Benchmark replacement rates are commonly used to set up saving plans or to assess retirement preparedness. An open question is whether high earners need the same replacement rate as low earners. In this paper, I apply the GAESE framework, an approach known from the equivalence scale literature, to assess how the replacement rate that maintains income satisfaction through retirement relates to income levels. Using longitudinal data from the German Socio-Economic Panel, and applying fixed-effects ordered logit models, I find that the benchmark replacement rate decreases with income. For singles, this finding is consistent across many modifications of the approach, whereas for couples the finding is sensitive to the composition of the retiree household, i.e. whether or not the retiree’s partner is still employed.
