So we spoke earlier this week about using either relative or absolute poverty as a predictor of child education. Now we're thinking a little more about poverty as a whole and I wanted to share some peculiarities.
Here's something to think about with regards to relative poverty: it's measured by anyone making less than half of the median wage of a country. Countries with less inequality will have less relative poverty; those with more inequality will have more.
In any country, if everyone's real income doubled or tripled, you'd still have the exact same rate of relative poverty--even though everyone were richer. If some country was mostly made up of people making less than $1.25/day (the UN's measure of extreme poverty), they might have a lower relative poverty rate than a first-world nation with greater income variance.
So I'd be very skeptical of using relative poverty to measure anything substantial or informative about people's real living situation.
So this has the potential to have teeth: you can determine poverty by basically asking, "can someone afford a reasonable amount of stuff?"
One way to do it is just by cash. If you make less than X (possibly adjusting for family size, etc), you're in poverty.
There's of course a major problem with this: the cost of living in Alabama and New York City are totally different. Yet the US Census's official poverty measure looks at your income and uses a national cash basis (equal to 3x the minimum national average food budget, adjusted for CPI inflation), so you might have a family making $22,282 (the 2012 national line) in Alabama doing reasonably well, and a family in New York making $40k that's really quite strained. So that's silly.
The Supplemental Poverty Measurement (SPM)--used by the Census but not official--seems to have a much better approach. They include both income and other government supplements (welfare, subsides, etc), and then subtract out taxes, medical expenses, food, housing, clothing, child care, etc. They basically make a basket of "what you need to live," and determine whether all of your inflows can cover it.
Huge improvement. But we generally don't report it, and don't use it when we're running around declaring who is and who is not in poverty. So using SPM instead, about 20 million people that would be in the official poverty line are taken out, and about 20 million that would be out of it are put in. You probably get a more accurate sense of who's there.
So the only major objective international comparisons are "who's living on <$1.25 day?" and "who's living on <$2/day?" For an advanced nation, that number is vanishingly small (especially if you include government benefits), so nobody bothers using it.
But most people are pretty lazy when making poverty comparisons. They'll take the reported poverty rates from each country and stack them side by side.
...But every country measures official poverty in its own way. Most countries don't happen to use, "3x the 1963 food price, adjusted for CPI."
Now you might say, "well, I'm sure it's pretty close." Let's take a look at that.
A quick look of Wikipedia gives us each country's official government poverty measurement: turns out all of these countries report poverty however their countries want! The US official poverty rate is 13.66%. Did you know Russia's poverty rate is lower than that of the US and Western Europe, at 13.4%, even though their GDP per capita is well under half that of these countries? Hungary's poverty rate is only 12%, despite their income being about the same as Russia's. Belarus' poverty rate is an impressive 7.3% despite having a GDP per capita of less than 1/3 that of the US and Western Europe--well done! And China seems to be doing quite well for itself, with only 5.1% of its population below the poverty line. The Western world really needs to step up its game! (Their GDP per capita, by the way, is just over 1/6 that of the United States.)
The Official OECD Measurement
Just to start this off, we can look at how the OECD measures poverty among its member nations. They've got a nice graph that of course shows the US as pretty high.
Here, we can see that poorer countries like Hungary, Chile, and Greece all have significantly lower poverty rates than the US!
But if you don't check their methodology, you might get the wrong impression. Turns out they also use a purely relative measurement--also half the median income. And it turns out this lines up pretty well with the GINI coefficient. If every country had the same income, this might give you some idea of how people on the lower end might be doing, but again--cost-of-living variance is high both between countries and within the US (the US actually has a very low relative cost of living), and for richer countries, people at its bottom end do better than in poor countries.
So really, I've not yet seen a comparison of poverty across nations that works. The US can use the SPM to get some idea of who domestically is living hard, but running about trying to say that we're doing better or worse than other countries using any of the conventional comparisons is just totally bunk.