Occupational Coding in Germany

1
Occupational Coding in Germany

• Present State and Problems
• Markus Promberger/IAB

2
Two systems of coding

• 1988 System of Occupations
• Used by the Labour Administration
• Two purposes Labour statistics and job placement
  
• 4-digit level for statistical purposes
• 7-digit level for job placement services

• 1992 System of Occupations
• Used by the Statistical Office
• Statistical purposes (mainly cross-sectional)
• Only on 4-digit level

3
Differences between 1988 and 1992 Systems

• Identical on 2-digit level
• Huge overlap on 3-digit level
• Almost completely different on 4-digit level
• Pipeline Engineer 6079 (1988) or
• 6089 (1992)?
• Common ancestor 1970 System

4
Sources of data

• 1988 System/BA
• Employers reporting to health insurance, 3-digit
  level

• 1992 System/destatis
• Censuses (rare), 4-digit level
• Mikrozensus
• (anual population survey), 4-digit level
• Social insurance data, 3-digit level

5
Main problems

• 1. Best reporting on general employment only on
  2-digit level
• Labour administration data are at 3-digit-level,
  but leave out some kinds of nonstandard
  employment
• Mikrozensus data are at 4-digit-level, but
  survey-based, and the Mikrozensus is not a real
  LFS

6
Main problems

• 2. Certain inadequateness concerning
  ICT-occupations
• Both systems reflect occupational structure of
  the 1980s at the 4-digit level
• Both reflect the 1960s in their overall structure
• Extremely high differentiated in manufacturing
  occupations
• Quite low differentiated in services and ICT
  occupations, if they appear at all, they are old
  fashioned

7
Main problems

• 3. Improper connection to ISCO-88
• Even the grandfather of both german coding
  systems (1970 system) missed the connection to
  ISCO-68
• Translation to ISCO-88 is possible, but only with
  a severe loss in exactness
• There is no exact equivalent in no direction.
  Often more than one possibility to translate from
  1992 Sys. into ISCO-88
• Experts do not regard it as impossible to develop
  a real translation key, but this is not the task
  of STILE WP3

8
Conclusions for WP3

• Drop the coding experiment for Germany
• Do double sector coding instead for Germany, one
  by the statistical office, one by the labour
  administration.
• NACE rev.1 and the german WZ2003 are well
  connected

9
...thanks for your attention!

• Occupational Coding in Germany Present State
  and Problems
• Markus Promberger/IAB