The Buried
Footnote Forensic Linguistics Analysis
on Corporate Disclosure Documents
Dossier · 2025-W51 Canadian Banks
← Back to dashboard

Canadian Banks — 40-F cohort dated 2025-12-02 through 2025-12-17

Lede

Five Canadian banks filed 40-Fs within a sixteen-day window, and the linguistic signatures have separated to the edges of the peer baseline. CIBC is hedging across nearly every axis the framework measures. Royal Bank, on the same axes, reads as the most direct document in the cohort. The story of this filing window is not what any single bank said — it is the spread between them.

Findings

1. CM's MD&A is the most defensive document in the cohort

Five same-direction signals converge in one filing: [CM Management's Discussion & Analysis Disclaimer density = 0.575 (above the 95th percentile)] against a cohort mean of 0.385, [CM Management's Discussion & Analysis Passive voice ratio = 1.543 (above the 95th percentile)], [CM Management's Discussion & Analysis Forward-looking framing = 10.278 (above the 95th percentile)], and [CM Management's Discussion & Analysis Hedging language = 3.037 (above the 95th percentile)]. Conviction reads as the lowest in the cohort at [CM Management's Discussion & Analysis Certainty / hedge imbalance = 0.278 (bottom 10%)]. Four hedging axes and one conviction axis pointing the same way in a single document is not drafting noise.

2. CM's AIF carries the same posture forward, with one inversion

The statutory document reads even more forward-looking and more disclaimed than the management commentary: [CM Annual Information Form Disclaimer density = 1.218 (above the 95th percentile)] against a baseline p95 of 1.154, and [CM Annual Information Form Forward-looking framing = 11.683 (above the 95th percentile)] against a p95 of 10.826. Both have cleared the ceiling of the industry baseline. The asymmetry sits in [CM Annual Information Form Certainty / hedge imbalance = 0.507 (above the 95th percentile)] — the AIF reads as more certain than the MD&A, which is the reverse of the usual ordering between statutory disclosure and management commentary.

3. RY sits at the opposite pole

The same axes that flag CM at p95 flag RY below p10: [RY Annual Information Form Disclaimer density = 0.000 (bottom 10%)], [RY Annual Information Form Passive voice ratio = 0.559 (bottom 10%)], [RY Annual Information Form Forward-looking framing = 2.541 (bottom 10%)], [RY Management's Discussion & Analysis Passive voice ratio = 1.195 (bottom 10%)], and [RY Management's Discussion & Analysis Forward-looking framing = 7.702 (bottom 10%)]. Five same-direction signals in the opposite direction from CM. The cohort is bookended by two of the Big Five.

4. BNS shows a split signature between its two documents

The MD&A reads assertive: [BNS Management's Discussion & Analysis Certainty / hedge imbalance = 0.431 (above the 95th percentile)] and [BNS Management's Discussion & Analysis Hedging language = 1.911 (bottom 10%)]. The AIF reads hedged on the same date stamp: [BNS Annual Information Form Passive voice ratio = 1.019 (above the 95th percentile)], [BNS Annual Information Form Hedging language = 3.085 (above the 95th percentile)], and [BNS Annual Information Form Vague attributions = 0.045 (above the 95th percentile)]. The two documents are not telling the same story.

5. BMO is the only amended filing, and its AIF language is anomalous

BMO filed a [BMO filing_metadata] 40-F/A on 2025-12-17 while the other four filed straight 40-Fs. Its AIF carries [BMO Annual Information Form Cliché density = 0.037 (above the 95th percentile)] and [BMO Annual Information Form Boilerplate callbacks to prior periods = 0.012 (above the 95th percentile)] — the only bank in the cohort above baseline p95 on either measure — alongside [BMO Annual Information Form Hedging language = 1.948 (bottom 10%)]. An amended filing paired with boilerplate-heavy AIF language is worth surfacing on its own.

Forensic note

The cohort baseline for Canadian bank disclaimer density in the AIF runs from p25=0.643 to p95=1.154. [CM Annual Information Form Disclaimer density = 1.218] has cleared the ceiling. The same filing pushes [CM Annual Information Form Forward-looking framing = 11.683] past a baseline p95 of 10.826. When a single statutory document is statistically more forward-looking and more disclaimed than 95 percent of its peer baseline on independent axes, that is a posture, not an artifact of drafting. RY's mirror-image position on the same axes — [RY Annual Information Form Disclaimer density = 0.000] and [RY Annual Information Form Forward-looking framing = 2.541], both below p10 — confirms that the spread is real and not a baseline calibration issue.

What to watch

Three measurable questions resolve over the next filing cycle. Whether CM's MD&A reverts toward cohort means on disclaimer density and tentative markers, or whether the hedging persists into the next quarter's commentary. Whether RY's low-disclaimer, low-future-focus posture holds through the next 40-F, or whether 2025 was the outlier. Whether BMO files any further amendments, and whether the cliche and thermo-callback densities normalize. Whether the BNS MD&A-versus-AIF gap closes in either direction. No bullets were extracted from any of the five filings in this window — the next read should target the risk-factor and critical-accounting-estimate sections in CM and BMO directly.