The individual data concerning each local unit and each employee should be provided in the form of two types of micro-data record:
Arecords for the local units
Brecords for the employees.
The employee records should be linked to the local unit records by a key which does not disclose the identity of the business. This could be an artificial number or an existing key, provided the same key is used for both the local unit and the employee records. Additionally, for easy identification of an employee, a key for each employee should be provided which does not identify the person.
All items for records A and B should be completed in full. There should be entries for all individual items, including optional variables (see below). ‘Blank’ information is not acceptable.
Complete information should be supplied for all mandatory variables on all micro-data records. Otherwise, the grossing-up factors supplied will not be suitable for all variables.
These should be coded strictly according to the following rules:
When information is available for an optional variable, data should preferably be provided (as for mandatory variables) for all observation units or employees.
When information is not available for an optional variable, insert ‘OPT’ for alphanumeric variables and ‘99999999’ for numeric variables (the number of ‘nines’ corresponds to the length of the field).
‘0’: this should strictly be used only for those variables where a zero value may sometimes genuinely occur (e.g. when an employee has no overtime in the reference month).
The SES variables listed in Annex I are either qualitative or quantitative variables. Most of the variables related to the local unit and to individual characteristics of the employees are qualitative variables. These SES variables are linked to a limited number of categories.
The quantitative SES variables are either count variables (number of employees, number of hours, days or weeks) or variables related to earnings.
The categories of the qualitative SES variables as well as bands for quantitative SES variables are set out by Eurostat in an implementation paper.
The content and sequence of the variables in records A and B and the codes to be used will be specified in separate documents.
The records should contain one field per variable. Numeric variables should be expressed in absolute terms, that is, giving the numbers in full (not in decimals, or in tens, hundreds, thousands, millions, etc.). However, because of the need for precision, the values for variables 4.3, 5.1 and 5.2 should be given to two decimal places.
There is one numeric variable that is different. This is the ‘share of a full-timer’s normal hours’ (variable 2.7.1), which should be expressed as a percentage and also given to two decimal places, e.g. 43,27.
Where numeric variables relate to money values (e.g. hourly, monthly annual earnings and bonuses, taxes, social-security contributions, payments in kind), these should be expressed in units of the national currency of the country concerned.
Member States shall transmit to the Commission (Eurostat) the data and metadata required by this Regulation in an electronic format compliant with an interchange standard proposed by Eurostat. Eurostat will make available detailed documentation in relation to approved standard(s) and will supply guidelines on how to implement this (these) standard(s) according to the requirements of this regulation.