By: Frank Armstrong
By: Frank Armstrong, CFP, AIF
One of the liveliest debates in modern finance revolves around defining the appropriate equity premium. Before you dismiss this as totally an academic concern, let me tell you that you should care. I’ll show you why later. But, first let’s define the terms.
The Equity Premium Defined
The equity premium is the extra return necessary to induce investors to abandon their safe investments in favor of more volatile ones. For instance, we know that since 1872, Treasury Bills have returned about 3%, while stocks have returned about 11%. So the investor has been awarded an equity premium of about 8% for taking on the additional risk.
Many economists and financial advisors (including DirectAdvice) use this calculation to predict future investment returns. At first blush, this equity premium appears reasonably stable over a long period. So, the 8% historical premium has become a default for many planning models.
But, not all economists are comfortable with that estimate. To them, the historical equity premium seems too high. And nobody can build a theory that convincingly explains why it should be so high. The central problem is that if the historical returns are reasonable then stock market returns are higher than corporate profits. If that were true, then every business should liquidate and invest in the market! Clearly, that would be an unstable and unsustainable condition.
There is no quicker way to catch an economist’s attention than to show him a system like this in dis-equilibrium. So, they will spend a lot of time and energy trying to resolve the contradiction.
A forthcoming paper, “The Equity Premium” by Gene Fama, and Ken French argues that rather than use the historical returns, the equity premium should be estimated from the corporate earnings or dividends (The Gordon dividend growth model). This approach makes a lot of sense. After all, there must be a link between corporate profits and stock returns. (You will need Acrobat Reader to view the paper. It can be downloaded from the above link.)
The authors argue that prior to 1950 both methods gave similar results. But since then market returns have been driven by expansion of price/earnings ratios rather than by growth of earnings. So, the two measurements produce sharply different equity premium estimates during the later years. While the historical returns method inexplicably rises during the second period, the Gordon dividend growth model gives consistent estimates during both periods.
If we believe that stock prices are linked to corporate profits, then we must prefer the Gordon model. Instead of a premium of about 8%, the appropriate premium appears to be about 3.5%. This would imply that a better “guestimate” of the future total return from stock investments should be around 6.5%, a very long way indeed from our present planning assumptions.
The implications for investors
This will not come as welcome news to investors who have come to believe that they are almost guaranteed an 11% or higher return. It follows that if we are going to earn less, we must save more in order to reach our objectives, and we may be able to draw down less from our nest egg during retirement.
Now you can see why you should care. If you think there is a good case for the lower estimate, perhaps you will want to make more conservative assumptions in your plans. After all, your own behavior is the only thing that you can directly control, and the key determinant of your financial success. If you save “too much”, the worst that will happen is that you will have much more at retirement. In the bigger scheme of things, this isn’t considered a critical problem.
Should you invest differently, or change your asset allocation? Probably not. No one is saying that stocks will return less than bonds, only that they may not return as much as we have become used to. Even if lower return estimates are right, you will still be getting a handsome premium over bonds.
Who is right?
Which method is right? I don’t know. The discussion certainly isn’t settled yet. Come back in thirty years and I will tell you. Meanwhile, please don’t shoot the messenger. I’m just reporting on the research.