / 469 = 24.7%; 469 - 116 = 353 vs published 352; alive listed funding ≈ $1.147B vs published $1.17B): same scrape.
3. “Since its establishment in 2009, ERA has committed $1.17 billion toward 352 projects”: ERA media release boilerplate, e.g. June 10, 2026, ERA Receives Additional $19.4 Million to Accelerate Methane Reduction Across Alberta.
4. Lifetime-claim trajectory and documented boilerplate drops: ERA /media-releases/ archive and funding-challenge pages, with Wayback snapshots for older URLs; working extracts held with the investigation notes.
5. ERA Financial Statements FY2022-FY2025 (KPMG): contribution agreements cancelled or terminated of approximately $19.9M, $22.6M, $53.75M, and $82.7M (four-year sum ~$179M); FY2025 subsequent events including On Hold. eralberta.ca.
6. Cumulative project expenses and outstanding commitments framing: The Float, drawing on CCEMC/ERA statements of operations FY2010-FY2025 and FY2025 Note 10 commitments of $374,993,802.
7. Tip line for cancelled or terminated projects: [email protected].