West Virginia, or the Bad News Takes a Long Time to Arrive

The blogger Evolutionistx had an interesting discussion of “What Ails Appalachia,” which discussed different hypotheses for West Virginia’s extremely high mortality rates for middle-aged whites, in the context of broader Appalachian poverty and social dislocation. West Virginia has the second highest death rate for middle-aged whites (after Mississippi), but is the poorest state for this group (income data taken from the 2013 CPS):


On the other hand, West Virginians have not done especially badly relative to their 1999 baseline in terms of income, even as their death rates have increased more than any other state:

image (1)

For another example, West Virginia incomes for this group (45-54 YO Non-Hispanic Whites) have increased a comparable amount to Pennsylvania incomes, even as death rates in Pennsylvania barely increased at all:


And employment rates in Pennsylvania have decreased more than employment rates in West Virginia:


In fact, West Virginia is among the very few states which have increased their labor force participation since 1999, probably due to the general boom in resource demand:

image (3)

The other major predictors of death rates I found, obesity and SSI-Disability receipt, tell a similar story: West Virginia has among the highest levels, but these values have not increased especially quickly in West Virginia relative to their increase elsewhere in the country, even as death rates have increased more quickly in West Virginia.

I don’t think this is a surprising finding, but it’s worth underlining: there’s a tendency to explain recent trends through recent changes, when in reality they may be the cumulative result of decades. The best book I read last year was Homer Hickam’s memoir Rocket Boys, which is about a West Virginia coal town coming together to push their most promising young men out of coalmining and likely out of West Virginia, because they know their town is going to die when the mine closes. The book takes place in 1958.

Age-Adjusted Drug Overdose Rates by State, CDC 2013


My belief is that the increases in death rates among middle-aged whites are exactly what they appear to be: suicide, drug overdose, and alcoholic cirrhosis, all three driven partly by chronic pain and isolation as well as easily available guns and legal and illegal drugs. Obesity may play a mediating role of increasing pain and making mobility difficult*, and SSI-Disability receipt may be both a consequence of actual pain and cause additional harm by enabling long-term isolation and detachment from the labor market. This does not mean that broader economic and social trends do not play a role. This could be the disappearance of the community ties that allowed Coalwood, West Virginia to come together so effectively in 1958 but might not do so today; or the decline of churchgoing or marriage, or other sad judgments that many have passed on working class whites. But these are not sudden or abrupt changes. They are bills that are being collected, with interest, after many years.

This is why I am much more pessimistic about the long-term effects of cash transfer programs than I am about their short-term costs (the kinds that can be measured through “well-identified” experiments or quasi-experimental designs.) The diminution of spirit caused by a world without work takes time, whether you are drinking yourself to death as an individual or, as a society, turning mere material insufficiency into squalor.

The bad news takes a long time to arrive.

*I’ve read actual transcripts of interviews with low-income parents where they say “it’s hard for me to get to school events because I am obese.” I found it a remarkable display of honesty, but for every one willing to make that admission, it seems likely there are many for whom it is true. As I’ve said elsewhere, schools in many ways are the last redoubt of in-person community for many people; getting parents involved is unlikely to boost test scores but it doesn’t mean it doesn’t have benefits for individual adults as well as kids.

9 thoughts on “West Virginia, or the Bad News Takes a Long Time to Arrive

  1. Obesity makes it hard to walk. I’ve been nine months pregnant a few times, and even then I was not “obese” by American standards, but I still the extra weight was a burden. There are a lot of people out there carrying a full extra person’s worth of weight with them all the time. It must be hard.

    Liked by 1 person

    1. The data does not show that “death rates among middle-aged whites are exactly what they appear to be: suicide, drug overdose, and alcoholic cirrhosis,”

      See table in page 49+ in http://www.cdc.gov/nchs/data/dvs/lcwk1_2013.pdf, and then compare with all the tables in http://www.cdc.gov/nchs/nvss/mortality/lcwk1.htm from 199 to 2014. Age adjusted, the percentage has increased slightly (age adjustment is necessary because the USA had a large increase in birth rate between 1956 and 1961, and those children reached the age of 50-60 in large numbers between 1999 and 2013). Malignant cancer, heart diseases, accidents, cirrhosis and diabetes dominate at 70%. Granted, drug overdoses are distributed between accidents, suicides, nephritis and other, here.

      Obviously, I am not talking about WV and about all of US, but such data for WV is also available in http://www.cdc.gov/nchs/pressroom/states/WV_2015.pdf but not by age group. Drug deaths and suicides are tiny at about 6%. Another bizarre point is that Cirrhosis is a top 5 cause of death in US from 1946, and has decreased by a factor of 3 from 49 to 16.9/100,000 in 70 years.

      My conclusion is a somewhat accelerated effect of aging of populations, amplified by economic difficulties. It could not be even argued that a lack of health insurance caused this, because, both, in WV and Pa, whites without health insurance is a lower percentage at 7% as compared to 10+ % in all of US. At this time, the data dos not show a clear reason for amplified (not explained by just aging) death rates in the 45-59 age group for whites. It cannot be easily attributed to evil things happening in Applachia.


      1. I’m aware of the age adjustment issues. As I mentioned in the earlier post, just comparing among US counties in 2013 there is huge variation, and figure 4 in the Case and Deaton paper suggests that these increases in suicide and drug-related deaths are not a cohort effect.


  2. Important responses to Case and Deaton paper, re: age-adjustment:
    Philip Cohen
    Andre Gelman
    (numerous, http://andrewgelman.com/2015/11/06/correcting-rising-morbidity-and-mortality-in-midlife-among-white-non-hispanic-americans-in-the-21st-century-to-account-for-bias-in/; and http://andrewgelman.com/2015/11/06/what-happened-to-mortality-among-45-54-year-old-white-non-hispanic-men-it-declined-from-1989-to-1999-increased-from-1999-to-2005-and-held-steady-after-that/)
    Summary = ““mortality rates among non-Hispanic whites aged 45-54 increased by an average of about 4% after controlling for age.””, and “What happened to mortality among 45-54-year-old white non-Hispanics? It declined from 1989 to 1999, increased from 1999 to 2005 (4%), and held steady after that.”

    Outside of abundance of Opioids as painkillers, there is virtually nothing to see here.


    1. Yes, I read all these, and mentioned them in the original post. The age-adjustment issue is relevant, but doesn’t change the main conclusions of Case & Deaton, since Europe, for example, had a baby boom after World War 2, too, but mortality rates are increasing for Americans of the same age group even as they are decreasing for Europeans. (Similarly, though the age structures are more different, mortality rates are falling for blacks and Hispanic Americans at the same time.) It’s simply not true that the phenomena Case & Deaton are observing (which again, include rapidly rising rates of death due to suicide, alcoholism, and drug overdose for all cohorts of whites, so clearly not a age-adjustment issue) are merely a statistical illusion.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s