Buettner (2003): Tiebout Visits Germany — Land Tax Capitalization in a Sample of German Municipalities

Overview

"Tiebout Visits Germany: Land Tax Capitalization in a Sample of German Municipalities" is a paper by economist Thiess Büttner (Centre for European Economic Research, ZEW, and Mannheim University; rendered "Buettner" in some citations), presented at the Third Norwegian-German Seminar on Public Economics, CESifo Conference Centre, Munich, 20–21 June 2003 (JEL H73, R21, C21).[1] The paper applies a Tiebout-style test to Germany's local land tax (Grundsteuer B), which municipalities set via a local multiplier (Hebesatz) on a national base rate, asking whether cross-municipal differences in the effective tax rate capitalize into land prices — as standard tax-capitalization theory predicts — and whether any of that burden instead shows up in rents paid by tenants.[1] The full conference-paper text was fetched and read in full this session (2026-07-18), superseding the earlier abstract-only version of this page.

Data and Method

The sample is the 675 municipalities of Baden-Württemberg (a major German state) with population above 1,000 and available 1987 land-transaction data, drawn from a base dataset of 1,111 municipalities in the state.[1] The land tax rate is defined by the Hebesatz multiplier on a basic land-tax rate of 0.35% (0.50% in 1964); across the sample the statutory rate averages 0.863%, ranging from 0.648% to 1.295%, while the effective rate averages 0.113%, ranging from 0.027% to 0.583%.[1] Land value averages 164.7 DM/m² (range 16.45–936.9) and monthly gross rent (including water, sewerage, and tax side-costs, but excluding heating) averages 6.29 DM/m² (range 4.12–9.42).[1] Because the current statutory tax rate is endogenous to unobserved local conditions, Büttner instruments for the effective tax rate using the current and 1961 statutory rates in a GMM specification, and adds a spatial-lag (spillover) structure for amenities across neighboring municipalities.[1]

Core Findings

• Land values fall with the local land-tax rate — more than full capitalization alone would predict. The GMM specification finds the land-value elasticity with respect to the effective land-tax rate is −0.314 (significant at 5%, N=675, R²=0.803); the simpler OLS specification on the statutory rate gives a similar but statistically insignificant −0.329.[1] Büttner works out what "full capitalization" alone would imply: at the mean effective tax rate of 0.113% and a 3% discount rate, the tax should be about 3.6% of the cost of holding land — a far smaller number than the estimated 31% elasticity, which Büttner reads as evidence of "substantial overcapitalization."[1] He offers two candidate explanations: (1) measurement error in the effective tax rate from Germany's "disappearing tax base" problem (uncertain assessment ratios, since land was last officially assessed in 1964); and (2) sample land prices rose roughly 9.6% annually between 1964 and 1987, so if landowners anticipate continued appreciation, the true cost of holding land net of expected appreciation is lower than the tax-only calculation assumes, mechanically amplifying the apparent capitalization response.[1]
• No rent effect — and the rent variable is defined in a way that makes this a stronger result than it first looks. Re-running the same specification with monthly rent as the dependent variable, the coefficient on the land tax rate is small and statistically insignificant under both OLS (0.005) and GMM (−0.002), against R²=0.791.[1] Büttner flags why this is notable in the German institutional context specifically: the land tax is legally listed as part of the Nebenkosten (side costs) that landlords routinely pass through to tenants and that are not subject to German rent-control limits — and the rent variable used is the gross rent, which explicitly includes those side costs.[1] Finding no relationship under a measure that would have captured a pass-through if one existed is a more demanding test than it would be under a narrower rent definition.
• Amenity spillovers, not the land tax, drive most of the cross-municipal variation in both land values and rents. The spatial-lag terms for public swimming pools, tennis courts, theaters, and highway access are significant in both the land-value and rent regressions, "pointing to significant spillovers from amenities and the provision of public goods across municipalities" — a secondary finding of the paper, orthogonal to the tax-capitalization result but part of its headline claim.[1]

Limits and Caveats

• Single-state, 1987 cross-section. The sample is Baden-Württemberg only, not a national German dataset, and the data are a single year (1987), with a fixed-tax rate history back to 1961 used only as an instrument. Generalizing the exact elasticity to other German states or later periods is not warranted by this design alone.
• The overcapitalization result is a puzzle the author himself flags, not a clean finding. An elasticity roughly nine times what simple full capitalization predicts is a strength for the qualitative claim (the tax capitalizes, per the significant GMM coefficient) but a complication for using this paper's coefficient as a quantitative capitalization-rate estimate; Büttner's own candidate explanations (assessment-ratio measurement error; anticipated appreciation) are plausible but not separately tested in the paper.
• A conference paper, not (as far as this session could confirm) a journal publication. This should not be confused with Büttner's separate, journal-published 2003 paper "Tax base effects and fiscal externalities of local capital taxation" (Journal of Urban Economics 54(1): 110–128), which concerns local capital taxation, not this land-tax-capitalization study.
• A partial-equilibrium, not general-equilibrium, exercise. Like the wiki's other reduced-form capitalization studies, this is a hedonic-style cross-section, not a model of the whole local economy; see the CGE cluster and specifically Choi & Sjoquist (2015) for the general-equilibrium analogue of the same no-rent-pass-through result.

See Also

• Landlords Cannot Pass a Land Value Tax on to Tenants
• Tax Capitalization
• Borge & Rattsø (2014), Norway capitalization · Choi & Sjoquist (2015), Atlanta CGE — the other members of this cluster
• Land Is a Big Deal
• Mieszkowski: The Property Tax — An Excise Tax or a Profits Tax?
• Objection: Land value can't be assessed accurately

Sources

1. Thiess Büttner (2003), "Tiebout Visits Germany: Land Tax Capitalization in a Sample of German Municipalities," presented at the Third Norwegian-German Seminar on Public Economics, CESifo Conference Centre, Munich, 20–21 June 2003. Free full-text PDF (conference-paper scan, gwern archive): gwern.net/doc/economics/georgism/2003-buettner.pdf — fetched and read in full (2026-07-18), superseding the earlier abstract-only citation to economicpossibility.org — used for every finding on this page: the 675-municipality Baden-Württemberg sample and descriptive statistics (Table 1), the land-value and rent regressions (Tables 2–3), the overcapitalization calculation and its two candidate explanations, and the Nebenkosten/rent-control institutional detail (footnote 5 of the original).
2. Economic Possibility, "In German municipalities, land value tax does not increase monthly rent levels" — a related insight summary corroborating the no-rent-effect finding. economicpossibility.org — used to cross-check the rent-incidence finding.
3. Lars Doucet, Land Is a Big Deal (2022), Ch. 21 — used for the discovery context and for listing this paper among the sources supporting full capitalization of property/land taxes. See this wiki's book page: Land Is a Big Deal.