A common nugget of wisdom passed down from on high to the masses is to "buy and hold," which really means several things: Set fixed allocations to stocks, bonds, and cash; dollar-cost average into the portfolio with periodic purchases; and, above all, stay the course. I tell investors to do these things all the time, too, because on the whole it’s good advice.
It’s also a noble lie, an untruth told to investors to keep them from hurting themselves.
You can detect a contradiction when juxtaposing the advice with the well-established fact that, in aggregate, individual investors are perverse market-timers, selling low and buying high. If investors manage the feat of reverse market-timing, wouldn’t doing the opposite of what they do lead to successful market-timing?
Either the market is predictable or individual investors are not reverse market-timers. Something has to give.
Which is it?
Here’s a hint: Eugene Fama and Robert Shiller shared this year’s Nobel Prize in Economics in part for showing that returns are predictable over long horizons. They found when valuations are low, expected returns are high, and vice versa.
Yes, Fama, the father of the efficient-market hypothesis, believes the market can be sorta-kinda timed. In an interview with The New York Times, he said, "If I were to characterize what differentiates me from Shiller or [Richard] Thaler, it’s basically we agree on the facts–there is variation in expected returns, which leads to some predictability in returns. Where we disagree is whether it’s rational or irrational." 
This wasn’t a recent change of heart. Fama and many other researchers discovered return predictability all the way back in the 1980s. Somehow, during the past 30 years, the academic research was translated into the dictum that you should never, ever change your asset allocation, except in response to changing risk tolerance.
So why do experts urge investors to stay the course, even though there’s good evidence that valuation can predict long-term returns? I suspect it’s exactly because the market is somewhat predictable. The fear is if you give investors leeway to adjust their allocations, they’re going buy high and sell low. "Stay the course" is simple, memorable, and easier to implement than "buy when there’s blood in the streets, even if the blood is your own."
Another reason is that there’s very little statistical evidence managers have been able to exploit market predictability; this has been misinterpreted to mean that no one can time the market. Statistical tests of whether a manager can time market allocations are weak. After all, a business cycle usually lasts around five years. Even if some investors can time the market based on long-term forecasts, it’s devilishly hard to prove they exist with the tools and data we have.
Finally, it’s because the level of predictability is fairly low for the stock market. The cyclically adjusted price/earnings ratio, or Shiller P/E, one of the best valuation signals, will give you only a ballpark range of returns over the next five to 10 years. Exhibit 1 shows a scatter plot of five-year forward real returns for the S&P 500 versus its starting cyclically adjusted earnings yield (the inverse of Shiller P/E), using data from 1926 to 2013. As you can see, the relationship is very noisy. The signal becomes more useful at the extremes, suggesting you should aggressively alter your stock allocation only on rare occasions.
Bonds, on the other hand, are very predictable. Over the next three to five years, your returns are going to be close to the starting yield. In Exhibit 2, I plot the forward three-year nominal annualized returns of the Ibbotson Associates SBBI Intermediate Government Bond Index versus its starting yield, using data from 1926 to 2013. The resulting plot is about as clean as you’re going to get: Returns move one-for-one with yield in high-quality bonds. One would have to be daft not to take into account valuation when deciding one’s bond allocations.
Even junk bonds, which act like both stocks and Treasuries, are predictable. In Exhibit 3, I plot the forward three-year nominal annualized returns of the Bank of America Merrill Lynch High Yield Master II Index against its options-adjusted starting yield minus 2.6% to account for average annual credit losses. The data spanned 1996 to 2013. The relationship is also surprisingly clean, partly because junk bonds are relatively illiquid. Illiquid stuff is really easy to time, because extreme price movements tend to quickly reverse themselves.
Yes, Virginia, you can time the market–if you have the brains and the courage to do so. I wouldn’t be surprised if lots of experts preaching buy and hold are closet market-timers. Academics can be notorious hypocrites. In a 1994 speech at the USC Marshall School of Business, Charlie Munger said, " … one of the greatest economists of the world is a substantial shareholder in Berkshire Hathaway and has been for a long time. His textbook always taught that the stock market was perfectly efficient and that nobody could beat it. But his own money went into Berkshire and made him wealthy."  According to Fortune magazine, that economist was Nobel Prize winner Paul Samuelson, whose work on efficient markets inspired Jack Bogle to launch the first Vanguard index fund. [3,4]
1 Sommer, J. "Eugene Fama, King of Predictable Markets." The New York Times, Oct. 26, 2013.
2 Munger, C. "A Lesson on Elementary, Worldly Wisdom as It Relates to Investment Management & Business." 1994.
3 Setton, D. "The Berkshire Bunch." Fortune, Oct. 12, 1998.
4 Bogle, J. "Eugene Fama and Efficient Financial Market Theory." The Wall Street Journal, Oct. 18, 2013.