House Price Uncertainty, Timing of Development, and Vacant Land Prices: Evidence for Real Options in Seattle

Summary

"House Price Uncertainty, Timing of Development, and Vacant Land Prices: Evidence for Real Options in Seattle" is a 2006 article by Christopher R. Cunningham, published in the Journal of Urban Economics, vol. 59, issue 1, pp. 1–31 (DOI: 10.1016/j.jue.2005.08.003). Cunningham, an urban/real-estate economist later a research economist at the Federal Reserve Bank of Atlanta, assembled a parcel-level dataset of vacant land characteristics and real-property transactions for the Seattle metropolitan area to test two predictions of real options theory applied to land markets: that greater uncertainty about future house prices should (1) raise the price of vacant land and (2) delay the timing of its development. The paper carries weight in the urban-economics literature as one of the clearest microdata tests of real-options theory in land markets — it is frequently cited alongside Titman's (1985) theoretical treatment of urban land prices under uncertainty as empirical confirmation that developers price and act on the option value of waiting.

The Core Argument and Findings

Real options theory treats an undeveloped parcel of land as analogous to a financial call option: the owner holds the right, but not the obligation, to "exercise" by building, and can instead wait for more information about future prices. Because development is costly and largely irreversible, greater uncertainty about the future payoff from development makes the option to wait more valuable, which should show up empirically as (a) a land price premium above the value implied by current use alone, and (b) a lower probability of development in any given period, holding current profitability constant.

Cunningham tests both predictions directly using Seattle parcel and transaction data, combined with a measure of house-price volatility/uncertainty constructed from the data. The journal's published abstract states the headline results verbatim: "A one-standard-deviation increase in uncertainty lowers the likelihood of development by 11 percent and raises vacant land prices by 1.6 percent. These findings suggest that developers consider their real option to future buildings when deciding to invest."[1] The paper's central, frequently cited quantitative results are therefore:

• Uncertainty raises vacant land prices. A one-standard-deviation increase in house-price uncertainty is associated with a roughly 1.6 percent increase in vacant land prices, holding other parcel and market characteristics constant.
• Uncertainty delays development. The same one-standard-deviation increase in uncertainty is associated with an estimated 11 percent reduction in the probability that a given vacant parcel is developed in a given period.

Both results are consistent with the real-options prediction and inconsistent with a simple net-present-value model in which uncertainty (with no correlation to expected returns) should not affect the land-price/development-timing relationship on its own. Verification note (2026-07-11, Hermes fact-check lane): the "1.6 percent" and "11 percent" headline figures are confirmed verbatim from the ScienceDirect published abstract ("A one-standard-deviation increase in uncertainty lowers the likelihood of development by 11 percent and raises vacant land prices by 1.6 percent") and independently from the abstract of Cunningham's own Syracuse PhD dissertation, Uncertainty, Zoning and Land Development (2005), whose first chapter is this paper. The precise regression specification, sample size/estimation period, and standard errors/significance levels remain in the paywalled JUE article and the ProQuest-gated full dissertation, neither of which yielded a readable full text this pass (no open-access copy exists via Unpaywall/RePEc; the Atlanta Fed working-paper guess returned 404); a future editor with full-text or ProQuest access should confirm the underlying tables.

Relation to the Georgist Case

Cunningham (2006) is not a study of land value taxation, Georgism, or any tax policy — it does not mention LVT and does not model a tax intervention. Its relevance to the Georgist case is indirect but structurally important: it is empirical, microdata confirmation of the mechanism that Georgist arguments about land speculation depend on. The paper establishes that:

1. Developers and landowners do, in practice, price and act on the option value of not developing — speculative waiting is a real, measurable phenomenon in an actual land market, not merely a theoretical possibility.
2. That option value rises with uncertainty and is capitalized into the price of vacant land itself, meaning part of the price of unimproved land reflects expected future appreciation and the right to wait, not just its value in current use.

Georgist arguments for land value taxation commonly assert that a recurring, annual tax on land value reduces the option value of holding land idle — because it converts the "free" wait into a costly one, shrinking (though not eliminating) the incentive represented by Cunningham's option-to-wait mechanism. Cunningham's paper is evidence for the mechanism a land value tax is designed to counteract, not evidence that a land value tax achieves that counteraction — the paper contains no tax variation and cannot speak to how large an LVT-induced reduction in speculative waiting would be. Readers should not read this paper as showing that LVT reduces speculation; it shows only that the speculative-waiting behaviour LVT proponents target is empirically real and non-trivial in magnitude (an 11 percent swing in development probability from one standard deviation of uncertainty).

Nuances and Limits

• Single metro area, one time period. The results are drawn from Seattle-area parcel data for a specific historical window; the magnitude of the option effect (1.6 percent price premium, 11 percent development-probability reduction) is local and time-bound, not a universal constant, and may not generalize to other markets, land types (e.g., agricultural vs. urban infill), or price-cycle regimes.
• Uncertainty, not taxation, is the treatment variable. The paper's identification strategy varies house-price uncertainty, not any tax rate. Any inference about how an LVT would change these results is an extrapolation from theory (a holding-cost tax should reduce the net value of waiting), not something the paper tests.
• Real options theory predicts a direction, not a policy-relevant magnitude. Even taking the paper's estimates at face value, translating "uncertainty raises the option value of waiting by X" into "an LVT of rate Y would reduce speculative withholding by Z" requires additional modeling assumptions the paper does not supply.
• Full-text verification gap. The paper's abstract and headline findings are confirmed across the ScienceDirect abstract, the author's Syracuse dissertation abstract, and multiple independent citing sources; the paywalled full text (regression specification, exact sample size, standard errors) was not obtainable this pass. Primary-text confirmation of the full regression tables and sample description remains available only in the paywalled JUE article and the ProQuest-gated Syracuse dissertation (2005), which a future editor with access should consult.

Bears On

• Outcome: Land value taxation dampens land speculation — Cunningham provides empirical confirmation of the underlying speculative-holding mechanism (option value of waiting rises with uncertainty and is priced into vacant land) that an LVT is theorized to blunt by taxing land value annually regardless of use; it does not itself test or measure an LVT's effect.
• Concept: Speculative Vacancy — the paper's finding that developers rationally delay development to preserve option value is a formal, priced version of the same withholding behaviour discussed on that page.
• Concept: 18-Year Land Cycle — uncertainty-driven option value and delayed development are consistent with (though not a direct test of) the speculative dynamics described in land-cycle accounts.

See Also

• Land value taxation dampens land speculation
• Speculative Vacancy
• 18-Year Land Cycle
• Land Value Tax

Sources

1. Christopher R. Cunningham (2006), "House Price Uncertainty, Timing of Development, and Vacant Land Prices: Evidence for Real Options in Seattle," Journal of Urban Economics, 59(1), pp. 1–31. DOI: 10.1016/j.jue.2005.08.003 — used for the paper's authorship, venue, and headline empirical findings (1.6% price effect, 11% development-probability effect). The publisher's abstract was directly fetched and quoted verbatim from the ScienceDirect article page; the full text behind the paywall was not obtainable — no open-access copy exists via Unpaywall or Semantic Scholar, no working-paper version is indexed on RePEc (the Atlanta Fed WP guess 404s), and the author's own site links only to the paywalled publisher page.
2. Christopher R. Cunningham (2005), Uncertainty, Zoning and Land Development, PhD dissertation, Syracuse University (advisor Stuart Rosenthal). Syracuse SURFACE listing — the dissertation's first chapter is this paper; its abstract independently corroborates the two headline findings. Full dissertation text is ProQuest-gated and was not readable this pass.

Sourcing note: the paper's existence, citation, and headline findings are verified via the ScienceDirect abstract and the dissertation abstract, but the full text (regression tables, sample construction, exact estimation period) was not fetchable through available web access this pass; it remains in the paywalled JUE article and the ProQuest-gated Syracuse dissertation, which a future editor with access should consult.