The Price of Residential Land for Counties, ZIP Codes, and Census Tracts in the United States

Summary

This FHFA staff working paper (19-01, 2019; published in the Journal of Monetary Economics, 2021) by Morris Davis, William Larson, Stephen Oliner & Jessica Shui constructs land prices and land shares for counties, ZIP codes, and census tracts across the United States, from millions of property appraisals over 2012–2019.[1] It is the granular successor to the aggregate land-price literature its authors built — Davis & Heathcote (2007) and Davis & Palumbo (2008), the series the Lincoln Institute maintains as "Land and Property Values in the U.S."[2] — and the modern data backbone under Larson's earlier national estimate.

Why It Matters Twice Over

• As data: it is the wiki's best citable source for where US land value sits — tract-level land shares covering most of the housing stock — feeding the scale-of-rent debate and the geography of the capital-share finding.
• As a proof of method: producing defensible land values separately from structures, at national scale, from appraisal microdata (including kernel-smoothing over vacant-land and teardown sales) is precisely what the assessment objection says cannot be done reliably. A federal housing agency doing it as routine research is the strongest institutional counter-exhibit on file — stronger than any advocacy claim, because assessment feasibility is incidental to the authors' purpose.
• It is cited in both the data and assessment installments of Doucet's ACX series; per the wiki's de-referencing rule, the wiki now carries it directly.

Nuances and Limits

• Residential only — the dataset covers residential land; commercial, agricultural, and public land require other sources.
• Appraisal-derived — estimates inherit appraisal-industry conventions; the authors' methods mitigate but cannot eliminate this.
• Headline magnitudes (direct read of WP 19-01). The authors put land's share of house value at about 40% "for the aggregate United States over 2012–2019" (used in their §4.3 measurement-error calibration). Land shares vary enormously across geography — from 7.7% at the 1st percentile of counties to 54.3% at the 99th (pooled cross-section) — and the paper's dynamic finding is that over 2012–2019 land prices (and hence land shares) rose fastest in the largest metros and near CBDs, while land shares fell in the smallest metros. The average price of residential land was about $152k/acre (median $53k/acre), with a right-skewed distribution. The dataset covers 960 counties, 7,742 ZIP codes, and 10,515 census tracts in the annual panel (more in the pooled cross-section). Scan depth now Full for these figures.

Bears On

• Outcome (data): Land rent could fund government
• Objection (method counter-exhibit): Land cannot be assessed
• Concept: Mass-Appraisal Methods
• Research: Davis & Heathcote (2007) · Larson (2015)

See Also

• Albouy, Ehrlich & Shin — Metropolitan Land Values
• Gwartney, Estimating Land Values — the practitioner counterpart
• Recording Land in the National Balance Sheet (London Group, 2009) — the national-accounts backdrop: why official balance sheets struggled to separate land from structures before granular projects like this one

Sources

1. Morris A. Davis, William D. Larson, Stephen D. Oliner & Jessica Shui, "The Price of Residential Land for Counties, ZIP Codes, and Census Tracts in the United States," FHFA Staff Working Paper 19-01 (2019); Journal of Monetary Economics (2021). FHFA · Publisher · full-text PDF — full text read directly in this pass; used for scope, data, method, and the headline magnitudes (≈40% aggregate US land share 2012–2019; 7.7–54.3% county percentile range; $152k/acre average land price; panel coverage counts).
2. Morris A. Davis & Michael G. Palumbo, "The price of residential land in large US cities," Journal of Urban Economics 63(1), 2008 (earlier FRB FEDS 2006-25). FEDS — used for the series lineage (A-claim).