To Own or to Rent? The Effects of Transaction Taxes on Housing Markets (Han, Ngai & Sheedy, 2026)

Summary

"To own or to rent? The effects of transaction taxes on housing markets," by Lu Han (University of Wisconsin–Madison), L. Rachel Ngai (LSE, CEPR), and Kevin D. Sheedy (LSE), appeared as an advance article in the Review of Economic Studies (Vol. 93, Issue 4, p. 2605, DOI 10.1093/restud/rdaf092). It is not a study of land value tax — it studies a very different instrument, a transaction tax (a one-off charge on the sale of a property, the same family as UK stamp duty) — but it matters to the wiki because it supplies a rigorous, quantified account of exactly the kind of distortion land value tax is designed to avoid, using a real natural experiment rather than a theoretical model alone.

The Natural Experiment

The paper's empirical core is Toronto's introduction of a city-level Land Transfer Tax (LTT) in 2008, layered on top of Ontario's existing provincial transfer tax. Using MLS sales and leasing microdata for the Greater Toronto Area (2006–2018) and comparing neighborhoods just inside the City of Toronto boundary (where the new tax applied) against neighborhoods just outside it (where it did not), the authors find: a 1 percentage-point higher effective LTT rate causes buy-to-rent purchases to rise 9.3%, owner-occupier purchases to fall 9.6%, the leases-to-sales ratio to rise 26%, and the price-to-rent ratio to fall 3.8%.

The Mechanism: Why a Tax on Both Sides Still Tilts the Market

The paper's central puzzle is why a transaction tax that applies equally to owner-occupier purchases and investor purchases would still push the market from owning toward renting. The answer, in the authors' own words, is about transaction frequency, not transaction rate: "Choosing to be an owner-occupier rather than a renter means expecting to pay the LTT every time a new property is purchased... However, landlords do not need to transact again in the ownership market just because a tenant no longer finds their property suitable and moves out... investors have less need to transact compared to owner-occupiers who face match-quality shocks." Because owner-occupiers move (and re-buy, re-paying the tax) far more often over their lives than buy-to-rent investors turn over their holdings, the same nominal tax rate is a larger expected lifetime cost for ownership than for renting-via-landlord. That asymmetry pushes credit-constrained households out of ownership; the falling price-to-rent ratio then attracts enough buy-to-rent investors to absorb the resulting rental demand. A second, "intensive margin" effect compounds this: existing owners become more tolerant of a poor match (move less often) specifically to avoid repeatedly re-paying the tax.

In the calibrated model, four years after a tax increase: owner-occupier transactions fall 14%, buy-to-rent transactions rise 35%, the lease-to-sales ratio rises 15%, and the homeownership rate falls 0.23 percentage points (rising to −2.4pp in the long-run steady state). The deadweight loss is 111% of the extra revenue raised — more than the entire revenue take is lost to distortion — with 60% of that loss coming specifically from distorting the own-vs-rent decision (13% from within-ownership-market distortions, 38% from within-rental-market distortions).

Relation to the Georgist Case

The paper is explicit, in its own literature review, about where this leaves transaction taxes relative to land value tax: it cites both the UK's Mirrlees Review and Australia's Henry Review as concluding that transaction taxes like stamp duty should be replaced by "a land value tax or a tax on housing consumption" precisely because of the mobility-lock-in and tenure-distorting effects this paper now measures directly. In other words, this is not a paper critical of LVT — it is a rigorous empirical demonstration of the specific failure mode (tax capitalization research shows LVT avoids by falling on an inelastically-supplied asset rather than on the act of trading it). A one-off transaction tax taxes a behavior (moving, buying, selling) and so distorts that behavior; a recurring tax on unimproved land value taxes a stock that cannot be avoided by transacting less, which is exactly why the mobility-lock-in mechanism documented here has no close analogue under LVT.

Nuances and Limits

• This is not an LVT study. The paper says nothing directly about recurring land value taxation's own incidence or efficiency; its relevance to the wiki is as a rigorous quantification of the transaction-tax alternative's costs, cited as a foil.
• Toronto-specific institutional setting. The LTT is layered on Ontario's existing provincial transfer tax, and Toronto's rental/ownership market structure (condo-heavy, investor-heavy) may not generalize cleanly to other cities' stamp-duty regimes.
• Full text read (A-claim). The working-paper PDF (Kevin Sheedy's LSE page, dated August 2025) was read directly, including the abstract, natural-experiment results, model mechanism, and welfare-cost breakdown quoted above.
• One author (Lu Han) conducted the underlying Toronto MLS data work while previously at the University of Toronto's Rotman School; her current affiliation on the August 2025 draft is University of Wisconsin–Madison.

Bears On

• Concept: Tax Capitalization — this paper's transaction-tax findings are the mirror image of the capitalization literature's central claim: a stock tax (LVT) capitalizes into price without much behavioral distortion, while a transaction tax visibly distorts the decision to trade at all.
• Research: Chu: Taiwan Property Tax Incidence — another paper directly contrasting a recurring property/land tax against a one-off transfer tax in a structural model, useful side-by-side reading.
• Concept: Land Value Tax — the instrument the paper's own cited reviews (Mirrlees, Henry) recommend in place of the transaction tax it studies.

See Also

• Land Value Tax
• Tax Capitalization
• Chu: Taiwan Property Tax Incidence
• Tubío-Sánchez & Reyes-Bueno: Ecuador LVT Capitalization

Sources

1. Lu Han, L. Rachel Ngai & Kevin D. Sheedy (2026), "To Own or to Rent? The Effects of Transaction Taxes on Housing Markets," Review of Economic Studies 93(4): 2605, DOI 10.1093/restud/rdaf092. doi.org — abstract confirmed against the published Oxford Academic listing; full working-paper PDF read directly at personal.lse.ac.uk/sheedy/papers/ToOwnOrToRent.pdf (dated August 2025) 2026-08-28 — used for the Toronto LTT natural-experiment estimates, the transaction-frequency mechanism (verbatim quotations above), the calibrated-model results, the 111%-of-revenue deadweight-loss figure and its decomposition, and the paper's own citation of the Mirrlees and Henry Reviews recommending LVT over transaction taxes (A-claim).