Location Based Inequality Is Rising

The urban-rural divide is growing larger

by Nate Griffin – November 16th, 2021

A recently published article by Benjamin Couillard, et. al. in the Journal of Economic Perspectives examines the disparities in mortality rates in the US by geographic location.

Key Points:

• Life expectancy declined by 1.5 years in 2020.

• There are large mortality differences between college educated and non college educated individuals, but changes in state wide education levels do not explain much of the differences.

• The negative correlation between income and mortality rates has been strengthening over time.

• State level life-expectancy rates likely stem from a combination of individual behaviors and state level health investments that also correlate with income and education levels.

---

The COVID-19 pandemic took the lives of nearly 500 thousand people in 2020, resulting in a reduction of life-expectancy by about 1.5 years. This was a striking reversal in what had been generally consistent gains over the last six decades, and was the worst decline since WWII.

But the pandemic was not the only health crisis of 2020, a recent report by the US Centers for Disease Control found that over 100,000 people had died of overdoses between the 12 month period of May 2020 to April 2021.

Also troubling is that life-expectancy has been leveling out since about 2010. And midlife mortality – defined as deaths between 24 and 64 – has been increasing due to two factors: medical conditions like heart disease and, secondly, drug and alcohol related deaths.

But differences in mortality rate across states have been increasing over time. This study examines the causes for the growing disparity.

Causes of disparities in mortality rates

The authors posit three potential causes of this statewide variation of mortality rates.

1. Variation in educational attainment across states

2. Disparities in income inequality

3. State level "place effects" such as individual's health behaviors as well as

Educational Attainment

Previous research has found a correlation between increased educational attainment and lower mortality rates among Americans. The author's calculations, using data from the National Center for Health Statistics (NCHS), confirm the large difference between college and non-college educated mortality rates.

For example, in West Virginia and Kentucky, mortality rates rose for non college-educated individuals while rates for the college-educated population declined between 1992 and 2016. For the majority of states, the gap in mortality rates for college educated and non-college educated individuals widened over those two decades.

But one of the challenges in understanding the growing disparity in mortality rates is that the population of college educated individuals has increased in most states over the years.

And when the authors control for this change, college education only explains a small fraction of the variation in mortality rates.

The authors find that educational mortality rate deviations are largely explained by two factors:

1. The college-educated residual: The difference between national and state mortality rates among college educated individuals. These differences could be considered "place effects".

2. The non college-educated residual: The difference between national and state mortality rates among non-college educated individuals. Again, assuming the explanation is "place effects".

The researchers propose that the interaction of numerous environmental factors influence an individual's health. They find that, "..the importance of both residuals [college and non college educated] in our framework of state-level mortality suggests that in some states, 'place effects' have evolved over time to the benefit of both college and non-college residents."

These place effects could be a many things, from better investments in healthcare, higher taxes on cigarettes and alcohol, attitudes toward exercise and nutrition, etc. All of these act together – at the state and probably at the very local level – to either improve or worsen individual health.

Income effects

Using the Census Bureau's per-capita income data for 1968, 1980 and 2019, the authors perform a series of regressions, comparing state level mortality data against state level per-capita income.

In both 1968 and 1980, the age adjusted mortality rate was not highly correlated with real income per person, a ρ value of -0.20 and -0.31 respectively.

In 2019, however, a much stronger negative correlation appears, with a ρ value of -0.71.

One clue to why this relationship increased over time might be that mortality in higher per-capita income states – like New York and California fell – while rates in lower per-capita income states were mostly flat. The authors state that, "mortality changes have been most favorable in those states that have tended to have high relative levels of income over the past three decades."

The map below shows Age Adjusted Mortality rates for 2017. There is a concentration of high mortality rates in the South, which has historically had lower per-capita incomes.

Age Adjusted Mortality Rates by State, 2017

Source: CDC. Does not constitute an endorsement or recommendation by the U.S. Government, or Centers for Disease Control and Prevention;

A proposed explanation for the increasing correlation of income effects with mortality could be the growing rate of "deaths of despair"; those which are from suicide and drug overdose.

The authors find, however, that when excluding deaths of despair – which account for about one sixth of all mid-life deaths – the coefficient of variation of mortality rates increased by 67.9 percent.

This finding implies that the rise in deaths of despair does not account for much of the income-mortality relationship.

State Level Place Effects

Building upon previous work from Grossman (1972) and Case and Deaton (2005) using the concept of "health capital." The authors propose that Individuals invest in their own health through positive behaviors like health diet, exercise and avoidance of harmful substances like cigarettes and alcohol. But individual behaviors are also influenced by state level effects like government policies and societal attitudes related to health.

The author's final hypothesis is that:

"The widening divergence in midlife mortality and the tightening relation ship between mortality and income reflect the long-run effects of varying behaviors and policies related to health capital during the last several decades. The data suggest that residents of high-income states have enacted policies and adopted behaviors with long-run payoffs to midlife mortality that are becoming increasingly apparent over time.

Several factors indicate why the state level income-mortality correlation continues to increase:

• Migrants to the North, including African-Americans who left the South during the Great Migration tended to have higher incomes, but also higher mortality rates, partially because social and health institutions were apparently overwhelmed during that time.

• Higher income states tended to expand Medicaid, lowering infant mortality.

• Higher income states also operated other health programs like Community Health Centers to care for medically under-served populations.

• States with declining mortality rates like California have enacted stricter environmental protections.

• Substantial excise taxes on cigarettes in places like New York have helped to lower rates of smoking.

While income and education levels are correlated with state level mortality rate outcomes, the relationships are more complex. State level policies and societal attitudes are important contributions that are harder to measure.

State level characteristics like smoking rates, Medicare and Medicaid investments and environmental protections can have an effect on overall health outcomes, but these effects likely compound over time.

The implication is that states may improve their overall health outcomes with proactive policy action but, observing a measurable improvement could take years or decades.

Reference:

Couillard, Benjamin K., Christopher L. Foote, Kavish Gandhi, Ellen Meara, and Jonathan Skinner. 2021."Rising Geographic Disparities in US Mortality."Journal of Economic Perspectives, 35 (4): 123-46.