Falkenbach et al. (2026): Information Capitalization in the Housing Market — Evidence from Helsinki Land Leases

Summary

Heidi Falkenbach, Oskari Harjunen, Erik Mäkelä and Elias Oikarinen, "Information Capitalization in the Housing Market: Evidence from Land Leases," Real Estate Economics (2026; open access under a Creative Commons Attribution licence; DOI 10.1111/1540-6229.70073), study what happens to home prices when a public landowner announces a large, scheduled rise in ground rent. In Helsinki, much of the housing built in the 1950s and 1960s stands on land the city owns and leases to condominiums. The old lease rents were indexed to the cost of living and had become "negligible in relation to the prevailing market rents" (p. 4). In January 2008 the city announced that renewed leases would be priced at roughly 4% of independently appraised land value, "a multifold increase in land rents" for sites due for renewal (p. 5).

The authors compare flats on leasehold land that would be renewed in 2021 with similar flats on freehold land in the same neighbourhoods, before and after the announcement. Their headline result is that capitalization of the higher rent into leasehold prices did occur and ended up roughly consistent with a plausible discount rate, but it was slow. In their words, the discount increase "is substantially slower than predicted by the present value model, implying only partial adjustment over the short run, but is in line with full capitalization eventually" (abstract, p. 1).

The paper concerns a contract rent charged by a city as landowner, not a tax. It is nonetheless a rare causal test of how a large change in the recurring charge on land is absorbed into prices, and of how much time that absorption can take.

The Core Argument and Findings

The setting. The 2008 announcement fixed the renewal rent at "approximately 4% of the land value appraised by independent experts," aligned with the policy already applied to new development sites (p. 5). After political pressure and threats of lawsuits, the city phased the increase in: half of it in the first year of the renewed contract, then a further 5% of the total each year, reaching the full level ten years after renewal. Renewed contracts run 50 years, with annual adjustment by the cost-of-living index (p. 5). The authors report that the expected increase raised operating costs of affected dwellings by over 40% (p. 2); for the sample, the expected rise in maintenance costs averaged €1,130 a year, ranging from €375 to €3,693 (p. 10).

The analysis uses the batch of more than 700 lots renewed in 2021 (p. 5). Two earlier batches (2011 and 2016) were too small and dispersed for a reliable comparison; they matter mainly because the city enforced the policy on them (p. 5, note 1; p. 13). Media coverage of the coming increases was, per the authors' newspaper count, essentially absent before January 2008 (p. 5, Figure A1), which is why they treat the announcement as a genuine information shock.

The design. A difference-in-differences hedonic regression compares 4,545 leasehold and 6,439 freehold condominium transactions (10,984 in all), 2003–2021, restricted to buildings from 1957–1963 within the same neighbourhoods (pp. 9–10). Freehold units serve as the control group. Since property rights and lease law did not change over the period, permanent level differences between the two tenures are absorbed by the pre-period comparison (p. 9).

Result 1: no discount, and no immediate reaction. Before 2008 the leasehold and freehold price paths are statistically indistinguishable; the 2007 baseline difference is about −0.9 log points and insignificant (p. 13, Figure 3 note). The authors read this as buyers having been inattentive to the coming rents, which "resulted in significant overpricing of leaseholds relative to corresponding freeholds" (p. 3). After the announcement there was no immediate effect. For comparison, the authors compute that if buyers had known nothing before 2008 and then fully priced the announced policy, a 3% real discount rate would have implied a 2008 effect of −14.5% (p. 12). The estimated 2008 coefficient is −0.009, indistinguishable from zero (p. 26, Table A1).

Result 2: a slow, then large, discount. The pooled estimates (Table 2, p. 14; treatment-dummy specification) are:

The continuous-treatment specification gives −0.053 and −0.095, or about −9.1% (−€24,422) per €1,000 of annual rent increase (p. 14). The yearly path shows discounts that "remain relatively small through the early 2010s and become consistently large and significant from 2015" (p. 13); for example the 2015 coefficient is −0.078 and the 2020 coefficient −0.117 (p. 26, Table A1). The authors find no significant difference between 2020 and 2021, when the new rents took effect, and read this as the information having been capitalized before the higher rents were actually paid (p. 13).

Result 3: implied discount rates. If the 2017–21 discount is treated as full capitalization, it implies a real discount rate of about 2.5% over a 50-year horizon (2.45% and 2.48% in the two specifications) and about 3.7% in perpetuity (3.68% and 3.70%), assuming real land values rise 0.7% a year (p. 15, Table 3). The 95% confidence intervals are wide: 0.69–5.49% for the 50-year figure and 2.67–5.97% in perpetuity (dummy specification). The authors judge these rates in line with earlier estimates for long-lease contracts in the UK, Singapore and Amsterdam (p. 15).

Result 4: why so slow? The authors test three explanations.

• Sluggish sellers. If sellers were slow to cut their asking prices, leasehold sales volume should have fallen and time on market lengthened. They find no trend in the leasehold share of sales and no trend in selling times, and conclude this explanation is unlikely to account for the pattern (p. 19).
• Dispersion. Leasehold price dispersion rose by about 15% in 2008–2016, stayed at a similar level in 2017–2020 (with a larger standard error), and returned to the pre-announcement level in 2021 (p. 19, Table 5, p. 20).
• A segmented market. Classifying each leasehold sale as "noncapitalized," "in-between" or "capitalized" against a benchmark discount rate of 3.7%, the authors find the noncapitalized share falling "from the level of 67%–72% in 2003–2006 to about 30% in 2018–2021," while the capitalized share rose correspondingly; the in-between group averaged 13% (p. 21). The average discount, on this reading, blends buyers who price the announced rent and buyers who do not, with the first group growing as information accumulated.

The authors' summary of the process: "capitalization is neither perfect nor absent but rather a gradual and slow process" (p. 22).

Why the authors think information was the issue. Land rent was bundled into monthly maintenance fees paid to the condominium, and before renewal old rents were negligible, so leasehold and freehold flats looked alike in observed fees (p. 7). Pricing the change required buyers to forecast land values and rent adjustments and discount them, a task the paper calls a "significant cognitive burden" (p. 7). The authors also consider doubt that the city would enforce the policy, but note that enforcement on the 2011 and 2016 batches makes this an unlikely explanation for the whole period (p. 13).

Relation to the Georgist Case

Capitalization of a recurring charge on land. Land value tax theory rests on the present-value logic set out on the Tax Capitalization page: a recurring charge on land value is reflected in a lower land price. The authors themselves connect the two, writing that "land leases by local authorities yield monetary costs similar to property or other taxes" (p. 4). Their estimated discount, roughly a tenth of the price of a flat, and the implied discount rates in a normal range are consistent with a large increase in a recurring land charge being priced in by the end of the sample. This is a leasehold result, not a tax result, but it adds a Nordic public-landowner case alongside property-tax studies such as Høj, Jørgensen and Schou on Danish land taxes.

Timing. The most useful feature for the capitalization argument is the time profile. Standard theory has the price adjust at announcement; here it did not, for years. That fits the caveat on the Tax Capitalization page that the size and speed of price effects depend on how well buyers understand and believe the change. It also has an implication that the paper itself only gestures at: if prices lag the announcement, the burden falls less on whoever owned when the change was announced and more on buyers who paid pre-discount prices in the interim. The authors flag possible distributional consequences if less financially literate buyers systematically fail to price future rent obligations (pp. 22–23). That distributional point is the authors'; its bearing on the transition wealth-shock objection is an interpretation, not a finding of the paper.

Public land leasing and rent capture. The case shows a public landowner moving contract rents that had fallen far below market levels toward a rent policy used for new sites, and collecting on renewal. That is the repricing-at-renewal step that the Public Land Leasing page identifies as the weak point of some leasing systems. It also illustrates the gap between contract rent and economic ground rent: cost-of-living-indexed contract rents had become negligible because market land rents in Helsinki "have increased substantially faster than the CLI" (p. 4).

Nuances and Limits

• A lease, not a tax. The rent is set by contract and by a city acting as landowner. The lessee has a statutory right to renew but the law prescribes no maximum rent (p. 5). A statutory land tax would differ in legal form, in political process and possibly in how salient it is to buyers. The paper's own literature review notes that buyers appear to internalize "direct and highly salient costs, such as property taxes" (p. 2), whereas here the charge was bundled into a maintenance fee; how a salient, separately billed land tax would behave is not tested.
• Capitalization and discount rate are not separately identified. The authors state that "the capitalization rate can be identified only if we know the discount rate, and vice versa" (p. 15). The claim of "full capitalization" by 2017–21 is therefore an inference that the implied rate looks reasonable, not a direct measurement. The confidence intervals on implied rates are wide.
• The segmentation story is suggestive. The classification of transactions assumes that every deviation from the predicted price is due to the leasehold (p. 20, note 19), and the authors say the sources of segmentation (financial literacy, attention, prior experience, credibility of enforcement) cannot be separated with their data (p. 22). Dispersion results are also measured with less precision in 2017–2020 (p. 20, Table 5).
• One city, one batch, one housing type. The estimates come from 1957–63 condominiums in Helsinki neighbourhoods and the 2021 renewal batch. Finnish property taxes are low, and the authors report robustness checks on the small differences in land taxation and mortgage-interest deductibility between freehold and leasehold units (pp. 6, 17). Effects elsewhere, or for other kinds of dwelling, are not shown.
• The phase-in is part of what was priced. The announced increase was staged over a decade after renewal and the final policy stayed politically uncertain until the first renewals (p. 5). Part of the initial muted reaction could reflect that uncertainty, although the authors argue enforcement on earlier batches makes it a weak explanation for the whole pattern (p. 13).
• Estimates vary by specification. A specification allowing all housing attributes to have different prices for leasehold units gives a smaller post-2017 discount, −0.077 versus −0.111, still significant at the 1% level (p. 17, note 17).

Bears On

• The timing and anticipation caveat in Tax Capitalization: direct evidence that a large increase in a recurring land charge was capitalized only gradually.
• Public Land Leasing: a working case of repricing at renewal, with the market-side response measured.
• Ground Rent: contract rent falling away from market rent under indexation, then being reset.
• The transition wealth-shock objection: slow capitalization changes who bears a charge in the years after announcement, though this case does not test an LVT transition.

See Also

• Tax Capitalization: the present-value mechanism this paper tests in a leasehold setting
• Public Land Leasing: the leasing model and its rent-renewal problem
• Ground Rent: economic versus contract ground rent
• Høj, Jørgensen and Schou (2017): a Danish land-tax natural experiment finding full capitalization
• Hilber (2017): a synthesis of the house-price capitalization literature, which this paper cites
• Lyytikäinen (2009): a separate Finnish natural experiment on land-favouring property taxation
• Land Value Capture: the wider family of instruments

Sources

1. Heidi Falkenbach, Oskari Harjunen, Erik Mäkelä and Elias Oikarinen, "Information Capitalization in the Housing Market: Evidence from Land Leases," Real Estate Economics (2026), pp. 1–37, DOI 10.1111/1540-6229.70073. https://doi.org/10.1111/1540-6229.70073 (open access, CC-BY; full text read; page numbers above are the article's own) — used for the Helsinki renewal policy (about 4% of appraised land value; 50% of the increase in the first year, then 5% of the total each year; 50-year renewed contracts), the 10,984-transaction difference-in-differences sample, the absence of a pre-2008 discount and of an immediate reaction, the estimated discounts of −0.052 (2008–16) and −0.111 (2017–21; about −10.5%, €28,331), the implied real discount rates of about 2.5% (50 years) and 3.7% (perpetuity), the price-dispersion result and the fall in the noncapitalized share from 67–72% to about 30%, and the authors' statement that capitalization and the discount rate are not separately identified (A-claim: peer-reviewed, open access, read in full).