Statistical processing

Source data

The primary data are obtained from a range of sources, including the e-income register, central business register, income statistics, statistics on persons receiving social benefits and the central population register.

Frequency of data collection

The statistics are compiled annually, but data for the statistics are continuously collected.

Data collection

The data basis for the labor market accounts are various statistical registers in Statistics Denmark, all of which are based on administrative data.

Data validation

There is uncertainty attached to the information on self-employed persons. This is due to the circumstance that the hourly information for the self-employed is imputed. The information is considered to be correct at a general level in relation to the variables on which the information is imputed upon, but the information – especially at detailed level – must be interpreted with caution.

The information on paid hours of work is obtained from the statistics on employees in employment, where paid hours of work are, in some cases, imputed and consequently the information is uncertain.

Employers are legally obliged to report information on the workplace at which the individual job is performed. However, there are a number of employees, especially within the public sector whose reports are insufficient in this respect. By means of, e.g. Statistics Denmark´s central business register, the data reported for these workplaces are subjected to editing, and the job is attached to a workplace, which is Statistics Denmark´s statistical unit for job attachment.

On the basis of the LMA, it is possible to compile the population´s labour market status at arbitrary points-in-time during the course of the year. However, Statistics Denmark has only tested the compilation of labour market status by the end of November during the different years. This implies that there may be some degree of uncertainty linked to the compilation of labour market status at other points-in-time during the course of the year, and uncertainty can especially be linked to the statistical compilations in December.

Data compilation

Data processing in the LMA is performed in several stages. They are as follows:

1. Input of data in the source database
2. Data subject to processing of overlaps 1
3. Data subject to processing of overlaps 2
4. Various linkages, etc.

Below, a short description is given of these 4 stages.

1) Input of data in the source database

The first stage of the data processing involves data from different sources are processed and entered into a coherent and homogeneous source database.

For example, data are entered from the statistics on peoples receiving public benefits, statistics on employees in employment, data for recipients of sickness benefits and maternity benefits and data for person receiving education and training, including course participants.

Data on self-employed persons and assisting spouses are also entered, where the data are obtained from various data sources. At the same time, data on paid hours of work for self-employed persons and assisting spouses are also imputed on the basis of a range of background variables. This information is also entered into the source database.

2) Data subject to processing of overlaps 1

In this stage, the various data sources from the source database are subjected to processing of overlaps. During the processing of overlaps, so-called “illegal overlaps” are corrected by deleting or reducing the various states or correcting the to- and from-dates. In this way, a better periodicity of the various states of the population is thus achieved.

An example of an illegal overlap is if a person receiving unemployment benefit and at the same time is registered as being in employment. In this context, it has been decided that the information on unemployment benefit is considered to be the most valid. In order to rectify this overlap, attempts are made to find periods during the month in which the person could, alternatively, have been employed, so that the to- and from-dates for the job can be changed. If there are no periods, where the person could, alternatively, have been employed, the number of hours of the employment is reduced.

A number of linkages of the various states are also performed. For example, the states are linked with subsidized employment to the job held by the person. This same also applies to, e.g. persons who are temporarily absent, so that the job from which the person is absent is known and whether the person is absent from employment or unemployment.

Furthermore, a selection of self-employed persons and assisting spouses is also performed on the basis of a wide range of information on these persons.

On the basis of the first processing of overlaps, the so-called AMR-UN register is set up, which constitutes a non-hourly standardized longitudinal register storing information on the population´s labour markets status.

3) Data subject to processing of overlaps 2

In the processing of overlaps 2, an hourly standard rate of the population is performed, implying that the population´s labour market status can be compiled in terms of full-time persons.

This hourly standard is based on the existing hourly standard, i.e. 37 hours per week. Subsequently, a full-time person corresponds to 37 hours and a person can at maximum contribute with 37 hours in the LMA. This implies that if a person, who is, e.g. receiving early retirement pay and who is at the same time holding a job, the person will, e.g. be included as a 0.9 full-time person in relation to the recipient of early retirement pay and a 0.1 full-person in relation to being in employment.

If a person works for more than 37 hours, a reduction/redistribution of the hours is performed on the basis of the validity that the information is assess to have. When the information is considered to be equally valid, employment is given the highest weight.

In connection with the compilation of full-time persons, a range of special conditions apply, depending on the person´s labour market status:

• Persons in employment are included with the number of hours in which they normally work, i.e. independent of whether the persons are temporarily absent from their job. However, there is a limit that hours above 37 are not included. If a person has several jobs, where the number of hours worked together exceeds 37 hours, the jobs are proportionally written down.

• For some types of temporary absence from employment, there are no data reports on the jobs. This applies to, e.g. temporary absence, due to sickness or maternity leave. In these cases, both jobs and hours worked are imputed. This implies that the persons, who are absent from work, are included on the basis of this state, together with the number of hours that they normally work.

• Persons in subsidized employment receive in some cases payment for more hours than they have actually worked. However, persons in subsidized employment in the LMA will only be included with the number of hours they have worked in the subsidized employment.

• A person can receive various types of public benefits at the same time. When the various types of public benefits are weighted, the number of hours is reduced in the LMA in accordance with the same principles as those applied in the statistics on persons receiving public benefits. This implies that in the case of an overlap, e.g. between an employee job, guidance and activities upgrading skills and social assistance (passive), a reduction of the number of hours is first conducted concerning social assistance.

4) Various linkages, etc.

There is no indication of the hours in the various states for persons outside the labour force, who are not included in the statistics on persons receiving public benefits. This applies to children, young people, persons receiving education and training and old-age pensioners.

In this stage, these hours are fixed in such a way that the persons are included as being in the state applying with 37 hours. If a person appears in several categories, a priority is conducted among these categories in accordance with the order of priority below.

1. The state in the data subjected to processing of overlaps
2. Children and young people
3. Old-age pensioners
4. Receiving education and training
5. Recipients of educational grants
6. Course participants
7. Students at production schools
8. Other persons outside the labour force

A wide range of person- and workplace-related information on the population obtained from other data sources is also linked to the register.

On the basis of this, the LMA are compiled, which constitute hourly standardized longitudinal register containing information on the population´s labour market status.

Adjustment

Only corrections described under the section ”Data processing” and ”Quality assessment” are performed to the data.