"Every 8 days as many children die from poverty as the number of people died in the Haiti earthquake."This comment was twittered by the must-follow Solarafrica. The quote is by Esther Duflo at the recent TED conference.
Some questions for you to ponder:
- We have many different ways of measuring poverty. The most basic measure is a basic headcount under a given threshold. Let's say the threshold is $2 a day. Imagine that two people die of the same thing. If the person who earns $1.99 dollars a day dies, and the cause of death is poverty, what caused the person that earned $2.01 to die?
- Let's argue that a child who dies because her family is unable to to pay for a basic health intervention has died of poverty. If someone you know (who is not poor by our basic measure) dies of cancer because they cannot afford the effective, but extremely expensive new treatment, did they die of poverty?
- What does these inconsistencies say about the way we implicitly feel about the severity of poverty or about the distribution of income around the poverty line?. Are we using the right measures of poverty when we frame our questions in terms of mortality? Is this "type" of poverty measurable?
2 Comments
Hmmmm....I thought she was smarter than that. A poverty measure is just an instrument to determine who is rich and who is poor economically and something for policymakers and people to understand where public and private resources can be allocated and useful.
The person who couldn't afford the cancer cure got screwed but could have afforded the basic health care many times over and did not die from that. The person that couldn't afford the basic health intervention also couldn't afford the cancer cure if she needed it so she was relatively more screwed. So if you don't want to sue the dollar a day measure, maybe someone can come up with a measure of who's more vulnerable to getting screwed meter.
I don't think the problem is neccesaryily the measurements themselves, although they are of course imperfect, but the ways that we use those measurements. Saying such and such amount of "children die from poverty" every eight days is a depressing little factiod that you can tease out of the data. But its practical usefulness is pretty low, perhaps good for fundraising, but not helpful in formulating good policy prescriptions. As your question points out, "dying from poverty" is simply too vague a phrase to be really meaningful beyond mildly shocking people. Furthermore such formulations present us with questions that aren't worth answering (such as your first bullet point) and so waste our time (which is not a minor consideration). On the other hand knowing the number of people who live on $2 a day is undoubtably helpful in other contexts, for example in measuring economic progress, GINI coefficients, etc.
The misuse of imperfect stats is a bigger problem than the stats themselves.