I think a per-capita statistic is a bit flawed in this case, although I understand why they use it -- the math is very simple. The problem with per-capita meas

I think a per-capita statistic is a bit flawed in this case, although I understand why they use it -- the math is very simple. The problem with per-capita measurement of income for very poor people is that what you really care about is per-family income because of the economies of scale involved in so many of the goods they consume -- cooking fuel, etc. In places where poverty like this exists, a family of 8 living on $8/day ($1 per capita) is often better-off than a family of 4 living on $5/day ($1.25 per capita). Regardless of whatever methodological flaws exist in various metrics, this is a positive trend that is (hopefully) only going to get better as technology increases the efficiency of various systems of distributing goods and services.