Rental market survey
Updated: 22 December 2022
Next update: 22 December 2023
About the statistics
The purpose of the survey is to measure rent levels in Norway grouped into different segments of the rental market.
Rents - The actual rent of the rental object. Monthly rents are selected. No ajustments are made for rents that include electricity and/or heating except for the predicted monthly rents (table 09897), where the markup for electricity and or heating are excluded.
Rent allowance - Economic housing benefit administered by the State and the municipalities, provided to cover all or part of rental charges.
SSB-Matrikkelen - Ground Property, Address and Building Register owned by the Norwegian Mapping Authority.
Population and housing census 2011 - In 2011, Statistics Norway carried out a Population and Housing Census in Norway (Census 2011). The purpose of this nationwide census is to illustrate how people live in Norway, provide information on the composition of the population and describe living conditions in the Norwegian society. In the 2011 census, all data are for the first time collected from administrative and statistical registers. Last census was conducted in 2001. Census information is used as weight information in the rental market survey.
Number of rooms - The rooms that are used in the calculations are the number of bedrooms and living rooms excluding kitchens, bathrooms and storage rooms. Rooms beyond 8 are omitted.
Regression model - A statistical method where a dependent variable (here: rent) is explained by a set of explanatory variables (here: dwelling characteristics). Based on actual observations in the main survey a mathematic function gives a connection between the rent and its characteristics.
Predicted monthly rents – Rents estimated by the regression model and the "price" of the different explanatory variables.
A variant of standard classification of urban settlements is used.
Name: Rental market survey
Topic: Prices and price indices
Division for Price Statistics
Annual survey of rent levels. The annual statistics is published around a month after the current period.
Anonymous data at micro level are stored in SAS datasets.
The purpose of the survey is to measure rent levels in Norway grouped into different segments of the rental market. The rental survey was first carried out in 2005 as an external commission and was based on the need for more detailed and improved rental statistics. The statistics was further established as official statistics in 2006. In 2012 the survey is expanded with more detailed figures.
The rental survey is aimed at lessors, tenants and various professional and industrial bodies, as well as users in the public sector (such as ministries) and others with an interest in the rental market. Within Statistics Norway, the National Accounts, the Consumer Price Index and the Household Budget Survey are important users. Primary data is also used in analysis and research within Statistics Norway.
The population is defined as all rental units inhabited by private tenants in Norway.
Rents are mainly collected by web questionnaires directly with households.
As registers of rental units and of tenants are incomplete, a potential population of rental units/tenants is therefore established by connecting different registers. To remove homeowners, persons/addresses from the Central Population Register (DSF) are connected to the Ground Property, Address and Building Register which is called SSB-Matrikkelen. In order to remove homeowners in cooperative dwellings and institutions, information from the Statistics Norway's Business Register is also connected. The sample is established through random selection of persons/addresses from this population of potential rental units.
The size of the gross sample is about 35 000 persons/addresses including an overrepresentation of the largest municipalities. The size of the net sample (the share of responded questionnaires) is about 10 000. Each year a new sample is selected without overlapping previous samples.
The data collection period is mainly carried out in October. Rents are collected by web questionnaires.
Questionnaires with missing values for rents and inconsistent answers are deleted. The prices are subsequently checked in order to identify mistakes and observations with major deviations from average levels stratified by different segments of the rental market. Households are not contacted during the editing process.
Average monthly rents and average annual rents per square metre are calculated for different segments of the market such as geographical areas, letting status, size (such as number of rooms) and period of tenancy. Average rents are weighted together into more aggregated levels and at national level by using weights based on geographical areas, letting status and to a certain degree number of rooms. Most average rent levels are entirely based on rents from the main survey, i.e. the questionnaires, while some average levels are also based on register data. The number of rental units within each average level stratum will therefore vary immensely.
Detailed predicted monthly rents for different geographical areas and size (number of rooms and utility floor space) are estimated based on a regression model.
Data collected from households and firms are subject to confidentiality and are to be kept or destroyed in a secure manner. Any use of the data must be in accordance with the rules set out by the Data Inspectorate.
The rental market survey is a price level survey. The survey is not conducted for price development. The average rent levels from different years are not directly comparable since the survey is based on unique samples each year that can differ according to variables that are important for the rent level. Unlike 2006, no register data are collected from major private letting agencies. As of 2012 detailed predicted monthly rents are estimated by hedonic techniques. A more detailed stratification is introduced and the rents are, unlike previous figures, rounded off.
The data collection is carried out by web questionnaires. Measurement errors can occur if the tenant provides wrong answers due to the difficulty in estimating the values for, for instance, the utility floor space.
Non-response: The non-response is high. The non-response is primarily caused by difficulties reaching the respondents; refusal is only a minor problem in the survey. The non-response increased as a consequence of changing the data collection method in 2013 (mainly web qustionnaires). In 2013 only about 10 per cent of the net sample was contacted by telephone. The share of elderly in the sample is somewhat lower as of 2013 compared to earlier years.
Total and partial non-responses are not imputed.
Skewness: The lack of registers of rental units and tenants makes it difficult to define the population and to control how representative the sample is in proportion to the population.
The share of tenants among the younger population between the ages 20-29 is high. One criterion for the selection of the sample is that persons are actually living at their registered residences. A high percentage of the younger population, especially students, fails to report changing residences. This results in the omission of an important tenant group.
The largest municipalities are overrepresented in the sample in order to be able to publish more detailed figures. Geographical skewness of the sample is corrected with population shares from the Population and housing census 2011.
The survey shows wide variation in rents and several estimated average rent levels are not published due to high uncertainty in the estimates. The estimated rents are published under the condition that with 95 per cent certainty the actual average rent lies within 10 per cent from the published estimate.
Predicted rents based on a regression model will never be without uncertainty. A model will never capture all the factors affecting the rents. The explanatory power of the model, measured by the R 2 , is 61-63 per cent. In other words, the regression model manages to explain well over 60 per cent of the variation in the rents. This is regarded as quite high given the heterogeneity of the Norwegian rental market. The explanatory variables behave with expected signs and with reasonable magnitudes.
The survey measures the actual rent of the rental object, excluding rent allowance and other housing benefits (i.e. the gross rent). Due to different rent allowances, the actual rents of the rental unit and the rent actually paid by the tenants might deviate (gross versus net rents). The degree of reported net rents is unknown.
No adjustments are made for rents that include electricity and/or heating.