Thursday, August 22, 2013

The ABC's of Dividend Investing, Part III: So Goes the Dividend, So Goes the Stock

I am reposting this blog.  Several people indicated there was something wrong with the formating that did not allow them to read two paragraphs near the bottom.  Hopefully this version is competely visible.

When I first thought about writing the ABC’s of Dividend Investing, the third edition was supposed be the big wrap-it-up-with-a-bow offering:  A summation of everything we have learned in all these years of managing money reduced to a couple of paragraphs.  The truth is I have spent all afternoon staring at Oregon’s snow-tinged, Three Sisters Mountains attempting to tap into a narrative that would help tell the story in a simple, understandable way.  I can’t do it.  The principles of dividend investing are reasonably straightforward; however, the process of determining what to buy or sell and when to do it is incredibly complex.
For that reason, I will offer additional ABC’s in the future on this blogspot, zeroing in on confusing markets or worrisome stocks to help our readers see how our team arrives at buy and sell decisions.

Let me take a running start at a intermediate conclusion by listing what was described in Parts I and II, with some side comments.

  • Dividends have represented between 40% and 50% of the total return of stocks over the last 53 years.  If we are trying to make money in stocks, we’d be crazy not to pay attention to dividends.
  • The Dow Jones Industrial Average (DJIA) price growth and its dividend growth have been surprisingly similar over the long-term.  Since 1960 to the present, DJIA prices have grown at an average annual rate of 5.9%, and dividends have grown near 5.5%.  This association combined with the previous point would suggest that you ignore dividends at your own peril. (Remember, I provided 1960-1989 figures, and not to 2013.)
  • Dividends are much more predictable than earnings or prices.  The standard deviation of DJIA annual dividends is about 9% a year, versus nearly 15% for prices and 20% for earnings.  For crying out loud, watch dividend growth as the best predictor of price growth.
  • Investors are willing to pay up for dividend growth if it is consistent and enduring. You will soon see how this is the single most important point that we can make about dividend investing. 
  • Stocks with high dividend yields and low dividend growth perform well in falling interest rate or weak economic environments.  Stocks with high dividend growth do best in rising interest rate environments, or when the economy is beginning to accelerate.  I see many dividend investors ignoring  price appreciation.  That might seem like a more pure way of looking at dividends, but it is not realistic.  Price appreciation must always be taken into consideration as you will see later.

Now let’s look at some of the tools we use to make investing decisions.  One of our mottos is “So goes the dividend, so goes the stock.”  If you think about this for a moment, you realize that if this were true, it would offer a way to value a stock.  It would be using historical relationships to do so, yet everyone knows that the past does not necessarily predict the future.  However, since the future is unknowable, understanding the historical relationship between dividend growth and price growth can provide a strong clue about the value of a security, particularly if the relationship is statistically significant.

Scatter plot of
DJIA Dividends v. Price
Since 1960


Zero in on this chart for a few moments.  It has four stories to tell. 

  1. It is clear from the plot points that as DJIA dividend has risen (horizontal axis) DJIA price has risen (vertical axis). 
  2. The line running through the middle of the plots is called the linear regression line, or the ordinary least squared line.  If dividends and prices were perfectly correlated, all the dividend and price plot points would all be on the line.  Since they are not, we know the relationship is not perfect.  However, the relationship is a lot tighter than it might seem at first glance mainly because I am using a regular scatter gram instead of a log version.  I am doing so because I want to use it to make a point later.  Suffice it to say that if I were using the log scatter plot, the statistical significance is even higher than what is shown on this regular plot.
  3. Look at the formula at the top left side of the chart.  This is the formula for the linear regression line.  Below the formula is the term R2 = 0.883.  This statistical test indicates how tight the fit is between the annual movements of DJIA dividends and price.  The score for a perfect fit is 1.0.  An R2 of .883 is highly significant.  What it is saying is that annual changes in DJIA dividends have been able to predict over 88% of the movements of DJIA price.  By adding a few more variables such as earnings, inflation, GDP, and interest rates, we can move R2 up to as high as .94 for the Dow Jones Industrial Average.  Thus, a single formula has been able to do a decent job over the years of estimating the growth rate of the DJIA. 
  4. Now look at top most point on the linear regression line at the upper right side of the chart.  It is resting on the 16000 price level for the DJIA (vertical axis).  This is saying that the model’s best guess for today's DJIA price, based on the current level of the DJIA dividends, is 16000, about 1100 points above today’s close.  However, there is that matter of the fact that the model has been able to predict 88% of annual DJIA price growth, not 100%.  Thus, there is another statistical term we’ll introduce called the standard error.  The standard error for the model is 1500 DJIA points.  So, the last word is that based on today’s DJIA dividend, the model says its best guess is that the DJIA should be trading within a range between 14500 and 17500.  At 14900, it is doing so, but there is more.  Using year-end dividend estimates, the midpoint moves up to 16500 and the range from 15000 to 17500.
In addition to regression models such as the one just described, we use dividend discount models, dividend and earnings models, and discounted cash flow models. It is essential to have some sort of model that can offer a general idea of the valuation of a company versus it selling price.  In this regard, we believe for most stocks the market prices a stock correctly about one of every two to three years.  Bargains are always available if you look hard enough, but you have to have the tools to dig and know what you are looking for.

As I said earlier, I will write additional Dividend ABC's in the months ahead and go into deeper discussions of  how our models work and some decisions we are making.  Let me close with an analysis of a stock we have been buying lately in our Capital Builder Investment Strategy.

Industrial Gas company Praxair (PX) trades on the New York Stock Exchange.  At first PX will not look attractive to many of you as a dividend stock because its current yield in only 2%. However, let me show you the results of our Dividend Valuation model. The chart below is a copy of the model for PX.



The red line is the annual price of PX over the last 17 years.  The blue bars are what our valuation model calculates as PX's value for each year.  The model we are using for PX has multiple variables, but the main driver is dividend growth. It is clear that our model has nailed PX for many years.  Indeed, the R2 is .97, which means that PX's annual dividend growth has been able to explain 97% of the annual price movement of the company.
The blue checkered bar out to the far right is our model's estimate of the value of PX one year from today. That figure is about 13% higher than PX's current selling price of $118. Most dividend investors would not give PX a second thought. But they would if they knew that its price growth is so highly correlated to its dividend growth.  In effect, PX's dividend growth over the last 17 years has been so efficiently priced into the stock that is has become as close to a cash dividend as any stock I know of. And with PX, the total rate of return has been much higher during this time than the average stock.

We love dividends, but we see our ultimate job as producing as high a total return as we can and still have a portfolio with less volatility than the overall market. Having said that, all three of our investment management strategies have higher than average dividend yields.  In two of them, Cornerstone and Capital Builder, we own many companies with 2.0%-2.5% yields but with double digit dividend growth.  We have witnessed dividend growth turning into price growth for a long time.  We see no reason it should stop.

Thank you for your readership. We appreciate you and your comments.  If you remember all the way back in Part I, we explained that we learned dividend investing from a client. So it goes without saying that we have a ready ear to your experiences and understandings.


Blessings!