RAS is based on the Labour Market Account (LMA) which is a longitudinal register. RAS is done by taking a status (on the populations primary attachment to the labour market) on the last working day in November based on LMA. However with LMA it is also possible to take a status on arbitrary days in the year, and AMR can also be used for various process analysis etc.
Since April 2015 Labour Market Account (LMA) is the data foundation for RAS. In connection to the publication RAS was revised back to November 2008. In the same context the dating of the statistic was changed so that its now dated according to the time of reference at the end of November. That means that the latest count is called end November 2019, where it would previously have been called 2020. The data sources in LMA are various internal and external registers, e.g.:
RAS is a yearly statistic.
The data collection is done by separate processing of each source register. After that a transverse data processing is done (also called treatment of overlaps) where information in the various registers are compared, and corrected when needed. Finally data are linked to other registers to add background information and form the population.
The data foundation for RAS has since April 2015 been The Labour Market Account (LMA). LMA is produced both with and without an hourly standardization. The non-hourly standardized longitudinal register (LMA-UN) is the data foundation for RAS, and therefore the data validation takes place in LMA-UN.
In connection to the production of LMA a comprehensive validation of each input data is done. The most important ones are:
The data source for wage earner jobs are the eIncome register. The wage earner jobs contains information about which workplace the job is at. The workplace is the foundation for the information about industry, sector and geography. In some cases the reporting from the employer are incorrect. In that case a correction of the errors are conducted. The eIncome register contains information about work function (DISCO-08) for persons employed at workplaces covered by the wage statistic. If the workplace is not covered by the wage statistic the information about work function comes from the work classification module when it's available here. In addition other errors are corrected by the eIncome register.
The data source for information about self-employed is the business register, the income statistic, the eIncome register and the unemployment statistic. These sources are individually validated at the formation of information about self-employed.
The data source for information about absence due to sickness and maternity leave is the statistic of maternity leave and sickness benefits. Data is processed a great deal compared to e.g. a temporary determination of whether the absence is from employment or unemployment.
Transverse data treatment/data validation
The purpose of the transverse treatment/validation of data is to erase, correct or create labour market conditions in cases where the various data sources do not coincide. This is done by a so called treatment of overlaps. The rules used for the treatment are complex. Here some of the most important areas are mentioned:
After that data is connected to other registers etc.
The data compilation in LMA takes place in several steps. The first step in the data processing is to identify and correct errors in data from various sources, and put data in one joint and homogenous source data base. From different statistics data on public benefits, wage earners, self-employed, assisting spouses, persons in education, maternity leave and sickness benefits are joint. An imputation of the paid hours for self-employed and assisting spouses are also done. Afterwards 'illegal overlaps' between conditions are being corrected, and connections between various conditions are made.
After the processing of overlaps a classification of the population’s attachment to the labour market are made on the basis of the international guidelines from ILO, which is further described in item 2.2 Classification system. The guidelines consist among others of a set of rules for prioritizing the primary attachment to the labour market. The guidelines dictates that employment is prioritized higher than unemployment, while unemployment is prioritized higher than conditions outside the labour force. Data is also linked the business register to get background information (industry, sector, address for the work place) concerning the work places where the employed persons work. Linking to the population register is also made with the aim of deciding whether the person is resident in Denmark at the time of reference.
From this processing of data the so called LMA-UN is made, and this non-hourly standardized longitudinal register with information about the populations connection to the labour market is the data foundation for RAS. In RAS the populations primary attachment to the labour market are calculated at the end of November, but on the basis of LMA-UN it is possible to calculate the attachment to the labour market at arbitrary times in the year.
LMA is also done in a hourly standardized variant (LMA-TN). In this register every person always has an attachment to the labour market on 37 hours per week. Some of the hours can e.g. be as employed while other hours can be in education, or simply hours where it's not possible to determine a socioeconomic category. Learn more about statistical processing in LMA.
No corrections of data besides what is described under data validation and data compilation.