VoxEU Column Development Health Economics

Socio-economic adversity kills men; women are more resilient

Socio-economic conditions like wealth and education have been demonstrated to be powerful predictors of life expectancy. Similarly, over the 20th century, a tendency for women to outlive men has emerged in much of the Western world. Using data for countries from all around the world, this column shows how socio-economic factors mediate the sex differential in the probability of survival to advanced age. These patterns seem to transcend temporal and cultural barriers, but rather emerge and evolve as countries develop, and standards of living rise.

Ever since reports of a social class pattern of death among British Civil servants 40 years ago, it has been accepted among social scientists and many in the medical community that socio-economic conditions – wealth, education, high status jobs – are powerful predictors of life expectancy, more so than genes or quality of medical care (Phelan et al 2004). This relationship has been demonstrated consistently throughout the world. We have recently reported on an aspect that has been far less studied, namely the difference in impact of this gradient on men and women (Cullen et al 2015). Although it is not news that women, on average, outlive men, the degree of that advantage is extraordinarily variable. The sex-differential likelihood of a new-born making it from birth to age 70 in the richest US counties, based on current mortality rates, is scarcely a few percent, while in the poorest counties there is a 35% difference – a difference that translates to an almost 10 year gap in female versus male life expectancy. Put another way, there is a strong correlation between a variety of indexes of socio-economic conditions and the difference in probability at birth of survival to 70. The figure below shows sex-specific (top) and differential (bottom) rates for white Americans[NdS1]  in 2010 on the Y-axis as a function of various socio-economic indicators – male survival is much more sensitive to adverse socio-economic conditions than female. Women appear to be ‘resilient’. Results for black Americans (not shown) are similar for counties in which blacks are reasonably represented in the population.

Figure 1. Sex-specifc (top) and differential (bottom) rates of the probability of survival to 70 for white Americans in 2010

When did this pattern emerge?

The same relationship pertains currently when comparing sex differences in mortality among and within western European countries. But these countries enjoy relatively comparable culture, lifestyles and governance so we explored the sources available to trace the male and female mortality differences back in time. For the US and Europe we were able to reach back to 1900, allowing us to ‘re-discover’ an observation well-articulated almost 50 years ago by Omran (Omran 1971) – In the US and Europe 120 years ago, men outlive women on average![NdS2]  Omran made a convincing case that this was attributable to high fertility and high maternal mortality. With the beginning of modern hygiene and nutrition, alongside rapid economic development after the turn of the century (the ‘transition’), evidence shows that women start to have many fewer pregnancies and thrive, rapidly surpassing men in survival and life expectancy. By 1940 the same patterns of resilience we saw in 2010 in the US and Europe were already becoming statistically evident, and a few decades later the correlation starts to become strong enough to see with the naked eye, as in the present-day situation depicted above for US whites.

The evolving global pattern

Because some analysts have been tempted to attribute much of this evolution to common behavioural differences between the sexes in developed western countries – such as differential smoking rates or violent behaviours – we attempted to extend our work to places where cultures and habits differ, such as Eastern Europe (including the former Soviet Union) and the Low and Middle Income Countries (LMIC’s). From the Soviet experience we observe two consistent, remarkable things – first, the precipitous drop in male survival that occurred shortly after the collapse of the Berlin Wall, while women appeared remarkably less affected by this shock. Second, we see the same pattern of female resilience in the Oblasts (Russian states) that, for cultural reasons, are less prone to alcohol excess. Among developing countries the pattern of sex differences in mortality is also evolving in a parallel way. In the now middle income countries that began to expand economically mid-20th century – places like Brazil, Turkey, and Lebanon – the pattern of male vulnerability to adversity combined with female resilience is evident by 1970 (the year of the earliest available data for these countries) and gets stronger every decade thereafter. In the next poorer group of countries, including those like Ghana and Pakistan, in which the epidemiologic transition did not begin until around 1970, the resilience pattern is well ensconced by 1990. Sadly there are still some very poor countries which have not even begun this transition, in which maternal mortality and fertility remains high; in a handful there is no female advantage in mortality overall.

What might this mean?

While these relationships are not, of course, causal, in the sense education or wealth directly causes excess deaths in men, the consistency and strength of the finding in so many different places and cultures suggest that the causal factors are likely strongly associated with these broad indicators of social advancement. In particular we draw two broad inferences regarding the sex-difference. First, it’s hard to imagine from the data that any constant large portion of the difference is biologically programmed (e.g., that X-chromosome cells live longer) given the wide variability in the size of the sex mortality gap. On the other hand, the consistency in the differential mortality patterns over place (i.e., culture) and time makes equally un-compelling the interpretation that any single environmental factor or behaviour is largely responsible. Some theoretical and empirical arguments have been put forth suggesting a so-called ‘socio-biologic’ difference – differential ways of coping and adapting to circumstances – allowing for the small fixed advantage in the richest places on the one hand while potentially explaining the much larger gap in poorer ones. The recent film Winter Bones, depicting modern life in the Ozarks, would offer a vivid depiction of what the differential response to adversity might look like, at least in a US context.

One small hopeful signal

One might conclude from these findings that women have a superior built-in mechanism for adaptation to adverse conditions (from a survival perspective), and that the prospect of bringing male mortality into line with that of females is a pipedream, short of raising the overall level of prosperity. Since few counties or countries wouldn’t prefer that for their own reasons, there is little public health value in promoting economic growth as a primary strategy to enhance male survival. But we did explore the data for other factors possibly more amenable to policy intervention that might have a modifying role. And in fact we find a very interesting one, though it must be clear we have too little evidence at present to note these as more than interesting observations. With that caveat, the figure below illustrates the relationship between occupational similarity (in which ‘not working outside the home’ is considered an occupation) between men and women in each county, and the size of the sex gap in survival to age 70.

Figure 2. Sex-specific (top) and differential rates of the probability of survival to 70 by occupational similarity

In this way occupational concordance likely ‘explains’ (in the statistical sense, not the causal sense) the very high sex gap in Alaska, one of the 10 richest and best-educated states, but which has a low relative male survival rate. It is home to several of the counties in the US with the most dissimilar occupational profiles. To look at the same issue among other developed countries we used the WHO equity index, and find a similar relationship to the sex difference. Notable in this regard are Korea and Japan, two countries with much higher sex differences than would be expected based on education or income; these countries also have a very low gender equity score relative to other rich countries. The idea that narrowing gender social equality may reduce the mortality gap is also consistent with the steady overall reduction in that gap since about 1970 in the US and western Europe, coinciding with the rapid expansion of women’s occupational and social opportunities, although these were not the only changes in our society during this period. So while we remain far from confident about this conclusion given the limitations we have noted, it just may prove true that removing the hurdles for women in a male society hasn’t turned out so badly for men after all.


Cullen, M R, M Baiocchi, K Eggleston, P Loftus & V R Fuchs (2015) "The weaker sex? Vulnerable men, resilient women, and variations in sex differences in mortality since 1900", National Bureau of Economic Research, NBER Working paper W21114.

Omran, A R (1971) "The epidemiologic transition: A theory of the epidemiology of population change", The Milbank Memorial Fund Quarterly, 49(4): 509–538.

Phelan, J C, B G Link, A Diez-Roux, I Kawachi & B Levin (2004) "'Fundamental causes' of social inequalities in mortality: A test of the theory", Journal of Health and Social Behavior, 45(3): 265–285.