It's not unbiased noise though. People with the highest likelihood to die soon (elderly men) have different voting preferences than the general public.
I'm not sure I'm thinking about this correctly, but I'm not sure that it's relevant that elderly men die younger than elderly women. I think it would only be a bias is they were dying at a faster rate, relative to their population size, than other groups. That is, the size of the elderly male population was shrinking compared with the elderly female population.