Citation

Neumayer E. Soc. Sci. Med. 2004; 58(6): 1037-1047.

Affiliation

Department of Geography and Environment, London School of Economics and Political Science, Houghton Street, WC2A 2AE, London, UK. e.neumayer@lse.ac.uk

Erratum On

Soc Sci Med 2004;59(9):1993.

Copyright

(Copyright © 2004, Elsevier Publishing)

DOI

unavailable

PMID

14723900

Abstract

In his article with the provocative title "Are Recessions Good for Your Health?", Ruhm (J. Health Econ. 21(4) (2000) 659) has found robust and consistent evidence that the total mortality rate, age-specific mortality rates as well as most specific mortality causes are pro-cyclical. His finding that high unemployment rates are associated with lower mortality and vice versa stands in stark contrast to Brenner's earlier work, who found the opposite effect, possibly after a time lag. Ruhm controls for state-specific effects in a panel of US states over the period 1972-1991, whereas Brenner's work is based on time-series analysis. Extending and improving upon Ruhm's original analysis, we analyse the effect of state unemployment and economic growth rates on mortality in the states of Germany over the period 1980-2000, both in a static and a dynamic econometric model. Controlling for state-specific effects, we find evidence that aggregate mortality rates for all age groups taken together as well as most specific age groups are lower in recessions. The same is true for mortality from cardiovascular diseases, pneumonia and influenza, motor vehicle accidents and suicides, but not for necessarily for other specific mortality causes. In particular, there is never a statistically significant effect on homicides, other external effects and malignant neoplasms. There are also few differences apparent between the effect on male and female mortality. If we do not control for state-specific effects, then we often arrive at the opposite result with higher unemployment being associated with higher mortality. This suggests that a failure to control for time-invariant state-specific effects leads to omitted variable bias, which would erroneously suggest that mortality rates move counter-cyclically. Overall, we can confirm Ruhm's main finding for another country: recessions lower some, but not all, mortality rates in the case of Germany.

Language: en