30-Year Stock Returns Starting at Recent Valuations

Research on Safe Withdrawal Rates

Moderator: hocus2004

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

30-Year Stock Returns Starting at Recent Valuations

Post by JWR1945 »

30-Year Stock Returns Starting at Recent Valuations

I have calculated 30-Year annualized real returns for portfolios containing 80%, 50% and 20% stocks when all dividends are reinvested. Expenses were set at 0.00%. All portfolios were rebalanced annually.

I examined portfolios with commercial paper as the non-stock component and with 2.2% TIPS as the non-stock component.

I have used Excel's plotting capability to fit straight lines of returns based on the data in 1923-1972 versus the percentage earnings yield 100E10/P. I have had Excel present the formulas of these lines along with the values of R-squared. I have estimated confidence limits visually.

I have calculated the results when P/E10 equals 27.0. This is close to recent levels. Looking at the latest update of Professor Robert Shiller's data, the S&P500 index in January 2004 was 1132.52 and P/E10 was 27.65. In June 2004, the S&P500 index was 1131.4 and P/E10 was 26.40. [Earnings have been increasing recently.]

The S&P500 index was 1112.84 when I wrote this article.

Here are the results.

80% stocks and 20% commercial paper:
return = [29.76 / (P/E10)] + 3.6738
R-squared was 0.3112.
Confidence limits equal plus and minus 1.5% (using an eyeball estimate).
At today's valuations (with P/E10 close to 27), the 30-year annualized real return is estimated to be between 3.3% and 5.3% (and 4.8% is the most likely return).

50% stocks and 50% commercial paper:
return = [11.18 / (P/E10)] + 3.4017
R-squared was 0.2232.
Confidence limits equal plus and minus 1.0% (using an eyeball estimate).
At today's valuations (with P/E10 close to 27), the 30-year annualized real return is estimated to be between 2.8% and 4.8% (and 3.8% is the most likely return).

20% stocks and 80% commercial paper:
return = [-8.45 / (P/E10)] + 2.9732
R-squared was 0.0721.
Confidence limits equal plus and minus 1.5% (using an eyeball estimate).
At today's valuations (with P/E10 close to 27), the 30-year annualized real return is estimated to be between 1.2% and 4.2% (and 2.7% is the most likely return).

80% stocks and 20% TIPS at 2.2%:
return = [34.21 / (P/E10)] + 3.6104
R-squared was 0.3039.
Confidence limits equal plus and minus 2.0% (using an eyeball estimate).
At today's valuations (with P/E10 close to 27), the 30-year annualized real return is estimated to be between 2.9% and 6.9% (and 4.9% is the most likely return).

50% stocks and 50% TIPS at 2.2%:
return = [22.31 / (P/E10)] + 3.2253
R-squared was 0.3049.
Confidence limits equal plus and minus 1.0% (using an eyeball estimate).
At today's valuations (with P/E10 close to 27), the 30-year annualized real return is estimated to be between 3.1% and 5.1% (and 4.1% is the most likely return).

20% stocks and 80% TIPS at 2.2%:
return = [9.33 / (P/E10)] + 2.6655
R-squared was 0.3039.
Confidence limits equal plus and minus 0.5% (using an eyeball estimate).
At today's valuations (with P/E10 close to 27), the 30-year annualized real return is estimated to be between 2.5% and 3.5% (and 3.0% is the most likely return).

Have fun.

John R.
Mike
*** Veteran
Posts: 278
Joined: Sun Jul 06, 2003 4:00 am

Post by Mike »

If I understand this correctly, high stock portfolios will outperform low stock portfolios, which is consistent with other research I have seen. Of course, this does not seem to address switching portfolio performance.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

Look carefully at the 50% stocks / 50% TIPS data. Its lower confidence limit is 3.1%. This is higher than both 80% stocks / 20% TIPS, which has a lower confidence limit of 2.9%, and 20% stocks / 80% TIPS, which has a lower confidence limit of 2.5%.

While it is true that higher stock allocations result in higher expected 30-year returns, differences in volatility [in the TIPS data] favor a middle level stock allocation when it comes to guaranteeing a minimum level of return.

Notice as well that all of my calculations are based on valuations similar to today's. That is, with P/E10 = 27.0, not with typical valuations and not at bargain levels.

This thread is really a spin-off from hocus2004's thread about Is the Raddr Methodology Analytically Valid? dated Friday, Oct 01, 2004.
http://nofeeboards.com/boards/viewtopic.php?t=2986

It is worthwhile to read that thread and especially the last few posts (on page 4) to see how this all fits in. At question is how to avoid the penalty of selling stocks when prices are depressed. As brought out on that thread, the penalty can be huge when valuations are high.

Discussions about portfolio returns over periods of 15, 30 and 50 years started around page 2 of that thread. In the context of that thread, we see that there is no advantage gained by rebalancing a portfolio at 30 years. It is better to have a separate 100% stock portfolio and to leave stocks untouched. The minimum return is at least as high as all of these portfolios and the upside is a good deal better. [With 100% stocks, the return at 30 years is between 3.3% and 7.3% and the most likely return is 5.3%.]

I have collected similar data at 15 years. I have not made my charts for those conditions yet.

Have fun.

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

Post by hocus2004 »

Thanks for putting this together, JWR1945. I find it helpful.

I have a few observations re things we might be keeping in mind as we proceed:

1) I have doubts whether it is necessary to always examine the results for commercial paper. I see this as being a holdover of the old approach to SWR analysis, in which only stocks and commercial paper were considered. It seems to me that TIPS or ibonds usually are a more effective choice for the non-stock portion of a portfolio.

2) There is obvious value in using current valuations for stocks and current returns for TIPS. That tends to be what people are most interested in seeing at this moment. But the reality is that we can best determine the realities of effective portflio construction by looking at a number of possible combination scenariios. TIPS offered a compelling value proposition when the SWR for TIPS was 5.85 and the SWR for stocks was 1.6. That was not all that long ago, but the circumstance have obviously changed quite a bit in that short amount of time. It's entirely possible that, in another 4 or 5 years, circumstances will have changed quite a bit once again. What we really want to learn is not just how to design portfolios in today's particular circumstances, but how to do so in general. We want to learn the principles of effective porttolfio construction. To do that, we need to examine various possibilities, not just those that apply at any one particular time.

One principle we might discover is that it is generally a good idea to hold both a high-growth asset (like stocks) and a stability asset (like TIPS). My guess is that there is almost no circumstance in which an investor should not be holding at least some of both types of asset classes.

A second principle might be that the proportional weighting of the two types of asset classes should vary over the lifetime of a portfolio. My guess is that the benefits of "timing" are so great that it is generally a mistake to decide to stick with a single mix from the day a portfolio is constructed until the day it is terminated.

A third principle might be that an investor needs to consider more than just the returns provided by the various mix possibilities when deciding on a mix allocation. You note in a recent post that many investors might not stick with a given allocation if they had no gains at the end of 15 years. The data that I have seen on this question (which is admittedly skimpy) supports your claim on this point. So I believe that it is unrealistic to ignore the factor of investor psychology, even though the data available on this aspect of the question is not what we would like it to be. The best thing to do might be to take reasonable guesses as to how most investors would respond to various possible outcomes, describe the basis of those assumptions in the write-ups, and then report in the write-ups the results that would be obtained presuming that the reasonable guesses in fact turned out to be a good reflection of the realities.

The downside of what I am suggesting here is that it is a step away from a pure data-based approach. The problem with not incorporating investor psychology into our analyses is that there is an implicit assumption being made when investor psychology is not considered, and that implicit assumption is very much at odds with historical experience.

We have seen that there is a huge penalty imposed for lowering one's stock allocation when prices fall. We know that middle-class investors historically have lowered their stock allocations when stock prices fell. The reality is that one of the big risks of going with high stock allocations is that you will experience the sort of return sequence that will cause you to be hit with the Stock Selling Penalty. This is the biggest risk of the game. SWR analysis is a risk-assessment tool. So obviously SWR analysis needs to be focused on the biggest risk factor. Yet the conventional methodology way of dealing with this factor is to assume it away, to set up the analytical framework as if it were not possible for investors with 74 percent stock allocations to lower their stock allocations when stock prices fall dramatically. The result of using far-fetched assumptions when examining the risk involved in high stock allocation is to skew the results heavily in favor of high stock allocations. The assumption is a fantastical one, and I think we should reject it out of hand.

There is no good purpose served by making assumptions that have no support in the historical record. The reality is that investors have emotions, and emotions affect the returns obtained from stock purchases. Where we have data to guide us as to what assumptions to make re the effect of investor emotions, we should construct our scenarios in accord with the data. When we lack data, we should use our common sense to construct the most realistic scenarios possible (always describing them in our write-ups, of course). The aim should be to be as realistic as possible, not to skew the results to favor one particular asset class over another. The ultimate goal is to help aspiring early retirees assess the risks associated with various investment choices, not to trick them into following the course followed by the individual who prepared the study. That's the distinction we should be keeping in mind when asking ourselves whether a particular way of proceeding is consistent with the rest of the data-based methodology or not.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

hocus2004 wrote:1) I have doubts whether it is necessary to always examine the results for commercial paper. I see this as being a holdover of the old approach to SWR analysis, in which only stocks and commercial paper were considered. It seems to me that TIPS or ibonds usually are a more effective choice for the non-stock portion of a portfolio.
It is always helpful to have a few standard conditions when making studies. It helps us narrow down the number of possible causes related to an effect.

With our calculators, we generally include the following standards:
1) a 30-year duration,
2) stock allocations of 80% and 50% and often 20% as well,
3) conditions with commercial paper as the component other than stocks,
4) expenses of 0.20% and/or 0.00%,
5) the annual rebalancing of portfolios,
6) including inflation as measured by the CPI-U (i.e., consumers price index for urban workers) and
7) adjusting withdrawals and describing balances in terms of inflation adjusted (real) dollars.

We usually include conditions in which we estimate what it takes to avoid depleting our assets. Sometimes, we set this threshold higher, such as having a minimum balance of 50% of the initial balance or having the ending balance at least as large as 50% of the initial balance.

If we change our standard conditions to replace commercial paper with TIPS, we must identify some standard interest rates. Three sets come to mind:
1) 0%, 2% and 4%,
2) 1%, 2% and 3% and
3) 1%, 2.5% and 4%.

My preference is for the third set. An interest rate of 1% has been typical of the real return from commercial paper in many instances. An intermediate interest rate of 2.5% is typical of what one was able to do in the recent past and what he might be able to get in the near future. An interest rate of 4% corresponds to the highest rates (after rounding) that have been made available.

An interest rate of 0% can be attractive since it is (almost always) easy to handle. An interest rate of 3% would be helpful for reasons of symmetry when the other rates are 1% and 2%. In addition, the rates of 2.5% to 3.0% defines when TIPS interest and dividends yields become more attractive than the S&P500 index (under stressful conditions).

Using commercial paper has the advantage of requiring only one set of data. It has the disadvantage that it is, in effect, a combination of all three real interest rates (high, low and medium) applied to different sections of history.

Are there better choices? What are some other opinions and preferences?

Have fun.

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

Post by JWR1945 »

hocus2004 wrote:The downside of what I am suggesting here is that it is a step away from a pure data-based approach. The problem with not incorporating investor psychology into our analyses is that there is an implicit assumption being made when investor psychology is not considered, and that implicit assumption is very much at odds with historical experience.
It is easy to include psychology and data-based information.

Psychological conditions define different things to look for in the data-based information.

For example, the psychological effect of severe market downswings cause us to collect data with 20% (or 30%) stocks as well as with other allocations. We no longer ignore or bypass such conditions. We evaluate them as thoroughly as other conditions. Psychological effects are introduced later when discussing the realism of using various stock allocations. This is in contrast to what has been done in the past. The complete set of data is made available. Discussions are no longer limited to a narrow subset.

Have fun.

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

Post by MacDuff »

hocus2004 wrote:The best thing to do might be to take reasonable guesses as to how most investors would respond to various possible outcomes, describe the basis of those assumptions in the write-ups, and then report in the write-ups the results that would be obtained presuming that the reasonable guesses in fact turned out to be a good reflection of the realities.
The downside of what I am suggesting here is that it is a step away from a pure data-based approach. The problem with not incorporating investor psychology into our analyses is that there is an implicit assumption being made when investor psychology is not considered, and that implicit assumption is very much at odds with historical experience.

We have seen that there is a huge penalty imposed for lowering one's stock allocation when prices fall.
From my limited understanding, an extensive bear market will automatically lower stock allocations. Say you have a $1,000,000 50:50 portfolio, and equities drop 20% in year t, while the fixed component prices are stable. You have withdrawn your $40,000 living expenses from fixed, because the market is falling. The market has depreciated your equities by 20% of $500,000, so you now have $400,000 in equities, and $460,000 in fixed. Say year t+1 is the same, and ignoring inflation you take another $40,000 from fixed, and the market whacks your equities by another 20%. You have at year end (t+1): $420,000 in fixed, and 80% of $400,000 =$320,000 in equities. Let's allow it to happen again- not common, but not unheard of especially if we were looking at real results. So now we have left $380,000 in fixed, and 80% of $320,000= $256,000 in equities. Now at this time, year end (t+2) our steadfast investor is down to an overall portfolio of $636,000, of which equities are $256,000 or 40%. This is without rebalancing. With rebalancing, we can keep topping up equities to 50%, but our overall portfolio is going down faster as a result of this.

I submit that even if our cojones are such that we do not panic and sell stocks, and maybe we can convince our wife that this is a wise course, no way is she going to allow any "throwing good money after bad". That is how she will characterize rebalancing in these circumstances.

I observed this exact situation in the 70s bear market. The guy, who was an experienced investor, wanted to buy good merchandise cheap, but the wife had seen the carnage of their recent experience, and she said no way! The upshot was that they got divorced, and he didn't have much left with which to rebalance anyway.

My feeling is that these boards are heavily populated with "investors" who have not yet seen combat. Who knows what they will do? But IMO it would be a very big mistake to try to formally incorporate psychology into the analysis and planning process. Hocus, you seem to be preoccupied with what you should teach others about the portfolio side of early retirement. I say lets get it right first. If psychology is put in, all utility will be destroyed. Each potential user can observe himself, and his reactions to adversity, and decide what allowances he needs to make.

Respectfully, Mac
Last edited by MacDuff on Sat Oct 16, 2004 11:51 am, edited 1 time in total.
unclemick
*** Veteran
Posts: 231
Joined: Sat Jun 12, 2004 4:00 am
Location: LA till Katrina, now MO

Post by unclemick »

Yep

Numbers first and then try to deal with the emotions. And I wouldn't waste a lot of time flogging 'conventional methodology'. I have this sister who has asked on and off for twenty years - and promptly did not hear. I gave the nephews Bogle's 1994 book after they left the Naval Academy - they heard - although the oldest is reading Four Pillars.

Some hear, some don't. And nobody but nobody likes - I told you so.

Break it down - run a good data set to illustrate - and hope for the best - especially if it's a relative.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"It is always helpful to have a few standard conditions when making studies. It helps us narrow down the number of possible causes related to an effect. "

I agree. My question is whether it continues to make sense to consider commercial paper the standard alternative to stocks. There's obviously no problem with looking at commercial paper when there is unlimited time to look at a wide number of different options. It appears to me, though, that comparisons of stocks with either TIPS or ibonds generally provide more interesting results.

"If we change our standard conditions to replace commercial paper with TIPS, we must identify some standard interest rates. Three sets come to mind...."

My preference would be to look at 2 percent TIPS, 3 percent TIPS, and 4 percent TIPS, since that covers the range of options we have seen so far. If it were possible to do four sets, I would at that point add in 1 percent TIPS. I doubt that we will ever see TIPS paying much more than 4 percent real.

"a 30-year duration"

I generally think it is good to stick with the 30-year duration. The exception is where the question we are trying to examine relates to the issue of how things look at other time-periods. We already have been making exceptions to the general 30-year rule in those sorts of cases.

"stock allocations of 80% and 50% and often 20% as well, "

I think that 30/50/70 would be a reasonable alternative. I think that 20/50/80 is fine too.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"Psychological effects are introduced later when discussing the realism of using various stock allocations."

I believe that we are talking about different things, JWR1945.

The purpose of SWR analysis is to provide numbers-based insights. I don't think that there are too many people who would take issue with the common-sense observation that there is more risk in holding a large stock allocation at times of high valuation than there is in holding a moderate or low stock allocation at times of high valuation. The problem is that most people are not able to quantify the effect of increasing their stock allocation. They know it is in one sense a bad thing to do, but they don't know how bad. They also know that there is another sense in which it is a good thing to increase one's stock allocation. Stocks are a high-potential-return asset class. So the more you have of them, the more growth potential there is built into your portfolio. The growth potential factor is quantified in conventional SWR studies. The risk factor is not. The unfortunate consequence is that a misleading impression is created, an impression that it is possible to tap into the growth potential without taking on any increased risk. The historical data shows that this is not so.

Think of this in terms of income streams. Say that you have $1,000,000 in savings and you are going to choose 10 investments of $100,000 each. You want to obtain the greatest growth potential possible, but there is a limit to how much risk you can afford to take on. The conventional studies tell you that you can count on an income stream of 4 percent from 7 segements of $100,000 invested in stocks. Can you?

We have shown that you cannot count on a 4 percent income-stream today because that is a number that assumes moderate valuation levels and we are now not at modetate valuation levels. But we have not adequately answered the question "Well, what can you count on then?" The SWR you reported for a 80 percent S&P allocation at today's valuation levels is about 2.5 percent. Knowing that number is a big help. But it seems to me that what we discovered in the "Is the raddr Methodlogy Analytically Valid?" thread is that there are circumstances where you can invest $100,000 segments of your accumulated capital in S&P stocks and count on an income stream a lot better than the one suggested by a 4 percent take-out number.

What if you invested only $100,000 in S&P stocks? That's only a 10 percent allocation. Your numbers show that you can count on a 3.3 percent real return from that money (you would never need to sell stocks to cover living expenses if you had 90 percent of your money in other asset classes). A 3.3 percent real return permits a withdrawal rate of a heck of a lot better than 2.5 percent. I don't know the precise number, but I believe that a 3.3 percent real return translates into an SWR of something near 5 percent, does it not? That's double the take-out number you got from stocks when going with a 70 percent allocation.

That's a big deal, in my opinion. That's something we should want aspiring early retirees to learn about. The power of SWR analysis is the ability it provides to come up with these sorts of insights.

Anyway, that's my take. It does not appear to me that there is a single SWR that applies for stocks. The amount that you can withdrawal safely from stocks depends on the extent to which you are dependent on gains from the stock portion of your portfolio to cover your living expenses. Portfolio construction is a process of mixing and matching different asset classes to create the most effective combination for an investor in your particular circumstances. It seems to me to be a highly significant finding that it is possible to take a much higher withdrawal from an income stream coming from a stock investment when the stock allocation is kept to moderate levels. This way of looking at things provides a very different take on the question of what stock allocation is "optimal" than the take provided by conventional methodology studies.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

? ? would be a very big mistake to try to formally incorporate psychology into the analysis and planning process."

Thanks for coming forward with your viewpoint, MacDuff.

I hope that no one has gotten the mistaken impression that because I am moderator of this board that I get to set its agenda. I have a vote like everyone else, and, to the extent that I participate more, my viewpoint probably gets more weight. But this is a community resource, and the community as a whole determines the direction in which our conversations move.

For that matter, there's no need for us to go in a single direction. There's no law of the universe that says that we can't have one group that goes off one way and another that goes off in another. As a practical matter we don't have enough posters here today to be exploring too many different directions all at one time. But I hope that will change in coming days. So I just wanted to put up a little note letting people know that you don't have to agree with my views (or JWR1945's view, which in some respects are in line with my views and which in some respects are not) to participate.

Anyone who has a question or comment relating in some way to SWRs is both entitled and welcomed to participate. I can't guarantee that you will get a lot of feedback if you do. But you never know until you try. And it's always possible that a comment that elicits little feedback today will elicit a lot six months down the road when some newcomer visits and discovers it. If you think things have gone off track in some respect, you are doing a service to the community to say so and to say why you think so. We very much need new posting blood here. It is not even a little bit healthy that my views and the views of JWR1945 are so dominant. I love this board, but I hate the fact that there are so few viewpoints expressed here. That reality greatly diminshes the value of the work we do here.

"Hocus, you seem to be preoccupied with what you should teach others about the portfolio side of early retirement."

Yes, that's a big issue with me. I am planning lots of writings outside of the discussion board context, and I of course want to get things right to the extent possible. I have found discussion boards to offer a good means of testing ideas before going public in a big way with them. That's one of the reasons why I like to hear feedback representing as many viewpoints as possible.

I'm also concerned about what the community as a whole (not just this site) says about the various FIRE/Retire Early/Passion Saving issues. Newcomers come into our movement all the time, and I want us to look good when they do. I don't think we need to be in agreement to look good. I think that strongly expressed differences of opinion gives our boards a feeling of life and open-mindedness that makes a good impression. I don't like to see ugly posting practices, however. I think that is a turn-off. It makes it hard for us to bring in new voices to the community, and that diminishes the learning resource that these boards represent.

"Each potential user can observe himself, and his reactions to adversity, and decide what allowances he needs to make. "

Many people take this viewpoint, MacDuff. You are certainly not alone in believing this. I disagree strongly. I see no evidence whatsoever that the average investor is able to assess the extent of the risk that he will fall victim to the Stock Selling Penalty.

I think we are watching a tragedy play out before us. I think that lots of middle-class investors are going to re-learn the lessons that have been taught by the historical data a number of times in the past by losing large percentages of their accumulated capital. My goal is to do what little I can to help people in the Retire Early community avoid unnecessary losses. There are obviously limits to what one person can do, and there comes a time when I have to accept the reality of those limits. I want to be sure that I have done what I can, however. I don't want to do less than that much.

It does not concern me too much if the facts are presented and people choose not to respond in effective ways to them. When that sort of thing happens, I just don't see it as my responsibility to do anything about it. It's one of those things. On the SWR question, I feel differently. On the SWR question, it is not a matter of the facts being presented and people electing not to respond effectively. Here we have a matter where a small number of community members have taken it upon themselves to decide what discussions their fellow community members may engage in. I think that is wrong and I think that is dangerous.

If it turns out that people lose large segments of their accumulated capital because of demonstrably false claims re SWRs that have been put before these boards, there are going to be some who do not find too much humor in it. There are going to be angry community members. When they look back through the archives and see that they were denied information that might have made a difference because a small number were uncomfortable with the idea of permitting reasoned discussions on the SWR question, I think we are going to see some not-so-good reactions. We have had over 100 community members express a desire that reasoned discussions of the realities of SWRs be permitted. I think that we should be responsive to those desires. I thnk that it is a black mark on the integrity of the boards that we have not been responsive to those desires thus far.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"I wouldn't waste a lot of time flogging 'conventional methodology'. "

Our Board Project is to develop and refine a new SWR methodology. How can we do that without comparing and contrasting the new approach with the old one? The flaws we have discovered in the conventional methodology were the impetus for the discussions we are having. I don't see how we can leave discussion of the flaws of the conventional methodology studies out of this.

It is the friction that has characterized the discussions that we should be trying to do away with, not reasoned discussions of the flaws of the conventional methodology. The friction is the result of abusive posting practices. All sorts of stuff that has no place on discussion boards created to help people achieve financial freedom early in life has appeared on our boards over the course of the past 29 months. We should all be working to see that stuff shut down. It reflects poorly on each and every one of us. It helps absolutely no one. There is not one community members who benefits from that stuff. Not one. That stuff should be brought to a quick stop. Today would be a good day for that to happen.

"Some hear, some don't. And nobody but nobody likes - I told you so. "

I agree with this statement. The problem today, as I noted in my response to MacDuff, is that the DCMs are engaging in all sorts of abusive posting practices to block the community's desire for reasoned debate. You can't expect people to hear what cannot even be spoken. Once the discussions are permitted to go forward, then it indeed becomes reasonable to just leave it to the various community members to determine how they will respond to the message. So long as reasonsed discussions are being actively blocked, the process question is of paramount importance, in my view. I don't think that people should need to agree with me on SWRs to agree that reasoned discussion should be permitted. All responsible community members should favor a reining in of the abusive posting practices of the DCMs. That stuff is a poison to a posting community. It does harm to us all.
unclemick
*** Veteran
Posts: 231
Joined: Sat Jun 12, 2004 4:00 am
Location: LA till Katrina, now MO

Post by unclemick »

1. Dream on. 2. You have not met my sister(lucky you - heh, heh). 3. A good data set tells the story unto itself. 4. Some get it, some don't.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

hocus2004 wrote:What if you invested only $100,000 in S&P stocks? That's only a 10 percent allocation. Your numbers show that you can count on a 3.3 percent real return from that money (you would never need to sell stocks to cover living expenses if you had 90 percent of your money in other asset classes). A 3.3 percent real return permits a withdrawal rate of a heck of a lot better than 2.5 percent. I don't know the precise number, but I believe that a 3.3 percent real return translates into an SWR of something near 5 percent, does it not? That's double the take-out number you got from stocks when going with a 70 percent allocation.
This is how you handle the mathematics.
1) You determine what the account balance will be after the specified number of years (in this case 30 years). The formula is (1+r)^N = [final balance / initial balance].
2) At r = 3.3% and N = 30, the final balance is 2.65 times the initial balance. This is worst case.
3) At r = 5.5% and N = 30, the final balance is 4.71 times the initial balance. This is the most likely result.
4) At r = 7.3% and N = 30, the final balance is 8.28 times the initial balance. This is the best case.
5) At the end of the savings period, the account will be put into service to provide income. Multiply the income that would be provided by the current balance by the ratios above. For example, if we wish to withdraw funds for another 30 years, TIPS at a zero percent interest rate would provide 3.33% per year (since 1/N = 1/30 = 3.33333%) and end up [after making withdrawals for 30 years] with a final balance of zero dollars. Multiply this by 2.65, 4.71 and 8.28. The results are 8.8%, 15.7% and 27.6%, respectively. These numbers show the benefit of placing the stock account to the side for the next 30 years (starting from today's valuations).
6) Expect to be able to do better than TIPS at a zero percent interest rate.

Have fun.

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

Post by JWR1945 »

Mike wrote:If I understand this correctly, high stock portfolios will outperform low stock portfolios, which is consistent with other research I have seen. Of course, this does not seem to address switching portfolio performance.
You are correct.

I would expect switching to increase returns. The difficulty is that switching would probably have us entirely out of stocks right now.

When I have investigated multiple, separate portfolios in the past, I have always ended up favoring combining everything into a single portfolio. That is, you would be better off withdrawing from a combined stock / commercial paper portfolio than withdrawing from a 100% commercial paper portfolio and, after it is depleted, withdrawing from a second portfolio (which has its own allocation of stocks and commercial paper).

In terms of actually implementing a strategy, there is quite a bit of emotional appeal to putting something on the side and leaving it alone and in automatic (e.g., a stock index fund with everything reinvested) for a number of years. For one thing, it is less likely to be raided if it has been separated from current accounts.

Have fun.

John R.
Mike
*** Veteran
Posts: 278
Joined: Sun Jul 06, 2003 4:00 am

Post by Mike »

The difficulty is that switching would probably have us entirely out of stocks right now.
Probably out of the S&P for most of the last decade, which would have been unproductive due to historically ununsual circumstances. My interpretation of the above data is that the 30 year fixed allocation tends to approach the efficient frontier at about the 60% equity level even today, which is similar to what has been published for years. S&P returns go down as P/E goes up (but there is a wide band), which has also been well documented. This will reduce returns, but static allocations that include equity will likely still outperform 100% static fixed income.

Still not fully explored are switching modifications to make them more useful today, carefully selected multiple uncorrelated asset classes, moving into under valued assets beyond the S&P, short term momentum, and value as alternatives. I currently approach high dividend equities as a subset of value investing. The topic is interesting.
...emotional...
I suspect this sums up the abysmal statistics for most active investors, and why so many experts recommend a static 60/40 allocation (e.g., Vanguard Balanced Index) It may not be optimal, but they feel it will keep the majority out of trouble. Those who can control their emotions, such as Warren Buffet, do much better with alternative approaches. However, they seem to be in the minority. Hmm.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"A good data set tells the story unto itself"

I half agree with you on this and I half don't, Uncle Mick.

When I did my first investigations into SWR stuff, I was only engaging in cursory examinations of the historical data. I certainly did not engage in the sorts of in-depth analyses you often see JWR1945 put forward at this board. Those cursory reviews of the data were indeed good enough to provide me some valuable insights. Nothing that I have learned in recent years has caused me to change impressions that I formed through those cursory investigations.

The other side of the story is that the understanding I possess today is far more sophisticated (although by no stretch of the imagination complete). A quick look at the data helps. But in-depth analyses helps more. So I agree with the suggestion that the data to some extent speaks for itself. I also believe, however, that in-depth analysis of the data (both from a conceptual standpoint and from a numbers standpoint) adds a lot to one's understanding.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"Expect to be able to do better than TIPS at a zero percent interest rate. "

Indeed. I find it fascinating how looking at the same data from different standpoints yields such different action-step possibilities. We have a number of switching threads that suggest that it might make sense to go with a zero stock allocation at today's prices. The "A Bright Future" thread makes a compelling case for TIPS. And now we have a few threads suggesting that a moderate S&P allocation even at today's valuation levels may not be at all a bad idea for most investors.

None of the different types of threads conflict with each other. The findings of the new threads do not reveal the findings of the old threads to be inaccurate. We are coming up with different "findings" because we are examining different sorts of things. The realities of the historical data are more complex than many of the existing analytical tools indicate. Cookie-cutter approaches are dangerous because they work in enough circumstances to gain credibility and then fail to work when circumstances change from those which applied when the analytical frameworks for these approaches were being developed.

My view is that we need to be tentative in all of our findings. We are still in the early innings of this, and there are probably a lot more surprises in store for us down the road. Surprises are exciting because, when we are surprised, we are not relearning old lessons but breaking new ground.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"I suspect this sums up the abysmal statistics for most active investors, and why so many experts recommend a static 60/40 allocation (e.g., Vanguard Balanced Index) It may not be optimal, but they feel it will keep the majority out of trouble."

I think you are right about this, Mike. Given what we have learned about the benefits of a long-term switching strategy, one of the great puzzles of all this is, why is it that the smartest money advisors do not generally make investors aware of what the historical data says on this point? I think that the explanation is what you suggest above. There is a worrry that, if the benefits of switching are publicized, many investors will misunderstand what is being said and will engage in all sorts of short-term timing exercises and will end up hurting themselves rather than helping themselves.

I don't think that the experts who refrain from publicizing the benefits of switching are doing the right thing by keeping people in the dark. I do understand the danger that they are concerned about, however. This stuff is not terribly complicated, but it is a little complicated. There are indeed a good number of people who would misunderstand the switching realities if they were better publicized.

The problem is that, as stock valuations went up and up, the penalty for not engaging in any switching whatsoever grew larger. We now are at valuations where there is genuine danger inherent in the reality that most investors think that even long-term timing is "impossible." It appears that a lot of people are going to lose a lot of money because of the efforts that a good number of experts engaged in to "protect" investors.

My guess is that switching strategies will become a lot better publicized after investors have sustained large losses from failing to consider switching options. At that time the benefits of switching will not be nearly as great as they are today, of course. So people are being kept in the dark at the time when switching offers its greatest possible rewards, and then will be made aware of switching's benefits only when the strategy becomes a far less profitable one. An unfortunate turn of events all the way around, in my view.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

Mike wrote:The difficulty is that switching would probably have us entirely out of stocks right now.
Probably out of the S&P for most of the last decade, which would have been unproductive due to historically unusual circumstances. My interpretation of the above data is that the 30 year fixed allocation tends to approach the efficient frontier at about the 60% equity level even today, which is similar to what has been published for years. S&P returns go down as P/E goes up (but there is a wide band), which has also been well documented. This will reduce returns, but static allocations that include equity will likely still outperform 100% static fixed income.

Still not fully explored are switching modifications to make them more useful today..

This may be of interest.

Look at projected 15-Year returns for the S&P500 and mixtures consisting of 80%, 50% and 20% stocks along with commercial paper or 2.2% TIPS.

15-Year Stock Returns Starting at Recent Valuations dated Sat Oct 16, 2004.
http://nofeeboards.com/boards/viewtopic.php?t=3010

In all cases the lower confidence limit is less than 2.2%. In fact, the most likely return at 15 years is less than 2.2% unless the portfolio has 50% TIPS or 80% TIPS. With 50% TIPS, the most likely return is 2.2%. With 80% TIPS, the most likely return is 2.3%.

A portfolio consisting of 100% TIPS at 2.2% interest always returns 2.2%: high, low and most likely.

Being 100% in TIPS at today's valuation provides a return at least as good as what is likely and often much better and all at zero risk.

It takes unusual circumstances for stocks to be attractive at today's valuations. A switching allocation of 100% makes sense.

In addition, an allocation of 100% TIPS at a 2.2% interest rate allows you to withdraw 4.0% (plus inflation) for 36 years with zero risk. Reference:
3% SWR for 56 Years dated Mon Oct 13, 2003.
http://nofeeboards.com/boards/viewtopic.php?t=1541

Have fun.

John R.
Post Reply