The Montreal land share — the metro anchor, and a third self-correction
2026-07-18 · wave 21 · ESTIMATE lane, deepening mode · Canada's second-largest city, parsed in full from Quebec's provincial open-roll host.
What was measured
Montreal's complete 2026 assessment roll (MAMH open XML, 795MB, 437,192 assessment units, $503B of assessed value) downloaded and parsed this wave:
| Segment | Units | Land value | Building value | Land share |
|---|---|---|---|---|
| All units | 437,192 | $189.9B | $313.6B | 37.7% |
| Residential (usage 1xxx) | 408,399 | $117.5B | $236.8B | 33.2% |
| Non-residential | 20,419 | $63.1B | $76.7B | 45.2% |
| Vacant land | 8,374 | $9.2B | $0.1B | 98.8% (parse self-check) |
The self-correction (third)
Wave 19 generalized the 60–80% band as "major metros only." Montreal refutes that: Canada's second-largest city sits at 38%. The measured spectrum is now:
| City | Land share (city-wide / residential) |
|---|---|
| Vancouver | 75% / — (85% single-family zones) |
| Victoria | 69% / — |
| Montreal | 38% / 33% |
| Sherbrooke | 30% / 27% |
| (economy-wide join, w3) | 53% |
The variable that predicts land share is not metro size — it is supply constraint. Vancouver (mountains, water, restrictive single-family zoning) capitalizes scarcity into ground; Montreal (flat geography, plex-dense housing tradition, historically permissive multiplex zoning, rent regulation) keeps ground cheap relative to what stands on it. This is precisely the mechanism the housing-supply literature documents, and our own data now shows it at parcel scale. The band language in NUMBERS.md row 17 is corrected to "supply-constrained metros," and the correction is logged on the site, per the standing rule.
What it means for Groundshare
- Montreal is net-recipient territory almost wholesale. With residential land at 33% of value, even substantial Montreal homes carry modest land value; the typical Montreal homeowner household sits far below the $456k crossover. Canada's second metro — 4M+ people in the CMA — lands overwhelmingly on the receiving side of the schedule, homeowners included. The paying class remains where waves 8–9 found it: supply-constrained west-coast (and central-Toronto) ground, plus entities.
- The Vancouver-exceptionalism story sharpens. The national politics are not "metros vs the rest"; they are "scarcity-priced ground vs everywhere else." That is a better story for the campaign and a truer one.
- A cross-regime caveat, promoted to a research question. Quebec and BC assessors allocate land vs building under different methodologies; part of the BC-Quebec gap could be methodological. This is now a concrete, testable instance of academic-path Q5 (site-vs-agglomeration and allocation methodology) — one more reason the Q1/Q5 measurement studies matter before any legislated schedule.
- Quebec's data regime keeps paying. Two cities parsed from one host with identical schema; the remaining Quebec metros (Laval 241MB, Gatineau 132MB, Quebec City) are routine follow-ups.
Sources
- MAMH / Données Québec, Rôles d'évaluation foncière du Québec — Montreal RL66023_2026.xml (794,991,963 bytes; 437,192 units), downloaded and parsed in full 2026-07-18. https://www.donneesquebec.ca/recherche/dataset/roles-d-evaluation-fonciere-du-quebec
- Spectrum comparisons:
analysis/vancouver-land-share.md(w9),analysis/victoria-land-share.md(w19),analysis/sherbrooke-land-share.md(w20),analysis/household-incidence.md(w3).