The news is flush with stories of "record highs" in the U.S. stock market, but all of the stories I see make the error of looking at nominal prices and excluding the effect of dividends (and inflation since they are looking at nominal prices).
If you look at real values, and assume dividends are fully re-invested, you can come up with a chart for the S&P500 that looks like this:
This graph shows you that the recent news saying we've reached record highs since 2007 is misleading. The true value of stocks has just barely reached a new high since September 2000, 13 years ago!
This means that an investor who, on September 2000, put money in an ETF like SPY and dutifully reinvested all their dividends for 13 years would have about the same purchasing power today as back then -- zero real return!
Let's take a different look at the historical data. The graph below shows how many months the market has gone without hitting a new real, dividend-reinvested high.
You can see that there have been many periods of "drought" in the market. Interestingly, the period in the 1970s and early 1980s was 12 years in length, and the period we have been in since 2000 lasted 12.5 years... two samples don't make for proof, but it is interesting that the two big droughts in modern times were of very similar lengths. It's also interesting that the famous "Death of Equities" article that appeared in Business Week was written after 7 or 8 of the drought years, and the recent financial crisis was also 7 or 8 years into the drought. I don't know as much about the history of the droughts in the '40s.
If the news of record highs is making you nervous and prompting you to sell (thinking that stocks have had a good run), look at this graph and ask yourself whether zero real return over 13 years is really a "good run."
If anything, the past history shows that long periods of zero real returns (the last being from 1972 to 1984) are followed by bull markets where the real return "catches up" to the long-term average of ~6%.