Wealth and Property Taxation in the United States (Dray, Landais & Stantcheva)

Summary

Sacha Dray (World Bank), Camille Landais (LSE/CEPR) and Stefanie Stantcheva (Harvard/NBER) reconstruct the long-run history of American wealth and property taxation from the administrative records of the General Property Tax — the broad, in-principle comprehensive levy that most U.S. states used to tax "all property" (real estate, business assets, money, stocks, bonds, and, before Emancipation, enslaved people) from the 1840s until its decline after the Great Depression.[1] Using newly digitized state annual reports and decadal U.S. Census "Wealth, Debt, and Taxation" compilations, the authors build annual state-level wealth series back to 1850 (or earlier), decadal county-level series to 1850, and a national series from the early 1800s — what they describe as the first comparably comprehensive, high-frequency, sub-national U.S. wealth dataset for this period (p. 1, p. 5).

The paper is not primarily a study of property taxation — its main contributions are a new wealth dataset and findings on long-run spatial inequality, Civil War/Reconstruction economics, and the county-level correlates of wealth growth. But Sections 2–3 give the General Property Tax the fullest documented institutional and quantitative history the wiki currently holds for any period of U.S. property taxation, including its revenue share of GDP, its effective rates, and the mechanisms behind its 20th-century decline.

The Core Argument and Findings

Property tax revenue as a share of GDP rose, then fell by half. In 1850, total property tax revenue across all levels of government — state, county, municipal, and special districts — was "somewhat below 2% of GDP." It "more than doubled to 5% of GDP in the 1920s," the eve of the Great Depression, before falling to "around 2.5–3% in the 1950s and beyond" (Figure 1, p. 11–12).[1]

The tax's share of total government revenue collapsed even faster than its share of GDP. Citing Benson et al. (1965), the authors report that property tax revenue fell from 38.8% of total government revenue in 1927 to 25.2% in 1938, then to 8.1% in 1946 — a decline the authors attribute to three interacting causes: the expansion of federal spending under the New Deal and Social Security (crowding in income-tax financing), new state revenue sources (automobile fees, motor-fuel, sales, and income taxes), and Depression-era and postwar exemptions — homestead exemptions for owner-occupied residences and statutory rate limits that "further accelerated the decline" (p. 12).[1]

Effective rates on wealth roughly doubled from 1850 to 1930. The authors compute effective property tax rates as the ratio of tax revenue to their own estimates of property value at each level of jurisdiction. Municipal and lower-level rates rose from about 0.3% in 1850 to about 1% in 1930, while combined county-and-state rates stayed roughly flat around 0.3%; the total effective rate rose from about 0.6% in 1850 to about 1.35% in 1930 (Figure 2, p. 13).[1] For 1920 specifically, the paper gives a full breakdown: "the average effective tax rate was 1.4%; the average city tax rate was 1%; the average county tax rate was 0.24%; and the average state tax rate 0.16%," with substantial geographic dispersion — from around 0.5% in the lowest-tax areas to more than 3% in the highest (p. 16).[1]

The base narrowed through classification and exemption, not primarily through non-compliance. Turn-of-the-century critics — "often spearheaded by economists" — attacked the General Property Tax on three grounds: local assessors could not keep pace with property that "became increasingly intangible and mobile (e.g., stocks, bonds, and other financial assets)"; assessment quality suffered as the economy grew more complex; and rising wage income made property a worse proxy for ability to pay (p. 9).[1] The response was a "classification movement," which replaced the uniformity clause with lower statutory rates specifically on intangible property (p. 9, n.10).[1] The authors' own exemption data — drawn from Census Bureau and National Industrial Conference Board records covering 1880–1937 — show the exemption ratio (exempt property as a share of total property) was "generally small and stable over time... around 6–7%" nationally (p. 14).[1] The tax's later collapse is therefore attributed mainly to institutional substitution (new revenue sources, homestead exemptions, and rate limits after the 1930s) rather than to a large pre-1937 exemption of intangible wealth from the recorded base.

Reconstruction-era Southern rates spiked, then collapsed with Jim Crow. Effective property tax rates in the antebellum South were about half those in the North. During Reconstruction, newly elected Republican legislators — facing a diminished tax base and the need to fund public schools — pushed rates up sharply: Southern effective rates "almost tripled in about five years, reaching a peak of 1.2% in 1870." As Democratic "Redeemer" governments regained control and Jim Crow was instituted, rates "quickly reverted to around 0.6%" (p. 31–32).[1] In three Deep South states — Georgia, Florida, and Alabama — enslaved people had accounted for more than 50% of recorded property value before the war; excluding that category, Southern wealth fell by more than 25% between 1860 and 1870 (p. 5).[1]

National wealth grew rapidly, and spatial inequality did not converge. After the Civil War, national wealth grew "much faster than income," and the U.S. wealth-to-income ratio rose from around 300% in the early 19th century toward almost 600% by the 1920s (p. 30, Figure area).[1] Despite internal migration and deepening national capital markets, county- and state-level wealth inequality shows no "sigma-convergence" (no decline in dispersion) through 1930: the share of national wealth held by the top-10% richest counties rose, and county wealth growth 1870–1930 is negatively correlated with a county's initial 1870 top-10% wealth concentration, robust to geographic and demographic controls (p. 5–6).[1]

Relation to the Georgist Case

This paper is a history of the taxation of wealth through property — of which land is a large but never separately measured part. The authors' own "real property" category bundles "land, buildings, and improvements" throughout (p. 27),[1] and the conclusion explicitly names "a finer analysis of different types of wealth" — including, implicitly, isolating land from structures — as future work the dataset could support but that this paper does not do (p. 38).[1] Readers looking for a land-specific effective tax rate, or a land share of GDP, will not find one here; what the paper documents is the taxation of a broad wealth base that included land alongside buildings, business capital, and (in one state, Connecticut, where detailed data exist) intangible financial assets.

That caveat is what makes the paper useful as historical background for the Georgist case rather than as direct land-tax evidence. The General Property Tax's founding ambition — to tax "all property," uniformly — is the fiscal-history opposite of a land value tax: rather than concentrating the levy on the one asset class that classical and modern public-finance economists single out for its fixed supply and administrative transparency, land, it tried to reach everything, including highly mobile financial wealth that proved administratively very hard to assess. The paper's account of the "classification movement" — carving out lower rates for intangible property because assessors could not reliably value or locate it (p. 9) — is a documented instance of exactly the base-erosion problem that a land-only tax largely avoids: land cannot be hidden, moved across a state line, or reclassified into a lower-rate intangible category. The paper's own data on this mechanism are limited (their national exemption-ratio series runs only 1880–1937 and puts the exempt share at 6–7%, smaller than the later, post-1930s narrowing driven by homestead exemptions and rate limits), so this reading is an inference from the institutional history the authors document, not a claim the paper itself makes about land taxation.

Nuances and Limits

• Land and buildings are never disaggregated in the main analysis. "Real property" combines land and improvements throughout; a state-by-state breakout of taxable land alone exists for only a subset of states with unusually detailed records (p. 27), and the paper does not report a national land-only series.[1]
• The text consulted is the NBER working paper, not the QJE typeset article. The Quarterly Journal of Economics published this paper as an advance article on 4 September 2026 (DOI 10.1093/qje/qjag038); the summary above was read from NBER Working Paper 31080 (revised May 2026), the version the QJE advance listing corresponds to. The two have not been compared line by line, and NBER working papers are, in the authors' own disclaimer, circulated for discussion and have not themselves been independently peer-reviewed.
• The exemption-ratio evidence covers 1880–1937 only, and the authors describe it as "generally small and stable" over that window (p. 14) — it does not, on its own, explain the much larger post-1930s decline in the tax's share of both GDP and total government revenue, which the paper instead attributes mainly to federal expansion, new state revenue sources, and postwar homestead exemptions and rate limits (p. 9–12).
• The growth-determinants results (county literacy, agglomeration, and the inequality- growth correlation) rest on regressions in Sections 5–7 that this summary does not walk through in full; they are cited here only for the headline direction the authors report.

Bears On

• Place: United States — the source for the "General Property Tax and Its Narrowing" background on that page's property-tax history.
• Research: Saez & Zucman, Wealth Inequality in the United States since 1913 — this paper is effectively that study's 19th- and early-20th-century predecessor, built from direct assessment records rather than capitalized income-tax flows.

See Also

• Saez & Zucman, Wealth Inequality in the United States since 1913
• Lincoln Institute, "Past Forward: Tracing the Property Tax Through Time and Place"
• United States
• Split-Rate Taxation
• Land Value Tax

Sources

1. Sacha Dray, Camille Landais & Stefanie Stantcheva, "Wealth and Property Taxation in the United States," Quarterly Journal of Economics, advance article, 4 September 2026, DOI 10.1093/qje/qjag038 — read from the NBER working-paper version, NBER Working Paper 31080, revised May 2026; the QJE typeset text was not consulted, and the two have not been compared line by line — used for the property-tax-revenue-to-GDP series (below 2% in 1850, 5% in the 1920s, 2.5–3% from the 1950s), the property-tax share of total government revenue (38.8% in 1927, 25.2% in 1938, 8.1% in 1946), the effective-tax-rate series (0.6% in 1850 rising to 1.35% in 1930, with the 1920 jurisdictional breakdown), the classification movement and exemption-ratio data (6–7%, 1880–1937), the Reconstruction-era Southern effective-rate spike to 1.2% (1870) and reversion to 0.6%, the enslaved-property share of Southern wealth before the Civil War, the wealth-to-GDP series, and the explicit statement that land is not separated from buildings in the paper's "real property" category. A-claim: peer-reviewed QJE article; all three authors are established academic economists (World Bank, LSE, Harvard/NBER).