In a simple 2-period model of relative income under uncertainty, higher comparison income for the younger cohort can signal higher or lower expected lifetime relative income, and hence either increase or decrease well-being. With data from the German Socio-Economic Panel and the British Household Panel Survey, we first confirm the standard negative effects of comparison income on life satisfaction with all age groups, and many controls. However when we split the West German sample by age we find a positive significant effect of comparison income in the under 45s, and the usual negative effect only in the over 45 group. With the same split in UK and East German data, comparison income loses significance, which is consistent with the model prediction for the younger group. Our results provide first evidence that the standard aggregation with only a quadratic control for age can obscure major differences in the effects of relative income.