Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Due to demographic change the replacement rates of the German statutory pension scheme will decrease over the next decades. Voluntary savings for retirement will therefore gain more and more relevance in order to maintain one's standard of living during retirement. This article examines the savings behavior for retirement on an individual level in Germany. As a first step the decision to save at all is analyzed, showing that the main determinants for saving are personal income as well as the disposable household income. Furthermore migrants and individuals living in the Eastern part of Germany turn out to be less likely to save additionally privately for retirement. In a second step the chosen gross saving rates are analyzed using a Tobit, a lognormal hurdle model and a Type II Tobit Model. The results suggest that the decisions to save at all and about the saving rate are independent of each other leading to a loss of information if only a standard Tobit model is used. For example personal income increases the probability to save for retirement but decreases the resulting saving rate. Modelling both decisions separately therefore leads to a better understanding of the determinants of saving for old-age.
Terms of use:
Documents in
JEL-Classification: D14, D91, H31Keywords: Savings, retirement, life-cycle, two-part model, tobit, exponential type II tobit *Institute for Public Finance I, Albert-Ludwigs-Universität Freiburg, Bertoldstr. 17, 79085 Freiburg,. For helpful comments I would like to thank Daniel Ehing and Stefan Moog.