There appears to be little difference between genders in the mortality rate due to drug suicide. In addition, the table indicates that less-educated adults are more likely than college graduates to be nonwhite, uninsured and to live in a rural area. Figures 1 plots average cause-specific death rates for less-educated adults over the 1999-2015 period. For all three causes, mortality has increased dramatically over the sample period. In particular, unintentional drug overdose deaths have increased dramatically among less-educated individuals. Over the sample period, drug non-suicide mortality rates increased nearly four-fold for this group. Non-drug suicides also increased substantially among this group. The increase is especially large for women, who experience a 50 percent increase in suicide rates over the sample period. To estimate the causal effects of minimum wages and the EITC on mortality, we adopt a quasi-experimental approach, leveraging panel variation in state economic policies over time. This approach allows us to control fully for time-invariant characteristics of each state as well as changes in mortality over time. Specifically, we estimate standard difference-in-differences regressions that control for state and year fixed effects. In addition, the models control for a number of state policies and the cell-level demographics as outlined above.Over the sample period, mortality rates have changed differentially by race . To account for this change, our models include interaction terms between calendar year and share Hispanic and share non-white. Educational attainment has increased considerably over this period: from 1999 to 2015 the share of U.S. adults aged 25 or older who had not completed high school or college fell by 30 percent,seedling grow rack from 16.6 percent to 11.6 percent. As a consequence, the average person without a high school degree is likely more negatively selected in the later years of the sample .
To account for this selection bias, our models include interaction terms between calendar year and the share of high school graduates. The two key independent variables are the minimum wage and the EITC. We use the natural logarithm of the minimum wage, which takes on the higher of the federal minimum wage or the minimum wage in the state . Researchers sometimes use the Kaitz index, defined as the ratio of the minimum wage to median full-time wage, or divide the nominal minimum wage by state level cost-of-living indices. In our view, both of these approaches complicate the interpretation of results, as these variables will reflect changes in minimum wage policy as well as changes to the respective denominators. To illustrate, suppose a state experiences a local economic boom, resulting in higher average nominal wages and local inflation. The boom would reduce both the Kaitz index and the COL-adjusted minimum wage, even in the absence of any change in the nominal minimum wage. If the improved local economic conditions also reduce deaths of despair, the Kaitz index and COL-adjusted minimum wage will be positively correlated with mortality, potentially leading to the erroneous conclusion that higher minimum wages increase mortality.In these models, the fundamental assumption is that we can obtain causal estimates of policy effects by comparing states that have different minimum wages and EITC rates within the same year. For this approach to be valid, the parallel trends assumption must hold; that is, changes in state minimum wages and EITC rates should be uncorrelated with unobserved drivers of mortality. For example, states with high minimum wages are geographically clustered, more likely to vote Democratic, and more unionized . Including state fixed effects in our regression models will control for time-invariant heterogeneity among states. However, these states may have different economic fundamentals or different changes in other policies, compared to lower minimum wage states.
A lack of parallel trends would violate our research design. To increase the likelihood that the parallel trends assumption holds, our models include controls for a range of potential confounders. Still, we may be concerned that unobserved spatial heterogeneity could bias the estimated effects. To address this possibility, one approach in the literature implements a border-county pair approach, exploiting variation in economic policies within pairs of contiguous counties that straddle a state border . However given the relatively low incidence of cause-specific deaths, we do not have the statistical power to use this estimator. Instead, we implement a dual approach to testing whether the parallel trends assumption appears to hold in this setting. First, we estimate effects on the cause-specific mortality of college graduates using the same empirical specification above . Since college graduates are much less likely to be exposed to minimum wage jobs or to be eligible for the EITC, any effect on this group is likely spurious, reflecting divergent trends between high and low minimum wage states. Second, we estimate event study models that capture the time path of effects around the time of minimum wage increases. The intuition behind these models is that higher minimum wages or EITC rates should not have any effects on mortality in the years leading up to the policy changes. We estimate separate event study models for each of the two policies. For the minimum wage, we define an event as a year-on-year increase in the state or federal minimum wage of 25 cents or higher . The event study sample includes all events occurring between 2004 and 2010; we require at least five years of pre-event data, during which we require that the state does not increase its minimum wage . Similarly, we include five years of post-event data, to estimate the path of any effects over time. Using this definition, 46 of the 47 states experience a qualifying minimum wage event. To study the effects of state EITC policies, we focus on the 15 states that introduced state EITC top-ups between 2000 and 2014.
We retain the 11 states that introduced state EITC earlier in the estimation sample, together with the 25 states that do not operate state EITCs during the sample period. We begin our analysis by estimating equation on the primary sample of less-educated adults as well as on the placebo sample of college graduates. Next, we present results from the event study model, together with a discussion of parallel pre-trends. Finally, we present results from a sub-sample analysis, using auxiliary data from the CPS ASEC, comparing estimated effects on mortality with corresponding effects on poverty rates. Table 2 presents the estimated models for the three causes of death. The upper panel shows effects for adults with high school or less, while the lower panel shows estimates for the placebo sample . Higher minimum wages and EITC credits have no statistically significant effects on drug deaths. While we estimate a marginally significant negative effect of the EITC on unintentional drug deaths, the point estimate is similar in magnitude to that found in the placebo sample, suggesting the effect may be spurious. Meanwhile, results in column 3 indicate both policies significantly reduce non-drug suicides. A ten percent increase in the minimum wage translates to a 3.6% reduction in suicide deaths for less-educated adults. For the EITC, a ten percent higher maximum credit reduces suicides by 5.5%. The coefficients are statistically significant at the five and one percent levels respectively. Reassuringly for our study design,greenhouse growing racks the placebo models fail to find significant effects of minimum wages or state EITC policies on suicides among adults with higher education levels. The regression models include a number of state characteristics and policy variables. Appendix table 2 summarizes the estimated effects of these covariates. We stress that the estimated coefficients of these covariates represent correlations only; we do not claim that the underlying variation is exogenous, and as such the estimated coefficients should not be given a causal interpretation. Both the share uninsured and the state unemployment rate predict significantly higher mortality from drug overdoses, both intentional and unintentional. The correlation between unemployment and drug deaths suggests a role for economic factors in explaining drug mortality, even if the economic policies we study do not significantly shift outcomes. At the same time, the positive coefficient could also reflect reverse causality: higher rates of drug abuse could lead to higher local unemployment rates. Research suggests that expanding access to healthcare could improve mental health and reduce depression . We include two measures of state healthcare coverage: an indicator variable equal to one for states that implemented Medicaid expansion to cover all low income adults after 2014, as well as the estimated share of individuals in each cell who are uninsured. Our models indicate that states that chose to expand Medicaid under the Affordable Care Act have higher mortality rates, although previous studies indicate this result reflects divergent trend between expansion and non-expansion states rather than causal impacts of Medicaid expansion . On the other side, a higher uninsured rate predicts higher drug mortality; though again interpretation of this coefficient is potentially complicated by omitted variable bias as our estimate likely reflects a combination of insurance impacts and effects of unobserved determinants of insurance status.
Our two measures of state drug policy – medical marijuana and state PDMP requirements – are not statistically significant in predicting drug mortality, although the point estimates of the PDMP coefficient are negative in for all three outcomes and marginally significant at the 10 percent level for non-drug suicides. With these exceptions, the covariates are not statistically significant in explaining variation in non-drug suicides. Table 3 presents results by gender. The upper panel now shows results for less-educated women, while the lower panel shows results for less-educated men. The effect of economic policies on suicide deaths appears to vary by gender. For women, a ten percent increase in minimum wages leads to a 4.6 percent reduction in suicide deaths. The estimates are significantly different from zero at the five percent level. For men, the point estimates are smaller, and the effect of the minimum wage is now only marginally significant at the 10 percent level. The relatively low precision of the estimates means we cannot reject that the male effect sizes are equal to the female effect sizes. Still, the gender difference is consistent with differences in exposure: compared to men, women are more likely to work minimum wage jobs and to be eligible for the EITC. During the sample period, mortality rates shifted differentially by race. While midlife mortality for less educated white non-Hispanics has increased, mortality rates of blacks and Hispanic adults continue to decline until 2013, when they begin to increase . Table 4 shows effects by race/ethnicity: the upper panel shows effects for white nonHispanics, while the lower panel shows the models estimated for racial minorities.The models do not detect any differential effects of minimum wages on suicide for white nonHispanic and other racial/ethnic groups. Comparing the estimates in Table 4 with the estimated effects on suicide for the pooled sample, the effects are estimated with less precision, but the effect sizes are remarkably similar. The EITC meanwhile has larger estimated effects on people of color, though once again precision issues suggest we should interpret this difference with caution. We have also estimated models by race/ethnicity and gender– see Appendix table 3. Among men, estimated effects of the minimum wage and the EITC are larger for African Americans, people of Hispanic origin, and people of other races, which is consistent with their greater exposure to minimum wage jobs relative to white non-Hispanic men.Among women, the reverse seems to be the case: the minimum wage significantly reduces suicides among white non-Hispanics, with no statistically significant reduction for other racial and ethnic groups. The negative effect of the EITC on female suicides does not appear to vary with race and ethnic origin. To assess the robustness of these findings, we estimate additional models, analyzing the effects on mortality rates rather than counts, as well as nonlinear models of mortality counts. These results, presented in Table 5, are consistent with our preferred specifications. All models find significant negative effects of minimum wage and EITC policies on non-drug suicides. Quantitatively, the effect sizes are somewhat smaller in these alternative specifications: the sample average non-drug suicide mortality rate is 17.6 per 100,000, implying that the estimated effects of a ten percent increase in minimum wages on mortality rates corresponds to a 2.1 percent relative reduction in suicides.