01
PRIMARY LAW
30% is a disclosure trigger, not a compensation cap.
Québec s.431 requires disclosure when distributor remuneration exceeds 30% of the product sale price. The historical note traces the provision to the 1998 enactment.
RED TEAM · AUDITED v2.0 · SEPTEMBER 16, 2026
What survives
red team.
A comparative evidence record for Québec, New Brunswick and British Columbia. The standard is simple: separate law from proposals, published data from causal inference, and unanswered questions from invented answers.
01 / SURVIVES
The findings that made it through.
02
REGULATOR RATIONALE
New Brunswick does publish a rationale for the 30% threshold.
FCNB describes compensation above 30% as a significant inducement and says disclosure supports informed decisions and consumer protection. The unresolved question is what evidence calibrated 30% as the breakpoint.
03
PUBLISHED DATA
Québec remuneration stayed materially above 30% in the reported products.
AMF reports dealer remuneration of 53-54% of premium for VSPED and 45-46% for F.P.Q. No. 5 in 2020-2022. AMF also states that the underlying insurer submissions were not independently verified.
04
UNKNOWN
New Brunswick does not yet answer the central dealer-economics question in the reviewed public record.
Public reporting shows licensing and compliance activity, but not a dealer-level series for F&I PVR, affected-product penetration, payroll/FTE, pretax profit or consumer price pass-through.
05
PRIMARY LAW
B.C. is not a clean January 1, 2027 point treatment.
The regulation takes effect January 1, 2027, but transition provisions can extend dealer-specific exposure beyond that date. Evaluation should use actual application, licensing and training dates.
06
02 / CORRECTIONS
The audit changed the argument.
Sharper, not louder.
FCNB does give a public rationale: compensation above 30% is described as a significant inducement. The unanswered question is the calibration of the 30% breakpoint.
Published ≠ independently verified.
The Québec figures remain useful, but they are regulator-published insurer-reported data. The source-quality caveat now travels with the numbers.
Legal commencement ≠ one-day behavioural switch.
New Brunswick had an application transition. B.C. has transition mechanics. Event time must follow actual exposure, not only the date printed in the regulation.
Natural comparators must still earn the job.
Nova Scotia and PEI are candidates, not automatic controls. Pre-treatment fit and product/legal comparability determine whether they belong in the counterfactual.
03 / PUBLICATION GATE
Three bins. No rhetorical blending.
Legal chronology; the distinction between a disclosure trigger and a compensation cap; exact AMF descriptives with the source-quality caveat; New Brunswick rule text and aggregate regulator activity; B.C. enacted regulation and transition mechanics.
Threshold provenance and calibration; treatment timing; licence-count interpretation; B.C. rule, fee and training status; Québec cross-channel comparisons; control selection.
New Brunswick causal F&I effects; B.C. job, profit or tax point forecasts; quantified consumer savings; causal attribution of Québec refusal-rate changes to the 30% disclosure rule.
04 / STUDY DESIGN
Build the evaluation so it can prove us wrong.
01
Treatment timing
Separate proposal and anticipation from legal commencement, transition and later amendments. Use actual B.C. application, licence and training dates where available.
02
Treatment intensity
Use product-level exposure: regulated-product share, >30% disclosure share, seller/licence/training status and product version - not only a province × post dummy.
03
Counterfactual
Treat Nova Scotia and PEI as candidate natural comparators, then test a broader donor pool using pre-treatment fit and legal/product comparability.
04
Inference
Use province-level placebo/permutation inference, synthetic or control-weighted counterfactuals and comparative interrupted time series. More dealer rows improve measurement; they do not create more independent policy clusters.
05
Falsification
Pre-trends, placebo dates, placebo provinces, negative-control products, negative-control outcomes, pandemic-window sensitivity, ownership/OEM stability and missing-not-at-random participation bounds.
06
Publication rule
Freeze primary outcomes before the B.C. post-period and publish nulls, mixed effects and trade-offs without collapsing them into a single policy verdict.
05 / ACQUIRE BEFORE 2027
The highest-value evidence is not another opinion.
| Priority | Recipient / dataset | Highest-value request |
|---|---|---|
| P0 | B.C. dealer baseline | Monthly 2024 onward dealer/product economics plus actual application, licensing, training and transition dates. |
| P0 | B.C. insurers / administrators | De-identified product/policy premium, compensation, cancellation, refund, claims and product-version data. |
| P0 | Insurance Council of B.C. | Final approved rules, fees, application opening, training/accreditation, licence counts and implementation metrics. |
| P0 | New Brunswick dealer / insurer panel | Monthly 2018-2026 matched outcomes with actual implementation dates where held. |
| P1 | FCNB | RIR roster by business type and status; complaints; compliance findings; enforcement; held outcome data; post-implementation evaluation records. |
| P1 | B.C. Ministry of Finance | Jurisdictional scans, problem definition, threshold analysis, alternatives, economic analysis and success criteria. |
THE CENTRAL RESULT
Unknown is a finding when the data are not there.
Public data reviewed do not establish New Brunswick’s causal effect on dealership F&I economics, employment, consumer prices or total social surplus. No B.C. point estimate for jobs, dealer profit, tax revenue or consumer savings should be presented as a factual forecast before the microdata and counterfactual exist.
That boundary is not weakness. It is the line that keeps a research page from becoming an advocacy spreadsheet.