The long-term historical data show the relative investment returns of various asset classes. Looking at the major asset classes over a very long period, stocks have performed the best, particularly small cap stocks, followed by bonds and Treasury bills. If you look at more recent periods, e.g. the last 20, 30 or 40 years, you get similar results.
In the long run, notice from the chart below that the value of the U. S. dollar has fallen in real terms (inflation adjusted). Its current value is equivalent to what 5 cents would buy in 1870. An ounce of gold, which is viewed as an inflation and currency hedge, was valued at $18.93 in 1870, which represents purchasing power in today’s dollars of 20 times that, or $379. As of this writing, an ounce of gold is worth about $1,540. Gold has actually grown its purchasing power over the very long run. The dollar has depreciated due to inflation, most of which was created by our Federal Reserve Bank’s policies.
Long Run Asset Class Performance
The tables below show long-term returns of various asset classes. A few observations:
- Small-cap stocks have outperformed large-cap stocks.
- Value stocks (both large and small cap) have outperformed their average stock counterparts.
- Emerging-market stocks have outperformed international stocks, including value and small company sectors.
- Precious-metals stocks were the bottom performer among those depicted. However, rather than conclude they are a bad investment, later we will explore this category in more detail.
- Private equity has had relatively high returns compared with all types of traditional stocks. However, there is debate about the relative risk-adjusted return superiority of this category.
- Gold has provided a reasonably good long term return.
We discuss many of these asset classes and their characteristics elsewhere on the site.
Warning: Past returns may not be indicative of future returns! Investors should weigh many factors, including risk of the given asset class when deciding how much to invest in a given asset. I strongly recommend investors educate themselves and understand that asset classes move in and out of favor over time. It is very difficult if not impossible to use tactical asset allocation to move in and out of the superior performing assets. I explore this in greater detail below and in Selecting Your Assets.
Caution: Your Results Will Vary with Time
A problem with selecting the historically best performing investments is that the performance of a given asset class or sub-class can vary widely. This can be the case over a year, three years, ten years or even 20 years. It is impossible to predict future returns on a consistent basis.
This shows that the best performing asset classes vary from year to year. It also holds true for longer periods. For example, from 2000 to 2009, the S&P 500 lost 24% of its value and delivered a total return (including dividends) of -9%. After the March, 2009 bottom the S&P went up by 17.8% per year for the next 10 years.
From 2001-2012, gold went up 12 straight years. That was followed by a drop of 28% in 2013. From 2014-2018, gold retreated further followed by a turnaround in 2019 when it was up 18%.
The lesson is that no one can reliably forecast asset class returns, or even relative asset class returns on a consistent basis, especially over shorter periods.
Time is on Your Side
The longer you hold an asset, the more likely you will earn positive returns. However, most investors don’t hold for the long run. They buy and sell at the wrong times. Too much activity is the enemy. The best investors choose solid investments and hold them a long time. The old saying is “set it and forget it.”
What does the Future Hold for Returns?
“It’s tough to make predictions, especially about the future.”
– Yogi Berra
In order to construct a portfolio it is useful to have a gauge of expected future returns of various asset classes. Naturally, there is much uncertainty in predicting returns. A good starting point is to look at historical returns over very long periods, noting that there can still be substantial variation in returns over many decades. For your convenience below we repeat the historical returns cited above, along with forecasts from select professional sources.
The majority of forecasters predict significantly lower returns for U. S. stocks and bonds in the next decade. Some reasons:
- High U.S. stock valuations by historical standards (more on this below).
- Slower expected long-term U.S. economic growth due to high and rising debt, an aging population, and stalled productivity growth.
- Historically low interest rates and higher economic and financial market uncertainty. As of this writing more than $16 trillion of worldwide debt has a negative yield.
Here is a good article citing other opinions. Many foreign markets also face potential headwinds from large debt burdens, unemployment, aging populations and economic structural problems. My opinion is that it is prudent to plan for more conservative asset class returns over the coming 5-10 years. This makes it imperative to build a diversified portfolio with exposure to select asset classes. However, the warning above from Yogi Berra applies.
What Can We Expect from US Stocks?
Since most investors have a large stake in U.S. stocks and it’s the largest stock market in the world, let’s take a look at what kind of returns might be expected in the future.
You’ll see that when stocks are cheap as measured by the P/E ratio, subsequent long term returns are higher than average. When P/E ratios are high, subsequent returns are low. As of mid-February, 2020 the trailing, as-reported P/E ratio of the S&P 500 was 26, compared to the long run average of about 16. Note that the majority of financial pundits and writers today report the P/E ratio based on forecasts of operating earnings for the coming 12 months. This consistently depresses the true P/E ratio of the market. For example, as of this writing, the S&P 500 P/E based on forecasted operating earnings is 19, making the market look reasonably valued. I don’t buy into this gauge. At 26 times as-reported trailing earnings, the P/E is historically very high.
When the P/E was much higher than average, for example at 29 in April of 2000, stocks tended to perform below average in the future. From year end 2001-2011, the S&P 500 returned only 1.5%/yr. When the P/E was well below average, for example in 1931 and 1981, stocks performed well in subsequent periods. However, from 1955-1975 and during the 1990’s, stocks performed well even though the P/E was above its long term average.
The chart above, courtesy of dshort and Advisor Perspectives shows the P/E10 ratio. The P/E10 looks at the S&P 500 index relative to its earnings over the previous ten years. The chart shows that the P/E10 dropped to a cycle low of 13.3 in 2009. This was followed by excellent returns of approximately 18% per year for the following ten years.
As late August, 2019 the P/E10 stands at 30. This is 50% above the regression path of a slowly rising long-term P/E10 and nearly double the long term average of 16.9. The current level is also well into the highest quintile of historical P/E10 ratios ranging from 21.2 – 44.2. This clearly signals the likelihood of lower returns ahead.
A similar and more well known gauge is the Shiller P/E ratio. The Shiller P/E also looks at stock prices vs. their average earnings over the previous ten years. While it uses a different methodology than the P/E10, the logic and concept is the same.
The table below shows the subsequent 10 year returns of the S&P 500 index based on different starting Shiller P/E ratios. The current level of 28.95 (orange row in table below) has historically produced average real returns over the following ten years of 0.9%/yr, with a range of -6.1% to +5.8%. Note that these returns exclude dividends. The current dividend yield on the S&P 500 is 1.9% giving a total real return of 2.8%/yr.
Based on the Shiller P/E, the mean expected annual return for U.S. equities over the next ten years is only 3.1% according to the excellent article, “The Sad, Stark Truth About Future Returns.”
Some make arguments that this is a flawed measure. But Research Affiliates wrote an outstanding article about the CAPE ratio, supporting its usefulness. The Shiller P/E is one tool that can be used, albeit with caution.
Bottom line, no forecasting tool or method should be relied upon to make major bets on the stock market’s future direction. All forecasts of future asset returns should be looked at skeptically. The historical and forecast returns can be used as an imperfect, yet “directional” aid to investment decisions.
Here is another very interesting article that looks at historical P/E ratios. It also suggests stocks have very high valuations now. Take a look at the chart, courtesy of Advisor Perspectives:
Other Measures Show a Richly Priced Market
There are two other metrics that show the U.S. market is priced high by historical standards. The Q Ratio measures the total price of the market divided by the replacement cost of all its companies. As of July, 2019, the ratio was 1.32, compared to a long term mean of 0.70. This is 87% above its arithmetic mean. The previous peak was in 1999, before the Y2K market bust and the “lost decade” of 2000-2009.
The other metric is the market cap to GDP ratio, popularized by Warren Buffet. Buffet said “it is probably the best single measure of where valuations stand at any given moment.” As July, 2019, the ratio stood at 146%. This compares to the peak value of 161% in 1999 and is well above its long term mean.
The conclusion we may draw from the P/E, P/E10, Shiller P/E, Q Ratio and Buffet Valuation metrics is that caution is in order.