We often hear that tourists push out tenants: owners let their flats on Airbnb, fewer are left for locals and rents rise. The state passed a new hospitality act. In the 14 municipalities on the government's list, short-term letting will be capped at 60 days a year in 2027 and 2028, and a municipality may set its limit between 30 and 270 days.
We do not have Airbnb's data. We have something better: the register of accommodation establishments (AJPES), in which anyone letting to tourists must be entered, with an address and a number of units. We set it against the filed leases and sales of the property-market register and against residents on the 100 m grid. Unregistered letting stays dark; every number here is about the registered part.
We looked at seven municipalities, Ljubljana, Piran, Izola, Koper, Bled, Kranjska Gora and Bohinj, and three towns with almost no tourists for comparison: Maribor, Celje and Kranj.
How many tourist flats there are
In Bled there are 41 registered tourist units for every 100 dwellings, in Bovec 37, in Bohinj 33, in Kranjska Gora 30. There tourism is not part of the place; it is the place. Piran has 18, Ljubljana only 1.6, but that is 2,262 units, the most in the country.
A city-wide average hides where they are. The map shows, for every 100 m cell, the units within 250 m per 100 dwellings within the same circle. In Ljubljana they cluster in the old centre; in Piran they cover almost the whole peninsula and Portorož.
A unit is a registered accommodation unit (flat, room, house), not a dwelling. Hotels, campsites and hostels are left out; unregistered letting is not visible.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
Cells with fewer than 25 dwellings within 250 m are left out. Roads: OpenStreetMap.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
Rents really did rise
First the fact nobody disputes: rents in Ljubljana rose fast. The median filed rent was €6.63/m² in 2014 and €9.48/m² in 2024 (+43%).
The last two years need care. Leases reach the register late, so for 2025 and 2026 it holds mostly the ones filed quickly, and those rent above the rest every year (in 2024 €12.50 against €9.19/m²). Only the series of quickly filed leases compares like with like: €8.83 in 2018, €10.63 in 2021 and €14.37 in 2026, +63% since 2018. Within a kilometre of the centre this year's leases stand at €16.32/m².
Over the same years Ljubljana's short-term letting operators grew from 373 to 761 active. Both lines rise. That alone does not say how much one causes the other.
Operators are counted by registration number, which is not published. Rents for 2025–26 are only leases filed in their own year, which rent above the rest every year; the orange line is the comparable one.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
Next to the tourists, prices did not rise faster
If tourist flats push rents up, it should show first where there are most of them. We compared the most-let quarter of dwellings with the rest at the same distance from the centre of the same town, in the same year. Such a comparison removes everything a town shares: inflation, interest rates, the Slovenian market and the European one.
Sale prices next to the tourists did not pull ahead. In Ljubljana they even fell behind after 2019 (−6.6%), on the coast too (−6.4%), and not in the control towns. Rents show no clear pattern: slightly higher in Ljubljana's tourist parts (+4.6%), lower on the coast, both with wide uncertainty.
This comparison has a blind spot, and we have to say it: if short-term letting raises rents across the whole city alike, the effect cancels out in it. That is one reason we also looked at whole towns.
The most-let quarter of dwellings against the rest of the same ring in the same city. An effect common to the whole city cancels out in this comparison.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
Tourist towns are no dearer than others
Across 104 municipalities with 60 or more sales, prices from 2016–18 to 2022–24 did not rise more where there are more tourist flats. The median growth of the price-to-value ratio was 61%: Bled 71%, Kranjska Gora 76%, Piran 50%, Maribor 77% and Celje 81%.
Slovenia did get dearer than Europe: prices are up 115% since 2015, the EU's 62%; rents 74% against 21%. But the tourist towns did not rise more than Maribor, Celje and Kranj. The gap to Europe is Slovenian, not touristic.
Municipalities with 60 or more sales. Growth of the price-to-value ratio, which removes differences in what sold.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
Eurostat's index for countries; for towns the price-to-value ratio within neighbourhoods. Bled has few sales, so its line jumps.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
Where tourists took the tenants' place
The effect the data shows clearly is not in the price but in the number. In Ljubljana's 458 most-let cells (13,533 dwellings), 51% more long-term leases were filed in 2014 than in 2019, compared with the rest at the same distance from the centre. During the Airbnb boom about a third of the leases there disappeared (34%). In the most-let quarter of dwellings the gap is smaller (30%), so the effect grows with density. During the pandemic some leases came back (+10%), then fell back again.
If there is a set number of dwellings to let and ever more short-term operators, must fewer remain for long-term letting? Only if the pool is fixed, and in Ljubljana it is not. The distinct dwellings with a long-term lease grew from 8,046 to 11,803 between 2018 and 2024 (+47%), the tourist units in residential buildings from 1,352 to 2,047. The tourist share fell to 13.3% in the pandemic and has since risen to 14.8%. Dwellings move between owners, emptiness, long lets and short lets, and new ones are built.
In Ljubljana companies carry the growth: units let by companies numbered 728 in 2018 and 1,344 this year; registered private hosts hold only 193 units. There are 120 operators with five or more units.
Compared within the same distance from the centre and the same year; the band is ±2 standard errors. The register starts in 2018, so the years before are dated by the leases.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
The long-term pool is the number of distinct dwellings with a lease that started in the last three years. Filing completeness changed before 2018, so the table starts there.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
How much of the rise they can explain
Short-term letting in Ljubljana is not small. The 2021 census counts 13,581 rented dwellings, and there are 2,063 registered tourist units in residential buildings: 15% of the rental stock. Since 2018, 711 have been added, 5.2% of the stock.
How far such a withdrawal can raise rents depends on how the market responds. Across municipalities we measured that more tourist units per rented dwelling go with slightly faster rent growth, but uncertainly; for Ljubljana that gives +0.2% (up to +0.6%). Even assuming a closed market, where no tenant can buy or move away (a demand elasticity of -0.5), we get +10%. Rents rose +63%.
Registered short-term letting therefore explains at most a sixth of the rise, probably much less. If unregistered letting is as large again, the upper bound doubles to +21%. Most of the rise is demand: more households, students and foreign workers, and buyers pushed into renting by interest rates.
The rental stock is the 2021 census. The upper bound assumes no tenant can buy or move away. Unregistered letting is not counted; if it is as large as the registered, the bound doubles.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
Fewer residents in the centres
Cells with more tourist units hold fewer registered residents per dwelling: in Ljubljana 2.20 under five units per 100 dwellings and 1.70 at 10 to 20, in Piran 1.36 in the most-let part. But Maribor, with almost no tourists, has 1.53 in its centre. Centres have fewer residents per dwelling on their own, so we cannot put this down to tourists alone. Piran's old town stands out.
An empty dwelling names no cause: a short let, a holiday home and an owner registered elsewhere are the same to the grid. Centres have fewer residents even without tourists.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
The law began to act before it applies
The cap on days applies from 1 January 2027, but the register changed this year already. Between the snapshots of 1 September 2026 and 17 October 2026, 761 establishments became inactive, 352 of them on one day, 30 September 2026. In September the register lost 782 units, the most in any month since it began.
Mostly registered private hosts left: 538 of 3,199 (17%) in seven weeks, against 2.6% of companies. And mostly on the coast: Izola lost 23% of its tourist units, Koper 16%, Piran 10%, elsewhere mostly a few percent.
The new act required existing private hosts to enter missing consents and data by 30 June 2026, or AJPES deletes them. A deletion on one day and in one region looks like those cases being processed; other regions may follow. The register does not say whether the flats went to long-term tenants, stayed empty or are let on without an entry. The filed leases for 2026 and 2027 will show it.
An entry is the registration, not the opening: registration became compulsory in 2018. An exit is the last change of the entry; between the 1 September and 17 October 2026 snapshots the date is exact.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
The register does not say why they left. The 30 June 2026 deadline for consents and data is known from secondary sources.
Source: AJPES RNO, ETN, SURS, Eurostat · bloki.si
What the data cannot say
Unregistered letting is not visible. Every number is about the registered part; if unregistered letting is as large again, the upper bound of the effect on rents doubles.
The register starts in 2018, when entry became compulsory. Growth before then is dated by the leases, not by the entries.
A comparison within a town cannot see a town-wide effect; that is why we added the comparison between municipalities and the estimate of the withdrawal from the rental stock. The response of rents to the stock is not measured directly here, so we give a range.
Bled, Bohinj and Kranjska Gora have no comparison: almost every dwelling is in a tourist cell, so there we stay with description.
An empty dwelling names no cause: a short let, a holiday home and an owner registered elsewhere are the same to the residents' grid.
§ How it is calculated
The register. AJPES's register of accommodation establishments in two snapshots (1 September 2026 and 17 October 2026). Short-term letting is flats, rooms and houses; hotels, campsites and hostels are left out. The stock in past years is rebuilt from the registration date and, for inactive establishments, the date of the last change. Establishments are placed at their own address.
Intensity. For every 100 m cell: tourist units within 250 m per 100 dwellings within the same circle, cells with too few dwellings left out. Tourist cells are each town's most-let quarter (and tenth) of dwellings as of 2019.
Comparison. Price is the sale price over the GURS generalised value, rent is €/m² a month, leases are the number of filed contracts. The comparison is within a town, a ring around its centre and a year, with 500 m neighbourhood effects; uncertainty is clustered by neighbourhood.
Attribution. Tourist units added in residential buildings, divided by the rented dwellings of the 2021 census, times the response of rents: the one measured between municipalities, an assumed range up to 0.4% per 1% of the stock, and a closed market (elasticity -0.5).
The study is short_term_lets in the bloki_si project; every number on the page comes from its export.
Can the Bloki server answer such questions?
For a given address or area the server says how many tourist establishments are around and what rents are; who lets and whether a flat is let without an entry, nobody knows.
- Yes
How many tourist establishments are around this address?
Registered establishments, beds and units nearby, and units per 100 dwellings within the same circle.
neighbourhood - Yes
What are rents in this neighbourhood?
Filed leases with a median and n, by size and period.
find_comparablesmarket_stats - Yes
How many people live around here?
Registered residents per dwelling within 500 m and 1 km, with the warning that an empty dwelling names no cause.
neighbourhood - Partly
Is this flat let to tourists?
Only if it is in the register; unregistered letting is not visible.
neighbourhood - No
Who lets, and what do they earn?
Private persons are never published, and the register holds no nightly prices or occupancy.
data_limitations