Bourassa (1987): Land Value Taxation and New Housing Development in Pittsburgh

Summary

Steven C. Bourassa's "Land Value Taxation and New Housing Development in Pittsburgh," published in Growth and Change 18(4) in Fall 1987 (pp. 44-56), is the earliest econometric study of the housing effects of Pittsburgh's split-rate property tax — the same system later examined by Oates & Schwab (1997) and, across a wider panel of Pennsylvania cities, by Plassmann & Tideman (2000). Bourassa built an econometric model using building-permit data as the dependent variable and local tax rates, alongside other determinants of housing supply and demand, as independent variables, testing separately for two theorized channels: a "liquidity effect" (a lower up-front tax burden on new construction) and an "incentive effect" (a heavier ongoing tax on idle land pushing owners toward development).[1]

Key Finding

Bourassa found that for new housing, "the incentive effect is significant but the liquidity effect is not," and that the incentive effect showed up as an increase in the number of new housing units built in Pittsburgh rather than in their average cost.[1] The study used monthly data over 1978–1984 and estimated a log-linear model (so coefficients are elasticities). In it, the improvement tax rate is a highly significant determinant of new construction — a 1 percent cut in the improvement tax rate is associated with a 2.36 percent rise in the dollar value of new housing, and the improvement-tax elasticity of the number of new units is −2.79, both significant at the 1 percent level. Bourassa reports that the 5 percent improvement-tax cut effective at the start of 1983 implies "about an 11.8 percent increase in the dollar value of new housing construction," equivalently roughly a 14 percent rise in the number of units (about 4.5 additional units per month against a mean of ~32.7).[1] When the average cost of new units is the dependent variable, neither tax rate is significant — so the effect operates on unit count, not on cost per unit.[1]

Notably, the land tax rate coefficient is not itself a significant determinant of new construction in the regressions, which Bourassa attributes to serious multicollinearity with the other independent variables; the "incentive effect" he identifies runs through the improvement-tax side of the split-rate system (lower taxes on structures) rather than through a directly estimated land-tax push.[1] This is an early data point for the claim, later tested at larger scale, that split-rate taxation increases construction — though the direct pathway it measures is the improvement-tax reduction, not the land-tax increase.

Context

The 1987 article was based on Bourassa's University of Pennsylvania dissertation of the same title (1988), and Bourassa went on to become a repeat contributor to the land-value-tax literature, later writing the U.S.-experience and political-economy-of-LVT chapters of the Lincoln Institute of Land Policy's Land Value Taxation: Theory, Evidence, and Practice (2009).[2] Lars Doucet's Land is a Big Deal (2022) lists Bourassa (1987) among 13-plus studies it surveys as supporting the capitalization and development-incentive effects of land value taxation (Ch. 21).[3]

Bears On

• Outcome: Split-rate taxation increases urban construction

Limits and Caveats

The regression specification, 1978–1984 monthly sample, and effect sizes reported above were verified this session against the full text of the article (an open-access scan of the Growth and Change original is hosted at gwern.net). Two honest caveats on the study's own terms: (1) building-permit applications (deflated dollar value, and a derived unit count) are the sole available monthly proxy for new construction, so the dependent variable measures intended rather than completed building; and (2) because the land-tax-rate coefficient is statistically insignificant (multicollinearity), the paper does not directly isolate a land-tax development push — its positive result is carried by the improvement-tax elasticity.

See Also

• Oates & Schwab (1997): The Pittsburgh Experience — the more-cited later study of the same city and tax system
• Plassmann & Tideman (2000) — the multi-city Pennsylvania panel extending this line of evidence
• Pittsburgh, Pennsylvania — the place page for the split-rate system Bourassa studied
• Land is a Big Deal (book) — the discovery source citing this paper
• Split-rate taxation increases construction (outcome)

Sources

1. Steven C. Bourassa, "Land Value Taxation and New Housing Development in Pittsburgh," Growth and Change 18(4) (Fall 1987): 44-56 — used for the study's method (log-linear building-permit econometric model, 1978–1984 monthly data), the improvement-tax elasticities (−2.36 on dollar value, −2.79 on unit count, both significant at 1%), the insignificance of the land-tax-rate coefficient, and the finding that the effect falls on unit count rather than average cost. Full text read this session via the open-access scan at gwern.net; abstract also at the Library of Economic Possibility.
2. Richard F. Dye & Richard W. England (eds.), Land Value Taxation: Theory, Evidence, and Practice (Lincoln Institute of Land Policy, 2009) — used to confirm Bourassa's standing as a repeat contributor to the land-value-tax literature (author of the "U.S. Experience" and "The Political Economy of Land Value Taxation" chapters) and Pittsburgh's 1913-2001 split-rate history. Free PDF (front matter/introductory chapter)
3. Lars A. Doucet, Land is a Big Deal (Shack Simple Press, 2022), Ch. 21 — discovery source; lists Bourassa (1987) among the studies it surveys as supporting capitalization and construction-incentive effects of land value taxation. Book page