These figures are asking prices from listings, not sale (transaction) prices.
- Source
- Active residential sale and rental listings published on Evlek. Land, commercial property and listings with a placeholder price are left out. Homes fall into two groups: flats (apartment, penthouse, studio, residence, duplex) and houses (villa, semi-detached villa, detached house, bungalow).
- Cleaning
- Duplicate listings of the same property count once. Listings with an area outside 20–1000 m² are left out of the sample. From methodology v5 (November 2026) the area basis is gross: a listing that does not state its basis is taken as gross, and a listing with a net area is left out of the per-m² figure and computed separately for comparison only; no gross-to-net factor is applied. v4 (October 2026) left gross areas out, and v2 and v3 made no such distinction. Outliers are removed with the Tukey IQR 1.5× rule both within the figure’s own sample and within the listing’s flat or house group. Listings older than 180 days whose price or status has not been renewed for more than 120 days are not counted; they stay on the site.
- Sample
- n is the number of listings left after cleaning and is printed next to every figure.
- Threshold
- A figure is published from n ≥ 10 and withheld below it; the price band (p25–p75) is shown only from n ≥ 20. There are two public labels: Established and Early data. A new figure becomes Established after two consecutive months with at least 30 listings, a 95% confidence interval whose half width is at most 10% of the median, and the office test (with the largest office taken out, at least 30 listings remain and their median moves by no more than 10%). An Established figure stays Established. Any other figure is shown as Early data, with at most two reasons: few listings, listings mostly from one office, or the figure is volatile. Figures that were mature in October 2026 start as Established; a figure published before the method changed stays published while its sample has at least 5 listings. The monthly table and the Q3 summary follow the same rule. A change on the previous month is given only in the frozen monthly report and the CSV: Established in both months, at least 100 listings in both months, the office test in both months (at least 20 listings left for an Established figure), at least 90% of the room × type strata weight covered, no series break in that month, and a change larger than its 95% confidence interval. The change is mix-adjusted with fixed room × type weights (November 2026 base, fixed for a year); the level is the raw median.
- Updates
- The index is recalculated every day; the date shown is the date of the data. Each month’s figure is frozen on the first day of the next month; once published it changes only to correct an error, with a correction note. When the method changes a new version starts and no change is computed across versions: at the move to v5 (November 2026) there is no comparison with earlier months, which is also why the Kyrenia (Girne) rent figure is not compared with the month before.
- Name
- Evlek Asking Price Index (Evlek İlan Fiyatı Endeksi). The figures are asking prices, not transaction (sale) prices.
- Area
- Gazimağusa is the Famagusta district of North Cyprus; the Famagusta area of the Republic of Cyprus and its series are not part of this index.
- Beta
- The index is in beta: its method and coverage may still develop in the first months, and every change is recorded as a new version. The November and December 2026 reports carry the beta note; the January 2027 report is the first without it.
For sale, each city cell is the median asking price per m² (price ÷ area); for rent, it is the median monthly asking rent. Only active GBP residential listings count, and outliers are removed per city and deal type with the Tukey IQR 1.5× rule on the £/m² distribution (for rentals, monthly rent ÷ area). A figure becomes “Established” after two consecutive months with at least 30 listings, 10% precision and the office test; any other figure is shown as “Early data” with its reason, and a city without a sufficient sample shows no figure. A median is published from n ≥ 10 cleaned listings; the typical-price band (the 25th–75th percentile of total price) from n ≥ 20. Figures refresh daily.
The file is the same in all five languages: current city figures and month-end values.