Monday, April 13, 2009

Best & Worst 20-Year Returns in Stock Market History

20-Year Rolling Returns

This post graphs not only the best and worst 20-years in stock market history, but returns for all 20-year periods for the last 100 years or more; these are so-called rolling 20-year returns.
(Note: This post covers returns through year-end 2008; for the most recent 20-year return, see this post.)

In a recent post we looked at a graph of yearly stock market performance since 1929; it appears to be completely erratic. Yearly returns varied from gains of close to 75% to losses of almost 50%, with no apparent rhyme or reason. We then looked at stock market rolling 10-year returns. The range of returns was dramatically smaller, roughly 0-20%, and much less erratic. What happens if we look at twenty to twenty-five year stock market returns?

Stock Market "Rolling" 20-Year Returns Graph

100 years of stock market (Dow) rolling 20-year returns (since 1900)
Dow 20-Year Rolling Returns
Above is a chart of the 20-year total return of the DJIA (Dow Jones Industrial Average) beginning around 1900. (Click on graph to expand it.) Each point on the graph represents the average annual return earned by an investor who bought the Dow at that year-end and sold 20 years later, reinvesting dividends in the interim. For example, the first point on the graph shows that an investor who bought at year-end 1901, reinvested dividends annually, and sold at year-end 1921 earned 7.1% per year. Note that the returns prior to 1929 are estimated (I have accurate closing prices, but have estimated the dividends based upon another market index).

As you might have expected, the maximum 20-year annual return is somewhat less than the maximum 10-year return -- 18% compared to 19.5%. The difference in minimum returns is somewhat larger; the minimum 20-year return was 2.5% compared to the worst-case 10-year loss of 1.3%.

The Best and Worst 20-Year Returns in Dow Jones History (since 1900)

What were the best and worst 20-year periods to own stocks? Well, if you bought in:
  • 1941: the return was about 15% per year for the next 20 years, or
  • 1979: 18% annual return

The worst years to buy were:
  • 1928: the return was about 2.5% for the next 20 years
  • 1958, 59 & 61: about 5-5.5% annual return

Average Returns, and the Impact of Start/End Years

On average, 20-year returns were the same as 10-year returns -- around 10% per year. But again the start year had a significant impact -- as much as 10% per year or more. It's worth noting that, whether you're looking at a chart of returns over 10-year periods or over 20-year periods, 1928 is at the bottom of the pile. The obvious reason is that those investors had the great misfortune of starting just before the 1929-1932 stock market crash.

Similarly, 10 or 20-year periods ending close to 1999 are near the top. Ten-year returns appeared to cycle between 0% and 20%. Twenty-year returns exhibit a similar, but more muted, cyclicality. However, unlike the 10-year returns, they appear to have a slight upward slope. That might be interesting to look into further at some point since it's somewhat counterintuitive.

A Better Way to Look at Stock Market Returns -- Especially for Retirement Planning

Since the 20-year returns are cycling up and down, it should be clear that you cannot count on receiving the average returns -- even over periods as long as 20 years (say, from age 45 to 65, or 65 to 85).  What if the market returns only 6% for the next 20 years instead of 10%? What are the chances of that? What impact would it have on your retirement portfolio? To get perspective, it's useful to look at What Will a $100,000 Investment be Worth in 20 Years? (showing the variability of the returns in dollars) or the distribution of 20-year market returns.

Finally, it's important to remember that the impact of expenses has not been included, and, in addition, these are nominal returns, not "real" (i.e. inflation-adjusted) returns. In the real world, expenses usually significantly reduce theoretical "market" returns. And, the real world experiences inflation. If you assume 3% inflation, the 2.5% nominal return for 1928 turns into a 0.5% loss in real terms, and the 5-5.5% nominal returns become 2-2.5% returns -- for 20 years. (For a more detailed discussion of nominal vs real rates of return, see this post.)

Twenty years of negative, or barely positive real returns is not what most people expect. It could be a very serious problem, especially if the 20 years start the year you retire....  And, if your expenses subtract another 2% per year (not unusual for a mutual fund investor).... Well, you get the idea.

Note: The above charts are based on DJIA (Dow Jones Industrial Average) data from my Stock Market Analysis Model. Results would be essentially the same if we used S&P 500 data. Dividends prior to 1929 have been estimated based upon another stock market index.

Other Related Posts:

10-Year Rolling Returns: Similar to this post.
20-Year Rolling Returns vs P/E Ratio also graphs p/e ratio at beginning of each 20-year period.
Best & Worst Returns in DOLLARS: 10-100 Years.  To see the range of returns in $$ -- another important perspective.
The 30 Best & Worst Years in Stock Market History: 1-year returns.
35-Year Rolling Returns, and Stock Market Rolling 5 (and 50) Year Returns Graphs
Range of Returns for 1 to 100-Year Holding Periods Best & worst returns for 1, 2, 3, ... 100 years.
Stock Market Compound Growth Calculator: Calculates growth/return rate between any 2 years -- e.g., between 1974 and 1999.
For a list of other popular posts and an index of stock market posts by subject area, see the sidebar to the left.

Share This Article

Bookmark this on Delicious
See the "Share" option in the Menu Bar at the top of the page to share via Facebook, Twitter, etc.
Copyright © 2009. Last modified 11/5/2011


  1. It appears to me you've posted some tremendously valuable data. I've not really encountered such long-term data. And I particularly like your semi-log plots.

  2. Thanks for the compliment. It's good to know that all this hard work is not going to waste. :-)

  3. Good stuff. Although I do want to point out that inflation was relatively low during that 1928 to 1948 period. Almost zero inflation for the first 15 years of that time horizon. From 1961 to 1981 prices basically tripled so in real terms 61 to 81 might have been worse.

  4. Excellent point. I think my original point was just that, assuming normal inflation, all of the "real" 20-year results would more disappointing than presented. But you're absolutely right. If we use actual inflation rather average inflation, '61-'81 is indeed even worse than '28-'48. Adjusting the numbers using actual CPI data, the 2.5% nominal return for '28-'48 becomes 0.7% because of less than average inflation; the 5.3% nominal return for '61-'81 becomes minus 0.6% because of the high inflation (the CPI more than tripled).

  5. agreed what way to view the markets. Would like to know how that data looked in March of 2009 as I would guess the returns after inflation is not as compelling, meaning rear view thinking on this should be measured from all conceivable market scenarios....... IE from 1970-march 2009 would look considerably less interesting than measured from 1970 - today, not to suggest stocks can't go up 5-6% per year from here....

  6. Hellboy,
    If I understand you, you're reiterating my main point -- that even over periods as long as 20 years returns can vary considerably. Since returns are not a constant 10%/year, it's useful to be aware of what the approximate p/e is when you buy in order to reduce your chances of poor performance. (See Projecting Market Returns)

    To do your own analysis, download my Stock Market Analysis Model. That way, you can choose whatever start years and ending years you like.

    Thanks for reading


No spam, please! Comment spam will not be published. See comment guidelines here.
Sorry, but I can no longer accept anonymous comments. They're 99% spam.