Differential mortality and poverty by age.

Summary In the presence of differential mortality by income, standard measures of poverty do not only capture true poverty, but also reflect the selection process induced by differential mortality. One solution to this type of selection bias is to extend the income profiles of the prematurely deceased to include them in the poverty measure using a dummy income. The present study proposes to analyze the effect of age on the size of the selection bias. To this end, we exploit age-specific poverty data in Belgium and compare conventional poverty measures with adjusted poverty measures obtained by extending the income profiles of the prematurely deceased. We show that, although the gap between adjusted and unadjusted measures is larger at higher ages of life, selection bias is nevertheless present at all ages.