Skip to content

Comparability

Contact info

Short Term Statistics, Business Statistics
Paul Lubson
+45 39 17 35 42

pal@dst.dk

Get as PDF

Construction

The greatest obstacle for the comparability over time is the change in 1981 from the survey-based statistic to the register-based statistic (BBR). It gives both a break in time series and a market extension of the descriptive variables, see the section on comparability over time.

Comparability - geographical

Every month figures for permitted construction are submitted to the statistical office of EU, Eurostat. Data for all EU countries can be found in the Eurostat database. The statistics are produced following the principles of an EU regulation, so the results are comparable

Comparability over time

Full data comparability over time is only possible from the year 1980. Before 1980 the statistics for construction were compiled on the basis of surveys and there are different geographic degrees of coverage and some differences in definitions. There are conducted different projects where attempts are made to carry the comparability longer back in time:

For the completed construction activities, estimates are available (the coverage compared with BBR):

  • January 1950 - March 1960: The metropolitan area plus the 5 largest towns and their suburbs (degree of coverage: around 35 p.c.).
  • April 1960 - December 1968: The metropolitan area plus the 18 largest towns and their suburbs (degree of coverage: around 50 p.c.).
  • January 1969 - March 1970: Urban areas with more than 10,000 inhabitants (degree of coverage: around 50 p.c.).
  • April 1970 - December 1972: Urban areas (degree of coverage: around 70 p.c.).
  • January 1973 - December 1979: All Denmark (degree of coverage: 100 p.c.)

Furthermore, for the year 1979 the statistics of construction activities, based on questionnaires, are compared with the statistics based on the Central Register of Buildings and Dwellings. The new method proved to have 20 p.c. more floor area and 5 p.c. more dwellings, measuring the start of constructions.

There are established some historical series in Statistikbanken, see Historical summaries:

  • BYGV04: Total construction back to 1939, by phase of construction and use
  • BYGV05: Residential construction back to 1917, by phase of construction, use and builders
  • BYGV06: Average floor area in new-constructed dwellings back to 1916, by use

Since the publication of 3th Quarter 2010, the statistics are based on a changed BBR, which has resulted in some correction in the figures back to 1998.

The main consequences are seen within construction of dwellings:

  • 1998-2002: a reduction in number of permitted, started and completed dwellings of about 1 p.c.
  • 2002-2005: a reduction in in number of permitted, started and completed dwellings of about 4 p.c.
  • After 2005: a rise in number of permitted, started and completed dwellings of under 2 p.c.

For the total construction the rise has only be under 1 p.c. for permitted, started and completed construction.

The change of method in connection with publishing 2nd quarter 2015 where temporary buildings are removed back to 1st quarter 2010 is the estimate that it will have minimal consequences because there are very few registered temporary buildings before 2010. The corrections back to 2010 are about 1 p.c. per. year for both construction of dwellings and area.

Since 31. August 2016 become the codes for use for both buildings and units more detailed and standardized. Until the changes are completely implemented, will the old codes for use been used, with some exceptions: - The level of building: The valuation is that the break in data will be minimal, because it by and large only is in connection with the creation of the new codes for annex there will be relocation between subcategories, e.g. from detached houses to Other residential buildings. For the main categories of use there will not occur changes. The statistics for construction is based on the level of building - The level of unit: It has been necessary to make a break in data - however inside the main category, commercial units, when 3 old codes (340, 350 and 360) are closed and are gathered in a new code (321, office)

In June 2017 came a new version of BBR. This has result in new methods, altered data structure and new, removed or changed variables etc. compared to the replaced version of BBR. This had resulted in two forms for data break (for the present, because the new version is first implemented in 2019): 1. A "temporary" data break concerning permitted and started construction, in particular for dwellings. This is a result of, that the information about "expected number of dwellings" first is re-establish in February 2018, i.e. after the publication of the construction statistics for 4Q 2017. It is expected ahead that it will increase the number of permitted and started dwellings in the first reporting, so that they - as minimum - will come on the level, as they have been before. 2. A minor data break in the time series, because there now come in more square metre in the construction statistics. This is the result of the changed method concerning small buildings (see 3.05) which lead to more square metre (catch more small buildings) and primary a changed duplicate searching, where fewer duplicates are removed. The changed duplicate searching has the effect that in the period 2006-15 there is coming 2.3 pct. more started square metre and 5.8 p.c. more started dwellings in average for all the years (test, February 2018). The consequence of the extension/change in including types of dwellings (all dwellings in residential institution and only legal permanent habitation in summer cottage) uncertain, because they pull different ways, namely more dwellings in residential institutions and fewer dwellings in summer cottages

Coherence - cross domain

The only statistics where comparability is possible are the statistics on the building stock and the housing censuses.

Because of several differences complete data comparability is not directly possible:

  • The problem of delays in the construction statistics.
  • The registrations of burnt and demolished buildings are not complete in the construction statistics. The statistics on constructions give a gross figure for the number of new dwellings and new floor area. The housing censuses give net figures
  • Comparing the housing censuses and the construction statistic the definition of dwellings differs. The housing censuses include buildings and units of buildings which are linked to an address in the Central Population Register (CPR). The construction statistics include dwellings registered as all-year-residential-dwellings, e.g. excluding weekend cottages.
  • Furthermore, data registered directly in the building stock, without building permits, are observed.

Coherence - internal

Seasonal adjustment is performed indirectly, which results in consistency between seasonal adjusted aggregates and their subcomponents. There are no other sources for possible internal inconsistency.