One explanation is that high-school graduation rates have increased, in part because high school has become less challenging. Those who fail to finish today are therefore more likely than ever to have done so due in part to problems directly and indirectly related to both poor health and underemployment. This is not, however, to minimise the direct contribution of unemployment on health. Unemployment can be humiliating, stressful and depressing, and studies consistently find that losing a job tends to put a hitch in one's health.

However, that unemployment is unhealthy for the unemployed does not necessarily mean a higher rate of unemployment is bad for health generally. According to Peter Orzsag, "Deep economic declines, such as the one we experienced in the U.S. a few years ago, probably lengthen life expectancy. This is exactly the opposite of what most people believe." He estimates that a one-point increase in the unemployment rate predicts a drop in the mortality rate of a third of a point. What gives?

A small part of the answer is that people drive less in recessions, reducing traffic fatalities. Mr Orzag is more interested in other, less obvious, mechanisms. A survey of Icelanders, who recently suffered a severe economic meltdown, showed that recession reduced smoking, heavy drinking, fast-food consumption and indoor tanning, while increasing the share of the population getting a good night's sleep. A different study, focused on America, found that recession improves air quality, leading to a decrease in deaths from respiratory and cardiovascular ailments. This accounts for about a third of the correlation between higher unemployment and increased mortality. Least intuitive of all is a result that turns on the unpleasantness of working in nursing homes. A weakening economy increases the staffing ratios of assisted-living facilities by squeezing health-care workers out of more desirable but scarcer jobs.

So unemployment kills, but the economic downturns that cause large increases in unemployment save more life-years than unemployment steals. What are we to make of this? Mr Orzsag says:

He's right. There's "some small consolation" in this, for some of us. But it's cold comfort for those nevertheless getting eaten alive by the stress of their joblessness. To my mind, the upshot is that the distributive consequences of recession are even worse than they may have seemed. During recessions, folks who have it good anyway—folks with good jobs and gas money—get lighter traffic, better air, and shorter lines at Whataburger, in addition to a few extra days of life. Those who bear the brunt of the downside can hardly afford their unhealthy pleasures, and anyway come to an accelerated end. It's a raw deal. Good countercyclical policy—generous automatic stabilisers and a central bank dogged about hitting a reasonable inflation target—not only shortens the season of suffering, but protects the most vulnerable among us from getting both objectively and relatively screwed.