Twenty years after that paper was published, the evidence now that there is at least some predictability in the markets is even broader and deeper. There are still no guarantees, and risk remains alive and kicking, but it's clear that stock price behavior isn't a pure random walk. Markets aren't completely random over, say, three-to-five year periods, economists now recognize. That means market valuations can be analyzed and there can be expectations about medium- and long-term investment horizons.

In coming to understand these issues, researchers have identified a time-varying risk premium for equities. This idea stands in direct contrast to the older belief that it's best to think of the expected equity risk premium as stable and unchanging through time, which implies there should be a constant asset allocation-for example, a mix of 60% equity and 40% bonds. The old wisdom says this mix should remain in place regardless of market conditions-if a pure random walk describes equity price behavior all of the time.

But such a static view of asset allocation no longer fits with the world as we know it. Expectations for the equity risk premium fluctuate. That's not necessarily evidence that the market is inefficient, although some economists-and quite a few investors-think no less than this. Regardless of what a time-varying risk premium says about the efficient market hypothesis, the fact that so many different strands of research have begun to agree on this point suggests that it's based on a fundamental market truth.

The devil, of course, is in the details for designing and managing an asset allocation strategy. Still, we can take some basic lessons from the last 20 years of financial research. Assume, for instance, that one-quarter of the variance in equity returns is predictable, as Fama and French report. That suggests that one-quarter of the equity allocation should be managed dynamically.

Yes, that sounds like market timing. But we should distinguish between trying to speculatively time markets in the short term and modestly adjusting strategic asset allocations according to current market fundamentals. The latter is already widely accepted in fixed-income investing. If you buy a ten-year Treasury note at, say, a 5% yield with the intention of holding it to maturity, the expected return is, of course, 5%. If the ten-year note's current yield falls to 4% a month later, no one bats an eye if investors lower their return expectations and change asset allocations.

Applying a comparable framework for stock allocations, by contrast, is too often dismissed as hopelessly speculative. But that's a misguided view, based on a 21st century reading of market behavior.

"There's a difference between trying to identify short-term direction and trying to identify a prospective long-term rate of return," says Hussman. Bonds as well as stocks, he explains, have a duration (the average date at which the cash flows arrive discounted). Estimating duration for equities is a bit tricky, of course, but it's worthwhile for estimating stock returns for seven years and beyond, he says. "The job of a financial planner is to align the duration of the investments held by their clients with the expected date that those investments will be called on."

A big part of that responsibility for the planner is to consider the duration of equities generally, which means looking at market valuation.

Dynamic asset allocation, as a result, can be justified as managing investment duration to match client liabilities. Some may call that market timing, but 20 years of financial research and the track records of value-oriented money managers suggest otherwise.

True enough, people are quite a bit less confident about projecting returns for equities using dividend yields, P/E ratios, etc. than they are modeling fixed-income performance. But there is plenty of room for conviction in this area, particularly when valuations are at extremes. And to ignore the possibility means ignoring the long arc of financial history.

James Picerno is a freelance financial writer and editor of The Beta Investment Report (BetaInvestment.com).