The Shifting of the Property Tax on Urban Renters: Evidence from New York State's Homestead Tax Option

Summary

"The Shifting of the Property Tax on Urban Renters" by David J. Schwegman and John Yinger (US Census Bureau Center for Economic Studies Working Paper CES-WP-20-43, 2020) is the modern US quasi-experimental counterpart to the German Löffler & Siegloch study — and it points the other way. Exploiting New York State's 1981 Homestead Tax Option (which lets municipalities tax larger rental buildings at higher rates than owner-occupied homes) with within-unit variation in restricted American Housing Survey data, the authors find landlords shift only about 14% of a property-tax increase onto renters — roughly six-sevenths of the burden stays with the owner.[1]

Yinger is the same empiricist behind Carroll & Yinger (1994), whose Boston-area estimate (~15 cents of rent per tax dollar) this modern design essentially reproduces a generation later with stronger identification.

Relation to the Outcome

Wired as supporting evidence on landlords cannot pass LVT to tenants: in a US institutional setting, even a tax aimed disproportionately at rental buildings — the component theory says can shift — mostly failed to reach tenants. Together with the German full-pass-through result, it shows property-tax rental incidence is institution-dependent; the land component's incidence (capitalization into owner prices) is the part the theory fixes.

Verification note

Direct fetches proxy-blocked; details corroborated via the Census Bureau listing and RePEc; finding magnitude via one corroborated rendering — treat the 14% as approximate until the PDF is read. Scan depth Light.

Bears On

• Outcome: Landlords cannot pass LVT to tenants
• Research: Carroll & Yinger (1994) · Löffler & Siegloch · Tsoodle & Turner

See Also

• Tax Capitalization

Sources

1. David J. Schwegman & John Yinger, "The Shifting of the Property Tax on Urban Renters: Evidence from New York State's Homestead Tax Option," US Census Bureau CES Working Paper 20-43, 2020. Census listing · RePEc — used for design and the ~14% shifting estimate (B-claim; magnitude corroborated via one rendering — [VERIFY against the PDF]).