Stewart (2022): How Important Are Land Values in House Price Growth? Evidence from Canadian Cities
A Canadian growth-accounting decomposition using directly observed (not imputed) land prices finds land is the dominant contributor to new-home price growth in some cities but not uniformly — the land share 'varies greatly by city,' disciplining the universal reading while confirming the.
Summary
"How Important Are Land Values in House Price Growth? Evidence from Canadian Cities," by Kenneth G. Stewart (University of Victoria), appeared in the Canadian Journal of Economics / Revue canadienne d'économique 55(1): 249–271 (2022). It applies a Cobb–Douglas growth-accounting framework to decompose the price of newly constructed homes across major Canadian cities into structure and land contributions.
Its methodological advantage over most of the decomposition literature is the data: as the abstract states, "the data set is unusual in that land prices are directly observed rather than having to be imputed, and quality change is carefully controlled for in the measurement of structures." Because land prices are observed rather than backed out as a residual, the study avoids the residual method's known tendency to over-attribute boom-bust price swings to land (a caution the wiki also carries via Kuminoff & Pope).
The Finding, With Its Nuance
The headline result is deliberately two-sided. Stewart writes: "Whereas standard analyses find land costs to be the dominant contributor to the growth in housing costs, I find that this varies greatly by city." In some Canadian cities land is indeed the dominant driver of house-price growth — consistent with the Knoll–Schularick–Steger and Glaeser–Gyourko findings — while in others the structure component carries more of the growth. This heterogeneity is not a refutation of the land-problem diagnosis so much as a sharpening of its scope: the land channel bites hardest in the most demand-pressured, supply-constrained cities, exactly the scope to which the wiki restricts the "housing unaffordability is a land problem" claim, and less so elsewhere. Notably, despite the city variation, Stewart reports that "the evidence supports other recent work that endorses the constant-returns Cobb–Douglas methodological framework" used across this decomposition literature.
Why It Matters
Stewart supplies an independent, directly measured, national test of the land/structure decomposition from Canada — a data source and methodology distinct from the US micro and 140-year cross-country evidence. It corroborates the framework and the land channel where demand pressure is high, while providing an honest empirical caution against over-generalizing the land diagnosis to markets where it does not hold. That combination makes it a supporter of the scoped land-problem claim rather than the unscoped one.
See Also
- Housing unaffordability is a land problem, not a construction-cost problem — the outcome claim this evidence supports, within its stated scope
- Knoll, Schularick & Steger (2017) · Glaeser & Gyourko (2018) — the cross-country and US decompositions Stewart tests against
- Davis & Heathcote (2007) · Kuminoff & Pope — the residual-method debate Stewart's directly observed land prices sidestep
Sources
- Kenneth G. Stewart (2022), "How Important Are Land Values in House Price Growth? Evidence from Canadian Cities," Canadian Journal of Economics / Revue canadienne d'économique 55(1): 249–271. DOI · RePEc/IDEAS listing — used for the growth-accounting decomposition, the directly-observed-land-price data advantage, and the "varies greatly by city" heterogeneity result; abstract quotations verified verbatim against the published CJE abstract this session (2026-07-13).