The survivorship bias applies to the U.S. data as a whole. I've found this paper [1] that adds data from 15 more countries, some of which experienced interruptions (germany) but still excludes those that suffered "permanent interruptions" (eastern bloc) to the sample. Needless to say, the results are not as exciting.
Memorable quote form the paper (quoted in turn from Samuelson):
> We only have one history of capitalism. Inferences based on a sample of one must never be accorded sure-thing interpretations. When a thirty-five-year-old lost 82% of his pension portfolio between 1929 and 1932, do you think it was fore-ordained in heaven that later it would come back and fructify to +400% by his retirement at sixty-five? How did 1932 Tsarist executives fare in their retirement years on the Left Bank of Paris?
Memorable quote form the paper (quoted in turn from Samuelson):
> We only have one history of capitalism. Inferences based on a sample of one must never be accorded sure-thing interpretations. When a thirty-five-year-old lost 82% of his pension portfolio between 1929 and 1932, do you think it was fore-ordained in heaven that later it would come back and fructify to +400% by his retirement at sixty-five? How did 1932 Tsarist executives fare in their retirement years on the Left Bank of Paris?
[1]: http://merage.uci.edu/~jorion/papers/risk.pdf