Changes in the Income Distribution of the Dutch Elderly between 1989 and 2020: a Dynamic Microsimulation


  • Note: The authors thank Stichting Instituut GAK and Netspar for their financial support. Furthermore, we wish to thank Wim Bos (CBS), Arjan Soede (SCP), and Centrum voor Beleidsstatistiek for tracking down the changes in the IPO data due to the revision in 2000. Finally, we thank Hans-Martin von Gaudecker, Didier Fouarge, Koen Caminada, participants of EALE 2009, participants of the conference “Labor Force Participation and the Well-Being of the 50+ Population” in Utrecht (2009), participants of the European Workshop on Dynamic microsimulation modeling in Brussel (2010), participants of the Econometric Society World Conference 2010, and two anonymous referees for their useful comments.

Correspondence to: Marike Knoef, Leiden University, Steenschuur 25, 2311 ES Leiden, the Netherlands (


This paper analyzes the income distribution of the Dutch elderly using a microsimulation model. Microsimulation models allow for detailed estimates of the income distribution. Our model deviates from traditional models by explicitly considering the persistency and heteroskedasticity of real income shocks. In this way, modeling all underlying processes influencing household income becomes less necessary, which can improve the trade-off between refinement and tractability of microsimulation models. We show the results of three model specifications with different levels of refinement. The results are in line and indicate that between 2008 and 2020, the highest predicted annual growth among the elderly is for median-income households (about 1.2 percent). High-income households have a somewhat lower predicted growth (about 1.0 percent) and low-income households only have a predicted annual growth of 0.5 percent. Inequality therefore seems to increase in the lower part of the distribution, while it will probably decline in the upper part of the distribution.