"Getting Going" or "Going Bust"?

Research on Safe Withdrawal Rates

Moderator: hocus2004

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

"Getting Going" or "Going Bust"?

Post by hocus2004 »

Here is a link to today's "Getting Going" column by Jonathan Clements of the Wall Street Journal.

http://online.wsj.com/article/0,,SB1100 ... 61,00.html?

He is writing about a study that suggests that a withdrawal rate well in excess of 4 percent is reasonable in many circumstances. I need to look at the study itself. But it appears from the column that the study starts with the famous conventional methodology 4 percent number and then makes some adjustments in assumptions that would permit higher withdrawals and puts forward the suggestion that such higher withdrawals are therefore also "safe."

If this is so, the advice in this column is nonsense gibberish. The conventional methodology studies are rooted in the far-fetched assumption that for the first time in history changes in valuation are going to have zero effect on the intermediate or long-term performance of stocks. Those studies all report inaccurately what the historical data says re what withdrawal rate is safe. Any studies that incorporate new features onto the base of the conventional methodology approach are of course also inaccurate.

This column illustrates the danger we have created for aspiring early retirees (and regular retirees too, for that matter) by passing up so many opportunities to talk straight about the realities of SWRs. Clements is a respected journalist. The Wall Street Journal is a respected business newspaper. If these sorts of sources are peddling aspiring retirees this junk science, how can we expect retirees to know what the historical data really says?

If we had spent more of our energies developing the data-based SWR tool, we could be taking that tool public right now. We could be getting the work of this board community written up in the Wall Street Journal instead of seeing columns like this being put forward and putting the retirements of thousands at great risk of failure.

The publication of this column bears on another point that we have discussed at earlier times. I have argued that my post of May 13, 2002, was the most important post put to a board in the history of FIRE/Passion Saving/Retire Early boards. That comment generated a lot of flak. We have had people put forward sarcastic comments that "hocus claims to have come up with the idea of value investing." I did not come up with the idea of value investing and I did not say that I came up with the idea of value investing. What I did in the mid-90s when I developed the data-based tool was to marry the concepts of SWR analysis and value investing. By doing that, I made it possible for SWR analysis to generate numbers that are at least reaosonably accurate.

I think it is fair to presume that Jonathan Clements has heard mention once or twice of the concept of value investing. But here he is writing a column in which he argues in favor of a study rooted in the conventional methodology, a methodology rooted in the presumption that changes in valuation have zero effect on long-term stock returns. The assumption is wrong, the conventional methodology studies are wrong, and the column is wrong. If the data-based methodology is "nothing new," as many have claimed, why is it that a respected journalist like Clements is making such basic errors? The reality is that Clements has been fooled by the surface plausibility of the conventional methodology studies just as many board members have been fooled by the surface plausibility of the conventional methodology studies. The data-based tool is indeed something important and new. The data-based tool takes the concept of SWR analysis and changes it so that it generates analytically valid results. That's a significant advance.

We should be doing what we can to set people like Jonathan Clements straght. We should be letting early retirees and regular retirees know that they cannot place their trust in nonsense studies using far-fetched assumptions to generate fantasy-land indications of what withdrawal rates are safe. The SWR is a data-based constuct. The SWR is whatever the historical data says it is. Studies that fail to take into account the historical reality that changes in valuation have always affected long-term returns are not science, they are science fiction.

Presuming that the study referred to in this column is based on some variation of the conventional methodology (I do not know that for certain, but the language of the column suggests that this is so), the advice set forth in this column is dangerous advice. I hope that after 30 months of this that most members of this community have a better understanding of the realities of SWRs than either Jonathan Clements or his editors at the Wall Street Journal possess today. I hope that in future days we will be doing what we can to bring Clements and his editors at the Wall Street Journal up to speed.

Clements Quote #1: "one influential study found that, if retirees want to be confident their savings will last 30 years, they need to limit their initial withdrawal rate to 4.1%, or $4,100 for every $100,000 saved. But a new study by Minneapolis certified financial planner Jonathan Guyton, which appeared in October's Journal of Financial Planning, suggests retirees don't have to be nearly so frugal."

Clements Quote #2: "Retirees may be able to withdraw as much as 6.2% initially, provided they follow three rules....If your portfolio loses money during the year, you can't give yourself a raise the following year.... No matter how high inflation gets, your maximum annual increase is 6%....You have to avoid selling hard-hit stock funds. Instead, each year, start by lightening up on winning stock funds."

Clements Quote #3: "All this, of course, is based on an analysis of historical returns. If we get wacky markets, miserably low long-run returns or you incur hefty investment costs, you may have to withdraw less than Mr. Guyton suggests."

Clements Quote #4: "Mr. Guyton's rules make a ton of sense to me -- and I suspect many retirees will find his approach appealing."
Mike
*** Veteran
Posts: 278
Joined: Sun Jul 06, 2003 4:00 am

Post by Mike »

You can take more than 4% if you don't adjust for inflation. Same old story. If you are willing to risk a lower standard of living later on if inflation takes off, or the market swoons, you can spend more at the start.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"You can take more than 4% if you don't adjust for inflation. "

Probably so. But how much more? Can you go as high as this Wall Street Journal article suggests or are the assumptions used in the study reported on in the article analytically invalid? I'd like to know for sure.

Is this something that you could determine by making reference to the historical data, JWR1945? You have determined that the SWR at today's valuation level is about 2.5 percent, assuming an inflation adjustment. Would you be able to say how much higher that number goes if there is no adjustment made for inflation? The research cited in the article says that you can go with a take-out of 6.2 percent if you don't adjust for inflation. Does the data support that claim or is this more nonsense gibberish?

The Wall Street Journal sells a lot of copies of its newspaper. Many thousands of aspiring retirees are going to be placed in danger of severe life setbacks if the numbers in this artilcle are wrong and are not corrected. We are a community concerned with educating people as to how to achieve safe retirements. It would seem to me that we would see it as part of our mission to correct junk science, if the number that is being advanced in this article is indeed the product of junk science.

We all need to start taking this thing seriously. It is not a joke when people see their retirements go bust. People have noted quite properly that I possess limited skills with numbers. One of the great ironies of all this is that someone with such limited numbers skills would be the one to discover that the conventional methodology was analytically invalid. It is not something you would intuitively expect to see happen. So what edge is it that I possessed that no one else possessed? The edge that I possessed was an intense determination to build a safe plan. When I use words like "safe," I am not kidding around. I mean for them to signify just what they are said to signify in their dictionary definitions. My plan has lots of risk in it in its non-investing aspects. My family's hopes for the future are riding on my ability to make my plan work. So I can't afford to play games with the word "safe" when I am making investment decisions. If someone tells me that some investment strategy or other is "100 percent safe," they had better know what they are talking about. If they don't, my kids might not be able to go to college. That's not a joke. That's a serious life setback for them and for me too.

I take offense at people who treat this as a joke. I learned the realities of SWRs because I took the time to get it right. I did my homework. There's a song by Graham Parker called "Fool's Gold" where he starts out with the line "I've been doing my homework now for a long, long time." I like that song.

I am not the only early rertiree who has something riding on the accuracy of SWR analysis. I have seen many, many posts on the various boards in which community members have made use of SWR studies to construct their plans. If they are being fed nonsense gibberish, we have a responsibility to let them know. I know that back in the day when I was putting together my plan, I would have wanted someone to tell me if there was someone pushing a study that contained numbers that he knew to be wrong.

I want to correct Clements on the errors he makes (unknowingly) in this column. I want to see the Wall Street Journal run an article saying what the historical data really says, letting aspiring early retirees know that the conventional methodology studies are "bogus" (raddr's term) or "highly misleading" (Bernstein's phrase) or "analytically invalid" (JWR1945's phrase) or whatever equivalent phrase that the editors of the Journal elect to use to get the essential point across. The essential point is, as Bernstein put it in his e-mail to Ataloss, that, if you are planning a retirement and you have given any thought to using the REHP study or any other conventional methodology for guidance about what is safe, you had best just "Fugettaboutit" given the flaws of the analytical approach employed in those studies.

To take the case to the Wall Street Journal effectively, I could use some help from fellow community members. Are you able to work up some numbers, JWR1945? In telling the Wall Street Journal that the Clements numbers are wrong (it is fair to call them the Clements numbers now that he has put his name to an endorsement of them, even though obviously without knowledge of the errors), I think that a good first step would be to tell them what numbers are right. Can we say what withdrawal rate is safe if you set things up in the way described in this column but use an analytically valid methodology to examine the question?
Mike
*** Veteran
Posts: 278
Joined: Sun Jul 06, 2003 4:00 am

Post by Mike »

I used FireCalc and the REHP calculator, and removing the inflation adjust made little difference in the 100% SWR. I don't know of any quick way to test the points in Jonathan's thesis.

Odds are that the study is accurate for the historical period covered, at the survival rate specified in the study. Adjusting for today's valuations, and bumping the survival rate up to 100% would likely produce a number a couple of percentage points lower. Just my wild guess.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"Adjusting for today's valuations, and bumping the survival rate up to 100% would likely produce a number a couple of percentage points lower. Just my wild guess."

That's my wild guess too.

Do you think that Jonathan Clements would have written that column in the way he wrote it if he knew about SWRs what the people who participate at the various FIRE/Passion Saving/Retire Early boards know about SWRs? I think that the answer is "no" and I think that's important.

If the answer is really "no," then it means that the DCMs' effort to block discussions is ultimately doomed. Even if we don't engage in honest and informed discussions of SWRs at the boards, people like Jonathan Clements and Scott Burns and William Bernstein and Rob Arnott and Andrew Smithers and Peter Bernstein and Robert Shiller are going to talk in honest and informed ways about them in other places. So in time word is going to get out anyway. Sooner or later aspiring early retirees are going to gain access to the information that a good number of them have been trying to gain access to at these boards for 30 months now.

So I just don't see what purpose is served by blocking disucssions at the boards. If this stuff is going to get out in the long run anyway, why not just allow it to get out now? It would do aspiring early retirees a whole bunch more good to be able to lower their stock allocations before prices drop. I honestly cannot figure out the DCMs' long-term strategy. It seems to me that they are reacting emotionally and not thinking this through at all carefully.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

hocus2004, quoting the article, wrote:Clements Quote #2:"Retirees may be able to withdraw as much as 6.2% initially, provided they follow three rules....If your portfolio loses money during the year, you can't give yourself a raise the following year.... No matter how high inflation gets, your maximum annual increase is 6%....You have to avoid selling hard-hit stock funds. Instead, each year, start by lightening up on winning stock funds."
If your portfolio loses money during the year, you can't give yourself a raise the following year.

We have this capability in our latest calculator. In fact, we can identify what percentage of the year-to-year gains that you wish to withdraw.
No matter how high inflation gets, your maximum annual increase is 6%.
We do not have this capability. Gummy has this kind of capability in his Monte Carlo models. He refers to this approach as Sensible Withdrawal Rates. IIRC, however, he sets the ceiling according to real returns (i.e., after adjusting for inflation), not nominal returns.
You have to avoid selling hard-hit stock funds..Instead, each year, start by lightening up on winning stock funds.
We do not have this capability. This is probably based on data sources dated no earlier than 1926. (The Ibbottson (sp?) data. Possibly after 1950, the Standard & Poors' Compustat data base.)

These rules strike me as being much too complex. They are highly likely to be sensitive to small errors in projections. That is, the future's being similar to the past but not identical.

You can see what I have done with removing various percentages of the year-to-year gains in the following posts. In them, I removed a constant percentage of the initial balance (plus inflation) and a varying amount that depended upon gains.

Tapping into Portfolio Gains dated Saturday, Oct 23, 2004.
http://nofeeboards.com/boards/viewtopic.php?t=3021
Tapping into Portfolio Gains: Portfolio Balances dated Sunday, Oct 24, 2004.
http://nofeeboards.com/boards/viewtopic.php?t=3023
Tapping into Portfolio Gains: Amounts Withdrawn dated Saturday, Oct 23, 2004.
http://nofeeboards.com/boards/viewtopic.php?t=3022

I determined maximum survival rates in the first post. I used withdrawal rates of 3.5% and 4.0% for the constant portion in the other two posts.

Removing some of the gains increased risk substantially.

Have fun.

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

Post by JWR1945 »

With our latest calculators, you can make withdrawals based upon any combination of a percentage of the initial balance (with or without adjustments for inflation), a second percentage of the current balance and a third percentage of any year-to-year increases in the balance (but none of the year-to-year decreases in the balance).

Our data reduction software lets us see the four and five year rolling averages of amounts withdrawn. [We have to be careful to include looking at balances. Our software allows us to make withdrawals even when the portfolio's balance is negative.]

If someone is interested in a particular approach, tell us so that we can run the numbers.

Have fun.

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

Post by JWR1945 »

hocus2004 quoting Clements: wrote:Clements Quote #1: "one influential study found that, if retirees want to be confident their savings will last 30 years, they need to limit their initial withdrawal rate to 4.1%, or $4,100 for every $100,000 saved..

Clements Quote #2:"Retirees may be able to withdraw as much as 6.2% initially..
..
Clements Quote #4: "Mr. Guyton's rules make a ton of sense to me -- and I suspect many retirees will find his approach appealing."
This is exceedingly bad advice. It is exactly the wrong thing to do.

The success of a retirement portfolio is critically dependent upon the early years. By the eleventh year, either one's portfolio has grown substantially as a result of holding stocks or the portfolio is already in danger. When you clip off early growth, you are destroying your margin of safety.

One should wait until the portfolio's growth is well established before increasing withdrawals.

Have fun.

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

Post by JWR1945 »

Clements Quote #3: "All this, of course, is based on an analysis of historical returns. If we get wacky markets, miserably low long-run returns or you incur hefty investment costs, you may have to withdraw less than Mr. Guyton suggests."
This is another version of unless the future is worse than the past.

The analysis is already deeply flawed. Today's market is already outside of the historical range. We see this both in P/E10 and dividend yields. To get accurate numbers, there must be adjustments for current valuations.

The study lacks critically important identification of cause and effect relationships.

We already knew that the reported withdrawal rates were in error from this quote:
Clements Quote #1: "one influential study found that, if retirees want to be confident their savings will last 30 years, they need to limit their initial withdrawal rate to 4.1%, or $4,100 for every $100,000 saved.
Any analysis of this nature, if presented without adjustments for valuations, needs to include the qualifier the withdrawal rate of XX is safe only to the extent that the conventionally reported withdrawal rates are safe. That is, they are known to be in error. They require adjustments for today's valuations.

Have fun.

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

Post by JWR1945 »

hocus2004 wrote:To take the case to the Wall Street Journal effectively, I could use some help from fellow community members. Are you able to work up some numbers, JWR1945? In telling the Wall Street Journal that the Clements numbers are wrong (it is fair to call them the Clements numbers now that he has put his name to an endorsement of them, even though obviously without knowledge of the errors), I think that a good first step would be to tell them what numbers are right. Can we say what withdrawal rate is safe if you set things up in the way described in this column but use an analytically valid methodology to examine the question?
We already have the before and after numbers for the conventional methodology: the reported 4.1% is really around 2.5% at today's valuations.

I can take some data by removing some of the year-to-year gains. I would try to line it up with the percentage earnings yield 100E10/P (or 100% / [P/E10]). To do that, however, would require me to collect some number (the value of Y) that is unique for each year (the value of X in the original data set). Most often, the value of Y has been the highest surviving withdrawal rate. Sometimes, it has been a portfolio balance.

An alternative that comes to mind would be to show how much safety has been reduced by removing a fraction of portfolio gains.

I would appreciate suggestions as to what might be the best kind of data to collect.

Have fun.

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

Post by JWR1945 »

I have taken a lot of data and I will need a couple of days to get everything together.

I looked at a portfolio with 50% stocks and 50% TIPS at a 2% interest rate. In one case, I removed 25% of any year-to-year gains (but only when there were gains). I determined the highest historical surviving withdrawal rates and the total amounts withdrawn for each year (from 1921-1980).

I did the same in a second case, but without removing any of the gains. That is, the second case had the conventional withdrawal methodology. Withdrawals were constant after adjusting for inflation.

I also calculated the equivalent withdrawal rate from a 100% TIPS portfolio at a 2% interest rate. It was 4.465% over 30 years.

This is a good baseline. For every $100000 in a 100% TIPS portfolio, you can withdraw $4465 plus inflation every year for 30 years. You have true 100% safety. Your final balance is zero dollars.

These are the results with 25% of gains removed.

An all-TIPS portfolio is superior than the calculated rates when P/E10 is 30 or more. That is, the most likely return when P/E10 is 30 or more is lower than that of an all-TIPS portfolio. An all TIPS portfolio is superior in terms of safety when P/E10 is 25 and higher. It is only slightly inferior when P/E10 is 20.

These are the results in the conventional case (with none of the gains removed).

At all-TIPS portfolio is slightly worse than the calculated rate when P/E10 is 25. An all-TIPS portfolio is better when P/E10 is 30 and higher. An all-TIPS portfolio is only slightly inferior when safety is required and P/E10 is 20. An all-TIPS portfolio is superior when safety is required and P/E10 is 25 and higher. [Of course, TIPS provide a true 100% level of safety. With stocks, these P/E10 levels are relative to the lower confidence limit about the calculated rate at a 90% confidence level (two sided or 95% one sided).]

The 5-year averages of the amounts withdrawn varied close to $800 when 25% of the gains were removed.

Removing 25% of the gains resulted in higher total amounts withdrawn during the first decade, a wash during the second decade and lower total amounts withdrawn during the third decade. The boost in the 5-year averages was around 5% to 10% during the first decade. It was given back during the third decade.

The P/E10 value of today's market is between 25 to 30. [Actually, it "has recently been"..I did not check it out for today.]

Have fun.

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

Post by JWR1945 »

I forgot to include some important numbers.

Based on an initial amount of $100000, 2% TIPS allow you to withdraw $4465 each year for 30 years with true safety.

If you remove 25% of your portfolio's year-to-year gains (but none of the losses), the calculated rates are:
$5177 at 5 years when P/E10 = 20.
$4642 at 5 years when P/E10 = 25.
$4285 at 5 years when P/E10 = 30.
$3839 at 5 years when P/E10 = 40.

At 20 years, the 5-year average withdrawals when 25% of year-to-year portfolio gains are removed are:
$4863 at 20 years when P/E10 = 20.
$4481 at 20 years when P/E10 = 25.
$4226 at 20 years when P/E10 = 30.
$3908 at 20 years when P/E10 = 40.

At 30 years, the 5-year average withdrawals when 25% of year-to-year portfolio gains are removed are:
$4531 at 30 years when P/E10 = 20.
$4137 at 30 years when P/E10 = 25.
$3874 at 30 years when P/E10 = 30.
$3546 at 30 years when P/E10 = 40.

The amounts that can be withdrawn safely are (roughly) $500 less than these numbers.

If you remove none of your portfolio's year-to-year gains (i.e., if you withdrawal a fixed percentage of the initial balance plus inflation), the calculated rates are:
$4599 when P/E10 = 20.
$4197 when P/E10 = 25.
$3931 when P/E10 = 30.
$3599 when P/E10 = 40.

The amounts that can be withdrawn safely are (roughly) $500 less than these numbers.

These numbers use data from 1923-1972. All of the historical sequences were completed. Data from 1923-1980 have partially completed sequences. They suggest that higher withdrawals are OK even though the calculator uses dummy data with heavy losses for the years 2003-2010.

Withdrawing 25% of gains smooths out the variation of the maximum surviving withdrawal amounts versus time. That is, the maximum amounts (averaged over 5 years) that you can withdraw in any given year is close to the amounts that would have been OK in the previous year or two.

Have fun.

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

Post by JWR1945 »

I have found an error in my data reduction software related to converting withdrawal amounts from nominal dollars to real dollars. This does not affect the highest surviving withdrawal rates. But when 25% of the portfolio's year-to-year gains are removed, its contributions to the 5-year rolling averages is reported incorrectly.

That is, only the 5-year rolling averages (the amounts withdrawn) are in error.

I will collect a correct set of withdrawal amounts.

Have fun.

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

Post by JWR1945 »

Going back to an earlier question:
hocus2004 wrote:To take the case to the Wall Street Journal effectively, I could use some help from fellow community members. Are you able to work up some numbers, JWR1945? In telling the Wall Street Journal that the Clements numbers are wrong (it is fair to call them the Clements numbers now that he has put his name to an endorsement of them, even though obviously without knowledge of the errors), I think that a good first step would be to tell them what numbers are right. Can we say what withdrawal rate is safe if you set things up in the way described in this column but use an analytically valid methodology to examine the question?
I have taken data for my new thread about What happens if you withdraw 25% of each year's gains? dated Sat Nov 20, 2004.
http://nofeeboards.com/boards/viewtopic.php?t=3113

My analysis is not yet complete. Here are some initial comments.

It is clear cut and obvious that a new strategy (such as what is mentioned in the first post in this thread) has virtually no effect on Safe Withdrawal Rates.

Whether one uses the conventional constant withdrawal rate (i.e., constant after adjusting for inflation) or an alternative that removes some of the year-to-year gains (but not losses) in one's portfolio, all results are similar. Earlier withdrawals are a little bit higher with the alternative, but later amounts (around 20 years) are lower.

The Historical Surviving Withdrawal Rates in my new thread are higher than the traditionally reported 3.9% with 50% stocks. The reason is that TIPS (at 2% interest) were used instead of commercial paper. Today's Safe Withdrawal Rates may be a little bit higher than previously calculated for a portfolio with 50% stocks and 50% commercial paper. They will still be very close to 2.5%.

Here is what happens when valuations are high.

For the sequence that began in 1965, the traditional HSWR [Historical Surviving Withdrawal Rate, which is not necessarily safe] was 4.2%. Assuming an initial balance of $100000, this amounts to $4200 plus inflation each year. When removing 25% of one's year-to-year gains (but not losses), the maximum amounts (5-year averages) that could have been withdrawn at 5, 10, 20 and 30 years changed to $4474, $4297, $ 4046 and $3901.

For the sequence that began in 1966, the traditional HSWR [Historical Surviving Withdrawal Rate, which is not necessarily safe] was 4.2%. Assuming an initial balance of $100000, this amounts to $4200 plus inflation each year. When removing 25% of one's year-to-year gains (but not losses), the maximum amounts (5-year averages) that could have been withdrawn at 5, 10, 20 and 30 years changed to $4363, $4305, $4016 and $3900.

The withdrawal rates mentioned in the referenced article are dangerously high in light of today's valuations. Even the traditional rates, which are based upon the conventional methodology, do not come anywhere close to being safe. Most likely, but not yet determined, their odds of survival will be around 50-50.

Have fun.

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

Post by Mike »

Earlier withdrawals are a little bit higher with the alternative, but later amounts (around 20 years) are lower.
So you are saying a person can't save too little and depend upon fancy tricks to bail them out.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

I have finished my major write up in my thread about What happens if you withdraw 25% of each year's gains? dated Sat Nov 20, 2004.
http://nofeeboards.com/boards/viewtopic.php?t=3113
http://nofeeboards.com/boards/viewtopic ... 998#p24998

Withdrawing anything above 5% in today's market is reckless. Withdrawing even 4% carries a significant element of risk.

Here is what I wrote in my conclusions. There is much, much more detail in the write up itself and it covers a much wider scope.
Withdrawing 25% of a portfolio's gains is likely to increase amounts that can be withdrawn in the early years of retirement. But this is at the expense of lower withdrawals in the later years.

At today's valuations (with P/E10 = 27 to 28 ) the Safe Withdrawal Rates of a 50% stock portfolio is 2.9% when it includes TIPS at a 2% interest rate (with none of the gains removed). When 25% of the gains are removed (but none of the losses), the Safe Withdrawal Rates (in terms of 5-year averages) vary from 3.5% to 2.7% to 3.1% to 3.1% at years 5, 10, 20 and 30. The Safe Withdrawal Rate for a portfolio of 50% stocks and 50% commercial paper is 3.1%.

[The relative Safe Withdrawal Rates with TIPS (SWR = 2.9%) versus commercial paper (SWR = 3.1%) may be an artifact of how the confidence limits were determined.]

One should use caution when considering withdrawals of 5.0% or more in today's market. Standard calculations indicate that there is a reasonable chance of success at 5.0% even though it is less than 50%. But if estimates using the special confidence limits are correct, a point that is an arguable, the odds of success are very low, less than 5%.
Have fun.

John R.
unclemick
*** Veteran
Posts: 231
Joined: Sat Jun 12, 2004 4:00 am
Location: LA till Katrina, now MO

Post by unclemick »

Excluding my dividend stocks, the SEC yield of my portfolio is running in the 2.7 - 2.9% range.

Handgrenade anyone???

Heh,heh,heh,heh,heh
unclemick
*** Veteran
Posts: 231
Joined: Sat Jun 12, 2004 4:00 am
Location: LA till Katrina, now MO

Post by unclemick »

Toss in dividend stocks and the yield is 3.1% plus. There is a hot rumor, I may be getting a real computer for Christmas (no more webtv). I may even swag whether, I've kept up with inflation the last ten years, although we seem to have survived.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

Here's Scott Burns' take:

http://www.dallasnews.com/sharedcontent ... 9fa68.html

Juicy Burns Quote #1: "Alas, before you start lining up for the spending parade, there is something you need to know. Other research, published in the September/October issue of the Financial Analysts Journal, will rain on your parade. Starting with the idea that future portfolio returns are influenced by valuation levels when you start, researcher Robert Arnott examines returns of typical portfolios and finds that real returns have averaged about 4.2 percent in the postwar period. A portion of that real return, however, came from rising valuation levels. When valuation level changes are stripped out, the average real return on a 60/40 stock/bond portfolio has been only 3.4 percent since World War II. "

Juicy Burns Quote #2: "It's very likely Mr. Guyton's rules give a dangerous sense of safety. "

Juicy Burns Quote #3: "Bottom line: Withdraw more than 5 percent a year at your peril."
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

Here's a link to a PDF copy of the September/October 2004 article by Arnott cited by Scott Burns in the column linked to above:

http://www.aimrpubs.org/faj/home.html

(The article cited by Burns is the one titled "Sustainable Spending in a Lower Return World.")

Juicy Arnott Quote #1: â"Relative to current yields on TIPS, the risk premium on stocks is dismayingly small (probably less than 1 percent at this writing)."Â￾

Juicy Arnott Quote #2: "The rolling 10-year returns range from -1.3 percent (for the decade ending September 1974) to +15 percent (for the decade through July 1992). When the portion of the return associated with changing valuation levels is removed, the range is much narrower--averging 5.7 percent, with a standard deviation of just 1.3 percent....Current pension return assumptions and public fund discount rates of 8-9 percent have never appeared in the entire history of the sustainable return data for a 60/40 passive portfolio."Â￾

Juicy Arnott Quote #3: "Mean reversion is widely evident in Figure 1. When we compare a 10-year rolling return for the 60/40 passive mix with the nonoverlapping prior 10-year return, we find an 85 percent negative correlation between one decade's return and the subsequent decade's return."

Juicy Arnott Quote #4: "Now, consider the taxable investor. Historical real returns since WWII for a 60/40 balanced portfolio have averaged 3.3 percent. When we strip out the effects of falling yields and rising valuation levels, that return falls to 1.9 percent. Given today's low yields, we cannot reasonably expect more. Because we are taxed on both our real return and the inflation component of our return, a reasonable expectation for the after-tax real return on a 60/40 portfolio is fairly close to zero."Â￾

Juicy Arnott Quote #5: "The arithmetic for taxable investors is remarkably simple if the after-tax return on a portfolio is roughly zero. If a retiree wants to maintain a lifestyle costing $40,000 a year, adjusted for inflation, for a life expectancy of 25 years, he or she will need $1 million."Â￾

Juicy Arnott Quote #5: "We have not done a good job of conveying this message, OR ANYTHING NEAR IT [Arnott had this phrase in italics], to our taxable clients."
Post Reply