spb/wp9_uqo Public

1% Author: Simon-Pierre Boucher — contact@spboucher.ai2% ============================================================================3\section{Conclusion}4\label{sec:conclusion}5% ============================================================================67Using a nationwide cross-section of 140{,}931 MLS listings and 1{,}153 absorbed8neighbourhood fixed effects, we estimate a grand hedonic model of the Canadian9residential housing market. The exercise yields three robust conclusions. First,10location dominates: resolving geography from the provincial to the neighbourhood scale11raises explained price variation from 57\% to 77\%, and location as a whole accounts for12roughly thirty percentage points of $R^2$---more than every structural attribute13combined, and a listing-level counterpart to the large land shares documented in14aggregate data \citep{davis2007price,knoll2017no}. Second, the structural implicit15prices behave as hedonic theory predicts and are strikingly stable across samples:16living area carries an elasticity near 0.55, each full bathroom adds about 11--15\%, and17bedroom counts are economically negligible once floor space is held18fixed---magnitudes squarely within the range catalogued by decades of hedonic work19\citep{sirmans2005composition}. Third, the neighbourhood premia are vast---a factor of20roughly nine separates the most and least expensive FSAs---and the model translates into21a credible valuation tool, predicting held-out prices with an out-of-sample $R^2$ of220.76 and a median error of 16\%. These conclusions survive an extensive battery of23checks: the implicit prices are stable across subsamples and across the price24distribution, the neighbourhood effects absorb 82\% of the spatial autocorrelation in25residuals, floor space displays diminishing returns, value decays with distance to major26metros as the monocentric tradition predicts, and the structural prices transfer across27provinces in leave-one-province-out cross-validation.2829The reading we propose is deliberately modest and, we believe, well supported: in30Canada, the single largest priced component of a home is its neighbourhood. That fact31is simultaneously a measurement result (for assessment, taxation and index32construction), a validation result (a transparent fixed-effects design removes the33spatial dependence that motivates far more elaborate machinery), and a policy result34(the premium dispersion that supply constraints and agglomeration generate is the35dominant term in the decomposition). Section~\ref{sec:discussion} details the36limitations---list prices, unobserved quality, noisy lot data, a single37cross-section---and the research agenda they imply. Even within those bounds, the model38provides a transparent, reproducible benchmark for automated valuation, market39monitoring and the welfare analysis of local amenities across the Canadian residential40market.4142\vspace{0.6em}43\noindent\textbf{Reproducibility.} Data engineering was performed in \texttt{DuckDB} and44\texttt{pandas}; estimation used \texttt{statsmodels} (cluster-robust OLS) and45\texttt{linearmodels} (absorbing least squares); figures were produced in46\texttt{matplotlib}. All tables and figures are generated programmatically from the47source database by the numbered scripts in the accompanying repository.48