Rent control,
as the evidence sees it.
A 2026 group debate — Mietendeckel, empty homes, land value taxes — turned into a literature review. 1,641 papers collected from OpenAlex, 33 studies extracted, cases from four continents (corpus, method and collection date in ch. 08). Every claim has its reference; every reference has its verification grade: ✓ citation verified against the full text, ~ partially verified, ? to be verified.
Why rents rise
No mystery and no conspiracy: it is demand bearing down on cities where supply cannot keep up — because the neighborhood electorate does not let it.
The Hankinson cycle (APSR 2018): high rents turn homeowners — and, in expensive cities, renters too — into neighborhood-level NIMBYs; supply does not keep up; scarcity feeds back into rents. Valves in blue and amber: the policies of chapter 04. Editorial synthesis ? of evidence ✓.
17–73% of the rise in US rents (2000–18) comes from demand bearing down on cities with inelastic supply ✓ — Howard & Liebman 2021, JUE. The rigidity is associated with land-use regulation, and tighter rent controls are linked to weaker supply elasticities ✓ — OECD 2019. The political link is the neighborhood electorate: “institutions can undersupply even widely supported public goods” ✓ — Hankinson 2018, APSR.
The classic remedy: the cap
Rent control is the attempt to have low rents without touching property values. It works — and it charges the price in other currencies: supply, spillover, distribution.
Reported magnitudes, in per cent; each bar has its own metric (hover). They are not directly comparable across studies — the order of magnitude is.
Eight decades of interventions in the corpus: Mumbai 1947 → Toronto 1975 → SF 1994 → Cambridge 1995 → São Paulo 2014 → Mietpreisbremse 2015 → Auckland 2016 → Mietendeckel and Catalonia 2020 → St. Paul 2021. The post-2014 acceleration is the “renaissance” documented by Kholodilin's 2024 meta-review.
Harvest plot generated from the chapter 08 table and its provenance filters — a snapshot of the selected corpus, not the consensus of the literature.
What the cap does well: it lowers rents consistently in the short run — Catalonia −4–6%, Mietpreisbremse up to −5%, NYC $468/month ✓. What it charges: the effect dies out (~1 year in the Mietpreisbremse, Breidenbach 2022 ~); supply falls (−15% SF; −10% in the 27-city panel, JHE 2025 ✓); and it pushes rents up outside its scope — robust spillover (Mense et al. ✓; Early & Phelps 1999 ✓; buildable land appreciates, Mense 2019 ✓).
Who wins, who loses
The blanket cap is not progressive. It benefits those already inside — and the record tells us exactly who.
St. Paul 2021 ✓ (NBER WP 30083): property values −6–7% ($1.6 billion). Who gains most: tenants with higher incomes, more likely white. Who loses most: landlords with lower incomes, more likely minorities.
New York ✓ (Chen 2023): average discount $468/month (39%), larger in Manhattan and gentrification; immigrants receive $151/month less than natives (REE 2024 ✓).
The premium is one of tenure, not of entry: new tenants of the stabilized sector paid more than in the free sector (1981); only those who stay come out ahead (1987) — Ault & Saba ✓.
Honest counterpoint: for Arnott (JEP 1995), the case against modern soft controls is weak and calls for case-by-case evaluation ✓ — and in near-total-coverage Catalonia rents fell without a fall in supply ✓. Design matters more than ideology (Malpezzi & Mayo 1993 ✓; a century of control in Argentina with no measurable effects, 2022 ✓).
What works
If the problem is elasticity, the serious solutions change the elasticity, protect incomes without touching prices — or discipline the market through public supply.
Editorial synthesis of the confirmed magnitudes: effect on rents (horizontal, left = lower) vs. effect on future supply/elasticity (vertical, up = more supply). Bubble = number of verified supporting studies. Tooltips cite the studies.
Broad upzoning: São Paulo — more permits, more supply, falling prices in the densified neighborhoods (spatial RDD, AEJ 2021 ✓); Auckland reclassified ¾ of the inner city (JUE 2023 ~; Chicago UAR 2019 ✓). Counterpoint: income elasticity is low — marginal projects are not enough (Fed Board 2018 ✓); what counts is the aggregate regime, not the single lot.
Subsidies with design: generous vouchers are captured by landlords; indexed to the neighborhood, they move families to better neighborhoods without inflating rents (Collinson & Ganong 2018 ✓).
Taxing land by value: LVT does not distort the base and, with recycling, reduces the net burden on low/middle incomes (IMF 2022 ✓); the obstacle is cadastral (Hughes 2020 ✓). On vacancy: a measured problem — 340 thousand low-use homes in England & Wales (Bourne 2019 ✓) — but the causal evidence on the taxes is still thin ?; outright confiscation has no empirical literature to support it.
The case worth studying seriously is Vienna: a giant social housing stock disciplining private rents through competition, without a cap (Kadi 2015 ~ (bibliographic)).
Verdict
Can rents come down without destroying the market? Yes — but not with a single instrument, not instantly, and not without choosing who pays for the transition.
1 · The main lever is supply, not administered prices
“Efforts based on subsidies and control are undermined by planning decisions that restrict supply even in the face of growing demand” — Metcalf 2018, JEP ✓. Control depresses supply and deregulation appreciates the entire neighborhood (Autor 2014 ✓).
2 · Protecting low incomes: targeted subsidy, not a blanket cap
The cap gives the largest discount to those already inside, regressively (ch. 03). The recommendation of vouchers as the preferred policy is implicit in Metcalf — a reading of the abstract, not of the full text ~.
3 · Tax land by value, don't confiscate
The fiscal route has literature (IMF, Hughes, Bourne); the confiscation of empty homes has none. In Portugal there is now the comparative legal analysis by Violante & Gameiro 2026 — unread ?.
4 · Structurally low rents: the model is Vienna
Discipline through the social supply side, not through an administrative cap — the institutional route that escapes the control vs. market dichotomy ~ (bibliographic).
The honest caveat
Design rules (Catalonia vs. partial designs; Argentina null). Zero Portuguese causal studies in the corpus — the gap to confess rather than hide ?. And the “impossible” is coalitional, not economic: the Cambridge 1995 and St. Paul 2021 referendums show the electorate choosing both directions, including the fall in property values ✓.
The seven questions from the debate
Each answer is a chain of claims; each claim has its reference and its level of verification.
Biases: where these studies may be skewed
Trust without audit is faith. Affiliations confirmed on OpenAlex, the press and the PDFs; citation counts from the API.
Who pays for the research — and the symmetry
The most critical study of Berlin comes from the IW in Cologne — an institute funded by the German employers' associations (Sagner & Voigtländer 2022 ✓ affiliation). The most favourable one (Catalonia) is funded by the IEB Foundation, with a chair co-funded by the Barcelona City Council (PDF ✓). The DIW (public, Leibniz) authors the in-between studies; Chen et al. declare no funding. The critical and the favourable poles have sponsors pointing in opposite directions — both must be read against each other before strong claims ?.
Citation bias · metrics · horizons · geography
426 citations in the most-cited anti-control study vs. 30 in the pro — consistent with the publication bias against null results ✓ counts / ? inference. Berlin measures listings, Catalonia registered contracts, NYC hedonic counterfactuals — magnitudes are not comparable across designs ✓. Most measure 1–3 years, and the Mietpreisbremse effect dies out at ~1 year: short horizons would overstate permanent effects ✓. The US and Germany dominate the causal corpus; transferring to Portugal as certainty is not defensible ✓+?.
Circulation has an owner too
Kholodilin's meta-review (2024) was promoted by the IEA — a free-market think tank — in a review that selects partial conclusions from a paper that is itself balanced. Distrust the paper and whoever passes it along in equal measure ?.
Data explorer
The evidence table and the case studies behind everything above.
Case studies, the world over
Evidence table
| Study | Place / intervention | Controlled rents | Supply / quality | Spillover (free/land) | Distribution |
|---|
Methodology, in three lines
Corpus collected from the OpenAlex API (2026-09-03, directed queries; full responses archived, 2026-09-03). Deduplication by DOI; 33 studies extracted by causal design and centrality; 17 verified against the full abstract or complete PDF (✓ key); the rest are marked ~ or ? and must be read as provisional. Charts in ECharts 5.5.0, served from the domain itself and pinned to that version — a library that changes under your feet changes the charts without warning.