1 Income-related health inequality in Belgium,

In this paper, we measure income-related health inequality over time in the Belgian population. We apply the methodology of Jones and Lopez (2003) for Belgium. They develop a long-term concentration index based on weighted short-term concentration indices and some measure for health-related income mobility to measure the longitudinal effects of changes in income on health. Then, they construct an index of health-related income mobility to measure how the longitudinal outcome differs from the cross sectional ones. Further, they decompose this mobility-index through simple linear regression into the contributions of different regressors. The empirical analysis is based on waves 3 (1994) till 10 (2001) of the Panel Survey of Belgian Households. The analysis is restricted to individuals of 16 years or older, who participated in all the waves and who filled in the questions on all the variables used in the regression. After all these omissions, the dataset contains a total of 23,152 observations, which means that there is a balanced panel of 2,894 observations per period. For the health variable, we used the recently developed interval regression approach proposed by van Doorslaer and Jones (2003) to make our categorical self assessed health variabl