Building Blocks

I have posted a lot of articles with lots of numbers. Those numbers are meant to help, to provide assistance, to supply insights. They tell you can do. The let you know what to look for. They do not tell you what you must do.

Lots of details are not found in numbers. Lots of details are unique to your own situation. This article will help you bring everything together.

Portfolios and Safe Withdrawal Rates

Baseline Portfolios

A 100% TIPS portfolio makes an excellent baseline inside of a tax sheltered account. I-Bonds do well in taxable accounts.

A 100% TIPS portfolio with a 2% (real) interest rate produces an income stream of 4.0% of the original balance (plus inflation) for 35 years. This comes with a true, 100% Government guarantee. This is more than competitive to the 30-year survivability claimed by early researchers using fixed allocations of stocks and commercial paper. In view of what we have learned about safety and valuations, our baseline is safe. But the traditional approach is exceedingly risky.

The disadvantage of using 100% TIPS is that you will run out of money. You are withdrawing principal as well as interest.

There are some practical details. Building a TIPS ladder gets around most of them.

To build a ladder, you invest your money over a range of maturities, possibly in one-year increments. When any bond matures (in this case, the bonds are TIPS), you make your withdrawal and reinvest whatever remains at the longest maturity of the ladder. For example, with a 10-year ladder, you would buy new 10-year TIPS to replace older TIPS that have matured. Each time, you purchase at the longest maturity of the ladder. This usually gives you the highest interest rate available.

An alternative baseline is to purchase high dividend stocks from high quality companies and never sell. DVY was an early choice. It is an Exchange Traded Fund (ETF). It currently yields 2.90%. There are now many choices among ETFs and low cost index funds that focus on dividends.

This alternative is attractive because the income stream should last far into the indefinite future. It can be a good idea to reinvest a little bit, especially in the early years, to cover occasional dividend cuts and bankruptcies.

This alternative is highly likely to keep up with inflation. It is likely to do even better. But this is not guaranteed. Judging from the S&P500 index during stagflation, you can always expect the (nominal) dividend amount to grow, but not necessarily as fast as inflation. After adjusting for inflation, the (real) dividend amount may dip by 10% before recovering. You can expect a full recovery within a few years.

You can mix these baseline portfolios together. TIPS provide a true guarantee, but over a limited number of years. High quality, high dividend stocks offer a continuing income stream, but at reduced rate.

Fixed Stock Allocations

The amount that you can withdraw safely from a portfolio varies tremendously with its initial valuation. Rebalancing helps when valuations are high. Rebalancing hurts considerably when starting valuations are typical or low.

When you rebalance, you should withdraw a larger amount from an investment that has increased its percentage in your portfolio. You should attempt to restore the original percentages. But costs are more important than rebalancing. Or, at least, it is doubtful that rebalancing helps when it increases costs. If your stocks have grown by more than your withdrawal amount, it is far from clear-cut that you should sell shares merely to rebalance. The advantage of rebalancing is small, even under favorable circumstances. Researchers seldom, if ever, include costs when it comes to rebalancing.

If you do not rebalance, you withdraw from individual investments according to their current percentages in your portfolio.

Central to rebalancing is the notion that you cannot measure relative value in a meaningful way. This notion is seriously flawed. Most of the research that is cited applies only to very short periods of time, typically no longer than two years. There is some evidence that short-term timing can be successful, provided that costs are very low. There is very little evidence that short-term timing can be taught successfully.

When you look at longer periods of time, of the order of 10 to 20 years, the story is different. The Gordon Equation helps. Professor Robert Shiller’s P/E10 helps.

Perhaps the most important flaw in stock market research is the mechanical treatment of time without regards to valuations. If you have ever established prices to buy and prices to sell, you are aware of the tremendous advantage of price discipline. If you have multiple stock holdings, be patient. You will be amazed at how well you will be rewarded.

High, fixed stock allocations at times of high valuations are reckless. What was claimed to be safe is no better than making a coin toss. Even cash equivalents can be expected to do better. This has happened before. TIPS and I-Bonds are almost guaranteed to do much better.

Variable Stock Allocations

Varying stock allocations in accordance with P/E10 works well. It is not critically sensitive to the precise details such as P/E10 thresholds and allocations.

These are gradual changes. I am NOT talking about getting 100% into stocks or 100% out of stocks except under the most extreme circumstances. I am NOT talking about selling stocks that you purchased last year just because P/E10 increases by one or two points (e.g., from 19 to 21). I am talking about reducing your stock allocation to low levels when valuations are sky high, as they are today, adding stocks when valuations return to normal (about one-half of today’s prices relative to 10-year earnings), and loading up on stocks when they fall to bargain levels (about one-third of today’s prices relative to 10-year earnings).

Variable Withdrawals

Two basic approaches are to withdraw a fixed percentage of a portfolio’s initial balance plus inflation or to withdraw a fixed percentage of the portfolio’s current balance.

The danger of withdrawing a fixed percentage of the INITIAL balance plus inflation is that you may run out of money. The danger of withdrawing a fixed percentage of the CURRENT balance is that your income after twenty years can be much lower than you need.

There are a variety of approaches to vary withdrawal rates. Gummy’s (retired Professor Peter Ponzo’s) Sensible Withdrawals is among the best. He sets a floor that matches inflation and withdraws a percentage of the excess when his investments do well. Overall, this is a good approach. It is natural. Most retirees cut back during bad times anyway.

There is also the approach of conventional annuities and pension plans. They supply a fixed percentage of the initial balance without adjusting for inflation. This is inherently dangerous. The recipient must take special steps to counter the erosion of inflation.

Variable withdrawal methods are disappointing. In essence, they take income from the future and move it to the present (or vice versa). They adapt to the effects of valuations after the fact. However, they do prevent bankruptcy.

To a first approximation, but only to an approximation, you can shift money around during the first decade of retirement. You can treat partial withdrawals as if you were making full withdrawals and adding what you left in as a contribution to your initial balance.

As a general rule, you will have a very good idea of how well your retirement portfolio is doing within the first eleven or twelve years. Either your portfolio will have grown dramatically or it will be clearly in danger.

Valuations

Measures of Valuation

I have found Professor Robert Shiller’s P/E10 to be the best single measure of valuation so far. It is the current (real) price of the S&P500 index level (price) divided by the average of the most recent (trailing) ten years of (real) earnings. Professor Shiller credits Benjamin Graham with the basic idea. Benjamin Graham averaged (real) earnings when he calculated the price to earnings ratios of individual companies. Professor Shiller applied this idea to the stock market as a whole, as represented by the S&P500 index.

P/E10 turns out to be useful when applied to the individual segments of the stock market. Its merit is not limited to the S&P500 itself. The S&P500 really does tell us about the market as a whole. P/E10 is helpful when looking at portfolios built from various combinations of market segments.

What I actually use is the reciprocal of P/E10 in the form of the percentage earnings yield 100E10/P or 100%/[P/E10]. Using the percentage earnings yield 100E10/P has been spectacularly successful for making predictions.

One reason is dividends. It turns out that the percentage earnings yield 100E10/P gives us better information about dividends than the dividend yield of the market. Because it (100E10/P) includes ten years of earnings, it protects us against the effects of surprise dividend cuts. It contains information about the quality of earnings and the quality of dividends.

In a recent breakthrough, I discovered that the payout ratio (dividends/earnings) based on the percentage earnings yield 100E10/P tells us a lot about the dividend growth rate. This is a bright spot in today’s market. Today’s payout ratio (based on 100E10/P) is at the lower end of its historical range. Today’s dividends, miniscule as they are, are more secure than they have been in the past. Their continued growth is more likely than they have been in the past.

Dividends and the dividend growth rate are the two essential ingredients of the Gordon Equation in its basic form. John Bogle refers to this as the Investment Return. Changes in what people will pay for an income stream make up the third element, what John Bogle calls the Speculative Return. P/E10 addresses this issue directly.

[The Gordon Equation is based on the Dividend Discount Model. The total return of stocks equals the Investment Return plus the Speculative Return. The Investment Return equals the initial dividend yield plus the (annual) dividend growth rate. The Speculative Return annualizes the effect of different prices at the beginning of a period and at the end of a period to pay for the same income stream. There are many variants, usually substituting earnings or earnings growth rates for dividends or dividend growth rates.]

History and the Gordon Equation

The Gordon Equation has been a spectacular success. History shows that it works. John Bogle presents convincing evidence in Common Sense on Mutual Funds.

History and mathematics show that the Gordon Equation cannot possibly be right. John Bogle presents convincing historical evidence in Common Sense on Mutual Funds. History shows that there is consistent long-term growth in the stock market. It is of the order of 6.5% to 7.0% (real, after adjusting for inflation). This is in the very long-term, 50 or 60 years.

Dividend yields have fluctuated much too much for there to be a consistent long-term growth rate in the stock market. This means that the mathematical structure supporting the Dividend Discount Model and the Gordon Equation breaks down. When you get to the very long-term, initial valuations (because they are bounded, both according to theory and according to history) have a smaller and smaller influence on the (annualized, percentage) Speculative Return.

The explanation is simple enough. The investments that are available after 10 or 20 years are seldom the same as were available at the start. They are either better (if the Gordon Equation suggested a lower than average return) or worse (if the Gordon Equation suggested a better than average return). Only if the initial projection was within the historical long-term range of 6.5% to 7.0% (real) would we anticipate that they might be the same. Even then, we would not be sure. They could be the same, but not necessarily so.

Taking Advantage of Valuations

If we visit Professor Shiller’s web site, download his S&P500 data and look at his plot of the price to earnings ratio (which is actually P/E10, not the single year P/E ratio), we can see the historical story about the Speculative Return. We need to make predictions. This picture helps us estimate how likely our predictions will come true. Demographics is an important reason, perhaps the most important reason, for the expanding (P/E10) multiples leading to the bubble. We can look at demographics as a reason for the continuing high (P/E10) multiples. Today’s multiples are very close to those just before the Great Depression (and higher than those in the worst years for starting retirement, the mid-1960s). Stagnant nominal stock prices and robust earnings growth coupled with a little inflation can bring valuations into line. Investor impatience from disappointing returns could erode the stock market’s popularity. This would drag prices down. A recession, even a mild recession, could cause a precipitous drop.

If we allow ourselves ten years, we can expect to see reasonable valuations again. If we allow ourselves 15 to 20 years, we are almost certain to see outstanding bargains.

Putting It All Together

We have two baselines and a lot of information about how valuations affect Safe Withdrawal Rates. More recently, we have gained new insights into dividend growth in terms of today’s payout ratio. The building blocks are in place.

Start with a baseline portfolio that meets your needs. This can be all-TIPS, all-dividends or a combination.

Build on this portfolio by taking advantage of intermediate-term timing. One approach would be to modify the switching of stock allocations.

In my original investigations, I developed a single algorithm. I added to and reduced stock allocations regardless of how well the portfolio was doing. Modify this by converting (at least, a portion of the portfolio) to a dividend-based strategy as soon as yields among high quality companies are high enough. Lock them in. Never sell. Harvest their income stream. You can expect dividends to grow faster than inflation. Today’s low payout ratio (in terms of the percentage earnings yield 100E10/P) shows that this is likely.

There are other factors. If you can get better returns than the S&P500, you are likely to do better when making withdrawals. If you can get more consistent returns than the S&P500, you are likely to be able to increase your withdrawal rate safely.

There are lots of people who can help you in this regard. I recommend the books by David Dreman, Lowell Miller and James O’Shaughnessy as among the best. My only caution is that the historical models have reported the performance of small capitalization, value stocks to be so high that they may have become too popular.

The key is having a solid baseline in place. The worst case would be if valuations never become favorable. You would be stuck with your baseline. This would not be a bad outcome. We have excellent baselines.

Have fun.

John Walter Russell
March 21, 2006