Running pattern detector against the corpus
The dataset shows entities receiving substantial federal funding with no recorded business identity — recipients the federal government cannot independently identify.
Glassbox surfaces correlations and observations, not findings of misconduct. Severity is a calibrated read on how strongly the dataset matches the pattern definition — a starting point for review, not a verdict.
The match meets the pattern threshold but the signal is the lower band. Normal review queue.
The match exceeds the routine threshold. Worth a closer look — pull the source rows, walk the citation.
The match is in the highest band of the calibrated scale. Recommend: prioritise for active review.
Severity is computed from the detection signal · recipient_business_number IS NULL or a placeholder ('0', all-zeros, '-', 'n/a', 'none') · is_amendment = false · total ≥ p90 of ALL recipient totals within the recipient's primary department AND ≥ $500K floor · publisher batch-report rows excluded · signal = total $
The dataset shows Council of Yukon First Nations (Whitehorse, YT) received $15M across 19 federal agreements from 6 departments — above the 90th percentile of recipients funded by Employment and Social Development Canada | Emploi et Développement social Canada — with no recipient business number recorded; pattern consistent with ghost capacity (entity not independently identifiable in the federal corpus).
Why this may be benign · Identity rests on the legal name alone (no business number); spelling variants can split one entity into several or merge homonyms (KNOWN-DATA-ISSUES F-6, C-3).
Once a match is identified, the dataset suggests these review actions. Glassbox does not prescribe outcomes — these are calibrated options for funders, auditors, and program officers, framed as correlations to investigate.
An entity receiving substantial federal funds without a recorded business number is identifiable only by name string. Verify identity by mapping to corporate / charity registries and adding the BN to the funder's master record.
TRACE's original definition of ghost capacity covers entities with no employees, no addresses, and revenue dominated by government transfers. For these matches, verify capacity (audited financials, employee count, physical presence) before further awards.
What did the entity actually produce for the funding? Pull deliverable reports, site visits, third-party verification. Ghost-capacity is a capacity question, not a fraud claim.
Many matches here are data-entry omissions (BN simply not recorded) rather than substantive ghost capacity. Use this list to clean up master records and tighten future intake.
These are calibrated suggestions, not directives. The dataset shows the pattern; the funder decides the response.
Aggregated from the matches below. Severity bands are calibrated per pattern; counts and dollar totals reflect this snapshot only.
The dataset shows Saint Mary's University (Halifax, Nova Scotia | Halifax, Nouvelle-Écosse, NS) received $15M across 37 federal agreements from 5 departments — above the 90th percentile of recipients funded by Employment and Social Development Canada | Emploi et Développement social Canada — with no recipient business number recorded; pattern consistent with ghost capacity (entity not independently identifiable in the federal corpus).
Why this may be benign · Identity rests on the legal name alone (no business number); spelling variants can split one entity into several or merge homonyms (KNOWN-DATA-ISSUES F-6, C-3).