Dividend Growth Projections

The traditional retirement strategy for stocks is to live off dividends alone. This is an excellent approach. Typically, dividend payments continue far into the future. Typically, dividends keep up with inflation. Even when dividends fail to keep up with inflation, any loss in buying power is gradual.

By living off dividends alone, a retiree avoids the most serious danger to portfolio survival: being forced to sell stocks at low prices.

Dividend investors are well aware that dividend payments grow with time. I was surprised when I realized I had not published a set of numbers. I had looked at what happens to the overall portfolio. I had glanced at dividend yields. But I had not studied the dividends themselves.

These are the numbers for the S&P500 index.

The Procedure

I took the real dividend amounts of the S&P500 index in Professor Robert Shiller’s database. I thinned the data to only January values for each year.

Professor Shiller’s Web Site

I used an Excel spreadsheet to calculate the ratio of the dividend amount at a later time (of 1, 5, 10, 15, 20, 25, 30, 35 and 40 years) to the initial amount to construct gain multipliers. I converted these ratios (or gain multipliers) into percentage increases. I also converted them into annualized percentage increases.

I used Excel’s plotting function to calculate regression equations (i.e., linear, straight-line curve fits) of the dividend amount at Year 10 and at Year 20 versus the percentage earnings yield 100E10/P. I used data from 1921-1980.

[P/E10 is the current (real) price of the S&P500 index divided by the average of the previous ten years of (real) earnings. The percentage earnings yield 100E10/P is 100/[P/E10]. Professor Shiller includes P/E10 in his tables.]

[Information based on 1881-1970 showed the same P/E10 anomaly in the early years as we have seen in other calculations. I have written several notes regarding it, including a reference to Gummy’s (Professor Peter Ponzo’s) findings. This P/E10 related anomaly occurs around 1900-1910.]

The Equations

Here is the equation for the total percentage increase in real dividends at Year 10.

y = 6.0722x-30.254 plus 70% and minus 60%.
R-squared = 0.2998.

Here is the equation for the total percentage increase in real dividends at Year 20.

y = 5.9354x-12.509 plus 130% and minus 50%.
R-squared = 0.1676.

Here is the equation for the annualized percentage increase in real dividends at Year 10.

y = 0.4715x-2.4195 plus and minus 5%.
R-squared = 0.263.

Here is the equation for the annualized percentage increase in real dividends at Year 20.

y = 0.2332x-0.5709 plus 3% and minus 2%.
R-squared = 0.2136.

Dividends are likely to grow by Year 10 if the earnings yield is 5% or more (i.e., P/E10<20).

I have placed tables and graphs into my Yahoo Briefcase in the Dividends folder.

Yahoo Briefcase

Analysis

Theoretically, dividend payments should be independent of price. Dividends come out of earnings and using E10, a decade of smoothed earnings, should reduce the influence of most uncertainties. Earnings, especially smoothed earnings, grow steadily, correlated strongly with the GDP (Gross Domestic Product).

The payout ratio (dividends = payout ratio*earnings) has varied, generally decreasing since the 1950s. It has not varied in a manner that obviously tracks the percentage earnings yield.

The fact is that dividend payments grow more when valuations are more favorable (higher dividend yields, lower P/E10, higher earnings yield 100E10/P). Dividend payments fall behind when valuations are high.

There are, at least, two causes. There were dividend cuts related to the 1930s. There was a failure of dividend payments to keep up with inflation related to the 1960s. NOTE: remember that these are related to Years 10 and 20.

Observe that the total dividend growth is higher at Year 20 than at Year 10. The slope of the Year 20 graph changes as a result of converting the growth into annualized values.

Notice that there have been times when the Year 10 dividend amount has fallen by 40%. There have been times when the Year 20 dividend amount has fallen by 20% to 25%. These occurred in the 1930s. The real-dollar declines of the 1960s were only half as bad.

Conclusions

The news is good EXCEPT during times of high valuations.

Unfortunately, this is a time of high valuations.

We know what to watch for: dividend cuts and roaring inflation. Fortunately, earnings are strong. Inflation, except for the momentary gasoline price spike, is subdued.

Have fun.

John Walter Russell
January 10, 2006