Human Development Map is Bunk
Andrew Gelman looks inside the methodology used to construct the widely-circulated “Human Development Index” map and concludes that it’s bunk.
It turns out that the “index” considers only three elements: Life expectancy at birth, adult literacy and education, and a variant of GDP per capita. It seems that the first two of these are so uniform across the 50 states as to be negligible, making the last the main determinant of the ranking. Additionally, Mississippi’s .799 makes it just barely fit into the arbitrary color breakdowns, making it appear to those looking at the map to be massively less “developed” than West Virginia and Louisiana, at .800 and .801, respectively.
What Gelman does not add is that “the natural logarithm of gross domestic product (GDP) per capita at purchasing power parity (PPP) in United States dollars” is actually a rather poor measure of “standard of living” given that it’s not normalized for local housing prices.
Does anyone really think that there’s a significantly different level of “development” between North and South Carolina? Or that Michigan is better off than Missouri?
This is yet another instance of a trend that I’ve long found aggravating: the ordinal ranking of relatively similar bodies to create the illusion of substantial disparity. We usually see it in the form of international comparisons, which have the United States ranked 35th on some attribute despite being essentially the same as the country ranked 1st. Here, we’re doing the same with states. It’s junk social science.
PPP is not normalized for local cost of housing? Isn’t that the entire point of PPP? Or does it only include food and consumer goods?
I concur with Andy regarding NC and SC, though I agree with James regarding MO and MI.
Yes, this is junk social science that serves no purpose other than to reinforce preconceived notions and bigotry.
I don’t know as I’d limit it to that, Charles. I noted in reading James’ article something he did not. Look at the states that are claimed as being ‘more developed’, again, and consider that the darker states on the map are with the possible exception of Alaska, liberal strongholds, if not Democrat strongholds. Places like Louisiana, for example are certainly Democrat strongholds, but I would argue are never considered, even by liberals, to be among the liberal elite.
Notice also, that the map (With again, the exception of Alaska) shows the states with the biggest cities as ‘developed’. right online with what the UN has been pushing for for decades… they want the populations in the cities (It’s currently being sold as being green)…
If as James suggests, and I agree, the report makes no sense in it’s stated context, might a logical explanation of the release of this report be that someone’s trying to make a connection between the liberal elite, their utopian wet dreams, and human development?
“You’re a (shudder) conservative? You’re from flyover country? Why, you backward putz, you…”
It is a junk study that is very subjective. Someone will spend as much on a cramp small apartment in Sacramento than what someone spends on a large house in some places in the Midwest. Which is a higher standard of living? Of course people from both places can argue their place is the best because of other reason but that is all subjective.
The BEA is working on a spatial cost of living index that makes cost of living adjustments for the major cities and the states.
It generates some interesting conclusion about the various states real income and real gdp adjusted for differences in the states cost of living.
I wrote about it here.
The adjusted cost of living ranks from Delaware with real per capita gdp at 160% of the national average to Mississippi at 79% of the national average. It has SC at 91% of the national average
and NC at 112% of the national average so there
is a big difference between North and South Carolina.
Maybe the big surprise in the data is how it moves New York and California down. California at 89% and New York at 80% of the national average are among the poorest states.
Of course it’s bunk. But since I’m in a dark green state I’m going to get a kick out of it.
Heh. Virginia’s dark green, too, thanks mostly to NoVa. But it’s still silly.
Consider the top 11 elected positions in LA: of the governor, lt. gov, 2 senators, 7 reps, 3 are Democrats – the lieutenant governor, 1 senator, 1 representative. LA also has not given its electoral votes to a Democratic presidential candidate since 1996, but voted Republican in 1988, 1984, and 1980.
I couldn’t find a partisan breakdown of the state legislature, but a Democratic stronghold, it sure ain’t.
Interesting link, Spencer.
The cost of living adjustments do appear to tend to move states like Massachusetts down from 6th to 25 in GDP per capita rank.
I assume that if Governor Palin got her way (drill, baby, drill), the human development of Alaska would simply have gone off the charts.
I think they’re using tools that might be useful for comparing the development of the U. S. to that of Rwanda to measure the difference between states within the U. S. As should be expected the margin of error of the measurement exceeds the differences observed.
As odograph observes above, that may be fun but it’s not useful for devising policy.
According to census projections the range of life expectancy by state ranges between 73.7 for Mississippi to 79.8 for Hawaii. Is that so uniform as to be negligible?
According to the NCES the percent lacking basic prose literacy skills in MS is ~16% and is ~7% in Vermont. Again is this really so uniform as to be negligible.
Looking at education states range from TX’s 78.3% HS graduation rate to MN’s 92.3%. For college the range extends from WV’s 16% with a BA or BS to MA’s 37.4%. So I don’t think they really qualify as being so uniform as to make differences negligible.
The study may be bunk, but apparently not because of the uniformity of life expectancy at birth, uniformity of education, or uniformity of adult literacy.
While your point is well taken, I’d point at NOLA. I’d also suggest there is something of a difference between being a liberal stronghold and a Democrat party one, which is why I double clutched that in my original response. As the party has moved left, they’ve been slowly losing power outside NOLA. Inside it…. well…
See, I suspect based on my short run down there years ago, that outside NOLA, they tend toward ‘Democrat’ in the ‘southern democrat’ mold, whereas NOLA seems to be unabashedly liberal. (Explains Ray Nagin, for example, and the city council, there, for that matter)
Admittedly, I haven’t researched that point lately,(And as you point out, the local stuff isn’t exactly easy to locate from our perspective) but then again, even with Katrina, I’ve never seen anything to cause me to think the assumption was in error, either.
Hmmm. Used to know a guy that worked at WWL. Wonder what ever happened to him, or if he’s still there. Maybe I can tap him for some flavor on this.
The life expectancy of African-Americans is about 73.1 and for whites it’s 78.3. I couldn’t find any comparable numbers for Asian Americans and Pacific Islanders, but they appear to have life expectancies in the eighties. When comparing Mississippi and Hawaii are we comparing “development” or race?
Both Mississippi and Hawaii consistently come in at the bottom of the scale for standard of living in the US.
I have no idea what the “Democrat party” is.
For those two states in particular I would guess it was more about race (and factors associated with same) than development. My point was that the factors including life span are not so uniform as was implied by the post.
Depends on how you define standard of living. I much preferred living in HI to living in AL (pretty close to the same as MS).
LOL…. Well, you just stick to that story, Rick.
Oh, that’s easy…a silly, childish slur used by people who think that simple name-calling is a legitimate way to advance an argument…