16
r∕*≈0.93 (0.02, я =4)
rʃ ≈2.74 (0.17, n= 2)
The elderly consider a small loss for the younger generation on their investments in the public pension
system as fair. The optimal fairness ratio for the youngest generation as conceived by the middle-aged
and older generation is about the same. For the rate of return for the elderly themselves we find on
average r00*≈ 1.6, but the old probably regard this rate of return not as an optimal value but as a
minimum required rate.
The elasticity of altruism of the elderly is not significant in variant 1. In variants 5 and 6 it is
significantly negative (-2). The estimation results for these latter variants are debatable, however, as
the coefficients of the fairness ratio with respect to the elderly themselves do not have the right sign
due to multicollinearity problems. There is no evidence for altruistic feelings among the elderly,
however. The elasticities of fairness with respect to the actual rates of return for the young and middle-
aged generation are respectively about 0.0008 and 0.0013. Note also that the elasticity with respect
to the old individual’s own income is very low in comparison with the young generation (0.54 versus
1.38, considering only estimates that are significant at a 1%-level). The estimated coefficients for the
background characteristics are rather robust and significant in almost all variants considered for the
elderly. Since we took the differences of the reported evaluations, this implies that there exists an
interaction effect for the elderly. This hardly occurred for the other generations. Apart from one case,
the intercept term is not significantly different from zero.
Comparing the results for the three generations shows among other things the following.
* The intercept is insignificant in almost all variants, as it should be if the specification chosen is
correct.
* Background characteristics have a different impact on the utility of the elderly compared with the
impact on other generations’ utility.
* For young couples changes in the pension system lead to significantly larger declines in utility and
significantly smaller increases in utility than for young singles, whereas this is just the other way
round for the elderly. The presence of a partner has no effect on the utility of the middle-aged
generation.