# How I Estimate Future Dividend Raises

|
Includes:
by: Robert Allan Schwartz
Summary

I present two different ways of estimating next year's dividend raises.

One requires making two reasonable assumptions.

The other repeats what was done last year.

There are probably many ways that an investor could use to estimate future dividend raises. I could estimate future revenues, profits, or cash flow. I could estimate future payout ratios. I could estimate what the Board of Directors might do. All such estimates are guesses about the future.

I choose to do it differently. I extrapolate from what a company has done in the past.

Now, we all know that "investing is about the future". And, we all know that "past performance is no guarantee of future results". But according to Mark Twain, "History doesn't repeat itself, but it does rhyme." So, perhaps it might be useful to look at a company's dividend history.

### Over what time interval should we look?

Wikipedia lists the most recent three recessions as occurring in 1990, 2001, and 2007, and I want to see how a company responds to recessions, so I will look from 1990 to 2016 (the last full year for which I have dividend histories), an interval of 26 years.

### Which ways of estimating are not sufficient?

It's not sufficient to simply compute the mean (i.e. arithmetic average) of the dividend raises during that interval. A company that raises its dividend by 10% each even year and lowers its dividend by 10% each odd year, would have a mean of 0%, which is not useful.

It's not sufficient to simply compute the range (i.e. difference between the smallest raise and the largest raise) of the dividend raises during that interval. A company with a history of a 1% raise, eight years of 5% raises, and a 10% raise (in any order), is different from a company with a history of a 1% raise, a 2% raise, a 3% raise, …, a 9% raise, and a 10% raise (in any order). They have the same ranges, but the former is far more predictable.

I do not know if the distribution of dividend raises is a normal distribution, but I suspect that it is, and I proceed on the assumption that it is. If that offends you, you can stop reading here, and I won't blame you a bit.

### How do I estimate future dividend raises?

My first way is by using the standard deviation, which measures "the amount of variation or dispersion of a set of data values. A low standard deviation indicates that the data points tend to be close to the mean (also called the expected value) of the set, while a high standard deviation indicates that the data points are spread out over a wider range of values."

I assume that a company is likely to continue to do what it has done repeatedly in the past, which means that the smaller the standard deviation (i.e. spread), the more probable it is that future dividend raises will be close to the mean. If that offends you, you can stop reading here, and I won't blame you a bit.

### How close to the mean?

The 68-95-99.7 rule says that:

1. 68% of the data points are within 1 standard deviation of the mean (i.e. (mean – 1 standard deviation) to (mean + 1 standard deviation));
2. 95% of the data points are within 2 standard deviations of the mean (i.e. (mean – 2 standard deviations) to (mean + 2 standard deviations));
3. 99.7% of the data points are within 3 standard deviations of the mean (i.e. (mean – 3 standard deviations) to (mean + 3 standard deviations)).

Let's look at some examples.

From 1990 to 2016, Bank of America (NYSE:BAC) has a mean of 15.022% and a standard deviation of 44.579. That means that:

1. 68% of the dividend raises are between -29.557% and 59.601%;
2. 95% of the dividend raises are between -75.136% and 104.18%;
3. 99.7% of the dividend raises are between -118.715% and 148.759%.

Clearly, a negative raise means a dividend cut, and indeed, BAC cut its dividend in 2008 (by -6.667%) and 2009 (by -98.214%).

That's a very large standard deviation. I can't begin to estimate what BAC might do next with its dividend.

24 companies had higher standard deviations than BAC did over that interval. The highest standard deviation was 671.960 for Superior Industries International (NYSE:SUP).

From 1990 to 2016, Southern Company (NYSE:SO) had a mean of 2.861% and a standard deviation of 1.401. That means that:

1. 68% of the dividend raises are between 1.46% and 4.262%;
2. 95% of the dividend raises are between 0.059% and 5.663%;
3. 99.7% of the dividend raises are between -1.342% and 7.064%.

There were five freezes during that interval, but no dividend cuts.

That's a very small standard deviation. I estimate the next dividend raise will be close to 2.861%.

I conjecture that utilities will tend to have small standard deviations, but I have not yet done the research.

### What have you done for me lately?

There is some evidence that dividend growth is slowing down, which means that metrics over a long time interval might be misleading or less accurate, so my second way is by using a company's most recent dividend raise as an estimate of its next dividend raise.

### How do I estimate my total dividend income for 2017?

Column A is the ticker.

Column B is the number of shares.

Column C is the total dividend amount paid per-share in 2016.

Column D is column B * column C.

Column E is the standard deviation from 1990 to 2016.

Column F is the mean – 2 * the standard deviation, the low end of the 95% rule.

Column G is the estimated total dividend income for 2017 using column F as the raise percentage.

Column H is the mean – 1 * the standard deviation, the low end of the 68% rule.

Column I is the estimated total dividend income for 2017 using column H as the raise percentage.

Column J is the mean from 1990 to 2016.

Column K is the estimated total dividend income for 2017 using column J as the raise percentage.

Column L is the mean + 1 * the standard deviation, the high end of the 68% rule.

Column M is the estimated total dividend income for 2017 using column L as the raise percentage.

Column N is the mean + 2 * the standard deviation, the high end of the 95% rule.

Column O is the estimated total dividend income for 2017 using column N as the raise percentage.

Column P is the 2016 raise percentage.

Column Q is the estimated total dividend income for 2017 using column P as the raise percentage.

It is 95% likely that the overall portfolio dividend income will go up between 1.88% and 19.71%.

It is 68% likely that the overall portfolio dividend income will go up between 6.34% and 15.25%.

Or, if the portfolio simply repeats what it did in 2016, the overall portfolio dividend income will go up by 4.62%.