B.C.’s $1.5 Billion Spreadsheet Error: What the Numbers Mean

B.C.’s roughly $1.5 billion spreadsheet error was an overestimate of future natural gas royalty revenue—not $1.5 billion in cash that disappeared. Officials put the error at about $1.46 billion across five years, averaging $292 million annually, according to Business in Vancouver’s September 1, 2026 reporting. Amounts here are in Canadian dollars.

If you are turning this story into a spreadsheet, chart or public data page, the first task is to separate the five-year correction from the annual forecast revisions. They describe different periods and, in some cases, different causes.

What went wrong?

Officials described four mistakes in the forecast inputs:

• An unnecessary currency conversion. A U.S.-dollar-to-Canadian-dollar conversion was applied to figures already expressed in Canadian dollars. Officials linked the largest error to a formula incorrectly dragged across spreadsheet cells.
• An incorrect gas-unit conversion. Gas volumes were converted incorrectly.
• Two stale-data errors. Staff used cost inputs from 2025 instead of 2026 in two instances.

These details come from CBC’s account of the ministry briefing. BIV reported that the errors affected half of the 22 third-party gas price forecasts used by government, with the annual royalty overestimate reaching $378 million in 2028–29. These are officials’ estimates reported by BIV, not a reconstruction of the workbook.

The cause was disputed. Earlier expert analysis pointed to transportation and processing costs; ministry officials said those costs were not responsible for the identified errors. Do not present that earlier explanation as the established cause. Victoria News documents the disagreement.

Why the later report shows a $531 million reduction

B.C.’s First Quarterly Report, released September 14, 2026, lowered the 2026/27 natural gas royalty forecast by $531 million compared with Budget 2026. The province’s backgrounder, updated September 16, attributes approximately $306 million to correcting calculation errors. The remainder reflects lower price expectations and updated assumptions for the proposed new royalty framework. Source: Ministry of Finance.

A useful public table separates those components:

Component of the 2026/27 revision	Change from Budget 2026 (CAD millions)
Calculation-error correction	-306
Other forecast changes	-225
Total natural gas royalty revision	-531

The First Quarterly Report, Chart 1.1 and Table 1.2 presents these figures. Negative values mean lower forecast revenue; the correction component is approximate.

Do not add $306 million to $531 million: it is already included. Likewise, do not compare the $531 million single-year revision directly with the roughly $1.46 billion five-year error total as if they measured the same thing.

Checks to apply before publishing your own forecast

The following are recommended controls, not claims about what B.C.’s workbook contained.

Make units and currencies explicit. Keep separate fields for original value, original currency, original unit, conversion factor and converted value. Test that an input already in CAD does not receive a USD-to-CAD conversion.

Review copied formulas at boundaries. Inspect the first and last cells of each filled range and any change between currencies, units or years. Excel’s relative references change when copied; absolute references remain fixed. Use the reference type that matches the intended calculation—not absolute references everywhere. Microsoft explains the distinction.

Record the input year. Store the source’s publication date and the year its values describe. Flag prior-year inputs for review rather than silently treating them as current.

Reconcile revisions independently. Have a reviewer calculate totals and selected conversions separately. Distinguish corrections from changes in market expectations or policy assumptions.

Publish a revision table, not just a headline number

For a reusable dataset, use one row per fiscal year and revision component. A practical header structure is:

fiscal_year,component,row_type,change_cad_millions,comparison_baseline,source_date,source_url,notes

Label approximate values, explain the negative-sign convention, and preserve source links and version dates. If you include totals, mark row_type as component or total so readers can exclude totals when summing the changes.

Publish only reviewed, non-sensitive data. On the public page, show the period, baseline and units beside the table, and retain a dated correction note when replacing figures. A sortable page helps readers inspect a dataset; it does not validate the calculations behind it. For the broader workflow, see how to prepare and publish a data table.