The magic dividend formula

What if I told you that there was a way to glimpse into the future? To somewhat accurately estimate the future returns on your investments? Well, with dividend investing  there is a way. Its called the Gordon Growth Model or as I like to call it, the Magic Dividend Formula.

You may be surprised to hear me talk about forecasting given that I’ve ranted before on the folly of forecasting. My rant had more to do with overall forecasting methods rather than the principle of it. After all, when we buy any investment we have an expectation of the future cash flows that the investment gives off. Investing is inherently about the future. The type of forecasting that I dislike uses earnings, relative prices, and macro economic themes all of which are subject to way too many unknowns. But there is a better way – forecasting based on dividends and dividend growth. Dividends are known at the time of investment so no forecasting necessary and dividend growth can be conservatively estimated using long term historical growth rates. That’s it. One unknown variable. Lets dive into the details of the Gordon Growth Model.

The Magic Dividend Formula

The simplest form of the Gordon Growth Model says that a dividend stock’s return is defined by;

Expected Return = Dividend Yield + Dividend Growth Rate

That’s it. The dividend yield is known so all you need to do is to estimate the dividend growth rate. And if one uses long term dividend growth rates the model becomes quite accurate. I’ve hinted at this model before. The analysis in my post on how dividends account for 80-100% of historical stock market returns is based on this model. Using the S&P500 historical data going back to 1871 is pretty telling. In 1871 the dividend yield on the index was 5.86%(can you imagine?). As of the end of 2009 the dividend growth rate on the index has been 3.28% annually. Thus the model would say that the total returns on the index during that time would be 9.13% (5.86% + 3.28%). The actual return of the index during that time period, 1871-2009, was 8.53%. Pretty close given the 138 year time frame. The difference in the two numbers is due to the change in valuation during that time. More on that in a minute.

Let use a more recent example from one of my favorite asset classes, MLPs. At the inception of the Alerian MLP index, it was yielding 8.98% in 1995. Over the next 15 years, through 2010,  the dividends from the index grew at a 7.78% annual rate. Given this, one would expect a total return over that time period of 16.76%. The actual total return on the index was 16.78%. Again, pretty darn accurate. The difference of .02% annually was due to valuation changes. Turns out you can add the change in valuation to the formula to arrive at a more complete version of the model which says

Expected Return = Dividend Yield + Dividend Growth Rate + Annual Change in Valuation

All in all pretty simple. No complex DCF analysis. No P/E ratios. No forecasting of interest rates. No ‘macro themes’ to worry about. Just tried and true long term investment principles. Going back to one of my favorite investment mantras from Ben Graham, ‘in the short term the market is a voting machine, in the long term it is a weighing machine’.

Using the Magic Dividend Formula

There are many ways to use the Gordon Growth Model. I use it to estimate a range of forward returns on individual dividend paying stocks for a holding period of at least five years. I use a five year holding period estimate because that is usually enough to ignore any changes in valuation. Then I estimate a dividend growth rate range for the next five years. My base assumption on growth rate is the long term historical growth rate of dividends for the overall market which as I stated above is about 3%. This number is remarkably consistent over history. The reason? It basically follows the long term historical growth rate of the economy. After all, how can any business grow dividends forever at a faster rate than overall economic growth. My upper assumption on dividend growth rates is usually an average of the last 3-5yr dividend growth rate for the company. Let me use Enterprise Products Partners (EPD) as an example. Its current dividend yield is 5.5%. Its dividend growth rate over the last several years is about 5%. Long term dividend growth rates for all stocks is 3%. Thus, my range of expected return over the next 5 years is 8.5% to 10.5%.

One note on stocks that are currently in your investment portfolio. The same model and analysis applies except that one should substitute yield on cost for current yield in the calculation of expected returns. Again, for EPD, a stock I have held for over 5 years my yield on cost is about 10%. My personal expected range of returns for the next 5 years is then 13% to 15%.

The last part of my process is to apply a margin of safety to my expected return calculations and then to compare them to my required rate of return. My long term goal is to compound my investments at an annual return of 10%. Also, I realize that while I’m using a model with only one unknown and that I’ve been conservative in estimating that unknown variable there are many things that can go wrong especially with the stock of a single company. Thus I apply a margin of safety to my return calculations. My 10% annual return then becomes 13% with a required margin of safety of 25%.

Summary

The Gordon Growth Model is a truly the Magic Dividend Formula that allows an investor to evaluate investments based on dividends alone. I’ve shown one way to use the model to estimate future returns. There are several other uses of the model which I’ll explore in later posts.