Dividends, the Key to Long Term Returns, Even for Growth Stocks

A good article on dividends and their importance in the Wall Street Journal. Two quotes stuck out for me:

First, stock returns don’t typically consist of exciting price gains with dinky dividends tacked on. Over the eight decades ended September 2010, dividends contributed 44% of S&P 500 total returns, according to research by Fidelity Investments. And that includes a long, anomalous stretch during the 1980s and 1990s, when valuations bloated and yields shrank. During the 1970s, when returns averaged 5.9% a year, dividends contributed 71% of that figure.

That is all true, but it actually understates the importance of dividends. If you take out inflation dividends have over the long term contributed more than 80% of the real return (return above inflation) and over 100% of the real return during the seventies. James Montier provided these charts (pdf) to illustrate:




At the end of the day this is as it should be. Investors often talk as if dividends don’t matter as long as we get capital appreciation, but that is extremely short sighted.

Certainly we want fast growing companies who can reinvest their earnings at high rates of return to do so. However, the bar should be high since management and analysts are overoptimistic about their ability to do so. More fundamentally remember what the point of owning stocks is ultimately. It is to receive dividends. What is the point of owning a company if you never receive any of their earnings?

Casual readers might reply that people have made fortunes investing in companies that do not pay dividends, but they would be missing the point. Why do they make a fortune? Why do investors (at least over time) reward growing companies with a higher stock price?  If it is just because earnings are going up, we have to ask the question, “how does that help the investor?” They receive no cash from that growth. Is it just a game? We all just sit around bidding up companies that grow even though it does not benefit us directly for the price to go up? Are we all just making side bets because we know everyone else will want the stock because they know everyone else will want the stock ad infinitum? Often investors do act that way. They think that way and the market becomes divorced from dividends, free cash flow or any relationship with any tangible benefit for the investor. That leads to bubbles and busts and investor irrationality.

It should not be that way though, because there are only two reasons we should allow a company not to pay us its earnings.

  1. Because some of those earnings are needed just to keep the company going. All companies need to maintain and replace existing equipment and other assets.
  2. Because we want them to grow (accrete to book value) and be able to pay a larger dividend in the future. Warren Buffet’s Berkshire Hathaway is an example of a company that has focused entirely on that since Warren still believes Berkshire can earn a high enough return on retained earnings to justify doing so.

That is it. Number two being the key. If a company can retain a dollar and reinvest it to grow future income at a rate of 15% per annum for example it would make more sense for them to do so than give it to us. However, investors need some cash at times in the meantime, and we can sell shares (which appreciate due to growth.) Without that promise that at some point in the future a dividend will be paid owning stocks would be silly, akin to betting on horses. They would not go up!

Wonderfully, over time investors do earn a return that equates to the dividends distributed over time as the charts above illustrate. It does get out of whack from time to time. When it does the market eventually corrects that discrepancy, which is what the awful returns of the last 11 years has been all about at the end of the day. JJ Abodeeley calls it The Value Restoration Project.  We call it a secular bear.

Here is a wonderful chart from Doug Short which shows just how correlated dividends and growth are over time:


This chart tells us a very interesting story (though it is from July.) We use trend growth in both the Real (after inflation) price of stocks and the real growth of dividends. Trend Real Growth in Earnings has been somewhere around 1.5%.

Some things to notice.

Trend real growth in the index has been 1.7% a year, which is far lower than people realize. The rest of the return from stocks has been inflation and dividends as discussed earlier. Growth has been a minor component. We want to emphasize this quite carefully. The return of stocks is defined by the dividend yield, the growth in real earnings, inflation and a rise in P/E ratios. One could argue for real dividends as a better guide than reported earnings, but over time those should even out. There is nothing else. If someone argues for higher returns than implied by estimating each of those numbers they are wrong. It is not a difference of opinion, they are wrong. If they argue for faster growth than the 1-2% above inflation mentioned above, remember they are saying things will be much better than the past. Maybe so, but it is certainly unlikely.

Since 1870 a gap has grown between the two numbers in the graph above, which should track, to about 64%! Almost all of which is accounted for by the period since 1982. In all fairness some of this is because dividend payouts have shrunk, and possibly that will result in larger payouts in the future. Some of it is that the price since 1982 has gone way too high. This gap will likely shrink both by payouts from stocks increasing and prices coming down.

Of course Wall Street has been busy disguising the lack of cash actually returned to investors through accretion to book value and dividends by getting us to focus on forward operating earnings. How has that worked out? John Hussman shows us:


Historically, the actual reported net earnings of the S&P 500 have averaged only about 72% of one-year forward operating earnings estimates by Wall Street analysts. The sum of dividends and increments to book value have been even lower, averaging just 60% of forward earnings estimates (and representing only about 84% of the net earnings reported to investors). The remaining portion of “earnings” reported to investors goes the way of the Dodo.

Not too well.

Since dividends have been given short shrift in recent decades it would be wrong to say that stocks were as overvalued in July as the 64% gap in the graph on Real Price and Real Dividend’s would lead one to believe (since with the market price well above trend the overvaluation would have been a lot more that 64% without that caveat.)

However, we should acknowledge that stocks using normal assumptions of growth of 1-2% above inflation plus dividends (about 2.12% today) are priced to deliver over the long-term only about 3-4% above inflation. Not exactly inspiring. Throw in about .5% to account for a low payout today (assuming the retained earnings are not squandered, a big if) and we get 3.5% to 4.5% above inflation. If we set fair value at a projected return of 6% above inflation stocks would need to fall by about 58% at a 1% real growth rate and 25% assuming a 2% real growth rate.

Note that a 2% real  growth rate is not only above long-term averages, it is also unlikely since we are already at peak profit margins. From peak profit margins long-term real growth rates have generally averaged between 0% and 1%. Because companies are loath to cut dividends, barring a major financial crisis we should expect the floor to be closer to 1% on dividend growth (and that rate possibly higher as payouts increase) though earnings would likely be much more volatile.

This leads me to the second quote that stuck out:

Second, from here, broad-market returns might be smaller than investors are accustomed to. Bradford Cornell, a finance professor at the California Institute of Technology who specializes in valuation, argued in a paper published last year in the Financial Analysts Journal that stock returns are inextricably tied to economic growth, which is necessarily slowing around the developed world. Stock investors, he says, should expect to collect their dividend yields plus about 1% a year in price gains after inflation.

The connection between dividends, long term economic growth and the growth in earnings and dividends we have been discussing is exactly why Bradford Cornell’s paper is mentioned in the article. Growth is lower than we think and not nearly as important as dividends in explaining your long-term return. You can find his paper here.

What about stock buybacks? Yes, they do count, though their worth is highly exaggerated and are not equal on a dollar for dollar basis to dividends. However, it is a way for companies to return cash to shareholders. What it doesn’t do is increase the value of the S&P 500 as a whole.

First of all the index is already adjusted for share buybacks. So adding them into the index numbers to assume a higher return is double counting.

Second, the value of share buybacks is they reduce the share count, meaning each individual share can receive more of the dividend, even if delivered far in the future.

If dividends are not high then returns are low once speculative mania has reversed. The positive side of that is negative overreaction will eventually arrive and stocks will be yielding 4-6% at some point in time, usually because of financial disorder or inflation. When it does long-term returns can be far above 6% real (such as in 1982 or 1974.)

At that point the Value Restoration Project will be finished as well as the secular bear. Then we can actually be long-term investors again in the US stock market and not speculators hoping for historically unusual outcomes to bail us out.