[as_i_fall]

Well obviously nothing is guaranteed, but something like 90% of 20 year windows in the last 100 years would in fact result in positive real returns.

http://archive.nytimes.com/www.nytimes.com/interactive/2011/...

Note that the light red color is actually still indicating a positive real return, this is particularly relevant in the 70s and early 80s when a 2% real return would be a much higher nominal return.

Agreed that some people expect returns for time horizons as short as 10 years though, which is clearly a mistake.

[BeetleB]

Interesting: I did my own analysis[1] and did not show a loss for any 20 year period:

http://blog.nawaz.org/posts/2015/Dec/pay-down-mortgage-or-in...

Must be some differences in the data they have vs what I could find. I did include inflation and dividends, but not taxes - wondering if that brought it to the negative.

[bryanlarsen]

The last 100 years is only a weak predictor of the future.

[as_i_fall]

If you find a stronger one I'll be all ears