• Income mobility and female life expectancy in two maps

    From a new study, which has received considerable attention:

    income mobility

    Compare to this, via Bill Gardner:


    Read the captions and/or follow links for details. Correlation doesn’t imply causation. (And, how correlated are these?) This warrants further study and comment, for which I’m not (at the moment) prepared or qualified. So, blog post for the taking.


    • You get a similar map if you look at obesity rates (male or female). I’d like an epidemiologist’s take on this. There are a lot of disease epidemics that follow very similar patterns in terms of how they affect marginalized groups (which, in turn, are the poorest groups with the least income mobility).

    • Just posted this on Twitter, but I’m not seeing a correlation between drops in income immobility and life expectancy drops here.

      I’m looking at southeast Ohio’s border with Kentucky and West Virginia, south Missouri, and south Oklahoma: Those are about the peaks in drops in life expectancy (red coloring), but also where income mobility was higher (light colors). Additionally, the broad swatch of income immobility on the SE coast is matched by (mostly) positive changes in life expectancy.

      Agreeing with Austin, I’d like to see a county-by-county comparison. But it looks to me like there’s not a one-to-one mapping here. At least, it’s not the dominant pattern.

      (Also, the life expectancy changes on OK’s edge counties not leaking over into the other states makes me a little nervous. Could there be sampling issues here?)

      • I was thinking about Oklahoma too. But I think the answer is that those lands on the border are still (as best I can tell) populated largely by Native American people. They are the sites of the former Indian reservations, and are still today considered “Tribal Jurisdictional Areas.” Again, this has a lot to do with marginalization of minorities.

        • I don’t know if Native Americans is the whole answer.

          I looked up a map of Native American reservations (http://bit.ly/13cXupk) and only one is in the southern swath of red. That is Bryan County, which is 80% non-Hispanic White and 12% Native American. For that 12% to be responsible for the entire drop in life expectancy while everyone else went up in step with across the Texan border, it’d have to be a drop of a few years.

          That may be what happened, but I’d like to see data breaking down those calculations by ethnicity to be sure.

        • I live in Oklahoma.

          The counties along the southern border of Oklahoma with Texas are overwhelmingly rural and overwhelmingly white. The only major city in those counties is Ardmore in Carter/Love counties in the central southern portion, on I-35 between Oklahoma City and Dallas. No other cities in that area have more than 30k population.

          The indian nations (5 civilized tribes) live mostly in the interior eastern portion of the state, from Durant up thru Muskogee and Tulsa, all the way thru Vinita in the northeast corner of the state.

          Also, the indian nations actually have a very good centralized healthcare system thru the IHS. Many studies of indian health in Oklahoma show improved results compared to the local white population. In other parts of the country the IHS is not as well organized and their healthcare statistics are not nearly as good.

          • Another Oklahoma factoid — Southern Oklahoma has the 2nd largest casino in the world (WinStar. owned by indian tribes).

            The largest casino is the Venetian in Macao.

          • Interesting. I was unaware of the IHS.

            Of course, with my theory dis-proven, that means that the Oklahoma-Texas border really is a strange anomaly.

    • I’m a big one on socio economic relationships. Lower life expectancy, poorer health, higher infant mortality, smoking, drug abuse, low income, low IQ, low education, etc. all seem to group together in one way or another.

    • “Poverty is an exam room familiar. From Bellevue Hospital in New York to the neighborhood health center in Boston where I used to work, poverty has filtered through many of my interactions with parents and their children.” Poverty as a Childhood Disease By PERRI KLASS, M.D.
      May 13, 2013,

    • Believe the book “The Spirit Level” made a similar case to this.


    • That’s definitively something to think about… Surely, there is no direct linking as you said correlation does not imply causation, but the maps are too similar to be written off as a simple coincidence.

    • Evaluate the geographic distribution of any correlate of poor health and poor social function in the US at the county level and you get virtually the same pattern. Where poverty is high, so is the disease burden, and so is the level of government expenditure dedicated to attenuating the impact of poverty and disease.

      The correlation between spending and disease burden is so striking and vivid that the decades long quest to correlate variations in spending to practice variation rather than the variance in disease burden is quite puzzling.

      “Geographic Variation in Fee-for-Service Medicare Beneficiaries’ Medical Costs Is Largely Explained by Disease Burden
      James D. Reschovsky jreschovsky@hschange.org
      Jack Hadley
      Patrick S. Romano


      Control for area differences in population health (casemix adjustment) is necessary to measure geographic variations in medical spending. Studies use various casemix adjustment methods, resulting in very different geographic variation estimates. We study casemix adjustment methodological issues and evaluate alternative approaches using claims from 1.6 million Medicare beneficiaries in 60 representative communities. Two key casemix adjustment methods—controlling for patient conditions obtained from diagnoses on claims and expenditures of those at the end of life—were evaluated. We failed to find evidence of bias in the former approach attributable to area differences in physician diagnostic patterns, as others have found, and found that the assumption underpinning the latter approach—that persons close to death are equally sick across areas—cannot be supported. Diagnosis-based approaches are more appropriate when current rather than prior year diagnoses are used. Population health likely explains more than 75% to 85% of cost variations across fixed sets of areas.”

    • “When analyzing local area variation, we continue to rank both children and parents based on their positions in the national income distribution. Hence, our statistics measure how well children do relative to those in the nation as a whole rather than those in their own particular community.” From the Findings .pdf file in the Executive Summary.

      I don’t understand what exactly they’re measuring; the criticism I’ve read elsewhere is that children in the lowest income areas (call them “Bangladesh”) have almost zero chance of getting to the highest income levels in the nation ( call them “Monaco”). If the parents were at $20,000 and the child moves to $40,000 (inflation adjusted) that should be seen as “income mobility” not just if the child moved all the way up to “Monaco levels” of $250,000.

    • You do not have to be a racist or like the fact that it is so but it appears that racial makeup has a lot to do with which areas do well mobility. It seems that areas with the lowest percentages of African Americans do well on mobility and areas with the highest percentages of African Americans do poorly on mobility. That is true despite the fact that African Americans do better in areas highest percentages of African Americans.

      • What to do about the lank of upward mobility among black Americas depends on what is causing it. If because of history of slavery or whatever black Americans distrust and so opt out of the the struggle for wealth only time will help. If it is due to racism then we should continue to fight racism in various ways.