Where Did All the Land Go? (Michael Hudson's critique of Fed land-value accounting)

Summary

"Where Did All the Land Go? The Fed's New Balance Sheet Calculations — A Critique of Land Value Statistics" is a March 2001 essay by economist Michael Hudson, arguing that the Federal Reserve's method for splitting real-estate value between land and structures systematically understates land, to the point of implying negative land values for US corporate real estate.[1] It is the "Fed-critique Doucet leans on" that Lars Doucet's Does Georgism Work? Part 1 cites but the wiki had not yet ingested; this page de-references it.

Hudson is a heterodox, advocacy-adjacent economist (a self-described critic of the "FIRE" — finance, insurance, real estate — sector), so this page tiers his claims carefully per the wiki's source hierarchy: his factual critique of the Fed's accounting mechanics is checkable and independently corroborated by mainstream land-value researchers; his estimates are back-of-envelope; and his political interpretations remain attributed to him, not asserted. The essay is published under a Creative Commons BY-NC-SA licence on Hudson's own site.[1]

What It Establishes

The checkable accounting critique. The Fed's "Balance Sheets of the American Economy" (the B-series of the Flow of Funds Z.1 release) valued economy-wide real estate at $13.4 trillion for 1994 — $4.4T land, $9T structures — with land "about a third of total real estate values."[1] Hudson's core point is about how that land figure is produced: the Fed imputes land as a residual, subtracting the estimated replacement cost of buildings from a property's overall market price. Because replacement cost is indexed up by construction prices year after year — even for obsolete or unmaintained buildings — the residual left for land is squeezed, and can go negative. Hudson reports the Fed's own result that "by 1993 the FRB estimated that the land held by all nonfinancial corporations had a negative value of $4 billion," and that the Fed "stopped publishing" land estimates in 1995 rather than defend the figure.[1] His memorable illustration: anyone offered "$4 billion" plus all US nonfinancial-corporate land "free and clear" would accept, becoming "instantaneously a multi-billionaire" — so the negative number is nonsense.[1]

The proposed correction (attributed to Hudson). He argues it is "more plausible to use a 'building residual' method of appraisal than to use a 'land residual'" — start from an independent land appraisal and treat structures as the residual — and calls for "a land-map of the U.S. economy." His own "sense of proportion" adjustments (valuing structures at historical rather than replacement cost, holding land shares steady, etc.) would roughly double the Fed's land total, from $4.4 trillion to nearly $9 trillion, implying land is "60 percent" of real-estate value rather than the "40 percent" rule of thumb.[1] These are explicitly rough estimates, not a formal econometric measurement.

Relation to the Georgist Case

The essay matters to the Georgist argument in one narrow, robust way and one broader, contested way.

Robust: if official land statistics understate land value — because the residual method mechanically transfers land's appreciation onto structures — then the fiscal base for a land value tax is larger than headline Fed numbers suggest, strengthening the premise behind land rent could fund a large share of government. Crucially, Hudson's accounting critique is not just a Georgist talking point: the same defect is documented by mainstream land-value economists. Albouy, Ehrlich & Shin (2018) note that the residual method can produce implausible, even negative, land values in the Fed Flow of Funds, which motivated their transaction-based alternative; the wiki's Davis & Heathcote page records the same negative-value problem in disaggregated residual series. And Kuminoff & Pope (2013) show, from the other side of the cycle, that pinning structures to replacement cost mis-splits land and structures during booms too. Hudson's central factual claim thus survives independent checking.

Contested: Hudson's interpretation of why the mis-measurement persists is heterodox advocacy and should travel as his view only. He attributes it to the "FIRE sector['s] self-interest in not tracking land gains more closely," to the maxim "what is not seen will not be taxed," and to tax machinery such as "over-depreciation" (a term he credits to Lowell Harriss) and the 1981 Reagan depreciation rules.[1] These are arguments about motive and politics, not measured findings.

Limits

• Advocacy source, not peer-reviewed. Per the wiki's source hierarchy, Hudson is used to represent his own position; his quantitative adjustments are self-described rules of thumb ("sense of proportion" questions), not a rigorous estimate, and should not be cited as a measurement of US land value.
• Dated (2001). The specific Fed vintages (1993–96 figures), the discontinued B-series line, and the tax-code details are of their time; the Fed's later Flow-of-Funds and the FHFA/BEA land series (Davis, Larson, Oliner & Shui, Larson) supersede the numbers even where they vindicate the method critique.
• Direction-of-bias is method-dependent. Hudson argues the residual method makes land vanish in downturns; Kuminoff & Pope show it can overstate land in booms. Both are right that the residual method mis-attributes value — but "official statistics understate land" is not an unconditional law, so the safest claim is the neutral one: the land/structure split is method-sensitive and residual accounting handles it poorly.
• Interpretive claims unverified. The political-economy narrative (FIRE-sector lobbying, deliberate statistical suppression) is asserted, not evidenced in the essay, and stays attributed.

Bears On

• Problem: Land rent could fund a large share of government — as context, not a load-bearing anchor: Hudson argues official statistics understate the land base, a claim whose accounting core is corroborated by mainstream sources even though his numbers are informal.
• Concept: Economic Rent · Unearned Increment — the essay is, at bottom, an argument that residual accounting hides land rent and the unearned increment.

See Also

• Does Georgism Work? (Doucet) — Part 1 leans on this Fed-critique
• Kuminoff & Pope — Value of Residential Land and Structures — the same residual-method mis-split, seen in a boom
• Albouy, Ehrlich & Shin — Metropolitan Land Values — mainstream corroboration of the negative-residual problem
• Davis & Heathcote — Price and Quantity of Residential Land — Federal-Reserve-affiliated residual-method series
• Killing the Host (Hudson) — Hudson's book-length treatment of finance and rent

Sources

1. Michael Hudson, "Where Did All the Land Go? The Fed's New Balance Sheet Calculations — A Critique of Land Value Statistics," 25 March 2001, michael-hudson.com (CC BY-NC-SA 3.0). Fetched and read in full (2026-07-10): michael-hudson.com/2001/03/where-did-all-the-land-go.... Used for: the Fed's residual/replacement-cost imputation method; the 1994 $4.4T land / $9T structures / $13.4T real-estate baseline; the 1993 negative-$4-billion corporate-land result; the 1995 discontinuation of Fed land estimates; the building-residual proposal and the ~$4.4T→~$9T (60%-of-real-estate) revaluation; and the attributed FIRE-sector / "what is not seen will not be taxed" / over-depreciation interpretations.
2. David Albouy, Gabriel Ehrlich & Minchul Shin, "Metropolitan Land Values," Review of Economics and Statistics 100(3), 2018 — used for independent, mainstream corroboration that the residual method can yield implausible/negative land values in the Fed Flow of Funds. See wiki page.
3. Lars Doucet (2021), Does Georgism Work? Part 1, Astral Codex Ten — identifies "Michael Hudson, Where Did All the Land Go? (2001)" as the Fed-critique his land-magnitude argument leans on; this page de-references it. See wiki page.