The Swedish income panel was originally set up in the beginning of the 90s to make studies of how immigrants assimilate in the Swedish labour market possible. It consists of large samples of foreign-born and Swedish-born persons. Income information from registers is added for nearly 40 years. In addition income information relating to spouses is also available as well as for a subset of mothers and fathers. This makes it possible to construct measures of household income based on a relatively narrow definition. However, starting in 1998 there is also more information making it possible to include children over 18 and their incomes in the family. By matching with some different additional registers information has been added for people who have been unemployed or involved in labour market programmes during the 90s, on causes of deaths for people who have deceased since 1978 and on recent arrived immigrants from various origins. It has turned out that the data-base is quite useful for analysing research-questions other than originally motivating construction of the panel. The panel has been used for cross country comparisons of immigrants in the labour market and to analyse income mobility for different breakdowns of the population, and analyses the development in cohort income. There have been analyses of social assistance receipt among immigrants as well as studies of intergeneration mobility of income, the labour market situation of young immigrants and the second generation of immigrants. On-going work includes evaluation of labour market training programmes and studies of early retirement among immigrants. Planned work includes studies of the economic transition from child to adulthood during the 80s and 90s as well as studies of how frequent immigrant children are subject to measures under the Social Service Act and the Care of Youth Persons Act. The potentials of the Swedish Income Panel can be understood if one compares it with better known income-panels in other countries. For example SWIP covers more years and has a larger sample than the German Socio-Economic Panel (GSOEP). On the other hand, the fact that information is obtained from registers only makes this Swedish panel less rich in variables. There are striking parallels between the Gothenburg Income Panel and the labour market panel at the Centre for Labour Market and Social Research in Aarhus for the Danish population.

The samples are taken from a register of the total population (RTB) kept by Statistics Sweden. This register is frequently updated and covers all persons registered as residing in Sweden. Excluded are asylum-seekers waiting for a residence permit. When a person emigrates or dies this should be recorded in the register. As every person residing in Sweden should have a personal number this can be used to identify persons in various registers. Merging such information requires a permit from the data protection agency. The merging has been made by Statistics Sweden. The research group has access to data-sets not containing personal numbers. It is widely believed that the register of the total population has a high quality in many, but not all respects. Undocumented immigration has not been a big issue in public debates in Sweden. However, there are reasons to believe that undocumented emigration to non-Nordic countries is underreported. This means that some (mainly foreign-born) persons registered as living in Sweden actually have left the country. From the register of 1978 a 1-percent sample of native born persons (about 77 000 individuals) was taken as well as a 10-percent sample of foreign born persons (about 60 000 individuals). Supplementary 10-percent samples of people immigrating each year 1979 until 1998 was also taken (sample sizes varies between 3 000-7 000 individuals). For immigration to Sweden during 1999 and 2000, it is not a sample but the data contain all cases. The changed design allows studies of immigrants who arrived those years to be made in larger dept. However, the changed design must be taken into account in case a user in interested in making repeated cross section analyses. In such cases it is advisable to take a 10 percent sample immigrants who arrived in 1999 and 2000. This research strategy enables an analyst to construct yearly cross-sections for the adult Swedish population for each year 1978, 1979, ... For each individual in the panel there is also information on the (present) spouse to the person sampled. Also cohabiting adults are regarded as spouses in case they are mother and father to the same child. This makes it possible to construct variables measuring family income and thus studies of family disposable income. Another attractive feature of the Swedish Income Panel is income information referring to parents for a subset of the sample. The subset constitutes of sampled persons born 1941 and later for whom mothers and fathers could be identified using a key. Identification was successful in most but not all cases. The possibility to link individuals to records for parents makes it also possible to identify native-born persons having a foreign born parent (´the second generation of immigrants´).