S&P 500 Rate of Return

The long term annual rate of total return (includes reinvested dividends) on an S&P 500 portfolio, from 1871 to today is 9.4%.

Note that there are 3 regression lines on this graph. The dotted line is the long term fit from January 1871 to the end of the third quarter in 2019. This is 9.4%.

One of the dashed lines is the fit from January 1871 to December 1949. This fit shows that the total return before 1950 was 7.1%.

The other dashed line is the fit from January 1950 through October 2019. This fit shows the rate of return after 1950 is 10.7%.

The rate of return over any given 10 year period varies quite a bit, depending on the economy and market condtions.

The following graph shows the Compound Annual Growth Rate (CAGR) over monthly rolling 10 year periods, starting with the period from January 1871 through December 1881, and continuing to the last period which covers the time span of November 2009 through the end of October 2019.

And you can see that the lowest rate of return over 10 years is a -4% CAGR for the period covering September 1929 through August 1939 - the Great Depression.

BTW, note that the 10 year period covering April 1999 through March 2009 is a close second in misery, with a -3.5% CAGR - the Great Recession.

Fortunately, these types of horific market conditions only seem to happen once every 70 years.

On the more positive side, the best 10 year period was from July 1949 through June 1959, with a CAGR of 21% (note: at this rate of return, your porfolio would have grown by 6.8 times over the 10 years!).

The 50s and early 60s were very good to investors, with 10 yr CAGRs averaging 15% or better.

Investors had another good run from 1986 through 2001, also with 10 yr CAGRs averaging 15%.

The period from 2002 through 2013 knocked the stuffing out of returns with a double whammy. First with the dot com crash when the 1995 to 2000 dot com bubble burst, followed by the 2008–2009 Great Recession when the housing bubble burst and credit markets tied to toxic mortgage securities evaporated.

The most recent 10 years has been pretty good though, getting back to the above average 15% rate of return.

You can see by visually examing the graph that 10 year returns are on average higher after 1950 than before.

You may also find it interesting to look at CAGRs over a 20 year period. Here’s that graph:

Here we see the lowest point in September 1949, exactly 20 years after start of the Great Depression in October 1929. But the interesting thing here is that the CAGR is still positive, weighing in at +2%.

In fact, there is no 20 year period from 1871 to today in which an S&P 500 portfolio would have lost money.

The best 20 year period was from May 1980 through April 2000 (the start of the dot com crash), with an eye popping 18% CAGR over the 20 year period. If one had the forsight (or a time machine) to invest in the S&P in May 1980 and then get out at exactly the right time in April 2000 (the peak of the dot com bubble), they would have seen their investment multiply by 26 times!

You can also see the difference in 20 year CAGRs before 1950 and after 1950. Before 1950, total returns over 20 years averaged in the 7% range. After 1950 they averaged in the 11% range.

And finally here’s what 40 year CAGRs look like - the investment horizon of someone who starts investing at 25 in the hopes of retiring at 65:

And now the 7% rate before 1950 and the 11% rate after 1950 becomes even clearer.

Also note that the longer the investment period, the less the uncertainty in the CAGR as calculated from the beginning of the investment to the end.

So for example, for the 40 year time periods after 1950 the 40 yr CAGR varies between 9% and 13%, and spends most of the time between 10% and 12%, a remarkably low level of uncertainty about the final value of your forty year investment.

Comparing that to the 10 year time period the story is quite different. Here we see CAGRs varying over a much wider range - from as low as minus 4% and going as high as 20%, spending most of the time between 5% and 15% - a much higher level of uncertainty.