Dividend Growth Basics

Research on Safe Withdrawal Rates

Moderator: hocus2004

Post Reply
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Dividend Growth Basics

Post by JWR1945 »

Dividend growth strategies are based upon the Dividend Discount Model presented in the form of the Gordon Equation. The Gordon Equation calculates a discount rate R from a steadily growing income stream. If suitable reinvestments can be found for the dividends, the R is the overall return of a stock.

The Gordon Equation is: R = initial dividend yield + the growth rate of the dividends.

Suitable reinvestments are those that maintain the same rate of return as the stock when it was originally purchased. In short, the Gordon Equation assumes that the stock's dividend yield and dividend growth rate remain steady on into the distant future. All dividends are reinvested to purchase additional shares.

Notice that it is neither the dividend yield nor the dividend growth rate that is important. It is their sum.

Add in some restrictions to assure quality and you have the foundation of a dividend growth strategy. All of this applies during the years of accumulation as well as during retirement. In addition, focusing on income streams helps psychologically. Steadily rising dividend income mitigates the unpleasantness often associated with short-term price fluctuations.

By focusing on income streams, a dividend growth strategy avoids the problem of selling shares during retirement. An income stream from dividends is inherently safe as long as the company is sound. Restrictions on quality require selling and reinvesting elsewhere before a company's outlook deteriorates too severely. There can be losses. For the most part, a reliable dividend stream is a Safe Withdrawal Rate with an indefinite, but very long lifetime.

A retiree can choose to withdraw all or part of his dividend income stream. The best approach appears to be to reinvest at least a small portion to cover the occasional losses that are bound to occur.

A natural question arises as to the rate of return that applies to the dividends that are reinvested. A good guess might be to adjust the Gordon Equation by scaling the dividend yield. That is, if one reinvests 25% of his dividends, he would add a quarter of his dividend yield to the dividend growth rate. It is a good guess. But it is wrong.

The problem is that the growth rate term is no longer constant with respect to the initial withdrawal. Suppose, for the sake of illustration, that dividends double every year, starting at $1. The growth rate is 100% per year. The dividend amounts are $1, $2, $4, $8 and so forth. If you withdraw 75 cents each year, reinvesting 25%, the amounts to be reinvested are $0.25, $1.25, $3.25, $7.25 and so on. This is not a steady rate and the Gordon Equation does not apply.

But what about our guess? For the Gordon Model to apply, the reinvested amounts would have had to be $0.25, $0.50, $1.00, $2.00 and so on to maintain a growth rate of 100% per year. The actual amounts reinvested are higher, especially at first. We end up buying a lot more shares than expected. Scaling the dividend yield produces a low estimate of the return from reinvested dividends.

I think that this subtle mathematical point has helped to cloud discussions of dividend growth strategies. If it is natural to think in terms of a stock's total return with reinvested dividends, it is natural to scale incorrectly when reinvesting only a portion.

Adding to the confusion is the insistence of those advocating dividend growth strategies to refer everything back to the initial investment. If there were no withdrawals, there would be only one number to apply. It would work in all cases and that would be that. When there are withdrawals, there is a different number for each withdrawal amount. The growth rate term has been affected. The corrections are pleasant. The income stream increases faster than anticipated (when scaling only the dividend yield).

What about risk? Dividends are much more amenable to analysis and projection than prices and total return. Notice that their expected return is highly predictable, especially in the short-term. This is in radical contrast to price behavior, which is almost entirely unpredictable in the short-term but reasonably predictable in the long-term. Dividend yields fluctuate considerably. Dividend amounts do not. It still makes sense to buy only at a good price. But retirees are under no pressure to sell when relying on dividend income. They may choose to do so when prices are ridiculously high.

I cannot over-emphasize this latter distinction. It is important. Some things are predictable in the short-term. Other things are predictable in the long-term. Safe Withdrawal Rate analysis is only beginning to take advantage of such distinctions.

Have fun.

John R.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"Some things are predictable in the short-term. Other things are predictable in the long-term. Safe Withdrawal Rate analysis is only beginning to take advantage of such distinctions."

I know that you already appreciate the point I am going to make here, JWR1945. I am making it for the benefit of any listening in who are unsure at this point of the power of an analytically valid SWR tool to inform one's investing stratagies.

Exceedingly loose standards apply to most investment analyses that are reported on in the general media. You have in recent posts put forward several striking examples of elementary but common analytical errors. The error discussed in the "Yahoo Finance Quiz" thread at the Early Retirement Forum was one. The creators of that Quiz "forgot" to include the effect of dividends when calculating the long-term gain from stocks. That's the sort of thing that people do when they want to put a negative "spin" on stock investing. When they want to put out a positive "spin," they do something like "forgetting" that changes in valuation levels have a critical effect on the long-term return you can expect to obtain.

My sense is that there are two primary reasons why standards are so loose. One is that much of the media relies on people selling stocks for a living as sources of information. Those who sell stocks for a living have an obvious bias. My guess is that the reason why many have come to believe that there is some sort of research somewhere that indicates that it is impossible to engage in long-term timing is that people in the stock-selling industry often "forget" to note that the studies that show timing not to work all examine the issue of short-term timing, not long-term timing.

The other big factor is that people get emotionally attached to their investment choices. Once you have put most of your life savings into one investment class, you want that investment class to do well. So without even being aware that you are doing it, you filter out any information you come across that suggests that your reasons for investing the way you did were not as sound as you once thought.

One the one hand, people talk about stock investing as if it were a rational enterprise. People are always making reference to studies and spreadsheets and numbers as if these things informed their stratagies. But many of the studies and spreadsheets and numbers are based on poorly thought-out assumptions or in some cases outright contradictions. Big buildings are erected on exceedingly weak foundations. So the end products appear to the newcomer to be serious and trustworthy and important. At the core, however, is some silly idea that dividends don't matter or that valuation does not matter or whatever.

The point of SWR analysis is to provide objectivity. Every retiree needs to decide on a take-out number, and it is a hard thing to do when much of your money is invested in a volatile investment class like stocks. The natural thing to do is to look at how stocks have performed in the past to assess how they are likely to perform in the future. SWR serves an extremely important function. I completely reject claims that SWR analysis should be ditched or that SWRs should be determined by guesswork instead of by reference to the data or whatever. I firmly believe that the SWR is whatever the historical data says it is.

The comment of yours that I quote above is the sort of insight that can be developed only by examining the historical data objectively. When all you are doing is guessing or engaging in spin, you cannot come to reasonably conclude that "some things are predictable in the short term and other things are predictable in the long term." How would you know when all of your calculations are corrupted by the spin or guesswork you engaged in to produce them?

The sort of insight that you put forward above is very important, in my view. Investors of all kinds (not just retirees) need to know what sorts of things are predictable in the short-term and which sorts of things are predictable in the long term. It's amazing to me that, despite all the money and energy that is devoted to investment analysis, few people know these sorts of things today. The reason why they don't know is not that they do not want to know or do not see the value in knowing. It is that most of the sources to which they turn for insights apply such loose standards when analyzing investment options. So-called experts define their terms in so many different ways to suit so many different agendas that it becomes impossible to gain any legitimate insights into what is going on. There is so much spin going on that people can't even get straight reports on what the long-term return on stocks is.

SWR analysis is not just about determinng what take-out rate is safe. The insights that this analytical process provides reach in all sorts of directions. The distnguishing factor of this analytical tool is that it is objective. I will never credit anyone who employs guessing games or rule-of-thumb type analyses as being engaged in legitimate SWR analysis. Those who are just taking guesses or just engaging in spin should call their studies something other than SWR studies. The entire point of SWR analysis is to cut through the spin and say WHAT IS, not what the researcher wishes were so.

The SWR is whatever the historical data says it is. That's my take. The amazing work you have done over the course of the past two years shows what can be accomplished when researchers are willing to discipline themselves enough to follow the data and not their own personal wish lists when reporting to the public what the data reveals. It is my belief that, so long as we stay on the path we are now on, there will continue to be insights to come in the future as powerful as many of those we have discovered in the first 26 months of these discussions.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

Safe Withdrawal Rates are determined primarily by short-term and middle-term effects. This shows up repeatedly in our investigations.

We learned from early investigations that portfolio survival is almost always established by the eleventh year. Either the portfolio has grown enough to assure safety or it is already in trouble.

The New Tool predicts (30-year) portfolio survival most accurately in the fourteenth year.

My investigation into The Stability of Estimates Based on Earnings Yield shows that that the effects of valuation are determined almost immediately. A single decade of data provides useful information. The only qualifier is standard for any extrapolation. Estimates are best when looking at valuations within the range of the input data. They become less and less reliable as one goes farther and farther away from the range of inputs.

In contrast, the conventional methodology requires a very large number of years before making useful predictions. It has to. It must assure that it includes some data at valuations relevant to what is being examined. It can break down, as it has, when current conditions fall outside of the historical range.

Strategies that focus on dividend income are inherently most accurate in the short-term. The big issues are whether dividends will continue and whether they will rise fast enough to compensate for inflation. Near-term dividend projections are easily subjected to analysis. Are earnings (averaged over several years) sufficient to cover dividends? Do the fundamentals hold up? What is management's record concerning dividends and dividend increases?

Companies with high dividend yields and high dividend growth have performed almost exactly [within 0.5%] as well as other companies in providing market returns. They have performed much better on a risk adjusted basis, typically twice as well as the market. One must insist upon quality and avoid reaching for the highest yields available. Once one is assured of quality, higher yields most often translate into higher total returns. That makes sense because it is indicative of a short-term price dip as opposed to a long-term problem.

Have fun.

John R.
unclemick
*** Veteran
Posts: 231
Joined: Sat Jun 12, 2004 4:00 am
Location: LA till Katrina, now MO

Post by unclemick »

1966, Geraldine Weiss started Investment Quality Trends - wrote her first book in 1988, Dividends Don't Lie. Basically follows 350 'blue chip' stocks and buys when dividends indicate a drop into the buy range. When they followed her formula on the 'old' Quicken, I used to watch the list. Her other metrics were very Ben Graham - as to P/E, debt/equity, div growth, etc. But she's not a find a good stock to hold forever type. A library copy of her book might be worth a scan. Also has a newer book - she hasn't deviated much, so whatever is cheaper - ie free from the library - heh,heh.

1995, The Dividend Connection (her second book) was probably exactly the wrong time to talk div.s - heh, heh - given the start of the stock bubble.
unclemick
*** Veteran
Posts: 231
Joined: Sat Jun 12, 2004 4:00 am
Location: LA till Katrina, now MO

Post by unclemick »

Hint. ? Geraldine uses a lot of Ben Graham type metrics to define 'her family of blue chips' (350?) and Mergent's uses a single ten years of rising dividends (332 in 1994, 284 in 2003). Both are fishing in small ponds? ? Thoughts on identifying durable quality - the rising dividend stock capable of lasting a thirty year ER?
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

It is likely that I will look at Geraldine Weiss's books later. My impression is that she searches for the best stocks to buy on a total return basis during accumulation.

What I consider different and unique about dividend strategies and dividend growth strategies during retirement is how they alter Safe Withdrawal Rates. Remember that it is the volatility of stocks that reduced the worst case Historical Surviving Withdrawal Rates down to 3.9% (before the bubble). That is more than 2.5% below the 6.5% to 7.0% long-term (real) total return of stocks.

If carefully selected dividend producing stocks can continue to come within 0.5% of the (nominal) total return of stocks at only one-half of their volatility, which is what they have done, their Historical Surviving Withdrawal Rates of the past (pre-bubble) might have been a reliable 5.5% to 6.0% or even higher.

At current valuations stock prices are likely to fall about 3% per year for the next decade (annualized). This reduces Safe Withdrawal Rates when stocks must be sold.

Current Safe Withdrawal Rates of high stock portfolios are lower than inflation-matched cash. That is, an 80% stock portfolio has a Safe Withdrawal Rate close to 2.5% but TIPS and/or ibonds at an interest rate of zero percent would allow 3.33% to be withdrawn annually. [Both Safe Withdrawal Rates are for 30-year portfolio lifetimes.]

If you can get dividends that start out at a 2.5% yield and that match inflation, you have equaled the Safe Withdrawal Rate of an 80% stock portfolio while extending its lifetime from 30 years on into the distant future. You would never have to sell any of your dividend producing stocks. You have also equaled TIPS that return 2.5% each year with reinvestments (assumed continually available) without the tax burden caused by increases in TIPS principal.

Have fun.

John R.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

unclemick wrote:1966, Geraldine Weiss started Investment Quality Trends - wrote her first book in 1988, Dividends Don't Lie. Basically follows 350 'blue chip' stocks and buys when dividends indicate a drop into the buy range. When they followed her formula on the 'old' Quicken, I used to watch the list. Her other metrics were very Ben Graham - as to P/E, debt/equity, div growth, etc. But she's not a find a good stock to hold forever type. A library copy of her book might be worth a scan. Also has a newer book - she hasn't deviated much, so whatever is cheaper - ie free from the library - heh,heh.

1995, The Dividend Connection (her second book) was probably exactly the wrong time to talk div.s - heh, heh - given the start of the stock bubble.
I have just checked Investment Quality Trends in the July issue of the Hulbert Financial Digest. It has been a consistent winner, ranking among the very best on a risk adjusted basis (at 5, 10 and 15 years) with annualized gains of 12.0% to 13.5%.

This includes a gain of 12.0% over the last 5 years when the market as a whole (Wilshire 5000) returned -1.1% per year.

Mark Hulbert has noted that only a few Newsletters have beaten buy-and-hold over the long-term. He is of the opinion that this is a bull-market phenomenon. When the market rises, Newsletters do not offer a great advantage.

Looking forward, however, he expects Newsletters to outperform buy-and-hold on a broad basis. That is because he expects subdued returns for the next decade. Over the last 5 years, for example, 83% of all Newsletters outperformed the market.

Have fun.

John R.
Post Reply