Gaffney (2015): A Real-Assets Model of Economic Crises — Will China Crash in 2015?

Overview

"A Real-Assets Model of Economic Crises: Will China Crash in 2015?" is a peer-reviewed paper by Mason Gaffney published in the American Journal of Economics and Sociology, Vol. 74, No. 2 (March 2015), DOI 10.1111/ajes.12093.[1] Unlike most of Gaffney's business-cycle writing on the wiki — the 1982 "Causes of Downturns" working notes and the short capital-destruction memos on How Land Booms Destroy Capital — this is a finished, refereed journal article that states a general model explicitly and then applies it to a testable, dated, out-of-sample prediction: that China's property-driven economy would enter contraction in 2015.[1] Before this wave the wiki cited the paper only once, as an external "See Also" bullet on concepts/18-year-land-cycle; this page mines it in full for its distinct mechanism, which is separate from — and complementary to — the empirical periodicity work of Fred Harrison and Phil Anderson that page otherwise relies on. Where Harrison and Anderson document when land cycles recur (roughly 18-year historical periodicity), Gaffney's model addresses why a land-price cycle produces a banking crisis, through a specific capital-theory mechanism.

The Four-Element Model

Gaffney states the hypothesis as four elements, developed in turn:[1]

1. A rise and fall in land prices, driven mostly by "autonomous real economic changes" rather than by monetary or fiscal policy, which reflects and can reinforce the real change but does not initiate it.
2. Investment at the geographic and value margins — marginal sites developed only because better locations are being held off the market by owners anticipating further price rises.
3. A shift in the structure of capital investment toward long-payout, low-turnover "fixed capital" (buildings, infrastructure) at the expense of fast-turnover "circulating capital."
4. Rising bank leverage as a consequence of (3), because slower capital turnover forces banks to hold increasingly illiquid, long-dated collateral — leaving many technically insolvent once the land-price bubble bursts.

This ordering is the paper's central methodological claim against the standard account: banks are not the autonomous cause of the cycle but a downstream amplifier responding to a real-side land-price signal. "Banks are responding to an external stimulus... rather than creating the conditions for a boom on their own."[1]

The Capital-Turnover Mechanism

The paper's most original technical contribution is a simple formalization of why a shift toward fixed capital drags on employment and growth, developed through the identity K·T = F — capital investment (K) times turnover (T) equals the flow of capital (F) available in a period.[1] A retail inventory turning over twelve times a year at $100,000 generates $1.2 million in annual flow; the same $100,000 sunk into a 40-year building at T = 1/40 generates a dramatically smaller flow, and — at an 8 percent interest rate — roughly 70 percent of the payments on that investment go to interest rather than to production. "[T]he ability of capital to sustain labor is as much a question of turnover as it is the size of the capital stock," Gaffney writes, attributing the underlying insight to Smith, Ricardo, Wicksell, and Turgot rather than claiming it as novel — only its application to the land-price cycle is new.[1] Gaffney identifies four distinct channels by which rising land prices push investment from circulating toward fixed capital: a circular (positive-feedback) effect (marginal-site development lowers returns elsewhere, pushing more capital toward land); a price effect (builders substitute capital goods for increasingly expensive land, following Alfred Marshall's 1890 observation); a wealth effect (landowners spend paper gains, displacing saving); and a sprawl effect (scattered development requires more connective infrastructure capital per unit of output).[1]

The 2008 Case and the Banking Mechanism

Applied to the 2008 U.S. crisis, the model treats the housing-price index (S&P/Case-Shiller, 100 in Jan. 2000, peaking at 206 in mid-2006, falling to 139 by April 2009) as a proxy for the underlying, unmeasured land-price cycle, since land itself has no national price index in the United States.[1] Residential construction rose 95 percent (2000–March 2006) then fell 66 percent by February 2011; nonresidential construction rose 232 percent (2002–Nov. 2008) then fell 29 percent — with a 32-month lag between the two peaks, a pattern Gaffney also documents for 1921–1933 (housing construction +167%/−93%; nonresidential +93%/−85%, roughly three years apart).[1] On the banking side, the paper argues repeal of Glass-Steagall let commercial banks package land-collateralized mortgages into collateralized debt obligations, obscuring the underlying leverage; it cites Goldman Sachs' 24.5 percent return on equity for the year ending June 2010 against an all-bank average of only 3.5 percent (versus a 14.5 percent S&P average) as evidence of how unevenly the "recovery" was distributed, and argues that mark-to-market accounting would have revealed many nominally healthy 2010 banks as insolvent.[1]

The China Forecast

Writing "as I write this in the fall of 2014," Gaffney treats China as an unusually clean test of the model because — unlike the U.S., Ireland, Spain, or Greece — it had not yet experienced a full property-driven bust, despite having the same underlying real-asset dynamics (rapid land-price appreciation, marginal-site overbuilding, rising local-government debt) operating through a different institutional structure (state land ownership, 50–70-year leaseholds functioning as de facto private tenure, state-owned banks rather than private ones).[1] The evidence he marshals: Beijing land prices up over 750 percent from 2003–2010 versus housing prices up only 100 percent over the same period (Gyourko et al. 2010), confirming land's greater volatility relative to housing; a price-to-rent ratio of roughly 30 in eight major cities (2007–2010), matching the U.S. ratio just before its 2007 peak; a housing inventory-to-sales ratio rising from 12 to 18 months (June 2013–June 2014) against a U.S. figure of 5.8 months; local-government debt growing from 10.7 to 17.9 trillion yuan (2010–2013), increasingly financed through "shadow banks"; and office vacancy rates of 15–21 percent in second-tier cities, with Chengdu cited as an extreme case (nearly half its office space vacant).[1] Gaffney's explicit methodological point — anticipating the standard objection that national aggregates in a country as large as China obscure local conditions — is that "overbuilding in small cities throughout China is a much better indicator of a coming decline... than indices that measure only the big cities," since the 200 or so mid-sized Chinese cities account for 70 percent of national residential sales.[1] The paper predicts, on this basis, "a likely recession" in China in 2015.

Outcome. The Shanghai Composite index fell more than 30 percent between June and July 2015, months after the paper's March 2015 publication — an outcome documented contemporaneously by Martin Adams on progress.org and already cited on concepts/18-year-land-cycle.[2] This page treats the forecast's confirmation as suggestive rather than dispositive: a single successful out-of-sample prediction, however striking in timing, is not independent statistical validation of the four-element model as a general theory, and the 2015 stock-market fall is not identical to the property-sector contraction the paper specifically forecast (China's property and local-government debt problems continued to develop over subsequent years rather than resolving in a single 2015 crash).

Policy Implications: Credit Controls vs. Taxing Land Values

The paper closes by comparing two classes of crash-prevention policy. Credit controls (interest-rate rules, capital requirements, capital controls on cross-border flows) are found only partially effective: Gaffney cites Kim and Yang (2008) finding capital inflows explain only "a relatively small part of asset price fluctuations" in emerging Asian economies, and Olaberría (2012) finding capital controls do not reliably reduce the probability of asset-price booms.[1] Gaffney's preferred alternative is taxing land values directly, on the reasoning that a tax targeted at the land-price signal itself — rather than at the credit that merely responds to it — addresses the root of the four-element sequence rather than one downstream symptom.

Standing and Limits

• Claim class. The four-element model and turnover mechanism are B/C-claims (a theoretical framework illustrated with descriptive statistics, not an econometric test with standard errors or a formal out-of-sample forecast evaluation). The China section is the paper's strongest evidentiary case precisely because it is a genuine ex-ante prediction, but it is a single case, not a panel test.
• Peer-reviewed but not independently replicated. This is the most academically credentialed single paper in the wiki's Gaffney business-cycle material — published in a refereed journal rather than as a working paper or book chapter — but the wiki has not identified a subsequent peer-reviewed paper that formally tests the K·T = F turnover mechanism against alternative crisis models.
• China data circa 2014. The paper's Chinese statistics are drawn from 2010–2014 sources (Gyourko et al. 2010; Nie and Cao 2014; Davis and McMahon 2013) and are now a decade old; readers using this page for a current view of Chinese property markets should treat the specific figures as historical, not current, per the wiki's standing dated-data convention.
• Native text, no OCR. The masongaffney.org copy and the Wiley-hosted published version were both consulted; the working text mirror is drawn from a cleanly pdftotext-extractable copy, no OCR needed.
• Priority note (2026-07-18). The K·T = F capital-turnover identity this paper states and attributes, without novelty claims, to "Smith, Ricardo, Wicksell, and Turgot" is not new to Gaffney's own corpus either: he derives the identical mechanism at much greater length and with full mathematical derivation — including how land rent shortens the optimal capital-recovery cycle (dR/dn = i(R+S)) — in his 68-page 1976 chapter "Toward Full Employment with Limited Land and Capital." That paper's employment framing is distinct from this one's business-cycle framing, but the underlying turnover mechanism is the same one Gaffney worked out nearly forty years earlier.

Bears On

• Concept: 18-Year Land Cycle — supplies the mechanistic "why" (capital-turnover and bank-leverage channels) alongside Harrison's and Anderson's empirical "when" (historical periodicity); the page's existing citation is upgraded from a bare bibliographic bullet to a substantively mined source.
• Outcome: Resource-rent capture works — cited lightly as a mechanism argument for why land-value taxation, rather than credit regulation alone, addresses crisis prevention at its root.

See Also

• Gaffney (1976), Toward Full Employment with Limited Land and Capital — the earlier, fuller mathematical derivation of the K·T = F capital-turnover mechanism this paper uses
• Gaffney (1982), Causes of Downturns: an Austro-Georgist Synthesis — the earlier, unpublished working-notes version of a similar land-price-to-credit-collapse sequence
• Gaffney, How Land Booms Destroy Capital — the short-form statement of the capital-turnover/capital-destruction mechanism
• After the Crash: Designing a Depression-Free Economy (2009) — Gaffney's book-length treatment this paper's model summarizes and extends to China
• 18-Year Land Cycle
• Mason Gaffney

Sources

1. Mason Gaffney (2015), "A Real-Assets Model of Economic Crises: Will China Crash in 2015?" American Journal of Economics and Sociology 74(2), 325–353. DOI: 10.1111/ajes.12093 — used for all claims, figures, and quotations on this page; read in full from a local mirror of the source PDF. Local mirror at sources/gaffney/text/I2015-RealAssetsModel-China.txt.
2. Martin Adams (2015), "Mason Gaffney Predicted the China Crash," progress.org — used for the contemporaneous documentation of the Shanghai Composite's June–July 2015 decline as an outcome check on the paper's forecast. Article