To what extent are individuals’ incomes linked to those of their parents? Unlike previous studies on France, the authors of this study measure income at household level, which provides a more accurate picture of socio-economic status than individual income alone.
Intergenerational income mobility in France: a comparative and geographical analysis
This paper examines intergenerational income mobility in France, focusing on children born in the 1970s.
Reference IPP Policy Brief No. 95
Presentation
Key Results
- France is characterised by low intergenerational income mobility compared with other developed countries. Only 9.7 per cent of children from the 20 per cent of households with the lowest incomes end up in the 20 per cent of households with the highest incomes as adults – four times fewer than children of parents in the 20 per cent of households with the highest incomes.
- The likelihood of obtaining a higher education qualification increases significantly with parental income. Children from disadvantaged families are 2.5 times less likely to obtain a higher education qualification than those from very well-off families.
- Intergenerational mobility varies considerably across the country and appears to be particularly closely linked to the local unemployment rate.
- Geographical mobility between childhood and adulthood is associated with a significant increase in upward mobility. Individuals born to parents on the lowest incomes who move to a high-income department achieve, on average, the same income level as children from well-off families who have not moved.
Method and Data
Sources
The source used for this study is the Permanent Demographic Sample (EDP), produced jointly by the National Institute of Statistics and Economic Studies (INSEE) and the Directorate-General for Public Finances (DGFiP). Since 1968, this database has included information from various administrative sources (civil registration, population censuses, salary data and tax returns from 2011 onwards) for a representative sub-sample of the French population (individuals born during the first four days of October).
Sample
The study focuses on individuals born in mainland France between 1972 and 1981, who were recorded with their parents in the 1990 census and who declared income between the ages of 35 and 45. The method used requires us to exclude children with a parent who is a working farmer or practises a liberal profession. In total, our analysis sample comprises 64,571 child-parent pairs.
Prediction of parents’ income percentile
The EDP makes it possible to track the trajectories of individuals born in the first four days of October, whom the authors refer to as ‘EDP individuals’. Their family members are not tracked, unless they too were born on an EDP day. Thus, EDP individuals observed as children in the 1990 census are tracked in the income data, but not their parents. Similarly, EDP individuals recorded as adults in the 1990 census are tracked in the income data, but not their children. The authors select EDP individuals who had children between 1972 and 1981 (in mainland France) and recorded in the 1990 census to econometrically model the link between the socio-demographic characteristics of the 1990 census and the average gross wage between the ages of 35 and 45.
The results of this modelling are used to predict the gross wages of the parents in the sample of children. The characteristics used to make these predictions are educational attainment, socio-occupational category, a set of demographic characteristics (year of birth, nationality, country of birth and household structure), as well as characteristics of the local authority area of residence in 1990 (unemployment rate, proportion of single mothers, proportion of foreign nationals, population size and population density). All these characteristics enable the authors to estimate the parents’ position in the income distribution with sufficient accuracy.
Validation of the method
The authors replicated this analysis using US data from the Panel Study of Income Dynamics, a source that not only allows the same prediction methodology used in the study to be applied, but also enables direct observation of parents’ income. A comparison of the results obtained using the two methods suggests that parents’ position in the income distribution can be reliably predicted on the basis of their observable characteristics, even though this approach leads to a slight underestimation of intergenerational income persistence for the indicators used in this note.
Louis Sirugue
Partners
This paper was supported by the Chair in Education Policy and Social Mobility. Established in 2021 as part of a partnership between the Ardian Foundation, the Directorate for Evaluation, Foresight and Performance of the Ministry of National Education and Youth (MENJ-DEPP) and PSE-Paris School of Economics, this Chair aims to promote high-level research and the dissemination of knowledge on education policy and social mobility.
This work has also received state funding administered by the French National Research Agency (ANR) under the ‘Investissements d’avenir’ programme, reference ANR-10-EQPX-17 (Centre for Secure Data Access – CASD), from the ANR JOCE programme (reference ANR-18-CE41-0003-02), and from the Convergences Migrations Institute, supported by the CNRS (reference ANR-17-CONV-0001)