Letters to the Editor

Updated multiple times on February 11, 2006.

More Comments: This Time Inspired by "Dollar-Cost Averaging Today"

Rob Bennett wrote this new Letter to the Editor. In many respects, it takes us in the traditional direction of Benjamin Graham. It also opens up new areas of research. Notice, for example, the focus on what happens over a 10-year period.

"Dollar-Cost Averaging Today" was in response to Rob Bennett’s January 29, 2006 Letter.

January 29, 2006 Letters to the Editor

I was excited to read the article on "Dollar-Cost Averaging Today."

First, it contains several compelling insights on how to invest successfully for the long term.

Second, I think it may end up doing more than that. I believe that it suggests a new direction in which to take your research efforts.

The early research on index investing has been aimed at making the case for investing in an index rather than picking individual stocks. I believe that the next stage of research will examine a different question, the question of HOW best to invest using indexes.

It has been ASSUMED that the best thing to do is to stick with a single stock-allocation percentage at all times. This assumption has been discredited in recent years. The historical stock-return data shows that it is best to CHANGE allocations with changes in valuations.

Once it becomes clear that the best thing to do is to change allocations, all sorts of questions that have never yet been examined come up. This new article of yours is putting a toe into a new pool of research questions, in my view.

It seems to me that a not uncommon scenario might turn out to be:

a) stocks provide the highest possible return over a 10-year period;

b) stocks provide the lowest possible return over a 10-year period; and

c) TIPS provide a moderate possible return over a 10-year period.

In those sorts of circumstances, it might make sense for some investors to invest a portion of their assets in stocks and a portion in TIPS. The trick is identifying the stock-allocation percentage that is not too hot and not too cold, but just right for that particular investor.

It's a risk-assessment question. If you aim for the best possible return, you take on a significant chance of ending up with the worst. But you unnecessarily give up upside potential if you go with the most predictable 10-year return available (the one obtained by investing 100 percent in TIPS).

You want to inch up the risk quotient of your portfolio until you get to the point where you would feel pain sufficient to cause you to sell stocks in the event of a worst-case-returns sequence turning up. This is not the same percentage for all investors. The particular financial circumstances and life goals of the particular investor influence the extent to which he can accept short-term losses without feeling a need to sell stocks in response.

The name I use to refer to these sorts of investigations is "Scenario Surfing." The conventional approach of determining a single "Optimal" allocation for all and that applies at all times is mumbo-jumbo la-la-land stuff, in my view. The idea of anticipating possible returns sequences, and taking into consideration the statistical probabilities of them turning up, makes a lot of sense, in my estimation.

I see the new article as being potentially a first step down an exciting new path.

HERE IS MY INITIAL RESPONSE.

I think that the following article gets us headed in the right direction. I contrast today’s actions, starting from today’s valuations, to those of the not-too-distant future, when valuations will be much more favorable.

What I did NOT do is analyze the effect of scatter.

If you look at the lower confidence limits, you will find that the worst case stock returns are as good as or better than 2% TIPS after a decade provided that you start at favorable valuations (P/E10 = 10). Today's story is much different.

Year 10 Choices: Edited
Year 10 Choices

I have added a new article about Risky Alternatives. I mentioned them first in the January 29, 2006 Letters to the Editor. Here is my latest offering.

More About Risky Alternatives

Comments Inspired by "Year 10 Choices"

Rob Bennett responded with this.

I have two questions:

1) Is there any argument that can be made for going with a stock allocation (I do not mean for those going with a high-dividend stock strategy, I am talking about those invested in a broad U.S. stock index) above 30 percent at today's valuations? I don't mean just for the typical investor. More risk is appropriate for some investors than for others. Are there any investor types for whom it would make sense to go with a stock allocation of higher than 30 percent at today's valuations? I believe that thinking about it this way can push us into seeing the numbers in new ways.

2) Do you have thoughts on how allocation shifts should be performed? In the research you have done, you have generally looked at three PE/10 levels and determined that it would be a good idea to make allocation shifts as the PE/10 level increased or decreased. Does it make sense for investors to make the shifts all at once as the new PE/10 level is reached? Or is that artificial? Would it be better to make gradual shifts in response to smaller changes in PE/10 levels? Is there a problem with gradual shifts that the numbers can go back and forth, causing lots of buys and sells? Should allocation shifts be made only after PE/10 changes have been "confirmed" by holding at the new level for a specified amount of time? Or should the shifts be lagging shifts, completed only after the actual PE/10 has moved a little further in the same direction as the change that prompts the allocation change according to the numbers?

HERE IS MY RESPONSE

Yes. There are several instances in which it can make sense to allocate more than 30% to stocks, even today. You have already touched on one: maintaining a high-dividend strategy. Add combinations such as intermediate-term timing with TIPS along with owning some high-dividend stocks. The worst case with intermediate-term timing is that you never buy stocks. The combination means that you always own some stocks even if they never become attractive enough to switch from TIPS.

Another thing to consider is the starting point. Many people start with a high stock allocation, especially if dollar cost averaging. They are likely to think of all price drops as buying opportunities. They are unlikely to be aware of today's risk. They need time to improve their allocations. It seldom makes sense to act abruptly.

My own thinking is that you should allow yourself about three years to execute any decision to sell. The idea is to get a decent price, to avoid a huge loss, not necessarily to get a good price. Prices fluctuate a lot. They go up as well as down, even when trending down. You should be able to avoid huge losses. Avoiding huge losses is important.

Another reason that someone would want to own more stocks, even at today's valuations, is to gamble based on price discipline. This is an instance where the thinking behind the risk-reward fallacy can make sense. This fallacy is the assumption that higher risk guarantees a higher reward (in a statistical sense, that is, on-average). The correct statement of a risk-reward tradeoff is that you should never accept a risk unless you are paid enough. That is, you accept a risk only if you are likely (in a statistical sense, on-average) to receive an adequate reward. In this case, we have a careful application of the greater fool theory.

Let us look at some numbers from "Year 10 Choices":

HSWR80T2, Withdrawal Rate = 4.0%, y = 68.1% of the initial balance (from -1.9% to 168.1%).

HSWR50T2, Withdrawal Rate = 4.0%, y = 72.7% of the initial balance (from 32.7% to 132.7%).

HSWR20T2, Withdrawal Rate = 4.0%, y = 75.1% of the initial balance (from 65.1% to 90.1%).

Knowing that prices fluctuate dramatically and that there are three or more upward price fluctuations within any decade, a disciplined gambler might make a bet. The odds are about 50%-50% that all three portfolios will end up with 68% to 75% of their original balances at year 10. The odds are about 20% that the balance at year 10 will be higher than one-half way between this level and the level of the high confidence limit. One-half way between 68.1% and 168.1% is 118.1%. One-half way between 72.7% and 132.7% is 102.7%. Might it not make sense to stick with HSWR80T2 or HSWR50T2 with a hope of catching an upward fluctuation big enough to restore your original balance by year 10? If you have enough discipline to get out and stay out after a bigger fool pays you his money, you have a chance to come out ahead.

There were many winners who got out before the bubble popped.

In terms of making such a bet, I advise you to be very careful about the statistics. Even though we can expect three or more upward price fluctuations, they are NOT independent. We always have to be prepared to live with our losses when a gamble fails. Price discipline helps us realize gains and mitigate losses. It does not prevent losses.

I have addressed varying allocations in several current research investigations:

Current Research A
Current Research B
Current Research G

Keep two thoughts in mind: price discipline and John Bogle's Cost Matters Hypothesis. Add this: our models are never EXACTLY right.

Moving in and out of stocks is a bad idea because of costs.

[In fact, when you consider costs, making special purchases to rebalance seldom makes sense. Rebalancing should affect only your choices of what new purchases to make or what to sell in order to make your normal withdrawals. When you include costs, rebalancing for the sake of rebalancing almost always loses money.]

I strongly recommend using limit orders, both to buy and to sell. It is amazing how well you can do. You can spread your purchases over several limit orders. That way, you mitigate any personal disappointment if you miss out on a great buying opportunity. You are likely to end up pleasantly surprised, even after missing out on a great price, when you end up with an even better price later on.

One of the reasons that I like having a TIPS baseline as part of intermediate-term timing is my experience with limit orders. You suffer very little if the timing strategy fails. You win big when it succeeds.

Our sensitivity tests have shown quite a bit of tolerance in terms of the exact thresholds and exact allocations. I stopped adding thresholds and allocations to my calculators because of diminishing returns. Extracting any greater detail would be pushing the data beyond what they can tell. If the extreme allocations had not been 100% below the lowest P/E10 threshold and 0% above the highest P/E10 threshold, I would have come up with locations for additional user-defined allocations.

Use limit orders and take your personal situation into account.

A final consideration is your own level of experience. But be very careful. Understand the consequences. Be sure that you understand how you ACTUALLY react in different situations, not just how you are supposed to react. Be sure that you have learned the RIGHT lessons.

Earlier Letters to the Editor (2006)

Unclemick about the Dow Jones Utilities. BillW (with Thanks!) about getting started. Rob Bennett about Dividend Theory versus Dividend Reality. This led to a breakthrough.

January 11, 2006 Letters to the Editor

Comments Inspired by Reading "The Story Behind the Numbers" by Rob Bennett. My response, including Risky Alternatives, Dollar Cost Averaging Today and DCA Today: The Point of Frustration.

January 29, 2006 Letters to the Editor