Many Different Objectives

Research on Safe Withdrawal Rates

Moderator: hocus2004

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

Many Different Objectives

Post by JWR1945 »

Much of what people think that they know about investing is false. Demonstrably so. Some of this can be attributed directly to advocacy. Key qualifiers are left unmentioned. Over-generalization is commonplace.

Different people have different needs.

The most commonly used, misleading assertion is that nobody can beat the market. Along with this is the observation that the average investor receives the average return of the market (before taking fees and expenses into account). That might be true if everybody shared the same goals subject to the same constraints. It is not true, because the phrase beat the market has been left undefined. The market consists of lots of groups with lots of different objectives. Each can do better than the overall market in terms of its own objectives, provided only that its objectives are realistic.

Even John Bogle, who strongly advocates the position that the best the average investor can do is to match the market, draws our attention to this. He points out that many investors are interested in accumulating assets for twenty or thirty years. Their objective is to end up with the largest final balance with risk measured in terms of the final balance. It does not matter how much their portfolios fluctuate along the way, provided only that they stick with a strategy. For them, the risk-adjusted return is based on twenty or thirty years. The annual risk-adjusted return is meaningless. Toss out the efficient frontier. It is a different issue.

We cannot even stop at this juncture. Most investors are interested in something other than the total return of an initial investment left untouched for a specified number of years. During accumulation, people continue to make deposits. During retirement, people make withdrawals. In the first case, dollar cost averaging is a blessing. In the latter, it is a curse.

What is worse is how misleading such reporting can be. Many mutual funds have started out with spectacular growth in their first few years, attracting many new investors, and then done poorly. Such a mutual fund's long-term performance continues to rank high, but only because its reported results start with an investment at the very beginning, when there were few dollars and fewer investors.

Hidden from many presentations is the issue of fees. Often, this is by necessity. Just be aware of the implications. For example, rebalancing costs money. There are always transaction fees and there may be taxes as well.

Perhaps the most complicated situation has to do with market timing. The time frame is usually left unmentioned. Great generality is claimed although the time frame is almost always restricted to one or two years. The effects of transaction costs are ignored or left unstated. It has been demonstrated that short-term market timing can beat a buy-and-hold strategy when fees are low enough. Earlier academic studies had assumed much higher transaction costs. The seemingly high potential advantage of short-term timing shrinks dramatically when models are restricted to using information available before the fact. Unstated, but important, is that a successful short-term timing model necessarily influences prices simply because its application results in large purchases and sales of securities. This limits its advantage further still.

Retirement Portfolios

As we study retirement portfolios, we place an initial screen on everything. We look at portfolio survival first. We insist upon high levels of safety.

We are still interested in portfolio balances along the way. We seek strategies that real people can live with. We adjust our approaches to improve the level of income produced and/or the final balance, but only after meeting a constraint on safety. Having such a constraint makes some important comparisons possible.

A recent example was an investigation into dividend strategies. TIPS can be used as a surrogate for high dividend stocks when it comes to safety. They cannot be used for determining the upward potential. Our investigations showed that the relative allocations between TIPS and stocks (i.e., the S&P500 index) would have been a toss up historically with TIPS interest rates of 2.5% to 3.0%. This tells us the dividend yield that we would have needed in the past. At yields of 3.0% and higher, a high dividend strategy would have been superior to an overall market strategy.

We would not have been able to reach that conclusion if it were not for the constraint on safety. There were too many cases when a final balance was higher or lower, sometimes favoring stocks in general (i.e., the S&P500 index) and sometimes favoring TIPS. The constraint limited the comparisons. In addition, an extrapolation from TIPS to high dividend stocks makes sense only when stock performance is poor and safety becomes an issue.

Our happy observation was that there are many high quality companies with dividend yields of 3.0% and higher. A dividend-based strategy is attractive even at today's valuations.

Have fun.

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

Post by hocus2004 »

"The market consists of lots of groups with lots of different objectives. Each can do better than the overall market in terms of its own objectives, provided only that its objectives are realistic. "

This is an insight rich in profitable implications. My expectation is that we will find value in exploring it not only for weeks and months but for years to come.

For now, I'll offer one reflection relating to some observations that have been put forward by UncleMick. He has noted that some of his investing ideas are in line with advice that was popular in earlier times but that has fallen out of favor in recent years.

What I believe has happened is that it became possible in the 80s and 90s to do a lot more to state in statistical terms the benefits of various stratagies. There are real benefits to quantification. Quantification permits a level of precision that is not obtained through use of vague recommendations to "buy low, sell high," or to "watch the dividends."

The reason why "much of what people think they know about investing is false" is that this trend toward quantification of investment insights took place during the greatest bull market in history. In bull markets, things are "spun" in a pro-stock direction for a whole host of reasons.

So a lot of what we "know" about investing is the product of statistical spin. As Bernstein notes in one of his essays, entire industries have grown up whose purpose is to "prove" various things about stocks that anyone with common sense knows cannot possibly be so.

My guess is that all of the spin of the past 20 years will be negated when investors come to full emotional acceptance of the fact that the bull market has really come to an (presuming that it has). At that point, the spin will be reversed, and the same quantification tools that were earlier used to "prove" that stocks are always the best investment will be employed to "prove" instead that investing in stocks is always a bad idea for the middle-class investor.

My hope is that the Data-Based SWR Tool will be used to counter the spin coming from both directions. My goal for this tool is that it be used not to support bull market psychology and not to support bear market psychology. I think that what people need is a tool that provides the benefits of quantification without the spin that has unfortunately usually been part of the quantification tools package for the past 20 years.

UncleMick has discovered that the historical data actually supports some of the conventional wisdom of earlier, pre-quantification days. I don't think that should be a surprise. Much of the earlier wisdom was pre-spin wisdom, closer to the mark than the "insights" we have seen generated by the quantification tools that have come along in recent decades.

That doesn't mean that quantification cannot be put to a good purpose, however. Statistical analysis is of great value. The problem is the spin that has been associated with it in its early days of development. What makes the Data-Based SWR Tool special is that it is the first quantification tool that is spin-free, that reports what is rather than what some people feel things to be or want things to be when they are living through the white heat of the greatest bull market in history.
unclemick
*** Veteran
Posts: 231
Joined: Sat Jun 12, 2004 4:00 am
Location: LA till Katrina, now MO

Post by unclemick »

JWR

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

Post by JWR1945 »

I wish to expand upon this point:
Even John Bogle, who strongly advocates the position that the best the average investor can do is to match the market, draws our attention to this. He points out that many investors are interested in accumulating assets for twenty or thirty years. Their objective is to end up with the largest final balance with risk measured in terms of the final balance. It does not matter how much their portfolios fluctuate along the way, provided only that they stick with a strategy. For them, the risk-adjusted return is based on twenty or thirty years. The annual risk-adjusted return is meaningless. Toss out the efficient frontier. It is a different issue.
Most of you are aware that the overall stock market's return is more predictable in the long-term than might otherwise be expected. The spread in returns narrows faster than the usual rate of 1/[square root of N], where N is the number of years. We sometimes refer to this as Reversion to the Mean. The reason is that stock prices are not entirely independent but they are related to earnings, at least loosely. John Bogle's modified version of the Gordon Equation captures this: the overall return = the initial dividend yield + the rate of growth in earnings + a speculative term related to the expansion or contraction of the P/E multiple.

Knowing this, you should not be overly surprised when you come across something like the following. It is from page 209 of The Single Best Investment by Lowell Miller.
Grant also pointed out the consistency between his findings and those of an earlier study by Fama and French. These two researchers concluded, in the Journal of Financial Economics (Vol. 22, 1988), that dividend yields can predict future stock returns and the forecasting power of dividends increases with the length of the holding period. Although yields explain less than 5% of monthly and quarterly return variances, Fama and French found that dividend differences explained fully 25% of return variances over a two- and four-year period, and that the level of impact is highly significant from a statistical point of view. They tested various factors as predictors and found..that the ability of [single year] earnings per share to forecast future price changes was inferior..because [single year] earnings are much less predictable..[Emphasis added.]
When you read about 5% of monthly and quarterly return variances, these are the values of R-squared that our computers calculate for us. The rise from 5% to 25% is indicative of an underlying bias (i.e., constant, always present) term. Its importance grows because it persists. [The formulas for calculating variances and R-squared do not guarantee that what they measure is caused by randomness. They may be indicative of bias terms and/or slowly varying terms instead.]

Be very careful about the time frame associated with any assertion about stock market returns. Conclusions do not automatically extend to different time frames. Most often, the statistical quantities that we run across are in terms of individual years. They do not necessarily apply to longer time periods.

Have fun.

John R.
MacDuff
* Rookie
Posts: 31
Joined: Wed Jun 18, 2003 8:20 pm
Location: Centralia, WA

Post by MacDuff »

JWR1945 wrote:Be very careful about the time frame associated with any assertion about stock market returns. Conclusions do not automatically extend to different time frames. Most often, the statistical quantities that we run across are in terms of individual years. They do not necessarily apply to longer time periods.

John R.
An excellent point. Some of the most helpful and insightful market discussions on the web take place on these pages. Keep up the very fine work.

Mac
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"Some of the most helpful and insightful market discussions on the web take place on these pages. Keep up the very fine work. "

Thanks for saying that, MacDuff. I very much agree with you re the significance of JWR1945's research. He has been smeared at a number of boards, but in the long run it is the work product that matters and the trash talk that gets blown away in the wind. There will come a day when JWR1945 will be recognized as a godsend to the entire FIRE/Retire Early/Passion Saving community. I am proud of the role that I played in kicking off the debate that got him posting and in creating the forum at which he has put forward so much outstanding work under such difficult circumstances.

I have a favor to ask of you, MacDuff. Please post here more often! We have earth-shaking insights coming out of our ears at this board. But we lack posters! Anyone who posts an occasional question or comment helps give this place a badly needed sense of life and energy. It only takes a few to step forward to give others the feeling that it is safe to do so themselves. Anyone who contributes here in even a small way is taking a large step toward bringing the various boards back to the state of Normalcy that most of us have been hoping to recapture for a long time now.
Post Reply