The Size and Scope of Commercial Property Tax Misvaluation (Howard & Wong, 2026)

Summary

Troup Howard (University of Utah) and Francis Wong (LMU Munich), "The Size and Scope of Commercial Property Tax Misvaluation," working paper dated 8 October 2026 (SSRN 7585598), is a national, parcel-level measurement of how accurately local assessors value commercial property, far broader than earlier studies, which covered one county or a handful. Compared with each parcel's later sale price, commercial assessments are much more dispersed than residential ones in almost every county examined, and the gap holds within a county, where assessor skill and local capacity are common to both classes.

Howard is a co-author of the residential counterpart, The Assessment Gap and the Black-White Wealth Gap. The commercial paper is a descriptive working paper; it has not been peer reviewed. It values whole properties, land and buildings together, and says nothing about site-value taxation.

Data and Measure

The authors use BlackKnight tax-roll and deed records for arm's-length sales in 2017-2019: about 121,000 commercial parcels and 3,995,000 single-family and townhome parcels in 900 counties. Counties enter only with at least 25 commercial and 100 residential sales, and counts are rounded for Census disclosure. Multifamily, agricultural, vacant-coded, exempt and partial-interest sales are excluded. A 2020-2023 sample serves as an out-of-sample check.

The core measure is the normalized assessment ratio: assessed value divided by sale price, then divided by the median ratio for that county and property class. A value of 1.1 means a parcel is valued 10 percent above its county-class norm. Because the median is divided out, a uniform over- or under-assessment of a whole class is invisible by construction; the authors therefore compare effective tax rates as a check. A county-level relative-error statistic divides aggregate commercial error by aggregate residential error, each scaled to its segment.

Findings

• Dispersion. The interquartile range of the ratio is 0.77-1.53 for commercial parcels against 0.94-1.34 for residential (p. 7): one quarter of commercial parcels are assessed more than 23 percent below sale price and one quarter more than 53 percent above. A commercial parcel is "more than two-and-a-half times more likely" than a residential one to be assessed at five times its sale price or more (p. 7).
• County comparison. In only 80 of 900 counties is aggregate error lower for commercial than for residential parcels; in 60 it is more than three times as high. The median county has a relative error of 1.67, about two-thirds more error for commercial parcels (p. 9). The 2020-2023 sample gives a median of 1.65.
• Regressivity. Lower-valued commercial properties "exhibit systematically higher assessments" (p. 2). The pattern is sharpest for the cheapest parcels, though a regressive trend runs across higher values too (p. 10, Fig. 3).
• Neighborhoods. Unlike the residential finding, the median ratio is little related to a tract's non-white share or income. Extreme over-assessment is larger in tracts with the highest non-white share and the lowest incomes (pp. 10-11).
• Owner influence. In a subsample of about 20,500 parcels matched to Census establishment records, parcels occupied by larger firms, multi-establishment firms or law firms are not valued lower; law-firm parcels are valued higher. Errors are also right-skewed, with more over- than under-assessment. The authors conclude that owner influence is "not the first-order explanation" for the national gap (p. 15) and point instead to the difficulty of valuing thin-market property, where assessors lean more on income capitalization.
• Burden. Misvaluation is "symmetric enough that it largely washes out of class-level medians" (p. 14). The cost is mainly unequal treatment among owners, not a systematic shift between classes. The pooled median effective rate is 1.54 percent for commercial and 1.16 percent for residential, which the authors attribute chiefly to homestead exemptions and geographic mix.

The authors read the results as inconsistent with the uniformity standards in the IAAO ratio-study standard and suggest that more assessment capacity may be needed (p. 16).

Relation to the Georgist Case

The paper never separates land from improvements, mentions no land value tax, and excludes vacant-coded parcels. It is not evidence for or against taxing land, and it should not be cited as either.

It bears on three nearby claims. First, mass appraisal and the IAAO standards treat assessment error as measurable against ratio-study benchmarks. The paper confirms that it is measurable; for commercial parcels the measured error is large. Second, the objection that land cannot be assessed points to thin markets as the hard case, and commercial property is where sales are scarcest. The paper's evidence concerns whole-property values in thin markets and does not test land values, so it neither confirms nor refutes the reply that land is easier to value. Third, the finding that appeals and capture get little support as the national driver of error is relevant to the corruption worry in assessment corruption, within the limits of this one measure.

As theory, not as a finding of the paper: a land-only base would change what has to be valued, namely land, which has comparables across uses, but would not remove the thin-market problem for unique parcels. The sister study on commercial property taxes and business location also measures a combined land-and-building tax.

Nuances and Limits

• Working paper. Unreviewed and descriptive, with no causal identification; the numbers may change in revision.
• Disclosure noise. Many quantities are pseudo-quantiles with noise injected for Census disclosure, and counts are rounded. The results use confidential microdata, so outsiders cannot replicate them.
• Sales-selected sample. Only parcels that sold in 2017-2019 are observed; parcels that did not sell are not observed.
• Normalization. Uniform class-level over- or under-assessment is hidden by design; effective-rate comparisons serve as the check.
• Atypical sales. Impaired properties, data errors and partial interests can distort ratios. The authors argue these cannot explain a commercial-specific excess but cannot rule them out parcel by parcel.
• Firm analysis. The occupant results rest on about 17 percent of the commercial sample (20,500 of 121,000), matched by exact address, and assume some of the tax is passed to tenants; no pass-through is estimated.
• Coverage. No list of the 900 counties or their states is given. The reported 66 and 67 percent figures are rounding of the same 1.666 median.

See Also

• Mass Appraisal Methods
• IAAO Standards: Ratio Studies
• Land Cannot Be Assessed
• Assessment Corruption Objection
• The Assessment Gap and the Black-White Wealth Gap
• Commercial Property Taxes and Business Microlocation

Sources

1. Troup Howard & Francis Wong (2026), "The Size and Scope of Commercial Property Tax Misvaluation," working paper dated 8 October 2026, SSRN 7585598. papers.ssrn.com; author-hosted PDF: trouphoward.com — used for the data, measure, findings and limits above; page numbers are the printed ones (B-claim; working paper, not peer-reviewed; read in full from the author's copy).
2. Carlos Avenancio-León & Troup Howard (2022), "The Assessment Gap: Racial Inequalities in Property Taxation," Quarterly Journal of Economics 137(3): 1383-1434, DOI 10.1093/qje/qjac009. doi.org — used for the residential counterpart as the paper describes it (A-claim for the fact of publication; cited as reported).
3. Gary C. Cornia & Barrett A. Slade (2006), "Horizontal Inequity in the Property Taxation of Apartment, Industrial, Office, and Retail Properties," National Tax Journal 59(1): 33-55 — used for the prior largest commercial study (about 5,000 parcels in one county) as the paper describes it (cited as reported).
4. International Association of Assessing Officers (2013), Standard on Ratio Studies. iaao.org — used for the uniformity standards the paper benchmarks against (A-claim; standards body).