The concept:
"The dividend blend is a cash management concept. You never sell any shares.
The dividend blend consists of a high yielding portfolio with little dividend growth combined with a portfolio with a lower initial yield that grows its dividend amount rapidly. Excess cash generated by the high yielding portfolio covers shortfalls during the middle years. The fast growing portfolio eventually takes over, providing a continuing income stream.
I assume that the high yielding portfolio does not grow fast enough to match inflation."
A portfolio: (one of many, it turns out!)
"I used DVY for my high growth portfolio. It currently has a dividend yield of 4.5% and a dividend growth rate in excess of 7.5%.
I used BAC (Bank of America) preferred shares (callable in 5 years, non cumulative) for my high yield portfolio. It has an 8.2% dividend yield (constant).
I assumed a TIPS interest rate of 2% (real) for cash management. This was necessary for purposes of analysis. In actuality, you would put excess cash back into your investment portfolios. I started with an initial TIPS balance of zero."
Comments:
The concept is that you withdraw, and spend, no more then the yield thrown off by your portfolio. With suitable choices of assets for that portfolio, the amount of yield thrown off will grow over time, at least by the rate of inflation.
Withdrawals, based on this strategy, have a near 100% probability of success, regardless of the yield or rate of withdrawal. If success is measured by having your withdrawals match inflation, then failure will be that the yield growth doesn't match it. If success is based on portfolio survivability, then, by never selling shares, the only way that the value of your portfolio would ever go to zero is if the per share/bond/fund price went to zero. The only way the yield goes to zero is if all dividends went away and all bonds defaulted.
Simply put, the probability of success of this strategy doesn't diminish with the amount of time (10, 20, 30, 40 years) that your withdrawal is required.
Orygunduck's strategy, as I understand it, matches the rate of withdrawal to the yield of the portfolio.
JWRs strategy withdraws less then the yield generated by the portfolio and 'saves' the excess yield in TIPs.
My strategy withdraws less then the yield generated by my portfolio and reinvests it back into my high yield assets.
In all three strategys, shares need never be sold, so capital/principal/seed corn price behavior doesn't factor into withdrawal considerations.
JWRs strategy mixes a high yield, low/no/negative yield growth asset (typically, debt related) with a low yield (never a no yield, and negative yields are not defined!), high yield growth asset (typically, equity related.). In this strategy (based, as I understand it, on Josh Peters work), the 'return' one expects from a portfolio is the sum of the percentage yield plus the percentage yield growth. That is, it is Dividend Discount Model based.
Some considerations:
First, there is nothing to suggest that a high yield, high yield growth equity assets (think ALD or ACAS type stuff!) doesn't have a place in JWRs, or any dividend based, portfolio.
Next, there is nothing to suggest that the dividend based portfolio be limited to 2 assets. What IS important is the overall weighted yield of the portfolio and the overall weighted yield growth rate.
Next, focusing on whether BAC preferred stock is a good, or bad, asset to hold is argueing how many angels can fit on a pin head!
Next, the strategy doesn't depend on whether the individual assets are equity or debt. The ONLY consideration is the yield and the yield growth rate. So, for example, if one considers real estate rents to be the equivalent of yield, then real estate ownership type assets would have a place in one's retirement portfolio. So, traditional 'asset allocation' takes on a whole, entirely different, meaning, regarding withdrawal strategies. (more on this in a minute)
Next, for the most part, yields and yield growth rates are fairly stable and predictable. That means that all of the statistics and probabilities associated with share price behavior (think efficient markets and frontiers, FF3F, P/E10, Monte Carlo, and so forth) are not necessary for future performance expectations with yield based strategies. Especially as it relates to portfolio withdrawals. In general, if yield shows up in your money market account, you can spend it. Whether you spend it on food and shelter, or portfolio reinvestment, is a decision you can make on the fly, each and every month/quarter/year.
Finally (it's been one minute!), traditional withdrawal studies (the likes of Trinity) all come to the conclusion that higher rates of withdrawal lasting longer periods of time with higher probabilities of success occur whenever one's equity allocation goes up, and the debt allocation goes down correspondingly. This is due to typical share price behavior, for these portfolio liquidation strategys. And concepts, like P/E10 come into play.
Yield based strategys, for the most part, depend on the size of the yield and the yield growth rate, so it's possible for an all debt allocation (think junk, for example) to support higher rates of withdrawal that last longer periods of time. This is the all debt portfolio (VWEHX) compared to the all growth equity portfolio argument.
So, for mind candy, think of a discussion with a diehard indexer, where you take the position that an all VWEHX portfolio, today, will support a real inflation adjusted 6% rate of withdrawal with a 100% probability of success to last 30 years. Ask him to define any indexed portfolio allocation that has these characteristics!
