Do Commercial Property Taxes Affect Business Microlocation Decisions? (Bulloch, Gaertner & Hoopes, 2026)

Summary

Brayden D. Bulloch and Fabio B. Gaertner (University of Wisconsin-Madison) and Jeffrey L. Hoopes (University of North Carolina at Chapel Hill), "Do Commercial Property Taxes Affect Business Microlocation Decisions?," The Accounting Review (published online 6 October 2026, pp. 1-31, DOI 10.2308/TAR-2025-0248), asks whether differences in commercial property taxes between neighboring counties change where businesses place their establishments. The study uses precise geolocations for 2.8 million US establishments, set against county-level effective commercial tax rates.

The answer in the published abstract is yes, at the margin. For-profit establishments, relative to nonprofits, "disproportionately locate on the low-tax side of county borders," and the effect grows with the size of the tax differential (abstract).

The tax measured is the combined rate on land and buildings. The paper does not mention land value or site-value taxation and does not test a land-only base.

Design and Data

The authors use what the abstract calls "a spatial discontinuity density design, comparing for-profit establishments to nonprofits, which are generally exempt from property taxes" (abstract). The county line is the discontinuity: the question is whether, near a border, for-profit establishments are denser on the lower-taxed side than a nonprofit benchmark would predict.

The authors' February 2024 working paper, "Borderline Tax Planning," an earlier draft whose estimates were revised down in the published version and are therefore not reported here, describes the data and design in more detail:

• Establishments. SafeGraph Places geolocations, restricted to within-state county borders and to establishments within 10 km of a border (2,805,455 establishments). Manufacturing and medical establishments are removed because they are more likely to receive abatements and to be tied to non-tax siting factors such as rail or ports (working paper, pp. 9, 11).
• Tax measure. CoreLogic commercial property records for 2015-2021. The effective rate is "dividing property taxes paid by market value for the commercial properties in the dataset" (working paper, p. 10), taken as the county median. The mean county rate is 2.1 percent of market value and the mean border differential is 0.5 percentage points (working paper, p. 11).
• Counterfactual. 303,265 nonprofit establishments, such as churches, schools and public buildings, stand in for where for-profits would sit absent property-tax considerations (working paper, p. 14).
• Inference. Bootstrapped standard errors from 200 draws for the density estimates, and regressions with county fixed effects and standard errors clustered at the border level (working paper, pp. 16, 21).

Findings

The published abstract reports the following.

• Size of the shift. "At borders with differentials in the top decile (quartile), we estimate a 6-7 (5-6) percent shift in establishment location toward the low-tax side over a 10-km band" (abstract).
• Dose-response. The effect strengthens as the differential grows (abstract).
• Distance pattern. Businesses "move from as far as 10 km away on the high-tax side but concentrate within 1 km on the low-tax side" (abstract).
• Customer access. "Firms trade off tax savings with customer access": low-foot-traffic establishments respond more than high-foot-traffic ones (abstract).
• Perceptions. A survey of US mayors finds that perceived importance of property taxes for business location rises with cross-border tax differentials (abstract).

The working paper's larger headline figures were revised downward in the published version; the abstract's numbers above are the article's findings.

Relation to the Georgist Case

What the paper shows is that differences in the combined land-and-building effective rate across a county line move for-profit establishments toward the lower-taxed side. It is evidence that a recurrent tax on improved commercial property is not locationally neutral, which is the kind of behavioral response the economic literature treats as deadweight loss. It is not evidence for a land value tax, and it should not be cited as such.

The Georgist inference runs through theory, not through this paper. Land cannot cross the county line, while buildings and firms can, so a tax on site value alone should offer no gain from relocating: the differential would be expected to be absorbed in land values, the logic of tax capitalization, and the sorting logic is the Tiebout model. The paper does not test this. Its measure cannot separate a land component from a building component, and it observes establishment locations, not land prices or rents, so capitalization is not examined. The design also leaves untouched the within-jurisdiction margin where Georgists locate neutrality, namely how intensively a site is built.

For related evidence see the Carroll & Yinger and Yang split-rate pages. A policy reading, that a uniform national land base would remove cross-border rate differentials, is analysis that goes beyond what the authors state.

Nuances and Limits

• "Microlocation" is coarse. It means which side of a county line within 10 km, not parcel-level or within-city siting.
• Density, not movement. The design compares cross-sectional densities. The abstract's language of businesses moving describes an inference; relocation of existing firms is not observed.
• Imperfect counterfactual. Nonprofits are not uniformly exempt, and schools or churches have their own siting logic. The working paper concedes the counterfactual is not perfect (working paper, pp. 14-15).
• County medians. Effective rates are county-level medians of taxes paid over market value, not parcel-specific rates, and abatements are only partly captured (working paper, pp. 10-11).
• Scope. The sample is mostly retail-type establishments, excludes manufacturing and medical sites, and covers within-state borders only.

See Also

• Land Value Tax
• Deadweight Loss
• Tax Capitalization
• Tiebout Model
• Carroll & Yinger: Is the Property Tax a Benefit Tax?
• Yang: Split-Rate Taxation and Spillover Employment
• Split-Rate Taxation

Sources

1. Brayden D. Bulloch, Fabio B. Gaertner & Jeffrey L. Hoopes (2026), "Do Commercial Property Taxes Affect Business Microlocation Decisions?," The Accounting Review, published online 6 October 2026, pp. 1-31, DOI 10.2308/TAR-2025-0248. doi.org — used for the design and headline findings as stated in the published abstract (B-claim; peer-reviewed; the full text sits behind the publisher's wall, so locators are to the abstract).
2. Bulloch, Gaertner & Hoopes (2024), "Borderline Tax Planning," working paper, February 2024, 49 pp. aaahq.org — used for data and design detail only (D-claim for its estimates, superseded by the published version); page numbers are the printed ones.
3. Bulloch, Gaertner & Hoopes (2024), "Borderline Tax Planning," SSRN abstract 4882083, posted 10 July 2024. papers.ssrn.com — used as a pointer to the working-paper version only; page not accessible at last review (2026-10-09).
4. UNC Kenan Institute of Private Enterprise, "Borderline Tax Planning" (10 July 2024). kenaninstitute.unc.edu — used for the fact of the working paper's circulation (A-claim).