It’s all about women in the South

by on January 19, 2016 at 12:40 pm in Current Affairs, Data Source, Medicine | Permalink

Wow!! Remember that increasing death rate among middle-aged non-Hispanic whites? It’s all about women in the south (and, to a lesser extent, women in the midwest). Amazing what can be learned just by slicing data.

I don’t have any explanations for this. As I told a reporter the other day, I believe in the division of labor: I try to figure out what’s happening, and I’ll let other people explain why.

That is from Andrew Gelman, there is more at the link.


1 Marchin January 19, 2016 at 12:45 pm

Media narrative collapses in 5, 4, 3…

2 charlie January 19, 2016 at 12:46 pm

USPTF killing breast cancer screening — I kid, I kid.

3 DZK January 19, 2016 at 12:47 pm

I don’t know how much of this recent trend can be explained by just poverty or labour force participation. Latina and black women are much more likely to be poor and to work in low-paid occupations. And white women in the south are more likely to be married and religious than in the northeast, so I don’t know if Ross Douthat’s normal blaming of secular decline works here.

What I do know is that it’s good this is garnering a lot of sympathetic attention from all sides. I don’t remember there being a similar reaction during the inner city crack epidemic

4 Dave Smith January 19, 2016 at 12:56 pm

Are there regional differences in drug laws? There’d be a bunch of endogeniety…

5 zach January 19, 2016 at 2:13 pm

There are dry counties, strict occupational licensure laws, and poor accessibility to healthcare.

6 rayward January 19, 2016 at 12:59 pm

It would be interesting to have the data that connects length of time in the South with the death rates. This was a period of great migration to the South from other regions, many migrating on a hope and a prayer rather than definite employment or family connections. That, along with the stress and lost dreams, is captured by George Packer’s book, The Unwinding.

7 charlies January 19, 2016 at 1:09 pm

Could there be a connection to recent polling suggesting a lot of the rise in “voter dissatisfaction” is concentrated among white, female, republicans?:

8 affenkopf January 19, 2016 at 1:10 pm

Paging Steve Sailer…

9 Jay January 19, 2016 at 1:17 pm

Meth adoption?

10 mulp January 19, 2016 at 1:20 pm


11 Mark Thorson January 19, 2016 at 1:37 pm

“You’ve got your own cigarette now, baby. You’ve come a long, long way.”

12 Cooper January 19, 2016 at 4:37 pm

As a percentage of the population, people in Kentucky are twice as likely to smoke as people in California. That’s bound to create a huge gap in life expectancy.

13 mulp January 19, 2016 at 1:22 pm

It’s all Obama’s fault.

14 JWatts January 19, 2016 at 1:33 pm

I blame it on Reagan. /ducks

15 Jan January 19, 2016 at 1:38 pm
16 HL January 19, 2016 at 1:41 pm


17 Bob January 19, 2016 at 2:49 pm

My guess too.

18 Urso January 19, 2016 at 3:24 pm

Not sure about that. Much of the obesity epidemic is driven by African-Americans (who are disproportionately obese), and this is solely looking at white people, who are also getting fatter, but at a slightly slower pace. If it was obesity, it’d be reflected in higher death rates among other races as well.

19 Cooper January 19, 2016 at 4:35 pm

The definition of obesity was based on a sample of white people in the 1840s. Different ethnic groups have different body types so what is considered overweight for a white woman might actually be a healthy weight for a black woman.

Body Mass Index is a decent proxy for “healthy weight” or “unhealthy weight” but I’m inclined to believe that it might miss some of the nuisances of calculating appropriate “healthy weight” for different ethnic groups.

20 The Original D January 19, 2016 at 6:55 pm

I grew up in semi-rural Georgia and return there every year for Thanksgiving. Trust me, there are a lot of obese white people there.

21 Urso January 20, 2016 at 11:38 am

Not disputing that, but per the CDC: “Non-Hispanic blacks have the highest age-adjusted rates of obesity (47.8%) followed by Hispanics (42.5%), non-Hispanic whites (32.6%), and non-Hispanic Asians (10.8%).” So if this were obesity-driven, presumably A-As and Hispanics would be feeling the effects as well. Cooper’s hypothet is interesting; I have no idea if there’s any medical support for it though.

22 The Original D January 20, 2016 at 2:26 pm

I’d want to see those numbers broken down regionally.

23 albatross January 19, 2016 at 2:31 pm

The common narratives to explain the statistics keep changing as further analysis comes along. I wonder if there is really anything real going on at all here, or if this is some kind of statistical artifact or change in how things are recorded.

24 Floccina January 20, 2016 at 12:46 pm


25 duderino January 19, 2016 at 2:41 pm

It’s because they all work in fast food restaurants, never advance to anything else, and can’t afford to quit. Service economy!

26 The Anti-Gnostic January 19, 2016 at 9:32 pm

Yep. And very soon now, instead of telling your order to a cashier who presses a picture on a keyboard, you’ll just push that picture on the keyboard yourself.

27 Urso January 19, 2016 at 3:04 pm

Horrifying. What does it look like if you move the series back a little bit – or is there insufficient data before the 90s?

28 Hazel Meade January 19, 2016 at 3:35 pm

Hmmm…. husbands with PTSD returning from the Iraq war? Opiate abuse?

29 JWatts January 19, 2016 at 3:41 pm

“Hmmm…. husbands with PTSD returning from the Iraq war?”

On the graph it’s a steady trend upward starting in roughly 1999. And the graph for men is much flatter. So, it’s unlikely to be anything to do with the Iraq War.

30 Hazel Meade January 19, 2016 at 5:54 pm

You can’t really judge anything by the two data points before 2001. Although it’s pretty fair that you wouldn’t see anything based on the iraq war until 2005 or so.
I’m just throwing random ideas around.

What could possibly be specific culturally to southern white women, starting around 1999?

31 JWatts January 20, 2016 at 12:44 am

I wouldn’t assume that the actual trend started in 1999 just because the graph does. This trend might have been going for some time.

32 ohwilleke January 19, 2016 at 3:54 pm

My guess would be that the gender difference is mostly time lagged impacts of changes in smoking habits. Also, note that the scale for the women is half the scale for the men so the growth in Southern female mortality is exaggerated, and that the South is much worse than the rest of the country (in large part because its diet and culture are not tracking the national media narrative while everyone else is on board).

33 Hazel Meade January 19, 2016 at 6:14 pm

I think I’m buying this the most. Opiate abuse being second in line.
Although I’m not sure why smoking deaths would be higher for white females than for white men.

34 Curt F. January 20, 2016 at 12:04 am

Deaths for white males are nearly twice as high as deaths for white females. Make sure to observe the y-axes.

35 ohwilleke January 20, 2016 at 12:37 pm

Smoking in women went from almost zero pre-1920s, to low-pre-1960s, to much more common in the 1970s as gender based taboos decreased. This trend was stronger in women than the countervailing trend against smoking as its health effects became known that has much more impact on men since far more men smoked and thus reduced overall smoking rates. The improvement in Southern male mortality may very well be due to declining smoking rates from a much higher base number in men.

36 MikeW January 19, 2016 at 4:02 pm

We’re talking about the “Stroke Belt” here, the one cause of death that distinguishes the South from the rest of the country. I would hypothesize that some factor which predisposes people to strokes increased a bit among white women in the past few years. Take your pick: smoking, obesity or untreated hypertension.

37 Andre January 19, 2016 at 4:20 pm

I thought the recent wave of articles were highlighting that the increase in drug abuse and suicide was canceling out all the last few decades medical gains, not that the rate was strictly increasing. A very interesting change. I love this take though:

“This is the smallest proportional and absolute gap in mortality between blacks and whites at these ages for more than a century,” Dr. Skinner said. If the past decade’s trends continue, even without any further progress in AIDS mortality, rates for blacks and whites will be equal in nine years, he said.

Only 9 years from being as bad off as black people, time to panic!

38 yo January 19, 2016 at 5:21 pm

1) the mass migration of old white people of both sexes to the South.
2) Women being more social creatures, perhaps loneliness in retirement homes after moving? (sounds unlikely, if anything retirement homes seem to be very social places)
3) Those immigrants can’t cope with the hot weather?

39 Cooper January 19, 2016 at 5:43 pm

We are looking at trends among middle aged people, not the elderly. I don’t think many 45-55 year olds are in nursing homes.

Extreme cold kills more old people than extreme heat. It’s cheaper to cool a room than to heat a house.

40 Steve Sailer January 19, 2016 at 5:41 pm

If you look at the county level maps of overdose deaths in today’s NYT, the heartland of the White Death is less the South per se than the Scots-Irish parts of the country, which are typically on the fringes of the South: For example, in 2002, high rates of overdose deaths were concentrated in the eastern Kentucky – western West Virginia coal region, but have now spread to Ohio and parts of Pennsylvania. Oklahoma is another problem spot.

Reporter Sam Quinones’s book “Dreamland” focuses on southern Ohio.

41 Jan January 19, 2016 at 9:01 pm

I was also thinking about the known hot spots–WV, KY, OK, parts of TN, and parts of FL. I believe the authors used US Census regions and all those states are categorized as being in the South. Really it is Appalachia that is driving a lot of it.

42 Steve Sailer January 20, 2016 at 2:02 am

One reason the lowland South looks better than Appalachia on the NYT’s new map of overdose deaths over time is because life expectancies for blacks have been improving, and the lowland South has a lot more blacks than West Virginia.

On the other hand, comparing Appalachia to the Upper Great Plains, both very white regions, shows the Upper Great Plains doing much better.

I don’t know whether this is economic or cultural. The North Central did well through 2013 due to energy (e.g., North Dakota) and the China boom driving up the price of farmland, while the Appalachians got hit hard by the collapse of home construction after 2007 by drop off in demand for wood.

Or there could be a cultural difference.

43 Sarah January 20, 2016 at 7:19 am

Appalachia has also been hit very hard by the reduction in coal mining.

44 Urso January 20, 2016 at 11:40 am

It should be easy to remove the Southern states without heavy Scotch-Irish influence – Texas and Louisiana – to test this.

45 Matthew Ernest January 20, 2016 at 1:39 am

Please note that the Y-axes differ in both scale and offset between the men and women graphs.

While the graph for men in the South is coming down form its peak in 2008, the current delta for the 1999 starting data is comparable between men and women in the south

46 Ognian Davchev January 20, 2016 at 2:15 am

Men and women in the South diverge in 2005. This happens to be the year of hurricane Katrina. Maybe this has something to do with the rise of ACM in southern women.

Just throwing this out there.

47 Mikko Särelä January 20, 2016 at 9:39 am

Anything happening with abortion on those states compared to the rest of the US?

48 Andrew Graves January 20, 2016 at 9:40 am

I don’t want to be unkind, but it is obvious that very few of the commenters here have ever lived in, perhaps ever visited the south. For the last f*cking time, the South is not just a section of the Yankee Zoo. Now to the data..

I suspect there are two main components here. The first is the easiest to discern, which is rural white women that are up in years often have poor health. The primary indicator here is low income, but not only that. Health as Preventative Maintenance is not a cultural fixture in the South, with many people (black and white) putting off going to the doctor unless absolutely necessary. This goes for even potentially major health issues. Then, once they get a diagnosis, they may not be able to afford everything they need, or they may live in a very rural area where care is not easily available.

Tangential to that, there is no culture of fitness in the rural South. If you are not engaged in manual labor every day there is a very high probability that you will be obese. There are no gyms, no health clubs, few convenient parks with walking trails. There’s always a way to exercise, of course, but the cultural inertia–like what you would find in the middle class in almost any city–doesn’t exist. This phenomenon cuts across most income strata.

Lastly–and I believe this may be the key to explaining this data–huge numbers older northerners have migrated to the suburban South. I don’t have the data close to hand, but I know it is large. And some of those are getting sick and dying.

49 Urso January 20, 2016 at 11:43 am

Again, it’s not enough to say “the South is fat” without showing that – between 2000 and 2015 – the South grew fatt*er* at a higher pace than the North did. More Southerners are dying; that’s clear throughout the data series. But it’s the increase in delta over that time period that is grabbing everyone’s attention.

Comments on this entry are closed.

Previous post:

Next post: