The objective of this paper is to analyze the relationship between income and living conditions and well-being of elderly. The best from two worlds is used for this purpose: a Danish survey covering 1440 elderly aged 72 and 77 from 1997, connected to reliable register information on income, 1988-1996. Indicators of physical activity, social and solitary activities, social contact, functional capability, loneliness and psychological well-being are constructed from the survey. Ordered logistic models are estimated controlling for demographics, education, previous occupation and whether the elderly are still working. It is found that elderly with low income levels have poorer functional capability, lower physical activity and poorer psychological well-being whereas social contact and social activities show no relationship with income. The relationship with income vanishes for solitary activities and loneliness when adding control variables. Having established that robust relations exist between income and some measures of well-being of elderly, we turn to an investigation of causality. Simultaneous models are estimated to assess whether the income relationship can be causally interpreted for functional capability and physical activity. The results show that it cannot be rejected that the income effects are causal, although care should be taken when interpreting results. We finally address our findings in the light of current and future changes regarding the size and economic well-being of the elderly population.