In my last post, I showed that U.S. stocks (the S&P 500) have offered positive inflation-adjusted returns 87% of the time and positive nominal returns 94% of the time for investors with a 10-year timeframe.  (By “positive return”, I mean that the investor made a profit, while a “negative return” means losing some money over the specified timeframe.)  And even over a one-year timeframe, U.S. stocks have provided positive returns 68% to 73% of the time, inflation-adjusted and nominal respectively.  Despite these facts, stocks have a “scary” reputation for being crash-prone and highly risky as compared to “safe” government bonds.

Data Limitations

However, a regular reader by the name of Mathew made this insightful comment about these statistics:

• This is interesting although it’s just for the S&P 500 in a period where the US had a great 100 years.

He went on to note that, for example, the prices of Japan’s stock market (as represented by the Nikkei 225) have gone essentially nowhere since 1990, as shown in this chart from Macrotrends.

 

Mathew makes a good point.  At the absolute most, there are only 150 years of reliable U.S. stock market data, and probably only the last 100 years are relevant to our modern economy and markets.  I’ve used Mathew’s same point myself to criticize factor researchers for relying too much on this rather limited dataset to find all sorts of weird anomalies that have outperformed the broader stock market in the past but aren’t guaranteed to outperform in the future.

So, am I making the same sort of mistake when I calculate positive return statistics based on just 100 years of U.S. stock data?  Will the next 100 years of U.S. stock market returns mimic, or even resemble, the last 100 years?  Perhaps the future of U.S. stocks will be scarier than its past.  And perhaps, stocks from other countries have been and will be scarier than U.S. stocks.

Beyond the United States

As Mathew’s observation suggests, comparing historical stock returns across many countries might provide:

1. A measure of how exceptional or unusual the last 100 years of U.S. stock returns have been.
2. A better sense of the risks involved with stock investing in general, as indicated by many different geographic and economic conditions.

Further, I’ve argued that it’s mindful to consider geographic diversification of stock portfolios.  For such geographically diversified investors, an analysis confined solely to U.S. stocks could be misleading.

Consistent with the idea of maximizing datasets, I searched for the longest annual return records for different countries and found that the MSCI dataset for Developed Market countries goes back to 1970 for 17 countries and nearly that far for 5 other countries.  The MSCI Developed Market dataset spans only about half the available period for the U.S. dataset.  But the 50 years from 1970 to 2020 include some pretty turbulent times such as the stagflation and the global oil crisis of the 1970s, the rise/fall of the dot-com bubble in the 1990s and early 2000s, and the global financial crisis of late the 2000s.

Using these data, I conducted a similar analysis to my last post, where I calculated the percent positive returns for various investor timeframes.  Specifically, I calculated the percent positive returns since 1970 for the 22 MSCI Developed Market countries on an annual, rolling 5-year, and rolling 10-year basis.  And as an overall comparison, I included the MSCI Developed Market Index that covers all these countries.  This index is tracked by many low-cost mutual and exchange-traded funds¹.

Also, all these returns are calculated in U.S. dollar terms.  So, this analysis reflects someone in the U.S. investing in index funds that include stocks from one or more of these foreign countries.  For this same reason, inflation-adjustments are calculated using U.S. inflation data.

How Scary Are Global Stocks?

In my last post on U.S. stocks, I showed that the shorter the investing timeframe, the greater the chances of losing money.  I found this same relationship for other developed market stocks as shown in these two summary tables.

Nominal Basis

Country	Annualized Return (CAGR²) 1970-2020	Annual Positive Rate	Rolling 5-Year Positive Rate	Rolling 10-Year Positive Rate
U.S. S&P 500	10.66%	80.39%	87.23%	95.24%
Median for 22 Dev. Market Countries	9.26%	66.67%	83.70%	95.18%

Inflation-Adjusted (Real) Basis

Country	Annualized Return (CAGR²) 1970-2020	Annual Positive Rate	Rolling 5-Year Positive Rate	Rolling 10-Year Positive Rate
U.S. S&P 500	6.55%	72.55%	70.21%	85.71%
Median for 22 Dev. Market Countries	5.32%	64.17%	72.83%	86.76%

Further, the positive rate for the annual and 5-year rolling timeframes for the median of non-U.S. stocks is lower than for U.S. stocks, indicating that non-U.S. stocks have a somewhat higher chance of generating losses.  But interestingly, on both a nominal and inflation-adjusted basis, the 10-Year rolling positive rates for U.S. and non-U.S. stocks are nearly identical, indicating similar chances of losses.

We can better see the relative standing of U.S. stocks among its global peers in these three graphs for annual, rolling 5-year, and rolling 10-year positive rates.  The U.S. is shown by the diagonally hatched bars, and the aggregate MSCI Developed Markets Index is shown by the horizontal hatched bars.

Again, we see that on an annual timeframe, the U.S. is near the head of the pack of developed market countries in terms of the inflation-adjusted percent positive rate.  But on a 5-year timeframe, the U.S. is actually in the back half of the pack.  And on a 10-year basis, the U.S. is right around the middle of the pack.

Given that mindful investors care more about longer timeframes (5 to 10 years a least), these percent positive rates over the last 50 years suggest that:

1. U.S. stocks are not particularly exceptional performers relative to the rest of the developed world.
2. The risks involved with stock investing in general, as indicated by many different geographic and economic conditions, are pretty similar to the risks of U.S. stock investing.

Putting the second point another way, my conclusion about the “scariness” of stock investing doesn’t change much from my last post.  Specifically, over the last 50 years, investing in stocks for 10 years almost anywhere in the developed world offered chances of a positive return that were usually higher than 75%, on both a nominal and inflation-adjusted basis.³  The few exceptions were Ireland, Italy, Japan, Portugal, and Austria.  Further, the MSCI Developed Market Index posted an inflation-adjusted positive rate for the 10-year timeframe of about 90%, which was higher than the analogous positive rate for U.S. stocks at about 86%.

Annualized Return Is Still The Bottom Line

A high chance of a positive return is nice, but it’s not much consolation if the annualized return is anemic.  The two tables above give a few summary statistics for the annualized returns for U.S. and non-U.S. stocks.  But I thought it would be helpful to examine the relationship between the positive rate and annualized returns across these countries.  Again consistent with a mindful focus on the long term, these two graphs plot the 10-year rolling positive rate against the annualized return for the entire 50-year period on both a nominal and inflation-adjusted basis.

We can see that higher annualized returns usually mean higher rates of positive outcomes, which makes sense at an intuitive level.  We can also see that the U.S., while a good performer, is not exceptional in terms of either long-term annualized return or chances of having a positive outcome.  The truly exceptional performers for the last 50 years have been Hong Kong, Denmark, Sweden, the Netherlands, and Switzerland.

Conclusions

Mathew’s original caution is correct.  Given these relatively short data histories, we still can’t say that past stock return results are necessarily a good predictor of future results.  Nonetheless, past returns suggest that long-term investing in global developed markets isn’t any scarier than focusing solely on U.S. stocks.  And if the last 50 years of history is any guide at all, the chances of losing long-term money in a globally diversified stock portfolio seem pretty low.

However, the stories of some countries, particularly Ireland, Italy, Japan, Portugal, and Austria, suggest that a concentrated bet on one or two developed market countries could open the door to higher risks.  This is yet another example of how diversification can potentially reduce investing risks.

Getting back to the Japan question, these calculations show that Japan represents one of the three worst scenarios for a 10-year investment in developed market stocks over the last 50 years.  Although an investor in Japanese stocks since 1970 netted an annualized return of 9.21% nominal or 5.16% inflation-adjusted, that’s mainly because Japanese stocks did very well from 1970 to 1990.

If we focus a moment on the worst period in Japan from 1990 to 2010, the nominal annualized return over those 20 years was a pitiful -1.35% nominal or -3.89% inflation-adjusted.  That’s a pretty bleak outcome for even the most mindful stock investor.

This is a good reminder that when we’re presented with statistics like a “90% chance of a positive return”, it still means there’s a 10% chance of losing money.  And some of those negative return periods could be particularly deep and/or long such as the Japan example.  But a global review of developed markets suggests that these really bad return scenarios are the exception rather than the rule when it comes to stock investing.

---

1 – The MSCI Developed Market data are different from, but comparable to, the developed market data I’ve used in past posts, which have mostly come from Portfolio Visualizer datasets.  I made this switch because the MSCI datasets extend back further in time.  So, if you compare the results in this post to my past posts on developed markets you may see some differences.  After working with both datasets, they seem pretty similar where they overlap in time, but they’re definitely not identical.

2 – Annualized return is also known as Compound Annual Growth Rate or CAGR.

3 – You may have noticed that the percent positive rates for the U.S. in today’s post are a little different than those presented in my last post.  That’s mainly because the U.S. data used in my last post extended back to 1928, and in this post, I’m using U.S. data back to 1970 to be consistent with the period available for the other developed countries.