Quantile treatment effects of Riester participation on wealth

In numerous industrialized countries the demographic change erodes the financial basis of traditional pay-as-you-go pension systems. To compensate for decreasing statutory pensions, many governments incentivize private saving by means of subsidized retirement plans. In this context, Germany introduced the so-called Riester pension plans. To assess its effectiveness, this paper analyzes the effects of participation in Riester plans on wealth at different points of the distribution. We employ an instrumental quantile regression approach using Riester eligibility as instrument for Riester participation. The analysis is based on microeconomic survey data from the German Socio-Economic Panel of wave 2012. Results suggest substantial heterogeneity in the effect of Riester participation on wealth. While Riester participation increases total net wealth in the lower tail of the conditional distribution, it does not have a significant effect on households in the middle part of the distribution. In the upper tail of the conditional asset distribution, we find negative treatment effects providing weak evidence in favor of a mere reallocation of households' asset portfolios.