Two new studies find that unemployment at older ages may shorten life and that the gap in life expectancy between less and more educated workers is widening. Though neither result may seem surprising, the first is at odds with some previous research, while the second reinforces earlier findings but provides shocking new statistics—notably the fact that the least educated white women have seen their life expectancy at birth fall by five years since 1990, as highlighted in a recent New York Times article.

A seminal paper by Christopher J. Ruhm (2000) found that recessions were associated with lower mortality rates, a counterintuitive result confirmed by later studies. Ann Huff Stevens et al. (2011) identified a possible reason: Reduced employment opportunities in the broader labor market appeared to leave nursing homes better staffed, explaining why the pro-cyclical mortality effect was concentrated among seniors.

In other words, while higher unemployment may be associated with lower mortality, this doesn’t necessarily mean working is bad for your health. Later research focusing on workers who lost their jobs (as opposed to economy-wide unemployment rates) found that displaced workers experienced elevated mortality, as you would expect after a loss of income and health benefits (Sullivan and von Wachter 2009).

In the latest twist, economists at Wellesley College (Coile et al 2012) took another look at the relationship between unemployment and mortality rates, this time considering long-term as well as contemporaneous effects. Exploiting state-by-state differences in unemployment to tease out small but statistically significant effects, they found that workers who were in their late 50s or early 60s during periods of high unemployment had shorter life expectancies, a finding consistent with reduced income, health insurance coverage, and health care utilization. Assuming job loss was the main explanation, the authors estimated that a worker who lost his or her job at age 58 might live three fewer years as a result.

Interestingly, the effect appears limited to workers who lost their jobs a few years before becoming eligible for Social Security—those aged 57 to 61. Workers who experienced recessions earlier (ages 55–56) or later (ages 62—65) had no significant long-term change in life expectancy. The authors found that workers who experienced recessions in their mid-50s were more likely to be employed at older ages, presumably delaying retirement in an effort to make up for lost income and offsetting the impact of earlier unemployment. Unemployed workers in their late 50s or in their 60s, however, may not have the option of reentering the workforce due to age discrimination or for other reasons. However, those who experienced recessions after becoming eligible for Social Security at 62 appeared able to weather recessions with no significant effect on mortality, presumably also assisted by eligibility for Medicare at age 65. The authors therefore cautioned that policy makers contemplating raising the age of eligibility for Social Security and Medicare consider the role these programs might play in mitigating the negative effects of unemployment shocks.

The latest research on life expectancy and educational attainment also has implications for retirement policy. Though the most startling results relate to life expectancy at birth, the authors also found growing disparities in life expectancy at 65, with life expectancy declining for white seniors with a less-than-high-school education. Specifically, Olshansky et al. (2012) found that between 1990 and 2008 life expectancy at 65 declined from 19.5 years to 17.6 years for (non-Hispanic) white females with 11 or fewer years of schooling, and from 15.1 years to 14.4 years for their male counterparts. Meanwhile, life expectancy at 65 rose from 19.1 to 21.7 years for white women with 16 or more years of education, and from 16.1 years to 19.4 years for their male counterparts (numbers taken from an appendix linked to the paper online). While educational disparities also grew among blacks and Hispanics, less educated blacks and Hispanics did not experience declines in life expectancy between 1990 and 2008.

This reinforces a growing body of research on life expectancy and socioeconomic status showing disadvantaged groups falling farther behind and sometimes even losing ground. For example, Meara, Richards and Cutler (2010) earlier found that life expectancy at 25 for less educated white and black women fell by 0.9 and 0.2 years respectively between 1990 and 2000 (here is a partial list of studies). The New York Times article reports that American women now have the shortest life expectancy among developed countries.