Experimental statistics

Methodological notes

Budapest district groups to classify districts:

Buda, hilly districts: I, II, XII

Buda, other districts: III, XI, XXII

Pest, inner districts: V, VI, VII, VIII, IX

Pest, transitional districts: X, XIII, XIV, XIX, XX

Pest, outer districts: IV, XV, XVI, XVII, XVIII, XXI, XXIII

The ingatlan.com database makes it possible to identify multiple-advertisements, and this solves the most difficult problem for advertising databases, the bias caused by repeated advertisements.

Rent indices showing changes in home rental prices are determined by hedonic regression analysis performed on different subdivisions. A similar modelling technique was used for each sub-index, and the range of variables involved was the same. All models cover the entire period. Data are available from May 2015. The 2015 data were considered as a control during the modelling; accordingly, the calculated indices are included as the 2015 base index. Changes in rent over time – so the rent index itself – are derived from the coefficients of time-dummies formed per month. To monitor the composition effect, we used duly completed housing quality characteristics available in the advertising database, as well as the indicators included in the HCSO databases. The target variable, the floor area and the housing market price level characteristic of the settlement (part) were included in the models by logarithmisation, and the categorial variables by transforming them into a dummy form.

By aggregating different sub-indices, covering all ads examined, we can calculate multiple indices. The present compilation uses an indicator calculated from regional sub-indices. Additionally, the attached tables include both indicators calculated by advertiser stratification and indicators calculated on the basis of one single model.

The detailed methodological description is available in the publication entitled Rent index was developed by HCSO and ingatlan.com.