Deregulation, Devolution, and State Preemption Laws’ Impact on US Mortality Trends

Open any introductory demography textbook and you will find the core explanations for the historical decline in mortality: industrialization, improved standard of living, better nutrition, and public health infrastructure. In other words, macrolevel explanations. These core explanations are widely accepted, even if their relative importance remains a matter of debate.

Fast forward to recent trends in mortality—trends that are less favorable, even worrisome. Since the early 1980s, declines in mortality in the United States have not kept pace with those in other high-income countries.¹ Moreover, some parts of the United States and certain population subgroups, such as low-educated adults, have reportedly experienced increases in mortality.² According to the US National Center for Health Statistics, between 2014 and 2015, life expectancy at birth for the nation as a whole declined by 0.1 year.³ Although that decline may ultimately prove to be a minor hiccup, it is a troubling indicator of the nation’s current health.

What explains the mortality trends of the post-1980 era? Curiously, as mortality trends changed from (historically) encouraging to (more recently) discouraging, the dominant hypotheses have also changed. By and large, they shifted from macrolevel to microlevel hypotheses. They shifted from hypotheses about industrialization and public health infrastructure down to hypotheses about individuals’ lifestyle “choices”—for example, in terms of smoking, sedentary lifestyles, opioids, and work-family situations. Why have scholars and policymakers “hypothesized upwards” to explain historical trends, yet “hypothesized downwards” to explain more recent trends?

In this editorial, I encourage scientists and policymakers concerned about relatively recent US mortality trends to hypothesize upwards—and, in doing so, to focus on the diverging policy contexts of US states.

NEW YORK AND MISSISSIPPI: TWO POLICY DIRECTIONS

Why focus on US states’ policy contexts? Compare New York and Mississippi, for example. In 1980, life expectancy at birth in these two states differed by just 1.6 years.⁴ By 2014, the difference was a striking 5.5 years. In terms of life expectancy, New York now resembles Denmark, and Mississippi resembles Romania. Why has life expectancy diverged across US states since the 1980s? And what can the answer to that question tell us about the reasons for relatively recent trends in mortality?

Here is one compelling upward hypothesis. Three macrolevel initiatives were launched around 1980: deregulation, devolution, and preemption. These initiatives have likely had far-reaching, yet underappreciated and unintended, consequences for population health. Deregulation of industries such as transportation has meant that rail and air travel in some parts of the country shriveled. Businesses soon followed. Devolution has provided states greater discretion over which policies and programs to fund and at what levels. State preemption laws—which forbid local governments from legislating on particular issues such as indoor smoking bans—have accelerated across several states.⁵

Consequently, states have devised markedly different policy contexts since the early 1980s. The comparison of New York and Mississippi is again instructive. New York has imposed a hefty excise tax on cigarettes, implemented its own Earned Income Tax Credit (EITC), participated in the Affordable Care Act’s Medicaid expansion, and has not preempted local governments from implementing health-promoting legislation such as paid sick days, a higher minimum wage, stricter firearm regulations, or requiring that calorie counts be posted. In sharp contrast, Mississippi has retained a negligible cigarette tax, does not offer its own EITC, did not participate in Medicaid expansion, and has preempted local laws in all four areas listed previously. The diverging mortality trends suddenly become much less mysterious.

RECENT EVIDENCE

States shape their population’s health and mortality, irrespective of residents’ own characteristics and local environments. Indeed, a large and robust literature has repeatedly shown the importance of states for their residents’ well being.

Two recent studies also show that states are particularly important for individuals who are socioeconomically disadvantaged, such as women and low-educated individuals. A 2016 study of adult mortality across states found compelling evidence that states matter more for women’s mortality than for men’s.⁶ The study estimated how much adult mortality varies across the 50 states, then partitioned that variation into the part attributable to individuals’ characteristics (e.g., race) and the part attributable to states’ characteristics (e.g., tax policy). For men, roughly 34% of their variation in mortality across states was attributable to men’s attributes and 23% to the states. For women, roughly 30% was attributable to women’s attributes and 53% to the states. A 2017 study of disability across US states found that states have their greatest impact on the probability of having a disability among the least-educated adults.⁷ States where having a low level of education is often a “poverty sentence” are particularly unhealthy places to live. The study concluded that the fewer resources that individuals have to build a healthy life, the more important their state of residence becomes in shaping their health.

WHERE DO WE GO FROM HERE?

The macrolevel mortality trends require macrolevel explanations. They also require interdisciplinary perspectives. Demography, sociology, and epidemiology remain central to the debate, but so too do political science, history, and geography. Longer time horizons must also be examined. Explanations for the historical mortality decline would not have been discernible from a decade-long comparison of causes of death; neither will explanations for more recent trends. Furthermore, macrolevel explanations such as deregulation, devolution, and preemption are not listed on death certificates. New and innovative study designs that can capture these broader factors are required. Lastly, better data are required. Publicly available mortality data often do not even contain state of residence. When they do, they do not contain state of birth or interstate migration histories. These data are available, however, in the restricted-use versions of many of these data sets (e.g., the National Health Interview Survey) and should be made publicly available. Despite these obstacles, we must rethink and reexamine the reasons for worrisome contemporary mortality trends. It is time to hypothesize upwards.