They say that no one has a crystal ball regarding the stock market. While this is certainly true in the short run, we can actually do a decent job on the long run.
Many people have asked me if I thought stocks are too expensive to invest in today. After U.S. stocks rose by over 30% last year, and with many people in the financial media calling stocks overvalued, I understand the sentiment.
However, in my view, such questions slightly miss the mark. Higher valuation of stocks just means lower future returns. The real question is, given today's valuation, how much will stocks return in the long run? If this number is sufficiently high, then we still ought to invest our dollars into stocks today. If this number is too low, we should look at alternatives, such as long term bonds.
This is an age old question. As such, many authors have written about the factors that influence long term stock returns. For instance, the Canadian Couch Potato wrote this piece on the subject.
However, I have yet to come across someone who answered the crucial question: why? Why do these factors explain future stock returns? In this article, I will answer the question, and we'll see that by answering, we can actually use it to roughly predict long term stock returns.
Predicting Bond Returns
In order to understand stock returns, it's worthwhile to understand bond returns first.
Compared to stocks, predicting future bond returns are much simpler. That's because with a bond, the issuer has promised a set schedule of payments.
For example, an issuer may have promised to pay $20 every half year (called coupons) until maturity in 5 years time. At maturity, the issuer will pay back a principal of $1,000. You will receive this much - no more and no less - unless the issuer defaults.
Since you know the market price of the bond, and since you know how much you'll likely receive, calculating the expected return on your bond merely involves putting these numbers into a financial calculator. In fact, many information sources even perform this last step for you, so you just have to look up "Yield to Maturity".
Applying The Same Logic For Stocks
In principle, estimating stock market returns work the same way - figure out how much you receive from stocks every year, and punch them into a financial calculator. However, with stocks, we have two main differences that make the calculations more difficult.
First, instead of coupons, we should punch in earnings as recurring cash flow.
For example, let's say that a stock is priced at $100 today, and we think it will earn $10 a year forever. Then, we can estimate that the stock will return roughly $10/$100 = 10% per year. Notice that this is the Earnings to Price ratio, which is the inverse of the P/E ratio. This will be useful later.
Now, some people might say that we should punch in dividends as opposed to earnings in this calculation, because stockholders only get to see the dividends pour into their accounts. However, I believe that's a mistake because A. corporations can also return money to shareholders using share buy backs, and B. because earnings that corporations retain don't disappear. Instead, good corporations will use retained earnings to increase dividends and buybacks in the future.
Unfortunately, punching in earnings into our financial calculator poses problems because unlike coupons, earnings are not predictable. One year, a corporation may have a good year and earn $30 a share. The next year, it may have a bad year and return $10 a share. The future is especially unpredictable for earnings many years into the future.
Yet unlike bonds, stocks have no maturities, which makes this the second big difference between stocks and bonds. Since stocks have no maturities, it forces us to at least guess what earnings far into the future looks like. Otherwise, we can't arrive at an estimated return for stocks. This is why picking individual stocks is so challenging.
Forecasting The Stock Market As A Whole
However, when it comes to analyzing the stock market as a whole, we benefit from aggregating a large number of stocks. Some stocks will grow their earnings quickly, while others' will decline. On the whole, they balance each other out. There are also compelling reasons why aggregate earnings growth will roughly track economic growth as well.
Therefore, for the stock market as a whole, we can look at past years' earnings data and use it as a rough guide for the future. Let's see what this means in practice.
Let's say that the stock market has a P/E ratio of 20. As we have seen, if we assume that the earnings will stay the same, we can take the inverse of the ratio and use that as a way to estimate returns. In this case, we'll arrive at an expected return of 5% per year. This is why according to many studies, P/E ratio acts as a predictor of long term stock market returns.
At the time of this writing, the P/E for the S&P 500 stood at 17, which implies that if earnings stay the same, we can expect a return of roughly 6% per year.
However, there's a big problem with the P/E ratio, and it's that earnings never stay constant. Over the past few years, companies have benefited from a growing economy, which enabled companies to generate higher earnings. When a recession hits, which will eventually happen either in the near term or in a few years, earnings will go down, which will increase the P/E ratio, and decrease our expected return on stocks.
In other words, since the P/E ratio looks back just one year, and since most business cycles last for more than that, the 'E' of the P/E ratio is unreliable. This gives us wild variations in expected long term returns.
To overcome this problem, it makes sense to use more of a long term average for earnings, and that's where CAPE comes in.
CAPE stands for Cyclically Adjusted Price-to-Earnings ratio. Basically, it takes the past 10 years of earnings history, and adjusts the older earnings up to account for inflation, and then takes the average of those earnings. CAPE is the price of today's stocks divided by this average earnings.
As at the time of this writing, CAPE stood at roughly 25. Inverting this ratio implies that stocks may return 4% per year going forward, IF earnings stay the same.
However, as I've mentioned in the past, corporate earnings will increase at a similar rate to the economy in the very long run. Therefore, we should also add the projected nominal rate of GDP growth to arrive at our prediction. If we think the economy will grow 2%, then our current projection becomes 4% + 2% = 6%.
Where Our Method Falls Short
That said, even using CAPE won't give us a bullet proof prediction for the future.
For one thing, like bonds, stocks are subject to interest rate risk. As interest rates rise, investors demand higher returns on every financial asset, so the same mechanism that depress bond prices (explained here) affects stocks as well. So even if earnings stay the same or even increase, rising interest rates can depress stock prices. This happened in the 70s.
But stocks are also subject to the violent mood swings of investors. Both irrational exuberance and irrational fear can last for years, and such stock movements caused by emotions can affect even 10 year returns.
So how do these shortcomings affect our view of the stock market today?
While we can't predict the emotional state of investors, we can make educated guesses regarding interest rates. Since interest rates are at historical lows and since people expect them to increase in a few years, we may see less than 6% per year returns for stocks over the next decade.
Recently, Ray Dalio has come out and said that he expects to see 4% returns per year on stocks over the next decade. It sounds about right to me.