Debate over the expected return on the stock market is a hardy perennial for two basic reasons. First, the true ex ante market return can never be known with certainty. Second, the true ex ante return is forever changing. The problem, of course, is that investors must make decisions, even with imperfect information about the future. Where to begin? One possibility is a simple Gordon growth model that equates equity market return with the sum of the growth rate of dividends plus the current dividend yield. It's really an identify rather than a model, but it's useful just the same.

"The long-term increase in stock market value is entirely the result of the sum of long-term dividend growth and dividend yield calculated from the Gordon Equation," writes financial planner Willliam Bernstein in his indispensable book The Four Pillars of Investing: Lessons for Building a Winning Portfolio. Critics will note that this approach to looking ahead isn't perfect (nothing is, of course). For example, Antti Ilmanen warns in Expected Returns: An Investor's Guide to Harvesting Market Rewards that dividend yield "has become too narrow a measure of carry because firms increasingly use means other than dividends to distribute cash to investors, including share repurchases and cash acquisitions."

Nonetheless, dividend yield is still a good place to start, if only as a first stab at developing some context about how the market's pricing assets and what that implies about the return outlook. Just don't confuse a starting point with an end point. In any case, history suggests dividend yield offers valuable information. Consider, for instance, the relationship through the decades between the S&P 500's current yield and the subsequent value of $1 invested after 10 years. (The underlying data, by the way, comes from Professor Robert Shiller's website.)

The chart below tracks how a $1 investment rose (or fell) after a decade and how the initial investment compared with the current dividend yield at the initial purchase point. For example, the latest entry (shown at the right-hand side of the chart) indicates that the current yield in May 2002 was roughly 1.5%. A $1 dollar investment in the S&P that month would have been worth around $1.24 ten years hence.

The larger point is that the previous 50 years suggests that there's a relationship between current yield and the subsequent 10-year investment. It's not a perfect relationship. In the 1990s, in particular, the connection between yield and subsequent 10-year return went a bit crazy, albeit in favor of buy-and-hold investors. As the surge in the red line in the chart above reminds, investment returns were unusually high. Call it market irrationality or inefficiency. Whatever you call it, don't dismiss it-it can and probably will happen again, and not necessarily in favor of investors.

In any case, looking at the connection between current yield and subsequent return is an obvious starting point for deeper analysis in the dark art of projecting the return on the equity market. Suffice to say, analysts have their work cut out for them. Even the simple Gordon growth model raises a number of questions without easy or obvious answers.

For instance, let's assume that the future long-run return on the stock market is the sum of current yield plus the expected dividend growth rate. Okay, but what expected growth rate should we use in the calculation? As a back-of-the-envelope estimate we might look to history as a guide. But how much historical data is optimal? Unsurprisingly, the results vary considerably with different ranges.

Consider that the implied return on the stock market based on the trailing 50-year dividend growth rate plus current yield was 7.3% last month. But if we use the last 10 years as a benchmark for dividend growth the performance outlook falls to 3.1%.

There are various econometric techniques to figure out what's "optimal," or at least what appears to be "optimal." But that's a subject for another day. What's clear from the chart above is that the market is telling us that the expected return for equities generally is much lower at the moment compared with just a few years ago. Notably, the expected return for stocks was unusually high in February 2009. As it turned out, that was an excellent time to buy, given the subsequent rally. In fact, the surge in dividend yield around February 2009 hinted that expected return had taken wing.

You can't blindly accept these implied return clues, but neither can you dismiss them. There's plenty of science (and a lot of models) to consider in developing expected return estimates, but a fair amount of art inevitably comes into play too.

Predicting, in short, is still hardâ€¦ especially about the future.