What was the life expectancy of an 18 year old at the time? That number was skewed by childhood mortality, which largely disappeared as a statistically significant factor starting with children after the 1940s but has no real effect on Social Security's budget.
[Typed a couple minutes later after actually searching for the info]
http://www.infoplease.com/ipa/A0005140.html has a table of remaining life expectancies for people of various ages.
I don't know if this table is predictive at the time (what actuaries in 1940 predicted) or if it has been corrected later with actual survival statistics since then.
Well, either way it lists the life expectancy of a newborn white male in 1939-1941 as 62.81 years, while a 20 year old's life expectancy is 47.76 more years (or 67.76 total). So five years of your original 60 year estimate is from childhood mortality.
An even bigger assumption was that the population would continue growing quickly, so one generation would rely on an even larger couple of generations following it to pay for its retirement. Between less natural growth and a longer retirement we now have fewer and fewer worker supporting each retiree.
but is your conclusion that the math of Social Security is basically something that does not work? That is my conclusion. I see it as a Ponzi scheme.