Contractor patterns
Bidding patterns worth a closer look
An anomaly is a statistical pattern worth a closer look. It is not a finding of wrongdoing. Ordinary explanations exist, such as related companies, shared offices, or how an office schedules its tenders.
Pattern checks are being verified before publication. They look for win rotation, shared business addresses and shared principals among contractors, using PhilGEPS award notices and PCAB license records. Scope analyzed so far: 3,966,103 awards, 27,841 contractors and 16,785 procuring offices (data as of December 26, 2025) · Source: PhilGEPS.
Method and thresholds
- Win rotation: Wald–Wolfowitz runs test on the date-ordered wins of two contractors at one procuring office (same-day wins dropped). Each needs ≥10 wins and ≥24 award days combined. Significant only if p < 0.01 ÷ 1,056,727 pairs tested (Bonferroni, p < 9.5e-9).
- Shared address: identical specific street address on public map business profiles, 2–4 firms, ≥10 awards combined at a common procuring office.
- Shared principal: identical full name of the principal on PCAB online license records, 2–4 firms, same common-office rule.
- Co-bidding: not computed. PhilGEPS open data lists only the winning bidder, not every bidder.
- Every signal that passes is re-checked against the raw records by an automated verifier that rejects weak or ambiguous cases. Signals below threshold are never shown. A dispute hides a signal until it is re-checked, and a sample of every automated decision is audited weekly.
4 passed the statistical thresholds; 4 were rejected by the verifier; 0 are hidden pending a dispute. Data as of December 26, 2025 (PhilGEPS award notices); computed October 26, 2026. Counts on this page cover verified signals.
Signals
Individual signals are not published yet. They will appear here once they pass verification.