Skip to content

Accuracy and reliability

Contact info

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

pal@dst.dk

Get as PDF

Construction

In general, the latest figures for the construction must be regarded with reservations, because the data source is uncertain. The main source of inaccuracy is the delays in the municipality's data registrations in the Central Register of Buildings and Dwellings. The estimates made by Statistics Denmark attempt to address this problem. The model estimate the delays for the latest 18 months and revises the figures in respect hereof. The estimate is encumbered with uncertainty, because there is not a steady pattern in the size and causes of the delays.

Overall accuracy

A survey of the overall accuracy of the statistics has never been conducted. In general, the latest figures for the construction must be regarded with reservations, because the data source is uncertain. The statistics on construction are based on the database (the Central Register of Buildings and Dwellings/BBR) which contain all building cases and buildings in Denmark. The main source of inaccuracy is the delays in the municipality's data registrations in the BBR.

Therefore, the "raw" figures (uncorrected figures) are marked by great uncertainty in connection with their first publications. The delays varied widely, depending on the building phase and if it is total construction or construction of dwellings. Analyses show that most of the delays are received after 6 months and after 18 months and figures are regarded as "final", because there are only few and minor revisions after that time.

Generally, the quality can be evaluated to be of a reasonable quality which with some reservations can be used to estimate the market trend in construction. Normally, there are some different degrees of underestimation until the figures are being "final" after 18 months.

Sampling error

Not relevant for these statistics.

Non-sampling error

The construction statistic is register based (BBR), which in principle covers the target population. Our population differentiated a little from BBR:, because there are removed some building cases, which is not regarded as relevant for the construction statistic (please consult the section "Coverage error" for further details ). Furthermore are there added some building cases, which of different reasons not are included (please consult the section "Coverage error" for further details).

  • Only recorded building project, i.e. constructions without building projects are not included - however there is a exception, construction of small buildings (garage, carport and outhouse) under 50 square meter, please consult the section "Data compilation" for further details.
  • A range of "wrong" building projects are erased, e.g. doublets, please consult the section "Data validation" for further details.
  • Only buildings cases with a increase in dwellings and/or total floor space are included, which imply that demolitions and repair- and maintenance work are not included, if they not result in a increase in dwellings and/or total floor
  • Measurement errors appears, typically errors in number of dwellings. Data validation catch this and they are been corrected in co-operations with the municipalities
  • The problem of non-response is primarily a question of delayed reporting to the register. A model for estimation seek to corrected for this
  • If any miscalculations they will be corrected and there will be published revised figures
  • The statistical model is generally speaking determined by BBR. But there are carried out some delimitations in relation to what are published, see section on "Model assumption error"

The problem concerning non-response hangs basically about delayed reports to the register (BBR). Through the estimation model do we make an attempt to take corrective actions (see section on "Non response error" ).

Quality management

Statistics Denmark follows the recommendations on organisation and management of quality given in the Code of Practice for European Statistics (CoP) and the implementation guidelines given in the Quality Assurance Framework of the European Statistical System (QAF). A Working Group on Quality and a central quality assurance function have been established to continuously carry through control of products and processes.

Quality assurance

Statistics Denmark follows the principles in the Code of Practice for European Statistics (CoP) and uses the Quality Assurance Framework of the European Statistical System (QAF) for the implementation of the principles. This involves continuous decentralized and central control of products and processes based on documentation following international standards. The central quality assurance function reports to the Working Group on Quality. Reports include suggestions for improvement that are assessed, decided and subsequently implemented.

Quality assessment

A survey of the overall accuracy of the statistics has never been conducted. In general, the latest figures for the construction must be regarded with reservations, because the data source is uncertain. There are several "elements of uncertainty" concerning a correct estimation of the provisionally figures for construction. For the same reason, the statistics on construction do not operate with final figures, the figures are revised back to 1998 in every publication.

The main sources of inaccuracy are:

  1. The observed delays in the municipality's registrations. This is the greatest element of uncertainty. The definition of a delayed report is that the report does not have a bearing on the actual month of reference. It is difficult to state on single quality assessment, because the series, permitted, started, under construction and completed construction, have different sizes of delays. Since 2006 the size of delays for the "raw" figures (the uncorrected figures) is made up for started and completed construction. The figures can be seen in Statistikbanken/Business sectors/Construction/BYGVFORS. The statement shows the percentage growth of the first reported figures for a quarter up until today. It shows - together with other analyses - that: - The greatest delays are fund within the started construction, especially the started dwellings with a average quarterly per cent of delay of about 230 p.c., - The smallest delays are fund within the completed construction, where there is no big difference between the total construction and construction of dwellings with an average percent of delay of about 50 p.c., - The size of delays in permitted construction is considered to be between the started and the completed construction.
  2. The quality of the reported figures: Concerning the central variables, primary areas and number of dwellings, there is a great effort to secure the quality of these figures. Systematic surveys of the quality has never been conducted, but more analyses indicate that the quality is reasonable.
  3. The quality of the model of estimations which try to estimate what the statistics would have been if there were no delays. The model of estimation is described in the section about data compilation. It has appeared to be very difficult to find a stable pattern in the delays, because there are a great variety of causes, e.g. few resources in the municipalities, inexpedient administrative procedures and the builders often lacking/delayed reporting to the municipalities. All models of estimation from 1996 had a basic problem, namely a lacking capacity to catch short time changes in the market trends. All experiences show that the higher construction activities, the more delays and vice versa. Therefore, the model will in the short term underestimate or overestimate changes in the building activity.

There is define a success criterion that the first estimate shall be inside +/- 5 p.c. of the "final" figure (after 18 months).

The total quality of the statistic must be considered as reasonable. There are a row of conditions which must be considered when the construction statistics shall be used. With these reservations bear in mind is the statistics however a reasonable indicator for the business trends in construction.

Data revision - policy

Statistics Denmark revises published figures in accordance with the Revision Policy for Statistics Denmark. The common procedures and principles of the Revision Policy are for some statistics supplemented by a specific revision practice.

Data revision practice

Because of the problem of delays of the report to the register on buildings and dwellings, final figures are not produced. The figures are revised in step with the received delayed building cases.