The other day I was thinking about my investing life. The end of this year marks the 20th year that I’ve been seriously investing in stocks. I can honestly say that it felt like an incredibly turbulent two decades, both at the time and now looking at it in retrospect. This graph from JP Morgan paints the picture of my investing woe (click to enlarge):
Less than a year after I started investing, the tech bubble burst and the market skidded for two years, ending 49% lower. It took five long years to claw my way back to where I started, and then the global financial crisis exploded and drove the market back down by 57% to a new low. In the subsequent recovery, stocks surged for 10 years to multiple new highs, with a few more bumps along the way.
The Rule or Exception?
How unusual has my investing life been? On the one hand, everyone probably thinks their investing life has been a crazy roller coaster because it’s hard to find a truly calm 20 years anywhere in the stock record. On the other hand, two crashes of 50% in just 20 years feels like an extra whack outlier to me. In Mindfully Investing fashion, I dug into the data to see which hand holds the truth.
First I calculated the rolling 19-year nominal annualized returns for the entire history of the S&P 500 going back to 1871. I wanted to compare that history to my own investing life as represented by the last 19-year period from the start of 2000 to the end of 2018.* Here’s the graph.
There was only one 19-year period (ending 1947) when the nominal annualized returns were clearly worse than my investing life, although some other periods were nearly as bad. I remember my grandfather talking about the stock doldrums of the 1960s, which bled into the stagflation of the 1970s. And yet that historically crappy episode still garnered better returns than my investing life. Maybe bad luck runs in my family?
It’s remarkable that the market climb of 307% from the 2008 lows over the last 10 years hasn’t completely repaired the damage of the 2000 and 2008 crashes. Long-term investors are only just now recovering from the hangover of the “lost decade” that started this century. Here’s a graph showing the progression of nominal annualized returns for each successive year of my investing life using S&P 500 data.
My annualized return was exactly 0% from 2000 to 2010. A “lost decade” indeed. And 10 years later my long-term annualized returns are still only a meager 4.8%.
What Will Happen Next?
If the stock market keeps going up, my annualized returns will continue to improve. But here’s the problem: almost no one is expecting above-average stock returns for the next decade. I regularly review and update long-term return predictions. The consensus expectation for the next 10 to 15 years is for nominal annualized returns in the 4 to 6% range, which is considerably less than the annualized return of 9% for the entire history of the U.S. stock market.
The margin of error around these predictions is very high. For example, StarCapital is currently predicting a median 10-year nominal annualized return for U.S. stocks of 4.7% based mostly on current valuations, but the entire range for their prediction is 0% to 12%**. In the face of this uncertainty, the best we can do is use a 5% nominal return estimate for the next decade, which is the central tendency across all the predictions that I recently compiled here.
Using this highly uncertain 5% return prediction, what will happen to my annualized returns over the next decade or so? Here’s a graph of 30-year rolling annualized returns up through 2029, assuming that the market consistently returns 5% nominal each year for the next 11 years (from 2019 to 2029). The blue line indicates historical data and the purple dotted line shows where the 5% assumption for future returns is included in the calculation.
Even though I’m not using a “worst case” future return estimate, I get the worst 30-year annualized return (4.9%) in history; worse even than investing through the crash of 1929 and the Great Depression. Using the same assumptions, my successive annualized returns will progress as shown in this chart. (Again, the purple dotted line indicates calculations that include the future 5% annual return assumption.)
Just to make my point abundantly clear, here’s another way to look at it by plotting the cumulative distributions of 19- and 30-year annualized nominal returns of the S&P 500 from 1871 to 2018.
My 19-year annualized return is in the lower 4th percentile range. In other words, 96% of all 19-year historical returns were better than my returns. And if a 5% annualized return turns out to be true for the next 11 years, my 30-year returns will be lower than any time in history. My miserable investing life could set a new record for suckiness!
Valuation Versus Future Returns
If the all-time annualized return for the U.S. stock market is about 9%, and we’ve already had at least 20 years at nearly half that level, why is everyone predicting that future 30-year annualized returns will turn out even worse?
All these return predictions rely on current stock valuations to some extent. The Shiller CAPE (Cyclically Adjusted Price Earnings) ratio is the most fashionable measure of stock valuations right now. The CAPE Ratio is the current stock price divided by the average of the last 10 years of earnings adjusted for inflation. Here’s a graph of the CAPE Ratio for the entire history of the S&P 500.
The median CAPE Ratio over this period was a remarkably low 15.7. The CAPE Ratio dipped briefly below this median only once in my investing life, after the crash of 2008. And the CAPE Ratio is currently higher than at any other time except preceding the 1929 and 2000 crashes.
People use valuations to predict future returns because of the historical correlation between high valuations and subsequent low returns. Here’s an example graph of CAPE Ratios versus returns in the subsequent 10 to 15 years for 17 countries, again from StarCapital. (Click to enlarge.)
The relationship between CAPE Ratio and future returns is pretty evident, but the scatter in the data is huge.
This relationship suggests that unless returns trend lower in the future, valuations are unlikely to drift down toward the historical average. We can see this by estimating potential future changes in the Simple PE (price earnings) Ratio, which is the current price divided by the current earnings per share. I examined how the Simple PE of the S&P 500 would change from its current value of 22 when I applied different future annual returns and earnings per share (EPS) growth rates. Here’s a table of those calculation results.
| S&P500 Annualized Return | EPS Growth | Simple PE Ratio | |
| Historical Data (1871-2018) | 9% | 6% (median) | 15 (median) |
| Future Scenario 1 (2019-2029) | 5% | 6% | 14 (in 2029) |
| Future Scenario 2 (2019-2029) | 9% | 6% | 22 (in 2029) |
Future Scenario 1 uses the same 5% expected annualized return discussed above and an annual EPS growth equal to the historical median of 6% for the S&P 500. These assumptions push the Simple PE in 11 years (2029) down to 14, which is below the historical median of 15. But Future Scenario 2 uses annualized returns close to the historical value of 9%, which causes the PE to levitate well above the historical median for 11 more years.
Who Should We Believe?
Scenarios 1 and 2 are highly plausible outcomes among many possibilities. And the outcomes of both scenarios are unprecedented in non-trivial ways:
- If current high valuations revert downward toward historical averages, then future 30-year returns will be at unprecedented low levels.
- If future 30-year returns revert upward toward historical averages, then valuations will have maintained unprecedented high levels for three decades running.
In other words, something unprecedented is likely to happen regardless of whether future annualized returns and valuations get better or worse.
All the researchers predicting future returns are essentially assuming that Scenario 1 will occur and 30-year returns will sink to historically low levels, whether they realize it or not. A few folks like Christopher Burnham are considering the possibility of Scenario 2 where valuations continue to hover well above historical levels for the long term. Hypothesizers of Scenario 2 point out that relatively high valuations may be reasonable for the 21st century considering things like:
- Corruption and market manipulation were rampant in the stock market prior to 1900, making comparisons to those data unreliable.
- In the early 1900s, only 1% of the U.S. population owned stocks, now it’s more like 50%.
- The U.S. economy in the first half of the 20th century was entirely different than today. For example, back then about one-third of people were farmers and now it’s less than 2%.
- The Federal Reserve didn’t exist or functioned very differently for at least half of U.S. stock market history.
You could probably come up with several other relevant contrasts between newer and older market/economic conditions. It seems that most predictions of low future returns are married to the history of valuations and divorced, or at least separated from, the history of annualized returns. Could all these predictions of low expected returns be misguided?
This Time It’s Different
This brings me to the title of my post. A brief search of the internet surfaces dozens of articles about Sir John Marks Templeton’s observation that:
- “The four most expensive words in the English language are, ‘This time it’s different.'”.
This influential stock investor of the 20th century was referring to our human bias toward ignoring evidence of an over-valued stock market. Because the stock market has been hitting all-time highs, journalists and bloggers have recently been getting a lot of mileage out of Templeton’s four most expensive words. The conclusion in most of these articles is simple: given that current stock valuations are historically high, a major market slide is likely, and those who say “this time it’s different” are fools. In other words, investors should beware of Scenario 1.
On the other hand, we all know that history doesn’t repeat itself exactly. There are always new variables and combinations of circumstances that make this time unique from all other times in the past. One way or another, isn’t this time always different? It’s interesting that warnings about “this time it’s different” mostly focus on the valuation side of the question. Shouting the folly of ignoring current high valuations and the potential for Scenario 1 fails to voice all the other ways that this scenario would still be different this time. An unprecedented low 30-year annualized return is one clear difference that nobody seems to be talking about.
I Know The Future
Instead of “This time it’s different”, my vote for the most expensive four words in investing are: “I know the future”. The thing is, no one ever comes right out and says those particular four words because it sounds pretty stupid. But that sentiment is buried in all sorts of educated statements and research about the future of the stock market, although it may be hidden behind complex calculations and rigorous slicing and dicing of historical data. There’s often hand waving about uncertainties and probabilities, but most stock market predictions imply that the investor should be concerned about the future. And in every case, at their core, those predictions rely on historical patterns that may never be repeated. Predictions of low future return scenarios are even more problematic because they rely on one historical pattern (valuations) and mostly ignore that another historical pattern (30-year annualized returns) will be broken.
Conclusion
As you might be able to tell from my writing on this topic, I’m not particularly worried about predictions of unprecedented low future returns. I certainly think something like Scenario 1 is possible. But I know it’s just an educated guess. Look again at that graph by StarCapital comparing CAPE Ratios to returns for the subsequent 10 to 15 years. Under similar past circumstances of high U.S. stock valuations, the subsequent returns exceeded 10% four different times. There’s no ironclad reason why that can’t happen again.
But I’m not planning for high future annualized returns. As I’ve detailed in other posts, for my own investing and retirement plan, I’ve conducted evaluations using the 5% return estimate as well as various historical sequences of return including the good, the bad, and the ugly. So, if some variation of Scenario 1 plays out, I still know that I didn’t make a big mistake by retiring early. It’s mindful to plan for the worst, and hope for the best.
* Some readers may wonder whether the record of the S&P 500 has any bearing on my own personal rate of return over the last 19 years. I have done some rough calculations in the past, which show my return was similar to the S&P 500. However, my personal annualized return would not be very applicable to anyone else, and my early investing life would not be illustrative of consistently mindful investing methods. So, I’m using the S&P 500 as a generally relevant example to determine how relatively hard or easy the last 19 years of stock investing has been for the mindful investor.
** StarCapital reports these as real returns, which I converted to nominal returns using their stated inflation assumption.
