We know that the much-discussed increase in mortality among middle-aged U.S. whites is mostly happening among women in the south.
In response to some of that discussion, Tim Worstall wrote:
I [Worstall] have a speculative answer. It is absolutely speculative: but it is also checkable to some extent.
Really, I’m channelling my usual critique of Michael Marmot’s work on health inequality in the UK. Death stats don’t measure lifespans of people from places, they measure life spans of people who die in places. So, if there’s migration, and selectivity in who migrates where, then it’s not the inequality between places that might explain differential lifespans but that selection in migration.
Similarly, here in the American case. We know that Appalachia, the Ozarks and the smaller towns of the mid west are emptying out. But it’s those who graduate high school, or who go off to college, who are leaving.
It’s possible, but obviously not certain, that the rising death *rates* are simply a reflection of this selectivity in migration.
I replied: This could be true, I’m not sure. I haven’t tried to crunch the numbers to see if mobility is enough to cause these changes, but on first glance it seems possible. One thing also to remember is that when comparing a particular age category over several years, we’re not comparing the same people. Today’s 50-yr-olds are not the same as next year’s 50-yr-olds. So the usual challenge is separating age, period, and cohort effects. But I agree with you that mobility is an issue too. On a related point, I questioned Case and Deaton’s comparisons by education category, because the proportion of people with college degrees etc. in different age groups has been changing over time too.
Worstall replied, “Not sure when the switch took place in the US but in my age cohort in the UK some 12% or so went to university, now it’s near 50%.” And then he followed up:
Further to the point that migration might be explaining something about these changes in average lifespans. Interesting new research from Glasgow. Seems that it at least part of the story there.
I guess the point is that death rates below age 65 are low enough that it doesn’t take much migration of at-risk people to move the numbers around.
P.S. As an aside, it’s kind of amazing that the big discussion of mortality trends was over a year ago. It seems so recent! There’s so much going on in statistics and social science, room for 400 or so posts a year, sometimes it’s hard to see how we can possibly keep it all in our heads at once.
The are a few plausible stories that could be tested. This wouldn’t explain the death rate issue but I have definitely seen people use migration data to show that mentally ill people are in-migrants into low income neighborhoods, which makes sense if you think about it. However, while the southern states are growing, I’m not sure that it is reasonable to think they are growing with sick people (though they could be growing with smokers). Although again if people who are sicker retire earlier and then migrate … but that seems to me like too complicated a scenario to think about. It’s also possible that the social isolation of the people left behind in the hollowed out small towns is having an impact.
It is interesting to me to realize that there are so many things that we still do not know about this classical topic, mortality rates. In many models that I know, migrations, inconsistency in reporting were almost always minor issues that are necessary to consider yet never actually have been considered. If we can show that the impact of minor issues is not minor predicting mortality rates, with some missing data methods, that will rock some people (well, actually me). I guess you might have thought about some approaches a little bit, when you post this story.
Lj:
I think there are also communication barriers. There must be many demographers and actuaries who understand these things very well, but they aren’t plugged into the news media, so instead we hear the half-assed guesses of economists and statisticians like me.
Just to widen that explanation of what I mean by selectivity in migration. And this is an example, only an example.
Say we have two segments of an extant population. One segment doesn’t graduate high school, by and large, ends up in bad jobs or no jobs, drug problems, single parent families and all the rest. And the shorter lifespans of populations rife with drugs, booze and so on.
The other segment graduates, many go to college, get careers not jobs, live a rather more standard American middle class life. With the associated life span.
Now we have selective migration. The first group largely stays put. The second largely migrates away. The split happening when each age cohort is reaching 18-20 or so.
The actual death rate has not increased, lifespans have not decreased, among the whole population. But as we record death rates and lifespan among people who are in a specific location at death, not track people from birth, our numbers will show a rising death rate, shorter lifespans. Not because of any change other than that the longer lived are moving away. The death rate will become apparent among people in their 40s and 50s, as it does take a couple of decades for drugs and booze to kill (my favourite explanation, on no evidence whatsoever, for the pop stars dying at 27 is that it takes about a decade for unlimited booze and drugs to kill a suddenly rich and famous 18 year old).
The stories from Appallachia are that pretty much everyone who does get to college stays away.
Similar claims are much easier to make. We’re not going to start claiming that certain areas of Florida are healthier because the average age at death is 90. Rather, we’re going to note that they’re retirement villages which people don’t move to until they’re 70, life expectancy at 70 being longer than that at birth (this one comes up in England every year, retirement towns like Cheltenham or Bournemouth have longer lifespans. Really?)
We know that the rich have longer lifespans, there aren’t many rich in the slums of Glasgow, they’ve moved away. Some part of the low life span in the Gorbals is due to this.
If we were looking at AIDS incidence in the 80s we’d want to note that half the small town boys who are gay have moved to SF. We can’t take incidence in SF or the small towns as being generally true.
I have absolutely no technical skills whatsoever so cannot test this in any manner. But I’d be absolutely fascinated to see someone who did have them test this a bit.
Migration would only explain regional shifts in mortality—not the claim that overall US non Hispanic White mortality has failed to decrease, unless a substantial number of the healthy population leaves the US entirely or the survey undercounts healthy people who move.