But even two factors aren't enough. In a world with imperfect information and mere mortals making judgments, strategic-minded investors should mine intelligence from multiple sources in search of corroboration and context. The limitation of any one variable is somewhat muted in a multifactor model, explains the study by David Rapach, et al. Why does a strategy of combining factors seem to work better? For several reasons, according to Rapach and his co-authors. One is that the multifactor approach also fares well in projecting crucial macroeconomic variables, including real GDP growth and real earnings growth.
In other words, the multifactor model for asset pricing identified by Rapach and colleagues appears linked with the business cycle, which lends additional credibility to the results. If the ebb and flow of the economy is ultimately the source of risk premiums, there should be evidence of the connection. Yet every financial variable stumbles from time to time as a robust predictor. The good news is that they don't all wilt at the same time. By combining factors that capture different aspects of the economic cycle, an investor can partially compensate for the shortcomings of any one factor.
Yet we shouldn't be so naïve to think that simply looking at a range of data offers an easy and enduring solution. A 1995 study, for instance, shows that while economic factors are useful for estimating equity returns, each factor's overall forecasting power fluctuates along with market and economic cycles (according to the report, "Predictability of Stock Returns: Robustness and Economic Significance," by M.H. Pesaran and Allan Timmermann in the Journal of Finance, September 1995). This research shows that the "predictability of stock returns may indeed be particularly pronounced in periods of economic 'regime switches' where the markets are relatively unsettled and investors are particularly uncertain of which forecasting model to use for trading." In the comparative calm of the 1960s, for instance, stock returns were less predictable; that was followed by the volatile 1970s, when predictability increased.
The basic message: Economic variables, including interest rates, inflation and the money supply, offer clues about expected stock returns. The trouble is that these variables' usefulness in making predictions generally changes over time along with market volatility. The reason, of course, is tied to the business cycle.
It's hardly surprising to find macroeconomic forces behind the equity risk premium-which can also be considered a reward for an investor assuming recession risk. The payoff will vary according to the starting date of each investment and the prevailing market conditions, including the price of risk. As a result, the anticipated return is predictable to some extent, though how much we can predict varies, as does the expected return.
The fluctuation in the expected payoff needn't be a sign that the market is hopelessly irrational. There's no reason why equilibrium prices shouldn't convey some information about future returns. Nor should we be shocked to learn that the degree of predictability is subject to change. When we have more confidence in the message from one or more variables, the general economic risk is usually higher. Perhaps that's because investors require more incentive to hold securities in periods of high volatility. Think of the extraordinarily high dividend yield (i.e., low prices) early in 2009 and then the low yield (high prices) in the late 1990s. The former was probably a better window on future returns than the latter, or at least a more timely view.
More generally, the predictions offered by the dividend yield, the yield curve, etc. seem to be less reliable in times of economic plenty. But this makes sense. In fact, it would be surprising and inefficient if it were otherwise. Investors are more willing to speculate during economic expansions. In periods of economic turmoil, by contrast, visibility is at a premium and the market tends to respond by offering a higher level of predictability, albeit one that's still well short of absolute.
Economists have been turning up evidence over the years in support of this evolving state of equilibrium pricing. Does that mean that the investing challenge has been "solved"? No, not even close. Financial economists have learned a lot about market behavior over the last several decades, but investing is still risky, particularly in the short run.
Yet even if the equity risk premium is partly predictable over medium- to long-term horizons, as it seems to be, that's still no guarantee either. That means investors should follow the old advice: diversify, use the market weights as an initial guide for designing portfolios and change asset allocation cautiously. How should we change it? And to what extent? The answer begins by keeping a close eye on the business cycle.
James Picerno is a freelance financial writer and editor of The Beta Investment Report (BetaInvestment.com).