Pension benchmarks: empirical estimation and results for the United States and Germany |
| |
Authors: | Christian Dudel Julian Schmied |
| |
Affiliation: | 1. Max Planck Institute for Demographic Research, Federal Institute for Population Research, Max Planck–University of Helsinki Center for Social Inequalities in Population Health;2. Max Planck Institute for Demographic Research, Freie Universität Berlin |
| |
Abstract: | Benchmark replacement rates are a key parameter for retirement plans. Often, a pension level of around 70 per cent of net income during working life is considered as an adequate choice. However, this heuristic value is left unjustified, and data-based benchmarks are limited. In this paper, we propose to estimate a pension-level benchmark based on keeping the living standard achieved during working life constant after retirement. Applying parametric, semi-parametric and non-parametric estimation methods to data from the United States and Germany, we find that a net pension income of around 100 per cent of the last net working-life income, plus or minus 10 percentage points, is required to keep the living standard constant. However, we also find that the outcome of the exercise can depend on how ‘living standards’ are measured. |
| |
Keywords: | replacement rate retirement income semi-parametric estimation non-parametric estimation |
|
|