The Admin4 Linked Employer–Employee Database in Hungary
by Rita Pető
This paper introduces and documents the Linked Administrative Panel Database (Admin4), a large-scale linked employer–employee dataset covering a representative 50% sample of the Hungarian population between 2003 and 2021. Admin4 links individual- and firm-level records from multiple administrative authorities and provides longitudinal information on employment histories, earnings, working time, occupations, unemployment, social transfers, education, health care, pensions, and firm financial characteristics. Unlike most linked employer–employee datasets, Admin4 also includes inactive individuals and registered jobseekers, enabling the analysis of labour market entries, exits, and transitions across employment status for the entire population.
Structure and content
The database is organised in a modular structure, consisting of a core employment file and several supplementary modules that can be linked through anonymised identifiers. This design allows researchers to adjust the scope of the data to specific research questions while maintaining population coverage and longitudinal consistency.
The core module contains information on employment relationships, earnings, working time, occupations, and employers. Supplementary modules extend the database with records on unemployment, social transfers, pensions, education, health care use, firm balance sheets, and selected family links. Observations are available at monthly frequency for most variables.
Population coverage
Admin4 is based on a random 50% sample drawn from the national social security register and is continuously updated for births, deaths, and migration. It includes not only employees in the public and private sectors, but also the self-employed, registered unemployed, students, parents on leave, and pensioners. This allows labour market transitions to be studied for the broader population rather than for workers only.
Research applications and limitations
The database supports empirical work on topics such as labour market mobility, wage inequality, firm–worker matching, policy evaluation, and the links between employment, education, health, and social benefits. At the same time, the dataset has limitations typical of administrative sources. Informal employment is not observed and earnings are limited to contribution-based income.
Outlook
Future updates (Admin5 and later) are planned to extend the time series and maintain the panel structure. This will allow continued analysis of labour market dynamics and policy changes in Hungary using a consistent administrative data infrastructure.
Read the full working paper here.
Authors: István Boza, Rita Pető, Melinda Tir
Authors: István Boza, Rita Pető, Melinda Tir
