the life course is most accurate. The sample drawn from all newly granted pensions is so large
that all social strata and many different types of life-courses are represented in suffiently large
numbers to enable empirical research on many different questions. Only pensions based on
own contributions are selected for this sample.1 These are old age and disability pensions.
For the data set ‘Completed Insured Life Courses’ all pension funds send the information
on the completed biography from school time up to the moment of retirement to the data
centre of the pension fund (Stegmann 2007). The scientific use file for social research
combines the longitudinal life-course information with the result of the pension calculation as
a cross sectional part of the data set. The cross-sectional part includes further demographic
information. This means that demographic variables mirror the social situation at retirement.
However, the socio-demographic position might in some cases have changed over the life-
time, e.g. immigrants with foreign nationality might be naturalised at retirement and also the
marital status might have changed over the lifetime. The longitudinal information is presented
on a monthly basis. For each month the data shows if the person was gainfully employed or
had another social situation like unemployment, care-giving or sickness. Child-care is
assumed to be the main occupation if the birth of a child is registered and no gainful
employment has taken place afterwards.2 Employment has priority status in the data and all
other social situations are second in rank. A lack of information means that a person is in none
of these social status situations at this time in Germany. Such a gap in information can stand
for self-employment without social insurance obligation, unemployment without being
entitled to benefits from the Federal Employment Agency or working abroad. However, in
most female biographies a missing information stands for a period of housekeeping.3 The
main drawback of these data is the lack of information on other sources of income before
retirement and after.
The data are accessible at the Research Data Centre of the German Pension Fund. A
smaller sample is drawn for the scientific Use File, which can be ordered for use in research
institutions, but larger samples up to full samples in the case of data on recorded pensions can
be use on the spot in the RDC.
1 This excludes survivors pensions.
2 The birth of a child is registered in the pension record of one of the parents. This is in most cases the mother, because there is an income
tap that hinders higher earners to profit from the child benefit. The child benefit in the German pension fund accounts for children born
before 1992 1 credit point, for children born after 1992 the contribution is 3 credit points.
3 This fact can be proven with the data from the AVID 1996 project, where process produced data were combined with survey data.