In a choice experiment we test income effects for different attributes and we test the effect of stated expected changes in future income on stated WTP. We find both present and future income to be significant determinants of stated WTP when included in the same model. We also find that the less use-related the attribute, the stronger the income dependency, both in terms of present and future income. The finding that expectations concerning future income affects WTP seem to suggest relevance of the life cycle-permanent income hypothesis also for environmental goods, and it suggest that the current practice of testing for income effects using current income only is likely to be flawed. This may be particular true if the good in question and/or the payment vehicle used have long time horizons.