Tenure Atlas

Nine households in ten own the home they live in

That is the median across 24 post-socialist countries. In Western Europe it is closer to one in two. The gap is not something a housing market produced: between 1990 and 1999 these states handed sitting tenants the deeds to the flats they were already living in, mostly for nothing.

Post-socialist region
median of 24
87.9
Western Europe
median of 7
50.2

37.7 points apart

24
of 30 countries covered
11
at or above 90%
Dataset rebuilt 2026-09-04

Look up a country

Pick one here, or click any bar or dot further down the page.

Armenia

Census 2022 · Households

90.4of households own their home

+0.4 since 2011

  • Owner90.4
  • Rent1.8
  • Other7.8

Two ways to divide the same census

89.4
owners over every household counted, which is how most tables quote it
90.4
owners over the households that answered the question, which is the figure used here

1.1% of households were counted but gave no tenure. Putting them in the denominator does not make them renters.

Permanent population. The 89.4% figure Armstat quotes divides owners into the raw total; on the stated base used here the rate is 90.4%.

Owner-occupancy, households

Share of households whose home is owned by someone in that household, out of the households whose tenure the census recorded. Latest census for each country; Western European anchors in grey.

Mongolia95.22025Tajikistan93.52020Kyrgyzstan93.42022Slovakia93.32021Belarus93.12019Kosovo92.82024Lithuania91.32021Georgia91.02024Romania90.92021Armenia90.42022Hungary90.32021Moldova89.32024Azerbaijan86.42019Croatia86.42021Bulgaria86.02021North Macedonia85.82002Albania82.52023Kazakhstan81.72021Montenegro77.02023Estonia69.62021Poland68.82021Slovenia66.12021Czech Republic66.12021Latvia63.62021France58.12021Netherlands57.22021Austria51.02021Denmark50.22021Germany43.92021Sweden42.22021Switzerland36.320210255075100

Census years differ; each bar carries its own. † the census left more than 5% of households with no tenure recorded, so the rate rests on an incomplete base.

The same country, two numbers

Two published series for one country rarely agree. A census counts households, EU-SILC counts people, and owner households are larger, which lifts the population figure by a few points. The largest gaps here are bigger than that effect can explain and point at classification instead: where a census puts cooperative housing outside the owner category and the survey puts it inside, the two series diverge by twenty points. Secondary tables quote whichever number they happened to find. The distance between the dots is what that choice is worth.

Census, householdsEU-SILC, population
Sweden+22.6Latvia+20.1Poland+18.3Netherlands+11.6Denmark+10.7Estonia+9.7Slovenia+8.7Czech Republic+8.6Switzerland+5.7Croatia+4.6Austria+3.5Romania+3.4Germany+3.3France+3.1Hungary+1.3Bulgaria+0.0Slovakia-0.2Lithuania-3.9405060708090

Every figure, with its source

“Not stated” is kept out of every denominator and given its own column, because how hard a census chased non-responses is not a fact about housing. Sort by any column, and switch the basis to reach the countries that publish on a different denominator.

Basis

Census counts of private households: the basis the ranking uses, and the only one on which these countries are comparable to each other.

TypeSource
Mongolia202595.21.52.30.9Census · mapped

The source prints thousands to one decimal (957.9 / 15.5 / 22.9 / 9.5 of 1,005.8); the counts here are those figures times 1,000, so they carry the source's rounding. The derived shares reproduce the published percentages (95.2 / 1.5 / 2.3 / 1.0).

Category mapping

'Own' to 'owns', 'rented from others' to 'rents', 'in others' dwelling, rent-free' to 'free' — one of the few censuses in the region that distinguishes rent-free occupancy explicitly — and 'other' to 'other'.

Tajikistan202093.52.04.6Census · mapped

No 'not stated' category; the three categories sum exactly to 1,577,832.

Category mapping

'In the private ownership of persons of these households' (1,474,503) to 'owns'; 'in state ownership' (31,055) to 'rents'. 'In the private ownership of other persons or organisations' (72,274) goes to 'other' rather than 'rents': it names the owner without saying whether rent is paid, merging private landlords and rent-free relatives into one column that cannot be split.

Kyrgyzstan202293.46.00.5Census · mapped

No 'not stated' category. The components sum to 1,314,933 against a published total of 1,314,934; the one-household difference is rounding acknowledged in the source.

Category mapping

'Private ownership' (1,228,545) to 'owns'. 'Rented premises' (58,094), 'state ownership' (17,045) and 'municipal ownership' (4,241) are summed into 'rents'. 'Other form of ownership' (7,008) to 'other'.

Slovakia202193.30.66.13.1Census · as published
Belarus201993.16.94.1Census · mapped

Belstat publishes this only as an OLAP cube, not a numbered table; the counts were read by executing an MDX query against the cube's ownership dimension using Belstat's own guest login. The cube's universe totals 4,098,924 households and is not necessarily every private household in Belarus. 'Not stated' is unusually large at 4.1% and roughly doubled since the 2009 census.

Category mapping

'Private ownership of a natural person' (3,662,261) to 'owns'. State ownership (210,238), private ownership of a non-state legal entity (33,748) and possession under a rental or sublet contract from a natural person (25,327) are summed into 'rents'. IMPORTANT: Belarus does not separate a dwelling owned by a member of the household from one owned by some other natural person, so this 'owns' figure combines the two categories that Armenia's census splits into owner-occupation (90.4%) and living in another private person's dwelling (7.8%). The Belarusian rate is therefore an upper bound on owner-occupation and is not like-for-like with censuses that make the distinction.

Kosovo202492.85.32.0Census · mapped

No 'not stated' category. The source notes the total counts cases where two households share one dwelling.

Category mapping

'Owned' (339,136) to 'owns', 'Rented' (19,229) to 'rents'. 'Other forms of use' (7,189) is mapped to 'other' rather than 'free of rent': the glossary defines it as a unit provided free of charge by an employer or relative, which UNSD would split across both, and the source does not publish the split.

Lithuania202191.33.05.70.0Census · as published
Georgia202491.07.81.22.4Census · mapped

Census date 14 November 2024. Components: rented 85,710 + interest-free loan 2,590 + state 3,074; condominium 1,494 + other 12,534.

Category mapping

'Property of one or more household members' to 'owns'. 'Rented', 'Rented in exchange of interest-free loan' and 'State property' are summed into 'rents' — the second is a Georgian arrangement where the occupant lends the owner money instead of paying rent, which is still tenancy. 'Property of housing-construction condominium' is put in 'other', not 'owns': the title sits with the cooperative rather than the household, and misfiling cooperative housing as ownership is what makes Sweden's census and survey figures differ by twenty points. 'Other types of ownership' also to 'other'.

Romania202190.93.35.80.7Census · as published
Armenia202290.41.87.81.1Census · mapped

Permanent population. The 89.4% figure Armstat quotes divides owners into the raw total; on the stated base used here the rate is 90.4%.

Category mapping

State (8,899) + community (4,578) + legal entity (1,889) mapped to 'rents'; 'other persons' (66,134) to 'other'.

Hungary202190.38.90.80.0Census · as published
Moldova202489.35.65.1Census · mapped

Census date 8 April 2024. Excludes the left bank of the Dniester and Bender, which were not enumerated. No 'not stated' category: the three categories sum exactly to the published total. The source's own sheet has a defect in the urban 'Nord' row, which repeats the national urban totals; the national, urban and rural rows used here are unaffected.

Category mapping

'Proprietar/titular' to 'owns'. The source defines it verbatim as 'unul din membrii gospodăriei, sau mai mulți, au drept de proprietate asupra locuinței respective' — one or more members of the household hold the ownership right — so it means owner-occupation in the strict sense and excludes dwellings owned by some other private person. 'Chiriaș' to 'rents'. 'Altă situație (cu titlu gratuit etc.)' to 'other': it mixes social and service housing with rent-free occupancy of a non-member owner's dwelling, and the source does not separate them.

Azerbaijan201986.413.6Census · mapped

No 'not stated' category; the five columns sum exactly to 2,293,812. The official stat.gov.az URL was unreachable, so the archived copy of the same file is cited.

Category mapping

'Owned by household' (1,982,422) to 'owns'. The four remaining columns are summed into 'rents': state property (251,154), municipal property (17,880), owned by a non-state legal entity (30,706) and rented from a natural entity (11,650). The census classifies by who owns the dwelling, not by whether rent is paid, so no 'free of rent' split exists to extract.

Croatia202186.45.48.30.0Census · as published
Bulgaria202186.05.09.011.8Census · as published
North Macedonia200285.82.611.6Census · mapped

The 2021 census published tenure only as percentages and only for two of its four categories (owner 96.05%, rented 1.16%), which is not a distribution and cannot be normalised, so this 2002 census remains North Macedonia's most recent usable tenure count. The publication is served by an ASP.NET postback; the URL is the landing page, not the file.

Category mapping

'Owner' (483,888) to 'owns'. The four tenancy columns are summed into 'rents': tenant of a state-owned dwelling (5,375), tenant of a whole privately owned dwelling (6,701), tenant of part of a privately owned dwelling (2,374) and social dwelling (452). 'Kinship' (64,748) and 'Other' (758) are summed into 'other'.

Albania202382.58.88.7Census · mapped

No 'not stated' category: the three categories sum exactly to the published 755,950.

Category mapping

'Owner or in process of acquiring legal act' (623,387) is mapped to 'owns' — it includes households holding a dwelling whose title deed is still being issued, which INSTAT counts as ownership. 'Tenant' (66,608) to 'rents', 'Other tenure status' (65,955) to 'other'.

Kazakhstan202181.713.94.40.5Census · mapped

The 26,832 recorded as 'unknown' is the residual between the published total (5,611,253) and the ownership columns. It equals exactly the 'Other premises' row of Table 4 — yurts, gatehouses, wagons, containers — for which the questionnaire skips the ownership question, so it is a skip artefact rather than non-response.

Category mapping

Kazakhstan splits every non-owner category into paid and free, so all four UNSD categories are recoverable. 'Owns' = private property of one respondent in the household (2,820,347) plus of several respondents (1,744,910). 'Rents' = individual paid (714,431) + legal entity paid (11,966) + state property (49,509). 'Free' = individual free (231,264) + legal entity free (11,994).

Montenegro202377.011.111.92.7Census · mapped

'No data' is not a questionnaire option — the form offers five — so it is a non-response residual added at tabulation.

Category mapping

'Owner is household member' (155,489) and 'Co-owner is household member' (5,761) are summed into 'owns' — both place the title with a member of the household. 'Tenant household' (23,218) to 'rents'. 'Lives in the dwelling of parents, children or other relatives' (21,295) and 'Other' (3,576) are summed into 'other'. 'No data' (5,888) to 'unknown'.

Estonia202169.619.610.91.3Census · as published
Poland202168.825.55.86.7Census · as published
Slovenia202166.19.024.90.0Census · as published
Czech Republic202166.121.112.810.5Census · as published
Latvia202163.617.419.00.6Census · as published
France202158.139.62.40.0Census · as published
Netherlands202157.242.80.01.5Census · as published
Austria202151.045.13.90.3Census · as published
Denmark202150.249.80.00.9Census · as published
Germany202143.956.10.00.0Census · as published
Sweden202142.233.224.63.0Census · as published
Switzerland202136.360.53.20.0Census · as published

Where a source needed interpretation, the row carries it: the note in plain text, and the exact category mapping one click away.

What is missing

Six of the thirty countries are absent from the ranking. Every one of those gaps is at source: the census either counted something other than households, or never asked about tenure at all.

Published on a different denominator

  • Bosnia and Herzegovina
  • Serbia

These censuses count occupied dwellings instead of households. That is a real figure about a different thing, so it is held in the dataset and kept out of the ranking.

Not yet covered

  • Russia
  • Turkmenistan
  • Ukraine
  • Uzbekistan

Russia has never asked: the 2020 census questionnaire runs to ten questions and none concerns ownership. Ukraine's only census since independence was held in 2001 and recorded dwelling type, not tenure. Turkmenistan's 2022 census ran 401 tables and never crossed households with tenure; its one published figure is housing stock by owner, which is a different thing and is deliberately not shown here as a tenure rate. Uzbekistan held its first post-Soviet census in 2026, with final results not due until July 2027.

Methodology

Housing tenure has no maintained global dataset. What circulates instead is a mixture of household rates, population rates and housing-stock percentages, quoted interchangeably, which is how Armenia acquired a widely repeated homeownership rate that no census supports. This atlas fixes the classification, states the denominator, and links every cell to the table it was read from.

Shares are taken of the stated base
Armenia's censuses left 2.4% of households unstated in 2001, 3.7% in 2011 and 1.1% in 2022. Dividing owners into the raw total gives 88.7 → 86.7 → 89.4: a dip and a recovery that never happened. Dividing into the households whose tenure was actually recorded gives 90.9 → 90.0 → 90.4, which is flat, which is what the country did. Every rate here uses the second denominator, and the excluded share is published beside it.
UNSD categories
Owner, tenant, occupied free of rent, other arrangements, not stated. These come from UNSD's census recommendations, the same five used in Compendium Table 15. Adopting a published UN classification means each figure can be checked against the 2010-round baseline: the Armenian 2001 mapping here reproduces the cell UNSD published, to the household.
Units are never mixed
Households, persons and dwellings are three different denominators and are not interchangeable. Each observation records which one it uses, no ranking combines them, and where both exist the atlas shows the gap.
Mappings are written down
Many censuses publish tenure by owner identity — household member, state, municipality, legal entity, another person — instead of by rental arrangement. Turning that into tenure is a judgement call, so every adjusted record carries its mapping in plain text, and the build refuses to ship one that does not.
Limitations
Census years differ, so the ranking compares nearby years and not a single one. Large “other arrangements” shares — Sweden, Slovenia, Latvia — should be checked against national metadata before being read as non-ownership, since some countries classify cooperative housing there. Coverage outside the EU is incomplete, and the missing countries are listed above.

Data sources