Unlike all the other hard sciences I know, it is very hard to demonstrate or contradict anything in evolutionary biology. Thus, one should be especially careful when reading about it.
1. Women can't have children when they are old.
2. Women have genes that promote healthy organs in old age.
So how exactly can those genes be selected through evolution, if the advantage they give doesn't have effect on the number of offsprings.
My point is not to prove my personal "theory" about women living longer than men. Rather, I contend that there is almost never One True Cause for anything in evolution. All we can do is propose many different competing factors, and say, with a lot of hand-waving, that factor X was the most important in such and such cases. Newspapers often translate that into "Y happened because of X", but that's just bad reporting.
Here's another hypothesis. Quoting the Wikipedia article on the Medawar theory of ageing:
"Nature is a highly competitive place, and almost all animals in nature die before they attain old age. Therefore, there is not much motivation to keep the body fit for the long haul - not much selection pressure for traits that would maintain viability past the time when most animals would be dead anyway, killed by predators or disease or by accident."
This is particularly true for men, who were (and still are, to an extent) usually doing more dangerous jobs than women. Thus, they die from accidents or predation earlier, so there is less selective pressure for genes that promote healthy organs in old age.
Another possible scenario: there is more variation in male traits than in females, and age is more than linearly dependent on these traits. Thus, even if E[X] is the same for males and females, for some trait X, the average age E[X^n] will be lower for males.
Evolutionary biology is full of reasonable-sounding explanations for just about anything. As far as I can tell, Occam does not often apply, unfortunately.