# PAPER_REVIEW — Critical assessment before the scholarly upgrade (2026-08-05)

Scope: `paper/` (v1.1, 26 pp., 32 references). Assessment only — no results questioned.

## 1. Core contribution: is it clearly stated?

The empirical contribution is clear and genuinely stated in one sentence: *a national-scale
decomposition of Canadian house prices into structure vs. location using 1,153 absorbed FSA
fixed effects on 140,931 listings*. Three supporting claims (elasticity ≈ 0.55, bathrooms
≈ +11%, bedrooms ≈ 0; OOS R² 0.764; premia span ×9) are quantified and reproducible.

**Weaknesses in the contribution statement:**
- The introduction motivates in three bullets ("first-order question", "granular data",
  "transparent benchmark") but never states *what is new relative to specific prior work*.
  No named paper is identified as the closest antecedent, so the gap is asserted, not shown.
- "Among the most geographically comprehensive single-equation hedonic exercises for
  Canada" is unverifiable as written because the Canadian empirical hedonic record
  (Des Rosiers/Thériault's Quebec program, Toronto and Vancouver studies) is never cited.
- The headline result — location ≈ 30 pp of R² — is never connected to the macro
  literature quantifying the land/location share of housing value (Davis–Heathcote;
  Knoll–Schularick–Steger), which is the natural external benchmark for its magnitude.

## 2. Literature review: gaps by subfield

Current coverage (good): Rosen framework and second-stage identification; semi-log
functional form; spatial econometrics vs. fixed effects debate; three amenity
capitalization classics; brief AVM motivation.

**Missing subfields (each currently has zero citations):**

1. **Urban spatial structure / monocentric city.** The paper estimates an urban price
   gradient (§ Results 5.6) with no reference to Alonso–Muth–Mills, nor to modern
   quantitative treatments (Ahlfeldt–Redding–Sturm–Wolf; Duranton–Puga). The gradient
   section cites only Combes et al. (2015).
2. **Land value and the location share.** Davis & Heathcote (2007), Knoll, Schularick &
   Steger (2017), Albouy's work on urban land values. Directly benchmark the 30-pp claim.
3. **House-price indices.** The hedonic/repeat-sales index literature (Bailey–Muth–Nourse;
   Case–Shiller; Hill's survey) — the paper claims relevance for "house-price indices" in
   the first paragraph without citing any index paper.
4. **Housing submarkets.** Goodman & Thibodeau; Bourassa et al. on submarket definitions —
   the FSA-fixed-effect design is a submarket argument and should engage this strand.
5. **Spatial dependence in hedonics (empirics).** Dubin (1988); Basu & Thibodeau (1998);
   and the Canadian spatio-temporal work of Dubé & Legros. Moran's I is used but only
   Moran (1950) is cited.
6. **Capitalization theory.** Tiebout (1956), Oates (1969), Roback (1982) — the FSA
   premium is interpreted as capitalized amenities with no reference to why amenities
   capitalize.
7. **List price vs. transaction price.** The central data caveat (list prices) has no
   citation — Genesove & Mayer (2001) on seller behaviour, Han & Strange (2015) on the
   listing/search microstructure.
8. **AVM / ML valuation empirics.** One citation (Mullainathan & Spiess 2017). Missing the
   applied mass-appraisal accuracy literature and the ML-vs-hedonic horse-race studies the
   OOS section implicitly competes with.
9. **Canadian housing.** One CMHC report. Missing the Canadian hedonic tradition
   (Des Rosiers and coauthors on Quebec; Toronto/Vancouver-specific studies) and the
   supply-side explanations for Vancouver/Toronto price levels (Saiz-style constraints;
   superstar-city dynamics of Gyourko–Mayer–Sinai).
10. **High-dimensional FE methodology.** Correia (2017) is cited, but the intellectual
    origin (Abowd–Kramarz–Margolis) and the standard computational references
    (Guimarães & Portugal) are not.
11. **Sample selection / trimming practice.** The 1% trim and its robustness variants have
    no methodological anchor.

## 3. Weak argumentation / unsupported claims

- "Competitive with commercial automated valuation models" (§ Robustness, Conclusion) —
  no AVM accuracy benchmark is cited; needs at least one accuracy-standard reference
  (e.g., IAAO mass-appraisal standards or a published Zestimate-type accuracy figure).
- "Credible national estimates for Canada are scarce" — unsupported; requires citing what
  *does* exist to establish scarcity.
- Half-bath negative coefficient interpretation ("older or more compartmentalised floor
  plans") is speculative and unflagged — should be presented as conjecture or supported.
- "This is the housing counterpart of the within estimator" — fine, but the AKM/FE lineage
  should be cited where the claim is made.
- The bedrooms≈0 result is called "a textbook hedonic finding" without a supporting
  citation (Sirmans et al. tabulate mixed signs — an easy, honest anchor).
- Urban gradient "classic urban theory predicts" — no Alonso/Muth/Mills citation.

## 4. Underdeveloped sections

- **Introduction** (~1 page): thin on stakes (housing = largest household asset class;
  assessment/taxation machinery; index construction) and has no explicit numbered
  contribution list.
- **Related work** (~1.5 pages): organized adequately by theme but shallow within themes
  (2–4 citations each) and never says *where the gap is* at the end of each strand.
- **Discussion: absent.** Results flow directly into robustness and a short conclusion.
  There is no section interpreting magnitudes against prior literature (is 0.55 a typical
  elasticity? is an 11% bathroom premium high? is ×9 premia dispersion large vs. the US?),
  no engagement with findings that could contradict (e.g., submarket papers arguing
  *coefficients*, not just intercepts, vary by location — which the paper's own
  heterogeneity section partially confirms), and no consolidated Limitations & Future Work
  (currently squeezed into two conclusion paragraphs).
- **Methodology**: several choices uncited (trimming, FSA pooling threshold, kNN Moran
  implementation choice, quantile regression estimator details).

## 5. Positioning: current vs. ideal

**Current:** a well-executed measurement paper that reads as a technical report — it
presents itself mainly as "we ran a big transparent regression."

**Ideal:** a contribution at the intersection of (i) the classic structure-vs-location
decomposition, (ii) the land-share macro literature, and (iii) the AVM benchmarking
debate — i.e., "the neighbourhood is the single largest priced component of Canadian
housing; we quantify it with a design that is transparent enough to audit and accurate
enough to rival black-box AVMs." The framing should name its closest antecedents (US
land-share estimates; submarket hedonics; ML-valuation horse races) and state exactly
what is new: national scope for Canada, listing-level granularity, absorbed-FE scale, and
a validated accuracy claim.

## 6. Consequences for the upgrade (implemented in Phases 2–3)

1. Add ~28 verified references across the 11 gaps above (target ≈ 60 total).
2. Restructure Related Work into thematic subsections, each ending with the gap.
3. Add a proper **Discussion** section: magnitudes vs. literature, agreements/
   disagreements, implications (assessment, taxation, indices, AVM), limitations & future
   work.
4. Deepen the introduction (stakes, numbered contributions, roadmap); strengthen the
   conclusion without overselling.
5. Ground every methodological choice in a citation; keep every number, figure and table
   untouched.