% Author: Simon-Pierre Boucher — simon-pierre.boucher@uqo.ca
\section{Institutional setting: valuation and transaction records in Quebec}
\label{sec:setting}

Quebec offers an unusually clean environment for studying valuation because three public records
describe the same dwelling from three different angles, each produced by a distinct institution with
its own timing. The first is the land register (\emph{Registre foncier du Qu\'ebec}), in which every
transfer of ownership is published by a notary. A registered deed records the parties, the
consideration paid, the cadastral lot and the date of the act; it does not record the dwelling's
attributes, the marketing history, or whether the transfer was at arm's length. The second is the
municipal assessment roll, prepared under the \emph{Loi sur la fiscalit\'e municipale} by municipal
or regional assessors and supervised by the Minist\`ere des Affaires municipales et de l'Habitation
(MAMH). The roll is triennial: each unit of assessment carries a land value and a building value that
are meant to reflect market value at a reference date set eighteen months before the roll comes into
force, together with the physical attributes used by the assessor (building type, construction year,
floor area, lot area, number of storeys and dwellings, and the land-use code CUBF). The third is the
brokerage record (the MLS/Centris system operated under the authority of the OACIQ), which contains
listing prices, descriptions, photographs and marketing dates, but only for the roughly three quarters
of transactions that go through a licensed broker; it is proprietary and is not used in the arms of
this study.

Valuation itself is a regulated profession: chartered appraisers (\'evaluateurs agr\'e\'es) are
members of the Ordre des \'evaluateurs agr\'e\'es du Qu\'ebec and produce appraisal reports for
lenders, courts and expropriations following the sales-comparison, cost and income approaches. The
agents studied here are simulated analytic roles modelled on this division of labour; none of them is
a chartered appraiser and nothing in the paper should be read as an appraisal within the meaning of
Quebec law.

Two institutional facts shape the identification strategy. First, the assessment roll is public,
stale by construction (its reference date precedes the sale by eighteen months to four and a half
years), and known to every market participant: it is a natural, information-set-consistent anchor
that any valuation method, human or artificial, legitimately observes. We therefore give it to every
arm, together with its reference date, and we treat the price-to-assessment ratio of recent nearby
sales as the canonical ``local market state'' signal. Second, registration in the land register lags
the signature of the deed by a few days, and publication in the data platform we use lags registration
by two to eight weeks \citep[measured in a companion paper,][]{boucher2026sharp}; a valuation dated at
the deed therefore cannot use sales registered after it, and the comparable pool of every arm is
truncated at the valuation date on the \emph{sale} date, not the publication date, which is the
conservative choice.

The sample period, January 2021 to October 2026, spans a complete cycle: the pandemic boom of
2021--2022, the correction that followed the rise of policy rates in 2022--2023, and the recovery of
2024--2026. This matters for the question at hand because a valuation method that merely
extrapolates recent comparables performs differently at turning points than one that reasons about
the state of the market, and because distribution shift between the training window and the test
window is a feature of the design rather than a nuisance.