Chart: Investment Returns Vary, Even Over The Long Run

Updated and revised 2013. You often hear that stock investing is a sure thing over the “long run”. But as this chart from the NY Times and Crestmont Research shows, there is still a lot of variability involved. The matrix below visually displays the annualized returns for the S&P 500 for every starting and ending year from 1920 to 2010, adjusted for inflation, taxes, and transaction costs.


(click to enlarge)

Your actual returns depend a lot upon when you start, and also when you finally withdraw:

After accounting for dividends, inflation, taxes and fees, $10,000 invested at the end of 1961 would have shrunk to $6,600 by 1981. From the end of 1979 to 1999, $10,000 would have grown to $48,000.

“Market returns are more volatile than most people realize,” Mr. Easterling said, “even over periods as long as 20 years.”

Some further observations:

  • As your holding period lengthens, the returns converge towards the median of about 4% above inflation, estimated taxes, and estimated fees. Anything higher than that is very rare.
  • The “long run” may be a lot longer than most people think. It can take 40+ years to get to that 4% real return, not just 15 or 20. Now, if you’re in your 20s or 30s, you probably will have a holding time of 40+ years for the money you’re investing now. But that is still a very long time.
  • Visually, investing from about 2000 onwards looks at least so far somewhat like investing from about 1970 onwards. (Both pockets of red in the early years.) Not the most exciting prospects. :) However, even if you start out strong, over the long run the returns also drift back towards the long-term median. It’s the money that you invest right before retirement that can be the most at risk, which is why it is often recommended to increase your allocation to bonds as you age. If stocks are doing well and you are nearing retirement, much of your new investments will be into bonds.
  • For the charts that include the effect of transaction costs (expense ratios, commissions, bid/ask spreads), these are significant. From 1900-1975, these are assumed to be 2.0%. From 1976-1990 they are assumed to be 1.5% and from 2000 onward they are assumed to be 1%. With the help of low-cost index funds and ETFs, the average investor can do much better than that these days.

Crestmont Research has helpfully updated their stock matrix to include return data through the end of 2012, but in my opinion the charts are not as easy to use as the NYT version above so I left that one up as the overall lesson remains the same. The updated versions do offer the ability to compare index-only returns (no dividends), total returns (including dividends), inflation-adjusted returns, and/or tax-adjusted returns.

Comments

  1. THeir chargs are neat. But I question how they figure that people lost money from 61 to 81. In 1961 the S&P 500 was trading at around 60 and by 81 it had just about doubled.

  2. According to that table, no long term investment over 20 years was good. It also implies no investment withdrawn in 2000-2010 was good.

  3. Check out inflation during the 60′s and it isn’t pretty… Not a good time to retire it seems.

  4. Of course nobody owns (or at least nobody should own) as 100% of their portfolio the S&P 500 for 40 years.

    It would be really neat to see this done for say a 60/40 S&P/total bond market portfolio rebalanced annually over similar time periods.

  5. While interesting, I don’t feel this chart helps a whole lot for a long-term investor. I don’t want to know what $x invested in a lump at the beginning of a 20 year period is worth, I want to know what $x invested every year over 20 years is worth. That’s a much more “real world” scenario when it comes to long-term investing.

  6. I find this chart to be very mis-representative. First the colors he chose are very misleading. Intuitively I expect gray to be neutral; however, on this chart the gray area actually represents +3% return after inflation. I assume light red to be slightly negative; but again the scale on this chart assigns light red to be 0 to 3% return adjusted for inflation. Very misleading.

  7. I would like to see a similar chart for the aggregate bond index to be able to compare the volatility of the S&P 500 to AGG.

  8. Who’s going to put money into the S&P 500 in 1961 and keep it there 40 years? Nobody. That’s not reality. The reality is the stock market is a casino not an investment.

  9. Not surprising

  10. I agree with ttfitz. I don’t understand why they make a graph on lump-sum, but not dollar-cost-averaging. Unless someone inherits a big chunk of money, we always save little by little every year.

    I suspect the result from the latter will be less “news worthy”.

  11. Agree with ttfits…. This chart is only applicable for someone who invests 100% of their nest egg on one day and withdraws 100% of the nest egg on some date X years later… in other words, nobody!

    Also note that the author subtracted “average taxes and fees” whatever that means. Most of my nest egg is in 401k and IRA’s which are tax advantaged and hopefully will be taxed in a low bracket when withdrawn.

  12. follow the vertical line from where 1960 intersects 1980 on down to 1980 along the left hand side

    that is 20 years of investing, with virtually each and every investment in the S&P 500 getting beaten by inflation (real return <0).

    It is is an invalid argument that this table only looks at lump sums – it shows lots of different things…

  13. Enonymous gets it; this chart is not just about lump sum investing. If you want to see how well yearly investments did over x amount of time, look at a vertical slice in the chart. The take away is that even with a long term perspective of 20 years, there are periods of time that will not return the average rate of return that many people assume in their retirement projections.

    The question is what will you do in order to ensure a higher probability that you will meet your retirement/financial goals?

  14. enoynmous – I agree, the table doesn’t just show lump sums if you work at it, but that’s all it shows for each data point. And yes, the example you give does show a bad 20 year period for regular investment, I wouldn’t expect otherwise given the 20 year lump-sum result. But what this table doesn’t show well is how each 20 year period (each of those lines like you pointed out) of periodic investment worked out. Just eyeballing it, like you did with the 1960-1980 period, it appears to me most turn out with decent returns, but I can’t tell for sure.

    Which is what my problem with the table is – I have to guess about what is the more common form of long-term investment.

  15. No, the table doesn’t show the effects of dollar-cost averaging or rebalancing. I agree, that would be cool as well. But hey, the data is out there and if you make it into a pretty graphic, you might get written up in the NY Times (and here) too. :)

    If you take a diagonal line, you can see what rolling 20 year time horizons look like, versus rolling 10-year returns, versus rolling 40-year returns… as well.

  16. What I see when I look at this chart is no matter when you invested, if you cashed out for Y2K you made money! Matter of fact, the only years that seem to display this (making money no matter when you invested) is roughly from 1998 – 2000.

    The funny thing is that at that time (1998-2000) I remember thinking it was a GREAT time to invest. But what I got from this chart is that it was a GREAT time to cash out.

    Hey, I learned something…thank you.

  17. Something does not make sense to me. If you go back to SOURCE data (http://www.crestmontresearch.c.....215;11.pdf ). It says dividends are NOT included in the chart. That point needs to cleared up before acting on this chart. NYT may be churning some very wrong data since other reports show dividends are a bit part of return.

  18. I would like to see a similar chart for bonds. Given that this is after taxes and inflation I’d expect to see a lot of red.

  19. Chart is very misleading… It actually shows that there are historically only 2 years where if you had invested, you would NOT have beaten inflation after holding that investment for 20 years. The colors don’t lead someone to believe that.

    It’s actually a fascinating chart as you can see lump sum AND periodic investments. If you look at any vertical line it gives you an idea how periodic investments would have fared over any duration time period.

    Great chart, the colors just need to be shifted to the left one spot to make it legit.

  20. It’d be nice to be able to tell how much of the loss was due to the market versus inflation, but you get what you get. Since if the issue was mostly inflation, I doubt there were any winners in bonds or just savings accounts either, maybe if you were a gold bug you’d have made out then.

  21. OK I guess inflation in the 60′s and 70′s added up to a lot more than I realized. $1 in 1961 was about the same as $3 in 1981 due to the impact of inflation.

    So if you started with $10,000 in 1961 and it doubled to $20,000 by 1981 then thats more like having $6600 in 1961 dollars after considering the impact of inflation.

    Makes sense to me now.

    Of course inflation would have had the same impact to undercut the growth of any other investment too.

  22. Wow, look at the fairly clear periods where the color shades diverge for periods of 15 or 20 years. Judging by what I see here it looks like we are right smack in the middle of a bear market.

    If this is any indication, the latest run-up in stocks over the last several months (after a massive correction) is likely the best of what we will see for maybe the next 10 years or so.

    Sure wouldn’t give me any confidence in the market going forward over any kind of short or intermediate time frame. Looks like it could be many years of poor returns going forward.

  23. this is exactly why I only invest in CDs now. Not worth the risk in the stock market. For an average of 4% returns, no way.

  24. Maybe I am missing something but I am sure the average investor is not investing all their retirement funds at one time and waiting (hoping) it will grow. We are adding to the portfolio on a monthly basis during up and down markets. Have these analysts taken this into account?

  25. There are many ways to use this matrix.

    If you pick a vertical line, let’s say 2010. You can then see the returns if you had started to invest in 1940, 1980, 2000, whatever. You are setting the end date, and moving the start date. That shows how your returns changed for each year you put in money.

    If you pick a horizontal line, you can see how the returns would have changed over time as you change the end date. You are setting the start date, and moving the end date.

    If you pick a diagonal line, you are fixing the time horizon, and then moving along rolling windows of time. (See my earlier comment.)

  26. @Thad – There are multiple versions of this graphic by Crestmont. Some are inflation-adjusted (real), while some are not (nominal). Some include dividends, and some don’t.

    The one redone by the NY Times is just one of them, and the details appear to be correct. No way the S&P 500 returned 4% real after taxes without dividends.

  27. These charts are awesome -Thanks Jonathan!

    For those who are disappointed they’re not seeing the effects of gradual investment from each paycheck, do this:

    1. Observe the color changes over time for the horizontal line for the first year you started working and contributing to retirement savings.

    2. Repeat for each year thereafter.

    I actually feel more optimistic about my chances for successful retirement looking at this chart. My peers and I appear to be in one of the worst periods to have started investing, yet when I check my chances versus actual historical stock market returns and investing each year on calculators such as Firecalc.com I’m not doing so badly.

    So, the way I look at it, if history holds, there should be better investing years ahead as my peer group and I are ratcheting down our stock allocations as we get older.

  28. Taking the advice given the last time this was up, I used the new data to create a spreadsheet to see the effects of investing a set amount every year. From this I found that if I invested $1000 every year since 1985 (the year I started working after college), I would have invested $27,000 by 2012, and would have almost $50,000 in real (inflation adjusted) terms in a tax exempt account.

    And since I’m such a nice guy, I put a Google spreadsheet up that anyone else who wants to look at such things can use. I added data back to 1945, so anyone under retirement age can use it – just delete the years before your start, and you can see what your account would look like. You can also vary the amount invested each year if you want.

    The spreadsheet can be accessed at this link:

    https://docs.google.com/spreadsheet/ccc?key=0AqzALZmJa0NzdFpkd1ZwNVZUdmFkYVpLdWZJb0FDbGc&usp=sharing

    It is set to only view, but I think Google Docs lets you copy it to your own account.

  29. Wonder if there is something like this for real estate appreciation/fall across geographies. Anyone know?!

  30. Jonathan – this chart is so amazingly useful. I think you posted it before and I remembered it in my mind, but couldn’t remember where I saw it in the first place. I’m so glad you posted this again! I find it particularly interesting that the average return is 4.1% per year, yet people in the investment world (brokers, financial planners, etc) tell their clients to expect an average return of 10% per year.

    Do you have a similar chart that shows corporate bond rates and/or government bonds? I’d be curious to see if taking the risk in the stock market for only 4% per year is really worth it.

  31. That’s why I don’t think stocks shouldn’t be the lion share of any portfolio. Too much risk. In very long term they may offer modest returns. But in very long term we may be all dead.

  32. @Serge, so what investment may I ask do you make the lion’s share of your portfolio that has less risk yet earns more than 4.1% after inflation and taxes? :)

  33. @Tommy Z – those 10% annual returns you speak of are nominal rates, ie without taking into account inflation. The 4.1% shown in the chart is after adjusting for inflation.

  34. @Serge – If not stocks than what?

    When compared to inflation housing has been really unlucky (unless you factor in places like NY / SF)
    http://www.nytimes.com/2013/04.....&_r=0

    Bonds seem like a terrible place to be right now, as soon as the fed stops its buying the prices are going to be extremely unpredictable.

    P2P lending sites return is now pretty low. And their returns are taxed as real income…every year….unless you open a IRA on one of thesites.

    If you invest in the market as a whole…and stay with it…it should work out. Tax wise you are just taxed at the end (and also by a measly amount on dividends).

    Every place you put your money is risky, and the most risky seems to be sitting in a bank account. When you by into the market you are buying real assets as well, the railroads, factories, their cash on hand. If the market tanks and companies go bankrupt you are going to have some real assets after liquidation. There were many companies in 2008 selling below their liquidation value…even a better deal!

Speak Your Mind

*