There is only one U.S. stock market, but there are countless strategies for estimating the expected risk premium for domestic equities.
The choices come in two broad flavors: fundamental and technical. We could, for instance, crunch the numbers using one of the many variations of discounting future cash flows to assess if stocks are “cheap” or “expensive.” Meanwhile, there are a number of so-called trend-following measures that are worth studying. How to choose? Actually, there’s a persuasive case for routinely analyzing a mix of so-called predictors for evaluating the outlook for stocks, and other asset classes, for that matter. To the extent that a broad and varied menu on this front represents opportunity, the world is brimming with possibilities.
In our research for The Beta Investment Report, one of the technical factors we consider is long-term return performance. One approach is putting the latest number in context with history. As an example, the chart below graphs the rolling 10-year annualized price return for the S&P 500 since 1910—a century of performance history.
We’ve identified the last three extreme points in the chart. The most recent is March 2009, when the rolling 10-year price return for the S&P was an annualized negative 5.1%. If you didn’t know anything about stocks or investing, simply looking at that point on the chart above tells you something valuable: an annualized loss of 5.1% for the previous decade is an unusually poor return relative to the historical record. That alone is a short cut the promised land, but neither is it chopped liver.
At the opposite extreme is October 2000. At that juncture in the autumn of the new century, the previous decade delivered an annualized price rise of 16.3%. Once again, a naïve review of the chart tells us that the decade that had just passed in October 2000 was extraordinarily good to investors in the S&P 500.
If we extend the history lesson back through time, the larger message is that returns fluctuate. The great question, of course, is whether they fluctuate with any degree of predictability? Yes, although this answer comes with a truckload of caveats.
That starts with the reminder that the future is always and forever uncertain. You can’t get blood out of a stone, or flawless predictions mined from historical data. But while no one can see the future with absolute clarity, estimating expected risk premiums isn’t always a random crap shoot that’s definitively hopeless. It’s taken decades of financial research to establish this fact, as we discuss in some detail in our recent book Dynamic Asset Allocation. Better late than never.
What’s the glitch? There are many. Arguably the first is that you can’t look to any one predictor and rest easy. In other words, every risk premium estimation methodology is destined to fail at some point. Indeed, everything fails regularly. It’s not always in advance, but as a general proposition, this caveat is a sure thing. As one example, the equity dividend yield is quite valuable for assessing the stock market’s outlook, but it falls well short of perfect. But as we’ve discussed over the years, including here, keeping an eye on this metric is productive for strategic-minded investors intent on minimizing risk and maximizing performance. Certain times are better than others, in fact.
The crucial issue is diversifying the set of predictors so as to survey the future using tools that a) offer some degree of robust forecasting power; and b) don’t all fail at the same time in the business cycle.
There’s a growing array of research studies that documents the power and necessity of using an array of predictors. In the next issue of the newsletter, we review several recent additions to this worthy corner of financial economics. One of the key lessons is that investors should spend time evaluating and assembling a diversified mix of predictors. All the more so for managing a multi-asset class portfolio. Different asset classes rely to some degree on different predictors.
As for rolling 10-year returns and its relevance to the stock market’s outlook: Yes, this too is one variable and it doesn’t always offer timely information about expected return. In fact, it’s fair to say that its forecasting powers are limited. But sometimes, those capabilities are unusually potent. There’s always some reason to wonder in real time, of course, but it’s a mistake to look away, particularly at extreme points.
The central challenge is looking for other variables that float when another predictor sinks. Alas, there is no optimal mix that fills all the gaps and delivers perfection in the aggregate. On the other hand, we can do better than simply throwing up our hands and hoping for the best. A careful review of the academic literature, along with real-world money management records, suggests no less. Therein lies our newsletter’s raison d’être: dynamic asset allocation isn’t a mug’s game after all.
It’s not always easy to fully explain why, or to cash in on the opportunity. But if you’re willing to put in the time, and crunch the numbers on a mix of variables (we regularly look at more than 30), there’s reason to think that enhancing the market portfolio’s risk premium is more than the stuff of dreams.