We've been underestimating our trading returns


I was looking at a long term chart of the Dow Jones Industrial Average, and something struck me as really weird.
The economy & the stock market move in the same direction in the long term. If this is the case, then why did the Dow go nowhere from 1900 to 1924?
There’s not a lot of economic data from that period, but most accounts demonstrate that the economy was growing during most of that period (minus a few recessions). America at the start of the 20th century was a BOOMING nation (crazy population growth, which is good for business).

Recently I read something about Australia’s stock market (I live in Australia). On the surface, it seems as if Australia’s stock market has a lower average annual return vs. the U.S. stock market. But that’s because Australian stocks pay higher dividends than U.S. stocks (that’s the culture here anyways).
And then it hit me.
The S&P 500 EXCLUDES DIVIDENDS. It does not include dividends. Hence, the S&P 500 IS NOT an accurate representation of the “U.S. stock market”. An accurate representation of “the U.S. stock market” would include dividends into the index. That’s what the S&P 500 Total Return Index does. It includes dividends.
*The official S&P 500 Total Return Index only goes back to 1988. But I’ve managed to calculate it back to 1927
Here’s how dividends work. Imagine Company XYZ’s stock price is $100. It issues a $1 dividend. The stock price tomorrow will fall to $99.
The S&P 500 does not take into account this $1 dividend. E.g. If company XYZ issues a $1 dividend, the S&P 500 will fall by $1. On the surface, it looks like investors who owned the S&P 500 (e.g. through $SPY) lost money ($1). But that’s not the reality! Investors now have $99 worth of the S&P 500 and $1 in cash. They can use that $1 and reinvest it back in to the S&P 500!
So straight away, you can see that if the S&P 500 is flat but the S&P 500 companies are issuing dividends, investors are actually making money.
That is why the stock market was flat for practically the first 25 years of the 20th century. Back then, corporate culture favored issuing dividends. Dividend yields on the average stock were much higher than they are today.
*Why don’t you notice the post-dividend downwards adjustment in the S&P 500? Because there are 500 stocks in the S&P 500, and companies don’t issue all of their divdends on the same calendar day. Hence, all the adjustments are spread out over the 252 calendar days. If all the companies hypothetically decided to issue dividends on the exact same day, the S&P would tumble significantly on that 1 day merely due to the ex-dividend adjustment.

How dividend yields have changed over time

Here’s a chart of the S&P 500’s dividend yield over time

Click here to download the data in Excel
As you can see, the stock market’s dividend yield has gone down over time, especially after the mid-1980s.
That’s why to calculate the stock market’s TRUE RETURNS, you need to include the dividend yield back into the S&P 500 Index.
The S&P 500 Total Return Index exists to 1988. I have used monthly data to create a S&P 500 Total Return Index all the way to 1927

Click here to download the data in Excel
Immediately the picture becomes much clearer.

  1. From 1927 – present, the S&P 500’s average annual return is 5.7%. The average annual return from 1950 – present is 7.6%. (1927 -1950 was a bad time for the stock market)
  2. From 1927 – present, the S&P 500 Total Return Indices’ average annual return was 9.8%. That’s 4% higher!
  3. It’s also worth noting that when including dividends, the S&P 500 recovered much faster from the 1929 crash than most people believe. By 1937 the S&P Total Return Index was almost where it was in 1929, but the S&P 500 index (excluding dividends) was still far below its 1929 peak.

The Total Return advantage exists in every single decade, because S&P 500 companies always pay a dividend that >0% (it’s impossible to issue a negative dividend).
From 1988 – present, the S&P 500 yielded an average annual return of 7.99%. The S&P 500’s Total Return Index has yielded an average annual return of 10.4%. This is 2.4% higher
*It’s surprising how 95% of professional investors, traders, and finance professionals think that “dumb” buy and hold yields 7-8% a year. It doesn’t. It’s higher.
So what does Warren Buffett mean when he says “the S&P 500 yields an average of 6-7% a year”?
He’s referring to the fact that the REAL RETURN (inflation-adjusted) of the S&P is 6-7% a year. That is true. Inflation has averaged 3-4% over the past 100 years.
No other asset class comes even close over the past 100 years.

What this means for investors and  traders

For starters, this means that the returns in all of our long-only trading models have been underestimated by the dividend yield. Our returns are based on the S&P 500, excluding dividends. Including dividends, returns would be higher.
*Anyone who owns $SPY, $SSO, or $UPRO would earn dividends
Moreover, this explains why it’s better to focus on medium term and long term investing.
All these short term traders are glued to their screens each day, and on a 30 year period they can barely beat buy and hold (7.6% a year) for the S&P 500. When you factor in the dividend yield, these traders are lagging buy and hold significantly. (i.e. if you can’t even beat 7.6% a year, there’s no way you can beat 10% a year)
You need to be more careful with turning bearish than with turning bullish
This is the problem with traders. They are bearish 50% of the time and they are bullish 50% of the time. (Because for a trader, the best thing is if the market goes nowhere and just fluctuates in a small range every year).
This is inherently illogical, because the stock market has a higher % probability of going up vs. down.

The above chart looks at the S&P 500 itself. It doesn’t factor in dividends.
Here’s the same statistic, using my new S&P 500 Total Returns Index from 1927 – 2018

Here’s the same statistic, using my new S&P 500 Total Returns Index from 1950 – 2018

It’s easy to see how bears have a massive disadvantage.

On a side note

And here’s the important point: S&P 500 companies have been decreasing their dividend yields over time.
So while all the permabears complain “this bull market is crazy, the S&P 500 is rallying too fast”, realize that this is partially because dividend yields have gone down. If you have a same total return yield and the dividend yield goes down, that means the capital gains yield goes up (which is what people are normalling refering to when they talk about the S&P 500).
Here’s an example.

  1. Total return yield is 10%. If in the past, the dividend yield was 4%, that means the capital gains yield (change in S&P 500, ex-dividends) was 6%.
  2. Total return yield is still 10%. But now, the dividend yield is 2%, which means that the capital gains yield becomes 8%

On the issue of dividend investing

I never understood the purpose behind dividend investing. People love it because “it generates income”. That doesn’t make sense to me.
*Please comment below in the comments section if I got this wrong. I don’t do dividend investing.
When a company issues a dividend, its stock price is automatically adjusted by the size of the dividend.
E.g. if XYZ company’s stock price was $100 pre-dividend, and then it issues a $1 dividend, its stock price will go down by approximately -$1 on tomorrow’s OPEN. Hence, an investor who used to own a $100 stock now has $99 in stock and $1 in cash. It’s the same thing.
So by buying dividend stocks, these investors aren’t earning an “income” for themselves at all. They could just as easily do the same thing by:

  1. Buying a stock that pays $0 in dividends, and….
  2. Sell e.g. 3% of the stock every single year, thereby fooling themselves into thinking that they earned a 3% “income” on their shares.

Am I understanding this correctly? Comment below!

17 comments add yours

  1. not sure I am on the same page Troy but theoretically once the ex dividend period is out of the way, wouldn’t the stock price again rise because the investor would be entitled to the next dividend? In other words, at the end of the xd period if the market is rising the investor would be back at $100 in stock plus having received the dividend of $1 ?

    • It would rise, but it wouldnt’ rise to as much as it could have without the dividend.
      Example:
      let’s assume that a stock without dividends would have gone from $100 to $108.
      The same company now decides to issue dividends. Its stock goes from $100 to $99, and the investor receives $1 in cash. Then the stock rises from $99 to $107. So the investor now has a $107 stock and $1 in cash

  2. Troy,
    It’s my understanding that SPY dividend yields are significantly higher than for SSO and UPRO. So I would guess that dividends will have a more pronounced impact on SPY. Will you be updating the historical model returns to incorporate dividends?
    Thanks,
    Zach

    • True. Just a quick glance at the numbers: SSO and UPRO don’t increase dividends.
      E.g. if SPY is paying a 2% dividend, that’ll be 1% for SSO, and 2/3% for UPRO
      No I won’t. It’s hard to know the exact month to month and day to day dividend yield. I’d rather leave the models understated and err on the safe side

  3. As far as I understand dividend/income investing, the point isn’t purely mathematical. It’s about pursuing a different class of opportunities.
    Dividend opportunities usually look like large installed infrastructure bases that are throwing off tons of cash (but have limited ability to utilize that cash to pursue additional growth). These companies don’t have anything better to do than payout dividends–and they tend to have stable, low risk cashflows. The price may swing wildly with market fluctuations, but you just look to get in at a good time and hold for the dividend. You ignored the capital gains component of returns, assuming they net to 0 if you enter over time. It’s kind of like buying an annuity, as opposed to stock. AT&T back was the classic example of this, back before telco deregulation.
    The above was in contrast to looking for companies that would specifically generate capital gains. This could be from buy-and-hold, momentum, growth, value, or other strategies. For example, finding undervalued companies that will mean revert to their fair value is still a capital gains strategy vs an income strategy.

    • Good point. But sometimes, the capital gain/loss can completely swamp dividends. E.g. AT&T is a favorite for dividend investors. It’s down -19% this year

  4. Troy, thank you for your additional work.
    Just a thought from my side that came to my mind today: dividends issued by companies are also a kind of a compensation to the shareholders for investing their money in the company business.
    This is a gross income for the investor, but it is also compensated (in a negative way) by inflation: getting a 2% dividend is basically recovering the svalutation of my money given by inflation and the net income is 2% -inflation rate.
    In the 80s inflation (and also all common interest yields) were high in Europe and Us. I remember in Italy to be something like 10-15%. So also the companies were forced to issue large dividends to get funds for their activities (without significant dividends investor should have withdrawn their money and invested in other assets)
    But you know, considering inflation in the whole history of markets would be a very complicated operation.
    The logaritmic graphics can simulate the inflation contribution (using an average rate, which is not the real inflation over time)
    So after all, it is true that S&P yielded 10% per year on average, but we should also consider what has been the average inflation since 1927, and substract it.

    • Good point. That’s why Warren Buffett says on average, the S&P returns 6-7% a year after inflation.
      But the reason we don’t consider inflation is because no one else does as a benchmark. E.g. hedge funds don’t say “I want to earn 15% a year after inflation”. They just say “I want to earn 15% a year”. Industry standard I guess.

  5. There is another component (in Australia at least) that favours dividends over capital gains, franking credits. Franking credits essentially give you gains with the tax paid on them already, ie they are More tax effective.

    • Good point. I think the U.S. double taxes dividends (once at the corporate level as corporate earnings, and once at the indiivdual level as income for shareholders)

  6. Hi Troy –
    Relative to your last coment: “So by buying dividend stocks, these investors aren’t earning an “income” for themselves at all. They could just as easily do the same thing by:……..”
    That actually doesn’t work over the long-term because you have to sell more shares during bearish periods to yield the same amount of income (if the stock is down 50% you have to sell twice as many shares to achieve the same amount of income). Eventually, you impair the value of your whole portfolio – whereas using dividends your share count never goes down.
    I have spent many, many hours contemplating this very issue – there are many other aspects to it as well.
    Thanks as always for your outstanding work!
    Chris B.

    • Good point Chris,
      But don’t dividends fall during bearish periods? I.e. the first thing that companies cut back on during -50% declines is dividends. I.e. do a lot of companies sort of peg their dividends as a % of stock price? So if the stock price falls 50%, dividends would also be down 50%? So it seems like these 2 things would be the same.

      • That is generally not the way it works in the universe of high-quality dividend payers, as evidenced by the fact that there are more than 250 companies out there that have raised their dividends for between 15 and 60 years in a row (see http://www.dripinvesting.org/Tools/Tools.asp ). You are probably correct with the marginal dividend payers, but not the companies for whom dividends are in their blood.

  7. I don’t like dividends! I prefer growth stocks that pay little or no dividends to those that pay good dividends – because, as an overseas investor/trader in US stocks, I get taxed at 30% on dividends, and 0% on capital gains. So for me, that is an added advantage to using SSO or UPRO.

  8. Hi Troy
    Just commenting on dividends in general.
    1: You are correct. US companies pay dividends before taxes. Except REIT’s.
    2: High dividend stocks (and short-term capital gains) are much more attractive in Retirement Accounts. Taxes will eat up these investments in Personal Accounts.
    Appreciate your help. You’ve turned into my number one source.
    Dan in Dallas

Leave a Comment