14

This is probably a simple question, but I don't know much about investing, so here goes:

Some time ago, though I can't find it again now, I read a couple of responses on this site saying that these index funds historically only "lose" for very short times, as far as I can remember they said that they always recover within a year or two even after a large market crash.

However, looking at the graphs for S&P and Dow Jones on Wikipedia, that seems wrong. In fact, it seems that if someone invested in an index fund based on S&P 500 around the year 2000, they might have been in a loss until around 12 years later (unless they exited within a short time during late 2007 / early 2008, and possibly not even then if corrected for inflation). Am I correct in reading these graphs at face value, or is this dependent on some other information and the actual value did rise faster that the graphs indicate?

ispiro
  • 1,253
  • 1
  • 11
  • 16

5 Answers5

27

Historically, 20 year market returns have always been positive but there have been a number of 10 year periods where it lost money. The most recent one was from 2000 to 2009 and is called the Lost Decade (a loss of about 7% for the SPY). For someone who invested at the end of 1999, it took almost 11 years to break even. The DJIA fared a bit better, breaking even in about 9-1/2 years.

The accuracy of your two graphs depends on if they accounted for dividends. While it's not a huge amount, the SPY yielded an average of about 2% during that decade.

Bob Baerker
  • 76,304
  • 15
  • 99
  • 175
  • Thanks. You answered my question. You mentioned dividends, though, and I was wondering if dividends offset whatever inflation was in those years. Because if they didn't, and maybe even if they did, the number of years to cut it even in adjust-for-inflation dollars might even be greater than ten years. And that's not even considering the costs for handling the index funds. – ispiro Mar 27 '22 at 16:39
  • 6
    Here's something that an awful lot of investors have no clue about. Dividends do not offset anything because they do not provide total return because on the ex-dividend date, share price is reduced by the exact amount of the dividend. Read this. Though taxed as income, they are not true income. Your account will not be worth more because you became eligible to receive one – Bob Baerker Mar 27 '22 at 17:46
  • Isn't the share-price reduction temporary? If not, are you saying that articles like Microsoft Remains a Top-Notch Dividend Stock total nonsense? (I'm asking. Perhaps the answer is yes.) – ispiro Mar 27 '22 at 19:02
  • 4
    The reduction because of the dividend is permanent. If there is a dividend of $2 on the ex-div date then share price drops $2 because of it. After that, share price can go up or down but that is a separate issue. And if you are so inclined, you can observe this for every stocks that offers a dividend. – Bob Baerker Mar 27 '22 at 20:38
  • 1
    And the S&P 500 has only existed for roughly 3 20-year periods. So it might just be coincidence that it went up in every 20-year period. Don't forget our economy is flying by the seat of our pants, relatively speaking. The last time major upheaval causing a "new economic order" to begin, was in 2008 - it's not tried and tested. – user253751 Mar 28 '22 at 08:59
  • 3
    @user253751 Yes, tThe S&P 500 has only been around for 65 years but on an annual basis, there have been 45 twenty years periods since then. And if you use indexes that have been around longer, you can test back to the 1870's. It's no coincidence. – Bob Baerker Mar 28 '22 at 11:24
  • 3
    @BobBaerker yeah but the 45 periods aren't all independent - you can't say "well the probability that they'd all go up by chance is 2^-45" for example. Counting it as 3 periods gives a moderate underestimate to the probability that they'd mostly go up; counting it as 45 periods give a big overestimate – user253751 Mar 28 '22 at 14:26
  • 3
    You can interpret the data any way you like but as a market metric, it doesn't work that way. When looking at historical performance, the statistics for 10/20/30 year performances are presented as a rolling return from the starting date. Google for some examples. You won't find any showing a graph of only 3 separate time periods in 45 years. That would be totally useless data for someone looking to begin investing in a mid-period. – Bob Baerker Mar 28 '22 at 14:50
  • 1
    This is only true for investors who are no longer making contributions to their accounts, or whose contributions make up a very small percentage of their total holdings. If you contribute throughout a down turn, you become profitable again significantly faster. – BlackThorn Mar 28 '22 at 15:40
  • 6
    "Historically, 20 year market returns have always been positive" in the USA, yes. Some other countries haven't been so lucky. – Peter Green Mar 28 '22 at 17:44
  • 10
    "Dividends do not offset anything because they do not provide total return because on the ex-dividend date, share price is reduced by the exact amount of the dividend." The point of bringing up dividends is that indexes don't account for the value of dividends that would be paid if you owned the underlying stocks. If you buy an instrument (mutual fund or ETF) that tracks such an index, you may or may not receive that value as a distribution that you have to reinvest yourself; or you may be able to set up an automatic reinvestment; or it may be reflected in the value of the instrument. – Karl Knechtel Mar 29 '22 at 04:50
  • 1
    "And if you are so inclined, you can observe this for every stocks that offers a dividend" Not always very clearly, though - because the effect due to the dividend will overlay with any price change that would have happened that day normally due to ordinary speculation on the stock. – Karl Knechtel Mar 29 '22 at 04:51
  • @Karl Knechtel - It is clearly observable if you understand what you are looking at. First, on ex-div, the close will have been adjusted by the dividend before trading resumes. For example, if XYZ closes at $100 and offers a $1 div the next day (ex -div), the adjusted close will be $99 in the morning. – Bob Baerker Mar 29 '22 at 09:15
  • cont. - Second, attempt to reconcile the day to day change. Let's say XYZ closes at $99.40 on ex-div. The closing quote will say up 40 cents. Now how is it possible for it to close at $100 one day and then close at $99.40 the next day, up 40 cents? $99.40 is less than $100. Yes, this is obscured due to 'ordinary speculation in the stock' but not so much so that you can't see what's going on. – Bob Baerker Mar 29 '22 at 09:16
  • Wrt dividends, it should be pointed out that some ETF's re-invest dividends and others pay out dividends. Obviously the former will have a higher apparent return - the 2% mentioned in the answer. – MSalters Mar 29 '22 at 09:53
  • @MSalters - I don't think that reinvesting the 2% in dividends would change the historical performance of the fund because share price is reduced whether you keep the dividend or reinvest it. If reinvested, it would affect total return: higher if positive compounding, lower if negative compounding. – Bob Baerker Mar 29 '22 at 12:52
  • 4
    @BobBaerker: My point is that the same applies to the ETF itself. When it pays the dividends forward, the ETF price goes down. But when an ETF receives and reinvests dividends, that ETF price does not go down. For most ETF's, the chosen strategy is fixed when that ETF is initially created. – MSalters Mar 29 '22 at 13:11
  • @MSalters - I don't invest directly in ETFs so I'm not familiar with this. Just to clarify, this is a different reinvestment process than when an investor receives the dividend on the Pay Date and reinvests it at a subsequent ETF share price? If that's the case then I understand your clarification and yes, the graphs would be different. – Bob Baerker Mar 29 '22 at 13:52
  • Any N-year period that ends with a crash will be a poor example of long-term trends. – Barmar Mar 29 '22 at 15:23
  • 1
    @MSalters I thought regulations require funds to pass along most dividends and CG, not reinvest them. The investor can choose to reinvest their dividends. And most statistics about funds assume that they do. – Barmar Mar 29 '22 at 15:25
  • @BobBaerker not sure that's a valid point, because it also goes up by the dividend amount in the time leading up to the ex-div date in anticipation of the payment. It's just realising some of the profits. If you're reinvesting it works out the same (minus the taxes). – Aequitas Mar 29 '22 at 22:44
  • 2
    "Dividends do not offset anything because they do not provide total return because on the ex-dividend date" This is either a false statement, or a true statement confusingly stated that is very likely to be interpreted in such a way as to yield a false statement. The action of issuing a dividend doesn't increase returns, but including dividends in calculations of returns absolutely does increase the calculated return. – Acccumulation Mar 30 '22 at 01:10
  • @Acccumulation - Perhaps if you read the detailed explanation in the Vanguard link that I provided in the comment that you're citing then you would be less confused and you would understand that the dividend MUST be included in the calculation of return in order to account for the capital loss created on an ex-dividend date. The Vanguard article is quite clear. The only way that a dividend increases total return is if share price increases subsequently (compounding). – Bob Baerker Mar 30 '22 at 10:16
  • "The only way that a dividend increases total return is if share price increases subsequently (compounding)" I think you are understanding "increase" differently. We do not mean "cause to be higher than if the company reinvested the money rather than paying a dividend" (you are talking about, I think, Miller & Modigliani's indifference theorem, which we do not dispute and is not at issue). We mean "cause to be higher than the value one might naively calculate, if one only looks at stock quotes, and neglects the fact that the owner of the stock (or index fund, etc.) will receive dividends". – Karl Knechtel Mar 31 '22 at 01:06
  • @Karl Knechtel - I think that you should read about what total return is so that you don't have to offer guesses about what you think I mean. The short answer? If a dividend causes an equal amount of capital loss, the total return is zero from that dividend. If it is a non-sheltered account, taxation of the dividend will cause negative total return. – Bob Baerker Mar 31 '22 at 03:48
  • Yes, I understood that completely, and I know what total return is. Do you understand that the value of the S&P 500 index, as reported by news agencies etc., does not reflect the total return of an index fund allocated the same way? Nobody arguing with you suggested that a dividend "increases total return". What we are saying is that it causes total return to be higher than the index number suggests, because it is a component of the total return which is not reflected in that index. – Karl Knechtel Mar 31 '22 at 10:37
  • Per your own link: "Total return, when measuring performance, is the actual rate of return of an investment or a pool of investments over a given evaluation period. Total return includes interest, capital gains, dividends, and distributions realized over a period." To say "the total return is zero from that dividend", therefore, makes no sense; you cannot say that the dividend has a total return of zero because "total return" is a metric applied to investments, not to sources of return. – Karl Knechtel Mar 31 '22 at 10:40
  • 1
    This is important because OP specifically mentioned looking at "graphs of S&P and Dow Jones", which ordinarily will be graphs of the corresponding indices. These will show share price, not total return. – Karl Knechtel Mar 31 '22 at 10:44
  • Your responses remind me of the story about someone asking what time it is and being told how to build a watch. It was a very simple statement about how ON THE EX-DIVIDEND DATE, the dividend does not create total return. As stated, total return is due to subsequent price return not becoming eligible to receive the dividend on the subsequent Pay Date. You're off on a tangent. – Bob Baerker Mar 31 '22 at 15:53
18

There are no guarantees on either the short or long term that any investment will continue to rise. That has just been the general trend historically. You can always take any two arbitrary dates and show that a strategy was brilliant or stupid, but you are still looking at the short term volatility. Contrary to what you may believe, 8-10 years is not the long term. You could have just as easily picked two other dates to prove the opposite. However, try that on a 20 or 30 year horizon and the trend becomes clearer.

Also, consider that most people invest gradually over time. They don't just drop a bunch of money in at year X and then take it all out at year Y. You would likely have been investing during those dips and seen pretty good returns on that money that evened things out.

JohnFx
  • 52,979
  • 12
  • 134
  • 245
  • Thanks. I had not considered that last point of investments being done over time instead of once in and once out. – ispiro Mar 27 '22 at 16:43
  • 12
    Periodically adding more money on a regular basis during the Lost Decade of 2000 to 2009 would have shortened the breakeven period. – Bob Baerker Mar 27 '22 at 17:47
  • 1
    You ended with a loss, if you started investing regulary in 1999 and stopped investing in late 2008. – Bernhard Döbler Mar 27 '22 at 18:14
  • 1
    @BernhardDöbler Sure, but hopefully you invested gradually over that time period and also sold gradually over that time period - in which case you still made out pretty well. If you sold your house on Jan 1 1999 and put all the money in there, and then needed to buy a house in 2008, I guess you're out of luck, but that's just poor money management. – Joe Mar 28 '22 at 16:28
  • @Joe so I should invest and uninvest at the same time? Or was the idea that I should plan ahead on when I need the money, and then gradually uninvest before that? – Paŭlo Ebermann Mar 28 '22 at 23:49
  • The latter @PaŭloEbermann . Investing for the long term means being nearly 100% in stocks until you’re five or ten years out of needing the money, then start to balance to a lower risk profile. If you’re talking retirement and are 40, you stay all in. When you hit 55, you start to adjust to more bonds or safer stocks over the next five to ten years. So someone retiring in 2008 should’ve been selling from 1998 on - so they did fine. Not as good as someone retiring in 1998! But still fine. – Joe Mar 28 '22 at 23:54
7

The "ten year" rule is not "you're guaranteed to make money in ten years", it's rather that "in a less than ten year period, invest more cautiously, in a ten or more year period, invest more aggressively," because normally 10 years is indeed enough to likely avoid any major crashes.

The 1999-2001 crash was exceptional, but mostly because the 1994-1999 period was also exceptional - the amazing growth of that period (tripling over five years!) caused by both lax regulations around speculation and a lack of understanding of the actual value of tech stocks. Yes, a 1999 investment would have taken a long time to show growth - but a 1997 investment would have been fine (around 700 for S&P, the 2008 crash only barely touched that and very quickly recovered from it to see a reasonable gain from that 1997 investment).

It's very, very unlikely that today is the peak of the bubble, and hence the advice that a ten year horizon is long enough to mostly disregard that possibility - and I can tell you that there were tons of people in 1996-1997 warning of the bubble (that was, of course, actually a bubble!) who, if you listened to them, you still lost out on tons of gains.

A 10-year rolling return chart is the better way to visualize this to fully understand what's happening here - in particular due to the insane growth in the 1990s making it very hard to see what happened in the prior periods (all of which are consigned to a more or less straight line on the S&P price chart due to resolution). And to add to that, recognize that the strategy for retirement would be to start moving out of the stock market around 10 years before retirement, with a more aggressive move out at 5 years - so even if you were heavily in stocks and retiring in that 1999-2008 period, you'd still be selling a ton of stocks at a gain during the peaks.

The real lesson is "don't plan to sell everything at once" - have a multi-year plan to slowly move from stocks to more safe investments.

Joe
  • 35,786
  • 6
  • 90
  • 128
  • Thanks. But it seems that the graph you linked to is not inflation-adjusted. Do you have a similar graph that is by any chance? – ispiro Mar 28 '22 at 16:43
  • 1
    @ispiro I'm sure you can find one, but for the most part you don't adjust for inflation in ten year windows like this - the percentage itself is not affected by inflation, just the relative year-to-year change, and most 10 years periods don't have enough inflation for it to be very significant (Late 1970s aside). – Joe Mar 28 '22 at 17:41
  • @ispiro Rolling 10 year real returns from 1800-2021. Whether you should correct for inflation depends what you want to visualize. If you want to see the effects of both inflation+stocks then use nominal prices. If you want to plot real growth in purchasing power, then use real prices. (If you don't correct for inflation, then Zimbabwe would be considered one of the best investments of the last 20 years...) – bain Mar 30 '22 at 09:35
5

It could take a lot longer

The US is something of a special case, having had large and relatively steady growth for over a century.

There is no guarantee this trend will continue.

There have been times when investing in the indexes of other major economies (Say 1980s Japan) would have taken 20-30 years to see a positive return.

Others, where you end up losing almost everything and never really recover.

Nothing is ever guaranteed or predictable. Especially in finance.

Zakatak
  • 51
  • 1
  • Another example, iirc the German stock market had two complete wipes in the first half of the 20th century. So if you were in the stock market when one of those happened, your money was just completely lost. – quarague Mar 29 '22 at 09:45
  • @quarague You might confuse that with the currency. In 1923, the "old" Mark was replaced with the "Rentenmark" by 1,000,000,000,000:1 (one trillion!). In 1948, the "Reichsmark" (which was quite the same as the "Rentenmark") was replaced by the DM by about 10:1. But the shares kept their value across 1923. According to this Wikipedia article (in German only, alas), trading with shares ceased in 1943 and started again in 1948 with about 10% of their value and only gained their 1943 level in 1954. So not a complete wipe. – glglgl Mar 29 '22 at 10:26
  • @glglgl Thanks for the link, no complete wipe but it still seems most of the first half of the 20th century was a bad time for investing in the German stockmarket. – quarague Mar 29 '22 at 10:30
1

The book Irrational Exuberance 3rd ed. lists several historical periods where the real (inflation-corrected) market price took a long time to recover:

1901 crash:

"After 1901, there was no pronounced immediate downtrend in real prices, but for the next decade, prices bounded around or just below the 1901 level and then fell. By June 1920, the stock market had lost 67% of its June 1901 real value. The average real return in the stock market (including dividends) was 3.4% a year in the five years following June 1901, barely above the real interest rate. The average real return (including dividends) was 4.4% a year in the ten years following June 1901, 3.1% a year in the fifteen years following June 1901, and -0.2% a year in the twenty years following June 1901."

(the book does not state the exact year that the market recovered and exceeded its 1901 peak, but it was some time in the 1920s)

1929 crash:

"the market tumbled from this high, with a real drop in the S&P Index of 80.6% by June 1932. The decline in real value was profound and long-lasting. The real S&P Composite Index did not return to its September 1929 value until December 1958. The average real return in the stock market (including dividends) was -13.1% a year for the five years following September 1929, -1.4% a year for the next ten years, -0.5% a year for the next fifteen years, and 0.4% a year for the next twenty years."

(29 years from peak to recovery)

1966 crash:

"Real prices bounded around near their January 1966 peak, surpassing it somewhat in 1968, then falling sharply back after 1973. Real stock prices were down 56% from their January 1966 value by December 1974 and would not be back up to the January 1966 level until May 1992. The average real return in the stock market (including dividends) was -2.6% a year for the five years following January 1966, -1.8% a year for the next ten years, -0.5% a year for the next fifteen years, and 1.9% a year for the next twenty years."

(26 years from peak to recovery)

2000 crash:

"The Dow peaked at 11,722.98 in January 14, 2000, just two weeks after the start of the new millenium. The market had tripled in five years. Other stock price indices peaked a couple of months later. The real (inflation-corrected) Dow did not reach this level again until 2014, and, as of this writing, the real Standard & Poor's 500 index has still not quite returned to its 2000 level."

(14 years for the Dow to recover, at least 15 for the S&P 500, as the book was published in 2015).

bain
  • 111
  • 2