It’s difficult to exaggerate the significance of estimating expected returns in the quest for long-term investment success. An obvious statement, perhaps, but the widespread evidence that many (most?) investors earn unecessarily low or even negative returns over time suggests that the focus on expected returns falls well short of practical necessity.
Yet the fact remains that developing robust performance forecasts–and acting on the information–is essential, perhaps on par with risk management. As Cliff Asness writes in the introduction to Antti Ilmanen’s newly published Expected Returns: An Investor’s Guide to Harvesting Market Rewards
Projecting expected returns is difficult, of course, and so there’s no silver bullet here. There’s also the problem of time horizon, which can be a distraction for clear thinking on the subject. As Asness notes, “Good forecasts of expected return add up over the long term, but don’t matter for squat next week.”
The good news is that there’s no shortage of opportunity for developing robust estimates of future returns, as llmanen’s book reminds. In fact, the empirical and theoretical foundation for thinking that returns are partly predictable under certain conditions is more than marginal. As Professor John Cochrane points out in a magnificent survey of how financial economics has evolved over the last 40 years: the “pattern of predictability is pervasive across markets.” Why? “Asset prices should equal expected discounted cashflows flows.” You can’t count on that truism to pan out next month, or even next year. But there’s a fair amount of historical evidence for anticipating no less as a general proposition through time.
The techniques for predicting returns are varied and the interpretations of how to apply them are vast, but we can start with some basic observations. Perhaps the first “fact” is that historical performance varies. If returns were completely random, there would be little point to estimating expected returns. But decades of research, along with simple observation, tells us otherwise.
The empirical record also shows that dividends and other fundamental metrics sometimes hold clues about expected returns. For example, consider how trailing dividend yield varies with subsequent 10-year performance on the S&P 500 through time:
Is the apparent connection a coincidence? Unlikely. Markets make mistakes in the short run, but eventually there are very few if any inefficiencies to exploit over, say, 10-year spans. Risk and return are related in something approximating the predictions embedded in modern finance when measured over long stretches.
Simple historical returns may also provide clues about the future. Monitoring how asset classes perform relative to one another, for instance, can provide useful context. As an example, note how U.S. equity (Russell 3000 Index) performance compares with U.S. bonds (Barclays Aggregate Index) on a rolling three-year basis. The trend over the past 20 years is one of alternating between periods of strong equity returns and lagging performance vs. bonds, and vice versa. The message is that when trailing returns are relatively extreme vs. recent history, it may be time to consider an alternative future vs. the one that the previous period just delivered.
Any one predictor is subject to failure, of course. In fact, you can count on it. That inspires forecasts of expected return that are built on a diversified mix of predictors. For instance, the dividend yield alone may not tell us much on a regular basis. But it’s a different story when it flashes a buy or sell signal that’s corroborated with other predictors. Recall that the dividend yield was crawling at post-World War II record lows in 1999 and 2000. That was a warning sign that future equity returns weren’t going to be as rich as recent history implied. The warning was stronger when the low yield accompanied high trailing returns for stocks over bonds.
There are, of course, many more tools for evaluating expected returns. The problem isn’t a lack of useful applications. Rather, the problem is us. In particular, the bigger challenge is discipline. It’s human nature to extrapolate the recent past into the future. That works for a time, but it’s almost always a strategy that eventually crumbles. There’s no guarantee that mean reversion will prevail, but you need more than intuition and guesswork for thinking it won’t.
But let’s not fool ourselves. The solution for thinking strategically and acting tactically on matters of expected returns may get us into trouble if applied haphazardly. Mindlessly buying and holding a broadly diversified portfolio of multiple asset classes can deliver reasonable results for those who truly have a long-term focus. Those investors, however, are the exception. Some form of dynamic asset allocation, as a result, is warranted.
In the end, market volatility offers opportunity and risk. “Your emotional response to the market’s gyrations may be one of your biggest costs as an investor,” notes Morningstar’s Karen Dolan. “The return lost to poor timing can even trump the cost of expense ratios.”
Fair enough. But sitting tight when expected returns rise and fall by more than trivial degrees may not be helpful either. As with most aspects of money management, the goal is finding a practical compromise that navigates between two extremes.