Panel Paper: The Geography Good Jobs and Good Health: Local Labor Market Conditions and Deaths from Drug Overdose, Suicide, and Alcohol-Related Liver Disease

Saturday, November 10, 2018
McKinley - Mezz Level (Marriott Wardman Park)

*Names in bold indicate Presenter

Cassandra Robertson, Cornell University and Katherine Morris, Harvard University

In this project, we examine the association of mortality from drug overdose, suicide, and alcohol-related liver disease (hereafter “DSA mortality”) with long-term changes in local labor market conditions for blacks and whites separately. We argue that the measures of the local labor market used in the existing literature do not capture the degree of geographic and temporal variation in economic opportunity, nor the degree of geographic variation in economic desperation, particularly among participants in the low-wage and unskilled labor markets. This project responds to the debate following Case and Deaton (2015), which found that, contrary to trends among other demographic groups and other industrialized countries, there has been an increase in all-cause mortality rates among white non-Hispanic adults at midlife, between ages 45 and 54. Case and Deaton also found that this increase is almost fully explained by increasing DSA mortality rates, which they speculated could be linked to increasingly bleak economic conditions and inter-generational mobility. However, recent research suggests that economic conditions are minor correlates of DSA mortality increases in comparison to changes pharmaceutical prescription practices and the geographic expansion and growth of the illicit drug market (Ruhm 2017, 2018). These studies have focused on measures such as median income, income inequality, un/employment rates, or economic mobility. However, these measures do not capture the degree of variation in economic conditions, both overtime and by geography. We use the National Vital Statistics System (NVSS) micro-data, which afford us more precise estimates and allow us to evaluate differences in the health effects of local labor market conditions for the different sectors of the labor market. We estimate least squares regression models with logged mortality rate as the dependent variable using both contemporaneous and lagged county-level variables constructed using publicly accessible data from the US Census and Bureau of Labor Statistics. The variables of interest include: high school and college graduation rates; unemployment rate; proportion jobs at or below minimum wage; proportion jobs requiring a college degree; occupational segregation; income growth by quintile; and average job offshorability. These macro-economic conditions not only signal the precariousness of poverty or job loss and unemployment, but also signal the strain of living in a labor market characterized by increasingly limited growth opportunities due to structural shifts in the distribution of jobs. This project will allow policy makers to more effectively understand the underlying trends of the opioid epidemic and better predict its trajectory, thus allowing for a better policy response.