Background As mortality is more and more concentrated at old age, it becomes critical to identify the determinants of old age mortality. It has counter-intuitively been found that mortality rates at all ages are higher during short-term increases in economic growth. Work-stress is found to be a contributing factor to this association, but cannot explain the association for the older, retired population.

Methods Historical figures of gross domestic product (Angus Maddison) were compared with mortality rates (Human Mortality Database) of middle aged (40–44 years) and older people (70–74 years) in 19 developed countries for the period 1950–2008. Regressions were performed on the de-trended data, accounting for autocorrelation and aggregated using random effects models.

Results Most countries show pro-cyclical associations between the economy and mortality, especially with regard to male mortality rates. On average, for every 1% increase in gross domestic product, mortality increases with 0.36% for 70-year-old to 74-year-old men (p<0.001) and 0.38% for 40-year-old to 44-year-old men (p<0.001). The effect for women is 0.18% for 70-year-olds to 74-year-olds (p=0.012) and 0.15% for 40-year-olds to 44-year-olds (p=0.118).

Conclusions In developed countries, mortality rates increase during upward cycles in the economy, and decrease during downward cycles. This effect is similar for the older and middle-aged population. Traditional explanations as work-stress and traffic accidents cannot explain our findings. Lower levels of social support and informal care by the working population during good economic times can play an important role, but this remains to be formally investigated.