Principles to be applied by Member States for the estimation of output, intermediate consumption and transactions with the rest of the world, regarding dwelling services
In national accounts, by convention, the output of dwelling services comprises not only the services produced by rented dwellings but also those provided by owner-occupied dwellings. With regard to the valuation of output of dwelling services, the European System of Accounts established with Regulation (EU) No 549/2013 (ESA 2010) stipulates in paragraph 3.75 of Annex A, that ‘the output of services of owner-occupied dwellings is valued at the estimated value of rental [For consistency with ESA 2010 the term ‘rental’ is used throughout this text.] that a tenant would pay for the same accommodation, taking into account factors such as location, neighbourhood amenities, etc. as well as the size and quality of the dwelling itself. In principle, several methods are available for estimating the value of services produced by owner-occupied dwellings:
the stratification method based on actual rentals, which combines information on the housing stock, broken down by various strata, with information on actual rentals paid in each stratum,
the user-cost method, where separate estimates are made of intermediate consumption, consumption of fixed capital, other taxes less subsidies on production and net operating surplus. The output of dwelling services is the sum of these components,
the self-assessment method, where owner-occupiers are asked to estimate a potential rental for their property,
administrative assessment methods, where a potential rental is determined by third parties, e.g. by government for fiscal purposes.
The stratification method based on actual rentals is the preferred method. This method can also be used for estimating the value of all actual rentals from a sample of actual rentals and for estimating the value of dwelling services from dwellings let at a zero or low rental (see Section 1.4.1 for detail).
The user-cost method should be applied only under certain conditions and only for those strata of the housing stock where actual rentals are missing or statistically unreliable.
The self-assessment method should not be used due to a largely subjective influence on the estimate which leads to substantial uncertainties in the results.
Administrative assessment methods, especially when related to taxation, may produce biased results. However, sometimes results from objective assessment methods may be available for some strata. It is acceptable to use the results when the objectivity of the method and the comparability of the results can be demonstrated.
The stratification method uses information about actual rentals from rented dwellings to obtain an estimate of the rental value of the stock of dwellings. This may be interpreted as a grossing-up procedure based on a price times quantity approach. A stratification of the housing stock is required to obtain a reliable estimate and to include relative price differences properly. Subsequently, the average actual rental per stratum is applied to all dwellings in that particular stratum. If the available information is derived from sample surveys, this grossing-up relates to both a part of the rented and all owner-occupied dwellings. The detailed procedure to determine a rental per stratum is usually carried out for a base year and is then extrapolated to the current years.
The rental to be applied to owner-occupied dwellings in the stratification method is defined as the private market rental due for the right to use an unfurnished dwelling. The rentals for unfurnished dwellings from all private market contracts should be used to determine imputed rentals. Private market rentals that are at a low level due to government regulation should be included.
If the information source is the tenant, it may be necessary to correct the observed rental by adding any specific rental allowance, which is paid directly to the landlord. If the sample size for the observed rentals as defined above is not large enough, also observed rentals for furnished dwellings may be used for imputation purposes, provided they are adjusted for the furniture element. Exceptionally, also increased rentals for public-owned dwellings may be used. Low rentals for dwellings let to relatives or to employees should not be used (see Sections 1.2.3 and 1.4.1 for details).
The stratification method may be used also for grossing up to all rented dwellings. The average rental for imputation as described above may not be suitable for some segments of the rental market. For example, scaled-down rentals for furnished dwellings or increased public rentals may not be appropriate for the respective actually rented dwellings. Especially, in the case of the actually rented dwellings that are furnished, the furniture cost is in fact part of the rental and it should be included in the calculations for the output. Separate strata for actually rented furnished or social dwellings combined with appropriate average rentals can overcome this problem.
In principle, the rental should exclude charges for heating, water, electricity, etc. Where the data sources do not allow this, coherence between the rentals and intermediate consumption must be ensured (see Section 2 for details).
Where for certain strata of owner-occupied dwellings a representative actual rental is missing, this can be overcome in most cases by applying extrapolation or regression techniques.
An alternative to the standard stratification method which relies on extrapolating average rentals per stratum is the use of hedonic regression methods. Briefly, such methods use the sample data for rented dwellings to determine a price of each dwelling characteristic (size, location, presence of a balcony, etc.). Output is obtained by multiplying each characteristic of (a representative sample of) the dwelling stock with the hedonic price of this characteristic. Regression methods permit an enlarged number of variables to be taken into account and can be particularly effective when for some strata observed rentals are missing.
Obviously this does not solve the problem in the extreme case of all dwellings being owner-occupied or in the absence of a developed rental market. As an objective assessment for such cases the user-cost method should be applied. The user-cost method should only be applied to owner-occupied dwellings.
The user-cost method should only be applied when the stratification method based on actual rentals cannot be used because the rental market is unrepresentative.
By convention, this is considered to be the case when the following conditions are met: (1) the volume measure (number of dwellings or square meters of dwellings) of privately rented dwellings represents less than 10 % of the total volume measure (number of dwellings or square meters of dwellings) of dwellings; and (2) if (1) is fulfilled and the share of the total rented dwellings – market and non-market – in the total dwelling stock exceeds 10 %, the disparity between market and other paid rentals exceeds a factor of three. Even when both conditions apply, a Member State may still choose to use the stratification method provided that the results are of sufficient quality. When the conditions are not met, the stratification method should be used unless it can be demonstrated that no representative market rentals are available for significant parts of the dwelling stock and that the user-cost method produces more comparable results. If justified, the user-cost method can be applied for all or parts of the dwellings stock. When deciding on the split of the dwelling stock into the parts for which the user-cost method or the stratification method is applied, factors such as data limitations and the country specific situation should also be taken into account.
Under the user-cost method, the output of dwelling services is the sum of intermediate consumption, consumption of fixed capital (CFC), other taxes less subsidies on production and net operating surplus (NOS). For owner-occupied dwellings, no labour input is recorded for work done by the owners [It is conceptually possible that owner-occupiers, individually or collectively, employ housekeepers without involving other statistical units such as housing corporations or management service companies. Where this situation occurs, output according to the user-cost method should include an allowance for compensation of employees.]. Experience suggests that CFC and NOS are the two largest items, each representing 30 to 40 % of output.
CFC should be calculated based on a perpetual inventory model (PIM) or other approved methods. A separate estimate for the owner-occupied residential buildings should be available.
The net operating surplus should be measured by applying a constant real annual rate of return of 2,5 % to the net value of the stock of owner-occupied dwellings at current prices (replacement costs). The real rate of return of 2,5 % is applied to the value of the stock at current prices since the increase in current value of dwellings is already taken account of in the PIM. The same rate of return should be applied to the value of the land at current prices on which the owner-occupied dwellings are located.
The value of land at current prices may be difficult to observe annually. Ratios of land value to the value of buildings in different strata may be derived from an analysis of the composition of the costs of new houses and associated land.
To compile the output of dwelling services Member States shall apply the stratification method based on actual rentals, either by direct extrapolation or by means of econometric regression. In respect of owner-occupied dwellings, this implies the use of actual rentals for similar rented dwellings. In the justified and exceptional case where actual rentals are missing or statistically unreliable, other objective methods, like the user-cost method, may be employed. No additional justification is required for compiling the output of owner-occupied dwellings with the user-cost method provided that two conditions are met: (1) privately rented dwellings represent less than 10 % of the dwelling stock; and (2) if (1) is fulfilled and the share of the total rented dwellings – market and non-market – in the total dwelling stock exceeds 10 %, the disparity between market and other paid rentals exceeds a factor of three.
A first set of variables determining the level of rentals concern the characteristics of the dwelling and the building. To start with, the size of a dwelling will be important, both in terms to the area and the number of rooms. The bigger the dwelling, the higher will be the rental. At the same time, the rental per square metre tends to fall with the size of a dwelling. However, for some categories of dwellings (e.g. flats in capital cities) there may be a U-shaped relation between the price per square metre and the size of the dwelling. Another important factor will concern the amenities of a dwelling. These may cover variables like the existence of a bathroom, a balcony/terrace, special flooring or wall covering, an open fire, central heating, air-conditioning, special glazing and other noise or thermal insulation measures; the layout of the dwelling is also relevant. With regard to the building, certain facilities may have an influence like a garage, a lift, a swimming-pool, a (roof-) garden or the position of a dwelling within the building. The type of the building (detached, semi-detached house, flat), the architecture, the age, or the number of dwellings in a building may also affect the rental.
A second set of variables relate to environmental characteristics. A well-known factor is the rental difference between a comparable dwelling in a city and a remote location. The distance to an economic centre or the form of the landscape (flat land, mountainous) may not be negligible factors. In addition, neighbourhood factors like the view, surrounding green areas, transport facilities and access, shops and schools or the reputation and security of a district tend to have an influence on the actual rental.
Another set of variables may be summarised as socio-economic factors. For instance, in most Member States rentals are influenced by general government regulations like rental restrictions or subsidies. Further, the age of the tenancy agreement, the type of contract (temporary, permanent), the number of inhabitants per dwelling (flat-sharing community), the type of owner (government, housing association, private, employer) or the rental policy of the landlord may also affect the rental.
Obviously, several additional variables may have an influence on rentals. But to collect all the abovementioned factors could lead to overburdened questionnaires. Therefore the use of capital values for stratification purposes may be considered. The rationale behind the use of the capital value of a dwelling is that it reflects all its important characteristics. The capital value can thus be considered as an implicit stratification factor. Using the ratio of the capital value to actual rental may be regarded as a feasible approach, especially in those Member States where the rented dwellings represent a minor part of the housing stock. Provided the ratio is stable, such a method would enable the rental value to be determined for those dwellings which only feature in the owner-occupied sector. In addition, capital values do not exclude the use of ‘physical’ stratification criteria; they may be combined. In this situation, capital values are assumed to reflect the missing ‘physical’ stratification criteria. In any case, capital values to be used in the rentals calculation have to be based on an objective assessment established for an up-to-date reference year.
In practice the stratification differs from one Member State to another in respect of both the number of strata and the precise criteria used to define them. Although at first sight this may cause some concern, it should be stressed that certain basic criteria, like the size and the (geographical) location of a dwelling are used almost everywhere. Moreover, the aptness of other features will vary between Member States and Member States themselves are in the best position to determine significant criteria.
For stratification purposes, Member States shall use important features of the dwellings. These may relate to the characteristics of the dwelling and the building, to environmental characteristics of the dwelling or to socio-economic factors. In addition, the use of up-to-date capital values is acceptable for stratification purposes, if these are based on an objective assessment.
Given the various characteristics affecting the rental of a dwelling, the first task is to investigate those variables having a significant impact. One way of detecting significant variables is to produce a tabular analysis of the statistical information available. To obtain an objective measure of assessment, the variance of the actual rentals within a stratum would seem to be useful. This would create an incentive for possible improvements to the stratification by choosing strata in order to minimise the within-stratum variance. It is therefore recommended that the variance per stratum is calculated at least in those cases where the stratification affects the level of both actual and imputed rentals.
A more sophisticated approach is offered by advanced statistical techniques like (multiple) regression analysis. Such a technique allows the influence of individual variables to be assessed, so that the variation in rentals can be assigned to certain characteristics. Briefly, the explanatory power of a variable can be quantified via the correlation coefficient. As a by-product, it enables the characteristics to be ranked according to their importance. This helps to determine where to stratify in greater detail. Combining the most important variables, using multiple regression techniques, shows their overall explanatory power. The use of advanced statistical techniques to select important variables is considered to be an efficient way of stratifying the housing stock. In addition, regression analysis can be used directly for estimating the rentals, e.g. in the form of hedonic models. The technique is also a useful tool for estimating the average rental for strata where there are no corresponding observations in the rented sector (empty strata).
A further advantage of selecting stratification criteria on the basis of an advanced statistical technique is that it avoids the need to prescribe uniform criteria for all Member States. To obtain a comparable result, it is sufficient to establish a ranking of the most important criteria in each Member States and to stipulate the required overall level of explanatory power. Obviously, such a regression analysis depends largely on the available statistical information. However, even in a situation of restricted statistical information, this could be an incentive for future improvements.
Given that the information about the different variables affecting rentals mainly depends on the development of basic statistics, the possibility of using advanced statistical techniques may be restricted at present. Therefore a standard method is recommended, i.e. Member States shall apply all significant criteria as derived from tabular analyses. As a minimum, the size, the location and at least one other important feature of a dwelling have to be used to stratify the housing stock; this stratification should produce a minimum of 30 cells. The breakdown of the housing stock has to be meaningful and representative of the total stock of dwellings. An advanced statistical technique may be used to determine the important explanatory variable(s) for selecting the strata.
In practice, however, a Member State may prefer to use fewer variables, or variables other than those prescribed by the standard method. This is acceptable, as long as a (multiple) regression analysis proves an acceptable level of explanatory power. To guarantee comparable results, a correlation coefficient of at least 70 % is recommended as a threshold. This threshold value would be acceptable in the context of a large sample, with zero and cheap rentals as well as outliers having been removed.
Member States shall use tabular analyses or statistical techniques to derive significant stratification criteria. As a minimum, the size, the location and at least one other important feature of a dwelling have to be used. A minimum of 30 cells are to be produced and at least three size classes and two types of location shall be distinguished. The use of fewer or other variables is acceptable if it has been proved previously that the (multiple) correlation coefficient reaches 70 %.
Imputed rentals are determined based on observed actual rentals. For imputation purposes, the rental is defined as the price due for the right to use an unfurnished dwelling. In order to meet this definition, observed rentals may have to be adjusted.
Charges for heating, water, electricity, etc. should be excluded from the rentals, although it may sometimes be difficult to separate them out in practice. To be in line with the valuation rules of the ESA 2010, the output of dwelling services should be at basic prices.
With regard to observed rentals, certain public supports are probably important. For instance, a specific household as a consumer is entitled to a general government transfer (e.g. housing benefit) but which, for administrative reasons, is paid directly to the landlord. Depending on the information source, the observed rental may differ. If the information source is the tenant, it may be necessary to correct the observed rental by adding back any specific rental allowance.
In addition, the use of actual rentals for imputation purposes requires the clarification of several fundamental questions having an impact on the harmonisation of the data. The first point relates to the question of whether use should be made of all actual rentals or only those from new contracts, for the imputation procedure. Depending on the purpose, different theoretical arguments can be put forward supporting actual rentals paid according to new contracts, contracts signed in the construction year or ‘average’ contracts. Applying the general rule, i.e. to use rentals of similar dwellings, it does not seem acceptable to limit the basis of imputation to rentals from new contracts. Given that ‘average’ rentals are used for the rented sector, the same should apply for owner-occupied dwellings. In addition, a different solution would probably pose great difficulties in many Member States when applying the stratification method. Briefly, the conclusion is that the average actual rentals from all contracts should be used to calculate imputed rentals. Consequently, private market rentals that are at a low level due to government regulation should also be included when calculating the average rentals.
The second question concerns the problem as to whether rentals of public-owned dwellings can be used for imputation purposes. Given that owner-occupied dwellings are mostly privately owned, in principle, only actual rentals from the private sector should be used for imputation purposes. However, if not enough observations of actual rentals of privately-owned dwellings are available to constitute a sufficient basis for the imputation, exceptionally, rentals of public-owned dwellings may be used, provided they are appropriately increased to serve as proxies for private market rentals.
A further question relates to the use of rentals from furnished dwellings to enlarge the basis for imputed rentals. In principle, the basis for imputing a rental value for owner-occupied dwellings is unfurnished rentals. Therefore, rentals for furnished dwellings cannot be used directly. In order to avoid imputing the wrong level of rental, they should be scaled down to exclude the payment for the use of the furniture.