Cohort factors impinging on suicide rates in the United States, 1990-2010

We use CDC microdata on cause of death and CPS data on populations by age to create suicide rates for five-year age groups at five-year intervals, further cross-classified by race/ethnicity, education, and marital status. We examine the suicide history 1990-2010 of U.S. birth cohorts, net of age and cohort linear trends. These de-trended cohort deviations follow familiar patterns: most pronounced in the Baby Boom, least pronounced during the Baby Bust, they illustrate the so-called Easterlin effect. Suicide rates for women show similar patterns as suicide patterns for men. We show persistence of those effects net of micro factors (especially education and marriage) implicated in suicide behavior and correlated at the macro level with relative cohort size. Analysis of suicide patterns over time for high- and low-educated men and women shows that white men with low education face a sharp increase, significantly above the linear time trends, in suicide rates among cohorts born between 1955 and 1970. This bump is mostly unrelated to secular trends of increasing average educational attainment rates, at least if no interaction between age and cohort is involved in the explanation. No obvious pattern related to cohort size is found for African-American high- and low-educated men, which makes sense given the very different historical dynamics for this minority sub-population.