One problem with cross-sectional statistics is that if people die "early", you need to count not just their loss, but also that of anyone who would otherwise have been born. It would clearly be wrong to count the lower birth numbers due to prior tragedy, as success.