Socio-economic inequality in ill-health amongst the elderly. Should one use current or permanent income?

In Belgium, income-related inequality in ill-health among the 65- is present, but reduces significantly as soon as one restricts the analysis to the 65+. The main goal of this paper is to check whether these findings (based on current income) are robust to the income concept used. The analysis is repeated using estimates of permanent income. Measured inequality among the 65+ remains very limited, suggesting that the observed findings are not heavily dependent upon the income concept. Finally, I suggest an explanation for the observed difference in inequality between the 65- and the 65+. I argue that this difference is likely to be correlated with differences across income groups of (1) sample attrition; and (2) mortality. Copyright 2003 Elsevier Science B.V.