Skip to content
Official and academic sources Verified

Empty dwellings count

bfs-vacancy

Publisher
OFS/BFS
Access method
api
Table id
T 09.03.04.01 (je-d-09.03.04.01)
Licence
opendata.swiss terms of use "terms_by": "Open use. Must provide the source."
Coverage
Geography national (1984–), cantons (1995–), communes (latest year only) · referenceDay 1 June · from 1984 · to 2026
Gaps
National series from 1984 (counts on 1 June; until 1983 the count was on 1 December and there was no count in 1983). Cantonal series start in 1995; DF_LWZ_1 publishes no rates for 1995 and 1996 (empty cells). Commune values only for the latest year. Holiday homes are not counted (BFS definition).
Used on
PROB housing.vacancy-rate, STORY-05, STORY-09, PROB housing, AGORA-20, AGORA-22
Known limits
Vacancy rate = empty dwellings ÷ total dwelling stock. Until 2010 the rate was computed on the approximate total stock of the previous year; since 2010 on the stock of the Buildings and Dwellings Statistics (GWS) — a method break (T 09.03.04.01, footnotes 2 and 6). Results for 2012–2015, 2017 and 2019 were revised (footnote 7). Counts are reported by the communes on 1 June; only dwellings offered for permanent rent or sale are counted, holiday homes are excluded.
Verification
verification date: 2026-10-31 · method: Fetched T 09.03.04.01 (XLSX) and DF_LWZ_1 (SDMX), cross-checked national counts 1995–2026 between both (identical), and compared 2026 values with the BFS press release of 10.09.2026 "Leerwohnungsziffer fällt unter 1 Prozent im Jahr 2026" (45 493 empty dwellings on 1 June 2026, 0,93 %; Zug 0,20 %, Genf 0,31 %, Obwalden 0,38 %, Solothurn 1,91 %, Jura 3,35 %). · checkedValue: 2026: CH 45493 (0.93 %); GE 0.31; ZG 0.20; JU 3.35 · checkedAgainst: https://www.admin.ch/de/newnsb/kgvqAd91MYZr

How current is this data?

Freshness
Up to date

According to its update rhythm, no newer edition is expected yet.

Period covered
1984–2026
Last change of the data
Last check of the source
Update rhythm
Yearly

Ledger note (in English): Yearly. Reference day 1 June; BFS publishes in September (2026 count — press release of 10.09.2026; T 09.03.04.01 "Stand der Datenbank 10.09.2026").

Votara checks the source
Every week, and every day from 3 days before to 14 days after an announced publication date

The state is computed each time the site is built; this page shows it as of . Results of the accuracy checks

Transformations

  1. National rows (geo_level ch, 1984–latest): BFS table T 09.03.04.01 "Leer stehende Wohnungen. Entwicklung" (XLSX, current edition via the DAM orderNr URL). Columns located by their labels "Total" and "Leerwohnungs-ziffer". Script: scripts/fetch/bfs-vacancy.ts
  2. Cantonal rows (1995–latest) and commune rows (latest year): stats.swiss SDMX dataflow CH1.LWZ/DF_LWZ_1 version 1.1.0 (https://disseminate.stats.swiss/rest/data/CH1.LWZ,DF_LWZ_1,1.1.0/…), measures OBS (vacant dwellings) and RATE (%), totals over number of rooms (_T) and type of vacant dwelling (_T). The commune query uses the latest year that has cantonal counts.
  3. The national rows of DF_LWZ_1 are only used to cross-check T 09.03.04.01; any difference is reported as a warning in the nightly summary (none on 2026-10-31).
  4. Rates rounded half-up to 2 decimals (the precision of the BFS publications; some XLSX cells hold unrounded quotients such as 1.8488 for 1998). Counts copied as published.
  5. geo_name = German label of the BFS codelist (e.g. "Genf", "Biel (BE)"); official commune names are in bfs-communes (join on geo_code = bfs_number). Communes are those of the dataflow's commune state (snapshot 2026); districts and greater regions are dropped.