Between 1991 and 1993 the Finnish economy suffered a deep recession and health and social services were cut by 25%. The result was a fall in mortality. Sweden had the same experience, and both countries thought that the falls in mortality might have been caused by falls in consumption of tobacco and alcohol. Can recessions and cuts in health and social services really be good for health? A major study in the Quarterly Journal of Economics (probably not read by many BMJ readers) provides strong evidence that this is the case.

Christopher Ruhm from the University of North Carolina studied aggregate data from 1972 to 1991 for all 50 states in the US and the District of Colombia. During that time total mortality fell from 943 per 100 000 to 860 with deaths from heart disease falling by 27% but deaths from cancer increasing by 23%.

During these years unemployment and total mortality were inversely related with a 1% increase in unemployment being associated with a 0.5% drop in total mortality. As might be expected, the biggest effect is in those who are most active in the labour force, those aged 20-44 ( a 2% fall), and lowest among those who are retired (a 0.3% fall).

Of the 10 major causes of death, eight fell during recessions. A 1% increase in unemployment was associated with a 3% fall in deaths from car crashes, a 1.6% fall in deaths from other accidents, and a 1.9% fall in homicide. Deaths from cardiovascular disease fell 0.5% with a 1% increase in unemployment, and deaths from pneumonia/influenza by 0.7%. The only cause of death that increased was suicide—by 1.3%. Cancer deaths did not change.

What might explain these changes? Ruhm used other datasets to examine changes in tobacco and alcohol use, BMI, physical activity, diet,and preventive medical care. He found that smoking falls during a recession but total alcohol use is not affected. Physical activity and consumption of fruit and vegetables increases, whereas fat intake and BMI both fall. There was no discernible effect on preventive medical care.

These patterns seem to me to fit together. Deaths from car crashes and accidents may fall because there is less driving and fewer work related accidents. Deaths from cardiovascular disease fall because of a drop in risk factors, particularly smoking—which fits with the surprisingly rapid falls in hospital admissions with heart attacks that have followed smoking bans. Deaths from cancer don’t change because cancer is a slow death, and unsurprisingly deaths in suicide increase because of the deterioration in mental health associated with unemployment.

So happily we can expect people in the recession hit countries of Europe to be living healthier lives and dying less.

I thank Bianca Frogner for pointing me towards Ruhm’s paper.

Richard Smith was the editor of the BMJ until 2004.

**An typo in the third paragraph has been corrected. Previously it read “a 1% increase in employment” It should read “a 1% increase in unemployment.”